The SAGE Handbook of Sports Economics [1 ed.] 1473979765, 9781473979765

Table of contents :
The SAGE Handbook of Sports Economics
Contents
List of Figures
List of Tables
Notes on the Editors and Contributors
PART I: The Nature and Value of the Sports System and Economy
1: Introduction
2: Origins and Developments of Sports Systems
3: The Economic Value of Sport
PART II: Amateur Sports Participation, Supply and Impact
4: Sports Participation
5: Sports Participation and Health
6: Sport and Social Capital Formation
7: Recent Evidence on the Effects of Physical Activity on Human Capital and Employment
8: Private Household Consumption in Sport
9: Sports Clubs in Europe: Organization
10: Volunteering in Sports Clubs and its Impacts
11: The Role of Money and Time Donations in the Supply of Amateur Sport
12: The Economics of the National Collegiate Athletic Association
PART III: Professional Team Sports
13: Economic Objective Functions in Team Sports: A Retrospective
14: European Sports Leagues: Origins and Features
15: Competition Policy in Sports Markets
16: Competitive Balance: Measurement and Relevance
17: Economics of Attendance
18: Exposure and Television Audience Demand: The Case of English Premier League Football
19: Ticket Pricing
20: Secondary Ticket Markets for Sport Events
21: The Economics of the Transfer Market
22: Team Production and Efficiency in Sports
23: Officials and Home Advantage
24: Franchise Relocation and Stadium Subsidies
PART IV: Professional Sports Leagues
25: The Economics of Professional Soccer
26: The Economics of Cricket
27: Rugby Union’s Late Conversion to Professionalism: An Economic Perspective
28: ‘The Answer’ and the Economics of Basketball: Perceptions vs Production
29: Economics and the National Football League
30: ‘The Baseball Players’ Labor Market’: An Update
31: Economic Issues of the National Hockey League: A Survey of the Literature
32: The Economics of Australian Rules Football
33: The Economics of Major League Soccer from the NASL to MLS: A Brief History of North American Professional Soccer
PART V: Sports Events
34: The Economic Impact Measurement of the Olympic Games
35: Major Events: Economic Impact
36: Olympic Games: Public Referenda, Public Opinion and Willingness to Pay
37: Olympic Performance
38: The Economics of Mega-Events: The Impact, Costs, and Benefits of the Olympic Games and the World Cup
39: Economic Impact of Minor Sporting Events and Minor League Teams
40: Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans'?
41: Willingness to Pay in Sports
42: Positive and Negative Externalities of Sport Events: From Well-Being, Pride, and Social Capital to Traffic and Crime
PART VI: Individual Sports
43: The Economics of Running
44: Hitting the Ball Forward: The Economics of Racquet Sports
45: The Economics of Road Cycling
46: The Economics of Golf
47: IRON(O)MICS: The Market for Long-Distance Triathlon
48: NASCAR Economics
PART VII: Future Research
49: Behavioral Economics and Sport
50: Is There a Gender Difference in the Response to Competitive Settings?
51: Dynamic Pricing in Sports
52: Sports Betting
53: The Economics of Doping in Sports: A Special Case of Corruption
54: Performance Analytics
Index

Citation preview

The SAGE Handbook of

Sports Economics

SAGE was founded in 1965 by Sara Miller McCune to support the dissemination of usable knowledge by publishing innovative and high-quality research and teaching content. Today, we publish over 900 journals, including those of more than 400 learned societies, more than 800 new books per year, and a growing range of library products including archives, data, case studies, reports, and video. SAGE remains majority-owned by our founder, and after Sara’s lifetime will become owned by a charitable trust that secures our continued independence. Los Angeles | London | New Delhi | Singapore | Washington DC | Melbourne

The SAGE Handbook of

Sports Economics

Edited by

Paul Downward, Bernd Frick, Brad R. Humphreys, Tim Pawlowski, Jane E. Ruseski and Brian P. Soebbing

SAGE Publications Ltd 1 Oliver’s Yard 55 City Road London EC1Y 1SP SAGE Publications Inc. 2455 Teller Road Thousand Oaks, California 91320 SAGE Publications India Pvt Ltd B 1/I 1 Mohan Cooperative Industrial Area Mathura Road New Delhi 110 044 SAGE Publications Asia-Pacific Pte Ltd 3 Church Street #10-04 Samsung Hub Singapore 049483

Editor: Matthew Waters Editorial Assistant: Umeeka Raichura Production Editor: Manmeet Kaur Tura Copyeditor: Sarah Bury Proofreader: Derek Markham Indexer: Cathryn Pritchard Marketing Manager: Lucia Sweet Cover Design: Naomi Robinson Typeset by: Cenveo Publisher Services Printed in the UK

Introduction & editorial arrangement © Paul Downward, Bernd Frick, Brad R. Humphreys, Tim Pawlowski, Jane E. Ruseski and Brian P. Soebbing, 2019 Chapter 1 © Paul Downward, Bernd Frick, Brad R. Humphreys, Tim Pawlowski, Jane E. Ruseski and Brian P. Soebbing, 2019 Chapter 2 © Wladimir Andreff, 2019 Chapter 3 © Themis Kokolakakis, Chris Gratton and Günther Grohall, 2019 Chapter 4 © Paul Downward and Cristina Muñiz, 2019 Chapter 5 © Jane E. Ruseski, 2019 Chapter 6 © Tim Pawlowski and Ute Schüttoff, 2019 Chapter 7 © Carina Steckenleiter and Michael Lechner, 2019 Chapter 8 © Fernando Lera-López, 2019 Chapter 9 © Christoph Breuer, Philipp Swierzy and Svenja Feiler, 2019 Chapter 10 © Pamela Wicker, 2019 Chapter 11 © Kirstin Hallmann and Lea Rossi, 2019 Chapter 12 © Allen R. Sanderson and John J. Siegfried, 2019 Chapter 13 © Rodney Fort, 2019 Chapter 14 © Nicolas Scelles and Jean-François Brocard, 2019 Chapter 15 © Oliver Budzinski, 2019 Chapter 16 © Tim Pawlowski and Georgios Nalbantis, 2019 Chapter 17 © Placido Rodriguez, 2019 Chapter 18 © Babatunde Buraimo, 2019 Chapter 19 © Brian P. Soebbing, 2019 Chapter 20 © Pascal Courty, 2019 Chapter 21 © Stefan Késenne, 2019 Chapter 22 © Mikael Jamil, 2019 Chapter 23 © J. James Reade, 2019 Chapter 24 © Dennis Coates, 2019 Chapter 25 © Daniel Weimar, 2019 Chapter 26 © Ian Gregory-Smith, David Paton and Abhinav Sacheti, 2019

Chapter 27 © Patrick Massey, 2019 Chapter 28 © David Berri, 2019 Chapter 29 © Benjamin Blemmings, 2019 Chapter 30 © Anthony C. Krautmann, 2019 Chapter 31 © Duane W. Rockerbie and Stephen T. Easton, 2019 Chapter 32 © Ross Booth and Robert Brooks, 2019 Chapter 33 © Nicholas Watanabe, 2019 Chapter 34 © Holger Preuss, 2019 Chapter 35 © Wolfgang Maennig, 2019 Chapter 36 © Wolfgang Maennig, 2019 Chapter 37 © Eva Marikova Leeds, 2019 Chapter 38 © Candon Johnson, 2019 Chapter 39 © Nola Agha and Marijke Taks, 2019 Chapter 40 © Peter Dawson, 2019 Chapter 41 © Johannes Orlowski and Pamela Wicker, 2019 Chapter 42 © Pamela Wicker and Paul Downward, 2019 Chapter 43 © Bernd Frick, Katharina Moser and Katrin Scharfenkamp, 2019 Chapter 44 © Julio del Corral and Carlos Gomez-Gonzalez, 2019 Chapter 45 © Daam Van Reeth, 2019 Chapter 46 © Stephen Shmanske, 2019 Chapter 47 © Joachim Prinz, 2019 Chapter 48 © Peter von Allmen, 2019 Chapter 49 © Yulia Chikish and Brad R. Humphreys, 2019 Chapter 50 © Michael A. Leeds, 2019 Chapter 51 © Rodney J. Paul, 2019 Chapter 52 © David Forrest, 2019 Chapter 53 © Eugen Dimant and Christian Deutscher, 2019 Chapter 54 © Bill Gerrard, 2019

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted in any form, or by any means, only with the prior permission in writing of the publishers, or in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be sent to the publishers. At SAGE we take sustainability seriously. Most of our products are printed in the UK using responsibly sourced papers and boards. When we print overseas we ensure sustainable papers are used as measured by the PREPS grading system. We undertake an annual audit to monitor our sustainability.

Library of Congress Control Number: 2019932126 British Library Cataloguing in Publication data A catalogue record for this book is available from the British Library ISBN 978-1-4739-7976-5

Contents List of Figures ix List of Tablesxi Notes on the Editors and Contributorsxiii PART I THE NATURE AND VALUE OF THE SPORTS SYSTEM AND ECONOMY

1

1

Introduction Paul Downward, Bernd Frick, Brad R. Humphreys, Tim Pawlowski, Jane E. Ruseski and Brian P. Soebbing

3

2

Origins and Developments of Sports Systems Wladimir Andreff

8

3

The Economic Value of Sport Themis Kokolakakis, Chris Gratton and Günther Grohall

18

PART II   AMATEUR SPORTS PARTICIPATION, SUPPLY AND IMPACT

31

4

Sports Participation Paul Downward and Cristina Muñiz

33

5

Sports Participation and Health Jane E. Ruseski

45

6

Sport and Social Capital Formation Tim Pawlowski and Ute Schüttoff

54

7

Recent Evidence on the Effects of Physical Activity on Human Capital and Employment Carina Steckenleiter and Michael Lechner

8

Private Household Consumption in Sport Fernando Lera-López

72

9

Sports Clubs in Europe: Organization Christoph Breuer, Philipp Swierzy and Svenja Feiler

82

10

Volunteering in Sports Clubs and its Impacts Pamela Wicker

92

11

The Role of Money and Time Donations in the Supply of Amateur Sport Kirstin Hallmann and Lea Rossi

102

12

The Economics of the National Collegiate Athletic Association Allen R. Sanderson and John J. Siegfried

112

64

vi

THE SAGE HANDBOOK OF SPORTS ECONOMICS

PART III  PROFESSIONAL TEAM SPORTS

123

13

Economic Objective Functions in Team Sports: A Retrospective Rodney Fort

125

14

European Sports Leagues: Origins and Features Nicolas Scelles and Jean-François Brocard

135

15

Competition Policy in Sports Markets Oliver Budzinski

144

16

Competitive Balance: Measurement and Relevance Tim Pawlowski and Georgios Nalbantis

154

17

Economics of Attendance Placido Rodriguez

163

18

Exposure and Television Audience Demand: The Case of English Premier League Football Babatunde Buraimo

19

Ticket Pricing Brian P. Soebbing

181

20

Secondary Ticket Markets for Sport Events Pascal Courty

190

21

The Economics of the Transfer Market Stefan Késenne

203

22

Team Production and Efficiency in Sports Mikael Jamil

210

23

Officials and Home Advantage J. James Reade

219

24

Franchise Relocation and Stadium Subsidies Dennis Coates

231

171

PART IV  PROFESSIONAL SPORTS LEAGUES

241

25

The Economics of Professional Soccer Daniel Weimar

243

26

The Economics of Cricket Ian Gregory-Smith, David Paton and Abhinav Sacheti

256

27

Rugby Union’s Late Conversion to Professionalism: An Economic Perspective Patrick Massey

268

28

‘The Answer’ and the Economics of Basketball: Perceptions vs Production David Berri

279

Contents

vii

29

Economics and the National Football League Benjamin Blemmings

289

30

‘The Baseball Players’ Labor Market’: An Update Anthony C. Krautmann

298

31

Economic Issues of the National Hockey League: A Survey of the Literature Duane W. Rockerbie and Stephen T. Easton

308

32

The Economics of Australian Rules Football Ross Booth and Robert Brooks

322

33

The Economics of Major League Soccer from the NASL to MLS: A Brief History of North American Professional Soccer Nicholas Watanabe

331

PART V  SPORTS EVENTS

341

34

The Economic Impact Measurement of the Olympic Games Holger Preuss

343

35

Major Events: Economic Impact Wolfgang Maennig

356

36

Olympic Games: Public Referenda, Public Opinion and Willingness to Pay Wolfgang Maennig

367

37

Olympic Performance Eva Marikova Leeds

377

38

The Economics of Mega-Events: The Impact, Costs, and Benefits of the Olympic Games and the World Cup Candon Johnson

39

Economic Impact of Minor Sporting Events and Minor League Teams Nola Agha and Marijke Taks

395

40

Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans’? Peter Dawson

405

41

Willingness to Pay in Sports Johannes Orlowski and Pamela Wicker

415

42

Positive and Negative Externalities of Sport Events: From Well-Being, Pride, and Social Capital to Traffic and Crime Pamela Wicker and Paul Downward

385

428

PART VI  INDIVIDUAL SPORTS

439

43

441

The Economics of Running Bernd Frick, Katharina Moser and Katrin Scharfenkamp

viii

THE SAGE HANDBOOK OF SPORTS ECONOMICS

44

Hitting the Ball Forward: The Economics of Racquet Sports Julio del Corral and Carlos Gomez-Gonzalez

452

45

The Economics of Road Cycling Daam Van Reeth

462

46

The Economics of Golf Stephen Shmanske

472

47

IRON(O)MICS: The Market for Long-Distance Triathlon Joachim Prinz

482

48

NASCAR Economics Peter von Allmen

492

PART VII  FUTURE RESEARCH

503

49

Behavioral Economics and Sport Yulia Chikish and Brad R. Humphreys

505

50

Is There a Gender Difference in the Response to Competitive Settings? Michael A. Leeds

516

51

Dynamic Pricing in Sports Rodney J. Paul

525

52

Sports Betting David Forrest

534

53

The Economics of Doping in Sports: A Special Case of Corruption Eugen Dimant and Christian Deutscher

544

54

Performance Analytics Bill Gerrard

553

Index563

List of Figures 3.1 3.2 3.3 4.1 11.1 20.1 21.1 21.2 21.3 21.4 23.1 23.2 23.3 23.4 25.1 25.2 25.3 25.4 25.5 25.6 27.1 27.2 27.3 27.4 27.5 30.1 30.2 30.3 30.4 30.5 31.1 31.2 31.3 31.4 34.1 34.2 34.3 34.4 34.5 34.6 34.7 34.8 34.9 36.1

The circular flow of income Three layers in the Vilnius definition of sport Sport employment in the EU (top five scores) (thousands, 2016) Sport participation demand Multi-level framework on the different levels of time and money donations Fraction of teams with a sponsored secondary ticket marketplace Transfer system and competitive balance in a profit-maximization league Transfer system and competitive balance in a win-maximization league Transfer market and profit maximization Transfer market and win maximization Stylised examples of home advantage expressed in terms of relative team strengths and recorded (average) outcomes Relative proportion of matches ending in wins for the favourite team (according to Elo ranking) Cricket match outcomes by relative quality Cricket match outcomes by Test number Number of publications per main stream of empirical soccer economic research Empirical soccer economic studies by year of publication (2017 only considered until June) Time trends within main streams (2017 only considered until June) Journals with three or more publications on empirical soccer economic studies Authors with 10 or more empirical soccer economic publications Leagues and nations considered in empirical soccer economic research Top14 payroll as a percentage of revenue (2004/05–2016/17) Top14 Clubs’ aggregate revenue and expenditure (2004/05-2016/17 € million) Composition of Top14 team revenues in 2016/17 Average attendances in the main European rugby leagues (2001/02–2017/18) Average Pro14 attendances by country (2003/04–2017/18) Contract extensions in Major League Baseball (2001–2014) Free Agents’ WARP (2012–2014) Contract extension players’ WARP (2012–2014) Journeymen’s WARP (2012–2014) Apprentices’ WARP (2012–2014) NHL salary distribution, 2003–04 season (2003 = 100) NHL salary distribution, 2013–14 season (2003 = 100) Turnover rate versus winning percentage, 1967–68 to 2016–17 NHL fights per game, 1960–2015 Olympic Keynesian approach Basic monetary streams in impact studies Matrix of relevant monetary streams in impact studies Three-scenario ex-ante study on the Olympic Games Paris 2024 impact for Ile-de-France region Factors influencing the net impact of the Olympic Games Time frame to measure the impact of the Olympic Games Three-scenario forecast of three periods for the Olympic Games in Paris 2024 Importance of the size of the region when measuring an impact Games versus non-Games effect by changes in import rates and different multipliers Results of the referendum on the bid for the Olympic Games in Hamburg 2024 (29 November 2015). Share of yes/no votes in the different districts

20 22 26 34 105 195 205 205 206 207 221 223 225 227 248 249 249 250 251 252 270 273 273 275 276 299 300 302 304 305 310 310 313 316 345 345 345 346 347 348 349 350 352 371

x

39.1 39.2 39.3 39.4 40.1 40.2 42.1 43.1 43.2 54.1

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Potential to find impact of events/teams using the ex post regression method 396 Event Resource Demand (ERD) continuum 397 Economic impact drivers 399 Optimum economic impact where Event Resource Demand (ERD) equals City Resource Supply (CRS) 399 UK performance funding and (Summer) Olympic medals won 2000–2016 406 Stages of change within the transtheoretical model (TTM) 408 The total economic value of sport events 430 Percentage of Africans and Europeans in the world’s marathon elite, 1973–2015446 Age and marathon performance of recreational runners by gender 448 The five stages of analytical competition555

List of Tables 2.1 Stylized organizational features of closed and open sports systems 3.1 The economic value of sport in the EU, 2012 and 2005 3.2 Main sport-related indicators, direct effects 3.3 Employment contributions of the 10 top sport-related sectors 3.4 UK indirect output multipliers of industries with ties to sport, 2014 3.5 EU-wide output multipliers of sport-related goods and services 4.1  Leisure demand in the ‘income–leisure’ model and Becker’s time allocation model 4.2 Econometric modelling approaches 6.1 Empirical studies on sport and social capital formation 8.1 Economic importance of consumption in sport made by households 10.1  Volunteering in sports clubs from a sports economics perspective: overview of topics studied 16.1  Studies analyzing the impact of market regulations and/or competition design elements on CB 18.1 Domestic rights fees for English Premier League, 1983–2019 18.2 Distribution of televised games in the 2000–01, 2007–08 and 2016–17 seasons 18.3 Summary statistics of dependent and independent variables 18.4 Model for television audience demand. Dependent variable is ln (television audience ratings) 20.1 Changes in ticketing practices 23.1 Proportion of cricket Test matches won by each team 25.1 Number of soccer economic publications per substream 25.2 Authors with three or more publications per substream (in descending order according to the number of publications) 26.1 Summary of LBW appeals in DRS matches 26.2 Summary of LBW wickets in DRS and non-DRS matches 27.1 ERC/ERCC performances by league and country (1995/96–2017/18) 28.1 Impact of various player and team factors on wins in National Collegiate Athletic Association Women’s and Men’s Basketball, the Women’s National Basketball Association, and National Basketball Association, and the American Basketball Association 28.2 Modeling NBA free agent salaries, 2007–2017 28.3 Economic significance of box score statistics, 2007–2017 28.4 Explaining team field goal attempts in the NBA, 1987–88 to 2016–17 30.1 Summary statistics: Free Agents (2012–2014) 30.2 Huber-White estimates of Free Agent salaries (2012–2014) – Dependent variable: ln (REALSal) 30.3 Means of contract extension players (2012–2014) 30.4 Means of Journeymen (2012–2014) 30.5 Means of Apprentices (2012–2014) 31.1 Games lost due to work stoppages 31.2 Coaching turnover and winning percentage, 1967–68 to 2016–17 34.1 Problems of data collection 34.2 Share of visitors at events based on economic importance 36.1 Positive referenda 36.2 Negative referenda 39.1 Results from the Standard Economic Impact Analysis of the 2005 Pan American Junior Athletic Championships: Economic Impact Summary – Combined Total (Visitor – Operational –Stadium) for the City of Windsor in $ CDN

10 23 24 25 26 26 35 38 57 73 93 157 173 175 176 177 191 226 247 251 262 263 272

282 283 284 284 300 301 302 304 306 309 312 347 351 368 369 401

xii

THE SAGE HANDBOOK OF SPORTS ECONOMICS

39.2 Results from the cost–benefit analysis (in $ CDN) of the 2005 Pan American Junior Athletic Championships 401 41.1 Overview of CVM studies estimating WTP (in chronological, then alphabetical order) 418 43.1 Prize purses, 2017, in USD 442 43.2 Time bonuses, men 2017, in USD 443 43.3 Time bonuses, women 2017, in USD 443 43.4 Prize money (in USD) in middle- and long-distance running events, 2017 444 45.1 The competition structure of professional road cycling in 2019 466 45.2 Some key financial data for professional road cycling (1990–2018, in nominal euro) 467 47.1 Top three annual prize money (US$) makers in selected sports, 2015 483 47.2 Worldwide IRONMAN race characteristics (2011–2015, daily averages) 485 47.3 Triathlon-portfolio 2015 (professional athletes) 486 47.4 Development of the IRONMAN market 2008–2016 486 47.5 The demand of IRONMAN races 490 Appendix 20.1 Sponsorship adoption timelines201

Notes on the Editors and Contributors

THE EDITORS Paul Downward is Professor of Economics in the School of Sport, Exercise and Health Sciences at Loughborough University in the UK. He is the current editor of European Sport Management Quarterly and President of the European Sport Economics Association, serving also on the editorial boards of the Journal of Sports Economics, Sport Management Review and the Journal of Sport and Tourism. Paul is the author of The Economics of Professional Team Sports and Sports Economics: theory, evidence and policy and co-editor of the handbook of Sport Management. Paul has published widely on all aspects of sports economics, focusing more recently on the determinants of sport and physical activity participation as well as its health, well-being and labour market outcomes. He has received funding for his research from a variety of organizations such as the DCMS, UK Sport, Sport England, The Health Foundation, The British Academy, Streetgames, Sustrans and the IOC. Bernd Frick is Professor of Organizational Economics in the Management Department at Paderborn University, Germany and Professor of Sports Economics at Seeburg Castle University in Seekirchen, Austria. His research interests are in organizational and personnel economics as well as in sports and cultural economics. He has published more than eighty refereed papers in e.g. Industrial and Labor Relations Review, Labour Economics, Scandinavian Journal of Economics, Education Economics, Applied Economics, European Journal of Operational Research, Industrial Relations, Labour, British Journal of Industrial Relations, Journal of Sports Economics and the International Journal of Sport Finance. When younger, he was a mediocre football player and a competitive long-distance runner. Although somewhat slower today, he continues to be a dedicated athlete. Brad R. Humphreys is a professor in the Department of Economics at West Virginia University. He holds a PhD in economics from the Johns Hopkins University. His research on the economics of gambling and the economics and financing of professional sports has been published in academic journals in economics and policy analysis, including the Journal of Urban Economics, the Journal of Policy Analysis and Management, the Journal of Regional Science, the Journal of Economic Behavior and Organization, Health Economics, Public Finance Review, and Regional Science and Urban Economics. He twice testified before the United States Congress on the economic impact of professional sports teams and facilities. Tim Pawlowski is Professor of Sport Economics and Head of the Research Group on Sport Economics, Sport Management and Media Research at the University of Tübingen (Germany). Moreover, he is Associate Member at the LEAD Graduate School and Research Network and founding Board Member of the European Sport Economics Association (ESEA). Tim’s empirical work follows three broader lines, i.e. ‘leagues and competitions’, ‘society and public policy’ and ‘media and management’, and was supported with research grants from the German Research Foundation (DFG), UEFA and FIFA. He was principal investigator in several projects – amongst others for the Federal Ministry of Finance in Germany and Major League Soccer in North America – as well as guest speaker invited by the United Nations Children’s Fund (UNICEF), the Council of the European Union and the Sports Committee of the German National Parliament.

xiv

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Jane E. Ruseski is an applied microeconomist with interests in health economics, health financing and policy, and sports economics. Much of her current research studies the socioeconomic determinants of health and (un)healthy behaviors; the effect of health behaviors on outcomes, including chronic health conditions, obesity, and health disparities; the mechanisms underlying health behaviors; and the effect of public policy on health. The overarching goal of this research is to inform the policy environment in an effort to implement policies and interventions that will promote health behaviors and reduce the burden of disease. Dr Ruseski has published in academic journals, including the Journal of Regional Science, Contemporary Economic Policy, Health Economics, the Journal of Behavioral and Experimental Economics, the BE Journal of Economic Analysis & Policy, Public Finance Review and the Journal of Sports Economics. She is a co-editor of Contemporary Economic Policy and an associate editor of the International Journal of Sport Finance. Brian P. Soebbing is an Associate Professor in Sport and Recreation Management in the Faculty of Kinesiology, Sport, and Recreation at the University of Alberta. Dr Soebbing’s main research interest focuses on the strategic behavior of sport and recreation organizations and their constituents. Within this research area, he examines issues related to sport facilities and events, decisions made by organizational leaders and its members, policies adopted/modified by sports and recreation organizations, and behaviors by consumers of sport and recreation. Currently, Dr Soebbing is a member of eight editorial boards and is the Associate Editor for both Sport & Entertainment Review and the International Journal of Sport Finance.

THE CONTRIBUTORS Nola Agha is an Associate Professor in the Sport Management Program at the University of San Francisco. Professor Agha’s research interests lie at the intersection of finance, economics, and strategic management. She studies public goods and externalities, the economic impacts of teams and stadiums, the efficiency and equity outcomes of stadium subsidies, and a variety of issues related to minor league baseball. She worked in international business operations for several years and has also consulted to the sport and fitness industry by conducting economic impact studies, competitive analysis, and feasibility studies for clients in MLB, NBA, minor league hockey, local organizing committees, and fitness organizations. Peter von Allmen is the David H. Porter chair and professor of economics at Skidmore College. He has previously served as vice-president and president of the North American Association of Sports Economists (NAASE) and as a Fellow of the American Council on Education. His research focuses on contracts, the role of agents in the bargaining process, and incentives in professional sports. He is the co-author of two textbooks: The Economics of Sports, now in its sixth edition, and Economics. He is also a member of the editorial board of the Eastern Economic Journal. Wladimir Andreff is currently Emeritus Professor at the University Paris 1 Panthéon Sorbonne; President of the Scientific Council at the Observatory of the Sports Economy, French Ministry for Sports; Honorary President of the International Association of Sport Economists and of the European Sports Economics Association; Honorary Member and former President of the European Association for Comparative Economic Studies, and former President of the French Economic Association. His research areas are international economics, transition economies, and sports economics. He authored 13 books (six in sports economics), 430 articles (74 peer-reviewed), of which 155 are in sports economics, and edited 17 books (five in sports economics), and has published in 18 languages. His most recent books in sports economics are Mondialisation économique du sport: Manuel de référence en Economie du sport (De Boeck, Bruxelles, 2012; Globalization of the sports economy: Reference textbook in sports economics) and, as an editor, Disequilibrium Sports Economics: Competitive Imbalance and Budget Constraints (Edward Elgar, Cheltenham, 2015). David Berri is a Professor of Economics at Southern Utah University. Dr Berri has spent the last two decades researching sports and economics, and has published works on a variety of topics, including the evaluation of players and coaches, competitive balance, the drafting of players, labor disputes, the NCAA, and

Notes on the Editors and Contributors

xv

gender issues in sports. He was author of two books, The Wages of Wins and Stumbling on Wins, along with the recent textbook Sports Economics. In the past, he has written on the subject of sports economics for a number of popular media outlets, including the New York Times, the Atlantic.com, Time.com, Vice Sports, and Forbes.com. Benjamin Blemmings is a PhD student in Economics at West Virginia University (WVU). He received a BS in Quantitative Economics from Miami (OH) University, where he graduated with departmental honours in 2015. His current work focuses on economic policies relating to higher education institutions, such as alcohol sales at college football games and merit-based scholarship aid. His primary fields are health and public economics, and he also has broad interests within empirical microeconomics. Ross Booth is a Senior Lecturer in the Department of Economics at Monash University, Melbourne, Australia. His main research interest is the economics of professional team sports leagues, especially the Australian Football League. His research has been published in Australian Economic Review, Economic and Labour Relations Review, Economic Papers, Economic Record, International Journal of Sport Finance, Journal of Sports Economics, Labour and Industry and Sport Management Review, as well as several international books. He has been a Vice President of the International Association of Sports Economists (IASE) since December 2014 and was a Vice-President of the North American Association of Sports Economists (NAASE) from July 2011 to July 2013. He is a member of the editorial boards of the International Journal of Sport Finance, Journal of Global Sport Management and Sporting Traditions. Christoph Breuer is Full Professor of Sport Management at German Sport University Cologne since 2004 (at W3 level since 2010) and Director of the Institute of Sport Economics and Sport Management. Moreover, he is the Vice President for resources, planning and quality management at German Sport University Cologne. From 2006 to 2011 he was simultaneously research professor at German Institute for Economic Research (DIW Berlin). His main research areas are economics, sociology and management of elite and grassroots sport, economics of match-fixing and return on marketing investments/economics of sponsorship. Jean-François Brocard is an Associate Professor of Economics at the University of Limoges, where he’s a member of the Centre de Droit et d’Economie du Sport (CDES). His research focuses on the regulation of professional sports. His recent work includes research on the transfer market of professional football players and its excesses and in particular on the labor market of professional athletes. He is the Secretary General of both the International Association of Sports Economists (IASE) and the French Seminar of “Dynamique Economique du Sport” (DESport). Besides, he’s a member of the Board of the French regulatory authority for online games. Robert Brooks is Deputy Dean (Education) in the Monash Business School, Faculty of Business and Economics and Professor in the Department of Econometrics and Business Statistics, Monash University, Melbourne, Australia. His main research interest is in applied econometric modelling of a broad range of financial and economic data, including data of relevance to sports economics. Oliver Budzinski is Professor of Economic Theory and Director of the Institute of Economics at Ilmenau University of Technology in Germany. Former research and teaching positions include the University of Southern Denmark, Campus Esbjerg, the New York University, US, and Philipps-University of Marburg, Germany. Oliver Budzinski received his PhD from the Leibniz-University of Hanover, Germany. His research interests cover competition policy and industrial economics, media economics, as well as sports economics. To date, he has published more than 40 articles in refereed journals, three books and more than 40 chapters in edited volumes. Babatunde Buraimo is a Senior Lecturer in sports economics and sports management at the Centre for Sports Business at the University of Liverpool. He has published widely on a range of issues including sports broadcasting, competitive balance and uncertainty of outcome, demand for sport, sports corruption, and more recently, suspense, surprise and shock in professional football. His papers have featured a range of journals including Economic Inquiry, European Sports Management Quarterly, British Journal of Management and Oxford Bulletin of Economics and Statistics. In addition to his academic work, Babatunde has also worked on a number of projects for government departments at national and European levels.

xvi

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Yulia Chikish is a PhD candidate in West Virginia University. She graduated from Far Eastern National University in Vladivostok, Russia, majoring in Mathematics and Economics. She received her Master degree from Central European University in Budapest, Hungary. Her research interests include sports economics, behavioural economics and urban economics. Dennis Coates is a Professor of Economics at the University of Maryland, Baltimore County, USA and Leading Researcher in the International Laboratory of Intangible-Driven Economy at the National Research University Higher School of Economics, Russia. His research interests are in the areas of sports economics, public choice, and public economics. Julio del Corral is an Associate Professor (with habilitation to Full Professor) in Economics in the Faculty of Law and Social Sciences (Ciudad Real) at the University of Castilla-La Mancha. He holds a PhD in economics from the University of Oviedo. His current research focuses on the economics of sports, specifically the topics of demand, competitive balance, productivity, discrimination, and betting. He also has broad interests within empirical microeconomics specifically in behavioural economics and measurement of efficiency and productivity. He could have become a table tennis star but he preferred to focus on training in economics rather than training in table tennis. Pascal Courty is a Professor of Economics at the University of Victoria who has received his PhD in Economics from the University of Chicago and works in the field of industrial organization. One of his ongoing interests over the past 20 years is ticket markets. He has studied how tickets for popular concerts are priced in the primary market, estimated the return from using multiple seating categories, and shown that popular bands do not always exploit market power. He has also argued that resale in secondary markets can benefit both consumers and artists. Recently his research has looked at changes in ticket pricing with the advent of the Internet. His work has been published in leading academic journals such as the American Economic Review, Review of Economics Studies, Journal of Economics and Statistics, Journal of Law and Economics, and Journal of Economic Perspective. Pascal Courty is Research Fellow of the Centre for Economic Policy Research. Peter Dawson is a Reader (Associate Professor) in Economics in the School of Economics at the University of East Anglia (UEA). He has also worked at the University of Bath and held external examining appointments at Imperial College London, University of Liverpool and Loughborough University amongst others. He has worked in the area of sports economics for nearly 25 years, having first been inspired whilst an undergraduate student. His main focus has been on the behaviour of agents (fans, referees, managers and players) in sporting contests but more recently his attention has turned to the impact of major sporting events, which has included the impact of the London 2012 Olympics on health, happiness and participation in sport. His research has been published in Journal of Economic Psychology, Journal of the Royal Statistical Society: Series A (Statistics in Society), Scottish Journal of Political Economy, Economic Modelling, Journal of the Operational Research Society, Journal of Applied Statistics, Journal of Sports Economics and International Journal of Sport Finance. Christian Deutscher is a Professor of Sport Economics at Bielefeld University, Germany. His research focuses on incentives, effort and sabotage in contests. Recent studies cover betting markets to determine demand for sports betting, efficiency of betting markets and match fixing. Eugen Dimant is a Senior Research Fellow in the Identity and Conflict Lab at the University of Pennsylvania. Dr Dimant’s research interests center on experimental behavioral economics with a particular focus on behavioral ethics, crime, corruption, migration, social norms, and terrorism. Stephen T. Easton is Professor of Economics at Simon Fraser University. He has published in fields as diverse as Economic History, International Trade, International Finance, Economics of Education, Law and Economics, the Economics of Crime as well as in Sports Economics. A full list of publications can be found at www.sfu.ca/~easton. Svenja Feiler is a researcher at the Institute of Sport Economics and Sport Management at the German Sport University in Cologne, Germany. Since 2011, she is responsible for managing a large-scale panel study on nonprofit sports clubs in Germany, the Sport Development Report. The project contains regular online surveys on nonprofit sports clubs in Germany and their various stakeholders (members, coaches,

Notes on the Editors and Contributors

xvii

board members, referees). She holds a diploma degree in Business Administration and a Master degree in Sport Management (MSc). She is doing her PhD on the various sources of funding for voluntary sports clubs. Her main research interests are nonprofit sports organizations, finances of nonprofit sports clubs, and sport ­development. Rodney Fort is Professor of Sport Management at the University of Michigan and is recognized internationally as an authority on sports economics and business. His work covers a wide variety of sports topics (sites.google.com/site/rodswebpages/academic/cv). Professor Fort is a regular speaker on sports issues and has been a keynote speaker at international sports congresses and a panellist at a variety of universities and institutes. He has testified before the US Senate, the New Zealand Commerce Commission, his signature is on Amicus Briefs to the Supreme Court of the US and government regulatory agencies. He appears frequently in press reports, on television and radio. David Forrest is Professor of Economics in the University of Liverpool Management School and Honorary Professor, Macau Polytechnic Institute. His research interests lie in both sports economics and the economics of gambling behaviour and he has published fifty journal papers in these fields over the last decade. He works with regulators and operators on issues related to problem gambling and with many stakeholders (including player associations, regulators and world governing bodies) in efforts to combat manipulation of sports events. Bill Gerrard is Professor of Business and Sports Analytics in the Business School at the University of Leeds, UK with degrees in economics from the University of Aberdeen, Trinity College, Cambridge and the University of York. His research focuses on the use of data analytics as an evidence-based approach to management. Much of his research has been in professional team sports on such topics as football transfer fees, shirt sponsorship, stadium naming rights, the win-wage relationship, the trade-off between sporting and financial performance, and evidence-based coaching. His most recent work has been on the impact on performance of team-specific human capital (with Prof Andy Lockett) published in the British Journal of Management. He is a former editor of the European Sport Management Quarterly. He has worked with a number of elite teams in both football (soccer) and rugby union including AZ Alkmaar, Saracens and London Irish. Carlos Gomez-Gonzalez is a PhD candidate in Economics and Business at University of Castilla-La Mancha (Spain). He also works as a researcher in the field of sport economics for Group IGOID. His research interests include the areas of sports management and sports economics with a focus on gender studies and discrimination against minority groups. Chris Gratton is Emeritus Professor of Sport Economics and Director of the Sport Industry Research Centre at Sheffield Hallam University, UK. He currently has six academic sports books in print, the latest being The Global Economics of Sport, published by Routledge in 2012. His main areas of research include the economic benefits of major sports events, measuring the economic importance of sport including the use of satellite accounts for sport, and the modelling of large sports participation surveys. He represented the UK on the EU workshop on Sport and Economics and the EU Expert Group on Sport Statistics from 2006 until 2014. Ian Gregory-Smith is a Senior Lecturer in Economics at the University of Sheffield. His primary research interests concern the executive labour market and related issues associated with gender, corporate governance, executive remuneration and shareholder voting. He is also interested in using sporting settings such as professional cricket as a laboratory for testing economic ideas associated with the labour market and decision making in firms. His research has been published in economic and management journals including: The Economic Journal, Oxford Bulletin of Economics and Statistics, British Journal of Industrial Relations and the British Journal of Management. Günther Grohall has been working on applied economic problems since 2002, with more focus on sport economics since 2007, when he joined SpEA. After studying “Business, Economics, and Computer science” at the University of Vienna and the University of Technology of Vienna from 1993 to 2000, he finished post-graduate courses on Quantitative Finance and Corporate Finance. In 2002 he joined the applied economic team of the Vienna Institute for Advanced Studies (IHS), in 2005 Economica, in 2007 SportsEconAustria and teaches since 2009 at the University of Applied

xviii

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Sciences in Vienna. His work focuses on quantitative methods, modelling as well as data analysis. He has been involved in numerous input-output analyses, especially, but not only, in the field of sport. Two examples are the “Study on the Contribution of Sport to Economic Growth and Employment” and its recent update “Study on the Economic Impact of Sport through Sport Satellite Accounts”. Kirstin Hallmann is a Senior Lecturer and Researcher at the Institute of Sport Economics and Sport Management, GSU. Her research interests include volunteer management, sport consumer behavior, elite sports, sport events, and sport tourism. Mikael Jamil is currently a Senior Lecturer within the Department of Science and Technology, University of Suffolk. He is also the Course Leader of BSc (Hons) Sport Performance Analysis and works closely with the Performance Analysis Department at Ipswich Town Football Club in the role of consultant. His research focuses primarily on Performance Analysis in football and his most recent publications include: Intra-system reliability of SICS: video-tracking system (Digital.Stadium®) for performance analysis in soccer (2018), Reliability of internal and external load parameters in 6 a-side and 7 a-side recreational football for health (2018) and The Reliability of Technical and Tactical Tagging Analysis Conducted by a Semi-Automatic VTS in Soccer (2018). He received his PhD in Sports Economics from Loughborough University in 2016. Candon Johnson is a PhD candidate at West Virginia University in Morgantown, West Virginia. His research focuses on sports economics, urban and regional economics, and health economics. He is currently working on a variety of sports economics topics, including racial wage discrimination (co-authored with Eduardo Minuci), game attendance and loss aversion across different Collective Bargaining Agreements, and evidence of present bias in free agency decisions (co-authored with Brad Humphreys) within the National Basketball Association (NBA). He is also conducting research on the impact of the Olympics Games on host cities and the surrounding areas. Outside sports, he is researching the impact of the legalization of recreational marijuana on the consumption of legal goods that are potentially risky, such as alcohol, cigarettes, and smokeless tobacco. Stefan Késenne is Emeritus Professor of economics at the University of Antwerp (UA) where he has been teaching Econometrics, Macroeconomics, Labour Economics and Sports Economics. His most important research area is Sport Economics. He has published in many Internationally reviewed journals and is the author of the well-known textbook: The Economic Theory of Professional Team Sports, an analytical treatment (2nd Ed., 2014). He is also a member of the Editorial Board of the Journal of Sports Economics and of the Scientific Committee of CIES in Switzerland. Themis Kokolakakis is a Reader of Sport Economics at Sheffield Hallam University. He has been working over a number of years with Chris Gratton and Günther Grohall on the development of SSAs in Europe. He has produced numerous studies on economic value of sport using different methodological perspectives, including National Income Accounting and SSAs. He has published widely on the economic value of sport and the social and health impact of sport participation. Research interests include: evaluating sport economies, examining sport participation in relation to economic growth, sport volunteering as a factor of economic growth, the economic impact of sport on health and wellbeing and the economic evaluation of health interventions. Anthony C. Krautmann is the Professor and Chair of the Economics Department at DePaul University in Chicago. He has served previously as Vice-President of the North American Association of Sports Economists (NAASE) as well as the President of the Illinois Economic Association. Professor Krautmann’s research include a focus on labor markets in professional sports as well the industrial organization of sport leagues. His publications have appeared in a number of journals, including the Journal of Sports Economics, Economic Inquiry, and the Southern Economic Journal. Michael Lechner is Professor of Econometrics at the University of St. Gallen and director of the Swiss Institute for Empirical Economic Research (SEW). He holds his PhD from the University of Mannheim (1994) and spent one year at the London School of Economics and at Harvard University. His research focuses on developing and improving econometric methods for causal analysis in the field of microeconometrics as well as applications in the areas of labour, health and sports economics. His recent

Notes on the Editors and Contributors

xix

work centers on the question how machine learning methods can be used to obtain credible causal effects in large (or Big) data settings. Michael Lechner is also affiliated with CEPR, London, CESIfo, Munich, IAB, Nuremberg, and IZA, Bonn. Eva Marikova Leeds is Professor of Economics at Moravian College in Bethlehem, PA, USA. She has written on privatization and mortgage markets in transition economies. Her work in sports economics with Michael Leeds includes the application of event analysis to stadium naming rights and finding determinants of national success in international soccer and at the Summer Olympics. She has also examined pay determination in Japanese baseball and the differences between men and women in how they respond to competitive settings. She is the co-editor of Handbook on Economics of Women in Sports (2013). Her work has appeared in such journals as the Journal of Sports Economics and the Review of Industrial Organization. Michael A. Leeds is Professor and Chair of the Economics Department at Temple University. His research expertise is in the Economics of Sport,Labor Economics, and Applied Microeconomics. His research has appeared in such journals as The Journal of Urban Economics, Economic Inquiry, Social Science Quarterly, and The Journal of Sports Economics. His textbook, The Economics of Sports, with Peter von Allmen and Victor Matheson is a prominent textbook in the field. He is co-editor, with Eva Marikova Leeds of The Handbook on the Economics of Women in Sport. His current research includes work on the economics of baseball in Japan and gender differences in the response to economic contests. Fernando Lera-López is Senior Lecturer in Economics at the Department of Economics of the Public University of Navarra (Spain), where he lectures on Economics and Innovation in Teaching. He is also member of the Institute for Advanced Research in Business and Economics at the Public University of Navarra. His research focuses on sports economics, and particularly, in the analysis of the sports participation and consumption and their social health and wellbeing, and economic effects. His research has been published in leading sport journals such as Journal of Sport Management, Journal of Sport Economics, European Sport Management Quarterly, Sport Management Review, International Journal of Sport Finance, Journal of Sports Sciences and European Journal of Sport Science. He is the treasurer of the Spanish Society of Sports Economics. Wolfgang Maennig is a Professor at the Department of Economics of Hamburg University. He was a visiting scholar at the University of California Berkeley, at MIT, at the American University in Dubai as well as at the Universities of Istanbul (Turkey) and Stellenbosch (South Africa), at the Federal University of Rio de Janeiro and at the University of Economics Bratislava. He was also visiting scholar at International Monetary Fund in Washington, DC, and at Deutsche Bundesbank in Frankfurt. His research concentrates on sport economics and urban issues and he has been published in numerous leading academic journals. He is co-editor of the International Handbook on the Economics of Mega Sporting Events. Wolfgang Maennig has worked as an expert for many bids of large sport events, e.g. the Olympic bids of Berlin 2000, Leipzig 2012, Munich 2018 and the Athletics World Cup Berlin 2009. He was Olympic Champion (rowing, eight with coxwain) at the Olympics 1988 in Seoul and president of the German Rowing Federation, 1995–2001. He holds the Olympic Order. Patrick Massey is a Director of Compecon – Competition Economics, which he established in 2001, based in Dublin and specializing in the economics of competition (antitrust) and regulation. He was previously a member of the Irish Competition Authority for 10 years. Prior to that he worked for the New Zealand Treasury and DKM economic consultants. His books include New Zealand: Market Liberalization in a Developed Economy (1995); Competition Law and Policy in Ireland, (1996) (co-authored with Paula O’Hare); and Competition and Regulation in Ireland: The Law and Economics (2003) (co-authored with Daragh Daly). He has authored/co-authored papers published in various journals including The Antitrust Bulletin; The Economic and Social Review; European Competition Journal; European Sports Management Quarterly; the International Journal of Sport Finance; Journal of Competition Law & Economics; Journal of Sports Economics and World Competition. He has lectured on the economics of competition and regulation at the University of Dublin (Trinity College) and NUI Maynooth. Katharina Moser is Research Assistant and PhD Candidate at the Department of Human Resource Management and Organization, School of Business and Economics at the University of Tübingen. Her research interests include gender diversity on corporate boards, sports economics, and economics of religion. She received her Bachelor degree in Social Anthropology and Gender Studies at the University of Basel and her Master degree in Economics, major in Management, at the University of Lugano, Switzerland.

xx

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Cristina Muñiz is an Associate Professor (Profesor Contratado Doctor) of Economics at the University of Oviedo, Spain. She has an expertise in applying econometric models to examine the engagement of individuals in leisure activities, with particular interest in sports and cultural activities. She has also recently published papers on Ecotourism and Labour Economics. Her research has been published in leading journals such as The European Journal of Health Economics, Economic Modelling, Journal of Cultural Economics, Ambio, and International Journal of Sport Finance. She is a member of the editorial board of the European Sport Management Quarterly. She received her PhD in Economics from the University of Oviedo. Georgios Nalbantis is Research Associate at the Department of Sport Economics, Sport Management and Media Research, at the University of Tübingen, Germany. Born in Greece, he was educated in Germany and Austria. He studied Business Law in Heidelberg, obtained an LLM in European and International Business Law in Vienna as well as a MSc in Sport Management in Cologne. In 2017 he received his PhD in Sport Economics with summa cum laude from the University of Tübingen. His research interests include the economics of league competitions as well as the econometric analysis of (international) sports demand and regulations of sport systems. He co-authored articles published in journals such as Economic Inquiry, Applied Economics and Journal of Sports Economics. Johannes Orlowski is currently a postdoctoral researcher at the Department for Business Administration at the University of Zurich. In 2017, he received his PhD in Sports Economics and Management at the German Sport University Cologne were he also received his MSc in Sport Management in 2013. Johannes is interested in uncovering natural experiment settings in sports data and utilizing them for economic and managerial research questions. In his research, he has addressed health-related aspects of participation in sport and physical activity, labour markets in sport, and monetary valuation of intangibles associated with sports. David Paton holds the Chair of Industrial Economics at Nottingham University Business School and is Associate Dean within the School. David gained his PhD in economics from University College London. His research covers a range of topics including the economics of teenage pregnancy, betting markets, gambling taxation, the economics of suicide and economics of sport especially cricket. David has published over 50 academic papers in journals such as the Economic Journal, Economica, Demography and The Journal of Health Economics. He is Co-Editor of the International Journal of the Economics of Business. David has acted as an advisor to Government Departments such as DCMS, HMRC, and the Audit Office and his research has been featured extensively in the media including appearances on Newsnight, the Today Programme, Women’s Hour, BBC Radio 5, Voice of Russia, Channel 4 news etc. Rodney J. Paul is Professor of Sport Management and Director of the Sport Analytics Program in the Falk College of Sport and Human Dynamics at Syracuse University. He completed his PhD in Applied Economics at Clemson University in 2000. He has over 100 publications related to the economics and finance of sport in various journals and book chapters. His main research interests are in sports betting markets, attendance and television viewership models of sports teams and leagues, and the economics of professional hockey. Holger Preuss is Professor of Sport Economics and Sport Sociology at the Johannes GutenbergUniversity in Mainz, Germany, and at the Molde University College, Norway. He is also adjunct professor at the University of Ottawa, Canada, and international scholar at the State University of New York (Cortland). He is past editor of the journal European Sport Management Quarterly and currently is associate Editor of the Journal of Sport & Tourism. He has published 17 books and more than 125 articles in international journals and books. His field of research is the socioeconomic aspects of sport. In particular, his research focuses on impact analysis of mega-sport events (Olympic Games since 1972 and FIFA World Cup since 2006). A main theme examines cost and revenue overruns at Olympic Games and he is constantly developing a framework and measurement of Olympic legacy events. He is a member of the IOC Legacy and Sustainability Commission and the UIPM Development Commission. Joachim Prinz is a Professor of Business Administration u6the Mercator School of Business at the University of Duisburg-Essen. Generally he is interested in all “exotic” economic issues but mostly his research and teaching considers topics professional team- and endurance sports.

Notes on the Editors and Contributors

xxi

J. James Reade is an Associate Professor of Economics at the University of Reading. Prior to this he was Lecturer in Economics at the University of Birmingham. He completed his DPhil at the University of Oxford in 2007. His research has appeared in a number of top economics, management and operations research journals, as well as chapters in a number of handbooks and other collections. His research is in applied economics, primarily with a focus on sport and gambling-related topics. Daam Van Reeth is Senior Professor of Economics in the Faculty of Economics and Business at the KU Leuven, Belgium. He teaches courses in Micro- and Macro-economics. His research interest is in the economics of sport, with a focus on professional road cycling and media attention for sport. Research topics he worked on include TV demand for the Tour de France, doping perception, gender balance in sports coverage, and fantasy sports. He is editor and co-writer of The Economics of Professional Road Cycling and he has published numerous articles on media interest in the Tour de France. Duane W. Rockerbie is a Professsor of Economics at the University of Lethbridge. He received his PhD at Simon Fraser University in 1990. He is a past editor of the Journal of International Financial Studies, a member of the Editorial Board of the Journal of Sports Economics and a past Vice-President of the North American Association of Sports Economists. His research has focused on sports economics, economics of higher education and international finance. Duane has published over 60 articles in economics journals and book volumes and has published two textbooks. He has also performed consulting work for the Canadian Soccer Association and CONCACAF. Placido Rodriguez is Professor EU of Economics in the Department of Economics at the University of Oviedo, Spain. Doctor in Economics and Law Degree. He is the co-editor of several books on Sports Economics and he has published several papers in specialized journals such as the Journal of Sports Economics, European Sport Management Quarterly, International Journal of Sport Finance, Journal of Cultural Economics, Journal of Media Economics, Economic Modeling, Rivista di diritto ed economia dello sport, International Journal of Sport Management and Marketing, Journal of Sports Economics & Management, Ambio, Revista de Economía Aplicada, Estudios de Economía Aplicada, Intangible Capital, Revista Asturiana de Economía or Revista de Psicología Aplicada. He was formerly President of Real Sporting de Gijon Football Club and currently is the Director of the Fundación Observatorio Económico del Deporte, Honorary President of the IASE (International Association of Sports Economists) and has been awarded with the Larry Hadley Service Award 2018 by the North American Association of Sports Economists (NAASE). Lea Rossi is a PhD candidate at the Institute of Sport Economics and Sport Management at the German Sport University Cologne. Her research interests include non-profit sport clubs, commercial sport providers, sport volunteers, and elite sports. Abhinav Sacheti works as an economist for a multinational professional services firm in the United Arab Emirates, where he uses applied economics to advise clients in the public and private sectors across a variety of settings. He obtained his PhD on the economics of cricket from the University of Nottingham and his research interests continue to center around the economics of sport, particularly cricket. He is especially interested in decision-making by players and officials. His research has been published in journals such as the Journal of the Royal Statistical Society, Applied Economics, Economic Record and the Journal of Sports Economics. Allen R. Sanderson is a Senior Lecturer in Economics at the University of Chicago. A graduate of Brigham Young University and the University of Chicago, he came to Chicago from Princeton in 1984; served eight years as associate provost of the University; and has also been a senior research scientist at the National Opinion Research Center, where his contributions include research on education, labor markets and affirmative action. In addition to his popular two-quarter sequence on principles of microeconomics and macroeconomics, he teaches a course and does research on the economics of sports. He has also led an interdisciplinary team-taught course on “Sport, Society and Science.” He has received a Quantrell Award for Excellence in Undergraduate Teaching and has the distinction of having taught more students at Chicago than anyone in the history of the University. Nicolas Scelles is Senior Lecturer in the Manchester Metropolitan University Business School. He is a member of the Sport Policy Unit sitting in the Department of Economics, Policy and International

xxii

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Business. He is a sport economist having published extensively in journals such as Applied Economics, Applied Economics Letters, International Journal of Sport Finance and Journal of Sports Economics. His main topics are competitive balance and intensity, determinants of stadium attendance and TV audience, determinants of financial value and insolvencies of European professional sports clubs, specifically in football. His recent work includes research on the development of women’s sports. Dr Nicolas Scelles is also a member of the European Sports Economics Association (ESEA) and the Boards of the International Society for Sports Sciences in the Arab World (I3SAW) and the Scientific Council of the French Observatory of the Sports Economy. He works with international sports organizations such as UEFA. Katrin Scharfenkamp is a postdoctoral researcher at the Mercator School of Management (University of Duisburg-Essen). Her research focuses on diversity in organizations and teams, and in particular on the mechanisms in corporate boards, political economy and sports economics. Her recent work centres on the effects of socio-demographic as well as functional differences in groups on the individual or team outcome. Her research has been published in leading journals such as the European Journal of Political Economy, Managerial and Decision Economics and the International Journal of Sports Finance. She received her PhD in Economics from the University of Münster. Ute Schüttoff works as Lecturer at the Institute of Sports Science at the University of Tübingen with focus on sports economics and sports media research. Her research interests include amateur and leisure sports participation, in particular the economic effects of physical activity or sports participation, and sport in the media. Further, she works on a project on (state-funded) sports promotion in Germany. Her latest studies have been published in journals such as Social Science Quarterly or Journal of Sports Economics. Stephen Shmanske is Professor of Economics Emeritus at California State University, East Bay. He earned a B S in Mathematics from Dartmouth College and a PhD in Economics from the University of California, Los Angeles. He has published dozens of articles in professional journals in many areas of interest including sports economics, transportation economics, price discrimination, and the economics of education. He is the author of Golfonomics and Super Golfonomics and was a pioneer in applying economics to the sport of golf and in using statistics from the golf industry to address issues of general economic concern such as discrimination and production theory. He is the co-editor of The Oxford Handbook of Sports Economics published in two volumes in 2012. John J. Siegfried is Professor of Economics Emeritus at Vanderbilt University. He earned a B.S. from Rensselaer Polytechnic Institute (1967), and an M.A. from Penn State (1968). He was on Vanderbilt University’s economics faculty from the time he earned a Ph.D. from the University of Wisconsin in 1972 until retiring in 2010. In 1975 and 1976 he served as a senior staff economist at the Federal Trade Commission, and on President Ford’s Council of Economic Advisers. He chaired the Vanderbilt Economics Department from 1980 to 1986. From 1997 through 2012 he was Secretary-Treasurer of the American Economic Association. His research has been in industrial organization, antitrust, economics of higher education, economics of sports, and the teaching of economics. He co-authored Economic Challenges in Higher Education (Chicago Press) in 1991. His latest book is Better Living Through Economics (Harvard Press, 2010). Carina Steckenleiter is a PhD student in Economics and Finance at the University of St. Gallen and works as a Research assistant at the Swiss Institute for Empirical Economic Research (SEW). Her research interests are in applied econometrics with a focus on labour, education and sports economics. Philipp Swierzy is a PhD student at the Institute of Sport Economics and Sport Management at the German Sport University Cologne. As part of his PhD, he conducts research projects with a focus on the relationship between individual behaviour and its organizational context. His main research interests include behavioural economics, economics of nonprofit organizations and labour economics. He has a bachelor degree in Business Administration and Economics and a master degree in Sport Management. Marijke Taks is a Professor in the Faculty of Health Sciences at the University of Ottawa. Her area of expertise is in socioeconomic aspects of sport and leisure. Her grant-supported research focuses particularly on impacts, outcomes and leveraging of small and medium-sized sport events, and their

Notes on the Editors and Contributors

xxiii

meaning for host communities. She also studies sport consumer behaviour of various groups in society. Mass participation and the ‘Sport for All’ philosophy guide her research. She has published her work in leading journals of sport management and related fields. She is past-editor of the European Sport Management Quarterly, associate editor of the Journal of Sport Management and the Journal of Global Sport Management, editorial board member of the European Sport Management Quarterly, and guest reviewer for a wide variety of journals in the field of sport management, sport marketing, sport finance and sport tourism. Nicholas Watanabe is an Assistant Professor of Big Data and Analytics in the Department of Sport and Entertainment Management at the University of South Carolina. Dr Watanabe’s research predominantly focuses on the intersection of economics, management and communications in the context of sport, with special focus on the sport marketplace in the digital era. His work has been published in top journals in sport management, including the Journal of Sport Management, International Journal of Sport Finance, and Sport Management Review. Currently, he serves on the editorial board for the Journal of Sport Management, Journal of Leisure Research, Journal of Issues in Intercollegiate Athletics, Sport, Business and Management: An International Journal, and Managing Sport and Leisure. In 2018, he was named a Research Fellow of the North American Society for Sport Management. Daniel Weimar is a postdoctoral researcher at Mercator School of Management at the University of Duisburg–Essen. His research concentrates on empirical investigations in the areas of sports economics, personell economics and education economics. Pamela Wicker is a Senior Researcher and Lecturer at the Department of Sport Economics and Sport Management at the German Sport University Cologne where she obtained her PhD in 2009 and her Habilitation in 2015. In 2011 and 2012, she was employed as a Senior Lecturer in Sport Management at Griffith University, Australia. Her research interests include non-profit sport organisations, monetary valuation of intangibles in sport, determinants and outcomes of sport participation and physical activity, and labour economics. She currently serves as Associate Editor at Sport Management Review, European Sport Management Quarterly, and the Journal for Study and Teaching in Sport Science, and is an editorial board member of another six journals (Journal of Sports Economics, International Journal of Sport Finance, Journal of Sport Management, European Journal for Sport and Society, Journal of Sport and Tourism, Managing Sport and Leisure).

Part I

The Nature and Value of the Sports System and Economy

1 Introduction Paul Downward, Bernd Frick, Brad R. Humphreys, T i m P a w l o w s k i , J a n e E . R u s e s k i a n d B r i a n P. S o e b b i n g

The Development and Scope of Sports Economics Sports Economics is now a well-established field of economics that contributes to the development of a theoretical and empirical understanding of sport in the economy and informs policy and management. From the modest beginnings of seminal applications of economic enquiry to the analysis of sport and the labour market in the 1950s to the 1970s (Rottenberg, 1956; Neale, 1964; Sloane, 1969), and analysis of the financial fragility of football (Sloane, 1971), the literature has grown to embrace textbooks covering the UK and Europe (e.g. Gratton & Taylor, 1985, 2000; Downward & Dawson, 2000; Downward, Dawson, & Dejonghe, 2009) and the US (e.g. Fort, 2011; Leeds & von Allmen, 2018); and both edited volumes (e.g. Humphreys & Howard, 2008; Rodriguez, Késenne & Humphreys, 2011; Goddard & Sloane, 2014) and research monographs (e.g. Szymanski & Kuypers, 1999; Szymanski & Zimbalist, 2005; Kuper & Szymanski, 2012; Késenne, 2014), and many more, as indicated by the growing title list of major publishers.1 The literature is now also supported by dedicated journals in sports economics (the Journal of Sports Economics, the International Journal

of Sports Finance) and many researchers publish regularly in economics discipline journals (such as Economic Inquiry, the Journal of Economic Behavior & Organization, Applied Economics, Labour Economics, Kyklos, the Journal of Economic Psychology, the Journal of Health Economics, Oxford Economic Policy, the Oxford Bulletin of Economics and Statistics, the  Economic Journal, the Scottish Journal of Economics, Contemporary Economic Policy, etc.). This academic field contains a strong research community represented by the North American Association of Sports Economists (NAASE), the European Sport Economics Association (ESEA) and the International Association of Sports Economists (IASE).2 These organizations hold regular conferences and are linked to the dedicated economics journals associated with sports, i.e. the Journal of Sports Economics and the International Journal of Sports Finance. Sports economics papers are also regular components of international sports management conferences and sports economics research regularly appears in the major sports management journals, such as European Sport Management Quarterly, The Journal of Sport Management and Sport Management Review. Each of these journals is also associated with the respective international associations: the European Association for Sport

4

The Sage Handbook of Sports Economics

Management (EASM), the North American Society for Sport Management (NASSM), and The Sport Management Association of Australia and New Zealand (SMAANZ).3 Some sports science journals also regularly publish sports economics papers (such as the European Journal of Sport Science and the Journal of Sports Sciences) as they add to relevant social scientific contributions to multidisciplinary sports science. Research also finds an outlet in area journals in which theoretical, empirical and policy outcomes of sports economics research lead to valuable insights (such as The International Journal of Sport Policy and Politics, The Journal of Physical Activity and Health and Voluntary and Non-Profit Sector Quarterly). This Handbook aims to produce an up-to-date benchmark collection of insights that both maps the field and identifies lines of enquiry for future research for both existing and new researchers. In such a vibrant area of developing research insight, the coverage of the volume cannot be exhaustive. However, the volume covers a wide range of aspects of sport that provide inroads into the literature. These range from: the nature and value of the sports system and economy (Part I); amateur sports participation and its supply and impact (Part II); professional team sports, which is the most traditional focus of sports economics (Parts III and IV); sports events and their impacts (Part V); individual sports, that also embrace team sports that are competed in by individuals (Part VI); and, finally, some areas of research that are emerging in the literature as important general potential fields for future research (Part VII). Each chapter, though having different emphases and structure, will nonetheless typically outline the main areas of research associated with their topic and indicate specific areas for future research. Authoritative contributors to the field, supplemented by new researchers, have contributed to the volume.

Overview Part I: The Nature and Value of the Sports System and Economy In Part I, Andreff identifies the emergence and development of sports systems as comprising two alternatives: closed leagues rooted in American professional baseball and open leagues in English professional soccer. The stylized organizational features of the two systems are compared and a future research agenda is identified based on the convergence of the organization of sports.

Of course, the organization of sport goes beyond professional sports. Consequently, Kokolakakis, Gratton and Groball explore how, since the mid1980s, systematic attempts to identify the value of sport to the economy have been developed to inform policy.

Part II: Amateur Sports Participation, Supply and Impact Part II examines direct engagement in sport as a participant as well as a volunteer, thus contributing to the supply of sporting opportunity. The expenditures and monetary donations of those who engage in sport are also explored in sustaining sporting systems. The key aspects and challenges facing European sports clubs are also investigated, along with the unique economic nature of US college sports. Downward and Muñiz initiate discussion of the analysis of the demand to participate in sport, showing that participation in sport can be understood as a time allocation decision across a range of leisure activities, but also in the context of the household and wider social engagement in the face of available supply of opportunity. The impacts of participation upon health are then explored by Ruseski, who argues that as well as seeking greater causal research designs, a more careful assessment of the dose–response relationship between sport participation and health outcomes is required. Pawlowski and Schüttoff also argue for a need for greater causal research in their review and analysis of the association between sports participation and social capital, which they argue is grounded in the associational nature of sport, such as playing in a team or belonging to a club. Finally, Steckenleiter and Lechner explore how participation can influence human capital and subsequent labour market success, also arguing for a need to have better causal insight into the processes. The final chapter on the demand for sport by Lera-López examines the expenditure behaviour associated with participation as well as wider sports-related expenditures by households and individuals. The remaining chapters in Part II explore the supply side of mass participation and engagement in sport. Breuer, Swierzy and Feiler explore nonprofit sports clubs across Europe, arguing that their impact on the European economy is considerable, although difficult to quantify. They argue that greater research is needed on how the organizational and resource aspects of sports clubs may differ across European countries. Wicker consequently begins the analysis of the key voluntary

Introduction

labour resource of such clubs. Wicker argues that while the micro-determinants of volunteers are well understood, more research is needed on the meso- and macroeconomic determinants of volunteering as well as its impacts. Hallmann and Rossi extend the discussion of volunteering, as a time donation, to examine how monetary donations are relevant to the resourcing of sports organizations. They also argue that greater institutional and national levels analysis is required. Sanderson and Siegfried close the section by exploring the uniquely American mass participation opportunity provided through college sports. They focus on the balance of costs and benefits required by US educational institutions in subsidising sport in an environment in which the academic benefits are not clear-cut.

Part III: Professional Team Sports The discussion of professional team sports begins with the consideration of sports leagues as markets. Fort surveys the variety of objective functions used in the analysis of team sports and identifies areas where the careful choice of objective functions in future work will prove particularly important. Fort argues that evaluating sports league outcomes depends, crucially, on the relevant objective function assumed. In turn this requires appreciating the context. In this regard, Fort argues that the growth of Asian sports leagues and e-sports need to be researched. Scelles and Brocard explore the origin and rise of European sports leagues, focusing on the changes encountered by European leagues since the 1980s, with a considerable increase in the economic and financial resources available. They argue that an appropriate league design lies in a European Super League. Budzinski also recognises the context specificity of the analysis of sports leagues, arguing that competition policy in sports markets depends on the respective national (or supra-national in the EU case) law and its enforcement. Subsequent discussion develops insight into the component parts of sports leagues as markets. With respect to demand, Pawlowski and Nalbantis explore the competitive balance (CB) and Uncertainty of Outcome Hypothesis (UOH) literatures. Sketching some major findings from both lines of research, they suggest avenues for future work to advance the understanding of the complex relations between CB, suspense and the demand for sport. Rodríguez surveys the demand for team sport events and argues that future research should draw on advances in econometric techniques and new theoretical ideas from behavioural economics.

5

Buraimo focuses on the complexities of the broadcast market and its interaction with the sport in the context of football, showing how the impact of exposure on audience ratings is important for clubs as well as broadcasters. Soebbing focuses on the (primary) ticket pricing decisions made by clubs, and insights associated with price discrimination and price elasticity of demand. Courty subsequently explores the development of and pricing in secondary markets for tickets. The remaining chapters of this section of the Handbook focus on the supply side of the sports market. Késenne explores the ­persistence and implications of the transfer market for in-contract players since the Bosman-verdict of the European Court of Justice. Jamil then explores how the labour inputs of players are harnessed in the production of sport and how efficiency is measured and evaluated. It is noted that there is a general lack of attention afforded to the opponent and their impact upon observed performances in the literature, which could be addressed. Reade then focuses on the other labour input to sports events – officials – in an analysis of home bias. Reade explores the various competing mechanisms proposed for its existence and illustrates this through the analysis of cricket. This part of the book concludes with Coates’ analysis of capital inputs to sport and the (potential) net benefits from the relocation of professional sports franchises both between or within cities. Mixed ­evidence is noted.

Part IV: Professional Sports Leagues In Part IV, a number of chapters focus on particular professional team sports, further developing the context of many of the issues outlined in the chapters of Part III. Weimar explores football (soccer), Gregory-Smith, Paton and Sacheti examine cricket, Massey explores rugby, Berri looks at basketball and Blemmings examines American Football. These chapters are followed by Krautmann’s analysis of the baseball players’ market, Rockerbie and Easton’s examination of the National Hockey League (NHL), Booth and Brooks’ analysis of Australian Rules football and Watanabe’s exploration of Major League Soccer (MLS), noting in particular its unique single entity model of ownership.

Part V: Sports Events Part V of the Handbook switches attention to sports events, that is, tournaments that are more

6

The Sage Handbook of Sports Economics

episodic and might rotate across venues. Preuss assesses economic impact measurement connected with the Olympic Games, to recommend best practices and to point to typical mistakes in many evaluations. Maennig then explores the need for such work as the well-known lack of ability of events to generate their proposed economic benefits noted in the economic literature is now finding traction in public discourse. Recent public referenda on Olympic bids have – with few exceptions – resulted in the voters deciding against the Olympic ambitions of the local governments and/or local bid committees. Maennig explores this development and its implications. Leeds concludes the sequence of chapters on the Olympics by identifying the factors that drive national performance in medal counts by different countries. Though drawing on similar themes as the analysis of the Olympics, subsequent chapters focus on major and minor events more generally. Johnson discusses the economics of hosting sports mega-events, the financing of these events, and the process through which the rights to host sports mega-events are assigned to cities and regions. The chapter documents the presence and persistence of large cost overruns associated with hosting mega-events and discusses their likely causes. Johnson also reviews the potential economic impacts associated with these events. Maennig further considers these issues, identifying that economic impact analyses are often part of public relations exercises which lack substance. Taks and Agha, in contrast, focus on the much more prevalent and less researched context of minor events and minor league teams (development teams) across the globe. They conclude that calculating the economic impacts of events and teams remains a challenge and is often incomplete. The above chapters essentially have the economic impact of events as the key policy outcome that is used to justify public investment in hosting event and building infrastructure. Other less tangible outcomes are also claimed to occur from such investment. Dawson examines if increased participation in sport and physical activity takes place and the extent to which elite athletes and elite sport provide inspiration to enact behaviour change among the general population. The context is to enhance the health and well-being of the nation. It is concluded that there is limited evidence of such effects. Orlowski and Wicker explore the methodological challenges facing attempts to measure the intangible effects, such as reputation and pride, in economic analyses because they are non-marketed estimates. Finally, Wicker and Downward reflect more generally on the research studying the intangible impacts of sport on health and well-being, pride, social capital, as well as traffic, congestion

and pollution, and different types of criminal behaviour. They finish by identifying lines of enquiry for future research.

Part VI: Individual Sports Part VI of the Handbook turns its attention to individual sports, or team sports that are essentially competed in by individual contests. Each chapter highlights issues for future research. Frick, Moser and Scharfenkamp examine both professional middle- and long-distance running as well as recreational running, exploring issues such as the incentive and selection effects of prize money in the former case and self-selection and (over-)confidence on recreational runners’ marathon performance in the latter case. Del Corral and Gomez-Gonzales examine tennis and other racket sports, exploring demand behaviour, the betting market and performance modelling. Van Reeth analyses professional road cycling, outlining its institutional setting and league structure as well as the main challenges road cycling currently faces. Shmanske examines golf, its production, decisions to participate in tournaments, and the elicitation of effort in tournaments. Prinz considers triathlon, describing the organization of the events. He illustrates their location and gives some insights into the triathlon market, specifically athlete effort and output and the demand for triathlon. Finally in this section, von Allmen explores NASCAR racing, its structure and the nature of demand.

Part VII: Future Research The final part of the Handbook covers issues that are considered to be new and general enough to require further systematic development and ana­ lysis. This is not withstanding the general sentiment of many of the contributions of the Handbook that better causal econometric work is needed in the field of sports economics and, for example, that greater attention to behavioural economics is needed. In this regard Chikish and Humphreys identify that a behavioural economic perspective does provide new insights into old problems in sports economics and also helps to explain many outcomes that cannot be easily explained using standard economic models that dominate the sports economics literature. This arises, for example, in the role played by biases reflecting non-standard preferences for athletes, referees, teams, and firms. Leeds focuses on potential gender differences, in whether contests

Introduction

induce different behaviour from women and men, and hence whether high-states settings form an insuperable barrier to gender equality. Gender issues, discrimination and female sports in general are largely unexplored areas of sport economics. This is unlike, for example, the issue of racial discrimination, where a large literature exists. Paul provides insight into the increased use of dynamic pricing by sports teams, extending the need for greater insight into endogenous prices compared to traditional insights. Forrest examines the constant historic dependence between sports and betting, and how interest now mainly focuses on the potential for the manipulation of sport by those looking for gains in the betting market, achieved by bribing players to under-perform. Dimant and Deutscher continue to analyze the theme of corruption by exploring the literature, focusing on the antecedents and effects of doping. It is argued that doping can distort fair competition and trust, and consequently corrupts the entire sports system. They conclude that the fight against doping can only succeed with strong regulatory bodies in place. Finally, Gerrard explores the rise and role of data analytics as a management tool in professional team sports, showing how the applications of data analytics to player recruitment is underpinned by research on player valuation and player rating in sports economics.

Notes 1 

www.e-elgar.com/shop/handbook-on-theeconomics-of-sport; www.routledge.com/sport/ 2  www.byuresearch.org/naasportseconomists/; http://sporteconomics.eu/; and www.iasecon.net/ 3  www.easm.net/; www.nassm.com/; and www. smaanz.org/

References Downward, P., & Dawson, A. (2000). The economics of professional team sports. London: Routledge. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports economics. London: Routledge.

7

Fort, R. D. (2011). Sports economics (3rd ed.). Englewood Cliffs, NJ: Prentice-Hall. Goddard, J. A., & Sloane, P. J. (2014). Handbook on the economics of professional football. Cheltenham: Edward Elgar. Gratton, C., & Taylor, P. (1985). Sport and recreation: an economic analysis. In Sport and recreation: an economic analysis. London: E and F Spon. Gratton, C., & Taylor, P. (2000). Economics of sport and recreation. In Economics of sport and recreation (2nd ed.). London: Routledge. Humphreys, B., & Howard, D. R. (2008). The business of sports: economic perspectives on sport. Greenwood. Késenne, S. (2014). The economic theory of professional team sports: an analytical treatment (2nd ed.). Cheltenham: Edward Elgar. Kuper, S., & Szymanski, S. (2012). Soccernomics: why England loses, why Spain, Germany, and Brazil win, and why the US, Japan, Australia, Turkey and even Iraq are destined to become the kings of the world’s most popular sport. New York: Nation Book. Leeds, M., von Allmen, P., & Matheson, V. (2018). The economics of sports. London: Routledge. Neale, W. C. (1964). The peculiar economics of professional sports: a contribution to the theory of the firm in sporting competition and in market competition. The Quarterly Journal of Economics, 78(1), 1–14. Rodríguez, P., Késenne, S., & Humphreys, B. R. (2011). The economics of sport, health and happiness: the promotion of well-being through sporting activities. Cheltenham: Edward Elgar. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Sloane, P. (1969). The labour market in professional football. British Journal of Industrial Relations, 7(2), 181–199. Sloane, P. J. (1971). The economics of professional football: the football club as a utility maximiser. Scottish Journal of Political Economy, 18(2), 121–146. https://doi.org/10.1111/j.1467-9485.1971.tb00979.x Szymanski, S., & Kuypers, T. (1999). Winners and losers: the business strategy of football. London: Viking. Szymanski, S., & Zimbalist, A. (2005). National pastime: how Americans play baseball and the rest of the world plays soccer. Washington, DC: Brookings Institution Press.

2 Origins and Developments of Sports Systems Wladimir Andreff

INTRODUCTION A sports system is an organization that regulates how athletes and teams are rewarded and allowed to move according to their sporting merit (the Olympics system) or economic criteria (the North American franchise system). A set of rules may organize the system as a vertical ladder on which each athlete/team moves bottom-up (promotion) from the lowest bar (division) to the highest and top-down (relegation). A prerequisite is that each division must be open upwards and downwards which defines an open or divisional league system. An alternative option is to organize a sports system as a series of horizontal layers (leagues) with fixed membership, hermetic to each other in that the entrance of players/teams from outside is tightly restricted in such a way that regulated mobility primarily operates within the same layers. This results in a closed league system where incumbent franchise holders have the last say about the entrance of new teams and new players limited to those registered on a draft. Szymanski (2004a, p. 36) claimed that ‘economists have failed to present any detailed reasoning on the pros and cons of alternative (sports) systems’. This chapter challenges such view and surveys, in the case of team sports, the origins and

the current state of the two systems. First, it traces back to the economic history of closed leagues rooted in American professional baseball and open leagues in English professional soccer; then the stylized organizational features of the two systems are compared. In the next section, a separate economic modeling of the two systems is exhibited based on different assumptions – teams are assumed to be profit-maximizing in closed leagues while win-maximizing in open leagues – and theoretical backgrounds – Walras versus Nash equilibrium. Different organization entails different regulation of the labor market for talent; however, in both systems a deregulation trend is at work though not with a same intensity. Finally, in regard to cartelized revenue sharing, redistribution is more extensive in closed than in open leagues, and more geared toward within the league in the former whereas the latter share revenues with lower-tier amateur sport.

ORGANIZING MOBILITY IN CLOSED AND OPEN SPORTS SYSTEMS The historical origins of sports systems date back to the first top professional sport league, which

Origins and Developments of Sports Systems

was formed in 1876 in the US with the National League (NL) baseball – the ancestor of MLB – excluding weak teams and granting territorial monopolies to the eight cities represented in the league. Entry and exit required almost unanimous support from incumbents. When rival leagues emerged, they were driven out of business or absorbed to preserve the single cartel of the (meanwhile renamed) MLB. The latter’s expansion in number of teams after 1903 was controlled within a cartel framework. The NHL (1917), NFL (1920), NBA (1949), MLS (1996) and Australian rules football in 1877 (Pomfret & Wilson, 2016) were launched as closed leagues. The closed system has spread namely to English cricket (1894), Japanese baseball (1934) and the Japanese X-League (1971) of American football, baseball (1981) and American football (1996) in South Korea, and the Indian Cricket League (2007). Neale (1964) stressed that, each game being the joint product of two teams, clubs must both compete and cooperate, thus a league is prerequisite to continuous production according to the same rules, and each team needs strong rivals in the league to attract spectators all season long. In 1879, NL team owners introduced a reserve clause over five players per team (Eckard, 2001), extended in 1889 to all players in a team, thus phasing out teams’ overbidding for players. The closed league became a monopoly on its product market and a monopsony on the labor market for talent. The restricted move of players and no downward move of teams became the two pillars of the system. The National Association of Baseball Players, established in 1858, banned the paying of players and transfers between teams, but these rules were disregarded as the sport gained spectator appeal. An open sports system emerged in England with the first knock-out soccer competition (the Football Association (FA) Cup) in 1871, then with the start-up of the English Football League (EFL) (1888) and the Rugby League (1895) running round-robin championships on a regular homeand-away basis (Vamplew, 2006). EFL teams accepted from the beginning to release players to play international competitions with the FA national squad. English soccer had to face a fastgrowing spectator demand in expanding urban areas linked to industrialization, and the subsequent creation of new clubs and rival leagues. To cope with such mushrooming, a second division was established in 1892 with 16 teams, a third division in 1920, then divided into a southern and a northern conference of respectively 22 and 20 teams in 1921 (totaling 88 teams overall), and a fourth division picking up the weakest teams of

9

the two regional conferences in 1958. The number of clubs in the league’s four divisions stabilized at 92 in 1950.1 The salient difference between these two histories is that North American professional team sports leagues maintained a single closed cartel system while English professional soccer remained in open access until 1950 when every medium and large city in England and Wales had a team in the EFL The development of the two sports systems over more than one century has fixed two sets of stylized organizational features covered in the literature (Cairns, 1987; Hoehn & Szymanski, 1999; Noll, 2003a; Szymanski, 2003; Szymanski & Zimbalist, 2005; Andreff, 2011). Table 2.1 summarizes features that differentiate closed and open sports systems. From an economic standpoint, both types of league are cartels monopolizing the supply side of the product market, though in the closed system the league, exempted from the antitrust legislation in the US, behaves as the cartel manager (Jones, 1969), coordinates teams and shares revenues. Such a league has sometimes been considered as a single firm – and its teams as plants2 (Neale, 1964), or as a joint venture3 (Noll, 2003a); though Noll (2006) contends that teams and leagues are vertically related enterprises in the production of a contest points. An open league is more unstable a cartel, subject to the provision of antitrust legislation in Europe, where independent clubs have become firms, i.e. competing (limited-liability or stockholding) companies in the markets for outputs and inputs. Competition operates in the market of an open system whereas it operates for market expansion in a closed league. A primary difference is that the closed cartel system allows teams to avoid competition within local markets by providing them a territorial exclusivity but permits market entry through expansion of a league or the formation of a new rival league.4 Rejected candidates to the major league are incited to launch a rival league. Since the inception of the four North American major team sports leagues, 13 rival leagues emerged, seven to challenge the NFL, three the MLB, two the NBA and one the NHL. In a closed system, the only potential competition shows up at the league level. New leagues are formed to take advantage of spectator excess demand for professional sports (Noll, 2002) since the franchise system generates shortage on the supply side. North American major leagues typically limit the number of franchises to ensure that some cities in which there would be sufficient demand to sustain a major league team miss out

10

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 2.1  Stylized organizational features of closed and open sports systems Features

Closed sports system

Open sports system

Overall organization

Organizational independence of the domestic major leagues, collectively controlled by team owners Self-defined rules in each domestic major league No player released by the teams to play in a national squad Potential for creating rival major leagues

Top leagues integrated in a global hierarchy ruled by an overarching governing body (e.g. FIFA) Global rules enforced by an international governing body (federation) National associations select prestigious national squads, teams release players No rival league allowed in the global/national hierarchy Premier league or top division open downwards to competitors Lower open professional and amateur divisions

Rules National squads Competition between leagues Top of hierarchy Lower tiers International contests Major league/top division League’s entry and exit

Closed major league (with possibly distinct conferences) Shallower hierarchy with closed minor and intercollegiate leagues No international (inter-major league winners) champions league Closed cartel power

Expansion franchise sales (super majority vote); high expansion fee; rare exit Membership Fixed number of teams (until the next franchise sale) Team location Exclusive territory determined by the league* (other team owners) Team relocation Allowed, submitted to league approval on economic (market) criteria Market supply Team’s local monopoly Market demand (inhabitants Organized shortage: 10 million (4 North per team/club**) American major leagues) Antitrust policy Antitrust exemption (USA) TV rights sale Pooling sale of national broadcast rights by the league Merchandising, sponsorship, Pooling sale of some commercial items parking, etc. Teams/clubs League franchised entities Access to capital market Restrictions preventing the stock market flotation of teams Team/club mobility Horizontal: territorial franchise mobility Competition across Limited substitution by consumers, high teams/clubs fan loyalty Team/club objectives 1/ profit; 2/ wins Governance Teams’ owners appointing a league’s central Commissioner Labor market National, few international developments Restrictive regulation then… Reserve clause Deregulation Free agency for veteran players Player mobility through Low; barter trade; limited player trading market operation for cash

International contest (e.g. UEFA’s) between domestic leagues’ winners Inner competition and downward openness alleviates cartel power Annual promotion and relegation on merit; league entrance for free Unstable (turnover). Free entry of new teams in the bottom division Nonexclusive territories, multi-team cities, determined by merit Unusual and contested (e.g. Wimbledon soccer club); no ability to relocate Oligopolistic competition across the clubs Less shortage: 3 million (European soccer ‘Big Five’ top divisions) No antitrust exemption (EU) Pooling or (club’s) individual sale of national broadcast rights Individual club merchandising and sponsors Independent companies No restriction on the stock market flotation of clubs Vertical: to lower/upper division Significant potential for substitution and changing fan loyalty 1/ wins; 2/ revenues Club managers reactive to (recruitment) fan pressures International/global since 1996 Retain and transfer system until 1995 Post-Bosman free mobility High; player trading for cash in sequential mercatos; arms race for superstars (Continued )

Origins and Developments of Sports Systems

11

Table 2.1  Stylized organizational features of closed and open sports systems (Continued ) Features

Closed sports system

Open sports system

Recruitment

Draft rules: teams’ monopsony rights in player acquisition Roster limits Monopsony Collective bargaining and strong unionization Soft cap: NBA (1984), MLS (1996); hard cap: NFL (1994), NHL (2006) Teams in minor and college sports leagues operate as nurseries Tough, long and costly

No entry draft: clubs competing for talent

TV revenues sharing Through taxation

Cartel redistribution across the teams within the major league Yes, except local TV rights Luxury tax: MLB (1997) and NBA (2003)

Gate sharing

MLB and NFL

Other revenue sharing

Intra-league, e.g. MLB tax on teams’ net local revenues (since 1996) Municipal investment in stadiums to attract (or keep) franchises

Redistribution of TV revenues in top division and lower division(s) Only national league’s TV rights revenues Unusual, e.g. French 75% tax on highest player salaries (paid by the clubs) Abandoned in European leagues since the late 1970s Tax redistributed to amateur sport (e.g. French 5% tax on soccer TV revenues) Decreasing municipal subsidization of clubs (EU competition policy)

Team size Market structure Salaries and working conditions Salary cap Talent training and education Work stoppages, strikes, lock-outs Revenue redistribution

Public subsidies

No roster limits Monopolistic oligopsony for superstars Limited collective bargaining and low unionization rate English soccer (1901–1960); English and French Rugby Union (2000s) Educational center in top division clubs; nursery clubs in lower divisions Rare

* Multi-team cities in a same league are the exceptions (New York, LA and Chicago). ** Dividing a country’s population by the number of teams in a sport major league or top division.

(Goddard, 2014). They monopolize franchises to maximize their value and do not place teams in all financially viable locations. This has two consequences: (a) potential expansion franchises act as a buffer to inhibit entry by rival leagues; and (b) leaving the least profitable locations unexploited condemns rival leagues, during their inception years, to lose or make less money than the existing major league, which is often the starting point for their collapse or merger. Entry in an open league is for free while closed league expansion requires super-majority approval from established teams and entails paying a substantial entrance fee. Entry in a closed league is only possible by purchasing an expansion franchise when the new entering team’s assigned market location is assessed as profitable by the league Commissioner. Consequently, a closed league is smaller than the number of teams that would exist under competition or the absence of league-created entry barriers. Floating team shares on the stock market is forbidden in the NFL and restricted in other closed leagues so that no unwanted or non-voted team

owner can permeate the entry barriers while over 40 top European soccer clubs have been listed once since the 1990s. Noll (2003a) contends that closed leagues are too small because existing teams have a strong incentive to stop the league expanding long before the number of viable members is exhausted and because leagues actually face no threat of entry. A vertically open system basically alleviates the league’s cartelizing power by tolerating some competition across incumbent clubs and with potential entrants from lower/upper divisions through the promotion–relegation mechanism. Competition cannot spring up between rival open leagues – whose establishment is not allowed by international sports federations5 – while it is the clubs’ everyday life. An entrepreneur can buy a lower division club or form a new team to enter into the bottom division, hire high-quality players and coaches, and earn promotion to the highest division. No approval and no fee payment are required. The number of viable clubs is larger in an open system because clubs that cannot survive in a higher division would survive in a lower one

12

THE SAGE HANDBOOK OF SPORTS ECONOMICS

with its lower costs. Since an open league does not impose territorial restrictions, it can have multiple teams in the best markets and more potentially viable teams than closed leagues. This translates into a larger number of clubs per inhabitants in the former than the latter’s top tier (Sloane, 2006). In fixed-membership leagues, changing cities can be achieved through expansion and team relocation. Over 50 team relocations occurred in the four North American major team sports leagues since 1950. Individual team location decisions have external impacts on the value of other franchises, thus closed leagues impose conditions to restrain such decisions. If its local market ceases to be profitable, a team can move to another urban area only after the super-majority league agreement. A necessary condition for relocation to be feasible is that a major league must leave some attractive markets without a major-league team. It is very hard to point to any motive other than profit maximization for these relocations (Sandy et  al., 2004). In open leagues, there is no such geographical team mobility; mobility is only vertical across divisions. There is neither territorial exclusivity nor local monopoly of a team in a given sport: in several European big cities, more than one team play the top soccer division. Promotion–relegation undermines the value of territorial exclusivity and the open system means freedom to establish a team wherever one wishes. Teams obtain financial benefits from promotion and financial penalties from relegation, giving teams a greater incentive to improve team quality than is the case in a closed league. Promotion– relegation allows the location of teams to follow demand without the painful and controversial process of relocation, but when a team in a bigger market is demoted while a team in a smaller market is promoted, all teams except the promoted one are likely to suffer financially (Noll, 2002). The open system reduces clubs’ ability to extract rents and subsidies through the threat of relocation (Szymanski, 2004a). Contraction in the number of a division’s teams is not frequent in open systems although it may happen (EPL from 22 to 20 clubs, Bundesliga from 20 to 18 clubs). Contraction is even rarer in closed leagues. The MLS contracted from 12 to 10 teams due to two teams’ recurrent financial losses and unpromising markets at the moment. The MLB intended to eliminate two weak teams in 2002; postponed to 2007, this never happened. The threat of contraction stands as a potentially useful weapon for extracting subsidies from the cities (Noll, 2003b). Some convergence between the two sports systems were noticed, especially in regard to their model of finance (Andreff & Staudohar, 2000).

Fort (2000) claimed that the differences are more apparent than real: closed and open systems produce similar outcomes because in Europe the leagues are locked up tight. No entry can occur except for strong clubs that can afford financial support to move up the ladder. In both systems, the best teams and talents migrate to where they are most valued, whether it be through franchise expansion, relocation or promotion. In practice, teams with the largest markets tend to gravitate to the higher tiers in the promotion–relegation system (Goddard, 2014).

MODELING CLOSED AND OPEN LEAGUES A standard Walras equilibrium model is used to analyze the closed league system with profitmaximizing teams since Fort and Quirk (1995) and Vrooman (1995), while, with different underlying assumptions, it also analyzes the open league system with win-maximizing clubs since Késenne (1996). In the former, crucial assumptions are: (a) teams are wage-takers in a competitive labor market for talents; (b) each team i chooses the level of homogenous talent ti6 to maximize profits; (c) the supply of talent is fixed and talent is measured in units such that an additional unit of talent increases win percentage by one unit: ∂w i = 1. The simplest Walras model is: ∂t i

Max πi = Max (R i − C i )(1)



(

R i = R i mi ,t i with

∂R i ∂mi

or



)

>0,

∂R i ∂t i

∂R i ∂t i

0

∂2 Ri ∂t i2

(2) 0

C i = s .t i + c i0 (3)

(π: team’s profit; R: revenues; C: costs; m: market size; s: market-taken wage; ci0 : fixed cost). Assumption (c) allows for the substitution of win percent by the quantity of recruited talent in team i’s revenue function Ri. A fixed supply of ­talent is justified by league regulation that closes

Origins and Developments of Sports Systems

the labor market. Thus, team owners internalize the following externality: recruiting an additional unit of talent deprives another team from this unit; but this deteriorates the league competitive balance (CB). Team revenue Ri is a function of local product market size mi and win percent substituted by ti, given assumption (c). This function is ­concave in win percent: wins have a decreasing marginal effect on revenues. Equilibrium wage is provided by the market invisible hand or Walras auctioneer. Assuming that teams have different product market sizes, the marginal revenue products will be incapable of being equalized across clubs (Sloane, 2006). At the market-given wage, teams share the total supply of talent in proportion to the ratio of their market shares. League economic equilibrium is associated with disparities in team payrolls, win percents are uneven and the league is unbalanced, which justifies regulation. The Walras model was adapted to open leagues by Késenne (1996 & 2000) through two assumptions: (d) teams are win-maximizers; (e) therefore they recruit as much talent as possible (arms race) within their budget constraint (breakeven). Teams still are wage-takers:

Max ti (4)



Ri (mi , ti ) − s. ti − ci 0 = 0.(5)

Using the Lagrangian, first-order conditions are: hence

 ∂R  1 + λ i  i − s  = 0 (6)  ∂ti  RM i = s − 1 / λ i < s(7) Ri − s. ti = 0.(8)

Marginal revenue of talent shows up as lower than its marginal cost (7). For a given unit talent cost, the team demand for talent that maximizes wins is higher than if the team were profit-maximizing. The relevance of Walras model has been questioned for analyzing open league systems. Conditional to assumption (c), marginal revenue of a win is equal for all teams, but this assumption does not hold any more after the Bosman case with labor market globalization in which free entry of players makes the supply of talent variable in all national leagues. When in reality teams of all national leagues recruit in a global market, an additional talent hired by one team is no longer lost for another team of a same league (and the aforementioned internalization does

13

not hold any more), namely if players are transferred from foreign leagues and lower divisions. Szymanski and Késenne (2004) argue that, given the limited number of teams in a league, it is more appropriate to use a game-theoretic approach. Additionally, Szymanski (2004b) stresses that some teams are influential (wage-makers) in the labor market. The relevant model must be a Nash conjecture in a non-cooperative game when a team determines its choice of talent level without knowing the effect of recruitment choices made by opponent teams. With profit-maximizing teams and fixed talent supply, compared to Walras equilibrium, Nash equilibrium delivers a more equal CB, but it is inefficient since the marginal revenue of winning is higher in the big team than in the small team and league overall revenue could be increased by moving talents from the small to the big team. The small team imposes a stronger externality on the big team than the big team externality on the small team because the big team generates more significant additional revenue than the small team for any increase in win percent. CB is more unequal than the ratio between market shares. With win-maximizing clubs and fixed s­ upply, Nash equilibrium is reached with a marginal cost of a win larger than its marginal revenue, distinct from the profit-maximization model. The distribution of talent and CB is the same in Nash equilibrium as in Walras equilibrium: if a team wants to win as much as possible within the ­limits of its budget constraint, it will spend all its money on talent regardless of the recruitment strategy of other teams in the league (Késenne, 2014). Again, with Nash equilibrium CB is more un­ equal under win maximization than under profit maximization. Now introducing a flexible supply of talent with profit-maximizing teams in Nash conjecture, CB is demonstrated to be the same in Nash flexible and fixed supply models. Comparing this result with the fixed supply Walras equilibrium, one concludes that the distribution of talent is more equal in the fixed supply Nash equilibrium, although again the small team imposes a larger negative externality on the big team than the latter on the former. A team’s demand for talent is lower in the Nash equilibrium model than in Walras equilibrium and the market clearing salary level is lower if the constant supply of talent is not internalized. Overall, Nash equilibrium models with both flexible and fixed supply of talent confirm that the sport contest will be more unbalanced in a win-maximizing open league than in a profitmaximizing closed league and that the market clearing salary level will be higher in the former.

14

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Moreover, some results derived from Nash equilibrium models jeopardize the convincing power of the standard model. In particular, the invariance principle7 is not valid in a Nash conjecture (Késenne, 2005).

LABOR MARKET FOR TALENT: FROM REGULATION TO FREER MARKET A sport system’s labor market for talent basically is a by-product of its overall organization. Starting from a tightly regulated labor market, the two sport systems have evolved on two different pathways toward deregulation. A strictly regulated market in the closed league system happens to be opposed to a deregulated free market in open leagues. In the former, team owners usually claim that labor market regulation is a major tool for improving the CB, while it maintains their monopsony position, so that new regulations are added over time. The open league system eventually ended up with a deregulated (and globalized) free market for talent. In fact, a kind of deregulation has also occurred in the closed system since the 1970s when the reserve clause was lifted. In 1970 in the MLB, a forceful baseball players’ union negotiated a collective bargaining agreement that brought into existence an impartial arbitrator to help sort out contractual grievances. By 1973, final offer arbitration (FOA) for salary grievances was instituted for players who were no longer rookies. In 1976, veteran players obtained a free agent status that is enforced after a defined number of years playing in the MLB; sometimes free agency was immediately countervailed by a salary cap like in the NFL. The great bulk of literature about closed leagues – which cannot be covered extensively here – deals with regulations such as the reverseorder-of-­finish draft for rookies, the (hard or soft) salary cap, the luxury tax, collective bargaining between team owners and player unions, FOA on salary, strikes and lock-outs about payrolls and salary caps. Major questioned issues are: the effect of collective bargaining agreement on drafted player compensation; reserve clause modifications impacting free agents’ mobility and salaries, namely the shift of star players from poorer to wealthier teams; how the draft affects the CB and team standings; the level of FOA salary compared to the player’s previous salary or the veterans’ average; whether a salary cap improves the CB; the effectiveness of a luxury tax in restraining spending by high revenue teams and reducing player transfers; how the luxury tax affects the

CB, team profits and social welfare; the impact of free agency and FOA on eligible players’ effort, performance, salary, contract duration, player mobility and inter-team allocation of players, therefore on the CB; possible team’s acquisition of a biased set of players – a winner’s curse (Andreff, 2014); large-market teams hiring more players than small-market teams; the effect of labor strikes on consumer demand; and finally, testing the relevance of Rottenberg’s invariance principle in this context. In open leagues, the initial retain and transfer system has evolved toward regulating the length of player contracts and transfer fees, and international federations limiting the number of foreign players allowed to be fielded by a team in a same match. Labor market deregulation after the Bosman case ruled out all restrictions to soccer players’ free choice on the European labor market for talent in order to align professional sports on to the article 177 of the Treaty of Rome, which enshrines free labor mobility within the European Union. The Bosman ruling also phased out quotas of national players (six out of 11 in 1995) that a professional soccer team had to field at any game. However, the free labor market is not fullyfledged since some regulation has been kept as regard the span of time when the free global labor market is allowed to operate, that is during two transfer ‘windows’ (mercato) per season, new FIFA rules (2001) fixing the maximum length of a player’s contract at five years, a compensation for youth training to be paid to nursery clubs on players transferred younger than 24, and restrictions on international transfers of teenage players – often circumvented in practice (Andreff, 2010). Players actually become free agents at the age of 24. Player mobility was already higher in pre-Bosman open leagues than in closed leagues since the promotion–relegation system was encouraging clubs from lower divisions to sell talented young players to better clubs, but international mobility was tremendously boosted after 1995. Up-to-date literature on labor market in open leagues focuses on the effects of Bosman jurisprudence (Frick, 2007), that is: international migration of players; career duration; the proportion of end(out)-of-contract free transfers; transfer fees and contract lengths;8 the determination of before end-of-contract transfer fees; the rent sharing component in transfer fees; pay (wage)-performance and transfer fee-performance relationships; the impact of foreign players on domestic teams’ performance at club level and on the CB; differences between countries (domestic leagues) importing players and those exporting players; investment

Origins and Developments of Sports Systems

in training young talent; player effort; and social welfare. Some analytical focuses distinguish closed and open league systems. The first one opposes an arms race with player sales paid in cash in open leagues to (barter) trade9 of players in a context of milder competition between teams in closed leagues. In European open leagues, most player transfers are transactions in cash or monetary settlement, barter and lending a player to another team being an exception. Promotion–relegation and win maximization compel teams to an arms race in which each team attempts to recruit the best players in view of improving its relative strength compared with opponent teams; in turn, the latter are led to overbid. In closed leagues, player mobility is all the more limited in that trading for cash is restricted or forbidden (since 1960 in NFL and 1976 in MLB). Inter-team player transfers basically are barter so that team competition for hiring a same player is low (Szymanski, 2004a). Barter trades are more common than player sales. The likelihood that a player is sold rather than traded increases with age and as the percentage of remaining career productivity decreases in the MLB, although the relationship between player quality and the likelihood of sale is found for hitters, but not for pitchers (Marburger, 2009). Team owners are not used to sell first-line players for cash in the MLB. Due to profit maximization in closed leagues, investment in sporting talent is only undertaken if it increases revenues more than costs. Small market teams lack profit incentives to build up competitive teams that will maximize league revenues; this can justify that big market teams subsidize small market teams (Fort & Quirk, 1995). A second distinction is that labor discrimination is more focused on in closed leagues while market differentiation of players due to their skills, talent and seniority is crucial in both kinds of leagues. Employee and salary discrimination by race, ethnic groups, gender, language (French Canadians in the NHL, Spanish-speaking players in the MLS) as well as positional, hiring and fan discrimination have fuelled a bulk of literature – not covered here – about closed leagues. It is much less so for European open leagues – the few exceptions are Reilly and Witt (1995), Szymanski (2000), and Preston and Szymanski (2000). Finally, since the 1930s minor closed league playing fields served increasingly as training grounds for prospective major leaguers (MLB); previous arm’s-length transactions were replaced by formal affiliations in the form of purchasing minor league teams (vertical integration) or entering into agreements that gave them exclusive rights to the players on a minor league team’s

15

roster. In the 1980s, ownership of players and rights regarding training remained in the hands of MLB clubs while activities involving local sources of revenue shifted increasingly back to local ownership of minor league teams (Hanssen et al., 2016). With free agency, owners attempted to recoup minor league training costs through underpayment of players, thus extracting a surplus from apprentices – those who cost the least to train compared with star apprentices eligible for FOA (Krautmann et al., 2000). However, college sport remains the link between amateur and professional status for players, and a privileged nursery for training talents to be finally recruited by major (minor) leagues. Last not least, strikes and work stoppages are more frequent and tough in closed leagues with strong player unions struggling against the owners’ monopsony than in open leagues with a lower rate of unionization and a superstar monopoly countervailing the owners’ market power. Player unions have been largely successful10 in their bargaining with owners (Rosen and Sanderson, 2001). Drewes (2005) explains that the difference is deeply rooted in industrial organization of both types of leagues and labor markets, North American professional players having to fight for preventing further mobility restrictions while the industrial structure (alleviated cartel) and actual players’ mobility limit the monopsonistic power of owners in open leagues’ labor market.

CONCLUSION Is one of the two systems economically better or more efficient? It is vain to focus on such a question that can only call for a balanced response. Each system has obeyed its specific logic, its economic rationale and its own inner consistency for more than one century. Planting some seeds of one system into the other’s ground may result in surprising and disappointing effects. For instance, a rookie draft in the open system with a free labor market would be ineffective as teams’ initial public offerings would jeopardize a closed cartel. Nevertheless, in the course of their evolution the two systems exhibit some converging trends triggered by the increased commercialization and globalization of professional team sports, such as increasingly diversified sources of finance, some deregulation of the labor market for talent, an increasing need for top-tier teams to cooperate with lower-tier nursery clubs, and experimentation of various revenue-sharing schemes.

16

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Many avenues for further comparative research on the two systems are still wide open.

Notes   1  Eventually, the Premier League withdrew from the nationwide Football League in 1992, renamed English Premier League (EPL) though keeping the promotion–relegation system with the lower division 1.   2  Since no single team can produce the league’s product, i.e. a season-long series of contests.   3  This is more the case of a single entity league (Flynn & Gilbert, 2001) like the MLS, established as such by several investors in 1996 to get around antitrust: the league has an ownership interest in all teams, thereby controlling player contracts, salaries, sponsorships and broadcasting (Jewell, 2014).   4  Entry and exit are substantially more common for minor league sports in the US.   5  An open league, like in European soccer, is integrated in a hierarchical structure where the national association which supervises the league is itself dependent on an international sport federation. The positive side of the coin is that winning clubs participate in international competitions which are absent in the closed league system; supranational competitions function as a promotion (and incentive) for the best top division clubs.   6  Implicitly, a number of units of homogeneous talent are embodied in each player, probably more in superstars than journeymen players.   7  Following Rottenberg’s contention (Rottenberg, 1956), the invariance principle states that, if teams are profit-maximizers, restrictions on the player labor market do not change the distribution of talent among the teams as compared to a league with a free competitive labor market; this principle compares to the Coase theorem (1960).   8  And player salaries and wage determination, when data are available, such as for the Italian soccer top division, are rather rare so far.   9  In closed leagues, teams mostly trade players either for other players or for prior rights in future drafts. 10  It might not be that true as regard the 2004–2005 year-long lock out in the NHL, which ended up with players accepting a salary cap after revenue losses assessed to be $1bn – and a $2bn loss on the owners’ side (Kahane, 2006).

REFERENCES Andreff, W. (2010). Why tax international athlete migration? The ‘Coubertobin’ tax in a context of

financial crisis. In J. Maguire & M. Falcous (Eds.), Sport and Migration: Borders, Boundaries and Crossings. Abingdon: Routledge, pp. 31–45. Andreff, W. (2011). Some comparative economics of the organization of sports: competition and regulation in North American vs. European professional team sports leagues. European Journal of Comparative Economics, 8(1), 3–27. Andreff, W. (2014). The winner’s curse in sports economics. In O. Budzinski & A. Feddersen (Eds.), Contemporary Research in Sports Economics. Frankfurt am Main: Peter Lang, pp. 177–205. Andreff, W., & Staudohar, P. (2000). The evolving European model of professional sports finance. Journal of Sports Economics, 1(3), 257–76. Cairns, J.A. (1987). Evaluating changes in league structure: the reorganization of the Scottish Football League. Applied Economics, 19(2), 259–75. Coase, R. (1960). The problem of social cost. Journal of Law and Economics, 3, 1–44. Drewes, M. (2005). Locked out: why work stoppages in major league sports are frequent in North America but rare in Europe. European Sport Management Quarterly, 5(1), 63–76. Eckard, E.W. (2001). The origin of the reserve clause: owner collusion versus ‘public interest’. Journal of Sports Economics, 2(2), 113–30. Flynn, M.A., & Gilbert, R.J. (2001). The analysis of professional sports leagues as joint ventures. Economic Journal, 111(469), F27–F46. Fort, R. (2000). European and North American sports differences. Scottish Journal of Political Economy, 47(4), 431–55. Fort, R., & Quirk, J. (1995). Cross-subsidization, incentives and outcomes in professional team sports leagues. Journal of Economic Literature, 33(3), 1265–99. Frick, B. (2007). The football players’ labor market: empirical evidence from the Major European Leagues. Scottish Journal of Political Economy, 54(3), 422–46. Goddard, J. (2014). The promotion and relegation system. In J. Goddard & P. Sloane (Eds.), Handbook on the Economics of Professional Football. Cheltenham: Edward Elgar, pp. 23–40. Hanssen, F.A., Meehan, J.W. Jr, & Miceli, T.J. (2016). Explaining changes in organizational form: the case of professional baseball. Journal of Sports Economics, 17(6), 523–57. Hoehn, T., & Szymanski, S. (1999). The Americanization of European football. Economic Policy, 14(28), 203–40. Jewell, T. (2014). Major league soccer in the USA. In J. Goddard & P. Sloane (Eds.), Handbook on the Economics of Professional Football. Cheltenham: Edward Elgar, pp. 351–67. Jones, J.C.H. (1969). The economics of the National Hockey League. Canadian Journal of Economics, 2(1), 1–20.

Origins and Developments of Sports Systems Kahane, L. (2006). The economics of the National Hockey League: the 2004–05 lockout and the beginning of a new era. In P. Rodriguez, S. Késenne, & J. Garcia (Eds.), Sports Economics after Fifty Years: Essays in Honour of Simon Rottenberg. Oviedo: Ediciones de la Universidad de Oviedo, pp. 107–24. Késenne, S. (1996). League management in professional team sports with win maximizing clubs. European Journal of Sport Management, 2(2), 14–22. Késenne, S. (2000). Revenue sharing and competitive balance in professional team sports. Journal of Sports Economics, 1(1), 56–65. Késenne, S. (2005). Revenue sharing and competitive balance: does the invariance proposition hold? Journal of Sports Economics, 6(1), 98–106. Késenne, S. (2014). The Economic Theory of Professional Team Sports: An Analytical Treatment (2nd ed.). Cheltenham: Edward Elgar. Krautmann, A.C., Gustafson, E., & Hadley, L. (2000). Who pays for minor league training costs? Contemporary Economic Policy, 18(1), 37–47. Marburger, D.R. (2009). Why do player trades dominate sales? Journal of Sports Economics, 10(4), 335–50. Neale, W.C. (1964). The peculiar economics of professional sports: a contribution to the theory of the firm in sporting competition and in market competition. Quarterly Journal of Economics, 78(1), 1–14. Noll, R. (2002). The economics of promotion and relegation in sports leagues: the case of English football. Journal of Sports Economics, 3(2), 169–203. Noll, R. (2003a). The organization of sports leagues. Oxford Review of Economic Policy, 19(4), 530–51. Noll, R. (2003b). The economics of baseball contraction. Journal of Sports Economics, 4(4), 367–88. Noll, R. (2006). Sports economics after fifty years. In P. Rodriguez, S. Késenne, & J. Garcia (Eds.), Sports Economics after Fifty Years: Essays in Honour of Simon Rottenberg. Oviedo: Ediciones de la Universidad de Oviedo, pp. 17–49. Pomfret, R., & Wilson, J.K. (Eds.) (2016). Sports through the Lens of Economic History. Cheltenham: Edward Elgar.

17

Preston, I., & Szymanski, S. (2000). Racial discrimination in English football. Scottish Journal of Political Economy, 47(4), 342–63. Reilly, B., & Witt, R. (1995). English League transfer prices: is there a racial dimension? Applied Economics Letters, 2(7), 220–22. Rosen, S., & Sanderson, A. (2001). Labour markets in professional sports. Economic Journal, 111(469), F47–F68. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–58. Sandy, R., Sloane, P.J., & Rosentraub, M.S. (2004). The Economics of Sport: An International Perspective. Basingstoke: Palgrave Macmillan. Sloane, P.J. (2006). The European model of sport. In W. Andreff & S. Szymanski (Eds.), Handbook on the Economics of Sport. Cheltenham: Edward Elgar, pp. 299–303. Szymanski, S. (2000). A market test for discrimination in the English professional soccer leagues. Journal of Political Economy, 108(3), 590–603. Szymanski, S. (2003). The economic design of sporting contests. Journal of Economic Literature, 41(4), 1137–87. Szymanski, S. (2004a). Is there a European model of sports? In R. Fort & J. Fizel (Eds.), International Sports Economics Comparisons. Westport, CT: Praeger, pp. 19–37. Szymanski, S. (2004b). Professional team sport are only a game: the Walrasian fixed-supply conjecture model, contest-Nash equilibrium, and the invariance principle. Journal of Sports Economics, 5(2), 111–26. Szymanski, S., & Késenne, S. (2004). Competitive balance and gate revenue sharing in team sports. Journal of Industrial Economics, 51(4), 513–25. Szymanski, S., & Zimbalist, A. (2005). National pastime: How Americans play baseball and the rest of the world plays soccer. Washington, DC: Brookings Institution. Vamplew, W. (2006). The development of team sports before 1914. In W. Andreff & S. Szymanski (Eds.), Handbook on the Economics of Sport. Cheltenham: Edward Elgar, pp. 435–9. Vrooman, J. (1995). A general theory of professional sports leagues. Southern Economic Journal, 61(4), 971–90.

3 The Economic Value of Sport Themis Kokolakakis, Chris Gratton and Günther Grohall

INTRODUCTION OVERVIEW OF DEVELOPMENTS IN THE UK AND THE EU Since the mid-1980s, a systematic study of the sport economy began in the UK and in Europe focusing on sport and its impact on the economy. Such an impact was defined primarily in terms of employment, output and consumer spending. In the UK the economic importance of sport was studied first by the Henley Centre for Forecasting (1986 and 1992) and the Leisure Industries Research Centre (1997). Many other European countries (e.g. the Netherlands, Belgium, Finland, Denmark, France and Germany) carried out similar studies in the 1980s and 1990s. Jones (1989) reviewed all the first round of European economic impact studies and Andreff (1994) reviewed developments in the early 1990s. Since 2007, the driving force in the estimation of the economic value of sport studies has been the European Union, actively promoting the modelling of national sport economies through the development of Sport Satellite Accounts (SSA). Nine EU states have their own national SSAs: Austria, Belgium, Cyprus, Germany (Ahlert & Heident, 2015), Lithuania,1 the Netherlands (CBS, 2015),

Poland (Liberda, Tomaszewicz, Świeczewska, & Tręska, 2015), Portugal,2 and the United Kingdom (Department for Culture, Media and Sport, 2016). Other countries that have developed SSAs based on the same guidelines include Switzerland and Japan. The EU has also funded two PanEuropean SSAs (SpEA et  al., 2012; European Commission, 2018) combining all the European information into a single SSA.

CONTEXT OF THE ECONOMIC VALUE OF SPORT The relevance of research establishing the economic value of sport is evident in the context of UK policy, and particularly the publication of the Sport Satellite Account (SSA) (Sport Industry Research Centre, 2011), its subsequent update (Department for Culture, Media and Sport, 2015, 2016), the Satellite Account of Golf (Sport Industry Research Centre, 2016a) and The Economic Importance of Olympic and Paralympic Sport (Sport Industry Research Centre, 2017a). All of these find a rationale in the context of the Government’s latest strategy for sport, Sporting Future (HM Government, 2015), which cites

The Economic Value of Sport

economic development through sport as one of five high-level outcomes. What is meant by economic development in this context is ‘a more productive, sustainable and responsible sport sector’ (HM Government, 2015): A more productive sport sector will be one that maximises its available resources and assets (including facilities, skills and workforce) and contributes directly to economic development. By ensuring it can be more productive, the sector can better deliver everything else in this strategy. (HM Government, 2015, p. 80)

The adopted key performance indicator for measuring economic development through sport is employment in the sport sector (from the Sport Satellite Accounts) (HM Government, 2015). Similarly, in the Pan-European SSA (SpEA et al., 2012), employment and Gross Value Added (GVA) were the key indicators of economic value. Other important outcomes of economic value that are highlighted in various reports include consumer spending, government income and foreign trade. GVA is a measure of economic activity very close to GDP. It has been used in most of the economic assessments as the key indicator of the value of sport. Its exact definition in the National Accounts is the sum of salaries, wages, surplus (including profit), consumption of fixed capital, social contributions, and taxes less subsidies on production. Thus, GVA is used to pay work (wages, salaries and social contributions), entrepreneurship (surplus), fixed capital and public services (production-based taxes less subsidies).3 Defined in this way, GVA approximates the difference between total turnover and the intermediate cost of production. Finally, in 2007, the EU Working Group on Sport and Economics developed a definition of sport activity, referred to as the ‘Vilnius definition of sport’,4 that formed the basis of all current sport evaluations in Europe.

LITERATURE REVIEW: THE MAIN THEORETICAL CONTRIBUTIONS Two methodologies are discussed here in association with the calculation of the economic value of sport: the National Income Accounting (NIA) and the Sport Satellite Account (SSA) approach. Both of these methodologies have been developed in European studies. They evaluate the sport economy in terms of consumer spending, GVA and

19

employment. Examples of publications using such methodologies include Gratton and Kokolakakis (2013), SpEA et  al. (2012) and Sport Industry Research Centre (SIRC) (2010, 2011). In the US, a combination of approaches has been considered based on sport-related spending, income or GVA. Examples include Humphreys and Ruseski (2008) and Milano and Chelladurai (2011). An important distinction between different methodologies is the scope of sport under consideration. For example, even though they are estimating the US sports economy in the same year, Humphreys and Ruseski’s (2008) estimates are far smaller than those of Milano and Chelladurai (2011) because the former used sports-related expenditures made by households, while the latter took a wider approach and included data for sports-related consumption and investment by firms and the government. Similarly, the definitions used by the SSAs and NIA methodologies would increase the sport estimates even further as they additionally include categories such as sport-related education and public administration. All approaches aimed at evaluating the sports sector usually incorporate a multiplier approach to enhance the basic direct impact effects. The multiplier essentially refers to effects of an additional injection of spending into the economy and could comprise both indirect effects, from income flowing to sectors supplying the sports sector, and induced effects as a result of expenditures of the generated incomes. In tourism impact studies, tourism expenditure is considered to be an addition to the normal flow of expenditure in a local economy. However, in the case of sport, it is not always appropriate to treat all expenditure as additional at the national level. Moreover, arguments against using a multiplier approach for a national economic impact study are also discussed in the Henley Centre for Forecasting’s (1986, 1992) studies and mainly relate to adverse effects of additional spending on financial markets effectively reducing the multiplier to zero. However, the multiplier is appropriate for the estimation of the economic impact of sport events, because they have the potential to generate additional expenditure, income and employment to the economy. National studies in the UK and in Europe usually present indirect multipliers but ignore any induced effects, mainly due to a lack of tax-modelling and behavioural indicators affecting savings rates, interest rates and foreign spending. For example, the latest PanEuropean SSA only reports indirect multipliers (European Commission, 2018) and the original UK sport economic study by the Henley Centre of Forecasting (1986) only introduces the issues of multipliers in a conceptual framework.

20

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The National Income Accounting Framework The NIA methodology has been used in UK sport research since the 1980s, modelling the flows of the initial demand of sport consumers and companies throughout the economy. In its origin, NIA was a branch of development economics for which the late Professor Richard Stone received the Nobel award in economics. Sport studies that have used the NIA methodology include the estimation of economic importance of sport for England (Cambridge Econometrics, 2002; Sport Industry Research Centre, 2010; Sport England, 2013), Scotland (sportscotland, 2014), Wales (Sport Wales, 2012), Northern Ireland (Gratton & Kokolakakis, 2013), and Republic of Ireland (Irish Sports Council, 2010). To appreciate the NIA framework, one has to be familiar with the circular flow of income, illustrated in Figure 3.1. The left side of Figure 3.1 (connecting factor payments with consumption) shows the circular flow of income in an economy at its simplest. Households spend their income on goods and services produced by firms. Hence there is a flow of consumer expenditure from households to firms and a flow of goods and services (output) from firms to households. At the same time, households sell their labour services to firms in return for wages (factor payments).

Figure 3.1  The circular flow of income

In this simple circulation, the total amount of income is spent on goods and services, so that total expenditure equals both total income and total value of goods and services produced. The circulation can be expanded to take into account the government sector (both central and local), the financial sector and overseas trade. Consequently, any flow of expenditure from households that does not go into purchasing goods and services from firms (e.g. savings, taxes and imports) is a ‘withdrawal’ or ‘leakage’ from the circular flow of income. Any flow entering the circular flow (e.g. investment, government expenditure and exports) is an injection into the system. If injections are greater than withdrawals, then total expenditure increases, as does output and income. Conversely, if withdrawals are greater than injections, then total expenditure declines. The circular flow of income model incorporating the above-mentioned sectors, together with the voluntary sector, is the conceptual model of the economy that lies behind the NIA approach. The NIA framework indicates how expenditure in one particular sector of the economy flows as income to other sectors and hence generates GVA and employment. Other key outcomes of the model include sport-related consumer spending, government net income and international trade. The NIA framework, in all the UK sport studies that have followed this methodology (starting from Henley

The Economic Value of Sport

Centre for Forecasting, 1986), has identified seven broad sectors associated with the sport economy: • • • • • • •

Central government Local government Commercial sport Commercial non-sport Consumer sector Voluntary sector Overseas sector

The commercial non-sport sector is the only one of these that is not self-explanatory. It consists of commercial firms supplying sports organisations with non-sports goods and services which are required by the sports organisations, e.g. catering for sports outlets, advertising of non-sport products, transport for sports trips. NIA categorises each element of total sportrelated final expenditure to one of the seven sectors above and shows the flows from expenditure to income, indicating how expenditure in any one sector can contribute to income in several other sectors. In this way, seven income and expenditure accounts are created. To move from these sectoral accounts to the economic importance of the whole sports economy, it is necessary to estimate the GVA created in each sector by the sport-related final expenditure. This is done primarily on the basis of the sum of wages and surpluses (profits). In this way, GVA represents the value added on the product at the examined stage of production. For example, a retailer buys sports goods at wholesale prices, then provides a retail outlet, staff to advise customers, promotion for the product, etc., and sells the product at a higher price than the wholesale price paid. The difference in price is equivalent to the value added, and is reflected in the retailer’s wages and profits. Value added estimation for sport in the UK uses information from relevant business publications, Monitors and Input-Output tables to identify the shares of the total turnover that correspond to wages and profits. These shares are then used to turn income into GVA. Similarly, NIA does not directly generate estimates of employment. Average wages in each sector are calculated from a survey (such as the Annual Survey of Hours and Earnings), and then the total wages in each sector are divided by their average wages to give employment. The main advantage of NIA is that is easy to adapt within small economies or regions without Input-Output tables. The sectoral flows can provide new insights into the working of the sport economy. For example, the history of NIA studies

21

in the UK has shown that a consistent feature of the sport economy is the significant surpluses generated at the central government level. This is partly due to the VAT associated with sport spending. However, the NIA’s methodological framework, in terms of the definition of sport, cannot match the detail provided by the SSAs.

Sport Satellite Accounts and EU Initiatives In the UK, the NIA methodology was followed by the development of the SSA through the Department for Culture, Media and Sport (DCMS) (Sport Industry Research Centre, 2011). Satellite accounts are extensions of the standard National Account system. SSAs permit all sport-related economic activities to show up explicitly, rather than keeping them concealed, in deeply disaggregated (low-level) classifications of the National Accounts. Accordingly, elements of the National Accounts, such as ‘manufacture of plastic products’ or ‘textiles’, were investigated for sport content. Textiles, for example, include sport shoes and sport wearing apparel, which can be isolated in a SSA, often through dedicated datasets or the study of the more detailed foreign trade. In most cases, through primary or secondary research, a sport share is derived which is then applied to the National Accounts categories. The UK is in the first wave of European countries that developed such an Account according to a common agreed definition used throughout the EU (Vilnius definition). The EU has now developed systematically calculated SSAs, the origins of which date back more than 10 years to the first Commission-chaired EU Working Group of Sport and Economics in 2006. This group created the Vilnius definition of sport, a harmonised definition of sport within the System of National Accounts, using the Classification of Products by Activity (CPA) system. The Vilnius definition, in its philosophy, is consistent with the approach used in the earlier UK studies using the NIA framework. However, the latter had a more practical outlook, starting from what was available through surveys or national statistics, rather than investigating all sport-related codes in the CPA or NACE (nomenclature statistique des activités économiques dans la Communauté européenne) classifications. The Vilnius definition, as Figure 3.2 illustrates, is based on three layers: (i) the core definition, (ii) the narrow definition, and (iii) the broad definition. (i) Sport, under the core definition, is explicitly identified in the National Accounts. It includes

22

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Figure 3.2  Three layers in the Vilnius definition of sport Source: SpEA, SIRC

the operation of sport facilities (93.11), fitness facilities (93.13), sport clubs (93.12) as well as other sports activities (93.19) (e.g. sports and recreational event promotion services; services of athletes; support services related to sports and recreation). All the above elements of the core definition are grouped under the general code CPA 93.1 ‘sporting services’. (ii) Sport in the narrow definition includes the core definition and all goods and services which are necessary inputs for taking part in sport (i.e. to produce sport as an output). Examples include manufacturing retail and wholesale of sport equipment, construction of sport infrastructure and sport education. (iii) Sport in the broad definition incorporates the narrow definition and all products and services which relate to sport activities, but without being necessary for taking part in sport (i.e. they use sport as an input). Examples include sport tourism, betting, sport R&D and sport media. Finally, the methodology was extended to generate an SSA for all the EU countries (funded by the EU Commission), leading to the publication of Study on the Contribution of Sport to Economic Growth and Employment in the EU (SpEA et al., 2012; European Commission, 2018). This study has been one of the most important research outcomes at the European level in this field. In this, a Multiregional Input-Output Model (MRIO) was constructed, based on 28 national Input-Output

tables, placing sport within each state’s National Accounts on the one hand and intra-EU trade on the other. The theoretical basis for such a model was provided by Chenery and Moses (Moses, 1955). Accordingly, the MRIO used was set up in two steps: (i) intraregional tables (one table for each country, i.e. the national SSAs); and (ii) construction of import and export flows. These steps involved the modelling of international trade of all EU countries (using the UN datasets) and the creation of (proxy) single country SSAs in member states. Detailed methodological and data sources sections for the creation of full SSAs can be found in the technical annexes of European Commission (2016, 2018). The Vilnius definition has been instrumental in the development of sport economic studies in the EU. Despite the uniformity of the research framework, countries have some autonomy in adjusting it to their specific requirements. On the other hand, Russel, Barios and Andrews (2016) criticised the definition’s applicability: the definition calculates a given sector’s ‘sports-related production’ as the total production value of that sector (from national statistics) multiplied by the sports-related share of the sector. Members of the working group determined the ‘sports-related share’ on an industryby-industry and country-by-country basis; this process depends on the researchers and data available and cannot be harmonised as in the case of

23

The Economic Value of Sport

the definition. Nevertheless, use of multiple methods, exchange of information among countries and questioning of the results moderate the subjective component of the research process. Since the original ‘wave’ of four European SSAs more countries have produced their own SSAs based on the Vilnius definition, including Belgium, Germany, Lithuania, the Netherlands and Portugal. By 2018, nine out of 28 EU countries had comparable SSA tables, which is a major achievement both in terms of research and of policy.

OVERVIEW OF RESEARCH FINDINGS This section presents the key indicators in the EU studies of the economic value of sport, emphasising the ability of sport to generate GDP and employment.

Latest Results in the EU There have been two Pan-European Accounts estimating the economic value of sport in the EU (SpEA et al., 2012; European Commission, 2018). The former, using 2005 data, estimated economic value in terms of GVA, while the latter, using 2012 data, in terms of GDP. The two measurements are closely linked: GDP equals GVA plus taxes less subsidies on production. Table 3.1 illustrates the shares of sport activity within the EU economy in terms of GVA (GDP in 2012) and employment. In 2012, sport contributed 2.12% and 2.72% to European GDP and employment respectively, reaching €280bn in value and employing 5.7 million people. Further, when the indirect effects from the required inputs of sport production are included (by using indirect multipliers), sport contributed 3.6% and 4.5% to European GDP and employment respectively, generating 9.35m jobs. A not dissimilar percentage structure is also applicable in the case of the UK.

Table 3.2 shows the detailed shares of GDP and employment among the EU countries. It indicates that in 2012, the best-performing sport economy in the EU was Austria with shares of sport GDP and employment reaching 4.1% and 5.6% of the overall Austrian GDP and employment, respectively. However, in terms of absolute values of GDP and employment, the strongest sport economy is Germany, generating €104.7bn of GDP and 1.76m of employment. By comparison, in the same year the UK economy generated €36.7bn of sport GDP, and 1.06m of employment, corresponding to 2.2% and 3.8% of total UK GDP and employment, respectively. The strength of Austria’s sport economy is due to her winter sports, which is mainly practised in the mountainous areas of western and central Austria. This region is easily accessible from Germany, giving Austria a relative geographical advantage in the development of winter sport tourism. Austria’s foreign tourism at 55m overnight stays (2012) is five times the level of domestic tourism, creating a leverage effect. Other countries have other relative advantages increasing the value of sport GDP. In the case of Germany, for example, there is a strong element of manufacturing, while in the UK there is a very wide network of voluntary clubs and golf resorts. An overall trend that can be identified in the Pan-European research is that high values of sport shares in GVA and employment tend to be associated with highly developed economies and high levels of average national income among the population. As national income increases, the sport share of national GVA tends to rise as well. It therefore becomes important to examine the relationship between sport and economic growth in a more detailed framework.

Sectors that Drive National Sport Economies The Pan-European SSA has identified, across the EU, the sectors that drive the sport economy, providing the leading shares in terms of employment

Table 3.1  The economic value of sport in the EU, 2012 and 2005

2012, % 2005, % 2012, Value

EU sport GDP (direct)

EU sport GDP (direct and indirect)

EU sport employment (direct)

EU sport employment (direct and indirect)

2.12% 1.76% €280bn

3.59% 2.98% €470bn

2.72% 2.12% 5.67m

4.50% 3.51% 9.35m

Source: SpEA et al. (2012) and European Commission (2018)

24

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 3.2  Main sport-related indicators, direct effects Member State

Sport-related GDP, m Euro

Share of sportrelated GDP (%)

Sport-related employment in heads

Share of sport-related employment (%)

European Union AT – Austria BE – Belgium BG – Bulgaria CY – Cyprus CZ – Czech Republic DE – Germany DK – Denmark EE – Estonia EL – Greece ES – Spain FI – Finland FR – France HR – Croatia HU – Hungary IE – Ireland IT – Italy LT – Lithuania LU – Luxembourg LV – Latvia MT – Malta NL – The Netherlands PL – Poland PT – Portugal RO – Romania SE – Sweden SI – Slovenia SK – Slovakia

279,697 13,066 4,494 338 361 2,055 104,707 3,973 159 1,784 14,984 3,264 39,923 676 1,252 1,804 21,217 283 630 142 129 7,973 8,952 1,879 1,389 5,949 609 956

2.12 4.12 1.16 0.80 1.85 1.27 3.90 1.56 0.88 0.93 1.44 1.63 1.91 1.54 1.26 1.03 1.32 0.85 1.43 0.64 1.81 1.24 2.30 1.12 1.04 1.41 1.69 1.31

5,666,195 226,129 71,440 44,756 7,813 84,803 1,761,369 64,082 13,656 47,486 261,839 50,634 582,709 27,908 75,771 30,008 389,120 20,043 4,336 12,611 3,306 150,687 332,939 59,330 100,279 109,191 21,916 47,095

2.72 5.63 1.59 1.55 2.08 1.76 4.60 2.45 2.31 1.31 1.50 2.09 2.29 1.83 2.00 1.68 1.76 1.62 1.89 1.48 1.98 2.04 2.17 1.39 1.22 2.43 2.43 2.03

UK – United Kingdom

36,750

–

1,064,939

3.75

Source: European Commission (2018)

and GDP. Table 3.3 identifies the top 10 sectors in terms of their sport employment. Overall the strongest sector is ‘Education services’ with more than 1.1m jobs. This is followed by ‘Sport services’ with around 749,000 jobs and the retail services with more than 586,000 employed persons. Both sport services and retail are sectors which generate a lot of employment for a given value of GDP. These first three sectors provide employment for nearly 2.45m persons or 1.17% of EU’s total employment. Sport tourism is largely responsible for the size of the accommodation and food services sector, generating 586,000 jobs across the EU.

Accommodation is followed by administration, entertainment, construction, health services and textiles. ‘Construction and construction works’ includes investments into sports infrastructure such as stadiums, fitness clubs and swimming pools. The sport association with sector R90-92 mainly comes from ‘gambling and betting services’, although there are many sport-devoted museums and other cultural institutions. In the UK, sport betting generates approximately a third of the overall betting-related employment and GVA. Finally, ‘human health services’ are twofold, including treatment of sport-related injuries and sport as a treatment.

25

The Economic Value of Sport

Table 3.3  Employment contributions of the 10 top sport-related sectors Goods and services P R93_1 G47 I O G46 R90-92 F Q86 C13-15

Employment in heads

Education services 1,110,882 Sport services 749,291 Retail trade services, except of motor vehicles and motorcycles 586,516 Accommodation and food services 585,892 Public administration and defence services; compulsory social security 503,059 services Wholesale trade services, except of motor vehicles and motorcycles 345,683 Creative, arts, entertainment, library, archive, museum, other cultural 240,952 services; gambling and betting services Constructions and construction works 179,414 Human health services 143,666 Textiles, wearing apparel, leather and related products 124,104

Share of total employment (%) 0.53 0.36 0.28 0.28 0.24 0.17 0.12 0.09 0.07 0.06

Source: European Commission (2018)

IMPORTANCE OF SPORT IN ECONOMIC DEVELOPMENT Employment Intensive Industry The Pan-European Accounts, but also the accounts in most countries, show that the percentage of sport employment exceeds the percentage of sport GVA (or GDP) in an economy (SpEA et  al., 2012). Hence, sport is a relatively labour-intensive industry, with growth in the sport industry likely to lead to additional employment. In fact, as the European Commission (2018) indicates, across the EU an increase of GDP by 1% is associated with an additional 1.35% of employment. This is an important conclusion, as it underlines the role that sport can play in countering unemployment. The extensive sport club network, together with the golf resort economy, differentiates the UK from the remaining EU sport economies, generating high levels of ‘core’ sport employment. Figure 3.3 shows sport employment in the top five countries in the EU(28) according to the Eurostat data of sport employment (statistical definition, variable sprt-emp, sport employment5). The latter is based on the European Labour Force Survey. This is a slightly different approach from the core level of the Vilnius definition as it also includes sport workers from across the economy. According to the Eurostat definition of sport employment, the UK generates more sport employment than any other EU country. The UK has 431,000 sport-related jobs (according to the Eurostat definition), corresponding to 177% of Germany’s sport employment (the next country in the table). Further, using the above definition,

the UK generates more than a quarter of the EU’s sport employment. Note, however, that in terms of percentages of sport employment out of total employment, the UK (at 1.37%-statistical definition) ranks third in the EU, behind Iceland (2.03%) and Sweden (1.53%).

Relatively High Multiplier Effect As noted earlier, multipliers can be used to identify the wider scale of the sports economy. In essence, they can also show how much total output is necessary to produce in order to satisfy a single unit of direct demand. As so defined, they shift the research investigation towards growth potential and illustrate differences among the sport sectors examined. As an example, in the case of the UK sport services, with an output multiplier of 1.7, a total of £1.7m of output needs to be generated throughout the economy for £1m of output in sport services. Some sectors, such as construction, hotels and restaurants, media, and transport, have a very strong influence on the production of sport services. Table 3.4 shows the size of some indirect output multipliers. As noted above, a particular characteristic of the UK market is the strength of the club network where much of the employment is generated. This relates primarily to sporting services and to construction since, as Gratton and Kokolakakis (2013) have shown, clubs spend most surpluses, if any, on construction projects. Major sport events can increase the importance of construction within the sports economy, as was the case in the UK before the Olympic Summer

26

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Figure 3.3  Sport employment in the EU (top five scores) (thousands, 2016) Source: Eurostat

Table 3.4  UK indirect output multipliers of industries with ties to sport, 2014 Sporting services

Construction

Hotel and Restaurants Media

Rail transport

1.747

1.859

1.563

1.966

1.549

Source: Office for National Statistics, analytical Input-Output tables for the UK (2018)

Table 3.5  EU-wide output multipliers of sport-related goods and services CPA

Goods and services

Multiplier

C10-12 C29 C25 C33

Food, beverages and tobacco products Motor vehicles, trailers and semi-trailers Fabricated metal products, except machinery and equipment Repair and installation services of machinery and equipment

2.55 2.5 2.34 2.32

N79 C18 C30 F K65 C13-15

Travel agency, tour operator and other reservation services and related services Printing and recording services Other transport equipment Constructions and construction works Insurance, reinsurance and pension funding services, except compulsory social security Textiles, wearing apparel, leather and related products

2.28 2.18 2.18 2.16 2.16 2.15

Source: European Commission (2018)

Games in 2012. The size of the indirect multipliers indicates the strength of the sport sector in terms of growth both in a national and European context. Similarly, Table 3.5 shows the top 10 EU-wide output multipliers that relate to sport goods and services. It shows that the highest multiplier is in the category food beverages and tobacco products (2.55) and that most of the top categories relate to manufacturing. This reflects the fact that these sectors are very well interconnected within

the economy. For example, food beverages and tobacco products are strongly related to agriculture, which also has a high multiplier. On the other hand, education services, which typically require highly skilled personnel and very little intermediate output, has the smallest EU sport-related multiplier at 1.36. Note that it is not the sport content that conveys the small multiplier, but the nature of the activity itself (as is also the case with the larger multipliers).

The Economic Value of Sport

Overall, the more interconnected the economy is and the fewer the products that are imported, the higher the multiplier. Consequently, large economies benefit as they can produce more goods and services on their own. The EU, for example, would have necessarily higher multipliers than the ones reported in individual countries such as the UK and Germany (as they are strict subsets of the EU). Overall, in the EU, the average multiplier of sportrelated goods and services equals 1.95, a little higher than the average 1.93 of non-sport sectors.

Economic Importance of Volunteering All the estimates provided in this chapter on the economic value of sport underestimate the true economic importance of sport due to the underestimation of three areas of economic activity: the voluntary sector, sports events and effects on the wellbeing of sport participants. In this section we discuss the volunteering sector; sport events and wellbeing effects are discussed elsewhere in this Handbook. Research on the economic value of sport often emphasises the importance of sport volunteering as a factor of economic growth. Gratton and Kokolakakis (2013) showed that volunteering acts as an economic resource that can help company growth, supplementing and reinforcing public investment. Studies and policy reports have underlined the importance of sport volunteering as a factor of sport participation (Reid, 2012; HM Government, 2015), linking it with further economic and social effects. The UK government, for the first time, recognised volunteering as ‘a specific form of engagement in sport that we want to encourage for its own sake’ (HM Government, 2015, p. 37). According to the Sport Industry Research Centre (2010), the voluntary sector generates around 11% of sport employment (fulltime equivalent) compared to the sport total. So far both the NIA and SSA methodologies include the voluntary sport sector as long as it is part of the formal economy, i.e. through the income and expenditure accounts of voluntary sector clubs. They take no account of the unpaid labour services of volunteers, which is the essential resource element of the voluntary sector. A true estimation of the resources involved in sport would include these unpaid labour services. Studies6 have shown that if we included such services in the estimate of the economic importance of sport this would add up to 50% of the sport output. Volunteers add value to the sport economy through two routes: first, their own unpaid labour, and second,

27

through enabling the operation of business and organisations that could not survive otherwise in the market environment. Under volunteering, the household sector becomes part of the production sector. It is equivalent to the Becker’s (1965) household production concept, except here several households combine together in the voluntary sector and ‘produce’ a product from which they can all derive benefit. A sport club within the voluntary sector is an example of such an informal economic activity. Members contribute their time and effort without payment for the ‘production’ of sporting opportunities. Sport volunteering is not only an economic factor highly correlated with sport participation, but also an important parameter in the performance of the club network with economic and social consequences. A SIRC study for Sport England found that sport volunteers come mainly from the pool of sport participants, i.e. volunteering usually follows participation (Sport Industry Research Centre, 2016b). It is therefore impossible to comprehend the strength of the sport sector and the character of the sport industry without referring to the voluntary sector and club network. According to the UK’s Active Lives Survey, these are often interlinked. Moreover, sport volunteering has become an important mechanism for clubs to generate surpluses above costs and wages (Gratton & Kokolakakis, 2013). In effect, sport volunteering ensures the existence of clubs and organisations that would have found it hard to survive within a free market economy.

CONCLUSIONS, FURTHER RESEARCH AND POLICY OUTCOMES Consistently, UK and EU-wide research on the size of the sport economy has shown that its most important element is the grassroots sector related to sport participation and sport education. Further, the survival of clubs depends on both sport participation and volunteering, which then shifts the focus on the factors that affect sport participation and the importance of volunteering for the economy. As the value of the sport economy is likely to increase under conditions of increased sport participation (via greater demand for goods and services), future research should increasingly investigate the causal links leading to rises in the level of participation, via longitudinal analysis if possible. Further research in the study of the sport economy has been planned already in Europe. The EU policy initiative for creating SSAs in all

28

THE SAGE HANDBOOK OF SPORTS ECONOMICS

EU countries is expected to continue and drive research in this direction. For this reason, in 2018, the EU funded technical support and a series of workshops and projects around SSAs and their methodology. This methodology should be developed to incorporate two ‘informal’ elements of the sport economy: • examination and monetisation of the informal voluntary sector at a European level; • examination and monetisation of the wellbeing effects associated with sports participation. These include the benefits of mental wellbeing, personal development and social trust (Sport Industry Research Centre, 2017b), as well as improvement in health and fitness, and reduction of anti-social behaviour. The existing methodologies could produce further results, such as to identify: • geographical clusters in sport production and services: often there are locations with highly concentrating sporting activities, enjoying economies of scale; • employment profiles associated with sport activities (including part-time or full-time, average wages, skills required, etc.), and analysing the connections in production (as in Input-Output tables) that lead to high multipliers and growth potential. When examining the sport economy at a national level, both the methodologies of SSA and NIA provide useful insights but each with their own individual advantages: the former lends itself to better calculations of multiplier effects and creating a consistent basis for Pan-European economic studies, and the latter for examining smaller economies without suffering the disadvantage of ‘older’ data, which is often the case with the Input-Output tables that underpin SSAs. An advantage of SSAs, however, is the compatibility with the National Accounts. The SSA methodology imposes a discipline on the definition and investigation of the sport content, investigating sport-related content in every single economic activity. Previous methodologies had a relatively practical character, starting from what was available in the public domain. On the other hand, NIA examines the major sectors of the sport economy separately, emphasising the importance of volunteering or the government sector, a qualitative differentiation not existing in SSAs. For example, the government sector is involved in sport because

the sport market has aspects of market failure: that is, sport can generate social benefits over and above the private benefit to the individual participants. Most prominent of these are the health benefits generated by sport so government spending on sport may reduce government spending on health care. This and other market failures lead governments to spend money to encourage more sport participation. Also, governments spend money to enhance performance at the elite level since international sporting success for a country can be classed as a public good. This context of analysis is readily examined through the NIA framework. As a conclusion of the chapter, the key findings from the studies of the economic value of sport are summarised, thus: sport is an important economic sector in Europe, contributing around 2% of GVA to the economy; within the EU, as national income increases, the sport share of national GVA tends to rise; sport is a labour intensive industry, contributing efficiently to increases in employment (sport share of employment is greater than the equivalent sport share of GVA); typically within the sport sector the government sport-related income would exceed sport-related costs. A large part of those incomes come in the form of VAT from sport product sales. Sport is closely integrated with other economic sectors; as a result, it is associated with high values of economic ­multipliers, fostering economic growth. Finally, volunteering acts as an economic resource that can help company growth, supplementing and reinforcing public investment.

Notes 1  https://osp.stat.gov.lt/informaciniai-pranesimai? eventId=143461 2  www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_ cnacionais2010&contexto=cs&selTab=tab3&perfi l=220677460&INST=220617355 3  Sometimes GVA is defined as the marginal productivity of the primary factors of production, including land. 4  http://ec.europa.eu/eurostat/documents/6921402/ 0/Vilnius+Definition+Sport+CPA2008+offic ial+2013_09_19.pdf/30838d11-01ea-431f8112-50786e187c1c 5  http://appsso.eurostat.ec.europa.eu/nui/submit­ ViewTableAction.do 6  Hidden Diamonds: Uncovering the true value of sport volunteers (2015), available at: www. activelancashire.org.uk/research/hiddendiamonds-uncovering-the-true-value-of-sportsvolunteers-2015

The Economic Value of Sport

REFERENCES Ahlert, G., & an der Heiden, I. (2015). Die ökonomische Bedeutung des Sports in Deutschland, Ergebnisse des Sportsatellitenkontos 2010, erste Schätzungen für 2012, Themenreport 2015/01, GWS mbH, Osnabrück. Andreff, W. (1994). The Economic Importance of Sport in Europe: Financing and Economic Impact. Committee for the Development of Sport, Council of Europe, Strasbourg. Becker, G.S. (1965). A theory of the allocation of time. Economic Journal, 75(3). Cambridge Econometrics (2002). The Value of the Sports Economy in the Regions in 2000. A report prepared for Sport England. Sport England, London. CBS (2015). De Nederlandse sporteconomie De bijdrage van sport aan de Nederlandse economie in 2006–2012. Centraal Bureau voor de Statistiek, CBS, Den Haag. Department for Culture, Media and Sport (DCMS) (2015). 2011–2012 Sport Satellite Account for the UK. DCMS, London. Department for Culture, Media and Sport (DCMS) (2016). UK Sport Satellite Account, 2012, 2014 and 2015. Statistical Release, DCMS, London. European Commission (2010). A Strategy for Smart, Sustainable and Inclusive Growth. European Commission, Brussels. European Commission (2016). Study on National SSAs in the EU. A report prepared by SpEA, SIRC and HAN. European Commission, Brussels. European Commission (2018). Study on the Economic Impact of Sport through Sport Satellite Accounts. A report prepared by SpEA and SIRC. Available at: https://publications.europa.eu/en/ publication-detail/-/publication/865ef44c-5ca1-11e8ab41-01aa75ed71a1/language-en/format-pdf Gratton, C., & Kokolakakis, T. (2013). Assessing the economic impact of outdoor recreation in Northern Ireland. Research Report: Sport Northern Ireland. Available at: www.sportni.net/sportni/wp-content/ uploads/2013/03/Economic-Impact.pdf Henley Centre for Forecasting (1986). The Economic Impact and Importance of Sport in the UK. SC study 30, Sports Council, London. Henley Centre for Forecasting (1992). The Economic Impact and Importance of Sport I the UK Economy in 1990. Sports Council, London. HM Government (2015). Sporting Future: A New Strategy for an Active Nation. Cabinet Office, London. Humphreys, B.R., & Ruseski, J. (2008). The Size and Scope of the Sports Industry in the United States. IASE/NAASE Working Paper Series No. 08–11. http://

29

college.holycross.edu/RePEc/spe/HumphreysRuseski_ SportsIndustry.pdf Irish Sports Council (2010). Assessment of Economic Impact of Sport in Ireland. Prepared by Indecon and SIRC. Irish Sports Council, Dublin. Jones, H. (1989). The Economic Impact and Importance of Sport: A European Study. Council of Europe, Strasbourg. Leisure Industries Research Centre (1997). A Review of the Economic Impact of Sport. Report for the Sports Council, London. Liberda, B., Tomaszewicz, Ł., Świeczewska, I., & Tręska, J. (2015). Rachunek Satelitarny Sportu dla Polski za 2010 rok. Raport wykonany na zlecenie i ze środków budżetowych Ministerstwa Sportu i Turystyki, Warsaw. Milano, M., & Chelladurai, P. (2011). Gross domestic sport product: the size of the sport industry in the United States. Journal of Sport Management, 25, 24–35. Moses, L.N. (1955). The stability of interregional trading patterns and input-output analysis. The American Economic Review, XLV(5), December. Office for National Statistics, UK input-output analytical tables (2018). In https://www.ons.gov.uk/ economy/nationalaccounts/supplyandusetables/ datasets/ukinputoutputanalyticaltablesdetailed Reid, F. (2012). Increasing sports participation in Scotland: are voluntary sports clubs the answer? International Journal of Sport Policy and Politics, 4(2), 221–241. Russel, S., Barrios, D., & Andrews, M. (2016). Getting the ball rolling: basis for assessing the sport economy. CID Working Paper, No. 321, July 2016. Available at: https://sports.growthlab.cid.harvard.edu/files/icss/ files/cidwp_321_assessing_sports_economy.pdf SpEA, SIRC, Statistical Service of Republic of Cyprus, Meerwaarde Sport en Economie, FESI, Ministry of Sport and Tourism of the Republic of Poland (2012). Study on the Contribution of Sport to Economic Growth and Employment in the EU. Research Report. European Commission, Directorate-General Education and Culture, Brussels. Available at: http:// e c . e u ro p a . e u / s p o r t / l i b r a r y / s t u d i e s / s t u d y contribution-spors-economic-growth-final-rpt.pdf Sport England (2013). Economic Value of Sport in England, 2010. Sport England, London. Sport Industry Research Centre (SIRC) (2010). Economic Value of Sport in England 1985–2008. A Research Report. Sport England, London. Source: Available at: www.sportengland.org/media/ 102753/economic-value-of-sport-in-england-1-.pdf Sport Industry Research Centre (SIRC) (2011). 2004– 2006 Sport Satellite Account for the UK: A Research Report. DCMS and UK Sport, London. Available at: www.gov.uk/government/statistics/2004-06-sportsatellite-account-report

30

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Sport Industry Research Centre (SIRC) (2016a). A Satellite Account for Golf in the UK. Report prepared for The Royal and Ancient Golf club (R&A). Sport Industry Research Centre (SIRC) (2016b). Longitudinal Analysis of Sport Volunteering Using the Taking Part Survey. Report prepared for Sport England, London. Sport Industry Research Centre (SIRC) (2017a). The Economic Importance of Olympic and Paralympic Sport. Report prepared for UK Sport, London.

Sport Industry Research Centre (SIRC) (2017b). Active Lives Survey, Mental Wellbeing, Individual and Community Development Analysis. Report prepared for Sport England, London. sportscotland (2014). Economic Importance of Sport in Scotland 1998–2012. Report prepared by Sport Industry Research Centre, Sportscotland. Sport Wales (2012). Economic Importance of Sport in Wales, 2010. Report prepared by Sport Industry Research Centre, Sport Wales.

PART II

Amateur Sports Participation, Supply and Impact

4 Sports Participation Paul Downward and Cristina Muñiz

INTRODUCTION This chapter explores sports participation research which, from an economic perspective, can be understood to represent the demand to practise sport. First, we briefly outline the context of the research and some important terminological distinctions. Then the main theoretical contributions to the literature are reviewed. The next section outlines the empirical measurement of sports participation in the literature as well as the econometric approaches that have typically been used to estimate models of participation. We then provide an overview of the main empirical results that are identified in the literature before the final section concludes and explores some areas for future research.

CONTEXT AND CONCEPTS Figure 4.1 illustrates that out of a combination of individual motives and constraints, utility maximising sport participation can emerge. Two main types of outcomes follow. The first is the derived demand for inputs to participation activity. The expenditures from these are central to understanding

the scale of sports activity in the economy and economic impact activity. The other is non-market individual and externality impacts of sport on individual subjective well-being, health, human and social capital. The literature has primarily explored the latter impacts, which are the subject of other chapters in this Handbook. The reason for this development is that from the 1960s sport became a branch of social welfare policy (Downward et  al., 2009). In Europe, the ‘European Sport for All Charter’ instigated in 1975, and its replacement by the ‘European Sport Charter’ in 1992, which was subsequently revised in 2001, are examples of this. The European Sport Charter states that, ‘Sport’ means all forms of physical activity which, through casual or organised participation, aim at expressing or improving physical fitness and mental well-being, forming social relationships or obtaining results in competition at all levels.1

Although EU countries are not bound by regulation to it, they have helped to develop and reflect its sentiment in policy. An important feature of the Charter is that it does not prescribe a list of activities that can be described as sport. Consequently, as Gratton and Taylor (2000, p. 7) note, definitions of sport involve ‘the criterion of general acceptance

34

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Figure 4.1  Sport participation demand

that an activity is sporting, e.g. by the media and sports agencies’. It follows that policy makers collect data on activities that can vary across surveys in different countries and even within countries (see Downward et al., 2009).

THEORIES OF PARTICIPATION Core Theory The application of the theory in Figure 4.1 to the analysis of sport has developed out of the theory of labour supply, in which the demand for leisure is implicitly derived from choosing work to fund consumption activity. In the context of sports participation this ‘income–leisure’ trade-off model has been reinterpreted to focus on the demand for leisure time (Gratton & Taylor, 2000). Becker (1965), however, identified that this model fails to recognise the indirect costs of consumption that emanate from the foregone cost of time. The income–leisure model formally ties the cost of leisure only to the direct price of time, but not

indirectly to consumption through expenditure from income, which itself also requires time. Consequently, consumption requires that goods need to be produced by the combination of other goods and time before they can be consumed. Consumers are thus members of households that produce the goods that they consume, often by specialisation in production roles. The demand for leisure implied by the approaches, as well as the underlying choice problems faced by individuals are provided in Table 4.1. In this table ‘U’ refers to utility, ‘x’ to goods purchased on markets, ‘px’ is the market price of goods, ‘T’ is total time available, ‘Tl’ is the time allocated to leisure and ‘Tw’ is the time spent working, ‘Tc’ is the time spent at consumption, ‘w’ is the wage rate, ‘I’ is income, ‘V’ is any unearned income and ‘z’ refers to ‘composite commodities’ that are produced in order to be consumed. In the production functions ‘tc’ and ‘b’ are production coefficients associated with using time and market inputs respectively to produce ‘z’. ‘W’ refers to wealth. Table 4.1 presents the leisure demands derived from both the income–leisure trade-off model and Becker’s time allocation model. In the

35

Sports Participation

Table 4.1  Leisure demand in the ‘income–leisure’ model and Becker’s time allocation model Income–leisure trade-off model

Becker’s time allocation model

Maximise Subject to

U(x,Tl) Time: T = Tl + Tw Income: pxx = I = wTw + V Implies: pxx = I = w(T – Tl) + V

Demand

Tl (w, I)

U(z,(x,T)) Time: T = Tc + Tw Goods: pxx = I = wTw + V Production: Tc = tcz   x = bz  Implies: (pxb + wtc)z = wT + V = W z(pz(px,w), W)

income–leisure model utility depends on the consumption of goods and leisure time as distinct ‘goods’. This distinction becomes manifest in the demand function in which the distinct cost of leisure time and income emerge as arguments. Other influences on demand would reflect differences in preferences. In the time-allocation model, in contrast, utility derives from the consumption of produced commodities that both use income and time, as implied in the extended set of constraints. The production functions in Table 4.1 indicate the input of time and goods per unit produced of ‘z’. Overall, in this model the consumer faces a wealth constraint that comprises the time that is available to them, which can be valued at the wage rate as the opportunity cost of time, and unearned income. The demand for leisure is thus fundamentally dependent upon its shadow price (pz), deriving from both the goods and time required to produce it, as well as wealth. In this model, demand will also be influenced by socio-economic circumstances (proxy variables for preferences) since to some extent the technology in production functions capture the behaviour of households.2

Developments for Sport The theoretical advantages of the time allocation model mean that it has received some attention in the specific context of sport. In a simplified model, Downward and Riordan (2007) extend Becker’s own development of his earlier work to analyse the development of rational addiction (Stigler & Becker, 1977) and social interactions (Becker, 1974), in which individuals allocate resources in production for consumption in such a way as to enhance consumption skills or desirable social characteristics. This suggests that sport consumption is likely to be manifest in individuals that engage in multiple sports and who share similar characteristics. Humphreys and Ruseski (2011) extend Becker’s analysis in a much fuller way. The innovation in

Humphreys and Ruseski is that they explicitly account for sports participation at the intensive and extensive margins; that is, they formally distinguish between the related but separate decisions of participating in sport and the time spent participating in sport.

EMPIRICAL ANALYSIS Measurement of Demand The previous section clearly indicates that the nature of sports participation is multifaceted and time-dependent, which is reflected in surveys. On the one hand, participation can be understood as a discrete decision to participate in the activity. On the other hand, participation can be understood as a frequency and intensity of engagement. Survey responses to questions about such activity are necessarily contingent on a particular time scale. A practical issue that arises, then, is over what period the activity is measured. This can vary from survey to survey and even not be addressed. For example, it is common in UK surveys, such as the Taking Part Survey3 and Active People/Active Lives Survey,4 to ask of respondents if they participated in a range of sports in the last four weeks prior to the interview, and over the previous 12 months. For the questions over the last four weeks, supplementary questions are then asked about the number of days that participation took place as well as the number of minutes in which the activity normally takes place. Questions are then asked about whether the exercise raised breathing or produced sweating, which loosely enables the identification of vigorous and moderate intensities (Downward & Rasciute, 2015). In the US, the Behavioral Risk Factor Surveillance System asks similar questions covering engagement in leisure time physical activities over the last four weeks. The number of times per month and the duration in hours and minutes are also

36

THE SAGE HANDBOOK OF SPORTS ECONOMICS

asked. The Australian Bureau of Statistics collects data on active sports participation, defined to exclude coaching, officiating or spectating, during the 12 months prior to the interview. Likewise, in Canada, the Canadian Community Health Survey and National Population Health Survey asks questions about regular participation in any sports during the past 12 months. An important feature of these surveys is that they collect data on individual activities so that both aggregate and disaggregate analyses are facilitated. It follows that the frequency and intensity data can be recoded to develop discrete categories of behaviour. In contrast, there are other surveys that measure sports participation in a more aggregate way without a precise temporal investigation. For example, in the Eurobarometer Survey, the question is asked ‘How often do you exercise or play sport?’. The responses are: ‘5 times a week or more’, ‘3 to 4 times a week’, ‘1 to 2 times a week’, ‘1 to 3 times a month’, ‘Less often’, ‘Never’. In the British Household Panel Survey, until it was integrated into the new Understanding Society Survey in 2010, and in the German Socio-Economic Panel, similar questions are asked. These variables are clearly ordered in nature, with the possibility of being recoded into discrete categories. Such questions, as well as those requiring deliberation of responses over the last 12 months, naturally carry with them the possibility of measurement error in that respondents may not recall behaviour correctly over longer periods of time. In this respect, it is not surprising that detailed questions about specific activities and their frequency and intensity are asked in connection with shorter time periods. However, it is also possible that a different form of measurement error could then arise. The last four weeks (or whatever short period) may be atypical, in which there was an inability to engage in the activity because of the seasonal nature of certain sports, vacation, temporary illness or other restriction. This is less likely to be the case with data collected over 12-month periods. For this reason, the Active People Survey and Taking Part Survey were designed to be of a rolling monthly nature such that representative sub-samples were developed over the year to help to ameliorate this potential problem (see Downward et al., 2016). Regardless of such sampling, however, the problem remains that respondents replying ‘no’ to a question about participation over a distinct period of time does not discriminate between a choice not to take part over that period of time by someone who normally takes part, and someone who has never taken part in the activity. To an extent this is not relevant to questions which have ‘never’ as a lower bound. However, the issues of recall noted above suggests that this issue is not

completely eradicated. Moreover, a no response generally, as noted above, could be a choice or the outcome of a restriction on behaviour. It is because of these issues that a variety of econometric modelling strategies have been adopted in the literature beyond Ordinary Least Squares (OLS).

Econometric Analysis Table 4.2 illustrates the different estimators typically employed in the literature for different measures of sports participation. In the table, ‘d’ refers to a dummy variable scored ‘1’ if respondents participate in sport or ‘0’ if not, ‘pr’ refers to probability, ‘y’ refers to an ordered or continuous measure of the frequency or intensity of participation in sport, ‘α’ and ‘β’ are coefficients to be estimated, and ‘W’ and ‘X’ are vectors of independent vari­ ables. The superscript * refers to an unobserved or latent variable, ‘e’ and ‘ν’ are random error terms,5 ‘ϕ’ is a normal density function, ‘Φ’ is the cumulative normal density function, ‘λ’ is the mean number of counts and ‘σ’ is the standard error. The first three of the estimators essentially deal with the measurement of sports participation as a discrete value, measuring either participates or not, participates in various ordinal categories, or is measured as a series of counts. Examples of the use of Logit/Probit estimators include Downward (2007), Hovemann and Wicker (2009), Van Tuyckom and Scheerder (2010), Lera-López et al. (2016), Borodulin et al. (2016), Vandermeerschen et al. (2016) and Marques et al. (2016). The ZeroInflated Ordered Probit (ZIOP) model is employed in Downward et  al. (2011) and Downward and Rasciute (2015), while Muñiz et al. (2011, 2014) and Van Cauwenberg et  al. (2017) make use of Zero-Inflated Negative Binomial (ZINB) and other count models. The last two models are extensions of ‘standard’ ordered (Probit) and negative binomial models for ranked and count data respectively but, as discussed above, allow for the fact that observed zeros in participation data could be due to never participating in sport, or not participating in a period investigated by a survey.6 Both of these models nest estimation of a model accounting for participation as a discrete choice and one that varies in magnitude. The final three estimators also model the participation decision and the intensity or duration of participation but treat the latter as a continuous outcome. The Tobit model treats the zero values as a censored point in the underlying distribution of behaviour. The fact that participation is logically bounded below by zero means that there is probability mass at this point that a standard OLS regression – drawing upon a normal distribution

Sports Participation

for its inferences – does not account for. The usual regression coefficients in the Tobit model consequently do not always make sense as they refer to latent outcomes, so additional marginal effects can be estimated to explore the impact of changes in independent variables on participation for values above zero, and for participation including a zero value. The probability of participation being nonzero is also estimated. Tobit models have been estimated in the literature by Ruseski et al. (2011), Dawson and Downward (2013) and Thibaut et al. (2017). One problem with the Tobit estimator is that it essentially constrains the modelling of participation in sport and the frequency or intensity of participation to involve the same set of independent variables. The Heckman and Double Hurdle estimators overcome this problem, but in different ways. In both cases the decision to participate or not in sport is estimated jointly with the frequency or intensity of participation, such that the variables that account for these behaviours can be different. However, the Heckman model focuses on estimating the frequency or intensity of participation on the sub-sample for which values of participation are greater than zero, while the Double Hurdle model allows frequency or intensity to be equal to zero.7 In this way, the Heckman model treats inference about participation frequency to be potentially influenced by a selection bias, which is controlled for by the use of the inverse Mills ratio in the frequency or intensity equation. In the Double Hurdle model, in contrast, zero frequency or intensity of participation is treated as a genuine choice that is distinct from a decision not to participate in sport per se. Downward and Riordan (2007) and Humphreys and Ruseski (2007) make use of the Heckman model while Humphreys and Ruseski (2011, 2015), Muñiz et  al. (2014), Caparrós Ruiz (2017) and Cheah et  al. (2017) make use of the Double Hurdle estimator.8 It is clear that choice between the estimators depends upon both the measurement of the dependent variable and theoretical reflection on the process by which behaviours emerge. As Table 4.2 makes clear, comparing the insights from across the estimators should proceed with caution as the marginal effects cannot be inferred directly from the estimated parameters of non-linear models. The marginal effects are non-linear functions of parameters and explanatory variables.

MAIN EMPIRICAL RESULTS In this section some of the main empirical results from the literature are presented. Space precludes

37

an exhaustive summary of contributions to understanding sports participation, so the focus is on giving indicative examples only. It is also important to recognise that because of the nature of the definition of sport, many studies that are relevant to understanding sports participation, including some of those cited, are drawn from the sport management, physical activity and health, sports science and sport sociology literature.

Socio-demographic Characteristics: Age, Gender, Education and Lifestyle The structuring of sports participation by sociodemographic characteristics can, according to extensions of the time allocation model, be understood as reflecting investment in consumption skills and social characteristics. For example, the basic model of time allocation was later expanded to demonstrate that an individual can invest in accumulating social characteristics (Becker, 1974). On the other hand, human capital acquisition could reflect the enhancement of skills required to consume sport and thus increases the potential for sports participation (Becker, 1964). This suggests lifestyles of behaviour and clustering of behaviour associated with socio-demographic characteristics is evident. Examination of the main findings in the literature reveals that the probability of sports participation decreases with age (Ruseski & Maresova, 2014). Changes in the types of activity also occur with age (Barber & Havitz, 2001) and decline in participation seems to affect males more than females (Bauman et al., 2009). García et al. (2011) report that for Spain sports participation follows a U-shaped curve with two peaks: youth and retirement. The empirical evidence has also found a positive relationship between age and the frequency of participation (García et al., 2011) with some evidence of a decline with age (Eberth & Smith, 2010). There is some consensus that males participate in sport more than females (Downward & Rasciute, 2015), with males also having a higher frequency of participation (Eberth & Smith, 2010). However, as García et al. (2016), note the results can depend on the activity. Differences of incidence and intensity of participation are also less marked among older adults (Bauman et  al., 2009). A positive relationship between education and sports participation has also generally been reported (Breuer & Wicker, 2008), but the results vary for the frequency of participation, with negative (Muñiz et al., 2014) and positive results identified (Humphreys & Ruseski, 2015). There is some evidence that smoking reduces sports

Models participation decision Combined with ‘j’ ordered categories of frequency (Ordered Probit) Two types of zero are accounted for

ZIOP

Tobit

and so on

y = 2 if µ1 < y* ≤ µ 2 and d = 1

y = 1 if 0 < y* ≤ µ1 and d = 1

y* = Xβ + υ, y = 0 if y* ≤ 0 or d = 0

d = 0 if d* ≤ 0

d* = Wα + ε, d = 1 if d*> 0

d = 0 if d* ≤ 0

d* = Wα + ε, d = 1 if d*> 0

Model

Models frequency of participation accounting for a concentration of zeros (indicating a truncation point ‘t’ in the distribution)

)

(

)

= gamma distribution

y* = Xβ + ε y = 0 if y* ≤ 0 y = y* if y* > 0

(

(‘μj’ are the boundary parameters ) Models participation decision d* = Wα + ε, d = 1 if d*> 0 Combined with counts of frequency modelled d = 0 if d* ≤ 0 as a Negative Binomial Two types of zero are accounted for E y X, υ = exp Xβ + υ = hλ ,h

Models participation decision

Logit/Probit

ZINB

Dependent variables

Estimator

Table 4.2  Econometric modelling approaches

) ( )

j−1

( ) ( ( )) ( ) ( ( )) ( )

if y ≥ 1

if y = 0

) = β + β ∂λ

()

imr c

∂X

∂c

with c =

( ) (( ) ) φ c Xβ and λimr c = σ Φ c

() ( the inverse mills ratio) − observed sub-population of y > t

∂E y X,y > t

(

∂X

∂E y* X

( ) = β Influence on unobserved latent participation

With f1 the density of the binary process and f2 the density of the count process Marginal effect: as above Marginal effects:

()

  f1 0 + 1− f1 0 f2 0 g y =  1− f1 0 f2 y 

Marginal effect: complex; depends on parameters, density functions, correlation between errors

j

pr y* = 0 X,W = 1− φ αW  + φ αW φ −βX  

( ) ( ) ( )( ) pr ( y* = j X,W ) = φ ( αW )  φ ( µ − βX ) − φ ( µ − βX )   

( )

Marginal effect: ∂d = αφ Wα ∂W

pr d = 1 W = φ Wα

(

Probabilities and marginal effects

38 THE SAGE HANDBOOK OF SPORTS ECONOMICS

Models participation decision Combined with the frequency or intensity of participation for the sub-populations that participates and accounts for that sample selection

Models participation decision Combined with the frequency or intensity of participation allowing for a choice of zero frequency

Heckman

Double Hurdle

if d * > 0

y * = Xβ + υ, y = y* if d* > 0 and y* > 0 but otherwise y = 0

d = 0 if d* ≤ 0

d* = Wα + ε, d = 1 if d*> 0

y = Xβ + υ

d = 0 if d* ≤ 0

d* = Wα + ε , d = 1 if d* >0

   σ σ

) = φ  Xβ  β Is the probability of being uncensored

   σ 

)

Independent (Cragg model) and dependent errors can be accounted for Marginal effects y ≥ 0 ∂y * =β ∂X

∂y =β ∂X

Where ρ is the correlation between ε and ν σν is the standard deviation of ν Marginal effects on frequency y > 0:

(

  Wα    φ    σ ω   E y d * > 0 = Xβ+ρσ ν     Wα   Φ       σ ω  

Accounts for sample selection through inverse Mills ratio

∂X

∂Pr y > t X

(

−observed y including t

   σ 

( ) = βΦ  Xβ  With Φ  Xβ  indicating the p (y > t)

∂X

∂E y X

Sports Participation 39

40

THE SAGE HANDBOOK OF SPORTS ECONOMICS

participation (Lechner, 2009) and that drinking alcohol is positively associated with sports participation (Downward & Riordan, 2007).

Income and Time The economic theory presented above naturally stresses the role of income and time on sports participation. The literature typically suggests that higher household or individual income raises participation in sport (Downward & Rasciute, 2010). However, it is also found that income has either no influence on the frequency of sports participation (Gratton & Taylor, 2000) or the influence is negative (Humphreys & Ruseski, 2011, 2015), indicative of substitution effects. Muñiz et  al. (2011) consequently report that sports activity is concentrated on weekends, when individuals probably have lower opportunity costs of participation. In as much that employment status proxies income and time effects, García et  al. (2016) find that working generally decreases or does not affect the probability of practising sports, while it reduces the frequency of participation. Less sports participation has also been observed for lower socio-economic groups and non-skilled workers (Lera-López & Rapún-Gárate, 2007). However, Humphreys and Ruseski (2015) indicate that although white-collar workers are less likely to participate in sports, they participate in sport longer per week than others. Potential leisure time constraints have also been explicitly examined in previous research. Kokolakakis et  al. (2017), Downward and Rasciute (2016), Dawson and Downward (2011, 2013) and Muñiz et  al. (2011) suggest a complementary relationship between attendance at cultural events, other leisure activities, watching sport on the media, and volunteering in sport and cultural activities, respectively. Moreover, Adachi and Willoughby (2016) and Kuo and Tang (2014) find that sports participation and social media use are complementary. However, the possibility of substitution between various sport and various leisure activities has been addressed in Løyland and Ringstad (2009) and Downward and Rasciute (2010). It is also identified that the growth in nonclub organised sport, allowing greater flexibility of schedules, is occurring because of time pressures (Borgers et al., 2016).

Household Structure The time allocation model explicitly identifies the household as the site of production and consumption. This has been explored in the literature in

relation to family structure. The literature shows that married people participate less in sport and dedicate less time to it, although there are significant differences by type of activity and the frequency of activity (Humphreys & Ruseski, 2007, 2015; Eberth & Smith, 2010). The negative effect of marriage on participation in sports is greater for females (Muñiz et al., 2014). The size of a household has been shown to be negatively associated with sports participation (García et al., 2016), although varies according to the sport (Downward, 2004). Ruseski et al. (2011) find that the time spent caring for children and relatives reduces both the likelihood that individuals participate and the time spent taking part in sports. However, the presence of children in the household has mixed effects, depending on the type of decision under consideration (i.e. participation or time spent) about sports (Humphreys & Ruseski, 2015). Children may limit the time available for adult sporting activities while increasing participation in child-oriented sports such as football or swimming (Downward, 2004; Humphreys & Ruseski, 2015). García et al. (2011) report that in general children decrease the likelihood of being involved in sport and physical activities but when people with young children decide to participate, they allocate more time to sport than others. Households are also a natural site in which consumption skills and social characteristics are invested. It is reasonable to assume that children’s behaviour is largely influenced by parents’ attitudes, with evidence of significant intergenerational transfers in relation to physical activity (Downward et  al., 2014). Consequently, children and adolescents who perceive parents to be active report higher sports participation rates (Dollman & Lewis, 2010). The same is the case for children whose parents were active growing up (Downward et  al., 2014; Vandermeerschen et  al., 2016). Finally, Downward and Rasciute (2016) show that the household peers’ physical activity on individual physical activity is important, especially in the case of women. However, more generally, Stratton et al. (2005) found that weekly contact with family and friends lowers individual participation in sports, which is indicative of time substitution.

Environmental Influences Outside the time allocation framework, it is clear that broader environmental influences may shape sports participation. Many studies investigate how macro-level variables such as the availability of sport facilities and/or sports programmes influence behaviour (Hallmann et  al., 2012; Wicker

Sports Participation

et al., 2012, 2013; Downward & Rasciute, 2015; Van Cauwenberg et  al., 2017). Wicker et  al. (2009) show that a poor supply of sports facilities reduces the regularity of sports activities. Deelen et al. (2016) provide similar evidence based on the distance to indoor sport facilities and neighbourhood desirability. Eime et  al. (2015) also show that the results can vary for different types of sport. Direct policy impacts on participation have also been explored in this context. For example, US state-level spending on parks and recreation have been shown to increase participation (Humphreys & Ruseski, 2007) as well as European spending on health and education (Lera-López et  al., 2016), but this is not the case for lotteryfunded capital expenditure in the UK (Kokolakakis et al., 2017). Indirect effects from ‘trickle down’ or ‘demonstration effects’ from hosting or being successful at sports events show mixed results (Weed, 2009). Mutter and Pawlowski (2014a, 2014b) suggest that the past success of professional national soccer teams has little impact on sports participation but conclude that professional triathlons induce an increased frequency of participation among amateur triathletes. Frick and Wicker (2016) show that only winning the World Cup has a positive effect on observed football participation. However, indicators of government quality and economic freedom have been shown to have positive effects on participation (Ruseski & Maresova, 2014; Wicker & Downward, 2017). Finally, and related to the seasonality of sport and difficulty of measuring sport participation in a time-specific context, Muñiz et al. (2011) and García et al. (2016) suggest that sports participation increased during the spring and summer seasons.

discussed above. This could be targeted to help to refine understanding of the time allocation of individuals across a range of leisure activities, but also in the context of the household and wider social engagement. Both time and expenditure dimensions should be explored. An important aspect of this research would also be to focus on the supply of sporting opportunities, which are still not that well researched. This would require linking individual data to the spatial features of areas of residence but also in connection to the workplace and typical travel zones. There has also been very little research that has a direct managerial focus. Exploring how the economic analysis noted above could help different sports organisations to understand the structure of, and behaviour in, their markets is a completely open field of enquiry,9 with some work currently drawing upon sociological theory only.

Notes 1 

2 

3  4  5 

6 

CONCLUSIONS AND FUTURE RESEARCH The aim of this chapter has been to provide an overview of the context of sports participation research, to review how economic theory has been used to understand sports participation, to indicate how sports participation is measured and the challenges that it raises for analysis. Typical econometric estimators have been reviewed and some key results from the empirical evidence presented in the context of the theoretical premises discussed. It is clear, however, that a number of avenues of research could provide fruitful lines of further enquiry. The first of these is that richer data could be collected that are not readily available in the secondary sources that are typically analysed and

41

7 

8 

9 

See https://search.coe.int/cm/Pages/result_details. aspx?ObjectID=09000016804c9dbb (last retrieved 3.4.17). In these models fixed unearned income will not affect optimal choices or sets of demands but would augment the level of any given demand. www.gov.uk/guidance/taking-part-survey (last retrieved 7.4.17). www.sportengland.org/research/active-livessurvey/ (last retrieved 7.4.17). For example, in the logit model it follows the logistic distribution, but the normal distribution in the Heckman model. For example, Mutter and Pawlowski (2014a) make use of an ordered logit estimator. There is also a Zero-Inflated Poisson (ZIP) estimator in count data models, but negative binomial models are often preferred because of overdispersion in the data. Though not strictly statistically necessary, identification in the Heckman model is often argued to depend on the presence of a variable in the participation equation that is excluded from the frequency or intensity equation. Other estimators have been used too. Eberth and Smith (2010) employ a copula estimator, Eime et al. (2015) simply employ multiple classification methods, Conway and Niles (2017) use a generalised two-part model, Downward et  al. (2014) apply a matching estimator and Kokolakakis et al. (2017) use the Dirichlet model. Some work currently has a marketing flavour but draws primarily on sociological theory (Taks & Scheerder, 2006).

42

THE SAGE HANDBOOK OF SPORTS ECONOMICS

REFERENCES Adachi, P.J., & Willoughby, T. (2016). Does playing sports video games predict increased involvement in real-life sports over several years among older adolescents and emerging adults? Journal of Youth and Adolescence, 45(2), 391–401. Barber, N., & Havitz, M.E. (2001). Canadian participation rates in ten sport and fitness activities. Journal of Sport Management, 15, 51–76. Bauman, A., Bull, F., Chey, T., Craig, C., Ainsworth, B. et  al. (2009). The international prevalence study on physical activity: results from 20 countries. International Journal of Behavioral Nutrition and Physical Activity, 6(21), 1–11. Becker, G.S. (1964). Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education. New York: Columbia University Press. Becker, G.S. (1965). A theory of the allocation of time. The Economic Journal, 75, 493–517. Becker, G.S. (1974). A theory of social interactions. Journal of Political Economy, 82, 1063–1091. Borgers, J., Breedveld, K., Tiessen-Raaphorst, A., Thibaut, E., Vandermeerschen, H., Vos, S., & Scheerder, J. (2016). A study on the frequency of participation and time spent on sport in different organisational settings. European Sport Management Quarterly, 16(5), 635–654. Borodulin, K., Sipilä, N., Rahkonen, O., Leino-Arjas, P., Kestilä, L., Jousilahti, P., & Prättälä, R. (2016). Socio-demographic and behavioral variation in barriers to leisure-time physical activity. Scandinavian Journal of Social Medicine, 44(1), 62–69. Breuer, C., & Wicker, P. (2008). Demographic and economic factors influencing inclusion in the German sport system: a microanalysis of the years 1985 to 2005. European Journal for Sport and Society, 5(1), 33–42. Caparrós Ruiz, A. (2017). Adolescents’ time use in Spain: does the parental human capital matter? Child Indicators Research, 10(1), 81–99. Cheah, Y.K., Azahadi, M., Phang, S.N., & Hazilah, N. (2017). Factors affecting participation decision and amount of physical activity among urban dwellers in Malaysia. Public Health, 146, 84–91. Conway, K.S., & Niles, D.P. (2017). Cigarette Taxes, Smoking – and Exercise? Health Economics, 26(8), 1019–1036. Dawson, P., & Downward, P. (2011). Participation, spectatorship and media coverage in sport: some initial insights. In W. Andreff (Ed.), Contemporary Issues in Sports Economics, Participation and Professional Sports (p. 15–42). Cheltenham: Edward Elgar. Dawson, P., & Downward, P. (2013). The relationship between participation in sport and sport volunteering: an economic analysis. International Journal of Sport Finance, 8(1), 75–92.

Deelen, I., Ettema, D., & Dijst, M. (2016). Too busy or too far away? The importance of subjective constraints and spatial factors for sports frequency. Managing Sport and Leisure, 21(4), 239–264. Dollman, J., & Lewis, N. (2010). The impact of socioeconomic position on sport participation among South Australian youth. Journal of Science and Medicine in Sport, 13, 318–322. Downward, P. (2004). On leisure demand: a post Keynesian critique of neoclassical theory. Journal of Post Keynesian Economics, 26, 371–394. Downward, P. (2007). Exploring the economic choice to participate in sport: results from the 2002 General Household Survey. International Review of Applied Economics, 21, 633–653. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports Economics: Theory, Evidence and Policy. Oxford: Elsevier. Downward, P., Dawson, P., & Mills, T.C. (2016). Sports participation as an investment in (subjective) health: a time series analysis of the life course. Journal of Public Health, 38(4), e504–e510, online first, 6 November 2015. https://doi.org/fdv164 Downward, P., Hallmann, K., & Pawlowski, T. (2014). Assessing parental impact on the sports participation of children: a socio-economic analysis of the UK. European Journal of Sport Science, 14(1), 84–90. Downward, P., Lera-López, F., & Rasciute, S. (2011). The zero-inflated ordered probit approach to modelling sports participation. Economic Modelling, 28(6), 2469–2477. https://doi.org/10.1016/j. econmod.2011.06.024 Downward, P., & Rasciute, S. (2010). The relative demands for sports and leisure in England. European Sport Management Quarterly, 10(2), 189–214. Downward, P., & Rasciute, S. (2015). Exploring the covariates of sport participation for health: an analysis of males and females in England. Journal of Sports Sciences, 33(1), 67–76. Downward, P., & Rasciute, S. (2016). ‘No man is an island entire of itself’: the hidden effect of peers on physical activity: John Donne, Meditation XVII. Social Science & Medicine, 169, 149–156. Downward, P., & Riordan, J. (2007). Social interactions and the demand for sport: an economic analysis. Contemporary Economic Policy, 25, 518–537. Eberth, B., & Smith, M. (2010). Modelling the participation decision and duration of sporting activity in Scotland. Economic Modelling, 27, 822–834. Eime, R.M., Charity, M.J., Harvey, J.T., & Payne, W.R. (2015). Participation in sport and physical activity: associations with socio-economic status and geographical remoteness. BMC Public Health, 15(1), 434. https://doi.org/10.1186/s12889-015-1796-0 Frick, B., & Wicker, P. (2016). The trickle-down effect: how elite sporting success affects amateur

Sports Participation

participation in German football. Applied Economics Letters, 23(4), 259–263. García, J., Lera-López, F., & Suárez, M.J. (2011). Estimation of a structural model of the determinants of the time spent on physical activity and sport: Evidence for Spain. Journal of Sports Economics, 12(5), 515–537. García, J., Muñiz, C., Rodríguez, P., & Suárez, M.J. (2016). Comparative analysis of sports practice by types of activities. International Journal of Sport Finance, 11(4), 221–231. Gratton, C., & Taylor, P. (2000). Economics of Sport and Recreation. London: Spon Press. Hallmann, K., Wicker, P., Breuer, C., & Schönherr, L. (2012). Understanding the importance of sport infrastructure for participation in different sports: findings from multi-level modeling. European Sport Management Quarterly, 12(5), 525–544. Hovemann, G., & Wicker, P. (2009). Determinants of sport participation in the European Union. European Journal for Sport and Society, 6(1), 51–59. Humphreys, B.R., & Ruseski, J.E. (2007). Participation in physical activity and government spending on parks and recreation. Contemporary Economic Policy, 25, 538–552. Humphreys, B.R., & Ruseski, J.E. (2011). An economic analysis of participation and time spent in physical activity. The BE Journal of Economic Analysis and Policy, 11, 1–47. Humphreys, B.R., & Ruseski, J.E. (2015). The economic choice of participation and time spent in physical activity and sport in Canada. International Journal of Sport Finance, 10, 138–159. Kokolakakis, T., Castellanos-García, P., & Lera-López, F. (2017). Differences in formal and informal sports participation at regional level in England. International Journal of Sport Policy and Politics, 9(3), 491–504. https://doi.org/10.1080/19406940.2017. 1287757 Kuo, T., & Tang, H.L. (2014). Relationships among personality traits, Facebook usages, and leisure activities: a case of Taiwanese college students. Computers in Human Behavior, 31, 13–19. Lechner, M. (2009). Long-run labour market and health effects of individual sports activities. Journal of Health Economics, 28, 839–854. Lera-López, F., & Rapún-Gárate, M. (2007). The demand for sport: sport consumption and participation models. Journal of Sport Management, 21, 103–122. Lera-López, F., Wicker, P., & Downward, P. (2016). Does government spending help to promote healthy behavior in the population? Evidence from 27 European countries. Journal of Public Health, 38(2), e5–e12. Løyland, K., & Ringstad, V. (2009). On the price and income sensitivity of the demand for sports: has Linder’s disease become more serious? Journal of Sports Economics, 10, 601–618.

43

Marques, A., Ekelund, U., & Sardinha, L.B. (2016). Associations between organized sports participation and objectively measured physical activity, sedentary time and weight status in youth. Journal of Science and Medicine in Sport, 19(2), 154–157. Muñiz, C., Rodríguez, P., & Suárez, M.J. (2011). The allocation of time to sports and cultural activities: an analysis of individual decisions. International Journal of Sport Finance, 6(3), 245–264. Muñiz, C., Rodríguez, P., & Suárez, M.J. (2014). Sports and cultural habits by gender: an application using count data models. Economic Modelling, 36, 288–297. Mutter, F., & Pawlowski, T. (2014a). Role models in sports: can success in professional sports increase the demand for amateur sport participation? Sport Management Review, 17(3), 324–336. Mutter, F., & Pawlowski, T. (2014b). The causal effect of professional sports on amateur sport participation: an instrumental variable approach. International Journal of Sport Finance, 9(2), 172. Ruseski, J.E., Humphreys, B.R., Hallmann, K., & Breuer, C. (2011). Family structure, time constraints, and sport participation. European Review of Aging and Physical Activity, 8, 57–66. Ruseski, J.E., & Maresova, K. (2014). Economic freedom, sport policy, and individual participation in physical activity: an international comparison. Contemporary Economic Policy, 32(1), 42–55. Stigler, G.J., & Becker, G.S. (1977). De gustibus non est disputandum. The American Economic Review, 67(2), 76–90. Retrieved from www.jstor.org/ stable/1807222 (17.7.18). Stratton, M., Conn, L., Liaw, Ch., & Conolly, L. (2005). Sport and related recreational physical activity. The Social Correlates of Participation and Non-participation by Adults. Sport Management Association of Australia and New Zealand, 11th Annual Conference, Canberra. Taks, M., & Scheerder, J. (2006). Youth sports participation styles and market segmentation profiles: evidence and application. European Sport Management Quarterly, 6(2), 85–121. Thibaut, E., Eakins, J., Vos, S., & Scheerder, J. (2017). Time and money expenditure in sports participation: the role of income in consuming the most practiced sports activities in Flanders. Sport Management Review, 20(5), 455–467. https://doi. org/10.1016/j.smr.2016.12.002 Van Cauwenberg, J., Cerin, E., Timperio, A., Salmon, J., Deforche, B., & Veitch, J. (2017). Is the association between park proximity and recreational physical activity among mid-older aged adults moderated by park quality and neighborhood conditions? International Journal of Environmental Research and Public Health, 14(192), 1–11. https:// doi.org/10.3390/ijerph14020192

44

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Van Tuyckom, C., & Scheerder, J. (2010). Sport for all? Insight into stratification and compensation mechanisms of sporting activity in the 27 European Union member states. Sport, Education and Society, 15(4), 495–512. Vandermeerschen, H., Vos, S., & Scheerder, J. (2016). Towards level playing fields? A time trend analysis of young people’s participation in club-organised sports. International Review for the Sociology of Sport, 51(4), 468–484. Weed, M. (2009). The Potential of the Demonstration Effect to Grow and Sustain Participation in Sport. Review Paper for Sport England. Centre for Sport, Physical Education & Activity Research, Canterbury Christ Church University. Wicker, P., Breuer, C., & Pawlowski, T. (2009). Promoting sport for all age-specific target groups:

the impact of sport infrastructure. European Sport Management Quarterly, 9, 103–118. Wicker, P., & Downward, P. (2017). Exploring spill­ overs between government quality and individual health production through sport and physical activity. European Sport Management Quarterly, 17(2), 244–264. https://doi.org/10.1080/1618474 2.2016.1257038 Wicker, P., Hallmann, K., & Breuer, C. (2012). Micro and macro level determinants of sport participation. Sport, Business and Management: An International Journal, 2, 51–68. Wicker, P., Hallmann, K., & Breuer, C. (2013). Analyzing the impact of sport infrastructure on sport participation using geo-coded data: evidence from multi-level models. Sport Management Review, 16(1), 54–67.

5 Sports Participation and Health Jane E. Ruseski

INTRODUCTION Regular sport participation and physical activity (SPPA) is widely accepted as an effective way to mitigate the risks of chronic diseases, obesity, and premature death. Lifestyle choices, in particular health-related behaviors like physical activity, smoking, and alcohol use, are recognized as important non-medical determinants of health. The conceptual model of health used by County Health Rankings to develop an overall health index for each county in the United States weights health behaviors’ (including sport participation and physical activity) contribution to overall health at 30% (www.countyhealthrankings.org). This chapter focuses on sport participation and health outcomes through the lens of economics. There is clearly enough convincing science to recommend that all individuals should habitually participate in sport and physical activity to improve health and reduce the risk of many health problems. In their systematic review, Warburton et  al. (2010) conclude that the recent scientific literature supports a dose–response relationship between physical activity, all-cause mortality, and seven chronic diseases (cardiovascular disease, stroke, hypertension, colon cancer, breast cancer, type 2 diabetes, and osteoporosis) that

are most often associated with physical inactivity. Physically inactive people are also more likely to be obese, which is itself an important risk factor for many chronic diseases (for reviews of the literature on the effects of physical activity on health and disease, see United States Surgeon General, 1996; Sherwood & Jeffery, 2000; Katzmarzyk & Janssen, 2004; Warburton, Nicol, & Bredin, 2006; Brown, Burton, & Rowan, 2007). Although the health benefits of regular physical activity are well known, questions about the frequency and intensity needed to provide health benefits remain. From an economics perspective, questions about individual choices to engage in physical activity to promote health are central. This chapter explores the role that economists play in developing causal evidence of the effect of sport participation on health and illuminating what this evidence suggests about potential policy and intervention responses to promote and sustain regular and sport participation and physical activity. The next section lays out the policy context for sport participation and health. The following section provides the economic context motivating much of the empirical analyses in the literature, and this is followed by a discussion of the econometric challenges and empirical approaches. The next section summarizes the main finding in the

46

THE SAGE HANDBOOK OF SPORTS ECONOMICS

literature and finally, we conclude with some challenges for future research.

POLICY CONTEXT Rationale for Physical Activity Guidelines The European Union, United States, World Health Organization (WHO), and Canada all have published physical activity guidelines (European Commission, 2008; US Department of Health and Human Services, 2008; World Health Organization, 2010; Tremblay et al., 2011). Specific guidelines exist for children, adults aged 18–64, adults aged 65 and older, and special populations. For example, for adults (aged 18–64), the guidelines are to accumulate at least 150 minutes of moderate- to vigorous-intensity aerobic physical activity per week, in bouts of 10 minutes or more, to achieve health benefits. The rationale for developing these guidelines emanates from three widely accepted observations: (1) there is convincing science to recommend that all individuals should engage in regular physical activity to improve health and reduce the risk of many health problems; (2) physical activity is a lifestyle choice, just as diet, smoking and alcohol use are, that has important and independent effects on health; and (3) survey data from many countries indicate that most (if not all) populations fail to meet guidelines for physical activity. The target audiences for these guidelines are policy makers and health professionals. The main messages to these audiences are that regular physical activity over months and years can produce longterm health benefits and that realizing these benefits requires physical activity each week (World Health Organization, 2010).

Consequences and Challenges of Physical Inactivity The WHO report notes that levels of physical inactivity are rising in many countries and this has major implications for population health. Physical inactivity has been identified as the fourth leading risk factor for global mortality. Seven chronic diseases are consistently associated with physical inactivity – coronary heart disease, hypertension, stroke, colon cancer, breast cancer, type 2 diabetes, and osteoporosis. Physical inactivity is principal cause for 21–25% of breast and colon cancer

burden, 27% of diabetes burden, and 30% of ischemic heart disease burden (World Health Organization, 2010). Obesity is also a worldwide public health problem that can be reduced with regular SPPA. For these reasons, among others, promoting regular SPPA is an important public policy priority. Health professionals and policy makers face numerous challenges in addressing the widespread failure of individuals to achieve the recommended levels of physical activity for sustained health benefits. Developing physical activity guidelines would be easy if everyone was the same. In this case, one could simply say ‘X minutes of intense activity per week is needed to be healthy’ or ‘It is better to walk 40 minutes three times per week than 20 minutes six times per week’. However, a number of issues related to dose–response need to be considered. A dose is comprised of the type, intensity, duration, and frequency of activity. Current physical activity status, fitness level, health status, age, sex, and major health and fitness goals, and genetic differences all influence one’s responsiveness to a specific dose of activity. Sports and health economists have played an important complementary role in addressing this challenge in their efforts to develop causal evidence about the heterogenous dose–response relationship between SPPA and health.

ECONOMIC FRAMEWORK The economic framework underlying the majority of the econometric analyses of SPPA and health is Grossman’s health production model (Grossman, 1972). The Grossman model builds on the concept of household production introduced by Becker (1965) to develop a model of the demand for health. The model links household production and investment in human capital theories to describe the demand for health as having a consumption motive (people receive utility from being healthy), an investment motive (by investing in better health, people can increase the amount of healthy time available to earn income), or both. This model motivates the relationship between nonmedical inputs, such as SPPA, and health outcomes. In the basic health production model, individuals are assumed to have preferences over their health (Ht) and the consumption of other commodities (Ct). Age and education are the key observable factors affecting the stock of health in Grossman’s health production model. Over time, the model has been expanded to allow other factors, such

Sports Participation and Health

as health behaviors (or lifestyle choices) like participation in physical activity (SPPAt), to affect the stock of health (e.g., Cawley, 2004; Humphreys & Ruseski, 2011). Given this, an individual’s preferences can be described by a utility function: U (Ct, Ht, SPPAt). The stock of health depreciates over time. Individuals can invest in their stock of health by producing health through the combination of medical inputs (Mt), non-medical inputs (Xt) and time. Individuals make choices about how to allocate their time and resources to health investments and other activities subject to monetary budget and time constraints. The health production function can be written as Ht = h(Ht−1, Xt−1, Mt−1, SPPAt−1) where Ht is the stock of health in period t; Xt−1 represents non-medical inputs to health in period t − 1; Mt−1 is purchased medical inputs, and SPPAt−1 is participation in sport and physical activity. Ht highlights the temporal relationship between health and inputs by recognizing that the influence of healthy behaviors on health is not immediate. Some activities, like smoking, may provide utility today but are expected to decrease the stock of health in the future. Conversely, other activities, such as healthy eating habits and regular exercise, may increase or decrease utility today but are expected to increase the health stock (net of normal depreciation) in the future. Assuming that the marginal product of SPPA is positive and diminishing, Ht tells us (all else being equal) that the health stock in period t will be greater with positive participation in sport and physical activity relative to no participation. That is E[Ht|SPPAt−1 > 0] > E[Ht|SPPAt−1 = 0]. The health production framework recognizes individuals are heterogeneous in terms of health production. Investment in health realized by any individual depends on their initial endowment of health, their health production efficiency, and their decisions about engaging in healthy behaviours. The optimality conditions arising from this type of model describe the tradeoffs individuals face between choices that provide direct satisfaction and choices that provide indirect satisfaction through the production of health. The optimality conditions provide the basis for constructing and estimating empirical models of health outcomes.

EMPIRICAL CHALLENGES AND ECONOMETRIC APPROACHES Empirical Challenges The main empirical challenge in identifying a causal impact of SPPA on health is overcoming

47

endogeneity. SPPA is considered an endogenous regressor in a health outcome equation due to omitted variables that are correlated with both SPPA and health (e.g., urbanization) or unobserved individual characteristics that drive both SPAA and health (e.g., time preferences and tastes for SPAA). Regardless of its source, failure to address endogeneity generates inconsistent estimates of the effect of SPPA on health and prohibits the identification of causal effects. Secondary empirical challenges include uncovering individual heterogeneity in the dose– response relationship between SPPA and health and estimating the intertemporal nature of SPPA on health. Both challenges require longitudinal data with variables capturing individual characteristics and elements of the dose–response (type, intensity, frequency, and duration) and sufficient sample sizes to reliably identify the effects of SPPA on health.

Econometric Approaches Several approaches have been taken in the literature to address the endogeneity of health behaviors, including SPPA in econometric models of health outcomes. One approach is an instrumental variable model, such as two-stage least squares (2SLS) (Kenkel, 1995; Lindahl, 2005; Forrest & McHale, 2011; Pawlowski, Downward, & Rasciute, 2011; Huang & Humphreys, 2012; Ruseski et al., 2014; Sarma et al., 2015; Downward & Dawson, 2016, are examples). The instrumental variables approach entails first estimating a reduced form equation in which the endogenous regressor is the dependent variable. The predicted values from the first-stage regression replace the endogenous regressor in the second stage. Applying the 2SLS method to SPPA and health outcomes means estimating a first-stage equation such as: SPPAi* = α + βX i + δZi + ui where SPPAi* is a latent variable describing sport participation for individual i, Xi is a vector of individual characteristics that influence SPPA but are uncorrelated with the error term in the secondstage health outcome equation, Zi is a vector of instrumental variables, b and d are unknown parameters to be estimated, and ui is the error term. The observed realizations of SPPAi* are often specified as a dichotomous variable that equals 1 if the individual reports participating in sport and physical activity and 0 otherwise. Angrist and Krueger (2001) argue that it is preferable to treat the dichotomous dependent variable as a linear

48

THE SAGE HANDBOOK OF SPORTS ECONOMICS

probability and estimate the first-stage equation using the linear probability model. Using the predicted probability of SPAA from a non-linear probit model in the second stage is not recommended because the first-stage functional form must be correctly specified in order to generate consistent estimates in the second stage. Next, a second-stage health outcome equation of the following form is estimated:  i + γX i + Ei Hi* = η + λSPPA where Hi* is the latent health outcome for indi i are the fitted values of SPPAi vidual i, SPPA obtained from estimating the first-stage equation, Xi is a vector of individual characteristics that affect health outcomes, h and g are unknown parameters to be estimated, and Ei is the error * term. Like SPPAi , the observed realizations of * Hi are often specified as a dichotomous variable where Hi = 1 if Hi* ≥ 1 and 0 otherwise. In order for the parameters of the health outcomes equation to be consistently estimated, a variable must be included in the first-stage SPPA equation that is not included in the second-stage health outcome equation. This instrumental variable Zj, should explain variation in sport participation but be uncorrelated with the error term, Ei, in the health outcome equation. A second approach commonly used to account for the endogeneity of health inputs is to model the production of health as recursive structures with reduced form equations for health behaviors and a structural equation for the health production function (examples include Contoyannis & Jones, 2004; Balia & Jones, 2008; Humphreys, McLeod, & Ruseski, 2014; Sarma et al., 2014, 2015). The models are estimated as either bivariate or multivariate probit models. As an example, consider the following recursive bivariate probit model: SPPAi* = α + βX iPPA + ui Hi* = η + λSPPAi + γX iH + ∈i where SPPAi* is a latent variable describing sport participation for individual i, X iSPPA and X iH are vectors of individual characteristics that affect sport participation and health outcomes respectively, and b, l, and g are unknown parameters to be estimated. ui and ei are possibly correlated error terms that capture unobserved factors that affect SPPA and health outcomes. If the error terms are uncorrelated, then the bivariate probit model is equivalent to two independent probit models. The bivariate probit model is recursive in that health outcomes depend on the exogenous variables X iH and sport participation SPPAi. In this context the

sport participation equation is a reduced form equation which depends on the exogenous vari­ ables, X iSPPA. The health outcome equation is a structural equation which depends on the exogenous variables, X iH , and sport participation. Identification in bivariate and multivariate probit models relies on exclusion restrictions (Maddala, 1983). An explanatory variable must appear in X iSPPA in the reduced form equation that is excluded from X iH in the structural equation. This regressor must satisfy the requirement that it influences sport participation but does not directly affect health outcomes. However, Wilde (2000) shows that an exclusion restriction is not required to identify the parameters in the reduced form * * equation, SPPAi , as long as SPPAi and X iH each contain one varying explanatory variable. This approach is commonly referred to as ‘identification by functional form’ and relies heavily on the assumption of bivariate normality. Since bivariate normality may be a strong assumption, exclusion restrictions are often imposed to improve identification (Jones, 2007). Matching estimators are another approach to estimating causal effects of SPPA on health outcomes (Lechner, 2009; Cabane, Hille, & Lechner, 2015; Sari & Lechner, 2015). In general, matching estimators can be applied to situations where one has a treatment group (in this context, individuals who participate in sport) and an untreated (or control) group (individuals who do not participate in sport). The basic idea is to find a group of non-participants (untreated or control group) who are similar to the participants (treatment group) with respect to a set of observable characteristics, X (Caliendo & Kopeinig, 2008). After matching, differences in health outcomes between the treated and control groups can be attributed to sport participation (the treatment). The underlying identifying assumption is that the treatment satisfies some form of exogeneity and is referred to as unconfoundedness (Rosebaum & Rubin, 1983), selection on observables (Heckman & Robb, Jr, 1985), or conditional independence (Lechner, 1999). This assumption essentially means that assignment to treatment is independent of the outcomes, conditional on the covariates. The most commonly used matching estimator in the sport participation literature is propensity score matching.

EMPIRICAL EVIDENCE Sport Participation and Physical Health Belloc and Breslow (1972) developed some of the first epidemiological evidence of the effect of

Sports Participation and Health

lifestyle choices on health using data from a random sample of 7,000 residents of Alameda County, California. Coined the Alameda Seven, Belloc and Breslow find that maintaining proper weight, not snacking between meals, never smoking cigarettes, regular physical activity, moderate or no use of alcohol, and getting 7–8 hours of sleep regularly were all associated with better health. Kenkel (1995) re-examines the importance of the Alameda Seven using the health production framework discussed earlier. This is one of the first studies in the economics literature to consider the effect of (un)healthy behaviors on health using an instrumental variables approach. Kenkel, like Belloc and Breslow (1972), finds that regular exercise is associated with better health. Still building on the work of Belloc and Breslow, Contoyannis and Jones (2004) examine the effects of socioeconomic status and lifestyle on health using data from the British Health and Lifestyle Survey (HALS). They estimate the structural parameters of a health production function together with the reduced form parameters for health behavior equations. Health is measured by a binary indicator of self-assessed health status. The endogenous health behavior variables are based on the Alameda Seven. They find that sleeping well, exercising, and not smoking had positive effects on the probability of reporting excellent or good health. Balia and Jones (2008) also use the British HALS to estimate the effects of health behaviours on self-assessed health status and mortality. Like Contoyannis and Jones, health is measured by a binary indicator of self-assessed health status and the healthy behaviour variables are based on the Alameda Seven. Regular exercise is not found to be an important determinant of health. Humphreys et al. (2014) is the first recent paper to evaluate the dose–response of SPPA on a range of health outcome measures using data from the 2005/2006 Canadian Community Health Survey (CCHS) public use file. They estimate bivariate probit models with a structural equation for health and reduced form equation for participation in physical activity. They find contemporaneous participation in physical activity reduces the incidence of diabetes, high blood pressure, heart disease, asthma, and arthritis as well as being in fair or poor self-reported health. There appears to be a diminishing marginal impact on adverse health outcomes from an increase in the intensity of physical activity above the moderate level. Sarma et  al. (2015) extend Humphreys et  al. (2014) by evaluating the effect of both leisure time (LTPA) and work-related physical activity (WRPA) on obesity, diabetes, high blood pressure,

49

and heart disease using data from three cycles of the CCHS Confidential Master Files. They estimate instrumental variables and bivariate probit models using monthly average temperatures as an instrument or exclusion restriction for the endogenous SPPA variable. In contrast to Humphreys et  al. (2014), Sarma et  al. (2015) do not find an effect of LTPA on the probability of having diabetes, high blood pressure, or heart disease. On the other hand, they find that WRPA had a negative effect on obesity and having diabetes, high blood pressure, and heart disease. In a separate paper, Sarma et  al. (2014) ask a similar question to Sarma et al. (2015) but focus on an individual’s body mass index (BMI) as the health outcome measure. They use longitudinal data from eight cycles of the Canadian National Population Health Survey (NPHS). Overall, LTPA is found to reduce BMI, with a larger effect for females. WRPA is found to have a modest effect (less than half a BMI point) on reducing BMI, again with a modestly larger effect for females. Sari and Lechner (2015) also make use of the NPHS to evaluate the effect of sports and exercise on labor market outcomes. This paper differs from the papers discussed above in terms of using propensity score matching to account for the endogenous physical activity variable. Although the primary focus of this study is on labor market outcomes, the authors posit that one possible channel for the labor market outcomes is the health effects of SPPA because physical activity has a positive effect on health and healthy individuals are more productive at work. Using self-reported health status as the measure of health, they find that going from being inactive to moderately active has no effect on self-reported health status but going from being inactive to ‘actively’ active has a positive effect on health.

Sport Participation and Mental Health A number of recent studies evaluate the impact of SPPA on mental health, typically defined as subjective well-being (SWB) or happiness (Rasciute & Downward, 2010; Downward & Rasciute, 2011; Forrest & McHale, 2011; Pawlowski et  al., 2011; Huang & Humphreys, 2012; Ruseski et al., 2014; Wicker & Frick, 2015, 2017; Downward & Dawson, 2016). The majority of these studies identify the effect of SPPA on happiness or subjective well-being using an instrumental variables approach (Forrest & McHale, 2011; Pawlowski et al., 2011; Huang & Humphreys, 2012; Ruseski et al., 2014; Wicker & Frick, 2015, 2017; Downward & Dawson, 2016). Rasciute and Downward (2010) estimate bivariate probit models,

50

THE SAGE HANDBOOK OF SPORTS ECONOMICS

while Downward and Rasciute (2011) estimate ordered choice models with heterogenous thresholds to account for unobserved heterogeneity. Forrest and McHale (2011) use data from the first wave of the UK Taking Part Survey. This survey asks questions about participation in 67 sports activities as well as questions about the frequency in days of participation and typical minutes engaged in participation. They address the endogeneity of SPPA using an instrumental variable (IV) approach where they exploit variation in the respondents’ proximity to a sports facility to identify SPPA. Proximity to a sports facility is defined as the ability to travel to such a facility from home in 20 minutes or less. Well-being in the Taking Part Survey is measured by asking respondents how happy they are on a scale of 1–10 where 1 indicates extremely unhappy and 10 indicates extremely happy. They find that women who participated in sport and physical activity reported higher levels of happiness than women who did not participate. Similar studies by Huang and Humphreys (2012) and Ruseski et al. (2014) find results that are consistent with Forrest and McHale (2011). Huang and Humphreys (2012) use data from the US Behavioral Risk Factor Surveillance Survey (BRFSS) and County Business Patterns to estimate the effect of SPAA on self-reported health status and self-reported happiness. They exploit variation in the availability of sports facilities to identify the effect of SPPA on health and happiness. Subjective well-being is measured based on a question about overall life satisfaction, with four possible responses: very satisfied, satisfied, dissatisfied, and very dissatisfied. They find that people who are physically active report better self-reported health status and self-reported happiness. Ruseski et  al. (2014) use data from a crosssectional survey of residents of Rheinberg, Germany (2009). They also take an instrumental variables approach to identify SPPA using two instruments – the proximity to a sports facility and belief that physical activity is important. They find that individuals who practice sport report higher levels of happiness. Happiness is measured on a 5-point scale where 1 indicates very unsatisfied and 5 indicates very satisfied. They find that individuals who participate in sport and physical activity are more satisfied with life than those who do not participate. Like Forrest and McHale (2011), Rasciute and Downward (2010) use data from the Taking Part Survey to look at the effect of specific activities on happiness. They find that cycling, walking and ‘general forms of physical activity’ are positively associated with physical and mental health. In a

related study, Downward and Rasciute (2011) collapse various activities into sports that have a social interaction aspect (team sports and sports that required an opponent) and those that do not (individual sports like swimming, cycling, and weight training). They find that SPAA in general is positively associated with happiness, but sports with a social interaction component have a larger effect on happiness than those that do not. Pawlowski et  al. (2011) use data on a sample of 19 European countries drawn from the crosssectional International Social Survey Programme (ISSP) to evaluate the effect of SPPA on happiness. In this survey, happiness is measured on a 4-point scale with 1 indicating not at all happy and 4 indicating very happy. They estimate models using Generalised Method of Moments (GMM), where the instruments are frequency of attending sports events and individuals’ participation in a sports association. They find that people who are physically active report being happier than those who are not. Wicker and Frick (2015, 2017) extend this literature by trying to get at a dose–response relationship between SPPA and SWB by developing crude measures of intensity and duration using data from the cross-sectional Eurobarometer survey of European Union countries. SWB is measured on a 4-point scale in the Eurobarometer with 1 indicating not at all satisfied and 4 indicating very satisfied. Like other studies, they take an instrumental variable approach to identify SPPA in the wellbeing equations. The instruments used in both studies are the perceived availability of sport opportunities, being a member of sports club, and the time spent sitting on a usual day. The sport participation variables are walking (reflecting light activity), the number of days participating in moderate and vigorous activities in the last week, the number of minutes engaged in moderate and vigorous activity in the last week, and a set of dummy variables indicating if the reported activity complies with WHO guidelines for the amount and intensity of physical activity needed to achieve sustainable health benefits. They find positive effects of moderate intensity and duration on SWB but negative effects of vigorous intensity and duration. Downward and Dawson (2016) also explore the dose–response of SPPA on happiness using the third wave of the UK Taking Part Survey. Like Wicker and Frick (2015, 2017), they develop measures of intensity based on answers to a question about the activity raising the respondents’ breathing rate. These responses, coupled with frequency and duration of participation, are then used to construct measures of intensity. They estimate models

Sports Participation and Health

using GMM where the instruments are a measure of sport supply, defined as being able to get to a sports facility within 20 minutes from home (same as Forrest & McHale, 2011), and interview month. They find that SPPA is positively associated with happiness. With regard to intensity, they find that lower levels of intensity of SPAA generate the highest overall well-being.

DISCUSSION Regular participation in sport and physical activity is now widely accepted as an effective measure for reducing the risk for a number of adverse health outcomes across all socioeconomic and demographic subgroups. Yet, despite this knowledge, the most recent global comparative estimates (2010) indicate that worldwide, 23% of adults and 81% of adolescents (aged 11–17 years) do not meet the WHO global recommendations on physical activity for health (World Health Organization, 2018). The economic burden of physical inactivity is immense. The global cost of physical inactivity in 2013 is estimated to be INT$54 billion per year in direct healthcare costs, with an additional INT$ 14 billion attributable to lost productivity (Ding et al., 2016). Not surprisingly, promoting regular participation in sport and physical activity remains an important public policy priority. A large literature studying the determinants of and the effects of sport participation and physical activity has emerged in the biological and social sciences over the years. An ongoing challenge in this literature is developing causal evidence of the effect of SPPA on physical and mental health outcomes. This chapter focuses on the contributions from economics seeking to develop causal evidence of these effects. Overall, the consensus in the literature thus far is that SPPA has a positive impact on health outcomes. The WHO global recommendation on physical activity for health, for adults aged 18–64, is at least 150 minutes of moderate-intensity activity (or equivalent) throughout the week (World Health Organization, 2010). The recommendation goes further to specify the frequency, intensity, and duration of episodes of activity needed to achieve substantial health benefits. Some of the recent studies reviewed in this chapter begin to explore the dose–response nature of the WHO guidelines (Humphreys et al., 2014; Sarma et al., 2015; Wicker & Frick, 2015, 2017; Downward & Dawson, 2016) and generate some thoughtprovoking results.

51

Humphreys et al.’s (2014) results suggest that there are diminishing marginal returns to SPPA since going from inactivity to begin moderately active has the largest marginal impact on improving health outcomes. In contrast, Sarma et al. (2015) do not find an effect of leisure time physical activity on health outcomes but do find that ‘moderate’ work-related physical activity reduces negative health outcomes. Wicker and Frick (2015, 2017) find a negative effect of intense SPPA on subjective well-being while Downward and Dawson (2016) find that SWB is higher for individuals who engage in less intense activities. One might be tempted to conclude from these results that recommendations for moderately intense activity, whether at leisure or at work, should be emphasized, since this is where health and well-being benefits are seen. However, it is important to note that the measures of activity in these studies do not correspond precisely with the WHO guidelines. A fruitful avenue for future research is to more carefully assess the dose–response relationship between SPPA and health outcomes. The majority of the studies reviewed in this chapter assess the average treatment effect of SPPA on health outcomes for the study population as a whole, controlling for socioeconomic and demographic characteristics such as age, gender, education, and income. Less attention is paid to identifying a social gradient in SPPA and decomposing the effect of socioeconomic and demographic characteristics on SPPA. This is another fruitful avenue for future research given the persistent worldwide trend in failing to meet the WHO recommendations on physical activity for health. A better understanding of a social gradient in SPPA and which equityrelevant characteristics underlie the gradient is needed for informing policies aimed at encouraging regular SPPA.

REFERENCES Angrist, J., & Krueger, A. (2001). Instrumental variables and the search for identification: From supply and demand to natural experiments. Journal of Economic Perspectives, 15(4), 69–85. Balia, S., & Jones, A. (2008). Mortality, lifestyle and socio-economic status. Journal of Health Economics, 27, 1–26. Becker, G. (1965). A theory of the allocation of time. The Economic Journal, 75(299), 493–517. Belloc, N., & Breslow, L. (1972). The relationship of physical health status and health practices. Preventive Medicine, 1, 409–421.

52

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Brown, W., Burton, N., & Rowan, P. (2007). Updating the evidence on physical activity and health in woman. American Journal of Preventive Medicine, 33, 404–411. Cabane, C., Hille, A., & Lechner, M. (2015). Mozart or Pelé? The effects of teenagers’ participation in music and sports. Labour Economics, 41, 90–103. Caliendo, M., & Kopeinig, S. (2008). Some practical guidance for the implementation of propensity score matching. Journal of Economic Surveys, 22(1), 31–72. Cawley, J. (2004). An economic framework for understanding physical activity and eating behaviors. American Journal of Preventive Medicine, 27(3), 117–125. Contoyannis, P., & Jones, A. (2004). Socio-economic status, health and lifestyle. Journal of Health Economics, 23, 965–995. Ding, D., Lawson, K. D., Kolbe-Alexander, T. L., Finkelstein, E. A., Katzmarzyk, P. T., Van Mechelen, W., Pratt, M., for the Lancet Physical Activity Series 2 Executive Committee (2016). The economic burden of physical inactivity: A global analysis of major non-communicable diseases. The Lancet, 388(10051), 1311–1324. Downward, P., & Dawson, P. (2016). Is it pleasure or health from leisure that we benefit from most? An analysis of well-being alternatives and implications for policy. Social Indicators Research, 126(1), 443–465. Downward, P., & Rasciute, S. (2011). Does sport make you happy? An analysis of the well-being derived from sports participation. International Review of Applied Economics, 25(3), 331–348. European Commission (2008). EU physical activity guidelines: Recommended policy actions in support of health-enhancing physical activity. Technical report. Brussels: European Commission. Forrest, D., & McHale, I. (2011). Subjective well-being and engagement in sport: Evidence from England. In P. Rodríguez, S. Késenne, & B. R. Humphreys (Eds.), The economics of sport, health and happiness: The promotion of well-being through sporting activities pages (pp. 184–199). Cheltenham, UK: Edward Elgar. Grossman, M. (1972). On the concept of health capital and the demand for health. Journal of Political Economy, 80(2), 223–255. Heckman, J. J., & Robb, Jr, R. (1985). Alternative methods for evaluating the impact of interventions: An overview. Journal of Econometrics, 30(1–2), 239–267. Huang, H., & Humphreys, B. (2012). Sports participation and happiness: Evidence from US microdata. Journal of Economic Psychology, 33(4), 776–793. Humphreys, B., & Ruseski, J. (2011). The economics of participation and time spent in physical activity.

BE Journal of Economic Analysis & Policy, 11(1), 1–38. Humphreys, B. R., McLeod, L., & Ruseski, J. E. (2014). Physical activity and health outcomes: Evidence from Canada. Health Economics, 23(1), 33–54. Jones, A. (2007). Applied econometrics for health economics (2nd ed.). Abingdon, UK: Radcliffe Publishing. Katzmarzyk, P., & Janssen, I. (2004). The economic costs associated with physical inactivity and obesity in Canada: An update. Canadian Journal of Applied Physiology, 29(1), 90–115. Kenkel, D. (1995). Should you eat breakfast? Estimates from health production functions. Health Economics, 4(1), 15–29. Lechner, M. (1999). Earnings and employment effects of continuous off-the-job training in East Germany after unification. Journal of Business & Economic Statistics, 17(1), 74–90. Lechner, M. (2009). Long-run labour market and health effects of individual sports activities. Journal of Health Economics, 28(4), 839–854. Lindahl, M. (2005). Estimating the effect of income on health and mortality using lottery prizes as an exogenous source of variation in income. The Journal of Human Resources, 40(1), 144–168. Maddala, G. (1983). Limited dependent and qualitative variables in econometrics. Cambridge, UK: Cambridge University Press. Pawlowski, T., Downward, P., & Rasciute, S. (2011). Subjective well-being in European countries: On the age-specific impact of physical activity. European Review of Aging and Physical Activity, 8, 93–102. Rasciute, S., & Downward, P. (2010). Health or happiness? What is the impact of physical activity on the individual? Kyklos, 63(2), 256–270. Rosebaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. Ruseski, J. E., Humphreys, B. R., Hallman, K., Wicker, P., & Breuer, C. (2014). Sport participation and subjective well-being: Instrumental variable results from German survey data. Journal of Physical Activity and Health, 11(2), 396–403. Sari, N., & Lechner, M. (2015). Long-run health effects of sports and exercise in Canada. Technical report, Canadian Centre for Health Economics. Working Paper #150018. Sarma, S., Devlin, R. A., Gilliland, J., Campbell, M. K., & Zaric, G. S. (2015). The effect of leisure-time physical activity on obesity, diabetes, high BP and heart disease among Canadians: Evidence from 2000/01 to 2005/06. Health Economics, 24(12), 1–17. Sarma, S., Zaric, G. S., Campbell, M. K., & Gilliland, J. (2014). The effect of physical activity on adult obesity: Evidence from the Canadian NPHS panel. Economics and Human Biology, 14, 1–21.

Sports Participation and Health

Sherwood, N., & Jeffery, R. (2000). The behavioral determinants of exercise: Implications for physical activity interventions. Annual Review of Nutrition, 20(1), 21–44. Tremblay, M. S., Warburton, D. E., Janssen, I., Paterson, D. H., Latimer, A. E., Rhodes, R. E., Kho, M. E., Hicks, A., LeBlanc, A. G., Zehr, L., et al. (2011). New Canadian physical activity guidelines. Applied Physiology, Nutrition, and Metabolism, 36(1), 36–46. United States Surgeon General (1996). Physical activity and health: A report of the Surgeon General. Technical report. Washington, DC: Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion. US Department of Health and Human Services (2008). Physical activity guidelines for Americans. Technical report. Washington, DC: US Department of Health and Human Services. Warburton, D., Charlesworth, S., Ivey, A., Nettlefold, L., & Bredin, S. (2010). A systematic review of the evidence for Canada’s physical activity guidelines for adults. International Journal of Behavioral Nutrition Physical Activity, 7(1), 39.

53

Warburton, D., Nicol, C., & Bredin, S. (2006). Health benefits of physical activity: The evidence. Canadian Medical Association Journal, 174(6), 801–809. Wicker, P., & Frick, B. (2015). The relationship between intensity and duration of physical activity and subjective well-being. The European Journal of Public Health, 25(5), 868–872. Wicker, P., & Frick, B. (2017). Intensity of physical activity and subjective well-being: An empirical analysis of the WHO recommendations. Journal of Public Health, 39(2), e19–e26, online first, July 13, 2016. https://doi.org/10.1093/pubmed/fdw062 Wilde, J. (2000). Identification of multiple probit models with endogenous dummy regressors. Economics Letters, 69, 309–312. World Health Organization (2010). Global recommendations on physical activity for health. Technical report. Geneva: World Health Organization. World Health Organization (2018). Global action plan on physical activity 2018–2030: More active people for a healthier world. Technical report. Geneva: World Health Organization.

6 Sport and Social Capital Formation Tim Pawlowski and Ute Schüttoff

INTRODUCTION Although not central to his analysis, one of the most influential scholars in social science who conceptualized the term ‘social capital’ suggests that sport might offer the opportunity to generate social capital (Putnam, 2000). In this regard, sport is considered to be a platform for people to meet, to enjoy being together and thus to create social networks. It is assumed by Putnam that structured activities like participation in sports clubs leads to an increase in social cohesion and strengthens communities and therefore might justify the public funding of sport or sports clubs. In addition, sports clubs are often described as a venue to attract and unite individuals from different backgrounds, including socially disadvantaged individuals. Regular meetings and the cooperation during sports participation or in sports clubs build or enhance relationships and social networks, which might in turn improve the well-being of the individuals and encourage volunteering, which is also viewed as a central component of a civic society. At the same time, participation in social networks as sports groups or teams is seen as tool for a reduction in crime and anti-social behavior among young individuals (Hoye & Nicholson, 2008). All these claims have hardly been tested empirically before, although they frequently serve to

legitimize the use of considerable public funds to promote sport in the western world.1 The social benefits of sport are also claimed when justifying the use of sport as a tool for development aid. For instance, the United Nation’s (UN) Office on Sport for Development and Peace states: Sport plays a significant role as a promoter of social integration and economic development in different geographical, cultural and political contexts. Sport is a powerful tool to strengthen social ties and networks, and to promote ideals of peace, fraternity, solidarity, non-violence, tolerance and justice. (UN, 2014)

Therefore, this chapter intends to summarize both the theoretical background as well as the empirical findings on the role of sports participation for the accumulation of social capital.

GENERAL CONCEPTS Concepts of Social Capital The term ‘social capital’ has been used by several authors and can therefore be described as an umbrella term that encompasses very different

Sport and Social Capital Formation

underlying concepts. Overall, various types of social relations are linked to the concept of social capital, such as informal support, contact with family and friends, or social norms and trust. In the following, the different approaches of social capital developed by Pierre Bourdieu, James Coleman, and Robert Putnam, are presented since their works can be seen as the most relevant contributions in this field of research. The French sociologist Pierre Bourdieu predominantly focuses on the mechanism of how power relations and hierarchies are maintained in an unequal society, developing the concept of cultural capital as a polar opposite to economic capital. In subsequent work, Bourdieu develops the concept of social capital, which he defines as: the sum of the resources, actual or virtual, that accrue to an individual or a group by virtue of possessing a durable network of more or less institutionalized relationships of mutual acquaintance and recognition. (Bourdieu & Wacquant, 1992, p. 119)

James Coleman dealt with the concept of social capital through originally working in the area of education as an American sociologist. Coleman focused on social capital to describe the link between social inequalities and academic performance and argued that students who received support from their family and communities performed better in school (Coleman, 1988). According to Coleman, social capital is a resource for individuals which results from different social networks, including families, communities, and schools. These social relations establish social capital such as trustworthiness, set norms, and creates channels for information. The concept of social capital developed by Robert Putnam, an American political scientist, is based on studies of the factors that contribute to democratic performance and good governance, finding that governments perform better when a stronger sense of ‘civic community’ exists. Accordingly, a civic community is characterized by public-mindedness, which manifests itself through associational life in the community and the expectations that other members of the community will probably also adopt this behavior. To sum up, Putnam defines social capital as all ‘form[s] of norms of reciprocity and networks of civic engagement’ (Putnam, 1993, p. 167). Putnam also highlights trust as an essential component of social capital and, over time, broadened the concept of social capital as embracing different forms of interpersonal networks, such as workplace connections and informal socializing, as well as social engagement, such as political, civic and religious participation, and norms and values, such as reciprocity, trust, and altruism (Putnam, 2000).

55

In general, the concepts of the three contributors mainly differ with regard to the question of whether social capital is a private good or a public/collective good. Social capital as a private good describes a resource for the individual who ‘owns’ it, which is Bourdieu’s perspective. In contrast, social capital as a public/collective good describes its benefits being generated for other members of the society, which is Putnam’s perspective. Finally, Coleman recognizes both the public good and private good aspects of social capital (Scrivens & Smith, 2013).

Four Dimensions of Social Capital Based on the different concepts of social capital, Scrivens and Smith (2013) establish four dimensions that reflect different views of what social capital ‘is’: (1) personal relationships; (2) social network support; (3) civic engagement; and (4) trust and cooperative norms. (1) Personal relationships refer to the individual networks of people, comprising those people they know, and includes the individual’s behavior to establish and maintain those networks, e.g., spending time with each other. In this dimension, the level and nature of social relationships are of special interest and not the outcomes or the value which can be derived from those relationships. The direct outcome of people’s personal relationships is covered in the dimension (2) social network support. Social network support includes emotional, material, practical, financial, intellectual or professional resources that are derived through relationships with others and are available to each individual in their personal social networks. Activities such as volunteering, political participation, group membership, and different forms of community action through which people contribute to civic and community life are summarized in the dimension (3) civic engagement. Higher individual levels of civic engagement can contribute to both collective and individual outcomes. On the one hand, it can enhance institutional performance and the trust and cooperative norms within a society and, on the other hand, it can positively impact the individual’s well-being, because the activities provide an opportunity to get in contact with new people which can bring enjoyment. Furthermore, through participating in new activities individuals can gain new skills. The last dimension (4) trust and cooperative norms comprises intangible factors such as trust, social norms, and

56

THE SAGE HANDBOOK OF SPORTS ECONOMICS

shared values that contribute directly to better social and economic outcomes. The intangible values, for example, foster societal functioning or facilitating mutually beneficial cooperation, and therefore constitute a collective resource. Trust can also be differentiated further to reflect trust towards already known people and trust between strangers. The latter is described as generalized trust (or social trust) and is of higher relevance than particularized trust in friends, family, and acquaintances as it can be translated into broader social outcomes and enables social interaction (Putnam, 2000). The second part of the fourth dimension includes cooperative norms. Overall, norms differ throughout groups and are defined by, for example, role models. Norms influence the behavior in the group, but also describe the form of the use of sanctions in the case of non-compliance. Therefore, to be more precise, norms describe the expected behavior of individuals within a group or community. In general, norms such as tolerance and non-discrimination are of particular importance since they foster fair and inclusive cooperation, especially towards people and groups of different background, appearance or beliefs (Scrivens & Smith, 2013). Next to this conceptualization, two different types of mechanism of accumulating social capital can be distinguished. Putnam (2000) differentiates between horizontal and vertical relationships. Horizontal relationships are derived when people with similar status or power are linked. This kind of relationship is promoted through active participation in civic-minded groups such as neighborhood associations, choral societies, and sports clubs and is referred to bonding social capital. The creation of vertical relationships is more difficult because socially perceived different groups, for example, different ethnicities or classes, have to be linked. This bridging social capital is relevant in reducing inequities because individuals develop responsibilities for individuals beyond their familiar environment (Baum & Ziersch, 2003).

Sports and Social Capital In work on democracy and social capital, Uslaner (1999) lists several issues why sports participation builds social capital, including the generation of self-confidence and learning respect for the rules, widening the social contacts of individuals, and spreading tolerance and egalitarian values as well

as providing lessons in morality as a by-product of sport. In the following discussion, the relevance of sport for social capital accumulation will be explained and examined in more detail. Theoretical assumptions on the influence of sport on the four dimensions of social capital are presented, before empirical studies and their findings are discussed (an overview is provided in Table 6.1).

SPORT AND PERSONAL RELATIONSHIPS Personal relationships refer to relationships with those an individual already knows. This dimension can be broadened by including potential new relationships. Sport practiced in its diverse organizational formats provides a platform to get into contact with people. To get acquainted with other people and to make new friends might be a motivation to commence participating in sport or to joining a sports organization. In contrast, already established relationships or networks can be recruited into the sport or sports organization to strengthen existing relationships. A key factor for both establishing new relationships or strengthening existing relationships is frequent participation in sport or frequent meetings in a sport environment. The more time spent engaged in sport or in a sports organization, the higher the probability of getting into contact with others, which might translate into stronger existing relationships or even life-long friendships (Ulseth, 2004). In addition, it is suggested that voluntary sports clubs and commercial sports facilities, such as fitness centers, fosters the creation or strengthening of relations in different ways. In voluntary sports clubs there are also opportunities to meet through voluntary work, which affects the relations between the individuals in an organization. Such voluntary work is not required in commercial facilities, implying that personal relations are established to a lesser degree than in sports clubs (Ulseth, 2004). There are several empirical studies that analyze whether sports participation helps in establishing new or strengthening existing relationships. For instance, Becker and Häring (2012) analyze the social integrative function of sport with regard to personal relationships based on cross-sectional data of 2,002 German adults. They find that participation in sport is associated with having more and stronger personal relationships – measured by the number of friends, frequency of contact to friends, and the number of social settings in which repeated contact takes place (e.g., in the job, in the living environment, or during other leisure

Political engagement (voting)

Civic involvement

(iii) Civic engagement

Schüttoff et al. (2018)

Helping out friends, neighbors, or relatives Social acceptance and respect by peers

Delaney & Keaney (2005)

Frisco et al. (2004)

Schüttoff et al. (2018)

Gerlach & Brettschneider (2013)

Makarova & Herzog (2014)

Pawlowski et al. (2018)

Felfe et al. (2016)

Becker & Häring (2012)

Ulseth (2004)

Reference

Integration of immigrants

(ii) Social network support

Contact with and support of friends/peers

(i) Personal relationships Building or strengthening relationships

Dimensions and measure of social capital

Matching

Matching, instrumental variable

Regression

Regression

Method

Germany Socio-economic panel (2003–2013, n = 1,111) USA National Educational Longitudinal Study (1988–1994, n = 10,839) UK (and Europe) European Social Survey, Home Office Citizenship Survey, Time Usage Survey (2000–2002, n = n/k)

Regression

Regression

Matching

Switzerland Correlations Survey of students with a migration background (n = 454) Germany Matching Socio-economic panel (2003–2013, n = 1,111) Germany, Variance analysis Longitudinal study of children in the city of Paderborn (2001–2011, n = 1,637)

Norway Customers of fitness studios (2001, n = 1,585), members of sports clubs (2000, n = 1,205) Germany Partner market survey (2009, n = 2,002) Germany Survey on the health of children and adolescents in Germany, panel of children, information on the sports infrastructure (2002–2006, n = 5,632) Peru Young Lives (2002–2009, n = 658)

Country and data

Table 6.1  Empirical studies on sport and social capital formation

(+)

(Continued )

(+) dependent on the organizational format and the type of sport (+) in non-school team sports

(+) dependent on the organizational format (+) in sports clubs

(+) in sports clubs

(+) in sports groups

(+) dependent on the organizational format and the type of sport (+) in sports clubs

(+) dependent on the organizational format

Results

Sport and Social Capital Formation 57

Schüttoff et al. (2018)

Delaney & Keaney (2005)

Downward et al. (2014)

Bartolomeo & Papa (2019)

Seippel (2006)

Note: n ≡ number of observations, n/k ≡ not known, (+) ≡ positive effect, (−) ≡ negative effect.

Trust in civil institutions

Generalized trust

Norway Johns Hopkins Comparative Nonprofit Sector Project (1998, n = 1,695) Italy Experiment with undergraduate students at the Sapienza University of Rome (2014, n = 120) Worldwide, 30 countries International Social Survey Programme (2007, n = 34,324) United Kingdom (and Europe) European Social Survey, Home Office Citizenship Survey, Time Usage Survey (2000–2002, n = ns)

Norway Johns Hopkins Comparative Nonprofit Sector Project (1998, n = 1,695) Canada National Survey of Giving, Volunteering and Participating (2000, n = 12,387) Germany Socio-economic panel (2003–2013, n = 1,111)

Seippel (2006)

Perks (2007)

Country and data

Reference

(iv) Trust and cooperative norms

Volunteering

Dimensions and measure of social capital

Table 6.1  (Continued )

(+)

(+) in voluntary sports organizations

(+) dependent on the organizational format

(+)

(+) in voluntary sports organizations

Results

Regression, instrumental (−) in voluntary sports associations variable approach Regression (+)

Regression

Regression

Matching

Analysis of variance

Regression

Method

58 THE SAGE HANDBOOK OF SPORTS ECONOMICS

Sport and Social Capital Formation

activities). In addition, they find some differences with regard to the type of sport and the type of organizational format. For instance, participants in team sports are less likely to have contacts in other social settings, while members of sports clubs are more likely to have contacts through a variety of social settings. Ulseth (2004) also finds effect differences with regard to the organizational format based on cross-sectional data on customers of fitness studios and members of sports clubs in Norway. Results show that sports clubs are venues in which new friendships are likely to be established, while fitness centers are more likely to be a place to meet and maintain existing friendships. In a more recent study, Felfe, Lechner, and Steinmayr (2016) analyze the effect of sports club participation on cognitive and non-cognitive skill development of children in Germany aged 3 to 10 years using data from the ‘German Health Interview and Examination Survey for Children and Adolescents’ (KiGGS) as well as the ‘German Child Panel’ (GCP). Matching methods are employed, and background characteristics are used to control for any selection effects. Results show that there is a positive effect of participation in sport on the measure ‘feeling better-off among friends’. Finally, Pawlowski, Schüttoff, Downward, and Lechner (2018) use panel data to identify causal effects between the sports group participation of children in Peru and different outcomes during childhood, including the ‘perceived support by friends in difficult times’, which indicates stronger relationships/friendships of children participating in sports groups.

SPORT AND SOCIAL NETWORK SUPPORT In general, it is supposed that the interpersonal activity and cooperation during sport creates respectful relations between individuals since sport is based on internationally recognized rules and concepts of fair play (Schwier, 1998). This might lead to a stronger sense of helpfulness towards others. Furthermore, sport is expected to facilitate the integration of individuals with different cultural and/or socio-economic backgrounds by simplifying interactions through the common and universal language of sport. In particular, participation in sports clubs only costs low membership fees, which enables individuals with low income to become involved. Especially among young people, sport is a popular leisure activity and therefore provides the opportunity for young immigrants to have frequent intercultural contacts with their peers.

59

Makarova and Herzog (2014) examine this based on information about 454 immigrant youths in Switzerland. They find a positive association of immigrants participating in sports clubs and the amount of personal contact with Swiss peers during sport. Such adolescents are also more likely to report having more intercultural contacts in their free time and among their close friends, which leads to a stronger feeling of being integrated. Gerlach and Brettschneider (2013) analyze the association of sports club participation of children and adolescents on social acceptance and respect by peers that can be seen as forms of social support. Based on a longitudinal study of 1,637 individuals in Germany, a significant positive association is established using variance analyses. More recently, Schüttoff, Pawlowski, Downward, and Lechner (2018) analyzed the effects of sports participation on social capital formation with a sample of 1,111 school-enrolled adolescents with German origin based on data coming from the socio-economic panel (SOEP). They also consider a measure of social network support, that is the frequency of helping out friends, neighbors, or relatives. A matching estimator is employed, and the panel structure of the data is exploited to consider possible endogeneity. Results show that regular sports participation positively impacts the willingness to help.

SPORT AND CIVIC ENGAGEMENT Civic engagement comprises those actions or behavior that positively contributes to the community. This includes general associational involvement in citizens, religious, or sports groups, political participation, and volunteering. The latter is the most often used measure of civic engagement (Scrivens & Smith, 2013). Volunteering is also described as prosocial behavior of individuals, who are motivated by altruism, that increases the well-being of other members of the community (Downward & Rasciute, 2011). In particular, being a member of a sports club often entails volunteer activities (Becker & Häring, 2012), such as participating at general meetings, accepting an office at the club, or helping out at competitions or (sports) festivals organized by the club. In general, civic engagement constitutes an important activity for individuals. In particular, engagement during adolescence is of importance since personal identity is shaped, and adolescents gain independence from the parents, while peers and other institutions have a greater influence on them in this stage of life (Frisco, Muller, &

60

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Dodson, 2004). Adolescents are more likely to adopt the values and morals represented in an organization (Eccles & Barber, 1999) which fosters, among other things, prosocial behavior. Frisco et  al. (2004) explore whether membership in different youth-serving organizations, such as religious youth groups, non-school team sports, or boys’ and girls’ clubs, is related to voting behavior in early adulthood based on a sample of 10,839 US American adolescents form the National Educational Longitudinal Study. They use a weighted logistic regression to predict voting behavior and find that engagement in non-school team sports is positively related to voting. Delaney and Keaney (2005) analyze several data sets (European Social Survey, Home Office Citizenship Survey, Time Usage Survey) and confirm this positive association of membership in a sports group and political engagement such as voting or signing a petition for UK citizens. In addition, Seippel (2006) uses the Norwegian data of the Johns Hopkins Comparative Nonprofit Sector Project. Regression results show that there is also a positive association of being a member of voluntary sports organizations and political voting. Positive associations between sport and civic involvement and volunteering are also found in the empirical literature. For instance, Perks (2007) analyzes survey data collected from a representative sample of Canadians (n = 12,387) and finds that youth sports participation is positively associated with adult involvement in community activities (in terms of informal volunteering). Moreover, in Schüttoff et al. (2018) a positive effect of regular sports participation during adolescence and civic involvement in a citizens’ group, political party, or local government is found to be independent of the type of organization and type of sport. With regard to voluntary work, a positive effect is only found for those adolescents organized in sports clubs. In contrast, the type of sport does not matter here.

SPORT, TRUST AND COOPERATIVE NORMS Trust and cooperative norms are cognitive factors which determine the behavior of individuals towards each other as members of society. People who freely cooperate with one another and trust in others bring benefits to all members of a community as they transcend personal relationships and are more open to interact with unknown individuals and groups (Scrivens & Smith, 2013). In general, the relation between sport, trust and

cooperative norms is dependent on the other dimensions of social capital, as Putnam (1995, p. 666) states that ‘civic connections and social trust move together.’ Consequently, it is suggested that sports participation creates personal relationships and networks that, in turn, generate trust within a community, which is the foundation for an active and engaged citizenship serving broader community interests (Perks, 2007). This assumption implies that the causation flows from participating to trust building. Brehm and Rahn (1997), however, argued that the more individuals participate in their communities, the more they build trust for each other and, consequently, the greater the level of trust that individuals hold for each other, the more likely they are to participate. Taking both arguments together, it remains unclear whether participation in sport or membership of sports organizations causes increased levels of trust in others, or whether higher levels of trust within communities make people more likely to participate in group activities like sport (Delaney & Keaney, 2005). This ambiguity in theory is also reflected in the empirical findings on sport and trust. For instance, Seippel (2006) finds a positive association of being a member of a voluntary sports organization on generalized trust, whereas the association is weaker here compared to voluntary organizations in general. Downward, Pawlowski, and Rasciute (2014) explore the impact of associational behavior (meeting with family or friends or participation in cultural, political, sport etc. associations) on trust in society for 30 countries, based on the data of the International Social Survey Programme. Possible endogeneity between trust and associational behavior is controlled for by using an instrumental variable approach. Estimation results show that participation in sports association might actually even reduce trust. Therefore, it is concluded that trust might underpin associational behavior, rather than be derived from it. Finally, Delaney and Keaney (2005) differentiate between two forms of trust. They find a positive association of membership in sports clubs and trust in civil institutions, such as trust in the police or in politicians, but the association with trust in other people (generalized trust) being insignificant. In a recent study, Bartolomeo and Papa (2019) carry out an experiment with 120 individuals to directly test for the first time the effects of pure physical activity on trust and trustworthiness. They compare the choices of individuals playing an investment game between those practicing 30 minutes of physical activity and those being involved in another task. Results show that individuals practicing physical activity are more likely to exhibit trust and prosocial behavior.

Sport and Social Capital Formation

CONCLUSION The objective of this chapter was to summarize both the theoretical background as well as the empirical findings on the role of sports participation for the accumulation of social capital. Overall, the few existing empirical studies looking at a more disaggregated level suggest that particularly voluntary sports organizations like sports clubs provide a platform where various forms of social capital can evolve, such as building new relationships (e.g., Ulseth, 2004), having more contact to peers (e.g., Felfe et  al., 2016), integrating immigrants (e.g., Makarova & Herzog, 2014), fostering civic engagement (e.g., Schüttoff et al., 2018; Seippel, 2006), or building trust (e.g., Seippel, 2006). With regard to the type of sport, positive associations with social capital are mainly found for practicing a team sport. Most notably, practicing sport in a team or a group (versus individual sport) fosters engagement in civic groups (e.g., Schüttoff et  al., 2018) and increases the likelihood of exhibiting civic engagement in terms of voting (e.g., Frisco et al., 2004). However, results also show that practicing a team sport reduces the likelihood of participating in other social settings, because of the high commitment and time spent for sport (e.g., Becker & Häring, 2012). In general, these findings suggest that the positive association between sports participation and social capital are grounded in the associational nature of sport, such as playing in a team or belonging to a club. In this case, however, sport can also lead to inequalities and social exclusion because of the strong bonds that exist in a club or a team, which is homogenous in its members. Here, the mechanism of bonding social capital applies where the strong bonds are beneficial to in-group members but negative for out-group members (Marlier et  al., 2015). This might generally limit the contribution of sport to social cohesion and integration. Collins (2004) even argues that sports participation is rather exclusionary since participation rates in sport decline with lower socio-economic status. However, there is some evidence that the mechanism of bridging does also apply in the sport environment. In particular, sports clubs are venues where the integration of individuals with migrant background are facilitated (e.g., Makarova & Herzog, 2014). Finally, and in contrast to the presumed assumption that being a member of a sports organization or a sports team where shared goals are pursued especially increases cooperation and also generates trust, Bartolomeo and Papa (2019) find that pure exercise can also generate trust and trustworthiness among individuals.

61

Although empirical evidence revealing causal links between sports participation and social capital has been growing recently, previous studies could not always clearly establish whether sport leads to the accumulation of social capital or whether individuals with a higher stock of social capital rather participate in sport (Theeboom, Schaillée, & Nols, 2012) due to a lack of appropriate data and/or empirical strategies. So far, only a few studies exist which arguably arrive at causal interpretations by using panel data and matching techniques (e.g., Pawlowski et al., 2018; Schüttoff et  al., 2018), an instrumental variable approach (e.g., Downward et al., 2014; Felfe et al., 2016) or experiments (Bartolomeo & Papa, 2019). Clearly, in order to provide further valuable findings for (sports) policy makers and sports officials, future research should increase efforts for establishing causal links. Furthermore, a relatively unexplored research field relates to the potential benefits of sport as a tool for development aid. While there are several ‘sport for development’ programs that use sport to increase the socialization of individuals, to foster social inclusion of the disadvantaged, or to facilitate intercultural exchange and conflict resolution (e.g., Schulenkorf, 2012), empirical evidence for the existence of such links based on large-scale survey data of carefully selected samples is largely missing. The only exception is Pawlowski et  al. (2018), although it remains unclear whether their findings on the positive effects of children’s sports group participation for social capital formation in Peru also hold in other less developed countries.

Note 1  An introduction into the public funding of sports is provided by Pawlowski and Thieme (2016). An overview on the public funding of sports in Germany is provided by Pawlowski and Breuer (2012, 2014).

REFERENCES Bartolomeo, G. D., & Papa, S. (2019). The effects of physical activity on social interactions: The case of trust and trustworthiness. Journal of Sports Economics, 20(1), 50–71. Baum, F. E., & Ziersch, A. M. (2003). Social capital. Journal of Epidemiology and Community Health, 57(5), 320–323. Becker, S., & Häring, A. (2012). Soziale Integration durch Sport? Eine empirische Analyse zum

62

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Zusammenhang von Sport und sozialer Integration [Social integration and sport? An analysis of the relationship between physical activity and social integration]. Sportwissenschaft, 42, 261–270. Bourdieu, P., & Wacquant, L. (1992). An invitation to reflexive sociology. Chicago, IL: University of Chicago Press. Brehm, J., & Rahn, W. (1997). Individual-level evidence for the causes and consequences of social capital. American Journal of Political Science, 41(3), 999–1023. Coleman, J. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94(Supplement), 95–120. Collins, M. (2004). Sport, physical activity and social exclusion. Journal of Sport Sciences, 22(8), 727–740. Delaney, L., & Keaney, E. (2005). Sport and social capital in the United Kingdom: Statistical evidence from national and international survey data. Dublin: Economic and Social Research Institute and Institute for Public Policy Research, 32. Downward, P., & Rasciute, S. (2011). Sport and social exclusion: An economic perspective. In R. Bailey & S. Dagkas (Eds.), Inclusion and exclusion through youth sport (pp. 40–56). London, UK: Routledge. Downward, P., Pawlowski, T., & Rasciute, S. (2014). Does associational behavior raise social capital? A cross-country analysis of trust. Eastern Economic Journal, 40(2), 150–165. Eccles, J., & Barber, B. (1999). Student council, volunteering, basketball, or marching band: What kind of extracurricular involvement matters? Journal of Adolescent Research, 14(1), 10–43. Felfe, C., Lechner, M., & Steinmayr, A. (2016). Sports and child development. PLoS ONE, 11(5), e0151729. Frisco, M. L., Muller, C., & Dodson, K. (2004). Participation in voluntary youth-serving associations and early adult voting behavior. Social Science Quarterly, 85(3), 660–676. Gerlach, E., & Brettschneider, W.-D. (2013). Aufwachsen im Sport: Befunde einer 10-jährigen Längsschnittstudie zwischen Kindheit und Adoleszenz [Growing up with sport: Findings of a 10-year longitudinal study between childhood and adolescence]. Aachen, Germany: Meyer & Meyer. Hoye, R., & Nicholson, M. (2008). Locating social capital in sport policy. In M. Nicholson & R. Hoye, Sport and social capital (pp. 69–92). Amsterdam and Boston, MA: Butterworth-Heinemann. Makarova, E., & Herzog, W. (2014). Sport as a means of immigrant youth integration: An empirical study of sports, intercultural relations, and immigrant youth integration in Switzerland. Sportwissenschaft, 44(1), 1–9. Marlier, M., Van Dyck, D., Cardon, G., De Bourdeaudhuij, I., Babiak, K., & Willem, A. (2015).

Interrelation of sport participation, physical activity, social capital and mental health in disadvantaged communities: A SEM-analysis. PLoS ONE, 10(10): e0140196. Pawlowski, T., & Breuer, C. (2012). Die finanzpolitische Bedeutung des Sports in Deutschland [Fiscal political importance of sport in Germany]. Wiesbaden: Springer-Gabler Research. Pawlowski, T., & Breuer, C. (2014). Sport und öffentliche Finanzen – Die sportbezogenen Einnahmen und Ausgaben öffentlicher Haushalte in Deutschland (Essentials) [Sport and public finances – Sports related revenues and expenditures in Germany]. Wiesbaden: Springer. Pawlowski, T., Schüttoff, U., Downward, P., & Lechner, M. (2018). Can sport really help to meet the Millennium Development Goals? Evidence from children in Peru. Journal of Sports Economics, 19(4), 498–521. Pawlowski, T., & Thieme, L. (2016). Sport und Staat [Sport and the state]. In C. Deutscher, G. Hovemann, T. Pawlowski, & L. Thieme (Eds.), Handbuch Sportökonomik [Handbook of sports economics] (pp. 315–328). Schorndorf: Hofmann. Perks, T. (2007). Does sport foster social capital? The contribution of sport to a lifestyle of community participation. Sociology of Sport Journal, 24(4), 378–401. Putnam, R. (1993). Making democracy work: Civic traditions in modern Italy. Princeton, NJ: Princeton University Press. Putnam, R. (1995). Tuning in, tuning out: The strange disappearance of social capital in America. Political Science and Politics, 28(4), 664–683. Putnam, R. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster. Schulenkorf, N. (2012). Sustainable community development through sport and events: A conceptual framework for Sport-for-Development projects. Sport Management Review, 15(1), 1–12. Schüttoff, U., Pawlowski, T., Downward, P., & Lechner, M. (2018). Sports participation and social capital formation during adolescence. Social Science Quarterly, 99(2), 683–698. Schwier, J. (1998). Chancengleichheit [Equal opportunity]. In O. Grupe & D. Mieth (Eds.), Lexikon der Ethik im Sport [Lexicon of sporting ethics] (pp. 80–84). Schorndorf: Hofmann. Scrivens, K., & Smith, C. (2013). Four interpretations of social capital: An agenda for measurement. OECD Statistics Working Papers, 2013/06, OECD Publishing. Seippel, Ø. (2006). Sport and social capital. Acta Sociologica, 49(2), 169–183. Theeboom, M., Schaillée, H., & Nols, Z. (2012). Social capital development among ethnic minorities in

Sport and Social Capital Formation

mixed and separate sport clubs. International Journal of Sport Policy and Politics, 4(1), 1–21. Ulseth, A.-L. B. (2004). Social integration in modern sport: Commercial fitness centres and voluntary sport clubs. European Sport Management Quarterly, 4(2), 95–115.

63

United Nation (2014). Why sport? Overview. Retrieved July 6, 2017, from www.un.org/sport/ content/why-sport/overview. Uslaner, E. (1999). Democracy and social capital. In M. Warren (Ed.), Democracy and trust (pp. 121– 150). Cambridge: Cambridge University Press.

7 Recent Evidence on the Effects of Physical Activity on Human Capital and Employment Carina Steckenleiter and Michael Lechner

INTRODUCTION There are many potential effects which are commonly attributed to sports participation. Some of these, such as health or subjective well-being, might be relevant for all stages in life. Other sets of potential effects are, however, tied more specifically to different age groups. Hence, for children and adolescents it is of interest whether and how sport is able to foster the accumulation of human capital. For people who have already entered or plan to enter the labour market, it is, in contrast, more relevant to assess whether and how sports participation impacts their labour market outcomes, which is the focus of this chapter. As a first step, it is crucial to examine why one could expect physical activity to have a positive influence on human capital formation, as this is one of the key drivers of future labour market success. In that respect, neuroscientists and researchers from related disciplines provide evidence on which ways physical activity can foster human capital accumulation. These studies find that physical activity induces processes which simplify neuroplasticity and, through that, in a second step cognitive functions (Hillman, Erickson, & Kramer, 2008; Hötting & Röder, 2013).1 Besides cognitive abilities, non-cognitive skills are a central component of human capital (see e.g. Heckman &

Rubinstein, 2001; or Heckman, Stixrud, & Urzua, 2006).2 It appears plausible that sports participation might foster the development of these skills, even though the exact mechanisms are not (yet) pinned down in the literature. In addition, there are of course also indirect ways through which sports participation influences human capital formation, such as health and fitness (see e.g. Lechner, 2009; Rooth, 2011). This chapter starts out summarizing the empirical literature on how sports participation impacts human capital formation in terms of cognitive as well as non-cognitive skills for children and adolescents. In the second part, the chapter will then provide an overview over the literature on labour market effects of sports participation.3 The chapter concludes by considering the future horizon for research in this area. Some challenges for obtaining reliable empirical results apply to both questions at hand. In every study, observed sports participation levels have to be related to the outcomes of interest. However, activity levels of people are very likely not random. Hence, it might be, for instance, the case that people who are very successful on the labour market and have high activity levels display similar observed or unobserved characteristics. They might be particularly driven or have a good family background.4 A key challenge for obtaining

RECENT EVIDENCE ON THE EFFECTS OF PHYSICAL ACTIVITY ON HUMAN CAPITAL

credible empirical results is hence taking care of these confounding variables and tackling the selection problem into sports through suitable empirical designs.5

HUMAN CAPITAL As well as being relevant for individuals, human capital accumulation during childhood and adolescence is a key policy goal of almost all societies. It is, however, a very expensive goal.6 In consequence, it is important to investigate the potential ‘side-effects’ of sports participation and explore if and how these impact the accumulation of human capital. If positive effects are observable, these ‘positive externalities’ of sports participation might offer a cost-efficient way of human capital enhancement. We divide human capital accumulation very broadly into non-cognitive and cognitive skills. Since this question is an active research area in various disciplines, such as epidemiology, sociology, sports science as well as economics, only a very small selection of studies can be presented with an emphasis on the economics literature.

COGNITIVE SKILLS AND EDUCATIONAL ATTAINMENT While there very likely exist quite substantial differences in how much leisure time children or adolescents devote to physical activity, every child or adolescent will take part in some form of physical education at school. Hence, schools are our starting point to examine the impact of physical activity on cognitive skills. In that respect, Lees and Hopkins (2013) provide an overview over eight randomized controlled trials. They document generally positive effects of sports participation on cognitive performance, although these are sometimes very small. Even if school hours are reallocated from normal lessons in classroom to sports, no negative effects are found. Resaland et al. (2016) conduct a randomized controlled trial in 57 schools in Norway to examine the effect of physical activity on academic performance measured by nation-wide standardized tests (reading, numeracy, English). The intervention increased physical activity by 165 minutes per week compared to the control group in the fifth grade. The study finds no overall effect on academic performance, although a significant effect on numeracy

65

for the children who initially displayed the poorest academic performance with respect to numeracy. Knaus, Lechner and Reimers (2017) find that an increase in sports hours in schools by one hour improves cognitive skills (measured by grades in math and German) with data for Germany. Singh, Uijtdewilligen, Twisk, Van Mechelen and Chinapaw (2012) report positive impacts of sports participation on cognition for school-aged children who are between 5 and 18 years old in their review article. Esteban-Cornejo, TejoroGonzalez, Sallis and Veiga (2015) look at adolescents between 13 and 18 years of age and differentiate between academic and cognitive performance in their review article of 20 studies. In both, performance dimensions associations, if found, were mostly positive. Moreover, while correlations with academic performance were found for different intensity levels of sports participation, significant associations with cognitive performance were only found for vigorous physical activity. Fedewa and Ahn (2011) assess 59 studies in a meta-analysis. Their findings suggest that programs with an emphasis on aerobic exercise lead to the largest positive effect on academic achievements and cognitive outcomes. They also indicate that, in particular, the very young school children and cognitively impaired children benefit strongly. Besides school sports, sport clubs are an important dimension where physical activity takes place in an organized form. In that respect, Felfe, Lechner and Steinmayr (2016) find positive effects of club sports on cognitive skills for children who are between three and 10 years old. Whereas most studies are conducted with data from developed countries, Pawlowski, Schüttoff, Downward and Lechner (2016) investigate the impact of participating in sports groups for children in Peru. They do not find significant effects of sports group participation on cognitive skills. There are several studies looking at high schools in the US. Lipscomb (2007) analyzes the impact of athletic and club participation and finds that athletic participation has a positive association with math and science test scores. Rees and Sabia (2010), on the other hand, find little impact of sports participation on grade point average and only a moderate association with educational aspirations. The last step in the transition from adolescence to adulthood is post-secondary education at a university.7 The question here is also whether being physically active has an impact on educational outcomes. Fricke, Lechner and Steinmayr (2017) try to answer this question with an experiment at a Swiss university. They randomize financial incentives to take part in on-campus sports activities.

66

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Their study finds that the incentives have a positive effect on grades of the students as well as lead to a rise in the frequency of weekly participation. Male students as well as students who have good grades are identified as the driving forces of their results.

NON-COGNITIVE SKILLS As for cognitive skills, the literature on the impact of physical activity on non-cognitive skills stems from various fields and is hence difficult to overview in its full range. Conventional wisdom attributes the integrative character as well as positive impacts on the development of social skills and character skills to physical activity. It is very interesting to investigate that link in the light of the fact that aspects of non-cognitive skills have been identified to be positively associated with economic outcomes and interactions.8 For example, from the Big Five model, which subsets personality into five different domains, the factor ‘conscientiousness’ has been found to have a strong correlation with labour market outcomes (Almlund, Duckworth, Heckman, & Kautz, 2011).9 Several studies in the area of sociology analyze the impacts of physical activity on non-cognitive skills. However, these studies stress the point that whether these potential positive influences of physical activity materialize in increased noncognitive skills is highly dependent on the context and circumstances (see, for instance, Gould & Carson, 2008; Coakley, 2011; Holt & Neely, 2011). Felfe et al. (2016) aim at identifying the effect of club sports participation on non-cognitive skills with German data. Children who take part in club sports have significantly less peer as well as emotional problems in their study. Their results for behavioural problems as well as antisocial behaviour point into the same direction, although these are not statistically significant. Fuchs and Osikominu (2016) analyze the impact of physical activity in sports clubs on skill formation with the German Socioeconomic Panel (GSOEP). Their analysis is carried out separately for young males and females on either the academic or vocational educational track. They investigate, among other things, the impact of physical activity on the Big Five personality traits noted earlier. For the sample of people in the vocational track, they find significant positive effects on conscientiousness and agreeableness for men. In addition, they find positive effects on openness, extraversion as well as conscientiousness for the pooled sample

of men and women. For adolescents in the academic track, significant positive effects are found for the pooled sample with respect to extraversion. Pawlowski et  al. (2016) investigate if participation in sports groups has an influence on different social capital measures. They find that children who take part in sports groups report significantly higher support by friends. No significant effects are found, however, for respect by other children. Knaus, Lechner and Reimers (2017) investigate the impact of one additional hour of school sports. They detect an interesting heterogeneity with respect to the impact on non-cognitive skills for girls and boys. Girls benefit from physical education by displaying significantly less emotional symptoms as well as peer relations problems. Results for boys, on the other hand, indicate a significant increase in peer problems. This is particularly interesting in the light of the fact that Felfe et al. (2016) find the exact opposite effects. Finally, besides overviewing studies on the link of muscular fitness and physiological benefits, Smith et al. (2014) provide an overview over six studies which investigate the correlation between muscular fitness and self-esteem. Five out of six studies report a significant positive impact of muscular fitness on self-esteem. The above-mentioned overview article by Lees and Hopkins (2013) also reports a positive impact of physical activity on self-esteem.

LABOUR MARKET Effects of Sports in Childhood and Adolescence on Labour Market Outcomes A strand of literature, which bridges the gap between childhood and adulthood, investigates the impact of physical activity during childhood and adolescence on labour market outcomes. Barron, Ewing, and Waddell (2000) use an instrumental variable approach and find that athletic participation has a significant positive impact on educational attainment as well as wages. Their instrument (health and family characteristics) is, however, problematic since it seems rather unlikely that health and family characteristics influence labour market outcomes only through physical activity. Eide and Ronan (2001) use height as an instrument for sports participation and find that engaging in sport activities during high school has a positive impact on earnings for black males, but not for other groups. It appears plausible that height impacts labour market

RECENT EVIDENCE ON THE EFFECTS OF PHYSICAL ACTIVITY ON HUMAN CAPITAL

outcomes not only through physical activity. If so, this would violate one of the assumptions of the instrumental variable approach. Case and Paxson (2008), for instance, show that height is positively correlated with cognitive ability and hence, in a second step, with labour market outcomes.10 Stevenson (2010) exploits the exogenous variation in sports participation induced through title IX in 1972 in the US. In consequence, she is able to apply a more credible identification strategy than the above two papers. Title IX legally prohibited that females were discriminated against in all institutions which received funding from the US government in the area of education. Necessary upward shifts in female participation in order to comply with the new law varied across states since this was particularly dependent on male pre-law participation. Stevenson (2010) exploits this variation and finds that physical activity during high school has a positive impact on employment as well as college attendance. Rooth (2011) investigates if physical fitness has a positive impact on earnings. He measures fitness information as well as a large set of individual characteristics at the age of 18 for almost all Swedish men. He is able to do so by using Swedish military enlistment data. These data are then merged with data on earnings in the age range of 28 to 38 years old. The study finds that a rise in the fitness level by one standard deviation leads to an increase in annual earnings by 4% when sibling fixed effects are included. Through inclusion of sibling fixed effects, Rooth (2011) aims at netting out unobserved confounders which are related to family background. The effect is getting smaller, if he additionally accounts for non-cognitive skills. Cabane and Clark (2015) use the Add Health data set from the US and include family fixed effects in their analysis, like Rooth (2011). They find that childhood sports participation is positively correlated with labour market outcomes 13 years later. Finally, Pfeifer and Cornelissen (2010) investigate the influence of extracurricular sports participation on professional as well as secondary degrees with data from Germany. Positive and statistically significant results are found for both outcomes of interest.

Labour Market Outcomes Human capital is a key factor for labour market success, and it is attributed to physical activity to impact the formation of human and social capital favourably through various channels. The literature commonly states enhanced health and wellbeing as intermediate factors. In addition, network

67

effects, enhancing cognitive as well as the development of non-cognitive skills, such as teamwork, self-discipline and endurance, are put forward as potential channels (for example, see Cabane & Lechner, 2015, for an overview). In the last couple of years, a growing literature has started to analyze the labour market effects of sports participation. Drawing causal conclusions was, however, often hampered by the fact that detailed information on sports participation is rarely available. Suitable information with respect to the intensity of exercising is also often lacking. In addition, is it rarely known in which types of sports people participate. This is quite unfortunate given that one would expect differential impacts of different sports when thinking about the channels mentioned above. Within the framework of a field experiment in Sweden, Rooth (2011) investigates call-back rates of fake applications which were sent out in response to real-life job openings. Over 8,000 fake applications were sent out for different occupations and with varying requirements in terms of skills. Contacted employers were also active in different sectors of the economy. All job applications contained information on sports participation where intensity of exercising as well as types of sports were randomly varied between applications. The study finds that an indication of the sporting activities undertaken in the application increases call-back rates by two percentage points. The effect sizes are not large, but largest for football and golf. Rooth (2011) acknowledges that it is not possible to distinguish whether this is due to the preferences of the recruiter or indeed the types of sports. A non-experimental study which is able to provide insights about effects of different types of sports is the study by Lechner and Downward (2017). They use the Active People Survey which entails extensive sports information, but unfortunately only rather coarse labour market information. One additional drawback is the crosssectional nature of the data set. The study finds that sports participation is positively correlated with earnings. Also, there is a negative relationship with unemployment for males. Lechner and Downward (2017) find that team sports are associated more with higher employability (though varying across age and gender), whereas outdoor sports as well as fitness show the strongest correlation with earnings (for men). Cabane (2014) analyzes how sports participation impacts unemployment duration using the German Socio-Economic Panel (GSOEP). The study reports positive correlations between being active at least once a week and transitioning from being unemployed to being employed

68

THE SAGE HANDBOOK OF SPORTS ECONOMICS

for the group of women who have at least three years of professional experience. However, the study points out that this might rather be due to lower psychological barriers instead of an effect of sports participation. Several studies analyze how exercising materializes in the short and long run in terms of earnings. Cornelissen and Pfeifer (2008) use a random effects model and data from the GSOEP. They find that men who exercise at least once a week have 5% higher earnings compared to men who don’t. In addition, the study investigates the influence of sport participation during adolescence on earnings. They report that being physically active at the age of 15 goes along with 6% higher earnings for women. Lechner (2009) uses the same data set while applying a quite different approach. He combines semi-parametric matching methods with exploiting the panel structure of the GSOEP to control for unobservable confounders which are time constant. The study concludes that being physically active on a regular basis leads to an average increase of €1,200 in earnings per year over a time horizon of 16 years compared to doing no sports or doing sports only sporadically. Lechner and Downward (2017) also provide some estimates how sports participation impacts income. They come to the conclusion that the effect of physical activity on annual household income is around €4,900–€7,400 per year for men in the age group of 26–45 depending on the type of sport. For women in the same age group, the effects amount to €3,900–€6,000.11 Hyytinen and Lahtonen (2013) use three cohorts of a Finnish data set on identical and non-identical male twins to analyze differences in earning between twins who have different activity levels. By doing so they control for unobservable genetic as well as parental background factors. The twin data are matched with a panel data set containing labour market information of working age individuals in Finland. The study finds that going from occasionally to regularly exercising correlates with a 14–17% increase in long-term income. Finally, Lechner and Sari (2015) use a very similar approach as Lechner (2009). However, in using the Canadian National Population Health Survey (NPHS), they have a much richer data set in terms of health indicators as well as physical activity information at hand. This allows investigation of the dose–response relationships. The study finds that going from being completely inactive to a moderate level of sports activities (where moderate translates approximately into 30 minutes per day at five days per week) does not have a significant effect on earnings. Going from moderate

to active sports participation, on the other hand, leads to earning increases between 10% and 20% in an 8–12-year timeframe. Overall, the studies indicate that there is strong evidence for positive labour market effects of physical activity. This is particularly the case for earnings.

CONCLUSION AND FUTURE RESEARCH This chapter provides a brief overview of the effects of sports participation on human capital accumulation and labour market outcomes. Even though each study individually has some weaknesses, they provide clear evidence for positive impacts of physical activity on educational and labour market outcomes.12 However, while results are distinct, there are still many issues which need further investigation. The first one is that the channels through which sports operates and interact are not well understood. Secondly, to be better able to draw causal conclusions for the question of interest, better and more credible identification strategies (i.e. a more credible research designs) are needed. This was in the past often limited by data availability. Panel data sets which provide extensive information with respect to type, aptitude and intensity of sports participation as well as labour market outcomes were scarce in the past. Having richness with respect to sports information is important, because effects are likely to vary across different kinds of sports and also intensity. Very little is known about optimal levels of exercising and how these levels might be different for people with different levels of education or preferences, and so on. Here, also time constraints of people come into play. From the perspective of the society as a whole, substantial amounts of money are spent on subsidizing leisure sports participation in numerous countries. However, even though a lot of money is spent in that area, rigorous evaluations of the effects of these public expenditures are scarce so far. While more and more studies have emerged in the last couple of years analyzing the effects of sports participation for the average person, it would be important to understand if and how vulnerable groups of the societies, such as unemployed, migrants and prison inmates, may benefit from physical activity.13 From a policy perspective, it would be important to thoroughly investigate the integrative potential which is attributed to sports. This is important, since this could provide a supplement of currently implemented measures if positive effects were found. Finally, it is

RECENT EVIDENCE ON THE EFFECTS OF PHYSICAL ACTIVITY ON HUMAN CAPITAL

worthwhile mentioning the (future) importance of big data for this research field. Not only do fitness apps and activity trackers in all sorts of forms offer new possibilities for interesting new empirical strategies, but they also provide the possibility of obtaining a lot more accurate data, since selfreported activity levels are most likely prone to heuristics as well as social desirability bias.

Notes   1  Neuroplasticity is the capability of reorganization by the nervous system to cope with new environment and tasks (Bavelier & Neville, 2002).   2  Non-cognitive abilities are a set of skills which has been found very difficult to measure in the past. Also, the full range of the concept is hard to define. For example, Kautz et al. (2014, p. 9) wrote that non-cognitive abilities comprise skills such as ‘perseverance (‘grit’), conscientiousness, self-control, trust, attentiveness, self-esteem and self-efficacy, resilience to adversity, openness to experience, empathy’.   3  For a survey article (in German) with a similar focus as this one, which in addition discusses effects on health and life satisfaction, see Lechner, Knaus and Cabane (2016).   4  There is a large body of literature which documents that physically active people are different from people who do not engage in sports participation in several dimensions (for instance, see Farrell & Shields, 2002; Schneider & Becker, 2005).   5  For a more detailed discussion on this, see for instance Lechner (2016).   6  When taking public spending on education as a coarse proxy for investment in human capital, one finds that OECD countries spent on average 3.4% of GDP on education in 2013. These numbers include primary to non-tertiary education and go up to an average spending of 4.8% of GDP if tertiary education is also included (OECD, 2017).   7  There are some studies which document a positive impact of college sports on earnings in some sectors for the US (e.g. Henderson, Olbrecht & Polachek, 2006; Caudill & Long, 2010). However, it is important to stress that US college sports are not comparable to university sports in other countries in terms of competitiveness and level or to on-campus recreational sports every student can engage in.   8  For an excellent overview over the relationship of personality traits and, among other things, labour market performance, see for instance Almlund et al. (2011).   9  The five factors are neuroticism, agreeableness, extraversion, conscientiousness and openness to experience.

69

10  Lundborg, Nystedt and Rooth (2014) find that about half of the correlation between height and earning can be traced back to taller people having a good family background and better test scores with respect to cognition. In addition, they find that non-cognitive skills accounts for around 20% of the premium of being tall. 11  Results in the paper are given in Pound sterling and were converted from Pound sterling to Euro for easier comparison with other studies. 12  See also the discussion on the effects of physical activity on labour market outcomes in Cabane and Lechner (2015). 13  Lange, Pfeiffer and van den Berg (2017) provide some preliminary results from an experiment where refugees were randomly assigned to an inclusive soccer project.

REFERENCES Almlund, M., Duckworth, A. L., Heckman, J. J., & Kautz, T. D. (2011). Personality psychology and economics (No. w16822). Cambridge, MA: National Bureau of Economic Research. Barron, J. M., Ewing, B. T., & Waddell, G. R. (2000). The effects of high school athletic participation on education and labor market outcomes. Review of Economics and Statistics, 82(3), 409–421. Bavelier, D., & Neville, H. J. (2002). Cross-modal plasticity: where and how? Nature Reviews Neuroscience, 3(6), 443–452. Cabane, C. (2014). Unemployment duration and sport participation. International Journal of Sport Finance, 9(3), 261. Cabane, C., & Clark, A. E. (2015). Childhood sporting activities and adult labour-market outcomes. Annals of Economics and Statistics/Annales d’Économie et de Statistique, 119–120, 123–148. Cabane, C., & Lechner, M. (2015). Physical activity of adults: A survey of correlates, determinants, and effects. Jahrbücher für Nationalökonomie und Statistik, 235(4–5), 376–402. Case, A., & Paxson, C. (2008). Stature and status: Height, ability, and labor market outcomes. Journal of Political Economy, 116(3), 499–532. Caudill, S. B., & Long, J. E. (2010). Do former athletes make better managers? Evidence from a partially adaptive grouped-data regression model. Empirical Economics, 39(1), 275–290. Coakley, J. (2011). Youth sports: What counts as ‘positive development’? Journal of Sport and Social Issues, 35(3), 306–324. Cornelissen, T., & Pfeifer, C. (2008). Sport und Arbeitseinkommen – Individuelle Ertragsraten von Sportaktivitäten in Deutschland. Jahrbuch für Wirtschaftswissenschaften, 59(3), 244–255.

70

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Eide, E. R., & Ronan, N. (2001). Is participation in high school athletics an investment or a consumption good? Evidence from high school and beyond. Economics of Education Review, 20(5), 431–442. Esteban-Cornejo, I., Tejero-Gonzalez, C. M., Sallis, J. F., & Veiga, O. L. (2015). Physical activity and cognition in adolescents: A systematic review. Journal of Science and Medicine in Sport, 18(5), 534–539. Farrell, L., & Shields, M. A. (2002). Investigating the economic and demographic determinants of sporting participation in England. Journal of the Royal Statistical Society: Series A (Statistics in Society), 165(2), 335–348. Fedewa, A. L., & Ahn, S. (2011). The effects of physical activity and physical fitness on children’s achievement and cognitive outcomes: A metaanalysis. Research Quarterly for Exercise and Sport, 82(3), 521–535. Felfe, C., Lechner, M., & Steinmayr, A. (2016). Sports and child development. PLoS One, 11(5), e0151729. Fricke, H., Lechner, M., & Steinmayr, A. (2017). The effect of physical activity on student performance in college: An experimental evaluation (CEPR Discussion Paper No. DP12052). Fuchs, B., & Osikominu, A. (2016). Quality leisure time and youth development (CEPR Discussion Paper No. DP11330). Gould, D., & Carson, S. (2008). Life skills development through sport: Current status and future directions. International Review of Sport and Exercise Psychology, 1(1), 58–78. Heckman, J. J., & Rubinstein, Y. (2001). The importance of noncognitive skills: Lessons from the GED testing program. The American Economic Review, 91(2), 145–149. Heckman, J. J., Stixrud, J., & Urzua, S. (2006). The effects of cognitive and noncognitive abilities on labor market outcomes and social behavior. Journal of Labor Economics, 24(3), 411–482. Henderson, D. J., Olbrecht, A., & Polachek, S. W. (2006). Do former college athletes earn more at work? A nonparametric assessment. Journal of Human Resources, 41(3), 558–577. Hillman, C. H., Erickson, K. I., & Kramer, A. F. (2008). Be smart, exercise your heart: Exercise effects on brain and cognition. Nature Reviews Neuroscience, 9(1), 58–65. Holt, N. L., & Neely, K. C. (2011). Positive youth development through sport: A review. Revista Iberoamericana de Psicología Del Ejercicio Y El Deporte, 6(2), 299–316. Hötting, K., & Röder, B. (2013). Beneficial effects of physical exercise on neuroplasticity and cognition. Neuroscience & Biobehavioral Reviews, 37(9), 2243–2257. Hyytinen, A., & Lahtonen, J. (2013). The effect of physical activity on long-term income. Social Science & Medicine, 96, 129–137.

Kautz, T., Heckman, J. J., Diris, R., Ter Weel, B., & Borghans, L. (2014). Fostering and measuring skills: Improving cognitive and non-cognitive skills to promote lifetime success. OECD Education Working Papers, No. 110. Paris: OECD Publishing. Knaus, M. C., Lechner, M., & Reimers, A. (2017). The effects of physical education on child development. Mimeo. Lange, M., Pfeiffer, F., & van den Berg, G. J. (2017). Integrating young male refugees: Initial evidence from an Inclusive Soccer Project (ZEW – Centre for European Economic Research Discussion Paper No. 17–016). Lechner, M. (2009). Long-run labour market and health effects of individual sports activities. Journal of Health Economics, 28(4), 839–854. Lechner, M. (2016). Empirical evidence on educational effects of physical activity: Four examples (No. 1619). St Gallen, Switzerland: University of St Gallen, School of Economics and Political Science. Lechner, M., & Downward, P. (2017). Heterogeneous sports participation and labour market outcomes in England. Applied Economics, 49(4), 335–348. Lechner, M., Knaus, M., & Cabane, C. (2016). Nachfrage II – Effekte durch das individuelle Sporttreiben. Handbuch der Sportökonomie. Beiträge zur Lehre und Forschung im Sport, 190, 279–298. Lechner, M., & Sari, N. (2015). Labor market effects of sports and exercise: Evidence from Canadian panel data. Labour Economics, 35, 1–15. Lees, C., & Hopkins, J. (2013). Effect of aerobic exercise on cognition, academic achievement, and psychosocial function in children: A systematic review of randomized control trials. Preventing Chronic Disease, 10, 1–8. Lipscomb, S. (2007). Secondary school extracurricular involvement and academic achievement: A fixed effects approach. Economics of Education Review, 26(4), 463–472. Lundborg, P., Nystedt, P., & Rooth, D. O. (2014). Height and earnings: The role of cognitive and noncognitive skills. Journal of Human Resources, 49(1), 141–166. OECD (2017). Public spending on education (indicator). Paris: OECD Publishing. doi: 10.1787/ f99b45d0-en (accessed on 25 September 2017). Pawlowski, T., Schüttoff, U., Downward, P., & Lechner, M. (2016). Can sport really help to meet the millennium development goals? Evidence from children in Peru. Journal of Sports Economics, 2016, 1–24. Pfeifer, C., & Cornelissen, T. (2010). The impact of participation in sports on educational attainment: New evidence from Germany. Economics of Education Review, 29(1), 94–103. Rees, D. I., & Sabia, J. J. (2010). Sports participation and academic performance: Evidence from the National Longitudinal Study of Adolescent Health. Economics of Education Review, 29(5), 751–759.

RECENT EVIDENCE ON THE EFFECTS OF PHYSICAL ACTIVITY ON HUMAN CAPITAL

Resaland, G. K., Aadland, E., Moe, V. F., Aadland, K. N., Skrede, T., Stavnsbo, M., & Kvalheim, O. M. (2016). Effects of physical activity on schoolchildren’s academic performance: The Active Smarter Kids (ASK) cluster-randomized controlled trial. Preventive Medicine, 91, 322–328. Rooth, D. O. (2011). Work out or out of work: The labor market return to physical fitness and leisure sports activities. Labour Economics, 18(3), 399–409. Schneider, S., & Becker, S. (2005). Prevalence of physical activity among the working population and correlation with work-related factors: Results from the first German National Health Survey. Journal of Occupational Health, 47(5), 414–423.

71

Singh, A., Uijtdewilligen, L., Twisk, J. W., Van Mechelen, W., & Chinapaw, M. J. (2012). Physical activity and performance at school: A systematic review of the literature including a methodological quality assessment. Archives of Pediatrics & Adolescent Medicine, 166(1), 49–55. Smith, J. J., Eather, N., Morgan, P. J., Plotnikoff, R. C., Faigenbaum, A. D., & Lubans, D. R. (2014). The health benefits of muscular fitness for children and adolescents: A systematic review and meta-analysis. Sports Medicine, 44(9), 1209–1223. Stevenson, B. (2010). Beyond the classroom: Using Title IX to measure the return to high school sports. The Review of Economics and Statistics, 92(2), 284–301.

8 Private Household Consumption in Sport Fernando Lera-López

INTRODUCTION In only 40 years there has been an incredible change in the social and economic importance of sport. Up until the 1970s, sport was predominantly a local activity, led by the voluntary sector (e.g., sports clubs). In the early 1970s, this situation began to change as the result from a variety of factors. For example, initial public investment in new sporting facilities led to an increase in sporting opportunities for many people during the 1970s and 1980s. In the 1990s, due to a period of stagnation in public expenditure, a shift from public to private investments took place. The private sector began to play a more prominent role. At the same time, the economic development offered larger sections of the population the financial resources to access sporting facilities (Lera-López & Rapún-Gárate, 2007). In recent years, the concern for well-being has increased as well as the awareness of the positive effects of sport and physical activity in health outcomes. Along with the increased participation in sport the concept of sport changed. Initially the concept was limited to organized and competitive practice only, but it has increased by including unorganized, non-competitive, recreational sporting activity, with new activities and settings such as sport in nature, multisport adventure, and sport tourism.

While the concept of sport broadened, and an increasing part of the population has become more involved in it, the economic importance of sport has expanded astoundingly as it became a growing economic sector and a global market, with a total revenue of $145.34 billion in 2015 (Statista, 2016). Among the different agents in the sports market, the consumer expenditure on sport made by households are of particular relevance. In the next section, the economic relevance of sports consumption made by households is highlighted by presenting relevant figures. The following section presents theoretical frameworks describing sports consumption made by households. The next two sections, respectively, collect the most important empirical evidence on consumption in sports participation and sports attendance. The chapter ends with some ideas for further research in terms of sports consumption by households.

ECONOMIC IMPORTANCE OF SPORTS-RELATED EXPENDITURES BY HOUSEHOLDS Different studies, particularly in Europe, have tried to quantify the contribution of sport as a

73

Private Household Consumption in Sport

Table 8.1  Economic importance of consumption in sport made by households Country (year)

Source

Consumption in % of total consumer expenditure

Australia (2009–10) Austria (2005) Belgium (1990) China (2012) Cyprus (2004) Denmark (1990) Finland (1990) Flanders (1996) Germany (2008) Greece (1994) Hungary (1990) Ireland (2008) Netherlands, the (2006) Poland (2006) Portugal (1990) Slovenia (2005) Spain (2006) Sweden (1990) United Kingdom (2015)

Australia Bureau of Statistics (2013) European Commission (2012, 2013) Andreff, Bourg, Halba, and Nys (1995) Jin (2014) European Commission (2012, 2013) Andreff et al. (1995) Andreff et al. (1995) Taks and Késenne (2000) Ahlert (2013) Kolimpalis (1999) Andreff et al. (1995) Indecon International Economic Consultants (2010) European Commission (2012, 2013) European Commission (2012, 2013) Andreff et al. (1995) Bednarik et al. (2010) Aguirre et al. (2008) Andreff et al. (1995) Department for Culture, Media & Sport (2016)

1.50 3.60 1.16 3.03 3.70 0.31 0.75 6.80 6.60 1.10 0.28 2.00 2.70 1.20 1.04–1.90 2.88 3.50 0.48 3.00(1)

(1)

Data for 2008.

‘sector’ in terms of the consumption made by households. Table 8.1 summarizes some of the most important studies, with figures ranging from 1.2% of total consumer expenditure in Poland, to 3.6% in Austria and 3.7% in Cyprus, up to 6.6% in Germany. Due to a lack of information provided by the national institutes of statistics in many countries, since the 1990s several studies were conducted to quantify the different categories of sports-related household expenditures based on consumer surveys (e.g., Aguirre, Lera-López, & Rapún-Gárate, 2008; Taks, Renson, & Vanreusel, 1999). These studies normally differentiate between active sports participation (clothing and shoes, club membership fees, lesson fees, sport material, etc.) and passive sports participation (attending sporting events, watching sport TV programs, gambling on sporting outcomes, etc.). Aguirre et al. (2008) estimated, for example, that in Spain 75% of this consumption is related to active sports participation, the remaining 25% are associated with passive sports participation. Similar results are shown in other European countries. For the UK, active participation is estimated to account for around 63% of the total sport consumption, with the remaining 37% being associated with passive participation (Sport Industry Research

Centre, 2008). In Germany, the relevance of active sports participation is higher, with estimations around 82% of total sports consumption (Preuß & Alfs, 2013). Nevertheless, a comparative analysis of the empirical estimation on the amount of money spent for sport by households highlights the significant differences among studies, ranging from EUR 100 per year to EUR 1,600 per year (Wicker, Breuer, & Pawlowski, 2010). There are several reasons for these differences. First, there is not a homogeneous definition of sport in the empirical studies, and there are large differences in terms of expenditure among sport (Taks et al., 1999; Wicker et al., 2010). Second, there is no common and general definition of the term ‘sports expenditure’. As Wicker et al. (2010) highlight, there is a general agreement that concepts such as membership and entrance fees, equipment, clothing, and training fees are part of sports expenditure. However, not all studies consider the travel costs or the medical costs. Third, the population under examination varies between studies. Some studies consider the total household expenditure (Desbordes, Ohl, & Tribou, 1999), or individual expenditures (LeraLópez & Rapún-Gárate, 2005a, 2005b, 2007; Taks et al., 1999), and others consider only sports club members (Wicker et al., 2010).

74

THE SAGE HANDBOOK OF SPORTS ECONOMICS

THEORETICAL APPROACHES AND METHODOLOGICAL ISSUES FOR SPORTS CONSUMPTION Sport is a special type of good because its consumption generates externalities. Sport does not only produce private benefits for participants, but also social benefits (e.g., economic development, local identity and prestige, reductions in antisocial behavior), and thereby justifies government intervention in the presence of market failure (e.g., Gratton & Taylor, 2000). From an economic perspective, sport has different good-characteristics. First, it is a non-durable consumer good, in that the benefit to the consumer is generated at the time of consumption. However, since participation in sport can generate benefits that are not immediate, it is also a durable consumer good, with benefits that accrue over time. Finally, sport may have some of the characteristics of a capital good. If sport makes a person fitter and healthier, then this improved health status may lead to a payoff in terms of increased productivity in the labor market and higher labor market income. Furthermore, the sports market consists of the sports goods sector, that includes all pro­ ducts bought for use in sport (clothes, shoes, etc.), and the sports services sector, including entrance charges, spectator sport, sport TV and video, and gambling on sport. In this context, sports demand can be described as a composite demand, involving the demand to take part in sport, and the demand for equipment, shoes and clothing, sport facilities, memberships, sports book and newspapers, TV-sports programs, and sports events. To deal with this complexity in sports demand, according to Gratton and Taylor (2000), and Downward, Dawson, and Dejonghe (2009), two different theoretical approaches are possible. Neoclassical approaches employ a rationalchoice framework to model individual sports decisions and are focused on the classical economic demand theory (Thibaut, Vos, & Scheerder, 2014). This is built upon the idea that individuals maximize their utility subject to certain constraints, mainly relating to budget and time. Implicitly, sport uses ‘non-economic’ and ‘nonobligated’ time, the demand for which reflects the trade-off between the utility derived from consumption of goods and the opportunity cost of an hour of sport. An example of this approach is the household production theory by Becker (1965). This theory integrates time allocation into the consumption decisions faced by households. Thus, monetary and time restrictions limit the opportunities of an individual’s sport activity (Wicker et  al., 2010). As Thibaut et  al. (2014)

argue, families can invest in sport in two different ways: by directly consuming sports goods, or by acquiring human capital. This individual human capital (in terms of higher education levels) could increase sports competencies, and thus extends the possibilities of sports participation. In conclusion, Becker’s theory assumes that household production depends on the income and time resources and the level of human capital. This theory is very popular in explaining expenditure on sports participation in empirical studies (e.g., Downward & Riordan, 2007; Wicker et al., 2010). Heterodox economic theories, in contrast, consider a wider set of theoretical principles, such as economic, sociological, and psychological approaches (Downward & Riordan, 2007). The first perspective, the post-Keynesian approach, emphasizes that individual behavior is linked to broader aspects of social behavior, such as the importance of social values, and that the consumption of sport involves learning-by-doing habits and spillover effects, maintaining that the preferences of the agent are endogenous (Lavoie, 2004). The sociological analysis of sports participation explains sports decisions in terms of concrete social situations and the construction of identities by individuals in choice situations (Scraton & Watson, 1999). According to this theory, sporting styles and individual preferences are linked not only to individual feelings, but also to social pressure and the influence of habitus. This perspective highlights the role played by (social) class (Bourdieu, 1984). In his study, Bourdieu presents two different factors to explain different tastes in sport: economic capital (income) and cultural capital (education). With regard to cultural capital, people from different social classes develop different tastes for sports products through the influence of habitus (Bourdieu, 1984). Furthermore, the life cycle plays a role. It is evident that needs, and thus sports consumption patterns, differ throughout the different stages of life. Finally, the psychological approach argues that the individual’s preferences and tastes are not fixed; they therefore evolve and change over the individual’s lifespan. Sensation-seeking, awakening, concern, pleasure, or anxiety can be potential sources of demand for sport and other leisure activities (Downward & Riordan, 2007). Thus, individual work-versus-leisure choices are based on interdependent individual preferences and motives that change over the life span due to situational influences in the personal environment. As a consequence, this approach focuses on constraints on sports decisions at the individual level, minimizing the role played by social constraints. Researchers working within the sociological and psychological frameworks have developed

Private Household Consumption in Sport

different approaches to emphasize the role of factors and constraints outside the individual that may affect sports behavior, in particular in order to explain sports attendance. For example, ecological models consider that a sports decision is mediated by individual intrapersonal factors (contextual factors such as family, friends, and peers), as well as cultural and environmental influences (Beaton, Funk, & Alexandris, 2009). In this regard, the Psychological Continuum Model, also known as the Theory of Participation is of particular relevance. The central tenets of this theory are that the psychological connection between an individual and his/her participation in sport or a sporting event can be identified as belonging to one of four different stages of motivation: awareness, attraction, attachment, and allegiance (Beaton & Funk, 2008). Another framework, the Self-Determination Theory, posits that subjective experience – and individual interpretation of it – determines attendance behavior. In these frameworks, different authors developed some motivation scales for sport consumption in sporting events (see Ha, Ha, and Han (2013) for a literature review) in order to develop multidimensional models which consider a multiplicity of factors, particularly social and psychological factors and behavioral (game) related factors, to explain the behavior of consumers of sporting events and fans’ motivation (Casper & Menefee, 2010; Stander & Zyl, 2016). Other authors consider the importance of perceived satisfaction of sporting events and team identification to explain future event participation intention. Woo (2016) systematically classifies the different theories to explain sports consumption in sporting events into six groups: identity theory, social identity theory, attitude theory, theory of planned behavior, decision making theory, and satisfaction theory. However, these theoretical approaches have been applied more usually to explain attendance behavior than consumption behavior. Finally, some authors emphasize the relevance of game characteristics such as the timing of an event, stadium and team characteristics, or weather conditions to influence sports consumption (Prior, O’Reilly, Mazanov, & Huybers, 2013). From a methodological approach, the majority of the studies use censored regression techniques to deal with zero observations, for example the standard Tobit model by Tobin (1958) (Eakins, 2016; Lera-López & Rapún-Gárate, 2005a, 2005b, 2007; Pawlowski & Breuer, 2012a; Thibaut, Eakins, Vos, & Scherder, 2017). This model considers households with zero levels of sports consumption, for example because they cannot afford the current prices, or they do not have enough income. Nevertheless, as Thibaut

75

et al. (2017) argue, the model does not represent individuals deliberately deciding to avoid sports consumption. Furthermore, it is unable to make a distinction between the decision of consumption (yes–no) and the amount of money spent for sport. Therefore, alternative models were developed, such as the two-step Heckman model, to deal with censored and categorical estimations of sports expenditure (Pawlowski & Breuer, 2011, 2012b; Scheerder, Vos, & Taks, 2011; Thibaut et  al., 2014) or the double hurdle model (Lera-López, Rapún-Gárate, & Suárez, 2011). However, hurdle models are difficult to estimate due to the necessity of developing a different set of variables for both decisions (consumption yes–no, and amount of money spent).

CORRELATES OF CONSUMPTION IN (ACTIVE) SPORTS PARTICIPATION All the above-explained theories are used to explain consumption that is associated with sports participation and usually follow the household production theory of Becker (e.g., Loyland & Ringstad, 2009; Thibaut et  al., 2014). Previous literature reviews on these correlates are presented by Breuer, Hallmann, Wicker, and Feiler (2010), Wicker et  al. (2010), Pawlowski and Breuer (2011, 2012b), and Thibaut et  al. (2017). The majority of studies consider individual expenditure on sports participation. Some studies analyze household expenditure (Eakins, 2016; Loyland & Ringstad, 2009; Pawlowski & Breuer, 2011; Thibaut et  al., 2014; Thibaut et  al., 2017). Furthermore, most studies consider sports consumption in sports participation in total while only few considered specific activities, such as triathlon (Wicker, Prinz, & Weimar, 2013), golf (Hallmann & Wicker, 2015), cycling (Thibaut, Vos, Lagae, Van Puyenbroeck, & Scheerder, 2016), and marathon events (Wicker, Hallmann, & Zhang, 2012), or specific categories such as entrance and membership fees for swimming pools and fitness centers (Pawlowski & Breuer, 2011, 2012b), sports clubs subscriptions and leisure classes fees (Eakins, 2016), and sports clothing and shoes (Scheerder, Vos, & Taks, 2011). Finally, only a few studies make a distinction between the decision to spend and the level of expenditure, with significant differences between both decisions (Thibaut et al., 2014). In spite of the difficulties explained in the previous section on a comparative analysis of sports expenditure, general tendencies can be derived from the study results. The early studies of the

76

THE SAGE HANDBOOK OF SPORTS ECONOMICS

1980s and 1990s primarily pay attention to gender, age, and income variables (e.g., Lamb, Asturias, Roberts, & Brodie 1992; Michon, Ohl, & Faber, 1987). In general, men spend more than women on sports participation (Eakins, 2016; Lamb et al., 1992; Lera-López & Rapún-Gárate, 2005a, 2005b, 2007; Loyland & Ringstad, 2009; Taks, Renson, & Vanreusel, 1995), although in some studies it varies with the type of sports (Michon et al., 1987; Wicker et al., 2010). As Breuer et al. (2010) argue, the general pattern can be explained by the fact that women, in general, are less active in sport than men. There is no obvious relationship with respect to age. In some studies, the youngest age groups appear to be greater spenders (Lamb et al., 1992), while other studies show that age is inversely correlated, usually in a non-linear way (Lera-López & Rapún-Gárate, 2007; Loyland & Ringstad, 2009; Statistical Working Group (SWG), 1995), and some report a peak in the middle (LeraLópez & Rapún-Gárate, 2005a; Pawlowski & Breuer, 2011). In general, as age increases there is a decrease in sports expenditure associated with a significant decrease in sports participation (Breuer et  al., 2010), although there are differences depending on the type of sports consumption (Eakins, 2016). Following the household production theory, income has been traditionally considered as a significant determinant of both participation in and expenditures for sport. However, the evidence is not conclusive. While income is found to be positively related to sports expenditure (Bloom, Grant, & Watt, 2005; Lamb et  al., 1992; Lera-López & Rapún-Gárate, 2007; Pawlowski & Breuer, 2011; Taks et  al., 1995, 1999; Wicker et  al., 2010), it remains unclear whether this relationship is linear or not. In this context, some studies analyze the income sensitivity of sports demand. For example, Loyland and Ringstad (2009) show that the sports-related income elasticity decreases over time, being elastic at the beginning (value of 1.25) and inelastic in more recent years. The authors argue that sport, like other time-consuming leisure activities, ‘becomes more expensive in terms of lost income when the wage rate goes up’ (Loyland & Ringstad, 2009, p. 616). Consequently, the positive income effect is compensated by the negative effect of increasing wages that indicates the existence of Linder’s disease. Késenne and Butzen (1987), for Flanders, find an income elasticity of 0.57, although with differences among different sports. Pawlowski and Breuer (2012a), for Germany, show the sensitivity of results based on the estimator used: a luxury good when the Tobit model is applied, and a necessity good when a Heckman method is developed. Eakins

(2016), for Ireland, estimates elasticities greater than 1. Finally, Thibaut et al. (2017) find a positive income elasticity of sporting activities (0.69), although with significant differences among the 23 sporting activities considered. This evidence suggests that the amounts of money spent on sports participation seem to differ depending on the type of activity (Taks, Vanreusel, & Renson, 1994; Wicker et al., 2010). As suggested by the household production theory, human capital, represented by education or occupation, is included in many studies. In general, evidence shows that low socio-economic groups (unskilled or semi-skilled manual) and the lower educational groups spend less on sport (Debordes et al., 1999; Lera-López & Rapún-Gárate, 2005b, 2007; Pawlowki & Breuer, 2012a). It could be argued that highly-educated people have more access to sports infrastructure and are more aware of the (health) benefits of sports participation (Thibaut et al., 2014). Nevertheless, some studies find a negative relationship with sports consumption (Taks et al., 1999; Wicker et al., 2010). Also, from the perspective of Becker’s theory, time availability may be considered as a relevant factor. In fact, some studies show that time availability poses no restriction in terms of expenditure but limits sports participation (Lera-López & Rapún-Gárate, 2005b). Other studies consider the time involvement in sports participation, such as the number of years practicing sport (Taks et  al., 1999; Wicker et  al., 2010), the frequency of participation (Lera-López & Rapún-Gárate, 2005a, 2005b, 2007), or the individual level of performance (Davies, 2002; Lamb et  al., 1992; Wicker et  al., 2010), with a positive relationship with sports expenditure. In general, empirical evidence shows that sports expenditure increases with a higher level of involvement. Even more, some studies consider that involvement in passive participation, such as watching sport on TV, is positively associated with higher consumption of sports participation (Scheerder et al., 2011). Scant attention is paid to the influence of other demographic variables, such as marital status, household size, and the presence of children. Marital status is found to have an insignificant effect on sports consumption (Eakins, 2016; Pawlowski & Breuer, 2011, 2012a). According to Pawlowski and Breuer (2011, 2012a), the number of people in the household is positively related to sports consumption. In contrast, Scheerder et  al. (2011) and Thibaut et  al. (2014) find a negative association. There is no conclusive evidence on the presence of children in a household. On the one hand, Taks et  al. (1999), SWG (1995), Loyland and Ringstad (2009), Thibaut et  al. (2017), and Pawlowski and Breuer (2011, 2012a)

Private Household Consumption in Sport

show that couples and families with independent children (aged between 6 and 18 years) spend more on sport, while Gratton and Taylor (2000) and Lera-López and Rapún-Gárate (2005a, 2005b, 2007) find no relationship. The presence of young children (aged under 6 years) is usually associated with lower consumption in sport (Eakins, 2016; Thibaut et  al., 2014). This relationship might be explained by the fact that parents have to take care of the children, which makes it more difficult for parents to practice sport and consequently spend money on that (Thibaut et al., 2014). Some studies focus on the role played by the environment (location) in sports consumption, usually considering the population size. For example, Pawlowski and Breuer (2011, 2012a) and Eakins (2016) report higher spending on sport in large populations than in small ones. These findings suggest that in larger populations there are more opportunities and infrastructure for practicing sporting activities. Nevertheless, there is also evidence showing a negative association (Loyland & Ringstad, 2009; Thibaut et al., 2017). The studies analyzing the role played by prices usually focus on sports attendance, but there are also some studies on sports participation. For example, Késenne and Butzen (1987) have estimated a price elasticity of −1, although with differences among different sports. More recently, Loyland and Ringstad (2009) estimate price elasticities between −0.15 and −0.35, suggesting that sports are price inelastic goods.

CORRELATES OF CONSUMPTION OF SPORTS ATTENDANCE Among the different activities considered as ‘sport’, attendance at sporting events has experienced an incredible increase in economic terms. For example, while sales of sports goods increased ‘only’ by 79% (from $50,725 million to $90,802 million) between 1990 and 2007 in the US, attendance at sporting events increased by 285% (from $4.8 billion to $18.5 billion) in the same period of time (US Census Bureau, 2010). While the majority of studies in the US focus on specific professional sports, such as soccer, baseball, basketball, and rugby (Byon, Zhang, & Connaugthon, 2010), studies developed in Europe regularly consider sports attendance in general. When comparing the empirical evidence of the consumption of sports attendance, some further difficulties occur. First, correlates for attendance and their consumption might differ compared to the analysis of sports

77

participation. In fact, a general mistake made by some empirical studies is to consider the frequency of sports attendance (Gau & James, 2014) or repurchase intentions (Kim, Byon, Yu, Zhang, & Kim, 2013) as equal to spectator sports consumption. Following the proposal of this chapter, the focus is only on studies explaining sports expenditure at attended sporting events. Second, some studies make a distinction between the likelihood of spending and the amount spent on the events (e.g., Lera-López, Ollo-López, & RapúnGárate, 2012; Pawlowski & Breuer, 2011). Third, not all studies focus on the same items. While some studies focus only on merchandise consumption (e.g., Casper & Menefee, 2010; Park, Suh, & Pedersen, 2016), others consider more items, such as entrance fees, travelling expenses, food and drink expenses, and merchandising products (Lera-López et al., 2011, 2012). These issues have to be considered when discussing the state of research on the factors associated with sports expenditures in attending sporting events as follows. Although age is positively related to sports attendance, this variable is not significant in explaining consumer expenditure on attendance at sports events (Eakins, 2016; Lera-López et al., 2011, 2012; Pawlowski & Breuer, 2012b). There is no consensus on the role of socio-economic variables. According to Eakins (2016), working status is positively associated with expenditure on sporting events, but it is not significant in other studies (Lera-López et  al., 2012; Pawlowski & Breuer, 2011). Income is positively associated with consumption of sporting events (e.g., Cannon & Ford, 2002; Lera-López et al., 2012), confirming that many professional sporting events can be defined as normal goods or luxury goods (Loyland & Ringstad, 2009). There is greater consensus on the influence of other socio-demographic variables. For example, in terms of gender, fewer women than men attend sporting events and they are less likely to spend on these events than men (Eakins, 2016). Having taken the decision of spending, there are no differences between men and women in the amount spent on these activities (Lera-López et al., 2012). Thus, it might be argued that gender role is still a form of social constraint on leisure participation (Culp, 1998), including sporting events, but it is not an economic constraint. In general, although education is positively related to sports attendance (e.g., Montgomery & Robinson, 2010), it is only positively associated with the probability of spending for attending sporting events, but not in terms of the amount spent on them (Lera-López et al., 2012). Family structure seems to be relevant to explain expenditure on sporting events except marital

78

THE SAGE HANDBOOK OF SPORTS ECONOMICS

status (e.g., Eakins, 2016; Pawlowski & Breuer, 2012a). Households with small children (Eakins, 2016; Loyland & Ringstad, 2009) and the presence of children at such events (Cannon & Ford, 2002) are negatively related to consumer expenditure on sports attendance. However, households with teenagers and young adults (below 16 years) are more likely to spend on events, though their share of expenditure is significantly less (Pawlowski & Breuer, 2012a). Some studies include local and regional characteristics in their analysis. For example, Pawlowski and Breuer (2012a) report that households settled in small towns spend less on sporting events than those settled in cities. This possibly reflects that a wider range of entertainment options are available in larger urban centers. Finally, some studies consider the relationship between consumption of sports attendance and other variables reflecting sports involvement, such as sports participation and watching sports programs on TV. It is expected that the time and amount spent on other sports activities competes with attendance at sporting events. However, the evidence does not confirm these arguments. For example, Davies (2002) shows that sports participation is not related to consumer expenditure on sports events, while Lera-López et  al. (2012) report a positive relationship between the amount of spending on sports participation and on sporting events.

FUTURE DIRECTIONS FOR RESEARCH ON SPORTS EXPENDITURES Different horizons for further research towards a better understanding of sports-related expenditure by households and individuals are suggested that result from the limitations described in the previous sections and/or the expected effect of new technologies in practicing and consuming sport and sporting goods. On the one hand, as previously mentioned, studies on the determinants of sports-related expenditure are very difficult to compare, not only due to differences in methodologies and samples, but also because they differ in their definitions on sport and the different consumption items under study. Consequently, in future research it might be necessary to establish a common framework to develop new and comparable empirical studies. The development of international studies including different countries might be an initial step in this direction. Furthermore, as Breuer et al. (2010) mention, most of the studies are cross-sectional

analyses. However, consumption of sports goods, as it is also found for other types of goods, is not constant over time, and changes according to new products, technologies, and the life cycle of households can be expected. On the other hand, new technologies may open up new opportunities for further research in sports consumption. For example, the use of smartphones and sports apps is increasingly associated with sports participation. This could be a way to boost sports consumption through, for example, receiving promotional offers, transmission of information about sports teams, new activities, and practices, etc. These actions might lead to the development of more effective marketing strategies to promote sports consumption (Ha, Kang, & Ha, 2015). Another example is the development of fantasy sports participation and consumption, which has experienced an incredible surge in popularity in many countries and is related to both positive and negative effects on traditional sports consumption (e.g., Karg & McDonald, 2011; Yuksel, McDonald, Milne, & Darmody, 2016). Finally, the review of the literature shows that empirical evidence explaining consumption of sports attendance is very limited compared to the evidence about consumption associated with sports participation in general. Moreover, special attention should be paid on the demand for amateur sports attendance, because travel expenses, time investment, and food/drink and merchandise consumption might be more important for fans than fees and entrance tickets for professional sporting events. Also, in terms of explanatory variables, apart from individual variables usually included in all models, it might be interesting to analyze the influence of sports consumption made by relatives and friends, and the impact of national, regional, and local variables to explain sports consumption made by households.

REFERENCES Aguirre, J., Lera-López, F., & Rapún-Gárate, M. (2008). Economic analysis of sports participation in Spain: Spanish Higher Council for Sport (CSD). Madrid: CSD. Ahlert, G. (2013). The German Sport Satellite Accounts (SSA). GWS Discussion Paper 2013/4. Osnabrück: GWS. Andreff, W., Bourg, J. F., Halba, B., & Nys, J. F. (1995). Les enjeux économiques du sport en Europe: Financement et impact économique. Paris: Dalloz. Australia Bureau of Statistics (2013). Value of sport. Australia. Canberra: ABS. Retrieved March 24,

Private Household Consumption in Sport

2017, from www.abs.gov.au/ausstats/[email protected]/ mf/4156.0.55.002 Beaton, A., & Funk, D. (2008). An evaluation of theoretical framework for studying physically active leisure. Leisure Sciences, 13, 205–220. Beaton, A., Funk, D., & Alexandris, K. (2009). Operationalizing a theory of participation in physically active leisure. Journal of Leisure Research, 41, 177–203. Becker, G. S. (1965). A theory of the allocation of time. The Economic Journal, 75(299), 493–517. Bednarik, J., Kokar, E., & Jurak, G. (2010). Analysis of the sports services market in Slovenia. Kinesiology, 42(2), 142–152. Bloom, M., Grant, M., & Watt, D. (2005). Strengthening Canada: The socio-economic benefits of sport participation in Canada. Ottawa: Conference Board of Canada. Bourdieu, P. (1984). Distinction: A social critique of the judgement of taste. Cambridge, MA: Harvard University Press. Breuer, C., Hallmann, K., Wicker, P., & Feiler, S. (2010). Socio-economic patterns of sport demand and ageing. European Review of Aging and Physical Activity, 7(2), 61–70. Byon, K., Zhang, J., & Connaughton, D. (2010). Dimensions of general market demand associated with professional team sports: Development of a scale. Sport Management Review, 13(2), 142–157. Cannon, T., & Ford, J. (2002). Relationship of demographic and trip characteristics to visitor spending: An analysis of sports travel visitors across time. Tourism Economics, 8(3), 263–271. Casper, J., & Menefee, C. (2010). Prior sport participation and spectator sport consumption: Socialization and soccer. European Sport Management Quarterly, 10(5), 595–611. Culp, R. H. (1998). Adolescent girls and outdoor recreation: A case study examining constraints and effective programming. Journal of Leisure Research, 30(3), 356–379. Davies, L. E. (2002). Consumers’ expenditure on sport in the UK: Increased spending or underestimation? Managing Leisure, 7(2), 83–102. Department for Culture, Media & Sport (2016). UK Sport Satellite Account, 2012, 2014 and 2015. Statistical Release. London: DCMS. Desbordes, M., Ohl, F., & Tribou, G. (1999). Marketing du sport. Paris: Economica. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports economics: Theory, evidence and policy. Oxford: Elsevier. Downward, P., & Riordan, J. (2007). Social interactions and the demand for sport: An economic analysis. Contemporary Economic Policy, 25(4), 518–537. Eakins, J. (2016). An examination of the determinants of Irish household sports expenditures and

79

the effects of the economic recession. European Sport Management Quarterly, 16(1), 86–105. European Commission (2012). Study on the contribution of sport to economic growth and employment in the EU. Final Report. Brussels: European Commission. European Commission (2013). Sport Satellite Accounts. A European Project: New results. Brussels: European Commission. Gau, L., & James, J. (2014). An empirical exploration of relationships between personal values and spectator sport consumption. International Journal of Sport Management Recreation & Tourism, 16, 37–55. Gratton, C., & Taylor, P. (2000). Economics of sport and recreation. London: Spon Press. Ha, J-P., Ha, J-H., & Han, K. (2013). Online sport consumption motives: Why does an ethnic minority group consume sports in a native and host country through the internet? International Journal of Sport Management, Recreation & Tourism, 11, 63–89. Ha, J., Kang, S., & Ha, J. (2015). A conceptual framework for the adoption of smartphones in a sports context. International Journal of Sports Marketing and Sponsorship, 16(3), 2–19. Hallmann, K., & Wicker, P. (2015). Determinants of sport-related expenditure of golf players and differences between light and heavy spenders. Sports, Business and Management, 5(2), 121–138. Indecon International Economic Consultants (2010). Assessment of economic impact of sport in Ireland. Dublin: Irish Sports Council. Jin, H. (2014). Disposable income and actual sports consumption expenditures correlation research based on co-integration theory and error correction models. Journal of Chemical and Pharmaceutical Research, 6(3), 920–925. Karg, A., & McDonald, H. (2011). Fantasy sport participation as a complement to traditional sport consumption. Sport Management Review, 14, 327–346. Késenne, S., & Butzen, P. (1987). Subsidizing sports facilities: The shadow price-elasticities of sports. Applied Economics, 19(1), 101–110. Kim, S., Byon, K., Yu, J., Zhang, J., & Kim, C. (2013). Social motivation and consumption behavior of spectators attending a formula one motor-racing event. Social Behavior and Personality, 41(8), 1359–1378. Kolimpalis, C. (1999, September). The economic importance of sports in Greece. Paper presented at the 7th Congress of the European Association for Sport Management, Thessaloniki. Lamb, L., Asturias, L. P., Roberts, K., & Brodie, D. A. (1992). Sports participation: How much does it cost? Leisure Studies, 11, 19–29.

80

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Lavoie, M. (2004). Post Keynesian consumer theory: Potential synergies with consumer research and economic psychology. Journal of Economic Psychology, 25, 639–649. Lera-López, F., Ollo-López, A., & Rapún-Gárate, M. (2012). Sports spectatorship in Spain: Attendance and consumption. European Sport Management Quarterly, 12(3), 265–289. Lera-López, F., & Rapún-Gárate, M. (2005a). The determinants of consumer expenditure on sports: A Tobit model. In G. Y. Papanikos (Ed.), International research on sports economics and production (pp. 61–78). Athens: Athens Institute for Education and Research. Lera-López, F., & Rapún-Gárate, M. (2005b). Sports participation versus consumer expenditure on sport: Different determinants and strategies in sports management. European Sport Management Quarterly, 5(2), 167–186. Lera-López, F., & Rapún-Gárate, M. (2007). The demand for sport: Sport consumption and participation models. Journal of Sport Management, 21(1), 103–122. Lera-López, F., Rapún-Gárate, M., & Suárez, M. (2011). Determinants of individual consumption on sports attendance in Spain. International Journal of Sport Finance, 6(3), 204–221. Loyland, K., & Ringstad, V. (2009). On the price and income sensitivity of the demand for sports: Has Linder’s disease become more serious? Journal of Sports Economics, 10(6), 601–618. Michon, B., Ohl, F., & Faber, C. (1987). Lex prix de la practique sportive pour le consommateur. Strasbourg: UFR Staps. Montgomery, S., & Robinson, M. (2010). Empirical evidence of the effect of marriage on male and female attendance at sports and arts. Social Science Quarterly, 9(1), 99–116. Park, J., Suh, Y., & Pedersen, P. (2016). Examining spectator motivations in Major League Baseball: A comparison between senior and non-senior consumers. Choregia. Sport Management International Journal, 12(2), 13–36. Pawlowski, T., & Breuer, C. (2011). The demand for sports and recreational services: Empirical evidence from Germany. European Sport Management Quarterly, 11(1), 5–34. Pawlowski, T., & Breuer, C. (2012a). Expenditure elasticities of the demand for leisure services. Applied Economics, 44(26), 3461–3477. Pawlowski, T., & Breuer, C. (2012b). Expenditures on sport products and services. In L. Robinson, P. Chelladurai, G. Bodet, & P. Downward (Eds.), Routledge handbook of sport management (pp. 354–372). New York: Routledge. Preuß, H., & Alfs, C. (2013). Economic dimension of sport consumption in Germany. Sportwissenchaft, 43(4), 239–252.

Prior, D., O’Reilly, N., Mazanov, J., & Huybers, T. (2013). The impact of scandal on sport consumption: A conceptual framework for future research. International Journal of Sport Management and Marketing, 14(1–4), 188–211. Scheerder, J., Vos, J., & Taks, M. (2011). Expenditure on sports apparel: Creating consumer profiles through interval regression modelling. European Sport Management Quarterly, 11(3), 251–274. Scraton, S., & Watson, B. (1999). Sport, leisure identities and gendered spaces. Eastbourne: LSA Publications. Sport Industry Research Centre (2008). Sport market: Forecast 2008–2010. Sheffield, UK: Sport Industry Research Centre (SIRC). Stander, F., & van Zyl, L. (2016). See you at the match: Motivation for sport consumption and intrinsic psychological rewards of premier football league spectators in South Africa. SA Journal of Industrial Psychology, 42(1), 1–13. Statista (2016). Global sports market. Retrieved March 24, 2017, from www.statista.com/ statistics/194122/sporting-event-gate-revenueworldwide-by-region-since-2004/ Statistical Working Group (SWG) of the Sport and Recreation Ministers Council (1995). Available data and sources for the sport and recreation industry. Adelaide, Australia. Taks, M., & Késenne, S. (2000). The economic significance of sport in Flanders. Journal of Sport Management, 14(4), 342–365. Taks, M., Renson, R., & Vanreusel, B. (1995). Social stratification in sport: A matter of money or taste? European Journal for Sport Management, 2, 4–14. Taks, M., Renson, R., & Vanreusel, B. (1999). Consumer expenses in sport: A marketing tool for sport. European Journal for Sport Management, 6, 4–18. Taks, M., Vanreusel, B., & Renson, R. (1994). What does sport really cost? A micro-economic study of the consumer cost of golf and soccer. The European Journal for Sport Management, 1, 22–34. Thibaut, E., Eakins, J., Vos, S., & Scheerder, J. (2017). Time and money expenditure in sports participation: The role of income in consuming the most practiced sports activities in Flanders. Sport Management Review, 20(5), 455–467. Thibaut, E., Vos, S., Lagae, W., Van Puyenbroeck, T., & Scheerder, J. (2016). Participation in cycling, at what cost? Determinants of cycling expenses. International Journal of Sport Management and Marketing, 16(3), 221–238. Thibaut, E., Vos, S., & Scheerder, J. (2014). Hurdles for sports consumption? The determining factors of household sports expenditures. Sport Management Review, 17(4), 444–454. Tobin, J. (1958). Estimations of relationship for limited dependent variables. Econometrica, 26, 24–36.

Private Household Consumption in Sport

US Census Bureau (2010). The 2010 Statistical Abstract, Arts, Recreation, and Travel. Washington, DC. Wicker, P., Breuer, C., & Pawlowski, T. (2010). Are sports club members big spenders: Findings from sport specific analyses in Germany. Sport Marketing Quarterly, 13(3), 214–224. Wicker, P., Hallmann, K., & Zhang, J. M. (2012). What is influencing consumer expenditure and intention to revisit? An investigation of marathon events. Journal of Sport & Tourism, 17(3), 165–182. Wicker, P., Prinz, J., & Weimar, D. (2013). Big spenders in a booming sport: Consumption capital as a

81

key driver of triathletes’ sport-related expenditure. Managing Leisure, 18(4), 286–299. Woo, B. (2016). Revisiting the model of sport consumption: Review of theories and suggestions. Asian Journal of Physical Education and Sport Science, 4(1), 19–36. Yuksel, M., McDonald, M., Milne, G., & Darmody, A. (2016). The paradoxical relationship between fantasy football and NFL consumption: Conflict development and consumer copying mechanisms. Sport Management Review, 20(2), 198–210.

9 Sports Clubs in Europe: Organization C h r i s t o p h B r e u e r, P h i l i p p S w i e r z y and Svenja Feiler

INTRODUCTION OCCURRENCE AND RELEVANCE OF SPORTS CLUBS IN EUROPE Sports clubs as local nonprofit organizations exist in all European countries. In 2014, 12% of EU citizens held a sports club membership (European Commission, 2014). Admittedly, membership rates vary largely among European countries, as sports clubs have different positions in national sports structures and policies, and because historical, cultural, and socio-economic differences led to country-specific developments (cf., Scheerder, Willem, & Claes, 2017). However, sports clubs largely share similar roots (Hoekman, van der Werff, Nagel, & Breuer, 2015b). The impact of sports clubs on the European economy is considerable, although it is difficult to quantify. According to calculations by Breuer and Feiler (2017a), the total annual income generated by German sports clubs sums up to €3.33 billion. Resulting governmental income through taxes amounts to 20% of the revenues (Breuer & Feiler, 2013). Taking indirect economic contributions into account, Breuer and Feiler (2017a) estimated an added value of voluntary engagement in German sports clubs of €4.1 billion annually.

Also, intangible effects of sports participation, such as mental health, physical health, or social capital accumulation, significantly contribute to the welfare of states (Breuer, Wicker, & Orlowski, 2014).

CONSTITUTIVE CHARACTERISTICS AND FUNCTIONS OF NONPROFIT SPORTS CLUBS The organizational attributes of nonprofit sports clubs, which distinguish them from for-profit organizations, can be divided into constitutive and economic characteristics (Horch, 1994b). Constitutive features are voluntary membership, orientation towards the members’ interest, autonomy due to an independence from third parties, democratic decision-making structure, and voluntary work (e.g., Heinemann, 1984; Horch, 1994b). Economic features of sports clubs are the nonprofit orientation, the role identity of members as they act as producers and consumers of the club’s goods or services, the autonomous revenue structure, and the principle of solidarity (Horch, 1994b). The resources that each member contributes, whether they are money or time, are pooled

Sports Clubs in Europe: Organization

and the democratic board decides on the disposition of them (Heinemann, 1984). Sports clubs produce socially desirable goods or services (e.g., sporting activity, civic engagement, social integration) and thereby justify public subsidization. In particular, they are attributed with several beneficial functions, as summarized by Heinemann and Horch (1981) based on previous, non-empirical research: an integrative function for being inclusive regarding different population groups, a political function as they produce welfare within democratic structures (e.g., through youth work), a social function via being an instrument for political socialization (‘schools of democracy’), an identity function as they provide possibilities to self-realization, a status function as they create or secure status differences within or outside the club, an economic function (e.g., because volunteering affects income and employment), a health function through the provision of physical education, and in addition, clubs are the basis for elite sports and the benefits associated with it. Altogether, sports clubs are needs- not profit-oriented corporative actors. They collectively produce goods or services which depend on actions based on reciprocity and relationships based on solidarity (Heinemann, 2004).

THEORETICAL BACKGROUND

83

services that are neither completely rival and excludable (private goods) nor completely nonrival and non-excludable (public goods). Buchanan (1965) describes that certain public goods exist for which exclusion is possible and which are subject to rivalry in form of congestion, but which are not provided by the state. They are offered by private, non-governmental organizations, such as nonprofit sports clubs, wherein members pool their resources, have mutual benefits, and exclude people who do not pay membership fees. They further share production costs and beneficial personal characteristics. It is assumed that the members have homogeneous interests, work voluntarily, and aim to maximize their utility. According to Buchanan (1965), private, non-governmental alternatives are suited for the provision of these club goods. In addition to Buchanan’s theory of clubs, Musgrave’s (1957) theoretical framework on merit goods describes the type of good or service sports clubs offer. According to Musgrave, there are certain goods and services where the expected individual benefits are not clear to the individual at the point of consumption. The consumption of these goods causes external benefits to society or parts of it, which is also not known or recognized at the time of consumption. As a result, the total supply for goods or services with these features would not achieve the socially efficient level of consumption. To reduce the costs of supply, governmental subsidies are given to providers of merit goods. Due to their aforementioned beneficial functions, services of sports clubs are considered as merit goods.

Types of Goods and Services According to neoclassical economic theory, private organizations aim to maximize profits under certain constraints such as production costs or limited available information. If private market provision of a particular good or service is not worthwhile economically, there will be an undersupply of this good or service by private firms. As this insufficient provision could be undesired to society or parts of it, the state intervenes by providing these goods or services. Samuelson (1954) characterizes these public goods as non-excludable and non-rivalrous in consumption. Thus, the consumption of a public good by one individual does not lead to subtraction from other individuals of the same good or service. Thereby, a sharp differentiation between purely private goods, which are excludable and rival in consumption, and purely public goods is undertaken. The economic theory of clubs by Buchanan (1965) sets in at the point where fundamental neoclassical theories, which focus on private goods, and the public good theory (Samuelson, 1954) are unable to explain the existence of goods or

Economic Theories of Nonprofit Organizations According to Anheier (2014), a major strength of the following microeconomic theories on the existence of nonprofit organizations in developed market economies is that they are largely complementary rather than rival. The public goods theory of nonprofit organizations (or ‘governmental failure theory’) by Weisbrod (1975) argues that the third sector exists because of a failure of the state. It is based on two concepts. First, different demands regarding public goods lead to the development of nonprofit organizations. The more population groups with differing demands, the more heterogeneous the society. The more diverse the society, the larger will be the number of nonprofit organizations. Second, some demands remain unfulfilled as the provision of public goods focuses on the ‘median voter’. The median voter is the largest segment of the demand for public and quasi-public goods. The crucial assumption is that

84

THE SAGE HANDBOOK OF SPORTS ECONOMICS

governmental officials seek to maximize the chances of re-election and therefore try to fulfil the demands of the median voter. Consequently, several demands remain unfulfilled. Nonprofit organizations arise from voluntary contributions of citizens that want to increase the output quantity or quality of a public good. They function as ‘gap fillers’ for public goods that are not provided by the public sector. However, Hansmann (1987) criticizes that this theory is not able to explain why for-profit organizations do not fill the unsatisfied demands. Whereas Weisbrod’s answer was that the market would only address those target groups who are sufficiently large and profitable (Weisbrod, 1988), Hansmann elaborated on a different explanation. By introducing the trust-related theory (or ‘market failure theory’), Hansmann (1987) emphasizes that for-profit organizations have an incentive to take advantage of information asymmetries between provider and consumer. Information asymmetries occur when either provider or consumer have more information on the true quality of a product or service. As a consequence of information asymmetries, the risk increases that the individual or organization with more information behaves irresponsibly or immorally (moral hazard). Therefore, marketcorrecting mechanisms, such as the prohibition of profit distribution, are required. As a consequence, trust in the provider is higher for organizations that have no incentive to take advantage of surplus information. This trustworthiness is signaled by nonprofit organizations which have a non-distribution constraint that prohibits the distribution of profit. Advantages of nonprofit organizations compared to for-profits are that monitoring, which is needed in for-profits, is expensive, e.g., due to existing transaction costs, and that profiteering is likely, e.g. due to moral hazard (Anheier, 2014). Stakeholder theories of the nonprofit sector (e.g., Ben-Ner, Ren, & Paulson, 2011) extend Hansmann’s argumentation by describing that especially demand-side stakeholders, particularly consumers, establish nonprofit organizations because they want to avoid being disadvantaged by informational asymmetries. The consumers become volunteers, which are driven by a stronger intrinsic motivation than usual employees who are financially rewarded. Thus, according to the stakeholder theories, nonprofit organizations send signals of trustworthiness that are stronger than those described by Hansmann. Furthermore, entrepreneurship theories of nonprofit organizations (James, 1987; Rose-Ackerman, 1996) describe that social entrepreneurs contribute to the development of nonprofit organizations. They are relatively innovative, creative, and opportunityoriented and differ from business entrepreneurs as they are primarily interested in immaterial,

non-monetary value maximization. Social entrepreneurs create social instead of monetary value. Salamon (1987) recognized that both the market and the governmental failure theory complement each other and developed the ‘interdependence theory’ (or ‘voluntary failure theory’, respectively ‘third-party government theory’) that refers to governments and nonprofit organizations as partners living a symbiotic relationship. This is based on a large share of public funds given to nonprofit organizations. Non-governmental entities carry out governmental purposes and, therefore, indirect and direct governmental support is given to nonprofit organizations. Thus, the relationship between governments and nonprofit organizations is collaborative and they compensate each other’s weaknesses (Heinemann, 1995). In addition, the low level of willingness to pay for many sports is likely to be a reason for the fact that mainly, or exclusively, sports clubs offer these sports. Sports clubs are the only type of service organization in which most of the work is provided on a voluntary basis, and which can break even at a low price.

Other Theories Two other theories, the theory of organizational capacity of voluntary and nonprofit organizations (Hall et  al., 2003) and the resource dependence theory (Pfeffer & Salancik, 1978), often serve as theoretical background for empirical research on sports clubs. Organizational capacity is the ability of organizations to fulfil their missions and mandates. For sports clubs, this is to serve the needs and interests of their members. Organizational capacity is multidimensional and distinguishes between human resources, financial, and structural capacities. The functioning of a nonprofit organization depends on its organizational capacity, which is in turn influenced by the resource situation of the club. The resource dependence theory shows that the autonomy of organizations decreases with an increasing dependence on external resources which are needed in order to provide goods or services. Thus, external stakeholders have the power to influence organizational behavior.

EMPIRICAL RESEARCH ON ORGANIZATIONAL ASPECTS OF EUROPEAN SPORTS CLUBS Administrative Level, Organizational Form, and Degree of Formalization Voluntary nonprofit sports clubs are the main providers of sports participation opportunities in

Sports Clubs in Europe: Organization

almost all European countries (European Commission, 2014). Their unique value, particularly as sports for all-provider and contributor to social welfare is recognized by most European governments, which is one reason why sports clubs are subsidized directly or indirectly (Hoekman, van der Werff, Nagel, & Breuer, 2015a). In Europe, the respective national sports federation and the regional organizations, which are part of the national sports confederation, form the umbrella organizations for sports clubs. The political administration policies from local authorities, governments, or municipalities represent the administrative framework for sports clubs in Europe (Ibsen et al., 2016). The level of formalization expresses the extent of how rules, procedures, and instructions govern an organization’s behavior independently from its personal attributes (Nichols, Wicker, Cuskelly, & Breuer, 2015). The degree of formalization in sports clubs varies significantly across and within countries. Whereas in the UK, for example, clubs are rather formal or semi-formal, German clubs range from highly formal to highly informal (Nichols et al., 2015). In 1983, Horch argued that German sports clubs were still rather informal and the higher the degree of voluntary engagement, the less formalized they were. In Scotland (Allison, 2001) and Greece (Papadimitriou, 2002), clubs are rather informal and focused on day-to-day operations. Horch (1994b) found that the degree of solidarity among members decreases with an increasing level of formalization. Concerning the organizational form of sports clubs in Europe, the comparative analysis by Hoekman et  al. (2015a) shows the majority of clubs are single-sports clubs, especially in Belgium (85%), Sweden (85%), the Netherlands (81%), and Spain (70%). Allison (2001) found that Scottish sports clubs are also mostly single-sports clubs because they fear to share their resources among more sports. Therefore, cooperations between single-sports clubs are preferred. Multi-sports clubs that offer more sports and often unite more members are frequently found in Austria, Finland, and Germany (Hoekman et al., 2015a). The aforementioned results are confirmed by a recent comparative study on sports clubs in Europe, which shows that single-sports clubs are particularly found in the Netherlands, Flanders, and England, while in Germany more than 40% of clubs are multi-sports clubs and in Norway and Hungary about one third of all sports clubs offer more than one sport (Breuer et  al., 2017). Connecting organizational size in terms of the number of members with the level of formalization, Nichols and James (2008) found that, while smaller clubs neither have the capacity nor the need for formality, larger ones are characterized by more formalized procedures.

85

In addition to single- and multi-sports clubs, niche clubs that offer opportunities to practice sport for specific population groups such as migrants exist. Stahl, Wicker, and Breuer (2011) found that migrant sports clubs, compared to an average German sports club, have been founded more recently, have fewer members and are characterized by a higher share of males and middleaged members. They offer fewer sports and are more likely to offer football. Furthermore, they are more seriously threatened by limited infrastructural availability, survival and financial problems, and their culture is largely based on conviviality (Stahl et al., 2011). In contrast, another minority group – people with disabilities – is not organized in specific clubs. Wicker and Breuer (2014) found that disabled people rather hold memberships at multi-sports clubs. The greater capacity of multisports clubs is ideal for the provision of additional, more specific programs.

Organizational Objectives, Performance, and Problems The major goal of sports clubs is the provision of club goods for the exclusive utility and interest of their members (Heinemann, 2004). Therefore, financial resources, mainly in the form of membership fees, and temporal resources, in the form of voluntary work, are delivered by members (Sandler & Tschirhart, 1980). However, the particular objectives of clubs are diverse and often several are pursued simultaneously. According to Nagel (2008), club goals range from achieving a sufficient provision of opportunities to practice sport for the local population, to offering competitive sports opportunities, or increasing sociability (cf., Balduck, Lucidarme, Marlier, & Willem, 2014). Correspondence between the goals of the club and the interests of its members is decisive for the individual commitment to the club (Nagel, 2008). Changes of club goals, and also structural changes, most likely depend on initiatives of a few members (Nagel et al., 2015). According to recent findings by Breuer and Feiler (2017b), goals of German sports clubs have significantly changed between 2005 and 2015. Particularly, quality-­ orientation, engagement in youth work and health sports, as well as talent promotion and the importance of competitive success have decreased. According to Ibsen et  al. (2016), sport for all provision is the focus among most European sports clubs. In contrast to for-profit organizations, which aim to maximize their profits, the goals of sports clubs are formed by the shared interests and preferences of their members. This implies that goal

86

THE SAGE HANDBOOK OF SPORTS ECONOMICS

achievement measurability is not as straightforward as in for-profit organizations. By analyzing Finnish sports clubs, Koski (1995) argues that the effectiveness of sports clubs in realizing their aims depends on their ability to obtain resources and to construct an efficient throughput process, their general level of activity, and their internal atmosphere. Several club characteristics, such as club size, organizational environment, or ideological orientation are also linked to effectivity. Concerning club size, Wicker, Breuer, Lamprecht, and Fischer (2014) found that economies of scope exist in German sports clubs with an increase in the number of sports, and in Swiss sports clubs with an increase in the number of members. This could be a reason why larger clubs receive more public funds (Wicker & Breuer, 2009). Thus, clubs that aim for production efficiency should diversify and offer more sports. However, Wicker et  al. (2014) argue that smaller clubs provide important niche opportunities, and thereby, they also justify receiving governmental subsidies. The scarcity of resources determines the organizational performance. Wicker and Breuer (2011) found that German sports clubs mainly suffer from scarce human and infrastructure resources. Infrastructure resources are relatively scarce as clubs acquire them from cooperations and thus, from contributions of organizations that can hardly be controlled by the club itself. Clubs from the UK, for instance, are susceptible to changes in charges, availability, and quality of infrastructure, as they rely on partner organizations to provide access to facilities (Taylor, Barrett, & Nichols, 2009). Concerning human resources, the recruitment and retention of volunteers is a problem in many European countries (e.g., Burgham & Downward, 2005; Koski, 2012). According to Breuer, Wicker, and von Hanau (2012), German sports clubs possess the capacity to compensate volunteer resource problems as they tend to substitute a reduction in volunteers by an increase in secondary, sporadically active volunteers in the short term. In the long run, volunteers are substituted by paid staff. In addition to the recruitment and retention of volunteers, the recruitment and retention of adolescent competitive athletes, coaches, and members threaten clubs (e.g., Breuer & Feiler, 2017a; Breuer et  al., 2017). Wicker and Breuer (2010) found that these problems are shaped by club size, the share of members participating in convivial gatherings, and the offered sports. Wicker and Breuer (2013) identified human resources capacities, such as the number of women on the board, and secondary volunteers, who engage sporadically at sport or club events without holding an official position, as decisive for the severity

of organizational problems regarding the recruitment and retention of volunteers, members, and coaches, as well as financial problems. Moreover, financial, infrastructural, and cultural resources have an influence on these problems. For instance, a larger share of women on the board and setting a high value on conviviality reduce organizational problems. By analyzing communities and sports clubs simultaneously, Wicker and Breuer (2015) found that the financial and economic condition of the community influence sports clubs’ organizational capacity. High unemployment rates lead to relatively fewer problems regarding the recruitment and retention of volunteers. Clubs located in larger communities face greater problems regarding facility and financial problems. Sports clubs located in communities that break even are also more likely to break even.

Financial Structure, Development, and Autonomy Nonprofit organizations and sports clubs in particular have multiple income sources (Young, 2007). Membership fees mostly account for the largest share of income (e.g., Lamprecht, Fischer, & Stamm, 2011; Taylor et  al., 2009; Wicker, Breuer, & Hennings, 2012a). Governmental subsidies, fundraising or sponsorship activities, and lotteries (e.g., in Austria, Germany, Sweden) are also important revenue sources (Hoekman et  al., 2015a). The largest income categories of German sports clubs, for instance, are membership fees, followed by donations, income from sport events, subsidies from local governments, income from club bars and restaurants, as well as income from social events (Breuer & Feiler, 2017a). Commercial income sources represent a growing share of total revenues for sports clubs in various European countries (e.g., Enjolras, 2002; Priemer, Labigne, & Krimmer, 2015). Furthermore, cost savings arise from non-monetary grants such as the provision of sports facilities, which are mostly provided by local or central governments (Wicker & Breuer, 2011; Hoekman et  al., 2015a). Other indirect subsidies, such as tax advantages, are also frequently granted by municipalities (Wicker, 2009). There are several factors that influence the generation of income from particular sources. According to Feiler, Wicker, and Breuer (2015), the reception of donations is positively influenced by certain club goals, such as the provision of elite sports and the promotion of young talents. Moreover, if clubs place value on social aspects, companionship, and conviviality,

Sports Clubs in Europe: Organization

they generate higher revenues from donations. Professionalization through hiring paid employees also increases donative revenues, but an orientation to commercial suppliers negatively influences the reception of donations as donors fear that clubs change their mission. Additionally, the legal structure influences the revenue composition of sports organizations (Wicker, Weingärtner, Breuer, & Dietl, 2012). Sponsorship revenue, for example, is better protected against hold-up for sponsors and customers in sports clubs compared to private firms. Different types of revenue sources are interrelated with problems of sports clubs. According to Coates, Wicker, Feiler, and Breuer (2014), sports clubs that rely on external funding, such as sponsorship income, experience larger problems concerning their financial situation and regarding the recruitment and retention of volunteers. Subsidy funding solely indicates larger problems concerning the latter, which is why subsidy funding should be preferred over sponsorship income. Generating sufficient financial resources is not a major organizational aim of sports clubs. Instead, it is only a constraint in achieving superior objectives. However, the type of income is decisive for the total amount of revenues generated (Kearns, 2007). Organizations that largely rely on membership fees have relatively little income. Thus, additional channels for income generation need to be found (Priemer et  al., 2015), but potential crowding-in and crowding-out effects, which are observed in sports clubs, must be considered. Income from public subsidies leads to crowdingin of donations and income from commercial activities and sponsoring in German sports clubs (Wicker et  al., 2012a). Furthermore, volatility of income is positively influenced by governmental subsidies, but negatively by membership fees (Wicker, Longley, & Breuer, 2015a). The main expense categories of German sports clubs are coaches, maintenance of infrastructure, and sports equipment (Breuer & Feiler, 2017a). In order to reduce uncertainty regarding income, revenue diversification is an appropriate management approach. Especially, club-specific volatility can be reduced by generating revenues from an increased number of sources (Wicker et al., 2015a). However, systematic volatility that is external to the club is only marginally influenced by the degree of revenue diversification. The levels of systematic and club-specific volatility are reduced among clubs that rely more on membership fees and less on subsidies. According to Wicker, Feiler, and Breuer (2013), revenue diversification is determined by the organizational mission. Goals like promoting elite sport, tradition, conviviality, non-sport programs, and youth

87

sport positively influence revenue diversification. On the other side, clubs with a mission referring to an orientation to commercial suppliers or a focus on leisure and health sport have less diversified revenue sources. In addition to a loss of certainty concerning revenues, clubs fear to lose autonomy, which decreases with an increasing influential power of external stakeholders. In most European countries, direct and indirect governmental subsidies are common (Hoekman et  al., 2015a). However, the share of direct public subsidies within the income portfolio of sports clubs varies largely between countries. While in the Netherlands, only 5% of all club revenue stem from public money, the share of public subsidies amounts to 41% in sports clubs in Poland and 28% in Hungarian sports clubs (Breuer et al., 2017). An increasing state influence led to a reduced autonomy of the nonprofit sports sector in most of the modern European societies (Heinemann, 2005). Although clubs are widely independent in a legal and ideological sense, even in states where the governing bodies have a large influence on the nonprofit sector, for instance in Italy and Spain (Hoekman et  al., 2015a), an increasing share of governmental funds led to a decrease of autonomy (Horch, 1994a). Although certain sports authorities consider subsidy regulation as an effective instrument for achieving policy goals (e.g., Vos, Breesch, Késenne, van Hoecke, Vanreusel, & Scheerder, 2011), sports clubs still have considerable decision-making autonomy (Vos, Wicker, Breuer, & Scheerder, 2012).

Management and Strategy Based on Porter’s generic strategies, Wicker, Soebbing, Feiler, and Breuer (2015) explored how that strategy affects organizational problems in German sports clubs. Cost leadership and focus strategy lead to smaller organizational problems. Focusing on leisure, mass sport, sport for older people, and non-sport programs lead to less serious organizational problems, but inconsistent effects are found for focusing on health sport, sport for immigrants, and programs similar to those of commercial providers. However, the decisions to pursue specific strategies are rarely made from proactive initiating. In most sports clubs, the decision-making processes are rather characterized by reactive behavior (Schlesinger, Klenk, & Nagel, 2015). Moreover, staff development is a crucial issue of club management. In several countries, such as Switzerland, Sweden, Norway, Italy, and Finland, paid employees are increasingly hired by

88

THE SAGE HANDBOOK OF SPORTS ECONOMICS

nonprofit sports clubs (Hoekman et al., 2015a). A concern regarding staff composition is that women are underrepresented in sports clubs, particularly in management and leadership positions. With a higher share of women on the club board, organizational problems are less severe, which is why more women should become board members (Wicker, Breuer, & von Hanau, 2012). Major challenges for club management are the recruitment and retention of volunteers, adolescent competitive athletes and coaches (Breuer & Feiler, 2017a). Another challenge is the growing competitive environment of sports clubs through private, commercial providers, such as fitness clubs or yoga studios (e.g., Gammelsæter, 2010; Nagel et al., 2015), whereby substitution effects are observed. Although the problem of local competition from commercial sports providers is the smallest in relation to other challenges, it has increased within the past 10 years (Breuer & Feiler, 2017a; Breuer & Feiler, 2017b). The existence of commercial providers decreases participation rates in sports clubs (Hallmann, Feiler, & Breuer, 2015). Although commercial providers complement nonprofit sports clubs in terms of extending the availability of different kinds of sports programs to different target groups (Hallmann et al., 2015), there is a need among sports clubs to react more flexibly to changes in market demand (Nowy, Wicker, Feiler, & Breuer, 2015). One trend in sports clubs is that they offer their initially excludable services or programs to non-members. Also, an increasing level of internal commercialization can be observed, whereby members have to pay for additional services that are not covered by their membership fee (Wicker, 2017).

TRENDS AND FUTURE HORIZONS IN SPORTS CLUB RESEARCH The level of knowledge on organizational aspects of sports clubs differs among European countries due to heterogeneous data availability. In some Western and Northern states, the empirical evidence is relatively detailed, for example in Germany, where large-scale primary panel data on sports clubs are collected and analyzed within the Sport Development Report (see Breuer & Feiler, 2017a). In the UK, there have been the Central Council of Physical Recreation (CCPR) Survey of Sports Clubs 2009 (Taylor et  al., 2009) and the Sports Club Survey 2013 (Sport and Recreation Alliance, 2013). A comparison of Danish and Norwegian sports clubs, based on prior surveys of

sports associations, was undertaken by Ibsen and Seippel (2010). However, a lack of information on sports clubs can be observed in Eastern and some Southern European countries (Hoekman et  al., 2015a). Hence, club surveys from these countries are needed. Furthermore, cross-country comparisons are difficult as comprehensive data, based on studies surveying sports clubs from several European states simultaneously, are rare (Hoekman et  al., 2015a). An exception is the Eurobarometer on sport and physical activity, a comparative survey by the European Commission (2014). With the intention to improve comparable knowledge on sports clubs among different European countries, the SIVSCE-research project (‘Social Inclusion and Volunteering in Sports Clubs in Europe’) has been implemented from 2015 to 2017 by researchers from ten European countries, surveying sports clubs from an environment-, club-, and individuallevel perspective. These multi-level approaches provide a deeper understanding of the connections between and within nonprofit sports clubs and should therefore guide future research.

REFERENCES Allison, M. (2001). Sports clubs in Scotland. Edinburgh: SportScotland. Anheier, H. K. (2014). Nonprofit organizations: Theory, management, policy (2nd ed.). London: Routledge. Balduck, A. L., Lucidarme, S., Marlier, M., & Willem, A. (2014). Organizational capacity and organizational ambition in nonprofit and voluntary sports clubs. Voluntas, 26(5), 2023–2043. Ben-Ner, A., Ren, T., & Paulson, D. F. (2011). A sectoral comparison of wage levels and wage inequality in human service industry. Nonprofit and Voluntary Sector Quarterly, 40(4), 608–633. Breuer, C., & Feiler, S. (2013). Finanzielle Situation und ökonomische Bedeutung des Vereinssports [Financial situation and economic importance of club sports]. In C. Breuer (Ed.), Sportentwicklungsbericht 2011/2012: Analyse zur Situation der Sportvereine in Deutschland (pp. 48–71). Cologne: Sportverlag Strauß. Breuer, C., & Feiler, S. (2017a). Sport Development Report 2015/2016: Analysis of the situation of sports clubs in Germany. Abbreviated Version. Cologne: Sportverlag Strauß. Breuer, C., & Feiler, S. (2017b). Sportvereine im Wandel – Zur Entwicklung der Sportvereine von 2005 bis 2015 [Sports clubs in transition – On the development of sports clubs from 2005 to 2015].

Sports Clubs in Europe: Organization

In C. Breuer (Eds.), Sportentwicklungsbericht 2015/2016: Analyse zur Situation der Sportvereine in Deutschland (pp. 243–263). Cologne: Sportverlag Strauß. Breuer, C., Feiler, S., Llopis-Goig, R., Elmose-Østerlund, K., Bürgi, R., Claes, E., et al. (2017). Sports clubs in Europe: Results of an online survey among sports clubs in ten European countries. Odense: Center for Sports, Health and Civil Society University of Southern Denmark. Breuer, C., Wicker, P., & Orlowski, J. (2014). Zum Wert des Sports [About the value of sports]. Wiesbaden: Springer Gabler. Breuer, C., Wicker, P., & von Hanau, T. (2012). Consequences of the decrease in volunteers among German sports clubs: Is there a substitute for voluntary work? International Journal of Sport Policy and Politics, 2(4), 173–186. Buchanan, J. M. (1965). An economic theory of clubs. Economica, 32(125), 1–14. Burgham, M., & Downward, P. (2005). Why volunteer, time to volunteer? A case study from swimming. Managing Leisure, 10(2), 79–93. Coates, D., Wicker, P., Feiler, S., & Breuer, C. (2014). A bivariate probit examination of financial and volunteer problems of non-profit sport clubs. International Journal of Sport Finance, 9, 230–248. Enjolras, B. (2002). The commercialization of voluntary sport organizations in Norway. Nonprofit and Voluntary Sector Quarterly, 31(3), 352–376. European Commission (2014). Sport and physical activity. Special Eurobarometer 412, Wave EB80.2. Brussels: European Commission. Feiler, S., Wicker, P., & Breuer, C. (2015). How to raise voluntary giving for nonprofit sports clubs: an analysis of factors influencing donations. Voluntas, 26(4), 1219–1239. Gammelsæter, H. (2010). Institutional pluralism and governance in ‘commercialized’ sport clubs. European Sport Management Quarterly, 10(5), 569–594. Hall, M. H., Andrukow, A., Barr, C., Brock, K., de Wit, M., & Embuldeniya, D. (2003). The capacity to serve: A qualitative study of the challenges facing Canada’s nonprofit and voluntary organizations. Toronto: Imagine Canada. Hallmann, K., Feiler, S., & Breuer, C. (2015). Design and delivery of sport for all programmes: Should it be market, non-profit, or state-run? International Journal of Sport Finance, 7(4), 565–585. Hansmann, H. (1987). Economic theories of nonprofit organisations. In W. W. Powell (Ed.), The nonprofit sector: A research handbook (pp. 27– 42). New Haven, CT: Yale University Press. Heinemann, K. (1984). Socioeconomic problems of sports clubs. International Review for the Sociology of Sport, 19(3–4), 201–214.

89

Heinemann, K. (1995). Einführung in die Ökonomie des Sports [Introduction to the economy of sports]. Schorndorf: Hofmann. Heinemann, K. (2004). Sportorganisationen: Verstehen und gestalten [Sports organizations: Understand and design]. Schorndorf: Hofmann. Heinemann, K. (2005). Sport and the welfare state in Europe. European Journal of Sport Science, 5(4), 181–188. Heinemann, K., & Horch, H.-D. (1981). Soziologie der Sportorganisation [Sociology of the sports organization]. Sportwissenschaft, 11(2), 123–150. Hoekman, R., van der Werff, H., Nagel, S., & Breuer, C. (2015a). A cross-national comparative perspective on sport clubs in Europe. In C. Breuer, R. Hoekman, S. Nagel & H. van der Werff (Eds.), Sport clubs in Europe: A cross-national comparative perspective (pp. 419–435). Baltimore, MD: Springer. Hoekman, R., van der Werff, H., Nagel, S., & Breuer, C. (2015b). Introduction. In C. Breuer, R. Hoekman, S. Nagel & H. van der Werff (Eds.), Sport clubs in Europe: A cross-national comparative perspective (pp. 1–5). Baltimore, MD: Springer. Horch, H.-D. (1983). Strukturbesonderheiten freiwilliger Vereinigungen: Analyse und Untersuchung einer alternativen Form menschlichen Zusammenarbeitens [Structural peculiarities of voluntary associations: Analysis and investigation of an alternative form of human collaboration]. Frankfurt: Campus. Horch, H.-D. (1994a). Does government financing have a detrimental effect on the autonomy of voluntary associations? Evidence from German sports clubs. International Review for the Sociology of Sport, 29(3), 269–285. Horch, H.-D. (1994b). On the socio-economics of voluntary organisations. Voluntas, 5(2), 219–230. Ibsen, B., Nichols, G., Elmose-Østerlund, K., Breuer, C., Claes, E., Disch, J., et  al. (2016). Sports club policies in Europe. Odense: Center for Sports, Health and Civil Society University of Southern Denmark. Ibsen, B., & Seippel, Ø. (2010). Voluntary organized sport in Denmark and Norway. Sport in Society, 13(4), 593–608. James, E. (1987). The non-profit sector in comparative perspective. In W. W. Powell (Ed.), The nonprofit sector: A research handbook (pp. 397–415). New Haven, CT: Yale University Press. Kearns, K. (2007). Income portfolios. In D. R. Young (Ed.), Financing nonprofits: Putting theory into practice (pp. 291–314). Lanham, MD: AltaMira Press. Koski, P. (1995). Organizational effectiveness of Finnish sports clubs. Journal of Sport Management, 9, 85–95. Koski, P. (2012). Finnish sports club as a mirror of society. International Journal of Sport Policy and Politics, 4(2), 257–275.

90

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Lamprecht, M., Fischer, A., & Stamm, H. P. (2011). Sportvereine in der Schweiz [Sports clubs in Switzerland]. Magglingen: Bundesamt für Sport. Musgrave, R. A. (1957). A multiple theory of budget determination. FinanzArchiv, 17(3), 333–343. Nagel, S. (2008). Goals of sports clubs. European Journal for Sport and Society, 5(2), 121–141. Nagel, S., Schlesinger, T., Wicker, P., Lucassen, J., Hoekman, R., van der Werff, H., & Breuer, C. (2015). Theoretical framework. In C. Breuer, R. Hoekman, S. Nagel & H. van der Werff (Eds.), Sport clubs in Europe: A cross-national comparative perspective (pp. 7–27). Baltimore, MD: Springer. Nichols, G., & James, M. (2008). One size does not fit all: Implications of sports club diversity for their effectiveness as a policy tool and for government support. Managing Leisure, 13(2), 104–114. Nichols, G., Wicker, P., Cuskelly, G., & Breuer, C. (2015). Measuring the formalization of community sports clubs: Findings from the UK, Germany and Australia. International Journal of Sport Policy and Politics, 7(2), 283–300. Nowy, T., Wicker, P., Feiler, S., & Breuer, C. (2015). Organizational performance of nonprofit and forprofit sport organizations. European Sport Management Quarterly, 15(2), 155–175. Papadimitriou, D. (2002). Amateur structures and their effect on performance: The case of Greek voluntary sports clubs. Managing Leisure, 7, 205–219. Pfeffer, J., & Salancik, G. B. (1978). The external control of organizations. New York: Harper & Row. Priemer, J., Labigne, A., & Krimmer, H. (2015). Wie finanzieren sich zivil-gesellschaftliche Organisationen in Deutschland? Eine Sonderauswertung des ZiviZ-Surveys [How do civil-society organizations in Germany finance themselves? A special analysis of the ZiviZ-survey]. Hamburg: Körber-Stiftung. Rose-Ackerman, S. (1996). Altruism, nonprofits and economic theory. Journal of Economic Literature, 34(2), 701–728. Salamon, L. M. (1987). Of market failure, voluntary failure and third-party government: Toward a theory of government-nonprofit relations in the modern welfare state. Nonprofit and Voluntary Sector Quarterly, 16, 29–49. Samuelson, P. A. (1954). The pure theory of public expenditure. The Review of Economics and Statistics, 36(4), 387–389. Sandler, T., & Tschirhart, J. T. (1980). The economic theory of clubs: An evaluative survey. Journal of Economic Literature, 18(4), 1481–1521. Scheerder, J., Willem, A., & Claes, E. (2017). Sport policy systems and sport federations: A crossnational perspective. London: Palgrave Macmillan. Schlesinger, T., Klenk, C., & Nagel, S. (2015). How do sport clubs recruit volunteers? Analyzing and developing a typology of decision-making

processes on recruiting volunteers in sport clubs. Sport Management Review, 18(2), 193–206. Sport and Recreation Alliance (2013). Sports Club Survey. Retrieved March 9, 2017 from www. sportandrecreation.org.uk/policy/researchpublications/sports-club-survey-2013 Stahl, S., Wicker, P., & Breuer, C. (2011). Strukturelle und kontextuelle Spezifika von selbstorganisierten Migrantensportvereinen [Structural and contextual aspects of self-organized migrant sports clubs]. Sport und Gesellschaft, 8(3), 197–231. Taylor, P., Barrett, D., & Nichols, G. (2009). CCPR Survey of Sports Clubs 2009: Project report. London: Central Council of Physical Recreation. Vos, S., Breesch, D., Késenne, S., Van Hoecke, J., Vanreusel, B., & Scheerder, J. (2011). Governmental subsidies and coercive pressures: Evidence from sport clubs and their resource dependencies. European Journal for Sport and Society, 8(4), 257–280. Vos, S., Wicker, P., Breuer, C., & Scheerder, J. (2012). Sports policy systems in regulated Rhineland welfare states: Similarities and differences in financial structures of sports clubs. International Journal of Sport Policy and Politics, 5(1), 55–71. Weisbrod, B. A. (1975). Toward a theory of the voluntary nonprofit sector in a three-sector economy. In E. Phelps (Ed.), Altruism, morality and economic theory (pp. 171–195). New York: Russell Sage Foundation. Weisbrod, B. A. (1988). The nonprofit economy. Cambridge, MA: Harvard University Press. Wicker, P. (2009). Price elasticity in sport clubs: Measurement and empirical findings. Saarbrücken: SVH. Wicker, P. (2017). Finanzierung von Sportvereinen [Financing of sports clubs]. In L. Thieme (Ed.), Der Sportverein – Versuch einer Bilanz (pp. 335–353). Schorndorf: Hofmann. Wicker, P., & Breuer, C. (2009). Sports clubs in Germany. In C. Breuer (Ed.), Sport Development Report 2007/2008: Analysis of the sports clubs’ situation in Germany. Abbreviated version (pp. 5–50). Cologne: Sportverlag Strauß. Wicker, P., & Breuer, C. (2010). Analysis of problems using data mining techniques: Findings from sports clubs in Germany. European Journal for Sport and Society, 7(2), 131–140. Wicker, P., & Breuer, C. (2011). Scarcity of resources in German non-profit sport clubs. Sport Management Review, 14(2), 188–201. Wicker, P., & Breuer, C. (2013). Understanding the importance of organizational resources to explain organizational problems: Evidence from nonprofit sport clubs in Germany. Voluntas, 24, 461–484. Wicker, P., & Breuer, C. (2014). Exploring the organizational capacity and organizational problems of disability sport clubs in Germany using matched pairs analysis. Sport Management Review, 17(1), 23–34.

Sports Clubs in Europe: Organization

Wicker, P., & Breuer, C. (2015). How the economic and financial situation of the community affects sport clubs’ resources: Evidence from multi-level models. International Journal of Financial Studies, 3, 31–48. Wicker, P., Breuer, C., & Hennings, B. (2012a). Understanding the interactions among revenue categories using elasticity measures: Evidence from a longitudinal sample of non-profit sport clubs in Germany. Sport Management Review, 15(3), 318–329. Wicker, P., Breuer, C., Lamprecht, M., & Fischer, A. (2014). Does club size matter? An examination of economies of scale, economies of scope, and organizational problems. Journal of Sport Management, 28, 266–280. Wicker, P., Breuer, C., & von Hanau, T. (2012). Gender effects on organizational problems: Evidence from non-profit sports clubs in Germany. Sex Roles, 66(1), 105–116.

91

Wicker, P., Feiler, S., & Breuer, C. (2013). Organizational mission and revenue diversification among non-profit sports clubs. International Journal of Financial Studies, 1, 119–136. Wicker, P., Longley, N., & Breuer, C. (2015a). Revenue volatility in German nonprofit sports clubs. Nonprofit and Voluntary Sector Quarterly, 44(1), 5–24. Wicker, P., Soebbing, B., Feiler, S., & Breuer, C. (2015). The effect of Porter’s generic strategies on organisational problems of non-profit sport clubs. European Journal for Sport and Society, 12(3), 259–285. Wicker, P., Weingärtner, C., Breuer, C., & Dietl, H. (2012). The effect of a sports institution’s legal structure on sponsorship income: The case of amateur equestrian sports in Germany. International Journal of Sport Finance, 7, 340–357. Young, D. R. (2007). Toward a normative theory of nonprofit finance. In D. R. Young (Ed.), Financing nonprofits: Putting theory into practice (pp. 339– 372). Lanham, MD: AltaMira Press.

10 Volunteering in Sports Clubs and its Impacts Pamela Wicker

INTRODUCTION Non-profit sports clubs in many Western countries, including the United Kingdom (UK) (Taylor, Panagouleas, & Nichols, 2012), Australia (Cuskelly & O’Brien, 2013), Canada (Lasby & Sperling, 2007), Germany (Breuer & Feiler, 2015), and Switzerland (Schlesinger, Klenk, & Nagel, 2015), are largely or exclusively run by volunteers. By definition, sports club volunteers are individuals who work out of free will, receive no remuneration or only a reimbursement of expenses or low pay, and work for the benefit of club members (Cnaan, Handy, & Wadsworth, 1996; Orlowski & Wicker, 2015b). Within sports clubs, two types of volunteer can be distinguished (Breuer & Feiler, 2015): core volunteers, who have a formal position, and secondary volunteers, who help informally and less regularly in the club. From a sports economics perspective, volunteers have attracted research interest because they are different from employees in professional sport organizations. Similarly, non-profit sports clubs that employ volunteers have been studied because they differ from professional and for-profit sport organizations. Starting with the peculiarities of volunteers, the phenomenon that individuals work for nothing or only a small reimbursement that is

not equivalent to a wage rate is of interest from an economic perspective. Following Freeman (1997, p. S141), ‘they must receive greater utility from the first hour of volunteering than from working for wages or from leisure’. Turning to non-profit sports clubs, their legal form has implications for organizational behavior. For example, the organizational goals and means of non-profit sports clubs are contrary to those of for-profit sport organizations (Coates & Wicker, 2016): while the goal of for-profit organizations is to make profit and sport programs are the means to achieve this goal, the goal of nonprofit sports clubs is to provide sport programs for their members and financial resources, including profit, represent the means for goal achievement. Importantly, non-profit sports clubs can make profit, but they are not allowed to distribute profits among members because they must be reinvested into the organization (Hansmann, 1986). Another difference is that for-profit organizations rely on financial resources to employ paid staff, while non-profit sports clubs do not necessarily need financial resources to employ human resources because they can also rely on volunteers. These circumstances affect volunteers that are engaged in sports clubs and must, therefore, be considered when studying volunteers from a sports economics perspective.

Volunteering in Sports Clubs and its Impacts

The purpose of this chapter is to map the field of volunteering in non-profit sports clubs from a sports economic perspective. Specifically, the aim is to review existing research covering the entire range of sports economics, including research contexts, key findings, and theoretical contributions, and to identify lines of enquiry for future studies based on this review. Notably, this chapter focuses on volunteering as an individual activity and not on volunteer management as an organizational task. Also, this review considers exclusively the sports economic perspective and neglects other perspectives, such as sociological, psychological, and management perspectives. Moreover, it focuses on non-profit sports clubs and neglects other contexts of volunteering, such as sports events. Other perspectives and contexts of volunteering and volunteer management have been integrated in earlier reviews by Wicker (2017) as well as Wicker and Hallmann (2013). The present review is an extension of these existing review essays on volunteering in sports through the lens of sports economics.

the volunteering experience (throughput) has not yet been examined for sports club volunteers.1 Hence, this review focuses on antecedents and outcomes of volunteering in sports clubs. It summarizes theoretical approaches, research contexts, and key findings. In doing so, it relies exclusively on papers that have been published in journals with peer review. Table 10.1 gives an overview of topics studied by stage in the process model.

Inputs (Antecedents) of Volunteering This section summarizes the antecedents of volunteering and attempts to answer questions like the following: who produces voluntary work in nonprofit sports clubs? What microeconomic factors affect the decision to volunteer and the time devoted to volunteering? What meso- and macroeconomic factors affect volunteering in sports clubs? Based on what theories have these determinants been advanced?

Microeconomic determinants VOLUNTEERING IN SPORTS CLUBS FROM A SPORTS ECONOMICS PERSPECTIVE The review of existing research is organized around a process model, which encompasses three main stages: input (antecedents of volunteering), throughput (volunteering experience), and outcome (output; consequences of volunteering). Similar process models have already been used in previous review essays to categorize volunteering research (Wicker, 2017; Wilson, 2012). Notably,

Different theoretical approaches have been used to explain microeconomic determinants of volunteering, such as time, money, and human capital. For example, Burgham and Downward (2005) used a heterodox economic approach and argued that a rational-choice framework employed by neoclassical approaches is not appropriate because individual behavior and tastes are linked to wider social behavior, making psychological and sociological aspects relevant as well. Consuming leisure activities like volunteering is not considered a mere choice over consumption of goods or time because it also requires consumption skills which

Table 10.1  Volunteering in sports clubs from a sports economics perspective: overview of topics studied Stage in process model

Topics studied

Concepts/parameters included

Input (antecedents)

Microeconomic determinants

Income Time (working time, employment) Human capital (educational level) Club characteristics (number of members, number of divisions, paid staff) Community characteristics (number of inhabitants) —

Mesoeconomic determinants

Throughput (experiences) Outcome (consequences)

Macroeconomic determinants — Microeconomic outcomes Macroeconomic outcomes

93

Social capital Health-related outcomes (subjective well-being) Monetary value of voluntary work (opportunity costs, replacement costs, societal benefits)

94

THE SAGE HANDBOOK OF SPORTS ECONOMICS

must be acquired through learning. The theoretical framework in Dawson and Downward (2013) was based on Becker (1965, 1974), assuming that individuals seek to maximize utility by allocating time to several activities, including volunteering. Similarly, Schlesinger and Nagel (2013) applied the economic theory of behavior (Becker, 1976), suggesting that individuals invest time and money in activities that raise their well-being. In the Becker (1965) logic, human capital supports the effective allocation of time and money. Hallmann (2015) combined these theoretical approaches and integrated the economic theory of behavior (Becker, 1976) into the heterodox approach. Koutrou and Downward (2016) also used a combination of economic, sociological, and psychological approaches. For instance, they argued that volunteering is a result of the rational balancing of costs and benefits with the aim of increasing well-being. Turning to empirical studies and research contexts, existing studies have examined the role of microeconomic factors, such as income, time, and human capital in individuals’ decision to volunteer and time devoted to volunteering in various research contexts, such as organized sports in Germany (Hallmann, 2015); swimming clubs (Burgham & Downward, 2005), rugby clubs (Koutrou & Downward, 2016), and sports clubs in the UK (Burgham & Downward, 2005; Taylor et al., 2012); and Swiss sports clubs (Schlesinger & Nagel, 2013). Empirical findings for the decision to volunteer are summarized first. Income has been found to have a significant positive effect in most studies (Dawson & Downward, 2013; Hallmann, 2015; Schlesinger & Nagel, 2013; Taylor et  al., 2012), while being insignificant in Burgham and Downward (2005). The role of working time is associated with mixed results: having a significant negative impact in Schlesinger and Nagel (2013) but being insignificant in other studies (Burgham & Downward, 2005; Hallmann, 2015). The findings on employment are more conclusive: individuals working part-time as opposed to fulltime (Burgham & Downward, 2005; Dawson & Downward, 2013; Taylor et  al., 2012), students (Dawson & Downward, 2013; Taylor et al., 2012), and unemployed people (Balish, Rainham, & Blanchard, 2016; Taylor et  al., 2012) are found to be more likely to volunteer. Human capital which has been measured by educational level has inconclusive effects: while some studies reported a significant positive effect on the likelihood of volunteering (Dawson & Downward, 2013; Taylor et al., 2012), the effect can also be significant and negative (Hallmann, 2015) or insignificant (Balish et al., 2016; Burgham & Downward,

2005; Koutrou & Downward, 2016; Schlesinger & Nagel, 2013). The time dedicated to volunteering has also been shown to be affected by several microeconomic determinants. With regard to income, most studies document a significant positive effect (Burgham & Downward, 2005; Dawson & Downward, 2013) with the exception of Hallmann (2015), who reports a significant negative association. Working time has a significant negative effect in one study (Burgham & Downward, 2005), while being insignificant in another (Hallmann, 2015). Concerning employment, some studies show that individuals working part-time spent more time on volunteering (Dawson & Downward, 2013; Taylor et al., 2012), while others report that persons working full-time dedicated more time (Burgham & Downward, 2005). Being a student was positive and significant in one study (Taylor et al., 2012), but insignificant in another (Dawson & Downward, 2013). Unemployed (Taylor et al., 2012) or retired individuals (Dawson & Downward, 2013) were found to spend significantly more time on volunteering. The findings on the role of human capital are again inconsistent: while the effect of educational level has been shown to be significantly positive (Dawson & Downward, 2013) or negative (Taylor et al., 2012) in some studies, others report an insignificant relationship (Burgham & Downward, 2005; Hallmann, 2015).

Meso- and macroeconomic determinants

Starting with theoretical approaches, existing studies have acknowledged that volunteering – although being individual behavior – is not only affected by individual characteristics, but also by mesoeconomic2 factors (e.g., club characteristics, organizational capacity) and macroeconomic factors (community characteristics) (Schlesinger & Nagel, 2013; Wicker & Hallmann, 2013). Thus, individuals are nested within organizations and communities, implying that organizational and community characteristics affect individual behavior as well. Schlesinger and Nagel (2013) conceptualize the relationship between organizational context and volunteering using an adjusted version of the macro-micro-macro scheme originally developed by Esser (1999). This scheme assumes that structural conditions of the sports club (macro; logic of situation) affect individual behavior in the sense that club members make decisions about volunteering by considering their resources and constraints (micro; logic of selection). In turn, these individual decisions affect the personnel structure of the sports club

Volunteering in Sports Clubs and its Impacts

(macro; logic of aggregation) (Schlesinger & Nagel, 2013). This specific application to sports clubs (meso level) makes it more of a mesomicro-meso scheme. A different multi-level model has been developed by Wicker and Hallmann (2013), assuming that the organizational capacity of sports clubs affects individuals’ voluntary engagement. This theoretical model has not yet been tested empirically. A few studies have investigated the impact of organizational factors (Schlesinger & Nagel, 2013) and community characteristics (Balish et  al., 2016) on volunteering in sports clubs. To provide some key findings, Schlesinger and Nagel (2013) document that club size in terms of number of divisions had a positive effect on individuals’ likelihood of volunteering in Swiss sports clubs, while the effect of paid staff is insignificant. Using a multi-country dataset, Balish et al. (2016) provide evidence that individuals are significantly more likely to volunteer in small (but not tiny) communities with between 2,000 and 20,000 inhabitants.

Research gaps and lines of enquiry for future studies

The overview reveals that several studies investigate the economic determinants of volunteering and that this literature has generated a substantial body of knowledge. Notably, meso- and macro­ economic determinants of volunteering have attracted less research interest than microeconomic determinants. Several conclusions can be drawn from the above summary in terms of research gaps and avenues for future studies. First, the overview reveals that the direction of effect and significance of determinants can differ between the decision to volunteer and the amount of time devoted to volunteering. Hence, the findings support the notion that the decision to volunteer (extensive margin) and the decision about the amount of time devoted to volunteering (intensive margin) are two distinct decisions. Second, the effects of the economic factors are not consistent across studies, suggesting that the economic determinants of volunteering may vary to some extent depending on the research context in terms of country, type of sports, examined time period, etc. Future studies should clarify to what extent this is the case. Third, this review shares some of the conclusions by Wicker (2017), which notes that research efforts have focused on the analysis of individual (microeconomic) factors, while studies investigating the role of higher-level determinants, such as meso- and macroeconomic factors, are scarce. Future studies should enhance our understanding

95

of the role of the organizational and community context. This is especially important for management implications because individual characteristics can hardly be changed by club managers. Moreover, Wicker (2017) concludes that existing research has focused on the mass of volunteers in general without distinguishing between different groups of volunteers, such as board members, coaches, and referees/officials. Hence, another rich perspective for future research is to investigate whether and to what extent micro-, meso-, and macroeconomic determinants of volunteering differ between volunteer roles. Fourth, further economic determinants should be studied that inform our understanding of volunteering in sports clubs and generate knowledge that is needed for club management. Specifically, more attention should be paid to role identity within sports clubs. Role identity means that club members can have different roles within the organization at the same time, including consumer, producer, decision-maker, and financier (Horch, 1994). Club members consume sport programs offered by the club. If they perform voluntary work in general or specifically work as voluntary coaches, they are also producers of sport programs in a broader or narrower sense. Given the democratic nature of sports clubs as non-profit organizations, club members are also decision-makers as they have the right to vote on club decisions in the annual members’ assembly following the principle of one person, one vote. Club members are financiers of sport programs through paying membership fees. Thus, the question is how does role identity affect volunteering? The relationship between active participation in sport (consumer) and sport volunteering (producer) has already been explored by Dawson and Downward (2013), who found that volunteering and sport participation are complementary activities. However, they did not explicitly examine the context of sports clubs, but sport in general. On the contrary, Koutrou and Downward (2016) showed that playing rugby was not significantly associated with continued volunteering in rugby clubs. Moreover, Cuskelly and O’Brien (2013) addressed this aspect when examining the transition from playing to volunteering in sports clubs. However, their study included a time-lag, i.e., individuals were first consumers and afterwards producers. They did not have both roles at the same time. Similarly, Burgham and Downward (2005) examined the effect of former participation (active swimmers) on volunteering. They documented a positive effect on both the decision to volunteer and the amount of time devoted to volunteering. While these initial studies gave some

96

THE SAGE HANDBOOK OF SPORTS ECONOMICS

hints, a fruitful avenue for future research is to study the relationship between the two roles of consumer and producer more systematically. The relationship between paying membership fees (financier) and volunteering (producer) should also be explored. From a theoretical perspective, financial resources in terms of revenues from membership fees and human resources in terms of voluntary work can be partially substituted in sports clubs (Coates, Wicker, Feiler, & Breuer, 2014). Partial substitution means that sports clubs need some of both resources to exist, but that it is possible in some cases to get tasks done using either financial resources or volunteers. For example, in a tennis club, the yearly preparation and cleaning of tennis courts at the start of the summer season can either be done by a group of volunteers or by paid maintenance staff. Applying the idea of partial substitution to role identity within sports clubs, the question is whether and to what extent club members would be willing to perform additional voluntary work in order to get a reduction in membership fee. Importantly, the question can also be reversed, i.e., whether and to what extent club members are willing to pay increased membership fees in order to be released from volunteering duties. This aspect has been touched upon by Kiefer (2015), who analyzed the determinants of willingnessto-pay and willingness-to-work among members of equestrian sports clubs but did not set the two concepts in relation to each other. Thus, a fruitful line of enquiry is to systematically examine the relationship between willingness-to-volunteer and willingness-to-pay in sports clubs.

Outcomes (Consequences) of Volunteering This section summarizes the consequences of volunteering and attempts to answer questions such as the following: What are microeconomic outcomes of volunteering? What are macroeconomic outcomes of volunteering? Based on what theories have these outcomes been examined?

Microeconomic outcomes

Some studies have investigated microeconomic outcomes of volunteering which are divided into two categories. The first category includes direct economic outcomes, such as effects on human or social capital. Social capital is also included in this review because it was found to be related with human capital and has, therefore, often been examined together with human capital in sports economics (Barros & Barros, 2005). The second

category represents health-related outcomes, such as effects on individual utility in general and subjective well-being in particular as a form of psychological health. These outcomes can also have economic consequences in terms of public health costs. For example, subjective well-being is considered an early indicator when screening for several psychological diseases, such as anxieties and depression (Koivumaa-Honkanen et  al., 2001), which are associated with significant costs for public health systems (Sobocki, Jönsson, Angst, & Rehnberg, 2006). Starting with direct economic outcomes, existing studies have applied different theoretical approaches to explain the relationship between volunteering and social capital. For example, Harvey, Lévesque, and Donnelly (2007) use a combination of Bourdieu’s (1986) conceptualization of social capital, which considers social capital as resources grounded in network connections and networks. The resulting two forms of social capital are access to resources through social networks and access to people with different social statuses, which in turn facilitates access to the resources those people possess. Kay and Bradbury (2009) were more inclined towards Putnam’s (1995) conceptualization of social capital as networks, norms, and trust. They favored a more structural functionalist approach to social capital with the two core dimensions of relational networks of sociability and civil society. However, when volunteering is considered a form of participation in civil society, which is one dimension of social capital, this theoretical approach is limited by the fact that volunteering and social capital are not regarded as separate concepts. In empirical studies, research efforts have focused on the development of social capital through volunteering in sports clubs (e.g., Harvey et  al., 2007; Kay & Bradbury, 2009). For example, Harvey et  al. (2007) surveyed volunteers in two Canadian sport associations and reported a link between volunteering and social capital. However, they noted that the direction of effect was not clear, i.e., whether volunteering affects the development of social capital or whether the level of social capital facilitates volunteering. Kay and Bradbury (2009) investigated the capacity of youth sport volunteering to contribute to the development of social capital in the UK. Their findings indicate that young people benefit from volunteering in terms of involvement and increased social connectedness. With regard to health-related outcomes, existing studies have acknowledged the relationship between volunteering and subjective well-being (e.g., Downward & Dawson, 2016; Wicker & Frick, 2015).

Volunteering in Sports Clubs and its Impacts

Hence, volunteering cannot only affect subjective well-being through the development of social capital, which, in turn, was found to contribute to subjective well-being (e.g., Orlowski & Wicker, 2015a). There may also be a direct relationship between volunteering and subjective well-being. From a theoretical perspective, volunteering is assumed to produce relational goods which affect well-being (Becchetti, Pelloni, & Rossetti, 2008). Relational goods recognize the fact that much economic activity takes place through encounters rather than market exchanges. These goods are based on interpersonal sharing and include aspects like companionship, emotional support, social approval, solidarity, etc. (Becchetti et  al., 2008). In empirical studies, the impact of volunteering on subjective well-being has not yet been studied systematically in the context of sports clubs. Some existing studies have found that volunteering (including volunteering in sport organizations) (Brown, Hoye, & Nicholson, 2012; Stukas, Hoye, Nicholson, Brown, & Aisbett, 2016) or membership and active involvement in a voluntary sport and recreation organization (Bruni & Stanca, 2008) has a significant positive association with subjective well-being. In the study by Wicker and Frick (2015), voluntary work in sport was not significantly related to subjective wellbeing. Thus, the volunteering measures in previous research were not specifically related to sports clubs. Also, potential issues of causality have not been addressed. Studies outside the sport context attempting to make causal inferences have reported a positive effect of volunteering in general on subjective well-being (Becchetti et al., 2008; Binder & Freytag, 2013; Meier & Stutzer, 2008; Thoits & Hewitt, 2001).

Macroeconomic outcomes

Policy makers and non-profit lobbyists are interested in highlighting the importance and contribution of the voluntary sector when trying to secure public funding. For this purpose, a quantitative assessment of value and the provision of concrete Euro values are considered helpful (Orlowski & Wicker, 2015b). Also, concrete Euro values facilitate comparisons with other industry sectors for which several figures, such as contribution to gross domestic product and value of the labor force, are routinely available. To address this need for concrete Euro values, a handful of studies attempted to assign a monetary value to volunteering in sports clubs (e.g., Davies, 2004; Orlowski & Wicker, 2015b, 2016; Vos et  al., 2012). The monetary values for one hour of voluntary work can be aggregated to give an idea about the

97

economic significance of volunteering in sports clubs in a specific region or country. From a theoretical perspective, the challenge lies in the fact that volunteers – by definition (Cnaan et  al., 1996) – receive no remuneration in terms of a wage rate or salary, which is normally used as an indicator for the value of labor (Orlowski & Wicker, 2015b). Thus, the selection of adequate monetary equivalents is at the core of this stream of research. Three main valuation approaches have been advanced that allow assigning a monetary value to volunteering (for an overview see Orlowski & Wicker, 2015b; Salamon, Sokolowski, & Haddock, 2011). First, the opportunity cost approach considers the time an individual devotes to volunteering and the foregone income. Hence, the wage rate in the individual’s regular job represents the monetary value of one volunteering hour. Second, the replacement cost approach focuses on the task volunteers perform and seeks to find a market equivalent wage for this specific task, which is also referred to as shadow price. While the former two approaches evaluate the input, the third approach, the societal benefits approach, is output-focused and estimates the value of goods or services that have been produced by the volunteer. Importantly, several subapproaches exist within each valuation strategy (Orlowski & Wicker, 2015b). Turning to research contexts and key findings, Davies (2004) suggests a monetary value of approximately £9 for one volunteering hour in sports clubs in the UK when applying the replacement cost approach. A study in Flanders uses the same valuation approach but distinguishes between different functions in sports clubs (Vos et  al., 2012). The respective monetary values are €8.50 (maintenance and support), €12.70 (coaches and instructors), and €13.50 (management). In their study on German sports clubs, Orlowski and Wicker (2015b) compare four valuation approaches empirically. The standard opportunity cost and replacement cost approaches yielded a monetary value of approximately €14 for one hour of volunteering. When applying the leisure-adjusted opportunity cost approach, which includes only those volunteering hours where the individual would work in his/her regular job and excludes volunteering hours that are substitutes of leisure time, the monetary value drops to under €2. This drop is because volunteers reported that they would volunteer during their leisure time and not work time for 91.7% of all voluntary activities. The volunteer judgment replacement cost approach, where the volunteer assigns a monetary value to the task he/she performs, yields a monetary value of approximately €10 (Orlowski & Wicker, 2015b). Another German study applied

98

THE SAGE HANDBOOK OF SPORTS ECONOMICS

the societal benefits approach to voluntary coaches in sports clubs and regarded participants of training sessions as beneficiaries (Orlowski & Wicker, 2016). The resulting monetary values of one volunteering hour vary depending on the elicitation method, ranging from approximately €18 (dichotomous choice and payment ladder) to €67 (open question).

Research gaps and lines of enquiry for future studies

The review indicates that some studies analyze economic outcomes of volunteering and that this literature has generated a solid body of knowledge. It also reveals that microeconomic outcomes have been studied less compared with macroeconomic outcomes. At least three conclusions can be derived in terms of research gaps and avenues for future studies. First, the microeconomic outcomes that have been addressed in existing studies should be studied more systematically. While the literature gives some hints about the effect of volunteering in sports clubs on the development of social capital and on subjective well-being, these relationships should be examined more systematically in the sense that that attention needs to be paid to reverse causality and endogeneity, and actions should be taken to address these issues. One way to deal with them is to use instrumental variable techniques – similar to research examining the effect of sport participation on subjective wellbeing (e.g., Downward & Dawson, 2016; Huang & Humphreys, 2012; Wicker & Frick, 2015). Research taking these aspects into account would expand our understanding of whether volunteering contributes to social capital and well-being or whether volunteers are simply happier and more socially connected people. Also, rather than relying on measures of volunteering in general, studies should use measures that specifically capture volunteering in sports clubs to allow analyzing the extent to which the specific context of volunteering matters. Second, given that research efforts have focused on the development of social capital, further types of microeconomic outcomes should be analyzed in more detail, including effects on physical health, human capital development, and various labor market outcomes. The review by Wilson and Musick (1999) reveals that these outcomes have already been studied from a nonsport perspective. Thus, examining the effect of volunteering in sports clubs on physical health outcomes represents a fruitful avenue for future sports economics research. Another rich perspective for future studies would be to investigate

the impact of volunteering in sports clubs on the development of human capital in more detail. This topic has only been a side note in an Australian study which was not dedicated specifically to the analysis of volunteering (Darcy, Maxwell, Edwards, Onyx, & Sherker, 2014). However, knowledge is needed about how the amount of time dedicated to volunteering, and what volunteer roles, affect an individual’s stock of human capital. A natural extension of such studies would be to examine the impact of volunteering in sports clubs on labor market outcomes. While this topic has not yet been studied from a sports economics perspective, existing non-sport studies have shown a significant positive effect of volunteering in other contexts on various labor market outcomes, such as hiring recommendations (Chen, Huang, & Lee, 2011), applicant employability (Cole, Rubin, Field, & Giles, 2007), and earnings (Sauer, 2015), although some studies do not support this relationship (e.g., Prouteau & Wolff, 2006). However, information about how volunteering in sports clubs affects labor market outcomes would provide valuable knowledge for sports clubs in an effort to recruit and retain volunteers and also for public bodies to support sports clubs and club volunteers. Third, the overview reveals that most existing studies examining macroeconomic outcomes, such as the monetary value of voluntary work, have applied the standard opportunity cost or replacement cost approaches, while the societal benefits approach has attracted less research. This neglect is surprising because beneficiaries of volunteering in sports clubs can be more easily identified than in other organizational contexts where non-members benefit from volunteering. For example, members consuming training sessions represent the direct beneficiaries of voluntary coaches and instructors in sports clubs. Also, spectators, athletes, and coaches are the direct beneficiaries of voluntary referees, judges, and officials. Hence, sports clubs represent an ideal setting for applying the societal benefits approach. Another rich perspective for future studies would be to move beyond the standard opportunity cost and replacement cost approaches and apply, for example, the volunteer judgment versions of the two approaches (Orlowski & Wicker, 2015b). These approaches, which are based on the volunteers evaluating the value of their time or task themselves, may produce more reasonable figures than the standard approaches which tend to overestimate monetary values (Salamon et al., 2011). Moreover, comparative studies should be conducted to put valuation approaches and empirical figures into perspective.

Volunteering in Sports Clubs and its Impacts

CONCLUSIONS This chapter set out to review the literature examining volunteering in sports clubs from a sports economics perspective and to provide lines of enquiry for future research. When taking a process model (input, throughput, outcome) as the basis, this review reveals that the input stage – determinants of volunteering – continues to attract the most research attention. However, research efforts have focused on microeconomic determinants, while meso- and macroeconomic determinants have received less attention. An analysis of the throughput stage – volunteering experience – has been neglected from a sports economics perspective. Research interest in the third stage – outcomes of volunteering – has increased over the last years, particularly with respect to macroeconomic outcomes. These conclusions resemble those made by earlier reviews on volunteering research in general (Wilson, 2012) and in sport (Wicker, 2017). However, much less research has been done on that topic from a sports economics perspective. Consequently, this review identifies research gaps and suggests fruitful avenues for future sports economics studies.

Notes 1  Hallmann and Zehrer (2016) studied the effects of perceived costs and benefits on satisfaction with the volunteering experience, but they focused on non-sport organizations. 2  Within a multi-level framework, the meso level typically encompasses organizations and institutions. It is located between the micro level (individuals) and macro level (societal level).

REFERENCES Balish, S., Rainham, D., & Blanchard, C. (2016). Volunteering in sport is more prevalent in small (but not tiny) communities: Insights from 19 countries. International Journal of Sport and Exercise Psychology (in press). DOI: 10.1080/1612197X.2015.1121510 Barros, P. C., & Barros, C. (2005). The role of human and social capital in the earnings of sports administrators: A case study of Madeira Island. European Sport Management Quarterly, 5(1), 47–62. Becchetti, L., Pelloni, A., & Rossetti, F. (2008). Relational goods, sociability, and happiness. Kyklos, 61(3), 343–363. Becker, G. S. (1965). A theory of the allocation of time. Economic Journal, 75(3), 493–517.

99

Becker, G. S. (1974). A theory of social interactions. Journal of Political Economy, 82(6), 1063–1093. Becker, G. S. (1976). The economic approach to human behavior. Chicago, IL: University of Chicago Press. Binder, M., & Freytag, A. (2013). Volunteering, subjective well-being and public policy. Journal of Economic Psychology, 34, 97–119. Bourdieu, P. (1986). The forms of capital. In S. Baron, J. Field & T. Schuller (Eds.), Social capital – critical perspectives (pp. 1–38). Oxford, UK: Oxford University Press. Breuer, C., & Feiler, S. (2015). Sport Development Report 2013/2014: Analysis of the situation of sports clubs in Germany. Abbreviated version. Cologne: Strauß. Brown, K. M., Hoye, R., & Nicholson, M. (2012). Selfesteem, self-efficacy, and social connectedness as mediators of the relationship between volunteering and well-being. Journal of Social Service Research, 38(4), 468–483. Bruni, L., & Stanca, L. (2008). Watching alone: Relational goods, television and happiness. Journal of Economic Behavior & Organization, 65(3–4), 506–528. Burgham, M., & Downward, P. (2005). Why volunteer, time to volunteer? A case study from swimming. Managing Leisure, 10(2), 79–93. Chen, C., Huang, Y., & Lee, M. (2011). Test of a model linking applicant résumé information and hiring recommendations. International Journal of Selection and Assessment, 19(4), 374–387. Cnaan, R. A., Handy, F., & Wadsworth, M. (1996). Defining who is a volunteer: Conceptual and empirical considerations. Nonprofit and Voluntary Sector Quarterly, 25, 364–383. Coates, D., & Wicker, P. (2016). Financial management. In R. Hoye & M. M. Parent (Eds.), The SAGE handbook of sport management (pp. 117–137). London: Sage. Coates, D., Wicker, P., Feiler, S., & Breuer, C. (2014). A bivariate probit examination of financial and volunteer problems of non-profit sport clubs. International Journal of Sport Finance, 9(3), 130–148. Cole, M. S., Rubin, R. S., Field, H. S., & Giles, W. F. (2007). Recruiters’ perceptions and use of applicant résumé information: Screening the recent graduate. Applied Psychology: An International Review, 25(2), 319–343. Cuskelly, G., & O’Brien, W. (2013). Changing roles: Applying continuity theory to understanding the transition from playing to volunteering in community sport. European Sport Management Quarterly, 13(1), 54–75. Darcy, S., Maxwell, H., Edwards, M., Onyx, J., & Sherker, S. (2014). More than a sport and volunteers organization: Investigating social capital

100

THE SAGE HANDBOOK OF SPORTS ECONOMICS

development in a sporting organisation. Sport Management Review, 17, 395–406. Davies, L. E. (2004). Valuing the voluntary sector in sport: Rethinking economic analysis. Leisure Studies, 23(4), 347–364. Dawson, P., & Downward, P. (2013). The relationship between participation in sport and sport volunteering: An economic analysis. International Journal of Sport Finance, 8, 75–92. Downward, P., & Dawson, P. (2016). Is it pleasure or health from leisure that we benefit from most? An analysis of well-being alternatives and implications for policy. Social Indicators Research, 126, 443–465. Esser, H. (1999). Soziologie: Spezielle Grundlagen. Vol. 1: Situationslogik und Handeln. Frankfurt am Main: Campus. Freeman, R. B. (1997). Working for nothing: The supply of volunteer labor. Journal of Labor Economics, 15(1), S140–S166. Hallmann, K. (2015). Modelling the decision to volunteer in organised sports. Sport Management Review, 18, 448–463. Hallmann, K., & Zehrer, A. (2016). How do perceived benefits and costs predict volunteers’ satisfaction? Voluntas, 27, 746–767. Hansmann, H. (1986). The role of nonprofit enterprise. In S. Rose-Ackerman (Ed.), The economics of nonprofit institutions: Studies in structure and policy (pp. 57–84). New York: Oxford University Press. Harvey, J., Lévesque, M., & Donnelly, P. (2007). Sport volunteerism and social capital. Sociology of Sport Journal, 24, 206–223. Horch, H.-D. (1994). On the socio-economics of voluntary associations. Voluntas, 5, 219–230. Huang, H., & Humphreys, B. R. (2012). Sports participation and happiness: Evidence from US microdata. Journal of Economic Psychology, 33, 776–793. Kay, T., & Bradbury, S. (2009). Youth sport volunteering: Developing social capital? Sport, Education and Society, 14(1), 121–140. Kiefer, S. (2015). Are riding club members willing to pay or work for overall quality improvement? Managing Sport and Leisure, 20(2), 100–116. Koivumaa-Honkanen, H., Honkanen, R., Antikainen, R., Hintikka, J., Laukkanen, E., Honkalampi, K., & Viinamäki, H. (2001). Self-reported life satisfaction and recovery from depression in a 1-year prospective study. Acta Psychiatrica Scandinavica, 103, 38–44. Koutrou, N., & Downward, P. (2016). Event and club volunteer potential: The case of women’s rugby in England. International Journal of Sport Policy and Politics, 8(2), 207–230. Lasby, D., & Sperling, J. (2007). Understanding the capacity of Ontario sports and recreation organizations. Toronto, Ontario: Imagine Canada. Meier, S., & Stutzer, A. (2008). Is volunteering rewarding in itself? Economica, 75, 39–59.

Orlowski, J., & Wicker, P. (2015a). The monetary value of social capital. Journal of Behavioral and Experimental Economics, 57, 26–36. Orlowski, J., & Wicker, P. (2015b). The monetary value of voluntary work – conceptual and empirical comparisons. Voluntas, 26(6), 2671–2693. Orlowski, J., & Wicker, P. (2016). The monetary value of voluntary coaching: An output-based approach. International Journal of Sport Finance, 11(4), 310–326. Prouteau, L., & Wolff, F. (2006). Does volunteer work pay off in the labor market? Journal of Socio-­ Economics, 35, 992–1013. Putnam, R. (1995). Bowling alone: America’s declining social capital. Journal of Democracy, 6, 65–78. Salamon, L. M., Sokolowski, S. W., & Haddock, M. A. (2011). Measuring the economic value of volunteer work globally: Concepts, estimates, and a roadmap to the future. Annals of Public and Cooperative Economics, 82(3), 217–252. Sauer, R. N. (2015). Does it pay for women to volunteer? International Economic Review, 56(2), 537–564. Schlesinger, T., Klenk, C., & Nagel, S. (2015). How do sport clubs recruit volunteers? Analyzing and developing a typology of decision-making processes on recruiting volunteers in sport clubs. Sport Management Review, 18, 193–206. Schlesinger, T., & Nagel, S. (2013). Who will volunteer? Analysing individual and structural factors of volunteering in Swiss sports clubs. European Journal of Sport Science, 13, 707–715. Sobocki, P., Jönsson, B., Angst, J., & Rehnberg, C. (2006). Cost of depression in Europe. Journal of Mental Health Policy and Economics, 9(2), 87–98. Stukas, A. A., Hoye, R., Nicholson, M., Brown, K. N., & Aisbett, L. (2016). Motivations to volunteer and their associations with volunteers’ well-being. Nonprofit and Voluntary Sector Quarterly, 14(1), 112–132. Taylor, P. D., Panagouleas, T., & Nichols, G. (2012). Determinants of sports volunteering and sports volunteer time in England. International Journal of Sport Policy and Politics, 4(2), 201–220. Thoits, P. A., & Hewitt, L. N. (2001). Volunteer work and well-being. Journal of Health and Social Behavior, 42, 115–131. Vos, S., Breesch, D., Késenne, S., Lagae, W., Hoecke, J. V., Vanreusel, B., & Scheerder, J. (2012). The value of human resources in non-public sports providers: The importance of volunteers in nonprofit sports clubs versus professionals in for-profit fitness and health clubs. International Journal of Sport Management and Marketing, 11, 3–25. Wicker, P. (2017). Volunteerism and volunteer management in sport. Sport Management Review (in press). DOI: 10.1016/j.smr.2017.01.001 Wicker, P., & Frick, B. (2015). The relationship between intensity and duration of physical activity

Volunteering in Sports Clubs and its Impacts

and subjective well-being. European Journal of Public Health, 25, 868–872. Wicker, P., & Hallmann, K. (2013). A multi-level framework for investigating the engagement of sport volunteers. European Sport Management Quarterly, 13, 110–139.

101

Wilson, J. (2012). Volunteerism research: A review essay. Nonprofit and Voluntary Sector Quarterly, 41(2), 176–212. Wilson, J., & Musick, M. (1999). The effects of volunteering on the volunteer. Law and Contemporary Problems, 62(4), 141–168.

11 The Role of Money and Time Donations in the Supply of Amateur Sport Kirstin Hallmann and Lea Rossi

INTRODUCTION Both volunteering time and giving money as donations are an important part of the global economy. Both forms of behaviour are forms of altruism, philanthropy or charitable giving (Bekkers & Wiepking, 2011), and while they may be given for the donor’s benefit, they nonetheless incur costs and are notably undertaken for the recipient’s benefit (West, Griffin, & Gardner, 2007). Money donations are very popular across the world, but there are differences across countries. The most common contexts for donating money are international aid, social welfare and religion (Observatoire de la Fondation de France, 2015). In Europe, the nine biggest countries (Belgium, France, Germany, Italy, Netherlands, Poland, Spain, Sweden, Switzerland, and the United Kingdom) accumulate €24.4 billion annual monetary donations (Observatoire de la Fondation de France, 2015). The value of time donations is harder to estimate as time donations have no straightforward market value. Data on voluntary engagement on a global level are relatively scarce as many countries do not investigate volunteering at all, or differing results are obtained (Salamon, Sokolowski, & Haddock, 2011). However, estimates on the global value of

voluntary work range between $1.1 trillion (Salamon & Anheier, 1999) and $2.2 trillion (Salamon, 2010). Country-specific estimates range between £38.9 billion in the UK (Low, Butt, Ellis, & Davis Smith, 2007) and $179.2 billion in the United States (McKeever, 2015). The question arises as to why time and money donations are so important for amateur sport. First, time donations in the form of voluntary work are crucial for the functioning of non-profit sport organisations and sport events (Adams & Deane, 2009; Fairley, Kellett, & Green, 2007; Vos et  al., 2012). Many positions in non-profit sport organisations are filled by volunteers – both on the administrative level (i.e. president, board member, accountant, event organiser) and the executive level (i.e. coaches, referees; Breuer & Feiler, 2017). Empirical evidence using different measures has shown that the monetary value of one hour of volunteer work in German sport clubs is estimated to range between €1.78 and €14.27 (Orlowski & Wicker, 2015). In total, time donations by volunteers and voluntary workers organised in sport clubs account for €4.1 billion of annual added value to the German economy (Breuer & Feiler, 2017). Second, a stable financial situation is just as important for non-profit sport organisations to provide affordable sport programmes to their members. One major source

The Role of Money and Time Donations in the Supply of Amateur Sport

of income, apart from membership fees, are money donations from the community (Feiler, Wicker, & Breuer, 2014). Moreover, increases in money donations can interact with increases in revenues from sport supply (e.g. membership and service fees), through a so-called crowd-in effect (Wicker, Breuer, & Hennigs, 2012). Generating a substantial amount of money donations is therefore important for sport clubs. In addition, amateur sport clubs rely heavily on time and money donations as well as non-monetary donations in organising sport events (Kerwin, Warner, Walker, & Stevens, 2015). Even though most research has focused on major sport events, this chapter will focus on amateur sport clubs and their respective events as they represent non-profit organisations. This chapter will provide an overview of the different theoretical approaches to altruism, including economics, psychology, sociology and socio-biology (Haski-Leventhal, 2009). The focus is put on economic theory and the individual level as most studies have relied on survey data from individual participants. In addition, a review of the current state-of-the-art literature on time and money donations will be given – including both donations of time and money in general and with a focus on amateur sport. The chapter concludes with an overview of the implications and limitations of the research and gives a short outlook on future research perspectives.

THEORETICAL FRAMEWORK Economics From an economic perspective, altruistic behaviour is based on a model of utility maximisation (i.e. Becker, 1976, 1981; Frank, 1987; Hammond, 1987). This model implies that an individual chooses the best alternative which maximises their utility (Aleskerov, Bouyssou, & Monjardet, 2007). At first glance, this concept contradicts the idea of altruism as a selfless act. However, economists have integrated altruism in the model by assuming that an altruistic person aims to not only maximise his/her own utility, but strives to maximise the utility of others as well (Becker, 1976). Thus, the utility maximisation function of an altruistic person includes his own utility i plus the utility of a wider population j (Hammond, 1987). Other major theoretical approaches which are based on the utility maximisation model and applied to altruism are the public goods model, the private consumption model and the human capital model (Emrich & Pierdzioch, 2015). The

103

public goods model assumes that public goods are created through voluntary work. An individual derives utility from his/her own private consumption and the provision of public goods. In this model, altruism explains the motivation of voluntary work as voluntary work and the resulting public good does not only provide benefits to the volunteer but also to other individuals (Wildman & Hollingsworth, 2009). If altruism were not existent, volunteers would not contribute to the creation of public goods. In contrast, the private consumption model focuses on the private benefits a person receives through voluntary work. This includes an increase in one’s social status and a better reputation in the social environment (Bauer, Bredtmann, & Schmidt, 2013). Last but not least, the human capital model assumes a link between voluntary work and future labour income as volunteers increase their social capital through voluntary work. Volunteering enhances an individual’s job-related skills and helps to create a social network which in turn serves as a catalyst for future job opportunities (Emrich & Pierdzioch, 2015). The economic concept of altruism has also been applied to different contexts (i.e. Collard, 1992; Kurz, 1978; Unger, 1991). Becker (1981) studied altruism in the family context and found that altruistic behaviour is more common there than in the marketplace as it is a more effective behaviour in families. In his theory, an altruistic benefactor supports a selfish beneficiary to maximise utility in the ‘rotten kid theorem’: even selfish family members try to maximise family income as it maximises both their own income as well as the family’s utility (Becker, 1981). Other academics have challenged the idea of pure altruism, stating that altruistic behaviour is not undertaken out of pure goodwill but to satisfy selfish intentions (Wichardt, 2009). Andreoni (1990) introduced the concept of ‘impure altruism’, in which people act selflessly to increase their social status and receive a so-called ‘warm-glow effect’ (Andreoni, 1990). This implies that people care about what their surrounding environment thinks about them and act accordingly to receive appraisal. Empirical evidence has shown that the social component does play a role in altruistic behaviour (Cappellari, Ghinetti, & Turati, 2011).

Psychology From a psychological perspective, altruism is defined as ‘intentional, voluntary behaviour that benefits others and is not performed with the expectation of receiving external rewards or avoiding external punishments or aversive stimuli’

104

THE SAGE HANDBOOK OF SPORTS ECONOMICS

(Eisenberg & Miller, 1987, p. 292). The psychological assumption is that humans are born egoistically and learn altruistic behaviour over time (Haski-Leventhal, 2009). A theory which has been applied frequently is the Social Learning Theory developed by Bandura (1971). It assumes that behaviour is learned through observation of the social environment, which includes the family, education and mass media (Bandura, 1971). The frequency of prosocial (i.e. altruistic) or antisocial behaviour is thus a function of the people’s social learning experiences (Rushton, 1982). This implies that governments can steer the behaviour of the people through incentivising family time spent together and stressing the importance of altruism in the educational sector. However, this argumentation can be challenged from a cognitive developmental view which focuses on the way humans think rather than how they learn (Krebs, 1982). While social learning theorists focus on observable external stimuli and how they lead to a response, cognitive approaches acknowledge that internal assessments and evaluations as well as moral judgements as to why one helps are crucial (Krebs, 1982). Concerning the motivation for altruistic behaviour, different theories have been employed. Selfdetermination theory assumes that individuals are motivated through intrinsic and extrinsic motivation (Deci & Ryan, 1985). It applies to altruism as a person can either perform philanthropic behaviour for the sake of doing something good (intrinsic) or to achieve some hidden, more selfish agenda (extrinsic). Ajzen’s (1985) theory of planned behaviour states that human behaviour is guided by three kinds of considerations: behavioural beliefs, which lead to an attitude towards the behaviour; normative beliefs, which result into a subjective norm; and control beliefs, which induce a perceived behavioural control (Ajzen, 1985). A combination of these three constructs (attitude, subjective norm and perceived behavioural control) leads to the intention of undertaking an action. Thus, it can be assumed that volunteers who have a positive attitude towards their altruistic behaviour and feel that it fits to their subjective norm are more likely to undertake this kind of behaviour.

species in which individuals give warning signs which in turn endangers them individually (Hoffman, 1981). Two concepts are important in this regard: kin-selection and group-selection. On the one hand, kin-selection implies that humans protect their close ones as they share similar genetics. Thus, the survival of more humans with similar genetics increases the chances that certain characteristics survive in evolution (Simon, 1992). On the other hand, humans are stronger in a group than individually to protect themselves from natural enemies like animals. Altruistic behaviour helps to sustain large group sizes which then ensures individual survival (Hoffman, 1981). Even though socio-biologic assumptions themselves seem rather hard to apply to the sporting context, there have been attempts to relate the socio-biological concept of altruism to economics and psychology (Hoffman, 1981; Simon, 1992).

Sociology Sociologists mainly focus on socio-demographic factors influencing altruistic behaviour. Altruism is defined as acting according to norms, expectations and values (Haski-Leventhal, 2009). However, studies on an institutional and societal level have been performed as well (i.e. Healy, 2004; Künemund & Schupp, 2007; Lim & Laurence, 2015; McCulloch, Mohan, & Smith, 2012). Levels of altruism in different cultures can be explained through social exchange and the principle of the Maussian gift (Sleeboom-Faulkner, 2014). It implies four conditions determining the success of a gift: timing, structure, custom and strategy. Analysing these conditions can be crucial in understanding differences of altruistic behaviour in different cultures (Sleeboom-Faulkner, 2014). Another theoretical approach is social integration theory. It assumes that members of a group who can provide valued services to the rest of the group are respected and integrated (Blau, 1960). Thus, individuals who act in the sense of others tend to be more socially integrated which in turn leads to the achievement of social goals. Rotolo, Wilson and Hughes (2010) have used this approach to study the influence of home ownership on the likelihood to volunteer.

Socio-biology In contrast to the psychological approach, sociobiologists see altruism as a human characteristic which has evolved to ensure self-survival and the survival of close others when they are struggling for existence (Haski-Leventhal, 2009). An example of altruistic behaviour in socio-biology are

LITERATURE REVIEW Altruistic behaviour which results in donating time and money can be studied from different perspectives. One major distinction in this

The Role of Money and Time Donations in the Supply of Amateur Sport

context is between the individual level and the organisational level. Wicker and Hallmann (2013) introduced a model for sport volunteering, which includes exactly these two dimensions. The individual level looks at differences in volunteering based on economic, demographic, social and psychological indicators and is nurtured from economics, psychology, and sociology in terms of theory, as outlined in the previous chapter. On the other hand, the sport organisation is assumed to influence volunteering through human resources, financial, planning and development, network and relationship and infrastructure capacities (Wicker & Hallmann, 2013). In addition to the model by Wicker and Hallmann (2013), Künemund and Schupp (2007) developed a model including micro-level (individual), meso-level (clubs, associations), macro-level (institutional framework), and political factors. The following literature review will be based on a combination of these two models (see Figure 11.1): at the individual level, differences in time and money donations, as well as evidence for organisational impact on levels of volunteering, are examined. In addition, a third dimension to include differences of volunteering on the regional and national level is added.

Individual Level Studies taking an economic perspective have addressed the interdependent nature of time and money donations: Are these two types of donations complementary or substitutes? The majority of studies find that time and money donations are complementary – people who donate money or

105

time are more likely to donate also in other ways (Brown & Lankford, 1992; Cappellari et al., 2011; Priller & Schupp, 2011), but there is also evidence for a reverse relationship (Feldman, 2010; Yang, 2016). Hartmann and Werding (2012) argue that people increase their money donations when they volunteer in an organisation as they are better informed about the organisation and thus have higher levels of trust. On the contrary, Lilley and Slonim (2014) find time and money donations to be net substitutes. Their study includes the warmglow effect as a driver for donations and finds that time donations are stronger when motivated by warm-glow than money donations. Many studies investigate differences across demographic characteristics, in particular gender and age. Decker, Winter, Brähler and Beutel (2008) find that women are more altruistically motivated to donate (in this case potential organ donation), while men are in general more willing to donate and also accept financial incentives to increase donor rates. However, empirical research in sports produces differing results. While Hallmann (2015) supports the general results that men volunteer more, Burgham and Downward (2005) find that gender has no influence on the decision to volunteer but women who do volunteer spend more time on it than men. So far, no studies seemingly have investigated gender differences with regards to money donations in the sporting context. With regard to age, studies conducted in Western Europe typically find a positive relationship between age and donation, i.e. older people are more engaged in money and time donations than younger people (Baruch, May, & Yu, 2016; Cappellari et  al., 2011; Kalargirou et  al., 2015). Lee and Chang (2008) also discovered that younger people were more likely to donate than

Figure 11.1  Multi-level framework on the different levels of time and money donations (cf. Künemund & Schupp, 2007; Wicker & Hallmann, 2013)

106

THE SAGE HANDBOOK OF SPORTS ECONOMICS

older people in Taiwan, which is in contrast to some Western studies. However, more research is needed to confirm the different donation characteristics across cultures. Strigas and Jackson (2003) find that volunteers at a charity sports event were predominantly between 35 and 50 years old. In contrast, Dote, Cramer, Diets and Grimm (2006) discovered that college students between the ages of 18 and 25 years were more likely to volunteer at sport events than other age groups. The findings of Kim, Zhang and Connaughton (2010) support the results from Dote et al. (2006). Economic indicators are often considered in donating behaviour. The relationship seems obvious: the more money a person has at his/her disposal, the more he/she can spend on charity. Studies which controlled for income generally find support for this relationship (Rotolo et  al., 2010; Simmons & Emanuele, 2007). However, the relationship between income and other forms of donations (i.e. volunteering) might be different. Unger (1991) finds that socio-economic status and disposable time have little influence on volunteering. However, socio-economic status seems to have a negative influence on disposable time. This relationship between socio-economic status and disposable time has been supported for sports (Burgham & Downward, 2005; Dawson & Downward, 2013). However, other studies have found no significant effect of income on the time allocated to volunteering in sports (Hallmann, 2015; Taylor, Panagouleas, & Nichols, 2012). Yet again, no studies seem to have investigated the effect of income on money donations in the sporting context. In general, it is found that time and money donations increase with a higher educational level (i.e. Cappellari et al., 2011; Priller & Schupp, 2011; Smith, 2006). This relationship exists in the sporting context as well (Taylor et al., 2012). Social indicators are often studied, drawing upon the concept of social capital and the ‘warmglow’ effect. Wildman and Holingsworth (2009) find social capital to be an important driver in the decision to donate blood. Results of Cappellari et al. (2011) support this conclusion and they confirm the ‘warm-glow’ effect as people seem to care substantially about what others think about them and thus engage in prosocial behaviour. Applying the concept of social integration to the sporting context leads to interesting results. Hallmann (2015) found that the motive of shaping society had a negative influence on the decision to volunteer but was positively correlated with the time committed to volunteering. This might lead to the idea that people who are not engaged in sports are sceptical of the importance of sport for society and thus engage in other institutions. However, people

who are already volunteering in sports organisations might see the impact their work has on others and thus decide to put more effort into their work to help shape society. Regarding psychological indicators, different concepts have been used, for example commitment, identification or satisfaction. While altruistic and empathic motivation seem to play no role in the commitment to volunteering (Veludode-Oliveira, Pallister, & Foxall, 2015), life satisfaction was found to positively influence the decision to volunteer (Kumnig et al., 2014). In the sporting context, satisfaction can be an important driver for event-volunteers to return as volunteers in consecutive years (Lee, Reisinger, Kim, & Yoon, 2014). Furthermore, the concept of identification is important in the sporting context. Joo, Koo and Fink (2016) studied the effect of team identification on attitude towards a cause-related marketing (CRM) campaign and a new sponsorship of a team. They find that fans who identify less with a team considered an altruistic reason for a CRM campaign as more important than fans who identified strongly with the team. This can be related to time and money donations as teams can try to address low identifier fans with a more altruistic reasoning to convince them to get engaged.

Institutional Level In contrast to the individual level, literature on the institutional level is less advanced. Moreover, literature corresponding to the key capacities introduced in the conceptual framework (cf. Künemund & Schupp, 2007; Wicker & Hallmann, 2013) is scarce. The human resource capacity of sporting organisations is mostly characterised by voluntary staff and paid staff. Cuskelly (1996) finds that sporting organisations that employed a paid administrator had higher levels of organisational commitment in volunteers. In addition, Feiler, Wicker and Breuer (2014) detected that sport clubs with paid staff generated higher levels of donations than sport clubs without paid staff. This result contradicted the hypothesis that donors would mistrust clubs who employed paid staff. A potential explanation could be that clubs which are run more professionally act more stringently towards their goals and missions, which is, in turn, valued by donors (Feiler et al., 2014). Another resource which has been studied in volunteering research is infrastructure. In his study of organ procurement organisation in the United States, Healy (2004) found the logistics of procurement to have a strong influence on the donor rate. The better the infrastructure for organ donations is the more donations can

The Role of Money and Time Donations in the Supply of Amateur Sport

be detected. In addition, the author points out that organisational factors and the individual’s capacity for altruistic actions are inseparable, as they are two aspects of the same process. The networking and relationship capacity of organisations has been studied by Schlesinger and Nagel (2013) in a survey of Swiss sports volunteers. It is found that volunteering in rural sports clubs is more probable due to their stronger community character. In addition, Chen and Lee (2014) explore the difference between public and non-profit employees on their level of volunteering. They find that volunteers in the non-profit sector are more likely to engage in voluntary activities as these organisations are typically smaller and thus more focused on personal relationships. Moreover, they found that employees in the nonprofit sector are often engaged in political parties as well, and this in turn increases the probability to volunteer. The authors conclude that altruism is not the only factor determining volunteering as jobs in both industry sectors (i.e. public and nonprofit) are taken for altruistic reasons. Other studies have focused on the relationship between the individual and the organisation, especially the concepts of service quality and value congruence. In their study of German sports clubs, Feiler et  al. (2014) investigated the influence of clubs’ values on donations. They find that clubs with values like social aspects, companionship and conviviality generate higher levels of donations. Findings by Schlesinger and Nagel (2016) supported these results. In addition, van Schie, Güntert, Oostlander and Wehner (2014) discovered that, on the organisational level, volunteers are mainly motivated by the perceived value congruence. This means that volunteers who believe that the organisation shares the same values as they do will be more motivated to donate their time and effort to this organisation. On the other hand, the authors find that in general, volunteers are motivated by the task they are doing: the more motivational potential (i.e. the more attractive) the task has, the more motivated is the volunteer to complete the task. To sum up, research on the organisational level of time and money donations has been scarce. The studies that exist relate to some extent to the resources that were identified in the conceptual framework. Thus, more research is needed to get a complete understanding on the impact of organisational factors on time and money donations.

National and Regional Level Donations on the national and regional level have often been studied to compare different levels of donating and to detect further determinants of

107

donations across regions. The concept that has been applied most frequently is social capital (i.e. Kennedy, Kawachi, & Brainerd, 1998; Miller, Scheffler, Lam, Rosenberg, & Rupp, 2006; Scheepers, te Grotenhuis, & Gelissen, 2002). Kääriäinen and Lehtonen (2006) discovered that the relationship between welfare states and social capital is diverse: while social networking based on family ties was stronger in Mediterranean and post-socialist regimes, voluntary and civic engagement and generalised-trust were stronger in social-democratic and liberal welfare state regimes. Furthermore, state institutions are found to have an impact on social participation but this impact differs based on different countries and state regimes (van der Meer, Scheepers, & te Grotenhuis, 2009). Another stream of research has focused on economic factors on the national and regional level which influence volunteering. Lim and Laurence (2015) studied the relationship between economic recessions and levels of volunteering. They find that volunteering declines when the economy is facing a recession. This result holds especially true for informal helping outside an organisation. Furthermore, the decline in volunteering is larger in communities that are more socially and economically disadvantaged as they have weaker norms of social trust. Looking at a more regional level, Dury et al. (2016) find that neighbourhood connectedness, neighbourhood satisfaction, home ownership and presence of services predict voluntary engagement at older ages. On the other hand, younger people are found to be more inclined to volunteer to improve employability in municipalities with area-level insecurities (Chum et al., 2015). All in all, the overview of studies on the national and regional determinants of volunteering does not lead to clear results. It can be said that national and regional institutions generally have an influence on community participation. However, the extent of this impact as well as the kinds of participation differs across the countries and regions. Yet, there is a lack of sport-related studies.

LIMITATIONS AND FUTURE RESEARCH While many studies have been conducted on the individual determinants of time and money donations, there is still a lack of literature on the institutional and national level. It is especially the case that demographic indicators have been studied extensively as they are easy to measure. However, these results generate the least insights for

108

THE SAGE HANDBOOK OF SPORTS ECONOMICS

practitioners as they are not subject to change. There are only a few studies looking at money donations in amateur sport. The multi-level framework introduced in this chapter can serve as a guide for future research. On the institutional level, more insights into the different capacities as drivers for donations might be helpful to get a better understanding of the way organisations can increase their levels of donations. Taking a look at financial resources and planning and development resources might be of special interest as the literature in this area is very limited (Feiler et  al., 2014). Moreover, a comprehensive look at all resources could generate insights into the importance of each resource and their relationship. This might help organisations to improve their organisational structures and create effective strategies to increase time and money donations. On the national and regional level, further research is needed to better understand how governments and local authorities can create frameworks which support the ability to donate. It might be helpful to look into countries which have not been of major interest so far but which have a very active donation culture.

CONCLUSION The existing studies on the individual determinants of donating time and money have shown that psychological and social indicators matter. This is important for sport organisations as it helps to target donors more efficiently. As an example, studies have shown that value congruence has an important impact (Feiler et al., 2014; van Schie et  al., 2014). Thus, sport clubs should communicate their vision and mission clearly to attract potential donors who can identify themselves with the values of the club. In addition, the literature has shown that donating behaviour differs based on demographic (e.g. Dote et al., 2006; Kalargirou et al., 2015; Strigas & Jackson, 2003) and economic indicators (e.g. Burgham & Downward, 2005; Hallmann, 2015; Taylor et al., 2012). Even though these cannot be changed by the sport organisations, strategies can be created to address demographic groups that have not been targeted before. On the institutional level, studies have shown that more professionalised clubs that employ paid staff generate a higher amount of donations (Feiler et al., 2014). Thus, it is important for organisations to take on a more professional approach to attract potential donors. While studies on the national and regional level are rather limited, it still becomes

apparent that regional and national institutions need to create a sufficient framework to make time and money donations possible. Creating synergies between organisations and decreasing the extent of income inequality might lead to higher levels of social participation. To conclude, the question of the drivers of money and time donations is a very complex and multidimensional one. There seem to be many ways in which organisations can try to increase their donations, depending on the environment they operate in. Identifying the drivers of donations is important, especially for amateur sport clubs to ensure sustainable levels of financial donations and volunteering as well as to create efficient strategies to address potential donors. Thus, research can provide valuable insights to support the long-term success of amateur sports.

REFERENCES Adams, A., & Deane, J. (2009). Exploring formal and informal dimensions of sports volunteering in England. European Sport Management Quarterly, 9(2), 119–140. doi: 10.1080/16184740802571401 Ajzen, I. (1985). From Intentions to Actions: A Theory of Planned Behavior. In J. Kuhl & J. Beckmann (Eds.), Action Control. Berlin: Springer-Verlag. Aleskerov, F., Bouyssou, D., & Monjardet, B. (2007). Utility Maximization, Choice and Preference (2 ed.). Berlin: Springer-Verlag. Andreoni, J. (1990). Impure altruism and donations to public goods: A theory of warm-glow giving. The Economic Journal, 100(401), 464–477. doi: 10.2307/2234133 Bandura, A. (1971). Social Learning Theory. New York City: General Learning Press. Baruch, A., May, A., & Yu, D. (2016). The motivations, enablers and barriers for voluntary participation in an online crowdsourcing platform. Computers in Human Behavior, 64(1), 923–931. doi: 10.1016/j.chb.2016.07.039 Bauer, T. K., Bredtmann, J., & Schmidt, C. M. (2013). Time vs. money: The supply of voluntary labor and charitable donations across Europe. European Journal of Political Economy, 32, 80–94. doi: 10.1016/j.ejpoleco.2013.06.006 Becker, G. S. (1976). Altruism, egoism and genetic fitness: Economics and sociobiology. Journal of Economic Literature, 14(3), 817–826. Becker, G. S. (1981). Altruism in the family and selfishness in the market place. Economica, 48(189), 1–15. Bekkers, R., & Wiepking, P. (2011). A literature review of empirical studies of philanthropy:

The Role of Money and Time Donations in the Supply of Amateur Sport

Eight mechanisms that drive charitable giving. Nonprofit and Voluntary Sector Quarterly, 40(5), 924–973. doi: 10.1177/0899764010380927 Blau, P. M. (1960). A theory of social integration. American Journal of Sociology, 65(6), 545–556. Breuer, C., & Feiler, S. (2017). Sport Development Report 2015/2016: Analysis of the Situation of Sports Clubs in Germany (Vol. 6). Cologne: Sportverlag Strauß. Brown, E., & Lankford, H. (1992). Gifts of money and gifts of time: Estimating the effects of tax prices and available time. Journal of Public Economics, 47(3), 321–341. Burgham, M., & Downward, P. (2005). Why volunteer, time to volunteer? A case study from swimming. Managing Leisure, 10(2), 79–93. doi: 10.1080/13606710500146100 Cappellari, L., Ghinetti, P., & Turati, G. (2011). On time and money donations. The Journal of SocioEconomics, 40(6), 853–867. doi: 10.1016/j. socec.2011.08.013 Chen, C.-A., & Lee, Y.-j. (2014). A Closer look at the difference between public and nonprofit employees’ volunteering. International Public Management Journal, 18(1), 108–129. doi: 10.1080/ 10967494.2014.972481 Chum, A., Carpenter, S., Farrell, E., Mook, L., Handy, F., Schugurensky, D., & Quarter, J. (2015). Does geographic context influence employability-motivated volunteering? The role of area-level material insecurity and urbanicity. The Canadian Geographer/ Le Géographe canadien, 59(3), 354–368. doi: 10.1111/cag.12167 Collard, D. (1992). Love is not enough. In Zamagni (Ed.), The economics of altruism. Aldershot: Edward Elgar. Cuskelly, G. (1996). The Organisational Commitment of Volunteer Administrators in Sport. Belconnen: National Sports Research Centre. Dawson, P., & Downward, P. (2013). The relationship between participation in sport and sport volunteering: An economic analysis. International Journal of Sport Finance, 8(1), 75–92. Deci, E. L., & Ryan, R. M. (1985). Intrinsic Motivation and Self-determination in Human Behavior. New York: Plenum. Decker, O., Winter, M., Brähler, E., & Beutel, M. (2008). Between commodification and altruism: Gender imbalance and attitudes towards organ donation. A representative survey of the German community. Journal of Gender Studies, 17(3), 251–255. doi: 10.1080/09589230802204290 Dote, L., Cramer, K., Dietz, N., & Grimm, R. (2006). College students helping America. Retrieved 21.03.2017, from www.nationalservice.gov/pdf/ 06_1016_RPD_college_full.pdf Dury, S., Willems, J., De Witte, N., De Donder, L., Buffel, T., & Verte, D. (2016). Municipality and

109

neighborhood influences on volunteering in later life. Journal of Applied Gerontology: The Official Journal of the Southern Gerontological Society, 35(6), 601–626. doi: 10.1177/0733464814533818 Eisenberg, N., & Miller, P. A. (1987). Empathy, sympathy, and altruism: Empirical and conceptual links. In Zamagni (Ed.), The economics of altruism. Aldershot: Edward Elgar. Emrich, E., & Pierdzioch, C. (2015). Testing economic models of volunteer labour supply: Some empirical evidence for the German Red Cross. Applied Economics, 47(40), 4247–4259. doi: 10.1080/ 00036846.2015.1026587 Fairley, S., Kellett, P., & Green, C. B. (2007). Volunteering abroad: Motives for travel to volunteer at the Athens Olympic Games. Journal of Sport Management, 21(1), 41–57. Feiler, S., Wicker, P., & Breuer, C. (2014). How to raise voluntary giving for nonprofit sports clubs: An analysis of factors influencing donations. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 26(4), 1219–1239. doi: 10.1007/s11266-014-9489-3 Feldman, N. E. (2010). Time is money: Choosing between charitable activities. American Economic Journal: Economic Policy, 2(1), 103–130. doi: 10.1257/pol.2.1.103 Frank, R. H. (1987). If Homo Economicus could choose his own utility function, would he want one with a conscience? The American Economic Review, 77(4), 593–604. Hallmann, K. (2015). Modelling the decision to volunteer in organised sports. Sport Management Review, 18(3), 448–463. doi: 10.1016/j. smr.2014.12.003 Hammond, P. J. (1987). Altruism. In Zamagni (Ed.), The economics of altruism. Aldershot: Edward Elgar. Hartmann, B., & Werding, M. (2012). Donating Time or Money: Are they Substitutes or Complements? CESifo Working Paper (3835, Category 13: Behavioural Economics). Haski-Leventhal, D. (2009). Altruism and volunteerism: The perceptions of altruism in four disciplines and their impact on the study of volunteerism. Journal for the Theory of Social Behaviour, 39(3), 271–299. Healy, K. (2004). Altruism as an organizational problem: The case of organ procurement. American Sociological Review, 69(3), 387–404. Hoffman, M. L. (1981). Is altruism part of human nature? Journal of Personality and Social Psychology, 40(1), 121–137. Joo, S., Koo, J., & Fink, J. S. (2016). Cause-related marketing in sports: The power of altruism. European Sport Management Quarterly, 16(3), 316–340. doi: 10.1080/16184742.2016.1143854 Kääriäinen, J., & Lehtonen, H. (2006). The variety of social capital in welfare state regimes: A comparative

110

THE SAGE HANDBOOK OF SPORTS ECONOMICS

study of 21 countries. European Societies, 8(1), 27–57. doi: 10.1080/14616690500491399 Kalargirou, A. A., Beloukas, A. I., Kosma, A. G., Nanou, C. I., Saridi, M. I., & Kriebardis, A. G. (2015). Attitudes and behaviours of Greeks concerning blood donation: Recruitment and retention campaigns should be focused on need rather than altruism. Blood Transfusion, 12(3), 320–329. doi: 10.2450/2014.0203-13 Kennedy, B. P., Kawachi, I., & Brainerd, E. (1998). The role of social capital in the Russian mortality crisis. World Development, 26(11), 2029–2043. Kerwin, S., Warner, S., Walker, M., & Stevens, J. (2015). Exploring sense of community among small-scale sport event volunteers. European Sport Management Quarterly, 15(1), 77–92. Kim, M., Zhang, J. J., & Connaughton, D. P. (2010). Comparison of volunteer motivations in different youth sport organizations. European Sport Management Quarterly, 10(3), 343–365. Krebs, D. (1982). Psychological approaches to altruism: An evaluation. Ethics, 92(3), 447–458. Kumnig, M., Schnitzer, M., Beck, T. N., Mitmansgruber, H., Jowsey, S. G., Kopp, M., & Rumpold, G. (2014). Approach and avoidance motivations predict psychological well-being and affectivity of volunteers at the Innsbruck 2008 Winter Special Olympics. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 26(3), 801–822. doi: 10.1007/s11266-014-9462-1 Künemund, H., & Schupp, J. (2007). Konjunkturen des Ehrenamts – Diskurse und Empirie. In G. Meran, G. G. Wagner, J. R. Frick, J. Schupp, C. D’Ambrosio, C. Breuer, A. I. Drever, F. R. Lang, J.-P. Schräpler, C. K. Spieß, M. Spieß, V. Steiner, & A. S. Zuckerman (Eds.), SOEPpapers. Berlin: Deutsches Institut für Wirtschaftsförderung (DIW). Kurz, M. (1978). Altruism as an outcome of social interaction. The American Economic Review, 68(2), 216–222. Lee, C.-K., Reisinger, Y., Kim, M. J., & Yoon, S.-M. (2014). The influence of volunteer motivation on satisfaction, attitudes, and support for a megaevent. International Journal of Hospitality Management, 40(1), 37–48. doi: 10.1016/ j.ijhm.2014.03.003 Lee, Y.-K., & Chang, C.-T. (2008). Intrinsic or extrinsic? Determinants affecting donation behaviors. International Journal of Educational Advancement, 8(1), 13–24. doi: 10.1057/ ijea.2008.2 Lilley, A., & Slonim, R. (2014). The price of warm glow. Journal of Public Economics, 114(SI), 58–74. doi: 10.1016/j.jpubeco.2013.12.004 Lim, C., & Laurence, J. (2015). Doing good when times are bad: Volunteering behaviour in economic hard times. The British Journal of Sociology, 66(2), 319–344. doi: 10.1111/1468-4446.12122

Low, N., Butt, S., Ellis, P., & Davis Smith, J. (2007). Helping Out: A National Survey of Volunteering and Charitable Giving. London: Cabinet Office. McCulloch, A., Mohan, J., & Smith, P. (2012). Patterns of social capital, voluntary activity and area deprivation in England. In TSRC (Ed.), Briefing Paper Series. Birmingham: University of Birmingham. McKeever, B. S. (2015). The Nonprofit Sector in Brief 2015. Washington, DC: Urban Institute. Miller, D. L., Scheffler, R., Lam, S., Rosenberg, R., & Rupp, A. (2006). Social capital and health in Indonesia. World Development, 34(6), 1084–1098. doi: 10.1016/j.worlddev.2005.11.006 Observatoire de la Fondation de France (2015). An Overview of Philanthropy in Europe. Paris: Observatoire de la Fondation de France. Orlowski, J., & Wicker, P. (2015). The monetary value of voluntary work: Conceptual and empirical comparisons. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 26(6), 2671–2693. doi: 10.1007/s11266-014-9542-2 Priller, E., & Schupp, J. (2011). Social and economic characteristics of financial and blood donors in Germany. DIW Economic Bulletin, 1(6), 23–30. Rotolo, T., Wilson, J., & Hughes, M. E. (2010). Homeownership and volunteering: An alternative approach to studying social inequality and civic engagement. Sociological Forum, 25(3), 570–587. doi: 10.1111/j.1573-7861.2010.01196.x Rushton, J. P. (1982). Altruism and society: A social learning perspective. Ethics, 92(3), 425–446. Salamon, L. M. (2010). Putting the civil society sector on the economic map of the world. Annals of Public and Cooperative Economics, 81(2), 167–210. Salamon, L. M., & Anheier, H. K. (1999). Civil society in comparative perspective. In L. M. Salamon, H. K. Anheier, R. List, S. Toepler, & S. W. Sokolowski (Eds.), Global Civil Society: Dimensions of the Nonprofit Sector. Baltimore, MD: The Johns Hopkins Center for Civil Society Studies. Salamon, L. M., Sokolowski, S. W., & Haddock, M. A. (2011). Measuring the economic value of volunteer work globally: Concepts, estimates, and a roadmap to the future. Annals of Public and Cooperative Economics, 82(3), 217–252. doi: 10.1111/j.1467-8292.2011.00437.x Scheepers, P., te Grotenhuis, M., & Gelissen, J. (2002). Welfare states and dimensions of social capital: Cross-national comparisons of social contacts in European countries. European Societies, 4(2), 185–207. doi: 10.1080/14616690220142781 Schlesinger, T., & Nagel, S. (2013). Who will volunteer? Analysing individual and structural factors of volunteering in Swiss sports clubs. European Journal of Sport Science, 13(6), 707–715. doi: 10.1080/17461391.2013.773089

The Role of Money and Time Donations in the Supply of Amateur Sport

Schlesinger, T., & Nagel, S. (2016). Individual and contextual determinants of stable volunteering in sport clubs. International Review for the Sociology of Sport. First online, April 10. doi: 10.1177/ 1012690216638544 Simmons, W. O., & Emanuele, R. (2007). Malefemale giving differentials: Are women more altruistic? Journal of Economic Studies, 34(6), 534–550. doi: 10.1108/01443580710830989 Simon, H. A. (1992). Altruism and economics. Eastern Economic Journal, 18(1), 73–83. Sleeboom-Faulkner, M. (2014). The twenty-first-century gift and the co-circulation of things. Anthropological Forum, 24(4), 323–337. doi: 10.1080/ 00664677.2014.948380 Smith, J. D. (2006). Poor marketing or the decline of altruism? Young people and volunteering in the United Kingdom. International Journal of Nonprofit and Voluntary Sector Marketing, 4(4), 372–377. Strigas, A. D., & Jackson, E. N. J. (2003). Motivating volunteers to serve and succeed: Design and results of a pilot study that explores demographics and motivational factors in sport volunteerism. International Sports Journal, 7(1), 111–123. Taylor, P. D., Panagouleas, T., & Nichols, G. (2012). Determinants of sports volunteering and sports volunteer time in England. International Journal of Sport Policy and Politics, 4(2), 201–220. Unger, L. S. (1991). Altruism as a motivation to volunteer. Journal of Economic Psychology, 12(1), 71–100. van der Meer, T., Scheepers, P., & te Grotenhuis, M. (2009). States as molders of informal relations? European Societies, 11(2), 233–255. doi: 10.1080/ 14616690802133293 van Schie, S., Güntert, S. T., Oostlander, J., & Wehner, T. (2014). How the organizational context impacts volunteers: A differentiated perspective on self-­ determined motivation. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 26(4), 1570–1590. doi: 10.1007/s11266-014-9472-z

111

Veludo-de-Oliveira, T. M., Pallister, J. G., & Foxall, G. R. (2015). Unselfish? Understanding the role of altruism, empathy, and beliefs in volunteering commitment. Journal of Nonprofit & Public Sector Marketing, 27(4), 373–396. doi: 10.1080/10495142.2015.1080504 Vos, S., Breesch, D., Késenne, S., Lagae, W., Van Hoecke, J., Vanreusel, B., & Scheerder, J. (2012). The value of human resources in non-public sports providers: The importance of volunteers in nonprofit sports clubs versus professionals in for-profit fitness and health clubs. International Journal of Sport Management and Marketing, 11(1/2), 3–25. West, S. A., Griffin, A. S., & Gardner, A. (2007). Social semantics: Altruism, cooperation, mutualism, strong reciprocity and group selection. [Research Support, Non-US Government Review]. Journal of Evolutionary Biology, 20(2), 415–432. doi: 10.1111/j.1420-9101.2006.01258.x Wichardt, P. C. (2009). A status-based motivation for behavioural altruism. International Journal of Social Economics, 36(8), 869–887. doi: 10.1108/ 03068290910967127 Wicker, P., Breuer, C., & Hennigs, B. (2012). Understanding the interactions among revenue categories using elasticity measures: Evidence from a longitudinal sample of non-profit sport clubs in Germany. Sport Management Review, 15(3), 318–329. doi: 10.1016/j.smr.2011.12.004 Wicker, P., & Hallmann, K. (2013). A multi-level framework for investigating the engagement of sport volunteers. European Sport Management Quarterly, 13(1), 110–139. doi: 10.1080/ 16184742.2012.744768 Wildman, J., & Hollingsworth, B. (2009). Blood donation and the nature of altruism. [Research Support, Non-US Government]. Journal of Health Economics, 28(2), 492–503. doi: 10.1016/j.jhealeco.2008.11.005 Yang, W. (2016). Are contributions of time and money substitutes or complements? Applied Economics, 48(37), 3526–3537. doi: 10.1080/ 00036846.2016.1139684

12 The Economics of the National Collegiate Athletic Association1 Allen R. Sanderson and John J. Siegfried

INTRODUCTION One of the truly unique aspects of American ‘exceptionalism’ is commercialized sports on college campuses. There is no evidence that bigtime intercollegiate athletics helps to create or disseminate knowledge, yet many leading American universities sponsor sports programs that generate over $8 billion annually. Their team names have become their institutions’ brands. The organization that governs these activities is the National Collegiate Athletic Association (NCAA). We trace the evolution of intercollegiate athletics and its governing body, and the rationale for their existence. We explain how the NCAA reduces costs and increases revenues for the sports programs of its member institutions. We identify important contributions economists have made to understanding how the NCAA and the college sports market operate and identify important remaining research opportunities.

A SHORT HISTORY OF COLLEGE SPORTS There were approximately 200 mostly small, private colleges and universities in the United States by 1860. The number grew substantially after the

Morrill Act of 1862 created the federal land-grant system that spawned large state universities, which today field most of the big-time intercollegiate football and basketball programs. As colleges and universities grew in number and size after the Civil War, competition for students intensified. Sports programs were used to attract more students. Rowing was the first sporting competition between students at different colleges; Yale met Harvard on Lake Winnipesaukee in 1852. The first intercollegiate basketball game using five players on each team occurred in 1896, between the University of Chicago and the University of Iowa. The first intercollegiate football game was staged in 1869 between Rutgers and Princeton. By the 1890s gate receipts had grown to serious levels. Harvard built the first permanent college stadium in 1903, with a capacity of 31,000 fans. Other universities followed. Once substantial fixed costs were invested, steady revenues became indispensable to pay the mortgage.

WHY DO COLLEGES AND UNIVERSITIES SPONSOR COMMERCIALIZED INTERCOLLEGIATE SPORTS? University sponsorship of commercialized sports is especially odd because providing commercial

THE ECONOMICS OF THE NATIONAL COLLEGIATE ATHLETIC ASSOCIATION

entertainment is not a stated purpose in the charter of any university. And while about 400 do operate commercial intercollegiate athletics programs, the vast majority of American colleges and universities do not. Moreover, only a score of institutions with big-time athletics programs reap a net financial surplus from them. If this occurred in manufacturing, there surely would be a steady exit. Yet only two major universities have dropped big-time football over the last century – Washington University and the University of Chicago (Clotfelter, 2011, pp. 49–50). Before we dismiss commercialized intercollegiate sports programs as a colossal mistake, we should ask why they have survived so long while operating mostly in the red. Most Americans believe that intercollegiate athletics contributes money to their universities.2 NCAA data, however, reveal that only 24 of the 128 top-level (called Football Bowl Subdivision) universities earned an operating surplus on intercollegiate athletics in 2015 (Fulks, 2016), and only a tiny portion of those surpluses were transferred to the academic side of their universities. Moreover, at most institutions, capital costs and ancillary expenses such as game security are not charged against the income statement of intercollegiate athletics. How can so many large intercollegiate sports programs survive while losing money? In 2013, USA Today reported that over $1 billion of student tuition or fees moved annually to the support of intercollegiate sports (Berkowitz, Upton, and Brady, 2013). The amounts redirected from academic to athletic purposes at most of the colleges and universities playing big-time sports are remarkable at a time of shrinking state support for public universities and incessant complaints about rising tuition from students and their parents. There are several possible explanations. First, intercollegiate athletics may attract greater appropriations from legislators concerned about their constituents’ perceptions of public universities in their states, especially in light of the fact that the median voter in most states is not a college graduate, and is likely to be more interested in the quality of the flagship university’s sports teams than its library or research. Among 570 public universities, Humphreys (2006) found that those with big-time football programs receive about eight percent more funding than comparable ones without football; participation seems to matter more than winning. Alexander and Kern (2010) found similar results for basketball. Second, intercollegiate athletics may boost private donations. Numerous studies have explored the effects of intercollegiate athletics on contributions to colleges: some find no effect, while others report a modest positive gain (for a review of these,

113

see Getz and Siegfried, 2012). Playing in football bowl games appears to stimulate the most donations. Because most of the incremental contributions are steered to athletic departments (Anderson, 2012), it is not clear whether this effect produces much benefit to universities in general. It would be helpful to know, however, not just whether the presence of big-time college sports increases donations, but whether it increases them by more than it costs. Third, as was anticipated at the inception of intercollegiate sports competitions, high-profile sports programs, like other campus amenities, may attract applicants and additional enrollment, which is especially beneficial if fixed-cost facilities (e.g. dormitories) are underutilized. Evidence confirms that participating in post-season competition generates student interest in a university, although the gains are modest and fleeting. It appears that simply fielding big-time sports teams matters more for student recruitment than winning, and football matters more than basketball, perhaps because it is played during the college application season. More spending on intercollegiate athletics may alter the institutions to which high school seniors apply and which they attend, but there is no evidence that bigtime athletics increases overall college attendance (Getz and Siegfried, 2012, p. 359). Fourth, spending on sports programs resembles an arms race. Successful programs bid aggressively for high-profile coaches and enhance their physical facilities to lure recruits. Small spending differences can yield big recruiting advantages and, subsequently, augment winning. Unprofitable programs have little option but to ratchet up spending, or they may drop even farther behind in the competition for coaches and players, with devastating effects on their financial condition. Thus, the positive net revenues of a few dozen teams drive up costs for all competitive teams, requiring universities with already unprofitable intercollegiate athletics programs to further expand internal subsidies.3 Fifth, many colleges set tuition well below operating expenses. Students with specific talents and characteristics (including financially successful parents) are admitted selectively, in the hope that some mature into appreciative multi-millionaires willing to share their good fortune with their alma maters (Hoxby, 2014). To enhance the prospects that financially successful graduates remember them at estate planning time, the institutions invest in developing and maintaining emotional ties. They arrange alumni cruises, send faculty to alumni club talks, and sponsor ‘homecoming’ events that feature a football game. The challenge to universities is to weigh on the margin the value of funds devoted immediately to improving teaching and research against the expected future value of a more visible and successful intercollegiate athletics program.

114

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The hope, of course, is that the successful athletic program might someday attract sufficiently larger contributions to the non-athletic side of the institution, such that, when discounted, it generates an even greater boost to teaching and research.4

DEVELOPMENT OF THE NCAA5 The development of the NCAA is closely tied to the complicated development of intercollegiate football. Over time rules changes morphed it from a game like soccer into one similar to rugby, and eventually into American football. In the 1880s, important rules changes occurred when a snap from scrimmage replaced the rugby scrum, teams had three tries to gain at least five yards, and blocking was permitted. Violence in football intensified after 1888, when tackling below the waist was legalized. A tactic in which blockers interlocked arms and ran together into defenders was introduced. One version was called the Flying Wedge, where the players took a running start before the ball was snapped. From 1890 through 1905, over 300 college and university students died from injuries sustained playing intercollegiate football (Zimbalist, 1999, p. 8). News reports about football deaths and injuries prompted Theodore Roosevelt to summon to the White House representatives from the then football powerhouses Harvard, Yale, and Princeton. They promised the President to change the rules to reduce violence in the game. Further impetus for reform arose in 1905 when a Union College player died after being hit by the New York University offensive line. A witness, NYU’s chancellor, resolved to end the violence. He assembled representatives from 62 colleges to consider reform; they organized the Intercollegiate Athletic Association of the United States (IAAUS) and established a rules committee, which promptly required that at least seven offensive players be on the line of scrimmage when the ball is snapped, thereby ending the Flying Wedge. Formal rulemaking authority was necessary to reduce the brutality in football because violence led to victories and winning generated gate receipts. No team would unilaterally refrain from dangerous practices because that would lead to the worst possible outcome for it, even though continued violence seemed likely to reduce overall demand for intercollegiate football. In 1906, 39 colleges and universities ratified the IAAUS constitution. In 1910, the IAAUS changed its name to National Collegiate Athletic Association (NCAA); by 1911 it had 95 members

and was entrenched as the self-regulatory body overseeing collegiate athletics. In 1935 it allowed athletic-related scholarships for players. The next significant changes in NCAA rules occurred after World War II. To better control costs, which were at risk of devouring the growing athletic revenues collected by universities, the NCAA forbid the payment of any compensation to players. However, by 1951 athletic scholarships returned to stay. While player compensation has been restricted to a grant-in-aid (tuition, room, board, books, and fees) since 1951, player costs were not really controlled until 1973, when limitations on the number of scholarships (but not the number of players on a team) were instituted. To ease student-athletes’6 adjustment to college, from 1939 through 1968 first-year students were ineligible to play in NCAA championships. To reduce costs, in 1972 they became eligible. In 1973 the NCAA divided into Divisions I, II, and III. Division I contains large universities competing at the highest level in most sports; Division II institutions are usually smaller, and offer fewer scholarships; Division III awards no athletic scholarships. In 1978 Division I was further split into three groups for governing (just) football, Football Bowl Subdivision (FBS) now containing 128 teams competing at the highest level, Football Championship Division (FCS) for about 120 teams that compete at a lower level, and about 100 universities that have competitive basketball teams, but do not play football. In addition to the approximately 350 Division I teams, there currently are about 300 in Division II and 450 in Division III. NCAA interest in economics increased after World War II. Once player safety improved, interest in limiting the cost of players intensified because greater revenues tempted universities to squander their new-found largesse competing for players. Once costs were under control, attention turned to expanding revenues further. A confluence of independent events boosted intercollegiate sports revenues after World War II. First, the G.I. bill augmented college enrollment by over a million students per year from 1940 to 1950, enhancing spectator demand for college sports. But that was only a start. The number of enrolled college students rose from 2.3 million in 1950 to 20.3 million by 2015.7 Second, the post-war baby boom increased the demand for college sports as the population of teenage boys spiked during the late 1960s. Third, the rapid post-war expansion of television receivers added broadcast demand to the steadily growing spectator demand at games. In 1940 the University of Pennsylvania (Penn) started televising its home football games. By 1950 Penn was earning $150,000 (1950 dollars)

THE ECONOMICS OF THE NATIONAL COLLEGIATE ATHLETIC ASSOCIATION

for annual broadcast rights. In 1951, however, the NCAA prohibited televising college football games because it might reduce gate receipts of other games. With so much money at stake, Penn refused to stop, whereupon the NCAA threatened to expel Penn by having its opponents refuse to play the Quakers. Penn backed down. To ‘stabilize’ college football broadcasting, in 1952 the NCAA initiated a ‘television plan,’ which lasted 32 years. It was an agreement among members to televise only one weekly Saturday afternoon game (or a set of simultaneously aired regional games) and restricted the number of appearances by a team. The value of television broadcasting rights rose rapidly over succeeding decades. When combined with the ‘March Madness’ basketball tournament, TV revenues reached $91 million annually by 1984 (Raiborn, 1986).8 In 2018, the men’s basketball tournament alone brought in $1.1 billion.

Basic Components of NCAA Regulation The principal components of the NCAA’s cartel operations are agreements to: (1) restrict the number of games available for sale to broadcast networks (through 1984, and thereafter subdivide the country by region for conferences to exploit market power); and (2) limit the compensation and demand for players. The latter lowers costs relative to what they would be in a competitive player market, while the former enhances revenues compared to their competitive level.

Broadcasting College Sports The NCAA ‘television plan’ limited football broadcasts to just one game per week, creating artificial scarcity. Bids for the rights escalated, with three over-the-air networks chasing a single source of game content. Team appearances were restricted because most of the rights revenue was distributed on the basis of appearances, and limiting appearances spread revenues over more teams, expanding political support for the plan. Dissatisfied with their share of revenues, eventually the universities of Oklahoma and Georgia filed an antitrust suit against the NCAA for conspiracy in restraint of trade in the market for college football broadcasts. After winding through the courts for several years, Board of Regents of the University of Oklahoma vs. NCAA9 was decided in June 1984, with the Supreme Court ruling that not-for-profit colleges are subject to the Sherman Antitrust Act, and that by agreeing to limit the number of games on offer

115

for broadcast each week, the colleges had violated the law. The decision ended the NCAA’s nationwide football television contract. Immediately following the decision, the number of televised games expanded, and rights fees per game plummeted (Siegfried and Burba, 2004, p. 807). However, the short-lived decline in broadcast revenues to about one-third the 1983 level was followed by three decades during which broadcast revenues exploded10 (Carroll and Humphreys, 2016) in spite of a more competitive market for rights. Several factors caused this growth in revenues, some orchestrated by the NCAA and its member institutions, some a result of evolving demographics, and others emanating from rapidly changing broadcast technology. Those who predicted that broadcast rights would decline sharply after 1984 failed to anticipate the rapid growth of television networks demanding football and men’s basketball game content, and the degree to which college football demand, in particular, is regional, preserving market power for the (largely regional) conferences that negotiated broadcast rights sales after 1984. Interest in college sports, especially football, is regional partly because many alumni reside near their alma maters,11 and they and current students constitute a substantial base demand for both live attendance and television broadcasts. Once broadcast rights sales devolved to the conferences, the five power conferences12 quickly expanded in order to solidify their regional dominance. All five added teams during the 1990s and early twentyfirst century. After a brief period of confusion at the beginning of the 1984 college football season, during which home and visiting teams sometimes each sold ‘exclusive’ rights to the same game to different broadcasters, a duopoly emerged. The College Football Association (CFA), formerly an internal NCAA interest group, negotiated television rights for teams in the Southeastern Conference (SEC), Atlantic Coast Conference (ACC), and Big Eight (which has since evolved into the Big-12), plus Notre Dame and Penn State, two successful independents at the time. The Big-Ten (now with 14 members) and Pac-10 (now with 12 members and a new name – Pac-12) combined to offer networks an alternative television package. This duopoly was not challenged by antitrust authorities, but eventually proved to be unstable. The CFA dissolved in 1995 amid internal wrangling over revenue shares (Siegfried and Burba, 2004). The Big-Ten and Pac-10 had separated in 1990, for similar reasons. Since 1995 each power conference has negotiated broadcast rights separately for its members. While the power conferences consolidated market power, technological developments increased

116

THE SAGE HANDBOOK OF SPORTS ECONOMICS

the relative value of televising events that viewers prefer to watch live – ‘breaking news’ and live sporting events. Because watching an event live and recording it to view later after skipping the commercials are mutually exclusive, viewers cannot easily avoid commercials in live sports broadcasts, thereby increasing the relative value of advertising. Consequently, live sports programming is among the most valuable content on television. Thirty seconds of advertising time during the 2015 NCAA men’s basketball championship game cost $1.55 million. The highest prime time show, ‘The Big Bang Theory,’ charged $290,000 for 30 seconds. Saturday night college football attracted $84,000 on Fox and $104,000 on ABC, while the highest non-football show on Saturday night (‘Dateline Mysteries’) commanded only $44,000 for a 30-second-advertisement (Steinberg, 2015).

Compensating ‘student-athletes’ In contrast to professional leagues that control player costs by revenue sharing, penalties on ‘excessive’ payrolls, aggregate payroll caps, and limits on individual player compensation negotiated with players’ unions, colleges have agreed among themselves simply to limit player compensation to a grant-in-aid. Because there is no players’ union involved, such an agreement is not protected from antitrust liability by the 1935 National Labor Relations Act, as are the practices that professional leagues use to control compensation. One might ask why players accept only a grantin-aid to work in an essentially full-time job. The answer lies partially in help provided by the NFL and the NBA, which have reduced paid alternatives available to talented young football and basketball players. Since 1990 the NFL has refused to employ players who are fewer than three years past high school. Since 2006, the NBA has declared ineligible players who are less than a year beyond high school. Other than play for college teams for a year, elite basketball players could play professionally overseas for a year, but few do. Football players do not have even that option, because American-style football is not played outside North America. In return for restricting young players’ alternatives, the professional leagues secure free employee training. The ban on younger players also reduces the risk to professional teams from hiring immature players whose behavior might sully the league’s reputation, or of contracting with injury-prone players. Coaches’ and athletic directors’ compensation is not limited by the NCAA. The NCAA does not control what an institution spends on athletic

facilities, such as stadiums, training facilities, or locker rooms. The NCAA does, however, restrict recruiting expenditures in order to preserve the benefits available to those whose compensation is not limited.

NCAA CARTEL OPERATIONS: FOUR SIGNIFICANT CHALLENGES Every successful cartel must: (1) reach agreement, (2) avoid the erosion of cartel profits by either nonprice competition or member cheating, (3) deter new entry attracted by the prospect of sharing cartel profits, and (4) distribute cartel benefits in a manner that is viewed as equitable by the participants. Achieving agreement can be challenging for heterogeneous cartel members. Colleges and universities incur different athletic program costs and have divergent goals. Costs differ because private and public universities face different opportunity costs of offering a grant-in-aid to a student-athlete.13 Objectives diverge because institutions differ in terms of emphasis on teaching, research, and public service. Small liberal arts colleges usually field a wide array of intercollegiate athletic teams so as to provide athletic opportunities for students. In contrast, large state universities usually field fewer athletic teams, excluding many sports that do not generate much revenue. Big-time programs focus on football and men’s basketball. Even within the (now) six separate NCAA football governance categories, substantial differences persist. Stanford, Northwestern, and Duke are each in a power conference, along with Washington State, Nebraska, and Florida State. These two sets of institutions are quite different in terms of tuition, student characteristics, emphasis on faculty research, academic programs offered, and public service responsibilities. With such differences among members, it is remarkable the NCAA has survived as a cohesive economic cartel.14 Reaching and maintaining an agreement among competitors to reduce competition in selling broadcast rights and in acquiring player talent might arouse antitrust challenges. But so far, the NCAA has avoided charges of conspiring to fix player compensation by promoting an image of its players as ‘student-athletes’ rather than employees. Since 1984, there has been no challenge to the sale of broadcast rights now coordinated at the conference level. Second, the cartel must protect its rents from erosion by either non-price competition or cheating. Because each unit sold earns excess profits, cartel

THE ECONOMICS OF THE NATIONAL COLLEGIATE ATHLETIC ASSOCIATION

members will try to add sales at the expense of rivals – called non-price competition. If all cartel members confront similar incentives, such non-price competitive efforts likely cancel out. Eventually the only change is that everyone’s costs rise. Opportunities for non-price competition to attract additional television broadcast sales consist of spending more money to elevate the attractiveness of teams to broadcast networks. On the labor side of operations there are many opportunities for individual cartel members to spend money to make themselves more attractive than rivals. Intense recruiting feeds the egos of high school players, their families, and their coaches. Better training and playing facilities, special academic tutoring, superior culinary experiences, travel to attractive locations for non-league games, and professional marketing of successful players seeking recognition (as well as cash payments, phony jobs, and benefits channeled to family members) can all sway a 17-year-old’s university choice. Absent direct salary competition, non-price benefits increasingly affect prospective players’ institutional preferences, which impacts skill balance among teams, although the direction is unclear. How cash compensation would affect competitive balance depends on the relative preferences of various talented players for cash versus their perceived value of the non-pecuniary benefits of playing for different institutions. Since there must be some talented players who favor cash, relaxing the compensation cap would likely divert some players to universities that could not attract them with just program prestige and modern facilities. Even if competitive balance were to decline, demand may not. Currently, intercollegiate athletics is popular despite considerable imbalance. The popularity of dominant teams and the enjoyment fans of non-dominant teams receive when their team periodically upsets an elite team may overshadow the value of more competitive balance (Coates, Humphreys, and Zhou, 2014). But a larger challenge to maintaining demand could arise if players receive cash compensation. The demand for college sports may depend on the fact that the players are perceived primarily as ‘students.’ Teams are tempted to cheat on cartel agreements regarding number of games played, games televised and player pay. If everyone cheats, however, the restriction on output necessary to support the elevated price or depressed player compensation evaporates. To deter cheating, it must be detectable so it can be discouraged by punishment. In sports, where two teams are required to produce a game, the offending team’s scheduled opponent can refuse to play the cheater. Competitors can drive a defector’s output to zero.

117

Cheating on the agreement to restrict player compensation and limit recruiting expenses is difficult to monitor. Because it is easier to hide illicit payments to players, furnish benefits indirectly (e.g. funneled to friends or relatives, or as payment for a fake job), and virtually impossible to monitor efficiently millions of contacts between coaches and prospective players, violating the compensation and recruiting cost rules of the cartel is more likely than selling additional games to broadcast networks. But it does not jeopardize cartel stability as directly as does cheating on broadcast limits, because the stakes in each transaction are much smaller. The enticement to cheat depends on balancing the value of the expected advantage the defection creates against the cost of being caught, which, depends, in turn, on the combined likelihood of being caught and the severity of punishment if caught. Over 50 years ago, George Stigler (1964) observed that the probability of defecting depends on the number of participants, because the effect of widely dispersed cheating in cartels with many participants is difficult for other members to detect. Because over 100 NCAA football teams compete for elite high school football players, and over 300 chase premier basketball players, it would seem that the NCAA cartel would swiftly collapse. To minimize the chances of such an outcome, the NCAA employs an investigation staff to identify and evaluate claims of rule violations. The real pain in the possible cost of cheating comes not from the likelihood of detection, however, but rather from the severe penalties the NCAA can levy, including bans on post-season play and the corresponding sacrifice of financial and status benefits from such appearances, as well as limits on the number of future scholarships, and thereby future team strength. In extreme cases, the NCAA can even close a program.15 The incentive to cartelize is irresistible in college sports. Because costs associated with fielding college sports teams are largely fixed, including stadium and player costs, extra revenue falls directly to the bottom line.16 A high ratio of fixedto-variable costs also facilitates agreement in order to prevent a broadcasting price war or player bidding war, because in such circumstances broadcast rights prices could be cut substantially or player pay elevated significantly while still leaving revenues comfortably above variable costs. The third challenge confronting a successful cartel is to prevent profit erosion by competition with new entrants attracted by its success. The NCAA’s Division I FBS, which receives most of the revenues, has successfully fended off entry. There are many major universities without

118

THE SAGE HANDBOOK OF SPORTS ECONOMICS

a commercialized intercollegiate athletics program (e.g. MIT, Carnegie-Mellon, University of Rochester) that could try to enter the big-time. But entry has not created much new competition for the elite programs because the teams that have upgraded were already NCAA members and agreed to abide by the rules. Moreover, upgrades are discouraged by NCAA requirements on the capacity of playing facilities, minimum number of games scheduled against Division I teams that are difficult to arrange, and a minimum number of scholarships that must be awarded (raising costs for entrants). Only 14 universities have gained NCAA Division I FBS status (the top 128 institutions) since 2000. The challenge of gaining access to one of the power conferences (none of the 14 entrants since 2000 is a member of a power conference), to bowl games, or to March Madness are additional hurdles facing new competitors. The fourth challenge to a successful cartel is to convince its members that the fruits of their efforts are distributed equitably. This can be daunting if members disagree about what is ‘fair.’ Those who do not contribute televised games are likely to favor more equal distribution, while those who do produce televised games are likely to favor distribution based on production. All potential competitors should share in the spoils, because even those whose output is constrained to zero help to elevate price. The NCAA used some of the revenues from its football broadcast rights cartel from 1951 through 1983 to cover the association’s operating costs, so all members benefitted by not having to pay dues.17 It also limited annual television appearances per team to force a wider distribution of revenues. But the wider distribution also limited the revenues accruing to teams that spent the most on their programs and felt they deserved a larger share. This is what eventually led the universities of Oklahoma and Georgia to sue the NCAA.

A TWENTY-FIRST-CENTURY CHALLENGE When the Supreme Court decided NCAA v. Board of Regents of the University of Oklahoma in 1984, Apple had just introduced the Macintosh computer. In 1984 the internet was largely unknown. With the advent of widescreen and high-definition television,18 cable TV and satellite TV, the internet, video games, smartphones, social media sites, and streaming capabilities, technological advances that were unforeseen in 1984 have altered broadcast viewing choices forever, upsetting the balance between in-venue versus at-home viewing options for fans.

Current broadcast rights for just one major December bowl game exceed the amount paid for the entire 1984 college football season. For the current football playoff system, begun in 2015, ESPN has a $7.3 billion, 12-year contract to televise seven football games a year – four December/ January bowl games plus a three-game national championship series each January. With 50,000 people in the stands in Tuscaloosa on a Saturday in fall 1980, the Alabama-Georgia game, not selected as ABC’s football game-of-theweek, was essentially a private good because in order to watch it you had to buy a ticket. But after 1984 that game took on characteristics of a public good, as millions of fans tuned in on commercial, cable or satellite TV to watch the contest, including many who did not buy a ticket or pay for television service. During the game in virtually real time, viewers can interact via social media with countless other fans, perhaps increasing the value of viewing the game. For the NCAA and universities like Georgia and Alabama, this widespread exposure and television rights fees now dominate the revenue flows. With the development of new institutional and conference-wide cable and network packages, part of the public-good can be converted to a private-good again, as some non-payers are turned away by new technologies and blocking services. Nevertheless, broadcasting college sporting events still exhibits the other characteristic of a publicgood – low marginal cost – so that the efficient price is practically zero, threatening the viability of a service that generates a large welfare surplus. The Fox network now rivals CBS, NBC and ABC, competing for broadcast rights for most premier sporting events. By the 1980s, live cable programming was widespread, with sports content playing a central role in its expansion. More recently, individual conferences, and sometimes even individual teams, have formed networks to televise college sports. The entry of Fox, ESPN and other cable sports networks and the formation of new networks by conferences and individual teams has intensified bidding for game content. This has redistributed broadcast revenues, concentrating more of the largesse among the elite teams, which may eventually lead these teams in power conferences to break-off from regulation by the NCAA.19

INTERNAL REFORMS AND LAWSUITS In April 2014 the NCAA’s Division I voted to allow members to offer unlimited meals and snacks to their athletes, a clear break from cost

THE ECONOMICS OF THE NATIONAL COLLEGIATE ATHLETIC ASSOCIATION

control efforts, spawning a new intercollegiate competition in food provision.20 Soon thereafter, the NCAA altered its structure to allow the power conferences and Notre Dame to operate under a different set of rules than other teams for governing football.21 This allows power conference teams to provide benefits to their athletes that would be difficult for other teams to afford. But tinkering with player compensation rules falls well short of a competitive labor market for college athletes (Sanderson and Siegfried, 2015), which is the goal of several contemporary lawsuits. First is the 2014 O’Bannon v. NCAA decision. Ed O’Bannon, a member of a UCLA national championship basketball team, argued that once players leave college they should benefit financially from the commercial use of their image. But the NCAA asserts lifetime control over such use. The case expanded to include television broadcast rights of players while they are in college. US District Court Judge Claudia Wilken ruled that the NCAA’s cap on player compensation is collusion in restraint of trade. She suggested that a stipend set $5,000 above current grants-in-aid might be legal. Upon appeal the Ninth Circuit agreed that the cap on player compensation is illegal but stipulated ‘full-cost-of-attendance’ rather than $5,000 over the traditional grant-in-aid as maximum compensation. Both decisions cap player compensation by the courts, rendering the NCAA agreement irrelevant. In September 2016, the US Supreme Court declined to hear an appeal of O’Bannon, leaving the conflict between the District Court and Appeals Court over appropriate compensation unresolved. The confusion precipitated by O’Bannon may be resolved by another case moving through the Ninth Circuit. In March 2014, sports labor attorney Jeffrey Kessler filed a lawsuit on behalf of Shawne Alston, a former West Virginia University football player. Alston specifically requested an injunction to end all collectively imposed limitations on player pay. In March 2019 Judge Claudia Wilken ruled again that an NCAA coordinated cap on player compensation is collusion in restraint of trade and therefore illegal, and that universities are free to offer players more than just a grant-in-aid covering tuition, room, and board. However, the additional benefits offered to players may not include cash, and must be education related, e.g. computers, musical instruments, science equipment. Neither plaintiff Alston nor the NCAA are happy with the decision. Both are considering an appeal. The eventual outcome of the Alston case is unknown at the time of this writing, but the precedents from two earlier NCAA legal defeats – the 1984 television broadcast rights price-fixing case and the fourth assistant basketball coaches’ 1998

119

wage-fixing case22 – suggest the NCAA is in dangerous territory with its agreement to restrict player compensation to a grant-in-aid even if supplemented with a computer, piano and Bunsen burner. What might happen if elite college athletes can sell their services in a truly competitive market? It is inevitable that with no pay restraint some universities will offer their better players financial incentives to remain on their team and will include cash in bids for new recruits. Other teams will follow suit. Deteriorating financial conditions will probably cause some colleges to drop big-time intercollegiate athletics. The roster size of many football teams is likely to shrink so as to free some revenue to pay the stars,23 and relatively more of the players will be ‘walk-ons,’ receiving neither cash nor a scholarship.24 A victory by Alston probably finishes off the NCAA’s current business model. Concurring with this conclusion, NCAA President, Mark Emmert, said that a victory by Kessler would ‘blow up college sports’ (Strauss, 2014).25

WHAT DOES THE FUTURE HOLD? Even if the NCAA’s collective agreement to limit the pay for players is declared illegal, because of multi-year contracts, expenditures that have been soaking up the rents are unlikely to diminish much in the short run. Revenues could grow to cover some additional costs of paying star players, but it is difficult to predict broadcast revenues even ten years in the future, as technology is likely to restructure the amount and distribution of revenues faster than that. Rights fees received by universities could increase or decrease. Alternative technologies that access live sports programming or that block advertisements may enhance or undermine the current value of rights. If player pay increases, the costs of big-time athletics programs will rise, and demand may wane if fans lose interest because the myth of the ‘student-athlete’ is deflated.26 Consequently, the surpluses earned by programs that currently have a positive balance before capital and indirect costs are considered could fade, while subsidies from other university funds to intercollegiate athletics at those institutions currently reporting a loss grow. University presidents will confront difficult questions: ‘How much of a subsidy is too much? When do the indirect benefits from fielding a competitive FBS football or a tournament-level basketball team begin to fall short of the value of the research and teaching forfeited to support the team financially?’

120

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Unless Congress legalizes the NCAA’s cartel behavior limiting player compensation by exempting it from antitrust liability, which is not implausible in view of the political hazards confronting Congressional representatives who vote to change college athletics substantially, it seems unlikely that the organization of big-time commercialized intercollegiate athletics ten years in the future will resemble today’s arrangement. Alternatively, it is possible that the NCAA’s umbrella control might be replaced by just the 65 teams in the five power conferences dictating the rules of big-time college athletics, while other colleges and universities are abandoned to fend for themselves.27 Such a structure might satisfy antitrust constraints because it ostensibly enhances competition among college sports conferences.

Notes 1 Sections of this essay originated in two prior essays by the same authors, published in The Review of Industrial Organization, Vol. 52, No. 2 (March 2018), “The National Collegiate Athletic Association Cartel: Why it Exists, How it Works, and What it Does, pp. 185–209, and “The Role of Broadcasting in National Collegiate Athletic Association Sports,” pp. 305–321. 2  Based on a Knight Commission (2006) survey, 78 percent of Americans believe intercollegiate athletics is profitable. 3  While median revenue generated from intercollegiate sports increased by 110 percent from 2004 through 2015 at the 128 largest programs, expenses rose by 129 percent at the same programs (Fulks, 2016, Table 2.1). 4  Fort and Winfree (2015) discuss within-institution trade-offs, such as between the athletic department and other parts of the university, arms races, and other factors that escalate expenditures within athletic departments. 5  For a comprehensive history of the development of the NCAA, see Grant, Leadley, and Zygmont (2008) or Fleisher, Goff, and Tollison (1992). 6  Walter Byers first called the players ‘studentathletes’ in 1951, when he became Executive Director of the NCAA, in order to justify their continued unpaid amateur status (Byers, 1995). 7  US Department of Education, Digest of Education Statistics, 2017, Table 301.2. 8  For 2016, Forbes estimates TV-related revenue, including ‘both rights fees and, for the conferences with network ownership stakes, estimated profit shares’ for the 10 Division I conferences, including the Southeastern Conference (SEC), the BigTen, the Big-12, the Pac-12, the Atlantic Coast

Conference (ACC), the American, the Big East, the Mountain West, the Conference USA, and the Mid-American as $1.38 billion, an increase by a factor of 15 over 32 years (Smith, 2016).  9  The 7–2 decision was announced on June 27, 1984. 10  In nominal dollars, the median intercollegiate athletics program among the 128 largest generated $48 million in 2015, 110 percent more than the median of $23 million in 2004 (Fulks, 2016, Table 2.1). 11  Between 71 (New England) and 88 (Pacific) percent of recent college graduates live in the same region as the college they graduated from (based on nine US geographic regions) one year after graduation (Sasser, 2008). 12  The Atlantic Coast Conference (ACC), Southeastern Conference (SEC), Big Ten Conference (Big-Ten), Big 12 Conference (Big-12), and Pac 12 Conference (Pac-12). 13  In addition to variation in the cost of a grant-inaid to universities with different tuition and fees, those institutions with excess capacity face only marginal cost of enrolling an additional athlete on a grant-in-aid, while institutions operating at capacity face losing the average net revenues from a non-athlete when they add a scholarship athlete. 14  Humphreys and Ruseski (2009) describe how the NCAA overcomes the difficulties of maintaining stability among heterogeneous members. 15  In 1987 and 1988 the NCAA closed the football program at Southern Methodist University because of ‘a lack of institutional control,’ when it discovered that athletic department employees were assisting in providing cash payments to players. 16  The largely fixed cost nature of the sports production function also encourages expanding the length of the regular sports season, and the addition of ever more post-season games. 17  Since the football television plan was dissolved in 1984, revenues from March Madness have covered NCAA operating costs. 18  Large flat-screen televisions were developed in the 1960s, but for practical purposes they are a twenty-first-century invention. 19  Sports marketing and sponsorship growth came as a natural by-product as television allowed for ‘product placements’ through which audiences could see what their favorite players and teams were wearing. Apparel became the largest category of sponsorships. Sports sponsorship was an $11 billion per year industry 20 years ago; today it exceeds $30 billion (Morgan, Johnson, and Summers, 2005). 20  This change was provoked by University of Connecticut basketball player Shabazz Napier announcing on national television immediately after the Huskies won the 2014 national collegiate men’s basketball title that he frequently

THE ECONOMICS OF THE NATIONAL COLLEGIATE ATHLETIC ASSOCIATION

went to bed hungry because of NCAA restrictions on ‘excess food.’ 21  This action added a sixth division to the NCAA for football governance purposes. 22  The NCAA once fixed the salaries of the fourth assistant basketball coach, but a 1998 Court of Appeals ruling held that this was collusion in restraint of trade, costing the NCAA a judgment of $66 million (Law v. National Collegiate Athletic Association, 134 F.3rd 1010 [10th Circuit 1998]). 23  Division II football teams seem able to field complete teams with a 63, rather than 85, as in Division I, scholarship limit. Brown (1993) estimated that a star college football player earned about half a million dollars per season for his team in the early 1990s. The amount would be much more today. 24  Lane, Nagel, and Netz (2014) successively relate player performance to winning, and winning to gate receipts to measure individual NCAA basketball players’ marginal revenue products. They find that the playing contributions of about 60 percent of the players generate revenues exceeding the value of their grants-in-aid. An analogous study of football players would help to identify the winners and losers if the pay cap were relaxed. 25  For a thorough discussion about the current market for college athletes, see Sanderson and Siegfried, 2015. 26  Fan reaction to paying players is probably the most important unanswered research question affecting college sports today. A first effort addressing fan reaction to paying players has found that consumer demand does not depend on preserving regulations that limit athletic compensation (Baker, Edelman, and Watanabe, 2018). 27  An assessment of changes in inequality within and among Division I FBS power conferences, Division I FBS non-power conferences, and Division I FCS institutions in terms of winning percentages, coaches’ salaries, attendance, bowl game and March Madness attendance and television ratings, might provide insight into the prospects of the NCAA disintegrating internally.

REFERENCES Alexander, Donald L. and William Kern. 2010. ‘Does athletic success generate legislative largess from sports-crazed representatives? The impact of athletic success on state appropriations to colleges and universities.’ International Journal of Sport Finance, 5, 253–267. Anderson, Michael L. 2012. ‘The benefits of college athletic success: An application of the propensity score design with instrumental variables.’ NBER Working Paper No. 18196 (June).

121

Baker, Thomas A., Marc Edelman, and Nicholas M. Watanabe. 2018. ‘Debunking the NCAA’s myth that amateurism conforms with antitrust law: A legal and statistical analysis.’ Tennessee Law Review, forthcoming. Berkowitz, Steve, Jodi Upton, and Erik Brady. 2013. ‘Most NCAA Division I Athletic Departments take subsidies. USA Today (July 1). Brown, Robert W. 1993. ‘An estimate of the rent generated by a premium college football player.’ Economic Inquiry, 31(4), 671–684. Byers, Walter. 1995. Unsportsmanlike Conduct. Michigan, MI: The University of Michigan Press. Carroll, Kathleen and Brad R. Humphreys. 2016. ‘Opportunistic behavior in a cartel setting: Effects of the 1984 Supreme Court decision on college football television broadcasts.’ Journal of Sports Economics, 17(6), 601–628. Clotfelter, Charles T. 2011. Big-Time Sports in American Universities. Cambridge: Cambridge University Press. Coates, Dennis, Brad R. Humphreys, and Li Zhou. 2014. ‘Reference-dependent preferences, loss aversion and live game attendance.’ Economic Inquiry, 52(3), 959–973. Fleisher, Arthur A. III, Brian L. Goff, and Robert D. Tollison. 1992. The National Collegiate Athletic Association: A Study in Cartel Behavior. Chicago, IL: University of Chicago Press. Fort, Rodney, and Jason Winfree. 2015. 15 Sports Myths and Why They’re Wrong. Stanford, CA: Stanford University Press. Fulks, Daniel L. 2016. NCAA Division I Intercollegiate Athletics Programs Report, 2004–2015: Revenues and Expenses. Indianapolis, IN: The National Collegiate Athletic Association (July). Getz, Malcolm and John J. Siegfried. 2012. ‘What does intercollegiate athletics do to or for colleges and universities? Chapter 19 in L. Kahane and S. Schmanske (Eds.), Handbook on Sport Economics. Oxford: Oxford University Press. Grant, Randy R., John Leadley and Zenon Zygmont. 2008. The Economics of Intercollegiate Sports. Singapore: World Scientific Publishing Company. Hoxby, Caroline. 2014. ‘The economics of online postsecondary education: MOOCs, nonselective education, and highly selective education.’ American Economic Review, 104(5), (May), 528–533. Humphreys, Brad R. 2006. ‘The relationship between big-time college football and state appropriations for higher education.’ International Journal of Sport Finance, 1, 119–128. Humphreys, Brad and Jane E. Ruseski. 2009. Monitoring cartel behavior and stability: Evidence from NCAA football.’ Southern Economic Journal, 72(4), 826–845. Knight Foundation Commission on Intercollegiate Athletics. 2006. Public Opinion Poll, January.

122

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Lane, Erin, Juan Nagel, and Janet S. Netz. 2014. ‘Alternative approaches to measuring MRP: Are all men’s college basketball players exploited?’ Journal of Sports Economics, 15(3), 237–62. Morgan, Melissa, Jane Johnson-Summers. 2005. Sports Marketing. Southbank, Vic.: Thomson. Raiborn, Mitchell H. 1986. Revenues and Expenses of Intercollegiate Athletic Programs: Report (Vol. 4). Indianapolis, IN: The National Collegiate Athletic Association. Sanderson, Allen R. and John J. Siegfried. 2015. ‘The case for paying college athletes.’ Journal of Economic Perspectives, 29(1), (Winter), 115–138. Sasser, Alicia. 2008. The Future of the Skilled Labor Force in New England: The Supply of Recent College Graduates. Boston, MA: Federal Reserve Bank of Boston, New England Public Policy Center (September).

Siegfried, John J. and Molly Gardner Burba. 2004. ‘The College Football Association Television Broadcast Cartel.’ The Antitrust Bulletin, Fall, 799–819. Smith, Chris. 2016. ‘Most valuable College Conferences 2016.’ Forbes, 26 July. Steinberg, Brian. 2015. ‘TV Ad Prices: Football, “Empire,” “Walking Dead,” “Big Bang Theory” Top the List.’ Variety, 29 September. Stigler, George. 1964. ‘A theory of oligopoly.’ Journal of Political Economy, 72(1), (February), 44–61. Strauss, Ben. 2014. ‘After ruling in O’Bannon Case, determining the future of amateur athletics.’ New York Times, 21 October. US Department of Education. 2017. Digest of Education Statistics. Washington, DC: Government Printing Office. Zimbalist, Andrew. 1999. Unpaid Professionals. Princeton, NJ: Princeton University Press.

PART III

Professional Team Sports

13 Economic Objective Functions in Team Sports: A Retrospective Rodney Fort

INTRODUCTION The choice of objective function is critical in order to generate believable explanations and predictions about behavior. We have known this in the area of the economics of team sports since Sloane (1971) urged utility maximization over profit maximization for the analysis of European football. Sometimes, it can take a while to settle such issues as witness the ‘discussion’ of relevant objective functions for team sports analysis in the 1990s (Fort & Quirk, 1995; Késenne, 1996; Szymanski & Kuypers, 1999; Fort & Quirk, 2004; Garcia-delBarrio & Szymanski, 2009). This chapter surveys the variety of objective functions used in the analysis of team sports and identifies areas where careful choice of objective function in future work will prove especially important.1 Referring back to the example in the preceding paragraph, the main lesson was that individual ownership in a world where profits are the objective will give different predictions about choices, different talent distributions, and different impacts of chosen league policy impositions such as revenue sharing and payroll caps than in a world where utility or revenue are the objective. The chapter consolidates the literature in this area around three main objectives – utility maximization, profit maximization, and revenue maximization.

It also covers the main original and closely proximate theoretical contributions and provides an overview of their main findings. The body of work is vast and updating to present is left to the interested reader and their skill with Google Scholar.2 The aim, as always, is to understand the context of objective function choice for modeling and to identify the future horizons for research. The chapter proceeds as follows. Sloane (1971) is covered in the next section alone because, well, he essentially laid out all of the details facilitating the choice objective functions in team sports. The next three sections review the works closest in history on the variety of objective functions set out by Sloane (1971), first with utility maximization and then on to profit and revenue maximization, respectively. The following section covers mixed approaches and social optimality. The chapter ends with conclusions and a summary.

THE ORIGINAL OVERVIEW: SLOANE (1971) Sloane (1971) eventually settled on utility maximization for European football after an exhaustive treatment of the extant possibilities – utility maximization, profit maximization, and revenue maximization.3

126

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The piece also needs to be recognized as doing for English football what Rottenberg (1956) did for baseball and Jones (1969) did for hockey.4 All of these incisive original works cover a multitude of league-specific labor, industrial organization, and regulation topics. These are the broadest shoulders to stand on from the very outset in almost any work on sports. Sloane (1971) first set out factors distinct to English football, compared to North American pro leagues, that would ultimately guide his choice of objective function. An element in Sloane (1971) that would endure and drive different paths in the literature on sports teams and leagues was in his opening observation (p. 1): ‘Whilst several North American contributions on the economics of sports exist … there appears to have been no attempt to apply economic theory to the particular case of British professional football….’ The different treatment of management objectives in different league structures in different countries, now commonplace, was the opening observation in Sloane’s (1971) seminal work. Mainly, these differences involve member club structures and national institutional arrangements unique to English football. Moving on to objective functions, he cites his inspiration on utility maximization as Williamson (1963) and Marris (1964) and gave careful consideration to the types of managerial objectives extant in the literature at the time. While rejecting profit maximization outright for English football, there are self-professed elements of security maximization (Rothschild, 1947) and sales (revenue) maximization (Baumol, 1957) included in his utility maximization framework. The utility formulation is reproduced here as a jumping off point for the next section. For Sloane (1971, p. 136), English football club directors are characterized as: max U [ P , A, X ,(π R − π 0 − T )] subject to π R + Fx ≥ π 0 + T U = P = A = X = πR = π0  = T = Fx =

(1)

The club director’s utility function Playing success Average attendance Health of the league Recorded profit Minimum after-tax profit Taxes External club financial resources

P translates most directly to winning. A is selfexplanatory, although the specification of attendance relative to winning, and its role in different places in the specification, are not and Sloane offers nothing on these except that they relate to

revenue. He had already discussed X on two dimensions, ‘mutual dependence’ in producing league outcomes and what modern followers of team sports would recognize immediately as ‘uncertainty of outcome’. In particular (Sloane’s, 1971, p. 136), ‘Utility is derived from the health of the league because it is better to win a keenly fought competition than to win easily.’ The constraint on the pursuit of utility was a portion of this specification that would play a large role in later work (see the subsequent sections on utility and revenue maximization) and Sloane (1971) wrestled with this one a bit. Surely, potential investors had to be given some signal of team vitality. Further, club directors would not bankrupt their private wealth position pursuing this utility. So, Sloane (1971) arrived at the idea of ‘acceptable’ profit as the constraint. Later works sometimes adjust this part of Sloane’s original specification to a simpler break-even constraint (again, see below). The remainder of his chapter suggests how focusing on utility maximization by club directors would prove insightful into the analysis of club talent choices as well as club finances. However, there is no appeal to the mathematics of optimization in Sloane (1971). Instead, he applies the basic marginal utility tradeoff implied in his specification between winning and profit.

UTILITY MAXIMIZATION Quirk and El Hodiri (1974) must be counted the first mathematically rigorous treatment of utility maximization.5 This objective function is not to be found again until Rasher (1997) and Vrooman (1997). For the novelty of it, I also cover an area unique to North America, namely, college sports in Carroll and Humphries (2000).6

Quirk and El Hodiri (1974) Quirk and El Hodiri (1974) admit early on that profit maximization may not fit some North American pro sports team owners very well (p. 42):7 The assumption that the actions of franchise owners are motivated solely by profits from operation of their franchise is admittedly somewhat unrealistic. Owning a major-league franchise carries with it prestige and publicity, and a wealthy owner might view it simply as a type of consumption; for such a ‘sportsman’-owner, winning games rather than making money might be the motivating factor.

Economic Objective Functions in Team Sports: A Retrospective

They offer a dynamic version of utility maximization. In choosing increments to talent over time, sportsman team owners:   max φ    

∑U t

∑ t

(Cti , pti ) (1 + ρ )− t ,   (2)  π ti (1 + δ )− t   i

and profit in the tradeoff. Outcome uncertainty is handled via the determination of winning probabilities incorporating relative team talent choices.8 There are, of course, differences in the distribution of talent across the league and the impact of some league policy impositions between profit maximization and utility maximization in the dynamic Quirk and El Hodiri (1974) specification. Those are left to the interested reader, with this preview from Quirk and El Hodiri (1974, p. 76):

subject to: I tj − I tj−1 =

∑x

jk t

In general, once the utility function contains as an element the probability of winning as a source of satisfaction distinct from its effect on profits, any earlier assertions about the relationships between the distribution of playing strengths and the drawing potentials of franchises must be severely qualified. In principle, a sufficiently wealthy owner concerned with ‘winning at all costs’ could attain his objective even if he owned a franchise in a small city, simply by spending enough money.

+ xtjN − α I tj−1 and

j≠k

π ti ≥ 0, Wti ≥ 0, Cti ≥ 0, I tj ≥ 0 ϕ = A general functional statement indicating the tradeoff between utility and profit Ui = Utility function of the owner of team i Cti = Consumption of other goods (than the probability of winning), time t pti = Vector of probabilities that team i wins against the rest, time t ρ = Positive rate of subjective discount π ti = Net cash flow from operating team i, period t δ = Market rate of interest per period I tj = Inventory of playing skills, team j, time t xtjk = Units of playing skill purchased from team k by team j, time t xtjN = Units of playing skill drafted by team j, time t α = Rate of talent inventory depreciation Wti = Wealth of the owner of team j, period t, not including the team The function ϕ sets up a marginal rate of substitution between utility (which will have its own marginal rate of substitution with consumption of other goods) and profit. For example, if ϕ is independent of Ui then only profit matters again. There are no restrictions at all on the form of the utility function. The constraints on ρ and wealth are difference equation in this dynamic setting. Profit depends on attendance revenues from playing skills (winning). In turn, playing skills are dynamically adjusted (purchased and drafted) to augment previous inventory, minus the costs (including depreciation of skill). Clearly, Quirk and El Hodiri (1974) are on the same wavelength as Sloane (1971). All of Sloane’s (1971) elements from expression (1) are in expression (2) – playing strength, attendance through the profit function, and non-negativity in both wealth

127

Rascher (1997) Rascher (1997) sets up utility maximization with an additively separable utility function in winning and profit:9 max Ui = α i Wini + (1 − α i ) π i



(3)

= Team owner utility = Proportion that owner i trades off winning and profit in their utility function Wini = Wins for team i πi = Profits from winning, required non-negative Ui αi

Comparing back to the Quirk and El Hodiri (1974) specification, this specification captures the essentials of the earlier formulations, albeit in a restricted way. The utility function is more restrictive. There is no ‘other consumption’ and it is linear in winning and profits. In an underlying specification, attendance makes its way into the owner’s consideration through the profit function. In addition, the formulation in Rascher (1997) adds a non-standard task for empirical assessment. It is challenging enough to get an empirical handle on the estimation of marginal effects dictated by utility maximization, but in addition the parameter αi, the owner’s personal tradeoff between wins and profits, requires measuring and estimating. Some of the results in Rascher (1997) hinge on the size of that parameter and any future empirical work would need to account for it.

128

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Vrooman (1997)

North American College Sports

Utility maximization is just one part of Vrooman (1997), an extensive treatment of all ownership, labor, and industrial organization topics for Major League Baseball. Vrooman (1997, starting on p. 598) refers to it as joint maximization of team value and satisfaction from winning by the ‘sportsman’ owner.10 The explicit listing of first-order conditions, and the graphical demonstration, allows the reader to backtrack to an optimization problem of the form:

Carroll and Humphreys (2000) were the first to apply utility maximization to college sports. Their athletic director (AD) maximizes utility subject to a break-even constraint (due to being part of a nonprofit organization):





 ( S M + SW ), U = U ( ST , G , R) = U   (5) G ( S ), R ( Q , S ) M M   subject to:

max S (V ( w), w) subject to V > V 0 (4) S = The sportsman owner’s satisfaction function V = The market value of the team w = Winning percentage

The ‘satisfaction’ function is clearly a utility function, right down to his graphical treatment of indifference curves between team value and winning percentage. The former is also the discounted present value of profit so the same tradeoff is in Vrooman (1997) as in the papers already covered. The constraint is a minimum acceptable team value to remain in the endeavor. Since this is the market’s determination of the profitability of the team, essentially this is a static version of the El Hodiri and Quirk (1974) constraint. Attendance is not formally included but clearly is subsumed in the idea of just how revenues are generated. The utility function underlying the analysis is completely general and the essence of the ‘sportsman owner’ tradeoffs is here (although the other usual marginal rate of substitution between winning and other consumption, in Quirk and El Hodiri (1974), is not). Vrooman’s (1997) aim with utility maximization is to show the changes that such an assumption makes on owner choices. In an equilibrium of sportsman owners, franchise values would be lower than would occur under profit maximization, since all attempt to increase winning relative to profit maximization but cannot do so simultaneously, driving up the price of players. He labels the difference in team value the ‘sportsman effect’. Vrooman (1997) also shows that this equilibrium has more competitive balance than under profit maximization. Vrooman (1997) extends the analysis of the sportsman owner’s tradeoffs in novel ways. First, he extends to the case of syndication. Second, echoing the earlier observation on the mysterious Fx in Sloane (1971), he covers the use of ‘other people’s money’, that is, financial leverage by the sportsman owner.

[ R(Q, S M ) − C (Q, S M , SW )] ≥ 0 and



SW − α S M ≥ 0, 0 ≤ α ≤ 1 U = Athletic director utility ST = SM + SW = Total quantity of men’s event and women’s event staff G = Prestige R = Revenue Q = QW + QM = Total quantity of women’s and men’ programs and events C = Total cost of all production α = Staff of women’s events proportional to staff of men’s events

Within the context of nonprofit sports at a university, the AD gains satisfaction from his department size, prestige, and revenue. The focus of the analysis is in the tradeoff confronting the unregulated athletic director between men’s and women’s sports. In the US, this is an important regulatory area under Title IX in the education laws.11

PROFIT MAXIMIZATION El Hodiri and Quirk (1971) is the original on profit maximization in North American professional sports.12 The closest proximate works are Atkinson, Stanley, and Tschirhart (1988), followed by Fort and Quirk (1995) and Vrooman (1995).13

El Hodiri and Quirk (1971) El Hodiri and Quirk (1971) do the profit maximization case explicitly for North American sports and in the rigorous fashion not found in Sloane (1971). The paper is rigorously dense and only a flavor can be given here. The decision problem is:14 max

∫

∞

0

 Ri − wI i − bN xiN − bi  



∑ x  e ij

j



−δ t

(6)

Economic Objective Functions in Team Sports: A Retrospective

subject to:

I j = x jN +

∑x

jk

− µI j , and

k

I j ≥ 0, j = 1, , n Ri = Team i revenue under a revenue sharing formula w = Wage rate per unit of playing skill (equal across all teams) Ii = Team i playing skill stock, i = 1, …, n bN  = Price per unit of playing skills acquired in the draft by team i xiN = Units of playing skill acquired by team i in the draft of new players bi = Price per unit of playing skills acquired in a purchase by team i from other teams xij = Units of playing skill acquired by purchase of contracts by team i from team j An interior solution to the Hamiltonian representing the first-order condition differential equations is found. They then show that the league will be unbalanced. They proceed to show that forbidding cash sales of contracts only makes matters worse. Their conclusion is, under rules at that time (remember, 1971 was prior to free agency), North American sports leagues would exhibit no tendency toward equalization of playing strength.

Atkinson, Stanley and Tschirhart (1988) For Atkinson, Stanley and Tschirhart (1988), the point is to cast profit maximizing team owners in their league decision-making format to investigate the incentives inherent in revenue sharing. Their model is:

max π i = TRi − ct i(7) πi = Profit for team i TRi = R  i(w(t), Ai) = Total revenue for team i w(t) = Vector of total wins for all teams, a function of each team’s talent choice, ti i A = Population of the ith team’s home city c = Unit talent cost facing all teams in the league

TRi is first analyzed without revenue sharing, revealing the talent choice externality inherent in paying attention only to own revenues and thus the problem for the league.15 They prove wages will equal MRP but that MRP is different with and without revenue sharing so there is an individual owner incentive to

129

reduce talent choice under revenue sharing. In the league optimal profit outcome, internalizing the externality, there is less imbalance. They reject the implications empirically for the NFL, concluding that utility maximization is likely in that league.

Fort and Quirk (1995) Fort and Quirk (1995) use a simplified static version of the profit maximization model in Quirk and El Hodiri (1974) (in turn based on El Hodiri and Quirk (1971)) and offer theoretical and empirical investigations of the impacts of league impositions (player drafts, revenue sharing, and payroll caps). There is a significant amount of underlying complexity concerning the elements of the following revenue function, but their optimization model is:

max π i = Ri [t; α , n, A] − ct i (8) πi = Team i profit Ri = Team i revenue t = Vector of talent for all teams in the league α    = Portion of revenue kept by the home team n = Number of teams in the league A    = Vector of drawing potential for each team, across all teams in the league c   = Unit cost of talent the same for all teams ti = Talent chosen by team owner i

Since there is sharing, revenue for any team depends on its own home revenue-generation and the same when the team play away from home. Local and national TV are also part of the revenue function. They first examine individual team optimality, revealing the externality of individual talent choice. They show that equilibrium will also exhibit talent imbalance. They go on to examine the imposition of various league policies. Although the model is later criticized for its assumptions and omissions, they confirm Rottenberg’s (1956) original speculation that drafts and free agency do not alter balance. Another unique finding is that revenue sharing will not alter the level of balance in a league (from their earlier graphical intuition in Quirk and Fort, 1992).16

Vrooman (1995) While he does not directly state it, his development leads to the following objective function:

max Ri − Ci(9)

130

THE SAGE HANDBOOK OF SPORTS ECONOMICS

β

Ri = R0 piα wi = Total revenues for team i R0 = Exogenous revenue for any team pi = Home market size for team i, with market size revenue effect exponent α, or market size cost exernality exponent γ wi = Winning percentage for team i, with revenue elasticity of winning exponent β, or cost elasticity of winning exponent δ γ Ci = C0 pi wiδ = Total costs for team i C0 = Exogenous cost for any team With ‘customary’ assumptions on the exponents of revenues and costs, he argues earlier work to be special cases of those assumptions. More importantly, he explores the impacts of the exponential elasticity specifications on relative winning percentages (competitive balance). He also finds that revenue sharing will depress player pay but not change balance and goes on to explore the payroll cap, free agency. He also estimates the cost and revenue functions.

REVENUE MAXIMIZATION As argued presently, Késenne (1996) should be recognized as the original on revenue maximization in team sports. Késenne also cornered the market on subsequent work on revenue maximization, but one need only cite his book to point the reader in the right direction (Késenne, 2014). In keeping with the presentation of novel approaches, again on North American college sports, Fort (2018a) is presented.

Késenne (1996) Késenne (1996) states that his win maximization framework is a simple variety of utility maximization where only wins matter, subject to the club manager breaking even. This is not entirely clear to this reader for two reasons. First, nearly any model can be cast as simplified utility maximization of some form or another. Second, going all the way back to Baumol (1957, 1972), all of his results are consistent with revenue maximization (market share) in an oligopoly setting with the value of winning set equal to unity. As sales maximization, Késenne’s (1996) work offers heretofore unexplored implications and insights about the pursuit of market share among pro sports leagues.

In any event, the problem in Késenne (1996) is not formally stated but matches:

max L subject to R( M , L ) − WL ≥ 0(10) L = Units of playing talent for the team, relative to the talent in the rest of the league R = Team revenue M = Team market size W = Talent cost per unit

Since wins follow directly from L, he simply maximizes talent subject to a break-even constraint. The impact of attendance is subsumed (but not formally treated) in the revenue function. Baumol’s (1957, 1972) lesson is Késenne’s (1996) direct outcome (portrayed graphically in his paper). Units of playing talent (thus, wins) is chosen to equate the average revenue from winning, rather than the marginal revenue from winning, to the cost of talent, W. Thus, talent investment is larger in the winmaximizing league than under profit maximization. Késenne (1996) also shows that revenue sharing will increase competitive balance in this framework.17

North American College Sports Following his earlier descriptive attempts (Fort, 2015a, 2016), Fort (2018a) offers a rigorous examination of the determination of the distribution of talent in a conference of revenue-maximizing athletic directors. As they do for all units on campus, university administrators (UAs) determine the level of athletic outcomes that satisfies their own objectives – research, teaching, and service. Their oversight responsibility leads them to decide how much to invest in athletic outcomes. They then budget the athletic department accordingly. At the highest level of college competition, athletic directors appear to best serve the objectives of university administrators by maximizing revenue. Given the size objective that UAs have for their athletic department, revenue maximization minimizes the budget allocation investment by UAs. In Fort (2018a) athletic directors invest in the talent market to produce observed talent during play. Winning is dictated by relative talent of competing teams and revenues follow from that winning outcome. Revenue maximization for an AD with a two-sport output is:

max R11 + R12 subject to R11 + R12 − z11 − z12 ≥ 0

(11)

Rij[Wij (t(z))] = Revenue for ADi, i = 1, 2, from winning in a sport j = 1, 2, which in turn depends

Economic Objective Functions in Team Sports: A Retrospective

on the vector of talent on the field across the athletic departments t derived from investment in that talent z. Reducing the problem to the optimal choice by each athletic director, given the optimal choice of the other athletic director, he shows that talent imbalance is to be expected in equilibrium. In addition, the absolute level of talent increases, and competitive balance improves, in the conference with revenue sharing. However, competitive balance is unaffected by the amateur requirement (a form of payroll cap) as it is implemented in college sports.

VARIATIONS ON THE THEME The ideas above lead naturally to a popular social welfare specification, the sum of consumers’ and producers’ surpluses. Dietl and Lang (2008) and Fort and Quirk (2010, 2011) provide the earliest use of this particular social planner’s objective function.18 At this point, the formulations are left to the interested reader to pursue and a basic description follows. All of the works use consumers’ surpluses and profit for producers’ surpluses. However, the level of generality is different across them all. Dietl and Lang (2008) use specific functional forms for demand and revenue (hence, profits) but do not distinguish leagues any further. In addition, preferences for particular elements of competition are parameterized (much as in Rascher, 1997). Their non-cooperative equilibrium exhibits competitive imbalance, and revenue sharing leads to less talent investment and a more unbalanced league. Under the social planner, the league is more unbalanced and revenue sharing increases social welfare relative to the non-cooperative outcome. Interestingly, winning percentages are the same for a league optimum (the cooperative outcome as opposed to the noncooperative) and for the social planner’s optimum. Fort and Quirk (2010, 2011) use general functional forms, but distinguish between leagues where single-game ticket sales dominate (baseball) and leagues where season-ticket sales dominate (football). They reach pretty much the same conclusion in either case (2010, p. 594): ‘The upshot of all of this is that the determination of a league’s choice of balance, compared to one that would maximize welfare as defined here, is ultimately a complicated empirical task.’ Fortunately, the tried and true social planner’s optimum yields conditions under which an increase in balance is welfare-improving. And these conditions depend only on economic relationships that all can be estimated. For the single-game league, the empirical task is to estimate how incremental impacts of improved

131

balance on fans (that is, the areas under the new demand curves for each home game) sum up across all home games in a given market. Then, empirically, to determine how those totals compare between markets. Things are simpler for the seasonticket league. The impact of improved balance rests solely on shifts in demand in the smaller-revenue market, compared to the larger-revenue market.

CONCLUSIONS This chapter surveys the variety and application of objective functions used in the analysis of team sports, namely, utility maximization, profit maximization, and revenue maximization. The original works in the area and those that closely followed are covered. The works since then are now readily available quickly through Google Scholar searches and are not reproduced here. Mainly, the lesson learned is that team sports outcomes depend crucially on the relevant objective function. This makes it abundantly clear that applying the appropriate objective function, in its appropriate context (e.g., North American versus world leagues), determines the usefulness of the modeling results. What happens varies by objective function. Here is an example, which is part of my own research agenda. Fort, Kang, and Lee (2015) and Fort (2018b) detail the historical evolution and current status of Asian professional baseball leagues and current challenges. In both Japan and Korea, these leagues were formed as whollyowned subsidiaries of nationalized companies under centrally planned post-war rebuilding of their respective economies. As such, they have served as primarily advertising arms for those parent companies. Given the Asian economic boom has ended, parent companies are re-evaluating this relationship. To formally model anything about these leagues, the situation demands attention to the role of input demand derived from the parent company’s objective function. The teams themselves do not maximize profit. Since they are a budgeted subdivision of the parent company, perhaps revenue maximization makes more sense. Finally, it should be true that the choice of objective function will loom large in the ever-changing sports economics landscape at large. The operation of stadiums and arenas is now a part of larger real estate and entertainment complex operations. The value of sports intellectual property rights seems to be now part and parcel of the operation of media enterprises like regional sports networks.

132

THE SAGE HANDBOOK OF SPORTS ECONOMICS

And the medium itself is changing dramatically, suggesting that something new may take the place of sports networks. Finally, what objective function governs fantasy sports or eSports? Surely, just as it has in the past, the choice of objective function – whether one well-used or a novel approach – will enrich the impact of theory on our understanding and on future empirical work.

Notes  1  A variety of other objective functions are out there that may brush up against team sports and there is always a bit of subjective judgment on inclusions. For example, see Szymanski (2003) in the area of contests more generally considered.   2  To give the reader a feel for why this choice was made, Google Scholar yields the following just for the ‘backbone’ papers reviewed here: Sloane (1971) – 516 cites, El Hodiri and Quirk (1971) – 555 cites, Quirk and El Hodiri (1974) – 165 cites, and Késenne (1996) – 165 cites.   3  Sloane (1969) had already bridged the gap from the more general ‘managerial’ economics to sports economics. Dabscheck (1975) and Cairns, Jennett, and Sloane (1986) show ample evidence of violation of profit maximization for many nonNorth American leagues.   4  Neale (1964) also needs recognition among the original sports papers but his point was the uniqueness of the production function in sports and the need for some form of competitiveness in the eyes of fans.   5  This section relies heavily on my earlier review of utility maximization in team sports (Fort, 2015b).   6  An additional recent nuance concerns the maximization of club members utility in the European football context (Madden, 2012; and Madden & Robinson, 2012).   7  They reprise their earlier profit maximization version in El Hodiri and Quirk (1971) with some extensions as well. This earlier work is the headliner of the next section of this chapter.   8  This is the well-known contest success function in the sports literature years before the usually cited Tullock (1980).   9  Rascher (1997) does not cite Quirk and El Hodiri (1974) and must be viewed as an independent development. 10  Vrooman (1997) cites Quirk and El Hodiri (1974) throughout his paper, but not on the ‘sportsman owner’ (see the offset quote, above). He does not cite Sloane (1971), either but then his explicit focus is Major League Baseball. 11  Since the college literature using utility maximization is small, it can be covered quickly.

The models that followed were very similar (Leeds, 2002; Leeds, Suris, & Durkin, 2004). Hoffer, Humphreys, Lacombe, and Ruseski (2015) added the spending choices of rivals to the same type of utility maximization approach. 12  Both Rottenberg (1956) and Jones (1969) simply took profit as the objective without much discussion and without any rigor. 13  There have been other applications in the unexpected place of college sports (Brown, 1994; Fort & Quirk, 1999). But at least one of the latter authors (yours truly) has moved on to more complete representations of college sports (see the next section). 14  A definition of the parameter μ was not found in the paper. It appears to be the proportion that determines salvage stock from the previous period. 15  The authors account for the league adding-up constraint, but not the underlying contest success function (so Nash conjectures are ignored) or the elasticity of talent (whether the league is open or closed). 16  Carrying on from the observation above about the shortcomings in Atkinson, Stanley, and Tschirhart (1988), Fort and Quirk (1995) suffer the same as carefully delineated in Szymanksi (2003, 2004) and Szymanski and Késenne (2004). This is mostly sorted out later by Eckard (2006), Szymanski (2006), Winfree and Fort (2012) and Szymanski (2013). 17  All other league-imposed policies are investigated in his subsequent win maximization work; again, see Késenne (2014). 18  Analysis using a social planner’s problem did appear earlier in Dietl and Hasan (2008) but it was not as general since it was aimed only at assessing the social welfare of a particular policy issue (the treatment of television rights sales under pay TV and free TV).

REFERENCES Atkinson, S.E., Stanley, L.R., & Tschirhart, J. (1988). Revenue sharing as an incentive in an agency problem: An example from the National Football League. RAND Journal of Economics, 19(1), 27–43. Baumol, W.J. (1957). Business behavior, value and growth. New York: Macmillan. Baumol, W.J. (1972). Economic theory and operations analysis (3rd ed.). Englewood Cliffs, NJ: Prentice-Hall. Brown, R.W. (1994). Incentives and revenue sharing in college football: Spreading the wealth or giving away the game? Managerial and Decision Economics, 15(5), 471–486.

Economic Objective Functions in Team Sports: A Retrospective

Cairns, J., Jennett, N., & Sloane, P.J. (1986). The economics of professional team sports: A survey of theory and evidence. Journal of Economic Studies, 13(1), 3–80. Carroll, K.A., & Humphreys, B.R. (2000). Nonprofit decision making and social regulation: The intended and unintended consequences of Title IX. Journal of Economic Behavior & Organizations, 43(3), 359–376. Dabscheck, B. (1975). Sporting equality: Labor market versus product market control. Journal of Industrial Relations, 17(2), 174–190. Dietl, H.M., & Lang, M. (2008). The effect of gate revenue sharing on social welfare. Contemporary Economic Policy, 26(3), 448–459. Eckard, E.W. (2006). Comment: Professional team sports are only a game: The Walrasian fixed-supply conjecture model, contest-Nash equilibrium, and the invariance principle. Journal of Sports Economics, 7(2), 234–239. El Hodiri, M., & Quirk, J. (1971). An economic model of a professional sports league. Journal of Political Economy, 79(6), 1302–1319. Fort, R. (2015a). College sports spending decisions and the academic mission. In E. Comeaux (Ed.), Introduction to intercollegiate athletics (pp. 135–146). Baltimore, MD: Johns Hopkins University Press. Fort, R. (2015b). Managerial objectives: A retrospective on utility maximization in pro team sports. Scottish Journal of Political Economy, 62(1), 75–89. Fort, R. (2016). Collegiate athletic spending: Principals and agents v. arms race. Journal of Amateur Sport, 2(1), 119–140. Fort, R. (2018a). Modeling competitive imbalance and self-regulation in college sports. Review of Industrial Organization, 52(1), 231–251. Fort, R. (2018b). Reorganization challenges facing Korean baseball. In D.H. Kwak, Y.J. Ko, I. Kang, & M. Rosentraub (Eds.), Sport in Korea: History, development, management (pp. 193–209). London, UK: Routledge. Fort, R., Kang, J.H., & Lee, Y.H. (2015). KBO and international sports league comparisons. In Y.H. Lee & R. Fort (Eds.), The sports business in the Pacific rim (pp. 175–194). New York: Springer. Fort, R., & Quirk, J. (1995). Cross-subsidization, incentives, and outcomes in professional team sports leagues. Journal of Economic Literature, 33(3), 1265–1299. Fort, R., & Quirk, J. (1999). The college football industry. In J. Fizel, E. Gustafson, & L. Hadley (Eds.), Sports economics: Current research (pp. 11–26). Westport, CT: Praeger. Fort, R., & Quirk, J. (2004). Owner objectives and competitive balance. Journal of Sports Economics, 5(1), 20–32. Fort, R., & Quirk, J. (2010). Optimal competitive balance in single-game ticket sports leagues. Journal of Sports Economics, 11(6), 587–601.

133

Fort, R., & Quirk, J. (2011). Optimal competitive balance in season ticket leagues. Economic Inquiry, 49(2), 464–473. Garcia-del-Barrio, P., & Szymanski, S. (2009). Goal! Profit maximization versus win maximization in soccer. Review of Industrial Organization, 34(1), 45–68. Hoffer, A., Humphreys, B.R., Lacombe, D.J., & Ruseski, J.E. (2015). Trends in NCAA athletic spending: Arms race or rising tide? Journal of Sports Economics, 16(6), 576–596. Jones, J.C.H. (1969). The economics of the National Hockey League. Canadian Journal of Economics, 2(1), 1–20. Késenne, S. (1996). League management in professional team sports with win maximizing clubs. European Journal for Sport Management, 2(2), 14–22. Késenne, S. (2014). The economic theory of professional team sports: An analytical treatment (2nd ed.). Northampton, MA: Edward Elgar. Leeds, M.A. (2002). Collegiate athletic directors as entrepreneurs. Journal of Entrepreneurial Finance, 7(2), 33–43. Leeds, M.A., Suris, Y., & Durkin, J. (2004). College football and title IX. In J. Fizel & R. Fort (Eds.), Economics of college sports (pp. 137–152). Westport, CT: Praeger. Madden, P. (2012). Fan welfare maximization as a club objective in a professional sports league. European Economic Review, 56(3), 560–578. Madden, P., & Robinson, T. (2012). Supporter influence on club governance in a sports league: A ‘utility maximization’ model. Scottish Journal of Political Economy, 59(4), 339–360. Marris, R. (1964). The economic theory of ‘managerial’ capitalism. Glencoe, IL, and London: Free Press. Neale, W.C. (1964). The peculiar economics of professional sports: A contribution to the theory of the firm in sporting competition and in market competition. Quarterly Journal of Economics, 78(1), 1–14. Quirk, J., & El Hodiri, M. (1974). The economic theory of a professional sports league. In R. Noll (Ed.), Government and the sports business (pp. 33–80). Washington, DC: Brookings Institution. Quirk, J., & Fort, R.D. (1992). Pay dirt: The business of professional team sports. Princeton, NJ: Princeton University Press. Rascher, D.A. (1997). A model of a professional sports league. In W. Hendricks (Ed.), Advances in the economics of sport (vol. 2, pp. 27–76). Greenwich, CT: JAI Press. Rothschild, K.W. (1947). Price theory and oligopoly. Economic Journal, 57(227), 299–320. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258.

134

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Sloane, P.J. (1969). The labour market in professional football. British Journal of Industrial Relations, 7(2), 181–199. Sloane, P.J. (1971). The economics of professional football: The football club as a utility maximiser. Scottish Journal of Political Economy, 18(2), 121–146. Szymanski, S. (2003). The economic design of sporting contests. Journal of Economic Literature, 41(4), 1137–1187. Szymanski, S. (2004). Professional team sports are only a game: The Walrasian fixed-supply conjecture model, contest-Nash equilibrium, and the invariance principle. Journal of Sports Economics, 5(2), 111–126. Szymanski, S. (2006). Reply: Professional team sports are only a game: The Walrasian fixed-supply conjecture model, contest-Nash equilibrium, and the invariance principle. Journal of Sports Economics, 7(2), 240–243. Szymanski, S. (2013). Some observations on Fort and Winfree’s ‘Nash conjectures and talent supply in sports league modeling: A comment on current modeling disagreements’. Journal of Sports Economics, 14(3), 321–326.

Szymanski, S., & Késenne, S. (2004). Competitive balance and revenue sharing in team sports. Journal of Industrial Economics, 52(1), 165–177. Szymanski, S., & Kuypers, T. (1999). Winners and losers: The business strategy of football. London, UK: Viking. Tullock, G. (1980). Efficient rent seeking. In J.M. Buchanan, R.D. Tollision, & G. Tullock (Eds.), Toward a theory of the rent-seeking society (pp. 97–112). College Station, TX: Texas A&M University Press. Vrooman, J. (1995). A general theory of professional sports leagues. Southern Economic Journal, 61(April), 971–990. Vrooman, J. (1997). A unified theory of capital and labor markets in Major League Baseball. Southern Economic Journal, 63(January), 594–619. Williamson, O.E. (1963). Managerial discretion and business behavior. American Economic Review, 53(5), 1032–1057. Winfree, J., & Fort, R. (2012). Nash conjectures and talent supply in sports league modeling: A comment on current modeling disagreements. Journal of Sports Economics, 13(3), 306–313.

14 European Sports Leagues: Origins and Features Nicolas Scelles and Jean-François Brocard

INTRODUCTION A sports league is ‘a group of teams that schedules games and develops other policies and rules for the purpose of determining a champion’ (Noll, 2003, pp. 530–531). European sports leagues appeared at the end of the second half of the 19th century. The English Football League was the first in 1888, followed by the County Championship in England and Wales (cricket) in 1890 and the French Rugby Championship in 1892. The first two were already a round-robin competition with teams annually playing one another home and away (Harris, 1975), while the latter was initially a knockout tournament. It became a round-robin competition in 1895–1896, concomitant with the establishment of the Northern Rugby Football Union in Northern England, which was also a round-robin competition. The initial evolutions of these four leagues were quite different: • The English Football League absorbed its rival Football Alliance (created in 1889) in 1892, with the latter becoming the Second Division and the two divisions being linked together by a system of promotions and relegations or ‘open’ system (Szymanski, 2003a). Thus, the Football League did not

break away from the existing structures and admitted all the major teams into its ranks (Inglis, 1988). • The County Championship accepted new counties rather than creating a second division or controlling its group size (Schofield, 1982). • The French Rugby Championship, initially opposing clubs from the Parisian area only, accepted provincial clubs from 1898–1899. More exactly, they formed two different groups and each group winner faced each other in final. In 1899–1900, the eight participants to the final elimination tournament were the eight regional winners. In 1900–1901, there were three regions with the winner of the Seine region directly qualifying for the final. The number of regions or groups evolved over time but the format with several regions or groups then their winners facing each other to determine the French champion remained the same for most of the 20th century. • The Northern Rugby Football Union split into two separate county competitions in 1896–1897 (Lancashire and Yorkshire) before the top seven sides from both counties in 1900–1901 resigned and merged into a new league in 1901–1902, joined by four additional clubs in 1902–1903, when a second division was established with the two divisions linked together by a system of promotions and relegations.

136

THE SAGE HANDBOOK OF SPORTS ECONOMICS

European basketball leagues appeared later, in the 1920s (1920 in Italy, 1921 in France, 1923 in the Soviet Union, 1927 in Greece). Continental club competitions appeared even later, in the 1950s. Based on these illustrations, and contrary to an expanded idea, we realize that European sports leagues were not all based on promotions and relegations from their beginning. This challenges the pyramidal structure that is usually considered as a feature of all European sports. Even when such a system existed between the first and the second divisions for a given sport in a given country, this was not automatically the case between the second division and a lower level. For example, the French football league was established in 1932. However, it was only from the 1970–1971 season that clubs from the second division could be relegated in the third division (Scelles, Szymanski, & Dermit-Richard, 2018). Nevertheless, it is still true that if the focus is on the main European sport (football) and on its main divisions (the first two) in a given country, such divisions have always been linked together by promotions and relegations. As such, this can be seen as a main distinctive feature of European sports leagues from their origins, especially when compared to US sports leagues and from a sports economics perspective. Research into the application of economic concepts to sporting activities has mushroomed in recent decades: whether it be the contribution of sporting activities to economic growth, competition for media rights, labor markets for sports stars or the economic incentives embedded in the structure of league rules. This chapter aims to establish the context in which the origins and features of European sports leagues arise in sports economics; review and update the understanding of the literature; and identify avenues for future research.

CONTEXT IN WHICH THE TOPIC ARISES IN SPORTS ECONOMICS Overview One might think that the origins and features of European sports leagues arise in sports economics with the first studies on European sports leagues. These studies were conducted by Peter J. Sloane (1969, 1971, 1976) on British football. Nevertheless, although they dealt with the issues encountered at that time (labor market, team objectives, restriction of competition) and as such with some features of European sports leagues, they did not focus on their origins but rather on the leagues’ efficiency. The first article to really envisage both origins and features was that of Schofield (1982)

on the development of first-class cricket in England. As underlined by Schofield (1982, p. 337) at the beginning of his article: Economists have given some attention to cricket in recent years, as they have to other professional sports, without focussing in detail on the history of the game. […] Economic historians have also shown interest in the game but this has been confined to aspects of the labour market in the early days of organized cricket. […] None of the above work traces the development of the game using the tools of economic analysis.

More generally, the origins and features of European leagues were first analyzed stricto sensu in the context of the comparisons of North American and European professional sports leagues developments. Sport economists and economic historians, observing that until 1892 the development of sport was similar on both sides of the Atlantic, wondered why sports leagues structures diverged afterwards (Cain & Haddock, 2005; Szymanski, 2003a). They analyzed them under the scope of industrial organizations and found some similarities despite their diverging structures and objectives: ‘It is of interest that, despite an emphasis on differing objectives of professional sports clubs and leagues in Britain and the United States, the development of league rules in cricket (Schofield, 1982) and baseball (Davis, 1974) have been analysed from the perspective of cartel organisation’ (Cairns, Jennett, & Sloane, 1986, p. 58). As for the creation of leagues, economists underline the fact that team sports require coordination among contesting teams because the main product – a game – involves at least two distinct entities with potential divergent interests (Neale, 1964; Noll, 2003; Sloane, 1976). Nevertheless, Noll (2003) underlines that if several teams regularly play each other, there is an incentive to create a league. The formation of the Football League provides an interesting illustration as it was motivated by the fact that prior to 1888, some teams would not ‘fulfil the friendly fixtures that they had promised if they became involved in the later stages of the various cup competitions’ (Vamplew, 2006, p. 435). As such, teams were grouped together to ensure production of the common product.

European Leagues as ‘Open’ Cartels Sports economists also focused on other aspects of leagues, such as their optimal features, and observed their cartel-like behavior: ‘Professional team sports leagues are classic, even textbook, examples of business cartels (Quirk, 1987)’ (Fort & Quirk,

European Sports Leagues: Origins and Features

1995). This suggested feature requires providing the objective of a cartel in general before observing whether sports leagues fit with such objective. Cairns et al. (1986, p. 56) note that ‘the objective of a cartel is to determine a structure of rules constraining the behaviour of the group’s individual members to act in the interests of the group as a whole’. The way sports leagues operate is consistent with this framework, as Cairns et  al. (1986, p. 56) stress: ‘Analytically, the rule-making activities of leagues can be seen as a form of cartel behaviour.’ In the general case, such a framework intends to generate profit. Indeed: ‘According to cartel theory, output limitation and the erection of effective barriers to entry are a sine qua non for profit-seeking cartel activities’ (Cairns et  al., 1986, p.  59). However, a successful sports league is not necessarily a league maximizing the joint profit for its members, but a league achieving both a distribution of playing talent, ensuring sporting competition, and a distribution of income, ensuring the economic survival of the weaker members of the group (Sloane, 1976). Besides: ‘That a successful league must be a cartel fails to tell whether the cartel will be open or closed’ (Cain & Haddock, 2005, p. 1144). European sports leagues are then seen as ‘open’ cartels in opposition to the North American ‘closed’ cartels, and economists then highlight the discrepancies. An interesting illustration lies in the different systems of reduction of horizontal competition. Indeed, it appears that members of a sports league may benefit from a low level of economic rivalry between teams. In this context, ‘teams have a strong incentive to organize leagues in a fashion that reduces the extent of horizontal competition among them in both input and output markets’ (Noll, 2003, p. 531). While that led to territorial rights in US closed leagues, in European open leagues a newly formed team can simply register to play in football’s lowest tier and needs no one’s permission to be there.

The Optimal Size of League and Geographical Location of Members Economists interested in the evolution of professional sports leagues also show how different geographic compactness between European and North American countries contributed to these substantial differences in league structures: The excitement of closely contested games required little travel in England, and compactness provided a way both to limit peripheral players’ wage demands and increase attendance. The potential additional revenues from territorial monopolies could not offset the increased costs.

137

The number, quality and density of established English teams made a closed cartel untenable, and promotion and relegation offered a way to sort the teams. (Cain & Haddock, 2005, p. 1144)

The last elements open the door to two other issues tackled by sports economists: the optimal size of league and geographical location of members. According to Sloane (1976, p. 19): ‘The optimal size of league […] will be a function not only directly of the total population of the area covered by the league, but also inversely of population inequalities among clubs.’ The author develops the opposition between the financial rationale for reducing the number of clubs and the better consumer welfare, with more clubs covering more geographical areas and, as such, minimizing mean distances. Sloane (1976) underlines that an alternative product market proposal, put forward by Demmert (1973), is club relocation, but this was unknown in Britain at the time of writing. In cricket, Schofield (1982) notes that, while the County Championship did not initially control its size, no counties were added to the first-class list from 1921, demonstrating an awareness of the league organizers that the number of clubs should be limited, consistent with Sloane (1976). Besides, Schofield (1982, p. 341) adds that: ‘As well as having control over the size of the group, it is necessary to preclude external competition in product and input markets in order to protect group profits and hence the viability of the group.’ This is confirmed by Cairns et  al. (1986), who nevertheless nuance the threat. Indeed, these authors stress the following elements: ‘As Noll (1982) points out, however, the formation of a new league is only possible if a sufficiently large number of cities possess excess demand for the sport in question. Further, there is some doubt as to whether rival leagues are a stable, long-run, equilibrium solution’ (Cairns et al., 1986, p. 59). Cairns et al. (1986) also point out that leagues conventionally determine the geographical location of members. They take the example of English cricket, where various counties are represented by only one team. In respect to football, they stress an ambivalent reality: ‘The geographical transfer of soccer clubs is unfamiliar, but league clubs retain control over election to and dismissal from the league (discussed briefly by Jennett & Sloane, 1985)’ (Cairns et al., 1986, pp. 59–60). Moreover, the rules set by leagues can prevent a club with limited economic potential from being promoted to the first tier. For example, in France, FC Gueugnon should have been promoted to the first division in 1979, based on its sporting performance, but was prevented from doing so because the club did not want to become professional due to its limited economic potential.

138

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The Labor Market As underlined above, most articles have been interested in the features of the leagues themselves as a product. However, this section would be incomplete without mentioning the labor market since the very first article on European leagues by Sloane (1969) is focused on the labor market in professional football. Sloane (1969) wrote his article in the context of the Chester (1968) report on English football, which discussed at some length the economic implications of the wages and the conditions of employment of the professional footballer and included recommendations for the revision of contractual relations between the player and his club. Sloane (1969) mentions in his introduction a comparison – more exactly, an opposition – to North America, which is important given the fact that studies on European sports leagues were originally conducted by comparison with their North American counterparts. Indeed, Sloane (1969, p. 182) stresses that his premises are different from Rottenberg (1956): the latter ‘argues that the football club is a profit maximizer, just like the business firm. This author [Sloane], on the other hand, has argued elsewhere that football clubs are essentially utility maximizers – the majority of clubs making a loss on their football activities only survive through subsidies derived from various sources.’ This early opposition between North American and European sports leagues is still largely accepted around 50 years later, even though a North American club is not automatically a profit-maker (Andreff, 2015; Lavoie, 2004) and a European club may seek to make a profit (Terrien, Scelles, Morrow, Maltese, & Durand, 2017). According to Sloane (1969, p. 182), team objectives and the need for strong competitors ‘are crucially important when assessing the conditions under which the professional footballer is employed and explain why such marked differences must exist in terms of employment as compared with the industrial worker.’ These conditions and differences are reviewed below.

REVIEW AND UPDATE OF THE UNDERSTANDING OF THE LITERATURE League Model until the Beginning of the 1980s The evolving model of European team sport leagues can be separated into two periods: until the beginning of the 1980s and from the 1980s

onward (Andreff & Staudohar, 2000; Downward, Dawson, & Dejonghe, 2011). Indeed, the 1980s marked the beginning of the growth of TV rights and the arrival of rich owners that transformed the model. Prior to this, European leagues developed by adopting the British model of league organization initiated by the Football League in 1888 (Inglis, 1988). To describe the features of such a model, the literature focuses on the type of governance, the structure of the competition, the features of the labor market and the model of finance.

The Type of Governance In respect to governance issues, the main feature of European team sports leagues is that they were historically run by their respective national associations which were legally independent from the professional clubs playing in the competition. This kind of governance is of a contractual nature as it involves vertically separated entities (Dietl, Franck, Lang, & Rathke, 2011). The creation of institutions plainly in charge of the professional competitions made the governance of European leagues more cooperative, which represents a convergence with what is observed in their North American counterparts. In French football, the ‘Groupement des clubs autorises à utiliser des joueurs professionnels’ (Group of the clubs authorized to use professional players) was created in 1946, i.e. 14 years after the championship. This type of governance is still the same nowadays and has even been extended, for example, in 1987 in French basketball and 1998 in French rugby (Scelles, Ferrand, & Durand, 2015).

The Structure of the Competition The competition has its own features. Among these characteristics are the matches scheduling and the determination of the champion. For European leagues, the norm is seen as a balanced round-robin competition: ‘In a round robin, the league creates a schedule of games for a championship season for each team, and every team plays a predetermined number of games against other league members. The champion is determined by aggregating the results of all matches’ (Noll, 2003, p. 532). The balanced schedule comes from the fact that all teams play all others an equal number of matches. Noll (2003, p. 532) stresses that ‘Sports purists regard a balanced schedule as superior because it produces a final league standing that is most likely to reflect the actual rank-ordering of teams by quality.’ However, not all European leagues were balanced round-robin competitions.

European Sports Leagues: Origins and Features

In France, for example, the rugby, handball, volley and hockey leagues had both a group stage and playoffs. As most of the literature focuses on football leagues that were balanced round-robin competitions, it is assumed that all European leagues had such a format, but this was not the case. Another interesting aspect of European leagues lies in the plurality of major leagues. However, ‘the national leagues of Italy, Spain, Germany or England have been seen as only imperfect substitutes’ (Szymanski, 2003a, p. 1152). From 1955 with the ‘Union des Associations Européennes de Football’ (UEFA) launching continental competitions, European leagues have also begun to be characterized by the cohabitation of national and pan-European championships, with certain teams participating simultaneously in both.

The Labor Market We can identify two different phases of the labor market of athletes in Europe until the 1980s, according to the balance of the bargaining power between clubs and players. In the first phase until 1960, the bargaining power was clearly in favor of clubs, and that led to restrictions of freedom for player movement: ‘The basis of the employment of the professional footballer in the Football League is the retain and transfer system’ (Sloane, 1969, p. 183). The transfer of a player employed in a club participating to the league required the signature of the releasing club, provided he was initially/beforehand added to the club’s transfer list. As such, ‘the club with whom the player is currently registered can be said to possess a monopoly over the services of that player’ (Sloane, 1969, p. 184). This system was adopted in 1925 by the French Football Amateur Federation (Lanfranchi & Wahl, 1998). By 1960, the bargaining power slightly shifted in favor of players, with the empowerment of players’ unions. In 1961 in France, the ‘Union Nationale des Footballeurs Professionnels’ (National Union of Professional Players, UNFP) was created with the official aim to give players the appropriate means to fight for their rights. In 1963 in England, the Professional Footballers Associations (PFA) managed to reduce the power of the clubs to hold a player who became out of contract. Both negotiations ended with the emergence of fixedterm contracts for footballers.

The Model of Finance The model of finance of the European leagues until the 1980s has been qualified as

139

Spectators–Subsidies–Sponsors–Local or SSSL (Andreff & Staudohar, 2000). Money came mainly from spectators: ‘Throughout most of the 20th century, the primary source of revenue to European professional sports was gate receipts’ (Andreff & Staudohar, 2000, p. 259). Depending on national peculiarities, the rest of the financing could come from local authorities’ subsidies and industrial patrons. Most of the financing had a common feature: it was local. Gratton and Solberg (2007, p. 4) highlight the following aspects: ‘In contrast to the USA there was little or no competition in the […] television market in the early post-war years.’ This is confirmed by Andreff and Staudohar (2000, pp. 259 and 263), who write that ‘Given the lack of competition among broadcasters […] the monopsony rights fee would not be sufficient to compensate for lost gate receipts.’ However, the model of finance changed from the 1980s, as did more generally the European model of sports leagues.

LEAGUE MODEL FROM THE 1980S Overview European leagues have encountered a number of important changes since the 1980s: the growth of TV rights and rich owners from the 1980s, as mentioned previously (Andreff & Staudohar, 2000), the Bosman case (1995)1 and subsequent internationalization from the 1990s as well as the idea of Financial Fair Play from the 2000s, with the concept being approved by UEFA in 2009 (Morrow, 2014). Andreff and Staudohar (2000) note that during the 1980s and even more so in the 1990s, new sources of revenue emerged, and old ones declined. The new sources form the basis of a model – replacing the previous SSSL model – based on four pillars: ‘Media–Corporations– Merchandising–Markets’ (or MCMMG model) (see Andreff & Staudohar, 2000). When looking at the 10 richest European football clubs in 2015–2016, it must be noted that gate receipts (matchday) are still an important source of revenue (between 13% and 29%; Deloitte, 2017). Nevertheless, it is clear that broadcast and commercial revenues are much larger and more and more globalized. As underlined by Andreff and Staudohar (2000), the elimination of restrictions on the player labor market following the Bosman case has accelerated this change. For example, it is worth mentioning that while there were only 11 players not from the United Kingdom (UK) and Ireland named in the starting line-ups for the first rounds of matches in the Premier League in

140

THE SAGE HANDBOOK OF SPORTS ECONOMICS

1992–1993 (Atkinson, 2002), there were about two-thirds of foreign players in 2015–2016 (Poli, Ravenel, & Besson, 2016).

Financial Fair Play In September 2009, UEFA approved a Financial Fair Play (FFP) concept, seeking to ensure the future well-being and health of professional football (Morrow, 2014), which follows the club licensing system (established from the season 2004–2005) with its manual published as early as March 2002 (UEFA, 2002). In its foreword, Lennart Johansson, then President of UEFA, wrote that ‘The experience of France, European and World Champion, has shown that it is possible to introduce, with success and on the long term, club licenses subject to various requirements’ (UEFA, 2002, p. 1). Indeed, it should be reminded that the French Football Federation (FFF) was the first football governing body to put in place, in 1990, a regulation system which seeks to prevent insolvency through its National Direction for Management Control (DNCG) (Dermit-Richard, Scelles, & Morrow, 2019). Then, the Bundesliga established its own clubs licensing system in 2001 with the foundation of the Deutsche Fußball Liga (DFL). Since 2010, UEFA has established its own financial regulation system, introducing FFP regulations for clubs qualified to participate in its European-wide club competitions, i.e. the UEFA Europa League and the UEFA Champions League. A number of recent publications have focused on UEFA’s FFP regulations. Some are concerned with the objectives of the system (Drut & Raballand, 2012; Durand & Dermit-Richard, 2013) and its legitimacy (Budzinski, 2014; Müller, Lammert, & Hovemann, 2012), while others are concerned about whether FFP is in conflict with the rules of the European Union Treaty (Lindholm, 2010). There are also some studies on the anticipated effects of FFP. For example, Peeters and Szymanski (2014) use econometric modelling to establish the anticipated consequence of the implementation of the break-even requirement on club payrolls and other measures. They find that the break-even constraint could substantially reduce average payrolls and wage-to-turnover ratios, while strengthening the position of traditional top teams. Using a game theory approach, Preuss, Haugen, and Schubert (2014) considered FFP effects that could be contrary to expected objectives. Franck (2014) and Franck and Lang (2014) focused on the opportunity to introduce hard budget constraints to promote or incentivize more responsible management of football clubs and reducing dependency of shareholders.

BETWEEN SPORTING TRADITION AND MODERN ECONOMIC AND FINANCIAL STAKES In the first half of the 2000s, Szymanski (2003a, p.  1181) wrote that ‘the European leagues have maintained a high degree of public interest and structural stability over the last half century’. This assertion is still true at the time of writing this chapter (the second half of the 2010s). Indeed, European leagues remain generally faithful to their origins and their tradition, e.g. football leagues were still organized based on a regular season without playoffs. In French football, Scelles (2009, 2010) highlights that actors are not open to the introduction of playoffs, considering it to be fairer from a sporting perspective that the most regular team becomes champion. However, some European football leagues have introduced playoffs since the beginning of the 21st century. For example, the Dutch league has operated with a playoff system from 2005–2006 (even if its extent was reduced in 2008–2009, maybe a sign that Dutch actors were not fully convinced by this system), followed by the Belgian league from 2009–2010. An even more remarkable counter-example of structural stability is the French rugby league, which seems paradoxical for a sport that uses its so-called ‘traditional’ values as a communication and marketing tool. Until 1991–1992, 80 (!) clubs belonged to the first division (eight groups of 10 clubs each); in 1992–1993, this number was reduced to 32 (four groups of eight clubs each) and then 20 (two groups of 10 clubs each) in 1995–1996. In 1998–1999, this number was again increased to 24 (three groups of eight clubs each), a ‘step back’ that can be interpreted as a way for Serge Blanco to satisfy as many clubs as possible and be elected as President of the new league created in 1998 (LNR, Ligue Nationale de Rugby) rather than a conviction that more clubs is better (Scelles, Durand, Ferrand, & Mishra, 2014). Indeed, during his 10 years as the President of LNR, Serge Blanco reduced the number of clubs to 21 (two groups) in 2000–2001, then 16 in 2001–2002, before reducing the number of groups to a single one in 2004–2005, then the number of clubs to 14 in 2005–2006 (Terrien, Scelles, & Durand, 2015). This evolution in the league design is consistent with the growing importance of the economic and financial stakes in French club rugby.

AVENUES FOR FUTURE RESEARCH The changes encountered by European leagues since the 1980s underline the growing importance

European Sports Leagues: Origins and Features

of the economic and financial stakes. An expected consequence may be the search for the most appropriate league design from an economic perspective. Such a design seems to lie in a European Super League. As early as 1998, the Italian group Media Partners, presided by Rodolfo Hecht (a former collaborator for Fininvest, Silvio Berlusconi’s holding), attempted to create a European Super League in football (Moatti, 1998). As a consequence, UEFA changed the format of its Champions League by allowing more clubs from the richest countries and the strongest leagues to participate in order to convince them not to join the new leagues. Nevertheless, Hoehn and Szymanski (1999) argued in favor of the creation of a European Super League and against teams both in the Super League and in national leagues due to its negative consequence on competitive balance (CB). According to Szymanski (2003a, p. 1181): ‘If competitive balance really matters then we should expect the European system to collapse.’ As we get closer to the 2020s, a European Super League still does not exist in football. This could mean that CB does not really matter. A better concept may be competitive intensity (CI), which focuses on outcome uncertainty related to sporting prizes (Kringstad, 2005; Kringstad & Gerrard, 2005). Szymanski (2003a, p. 1181) opened the door to the consideration of prizes: ‘The role of prizes in providing incentives has been largely ignored in the team sports literature.’ However, he had rather economic prizes in mind (Szymanski, 2003b). During recent years, a considerable number of articles were published that tested the relevance of sporting prizes in various settings2 and came to the conclusion that a European Super League makes sense from an economic (TV audience) perspective (Scelles, 2017) since TV viewers want to watch balanced strong teams, but less from a more social (stadium attendance) perspective (Scelles et al., 2016), given that all sporting prizes (including relegation) attract stadium attendees. Further research is needed to enlighten this debate about the economic versus social perspective, which may determine the future features of European leagues and whether they will maintain their historical structure. Such research should focus not only on the main leagues that are usually studied, but also on the smaller leagues that are often organized as quadruple round-robin tournaments with teams playing each other four times (Pawlowski & Nalbantis, 2015). This should also include women’s leagues, which have been largely ignored in the literature (Valenti, Scelles, & Morrow, 2018). Last, this should not only deal with football, the archetypal European team sport (Szymanski, 2003a). This chapter provides some elements about the French

141

rugby league, which is experiencing a growth crisis at the time of writing (Renvoi aux 22, 2017). This may indicate that the evolution of its features has gone too far from their origins, generating an unbalance. Before a swing of the pendulum? The ability of league organizers and club managers to find a balance between economic development and respect of history seems a promising avenue for future research.

Notes 1  The Bosman case refers to the European Court of Justice having abolished not only the player reservation system and the payment of transfer fees for end-of-contract players but also the existing restrictions to the number of foreign European players that can be fielded (Késenne, 2000, 2007). 2  See Andreff (2009), Scelles (2009, 2010, 2017), Scelles, Desbordes and Durand (2011), Scelles, Durand, Bah and Rioult (2011), Scelles et al. (2013a, 2013b, 2016) as well as Andreff and Scelles (2015).

REFERENCES Andreff, W. (2009). Équilibre compétitif et contrainte budgétaire dans une ligue de sport professionnel. Revue Économique, 60(3), 591–633. Andreff, W. (2015). Governance of professional team sports clubs: Agency problem and soft budget constraint. In W. Andreff (Ed.), Disequilibrium sports economics: Competitive imbalance and budget constraints (pp. 175–227). Cheltenham, UK and Northampton, MA: Edward Elgar. Andreff, W., & Scelles, N. (2015). Walter C. Neale fifty years after: Beyond competitive balance, the league standing effect tested with French football data. Journal of Sports Economics, 16(8), 819–834. Andreff, W., & Staudohar, P. (2000). The evolving model of European sports finance. Journal of Sports Economics, 1(3), 257–276. Atkinson, R. (2002). England need to stem the foreign tide. The Guardian, 23 August. Retrieved from www.theguardian.com/football/2002/aug/23/ sport.comment. Budzinski, O. (2014). The competition economics of financial fair play. In O. Budzinski & A. Feddersen (Eds.), Contemporary research in sports economics (pp. 75–96). Frankfurt, Germany: Peter Lang Academic Research. Cain, L. P., & Haddock, D. D. (2005). Similar economic histories, different industrial structures: Transatlantic contrasts in the evolution of

142

THE SAGE HANDBOOK OF SPORTS ECONOMICS

professional sports leagues. Journal of Economic History, 65(4), 1116–1147. Cairns, J., Jennett, N., & Sloane, P. J. (1986). The economics of professional team sports: A survey of theory and evidence. Journal of Economic Studies, 13(1), 3–80. Chester, D. N. (1968). Report of the committee on football. London: Department of Education and Science. Davis, L. E. (1974). Self-regulation in baseball, 1909–71. In R. G. Noll (Ed.), Government and the sports business. Washington, DC: Brookings Institution. Deloitte (2017, January). Planet football: Football money league. London: Deloitte. Demmert, H. (1973). The economics of professional team sports. Lexington, MA: Lexington Books. Dermit-Richard, N., Scelles, N., & Morrow, S. (2019). French DNCG management control versus European UEFA financial fair play: A divergent conception of financial regulation objectives. Soccer & Society, 20(3), 408–430. Dietl, H., Franck, E., Lang, M., & Rathke, A. (2011). Organizational differences between US major leagues and European leagues: Implications for salary caps. Unpublished Working Paper Series, Paper No. 1105. Downward, P., Dawson, A., & Dejonghe, T. (2011). Sports economics: Theory, evidence and policy. Abingdon, UK: Routledge. Drut, B., & Raballand, G. (2012). Why does financial regulation matter for European professional football clubs? International Journal of Sport Management and Marketing, 11(1/2), 73–88. Durand, C., & Dermit-Richard, N. (2013). La régulation du sport professionnel en Europe: Le fair play financier de l’UEFA, annonciateur d’une révolution culturelle? International Review on Sport and Violence, 7, 74–89. Fort, R., & Quirk, J. (1995). Cross-subsidization, incentives, and outcomes in professional team sports leagues. Journal of Economic Literature, 33(3), 1265–1299. Franck, E. (2014). Financial fair play in European club football: What is it all about? International Journal of Sport Finance, 9(3), 193–217. Franck, E., & Lang, M. (2014). A theoretical analysis of the influence of money injections on risk taking in football clubs. Scottish Journal of Political Economy, 61(4), 430–454. Gratton, C., & Solberg, H. A. (2007). The economics of sports broadcasting. Abingdon, UK: Routledge. Harris, H. A. (1975). Sport in Britain: Its origins and development. London: Stanley Paul. Hoehn, T., & Szymanski, S. (1999). The Americanization of European football. Economic Policy, 14(28), 202–240. Inglis, S. (1988). League football and the men who made it. London: Willow Books.

Jennett, N., & Sloane, P. J. (1985). The future of league football: A critique of the report of the Chester committee of enquiry. Leisure Studies, 4(1), 39–56. Késenne, S. (2000). Revenue sharing and competitive balance in professional team sports. Journal of Sports Economics, 1(1), 56–65. Késenne, S. (2007). The peculiar international economics of professional football in Europe. Scottish Journal of Political Economy, 54(3), 388–399. Kringstad, M. (2005). Competitive intensity in European football. Communication abstract to the 13th European Association for Sport Management Conference (pp. 165–167), September 7–10, Newcastle, UK. Kringstad, M., & Gerrard, B. (2005). Theory and evidence on competitive intensity in European soccer. International Association of Sports Economists Conference Paper, 0508. Lanfranchi, P., & Wahl, A. (1998). La professionalisation du football en France (1920–1939). Modern & Contemporary France, 6(3), 313–325. Lavoie, M. (2004). Faut-il transposer à l’Europe les instruments de régulation du sport professionnel nord-américain? In J. J. Gouguet (Ed.), Le sport professionnel après l’arrêt Bosman: Une analyse économique internationale (pp. 61–84). Limoges, France: Presses Universitaires de Limoges. Lindholm, J. (2010). The problem with salary caps under European Union law: The case against financial fair play. Texas Review of Entertainment and Sports Law, 12(6), 189–213. Moatti, É. (1998). Superligue ou superfric. Stratégies, 30 October. Morrow, S. (2014). Financial fair play: Implications for football club financial reporting. Edinburgh: ICAS. Müller, J. C., Lammert, J., & Hovemann, G. (2012). The financial fair play regulations of UEFA: An adequate concept to ensure the long-term viability and sustainability of European club football? International Journal of Sport Finance, 7(2), 117–140. Neale, W. C. (1964). The peculiar economics of professional sports. The Quarterly Journal of Economics, 78(1), 1–14. Noll, R. G. (1982). Major league sports. In W. Adams (Ed.), The structure of American industry (6th ed.). New York: Macmillan. Noll, R. G. (2003). The organization of sports leagues. Oxford Review of Economic Policy, 19(4), 530–551. Pawlowski, T., & Nalbantis, G. (2015). Competition format, championship uncertainty and stadium attendance in European football: A small league perspective. Applied Economics, 47(38), 4128–4139. Peeters, T., & Szymanski, S. (2014). Financial fair play in European football. Economic Policy, 29(78), 343–390. Poli, R., Ravenel, L., & Besson, R. (2016, February). Foreign players in football teams. CIES Football Observatory Monthly Report, 12.

European Sports Leagues: Origins and Features

Preuss, H., Haugen, K. K., & Schubert, M. (2014). UEFA financial fair play: The curse of regulation. European Journal of Sport Studies, 2(1), 33–51. Quirk, J. P. (1987). Intermediate microeconomics (3rd ed.). Chicago, IL: Sciences Research Associates. Renvoi aux 22 (2017). Crise de croissance, crise de confiance. Renvoi aux 22, 28 March. Retrieved from http://renvoiaux22.fr/WordPress3/crise-decroissance-crise-de-confiance/ Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Scelles, N. (2009). L’incertitude du résultat, facteur clé de succès du spectacle sportif professionnel: L’intensité compétitive des ligues: Entre impacts mesurés et effets perçus. PhD dissertation, University of Caen Basse-Normandie, Caen, France. Scelles, N. (2010). La glorieuse incertitude du sport. L’intensité compétitive des ligues professionnelles: Entre impacts mesurés et effets perçus. Sarrebruck, Germany: Éditions Universitaires Européennes. Scelles, N. (2017). Star quality and competitive balance? Television audience demand for English Premier League football reconsidered. Applied Economics Letters, doi: 10.1080/13504851.2017.1282125. Scelles, N., Desbordes, M., & Durand, C. (2011). Marketing in sport leagues: Optimising the product design. Intra-championship competitive intensity in French football Ligue 1 and basketball Pro A. International Journal of Sport Management and Marketing, 9(1/2), 13–28. Scelles, N., Durand, C., Bah S. T., & Rioult, F. (2011). Intra-match competitive intensity in French football Ligue 1 and rugby Top 14. International Journal of Sport Management and Marketing, 9(3/4), 154–169. Scelles, N., Durand, C., Bonnal, L., Goyeau, D., & Andreff, W. (2013a). Competitive balance versus competitive intensity before a match: Is one of these two concepts more relevant in explaining attendance? The case of the French football Ligue 1 over the period 2008–2011. Applied Economics, 45(29), 4184–4192. Scelles, N., Durand, C., Bonnal, L., Goyeau, D., & Andreff, W. (2013b). My team is in contention? Nice, I go to the stadium! Competitive intensity in the French football Ligue 1. Economics Bulletin, 33(3), 2365–2378. Scelles, N., Durand, C., Bonnal, L., Goyeau, D., & Andreff, W. (2016). Do all sporting prizes have a significant positive impact on attendance in a European national football league? Competitive intensity in the French Ligue 1. Ekonomicheskaya Politika / Economic Policy, 11(3), 82–107. In Russian, English version at https://mpra.ub.­ uni-muenchen.de/73844/.

143

Scelles, N., Durand, C., Ferrand, A., & Mishra, G. P. (2014). Serge Blanco: Le French Flair du terrain au management. In E. Bayle (Ed.), Les grands dirigeants du sport: 23 portraits et stratégies de management (pp. 345–357). Louvain-La-Neuve, Belgium: De Boeck. Scelles, N., Ferrand, A., & Durand, C. (2015). Identification et maîtrise des facteurs clés de succès par les dirigeants: Le cas des ligues sportives professionnelles (basket, football et rugby masculins en France). La Revue des Sciences de Gestion, 272(2), 55–65. Scelles, N., Szymanski, S., & Dermit-Richard, N. (2018). Insolvency in French soccer: The case of payment failure. Journal of Sports Economics, 19(5), 603–624. Schofield, J. A. (1982). The development of first-class cricket in England: An economic analysis. Journal of Industrial Economics, 30(4), 337–360. Sloane, P. J. (1969). The labour market in professional football. British Journal of Industrial Relations, 7(2), 181–199. Sloane, P. J. (1971). The economics of professional football: The football club as a utility maximiser. Scottish Journal of Political Economy, 18(2), 121–146. Sloane, P. J. (1976). Restrictions of competition in professional team sports. Bulletin of Economic Research, 28(1), 3–22. Szymanski, S. (2003a). The economic design of sporting contests. Journal of Economic Literature, 41(4), 1137–1187. Szymanski, S. (2003b). Incentives and competitive balance in team sports. European Sport Management Quarterly, 3(1), 11–30. Terrien, M., Scelles, N., & Durand, C. (2015). L’ouverture de la Ligue Nationale de Rugby: «Le french flair» d’une instance récente. In P. Chaix (Ed.), Le nouveau visage du rugby professionnel français: Argent, succès et dérives (pp. 55–73). Paris: L’Harmattan. Terrien, M., Scelles, N., Morrow, S., Maltese, L., & Durand, C. (2017). The win/profit maximization debate: Strategic adaptations as the answer? Sport, Business and Management: An International Journal, 7(2), 121–140. UEFA (2002). Guide de procédure pour l’octroi de licence aux clubs à l’horizon de la saison 2004– 2005. Nyon, Switzerland: UEFA. Valenti, M., Scelles, N., & Morrow, S. (2018). Women’s football studies: An integrative literature review. Sport, Business and Management: An International Journal, 8(5), 511–528. Vamplew, W. (2006). The development of team sports before 1914. In W. Andreff & S. Szymanski (Eds.), Handbook of sports economics (pp. 435–439). Cheltenham, UK and Northampton, MA: Edward Elgar.

15 Competition Policy in Sports Markets Oliver Budzinski

INTRODUCTION: WHY, WHEN, AND HOW DO COMPETITION RULES APPLY TO SPORTS? The virtue of competitive markets is the decentralized coordination mechanism, driving supply and demand towards each other. As a consequence, competitive markets coordinate economic relations and while doing so promote efficiency of (i) allocation (static welfare), (ii) innovation (dynamic welfare) as well as (iii) reactive capacity (evolutionary welfare). Furthermore, competition is simultaneously a precondition and a consequence of the individual freedom to act in economic affairs (freedom of consumption choice, freedom of supply). However, competitive markets can only be workable and effective if anticompetitive arrangements and conduct by the market participants undermining competition forces is prevented.1 It is the task of competition policy to combat and deter anticompetitive arrangements and conduct from within the market. Virtually all competition policy regimes include three fundamental elements: (i) combating anticompetitive cartels and cartel-like arrangements, (ii) controlling monopolists and dominant companies, at least combating the abuse of market power, and (iii) limiting market concentration by

controlling mergers and acquisitions for anticompetitive effects. Some additionally seek to combat (however defined) unfair business practices from non-market powerful companies or engage in consumer protection regulation. Eventually, the international competition policy regime of the European Union (EU) additionally combats anticompetitive state aid for companies competing in the common market. Generally speaking, competition policy addresses all types of commercial activities and, thus, all industries and businesses within an economy. Sport is no exception as soon as it involves commercial activities. Their existence is obvious from an economic perspective, if we look into professional sports, which is often a billion-dollar business: for instance, revenues of the American National Football League (NFL) and the English Premier League were US$13 billion (2015) and €4.838 billion (2014–2015) respectively (Budzinski & Müller-Kock, 2018). The economic impact of sport on the EU economy sums up to almost 4% of GDP and more than 5% of the labor force.2 Beyond clear-cut cases, however, it is difficult to exactly delineate commercial sports business from non-commercial sporting activities like grassroots sports conducted by people in their leisure time. Between these two extremes, a large variety of semi-professional and semi-commercial sports

Competition Policy in Sports Markets

exists and it may require a case-by-case assessment whether they should be subject to competition policy or not. For instance, a recent controversy revolved around US college sports, which tends to view itself to be an amateur, non-commercial sport, but which creates considerable monetary turnover – and, thus, was found to be subject to competition rules and policy in several recent court decisions. If any given sport has a relevant commercial side, then every activity affecting its commercial performance may be subject to competition policy, irrespective whether this activity is directly commercial itself (like the sale of media rights) or ‘just’ influences the attractiveness of the sport for spectators (like a change in the sporting rules). However, what about the special characteristics of sports, in particular the need for cooperation among the competitors, which is a constituting character for sports economics as a discipline (Rottenberg, 1956; Neale, 1964)? In order to set up a championship or league as a marketable product, the competitors in the sport market must cooperate on the rules of the game, the schedule, enforcement, etc. In economic language this implies that a governance structure with market-internal institutions (rules of the game) that are enforced by a market-internal regulator is required. The market-internal institutions virtually always affect competition – regardless of whether they are of a sporting nature (like the dimensions of a football pitch or the number of participants and the conditions of their qualification) or commercial at heart (like the marketing of the championship brand or the sale of media rights). In contrast to most other industries, the market-internal regulators enjoy comprehensive power: they, inter alia, set and enforce the rules of the game, limit and control participation (market entry), market the common product (e.g. bundling and selling broadcasting rights), and organize the sharing, distributing and re-allocation of revenues. Thus, the market-internal regulator combines championship management with government-like regulation. While this would be viewed to be blatantly anticompetitive and against fundamental market economy principles in virtually every other industry, the existence of the market-internal regulators and institutions is essential for the business of professional sports since cooperation on rules among the participants is elementary for the common product to be produced. Thus, on the one hand, they are constitutive for establishing sports market competition and, at the same time, they restrict competition. From a competition policy perspective, this turns the antitrust evaluation of changes in the market-internal institutions or practices by market-internal market regulators into a challenging task. The role of the market-internal regulator is particularly delicate because it usually finds itself in a

145

monopoly(-like) position. The majority of professional sports organizes itself in a way that there is one single championship, tournament, or league at the top level.3 Such an organization promotes consumer (fan) welfare by offering a competition of the best teams, athletes, etc. and allows them to identify the very best, which is – next to the top sports quality (including athletic prowess) – an important motivation for sports consumption.4 At the same time, the monopoly position of the sports association creates market power that may be abused at the detriment of customers (high prices, reduced quantities) but also participants (player wages, market entry, cheerleader compensation, etc.). If the participants directly control the market-internal regulator, for instance when the teams of a league form and own the sports association, then anticompetitive conduct originates from cartelization (i.e. the competitors are colluding on parameters of competition). If the participants’ control of the market-internal regulator is not existent or so indirect that it is ineffective, then the sports ‘authority’ represents a monopolistic bottleneck with vertical market power (for more details, see Budzinski & Szymanski, 2015). In any case, again, the market power is inevitable and cannot be prevented but, at the same time, it may create antitrust problems. Notwithstanding, a general market failure of sports markets cannot be identified from an economic perspective and as such no sector regulation replacing ‘ordinary’ competition policy is required. Competition policy is inevitably an endeavor in law and economics. While its goals are economic at heart – protecting the competitive process – its codification and means are law. Logically, the actual law may fit to or deviate from the underlying economics. Therefore, competition policy in sports markets depends on the respective national (or supra-national in the EU case) law and its enforcement. In other words, anticompetitive practices that should fall under the rule of competition law from an economics perspective may be exempted in a given jurisdiction or vice versa. So, while from an economic perspective sports markets are subject to competition policy, they are still characterized by special features that complicate the competition policy assessment of cases in question and, furthermore, their actual treatment depends on the competition law regime of the competent jurisdiction. Since this chapter is economic in nature, it cannot give a comprehensive overview on sportsrelated competition law and case law (for overviews, see Pelnar, 2007; Papaloukas, 2008).5 Instead, it aims to reveal the economic line of reasoning underlying the selected sports antitrust cases and provide a critical discussion of them.

146

THE SAGE HANDBOOK OF SPORTS ECONOMICS

THE BASEBALL ANTITRUST EXEMPTION IN THE US Probably, the most famous sports-related antitrust exemption relates to baseball in the US. Baltimore Baseball Club of the Federal League brought antitrust action against the National League and the American League in 1922, claiming attempted monopolization of professional baseball in the US and restrictions of trade, for instance, by the implementation of the so-called reserve clause.6 However, the Supreme Court ruled that US antitrust laws are not applicable to professional baseball because it was not viewed to be commercial and because of competitive balance (CB) considerations, i.e. the anticompetitive structures were deemed to be necessary in order to make it a commercial success.7 Since then, professional baseball, most notably Major League Baseball (MLB), enjoys a comprehensive antitrust exemption in the US. From an economic perspective, this exemption can hardly be justified. Not only does it seem to be inappropriate nowadays to claim a non-commercial character of MLB in the face of annual revenues in excess of $9 billion.8 Even though these numbers had been much lower in the 1920s, the judgment itself reasoned that restrictive practices by the leagues were necessary to make the game more attractive and the success of the clubs more certain, thus, reverting to commercial reasoning. Furthermore, the strange situation results that one of the US major leagues is exempted from antitrust (baseball), whereas the others are not (American football, basketball) (Classen, 1988). Not surprisingly, many authors argue that several of the current arrangements in US baseball are anticompetitive in nature and increase ticket and broadcasting prices as well as deter market entries and further cement competitive advantages of the bigger incumbents, for instance, by blocking MLB teams with smaller markets to relocate to bigger and more profitable markets (inter alia, see Hamilton, 1998; Mehra & Zuercher, 2006; Mozes & Glicksman, 2011). These severe restrictive effects are hardly outweighed by claimed benefits like financial contributions to minor leagues (Grow, 2012), which supposedly should be possible with less restrictive practices as well. Competitive balance considerations also represent a difficult justification for the MLB antitrust exemption. First, their economic virtue is controversial at best (Pawlowski & Nalbantis, 2019). Second, there are several CB-enhancing regulations that are less restrictive in competition terms. Third, instead of granting an overall exemption, the nature of the CB problem would be more

adequately treated by examining a specific practice by the market-internal regulator and weighing its beneficial and restrictive effects.

THE JUDO CASE: ABUSE OF MARKET POWER? Since sports associations, by their nature as a market-internal regulator, enjoy a dominant position, their actions may be suspicious regarding an abuse of market power. While this is obvious if they engage in commercial activities such as selling media rights (see later section), it also extends to their core ‘business’: setting and enforcing the rules of the games. Even decisions that may appear to be purely sporting at first sight may actually include commercial calculus. For instance, the world football association FIFA changed the specifications of the football in time for the FIFA World Cup in South Africa (2010). The competition ball was made harder so that it would make the job of the goalkeepers more difficult, thus leading to more goals, thus making the sport more attractive – and more attractive implies economically more successful. Similarly, attempts to make sports more television-friendly in order to increase demand and revenues from media represents examples where ostensibly purely sporting rules are shaped according to commercial considerations. Many recent changes in the combating rules but also in the points system of Judo enacted by the International Judo Federation (IJF), for instance, were clearly motivated by the desire to enhance the television-attractiveness of Judo. And it is no stretch to assume that also the enforcement of rules may be subject to commercial considerations. Among the core sporting tasks of sports associations lies the definition of the number of participants in premier-level tournaments and the selection criteria. Logically, it is impossible to have open top-level championships in most sports, so some limitation – and subsequently the implementation of a selection system – is necessary. Notwithstanding, limiting the participants of premier-level championships represents (i) an artificial reduction of quantity and (ii) the deterrence of a number of wannabe-participants from the commercially most attractive tournaments. Both are typical abuse-strategies of market dominators and monopolies. Still, they are necessary in the sports market to some extent – but, of course, may also be abusively exaggerated (Grow, 2015). In February 1996, Christelle Deliège, a Belgian Judo fighter in the under 52 kg-class, was prevented from participating in the prestigious Paris

Competition Policy in Sports Markets

International Judo Tournament, a tournament where qualification points for the 1996 Atlanta Olympic games could be obtained. Obviously, for Deliège, such participation would have been extremely relevant – not only career-wise but also commercially.9 However, her home sports association, the Ligue Belge de Judo (LBJ), nominated two other athletes for the Paris tournament. According to rules implemented by the European Judo Union (EJU), each national federation was limited to a maximum of two judoka per weight class for such qualification tournaments as the one in Paris. Christelle Deliège contested (i) the EJU-quota as being unnecessarily restrictive and (ii) the selection decision of the LBJ as being based on non-objective, non-transparent criteria. The involved sports associations, in contrast, disputed the economic relevance of their selection decisions (because, according to their definition, Judo was an amateur sport) and insisted on their autonomy in such decisions. However, the European Court of Justice (ECJ)10 ruled that sporting rules, including participation and qualification rules, fall under the jurisdiction of business and competition law and ‘the mere fact that a sports association or federation unilaterally classifies its members as amateur athletes does not in itself mean that those members do not engage in economic activities within the meaning of Article 2 of the Treaty’ (European Court of Justice, 2000, para. 46). The ECJ gave the opinion that premier-level sport usually involves economic activities (sponsorships, advertising revenues, media revenues, ticket sales, grants for athletes, etc.). However, it also ruled that selection systems for premier-level competitions are not per se a violation of law as long as they are inherent to the organization of sport (European Court of Justice, 2000, para. 69). However, a specific shape of a selection system may still be unnecessarily restrictive. Furthermore, qualification tournaments for the Olympic Games are not events between national teams (European Court of Justice, 2000, para. 44), so that a selection based upon nationality quotas is not self-evident. From an economic perspective, the Deliège case (which enjoys an important role as a precedent in law; Bell & Turner-Kerr, 2002) is not only interesting because of its clarification that high-level sports usually involve economic activities. Instead, the questions brought up by the national court about the selection system applied in Judo are sports economic at heart: how to evaluate the trade-off between the softening and biasing effect on competition – here from artificial quantity reduction and access deterrence – and the specific requirements and characteristics of sports – here the necessity to limit participation in premier-level Judo

147

tournaments? From an economics perspective, this trade-off requires that access rules and selection systems must be designed so that they minimize the (inevitable) competition-lessening effects as much as possible. When sports associations define participation, selection and qualification systems and criteria, they (inevitably) enjoy monopoly power. Thus, they must be careful not to abuse this power. This can best be safeguarded by objective and transparent qualification, selection, as well as promotion and relegation (regarding leagues) criteria. For instance, participation in premier-level competitions could be open to anyone satisfying objective requirements in terms of sporting skills. Alternatively, if too many athletes/teams would meet this criterion, a selection based on objective and transparent performance criteria would be acceptable as well. However, scope for associations to select athletes/teams based upon non-objective or non-transparent, arbitrary or discretionary criteria represents an anticompetitive abuse of power from an economic perspective. Moreover, the total number of participants must not be unnecessarily restrictive, i.e. there must be reasons inherent to the nature of the respective sports for any given limitation of athletes or teams allowed to participate. Obviously, closed leagues represent an issue according to this line of reasoning as they deter outside teams from participating in a commercially lucrative league on non-performance-related grounds. For instance, even a better performing, better managed, more fan-attractive team cannot enter the closed league on merit of its performance.11 More difficult to evaluate is whether top-level sports leagues really allow as many teams to compete as would be inherent to the sports in question. Do US major leagues really exploit the maximum participant numbers (skeptical: Grow, 2015)? Are 18 teams (like in the German Bundesliga) for a European-style football league a necessary limitation or unnecessarily restrictive? What about 20 teams like in the Spanish La Liga or in the English Premier League? Similar reasoning applies to tournaments where individual athletes compete with each other. And what about national quotas for competitions that are not matches between national teams? For instance, two of the top three judoka in the IJF World Ranking List in the class of men under-100 kg before the 2016 Rio de Janeiro Olympic Games were German. However, due to the selection rules of the Olympic Committee (allowing only one judoka per class per country per sex to compete, given that she/he matches some minimum performance requirements), one of them was deterred from competing in Rio, whereas many lower-ranked judoka from other nationalities qualified since they were the best of their countries – despite lower-level performances.

148

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Do such selection systems and decisions display anticompetitive effects or are they necessary and inherent to the sport in question? Unfortunately, the sports economics literature remains predominantly silent on such questions so far. Following up on the Deliège and other court rulings,12 the European Commission (EC) developed a procedure that sporting rules and conduct by sports associations may not be restrictive in economic terms unless they (i) pursue a legitimate objective, (ii) its restrictive effects are inherent to the pursuit of that objective, and (iii) proportionate to it (European Commission, 2007a, 2007b; see also Budzinski, 2012). If these conditions are not met, the rule, arrangement or conduct in question is deemed to be anticompetitive and thus prohibited by competition policy. However, if an objective justification exists, i.e. if beneficial effects of the rule or conduct in question outweigh the anticompetitive effects, an exemption from prohibition is possible. Such an exemption usually requires an in-depth analysis and evidence that consumers (fans) enjoy a net benefit from it. Other jurisdictions typically rely upon case-bycase analyses without such detailed guidelines. From an economic perspective, the three conditions (legitimate objective, inherence of restrictive effects, proportionality) of the EC fit well to economic reasoning of transparency and of limiting restrictive effects to the necessary minimum (inherence, proportionality), even though the criterion of proportionality may be a bit vague and as minimal as possible would be a preferable concept. It can be questioned, however, whether an additional case-by-case exemption option is really necessary or merely creates scope for anticompetitive lobbyism.13

CENTRALIZED SALE OF BROADCASTING RIGHTS: A HARDCORE-CARTEL AGAINST MEDIA AND FANS? Sports associations more often than not do not restrict their activities to what is necessary to organize the sport in question. Instead, they often directly engage in commercial activities, with the sale of media rights (TV broadcasting, online broadcasting, etc.) in professional ball-game leagues representing a particularly exemplary case. Principally, two different models exist: either every club in the league sells the broadcasting rights individually or all the participants (often plus the competent sports association) sell the rights collectively. Currently, virtually all major professional leagues practice a collective sale of

media rights. Until the season 2016–2017, the premier-level Spanish European football league, La Liga, represented a prominent exemption where broadcasting rights were sold individually by the teams. From the viewpoint of the league and the participating teams, a collective sale is attractive because it creates a monopoly-like situation and consequently results in higher total media revenues (monopoly rent) compared to a situation where the teams compete with each other for broadcasting deals. However, the collective deal may not maximize revenues for each single team: maybe the most attractive teams could secure even more profitable deals on their own in contrast to less attractive teams. The distributional effect on the teams of the league also depends on how the common revenue is allocated among the teams (Budzinski, 2018; Budzinski & MüllerKock, 2018): • equal allocation, i.e. each team receives the same share of the collective revenue • performance-based allocation, i.e. teams receive different shares according to their performance: either better performance implies higher shares or, in a reverse-performance-based system, better performance implies lower shares and • brand value-based allocation, i.e. teams with larger fan-base or higher marketing potential (past success, tradition, etc.) receive higher shares The non-equal allocations can encompass different degrees of skewness and elements of different types of allocation can be combined. While a brand-value allocation, in tendency, mimics the distributional effects of individual sale systems, the other systems benefit some teams (for instance, less successful teams in systems of equal or reverse-performance-based allocation) at the expense of others (more successful teams). Despite the distributional effects and consequent continuous internal conflicts about the ‘right’ allocation scheme, the monopoly rent appears to be so attractive that virtually all major professional leagues converted to collective sale systems of media rights in the course of time. From a competition economics perspective, a collective sale of media rights represents a hardcore cartel where the competitors of a market collude to extract rents from the other market side by increasing prices and limiting output. Usually, the cartelists auction the bundled broadcasting rights (often in one or a few exclusive packages14), like a monopolist with media companies on the other market side competitively bidding for the rights (Cowie & Williams, 1997). Premier sports

Competition Policy in Sports Markets

represent important content for media companies, considerably influencing their competitiveness (Toft, 2006, p. 3), so their demand is usually relatively price-inelastic. Their desire to ‘win’ the auction for exclusivity deals with the most attractive sports leagues leads to a dynamic bidding competition, driving up prices. Often, the cartel also reduces the number or extent of available rights in order to further increase prices. For instance, limitations on geographical reach are usually included with several geographical areas not being served despite the existence of a positive willingnessto-pay from local media. Furthermore, rights for (innovative) online broadcasting and social media exploitation are sometimes severely limited or even completely excluded (e.g. in Formula One). Bundling broadcasting rights into one monopolistic package may lead to market foreclosure in media markets if only a few sports leagues represent premier content. While in the US several major leagues compete with each other for fanattractiveness, in Europe, for instance, Europeanstyle football dominates the market. Furthermore, the cartel arrangement can hamper the development of certain sub-markets (e.g. new media markets or other regional markets in order to protect pay-TV revenues). If media companies have to pay a supra-competitive price for sports broadcasting rights, consumers are indirectly harmed as well. Either media companies need to increase their prices in turn (directly in the case of paid content like pay-TV or tax-/fee-based content like public-service television, indirectly in the case of advertised-financed free content where users/ viewers will have to endure more advertising15) or they can spend less money on other sports and non-sports programs. In any case, considerable harm to consumer welfare must be expected. The welfare effects of collective (or centralized) sale systems of sports media rights have been extensively analyzed in the sports economic literature (Atkinson, Stanley, & Tschirhart, 1988; Késenne, 2000, 2001, 2009; Falconieri, Palomino, & Sákovics, 2004; Palasca, 2006; Gürtler, 2007; Noll, 2007; Peeters, 2011, 2012; Budzinski, 2018). While there is a notable tendency to derive or conclude negative total effects on welfare, some benefits for consumer (fan) welfare are discussed as well: • the CB defense • the creation of a single point of sale provides efficiencies by reducing transaction costs for clubs and media companies • the creation of a common brand provides efficiencies as it increases recognition and distribution of the product and enhances its attractiveness for the fans (consumers)

149

Despite widespread skepticism from the sports economic literature, competition authorities and courts have frequently followed these rationales to grant antitrust exemptions to collective sale systems in sports. In the US, the ‘competitive balance argument is the main pro-competitive justification that sports leagues offer to defend agreements otherwise prohibited by antitrust laws’ (Mehra & Zuercher, 2006, p. 1505). The CB defense has been brought forward in more than 40 cases in order to justify various restrictions of competition, such as cartel­ ization among incumbent league participants to prevent market entry of new investors, the granting of regional monopoly privileges (exclusive territories), restricting entry and transfer of players, defining the terms of player employment and (maximum) salaries, or limiting (live) broadcasting and other media coverage (with detailed case references, see Mehra & Zuercher, 2006). Even though courts do not always uphold the CB defense in any single case, they fundamentally ‘acknowledge that some otherwise anticompetitive restraints may be necessary to encourage competitive balance among the league’s teams’ (Grow, 2015, p. 590; see also Ross, 2003; Mehra & Zuercher, 2006). Interestingly, the important role of the CB defense in antitrust policy contrasts with an increasing skepticism in sports economic literature about its welfare-promoting effects. Even though CB was originally viewed to be an important driver of audience demand, the empirical literature struggles to find a significant positive effect of more CB on audience demand, both in terms of attendance and TV-viewership.16 Therefore, an unconditioned and widespread use of the CB defense to exempt commercial sports from competition law cannot be supported by the state of economics research. And even if CB was accepted as a defense, then it would only be eligible for arrangements that internally allocate the common revenues equally or according to a reverseperformance system (Budzinski, 2018). Performance-based or brand value systems are much more unlikely to create considerable procompetitive-balance effects. Surprisingly, however, competition authorities who accepted the CB defense have usually shied away from setting conditions on the internal allocation mechanism of the collectively earned revenues. Also, in the EU, involved parties have brought forward the CB defense in many competition policy cases on a community level. The EC and the European courts acknowledged CB justifications in various instances (Ross, 2003). The Commission’s 2007 White Paper and accompanying documents explicitly list the need to preserve

150

THE SAGE HANDBOOK OF SPORTS ECONOMICS

CB as a specificity of sports and, thus, a legitimate goal of (restrictive) interventions by sports associations (European Commission, 2007a, 2007b; Weatherill, 2012). However, the restrictive effects of the intervention must be inherent to pursuing the legitimate objective (preserving or enhancing CB) and proportionate (Budzinski, 2012). Thus, whether an anticompetitive arrangement or practice by a sports association can be justified by CB considerations is assessed case by case. For instance, it was accepted, inter alia, with respect to the implementation of transfer windows or the promotion of home-grown talent (Weatherill, 2012) but it was rejected in the majority of cases dealing with joint-selling arrangements of broadcasting rights (European Commission, 2003). The single-point-of-sale efficiency defense (Kienapfel & Stein, 2007) rests on transaction cost reasoning, i.e. having a single point of sale (the cartel) reduces the transaction costs for the buyers (media companies). However, without further qualifications this does not qualify from a competition economics point of view: having a monopoly supplier indeed always reduces transaction costs in the sense that costs of searching and selecting disappear. Still, it reduces welfare because the allocative and dynamic inefficiencies of monopoly (or cartelized supply in general) easily overcompensate the transaction cost decrease. So, there must be some specific aspects of selling a league’s or championship’s broadcasting rights (European Commission, 2003, rec. 139–153). The European Commission (2007b, p. 83) argues in its UEFA Champions League (European football) decision: [t]he single point of sale enabled the acquisition of coverage for the whole UEFA Champions League season, allowing programming to be planned in advance. […] [D]ue to the knock-out nature of the UEFA Champions League […] a broadcaster could not know in advance which clubs would make it through to the end.

This reasoning emphasizes the knock-out character (cup system) of the UEFA Champions League (European Commission, 2003, rec. 145). And, indeed, it is difficult to sell the coverage of a whole cup in advance with a decentralized system since nobody knows in advance who will survive the knock-out rounds. However, from an economics perspective, it is not clear at all why the complete coverage must be sold in advance of the season – and cannot be offered in sequences corresponding to the knock-out rounds (Budzinski, 2012). Actually, selling broadcasting rights sequentially round-by-round, when the respective team pairings are actually known, should lead to a more efficient price of the media rights.

Moreover, the specific reasoning only covers cup systems. However, the EC also granted conditional exemption from the cartel prohibition to centralized broadcasting selling systems of the English Premier League and the German Bundesliga (European Commission, 2005, 2006) – in both cases with reference to the reasoning of the earlier UEFA Champions League decision. Both leagues work with a playing schedule that is fully determined in advance of the season. So, here the single-point-of-sale defense – which is questionable from an economics perspective anyway – does not apply at all. Another line of reasoning refers to the efficiency effects of creating a common brand (increasing recognition and distribution of the product). The creation of a coherent league product may increase its attractiveness for the fans (consumers) as the product is focused on the competition as a whole rather than the individual clubs participating in the competition (Kienapfel & Stein, 2007). Notwithstanding, a joint-selling arrangement would need to be inherent to create a common brand (insofar as this represents a legitimate objective). It is difficult, however, to understand why centralized sales systems would be inherent to a common brand, i.e. a common appearance could not be safeguarded otherwise (common design, lower-level contractual obligations to follow certain standards for broadcasting, etc.) in an individualized system, except maybe of the broadcasting of comprehensive highlights programs of match-days (European Commission, 2003, rec. 146) – if this is viewed to be an essential service. Despite the acceptance of the single-pointof-sale defense as well as the common-brand defense, the European Commission (2007b) has established a list of remedies that must be fulfilled in order to exempt joint-selling arrangements from the cartel prohibition: • non-discriminatory and transparent competitive tendering • limitation of the duration of exclusive vertical contracts (max. three seasons), i.e. employing a ‘sun-setting mechanism’ • limitation of the scope of exclusive vertical contracts, i.e. unbundling media rights into several separate and meaningful packages in order to prevent market foreclosure, exclusion of conditional bidding • a fall-back option • use of obligation and parallel exploitation in order to remedy output restrictions, i.e. unused rights fall back to the individual clubs for parallel, competitive exploitation

Competition Policy in Sports Markets

• ‘no single buyer obligation’ in the case of already-existing dominance of one television operator and • trustee supervision of the tender procedure These conditions seek to remedy foreclosure effects on downstream media markets. As such they are certainly quite effective. However, whether they are sufficient to safeguard positive consumer welfare effects is doubtful. Furthermore, the dynamic development and digitization of media markets challenges these established conditions (as platform- and data-driven business models generally challenge traditional competition policies; see Budzinski & Stöhr, 2018; Budzinski & Kuchinke, 2019). While selling the media rights in several packages to different broadcasters (unbundling condition and no-single buyer rule) used to be comparatively consumer-friendly in traditional television markets (consumers just needed to switch to another channel for the next game and channels were usually accessible without charge), the growing importance of paid-for online broadcasting changes this assessment. Consumers now have to make contracts with several broadcasting platforms and deal with their differing technical requirements (e.g. access and software requirements, program scheduling) if they want to follow the league or cup as a whole. This burdens considerable costs on the consumers.

Notes 1  Additionally, there are some institutional preconditions for workable markets like property rights definition and enforcement as well as the absence of market external (for instance, governmental) torpedoing of competition. 2  See http://ec.europa.eu/competition/sectors/sports/ overview_en.html (accessed 29 May 2017). 3  A prominent exception would be commercial boxing. 4  While most sports economists would subscribe to the single entity concept, Ross (1989) emphasizes the virtues of a competition among championships in the same sports discipline. 5  Budzinski (2012) provides a comprehensive list of antitrust and competition policy cases in sports markets in the EU. 6  The reserve clause assigned far-reaching rights to the clubs regarding players, even after their contract had expired. It was part of the instruments allegedly used to protect the major leagues against upcoming minor leagues, their clubs or their players (Classen, 1988; Brand & Giorgione, 2003).

151

7  ‘If the reserve clause did not exist, the highly skillful players would be absorbed by the more wealthy clubs, and thus some clubs in the league would so far outstrip others in playing ability that the contests between the superior and inferior clubs would be uninteresting, and the public would refuse to patronize them. By means of the reserve clause and provisions in the rules and regulations, said one witness, the clubs in the National and American Leagues are more evenly balanced, the contests between them are made attractive to the patrons of the game, and the success of the clubs more certain’ (National League of Professional Baseball Clubs v. Federal Baseball Club of Baltimore, Inc., 269 F. 681, 687; D.C. Cir. 1920; cited after Mehra & Zuercher, 2006, p. 1506). 8  See, for example, www.forbes.com/sites/maurybrown/2016/12/05/mlb-sees-record-revenuesapproaching-10-billion-for-2016/#5d9ba5e67088 (accessed 14 October 2017). 9  Up to this point, Christelle Deliège had won the Belgian championships in her class several times as well as the European championship and the under-19 world championship once. Thus, she clearly does not represent some random judoka wanting to fight beyond her sporting scope. 10  The case was referred to the ECJ by a national court (the Tribunal de Première Instance de Namur, Belgium) for obtaining an opinion. 11  Very recently, two lower-tier US soccer clubs (Miami FC of the North American Soccer League) and Kingston Stockade FC of National Premier Soccer League) brought a case against the US Soccer Federation because the absence of a promotion and relegation system deters them from having the sporting and commercial option to enter higher league levels on sporting performance grounds. (At present, the only way to gain entry to Major League Soccer, North America’s top league, is by paying about $150 million and being selected by an expansion committee as a viable location for a new team.) Thus, they are excluded from competing in the commercially most relevant soccer markets in the US (www. reuters.com/article/us-soccer-usa-cas-idUSKBN1AJ2YL, accessed 17 August 2017). 12  In 2007, the ECJ confirmed and extended his line of reasoning in the Meca-Medina-case, involving prohibitions to participate because of doping suspicions. It clarified that every area of sports association activity may be subject to competition law according to EU law (Weatherill, 2006). Budzinski (2012) provides a list of sporting rules that have been evaluated by European competition authorities. 13  For instance, UEFA’s so-called Financial Fair Play regulations were exempted simply by a common note from the then UEFA president Platini

152

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and the then EC-commissioner Almunia, despite economic analysis showing that the conditions of inherence and proportionality were hardly met (Budzinski, 2014). 14  Note that the exclusivity character in itself represents an output limitation since media rights are not rival in nature (and for non-video contents, exclusivity is most typically not applied). 15  On the economics of advertised-financed contents (platform economics), see Anderson and Gabszewicz (2006) and Budzinski and Satzer (2011). 16  See Pawlowski and Nalbantis (2019) and the comprehensive literature cited in that chapter.

REFERENCES Anderson, S. P., & Gabszewicz, J. (2006). The media and advertising: A tale of two-sided markets. In V. A. Ginsburg & D. Throsby (Eds.), Handbook of the Economics of Art and Culture (1st ed., pp. 567–614). Amsterdam, Netherlands: Elsevier. Atkinson, S., Stanley, L., & Tschirhart, J. (1988). Revenue sharing as an incentive in an agency problem: An example from the national football league. Journal of Economics, 19(1), 27–43. Bell, A., & Turner-Kerr, P. (2002). The place of sport within the rules of community law: Clarification from the ECJ? The Deliège and Lehtonen cases. European Competition Law Review, 23(5), 256–260. Brand, S. M., & Giorgione, A. J. (2003). The effect of baseball’s antitrust exemption and contraction on its minor league baseball system: A case study of the Harrisburg senators. Villanova Sports & Entertainment Law Journal, 10(1), 49–68. Budzinski, O. (2012). The institutional framework for doing sports business: Principles of EU competition policy in sports markets. International Journal of Sport Management and Marketing, 11(1–2), 44–72. Budzinski, O. (2014). The competition economics of Financial Fair Play. In O. Budzinski & A. Feddersen (Eds.), Contemporary Research in Sports Economics. Proceedings of the 5th ESEA Conference (pp. 77–98). Frankfurt a.M., Germany: Lang. Budzinski, O. (2018). Financial Regulation as an Anticompetitive Institution. In M. Breuer & D. Forrest (Eds.), The Palgrave Handbook on the Economics of Manipulation in Professional Sports (pp. 159–179). Basingstoke: Palgrave. Budzinski, O., & Kuchinke, B. A. (2019). Industrial Organization of Media Markets and Competition Policy. In B. von Rimscha (Ed.), Handbook Economics and Management of Media and Communication (forthcoming). Berlin: DeGruyter.

Budzinski, O., & Müller-Kock, A. (2018). Is the revenue allocation scheme of Formula One motor racing a case for European competition policy? Contemporary Economic Policy, 36(1), 215–233. Online first, 21 July 2017. doi:10.1111/coep.12247. Budzinski, O., & Satzer, J. (2011). Sports business and multisided markets: Towards a new analytical framework? Sports, Business, Management: An International Journal, 1(2), 124–137. Budzinski, O., & Stöhr, A. (2018). Competition Policy Reform in Europe and Germany - Institutional Change in the Light of Digitization. European Competition Journal, 15(1), doi 10.1080/17441056. 2018.1555942. Budzinski, O., & Szymanski, S. (2015). Are restrictions of competition by sports associations horizontal or vertical in nature? Journal of Competition Law & Economics, 11(2), 409–429. Classen, H. W. (1988). Three strikes and you’re out: An investigation of professional baseball’s antitrust exemption. Akron Law Review, 21(4), 369–390. Cowie, C., & Williams, M. (1997). The economics of sports rights. Telecommunications Policy, 21(7), 619–634. European Court of Justice (2000). Judgment 11.4.00 in Joined Cases C-51/96 and C-191/97 ‘Deliège’. Kirchberg-Plateau, Luxembourg. European Commission (2003). COMP/C.2-37.398 – Joint Selling of the Commercial Rights of the UEFA Champions League, Commission Decision, 2003/778/EC. Brussels, Luxembourg: EC. European Commission (2005). Case COMP/C2/37.214 – Joint Selling of the Media Rights to the German Bundesliga, Commission Decision. Brussels, Luxembourg: EC. European Commission (2006). COMP/C-2/38.173 – Joint Selling of the Media Rights to the FA Premier League, Commission Decision, C(2006)/868. Brussels, Luxembourg: EC. European Commission (2007a). The White Paper on Sport, COM (2007) 391 final. Brussels, Luxembourg: EC. European Commission (2007b). The EU and sport: Background and context, Commission staff. Working Paper accompanying the White Paper on Sport, SEC(2007)935. Brussels, Luxembourg: EC. Falconieri, S., Palomino, F., & Sákovics, J. (2004). Collective versus individual sale of television rights in league sports. Journal of the European Economic Association, 2(5), 833–862. Grow, N. (2012). In defense of baseball’s antitrust exemption. American Business Law Journal, 49(2), 211–273. Grow, N. (2015). Regulating professional sports leagues. Washington and Lee Law Review, 72(2), 573–652.

Competition Policy in Sports Markets

Gürtler, O. (2007). A rationale for the coexistence of central and decentral marketing in team sports. German Economic Review, 8(1), 89–106. Hamilton, J. (1998). Congress in relief: The economic importance of revoking baseball’s antitrust exemption. Santa Clara Law Review, 38(4), 1223–1254. Késenne, S. (2000). Revenue sharing and competitive balance in professional team sports. Journal of Sports Economics, 1(1), 56–65. Késenne, S. (2001). The different impact of different sharing systems on the competitive balance in professional team sports. European Sports Management Quarterly, 1(3), 210–218. Késenne, S. (2009). The impact of pooling and sharing of broadcasting rights in professional team sports. International Journal of Sport Finance, 4(3), 211–218. Kienapfel, P., & Stein, A. (2007). The application of articles 81 and 82 EC in the sport sector. Competition Policy Newsletter, 1(3), 6–14. Mehra, S. K., & Zuercher, T. J. (2006). Striking out competitive balance in sports, antitrust, and intellectual property. Berkeley Technology Law Journal, 21(4), 1499–1545. Mozes, M. J., & Glicksman, B. (2011). Adjusting the stream-analyzing major league baseball’s antitrust exemption after American needle. Harvard Journal of Sports and Entertainment Law, 2(2), 265–296. Neale, W. C. (1964). The peculiar economics of professional sports: A contribution to the theory of the firm in sporting competition and in market competition. Quarterly Journal of Economics, 78(1), 1–14. Noll, R. G. (2007). Broadcasting and team sports. Scottish Journal of Political Economy, 54(3), 400–421. Palasca, S. (2006). Collective selling of broadcasting rights in team sports. In W. Andreff & S. Szymanski

153

(Eds.), Handbook on the Economics of Sport (1st ed., pp. 719–729). Cheltenham, UK: Edward Elgar. Papaloukas, M. (2008). Sport: Case law of the Court of Justice of the EC. Papaloukas Editions, doi:10.2139/ssrn.1311952. Pawlowski, T., & Nalbantis, G. (2019). Competitive balance: Measurement and relevance. In P. Downward, B. Frick, B. R. Humphreys, T. Pawlowski, J. E. Ruseski & B. P. Soebbing (Eds.), The SAGE Handbook of Sports Economics (pp. 154–162). UK: Sage Publishing. Peeters, T. (2011). Broadcasting rights and competitive balance in European soccer. International Journal of Sport Finance, 6(1), 23–39. Peeters, T. (2012). Media revenue sharing as a coordination device in sports leagues. International Journal of Industrial Organization, 30(2), 153–163. Pelnar, G. J. (2007). Antitrust analysis of sports leagues. doi:10.2139/ssrn.1021365. Ross, S. F. (1989). Monopoly sports leagues. Minnesota Law Review, 73(3), 643–761. Ross, S. F. (2003). Competition law as a restraint on monopolistic exploitation by sports leagues and clubs. Oxford Review of Economic Policy, 19(4), 569–584. Rottenberg, S. (1956). The baseball player’s labour market. Journal of Political Economy, 64(3), 242–258. Toft, T. (2006). Developments in European law. Congress Sports & Law, Berlin, Germany. Weatherill, S. (2006). Anti-doping revisited: The demise of the rule of ‘purely sporting interest’? European Competition Law Review, 27(12), 645–657. Weatherill, S. (2012). EU sports law: The effect of the Lisbon treaty. In A. Biondi, P. Eeckhout, & S. Ripley (Eds.), EU law after Lisbon (1st ed., pp. 403–420), Oxford: Oxford University Press.

16 Competitive Balance: Measurement and Relevance Tim Pawlowski and Georgios Nalbantis

INTRODUCTION Rottenberg (1956) and Neale (1964) were the first to elaborate in their seminal works on the supposed positive relation between the level of balance within a sports competition and its attractiveness for spectators. For six decades, their hypothesis is stimulating predominantly empirical economic research which can be divided into two lines, as proposed by Fort and Maxcy (2003), that is, the Analysis of Competitive Balance (ACB literature) and the test of the Uncertainty of Outcome Hypothesis (UOH literature). This chapter intends to sketch some major findings from both lines of research during recent years, to discuss our current understanding of this topic and to present some directions for further research.

CONCEPTS AND DEFINITIONS A certain level of competitive balance (CB) between athletes and/or teams translates into a certain degree of uncertainty about the outcome in a sports competition. Uncertainty in this regard might refer to outcomes of games, in-season

sub-competitions (such as the championship race, the fight for securing a place in the playoffs or the fight against relegation) or the degree to which a league is dominated (or not) by a few teams over time. Accordingly, Szymanski (2003) proposed a distinction between match/game uncertainty, seasonal uncertainty and championship uncertainty, respectively. In this chapter we stick to the more general wording, that is, the short-, mid- and longterm dimensions of CB, since a couple of recently developed concepts, which are important in this context, go somewhat beyond pure uncertainty.

Short-term CB The core concept in the context of the short-term dimension of CB is game uncertainty. Game uncertainty is often measured by using the firstand second-order terms of home win probabilities, which can be derived from betting odds (Peel & Thomas, 1988). Notably, these ‘objective’ home win probabilities are highly correlated with perceived home win probabilities by fans (Pawlowski, Nalbantis, & Coates, 2018). Another popular index to measure game uncertainty is the Theil (1967) measure, which makes use of information on home win, away win and draw probabilities. A number of studies also

Competitive Balance: Measurement and Relevance

focus on the absolute difference between home and away team win probabilities instead (e.g. Buraimo & Simmons, 2015). More recently, Ely, Frankel, and Kamenica (2015) introduced the concepts of ex-ante experienced suspense and ex-post experienced surprise. Both concepts are related to uncertainty although they go somewhat beyond by measuring two complementary dimensions of excitement. In this regard, higher suspense is attributed to greater variance in the next period’s beliefs. In contrast, higher surprise results from an outcome that contradicts anterior beliefs. In other words, a game involving teams with the same chance to win is expected to be more suspenseful than a game involving a clear favorite. A game involving a clear favorite might, however, offer a large surprise in case the underdog wins. Such kind of (surprising) upsets only occur in games with a (ex-ante expected) relatively certain outcome (Coates, Humphreys, & Zhou, 2014).

Mid-term CB In the context of the mid-term dimension of CB, the focus is put on seasonal uncertainty. Seasonal uncertainty is often measured by using league points and rankings in order to depict the closeness of sub-competitions, such as the championship race, the fight for securing a place in the playoffs/European club competitions or the fight to avoid relegation. According to Budzinski and Pawlowski (2017), seasonal uncertainty seems to be more relevant to fans than the short- or longterm dimension of CB, as it frames the fans’ perceptions about the (in)balance of the whole league. Following Pawlowski et al. (2018), it is important to note that single games might not only be characterized by game uncertainty. Rather, seasonal uncertainty might also unfold at the level of a single game, which is differently conceptualized in the literature. For instance, match relevance by Jennett (1984) approximates the ex-ante championship significance of each game to both contestants by using ex-post information about the number of games they still have to win in order to become the champion in the given season. Closely related to this is the championship uncertainty measure, introduced by Janssens and Késenne (1987) and slightly modified by Pawlowski and Anders (2012), to measure UEFA Champions League qualification uncertainty. Similar to Jennett’s measure, this measure makes use of expost information to proxy ex-ante probabilities.1 It is based on the points required to become the champion (or to qualify for the UEFA Champions League, accordingly), the points already collected

155

in the season so far, the maximum number of points that can be achieved during the season as well as the maximum number of points that can be collected during the remaining season. A third related measure is the playoff uncertainty (PU), which is differently defined in the literature. For instance, Fort and Lee (2006) and Lee and Fort (2008) implemented a PU measure based on the difference in win percentages between the teams qualifying for the playoffs and their nearest runner-ups, while Krautmann, Lee, and Quinn (2011) utilized a PU measure that takes into account the distribution of all playoff teams in contention. The decisiveness of a game by Geenens (2014) is measured on the tournament level, such as the FIFA World Cup, taking into account the playing strength of both contestants in a game and the temporal position of this game in the tournament. The measure is based on the variation in the entropy of the probability distribution of winning the tournament for each participating club. Quite recently, Humphreys and Zhou (2015) revived the league standing effect first elaborated by Neale (1964). Their measure is built upon three components: (i) the variation in total daily changes in rank order, (ii) the cumulative changes in rank order, as well as (iii) the standard deviation of winning percentages in a given league on each day of a given season. Finally, the competition intensity measure, as developed by Scelles, Durand, Bonnal, Goyeau, and Andreff (2013a, 2013b), is focused on the points needed to reach different sporting prizes for the club, which is the closest to a specific sporting prize. Interestingly, the ‘objective’ measure of seasonal uncertainty by Scelles et al. (2013a, 2013b) is also closely related to the fans’ subjective perceptions about the suspensefulness of games (Pawlowski et al., 2018).

Long-term CB In the context of the long-term dimension of CB, the focus is put on inter-seasonal uncertainty. This dimension entails a static component, which focuses on performance differences (such as team rankings, points scored or winning percentages) over time, and a dynamic component, which focuses on the domination (or not) of specific teams over time. The two most popular measures of inter-seasonal uncertainty are the ratio of standard deviations (RSD) (Noll, 1988; Scully, 1989) and the Herfindahl-Hirschman Index (HHI) (Hirschman, 1964). The RSD compares the ‘actual’ standard deviation of win percentages (or points) to the idealized standard deviation (ISD) and is frequently applied to North American sports leagues (e.g. Fort & Quirk, 1995). The HHI

156

THE SAGE HANDBOOK OF SPORTS ECONOMICS

index, adopted from the industrial organization literature, was first applied in a sports context as measure for uncertainty by Depken (1999). A frequently used version of the HHI is the normalized HHI (Owen, Ryan, & Weatherston, 2007), which captures the concentration of points and/or wins by clubs in a season (static) and the HHI of champion imbalance (Leeds & von Allmen, 2008), which measures how the championship titles are spread among the clubs of a league over time (dynamic). Quite recently, Pawlowski and Nalbantis (2015) introduced the so-called HHI of championship uncertainty (HIUCS), which focuses on seasonal uncertainty (mid-term CB) measured at the level of a single game by calculating the quadratic share of championshiprelevant games played by each club in a league in order to account for the degree of scale distribution and concentration of championship-relevant games within a season. Other popular (static) measures of long-term CB involve inequality measures such as Gini coefficients (Schmidt & Berri, 2001; Utt & Fort, 2002), the competitive balance ratio (CBR) by Humphreys (2002), which takes account of both team- and leaguespecific variations in winning percentages over time, as well as concentration ratios, such as the one proposed by Koning (2000), which is based on the points obtained by a pre-defined number of top teams divided by the maximum number of points that they could have gained.

ANALYSIS OF COMPETITIVE BALANCE (ACB LITERATURE) In general, the ACB literature ‘focuses on what has happened to competitive balance over time or as a result of changes in the business practices of pro sports’ (Fort & Maxcy, 2003, p. 155). The vast majority of studies that fall within this strand of literature has focused predominately on the longterm dimension of CB (e.g. Fort & Quirk, 1995), with only a few studies looking at the development of the short-term (e.g. Judde, Booth, & Brooks, 2013) or mid-term dimensions of CB (e.g. Pawlowski & Nalbantis, 2015). This strand of literature can be further subdivided into studies just looking at the evolution of CB over time, and studies testing for structural breaks in the development of CB over time as a result of certain institutional changes. The first set of studies makes use of simple graphical depictions (e.g. Koning, 2000), including basic trend analyzing techniques such as moving averages (e.g. Groot, 2008) or linear and polynomial trend lines (e.g. Pawlowski &

Nalbantis, 2015). This literature is vast and was already subject to comprehensive reviews (for a recent review, see Pawlowski, 2016). Therefore, we restrict our chapter to briefly introduce the second strand of research. To address structural changes in CB, previous studies either used structural change dummies as explanatory variables in their regression models (e.g. Fort & Quirk, 1995; Flores, Forest, & Tena, 2010) or employed time series analysis by using break point (BP) detection techniques (e.g. Lee & Fort, 2005, 2012; Fort & Lee, 2007). The main objective of this set of papers is to evaluate whether market regulations such as salary caps (e.g. Fort & Quirk, 1995), budget caps (e.g. Judde et al., 2013) or revenue-sharing devices (e.g. Peeters, 2011) – otherwise prohibited by antitrust laws (Mehra & Zuercher, 2006) – do indeed increase CB as claimed for. Other studies focus on the effects of labor market interventions, such as draft systems (Fort & Quirk, 1995) or foreign player restrictions (e.g. Flores et al., 2010) on CB. Finally, previous studies (predominantly those conducted on European soccer) examined whether particular competition formats, such as promotion and relegation (e.g. Buzzacchi, Szymanski, & Valletti, 2003), the point score systems (e.g. Haugen, 2008), league/tournament size (e.g. Groot, 2008; Geenens, 2014), (un)balanced league schedules (e.g. Lenten, 2008), or quadruple (instead of double) round-robin tournaments (e.g. Pawlowski & Nalbantis, 2015), might have an impact on CB. Table 16.1 provides a brief overview on selected studies from this strand of literature.

UNCERTAINTY OF OUTCOME HYPOTHESIS (UOH LITERATURE) In contrast to the ACB literature, the UOH literature is focused on empirically testing the link between the different dimensions of CB and the demand for sports. Relying predominately on econometric models, CB measures are treated as determinants (along with various control variables capturing standard economic characteristics, quality aspects, opportunity costs and others) in demand equations in order to investigate their impact on both in-stadium attendance and/or TV viewing figures.2 The vast majority of studies testing the UOH have focused on the relevance of short-term CB. In the following, we sketch some recent empirical findings regarding the (ir)relevance of the UOH, divided by the three dimensions of CB.

Competitive Balance: Measurement and Relevance

157

Table 16.1  Studies analyzing the impact of market regulations and/or competition design elements on CB

Market regulations Salary cap Free agency Rookie draft Bosman ruling Collective sales of TV rights RRA Break-even rule (FFP) Competition design elements Promotion and relegation 3–1–0 point system Reducing league size Unbalanced schedules Seeding system General rule changes UCL bonus payments Substitutions Red and yellow cards Back–pass rule Increasing UEFA EURO teams Quadruple round–robin

Sport (country)

Effect on CB (dimension)

Source

Basketball (US) Baseball (US) Football/Baseball (US) Soccer (EUR) Soccer (EUR) Racing F1 (INT) Soccer (EUR)

None (long) None (long) Positive/none (long) Positive (long) None (long) Positive (short & long) Positive (long)

Fort & Quirk (1995) Fort & Quirk (1995) Fort & Quirk (1995) Flores et al. (2010) Peeters (2011) Judde et al. (2013) Peeters & Szymanski (2014)

Soccer (EUR) Soccer (EUR) Soccer (SCT) Soccer (SCT) Tennis (INT) Racing F1 (INT) Soccer (EUR) Soccer (EUR) Soccer (EUR) Soccer (EUR) Soccer (EUR) Soccer (EUR)

Negative (long) Negative (long) Positive (long) Negative (long) Mixed (long) Positive (mid) Negative (long) Positive(long) Mixed (long) Negative (long) Negative (mid) None (mid)

Buzzacchi et al. (2003) Haugen (2008) Groot (2008) Lenten (2008) Del Corral (2009) Mastromarco & Runkel (2009) Pawlowski et al. (2010) Kent et al. (2013) Kent et al. (2013) Kent et al. (2013) Geenens (2014) Pawlowski & Nalbantis (2015)

Notes: EUR: Europe; F1: Formula One; FFP: Financial Fair Play; INT: International; RRA: Resource Restriction Agreement; SCT: Scotland; UCL: UEFA Champions League; UEFA: Union of European Football Associations; US: United States.

The Relevance of Short-term CB In contrast to the popular claim that competitions need to be tight in order to be attractive for fans, the majority of studies analyzing the impact of game uncertainty on stadium attendance report the opposite effect, that is, stadium attendance is maximized when either the home or the away team has a significantly higher chance of winning (for recent reviews, see Pawlowski, 2013; Coates et al., 2014; Schreyer, Schmidt, & Torgler, 2018). This gap between predictions of the standard UOH and the empirical evidence has motivated further research building upon different behavioral economic concepts.3 One approach is built upon the concept of reference-dependent preferences (RDPs) and loss aversion and was introduced by Coates et  al. (2014). Their model distinguishes between pure consumption utility and gain-loss utility generated by deviations between expected and actual game outcome and predicts that the UOH only emerges when the marginal utility of an unexpected win exceeds the marginal utility of an unexpected loss. Otherwise, the consumers’

preferences for game uncertainty are dominated by loss aversion. A second approach is built upon the idea that subjective evaluations of uncertainty and suspense by fans might deviate from ‘objective’ measures frequently used in the literature. While some studies suggest that such differences might indeed exist with regard to the mid-term dimension of CB (e.g. Pawlowski, 2013; Pawlowski & Budzinski, 2013; Nalbantis, Pawlowski, & Coates, 2017), Pawlowski et  al. (2018) did not find any evidence for such differences with regard to the short-term dimension of CB. In addition to these theoretical explanations for the gap between what the theory posits and empirical studies find, Humphreys and Zhou (2015) noted an (under-) identification problem in standard demand models: by simply using home win probability and its second-order term, previous studies fail to consider all relevant preferences of fans, that is, preferences for home wins, for game uncertainty and for loss aversion. Compared to the overwhelming evidence that more game uncertainty does not translate into a significant overall increase in ticket sales,

158

THE SAGE HANDBOOK OF SPORTS ECONOMICS

however, studies analyzing the impact of game uncertainty on TV viewing figures are ambiguous (for a recent review, see Nalbantis & Pawlowski, 2016). In this regard, two interesting observations occur. First, there seems to be a divergence in findings between different sports: for North American sports (e.g. National Football League, Major League Baseball), the majority of findings suggests a positive relation between game uncertainty and TV demand (e.g. Paul & Weinbach, 2007), while the majority of European studies (predominantly focusing on soccer) fails to provide evidence on the relevance of game uncertainty (e.g. Buraimo & Simmons, 2015). Second, several papers point towards the existence of moderating aspects. For instance, the impact of game uncertainty was found to depend on the half of the season (Forrest, Simmons, & Buraimo, 2005), the broadcasting platform (García & Rodríguez, 2006) and the market perspective, that is, local versus non-local consumption (Tainsky, Xu, & Zhou, 2014). In contrast to the widely spread belief, however, Pawlowski et al. (2018) could not detect any moderating effect of fanship status in this context. Finally, Bizzozero, Flepp, and Franck (2016) tested Ely’s et al. (2015) concepts of ex-ante experienced suspense and ex-post experienced surprise empirically, by using in-play betting odds and minute-by-minute TV viewing figures of tennis matches. Their results show that both suspense and surprise are associated with increased demand, with surprise, however, having a larger impact on viewing behavior than suspense. Interestingly, the authors found no evidence for a patriotic bias or gender-related differences in this regard, while (as anticipated) the effect of both measures is higher in the later stages of a match.

The Relevance of Mid-term CB Although the concept of mid-term CB was already introduced in 1984 by Jennett, up till now there are surprisingly few papers that have examined empirically its relevance for stadium attendance (e.g. Pawlowski & Anders, 2012) and TV viewing (e.g. Scelles, 2017). However, the few existing studies are largely supportive for the relevance of championship uncertainty across leagues and countries (e.g. Austria and Switzerland: Pawlowski & Nalbantis, 2015; Belgium: Janssens & Késenne, 1987; England: Scelles, 2017; France: Scelles et  al., 2013a; Germany: Pawlowski & Anders, 2012; Scotland: Jennett, 1984). For the other subcompetitions, however, the empirical evidence is mixed. For instance, games involving teams with chances to qualify for European club competitions

(Pawlowski & Anders, 2012; Scelles, 2017) or which are fighting against relegation (Jennett, 1984; Scelles, 2017) are not found to be associated with an increased demand. In the context of North American sports, previous studies fail to provide clear evidence on the relevance of PU, pointing towards the sensitivity of results with regard to the choice of the PU measure and the level of analysis. For instance, Lee and Fort (2008), focusing just on the differences between first- and second-ranked teams and utilizing annual league-level attendance, found that PU is associated with increased audiences in Major League Baseball (MLB). Similarly, positive effects are reported by Krautmann et  al. (2011), who used monthly league-level attendance data from the MLB and a measure based on information about all teams. However, using the same measure as Krautmann et  al. (2011) but examining annual league-level attendance data from the Big Four Major Leagues (National Basketball Association [NBA]; National Football League [NFL]; National Hockey League [NHL]; and MLB), Mills and Fort (2014) could attest a positive relation between PU and attendance only for NFL games.

The Relevance of Long-term CB Empirical evidence on the links between longterm CB and sports demand is sparse and again ambiguous, pointing towards the sensitivity of results with regard to the choice of the measure, as well as a divergence concerning different sports. Based on aggregate league attendance figures of the MLB, Schmidt and Berri (2001) found that attendance decreases by temporary improvements in CB. However, overtime, fans react negatively to persistent competitive imbalance (as expected). This is in line with Humphreys (2002), who found that variations in the CBR are positively related to variations in MLB attendance, indicating that total league attendance increases (decreases) with increasing (decreasing) CB. In contrast, however, Krautmann and Hadley (2006), who focused explicitly on the impact of dynasties in MLB, found that intra-seasonal balance (based on the relative standard deviation of winning percentages) does not affect attendance, while interseasonal balance, that is, the probability of previous playoff participants making it again into the playoffs, only marginally affects attendance in the American League (AL). Concerning European soccer, despite the longterm dominance of single teams within the leagues (a prominent example includes FC Bayern Munich in the German Bundesliga), both attendance and

Competitive Balance: Measurement and Relevance

TV viewing figures seem to be steadily increasing throughout the years (see Flores et al., 2010; Pawlowski et al., 2010). In line with this, Brandes and Franck (2007) could attest a positive relation between long-term CB and fan attendance only in one (i.e. the French Ligue 1) out of six European soccer leagues.4 By applying vector autoregressive (VAR) models to avoid a priori classification into endogenous and exogenous variables, they could even attest an inverse relation in some leagues (i.e. German Bundesliga and the English Championship), indicating a significant influence from stadium attendance on the level of CB (and not vice versa, as suggested by the UOH).

CONCLUSION In this chapter, we have tried to summarize some major findings from both the Analysis of Competitive Balance (ACB) and the test of the Uncertainty of Outcome Hypothesis (UOH) literature and to sketch some recent developments of theoretical and empirical research, which builds upon different behavioral economic concepts, in order to close the gap between predictions of the standard UOH and the empirical evidence. In the following, we suggest six avenues and topics that might merit further investigations in the future in order to advance our current understanding of the complex relations between CB, suspense and the demand for sport. First, while the framework of RDPs combined with loss aversion offers a plausible theoretical basis to explain the u-shaped relation between home win probabilities and sports demand (thus contradicting the UOH), current empirical tests are unable to distinguish between preferences for game uncertainty and loss aversion. Rather than just inferring their relative size (see Humphreys & Zhou, 2015), future studies should incorporate measures of all relevant preference parameters, that is, preferences for home wins, for game uncertainty and for loss aversion together. Second, although initial findings suggest that having a sufficient number of teams fighting for the championship until the end of a season might be more important than having an overall balanced league (Budzinski & Pawlowski, 2017), it remains unclear what a sufficient number of teams might be. Providing insights on this issue is important for both league organizers and competition designers. Third, studies using fan surveys report the presence of threshold effects with regard to attendance demand (e.g. Pawlowski & Budzinski, 2013; Nalbantis et al., 2017), suggesting that unexpected drops of CB levels below certain thresholds might

159

be able to trigger strong (demand) reactions. Given that efforts to increase uncertainty comes at a cost, knowing concrete figures of such threshold values is important (Budzinski & Pawlowski, 2017). Fourth, since several soccer leagues are currently ‘suffering’ from the long-term dominance of single teams, such leagues offer a natural (and practically highly relevant) setting to further investigate how consumers react to (sporting) dynasties and to what extent such dynasties are favored (or not) by current market distortions. Besides that, and given the sensitivity of results with regard to the choice of the measure, it is advisable for scholars to use different CB measures in their analysis and to provide some assessment/rationale of possible divergences in results in this regard. Fifth, it seems fruitful to build upon the concepts of suspense and surprise (Ely et  al., 2015) and examine audience reactions in team sports. Combining in-play viewership (or emotional response) data with individual-level information (e.g. on local versus non-local viewers, gender, age or income levels) might constitute a major contribution towards a better understanding of the relations of interest. Sixth, despite the increasing efforts in internationalizing sports leagues and competitions, close to nothing is known about the relevance of the UOH in cross-country settings. Using such crosscountry settings appears to be highly relevant from both a managerial and a theoretical perspective, since risk and uncertainty attitudes might generally vary between countries and cultures (Vieider et al., 2015).5

Notes 1  This divergence is also the core criticism of both measures when used for studying demand responsiveness to variations in mid-term uncertainty. 2  For a detailed discussion on the demand determinants for in-stadium attendance and TV viewing figures, see Borland and MacDonald (2003) and García and Rodríguez (2009) or Nalbantis and Pawlowski (2016), respectively. 3  For a recent overview on the behavioral economics literature on CB, see Budzinski and Pawlowski (2017). 4  Next to the French League 1, the German Bundesliga, the Italian Series A and B, the English Premier League and the English Championship had been investigated in their study. 5  In this regard, Nalbantis and Pawlowski (2018) have recently focused on a between-country setting, examining the demand for European soccer telecasts in the US. Using a similar survey design

160

THE SAGE HANDBOOK OF SPORTS ECONOMICS

as Pawlowski et al. (2018), their findings suggest that when examining the same sport (i.e. soccer), European and North American consumers’ game uncertainty preferences seem to be quite similar.

REFERENCES Bizzozero, P., Flepp, R., & Franck, E. (2016). The importance of suspense and surprise in entertainment demand: Evidence from Wimbledon. Journal of Economic Behavior & Organization, 130, 47–63. Borland, J., & MacDonald, R. (2003). Demand for sport. Oxford Review of Economic Policy, 19(4), 478–502. Brandes, L., & Franck, E. (2007). Who made who? An empirical analysis of competitive balance in European soccer leagues. Eastern Economic Journal, 33(3), 379–403. Budzinski, O., & Pawlowski, T. (2017). The behavioural economics of competitive balance: Theories, findings and implications. International Journal of Sport Finance, 12(2), 109–122. Buraimo, B., & Simmons, R. (2015). Uncertainty of outcome or star quality? Television audience demand for English Premier League football. International Journal of the Economics of Business, 22(3), 449–469. Buzzacchi, L., Szymanski, S., & Valletti, T. M. (2003). Equality of opportunity and equality of outcome: Open leagues, closed leagues and competitive balance. Journal of Industry, Competition and Trade, 3(3), 167–186. Coates, D., Humphreys, B., & Zhou, L. (2014). Reference-dependent preferences, loss aversion, and live game attendance. Economic Inquiry, 52(3), 959–973. Del Corral, J. (2009). Competitive balance and match uncertainty in grand-slam tennis: Effects of seeding system, gender, and court surface. Journal of Sports Economics, 10(6), 563–581. Depken, C. A. (1999). Free-agency and the competitiveness of Major League Baseball. Review of Industrial Organization, 14(3), 205–217. Ely, J., Frankel, A., & Kamenica, E. (2015). Suspense and surprise. Journal of Political Economy, 123(1), 215–260. Flores, R., Forrest, D., & Tena, J. D. D. (2010). Impact on competitive balance from allowing foreign players in a sports league: Evidence from European soccer. Kyklos, 63(4), 546–557. Forrest, D., Simmons, R., & Buraimo, B. (2005). Outcome uncertainty and the couch potato audience. Scottish Journal of Political Economy, 52(4), 641–661.

Fort, R., & Lee, Y. H. (2006). Stationarity and major league baseball attendance analysis. Journal of Sports Economics, 7(4), 408–415. Fort, R., & Lee, Y. H. (2007). Structural change, competitive balance, and the rest of the major leagues. Economic Inquiry, 45(3), 519–532. Fort, R., & Maxcy, J. (2003). Comment: CB in sport leagues: An introduction. Journal of Sports Economics, 4(2), 154–160. Fort, R., & Quirk, J. (1995). Cross-subsidization, incentives, and outcomes in professional team sports leagues. Journal of Economic Literature, 33(3), 1265–1299. García, J., & Rodríguez, P. (2006). The determinants of TV audience for Spanish football: A first approach. In P. Rodríguez, S. Késenne, & J. García (Eds.), Sport economics after fifty years: Essays in honour of Simon Rottenberg (pp. 147–167). Oviedo, Spain: Ediciones Universidad de Oviedo. García, J., & Rodríguez, P. (2009). Sports attendance: A survey of the literature 1973–2007. Rivista di Diritto ed Economia dello Sport, 5(2), 111–151. Geenens, G. (2014). On the decisiveness of a game in a tournament. European Journal of Operational Research, 232(1), 156–168. Groot, L. (2008). Economics, uncertainty and European football: Trends in competitive balance. Cheltenham, UK: Edward Elgar. Haugen, K. K. (2008). Point score systems and competitive imbalance in professional soccer. Journal of Sports Economics, 9(2), 191–210. Hirschman, A. O. (1964). The paternity of an index. The American Economic Review, 54(5), 761–762. Humphreys, B. R. (2002). Alternative measures of competitive balance in sports leagues. Journal of Sports Economics, 3(2), 133–148. Humphreys, B. R., & Zhou, L. (2015). The Louis– Schmeling paradox and the league standing effect reconsidered. Journal of Sports Economics, 16(8), 835–852. Janssens, P., & Késenne, S. (1987). Belgian football attendances. Tijdschriftvoor Economie en Management, 32(3), 305–315. Jennett, N. (1984). Attendances, uncertainty of outcome and policy in Scottish league football. Scottish Journal of Political Economy, 31(2), 176–198. Judde, C., Booth, R., & Brooks, R. (2013). Second place is first of the losers: An analysis of competitive balance in Formula One. Journal of Sports Economics, 14(4), 411–439. Kent, R. A., Caudill, S. B., & Mixon Jr, F. G. (2013). Rules changes and competitive balance in European professional soccer: Evidence from an event study approach. Applied Economics Letters, 20(11), 1109–1112. Koning, R. H. (2000). Balance in competition in Dutch soccer. Journal of the Royal Statistical Society, 49(3), 419–431.

Competitive Balance: Measurement and Relevance

Krautmann, A. C., & Hadley, L. (2006). Dynasties versus pennant races: Competitive balance in Major League Baseball. Managerial and Decision Economics, 27(4), 287–292. Krautmann, A. C., Lee, Y. H., & Quinn, K. (2011). Playoff uncertainty and pennant races. Journal of Sports Economics, 12(5), 495–514. Lee, Y. H., & Fort, R. (2005). Structural change in MLB competitive balance: The depression, team location, and integration. Economic Inquiry, 43(1), 158–169. Lee, Y. H., & Fort, R. (2008). Attendance and the uncertainty-of-outcome hypothesis in baseball. Review of Industrial Organization, 33(4), 281–295. Lee, Y. H., & Fort, R. (2012). Competitive balance: Time series lessons from the English Premier League. Scottish Journal of Political Economy, 59(3), 266–282. Leeds, M., & Von Allmen, P. (2008). The economics of sports (3rd ed.). New York: Routledge. Lenten, L. J. (2008). Unbalanced schedules and the estimation of competitive balance in the Scottish Premier League. Scottish Journal of Political Economy, 55(4), 488–508. Mastromarco, C., & Runkel, M. (2009). Rule changes and competitive balance in Formula One motor racing. Applied Economics, 41(23), 3003–3014. Mehra, S. K., & Zuercher, T. J. (2006). Striking out ‘competitive balance’ in sports, antitrust, and intellectual property. Berkeley Technology Law Journal, 21(4), 1499–1545. Mills, B., & Fort, R. (2014). League-level attendance and outcome uncertainty in US pro sports leagues. Economic Inquiry, 52(1), 205–218. Nalbantis, G., & Pawlowski, T. (2016). The demand for international football telecasts in the United States, edited by W. Andreff and A. Zimbalist. Basingstoke, UK: Palgrave. Nalbantis, G., & Pawlowski, T. (2018). US demand for European soccer telecasts: A between-country test of the uncertainty of outcome hypothesis. Journal of Sports Economics, 1527002518817598. Nalbantis, G., Pawlowski, T., & Coates, D. (2017). The fans’ perception of competitive balance and its impact on willingness-to-pay for a single game. Journal of Sports Economics, 18(5), 479–505. Neale, W. C. (1964). The peculiar economics of professional sports: A contribution to the theory of the firm in sporting competition and in market competition. Quarterly Journal of Economics, 78(1), 1–14. Noll, R. G. (1988). Professional basketball. Studies in Industrial Economics Paper No. 144. Stanford, CA: Stanford University Press. Owen, P. D., Ryan, M., & Weatherston, C. R. (2007). Measuring competitive balance in professional team sports using the Herfindahl–Hirschman index. Review of Industrial Organization, 31(4), 289–302.

161

Paul, R. J., & Weinbach, A. P. (2007). The uncertainty of outcome and scoring effects on Nielsen ratings for Monday Night Football. Journal of Economics and Business, 59(3), 199–211. Pawlowski, T. (2013). Testing the uncertainty of outcome hypothesis in European professional football: A stated preference approach. Journal of Sports Economics, 14(4), 341–367. Pawlowski, T. (2016). Competitive balance. In C. Deutscher, G. Hovemann, T. Pawlowski, & L. Thieme (Eds.), Handbuch Sportökonomik [Handbook of sports economics] (pp. 217–234). Schorndorf, Germany: Hofmann. Pawlowski, T., & Anders, C. (2012). Stadium attendance in German professional football: The (un)importance of uncertainty of outcome reconsidered. Applied Economics Letters, 19(16), 1553–1556. Pawlowski, T., Breuer, C., & Hovemann, A. (2010). Top clubs’ performance and the competitive situation in European domestic football competitions. Journal of Sports Economics, 11(2), 186–202. Pawlowski, T., & Budzinski, O. (2013). The (monetary) value of competitive balance for sport consumers: A stated preference approach to European professional football. International Journal of Sport Finance, 8(2), 112–123. Pawlowski, T., & Nalbantis, G. (2015). Competition format, championship uncertainty and stadium attendance in European football: A small league perspective. Applied Economics, 47(38), 4128–4139. Pawlowski, T., Nalbantis, G., & Coates, D. (2018). Perceived game uncertainty, suspense and the demand for sport, Economic Inquiry, 56(1), 173–192. Peel, D. A., & Thomas, D. A. (1988). Outcome uncertainty and the demand for football. Scottish Journal of Political Economy, 35(2), 242–249. Peeters, T. (2011). Broadcast rights and competitive balance in European soccer. International Journal of Sport Finance, 6(1), 23–39. Peeters, T., & Szymanski, S. (2014). Financial Fair Play in European football. Economic Policy, 29(78), 343–390. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Scelles, N. (2017). Star quality and competitive balance? Television audience demand for English Premier League football reconsidered. Applied Economics Letters, 24(19), 1399–1402. Scelles, N., Durand, C., Bonnal, L., Goyeau, D., & Andreff, W. (2013a). Competitive balance versus competitive intensity before a match: Is one of these two concepts more relevant in explaining attendance? The case of the French football Ligue 1 over the period 2008–2011. Applied Economics, 45(29), 4184–4192. Scelles, N., Durand, C., Bonnal, L., Goyeau, D., & Andreff, W. (2013b). My club is in contention?

162

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Nice, I go to the stadium! Competitive intensity in the French football Ligue 1. Economics Bulletin, 33(3), 2365–2378. Schmidt, M. B., & Berri, D. J. (2001). Competitive balance and attendance: The case of major league baseball. Journal of Sports Economics, 2(2), 145–167. Schreyer, D., Schmidt, S. L., & Torgler, B. (2018). Game outcome uncertainty and television audience demand: New evidence from German football. German Economic Review, 19(2), 140–161. Scully, G. W. (1989). The business of Major League Baseball. Chicago, IL: University of Chicago Press. Szymanski, S. (2003). The economic design of sporting contests. Journal of Economic Literature, 41(4), 1137–1187.

Tainsky, S., Xu, J., & Zhou, Y. (2014). Qualifying the game uncertainty effect: A game-level analysis of NFL postseason broadcast ratings. Journal of Sports Economics, 15(3), 219–236. Theil, H. (1967). Economics and information theory. Amsterdam, Netherlands: North-Holland. Utt, J., & Fort, R. (2002). Pitfalls to measuring competitive balance with Gini coefficients. Journal of Sports Economics, 3(4), 367–373. Vieider, F. M., Lefebvre, M., Bouchouicha, R., Chmura, T., Hakimov, R., Krawczyk, M., & Martinsson, P. (2015). Common components of risk and uncertainty attitudes across contexts and domains: Evidence from 30 countries. Journal of the European Economic Association, 13(3), 421–452.

17 Economics of Attendance Placido Rodriguez

THE BEGINNINGS When Rodney Fort (2005) nominated Simon Rottenberg as the father of sports economics, he simply confirmed a fact that all of us, who are devoted to this field of study, would agree upon. Even though the article of Rottenberg (1956) deals with the baseball players’ labor market and introduces two of the main ideas in the field of sports economics, namely the invariance proposition and the uncertainty of outcome hypothesis (UOH), it also offers initial insights on the economics of attendance, which is the objective of this chapter. On page 246 of his seminal work he states: Attendance at baseball games, as a whole, is a function of the general level of income, the price of admission to baseball games relative to the prices of recreational substitutes, and the goodness of substitutes. Attendance at the games of any given team is a positive function of the population size of the territory in which the team has the monopoly right to play; the size and convenience of location of the ball park; and the average rank standing of the team during the season in the competition of its league. It is a negative function of the goodness of leisure-time

substitutes for baseball in the area and of the dispersion of percentages of games won by the teams in the league.

In the 1950s, when Rottenberg wrote his article, most club revenues, i.e. more than 70%, came from admissions (home and away games). Therefore, studying the economics of attendance was very important for clubs. However, the situation has changed. The value of sports broadcasting rights started to rise in the late 1960s in the USA and around the 1990s in Europe. Nowadays, particularly in Europe, TV revenues in many countries and top-tier teams constitute between 33% and 75% of total annual income for big and small clubs respectively. Besides that, when sports started to turn into a global phenomenon by the end of the last century, merchandising revenues gained more relevance than in-stadium attendance revenues for some of the more popular teams. This change in revenue generation was not foreseeable in the 1950s. Therefore, according to Rottenberg (1956), maximizing attendance figures was the key for the industry to be successful. To this end, the competitors had to be of approximately equal size. In addition, the reserve clause was necessary to assure an equal distribution of playing talent among opponents, because the

164

THE SAGE HANDBOOK OF SPORTS ECONOMICS

more equal the distribution of talents the greater the uncertainty of the game and seasonal outcome. Nowadays, however, this proposal is being challenged by various scholars who point out, that perfect competitive balance (CB) in a league is not necessary for the sports industry to be successful. The second important reference on sports economics was made by Walter Neale. Neale’s paper (1964) is important since it is the first to address the economics of the organization of professional sports leagues by concluding: first, that the unit of economic analysis is the league, neither the teams nor the individual athletes, ‘in a peculiar position vis-à-vis our accepted way of looking at the firm in a competitive market’ (p. 1); and second, that the league is a natural monopoly.1 While there are not so many references concerning in-stadium attendance in Neale’s paper, his explanation about the equilibrium in the market on page two appears to be important in this context: ‘The first peculiarity of the economics of professional sports is that receipts depend upon competition among the sportors or the teams, not upon business competition among the firms running the contenders, for the greater the economic collusion and the more the sporting competition the greater the profits’ (p. 2). He further argued that the sport product is a consequence of the indivisible act of two or more companies, and the fundamental fact in professional sports – due to its heterogeneous product – is that fans decide to attend a game, ‘the closer the standings, and within any range of standings the more frequently the standings change, the larger will be the gate receipt’ (p. 3). For Neale, standard economic variables such as income or prices do not seem to be of major importance for the analysis of attendance demand. Rather, the quality of a match is seen as the most important explanatory variable of in-stadium attendance. Moreover, and related to the latter, the input-enthusiasm effect increases ‘the demand for game admissions and therefore the derived demand for skilled players and hence their salaries, so that the monetary cost of each unit of the larger supply of higher quality players rises’ (p. 9). In this regard, quality also captures the effect of superstar players. This point of view is similar to the dilemma mentioned by El-Hodiri and Quirk (1971), who pointed out that teams try to hire as much talent as possible in order to win as many games as they can, but this comes with a downside: if a team is too successful, profits are not maximized, since the demand for tickets decreases considerably when the probability of winning gets close to one. The third father of sports economics is Roger Noll. When he organized the first conference on the economics of sports at the Brooking Institution

in 1971, something relevant was about to happen in the field of sports economics. That congress prompted the publication of the book Government and the Sport Business (Noll, 1974), which is an essential text for future researchers. In this book, Noll wrote a chapter titled ‘Attendance and price setting’, which is a genuine study of the demand of the four most relevant North American sports at that time, i.e. baseball, basketball, American football and ice hockey. Noll pointed out that the knowledge of the industry’s sources of demand and pricing practices in professional sports is important since it allows us to understand the extent of viable economic demand for professional sports, or, in other words, how many teams could (economically) survive and which cities could sustain them. This knowledge leads us to the conclusion that not all teams are focused on winning the competition, but that team owners rather look forward to managing them in a ‘profit-oriented way’ (Noll, 1974, p. 115). To outline the differences between each sport, Noll considered a list of factors that affect the demand for games and measures their correlation with attendance figures. These factors are: (i) the average ticket prices (i.e. the price per seat), (ii) the income measured by the average income earned per person in a given area (city, region, etc.) for a specified year/season, (iii) the population in the city of the home team, (iv) the number of stars in a team, (v) the team quality and playing success measured by the percentage of games won, the number of points/wins behind the team leading the standings, success in the previous season and the amount of titles won in the past, (vi) the amount of entertainment competitions (substitutes), (vii) the quality of stadiums and arenas (including novelty effects and capacity constraints), (viii) the closeness of a competition/game, (ix) the weather conditions measured by the number of sunny days in a region or the temperature, and (x) the racial composition of a city measured as a percentage of black people in the population. His results showed that most of the variables have a different impact for each sport. For example, the presence of a superstar in a baseball team translates into an increase of 150,000 spectators per season in a city of 3.5 million people, while there is no significant superstar effect in the case of football. In contrast, however, population – probably the main factor that assures teams’ economic success as the other variables are multiplied by it – is always positively related to in-stadium attendance, i.e. the greater the population the greater the number of attendees. In general, Noll found that some variables were not in line with a priori expectations. For instance, income showed a negative correlation

Economics of Attendance

with the demand for baseball tickets, which appears to be a working-class sport. On the other hand, the positive correlation between income and the number of basketball tickets sold suggests that this sport is more popular in richer cities. With respect to ice hockey, findings showed no relation between income and attendance. Also, concerning weather variables, the results contradicted a priori expectations. For instance, it was found that a sunny day decreases the attendance at American football games. Other variables, however, had the anticipated impact on in-stadium attendance. In this regard, it was found that the percentage of games won, the novelty of a stadium, or the closeness of the competition increases the demand for tickets. This is particularly true for ice hockey due to the reduced number of teams in the National Hockey League (NHL). While the aforementioned variables were positively related to in-stadium attendance, the most prominent example of a negative correlation was the price. Importantly, however, the price did not include the effect of other expenses, such as parking or consumption inside the stadium. Finally, it was found that the racial composition of the neighborhood around the stadiums affects consumer attendance behavior. Summing up, from these findings we can derive several conclusions: First, attendance is higher in bigger cities. This converts population into the most prominent variable of sports demand. Second, pricing policies of the teams are motivated by profits. Third, the quality of the team is an important driver of in-stadium attendance. A successful team and better players in the squad attract more fans. In this regard, the effect of star players and teams is greater in those cities with larger populations. Regarding the star players, it could also be argued that they were being paid less in monetary terms than their actual contributions to overall team revenues. In general, the sports where the uncertainty of outcome is most valued are baseball and American football. Nevertheless, it should be noted that American football and baseball were the most balanced of the four leagues considered in the study.

THE DEVELOPMENT The evolution of sports economics studies, and thus the examination of in-stadium attendance, started in 1969 with the publication of the articles by Jones on the National Hockey League (NHL), and those of Peter Sloane (the European father of sports economics) on the labor market of

165

professional football. Sloane (1971) concluded that soccer clubs are utility maximizers. Noll (2006) indicated, by the end of 1971, that the aforementioned ‘six original pieces had built a firm foundation for creating the field of sports economics, but there were some conspicuous gaps. The most obvious is the absence of more than casual empirical research. The original six papers do not contain a single regression’ (p. 47). This started to change in the 1970s when various books and articles about attendance emerged. The first sport economics book published was that of Henry Demmert (1973), who elaborated a system of equations apparently not related to the analysis of distinct types of attendance. Demmert established a general system of five equations whose dependent variables were: total home attendance, price of admissions, number of televised games, stock of talent and the relative quality of the club. In this system he used as explanatory variables the average ticket price and the household’s real disposable income. This study was in fact the first one that estimated a demand equation. In those years, the North-Holland Publishing Company (Elsevier) published two further books. The first one, edited by Ladany (1975), contained four articles about the sports industry, one of which (written by El Hodiri and Quirk) dealt with stadium capacities and attendance. The second book, consisting exclusively of sports articles, was published in 1976, and was edited by Machol and Ladany (1976). One of the fifteen book chapters contained an article written by Heilman and Wendling (1976), who elaborated on pricing strategies for sports events. Unfortunately, after these two books North-Holland did not continue editing books about sports. Furthermore, a symposium on professional sports celebrated at the University of New Hampshire resulted in the publication of a book edited by Jones (1980). The book contained a preface and nine chapters. However, none of them referred to in-stadium attendance. In the 1970s, a decisive decade for the evolution of sports economics, several relevant articles were published dealing with topics such as the labor market for baseball players (Scully, 1974) or the production efficiency of US professional basketball (Zak, Huang, & Siegfried, 1979). At the end of this decade and in tandem with the already mentioned articles of Demmert (1973) and Noll (1974), another five studies dealing with instadium attendance emerged, i.e. Hart, Hutton and Sharot (1975), Hunt and Lewis (1976), Gärtner and Pommerehne (1978), Siegfried and Hinshaw (1979) and Siegfried and Eisenberg (1980). All of these studies were focused on team sports. While studies dealing with attendance behavior in North America had focused on the four popular

166

THE SAGE HANDBOOK OF SPORTS ECONOMICS

professional sports, i.e. basketball, baseball, ice hockey and American football, European studies had solely focused on soccer (in particular for England and Germany). Now, 40 years later, it seems that these professional team sports continue to dominate the literature without any rivalry from other sports. The authors of these studies frequently used data from the teams and/or the games and employed ordinary least squares regressions. These pioneering empirical studies, however, should not be considered as ‘true estimations’ of attendance demand, since they did not include explanatory variables such as price or income. Nevertheless, they presented a significant advancement of our understanding with respect to the attendance behavior of sport fans.

THE CONSOLIDATION Since the 1980s, the number of studies dealing with in-stadium attendance started to grow. From that period onwards, the analysis of sports attendance – together with the analysis of CB in sports – became a core topic in the sports economics literature. This growing number of studies has motivated scholars to start organizing the empirical findings in order to provide a systematic overview of the literature. The first important survey of the attendance demand literature was provided by Cairns, Jennett and Sloane (1986). Their article was published in the Journal of Economic Studies and dealt with the analysis of both supply and demand. The supply side encompassed topics such as the labor market and the rules adopted by leagues, team revenues, the effects on players’ salaries and the monopsonistic exploitation or discrimination of players. The demand side dealt with the analysis of the nature of the sport product and the demand for team sports. Moreover, this section covered the uncertainty of outcome hypothesis and how it relates to the objectives of professional sports clubs. Finally, the paper contained several considerations about the league as a cartel as well as public policies. Noteworthy in the context of this chapter, was the different classification of the variables used to explain in-stadium attendance, i.e. economic variables (e.g. price and income), uncertainty of outcome, team quality and opportunity costs (e.g. weather). Since the aforementioned article of Cairns et  al. (1986), the majority of the papers have tended to use similar classifications for their analysis. The second relevant survey, which is (to date) the most cited survey in sports economics, was developed by Jeff Borland and Robert Macdonald in 2003 and published in the Oxford Review of

Economic Policy. In this article, the authors examined more than 60 studies and suggested some lessons to be learnt by decision-makers in both the professional sports industry and governmental spheres. They distinguished between ‘direct’ demand, i.e. live attendance and TV viewing, and the ‘derived’ demand when sporting events serve as an input for producing other goods and services. Moreover, they explained the different variables and databases used by the different authors and devoted special attention to the findings concerning uncertainty of outcome using the classification proposed by Cairns et  al. (1986), i.e. uncertainty of match outcome, uncertainty of seasonal outcome and the absence of long-run domination by one club. The third survey was published in the Rivista di Diritto ed Economia dello Sport by Garcia-Villar and Rodríguez-Guerrero in 2009. The authors examined more than 80 papers and included only those considering price as an explanatory variable. In line with the classification of variables proposed by Cairns et al. (1986), they paid special attention to the econometric techniques used in the estimations of demand equations, putting emphasis on the definition of the dependent variable, the values of the coefficients, as well as the elasticities of the price and income variables (when available). In this regard, they found Cairns’ (1990) article to be the best precedent due to his discussion of ticket price setting for the inelastic range of the demand curve and his consideration of a wider definition of costs for attending sport events. Further overviews and introductions into the economic analysis of attendance are provided in recent textbooks by distinguished scholars. These books deal with the supply and demand in sports and contain further specific subjects depending on the specialization of the authors. The relevant books are: Downward and Dawson (2000), Dobson and Goddard (2001), Fort (2003) and Downward, Dawson and Dejonghe (2009). Finally, it is worth mentioning the following diverse handbooks and books with chapters by different authors: Fizel, Gustafson and Hadley (1996), Andreff and Szymanski (2006 and 2011), Fizel (2006), Andreff (2011), Kahane and Shmanske (2012), Shmanske and Kahane (2012), Rodríguez, Késenne and Garcia (2013) and Goddard and Sloane (2014).

NEW SPORTS, DATA, VARIABLES AND ESTIMATION METHODS Up until mid-2016, we discovered 304 articles published on the subject of in-stadium attendance, which serve as the basis for this chapter.

Economics of Attendance

These studies are focused on the following professional sports: baseball, American football, soccer, ice hockey and hockey, basketball, Australian football, rugby, cricket, college sports, tennis, horseracing, bowls, indoor soccer, car racing and Formula 1, lacrosse and martial arts. Moreover, the studies focus on some new countries and leagues, even though studies about sports in the United States, United Kingdom, Germany and Australia still continue to dominate the attendance demand literature. However, today articles on sports attendance are available for most of the European countries as well as for countries such as China, Japan, Korea, South Africa and Brazil. In recent years there has also been a significant change in the number of seasons that authors use to analyse demand. In contrast to early empirical works, nowadays, it is difficult to find a paper using data for a single season. This is not surprising taking into account that in many countries there has been an improvement in statistical sources. While scholarship still continues to use Ordinary Least Squares (OLS), which was the most common method of estimation in the early years of sports economics, studies using Generalized Least Squares or models such as Tobit or Logit estimator are also becoming increasingly popular. Moreover, nowadays, many studies frequently use fixed effects and instrumental variable techniques. With regard to the dependent variable, most of scholarship uses game-level attendance data, even though several recent papers still use annual attendance and, to a lesser extent, ratios weighted by the stadium capacity or by the average of the league attendance. The specification of these demand equations is a standard function using economic variables (such as price and income) and other specific variables in line with the classification proposed by Cairns et al. (1986). Regarding the price and income variables, there has been an important change over recent years. While in most of the early empirical work the ticket price and income were not included, nowadays most of the articles do include them in their estimations. For instance, the average price of the tickets for sale, and the minimum price of all the tickets for sale are frequently used. Moreover, the authors utilize deflated prices, Fan Cost Indices, price grow rates, prices adjusted for living costs, real price indices or prices for a particular seat. Over the last few years most of the studies have continued in the line of obtaining price elasticities lower than one in absolute value, this being due to specification problems and the fact that the ticket price does not represent the true cost of attending a game, as explained by the different authors. The first income variables included in the estimations were the household’s real disposable

167

and per capita income. As in the case of the price variable, however, the authors have started using alternative approaches, such as real consumer spending, weekly earnings, monthly earnings, regional income, city income, metropolitan area income, county income, wage averages, change in per capita income or gross domestic product. Garcia-Villar and Rodríguez-Guerrero (2009) argued that a priori expectations about the sign of the income variable coefficients are vague, which is confirmed by the ambiguous findings of the various studies. The effect depends on the sport practiced and the country analysed. The variables used to control the expected quality of the match partly capture the heterogeneity of the product, since all the matches are different from one another because competitor teams are different and their standing changes throughout the competition. The main variable used in this regard is the current and/or past season winning percentage of each team. Other authors control for the number of goals scored in previous matches. Moreover, the quality of players is controlled by the number of superstars in a team, the home and away team budget or the teams’ reputation, the number of games (points) behind (or above) a team’s sports target and the use of dummies classifying a match as being irrelevant or indicating a rivalry between teams. In specific sports the authors even control for the violence of the contenders. Overall, three forms of uncertainty of outcome have been distinguished in the literature: match uncertainty, seasonal uncertainty and the absence of long-run domination of the championship by a particular club (Cairns et al., 1986). Other authors, such as Borland and Macdonald (2003), use the same classification. Szymanski (2003) proposed the use of championship uncertainty instead of long-run domination by a team. Fort (2006) uses measures of game uncertainty, playoff uncertainty, and consecutive-season uncertainty. These citations reveal a common treatment by the authors in this field. In this chapter, I refrain from discussing the various variables used by scholars to capture uncertainty because it forms part of another chapter of this handbook (see Pawlowski & Nalbantis, Chapter 16). The only thing that I wish to point out, is that probably the most accurate type of information to measure game uncertainty comes from betting odds. While measures such as league standing capture the position several days or weeks in advance, betting odds capture multiple issues, such as injuries of the best player, as well as aspects such as sanctions just before the beginning of the match. The fourth group of variables measures the opportunity costs of going to stadiums and arenas, such as weather conditions, live TV broadcasts,

168

THE SAGE HANDBOOK OF SPORTS ECONOMICS

the day and time of the match, competition with other sports and the distance between the cities of the two teams. So far, however, there are very few studies that control for the possibility of watching matches on mobile phones, tablets or personal computers. The majority of the aforementioned variables are modelled by dummies, except for the variable distance. To control for the impact of weather conditions on attendance, scholars tend to use temperature measured in degrees or rain in inches. In this regard, one might assume that good weather would favor attendance. However, on several occasions when regressions include variables that control the course of the season, it is difficult to interpret the results. Last but not least, it is worth mentioning that further explanatory variables had been tested which cannot be classified to either of the previous groups. These include public or national holidays (e.g. Memorial Day, 4th of July), public school summer vacations, percentage of Hispanics (or other ethnic groups) in the local population, years of economic downturn, cities, hooliganism related to the home team, game-fixing scandals, first year in a city and Homeland Security Advisory System Alert.

NEW SIGHTS TO DISCOVER: WHAT ARE WE DOING? During the last five decades there has been a notable improvement on the econometric techniques employed by scholars focusing on the analysis of in-stadium attendance. However, today there are still many articles that do not use important vari­ ables such as price and income, and the variables used to capture quality and opportunity costs do not differ substantially from those already introduced decades ago. While there has been a replication of the empirical work for new countries and/or for minor leagues, the related literature is currently in a (more or less) stationary state with the most cited empirical article (i.e. Garcia & Rodríguez, 2002) dating some 15 years back. I believe, however, that a profound change in the empirical examination of in-stadium attendance is required and would improve considerably our understanding of the subject, i.e. the application and empirical testing of behavioral economic concepts. This century, the Swedish Academy has placed special interest in granting the Nobel Prizes of Economics to authors who analyse the behavior of economic agents, i.e. Daniel Kahneman in 2002 and Richard Thaler in 2017. In 2002, the psychologist Kahneman was awarded a Nobel Prize for his innovation in applying psychological concepts to

empirical economic research. Together with Amos Tversky, breaking with the view of economics just relying on pure rationality and being critical with the traditional theory of expected utility, they demonstrate that people tend to prioritize guaranteeing a scenario where they do not lose as opposed to one where they take a risk to win. Also, Thaler won the Nobel Prize in 2017 for his application of psychological concepts to empirical economic research in order to better understand irrational decisions. In his research, he showed how economic decisions are biased by several aspects, such as limited rationality, absence of self-control, the tendency of people to simplify financial decisions or the endowment effect. Clearly, the work of these authors (and other scholars working in the field of behavioral economics) might and should stimulate further empirical research in sports economics. In this regard, it is important to emphasize that some of our colleagues are already applying and testing behavioral economic concepts in the context of sports demand. Due to space constraints, I will only refer to two recent studies in the following. The first one is an article written by Coates, Humphreys and Zhou (2014), who developed a consumer choice model for decisions to attend sporting events that includes uncertainty and reference-dependent preferences following the prospect theory developed by Kahneman and Tversky (1979). Their model is based on the axiomatic framework introduced by Card and Dahl (2011) and captures ‘the idea that the outcome of the choice to attend a sports event depends on the actual result of the game relative to a reference point that reflects the consumer’s expectation of the game outcome’ (p. 960). In this regard, their model is able to explain why consumers might prefer watching winning teams, and why the former might have an interest in watching upsets (i.e. the desire to see the home team achieving a totally unexpected win), two outcomes that cannot be explained by the UOH. The authors estimated several models on game attendance and came to the conclusion, that – in the presence of loss aversion – attendance figures increase with increasing expected ex-ante certainty about a game outcome. The second article was written by Jeffrey Ely, Alexander Frankel and Emir Kamenica and published in the Journal of Political Economy in 2015. It is probably one of the most talked-about papers in economics of the last couple of years. The authors ‘formalize the idea that information provides entertainment and analyze the optimal way to reveal information over time so as to maximize expected suspense and surprise experienced by a rational Bayesian audience’ (p. 216). They consider that suspense and surprise are two main factors for attendance demand as they are determinants

Economics of Attendance

of consumer utility. In this regard, a moment is loaded with suspense if some crucial uncertainty is about to be resolved. This is the emotion felt when someone is curious about what will happen next. Surprise is experienced when something happens and the belief of the state of the world changes dramatically. In their model they distinguish between two players: the principal and the agent. The former reveals information to the latter and the agent observes the realization provided by the principal and formulates beliefs. The agent has preferences for his belief path, and his utility is an increasing function of suspense or surprise. Depending on whether the agent has a preference for surprise or suspense, the principal will solve different equations for this purpose. According to their analysis, soccer is a game abundant in suspense, whereas basketball is a game full of surprise. From my point of view, following up on such studies is vital in order to better understand some still unanswered questions on the economics of attendance demand. I consider that progress will only occur if forthcoming papers are not turned into purely repetitive texts on the subject.

Note 1  Neale’s arguments have spawned substantial additional research, but unlike Rottenberg the consensus view (although not unanimous), is that both of his conclusions are incorrect (Noll, 2006).

REFERENCES Andreff, W. (Ed.) (2011). Contemporary issues in sports economics: Participation and professional team sports. Cheltenham, UK: Edward Elgar. Andreff, W., & Szymanski, S. (Eds.) (2006). Handbook on the economics of sport. Cheltenham, UK: Edward Elgar. Andreff, W., & Szymanski, S. (Eds.) (2011). Recent developments in the economics of sport. Cheltenham, UK: Edward Elgar. Volume I and II. Borland, J., & Macdonald, R. (2003). Demand for sport. Oxford Review of Economic Policy, 19(4), 478–502. Cairns, J. (1990). The demand for professional team sports. British Review of Economic Issues, 12(28), 1–20. Cairns, J., Jennett, N., & Sloane, P. (1986). The economics of professional team sports: A survey of theory and evidence. Journal of Economic Studies, 13(1), 1–80. Card, D., & Dahl, G. B. (2011). Family violence and football: The effect of unexpected emotional cues

169

on violent behavior. The Quarterly Journal of Economics, 126(1), 103–143. Coates, D., Humphreys, B., & Zhou, L. (2014). Referencedependent preferences, loss aversion, and live game attendance. Economic Inquiry, 52(3), 959–973. Demmert, H. (1973). The economics of professional team sports. Lexington, MA: Lexington Books. Dobson, S., & Goddard, J. (2001). The economics of football (2nd ed., in 2011). Cambridge, UK: Cambridge University Press. Downward, P., & Dawson, A. (2000). The economics of professional team sports. London: Routledge. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports economics: Theory, evidence and policy. Oxford, UK: Elsevier. El Hodiri, M., & Quirk, J. (1971). An economic model of a professional sports league. Journal of Political Economy, 79(6), 1302–1319. El Hodiri, M., & Quirk, J. (1975). Stadium capacity and attendance in professional sports. In S. Ladany (Ed.), Management science applications to leisuretime operations (pp. 246–262). Amsterdam, NL: North-Holland Publishing Company. Ely, J., Frankel, A., & Kamenica, E. (2015). Suspense and surprise. Journal of Political Economy, 123(1), 215–260. Fizel, J. (Ed.) (2006). The handbook of sports economics research. Armonk, NY: M.E. Sharpe. Fizel, J., Gustafson, E., & Hadley, L. (Eds.) (1996). Baseball economics: Current research. Westport, CT: Praeger. Fort, R. (2003). Sport economics (3rd ed., in 2011). Upper Saddle River, NJ: Prentice-Hall. Fort, R. (2005). The Golden Anniversary of ‘The Baseball Players’ Labor Market’. Journal of Sports Economics, 6(4), 347–358. Fort, R. (2006). Competitive balance in North American professional sports. In J. Fizel (Ed.), The handbook of sports economic research (pp. 190–206). Armonk, NY: M.E. Sharpe. Garcia, J., & Rodríguez, P. (2002). The determinants of football match attendance revisited: Empirical evidence from the Spanish football league. Journal of Sports Economics, 3(1), 18–38. Garcia-Villar, J., & Rodríguez-Guerrero, P. (2009). Sports attendance: A survey of the literature 1973–2007. Revista di Diritto Ed Economia dello Sport, 5(2), 111–151. Gärtner, M., & Pommerehne, W. (1978). Der Fußballzuschauer – ein Homo Oeconomicus [The soccer spectator – A homo oeconomicus]. Jahrbuch fur Sozial Wissenschaft, 29(1), 88–107. Goddard J., & Sloane, P. (Eds.) (2014). Handbook on the economics of professional football. Cheltenham, UK: Edgar Elgar. Hart, R. A., Hutton, J., & Sharot, T. (1975). A statistical analysis of association football attendances. Applied Statistics, 24(1), 17–27.

170

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Heilmann, R., & Wendling, W. (1976). A note on optimum pricing strategies for sport events. In R. Machol & S. Ladany (Eds.), Management science in sports (pp. 91–99). Amsterdam, NL: NorthHolland Publishing Company. Hunt, J., & Lewis, K. (1976). Dominance, recontracting, and the reserve clause: Major League Baseball. American Economic Review, 66(5), 936–943. Jones, J. H. C. (1969). The economics of the National Hockey League. The Canadian Journal of Economics, 2(1), 1–20. Jones, M. (Ed.) (1980). Current issues in professional sport. Durham, NH: Whittemore School of Business and Economics. Kahane, L., & Shmanske, S. (2012). The Oxford handbook of sports economics (Vol. 1). New York: Oxford University Press. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–292. Ladany, S. (Ed.) (1975). Management science applications to leisure time operations. Amsterdam, NL: North-Holland Publishing Company. Machol, R., & Ladany, S. (1976). Management science in sports. Amsterdam, NL: North-Holland Publishing Company. Neale, Walter (1964). The peculiar economics of professional sports: A contribution to the theory of the firm in sporting competition and in market competition. The Quarterly Journal of Economics, 78(1), 1–14. Noll, R. (1974). Government and the sports business. Washington, DC: Brookings Institution. Noll, R. (2006). Sports economics after fifty years. In P. Rodríguez, S. Késenne & J. Garcia (Eds.), Sports

economics after fifty years: Essays in honour of Simon Rottenberg (pp. 17–49). Gijón, ES: Ediciones de la Universidad de Oviedo. Rodríguez, P., Késenne, S., & Garcia, J. (2013). The econometrics of sport. Cheltenham, UK: Edward Elgar. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Scully, G. (1974). Pay and performance in Major League Baseball. American Economic Review, 64(6), 915–930. Shmanske, S., & Kahane, L. (2012). The Oxford handbook of sports economics (Vol. 2). New York: Oxford University Press. Siegfried, J., & Eisenberg, J. (1980). The demand for minor league baseball. Atlantic Economic Journal, 8(2), 59–69. Siegfried, J., & Hinshaw, C. E. (1979). The effect of lifting television blackouts on professional football no-shows. Journal of Economics and Business, 32(1), 1–13. Sloane, P. (1969). The labour market in professional football. British Journal of Industrial Relations, 7(2), 181–199. Sloane, P. (1971). The economics of professional football: The football club as utility maximizer. Scottish Journal of Political Economy, 18(2), 121–146. Szymanski, S. (2003). The economic design of sporting context. Journal of Economic Literature, 41(4), 1137–1187. Zak, T., Huang, C., & Siegfried, J. (1979). Production efficiency: The case of professional basketball. The Journal of Business, 52(3), 379–392.

18 Exposure and Television Audience Demand: The Case of English Premier League Football Babatunde Buraimo

INTRODUCTION The role of broadcasting to sport cannot be understated and, similarly, the role of sport to the broadcast market is of great importance. Sport and broadcasting share a level of mutual dependency and the complexity of this dependency is such that it is difficult to assert which entity needs the other more. The extent to which sport depends on broadcasting can be highlighted by the economic rents that flow from broadcasting to sport. This is evident in the premium sports rights market, particularly professional football, which will be the focus of this chapter. For example, the broadcasting rights fees that the English Premier League receives from broadcasters, domestic and international, only serves to highlight the extent to which the league is dependent on such revenue. For the 2016–17 season to the 2018–19 season, the English Premier League stands to generate over £5 billion from the domestic market alone. This will be further supplemented by over £3 billion from the sale of rights in non-domestic markets. The payment of substantive rights fees by broadcast companies is also replicated in other football premier leagues in countries like Germany (€717 million in 2013–14) and Spain (€949 million in 2013–14) (Deloitte, 2015).

The main consequence of these large rights fees is that leagues and their constituent clubs can spend these large sums in the football players’ labour market and attract the best players from across the globe. While attracting the best players allow clubs to fulfil their objectives of improving performances and maximising wins, these objectives are mutually compatible with those of broadcasters. Assembling the best players means that the quality of the sporting product is substantially high and this makes for an attractive television offering, which in turn can be used to attract large numbers of television subscribers and viewers. Furthermore, companies willing to pay advertising rights, on television or in the stadia, are also attracted. This short appraisal highlights how sport is dependent on broadcasting, and in turn how broadcasting is dependent on sport. The dependency of football on broadcasting (and of broadcasting on football) is predicated on one important element: exposure. Broadcasters require football to provide high-quality sporting products that allow broadcasters to meet the demand of its principal consumers: television audiences and advertisers. In this regard, the broadcaster has the incentive to select the most desirable offering and give it the greatest level of exposure on its channels. Consequently, the relationship

172

THE SAGE HANDBOOK OF SPORTS ECONOMICS

between exposure of a sports product and the size of the audiences is of interest. If consumers respond positively to a sports product (say, a team or a particular match), then the product is likely to get a high level of exposure, providing consumers with the opportunity to consume more of it. The likelihood of continued selection of the product for broadcast increases and resources are likely to be allocated accordingly. In this regard, the broadcaster may rely on the persistent transmission of certain teams and matches so that audiences might continue to respond positively. In markets where there is flexibility in how rights fees are allocated to leagues and their constituent clubs, there may be an incentive to ensure that right fee is a function of how often teams feature in transmission of games. Consequently, teams have an incentive to make sure that their offerings are attractive to broadcasters, particularly in markets where not all games are televised and broadcasters have choice and some degree of flexibility. This chapter considers how exposure of football teams influence the size of television audiences. The impact of exposure on audience size should influence how broadcasters select games for transmission. In turn, teams may need to consider how to maximise the chance of selection for broadcasting particularly in markets where there is a direct link between revenue and exposure, for example, the English Premier League. Even in football leagues in which there is no direct link between broadcast revenue and exposure, football clubs may still have an interest in maximising their exposure since clubs can offer their sponsors greater audience reach, i.e. exposing the millions of audiences watching on television to stadium advertising. The remainder of the chapter is organised as follows: the next section considers the developments and evolution that has occurred in the football broadcast market and takes account of the levels of exposure and their changes. The following section reports on the data used in the study and the specification of the models. Then the results and discussion are considered, and the final section offers concluding remarks and consideration for future research.

THE EVOLUTION AND EXPOSURE OF FOOTBALL ON TELEVISION The dependency of the broadcast market on sport is in part because of the advances in broadcast technology that has resulted in the expansion of

broadcast spectrum. In the mid-1980s, broadcast technology was limited to analogue transmission, which limited the broadcast spectrum available. Consequently, the scarcity of broadcast spectrum and the relative abundance of programming meant that sports had to compete with other forms of programming for the limited space that was available (Gratton and Solberg, 2007). In English football, live transmission of league football became a regular feature in the early 1980s. During that period, only a small number of games were televised (Buraimo and Simmons, 2015). Over the decades, the broadcast spectrum has expanded, first, as a result of the introduction of satellite broadcasting in the late 1980s and early 1990s (Gratton and Solberg, 2007) and, second, because of the advent of digital broadcasting. While satellite broadcasting served to increase the range of platforms available to viewers, digital broadcasting further increased the range of platforms and the spectrum available across those platforms. More recently, the greater use of the internet as a means of transmission has further increased the broadcast spectrum and broadcasters transmit across a multiplicity of platforms, including social media platforms, to attract audiences in greater numbers. The expansion of the broadcast spectrum has therefore meant that, compared to periods prior to the mid-1980s, premium football, as a broadcast product, is scarcer while broadcast space is more abundant. The resultant effect is that rights values of premium sports continue to rise and there has been a shift in economic power from the broadcaster to the rights owner. This is illustrated by the evolution of rights fees in English football and the English Premier League from the early 1980s to the present, as illustrated in Table 18.1. Another interesting feature of Table 18.1 is the number of live televised games per season. Since the early 1980s, this has increased from a low of 10 matches per season to a high of 168. There are a few reasons for this increase. First, the cartel between the state broadcaster, the BBC, and the commercial broadcaster, ITV, suppressed rights fees and the number of televised games was far from optimum. However, the fallout between the cartel members saw exclusive rights being awarded to ITV and, along with it, the rights to televise more games for a then-record period of years, which foreclosed the market to other broadcasters for four years. ITV’s enthusiasm was in part to restrict entry into the market by the emerging satellite broadcasters, which were in search of content for their ailing audience ratings (Williams, 1994). Second, the early years of live television were influenced by the uncertainty of the effects of live

173

EXPOSURE AND TELEVISION AUDIENCE DEMAND

Table 18.1  Domestic rights fees for English Premier League, 1983–2019 Year

Duration of contract (years)

Broadcasters

Live matches per season

Mean annual rights fee (£m)

Rights fee per match (£m)

1983 1985 1986 1988 English Premier League

2 0.5 2 4

BBC/ITV BBC/ITV BBC/ITV ITV

10 6 14 18

2.6 1.3 3.1 11

0.26 0.22 0.22 0.61

1992 1997 2001 2004 2007 2010 2013 2016

5 4 3 3 3 3 3 3

BSkyB BSkyB BSkyB BSkyB BSkyB/Setanta* BSkyB/ESPN BSkyB/BT Sport BSkyB/BT Sport

60 60 66 138** 92 and 46 115 and 23 116 and 38 126 and 42

43 199 371 341 569 594 1,006 1,712

0.71 3.32 5.62 2.47 4.12 4.30 6.53 10.19

All four tiers

*The rights reverted to ESPN in the final of the three seasons as Setanta fell into administration. **Includes the rights to 50 pay-per-view matches.

broadcast of games on attendance in the stadia. The Football League, the body responsible for league football across all four professional tiers in England, was of the view that live transmission of football would have an adverse effect on the stadium attendances (Dobson, Goddard and Dobson, 2001). Consequently, the exposure of live football on television was low and the involvement of the broadcast market was limited. However, the five biggest clubs at the time (Arsenal, Everton, Liverpool, Manchester United and Tottenham Hotspur) were displeased with how the Football League engaged with the broadcast market and with the allocation of the rights fee. The might of the big five led to the breakaway of the top tier from the Football League and the establishment of the Premier League, and with it an increase in the number of live games transmitted and an end to the cross-subsidisation, which saw rights fee being allocated to clubs in tiers 2, 3 and 4 of the English football (Dobson et  al., 2001). This marked the advent of BSkyB’s long relationship with the Premier League. There was a period of stability in so far as the number of televised football games and the exposure of the Premier League in the domestic market. However, pressure from the competition authorities meant that the relationship between the English Premier League and the monopoly broadcaster, BSkyB, was under pressure and

this is noted as the third reason for the increased exposure and the number of games. The sale of rights fee was argued to be uncompetitive for two reasons: first, the limited number of the televised games and, second, the bundling of rights as a single rights package (Harbord and Szymanski, 2004). Even though some 66 games were being transmitted during the period 2001–02 to 2003–04, this represented only 17% of all games. Furthermore, all 66 games were sold as a single package to one broadcaster, which effectively excluded other broadcasters and retained the BSkyB monopoly of live Premier League football games. The monopoly arrangement attracted the attention of the European Competition Authority (Harbord and Szymanski, 2004) and to fend off any intervention that might severely disrupt the current market arrangements, the Premier League sold the 2004–05 to 2006–07 rights not as a single package, but as three, and for a greater number of games.1 The intention was to attract other broadcasters but any anticipated entry in the Premier League live rights market did not materialise. Attempts to keep the Competition Authority at bay was temporary as in 2005, the European Commission ruled that no single broadcaster should acquire exclusive live rights to all games. It was such pressure that created the market space for the new entrant, Setanta. With the increased attention of the Competition Authority was the

174

THE SAGE HANDBOOK OF SPORTS ECONOMICS

increased offering of more live games to attract additional broadcasters. The increased exposure and collapse of the broadcast monopoly created interests from broadcasters of varying sizes. ESPN has had the opportunity to broadcast live Premier League games, and so has BT Sport, a large but new entrant to the sports broadcast market. Whatever the catalysts and drivers for more football, Premier League games have seen an increase in exposure within the domestic market. Currently, there are 168 games televised across Sky’s and BT Sport’s platforms (and other platforms are acquiring channels in the wholesale market), but this still represents only a fraction (44%) of all games per season. There may well be a period in the future when all Premier League games are televised across a multitude of broadcast platforms. However, until then, it is important for the market to understand the effects of match selection and how exposure affects audience demand. The change in exposure is not just a matter of increasing games, but how the increased number of games are distributed across teams. During the early season of the Premier League, the obligation of the broadcaster was to broadcast each team’s matches on at least one occasion. In practice, teams had their matches transmitted on at least three occasions. With 168 games to televise, the distribution of games is such that a team could expect to have at least eight of its games selected for live transmission. The constraints over selection are such that broadcasters must adhere to certain criteria. Table 18.2 shows the distribution of games across teams in the seasons 2000–01, 2007–08 and 2016–17. The selection of games for broadcast is dependent on a few contemporaneous issues, as noted by Forrest, Simmons and Buraimo (2005), and these include a positive and significant effect for matches between local rivals (derby matches), games televised at the weekend, the combined quality of the two teams (as measured by the combined relative wages. Forrest et al. model this likelihood of selection using a probit model. The criteria for selection agreed between the league and broadcaster, and the various factors determining selection give rise to the distribution of matches and exposure noted in Table 18.2. The exposure that results from this selection means that audiences become more familiar with those teams that are afforded a greater volume of television exposure. Furthermore, these teams generate higher facility fees, compared with their rivals, as can be seen in Table 18.2; in 2016–17, Liverpool accrued £34 million from having 29 of its games televised compared with its neighbour, Everton,

which received £21.5 million for its 18 transmitted games. The benefits to clubs are not just directly from facility fees. If sponsors and club partners are confident that a club is going to generate a higher level of exposure, particularly from the transmission of home games, they are likely to be willing to pay a greater value for any sponsorship or partnership rights in the knowledge that exposure will be high.

DATA AND MODEL SPECIFICATION The principal focus of this chapter is to consider the effects of exposure on audience ratings and data from the 2000–01 season to the 2007–08 season are used. As noted in the previous section, as audiences become more accustomed to seeing teams, and this might be in addition to any aura that might surround such teams, their propensity for tuning in to watch matches involving those teams might be higher. In capturing the effect of exposure on audience rating, a linear regression model is used in which the dependent variable is the (logarithm) of television audience ratings. The independent variables include combined relative wages, which is the sum of the relative wages of the two teams involved in the match. The relative wage for a given team is that team’s wage for that season divided by the mean wage for that season. So as to capture how well the teams are matched up with respect to talent, the absolute difference in relative wages is also included. With reference to performance, the sum of the performances to date of the two teams is included and this captures the form to date of the two teams in the contest. This is the sum of the ratio of points per game of the two teams. In order to capture the competitive nature of the match, the difference in the points per game of the two teams is also included. As is common in the demand literature on football, the effect of uncertainty of outcome is included. The Theil measure (Peel and Thomas, 1992; Czarnitzki and Stadtmann, 2002; Buraimo and Simmons, 2008) is used to capture uncertainty and accounts for all three outcomes in the match. The Theil measure is computed by converting bookmaker odds to probabilities. These probabilities are adjusted to account for the bookmakers’ overround. The adjusted probabilities sum to one and is a very good reflection of the likely outcome of the match. See Forrest, Goddard and Simmons (2005) for an appraisal of bookmakers’ odds and the efficient use in predicting the outcome of football matches. The measure is computed as follows:

175

EXPOSURE AND TELEVISION AUDIENCE DEMAND

Table 18.2  Distribution of televised games in the 2000–01, 2007–08 and 2016–17 seasons 2000–01 Team Arsenal Aston Villa Birmingham City Blackburn Rovers Bolton Wanderers Bournemouth Bradford City Burnley Charlton Athletic Chelsea Coventry City Crystal Palace Derby County Everton Fulham Hull City Ipswich Town Leeds United Leicester City Liverpool Manchester City Manchester United Middlesbrough Newcastle United Portsmouth Reading Southampton Stoke City Sunderland Swansea City Tottenham Hotspur Watford West Bromwich Albion West Ham United Wigan Athletic

3

∑ t =1

2008–09

2016–17

Facility Position fee (£m)

Games Facility fee Position (£m)

Position

Games

2 8

8 6

4.8 3.4

20

3

1.8

9 6 19

6 7 4

3.4 4.1 2.6

17 16

3 4

1.8 2.3

5 4 13 3 18 1 14 11

7 9 5 11 6 13 3 5

4.0 5.5 3.0 6.1 3.6 7.4 1.9 3.0

10

4

2.3

7

6

3.6

15

12

6.5

12

5

2.9

11

11

6.1

15

5

3.0

10 14

11 10

6.1 5.7

1 Pi × ln    pi 

where Pi is the probability of outcome i, which is any one of the three possible results of home win, away win or draw.

3 6 19 7 16

23 16 10 14 10

11.4 8.7 5.7 7.4 5.7

2

18

9.2

20 5 17

10 15 10

5.7 7.9 5.7

4 9 1 13 12 8 18

21 16 25 10 20 16 10

10.5 8.3 12.2 5.7 10.0 8.3 5.7

Games Facility fee (£m)

5

25

29.4

9

13

15.8

16

10

12.4

1

28

32.8

14

14

16.9

7

18

21.5

18

8

12.4

12 4 3 6 19

16 29 28 28 13

19.2 34.0 32.8 32.8 15.8

8 13 20 15 2 17 10 11

15 9 8 10 25 13 11 15

18.1 12.4 12.4 12.4 29.4 15.8 13.5 18.1

A dummy variable for derby matches is included and takes the value of 1 for matches involving teams that are historical rivals and 0 otherwise. Another dummy variable used in the model is weekend and is intended to capture the differences in matches transmitted on a weekday compared with those transmitted at the weekend.

176

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 18.3  Summary statistics of dependent and independent variables Variable

Mean

Standard deviation

Minimum

Maximum

Television audience ratings Combined relative wage Absolute difference in relative wage Combined points per game Absolute difference in points per game Theil measure Derby Weekday Setanta Arsenal Chelsea Liverpool Manchester United Lagged exposure

1.252 2.531 0.666 3.250 0.656 1.023 0.062 0.277 0.077 0.200 0.163 0.180 0.235 18.826

0.571 0.888 0.564 0.801 0.507 0.092 0.241 0.448 0.267 0.400 0.369 0.385 0.425 6.205

0.204 0.949 0.000 0 0 0.662 0 0 0 0 0 0 0 7

3.436 5.067 2.598 6 3 1.096 1 1 1 1 1 1 1 35

n = 455

Setanta is a dummy variable intended to capture the effects of transmitting a match on Setanta’s channel as opposed to any of Sky’s channels; Setanta was the junior broadcaster during the sample period. Additional dummy variables are included, and these are for the big four teams during the period of analysis – Arsenal, Chelsea, Liverpool and Manchester United. The above independent variables are control variables. The focus and main variable of interest is lagged exposure. The lagged exposure is the total number of televised games of the two teams in the previous season. The mean is 18.8 occasions; the highest is 35 and is a match involving Arsenal and Manchester United, while the smallest is seven and are matches involving a combination of teams (Charlton Athletic, Blackburn Rovers, Middlesbrough and Southampton). In order to ensure the robustness of the modelling approach, fixed effects for the different seasons and the different months of the season are incorporated. Additionally, the model’s standard errors are clustered by the rounds of matches, as the choice of matches to be televised is constrained by the fixture list and matches available in any given round. For a given set of matches in each round, the broadcasters must select a proportion of matches for that week’s transmission. This is a significant constraint, as the broadcasters cannot necessarily select the best n matches from 380 games. In this sense, the round of the fixture list is binding, and a proportion of games from

those available that weekend must be chosen. This is the condition for which the models are clustered, and the standard errors of the models are therefore adjusted to reflect this. The summary statistics of the key variables used in the model and shown in Table 18.3. The model to be estimated is as follows: Ln(audience rating) = f(combined relative wage, absolute difference in relative wage, combined points per game, absolute difference in points per game, Theil measure, derby, weekday, Setanta, Arsenal, Chelsea, Liverpool, Manchester United, lagged exposure, months dummies, season dummies)

RESULTS AND DISCUSSION The model is estimated using ordinary least squares (OLS). The use of OLS in regression analysis requires checks for robustness. One such check is for multi-collinearity, a case in which the correlation between the independent variables is high. Checks for this include an examination of the correlation coefficient matrix. An alternative and more robust approach is the variance inflation factor, in which each independent variable is regressed on the other independent variables. The resulting coefficient of determinations is then used to generate the variance inflation factor (VIF) for each variable using the following formula:

177

EXPOSURE AND TELEVISION AUDIENCE DEMAND

Table 18.4  Model for television audience demand. Dependent variable is ln (television audience ratings) Model 1 Combined relative wage Absolute difference in relative wage Combined points per game Absolute difference in points per game Theil measure Derby Weekday Setanta Arsenal Chelsea Liverpool Manchester United Lagged exposure Constant Adjusted R2 Observations

Model 2

***

0.163 −0.104*** 0.050** 0.021 −0.029 0.062 −0.082*** −0.975***

(5.19) (3.13) (2.39) (0.68) (0.14) (1.37) (2.90) (16.14)

0.025*** 12.904*** 0.690 455

(5.87) (53.33)

0.166*** −0.081** 0.040* 0.021 0.059 0.058 −0.076*** −0.984*** 0.082 −0.001 0.062 0.127* 0.016*** 12.896*** 0.691 455

(3.43) (2.36) (1.80) (0.68) (0.27) (1.24) (2.69) (16.58) (1.37) (0.01) (1.27) (1.87) (3.22) (48.95)

Absolute t statistics in parentheses. Month and season and fixed effects are included. * p < 0.10, ** p < 0.05, *** p < 0.01

VIF =

1 1− R

2

Generally, if the VIF values are greater than 10, then multi-collinearity is a concern. The highest VIF value for the model is 4.98 and the mean VIF is 2.31. Given these values, multi-collinearity is not a problem. The regression results are presented in Table 18.4. The first of the independent variables is the combined relative wages, which is positive and significant and indicates that audiences respond positively to matches that involve teams with high relative wages, which is a proxy for the quality of talent involved in the match. As the quality of talent, as captured by the relative wages, improves by say 10% at the mean value for relative wages, the audience rating improves by 3%. At the mean value of audience ratings, this is an improvement of approximately 38,000 viewers, which is not a trivial increase in audience size for a pay-television platform. The difference in the teams’ relative wages is also significant but negative, indicating that audiences prefer matches in which there is a good matchup between the two teams with respect to the quality of talent they have on their rosters. With respect to the performances to date of the teams, the combined points per game of the two teams have a positive and significant effect on audience ratings. However, the absolute

differences in points per game are not significant, suggesting that closeness in performance of the two teams is not of interest to the audience. This offers insight into the perspective on close contests and the uncertainty of outcome concept and suggests that closeness of contest is not of interest. A more direct measure intended to capture uncertainty of outcome is the Theil measure and the results suggest that audiences are not interested in a close contest. The results are in line with the findings of Buraimo and Simmons (2015) and Caruso, Addesa and Di Domizio (2017). Focusing on the dummy variable derby, television audiences are not interested in matches that have local rivalry attached to them. This is most likely to be because the majority of audiences watching are less likely to have strong affinities with either of the team and are more likely to be neutrals. Therefore, the issue of derby is a local matter for the supporters of the two teams, who are dominated by neutral fans in the television audience demand market. The dummy variable weekday is negative, showing that audiences for weekday matches are less compared with weekend matches by 8.2% (or 7.6%). This is a reflection of the greater abundance of leisure time available at the weekend. There is a significant difference between matches televised in BSkyB’s platform and that of Setanta; Setanta’s audience reach, controlling for a number of factors, is 98%

178

THE SAGE HANDBOOK OF SPORTS ECONOMICS

fewer viewers compared with BSkyB. This in part explains why Setanta eventually defaulted and could not meet payments to the English Premier League. This clearly shows that the value of football rights to Setanta is clearly of a much lower value compared with the principal broadcaster. The significance of the dummy variables for the big four teams is of minor significance in model 2 with that for Manchester United showing any level of significance at conventional levels. This suggests that the effects of big teams are generally captured by the combined relative wage variable. In model 1, dummy variables for the big four teams are not included. With their absence, the lagged exposure is significant at the 1% level. The size of the coefficient shows that for an additional game of broadcast in the previous season, audiences ratings improve by 2.5%. Put another way, for an additional six games (approximately one standard deviation of the measure of exposure), the improvement in audience rating is 15% or an improvement of 187,800 viewers at the mean, which is a substantive improvement in audience ratings. These are the interpretations of the coefficients in model 1. With the inclusion of the big teams’ dummy variables, the coefficient of exposure falls but only slightly and is still significant at the 1% level. The significance of the impact of the exposure means that broadcasters and clubs can optimise their decision-making. For the broadcaster, it is important to maximise the television audience ratings. The results suggest that consumers develop familiarity and maybe affinity with the teams they have been accustomed to viewing. Holding other factors constant, a broadcaster would be better offering consumer games of teams that were involved in a higher number of televised matches from the previous season. For the broadcaster, the benefits are not only greater audience ratings, but also the likelihood of greater advertising revenue from those companies looking for their brands and products to reach a greater audience. So, for a broadcaster to appease audiences and advertisers, a sure way of achieving this is to transmit matches that have had a legacy of transmission from the previous season. For clubs, there are important considerations also. Selection by the broadcaster is accompanied by a facility fee, which consistently grows from one contract period to the next. In 2008–09, the mean facility fee was £530,000 per match but this increased to £1.2 million per match in 2016–17. As such, there is an incentive for clubs to maximise the likelihood of their matches being televised. As Forrest, Simmons and Buraimo (2005) noted in their analysis of the broadcaster’s demand, the quality of talent as proxied by the wage bill is key. As this increases, so does the likelihood of

the match being selected for broadcasting. This is the direct benefit from exposure and is immediate. There is also a long-run benefit in that once a team’s matches have had the benefit of being selected during the current season, the broadcaster has an incentive, all things being equal, to select that team’s match the following season and therefore there is an increased likelihood of future facilities fees. There is also an indirect effect. Sponsors and partners of clubs, in the knowledge of the greater effects of exposure on audience ratings, can maximise the reach of their brands by engaging with clubs. With the likelihood of the increased probability of selection and a greater audience reach, partnership and engagement with clubs are just as viable as advertising and engagement with broadcasters. This provides the potential opportunity for brand awareness or product promotion throughout the duration of a match and on a greater number of occasions per season. The exposure of teams, with the view of maximising audience ratings, has not been a feature of the empirical analysis of televised football. The findings of this chapter highlight its importance and how broadcasters and clubs should respond, considering the empirical findings. The final section that follows provides concluding remarks.

CONCLUSIONS The relationship and interplay between football and broadcasting have evolved over the decades into a complex market. The nature of broadcasting in an age in which the broadcast spectrum is abundant is such that premium content is desirable in order to provide an offering that allows broadcasters to differentiate themselves within the market and also to charge a premium fee. As such, sport, and particularly football, has provided broadcasters with this unique offering. Premium football and the likes of the English Premier League have provided broadcasters with the exclusive offerings which in turn has allowed some broadcasters to dominate the market. The evolution of Premier League football in the domestic market has seen an increase of televised matches from a low of 13% in 1992–93 to a high of 44% in 2016–17. While the majority of matches are not televised, there is now a sizeable televised offering. This greater offering of televised games comes at a substantial cost to the broadcast market and, currently, the league will receive over £8 billion from the domestic and overseas market for televised rights. For the domestic broadcasters which pay the substantive proportion of this fee and televise

EXPOSURE AND TELEVISION AUDIENCE DEMAND

only a proportion of all games, it is important that the portion of games they elect to televise maximises the audience ratings; any maximised audience ratings are likely to be accompanied by increased revenue from television advertising. In order for this to work for the broadcaster, considering the right determinants is important. One such factor is the impact of exposure. New empirical findings in the chapter have shown that the television audience responds favourably to exposure, that is teams that have had a greater number of their matches televised in the previous season. This suggests that audiences develop familiarity and possibly an affinity for these teams and, holding all factors constant, would be willing to see such teams over and above less familiar teams. This preference is over and above the popularity of the team as the results control for the big teams like Arsenal and Manchester United. As such, the broadcaster, in maximising its output, needs to exercise some degree of bias. This bias is likely to be at the expense of smaller teams and teams newly promoted to the league. Teams can, however, respond by considering factors that might drive the likelihood of the selection. While this chapter suggests that broadcasters must consider exposure, teams can elevate the selection probability by engaging positively in the players’ labour market as broadcasters respond favourably to this also. However, such an engagement must be efficient and expenditure in the player market should be congruent with the performances that are likely to be generated from such spending as inefficient spending is likely to be punished with underperformance and potential relegation to tier 2. The increased exposure highlights not only what broadcasters should do but also how football clubs can optimise their revenue. The likely exposure that football clubs are to receive should be factored into sponsorship contracts and partnership deals. A football club should be able to estimate the likelihood of selection and the audience ratings that are likely to be watching its matches. As such, the clubs should be able to provide its partners with audience ratings and information on the exposure the clubs will receive. This chapter has not only served to highlight the complexities of the broadcast market and its interaction with the sport, it has also shown the impact of exposure on audience ratings and how football clubs, as well as broadcaster, might look to optimise their decisions accordingly. This is just one of the many complex issues in broadcasting and sports. This particular issue is predicated on the fact that not all games are televised. Further research and consideration should also be given to potential developments within the market. At

179

present, the increased number of televised games in the domestic market over the decades suggests that the force of the market is in the direction of televising more games. At present, the main restrictions are institutional arrangements. More research is necessary to consider how strong the market forces are in pushing against these institutional constraints and if the market for televised Premier League football were to be relaxed (e.g. to televise all games), what effect will that have on football? The impact of such a move would be significant across many stakeholders, including the broadcast market, comprising the incumbent broadcaster and potential entrants, the League and its constituent clubs, the clubs in tiers 2 to 4 of the Football League, stadium goers across all English football, and television audiences.

Note 1  The Gold package comprised 38 games on Sunday at 4pm, the Silver package 38 Monday night games, and the Bronze package comprised two 62 games on Saturday at 1pm and 5.15pm.

REFERENCES Buraimo, B., and Simmons, R. (2008). Do sports fans really value uncertainty of outcome? Evidence from the English Premier League. International Journal of Sport Finance, 3(3), 146–155. Buraimo, B., and Simmons, R. (2015). Uncertainty of outcome or star quality? Television audience demand for English Premier League football. International Journal of the Economics of Business, 22(3), 449–469. Caruso, R., Addesa, F., and Di Domizio, M. (2017). The determinants of the TV demand of soccer: Empirical evidence on Italian Serie A for the period 2008–2015. Journal of Sports Economics (online first, 12 July). https://doi.org/10.1177/ 1527002517717298 Czarnitzki, D., and Stadtmann, G. (2002). Uncertainty of outcome versus reputation: Empirical evidence for the First German Football Division. Empirical Economics, 27(1), 101–112. Deloitte (2015). Annual review of football finance. Manchester: Deloitte. Dobson, S., Goddard, J. A., and Dobson, S. (2001). The economics of football. Cambridge: Cambridge University Press. Forrest, D., Goddard, J., and Simmons, R. (2005). Odds-setters as forecasters: The case of English

180

THE SAGE HANDBOOK OF SPORTS ECONOMICS

football. International Journal of Forecasting, 21(3), 551–564. Forrest, D., Simmons, R., and Buraimo, B. (2005). Outcome uncertainty and the couch potato audience. Scottish Journal of Political Economy, 52(4), 641–661. Gratton, C., and Solberg, H. A. (2007). The economics of sports broadcasting. Abingdon, UK: Routledge.

Harbord, D., and Szymanski, S. (2004). Football trials. European Competition Law Review, 25(2), 117–121. Peel, D. A., and Thomas, D. A. (1992). The demand for football: Some evidence on outcome uncertainty. Empirical Economics, 17(2), 323–331. Williams, J. (1994). The local and the global in English soccer and the rise of satellite television. Sociology of Sport Journal, 11(4), C376–C397.

19 Ticket Pricing B r i a n P. S o e b b i n g

INTRODUCTION The revenue gained from the sale of tickets is one of the most important sources of revenue for sports organizations (Borland & Macdonald, 2003), and has undergone a transformation over the past several decades. An activity that was first relatively static once ticket prices were set by the organization well before the start of the event or the beginning of a season is now a dynamic process which, through technology, is integrated in both the primary and secondary ticket markets. Seminal research in sports economics by Rottenberg (1956) commented on the importance of not only the price of tickets, but also those ticket prices ‘…relative to the prices of recreational substitutes…’ (p. 246). Other chapters in this Handbook review both the secondary ticket market and dynamic demand ticket pricing behavior. Thus, the purpose of this chapter is to review the literature on ticket pricing decisions made by clubs on how to price their tickets (the primary ticket market) prior to the introduction of dynamic ticket pricing. In the review of the literature, there are three main areas of study that emerge: price discrimination, price elasticity and methodological issues.

As such, the structure of this chapter is as follows. First, the chapter presents an overview of past work reviewing ticket pricing outcomes in sports economics research. Then, I proceed to outline those three main themes of recent ticket pricing research. Finally, the chapter concludes by suggesting areas of future research.

PREVIOUS REVIEWS Over the years, sports economists examined pricing in sports leagues around the world. This review of pricing has been undertaken within the larger context of summarizing the findings on attendance demand in sport. Within this literature, it is important to note that sports organizations are generally viewed as monopolies (Simmons, 2006). There are two main reviews to highlight as it pertains to this subject area, Borland and Macdonald (2003) and Villar and Guerrero (2009). In reviewing these prior reviews, the purpose is to both summarize the key findings as well as highlight important literature to provide context to those findings. The first review was conducted by Borland and Macdonald (2003). Their summary of close to

182

THE SAGE HANDBOOK OF SPORTS ECONOMICS

60 attendance demand studies in economics journals from 1974 to 2003 noted the primary sports studied were baseball and soccer with the primary geography being the United States and the United Kingdom. Borland and Macdonald (2003) stated that most research they reviewed were crosssectional studies compared to time series or panel data. Their discussion of these studies provided two pertinent conclusions as it relates to the empirical findings regarding ticket pricing. First, the literature predicts that the ticket price along with the opportunity cost of a consumer to come and attend the game was hypothesized to be negatively related to attendance. Most of the empirical studies that included ticket price in their analysis supported this hypothesis. Borland and Macdonald (2003) said that close to one-third of the studies examined had a negative and significant impact between ticket price and attendance. Alternatively, only three studies reviewed found the opposite relationship. The second conclusion was related to price elasticities and attendance. Findings, particularly in studies on European football, conclude an effect across both time and teams. Furthermore, Borland and Macdonald (2003) noted that these findings regarding elasticity are affected by the number of substitutes within the same area in which the observed team is based. Villar and Guerrero (2009) published the other critical review of ticket pricing studies. Similar to Borland and Macdonald (2003), Villar and Guerrero’s (2009) primary focus was to summarize the existing research on attendance demand in sport from 1973 through 2007. However, they did provide some in-depth summaries regarding the literature on pricing during this time. First, they summarized a discussion regarding the general behavior of a sports firm as it relates to the findings of inelastic ticket prices. Villar and Guerrero (2009) said that prior literature provides two rationales for this conclusion. The first is the belief that behaviors by owners that are assumed to be profit maximizers may not follow typical behaviors of profit maximization. The second rationale Villar and Guerrero (2009) discussed as to why ticket prices may be priced in the inelastic range is related to the team control of other revenues, such as media, concessions, and merchandise. They reviewed over 80 empirical attendance demand studies, all of them including some form of price to attend the sporting event. The price variable in these studies was generally some version of the ticket price (e.g., minimum price, average price). As it pertains to attendance impacts based on the price, the studies reviewed reported mixed results (i.e., positive, negative, or no statistical impact). As it pertains to price elasticities, many

reported inelastic prices while some reported nonstatistically significant results. In summary, these two reviews summarize prior research looking specifically at the role that pricing plays as it relates to attendance demand at live sporting events. Findings over a three-decade period generally show consistent evidence of teams setting ticket prices in the inelastic range of the demand curve as well as a negative and statistically significant impact on attendance at these sporting events. Since these two reviews, a number of studies over the last 10 to 15 years placed more focus on the prices to these events. The remaining part of this chapter is going to focus on three themes: price discrimination, price elasticity, and methodological challenges.

PRICE DISCRIMINATION Rosen and Rosenfield (1997) noted research on price discrimination ‘represent some of the most interesting and challenging problems in microeconomics’ (p. 351). In economics, the concept of price discrimination can be traced back to the early 1900s to Arthur Pigou (Rascher & Schwarz, 2012). Price discrimination can be simply defined as identical or nearly identical goods being sold to consumers at different prices where changes in price do not reflect the same changes in cost (e.g., Crompton, 2016; Howard & Crompton, 2004; Rascher & Schwarz, 2012). Rosen and Rosenfield (1997) noted that ‘[p]rice [d]iscrimination tends to be observed in activities where inventory/capacity constraints make the marginal costs of providing service to any one user smaller than the average cost’ (p. 357). Sport represents a good empirical setting to examine issues of price discrimination. In their theoretical work, Rosen and Rosenfield (1997) noted several considerations that apply to sport. First, they noted that pricing of repeat events (e.g., games during a regular season or theater shows) should decline to sort consumer tastes and preferences. Second, they looked at the pricing of concessions and other items sold in the facility during the event. The setting of concession and ticket prices should be done in considering the amount of concessions the average consumer buys in relation to the marginal consumer in the event. In research specifically looking at ticket price discrimination in sport, an initial key question is defining the good. Within professional team sport, several definitions have been presented. For example, Borland and Macdonald (2003) defined the good simply as the contest on the field of play. However, the authors noted that the quality of the

Ticket Pricing

viewing from the seat in which the consumer purchased cannot be discounted. Besides the physical quality of seat in which the consumer will sit in for the match, quality of viewing also refers to such elements as distance to the field, the food vendors that are available within the area, and the distance to the restrooms. Simmons (2006) stated the good sold consists of two specific elements. The first element is the uncertainty regarding the outcome of the contest itself. Uncertainty of outcome can be defined as ‘a situation where a given contest within a league structure has a degree of unpredictability about the result and, by extension, that the competition as a whole does not have a predetermined winner at the outset of competition’ (Forrest & Simmons, 2002, p. 229). The second element of the good is a level of suspense and the consumer experience surrounding the suspense. Bizzozero, Flepp, and Franck (2016) defined suspense as the emotion one feels while waiting in anticipation for something to occur within the match or event. Recent research looking at ticket pricing behavior acknowledges both of these broad conceptualizations of the good within sport. For example, Humphreys and Soebbing (2012) stated the core part of the good is a game with an uncertain outcome. They also acknowledged that each seat can and will have different viewing angles for the match, which may bring different experiences for fans. In other words, fans sitting closer to the field will experience different elements of the match in comparison to fans seated in the top row of the stadium. Rascher and Schwarz (2012) outlined the distinct elements of the three levels of price discrimination along with reviewing the available sport research. Their sport context is baseball, although the outline can be applied to other sports. In their review, Rascher and Schwarz (2012) noted that first-degree price discrimination does not occur within sport. Instead, the interaction between organizations and consumers as it relates to ticket pricing within sport generally fall within second- and third-degree price discrimination. The definitions of both second- and third-degree price discrimination include offering a number of ticket pricing options to consumers and allowing consumers to select at which price they will purchase a ticket. The key distinction between second- and third-degree price discrimination is third-degree price discrimination presents an opportunity for the seller to prevent certain groups for purchasing tickets at certain prices. For example, specific ticket prices offered for seniors and youths present easily identifiable groups in which certain ticket purchasers can be excluded (Rascher & Schwarz, 2012).

183

From a research perspective, most of the research undertaken within sports economics occurs within the realm of second-degree price discrimination. Rascher and Schwarz (2012) outlined broad categories within second-degree price discrimination. These categories are volume discounts (e.g., family ticket), bundled pricing (e.g., season tickets), two-part tariff (personal seat license that allows access to purchasing a ticket), and quality discrimination (e.g., variable ticket pricing). The published research in economics generally looks at bundled pricing, two-part tariff, and quality discrimination. As it relates to season ticket pricing, both Simmons (2006) and Schreyer, Schmidt, and Torgler (2016) noted that season ticket holders are looked at by sports organizations as their most committed and loyal fans. Anecdotally, season ticket prices tend to cost less on a per game basis than single game tickets. Recent research by Schreyer et al. (2016) looked at season ticket holder behavior of a German professional soccer club. For their research, they were looking at whether a season ticket holder attended the observed football match and what time they arrived at the match. For the purposes of this chapter, the focus is on two explanatory variables: the number of season tickets bought and the cost of the season ticket. Results of their study indicated that the number of tickets bought for the season led to a lower likelihood that the person attends a match. The higher the season ticket cost means a higher likelihood that a person attends a match. Both the number of tickets bought and the cost of a season ticket do not have any statistically significant effect on when a consumer shows up to the game. Traditional two-part tariffs in sport include a consumer paying an amount of money for the access to purchase tickets in that section of pricing. However, Rascher and Schwarz (2012) pointed out that a ticket price that bundles admission to the event and concessions together (e.g., a ticket which includes all-you-can-eat food) is also a two-part tariff. Depken and Grant (2011) looked at several different goods available for purchase to consumers of Major League Baseball (MLB) games once the ticket is purchased. They partitioned these potential purchases into three categories: obligatory purchases, food purchases, and nonfood purchases. Using primarily principal component analysis, their findings broadly suggest that a team’s method for pricing tickets and other ancillary goods, such as concessions, are much more advanced than previous literature credited MLB clubs. More specifically, one of the key findings suggest that teams price discriminate in such a way consistent with the idea of a two-part tariff.

184

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The majority of the studies examining price discrimination address some form of quality discrimination. Within this area, Rascher and Schwarz (2012) outlined several areas of quality discrimination. Two of these areas are dynamic pricing and secondary market pricing, which are addressed in other chapters of this Handbook. The two other categories are multi-tiered pricing and variable ticket pricing. Multi-tiered pricing is defined as a team setting different price levels for a single game throughout the stadium. Rascher and Schwarz (2012) specifically mentioned price dispersion as a type of multi-tiered pricing. Pan, Ratchford, and Shankar (2002) defined price dispersion as ‘the distribution of prices of an item with the same measured characteristics across sellers’ (p. 434). Early theoretical work by Borenstein and Rose (1994) outlined many sources of price dispersion within organizations and industries. One key source for sport is uncertainty. Dana (2001) proposed that for a monopoly organization which has to list its price prior to when it knows the actual demand for the good or service, price dispersion is an appropriate strategy. He noted that the aforementioned description was consistent with the behavior of sports organizations. Thus, sports represent a good empirical setting to study price dispersion. Humphreys and Soebbing (2012) looked at determinants of price dispersion within MLB from 1975 until 2008. They obtained the number of ticket pricing levels and prices from the Red and Green books that are the official yearly media guides of MLB. Results from the estimation revealed that the more uncertain demand is for a team as measured by team quality over the last five years, the more teams practiced price dispersion. This finding is consistent with Dana’s (2001) early theorization. However, other results also revealed interesting findings as it relates to price dispersion. First, the older the facility meant larger price dispersion for the clubs. Second, the presence of a second MLB team in the market meant lower price dispersion. Finally, the presence of other franchises in the Big Four professional sports leagues (i.e., National Basketball Association, National Football League, and National Hockey League) did not have a statistically significant impact on price dispersion. Watanabe, Soebbing, and Wicker (2013) further examined price dispersion behavior of MLB clubs. Using a similar dataset to Humphreys and Soebbing (2012), Watanabe et al. (2013) sought to understand how the MLB agreement with secondary ticket reseller StubHub affected price dispersion. The logic was that both increased revenue and additional information reduces uncertainty for clubs. Their findings suggested the StubHub

agreement increased price dispersion. Other findings found support for Humphreys and Soebbing’s (2012) earlier research as it relates to demand uncertainty and price dispersion. Additionally, Watanabe et al. (2013) found that the creation of MLBAM, the league’s internet platform, had a statistically significant and positive impact on price dispersion. Soebbing and Watanabe (2014) explored the role price dispersion had on regular season team attendance of MLB clubs. Again, using a similar time frame as the previous two studies, Soebbing and Watanabe’s (2014) results surprisingly found that price dispersion had a negative and statistically significant impact on regular season team attendance. However, a plausible explanation was that the practice of price dispersion increased revenue. Recent research by Soebbing, Watanabe, and Seifried (2017) examined the price dispersion and revenue relationship. Their results broadly suggested that price dispersion did not have any impact on team revenue. More specifically, Soebbing et  al. (2017) looked to understand the impact that price dispersion had on team revenue when looking at different facility types. For them, they highlighted those facilities that were less than 10 years of age and those facilities older than 48 years. Their results concluded that price dispersion in new stadiums had a positive and statistically significant increase in total revenue. They did not find a statistically significant relationship for price dispersion in old facilities. Unlike multi-tiered pricing, variable pricing means charging different prices for different games. As Howard and Crompton (2004) noted, variable pricing generally occurs on one of three levels, the quality of the opposing team, the time, and the place. Time can refer to the time of day, day of the week, or a particular part of the season (e.g., summer), while place refers to the location in the facility. Research by Rascher, McEvoy, Nagel, and Brown (2007) looked at the revenue MLB teams would gain by adopting variable ticket pricing. Using data and information from the 1996 season, their findings indicated that revenues from tickets would have increased by almost 3 percent. Later research by Pelnar (2009) looked at determinants of variable ticket pricing. In his research, he defined ticket price using a weighted average ticket price for each team in the regular season. His main research focus was twofold. First, he wanted to understand if the presence of other teams in the market impacted the weighted average ticket price of the focal teams. Second, he wanted to know if an increase in horizontal integration (i.e., one individual owning multiple professional sports teams in the same market) impacted ticket price

Ticket Pricing

determinants. Estimating a number of regression models for the 2008 season for all the Big Four North American professional sports leagues, Pelnar (2009) did not find any statistically significant results for his two main research questions. In summary, the literature regarding price discrimination has largely focused on second-degree price discrimination and, more specifically, quality discrimination. From this area of research, we find that teams price discriminate in order to capture additional consumer surplus. While some research found a positive increase on revenues as it relates to price discrimination (e.g., Soebbing et  al., 2017), there was also research indicating that teams could do a better job with setting ticket prices (Rascher et al., 2007).

PRICE ELASTICITY Understanding ticket price elasticity within sports economics has been common practice in the literature. Theory indicates that profit-maximizing firms would price tickets to be unit elastic (Borland & Macdonald, 2003; Simmons, 2006; Villar & Guerrero, 2009). An early review of the literature by Cairns (1990) said that studies found tickets to be priced in the inelastic range of the demand curve. He believed this finding was due to poor data and once data regarding the full cost of attendance was included, those findings would be different. Later, Fort (2004) noted in his research that over the previous three decades, studies have overwhelmingly found that teams price tickets in either the inelastic or unit elastic range of the ticket demand curve. Instead of blaming poor data, the literature focused on two main factors that Fort (2004) outlined. They are the inclusion of other stadium revenue and fans travel to attend matches. These two factors are explored below. Fort (2004) referred to early research which noted that game day revenues, such as concessions, could be a factor in teams setting prices in the inelastic range. He further contributed to the discussion by outlining that inelastic pricing makes sense in the North American professional team sport due to local revenues related to the focal teams’ media revenue, the other league members’ media revenue, and the marginal cost of player talent in the league. As it relates to concessions, the review by Villar and Guerrero (2009) noted the important early work by Marburger (1997), who found that setting ticket prices in the inelastic range of demand is consistent with profit maximization by organizations in so far as they receive some of the revenues from the sale of concessions.

185

More recent research seeks to further understand the role of concessions as it relates to price elasticity. Krautmann and Berri (2007) looked at concession prices in the Big Four North American professional sports leagues. They used the concession prices posted by Team Marketing Report. Their results indicated that teams reduce the face value of their tickets to hopefully capture additional spending by consumers in other components of stadium revenue, such as concessions. Coates and Humphreys (2007) further explored concession prices using the Fan Cost Index from Team Marketing Report for MLB, NBA, and NFL. They concluded MLB and NBA teams price tickets within the inelastic portion of the demand curve, consistent with earlier findings. Looking at the Fan Cost Index, they reported that the Fan Cost Index price elasticity is higher than that of the ticket price elasticity. Coates and Humphreys (2007) stated that this finding is consistent with teams pricing concessions closer to unit elastic than tickets in the hope that fans will spend money on other sources of stadium revenues, such as concessions and merchandise. For the NFL, they found no statistically significant results for ticket prices and the Fan Cost Index. However, most NFL games are sold out due to its popularity as well as the NFL’s policy of not showing the game on television in the local market if the game is not considered a sell-out. Thus, the results for the NFL results are not out of the ordinary (Coates & Humphreys, 2007). Besides taking concessions into account, other research highlighted the need to explore the full cost of a fan attending a match. Simmons (2006) noted this omission of the travel cost could be one reason why the literature consistently finds ticket prices in the inelastic range. In general, most studies are limited in not being able to gather the appropriate data to look at these factors. Villar and Guerrero (2009) noted the important early work in this area by Bird (1982). Bird (1982) estimated total admission and travel cost to attend English football matches. When incorporating the full cost of travel, he found that price elasticity of demand did fall within the inelastic range of the demand curve. Forrest, Simmons, and Feehan (2002) looked at price elasticity in European Football. Contrary to most early work, they used survey data to determine locations and fans. In doing so, they were able to include the cost of travel in their analysis. While Forrest et  al.’s (2002) results were still in the inelastic range of demand, they noted that their estimates were closer to unit elastic than previous work. Pawlowski and Anders (2012) incorporated travel costs in their examination of the German Bundesliga. They determined cost by calculating the distance between home and away clubs

186

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and multiplied it by the German government law regarding the cost for reimbursement for travelers. Their findings were negative and significant for cost, indicating a similar initial finding to previous research. Finally, Wicker, Whitehead, Johnson, and Mason (2017) surveyed fans of 28 German Bundesliga clubs across multiple divisions. The surveys took place online, and asked participants a variety of questions. Of importance to this review are their estimates of ticket price and travel cost and how they impact the probability of attending. Their findings indicate that costs (in terms of ticket price and travel) reduce the probability of attendance. Furthermore, consistent with earlier research, their estimated price elasticities for ticket price indicated prices in the inelastic range. They did not report elasticities as it related to travel cost. Wicker et  al. (2017) also estimated the average consumer surplus taking into account ticket and travel cost. Their estimates revealed the average consumer surplus being approximately 350 Euros.

METHODOLOGY A review of ticket pricing cannot be complete without discussing methodology. In this section, I discuss both sources and measures of data as well as estimation issues. Both Borland and Macdonald (2003) and Villar and Guerrero (2009) outlined the challenges from a measurement standpoint when discussing ticket price within attendance demand studies. The first is how to exactly measure price. Borland and Macdonald (2003) articulated several other dimensions of price that scholars may want to consider. The first dimension would be including the opportunity cost of attending that event. Included within opportunity costs could potentially be the cost of an attendee traveling to the event. The second dimension looks at the accuracy of the ticket price variable for a particular game. While there are many pricing options at which consumers can purchase a ticket for the individual game, there are also the prices for the full season ticket package and any half or mini season ticket packages in which consumers may have purchased their ticket to the game. Thus, these various options in which the ticket to single match could have been purchased leads to the question that Borland and Macdonald (2003) articulated, which is ‘…how should admission price be represented?’ (p. 485). Many studies used the Fan Cost Index or the average ticket price published by Team Marketing Report. Both Rishe and Mondello (2004) and Soebbing and Watanabe (2014) noted some

limitations with Team Marketing Report ticket data. The first limitation is that the average ticket price produced by Team Marketing Report is not weighted based upon the number of seats available at that price. Furthermore, the average ticket price is generally not a price point at which consumers can purchase a single game or season ticket. As Rishe and Mondello (2004) noted, the ticket information from Team Marketing Report has been the best price data that academics have access to in order to conduct studies. As scholars conduct more research in the current environment, where most teams use dynamic demand pricing, better ticket price data should be available. The second methodological issue raised by Borland and Macdonald (2003) and Villar and Guerrero (2009) is endogeneity, as it relates to the presence of price as an explanatory variable within an attendance demand model. Coates and Humphreys (2007) outlined the underlying problem in their study of North American professional sports leagues. They stated that teams are local monopolies and their choice of the menu of ticket prices offered to fans is a direct result of that monopoly power. This power, in turn, results in ticket price being correlated with the equation error term in an attendance demand study (i.e., endogeneity). Villar and Guerrero (2009) noted that ‘some studies propose modelling an equation system, either as a simultaneous equation system to take into account the potential endogeneity of some of the explanatory variables such as price or the winning percentage, or as a system where the equations are not apparently related to analyse different types of attendance’ (p. 126–127). Thus, instrumental variables are used to correct for this issue. A valid instrument is a variable that can explain price but is not a significant explanatory factor for attendance. For Coates and Humphreys (2007), the instruments used to correct for the endogenous variables were the stadium capacity that the team played in and a time trend reflecting the number of seasons the team played in their home facility in the season. For Soebbing and Watanabe (2014), the variables used were the natural log of the team’s stadium capacity, the natural log of the time trend for the sample time period in the study, and the natural log of the per capita personal income of the team’s Metropolitan Statistical Area.

CONCLUSION This chapter reviews the research of ticket pricing research conducted in sports economics that looks

Ticket Pricing

at tickets in the primary market prior to dynamic demand pricing. Examining past reviews and more recent research, the chapter provides insight in three areas: price discrimination, price elasticity, and methodology. While the price of tickets has been a popular explanatory variable within attendance demand studies, recent research also examines ticket pricing or levels as the dependent variable, seeking to understand the role that team, league, and market factors play in the behavior of sports teams. As Courty (2015) noted, ‘Ticket pricing is about selling the right seat to the right individual at the right time. The decreasing costs of data analytics and ticket distribution, together with the general frustration from the large gains captured by professional resellers in highly visible online secondary markets, have stimulated a wave of pricing innovation by event organizers, ticket distributors, and online marketplaces for the sport and entertainment industries’ (p. 35). Within his commentary, Courty (2015) noted five challenges for the live performance industry, three of which are important areas for future research within the context of this chapter. First, he discussed how to handle what he calls the ‘three pillars of ticket pricing’ (p. 36). His three pillars include tier pricing, defined as the number of price points offered by a team in the facility, variable pricing, defined as the number of levels applied to each tier over the course of a season/event, and dynamic pricing. In looking at each of these three pillars, future research needs to continue to explore all three individually as it relates to attendance, ticket revenue, and other facility revenue. However, future research also needs to look at the integration of these three pillars to understand both producer and consumer behavior. This integration is something not seen within the literature so far. Courty’s (2015) second challenge is to further understand market segmentation. As he noted, organizations can offer only a select number of tickets to the general public. In turn, they might be able to charge higher prices because only certain sections are available. Furthermore, organizations can practice third-degree price discrimination and offer different prices to certain segments of the population. These different prices are generally lower than prices offered to the general public. Future research should seek to further understand these two areas of market segmentation, in particular the use of third-degree price discrimination. In order to examine these areas in more depth, researchers will need better data than what are currently available in the major professional leagues. Researchers may want to consider minor leagues such as baseball and hockey to gain access to better ticket price data for these two concepts.

187

Anecdotally, minor league teams generally provide more incentives to encourage families and students to attend games. In their examination of minor league hockey attendance, Paul, Weinbach, and Robbins (2015) noted the frequent number of reduced-price ticket promotions as well as family nights that teams in their sample had. Finally, Courty’s (2015) third challenge that applies to this chapter looks at ensuring fairness for fans who purchase tickets. Within this challenge, Courty (2015) noted that one generally sees little variation in season ticket prices from one year to the next in comparison to single-game ticket pricing. In addition, he noted that teams using dynamic demand pricing generally do not allow a single-game ticket price to fall below the per-game cost for a season ticket in the same section. The rationale he provided for these observations is that teams try to develop season ticket holders into consumers who will come back each year to purchase season tickets. This challenge presents several areas of future research. First, future research should attempt to try to gauge the amount of consumer surplus generated for season ticket holders. Rascher and Schwarz (2012) noted that one reason to use price discrimination is for sports organizations to capture some of the consumer surplus from fans. As a result, future research should attempt to look at differences in consumer surplus generated from season ticket and single-game ticket purchases. Second, future research should follow the recent work by Schreyer et  al. (2016) to further assess season ticket holders’ behavior as it relates to attending games. Within future research exploring season ticket holder behaviors, one can integrate the single-game ticket price to determine how much fairness is considered in season ticket holders’ decision to attend matches. This work, in turn, would also support Rosen and Rosenfield’s (1997) earlier call for additional research on season ticket price setting and behavior. More broadly, there is need for additional research regarding season ticket holders. As mentioned in this review, season ticket holders represent long-term customers for the club. Furthermore, these prices are generally set early in the year and do not change in relation to changes in demand over the period of time leading up to the beginning of the season. Early work by Simmons (1996) found different price elasticities for singlegame ticket purchasers compared to season ticket purchasers with the single-game ticket purchasers having a higher elasticity. The recent work by Schreyer et  al. (2016) is a positive step forward in this direction. However, more work needs to be conducted in understanding this critical group’s behavior as it relates to both travel cost and the

188

THE SAGE HANDBOOK OF SPORTS ECONOMICS

purchasing of other complimentary goods, such as concessions.

REFERENCES Bird, P. J. (1982). The demand for league football. Applied Economics, 14, 637–649. Bizzozero, P., Flepp, R., & Franck, E. (2016). The importance of suspense and surprise in entertainment demand: Evidence from Wimbledon. Journal of Economic Behavior & Organization, 130, 47–63. Borestein, S., & Rose, N. L. (1994). Competition and price dispersion in the U.S. airline industry. The Journal of Political Economy, 102, 653–683. Borland, J., & Macdonald, R. (2003). Demand for sport. Oxford Review of Economic Policy, 19, 478–502. Cairns, J. (1990). The demand for professional team sports. British Review of Economics Issues, 12, 1–12. Coates, D., & Humphreys, B. R. (2007). Ticket prices, concessions, and attendance at professional sporting events. International Journal of Sport Finance, 2, 161–170. Courty, P. (2015). Pricing challenges in the live events industry: A tale of two industries. Sport & Entertainment Review, 1, 35–43. Crompton, J. L. (2016). Implications of Prospect Theory for the pricing of leisure services. Leisure Sciences, 38, 315–337. Dana, J. D., Jr (2001). Monopoly price dispersion under demand uncertainty. International Economic Review, 42, 649–670. Depken, C. A. II, & Grant, D. (2011). Multiproduct pricing in Major League Baseball: A principal component analysis. Economic Inquiry, 49, 474–488. Forrest, D., & Simmons, R. (2002). Outcome uncertainty and attendance demand in sport: The case of English soccer. Journal of the Royal Statistical Society Series D (The Statistician), 51, 229–241. Forrest, D., Simmons, R., & Feehan, P. (2002). A spatial cross-sectional analysis of the elasticity of demand for soccer. Scottish Journal of Political Economy, 49, 336–355. Fort, R. (2004). Inelastic sports pricing. Managerial and Decision Economics, 25, 87–94. Howard, D. R., & Crompton, J. L. (2004). Tactics used by sports organizations in the United States to increase ticket sales. Managing Leisure, 9, 87–95. Humphreys, B. R., & Soebbing, B. P. (2012). A test of monopoly price dispersion under demand uncertainty. Economics Letters, 114, 304–307.

Krautmann, A. C., & Berri, D. B. (2007). Can we find it at the concessions? Understanding price elasticity in professional sports. Journal of Sports Economics, 8, 183–191. Marburger, D. R. (1997). Optimal ticket pricing for performance goods. Managerial and Decision Economics, 18, 375–381. Pan, X., Ratchford, B.T., & Shankar, V. (2002). Can price dispersion in online markets be explained by differences in the e-tailer service quality? Journal of the Academy of Marketing Science, 30, 433–445. Paul, R. J., Weinbach, A. P., & Robbins, D. (2015). Fighting, winning, promotions, and attendance in the ECHL. Sport, Business, and Management: An International Journal, 5, 139–156. Pawlowski, T., & Anders, C. (2012). Stadium attendance in German professional football: The (un) importance of uncertainty of outcome reconsidered. Applied Economics Letters, 19, 1553–1556. Pelnar, G. (2009). Competition and cooperation between professional sports franchises: The impact on ticket prices. MPRA Working Paper No. 17787. Retrieved from: http://mpra.ub.uni-muenchen.de/ 17787/ Rascher, D. A., McEvoy, C. D., Nagel, M. S., & Brown, M. T. (2007). Variable ticket pricing in Major League Baseball. Journal of Sport Management, 21, 407–437. Rascher, D. A., & Schwarz, A. D. (2012). Illustrations of price discrimination in baseball. In S. Shmanske & L. H. Kahane (Eds.), The Oxford Handbook of Sports Economics (Vol. 2, pp. 380–399). Oxford, UK: Oxford University Press. Rishe, P., & Mondello, M. (2004). Ticket price determination in professional sports: An empirical analysis of the NBA, NFL, NHL, and Major League Baseball. Sport Marketing Quarterly, 13, 104–112. Rosen, S., & Rosenfield, A. M. (1997). Ticket pricing. Journal of Law and Economics, 40, 47–63. Rottenberg, S. (1956). The baseball players labor market. The Journal of Political Economy, 64, 242–258. Schreyer, D., Schmidt, S. L., & Torgler, B. (2016). Against all odds? Exploring the role of game outcome uncertainty in season ticket holders’ stadium attendance demand. Journal of Economic Psychology, 56, 192–217. Simmons, R. (1996). The demand for English League football: A club-level analysis. Applied Economics, 28, 139–155. Simmons, R. (2006). The demand for spectator sports. In W. Andreff & S. Szymanski (Eds.), Handbook on the Economics of Sport (pp. 77–89). Cheltenham, UK: Edward Elgar. Soebbing, B. P., & Watanabe, N. M. (2014). The effect of price dispersion on Major League Baseball

Ticket Pricing

team attendance. Journal of Sport Management, 28, 433–446. Soebbing, B. P., Watanabe, N. M., & Seifried, C. S. (2017). The impact of price discrimination, price dispersion, and facilities on organizational revenue: Lessons from Major League Baseball. Managing Sport and Leisure, 22, 442–457. Villar, J. G., & Guerrero, P. R. (2009). Sports attendance: A survey of the literature 1973–2007. Rivista di Diritto e di Economia dello Sport, 5, 112–151.

189

Watanabe, N. M., Soebbing, B. P., & Wicker, P. (2013). Examining the impact of the StubHub agreement on price dispersion in Major League Baseball. Sport Marketing Quarterly, 22, 129–137. Wicker, P., Whitehead, J. C., Johnson, B. K., & Mason, D. S. (2017). The effect of sporting success and management failure on attendance demand in the Bundesliga: A revealed and stated preference travel cost approach. Applied Economics, 49, 5287–5295.

20 Secondary Ticket Markets for Sport Events Pascal Courty

INTRODUCTION Secondary resale markets have gone through unprecedented changes over the past 20 years. While resale used to be local and each deal was idiosyncratic, most transactions now take place in one of a few centralized marketplaces. The two main players, StubHub and Ticketmaster, report steady growth. Laws restricting resale have been repealed or are no longer enforced. Many sport organizations now endorse resale and sponsor official secondary marketplaces. Some teams have even started to integrate the primary and secondary markets to offer fans a single place where they can browse through large inventories of tickets, sold by the teams themselves, brokers or other fans. Several factors have contributed to the growth of secondary markets. The creation of online ticket resale marketplaces has changed the way tickets are exchanged. Brokers and fans are better connected than they ever were. Brokers list their inventories on multiple platforms and tickets aggregators offer fans a place to browse for tickets from a wide variety of sources. Price transparency has increased. Scale economies in online marketplaces have led to consolidation, centralization of inventories, reduced search costs and have

ultimately brought convenience and value to fans. The introduction of electronic tickets has delivered the long-needed technology to fight fraud and counterfeit. Online marketplaces have also brought to the table powerful players that leverage economies of scale to innovate, shape markets, and lobby governments, sometimes on behalf of fans, for less regulation, and more freedom to exchange tickets. StubHub, for example, pushes for resale deregulation, offers sellers tools to optimize prices, and engages in partnerships with sports leagues, individual teams, and event organizers. At the same time, teams and leagues have changed the way they sell tickets in the primary market, adopting variable and dynamic pricing. Primary and secondary markets are evolving together. What happens in one market shapes the other. Some event organizers have even decided to integrate the primary market and the sponsored secondary market. This chapter focuses on the main developments that have taken place in secondary markets. We define secondary markets, explain why they exist, how they have evolved, and discuss the legislations that limits resale. We then discuss the emergence of sponsored resale marketplaces. Finally, we review the literature on price determination in secondary markets with an interest on price dynamics.

Secondary Ticket Markets for Sport Events

PRIMARY AND SECONDARY MARKETS THE TRADITIONAL AND MODERN SECONDARY RESALE MARKETS Table 20.1 presents the main characteristics of the primary and secondary markets and the changes that have taken place in ticketing practices. To capture the major changes in the way tickets are sold and resold over the past 20 years, the table introduces two stylized benchmarks: the ‘traditional’ model that characterizes ticketing prior to the advent of online marketplaces and the ‘modern’ model that accounts for online ticket sales and resale. In the traditional model, the event organizer offers tickets for the seats that remain after accounting for season tickets, guest tickets and other holders. This is called the primary market. Fans can buy tickets from the box office, the team club, or authorized ticketing agents (e.g. Ticketmaster). Traditionally, all tickets were released at the same time, typically a long time prior to the event date and at a fixed price, called face value, which is printed on the ticket. The face value is often based on cost consideration (Shapiro, Dwyer, & Drayer, 2016) and does not respond to changes in game interest, local conditions, and other demand shifters that happen between the time tickets are first released and game time. The price of tickets would also vary little from season to season. The price paid in primary markets is the face value plus

191

various fees, depending on the distribution channel. Tickets bought in primary markets are nonrefundable but transferable. Tickets are resold in secondary (or resale) markets. The prices in secondary markets are agreed upon by the buyer and seller, may vary from transaction to transaction, and do not have to match the ticket’s face value. Buyers are fans and sellers are either fans, brokers or scalpers. The distinction between broker and scalper can be blurry. Brokers are licensed, often have official brick-and-mortar offices and may form networks to offer a broad inventory of tickets for various events and in a wide range of seating sections. Scalpers are not licensed. They may sell tickets through classified ads or at the last minute around the stadium, possibly violating resale laws. With the advent of the internet, resale moved online to generalist auction sites such as eBay or to specialized online marketplaces such as StubHub. Brokers list their ticket inventory on multiple marketplaces and fans use ticket aggregators to browse through a wide variety of tickets. In the modern model, there are four main actors in the primary and secondary markets: teams, fans, brokers and online marketplaces. Online marketplaces help fans exchange tickets (fan-to-fan sale) or buy from brokers (broker-to-fan sale). Online marketplaces compete over user-friendly event maps, ticket selection, guarantees to the buyer that aim to reduce concerns over fraud and counterfeit, price setting options for the seller, and fees charged to the buyer and seller. Fees are around

Table 20.1  Changes in ticketing practices Traditional Ticketing Model

Modern Ticketing Model

Primary

Secondary

Sellers

Teams

Brokers, scalpers and fans.

Buyers Intermediation

Brokers and fans Ticket box office Team club Season ticket.

Fans Traditional methods such as classified ads. Most brokers deal with local events.

Tickets

Printed ticket with face value. Physically exchanged (handling and mailing costs).

Regulation and ticket restrictions

Tickets are non-refundable and transferable. Regulations restricting the resale of paper tickets are difficult to enforce.

Price determination

Fixed in advance and often cost based. Prices vary little with seat location and game’s specifics.

Brokers price according to market conditions.

Teams sponsor secondary online marketplaces. Primary and secondary markets are getting integrated. Online resale marketplaces and aggregators reduce search costs. Teams sponsor resale marketplaces. Electronic paperless tickets can be transferred at no cost until the event starts. Resale restrictions are enforceable on sponsored sites. Teams incorporate information from secondary market sales to optimize primary market prices for current and future events and for season tickets.

192

THE SAGE HANDBOOK OF SPORTS ECONOMICS

25%, shared between the buyer and seller, and this does not include the shipping cost for paper tickets. The main online ticket marketplaces are StubHub (acquired by eBay), TicketExchange and TicketsNow (part of Ticketmaster), and RazorGator. Seatgeeks and TicketiQ are ticket aggregators. They use metasearch engines to report inventories from multiple sources and offer price forecasts and other services. With the advent of modern resale markets, price transparency has increased, and search costs have decreased. The notion of primary and secondary markets has become blurred since some event organizers have started to dynamically manage prices in the primary market. They manage prices and inventory in a similar way as some large resellers do in secondary markets. The notion of face value becomes vague. Some event organizers even bypass the primary market and sell tickets directly in the secondary market. In fact, Drayer (2011b) made the case to fully integrate the primary and secondary markets. The Philadelphia 76ers were the first team to do so in 2016. Most tickets available for sale are displayed on the same online platform. Only the tickets of sellers who prefer to operate outside the sponsored resale marketplace are excluded. This greatly facilitates the search for the best seats and the best price. Other important changes have been facilitated by the introduction of electronic paperless tickets. Electronic paperless tickets can reduce fraud because the block chain of ticket ownership can be centrally recorded by the event organizer. It also allows teams to regain some control of the secondary markets because they hold a monopoly of the certification of ticket legitimacy. Most importantly, paperless tickets give teams valuable information about who sells tickets, who attends the event, and how much the attendees are willing to pay. This information can be used to adjust the prices of season tickets, sell ancillary goods and better dynamically price each seat, for each game, aimed at specific fan demographic.

THE SIZE OF THE SECONDARY MARKET Information about the volume and dollar value of ticket resale is scattered. A measure often used is gross transaction value (GTV). It captures the total value of all resold tickets and is composed of four distinct parts: the number of tickets resold, the ticket face value, the markup (or sometimes markdown) net of transaction fees, and the fees charged by resale marketplaces. Multiple sales of

the same ticket appear as separate transactions. In algebraic terms: GTV = Number of Tickets Resold × Face  Value × (1 + Markup) × (1 + Fee) The first two components are the most important in a world where tickets are not significantly priced below market value. The number of tickets resold can be further decomposed into broker and fan resale (see discussion below). The profits earned by resellers are equal to (Number of Tickets Resold) × (Face Value) × Markup. The amount of money at stake for teams contemplating running their own resale marketplace, or sponsoring one, is Fee × GTV/(1 + Fee). Taking the upper bond of 25% for the fee (see above), we obtain that the value captured by the resale marketplace is at most one-fifth of GTV. The secondary markets for sports event was estimated to be around US$6 billion in 2011 (Drayer et  al., 2014). Ticketmaster’s parent company, Live Nation, reports in its financials (http://investors.livenationentertainment.com) that: ‘In 2015, secondary ticketing continued to be a major focus, now operating in 13 countries and delivering 34% growth in GTV for the year to $1.2 billion.’ Forbes (2017) reports that: ‘StubHub revenues were up 30% and 15% on a y-o-y basis through 2016 and in Q1’17 to $937 million and $204 million, respectively.’ Although these figures include non-sports events, what is clear is that ticket resale is growing at a fast pace.

WHY DO SECONDARY MARKETS EXIST? A fan is a person who intends to attend the event and a broker is a person who does not intend to attend, keeping in mind that the distinction is not always clear-cut, especially in the case of season ticket holders. Ultimately, the secondary market exists because fans sometimes want to (a) buy tickets that are not available in the primary market or (b) sell tickets that they do not want to use anymore. To start, people’s plans change. Many fans need to book in advance to make travel and accommodation arrangements. They have to resell their tickets when something comes up that prevents them from attending. Using a large dataset from 56 rock concerts, Leslie and Sorensen (2013, p. 268) find that ‘46% of the resale transactions in our data were sold by non-brokers (i.e. consumers)’. Other consumers find out only at the last minute that

Secondary Ticket Markets for Sport Events

they want to attend the event. Courty (2003b) calls these consumers the ‘busy professionals’. They find out only at the last minute whether they will attend the event. Another source of resale is season ticket holders. Season ticket holders buy a subscription for the entire season or for part of the season and may not want to attend all the games or cannot attend specific games. They use the secondary market to resell the tickets for the games they do not attend. Most teams do not disclose information about season ticket sales but there is no question that season tickets account for a large fraction of total attendance, if not the majority of ticket sales for many teams. For example, season ticket sales account for about 50% of total NBA tickets.1 Another reason for secondary markets is that matching frictions in the primary market prevent an efficient allocation. Seats are highly differentiated products and consumers have preferences over various seating options. Moreover, some fans come in groups and they want to sit together. Large groups pre-order blocks of seats and may not end up using all the seats. As mentioned earlier, a large fraction of tickets is owned by season ticket holders who may choose to stay home if someone offers a high price for their ticket. Some fans may not find their preferred seating section in the primary market. They may still buy their second-choice option and use the secondary market to upgrade to their desired section. The point is that the efficient matching of supply and demand may take iterations and the secondary market helps achieve this goal. Secondary markets correct imperfection in the initial primary market allocation. Finally, secondary resale markets cannot be understood outside the context of what teams do in the primary market. Secondary markets exist because event organizers can mis-price, or sometimes systematically underprice, tickets in the primary market. It may be that the event organizer makes pricing mistakes or that prices are not flexible enough to capture all differences across seats. MBL, for example, offers about 100 million seats each season for different games, in different stadiums, and different locations. It would be amazing if each seat was optimally priced in the primary market with no scope for arbitrage. In practice, the overall level of price may be too low, there may be too few seating sections leaving the best seats in some sections grossly underpriced, or the price differences across seating sections may not capture the actual differences in seating experience. There is also some evidence that sport leagues price in the inelastic part of the demand and the

193

most profitable leagues in North America are no exception (Fort, 2004; Diehl, Maxcy, & Drayer, 2015). Drayer, Shapiro and Lee (2012, p. 184) argue that ‘the sport industry has traditionally underpriced tickets using a cost-based approach in order to maximize attendance and promote fan satisfaction’. This is known as the inelastic ticket pricing hypothesis. There is no doubt that tickets for some events, such as the Superbowl, are systematically underpriced (Krueger, 2001). Underpricing may be due to cost-based pricing, fairness concerns putting constraints on profit seeking, brand building, or lack of seller sophistication. For example, Rishe, Mondello and Boyle (2014) show that it is difficult for the National Collegiate Athletic Association (NCAA) to predict demand for the ‘March Madness’ basketball tournament because there are multiple event locations and qualifying teams are not known until 5–6 days prior to the event. The pricing of tickets in the primary market has also changed dramatically. MLB initiated the main changes in pricing among the major leagues. Variable pricing, which sets the price for each game as a function of the day of the week and the opponent, was introduced in the MLB in 1999. Dynamic pricing, which changes prices for a given game from day to day, was also introduced in MLB about 10 years later. Other leagues were quick to follow. Courty (2015) reports that most teams in all four leagues use variable pricing and a large number of teams use dynamic pricing. Xu, Fader and Veeraraghavan (2016) shows that dynamic pricing could increase revenue by 15% for a given Major League Baseball (MLB) franchise. Brokers offer liquidity to the fans who do not want to use their ticket, to those who could not find the seat they wanted in the primary market, and to those who want to change seating section. Brokers offer ticket inventory over a wide range of seats. Some brokers also bundle tickets with travel hospitality packages. Sweeting (2012, p. 1146) reports that 88% of the sellers of MLB tickets on eBay sold a single ticket and concludes that ‘many sellers are season tickets holders who do not want to attend all 81 home games’. Sweeting and Sweeney (2015) study resale transaction for a specific MLB franchise. They show that most buyers only buy once and the few buyers who make multiple purchases do so in adjacent rows. This suggests that most buyers are fans. This is not the case for sellers. They have one large broker who accounts for 31.5% of all transactions. Two other sellers account for 3–6% of transactions and they also observe hundreds of ‘small’ sellers.

194

THE SAGE HANDBOOK OF SPORTS ECONOMICS

RESALE LAWS AND RESALE RESTRICTIONS Resale may be restricted by laws and by specific restrictions instituted by the event organizer. Broadly speaking, resale laws try to prevent intermediaries from capturing surplus that was meant for fans, prevent fraud and counterfeit tickets, and reduce harassment of fans by street scalpers. Moreover, a ticket is a revocable license and the seller can in principle print restrictions that limit resale and transferability. Thus, resale depends on resale laws, but also on courts’ decision, through the enforcement of resale restrictions. Such restrictions have been challenged on the ground of being unfair or unreasonable (Moore, 2009). No US federal law restricts the resale of event tickets.2 Resale legislations are issued by state legislatures or through local ordinances and have been in place since New York introduced an antiscalping statute in 1922. In practice, there is much variation in resale laws in terms of the event covered (sports, arts, concerts, etc.), where the trade takes place (classified ads, online, physical exchange close to the venue, etc.), the premium charged relative to face value, the identity of the seller (broker and fans may be treated differently), how long prior to the event the ticket is resold, and whether approval from the primary seller is required. Seatgeek.com lists the resale laws and statutes by US states and classifies resale laws into three broad categories: prohibition of resale for profit, regulated resale, and permitted resale. Regulations can be strict if resellers have to be licenced, accountable, and prices are constrained. It is weak if it restricts, for example, only resale at the event site a few hours before the game starts. According to Wikipedia, 38 of the US states have laws allowing the reselling of event tickets as long as the sale does not take place at the event site. The other 12 states have varying degrees of regulation, including broker licensing and maximum markups. Resale laws have strong opponents. The economic argument against such legislation is that ticket resale increases welfare and corrects the inefficiencies created by event organizers who do not price each seat at its market value (Happel & Jennings, 1995). Teams, leagues, and secondary resale platforms lobby lawmakers and law enforcement agencies to design the secondary market that suits them (The Toronto Star, March 31, 2017). The advent of electronic tickets and sponsored resale websites has greatly expanded teams’ ability to influence resale. Teams may restrict resale to only some ticket categories (e.g. exclude season tickets) or may require that resale takes place through the

sponsored exchange. For example, the Yankees tried in 2006 to revoke the season ticket subscriptions of fans who sold unused tickets on secondary websites. In 2015, the Golden State Warriors also threatened to cancel fans’ subscription if they would not use the official Ticketmaster secondary ticket exchange. Punishing fans for not using the sponsored website is controversial. Elfenbein (2006) estimates the impact of state resale laws on eBay transactions of NFL tickets between 2002 and 2005. He notes that state laws may not be binding in the age of internet commerce because it applies only to within-state transactions. Only federal laws can regulate interstate commerce and there are no federal laws about resale. Moreover, he argues that state laws are rarely enforced for online transactions. Despite this, he finds that resale laws reduce online transactions, increase markups and the chance of interstate transactions. However, he also finds that prices and quantities in regulated states became more similar to those in unregulated ones over the period he studied. Drayer (2011a) and Moore (2009) question the relevance of anti-scalping laws in the modern age of online ticket marketplaces. They conclude that laws remain despite being mostly obsolete and not accomplishing the stated goal of protecting consumers. Moore notes that resale laws are not being enforced because resale markets for sports tickets have gained acceptance and teams and leagues are now active participants in online ticket marketplaces, making ticket resale a mainstream and legitimate practice. To sum up, resale laws help prevent fraud and counterfeit. But price caps have not been very effective at dealing with the issue of resale for profits, which is less controversial in sport than for other events, and this is especially the case in recent years with the advent of sponsored resale marketplaces. That being said, courts play a key role in influencing what happens in resale markets, through the enforcement of ticket restrictions placed by event organizers. The real issue for modern resale is whether the court will enforce restrictions that limit ticket transferability and/or resale prices.

SPONSORED RESALE MARKETPLACES After years of battling resale for profits, the sports industry has largely changed its position. It followed the proverbial advice, ‘If you can’t beat ‘em, then join ‘em’, and now embraces resale. Take the case of the NLF: ‘For years, the NFL did

Secondary Ticket Markets for Sport Events

whatever it could to combat scalping, routinely pressing police for arrests and scouring classifieds for brokers scalping season tickets. Team owners demonized ticket brokers as greedy, unscrupulous predators’ (The Washington Post, December 20, 2009, J. Grimaldi). This is not the case anymore. In 2017, all NFL teams sponsored a resale marketplace. This new development started in 1999 with baseball when the San Francisco Giants partnered with Tickets.com to create the ‘Double Play Ticket Window’. The first sponsored resale marketplace was born. The Giants collected 10% from each transaction that took place on the sponsored marketplace. Ten years later, most teams in the four major leagues have sponsored resale marketplaces. Sponsored resale marketplaces aim to offer fans a frictionless experience to exchange tickets until game time. The agreement can be negotiated by the league, in which case teams can opt in or out, or by teams directly. An agreement typically includes a fixed guarantee from the sponsored marketplace to the team or league, an agreement to share price data back to teams, cross-advertising, and an exclusive certification of ticket authenticity. The sponsored marketplace may offer seller various pricing options (fixed price, decreasing price with daily increments, or market price matching current transactions) and may also set price floors and ceilings agreed upon with teams. Figure 20.1 illustrates the adoption of a sponsored secondary marketplace for the four main US sports leagues. (See the Appendix 20.1 for detailed team and league adoption timelines.) Each curve plots the fraction of teams in the league that have a sponsored marketplace.3 The

195

grey bars represent league-wide deals: MLB and the National Basketball Association (NBA) signed with StubHub and Ticketmaster respectively in 2007, the National Hockey League (NHL) and the National Football League (NFL) signed with Ticketmaster in 2008. The leagues all extended their deals in 2012, except the NHL, which did so in 2013.4 Sponsorship agreements are typically exclusive. The event organizer offers the exclusive right to keep track of the chain of ownership which is the only effective measure against counterfeit and fraud. The advent of sponsored secondary markets is in part a response to buyers’ fears of purchasing fraudulent tickets. Teams receive a share of the profits from the resale transactions that take place on sponsored marketplaces. Financial figures are rarely made public but there are exceptions. For example, it was reported that the Redskins received $5 million from their 2006 deal with StubHub. MLB received a $250 million split of revenue for its sponsorship with StubHub in 2007. Ticketmaster agreed to pay $100 million to create the official NFL resale website. Sponsored secondary markets serve several economic purposes. First, fans want a place where they can search for non-fraudulent tickets. Second, it reduces search costs. Fans value the ability to view large ticket inventories with multiple sections and seating options. For example, in 2012 the NBA made a deal with Ticketmaster to offer teams the option to present both primary and secondary ticket inventories next to one another. Third, teams have access to secondary markets resale data. This information is used to better predict demand and adjust prices in real time in

Figure 20.1  Fraction of teams with a sponsored secondary ticket marketplace

196

THE SAGE HANDBOOK OF SPORTS ECONOMICS

response to demand shocks revealed by secondary market prices. Data analytics also tell teams when season ticket holders attend games and when they resell. They can better target potential buyers of season tickets, enhance relationships with sponsors, and improve consumer analytics for the sale of merchandizing.

EMPIRICAL ANALYSIS OF SPONSORED MARKETPLACE With sponsored resale marketplaces teams can better manage the pricing of season and game tickets. They find out information about the value of each seat in the stadium and can use secondary market information to dynamically price game tickets in the primary market. Zhu (2014) estimates a structural model of demand with data from an MLB franchise in season 2011. He shows that the franchise could increase profits by about 7% by changing prices dynamically to compete with the secondary market. Shapiro and Drayer (2012) show that primary ticket prices increased after the introduction of dynamic pricing by the San Francisco Giants but remained below the level of prices in the secondary market. Watanabe, Soebbing and Wicker (2013) study the impact of the 2007 StubHub agreement to become the official online seller of second-hand tickets on price dispersion in the primary market. Price dispersion is measured by the number of price levels offered each season and the differences among levels. They find that the StubHub agreement had a significant impact on price dispersion. They argue that the increased price dispersion is due to a better understanding of fan demand and a more extensive use of variable pricing.

Fairness and Legitimacy Resale markets have traditionally been perceived as fraudulent schemes to take surplus away from fans. Teams have tried to reduce the negative image of secondary markets by following two legitimacy strategies: lobbying for the deregulation of anti-scalping laws and entering into strategic partnerships with secondary marketplaces. Drayer et al. (2014) documents StubHub’s efforts to build partnerships with event organizers and to lobby governments to terminate anti-scalping regulation. They study the impact of these legitimizing strategies on the volume of tickets

exchanged and the traded prices and find an impact on the former. Important concerns with sponsoring secondary marketplaces are that they may reduce brand loyalty, antagonize consumers, and discourage fans from buying season tickets. Courty and Pagliero (2013) present evidence that some artists for popular music do not vary prices in response to demand. They argue that the systematic transfer of surplus to consumers is consistent with fairness concerns. There is some evidence that teams as well leave surplus on the table, but the common use of demand-based pricing in sports (Courty, 2015) indicates that such concerns are not overarching. Shapiro et al. (2016) study the impact of moving from cost-based pricing to demand-based pricing on consumer fairness perceptions. As much of the price fairness literature in economics, they follow a survey approach. Consumers perceive the primary market as fairer than the secondary market. Familiarity with demand-based pricing increases consumer fairness perception. Sponsored resale platforms can increase consumer acceptability of secondary markets. This explains the recent endorsement of the secondary market by teams and leagues and the emphasis on price transparency and consumer protection.

Design Issues The event organizer has exclusive control over the certification of ticket legitimacy. This gives sponsored resale marketplaces a competitive advantage over non-sponsored ones. However, this advantage is limited because tickets holders can always trade in non-sponsored marketplaces. Some event organizers have tried to use paperless tickets in order to restrict resale to the sponsored marketplace, by requiring credit card or identification for admission, but such actions have been challenged in court. This raises a broader set of questions: (a) Should the event organizer promote or deter resale markets? (b) Should the event organizer control prices in the secondary markets? (c) Should the event organizer restrict resale in sponsored marketplaces, for example, the resale of season tickets? (d) What is the optimal fee per transaction to charge the buyer and seller? Several models have shown that the seller’s attitude toward resale is mixed. In Courty’s analysis (2003a), consumers are uncertain about their valuation ahead of the game. The event organizer can charge fans their expected value by releasing tickets early. But this strategy works only if

Secondary Ticket Markets for Sport Events

resale is not allowed. In this analysis, the decision to allow resale is closely linked to the timing for ticket release. Karp and Perloff (2005) reaches the same conclusion that the monopolist’s attitude toward resale is mixed. In their model, resellers can price discriminate while the primary seller cannot, but there are transaction costs associated with resale. Cui, Duenyas and Şahin (2014) consider a monopolist who sells to consumers who learn new information about their valuation close to the event date. The monopolist uses multi-period pricing and can also sell ticket options. A fan who buys a ticket option (which is non-transferable) can purchase the ticket for an additional fee later. If she chooses not to do so, she loses only the value of the option. Resale always dominates noresale. Selling ‘buying options’ does even better because there is a transaction cost associated with secondary markets. One use of ticket option we are aware of is the Baltimore Ravens.5

PRICE DYNAMICS IN SECONDARY MARKETS A ticket can be resold until the event date, after which it is worthless. Price dynamics describe how prices change from the time tickets are issued in the primary market up to the event date. A given market has many independent sellers with possibly some dominant brokers who hold multiple listings (Sweeting & Sweeney, 2015). Each seller manages a perishable inventory, makes intertemporal trade-offs, while dealing with strategic consumers, demand uncertainty, and competing against other sellers. The environment is nonstationary because tickets are perishable. Finally, resale may entail important transaction costs due to imperfect information, imperfect certification, and costly search.6 On the empirical front, the good news is that sports data are widely available and of high quality. The empirical literature has produced some insightful descriptive analyses. However, many important economic questions are beyond the reach of empirical work: What is the social benefit of secondary markets? Do sponsored ticket marketplaces benefit teams and fans? The empirical challenge is that the existence of secondary markets and the level of activity in these markets are largely endogenous. It depends on a team’s pricing decisions in the primary market (e.g. ticket split between season and regular), its choice of sponsorship deal, whether it dynamically manages prices in the primary market, and so on.

197

THEORY The theoretical literature on dynamic pricing is largely influenced by the airline revenue management problem. Most models make simplistic assumptions about demand (van Ryzin & Talluri, 2005; McAfee & Te Velde, 2006). Consumers are largely myopic and passive: they arrive randomly and do not anticipate future prices. Such simplicity is necessary to manage the complexity inherent with the intertemporal trade-offs associated with the allocation of a fixed and perishable capacity. The literature assumes that tickets are non-­transferable, leaving no room for secondary markets, and explains how prices change over time when a seller manages a fixed and perishable inventory. Price changes both as a function of inventory and time remaining. Deneckere and Peck (2012) present a competitive equilibrium model of price dynamics with aggregate demand uncertainty. The approach is general and captures the main features of sports secondary markets: competition, sequential arrival of strategic consumers, fixed capacity, and perishable good. The price in their model increases within the trading period and then falls. The lowest price in each period follows a martingale and in expectation the ‘law of one price’ holds. As we will see, this prediction is at odds with much of the evidence from sports events showing that prices often decrease close to the event date. Sweeting (2012) assumes a broker owns a single ticket. The broker sets pt to maximize value function Vt = Maxpt[ptqt + (1 − qt)EtVt+1] where qt = qt(pt) is the probability to sell when the price is pt. Sweeting reviews the revenue management literature and argues that the price path over time depends on the seller’s opportunity cost of postponing a sale, EtVt+1, which is the expected value of keeping the ticket. He shows that under fairly general assumptions, the expected value of a ticket falls over time – a prediction that is supported by his data. His MLB data indicate that consumers are drawn from a time invariant distribution and do not strategically postpone purchase. The seller’s optimal price decreases because the option value of waiting vanishes once the event starts. He concludes that a simple dynamic pricing model explains the large observed price decrease in the last month prior to the event date. Dynamic pricing increases revenue by about 16%. Sweeting and Sweeney (2015) extend the anal­ y­sis to the case where there is a dominant seller

198

THE SAGE HANDBOOK OF SPORTS ECONOMICS

(a broker with a large ticket inventory) in the secondary market. They are interested in studying a specific feature of tickets markets – prices are not set simultaneously, as is assumed in many industrial organization models. A seller who changes her price assumes instead that other sellers are not also changing their prices, although they may respond quickly to any price change. They show that the ‘staggered’ price setting can generate a rich set of price dynamics. Sweeting’s data from MLB suggest that consumers do not strategically delay purchase. This is strange because prices typically decrease by significant amounts. Many consumers could benefit from postponing purchase. Others have shown in the airline context that some consumers are strategic and understand the benefit of waiting. According to Li, Granados and Netessine (2014), about 19% of airline travellers strategically delay purchase. Dwyer, Drayer and Shapiro (2013) show that NHL fans are aware of the existence of price dynamics. Search aggregators such as seatgeek.com offer price forecasts and advice to buy tickets at the right price.

EMPIRICAL STUDIES The main finding from empirical studies is that secondary markets prices decrease on average toward the event date. Sweeting (2012) uses MLB data from eBay and StubHub on tickets resold by fans and brokers. He reports that prices decrease by 40% as the event approaches. There is also evidence that prices decrease toward game day for events that are sold out. Harrington and Treber (2014) study resale prices for the 2013 Super Bowl. They show that prices tended to decrease in the last few days prior to the event. They also argue that there is a benefit from searching because there is much variation in the posted prices. Sweeting and Sweeney (2015) study an MLB resale market where there is a large broker who accounts for about 30% of all transactions. They show that sellers tend to cut prices as the game approaches, that the dominant broker tends to cut prices more dramatically than others, and that this influences the prices of the other sellers. They conclude that the dominant broker can ‘move the market’. There is much evidence in support of the inelastic ticket pricing hypothesis in primary markets. In secondary markets, however, price elasticities are much higher. See Sweeting (2012) for evidence from the MLB and Diehl et al. (2015) for evidence

from the NFL, where secondary prices are being sold on average at 34% above face value. Shapiro and Drayer (2012) study the introduction of dynamic pricing by the San Francisco Giants in the 2010 season. Prices in the secondary market are always higher than in the primary market. They report that the average season ticket was $31, the average price under dynamic pricing was $45 and the average price in the secondary market was $64. Moreover, the average price is higher in the secondary market relative to the primary market independently of the time to game. Prices in the secondary market decrease as game day approaches. Long before game day, prices may increase or decrease.

CONCLUSION Resale markets for sports tickets have been thriving in the past 20 years. Resale is not anymore about frowned-upon greedy scalpers grabbing undeserved profits and feeding off fans and event organizers. Resale has become mainstream in most sports. It is endorsed by event organizers. The emergence of large resale marketplaces has greatly reduced search costs and has also made the price determination process more efficient. Fans can access large ticket inventories in one click. Resale markets provide liquidity for fans who change their mind and help teams price tickets in the primary market. Moreover, sponsored marketplaces deliver unique data on who attends a game, when a season ticket holder resells their ticket, and on the value of particular seats and specific events. This information is used to price tickets, sell merchandise, and negotiate deals with sponsors. The growth of resale markets raises new economic questions. Teams and leagues have regained significant control over secondary markets because they hold a monopoly over the certification of ticket authenticity. Many event organizers have asserted this power by renegotiating the terms of exclusive sponsorship. Some teams have tried to leverage their market power by imposing resale restrictions such as price floors and caps and excluding season tickets from the secondary market. But fans and competing resale platforms have challenged such restrictions. The enforcement of these restrictions is controversial and raises interesting welfare issues for economists. The research on ticket pricing in secondary markets is in its infancy. With the progressive integration of primary and secondary markets, the quality of transaction data available to researchers is constantly improving. A stylized fact from existing

Secondary Ticket Markets for Sport Events

studies is that ticket prices tend to decrease close to the event date. Will this still be the case when more fans understand the opportunities available close to the event date? More generally, sports resale markets offer a unique laboratory to study competition in perishable good markets. The event organizer competes with brokers that have various levels of inventories and fans who often resell a single ticket. Each resale market for each team within a league can be analysed independently. It has its own dedicated set of brokers, giving it a distinct market structure and, more often than not, its own sponsored marketplace. The trading of a team’s tickets, season after season, offers a unique source of resale data to study price dynamics in perishable good markets.

Notes 1  The NBA reports selling 295,000 season tickets in 2015, which represent about 55% of total tickets sold (www.sportsbusinessdaily.com/Journal/ Issues/2015/10/26/Leagues-and-GoverningBodies/NBA-tickets.aspx). Team-specific evidence from the other three major North American sports leagues suggests that season tickets represent half or more of total ticket sales. 2  The Better Online Ticket Sales (BOTS) Act of 2016 makes it illegal to use software, or (ro) bots, to purchase tickets or to resell tickets that were bought by bots (www.congress.gov/bill/ 114th-congress/senate-bill/3183). The law aims to protect fans and increase ticket availability for events that are significantly under-priced. This happens for some sports events, but it is not as common in sports as it is for concerts. 3  The 2017 data were obtained by researching each team’s website. When a league agreement is made, it is assumed that all teams agreed to it unless information was found that they opted out. 4  The Las Vegas Golden Knights were excluded from the NHL analysis, as they officially entered the league while this article was being written. 5  The Daily Record, April 4, 2010, ‘Reserving Baltimore Ravens tickets early, for a price.’ 6  Leslie and Sorensen (2013) document search frictions due to inefficient allocations and transaction costs in the context of concert tickets.

REFERENCES Courty, P. 2003a. Ticket pricing under demand uncertainty. The Journal of Law and Economics, 46(2), 627–652.

199

Courty, P. 2003b. Some economics of ticket resale. The Journal of Economic Perspectives 17(2), 85–97. Courty, P. 2015. Pricing Challenges in the Live Events Industry: A Tale of Two Industries. Sport & Entertainment Review, 1(2), 35–43. Courty, P., & Pagliero, M. 2013. The pricing of the art and the art of pricing: Pricing styles in the concert industry. In V. Ginsburgh & D. Throsby (Eds.), Handbook of the economics of art and culture (Vol. 2, pp. 299–356). Amsterdam, NL: NorthHolland Publishing Company. Cui, Y., Duenyas, I., & Şahin, Ö. 2014. Should event organizers prevent resale of tickets? Management Science, 60(9), 2160–2179. Deneckere, R., & Peck, J. 2012. Dynamic competition with random demand and costless search: A theory of price posting. Econometrica, 80(3), 1185–1247. Diehl, M.A., Maxcy, J.G., & Drayer, J. 2015. Price elasticity of demand in the secondary market: Evidence from the National Football League. Journal of Sports Economics, 16(3), 557–575. Drayer, J. 2011a. Examining the effectiveness of antiscalping laws in a United States market. Sport Management Review, 14(3), 226–236. Drayer, J. 2011b. Making a case for the integration of the primary and secondary ticket markets for professional team sports in the United States. International Journal of Sports Marketing and Sponsorship,12(3), 2–11. Drayer, J., Frascella, V.P., Shapiro, S.L., & Mahan, J.E. III. 2014. Examining the relationship between legitimacy-building strategies and firm revenues. European Sport Management Quarterly, 14(5), 464–484. Drayer, J., Shapiro, S.L., & Lee, S. 2012. Dynamic ticket pricing in sport: An agenda for research and practice. Sports Marketing Quarterly, 21(3), 184–194. Dwyer, B., Drayer, J., & Shapiro, S.L. 2013. Proceed to checkout? The impact of time in advanced ticket purchase decisions. Sports Marketing Quarterly, 22(3), 166–180. Elfenbein, D.W. 2006. Do anti-ticket scalping laws make a difference online? Evidence from internet sales of NFL tickets. SSRN Electronic Journal, June. doi: 10.2139/ssrn.595682 Forbes. 2017. StubHub to drive eBay’s revenue growth, profits could remain low. Forbes, July 18. Fort, R. 2004. Inelastic sports pricing. Managerial and Decision Economics, 25(2), 87–94. Happel, S.K., & Jennings, M.M. 1995. The folly of anti-scalping laws. Cato Journal, 15, 65. Harrington, D.E., & Treber, J. 2014. Does it pay to wait? The paths of posted prices and ticket composition for the final four and Super Bowl. Journal of Sports Economics, 15(5), 559–576.

200

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Karp, L., & Perloff, J.M. 2005. When promoters like scalpers. Journal of Economics & Management Strategy, 14(2), 477–508. Krueger, A.B. 2001. Supply and demand: An economist goes to the Super Bowl. Milken Institute Review, 3, 22–29. Leslie, P., & Sorensen, A. 2013. Resale and rentseeking: An application to ticket markets. Review of Economic Studies, 81(1), 266–300. Li, J., Granados, N., & Netessine, S. 2014. Are consumers strategic? Structural estimation from the air-travel industry. Management Science, 60(9), 2114–2137. McAfee, P.R., & Te Velde, V. 2006. Dynamic pricing in the airline industry. In T.J. Hendershott (Ed.), Handbook on economics and information Systems. Amsterdam, NL: Elsevier. Moore, D. 2009. The times they are a changing: Secondary ticket market moves from taboo to mainstream. Texas Review of Entertainment & Sports Law, 11(2), 295–308. Rishe, P., Mondello, M., & Boyle, B. 2014. How event significance, team quality, and school proximity affect secondary market behavior at March Madness. Sport Marketing Quarterly, 23(4), 212–224. Shapiro, S.L., & Drayer, J. 2012. A new age of demand-based pricing: An examination of dynamic ticket pricing and secondary market

prices in Major League Baseball. Journal of Sport Management, 26(6), 532–546. Shapiro, S.L., Dwyer, B., & Drayer, J. 2016. Examining the role of price fairness in sport consumer ticket purchase decisions. Sport Marketing Quarterly, 25(4), 227–240. Sweeting, A. 2012. Dynamic pricing behavior in perishable goods markets: Evidence from secondary markets for Major League Baseball tickets. Journal of Political Economy, 12(6), 1133–1172. Sweeting, A., & Sweeney, K. 2015. Staggered vs. simultaneous price setting with an application to an online market. Working paper. van Ryzin, G.J., & Talluri, K.T. 2005. An introduction to revenue management. In Emerging theory, methods, and applications, September, pp. 142–194. Published online in INFORMS, October 14, 2014. https://doi.org/10.1287/educ.1053.0019. Watanabe, N.M., Soebbing, B.P., & Wicker, P. 2013. Examining the impact of the StubHub agreement on price dispersion in Major League Baseball. Sport Marketing Quarterly, 22(3), 129–137. Xu, J.J., Fader, P., & Veeraraghavan, S.K. 2016. The revenue impact of dynamic pricing policies in Major League Baseball ticket sales. Working Paper. Zhu, J.D. 2014. Effect of resale on optimal ticket pricing: Evidence from Major League Baseball tickets. Working Paper.

Secondary Ticket Markets for Sport Events

APPENDIX 20.1  SPONSORSHIP ADOPTION TIMELINES MLB 1999 2004 2006 2007 2007 2008 2012

San Francisco Giants partner with tickets.com to create ‘Double Play Ticket Window’ Los Angeles Angels of Anaheim sign with Ticketmaster to create their own tickets exchange Pittsburgh Pirates have a sponsorship deal with RazorGator but tickets.com is their official reseller Boston Red Sox sign with Ace Ticket MLB signs deal with StubHub Boston Red Sox opt out of StubHub deal MLB extends deal with StubHub for another five years

2012 2013 2013 2016 2016 2017 2017

New York Yankees, Los Angeles Angels of Anaheim, and Chicago Cubs opt out of StubHub deal Los Angeles Angels of Anaheim re-sign with Ticketmaster to create their own tickets exchange New York Yankees sign with Ticketmaster Boston Red Sox opt out of StubHub deal New York Yankees resign with StubHub Toronto Blue Jays sign with StubHub Los Angeles Angels of Anaheim sign with StubHub

NFL 2003 2005 2005 2005 2005 2005 2005 2006 2006 2006 2006 2007 2008 2012

New York Giants sign with Ticketmaster to create Ticketmaster’s Team Exchange Service Indianapolis Colts sign with StubHub Houston Texans sign with StubHub San Diego Chargers sign with StubHub Chicago Bears sign with StubHub Detroit Lions sign with StubHub Atlanta Falcons sign with StubHub Washington Redskins sign with StubHub Denver Broncos sign with Ticketmaster to create ‘Broncos Ticket Exchange’ Baltimore Ravens sign with TicketsNow Cincinnati Bengals sign with StubHub Pittsburgh Steelers sign with Ticketmaster to create ‘Steelers Ticket Exchange’ NFL partners with Ticketmaster to create The NFL Ticket Exchange NFL and Ticketmaster sign a five-year extension deal of their NFL Ticket Exchange partnership

NHL 2002 2003 2006 2008 2008 2009 2010 2010 2012 2013 2013 2013

Columbus Blue Jackets begin using Ticketmaster’s exchange Vancouver Canucks sign with Ticketmaster to create ‘Prime Seat Club’ Toronto Maple Leafs create a ticket exchange with Ticketmaster Pittsburgh Penguins sign with Ticketmaster to create their own exchange NHL signs with Ticketmaster Florida Panthers and New Jersey Nets (NBA) create a ticket swap program Ottawa Senators create their own exchange, Capitaltickets.ca Ottawa Senators sign with Ticketmaster Philadelphia Flyers sign with StubHub Ottawa Senators sign with StubHub Montreal Canadiens add ‘Ticket Vault’ to their website for reselling NHL re-signs with Ticketmaster

201

202

NBA 2002 2004 2006 2006 2006 2007 2009 2012 2012 2012 2012 2015 2016

THE SAGE HANDBOOK OF SPORTS ECONOMICS

New York Knicks create their own ticket exchange Los Angeles Lakers create their own ticket exchange New Jersey Nets sign with StubHub Washington Wizards sign with StubHub Cleveland Cavaliers’ owner buys Flash Seats for their secondary market NBA signs with Ticketmaster Florida Panthers (NHL) and New Jersey Nets create a ticket swap program Ticketmaster signs with Barclays Center and therefore indirectly with Brooklyn Nets Golden State Warriors sign with Ticketmaster Oklahoma City Thunder sign with TicketsNow NBA signs with Ticketmaster Minnesota Timberwolves sign with Flash Seats Philadelphia 76ers sign with StubHub

21 The Economics of the Transfer Market Stefan Késenne

INTRODUCTION The well-known Bosman verdict of the European Court of Justice (ECJ) in December 1995 abolished the existing transfer system for end-ofcontract players. By this verdict, professional football players are free to move to another team at the end of their contract without any transfer fee. However, the transfer system for players under contract is still alive with increasing and sky-rocketing transfer fees. If players are not free to move at the end of their contract, the labor market of professional football players cannot function. Another regulation system is needed to organize the allocation of talent and the determination of the price of talent. After the abolition of the transfer system for end-ofcontract players, a free and competitive European player market can take care of the allocation and the price of talent, so the transfer market is superfluous and obsolete. Stefan Szymanski (2015) concludes from his analysis: As it currently operates, the transfer system sustains the dominance of the elite clubs by ensuring that they are the only ones with the financial muscle to afford the transfer fees payable for the

very best players. Thus, as it currently operates, the transfer system is not only unfair to players, it also promotes the opposite of what was intended.

FACTS AND FIGURES The abolition of the transfer system for end-ofcontract players in Europe has hardly had any effect on the ongoing practices in the transfer market. The reaction of many football teams after the Bosman verdict was the lengthening of player contracts, in some cases up to ten years and more. After a negative reaction to this practice by the European Union (EU), the maximum length of a football contract was set at five years by the Fédération Internationale de Football Association (FIFA) in 2002. But also, five years is still too long in the short career of a football player. Moreover, some teams forced their players to sign a new contract before the old contract expired. By this obviously illegal practice, many players may never reach the end of their contract, and so the buying and selling of players on the transfer market continued, with higher transfer fees than ever before, up to more than €100 million paid by Real Madrid C.F. for Cristiano Ronaldo and Gareth Bale.

204

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The number of transfers in European football after the Bosman verdict increased from 5,700 in 1996 to 18,000 in 2011 – so much for the alleged stability of employment. Over the same period of 15 years, the total amount of transfer spending went up from €400 million to €3,000 million, with less than 2% of all transfer spending filtering down to smaller clubs. However, the net spending on playing talent in the transfer market (i.e. paid transfer fees minus received transfer fees) has not increased. These figures show that the transfer spending of the European top teams has become a quasiclosed money circuit among a few top teams from England, Spain, Germany or Italy. As has been mentioned by the KEA study on football transfers, commissioned by the European Commission (KEA-CDES, 2013), European football has evolved to a ‘de facto’ closed league comparable to the officially closed major leagues in North America. A market regulation such as the transfer system exists in no other industry. A manager or a worker who moves to another company is not paid for by a transfer fee. If a researcher or a professor with a PhD from one university moves to another university, no compensation is paid. There are no convincing arguments why this should be different in professional football. It is not clear what these transfer fees are supposed to compensate. Furthermore, this trade of player contracts also raises ethical questions. A human being cannot be owned by another human being and sold like a fighting bull in a cattle market. Already very young child players are traded in the transfer market. It is a form of modern slavery, be it that some slaves are very well paid. This raises the question if, after the Bosman verdict, football would not be better off without a transfer market for end-ofcontract and in-contract players. In this chapter, we investigate what the impact of the transfer market is on competitive balance (CB), average player salary levels and salary distribution, compared with the impact of a competitive player market.

TRANSFER MARKET AND COMPETITIVE BALANCE The transfer market is necessary to prevent large and rich teams from attracting the best players, leaving the smaller clubs behind with less talented players without any compensation, so the story goes. In the transfer market, small clubs receive transfer fees which allow them to engage new

players and ensure their sporting and financial survival. However, as transfer fees are generally determined on the basis of the players’ earnings, and given that bigger clubs pay higher salaries, the smaller clubs are not in a position to acquire good players. Small clubs cannot pay the salaries of top players. But without transfer fees, mid-sized teams might be able to offer short-term contracts to top players, paying their high salary, which is only a fraction of their transfer fee. So the transfer market actually worsens the competitive imbalance. Without transfer fees, the questionable role of player agents in agreements on transfer fees will also be reduced. In any case, no improvement of the CB whatsoever can be expected in a profitmaximization league. This can be illustrated in Figure 21.1, which shows the equilibrium in the player market for a two-team league with a large-market team (x) and a small-market team (y). Under profit maximization with a fixed supply of talent, the demand curves for talent are given by the marginal revenue curves (MRi). The point of intersection of the marginal revenue curves indicates the free market equilibrium (E) with distribution of talent (te). The impact of the transfer system can be analyzed by starting from an initial distribution of talent (ta) that is different from the free-market distribution (te). This talent distribution (ta) will not be stable because both profit-maximizing teams can increase their profit by trading players. Given the unit cost of talent (ce), the MR of talent in the large team is higher than the marginal cost, and profits can be increased by buying more talent. The opposite is true for the small team, which can increase its profits by selling players. Both teams will agree on a transfer fee to trade players in the transfer market, and playing talent moves from the small team to the large team. This trade of players’ talent will continue until the difference between the teams’ marginal revenues of talent is eliminated and, for both teams, it holds that MRi = ce. It follows that the distribution of talent will be the same with or without the transfer system. Players will move freely to another team in a free market and will be sold to another team in the transfer market. Profit-maximizing teams will not use the received money from transfer fees to acquire more talent; they will pocket the money. This is an illustration of the well-known Invariance Proposition of Simon Rottenberg (1956). However, most sports economists agree that, in European football, teams are not behaving as profit maximizers, but rather as win maximizers. Garcia-del-Barrio and Szymanski (2009) have shown empirically that the decisions taken by team managers in the English Premier League and in

The Economics of the Transfer Market

205

Figure 21.1  Transfer system and competitive balance in a profit-maximization league the Spanish Liga are driven by win maximization. If it is assumed that teams maximize their winning percentage under the breakeven constraint, it can be shown that the transfer system can have only a very small positive effect on CB. Figure 21.2 illustrates this improvement of the CB. The demand curves for talent of win-maximizing teams are given by their average revenue (ARi) or net average revenue (NARi), depending on the zero or nonzero capital cost, with market equilibrium (E). Starting again from a distribution of talent (ta) that deviates from the free-market equilibrium (te), we can see that the AR of talent in the large-market team is higher than the average cost (arx > ce). It

follows that team x is profitable. But a win-­maxi­ mizing team is not interested in profits. The team prefers to use its profit to buy more talent. In the small team, the opposite is true, the AR of talent is lower than the average cost (ary < ce), and the team is making a loss. In order to break even, it will prefer to sell talent. Both teams will agree to trade players on the transfer market, and talent will move from the small team to the large team until AR equals AC. The difference with the profit-maximizing case is that a win-maximizing small team will spend the received transfer fees on more talent, which causes an upward shift of its demand curve in Figure 21.2.

Figure 21.2  Transfer system and competitive balance in a win-maximization league

206

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The large team will be forced to reduce its demand for talent because it has to pay transfer fees, with a downward shift of its demand curve. The result is an improvement of the CB in te* (see Lavoie, 2000). However, this improvement can only be marginal or non-existent, because the sale of an occasional top player will weaken the small team which can only buy one or more average players with the received transfer fee. This will worsen the competitive balance (see Késenne, 2014).

TRANSFER MARKET AND AVERAGE PLAYER SALARY The strict regulation of player mobility by a transfer system turns the player labor market into a monopsony, that is, a market with only one demander. In its most extreme form of a singleentity league, where the league is the only employer of professional football players, the individual teams are just the local plants or branches of the league (see Neale, 1964). But also, in the less extreme case of a strict transfer system, teams have monopsony power in the player labor market. In a monopsony labor market, the sole employer of labor has the power to discriminate among players, and to exploit players. Economic theory shows that, under profit maximization, players in a monopsony labor market can be exploited, even by a non-discriminating monopsonist. Because a monopsonist in a labor market is facing an upward sloping labor supply curve: ∂c > 0 , the first-order condition for c = f(t) with ∂t maximizing profits π = R(t ) − c(t )t − c 0 can be

∂R ∂c ∂R = c + t , so > c, indicating ∂t ∂t ∂t that the MR of talent is higher than the salary level. Players are underpaid, as can be illustrated in Figure 21.3. For a non-discriminating monopsonist, the MCcurve is above, and steeper, than the supply curve, because the higher salary paid to the last-hired talent has to be paid to all present talents in the team. Given the profit-maximizing condition, MR = MC, only t* talents will be hired, and they are only paid a salary c*, which is lower than mr, because salary c* is high enough to attract the number of talents that maximizes profits. It follows that players are exploited. An empirical verification of this player exploitation in Major League Baseball (MLB) can be found in Scully (1989). This exploitation of players does not occur under win-maximization. Rather the opposite is happening; with or without the transfer system, players will be overpaid. If a strict transfer system grants monopsony power to a win-maximizing team, a non-discriminating team, facing an upward sloping supply curve of talent, will hire talent until its total revenue is spent. Given the breakeven constraint under win maximization with zero capital cost, R = ct, the unit cost of talent is equal to the AR. Also, by definition, it holds that R = AR.t, and δ AR so MR = AR + t . If the slope of the AR–curve ∂t * is negative, c = AR > MR, players are overpaid. With a non-zero capital cost, playing talent can still be paid below MR, depending on the size of the wage–turnover ratio (see Késenne, 2010). The zero-capital cost case is illustrated in Figure 21.4. At the point of intersection (E) of the AR and the talent supply, which is also the found to be:

Figure 21.3  Transfer market and profit maximization

207

The Economics of the Transfer Market

Figure 21.4  Transfer market and win maximization

free-market equilibrium, playing talent is paid c*, which is above marginal revenue mr.

TRANSFER MARKET AND SALARY DISTRIBUTION The transfer system also changes the distribution of player salaries. In a free and competitive player market, apart from taxes and social security contributions, the unit cost of playing talent (ci) is simply equal to the salary of a talent (si), ci = si. If European transfer spending has become an almost closed money circuit among a limited number of European top teams, the sum of all paid transfer fees is more or less equal to all received transfer fees, apart from the money that is lost from the circuit in the pockets of players’ agents. The unit cost of a playing talent can be said to equal its yearly salary plus the paid transfer fee (tpi) minus the received transfer fee (tri), divided by the contract duration (mi), so:

ci = si + (tpi − tri ) / mi (1)

If the transfer market is efficient, one single marketclearing unit cost of talent c* will emerge, and:

si = c* − (tpi − tri ) / mi(2)

It follows that salaries per unit of talent can deviate, depending on paid and received transfer fees and on contract duration. So equal talents do not receive equal pay, which is unfair. For a young and promising player, it holds that tpi < tri and si > c*.

For older players, past their performance peak, tpi > tri and si < c*. So, older players might be worse off with a transfer system. The average salary per unit of talent, with m being the average contract duration, can be derived as:

s = 1 / n∑ si = c*−

∑ (tpi − tri ) (3) nm

If much money from transfer fees disappears in the pockets of player agents, it holds that, in the quasi-closed money circuit, ∑ tpi > ∑ tri , and so, s < c* . The average player salary will be higher in a free and competitive player market without a transfer system. Without player agents, ∑ tpi = ∑ tri and s = c* , the average salary level per talent is not different from the average salary per talent in a competitive * player market where si = c* for all i and s = c . Furthermore, the equality ∑ tpi = ∑ tri does not hold for an individual team. If a team is a net-seller of playing talent in the transfer market, as most small-market teams are, ∑ tpi < ∑ tri and so, according to equation (2), si > c*, that is, small teams will pay higher salaries per talent than large teams. This does not mean that small teams pay higher player salaries than large teams; it is obvious that large teams will employ the most talented players.

WHAT IS THE ALTERNATIVE? The alternative to the transfer system is the labor market practice in other industries where transfer fees do not exist. The abolition of all transfer fees in football does not imply that contracts between

208

THE SAGE HANDBOOK OF SPORTS ECONOMICS

teams and players should not be respected. Also, in football, a contract is a contract between two partners, as in any other industry. A contract cannot be broken one-sided without a punishment of the contract-breaking party. In football, this punishment should not be a money fine, because this would reintroduce the transfer fee by the back door, as the fine can be paid by the new team. A contract-breaking player should be suspended for a number of games or for the rest of the season. A contract-breaking team should be punished by the loss of points in the ranking or by relegation to a lower division. Obviously, it should still be possible that team and player open their contract by mutual consent, but leagues have to control and punish inadmissible pressure on players. A contract can also be broken by court decision if one of the parties is not honoring the contract. Players should not be allowed to change team during the season, because this can falsify the competition. However, one of the implications of the abolition of all transfer fees is that some teams, in particular net-selling small-market teams with a well-established and successful youth training program, will suffer because the compensation of training costs by receiving transfer fees is lost. But this does not justify a violation of the EU competition laws by limiting competition and paying transfer fees. This problem can be taken care of in the youth training compensation proposal, which will be discussed in the next section.

TRAINING COMPENSATION AS AN ALTERNATIVE TO THE TRANSFER SYSTEM One of the major concerns of sports federations, in football as well as in other professional team sports in Europe, is the development and training of young talent. Youth training has been under pressure because of rising costs and changing regulations regarding the mobility of players in the EU. An unwanted side effect of the Bosman ruling was that many teams reduced, or even stopped, their investment in youth training, because trained players could leave the team without any compensation. By the abolition of all transfer fees, as advocated above, football clubs that spend a lot of money on youth training, and consequently become net-sellers of playing talent, would suffer. But the transfer system is not a necessity, nor a suitable system to compensate youth training. A simple model can show the positive effects of a youth compensation system that is not linked to the transfer of players.

Assume that there are n clubs in a league and that each club has a different budget (Ri). Each club contributes the same percentage (μ) of its budget to a youth training fund. The collected money of the fund, which then equals µ ∑ nj R j , is redistributed according to the clubs’ relative effort in youth training. So, each team receives a share (si) from the fund. Under these assumptions, the after-sharing club revenue (R*) can be written as: Ri* = (1 − µ ) Ri + si µ ∑ nj R j or Ri* = Ri + µ (nsi R − Ri ) with ∑ nj s j = 1  (4) In general, the budgets of small and large teams, before and after the youth training compensation, R will be unchanged if nsi R = Ri or si = i , that is, nR if their relative effort in youth training will be the same as their relative budget. So, a club will see its R revenue go up if si > i . nR In the particular case that si = 1 / n or nsi = 1, each club puts the same effort in youth training and receives the same amount of money µ R from the fund, the budget of a small club with Ri < R, will increase because it receives more money from the fund than it has contributed. This youth training compensation system will also have consequences for the distribution of talent in a league and the CB, because it is at the same time a revenue-sharing system. However, with this system, the negative effect of revenue sharing on talent investment is eliminated, given the strong incentive to invest in young talent that is built into the system. It can be derived that, if all teams in a league are win maximizers, as in European football, this youth training improves the CB if the small teams put more or equal relative effort in youth training compared with their relative budget. If teams are win maximizers under the breakeven constraint, the demand for talent is given by the net-average revenue NAR = (Ri – ci°)/ti (see Késenne, 2011). After sharing, the NAR is:

NARi* =

1 (1 − µ ) Ri + nsi µ R − ci0 (5) ti 

And the impact of the share parameter μ can be seen to be:

∂ NARi* 1  = nsi R − Ri (6) ∂µ ti

which is positive if: nsi R > Ri or if si > Ri / nR . The small-budget teams will increase their demand for talent if their share in youth training is larger than their budget share in the league, and

The Economics of the Transfer Market

the large-budget teams will lower their demand for talent. So the competitive balance will improve without reducing total talent demand. In a simplified two-team league, which a large team (x) and a small team (y), the change in total talent demand can be found to be:

(

)

(

t y ns x R − Rx + t x ns x R − Rx

)

Ry

REFERENCES , because tx > ty and

nR ns x R − Rx = − ns y R − Ry . This youth training compensation system is just one specific revenue-sharing system. Given the peculiar characteristics of the industry of professional team sports, and the need of a reasonable CB, more sharing of revenue is necessary. As distinct from a profit-maximization league, it has been shown that revenue sharing in a win-­maximi­ zation league does improve the CB and increases the player salary level.

(

) (

system that is not linked to the transfer of players is set up. This will also improve the CB and create a strong incentive to invest in youth training, because teams will have to fight for their share of the training fund. This would have solved the dilemma of the ECJ in the 2010 Bernard case (see Késenne, 2011).

(7)

tx t y which is positive if s y >

209

)

CONCLUSION Many football insiders and outsiders are unhappy with the direction European football is heading: heavy losses and debts in most football teams, increasing child trade and transfers of very young players, huge transfer fees of top players in a quasi-closed money circuit of a handful of top teams, and a fast-growing gap, in both budget and performance of European top teams in large and small countries, locking out the majority of European teams from the European championships, which has de facto made European football a closed league. The abolition of all transfer fees with an adjusted youth training compensation system can cure many abuses and problems in football. It puts an end to the unethical and unwanted trade of players and the lack of transparency in player contracts, eliminating also the doubtful practices of money-grabbing player agents. It will improve the CB, increase the average player salary and correct the unfair player salary distribution. Players will be free to move in accordance with the EU competition laws, without discouraging youth training if an alternative youth training compensation

European Court of Justice (1995). The J.M. Bosman Case C-415/93, European Court reports 1995, p. I-04921, Luxemburg. Garcia-del-Barrio, P., & Szymanski, S. (2009). Goal! Profit maximization versus win maximization in soccer. Review of Industrial Organization, 34(1), 45–68. KEA-CDES (2013). The Economic and Legal Aspects of the Transfer of Players. Brussels: European Affairs. Retrieved from www.keanet.eu Késenne, S. (2010). The financial situation of the football clubs in the Belgian Jupiler League: are players overpaid in a win-maximization league? International Journal of Sport Finance, 5(1), 67–71. Késenne, S. (2011). Youth development and training after the Bosman verdict (1995) and the Bernard Case of the European Court of Justice. European Sport Management Quarterly, 11(5), 547–553. Késenne, S. (2014). The economic theory of professional team sports: an analytical treatment (2nd ed.). Northampton, MA: Edward Elgar. Lavoie, M. (2000). La proposition d’invariance dans un monde ou les equipes maximisent la performance sportive. Réflets et Perspectives de la Vie Economique, 39(2–3), 85–94. Neale, W. (1964). The peculiar economics of professional sports. The Quarterly Journal of Economic, 78(1), 1–14. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Scully, G. (1989). The business of Major League Baseball. Chicago, IL: University of Chicago Press. Szymanski, S. (2015). The economic arguments supporting a competition law challenge to the transfersystem. Retrieved from: www.fifpro.org/ attachments/article/6241/Embargoed%20 S t e f a n % 2 0 S z y m a n s k i % 2 0 Tr a n s f e r % 2 0 System%20Analysis.pdf

22 Team Production and Efficiency in Sports Mikael Jamil

INTRODUCTION Seminal contributions to literature in the economics of professional team sports, such as Rottenberg (1956) and Neale (1964), have emphasised the importance of a key externality known as the uncertainty of outcome hypothesis (UOH). The UOH was first introduced by Rottenberg (1956), who argued that, other things being equal, the closer the competition between two competing teams in any sporting event, the greater the interest in the event and therefore the greater the attendance and revenue, thereby illustrating the joint production element in sports. Neale (1964) further emphasised the importance of UOH and provided a boxing example now commonly known as the Louis–Schmeling paradox. Neale states that Joe Louis: wants to earn more money, to maximise his profits. What does he need in order to do so? Obviously a contender, and the stronger the contender the larger the profits from fighting him … since doubt about the competition is what arouses interest. … Pure monopoly is a disaster. Joe Louis would have no one to fight and therefore no income. (Neale, 1964, pp. 1–2)

In other words, Neale (1964) proposed that earnings for heavyweight champion Joe Louis would increase if he fought against an evenly matched opponent rather than a relatively weak contender. This quote from Neale (1964) also illustrates the joint nature of production in sports, where contests depend on the exertion of efforts and economic value is likely to increase in cases where greater joint effort occurs between two evenly matched opponents with comparable skills. In theory, public interest in sport and hence revenue and attendance will increase when teams are as closely competitive as possible, all other things equal. For example, close competition between two teams (team A and team B) in a league system would benefit the entire league as not only would attendance at fixture A v B increase but also attendances involving other teams in the league, i.e. teams C, D, etc. (hence the externality). On the other hand, domination of the league by team A would likely reduce interest and attendance at games involving the remaining teams. Although attendance in games involving team A may increase, in the long run team A would also suffer if the overall standard of competition declined. This chapter outlines the analysis of production in sport, and how efficiency is measured and evaluated.

Team Production and Efficiency in Sports

TEAM PRODUCTION AND SPORT Production in team sports requires the cooperation of at least two teams. This requirement that clubs cooperate to produce the product (the contest) is almost unique to sports and it is why Rottenberg (1956) first proposed the notion of a sporting production function when he defined the product (the contest) as a function of the players of one team (as factors of production) and the players of the other team (other factors of production) as well as additional factors such as team management, transportation and stadiums. Estimating production functions in team sports enables the examination of team success or failure and can help identify team and player strengths and weaknesses (Carmichael & Thomas, 2014). A production function analysis can thus inform team selection and its management and coaching in the short term by way of preparation of tactics and strategies within a match, over a tournament or league season. In the long term a production function analyses can inform management of squad requirements and the purchase and sale of players. The commercial relevance of statistical analyses for the purposes of monitoring player performance and informing team selection and player recruitment has been described by Lewis (2003) in his book entitled Moneyball and further publicised in the popular 2011 film of the same name. Two strands of literature exist with regards to team performance in sports. The first strand of research employs average production functions that model team outputs as a function of playing and non-playing inputs. The second strand of literature consists of the associated production frontier analyses and efficiency measurements that enable teams to assess whether they are making full use of their financial and sporting resources, relative to their potential (Carmichael & Thomas, 2014). Both strands of literature are reviewed below to illustrate the distinct agendas between the two strands of research in terms of assessing the relationship between team performance and production and team performance and efficient use of resources in team sports.

PRODUCTION FUNCTIONS IN SPORT Gerrard writes that: The core economic process within any professional sports team is the sporting production process in which ‘raw’ playing talent is transformed into

211

player and team performance and, in turn, match outcomes. Formally, this process can be modelled as a sporting production function. (Gerrard, 2001, p. 220)

One of the first empirical applications of production functions in sport was presented by Scully (1974), who investigated monopsonistic exploitation in baseball. Scully (1974, p. 915) aimed ‘to crudely measure the economic loss to the players due to the restrictions of the reserve clause’ and set out to accomplish this by obtaining estimates of the marginal revenue product (MRP) of pitchers and hitters performing in Major League Baseball (MLB) before comparing these estimates to player earnings in order to determine what percentage of a player’s MRP was retained by them. Scully (1974) proposed a causal chain (recursive) model consisting of two equations, the first of which directly determined the marginal effect of player performances upon team success and the second the marginal effect of team success upon team revenues. Scully (1974) adopted an Ordinary Least Squares (OLS) regression estimation method and applied it to both equations to estimate these marginal effects, which in turn enabled Scully to estimate player MRP. The first equation of Scully’s model consisted of the dependent variable PCTWIN (a calculation of the percentage of team wins), which Scully hypothesised to be a measure of team success, and several other independent variables. The independent variables included measures of player performance, such as the team slugging average (TSA) and team strikeout-to-walk ratio (TSW). TSA was hypothesised to be a measure of batting (attacking) prowess and measured the number of bases gained per match, whereas TSW measures a pitcher’s ability to control pitches and was therefore used as a measure of the pitching (defensive) contribution to team wins. In the second stage of Scully’s model, PCTWIN (now specified as an explanatory variable) and other independent variables that accounted for local population, team fan base, stadium location and racial prejudice were used to determine team revenues. Scully then used the estimated coefficients obtained for TSA and TSW (from the first stage) in conjunction with the estimated coefficient obtained for PCTWIN (from the second stage) to estimate average MRP. To convert these average MRP estimates to MRP per player, Scully assumed that the team’s performance was the sum of the individual’s performances. Furthermore, Scully’s estimations did not incorporate a team’s full roster but only 12 out of 15 hitters (that were expected to have played in a match) and 8 out of

212

THE SAGE HANDBOOK OF SPORTS ECONOMICS

10 pitchers (expected to have played in a match). Scully then presented ‘gross’ MRPs that did not account for costs associated with placing players on the field (i.e. cost of training players, cost of equipment and transport costs) before calculating ‘net’ MRPs after the deduction of these costs. Finally, Scully estimated player salaries that would have been paid to players (in various categories of skill) to determine whether players were remunerated in accordance with their MRP or if they were exploited. These estimates were obtained from separate salary regressions for 148 players (hitters and pitchers) performing in the 1968 and 1969 seasons. Scully’s (1974) results ultimately suggested baseball players were being exploited to a high degree, i.e. approximately 80%. Scully’s player performance model has been criticised by others. Medoff (1976), for instance, argues that Scully’s (1974) model was incorrectly specified as the recursive model did not account for simultaneity between win percentage and revenues in the two equations. Consequently, Medoff (1976) conducted a follow-up study utilising a Two Stage Least Squares estimator but revealed similar results that confirmed the exploitation of baseball players. Furthermore, Gerrard (2001) highlights the limitations of Scully’s model by claiming the model is only applicable in those team sports where comprehensive player performance data can be captured easily, such as batting and fielding games like baseball and cricket. In another study on baseball, Zech (1981) employed a Cobb-Douglas production function1 in order to identify those factors that contributed to team success in the MLB. Once identified, Zech empirically estimated the effect of these factors and used the results of the empirical testing to construct a measure of the league’s most valuable player (MVP). To develop this measure, Zech (1981) summarised player skills into four main categories: hitting, running, defence and pitching. Measures used to account for these player skills included players batting averages (as measures of hitting frequency), number of homeruns (as measures of player power), a team’s stolen base total (as a measure of team speed), a player fielding percentage (as a measure of catches taken), total chances taken (as a measure of difficult catches taken) and pitchers strikeout-to-walk ratio (as a measure of pitcher ability). In addition, Zech also incorporated two measures of head coach ability in the form of a manager’s lifetime won–lost percentage and the number of years managed in the major leagues. Having estimated a production function based on these factors, Zech (1981) computed each player’s marginal product by calculating each player’s

contribution to team victories. The top ten most valuable players’ (MVPs) in the league computed by Zech differed significantly from the official voting list of sportswriters (who selected MVPs for each season using an informal weighting system for each of the player’s different skills). Zech suggests that this discrepancy may have been due to player publicity and personality playing a role in the sportswriters’ weighting schemes. Zech (1981) estimated the model using data from 26 MLB teams competing in the 1977 baseball season and estimated four specifications that captured all possible combinations of defensive and managerial ability in order to ensure no multi-collinearity. Ultimately, Zech revealed that no measures of a team’s defensive skills (player fielding percentage and total chances taken) were significant and the contribution of the manager was also revealed to be insignificant. Furthermore, hitting was revealed to contribute almost six times as much to team success than pitching and the number of homeruns contributed about twice as much as the number of stolen bases to a team’s success. Ultimately, therefore, Scully’s (1974) methodology was adopted as the standard approach for much of the subsequent research in sporting production functions, the vast majority of which focused mainly on US-based sports (Scully, 1989, 1995). Aside from Medoff (1976) and Zech (1981), other studies on baseball include Porter and Scully (1982), Krautmann (1990) and Ruggiero, Hadley and Gustafson (1996). Studies of a similar nature focusing on basketball have been conducted by Zak, Huang and Siegfried (1979), Scott, Long and Somppi (1985), McCormick and Clement (1992) and Chatterjee, Campbell and Wiseman (1994). A study focusing on American football was conducted by Atkinson, Stanley and Tschiart (1988). US-based sports have been the focus of much of this research due mainly to data being publicly available, easily recorded and easily categorised into measurable match play statistics, including both team and individual player contributions (Carmichael, Rossi, & Thomas, 2017). In contrast, there has been a relative dearth of empirical research on other competitive sports in other countries due mainly to the lack of comprehensive data sets publicly available and the difficulty associated with recording, measuring and categorising team and player contributions, due to the increased interaction between team members in sports such as European football (soccer). Despite these difficulties, similar methods of modelling the sporting production process implemented by Scully (1974) and Zech (1981) have since been applied to other non-US sports. Schofield (1988) estimated production functions

Team Production and Efficiency in Sports

in cricket using data from the 1981 to 1983 cricket seasons in the County Championships and John Player League in England. As for some of the studies detailed above, Schofield specified output to be team success and a function of player performance and weather. Player performance (measured by batting, bowling, captaincy and fielding variables) was further composed of a subset of variables, namely: coaching skills, team management skills, player availability, scouting, form, experience, the quality of practice and training facilities. Schofield therefore adopted a recursive system in which endogenous variables of player performances and team success were sequentially determined. Schofield’s (1988) study involved collecting cross-sectional data from 17 competing teams and data was initially normalised to account for random variation in the mean caused by external factors that had an influential impact on input and output variables. Team success (dependent variable) was measured in two different ways: the first measure was the number of wins achieved throughout a single cricket season and the second measure was the number of points acquired. Schofield’s analysis was broken into three parts: the first part assessed the general importance of batting and bowling and the second and third parts assessed the relative importance of different types of batting and bowling skills. Batting performance was measured by calculating the number of runs scored per game and per over and the bowling variables included in the production function were the calculation of a bowling average (runs conceded/wickets taken), a composite measure of both attacking and defensive bowling skills. Furthermore, separate measures of attacking bowling, such as the number of wickets taken per game and the number of overs bowled per wicket taken, as well as measures of defensive bowling, such as the number of runs conceded per over, were also included as bowling inputs. Using these calculated values, Schofield (1988) developed a linear production function using OLS and ultimately concluded that, generally, bowling had a greater impact upon match success than batting. Bairam, Howells and Turner (1990) also constructed a cricket production function in their study on cricket played in Australia and New Zealand. Although this study was largely based on Schofield (1988), it differed in a few key areas. Output was defined as points percentage rather than number of wins or points achieved. Furthermore, two measures of bowling ability were utilised in this study the first was a measure of attacking bowling, balls per wicket, and the second was a measure of defensive bowling identified as the number of runs scored by the opposition.

213

Runs scored per wicket was another dimension that was included and reflected attacking and defensive bowling. Runs per innings and runs per over were two calculated values that represented the batting performance of teams. Bairam, Howells and Turner (1990) chose to omit the element of fielding to avoid multicollinearity. Pooled data were collected and normalised by taking ratios of the variables used to their seasonal means and multiplying the ratios by 100. This was done to account for random events such as differences in weather conditions that could have influenced both input and output variables. Bairam, Howells and Turner revealed that both batting and bowling variables were important in explaining match success. In order to maximise the probability of match success, Bairam, Howells and Turner (1990) concluded that New Zealand should adopt a combination of attacking batting and attacking bowling strategies. In a comparative analysis with England, it was revealed that, relative to England, batting was marginally more important than bowling for New Zealand. The results obtained for Australian cricket suggested the probability of match victory was maximised by adopting an attacking batting strategy and a defensive bowling strategy (Bairam, Howells and Turner, 1990). Carmichael and Thomas (1995) formulated a production function for rugby league football in the form of a recursive model in which both player performances and team success were determined sequentially. Team success (the output) was stated as a function of three measures of attacking performance – tries for, kicked goals for (conversions following a try or penalty) and drop goals for – and three measures of defensive performance – tries against, kicked goals against and drop goals against. Carmichael and Thomas assumed these six performance inputs were influenced by variables such as player fitness, inherent ability, player experience, team organisation and coaching skill. Output was measured by the percentage of games won during the season (expressed relative to the divisional average). Attacking performance was measured by the average number of tries, kicked goals and drop goals scored per game and defensive performance was similarly measured by the average number of tries, kicked goals and drop goals scored against. Carmichael and Thomas included height and weight variables as a proxy for player strength and measured experience by squaring the average player age. Inherent ability was proxied by the percentage of full international players and the percentage of players from overseas. The average number of appearances made per squad member measured team organisation and the ability of the coach was proxied by the

214

THE SAGE HANDBOOK OF SPORTS ECONOMICS

number of years spent as a full-time rugby coach prior to the season of investigation. Carmichael and Thomas (1995) estimated their model on cross-sectional data collected for 35 rugby league clubs competing in the 1990–1991 season. Data was collected for 1,214 individual players and team input and output measures were also calculated as seasonal totals. In conclusion, Carmichael and Thomas stated that successful performance in rugby league was more dependent on defensive performance than attacking performance and suggested a disaggregated data set with more player-specific data could potentially produce improved analysis of production functions in rugby league. Carmichael, Thomas and Ward (2000, 2001), Carmichael and Thomas (2005) and Carmichael, McHale and Thomas (2011) have conducted production function studies on English Premier League football (EPL). Carmichael, Thomas and Ward (2000) employed disaggregated player performance statistics for matches played by teams during a single season (1997–1998) to estimate a production function for match performances. The output was specified as the match score and was expressed as a goal differential as this measurement allowed for the easy interpretation of OLS estimates. The independent (input) variables consisted of various attacking and defensive play actions and team characteristics. Measures of play included: shots on target, percentage of successful passes, the number of tackles made, clearances, blocks, interceptions, dribbles won or lost, the number of free kicks conceded, yellow and red cards. Measures of the goalkeepers’ performance, such as the percentage of successful distributions and number of catches/drops, were also included as independent variables. Ultimately, the results obtained by Carmichael, Thomas and Ward (2000) emphasised the influence of attacking skills, such as accurate shooting and passing, as well as the importance of defensive skills as reflected in tackles, clearances and blocks made upon match outcomes. In a follow-up study, Carmichael, Thomas and Ward (2001) utilised aggregated match plays for each team over a full league season (1997–1998) to examine team performance and efficiency by estimating a season-based production function. Carmichael, Thomas and Ward adopted a multiple equation (recursive) system where output (overall team success) was expressed as percentage points. The results obtained from this analysis emphasised the importance of defensive aspects of play and identified shots on the opposition goal, goals scored and accurate passing to be of particular significance with regards to output determination.

Carmichael and Thomas (2005) use match play statistics from the 1997–1998 EPL season to examine the home field effects (advantages) on withinmatch performance of home and away teams. A variety of team play measures are included as inputs in a team production function in which the number of shots and goals scored by each team in each match are specified as outputs. The model utilised by Carmichael and Thomas incorporates a recursive system, with a team’s success in terms of match plays. Team play measures included actions such as dribbles, runs, passes, fouls, tackles, clearances, blocks and saves made by the goalkeeper. The regression results revealed that the statistical significance of attack-related team play measures was stronger for the home team and the significance of defence-related measures were stronger for the away team, implying that attacking play is more important at home and defensive play is more important away. Carmichael, McHale and Thomas (2011) study team performance in the EPL over the seasons 1998–1999 to 2004–2005 and investigate the link between playing success and commercial success. Employing a data set that combines financial measures and performance data, Carmichael, McHale and Thomas based their empirical analysis on three behavioural equations. The first equation is a sports production function where league success (output) was interpreted as a function of playing inputs (performance statistics) and nonplaying inputs, such as managerial contribution and home support. The second equation was a standard revenue equation in which revenue was specified as a function of output and managerial specific inputs. Equation three was a hedonic2 wage-price or earnings function that related player performances to wage expenditure. Carmichael, McHale and Thomas (2011) revealed on-field success to be directly related to player skills and abilities and revenue was positively influenced by on-field success. In a study on the Serie A in Italy, Carmichael, Rossi and Thomas (2017) estimated a production function for the league and the relative efficiency on the teams participating in it. Carmichael, Rossi and Thomas constructed a panel data set consisting of season aggregated statistics for all 36 participating clubs over the 10-season period from 2000/01 to 2009/10. This study included data from the seasons (and teams therein) affected by the Calciopoli corruption scandal. The authors used factor analysis to construct composite measures of direct inputs that reflected playing performance and incorporated this into their team production function. The estimating model consisted of a recursive system comprising two behavioural equations. Output was expressed as percentage

Team Production and Efficiency in Sports

points and other measures incorporated goal difference as well as specific team and managerial characteristics. Carmichael, Rossi and Thomas revealed that to obtain a high ranking in the Serie A, offensive performance is of greater significance than defensive performance. Home reliance and managerial changes were revealed to be negatively associated with league success in the Serie A. In a study on the EURO 2004 international football tournament, Carmichael and Thomas (2008) estimate an aggregated production function model for competing teams using a match play data set. Carmichael and Thomas constructed four play variables designed to capture the quantity and quality of attacking and defensive plays and included these variables as independent variables in a regression where the dependent variable was a team’s average goal difference throughout the tournament. In addition to the estimated average production function, Carmichael and Thomas also estimated a production possibility frontier in order to examine team efficiency. The use of production frontiers in order to measure team efficiency is also evident in literature on team performance. The next section describes the production frontier approach and presents the findings of some studies that have adopted this approach.

PRODUCTION FRONTIERS AND EFFICIENCY MEASUREMENTS IN TEAM SPORTS In contrast to production function studies that identify the contributing factors to success, two distinct approaches have been adopted to measure relative team efficiency. These are the econometric SFA3 (stochastic frontier) approach, which builds on concepts of regression analysis, and the deterministic non-parametric frontier approach (DEA),4 which consists of mathematical programming techniques. As for production function ana­ lyses, efficiency studies have covered match-level performance and seasonal team performance in leagues, hybrid cup tournaments as well as assessing managerial and coaching efficiency. Barros and Leach (2006a, 2007) use SFA to measure efficiency scores for 12 English Premier League teams between 1999 and 2003. The cost function consisted of three factor input prices (labour and two capital inputs) and three outputs (points, attendance and turnover). An average (time-invariant) efficiency score of 88% is reported, suggesting that the average club operating at maximum efficiency could reduce costs by around 12% without affecting outputs.

215

Similarly, Kern and Süssmuth (2005) use SFA to estimate team production functions for teams competing in the German Bundesliga. The ex-ante inputs (pre-seasonal estimates) of player and coach wages are transformed during the production process of a season into ex-post (actual) pecuniary revenues and sporting success. The output measure takes into account performance across numerous domestic and European competitions and they report a wide variation in team efficiency scores. In standard SFA, teams are assumed to have the same technological possibilities, although if technological possibility between teams differ, technical efficiency may be overstated. In order to overcome this inaccuracy in estimation, some studies employ a variant of SFA known as the random frontier model (Greene, 2005). The random frontier model accounts for heterogeneity as it separates technical inefficiency from technological differences between teams. Barros and Garcia-Del-Barrio (2008) compare the efficiency scores obtained for several cost function specifications (including a stochastic frontier model and a random frontier model) in their analysis of 12 teams competing in the EPL between 1999 and 2004. According to Barros and Garcia-del-Barrio (2008), the random frontier model is a better representation of the data than the standard SFA. In further studies on football, Barros, Del-Barrio and Leach (2009) employ random frontier models to estimate efficiency scores for teams competing in the Spanish La Liga (top tier) between 1996 and 2005. Furthermore, Frick and Lee (2011) estimate time varying efficiency scores for German football teams between 1982 and 2003, and they reveal that on average, relegated teams suffer a considerable drop in technical efficiency compared to the previous season. Deterministic frontiers have also been employed to estimate production and cost frontiers based on team sports data, fitted using the DEA linear programming technique (Dobson & Goddard, 2011). Studies on North American sports utilising this methodology include Mazur (1994), Anderson and Sharp (1997) and Volz (2008) for the MLB and Einolf (2004) for the NFL and MLB. In football, Haas (2003b) uses DEA to measure the productive efficiency of EPL teams during the 2000–2001 season using team input measures such as team payroll and coach salary (as proxies for playing talent and manager contribution) and outputs such as league points and revenues. DEA frontiers were fitted based on both constant returns to scale (CRS) and variable returns to scale (VRS) assumptions concerning the production technology. Haas (2003b) concluded that wages were the main driver of team/club performance both on and off the pitch.

216

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Using EPL data from the seasons 1998–1999 to 2002–2003, Barros and Leach (2006b) combined sporting and financial variables in order to estimate a DEA cost frontier. Outputs were league points, turnover and attendances and inputs consisted of the number of players, net assets, payroll and expenditure on the stadium. Efficiency scores for EPL teams were estimated using both CRS and VRS specifications and Barros and Leach (2006b) revealed mixed efficiency scores for teams. Haas, Kocher and Sutter (2004) estimated a DEA frontier using data on the German Bundesliga for the 1999–2000 season. Player and manager payroll were used as input measures and outputs were league points, total revenues and a ratio of attendance to stadium capacity. CRS and VRS specifications were estimated and Haas, Kocher and Sutter (2004) revealed efficiency scores were not correlated to league position, indicating performance and efficiency are not synonymous. In a study on American Major League Soccer (MLS), Haas (2003a) estimated a DEA frontier using data from the 1999–2000 season. Input measures included player and coach payroll and the outputs consisted of league points, absolute number of spectators and revenues. Efficiency scores were revealed to be highly correlated with league performance. When inefficiency was decomposed into technical inefficiency and scale inefficiency, it was shown that the suboptimal scale of production largely explains inefficiency. Espita-Escuer and Garcia-Cebrian (2006) used DEA to measure the efficiency of teams in the conversion of attacking moves during a match into sporting success by applying CRS and VRS specifications to Spanish La Liga data from the 1998–1999 season to the 2000–2001 season. Input variables included the number of players used, the number of attacking moves, the number of minutes during which the teams had possession of the ball and the number of shots and headers (all measured throughout the course of the season). The output was the number of league points achieved throughout the season. The results revealed that efficient teams did not always correspond with those that finish highest in the league, therefore implying that highly placed inefficient teams could have achieved the same results with fewer resources or could have achieved more with the same resources (Dobson & Goddard, 2011). Bosca, Liern, Aurelio and Ramon (2009) used DEA to compare attacking and defensive efficiency in Italian and Spanish football between the three seasons from 2000–2001 to 2002–2003. Attacking inputs were shots on target, attacking plays created, balls kicked into the opposing penalty area and minutes of possession. Defensive

inputs were the inverse of the attacking inputs. The outputs for attacking and defensive production were the number of goals scored and the number of goals conceded. Bosca et al. revealed teams in Spain were more closely matched in terms of ability than teams in Italy. In the Italian league, the best-ranked teams scored more goals, conceded fewer goals and, on average, obtained more points. In Spain, the correlations between these indicators were revealed to be smaller. Furthermore, Bosca et al. (2009) revealed league rankings were highly correlated with measures of defensive efficiency in Italy, whereas the opposite was true for Spain.

CONCLUSION As demonstrated above, one of two approaches has generally been adopted for the measurement of team performance in football: the production function approach and the production frontier approach. The literature reviewed above highlights the general lack of attention awarded to the opponent and their impact upon observed performances. In previous literature specific to production in football, Carmichael, Thomas and Ward (2000) and Carmichael and Thomas (2005) both include variables accounting for opponent interactions, highlighting their importance in the production process, although only at a team level. Even at a more disaggregated level, studies that have analysed technical and physical aspects of play in football have generally overlooked the impact of the opponent. Out of 44 articles reviewing technical ability in football, Mackenzie and Cushion (2013) critically reveal that only eight studies included the opposition in their analysis. A similar pattern was revealed for performance analysis research concerning the physical demands of football, where out of 15 articles reviewed only three acknowledged the opposition. Future research could focus on the impact of the opponent’s performance on an observed team’s performance at a more disaggregated level (during matches and in-between matches).

Notes 1  A Cobb-Douglas function is a common function used by economists when turning to empirical work (Gartner, 2006). The function is used to describe the relationship between two or more inputs (such as capital and labour) and the amount of output produced.

Team Production and Efficiency in Sports

2  Hedonic models are often used to value assets by breaking them down into their component parts and obtaining estimates of the contributory value of each component (Brooks, 2008). 3  SFA is a regression based approach used to evaluate performance efficiencies and it is regarded as an alternative to DEA. SFA decomposes the usual error term into two components, an inefficiency component and a random component, that measure things like measurement error and environmental influences and is therefore useful for identifying factors that influence performance (Cooper, Seiford, & Tone, 2007). 4  DEA is used for evaluating performance of various kinds of entities engaged in different activities in different countries. DEA involves mathematical programming techniques that can handle large numbers of variables and relationships between these variables. This relaxes the requirements that one encounters when having to choose limited inputs and outputs because the techniques employed otherwise encounter difficulties (Cooper, Seiford, & Tone, 2007).

REFERENCES Anderson, T. R., & Sharp, G. P. (1997). A new measure of baseball batters using DEA. Annals of Operations Research, 73(0), 141–155. Atkinson, S. E., Stanley, L. R., & Tschirart, J. (1988). Revenue sharing as an incentive in an agency problem: an example from the National Football League. The RAND Journal of Economics, 19(1), 27–43. Bairam, E. A., Howells, J. M., & Turner, G. M. (1990). Production functions in cricket: the Australian and New Zealand experience. Applied Economics, 22(7), 871–879. Barros, C. P., Frick, B., & Passos, J. (2009). Coaching for survival: the hazards of head coach careers in the German ‘Bundesliga’. Applied Economics, 41(25), 3303–3311. Barros, C. P., & Garcio-del-Barrio, P. (2008). Efficiency measurement of the English Football Premier League with a random frontier model. Economic Modelling, 25(5), 994–1002. Barros, C. P., & Leach, S. (2006a). Analyzing the performance of the English FA Premier League with an econometric frontier model. Journal of Sports Economics, 7(4), 391–407. Barros, C. P., & Leach, S. (2006b). Performance evaluation of the English Premier League with data envelopment analysis. Applied Economics, 38(12), 1449–1458. Barros, C. P., & Leach, S. (2007). Technical efficiency in the English Football Association Premier League

217

with a stochastic cost frontier. Applied Economics Letters, 14(10), 731–741. Barros, C.P., Garcia-del-barrio, P., & Leach, S. (2009). Analysing the Technical Efficiency of the Spanish Football League First Division with a Random Frontier Model. Applied Economics. 41(25), pp. 3239–3247. Bosca, J. E., Liern, V., Aurelio, M., & Ramon, S. (2009). Increasing offensive or defensive efficiency? An analysis of Italian and Spanish football. Omega: The International Journal of Management Science, 37(1), 63–78. Brooks, C. (2008). Introductory econometrics for finance. Cambridge, UK: Cambridge University Press. Carmichael, F., McHale, I., & Thomas, D. (2011). Maintaining position by investing in human capital: the case of the Premier League. Bulletin of Economic Research, 63(4), 464–497. Carmichael, F., & Thomas, D. (1995). Production and efficiency in team sports: an investigation of rugby league football. Applied Economics, 27(9), 859–869. Carmichael, F., & Thomas, D. (2005). Home-field effect and team performance: evidence from English Premiership football. Journal of Sports Economics, 6(3), 264–281. Carmichael, F., & Thomas, D. (2008). Does the best team win? An analysis of team performances at EURO 2004. European Sport Management Quarterly, 8(3), 211–228. Carmichael, F., & Thomas, D. (2014). Team performance: production efficiency in football. In J. Goddard & P. Sloane (Eds.), Handbook on the economics of professional football (pp. 143–165). Cheltenham, UK: Edward Elgar. Carmichael, F., Rossi, G. G., & Thomas, D. (2017). Production, efficiency, and corruption in Italian Serie A football. Journal of Sports Economics, 18(1), 34–57. Carmichael, F., Thomas, D., & Ward, R. (2000). Team performances: the case of English Premiership football. Managerial and Decision Economics, 21(1), 31–45. Carmichael, F., Thomas, D., & Ward, R. (2001). Production and efficiency in Association Football. Journal of Sports Economics, 2(3), 228–243. Chatterjee, S., Campbell, M. R., & Wiseman, F. (1994). Take that jam! An analysis of winning percentage for NBA teams. Managerial and Decisions Economics, 15(5), 521–555. Cooper, W. W., Seiford, L. M., & Tone, K. (2007). Data envelopment analysis: a comprehensive text with models, applications, references and DEASolver software (2nd ed.). New York: Springer Science and Business Media LLC. Dobson, S., & Goddard, J. (2011). The economics of football. Cambridge, UK: Cambridge University Press.

218

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Einolf, K. W. (2004). Is winning everything? A data envelopment analysis of Major League Baseball and the National Football League. Journal of Sports Economics, 5(2), 127–151. Espitia-Escuer, M., & Garcia-Cebrián, L. I. (2006). Performance in sports teams: results and potential in the Professional Soccer League in Spain. Management Decision, 44(8), 1020–1030. Frick, B., & Lee, Y. H. (2011). Temporal variations in technical efficiency: evidence from German soccer. Journal of Productivity Analysis, 35(1), 15–24. Gartner, M. (2006). Economics – Macroeconomics (2nd ed.). Harlow, UK: Pearson. Gerrard, B. (2001). A new approach to measuring player and team quality in professional team sports. European Sport Management Quarterly, 1(3), 219–234. Greene, W. (2005). Fixed and random effects in stochastic frontier models. Journal of Productivity Analysis, 23(1), 7–32. Haas, D. J. (2003a). Financial efficiency in major league soccer. Journal of Sport Economics, 4(3), 203–215. Haas, D. J. (2003b). Productive efficiency of English football teams: a data envelopment analysis approach. Managerial and Decision Economics, 24(5), 403–410. Haas, D. J., Kocher, M. G., & Sutter, M. (2004). Measuring efficiency of German football teams by data envelopment analysis. Central European Journal of Operations Research, 12(3), 251–268. Kern, M., & Süssmuth, B. (2005). Managerial efficiency in German top league soccer: an econometric analysis of club performances on and off the pitch. German Economic Review, 6(4), 485–506. Krautmann, A. C. (1990). Shirking or stochastic productivity in Major League Baseball? Southern Economic Journal, 56(4), 567–579. Lewis, M. (2003). Moneyball: the art of winning an unfair game. New York: W.W. Norton. Mackenzie, R., & Cushion, C. (2013). Performance analysis in football: a critical review and implications for future research. Journal of Sports Sciences, 31(6), 639–676. Mazur, M. (1994). Evaluating the relative efficiency of baseball players. In A. Charnes, W. Cooper, A. Lewin, & L. Seiford (Eds.), Data envelopment analysis: theory, methodology, and applications (pp. 369–391). Boston, MA: Kluwer.

McCormick, R. E., & Clement, R. C. (1992). Intra-firm profit opportunities and managerial slack: evidence from professional basketball. In G. W. Scully (Ed.), Advances in the economics of sports (Vol. 1, pp. 3–35). Greenwich, CT: JAI Press. Medoff, M. (1976). On monopsonistic exploitation in professional baseball. Quarterly Review of Economics and Business, 16(2), 113–121. Neale, W. C. (1964). The peculiar economics of professional sports: a contribution to the theory of the firm in sporting competition and in market competition. The Quarterly Journal of Economics, 78(1), 1–14. Porter, P. K., & Scully, G. W. (1982). Measuring managerial efficiency: the case of baseball. Southern Economic Journal, 48(3), 642–650. Rottenberg, S. (1956). The baseball players’ labor market. The Journal of Political Economy, 64(3), 242–258. Ruggiero, J., Hadley, L., & Gustafson, E. (1996). Technical efficiency in Major League Baseball. In J. Fizel, E. Gustafson, & L. Hadley (Eds.), Baseball economics: current research (pp. 191–200). Westport, CT: Praeger. Schofield, J. A. (1988). Production functions in the sport industry: an empirical analysis of professional cricket. Applied Economics, 20(2), 177–193. Scott, F. A. J., Long, J. E., & Somppi, K. (1985). Salary vs. marginal revenue product under monopsony and competition: the case of professional basketball. Atlantic Economic Journal, 13(3), 50–59. Scully, G. W. (1974). Pay and performance in Major League Baseball. American Economic Review, 64(6), 915–930. Scully, G. W. (1989). The business of Major League Baseball. Chicago, IL: University of Chicago Press. Scully, G. W. (1995). The market structure of sports. Chicago, IL: University of Chicago Press. Volz, B. (2008). Efficient production of wins in Major League Baseball. Economics Working Papers, No. 200850. Zak, T. A., Huang, C. J., & Sigfried, J. J. (1979). Production efficiency: the case of professional basketball. Journal of Business, 52(3), 379–392. Zech, C. E. (1981). An empirical estimation of a production function: the case of Major League Baseball. American Economist, 25(2), 19–30.

23 Officials and Home Advantage J. James Reade

INTRODUCTION Economics, the study of outcomes given a set of decisions regarding scarce inputs, is interested in systematic patterns in outputs. Sport provides a range of outcomes, or outputs, that are very measurable, and attract great interest. As well as the participants in a contest, be they individuals or teams, contributing to its outcome, officials also contribute. Officials are agents responsible for regulating activity to ensure that participation is fair. Such officials are granted power to penalise contestants who do not act fairly. Officials operate usually on two levels: a team of officials who operate on the field during a contest, and sporting associations, or governing bodies, who operate more broadly, stipulating rules and regulations to be enforced. One persistent pattern in sporting outcomes is the home advantage; participants performing at their home venue win more often than they ought to, given their strength relative to visiting participants. If two identical competitors play against each other, the team playing at home would win more than 50% of the time. Pollard and Pollard (2005) identify its existence over more than a century. In this chapter we explore the various competing mechanisms proposed for its existence

and pay particular attention to explanations that involve officials. We illustrate the results found in the literature by analysing a dataset hitherto not before used in the analysis of home advantage, namely every single cricket match recorded on the Cricket Archive website between 1731 and 2017. We then introduce the concept of home advantage in more detail and outline a method of identifying it using data. We also review the key papers on the subject of home advantage, using cricket data to illustrate a number of the points, before reaching some conclusions.

THE CONCEPT OF HOME ADVANTAGE We define home advantage as the home team winning more often than it ought to, given the relative ability of the teams competing. Provided that the ability of participants can be measured, home advantage can be identified. We first define the outcome of event i to be yi, and conventionally we may think about yi as being zero for an away team victory, one for a home team win, and some third value (say, a half) for a draw:

220

THE SAGE HANDBOOK OF SPORTS ECONOMICS

 1 if home team wins,   yi =  0.5 if match drawn,  0 if away team wins. 

(1)

As such, we think about events from the perspective of the home team, enabling a relatively straightforward recognition of home advantage. Next, we define the relative strength of the home team to be a function of the strengths of the two teams competing, xi = f (xH,i, xA,i), where 0 ≤ xi ≤ 1, and, as with outcome, we think of xi → 1 reflecting a relatively stronger home team, and xi → 0 a relatively weaker home team. We can then run a linear regression model of:

yi = α + β xi + ei .(2)

Departures from α = 0 and β = 1 enable the identification of home advantage, since if both conditions are imposed, then outcomes are exactly proportional to the relative strengths of the two participants: an absence of any home advantage. This approach equates to a Mincer and Zarnowitz (1969) regression test on the forecasts produced by the relative strengths of the two teams competing. The Mincer–Zarnowitz approach is commonly used to evaluate the efficiency of forecasts: whether they incorporate all information available at the forecast origin. It is perhaps easiest to think about the implied relationship between relative strengths xi and outcomes yi in a two-dimensional set of axes, with outcomes on the vertical axis and relative strengths on the horizontal axis. Three examples of such a cross plot are provided in Figure 23.1. The absence of home advantage would see points in the axes congregated around the 45-degree line, with outcomes exactly proportional to the relative strength of teams. This is the example in the leftmost panel of Figure 23.1; over the entire range of relative team strengths, from zero to one, the outcome is perfectly predictable based on this information alone. A simple home advantage would see the points all lying above the 45-degree line, implying that for any set of relative strengths, the home team wins more often than they should. This is the middle panel of Figure 23.1, where for all ratios of team strengths, the home team wins more often than a pure strength difference would predict. Equally, all points below the 45-degree line would be indicative of a home disadvantage. Patterns in home advantage can be more complicated if the points intersect the 45-degree line. If β < 1, the slope is less than 45 degrees and home advantage favours weaker teams, as weaker teams win more often than their relative strength

suggests they should, but stronger teams win less often. Conversely, if β > 1 it favours stronger teams. The point at which the points cross the 45-degree line also provides information on the nature of the implied home advantage. As such, the Mincer– Zarnowitz method allows a rich characterisation of home advantage. Nothing guarantees that either slope will intersect the 45-degree line, but if it does, then this yields the point where a home advantage becomes a disadvantage, and vice versa. The third panel of Figure 23.1 provides an example of a home advantage favouring weaker teams, as β < 1 hence the slope is flatter than the 45-degree line and points below about 0.5 are above the line, indicating that weaker home teams (with less than a half share of quality) win more often than their relative strength implies, while above 0.5 points are below the 45-degree line, and hence relatively strong home teams win less often than they should. It might be argued that we are expecting a lot of our xi variable, that of relative team strengths; namely that it ought to be an efficient predictor of outcomes. Sports often have inbuilt mechanisms that promote competitive balance, which imply that the strongest teams will win less than their relative quality implies. Our method identifies such mechanisms only in so far as they facilitate weaker home teams winning more (less) often than their quality suggests, rather than just weaker teams per se winning more (less) often. We adopt (1) as our dependent variable when measuring home advantage in this chapter.1 Koop (2004) considers multinomial and ordered models for baseball, Vlastakis, Dotsis and Markellos (2009), Goddard and Asimakopoulos (2004), and Goddard (2005) for football, and Scarf and Shi (2005) and Akhtar and Scarf (2012) for cricket. In general, it might be argued that the distinction between a home or away win has no obvious ordering, and hence that a multinomial model is appropriate. However, when the metric is the relative performance of the home team, then the ordering is clear: away win, draw, home win. Thus, we measure home advantage at eventlevel detail. Many studies calculate an average at a sports-league-seasonal level: the percentage of matches won by the home team in all matches during a season. The implicit argument is that relative team quality is fixed throughout a season, and a balanced schedule exists where each team plays every other team once at home and once away. If this is not the case (for example, if teams can alter playing squads mid-season, if teams or players develop during a season, or if the schedule is not balanced), then it may be that such an aggregation provides a biased measure of home

221

Officials and Home Advantage

No Home Advantage

Home Advantage

Home Advantage for Weaker Teams

0.8 0.6 0.4 0.2 0.0

1.0 Proportion of outcomes

1.0 Proportion of outcomes

Proportion of outcomes

1.0

0.8 0.6 0.4 0.2 0.0

0.0

0.2

0.4

0.6

0.8

Relative strength of participants

1.0

0.8 0.6 0.4 0.2 0.0

0.0

0.2

0.4

0.6

0.8

1.0

Relative strength of participants

0.0

0.2

0.4

0.6

0.8

1.0

Relative strength of participants

Figure 23.1  Stylised examples of home advantage expressed in terms of relative team strengths and recorded (average) outcomes

advantage α since it implies the restriction β = 1. Unbalanced schedules are very common in sports leagues; Pollard and Pollard (2005) note various unbalanced schedules that have existed for North American sports, Lenten (2008) considers the impact of unbalanced schedules on conventional measures like competitive balance in the top division of Scottish football, and Lenten (2011) investigates unbalanced scheduled in Aussie Rules Football. Few, if any, cricket tournaments consist of balanced schedules. In their subsequent analysis of golf and tennis tournaments, Holder and Nevill (1997) control for the strength of contestants in a competition in order to identify home advantage. If we define home advantage to be home teams winning more often than their ability suggests, then controling for team strength is essential. The xi variable of relative strength needs to be constructed. A common strategy to do this is to use betting odds since these contain information on the expected outcome of a match (see, for example, Forrest, Goddard and Simmons (2005) on the information efficiency of betting odds, but also Sobel and Ryan (2008) on the known biases of betting odds). However, bookmaker odds only exist so far back, historically, and only for matches in a small fraction of competitions that exist even in recent history, and hence in order to be able to measure relative quality more broadly we need an alternative measure of the strength of teams. As we consider teams across a wide range of countries and competitions, we employ Elo rankings.2 Elo rankings (Elo, 1978) provide a strength rating for each participant in a contest, and were developed for the purpose of ranking chess players. These strength ratings are then used to create predictions of future event outcomes. Each team has a strength, xi,H for the home team, and

xi,A for the away team. Conventionally, a team has strength 1000 in its first match.3 For match i, a prediction of the outcome can be generated in terms of the expected score for each team, according to the formula:

xi =

1 1 + 10

( xi , A − xi , H )/400

.(3)

Hence xi is symmetric in relative team strength, and 0 < xi < 1, where the extreme cases represent one infinitely strong team against an infinitely weak team, and in the case of two equally matched teams, xi,H = xi,A, then xi = 0.5. As such, Elo predictions provide an ideal measure of relative strengths of teams. In our cricket dataset, used to illustrate issues surrounding home advantage, xi ranges between 0.014 and 0.985, and as such spans the unit interval as desired. In practice, the largest difference between two teams in our dataset is around 700. We use the outcome variable yi as in (1) to update the Elo strengths for each team by the formula: xi+1, j = xi , j + K ( yi − xi ),

j = H , A.(4)

Here, K is chosen to determine the sensitivity of the ratings to each individual match. A range of values have been used for this. We simply use K = 40. There is no explicit role for the existence of home advantage in calculating either the strengths, or event predictions, but both will reflect home advantage implicitly, if participants win more often when playing at home. Assuming that teams play a roughly proportionate number of matches at home as away, however, their Elo strength should not reflect home advantage, and the Elo prediction should reflect the actual quality of teams.

222

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Our interest is in the size of the home advantage, which is dictated by the α and β coefficients exemplified in Figure 23.1. In an ordered probit/logit model, however, the constant and slope terms are somewhat trickier to interpret. Given that ordinary least squares estimation of (2), the linear probability model, will still yield unbiased estimates in the presence of a discrete dependent variable, and given that it is most intuitive to understand, we use this method. We present some results as equations, and in other places comment on the value of the α and β coefficients, and also plot the implied regression lines.

ILLUSTRATING HOME ADVANTAGE In this section, we highlight the historic persistence of home advantage. We look at data for cricket going back to the 18th century and consider salient contributions from elsewhere in the literature. The Cricket Archive website4 provides scorecards for cricket matches played throughout the world of cricket, from games at the very top of the game down to regional leagues in cricketplaying nations. Most previous studies consider a single sports league over a period of time. Rather than restrict attention, we consider all recorded cricket matches, which covers 11,441 leagues and cups in 109 countries, as well as a huge number of ad hoc matches, such as tour matches. There are men’s and women’s matches, youth, full, and also veterans’ matches. In total, Cricket Archive has 610,208 cricket matches that we can use to conduct our analysis of home advantage. Naturally, the number of matches has increased over the years; there are 379 recorded scorecards from the 18th century, 29,961 from the 19th century, 178,061 from the 20th century and 401,809 in the 21st century. Cricket before the 19th century was generally played on an ad hoc basis, but during the course of the 19th century Test cricket between cricketing nations and domestic leagues began to form, and this development continued into the 20th century. The first step is to establish the existence of home advantage. Previous studies of home advantage have considered a wide range of sports and sports leagues. Schwartz and Barsky (1977) consider baseball, basketball, American football and ice hockey in 1971, and Pollard and Pollard (2005) add English football and a substantially larger set of years to their study. Schwartz and Barsky (1977) find home advantage to be strongest in basketball and hockey, and weakest in baseball and American football. Pollard and Pollard (2005)

show that home advantage has been declining across all sports except baseball and American football, where the latter shows a great degree of volatility. As such, the distinctions observed by Schwartz and Barsky (1977) have largely disappeared, although baseball still does retain the largest (and most consistent) advantage. Gómez, Pollard and Luis-Pascual (2011) investigated nine team sports in Spain: baseball, basketball, handball, indoor soccer, roller hockey, rugby, soccer, volleyball, and water polo. They found a significant home advantage in all, although it was strongest in rugby. The outcomes in our dataset do not reflect a balanced schedule, so as Holder and Nevill (1997) note, we need to control for team strength. In Figure 23.2 we plot the annual average forecast error from using Elo predictions to forecast outcomes of cricket matches; that is, ei = yi − xi from (1) and (3). If this is positive, then home teams win more often than their relative strength suggests they should, and if negative the home team wins less often, on average, than it ought to, based on relative strengths. The black line is for all matches, while the grey line is for recognised Test matches between national teams, played on non-neutral venues. The black line for all matches shows greater variance prior to 1850, reflecting the smaller number of matches in those years, and settles down at around 3% after this. This reflects that home teams win about 3% more often than their relative strength implies – a home advantage. In the post-war period, it begins to fall, and after 1979 is negative, suggesting that since the late 1970s there has been no home advantage in cricket. This is consistent with Pollard and Pollard (2005), who found that across a range of major team sports, home advantage was strongest in the earlier years of their existence. The reduction particularly since the late 1970s is consistent with the home advantage pattern in English football that Koyama and Reade (2009) identify: in football, since the late 1980s home advantage has fallen. They attribute that to television, which enabled football spectators to better monitor the effort levels expended by players in away matches. Test matches on non-neutral venues are also plotted in grey in Figure 23.2, as they represent what might be regarded as the most high-profile matches in cricket, but they also represent matches between countries taking place in one of the two countries, hence a natural setting to observe home advantage. There are 11,637 matches between two country teams in our dataset, a further 2,731 matches where the home team is not a national team but the visiting team is (usually a tour match), and 2,469 where the away

Officials and Home Advantage

223

Figure 23.2  Relative proportion of matches ending in wins for the favourite team (according to Elo ranking) team is not a national team but the home team is. Importantly for identifying home advantage, of those 14,368 matches with a national team listed first, 9,500 were played in that country, meaning that the remaining 4,868 were played on neutral territory. Of those played on neutral territory, there is no home advantage. We can say that historically home teams have won 8% more often than away teams of identical relative quality in cricket. As such, we are able to identify a strong home advantage after controlling for team ability, and accounting for the fact that many matches are played on neutral territory. This is reflected also in the grey line in Figure 23.2, which plots the forecast error for Test matches in non-neutral venues and has averaged 5% since the interwar period.5 This is indicative of a strong and persistent home advantage, as home teams are 5% more likely to win on average than their relative strength suggests they should. This is not particularly interesting without an explanation. Studies into home advantage have proposed explanations that can be classed under four broad headings, Nevill and Holder (1999) suggest: crowd, learning, travel and rules. Crowdbased mechanisms tend to focus on the size of the crowd in attendance and its impact on outcomes, focusing on two mechanisms: influencing player performance via encouragement, or influencing the actions of officials. As such, we bundle consideration of the crowd with rules-based explanations. Under learning, familiarity with local conditions (pitch size, stadium, climate/geographic factors) is important. The fatigue associated with travel,

and the disruption to normal routines, is argued to influence performance and hence may well contribute to home advantage. Finally, it may be that rules influence outcomes in favour of home teams, not least by mitigating the potential impact of travel on visiting participants. We thus also bundle consideration of travel in with rules-based explanations. In the next two subsections we explore these explanations.

Familiarity Familiarity with the venue at which a match is being played is argued to matter for outcomes and can certainly be influenced by officials. The dimensions of the playing area, the facilities and the local area may all play some part consciously or otherwise. Pollard and Pollard (2005) note that after the Second World War home advantage was significantly lower in English and Italian football, when home players would have had less familiarity given a long break. Pollard (2005) found that teams moving to a new stadium in baseball, basketball and ice hockey in North America lost 24% of the advantage associated with playing at home initially. In cricket, there are plenty of reasons to suspect that familiarity plays a role. Schwartz and Barsky (1977) list a set of reasons from baseball that are readily applicable: the ground staff may prepare a pitch that favours home players, and the size of the playing area ensures that players

224

THE SAGE HANDBOOK OF SPORTS ECONOMICS

must be familiar with its dimensions and character. All of these could be factors in why we observed a significant home advantage earlier. However, there is no significant postwar effect on home advantage, nor is there much experience with new stadiums to compare with other sports. County teams in England do play at a number of venues, conventionally. Counties play the majority of their matches at their primary venue, but a proportion of matches at one or more out fields, that is smaller grounds elsewhere in the county. Attendance analyses (Schofield, 1983; Paton & Cooke, 2005) have suggested that such out fields do attract greater crowds, not least because they are often associated with festivals and are one-off annual events. Nonetheless, they represent alternative venues from the ones that home players should be most familiar with. Inserting a dummy for matches at such out fields yields marginally significant results, suggesting that, indeed, home advantage is slightly smaller where familiarity is lower. But the effect of any lack of familiarity will be confounded with larger crowds that such matches attract. Familiarity must vary with the length of matches. Cricket matches vary from the short to the very long; historically, even timeless matches existed (1,339 such matches are recorded), but the most common match is a single-innings one-day match (more than half all observations). There are 61,879 three-day matches, 41,201 two-day matches, and 20,234 four-day matches. Conventionally, Test matches are five days in length, and there are 2,610 such length matches recorded.6 With a multipleday match, visiting teams get more opportunity to familiarise themselves with local conditions, and as such we might anticipate that home advantage is smaller in longer matches. Equally, of course, it may be that by the time familiarity has been acquired, the outcome of the game has already been determined due to the cumulative nature of scoring. Shorter cricket matches are distinguished by being played over limited overs (a fixed limit on the number of balls to be bowled per innings), rather than an innings continuing until a team has been bowled out (or the batting team declaring). In Figure 23.3 we plot outcomes of cricket matches organised by the relative quality difference of the teams involved, as measured by the Elo prediction (3). This is expressed along the horizontal axis, from 0 (visiting team infinitely stronger than home team) to 1 (home team infinitely better than visiting team). In the absence of home advantage, two equally matched teams would win repeated contests 50% of the time, and we should expect points in Figure 23.3 to be congregated around the 45-degree line. We plot the line of best fit for the points plotted.7 We find that home advantage

is reduced in limited-overs matches. The dashed line is limited overs matches, and the solid line is unlimited overs matches. The limited-overs line has a steeper slope, and a smaller vertical intercept, indicating a smaller home advantage. The two lines pivot around 0.5, meaning that at the upper end of the scale where the home team is much stronger, outcomes are closer to what quality differences imply they should be. Hence home advantage is stronger in longer matches, where visiting teams have more time to become familiarised with a venue. This suggests that familiarity may not be an important factor in determining home advantage, but it also may indicate that by the time familiarity has been acquired, the match is already too far developed for the outcome to be influenced.8 In the last half century or so, the number of concurrent tournaments taking place in countries has led to some teams playing each other in different formats at very similar points of the season. For example, in England the advent of a one-day tournament on Sundays in 1969 added to an already busy weekly calendar of three-day matches starting on Saturdays and Wednesdays (with a break on Sunday, traditionally). Hence it was often the case that two teams would begin a three-day match on a Saturday, then on a Sunday play a one-day match at the same venue, and on Monday resume the three-day match. If familiarity plays a role in home advantage, it should be that in such one-day matches the home advantage is reduced. More generally, we might anticipate that home advantage is increasing in the number of days between meetings between teams at a particular venue. Inserting a variable measuring the number of days between meetings between teams at a particular venue, the effect is opposite to that hypothesised: the more days that elapse, the smaller is home advantage. Considering one-day games that begin on a Sunday between the same two teams that started on the Saturday, we find that for both the one-day and three-day games paired together like this, home advantage is significantly stronger than in standard matches. This would appear to argue against familiarity as a cause of home advantage, since, in particular, the one-day match begins the day after a visiting team has had ample opportunity to become familiar with the ground (certainly relative to if the one-day match did not begin the day after a three-day match had started). Finally, it may be that familiarity is with weather and climatic conditions rather than physical surroundings. There are anecdotal stories of particular sporting venues with very particular weather conditions (lots of rain, very cold), but also about the local geography. McSharry (2007) investigated the impact of playing football at high

Officials and Home Advantage

225

Figure 23.3  Cricket match outcomes by relative quality Note: Plot of home advantage regression result from all cricket matches, grouped by whether a match is limited in overs or not. Outcomes are summarised by the relative quality difference of a match as measured by the Elo prediction in (3). The further rightwards along the horizontal hypothesis, the stronger is the home side and hence the more likely is a home victory

altitude in response to claims that Bolivia gained from an unfair advantage in South American football matches, finding that indeed altitude mattered.9 Conversely, Chumacero (2009) find that altitude does not matter for determining home advantage, although heat and humidity do.

Rules The purpose of rules is to make contests as fair as possible between the teams participating, and hence, in principle, should mitigate against home advantage. The purpose of officials is to ensure the rules are implemented such that contests are fair. Hence officials ought to mitigate against home advantage. Nonetheless it is possible that at times rules have failed to mitigate against home advantage, and also that officials have. One such aspect of this is the role of the crowd; the hypothesis is that the crowd influences the decisions of officials that then influence match outcomes. Nevill, Newell and Gale (1996) note that red card and penalty decisions appear influenced by crowd size, although the direction of causation is not necessarily clear, since if home sides attack more, that pattern of play is consistent with more red cards for away teams, and more penalties for home teams. Buraimo, Forrest and Simmons (2010) attempt to control for this by considering a minute-by-minute analysis of yellow and red cards in-match. They do find that home teams get

fewer disciplinary sanctions, but in addition find that this effect is absent for teams (in Germany) who had stadiums with running tracks, where the crowd would be much further from the field. Pettersson-Lidbom and Priks (2010) consider a special case of football matches in Italy played without crowds. They find, consistent with the literature, that referees award more fouls and cards against visiting players when crowds are present. Data on crowds in cricket is difficult to procure, hampering any analysis using the dataset we have collected on cricket outcomes. Nonetheless, in cricket the decision on whether to bat or bowl first is determined by coin toss, with the home team being much more knowledgeable should they win the toss than the visiting team. In 2016, the English County League began an experiment with uncontested tosses, where the visiting team can choose to bowl first, or have a standard coin toss. As this experiment is in its infancy, it is probably too early to determine whether it has had an impact on home advantage. Another example of rules is permitting home teams discretion in important aspects of the hosting of an event. Officials can also create further home advantage. For example, the selection of a venue at which to play may be determined based on factors that make a home win more likely. In tennis’s Davis Cup, for example, the host nation can choose both the venue and the surface that a match will take place on. Interestingly, Gayton et al. (2009) investigate the home disadvantage in Davis Cup matches, finding that based on Final

226

THE SAGE HANDBOOK OF SPORTS ECONOMICS

matches between 1900 and 2006, in decisive fifth matches the away team wins more often than the home team.10 It is worth noting that there are only 21 such matches as all other Finals were concluded without requiring a fifth match. Nonetheless 13 of those 21 resulted in wins for the away team. The authors suggest that this is consistent with the idea that the pressure of the crowd inhibits players from carrying out skilful activities, and is something that Baumeister and Steinhilber (1984) also found in baseball’s World Series decisive playoff games. However, Gayton et al. (2009) do not control for the strength of home or away teams. The nature of Tests series in cricket enable an investigation of this with our data. Test series are a succession of matches between two teams. For example, conventionally England play Australia over a series of five matches.11 In our dataset, we have 1,947 Test matches, of which 691 are first Tests, 554 are second Tests in a series, 382 are third Tests, 183 are fourth Tests and 130 are fifth Tests, reflecting that different series will be of different lengths. Table 23.1 shows the proportion of Tests won by the home side, the away side, and how many were drawn. On the basis of these proportions, home advantage appears to decline slightly, as 42% of first Tests are won by the home side to just 26.6% by the away sides, whereas 39.7% of fifth Tests are won by the home side and 25.2% by the away side. However, this analysis does not factor in the strength of the teams involved and, judging by the falling number of games as we move from the first to the second Test, the strength of the teams competing will not be identical. Figure 23.4 is of the same format as Figure 23.3, except that here we consider different matches in the Test series. In the plot, 1 refers to first Test matches, 2 to second Test matches, and so on. Again, in the absence of home advantage, points should be congregated around the 45-degree line (highlighted), and we plot lines of best fit for each type of match. This controls for team strength and presents a more nuanced picture. The red line is the 45-degree line, which represents an absence of home advantage. The home advantage in first Test matches (solid line) is consistent throughout the range of relative quality levels between the teams

and suggests that home teams in Tests win about 9% more than their relative ability suggests they should. For second Tests, the slope flattens such that for Test matches with a stronger home team, the home advantage lessens. For third, fourth and fifth Tests, the home advantage shows a different relationship: weaker home teams gain a significantly greater advantage, as the regression lines are above the 45-degree line until the relative quality gets to around 0.65, and after that, for particularly strong home teams, there is a home disadvantage. Hence, this lends some support to the home disadvantage theory set out by Baumeister and Steinhilber (1984) and Gayton et al. (2009), as stronger teams perform worse at home.12 One feature of Davis Cup tennis is that although national teams face each other, the event consists of up to four singles matches and one doubles match. Hence, it represents a mix of team and individual sports. The majority of investigations of home advantage focus on team sports such as football, baseball and basketball. A small number have focused on individual sports. Koning (2011) looks at tennis and finds that home advantage is strongest when two highly ranked players face each other, and disappears completely when two weak players face each other. Holder and Nevill (1997) consider tennis and golf, finding little evidence of home advantage. Officials in charge of matches are expected to show neutrality in their application of the rules. Convention dictates that officials should thus not hail from the region in which a team is based. Although this does not guarantee an official will be unbiased, it may make it more likely. Gallo, Grund and Reade (2013) show that referees in football implicitly discriminate. That is, subconsciously they make decisions that discriminate against particular types of player. In addition, Dawson and Dobson (2010) document that referees in football of different nationalities display distinctly different patterns in their implementation of the rules in international football. In many international matches in cricket, officials are drawn from the home country. Before 1992, all officials would be from the home country, but after 1992 one of the two umpires in Test cricket

Table 23.1  Proportion of cricket Test matches won by each team Test number 1 2 3 4 5

No. observations

Away

Draw

Home

691 554 382 183 130

26.6 27.6 23.5 25.0 25.2

30.9 34.4 37.1 34.8 34.4

42.3 37.8 39.2 39.7 39.7

Officials and Home Advantage

227

Figure 23.4  Cricket match outcomes by Test number Note: Regression lines for different Test matches

matches was drawn from a third country (on a trial basis, with it being confirmed in 1994), and after 2002 both were from a third country. In one-day international (ODI) matches, until 2001 all officials were drawn from the home country, and since then one of the two umpires has been drawn from a third country. In T20 international matches, all officials are again drawn from the home country. The possibility that this institutional arrangement contributed to home advantage can be investigated using the framework we have set up. We restrict attention to only Test matches, and include dummy variables for 1992, 1994 and 2002, and allow these dummies to affect the slope as well as the intercept. We find only marginal evidence of significance, albeit with coefficient signs in the anticipated directions (home advantage falling). We also investigate the impact of the 2002 change in ODI matches and again find coefficient signs consistent with declining home advantage with the change, but no significance. Ringrose (2006) and Sacheti, Gregory-Smith and Paton (2015) consider the Test change in more detail. They both consider leg before wicket (LBW), a method of a player being bowled out that is open to greater levels of discretion, as it historically involved the umpire determining a counter-factual: what would have happened to the ball had it not hit the batsman’s leg? Both studies found that dismissals by LBW fell for the away side around these two changes (partially falling after 1994, then falling to no difference between the teams after 2004). Hence the impact on individual decisions can

be clearly shown, whereas the impact on overall outcomes of matches is more muted. This suggests that the observed home advantage was not determined solely by this particular aspect of bias towards a home side. It would appear that other factors mitigated this bias. Garicano, Palacios-Huerta and Prendergast (2005) find that in Spanish football, referees systematically add more injury time to matches in which the home team is losing, and less injury time when the home team is winning. PetterssonLidbom and Priks (2010), Nevill et  al. (1996), Sutter and Kocher (2004) and Boyko, Boyko and Boyko (2007) all find evidence that refereeing decisions contribute to home advantage. Along with injury time decisions, incidence of penalties and cautions appear to favour the home side also. Nevill, Balmer and Williams (2002) showed qualified referees videos of incidents in football matches, and found that when the sound was turned on, the referees called 15.5% fewer fouls against the home team compared to when watching the videos in silence. There is thus a considerable body of evidence to suggest that officials are unable to act in a neutral manner, whether implicitly or otherwise. As such, even neutral officials in international matches will likely influence outcomes subconsciously. A consistent pattern in home advantage across sports is its decline through time, and hence if the actions of officials are responsible for home advantage, this requires that via some mechanism the impact of these actions is being gradually mitigated over time.

228

THE SAGE HANDBOOK OF SPORTS ECONOMICS

In many sports in recent years video technology has been implemented to assist officials, yet the decline in home advantage precedes any such adoption of technology. In cricket, an increase in draws in the 19th century corresponded with a decrease in home advantage, while a decrease in the frequency of draws since the 1960s has also occurred in line with an increase in home advantage. Rule changes that provoke more attacking play and thus reduce the likelihood of a draw have been cited as reasons for declining home advantage in football (e.g. additional substitutions, and three points for a win; see Jacklin, 2005). However, as Koyama and Reade (2009) note, incentives for more attacking play affect both the home and away side equally and hence ought not to be reasons for a change in home advantage.

CONCLUSIONS In this chapter the literature on home advantage has been surveyed, and new evidence from the sport of cricket has been presented to illustrate the findings of the literature. Home advantage exists and varies across sports, and in general appears to be in (slow) decline. As with the literature, our analysis using a wealth of cricketing data provide nothing conclusive in terms of a single factor explaining home advantage. There is some evidence for familiarity, there is evidence for rules and the role of officials, and potentially evidence surrounding crowd, psychological and biological factors.

Notes * Thanks to Sarah Jewell, Liam Lenten and Owen Gittings, and the members of the Association of Cricket Statisticians and Historians for their invaluable insight and comments during the writing of this chapter. All remaining errors are mine. 1  A tied outcome in cricket is not the same as a tied outcome in many other sports. A tie would represent both teams scoring an identical number of runs. As with high-scoring games like basketball and rugby, this is a very rare outcome – just 0.05% of matches in our dataset are ties. Much more common is a draw (21% of matches), where in the allotted time available (which may be reduced by adverse weather), neither team was able to bowl the other team out and claim victory.

2  The slight drawback of this approach is that it provides a single ranking for cricket teams that play different forms of the game, while it may be that some teams specialise in one form (say, oneday cricket) relative to another. 3  Naturally, a small number of matches is needed for each team in order for their strength to calibrate to something nearer to their actual strength.  4  See http://cricketarchive.com.  5  Using the indicator saturation method of Hendry, Doornik and Pretis (2013) for detecting step shifts in variables, coded in R Core Team (2014) by Pretis, Reade and Sucarrat (2017), shows that since 1923 the variation around a mean of 5% has been random.  6  There are proposals to reduce five-day Test matches to four days.  7  This is the outcome of running the linear regression in (2) for the particular subset of matches under consideration.  8  In cricket, pitches are well known for deteriorating as play progresses.  9  Bolivia is not one of the 109 countries in our dataset. 10  Davis Cup ties take place over up to five matches, or rubbers. The team to reach three wins first wins the overall event. The first two matches are singles matches, the third is a doubles match and the last two are singles matches. 11  And, similar to tennis’s Davis Cup, the host nation will choose the venues for matches. Interestingly, in recent years cricketing authorities have experimented with day/night Test matches, where play takes place into the evening each day. Of the eight that have taken place at the time of writing, seven have been won by the home team. 12  An alternative explanation is that particularly by the fourth and fifth Tests, the result may be beyond doubt, if a stronger team has won the first three Tests, and as such the stronger team may try less hard or experiment in the final two matches.

REFERENCES Akhtar, S., & P. Scarf (2012). Forecasting test cricket match outcomes in play. International Journal of Forecasting, 28(3), 632–643. Baumeister, R.F., & A. Steinhilber (1984). Paradoxical effects of supportive audiences on performance under pressure: The home field disadvantage in sports championships. Journal of Personality and Social Psychology, 47(1), 85. Boyko, R.H., A.R. Boyko, & M.G. Boyko (2007). Referee bias contributes to home advantage in

Officials and Home Advantage

English Premiership football. Journal of Sports Sciences, 25(11), 1185–1194. Buraimo, B., D. Forrest, & R. Simmons (2010). The 12th man? Refereeing bias in English and German soccer. Journal of the Royal Statistical Society: Series A (Statistics in Society), 173(2), 431–449. Chumacero, R.A. (2009). Altitude or hot air? Journal of Sports Economics, 10(6), 619–638. Dawson, P., & S. Dobson (2010). The influence of social pressure and nationality on individual decisions: Evidence from the behaviour of referees. Journal of Economic Psychology, 31(2), 181–191. Elo, A.E. (1978). The rating of chessplayers, past and present (Vol. 3). London: Batsford. Forrest, D.K., J. Goddard, & R. Simmons (2005). Odds-setters as forecasters: The case of English football. International Journal of Forecasting, 21, 551–564. Gallo, E., T. Grund, & J.J. Reade (2013). Punishing the foreigner: Implicit discrimination in the Premier League based on oppositional identity. Oxford Bulletin of Economics and Statistics, 75(1), 136–156. Garicano, L., I. Palacios-Huerta, & C. Prendergast (2005). Favouritism under social pressure. Review of Economics and Statistics, 87(2), 208–216. Gayton, W.F., W. Steinroeder, C. Bonnica, & A.C. Loignon (2009). An investigation of home disadvantage in Davis Cup tennis. Perceptual and Motor Skills, 109(2), 382–386. Goddard, J. (2005). Regression models for forecasting goals and match results in Association Football. International Journal of Forecasting, 21, 331–340. Goddard, J., & I. Asimakopoulos (2004). Forecasting football match results and the efficiency of fixedodds betting. International Journal of Forecasting, 23, 51–66. Gómez, M.A., R. Pollard, & J.-C. Luis-Pascual (2011). Comparison of the home advantage in nine different professional team sports in Spain. Perceptual and Motor Skills, 113(1), 150–156. Hendry, D.F., J.A. Doornik, & F. Pretis (2013). Stepindicator saturation. Economics Series Working Papers 658, University of Oxford, Department of Economics, June. Holder, R.L., & A.M. Nevill (1997). Modelling performance at international tennis and golf tournaments: Is there a home advantage? Journal of the Royal Statistical Society: Series D (The Statistician), 46(4), 551–559. Jacklin, P.B. (2005). Temporal changes in home advantage in English football since the Second World War: What explains improved away performance? Journal of Sports Sciences, 23(7), 669–679. Koning, R.H. (2011). Home advantage in professional tennis. Journal of Sports Sciences, 29(1), 19–27.

229

Koop, G. (2004). Modelling the evolution of distributions: An application to Major League Baseball. Journal of the Royal Statistical Society: Series A (Statistics in Society), 167(4), 639–655. Koyama, M., & J. J. Reade (2009). Playing like the home team: An economic investigation into home advantage in football. International Journal of Sport Finance, 4(1), 16–41. Lenten, L.J.A. (2008). Unbalanced schedules and the estimation of competitive balance in the Scottish Premier League. Scottish Journal of Political Economy, 55(4), 488–508. Lenten, L.J.A. (2011). The extent to which unbalanced schedules cause distortions in sports league tables. Economic Modelling, 28(1), 451–458. McSharry, P.E. (2007). Effect of altitude on physiological performance: A statistical analysis using results of international football games. British Medical Journal, 335(7633), 1278–1281. Mincer, J.A., & V. Zarnowitz (1969). The evaluation of economic forecasts. In Economic forecasts and expectations: Analysis of forecasting behavior and performance (pp.1–46). Cambridge, MA: National Bureau of Economic Research. Nevill, A.M., N.J. Balmer, & A.M. Williams (2002). The influence of crowd noise and experience upon refereeing decisions in football. Psychology of Sport and Exercise, 3(4), 261–272. Nevill, A.M., & R.L. Holder (1999). Home advantage in sport. Sports Medicine, 28(4), 221–236. Nevill, A.M., S.M. Newell, & S. Gale (1996). Factors associated with home advantage in English and Scottish soccer matches. Journal of Sports Sciences, 14(2), 181–186. Paton, D., & A. Cooke (2005). Attendance at county cricket: An economic analysis. Journal of Sports Economics, 6(1), 24–45. Pettersson-Lidbom, P., & M. Priks (2010). Behavior under social pressure: Empty Italian stadiums and referee bias. Economics Letters, 108(2), 212–214. Pollard, R. (2005). Evidence of a reduced home advantage when a team moves to a new stadium. Journal of Sports Sciences, 20(12), 969–973. Pollard, R., & G. Pollard (2005). Long-term trends in home advantage in professional team sports in North America and England (1876–2003). Journal of Sports Sciences, 23(4), 337–350. Pretis, F., J.J. Reade, & G. Sucarrat (2017). General-toSpecific (GETS) modelling and indicator saturation methods. Retrieved from: http://CRAN.R-project. org/package=gets. R package version 0.12 R Core Team (2014). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Retrieved from: www.R-project.org/ Ringrose, T.J. (2006). Neutral umpires and leg before wicket decisions in test cricket. Journal of the

230

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Royal Statistical Society: Series A (Statistics in Society), 169(4), 903–911. Sacheti, A., I. Gregory-Smith, & D. Paton (2015). Home bias in officiating: Evidence from international cricket. Journal of the Royal Statistical Society: Series A (Statistics in Society), 178(3), 741–755. Scarf, P., & X. Shi (2005). Modelling match outcomes and decision support for setting a final innings target in test cricket. IMA Journal of Management Mathematics, 16(2), 161–178. Schofield, J.A. (1983). The demand for cricket: The case of the John Player League. Applied Economics, 15(3), 283–296. Schwartz, B., & S.F. Barsky (1977). The home advantage. Social Forces, 55(3), 641–661.

Sobel, R.S., & M.E. Ryan (2008). Unifying the favorite– longshot bias with other market anomalies. In D. B. Hausch and W. T. Thiemba (Eds.), Handbook of sports and lottery markets (pp. 137–160). Amsterdam, Netherlands: Elsevier. Available at: doi: 10.1016/B978-044450744-0.50011-1 Sutter, M., & M.G. Kocher (2004). Favoritism of agents – the case of referees’ home bias. Journal of Economic Psychology, 25(4), 461–469. Retrieved from: http://ideas.repec.org/a/eee/ joepsy/ v25y2004i4p461-469.html Vlastakis, M., G. Dotsis, & R.N. Markellos (2009). How efficient is the European football betting market? Evidence from arbitrage and trading strategies. Journal of Forecasting, 28(5), 426–444.

24 Franchise Relocation and Stadium Subsidies Dennis Coates

The world-wide stadium and arena construction boom over the last several decades is well known. This boom has been discussed in numerous popular and academic books (e.g., Bennett, 2012; Delaney & Eckstein, 2003; deMause & Cagan, 2008; Euchner, 1993; Long, 2014; Rosentraub, 1999, 2014; Shropshire, 1995) and in academic journal articles (Coates, 2007; Coates & Humphreys, 2008). Historically, cities without a team promise new stadiums as an enticement for a league to expand or grant permission for an existing team to move away from its current location. Usually the promise of a new stadium is the bait, but St. Petersburg, Florida, actually built a stadium without first securing a commitment from any team. Opened in 1990 as the Florida Suncoast Dome, the stadium was leverage for both the Chicago White Sox and the San Francisco Giants of Major League Baseball (MLB) to get new stadiums in their home cities. The case of the Suncoast Dome exemplifies two distinct types of relocation connected to stadium subsidies. The first type is the possibility of a franchise relocating to a new city where there are greener pastures, in the shape of a new stadium with new and expanded revenue streams. While neither the White Sox nor the Giants relocated, teams from the other three major leagues frequently move, most recently the 2017 season.

In fact, Los Angeles Rams of the National Football League (NFL) originated as the Cleveland Rams in the 1940s, spent about four decades as the Los Angeles Rams before moving to St. Louis for about two decades, and are now back in Los Angeles. The last two moves occurred in return for promises of new stadiums, which were sold to the public in part on the basis of economic development benefits. But movement from one city to another is not the only relocation that occurs; clubs can also relocate within a city or metropolitan area from one facility to another. Indeed, that is precisely what the Giants did, moving from Candlestick Park in the southern part of the city on Candlestick Point to the central part of San Francisco. The White Sox, on the other hand, moved across the street into a brand new stadium. In relocations within a city, the question arises as to where the new stadium should be built. The answer in recent years is frequently motivated by a desire to revitalize downtown or to spur development in impoverished or run-down areas. This chapter focuses on evidence of net benefits from relocations of professional sports franchises, whether between or within cities. The purpose is to understand whether these relocations successfully revitalize the city. The discussion begins with a brief review of the subsidies cities use to

232

THE SAGE HANDBOOK OF SPORTS ECONOMICS

attract or retain franchises. The theoretical motivation and empirical evidence for stadiums and arenas as tools for development and revitalization of city neighborhoods makes up the remainder of the chapter.

PURPOSE OF SUBSIDIES One reason for local governments to provide stadium and arena subsidies to professional sports franchises is the threat of teams relocating. To a non-American fan, the idea of a sports club moving is nearly inconceivable. Since the mid1950s professional sports franchises have found it beneficial to move from one city to another. Early on, the moving clubs were often the weaker franchise in a city with two teams in the same sport. For example, the Braves went from Boston to Milwaukee in 1953 leaving the Red Sox as the sole baseball team in the city. The first three moves in the 1950s were this type, the clubs moved into new or newly renovated stadiums owned by the host cities at team-friendly costs (Coates & Humphreys, 2008). The most prominent early cases of teams relocating are the 1957 moves of the MLB’s Brooklyn Dodgers to Los Angeles and New York Giants to San Francisco. These moves do not fit the model of the previous relocations as both the Giants and the Dodgers were quite successful despite sharing New York with each other and with the New York Yankees. Since these moves, the MLB has seen six clubs move to new cities. NFL teams changed cities 12 times since World War II, with one more move authorized for the 2019 or 2020 season. National Basketball Association (NBA) franchises have relocated 20 times since 1951. The first National Hockey League (NHL) relocation came later (1976), but the league has since had 10 relocations, with the most recent one in 2015 when the New York Islanders moved about 30 miles. A second reason for cities to subsidize stadium or arena construction is to entice a relocating franchise to select the city as its new home. The earlier example of the Suncoast Dome is an instance of this rationale. Closely related to this motivation is the intention to convince a league to award the city with an expansion franchise. Regardless of whether the city hopes to entice a team to relocate, to stay put, or to win an expansion franchise, part of the justification will involve one or all of faster economic growth, job creation, and expanded tax revenue. All of these, it is argued, will come from the increased visitors to the city, more hotel stays and more dinners in restaurants plus all

the spillover benefits from these visitors as their spending ripples out into the community, causing greater incomes via the multiplier effect. In addition, citizens may benefit from their city achieving ‘world class’ or ‘major league’ city status. A plethora of academic research demonstrates convincingly that such effects are, at best, quite small and, at worst, negative (see Coates, 2007; Coates & Humphreys, 2008). Subsidies may come in a variety of forms and are often not clearly identified as subsidies. Long (2005) and Zimbalist and Long (2006) make clear the typical news account of a stadium deal may reflect the price tag of the facility but often does not identify other costs to the government (e.g., land, infrastructure). Moreover, the publicized values may reflect initial estimates rather than the final costs of the facility. The upshot is that ‘[g]overnments pay far more to participate in the development of major league sports facilities than is commonly understood’ (Long, 2005, p. 119). Long (2002, 2005) computed the total development cost of the 99 facilities in use in the four major US sports leagues in 2001. Her total development cost includes the cost of constructing the facility, and expenditures for the land, and new or improved infrastructure. The average facility cost was about $222 million in 2001 dollars, about $19 million more than claimed by industry sources. The public share was 79%, rather than the 56% reported by industry sources. The difference comes from land and infrastructure, accounting for $17 million, and ‘net annual costs’ (operating costs not covered by revenues from the facility) that amount to $34 million. Subsidies can take hidden forms. One form is the city getting a smaller share of revenues than circumstances warrant. This type of subsidy is not easily recognized but it is reflected in Long’s (2002, 2005) calculations. One example of this form of subsidy is granting the club the right to sell the name of the facility and then use the revenues to count as part of its contribution to the construction costs. A more obvious, but also hidden, subsidy is the foregone property taxes on the land under the stadium or arena. Long (2002) indicated 85 of the 99 facilities involve the city losing property tax revenues on the facility location. The average annual cost is over $4.5 million for stadiums and over $3.1 million for arenas. Taken all together, the evidence Long (2002, 2005) and Zimbalist and Long (2006) provided suggests the true cost to the public sector of subsidies for the construction of stadiums and arenas is substantially larger than is commonly thought. Whether these subsidies are warranted because of the benefits produced by having a team in the city is doubtful.

Franchise Relocation and Stadium Subsidies

REVITALIZATION The large literature finding no effect of stadiums, arenas or the presence of professional sports on incomes, jobs and tax revenues received criticism by a number of authors (e.g., Austrian & Rosentraub, 2002; Santo, 2005). The chief complaint was most of the facilities studied were part of the large wave of construction in the late 1960s and 1970s, which often put stadiums in suburban areas surrounded by parking lots. According to this critique, these facilities could not possibly have the same impact as building a stadium or an arena in the central city where spillover development was much more likely. The argument ultimately is built upon advances in the urban economics literature, which Glaeser and Gottlieb (2009) summarize. This literature explains the wealth of cities as arising from agglomeration economies, the benefits of proximity in reducing the costs of moving goods between firms in the supply chain, in the ability of people to move between firms, and in the increase in the speed of transmission of ideas which facilitates innovation and accumulation of human capital. In this literature, subsidies to firms can be efficient, since if the firms cannot reap the benefits of these agglomeration economies, they will under-invest; the subsidies induce firms to internalize these external benefits (Garcia-Mila & McGuire, 2002; Glaeser, 2001). Glaeser (2001) suggested that subsidies are bids to attract firms that will produce consumer or producer surplus for current residents and are hard to capture. Rosentraub (2014) emphasized this last benefit of proximity. ‘[C]ities have to be designed with space or neighborhoods where people can interact and where ideas are put into motion. Ideas are not only exchanged in the workplace’ (Rosentraub, 2014, p. 64). He argued that stadiums and arenas provide such space or are an amenity that attracts people with large human capital to a city. Furthermore, cities must engage in activities like stadium and arena subsidies to entice people back from the suburbs. The move to the suburbs that occurred in the last two-thirds of the 20th century, which was especially rapid after World War II, resulted from several factors: high taxes in the city, cheap land in the suburbs, and the lower cost of private transportation in the form of personal cars. The upshot of suburbanization is the central city lost the wealthy citizens who migrated to the suburbs, leaving predominantly the poor as city residents. This economic segregation led to a deterioration of the public services within the city, services on which the low-income population depended. Rosentraub (2014) noted this deterioration would not be a

233

problem if there was revenue sharing between poor and wealthy cities, between the prosperous suburbs and the impoverished central city. Since such redistribution is unlikely within the American system of fiscal federalism, his conclusion is that central cities must develop initiatives that focus ‘on changing the demand for land within their boundaries’ (Rosentraub, 2014, p. 45). One such initiative is stadium and arena subsidies. For Rosentraub (2014, p. 50), the focus was ‘on the best ways to undertake revitalization efforts that are anchored by big-ticket investment in sports, entertainment and cultural facilities.’ If these efforts are successful, they will produce’ (1) a deflection of regional economic activity into central cities, (2) an improved attractiveness of the downtown area as a place where people and businesses want to locate, and (3) an enhanced fiscal position for the central city’ (Rosentraub, 2014, p. 50). Before discussing the evidence on sport facilities as catalysts for local revitalization, there are three points to make about Rosentraub’s basic argument. First, his argument relies upon the ability of the central city in an urban area to enact policies that deflect regional economic activity their way. Evidence from a large array of tax exemptions, tax increment financing (TIF), enterprise zones, capital and labor subsidies, infrastructure improvements and exemption from regulations utilized as spatially targeted economic development incentives calls into question this first point. Greenbaum and Landers (2009, 2014) evaluated the effectiveness of enterprise zones and TIFs. Regarding enterprise zones, the evidence is mixed owing in part to different methodologies, differences in the programs, and differences in the outcome of interest – employment, business climate, resident welfare. The effectiveness of TIFs is also unclear. Some research shows that TIF areas see an increase in property values, for example, but also indicate that the non-TIF areas lose value. Moreover, even where the property values rise, the increment to tax revenues may be too small to finance the government services, loans or other benefits driving the property value increase. Additionally, Glaeser (2007) points out that increased property prices suggest benefits accrue to property owners, generally the relatively wealthy, and may force the poor to relocate, inducing an even greater concentration of poverty. Calcagno and Hefner (2017) reviewed targeted tax incentives used to attract or retain businesses, especially automobile manufacturers and film companies. They found that these incentives are very expensive, rarely meet the stated goals in terms of jobs created, and complicate and reduce

234

THE SAGE HANDBOOK OF SPORTS ECONOMICS

the transparency and the fairness of the tax system. Additionally, these firm-specific incentives enable firms to play one jurisdiction off against another to sweeten the incentive package. This strategy is consistent with the White Sox and Giants using the threat of the Sun Coast Dome in St. Petersburg to extract stadium deals from Chicago and San Francisco, respectively. Moreover, there is evidence that the wider use of targeted incentives is positively correlated with greater corruption in state and local government. Ahlfeldt, Maennig, and Richter (2017) evaluate place-based policy related to urban renewal in Berlin after the fall of the Berlin Wall. The policy in question aims to improve the quality of the existing housing stock, which was in particularly poor shape as a result of Berlin’s isolation (in West Berlin) and being subject to the command and control of economic organization (in East Berlin). They find that the policy led to an improved housing stock but the evidence of spillover into the value of unrenovated properties is weaker. Importantly, they conclude that the program ‘has also primarily been a cash transfer to those landlords participating in the program’ (Ahlfeldt et al., 2017, p. 152). Second, Rosentraub’s (2014, p. 50) desire to remake ‘the downtown area as a place where people and businesses want to locate’ needs justification. It is telling that Rosentraub (2014) identified Indianapolis as a model development plan and states that it is ‘important to point out that Indianapolis’ plan was not the product of a review process that incorporated public input or encouraged reviews and comment’ (2014, p. 129). In other words, it seems the city government, like Rosentraub, decided to pursue downtown renewal independently of the wishes of the citizens of Indianapolis. People and businesses moved out of the central city for any number of reasons that may have little to do with public policy. Moreover, it is not obvious that cities should or do exist to be where people and businesses want to locate, but rather that cities form where people have chosen to locate. Related to that is the third point. If the city cannot finance the activities it undertakes, the best solution may be to reduce the size and number of such activities. People moved to the cities when doing so was advantageous; they can move out of them when staying becomes worse than leaving. Of course, such an approach assumes individuals and firms are in the best position to decide what is best for them and does not concern itself with the outcome for the city. Glaeser (2007, p. 5) puts it this way, ‘Economics judges policies by whether they increase the choices available to people, not on whether they help rebuild a particular locale.’

EVIDENCE ON PLACE-BASED POLICY AND STADIUMS AND ARENAS Like the evidence on TIFs and enterprise zones, evidence on the effectiveness of stadiums and arenas at revitalizing neighborhoods or bringing people and businesses back downtown comes in several forms. First, one can look for the stadium or arena to have affected property values or rents. Second, one could assess the impact of the facility on employment, wages or new businesses. Finally, one could address potential negative consequences, like additional crime or traffic congestion. These studies also can be separated by research design, quasi-experimental designs or not. The quasi-experimental studies estimated changes in the outcome of interest across time periods identified by construction of a stadium or arena in an area; the other studies examine outcomes at different locations within a city. Indeed, some of this research considers the entire city rather than the neighborhood around a sports facility, so in that regard the implications are only suggestive of the relevant outcomes for stadiums or arenas as anchors for development.

PROPERTY VALUES OR RENTS Not Quasi-experimental Studies Carlino and Coulson (2004) evaluated the presence of an NFL team on a city by estimating hedonic rent regressions on two years of data from the American Housing Survey. They concluded that an NFL team in a city induced about an 8% increase in monthly rents for housing in the central city. They considered this increase an estimate of the social benefits of the NFL. Importantly, the NFL effect is generally insignificant when the data are expanded to include the entire Metropolitan Statistical Area (MSA) or Consolidated Metropolitan Statistical Area (CMSA), leading them to conclude that ‘football is more of a central city amenity’ (Carlino & Coulson, 2004, p. 42). Coates, Humphreys, and Zimbalist (2006) were critical of their findings and made modest changes to the estimations, which resulted in the NFL effect dropping in size and becoming statistically insignificant even in the central city sample. Neither Carlino and Coulson (2004) nor Coates et  al. (2006) were able to address the very localized benefits that Rosentraub (2014) touted as an outcome of stadium and arena subsidies. Feng and Humphreys (2012) used census block group information from the 1990 and 2000 censuses to evaluate the impact of proximity to a sport

Franchise Relocation and Stadium Subsidies

facility on property values. Their data included all the sports facilities in operation in either or both of 1990 and 2000 in 45 Metropolitan Statistical Areas. Data at this fine geographical level allowed them to use spatial econometric methods. They found the spatial lag is positive and statistically significant, indicating that the higher is the value of nearby property the higher is the value of this property. The coefficient on the distance variable is negative and statistically significant in three of four specifications. Feng and Humphreys (2012) interpreted this result as indicating that the facility affords nearby properties with an amenity value that dissipates with distance. Interestingly, Feng and Humphreys (2012) did not comment on the coefficients on the NFL, MLB and NBA dummies, each of which is negative and statistically significant in at least one specification, indicating that there may be some spillover benefit to properties close to the facility. There is also evidence that all property values are lower from the mere presence of a team. They showed with the 2000 Census data that moving a house one mile closer to the facility raises its value $793. From that same equation, the presence of an NFL team lowers the value of the house $28,202. As a policy for downtown revitalization, if home values in the neighborhood of the facility are the metric by which success of the revitalization is measured, the results surely indicate building a football stadium is ineffective. Feng and Humphreys (2018) examined the effects of the Nationwide Arena and Crew Stadium in Columbus, Ohio, on residential property values in the city in relationship to the distance between them. The distance variables have negative and statistically significant coefficients, meaning a 1% decrease in distance between a house and the Nationwide Arena increases property value by about 0.175%. They used their coefficient estimates to infer that the Nationwide Arena generates aggregate willingnessto-pay of about $225 million on residential properties within a mile of the facility. Crew Stadium produced far less, about $35.7 million.

Quasi-experimental Studies Tu (2005) evaluated the impact of the opening of FedEx Field outside Washington, DC on the resale value of houses in the neighborhood of the facility using a differences-in-differences methodology. His results indicate that homes closer to the stadium sold at a discount relative to houses farther away, but that discount became smaller after the stadium opening. He concluded that housing values improved as a result of the stadium construction, to a distance of nearly 2.5 miles, by

235

about $42 million. The state of Maryland spent about $70 million on infrastructure improvements to support the stadium construction. At the same time, an arena named the Capital Center or US Airways Arena was demolished. Many of the houses within 2.5 miles of FedEx Field were also close to the site of the demolition, raising a question about whether it was the construction of FedEx or the destruction of US Airways Arena that produced the improved house values. Ahlfeldt and Maennig (2009) used block structure data from Berlin to assess the impact on the growth in property values around three sport facilities that opened since the reunification of Germany. The analysis focused on the growth rates in property values before and after the opening of the facilities, and found positive effects on property values, which came from the sophisticated architecture rather than increased economic activity in the neighborhood. Ahlfeldt and Kavetsos (2014) also support these findings of stadium architecture influencing neighborhood property values in London, using localized sales data to assess the influence of two stadiums. Following a differences-in-differences approach they match properties near to one of the stadiums with properties distant enough from that stadium to be judged unaffected by the stadium. A unique aspect is that they were able to identify changes in property prices along paths to and from subway stations before and after construction of Emirates Stadium. The result was that routes taken by the stadium crowds going to or from subway stations changed. Their evidence shows that property prices rose about 30% on the abandoned routes and fell by about 12% on the new higher traffic paths. Aggregating willingness-to-pay, they found the new facilities generated between £1 billion and £2 billion, but the loss of the old stadium reduced property values £1.41 billion. These external benefits and costs of the new facility are highly localized, raising important distributional questions. Humphreys and Nowak (2017) took advantage of a natural experiment in which two NBA teams changed cities, the Supersonics from Seattle to Oklahoma City, and the Hornets from Charlotte to New Orleans, while the facilities they left remained in operation. Their results indicated that property prices appreciated after the departure of the team, in Seattle by 6–7% and in Charlotte by 7.5–14% depending on model specification and distance from the arena. Propheter (2017) focused specifically on commercial rents in the neighborhood of the Barclays Center in Brooklyn, New York, to see whether location near the Barclays Center matters for profitability of businesses. The evidence suggested higher commercial rents closer to the facility, but that

236

THE SAGE HANDBOOK OF SPORTS ECONOMICS

declined as distance increased. Propheter (2017) emphasized that this finding may mean that building owners, and not the businesses leasing space in those buildings, are the primary beneficiaries of the location near the arena. Chikish et  al. (2017) evaluated the impact of home NHL and NBA games as well as concerts on hotels near the Staples Center and L.A. Live complex in Los Angeles, California. The novel aspect of the research project is hotels adjacent to L.A. Live received 20-year waivers from the local hotel occupancy tax, which is not available to other hotels. The research question is, then, does this special treatment of the hotels affect their performance and does it differ between sporting event and concert dates and non-event dates. Performance is measured using the occupancy rate, average daily room rate and the revenue per available room. Rooms are more expensive close to the Staples Center than rooms in hotels more than a mile away, but rates in the closest hotels are not different on game days than non-game days. The occupancy rate is higher in the closest hotels on game days for the NBA but not for the NHL. Perhaps unexpectedly, room rates and revenue per available room are both higher during work stoppages in the NHL or NBA than when the two leagues are operating. The evidence clearly suggests that the benefits for the hotels close to the Staples Center are offset by losses to the hotels farther away.

EMPLOYMENT, WAGES OR NEW BUSINESSES Few papers address the question of whether employment, wages or new businesses increased following the opening of a new arena or stadium. Coates and Humphreys (2003) evaluated the impact of stadium openings and the presence of sport franchises on wages and employment by sector. Their results indicate small increases in earnings per employee in the amusements and recreation sector but decreases in earnings and employment in other sectors. Jasina and Rotthoff (2008, 2016) and DeAntonio (2016) examined employment impacts. Jasina and Rotthoff (2008) looked at employment at the county level and found mixed evidence on employment and payrolls in various industries but no evidence of any effect of franchises on average wages per employee in any industry. Jasina and Rotthoff (2016) returned to the issue of employment and payroll effects of professional sports, this time taking advantage of a natural experiment, the 2004/05 season-long lockout in the NHL. Their results indicated no impact on employment, but payroll

was reduced in some sectors of the economy. DeAntonio (2016) addressed employment and payroll issues with the focus largely on venues for minor league hockey and the NBA developmental league. He found no evidence that a new arena has beneficial labor market effects. Harger, Humphreys, and Ross (2016) assessed the influence of new sports facilities on new businesses and employment. Their analysis tests prior findings from Humphreys and Zhou (2015), indicating how the composition of the business population adjusts to the opening of a new sports facility. Harger et al. (2016) reported no evidence of any influence on the number of businesses near the new facility compared to Census tracts more than one, three or five miles away. Even when looking at specific industries, by two-digit SIC code, there is no evidence of more new businesses near to the facility relative to farther away. The lone piece of information favorable to the placebased development argument is that eating and drinking establishments within a mile of the facility have about one more employee after the facility opens than bars and restaurants more than a mile from the newly opened sport venue.

CRIME AND CONGESTION Sports facilities also introduce new costs on the neighborhoods where they are built. Every fan attending an event passes through the community, raising congestion in the area. Residents bear some of the brunt of these external costs. Similarly, fans leaving the game have frequently drunk too much, leading to petty crime, public urination, vandalism and sometimes even violence. Businesses and residents in the area of the facility incur costs associated with these types of unruly fan behavior. Empirical studies on crime related to sports mostly cover larger geographic areas, such as the city or metropolitan area. Unfortunately, such studies do not have a consistent definition of the crimes nor do they find consistent results. Rees and Schnepel (2009) found that all offenses are statistically significantly higher on the day of a home US college football game. Assaults were higher by about 9%, and liquor law violations by 76%. The results are somewhat sensitive to whether the home team wins or loses and whether the outcome was an upset. Kalist and Lee (2016) estimated increases in larceny and car theft on game days of 4.1% and 6.7%, respectively. Pyun (2019) found that assaults increase but there is no effect on vehicle theft, total offenses, or larceny. Pyun and Hall (2017) indicated larceny occurs more often before the NFL’s Detroit

Franchise Relocation and Stadium Subsidies

Lions games, but no evidence of changes in any other crimes. Baumann, Ciavarra, Englehardt, and Matheson (2012), found little evidence of a linkage between criminal activity and the presence of professional sports. Montolio and Planells-Struse (2016) studied the time distribution of crime relative to the soccer matches of FC Barcelona. They find that some types of crime occur more frequently prior to home matches and some increase in frequency after home matches. For example, thefts are higher before home matches but are unchanged prior to away matches. They suggest that this provides evidence of greater crime around the stadium prior to the home match as pick-pockets work the crowds congregated in and around the stadium, but they do not have geographic data to verify this. Marie (2016) has the best focus on neighborhood crime associated with sporting events. Using soccer matches in 31 boroughs of London, he evaluated day and time, and home and away effects on crime. This approach is possible because there are at least nine soccer clubs with stadiums in seven different boroughs and crime statistics are reported by borough. The primary finding is that property crimes, theft, burglary and criminal damage, increase in the borough hosting a match. There is some evidence that such crimes are lower when the match is away. There is no evidence that violent crimes are affected. Finally, Humphreys and Pyun (2017) examined the influence of MLB games on traffic congestion. They found that a downtown stadium increases miles traveled, and an increase in baseball attendance increases both miles traveled and vehicle hours of traffic delay. Besides the external costs to drivers stuck in traffic longer and driving farther, they suggested the added driving distance and time may increase greenhouse gases. These are external costs that communities subsidizing stadiums should include in their decision calculus; unfortunately, they do not.

SUMMARY Economists and other social scientists have studied the claims of large economic benefits to building sports facilities for professional teams. The early studies examined entire metropolitan areas and found little evidence of income increases, job creation or expanded tax revenues. Advocates of stadium and arena construction argued that these broad areas were inappropriate for the question. Indeed, some suggested that the whole point was to build up specific neighborhoods, revitalizing downtown, for example, rather than to produce

237

widespread benefits across the metropolitan area. Drawing upon urban economics and the idea of agglomeration economies, stadium advocates argued that a stadium could anchor and leverage other types of development in retail and residential activities. The subject of this chapter is the attempts to evaluate these claims. The evidence is mixed. Looking at the impact on property values, one concludes that there is some indication that property near to the newly opened facility becomes more valuable. Less clear is if the increased value generates enough in property tax revenue to cover the subsidies and other costs incurred by the local government. Moreover, an increase in the value of property benefits property owners whose wealth has increased but does nothing for the general public. Indeed, if property is more valuable, rents will rise, which may push out lower-income families. Evidence on job creation, wages and new businesses is far less supportive. The evidence is that very few jobs are created and they are largely limited to eating and drinking establishments. There is no evidence that new businesses spring up at a faster rate near to a new sports facility than in other areas of the city. Evidence on crime and other external effects of stadiums and arenas on neighborhoods is also mixed. Numerous studies find crime of some sort higher on game days than other days, but studies are not consistent in what crimes those are or even in addressing the same set of crimes. Attendance at sporting events and, to a lesser extent, a downtown stadium, do increase traffic congestion measured as travel delays and in additional miles traveled. The lack of strong results in favor of stadiums and arenas as powerful tools of neighborhood revitalization is consistent with the broader literature on place-based policy preferences. From a positive perspective, one must conclude that such policies are not the great success their proponents claim. From a normative perspective, the evidence is fairly clear that providing targeted development incentives has significant redistributional implications. To garner support for stadium and arena subsidies, advocates of those policies must provide a convincing case for the importance or appropriateness of the implied redistribution.

REFERENCES Ahlfeldt, G. M., & Kavetsos, G. (2014). Form or function? The effect of new sports stadia on property prices in London. Journal of the Royal Statistical

238

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Society: Series A (Statistics in Society), 177, 169–190. Ahlfeldt, G. M., & Maennig, W. (2009). Arenas, arena architecture and the impact on location desirability: The case of ‘Olympic Arenas’ in Prenzlauer Berg, Berlin. Urban Studies, 46, 1343–1362. Ahlfeldt, G. M., Maennig, W., & Richter, F. J. (2017). Urban renewal after the Berlin Wall: A place-based policy evaluation. Journal of Economic Geography, 17, 129–156. Austrian, Z., & Rosentraub, M. S. (2002). Cities, sports, and economic change: A retrospective assessment. Journal of Urban Affairs, 24, 549–563. Baumann, R., Ciavarra, T., Englehardt, B., & Matheson, V. A. (2012). Sports franchises, events, and city livability: An examination of spectator sports and crime rates. Economic and Labour Relations Review, 23, 83–97. Bennett, J. T. (2012). They play, you pay: Why taxpayers build ballparks, stadiums, and arenas for billionaire owners and millionaire players. New York: Springer. Calcagno, P. T., & Hefner, F. (2017). Economic development tax incentives: A review of the perverse, ineffective, and unintended consequences. In A. J. Hoffer & T. Nesbit (Eds.), For your own good: Taxes, paternalism, and fiscal discrimination in the twenty-first century (pp. 231–242). Arlington, VA: Mercatus Center at George Mason University. Carlino, G., & Coulson, N. E. (2004). Compensating differentials and the social benefits of the NFL. Journal of Urban Economics, 56, 25–50. Chikish, Y., Humphreys, B. R., Liu, C. H., & Nowak, A. (2017). Professional sports events, concerts, and urban place-based policy: Evidence from the Staples Center. Working Paper 17-32. Morgantown, WV: West Virginia University. Retrieved from: https://business.wvu.edu/files/d/6ff52891-bd3349e2-be43-6f2fd8cd42e8/17-32.pdf Coates, D. (2007). Stadiums and arenas: Economic development or economic redistribution? Contemporary Economic Policy, 25, 565–577. Coates, D., & Humphreys, B. R. (2003). The effect of professional sports on earnings and employment in the services and retail sectors in US Cities. Regional Science and Urban Economics, 33, 175–198. Coates, D., & Humphreys, B. R. (2008). Do economists reach a conclusion for sports franchises, stadiums, and mega-events? Econ Journal Watch, 5, 294–315. Coates, D., Humphreys, B. R., & Zimbalist, A. (2006). Compensating differentials and the social benefits of the NFL: A comment. Journal of Urban Economics, 60, 124–131. DeAntonio, D. (2016). Three essays in labor economics with applications to sports. Doctoral Dissertation, Lehigh University. Retrieved from: https://preserve.

lehigh.edu/cgi/viewcontent.cgi?article=3567 &context=etd Delaney, K. J., & Eckstein, R. (2003). Public dollars, private stadiums: The battle over building sports stadiums. New Brunswick, NJ, and London: Rutgers University Press. deMause, N., & Cagan, J. (2008). Field of Schemes: How the great stadium swindle turns public money into private profit (revised and expanded ed.). Lincoln, NB: Bison Books. Euchner, C. C. (1993). Playing the field: Why sports teams move and cities fight to keep them. Baltimore, MD: Johns Hopkins University Press. Feng, X., & Humphreys, B. (2012). The impact of professional sports facilities on housing values: Evidence from census block group data. City, Culture and Sociey, 3, 189–200. Feng, X., & Humphreys, B. (2018). Assessing the economic impact of sports facilities on residential property values: A spatial hedonic approach. Journal of Sports Economics, 18, 188–210. Garcia-Mila, T., & McGuire, T. J. (2002). Tax incentives and the city. Working Paper No. 631. BrookingsWharton Papers on Urban Affairs. Retrieved from: http://dx.doi.org/10.2139/ssrn.394284 Glaeser, E. L. (2001). The economics of locationbased tax incentives. Discussion Paper 1931. Cambridge, MA: Harvard Institute of Economic Research. Retrieved from: http://scholar.harvard. edu/files/glaeser/files/hier1932.pdf%3Fm= 1360042861 Glaeser, E. L. (2007). The economics approach to cities. Working Paper 13696. Cambridge, MA: National Bureau of Economic Research. Retrieved from: www.nber.org/papers/w13696 Glaeser, E. L., & Gottlieb, J. D. (2009). The wealth of cities: Agglomeration economies and spatial equilibrium in the United States. Journal of Economic Literature, 47, 983–1028. Greenbaum, R. T., & Landers, J. (2009). Why are state policy makers still proponents of enterprise zones? What explains their action in the face of a preponderance of the research? International Regional Science Review, 32, 466–479. Greenbaum, R. T., & Landers, J. (2014). The tiff over TIF: A review of the literature examining the effectiveness of the tax increment financing. National Tax Journal, 67, 655–674. Harger, K., Humphreys, B. R., & Ross, A. (2016). Do new sports facilities attract new businesses? Journal of Sports Economics, 17, 483–500. Humphreys, B. R., & Nowak, A. (2017). Professional sports facilities, teams and property values: Evidence from NBA team departures. Regional Science and Urban Economics, 66 (Supplement C), 39–51. Humphreys, B. R., & Pyun, H. (2017). Professional sporting events and traffic: Evidence from US

Franchise Relocation and Stadium Subsidies

cities. Retrieved from: https://papers.ssrn.com/ sol3/papers.cfm?abstract_id=2940762 Humphreys, B. R., & Zhou, L. (2015). Sports facilities, agglomeration, and public subsidies. Regional Science and Urban Economics, 54 (Supplement C), 60–73. Jasina, J., & Rotthoff, K. W. (2008). The impact of a professional sports franchise on county employment and wages. International Journal of Sport Finance, 3, 210–227. Jasina, J., & Rotthoff, K. W. (2016). The impact of the NHL lockout on county employment. International Journal of Sport Finance, 11, 114–123. Kalist, D. E., & Lee, D.Y. (2016). The National Football League: Does crime increase on game day? Journal of Sports Economics, 17, 863–882. Long, J. G. (2002). The real cost of public subsidies for major league sports facilities. Cambridge, MA: Harvard University Press. Long, J. G. (2005). Full count: The real cost of public funding for major league sports facilities. Journal of Sports Economics, 6,119–143. Long, J. G. (2014). Public-private partnerships for major league sports facilities. New York: Routledge. Marie, O. (2016). Police and thieves in the stadium: Measuring the (multiple) effects of football matches on crime. Journal of the Royal Statistical Society: Series A (Statistics in Society), 179, 273–292. Montolio, D., & Planells-Struse, S. (2016). How time shapes crime: The temporal impacts of football matches on crime. Regional Science and Urban Economics, 61 (Supplement C), 99–113. Propheter, G. (2017). Estimating the effect of sports facilities on local area commercial rents: Evidence

239

from Brooklyn’s Barclays Center. Journal of Sports Economics (in press). doi: https://doi.org/10.1177/ 1527002517723048 Pyun, H. (2019). Exploring causal relationship between Major League Baseball games and crime: A synthetic control analysis. Empirical Economics, 57, 365–383 Pyun, H., & Hall, J. C. (2017). Does the presence of professional football cause crime in a city? Evidence from Pontiac, Michigan. Working Paper 16-02. Morgantown, WV: West Virginia University. Retrieved from: https://business.wvu.edu/files/d/28db83458b12-4659-9ce2-0e8eee5ee73e/16-02.pdf Rees, D. I., & Schnepel, K. T. (2009). College football games and crime. Journal of Sports Economics, 10, 68–87. Rosentraub, M. S. (1999). Major league losers: The real cost of sports and who’s paying for it (revised ed.). New York: Basic Books. Rosentraub, M. S. (2014). Reversing urban decline: Why and how sports, entertainment, and culture turn cities into major league winners (2nd ed.). Boca Raton, FL: Routledge. Santo, C. (2005). The economic impact of sports stadiums: Recasting the analysis in context. Journal of Urban Affairs, 27, 177–192. Shropshire, K. L. (1995). The sports franchise game: Cities in pursuit of sports franchises, events, stadiums, and arenas. Philadelphia, PA: University of Pennsylvania Press. Tu, C. C. (2005). How does a new sports stadium affect housing values? The case of FedEx Field. Land Economics, 81, 379–395. Zimbalist, A., & Long, J. G. (2006). Facility finance: Measurement, trends, and analysis. International Journal of Sport Finance, 1, 201–211.

PART IV

Professional Sports Leagues

25 The Economics of Professional Soccer Daniel Weimar

INTRODUCTION Since the first sports economics analysis by Rottenberg (1956), sport economic investigations are of rising interest (Santos & Garcia, 2011). Within the sports economics literature, soccer is among the most researched fields. The discipline of ‘soccernomics’ (Kuper & Szymanski, 2009) started with the seminal articles by Sloane (1969, 1971). Since then, literature on the topics of soccer and/or European soccer increased significantly.1 This increasing attention can be explained by two phenomena: the rising ­economic importance of soccer competitions and the awareness of soccer competitions being an excellent economic laboratory for gaining new insights into general economic questions (Sloane, 2015). As a consequence, there are several books providing a comprehensive overview on the economics of soccer (e.g. Andreff & Szymanski, 2006; Dobson & Goddard, 2011; Goddard & Sloane, 2014; Kuper & Szymanski, 2009). Over time, different key topics which are repetitive or temporarily ‘en vogue’ have been evolved in empirical soccer economics research. For researchers at the beginning of their sports economic research career it might be important to understand the streams (and temporal trends) of research within soccer economic research. It might be also of importance to know the most

active researchers per stream to build up networks or to have an overview on journals publishing soccer economic research for identifying an appropriate journal. Nowadays, however, most existing review studies only offer a literature review about one specific research substream. Thus, instead of economically analyzing  the ­soccer industry, this chapter categorizes the  ­ soccer economic literature by identifying main streams and substreams of research, and discussing temporal trends, researchers and journals open for the topic. Since the literature on soccer economics and management is impressively large, this investigation focuses on international journal publications using empirical approaches and actual outcomes.2 Studies exploring questions with respect to the pure management of professional and non-profit soccer clubs are not considered in this chapter (e.g. marketing, sponsorship, fans, governance, memberships, and volunteering).3 The main objective of the content structuring was to cluster past studies into main streams, substreams and special topics within substreams to achieve clusters of an almost equal number of observations. The attempt was to give a mostly complete overview on the empirical soccer economic research. However, a complete census is beyond realization. Thus, the selection of articles should cover the basic population in a great extent.

244

THE SAGE HANDBOOK OF SPORTS ECONOMICS

In sum, 486 studies (615 different authors) considering soccer economics were categorized into six main streams of research, 18 substreams and one miscellaneous cluster. The main streams of research were identified as: team productivity, individual productivity and soccer labor market, soccer finance, demand for soccer, betting and forecasting and penalties.

CONTENT STRUCTURING Team Productivity The first main stream, namely ‘Team productivity’, includes research on team outcome and efficiency of soccer teams. ‘Team outcome’ (e.g. goal outcome, match outcome, season outcome) comprises studies considering the outcome of league teams or of national teams. More precisely, mainly determinants of team success are explored. Within the special topic of league outcome, the impact of managerial change on team ­performance, the relation between team performance and payment, and investigations on the home advantage are of main interest. Research on ‘Team efficiency’ are the second subcluster of research. Such investigations mainly use production functions and special econometric estimations, which were mainly nonparametric data envelopment analysis (DEA) models or stochastic/deterministic/random/bey­ esian frontier (SFA) models.4 Efficiency analyses measure the relation of input to output to assess whether the input factors had been used efficiently to produce the outcome. Another application is the ­estimation of a potential outcome (e.g. end season rank, match outcome) if all resources would have been used efficiently (Villa & Lozano, 2016).

Individual Productivity and Soccer Labor Market The second category is labeled as ‘Individual productivity and soccer labor market’, since both constructs are closely connected. The empirical research done in this field is divided into investigations on ‘Managers and coaches’, ‘Referee’ and ‘Player’. Empirical articles considering managerial efficiency, career duration of managers or risk-taking of managers are concluded within the stream ‘Managers and coaches’. The research on managerial aspects has a long tradition and has its beginnings in the seminar article of Wilders (1976), who already

claimed the position of a soccer manager a ‘precarious occupation’. Investigations surrounding referee performance in soccer primarily investigate the ‘referee bias’. Dohmen and Sauermann (2016) provide a comprehensive review of the literature on the referee bias. With regard to the referee bias, two different types of favoritism are possible: favoring the home team over the guest or favoring certain players over others. The bias towards the home team has several facets and is generally explained by stake size, the attendance (crowd pressure), distance to the field, financial i­ncentives or referee-specific-effects such as age or cultural ­ differences (Dohmen & Sauermann, 2016). In addition, there are also studies on the career of referees and the performance without focusing on the ‘usual’ referee bias. The third cluster ‘Player’ is, besides research on team outcome, the most researched category within the group of empirical soccer economics. Since Sutherland’s (1988) seminal article on the soccer player’s labor market, this topic has been of sustained interest among researchers. Up to today, there are 68 empirical papers published as well as a couple of review articles (e.g. Goddard & Sloane, 2014; Frick, 2007; Sutherland, 1988). Most of the research on players can be subdivided into explorations on players’ careers, migration and discrimination, the performance and effort provision of players, the importance and functioning of the transfer market, wage determination, market values of players and insights into youth soccer. A general drawback of soccer research concerning wage determination is the common disclosure of individual salary information. Since high-quality soccer players’ talent is an incomplete substitute for poor talent, salaries in soccer are found to be rising disproportionally to marginal productivity, which is referred to as the ‘Superstar effect’ (Rosen, 1981). Therefore, seven out of 20 publications considering wage determination investigate superstar salaries. The so-called ‘transfer market’ is a construct summarizing the player transfer regulations (especially within the European Union) and the transactions occurring within these regulations (Feess & Mülheußer, 2002). These regulations affect contract length, wages, profits, investments into player development and turnover probabilities. The empirical literature on the soccer transfer market is in great parts associated with disruptive ‘Bosman judgement’5 (1995), which fundamentally changed transactions on the transfer market and thus had great economic impact on worldwide soccer (Ericson, 2000; Fees & Mülheußer, 2002; Simmons, 1997). The Bosman case was also the first great impact of the European Law and the European Court of Justice on sports

THE ECONOMICS OF PROFESSIONAL SOCCER

in Europe. Since the Bosman regulations emerged ‘over night’, it had the nature of a large field experiment and, thus, was and still is an ideal setting for causal investigations of various phenomena (e.g. Frick, 2009; Marcén, 2016; Radoman, 2015). The topic of youth soccer is poorly researched, with the most empirical studies of the field investigating the relative age effect (Barnsley, Thompson & Barnsley, 1985). This seems surprising, since the quality and efficiency of youth education determines the later success and financial flexibility by trading player rights (Prinz & Weimar, 2017). More specifically, soccer talent is developed in socalled talent identification and ­development (TID) processes, which provide development of soccerrelated skills (e.g. passing, shooting), physical forces (body strength, muscle), sociological traits (e.g. team playing) and psychological abilities (e.g. decision making, personality, creativity, game intelligence) as well as basic general education (Williams & Reilly, 2000). Youth soccer education is very peculiar from other labor markets in several terms, such as the influence of parental utility (Prinz & Weimar, 2017).

Soccer Finance The discussion about soccer clubs and finance has always been controversial. Since Sloane (1971), researchers are in debate about whether European soccer clubs are profit-maximizing organizations (as it is generally assumed for North American major leagues) or utility maximizers, with the ­primary aim of achieving titles and wins. One argument for the utility-maximizing assumption is the critical effect of the missed expectations of fans (e.g. relegations or missed qualification for international competitions), which shed a critical light on club managers and lead to firings. Today, the conclusion is that there has been a shift from utility to being profit maximizer. This view is now shared by a higher percentage of researchers than it was at the time of Sloane’s (1971) article (e.g. Késenne, 2006; Leach & Szymanski, 2015; Szymanski, 2017). Despite the shift to more profit-maximizing behavior, soccer clubs have always been criticized for financial mismanagement. Already in 1983, Sloane published a paper with the title ‘Economic crisis in professional football’. Since then, various authors emphasized the financial instabilities and financial peculiarities of soccer clubs. Among various factors of explaining the strong incentives to financial overconfidence, especially the market characteristics of a ‘Winner-take-all-market’, weak governance, resulting rat races among competing clubs and the nature of soft budget

245

constraints are seen as the main causes (Franck & Müller, 2000; Storm & Nielsen, 2012). The financial critical situation has been discussed for various European professional leagues (e.g. see the special issue of the Journal of Sports Economics in 2006 ‘The Financial Crisis in European Football’). Even if some authors see fewer financial miseries in German soccer (e.g. Frick & Prinz, 2006; Gouguet & Primault, 2006), insolvency declarations in Germany seem very similar to the English or the French case, leading to the conclusion of similar patterns of financial instability (Scelles, Szymanski & Dermit-Richard, 2016; Szymanski, 2017; Szymanski & Weimar, 2017). Apart from that, financial investments in bonds or shares are of rising interest to fans and/or club managers (Weimar & Fox, 2012). Based on this intense discussion over the past years, ‘Financial performance’ is one of two substreams of soccer financial investigations. A recent trend is Financial Fair Play as a new institutional change in European soccer. Due to increasing money injections by external investors and an expected negative impact on long-run competitive balance as well as negative impacts of financial escalations on the enjoyment from soccer, the UEFA set up restrictions regarding outside investments in European soccer (Franck, 2014). Another argument is that money injections might let clubs take more precarious options and investments strategies (Franck & Lang, 2014). However, taking the ‘Matthew-effect’ into account, theoretical discussions already indicated the reciprocal effect by which larger clubs can take even more market share, since small clubs are prevented from breaking out of the ‘Matthew-spiral’ using money injections from investors. The current research on the Financial Fair Play regulations is still of a theoretical nature. However, since the change is relatively recent, there has only been one empirical investigation on the potential effects of Financial Fair Play so far (Peeters & Szymanski, 2014). The second substream of research on soccer finance can be labeled as ‘Soccer and stock Exchange’. Research within this substream is tied into investigations on the determinants (especially sporting success) of soccer club share prices and investigations on the impact of soccer on general share prices. Generally, Initial Public Offerings (IPO) of soccer clubs were rather popular in the 2000s. Today, there are only 22 clubs listed at the stock exchange (see STOXX Europe Football). Regarding the impact of soccer on the non-sport stock market, the literature is divided whether the correlation between soccer success or soccer events and stock market reactions is really a causal relationship (Klein, Zwergel & Henning Fock, 2009). One reason for soccer being popular

246

THE SAGE HANDBOOK OF SPORTS ECONOMICS

for stock market research is the exogenous event of sport success, by which market efficiency and investment behavior research can be done adequately (Berument & Ceylan, 2012).

Demand for Soccer What brings fans to the stadium? This question has always been of interest to both club managers/ owners and researchers. Besides the practical importance, the topic of soccer demand was and still is popular among researchers since proxies for demand in the sense of stadium attendance is recorded back to the beginnings of professional soccer. Available data are an essential requirement for empirical analysis, thus researchers are attracted by such panel data. Furthermore, attendance data are often normally distributed, which makes data analysis straight­forward (except for sold-out stadiums, where censored models such as Tobit models have to be applied). Since 1999, when data on TV audiences became available, empirical research on TV spectators (as a different proxy for demand) is rising. Both TV demand and stadium attendance are to a great extent connected with the concept of uncertainty of outcome.6 Consequently, studies on stadium attendance and TV audience can be further subdivided into research explicitly focusing on the impact of outcome uncertainty and research focusing on the impact of other factors (e.g. the quality of teams, superstars, market sizes, revenues, new stadiums, running tracks, or weather conditions).7 Moreover, since research on uncertainty can be divided into research on uncertainty as a determinant of stadium and TV demand, i.e. the test of the uncertainty of outcome hypothesis, and research on uncertainty at league level as a dependent variable, i.e. the analysis of changes in competitive balance over time (see Fort & Maxcy, 2003), three substreams are distinguished here: ‘Stadium attendance’, ‘TV audience’ and ‘Competitive balance’. Jennett (1984) was the first to test whether outcome uncertainty of soccer matches might affect attendance. Unlike Jennett (1984), Peel and Thomas (1988) found that close games are undesirable. They were also among the first using win probabilities derived from betting odds as a measurement of ex-ante game uncertainty, which has become state of the art when investigating the uncertainty of ­ outcome hypothesis. Afterwards (and still ongoing), a debate on the effects and importance of uncertainty and demand emerged. In  2000, Koning was the first to connect the concept of competitive balance to soccer, which

suggests that a league should have an almost equal distribution of talent among the participating clubs (see Budzinski and Pawlowski (2017) for a recent review of the literature on competitive balance and the ­relevance of uncertainty of outcome).

Betting and Forecasting The soccer betting market has become popular to researchers for three reasons: the availability of data, the worldwide increasing demand for betting and it being the ideal setting to investigate market efficiency. In this regard, Hvattum and Arntzen state (2010, p. 460): ‘Sports betting markets are becoming increasingly competitive. These markets are of interest when testing new ideas for quantitative prediction models.’ Thus, the literature on betting is primarily tied into research on betting odds, namely the use of betting odds for an investigation of ‘Market efficiency’ hypotheses, research on ‘Forecasting’ soccer outcomes to predict and improve betting odds and on ‘Sentiment and soccer pools’ in the betting market.

Penalties While the main streams mentioned above mainly investigate sports data to achieve insights into soccer market behavior, the topic of investigating ‘Penalties’ especially became popular among game theory researchers. The penalty setting seems very fruitful for this research topic, since the setting is almost equal in every game (one ­ goalkeeper, one shooter, 11 metres of distance, a shooting angle of 90°, strict rules on how to execute penalties). In this form, p­ enalties are very similar to repetitive laboratory experiments. Mainly, these game theory papers investigated whether players choose random strategies or not (the Nash equilibrium and mix strategy play). Other papers looked at the first mover advantage, cultural differences or avoidance behavior. The category of penalties is the smallest main stream, however, since the investigations are very specialized, but a separate classification seems reasonable.

Miscellaneous Beyond the main streams identified above, other topics emerged which are generally important although they attract far less research. The first

THE ECONOMICS OF PROFESSIONAL SOCCER

topic that can be identified in this regard is ‘Mega soccer events’, with the majority of studies focusing on the economic and touristic impacts of mega soccer events. This kind of research is of great importance to local and regional economies, since mega events such as the FIFA World Cup or continental championships require changes of urban infrastructure which is mainly financed with public subsidies. Overall, these studies come to the conclusion that promises and ex-ante estimations ­regularly differ from ex-post realities, leading to long-run fixed costs for the public (Schausteck de Almeida, Bolsmann, Marchi Júnior & de Souza, 2015). Second, due to the current rise in v­ iolence before, within and after European ­soccer games, research into hooliganisms and ­ game-day ‘Violence’

247

is beginning to leave theoretical considerations and to focus on empirical works. Third, even though the introduction of the ‘Three point rule’ dates back to 1995, the debate about whether the adoption of the new scoring model from two to three points was effectively more exciting is still ongoing (Hon & Parinduri, 2016). However, while some authors find no effects on excitement measures, such as goals scored or more wins, others find evidence for positive effects (Hon & Parinduri, 2016; Moschini, 2010). Other papers looked at topics such as dynamic pricing, fantasy soccer, social media, beer prices or referendums for stadiums and are classified as ‘Diverse’ in Table 25.1, which provides an overview on the number of soccer economic publications per substream.

Table 25.1  Number of soccer economic publications per substream Main Stream

Substream

Special Topic

Team Productivity

League Teams

Aggressive Play, Cards, Sabotage Home Advantage Managerial Change Divers

Individual Productivity & Soccer Labor Market

National Teams Efficiency Managers and Coaches Referees Players

Soccer Finance Demand for Soccer

Betting and Forecasting

Penalties Miscellaneous

Financial Performance Soccer and Stock Exchange Competitive Balance Stadium Attendance TV Audience Market Efficiency Forecasting Sentiment and Soccer Pools Mega Soccer Events Soccer and Violence Three Point Rule Diverse

Career, Migration and Discrimination Player Performance Transfer Market Wage and Market Value Youth Soccer

Uncertainty of Outcome Other Factors of Stadium Attendance

Obs. 10 13 17 30 26 39 12 25 15 11 12 20 10 12 20 14 25 37 17 15 26 9 19 10 6 8 11

248

THE SAGE HANDBOOK OF SPORTS ECONOMICS

160

Main Streams of Empirical Soccer Economic Research

120

Number of Publications

140

152

100

105

60

80

93

50

40

35

32

Pe

na

lti

es

e nc na Fi er cc So

ce is M

re Fo d

Be

tti

ng

an

lla

ca

ne

st

in

ou

s

g

r ce oc rS fo d an

em D

er cc So

Te a

m

La

Pr

bo

od

rM

uc

ar

tiv

ke

ity

t

0

20

19

Figure 25.1  Number of publications per main stream of empirical soccer economic research

DESCRIPTIVE LITERATURE STATISTICS As depicted in Figure 25.1, most research was done on team productivity, with 152  ­ publications. Research on the soccer labor market (individual productivity and wages of players, transfer market, referee performance, manager and coach performance) and on the demand for soccer (competitive balance, stadium attendance, TV audience) were of similar interest, and were found in 105 and 93 publications. Trailing far behind the leading categories, one can find literature on ­betting and forecasting (50) and soccer finance (32). Studies on penalty shootouts can be classified as the smallest cluster of research. The biggest s­ ubstreams are investigations on league ­performance (87) and team efficiency (39).

Time Trends Looking at the chronological distribution of the identified articles, a constant increase in the interest on soccer economic research is evident (Figure 25.2). In retrospect, the early peaks in 1997, 2000 and 2002 are noticeable. Although any causal explanation is not possible, three milestones of sport economic research occurred in these years:

the first special issues on sports economics by Economic Affairs (1997) and The Statistician (2002) and the first issue of the Journal of Sports Economics (2000). After 2007, there were permanently more than 20 publications per year on empirical soccer economics, indicating the popularity of soccer economics research. While there seems to be an almost linear positive trend on the number of publications in the basic population, some special temporal ‘en vogue’ trends are visible in the chronological distribution of the main streams and substreams (Figure 25.3). In contrast to the interest on team productivity, soccer labor market and demand for soccer (which were already of importance in the early 2000s), other main streams show different ­patterns. In this regard, research on betting and ­forecasting had its peaks around the years 2008–2011. The research setting of penalties was only ‘discovered’ after the year 2007 and the effort shown remains constant up to 2017. Probably rooted in the disclosure of financial information, the topic of soccer finance only became relevant in the late 2000s, had a peak in 2009 and is progressing at the same level since then. Other temporal trends can be seen for the substreams of referee performance (peak 2007–2010) and soccer and the stock exchange (peak 2009–2012). In contrast, research on efficiency peaked from 2006 to 2010, but has had

4

6

8

10

12

2

4

6

8

10

12

14

14

Penalties

Number of of Publications

2

Number of of Publications

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 20 17

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 20 17

0

0

2

4

4

6

6

8

8

10

10

12

12

Number of of Publications

2

Number of of Publications 14

14

Soccer Finance

0

0

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 20 17

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 20 17

0

0

2

4

4

6

6

8

8

10

10

12

12

Number of of Publications

2

Number of of Publications 14

14

Team Productivity

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 20 17

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 20 17

19 0 7 19 5 8 19 2 8 19 4 8 19 7 88 19 8 19 9 92 19 9 19 3 94 19 9 19 5 9 19 6 9 19 7 9 19 8 9 20 9 0 20 0 0 20 1 0 20 2 03 20 0 20 4 0 20 5 06 20 0 20 7 0 20 8 0 20 9 1 20 0 1 20 1 1 20 2 1 20 3 14 20 1 20 5 1 20 6 17

20

30

Number of of Publications

10

40

50

THE ECONOMICS OF PROFESSIONAL SOCCER

Figure 25.3  Time trends within main streams (2017 only considered until June)

249

Year of Publication

Figure 25.2  Empirical soccer economic studies by year of publication (2017 only considered until June) Individual Productivity and Soccer Labor Market

Demand for Soccer

Betting and Forecasting

250

Journals with Empirical Soccer Economic Publications

Sc

ot

tis

Ap

pl App ie d lied Ec 0 M h an Jo on Eco JS ag ur om no E er na ic mi ia l o s la f Le cs nd Po tte rs D litic ec a is l E IJS Eu io n co F ro Ec Ec no pe m o J an o no ono y m m Sp urn T i a or l o he c I ics tM f S nq an Spo tat uir Jo ag rt ist y In urn em s S icia te al n S rn e c at of E tat nt ienc io co isti Qu es n al no cs ar Eu Jo m in ter ro ur ic pe S ly na P oc an l o syc iet Jo f h y u So Fo olo Sp rna cc rec gy l or o La e t, f O b r & ast in Bu p ou si era r E Soc g IM ne t i c e A ss ion on ty Jo an al om ur Jou d Re ic na rn M se s l o al an a f M of ag rch an Sp em ag or e em t M K nt a y e k Jo n n ur Em t M age los na pi at m r Ea l of ica hem ent st Me l E er c at n dia on ics Ec E om on co ic om no s ic mic Jo s ur na l

20

40

60

80

Number of Publications

100

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Figure 25.4  Journals with three or more publications on empirical soccer economic studies a ‘comeback’ since 2015. Due to more and better data on individual performance, investigations on players (wages, productivity, the transfer market) has increased remarkably since 2013. In sum, the research on empirical soccer economics has had a linearly increasing trend, although the focus of research has had several temporal trends and ‘en vogue’ topics.

Journals Figure 25.4 gives an overview of the number of publications per journal. Since Sloane (1971), the topic of soccer had always been of general interest. Therefore, it is not surprising that around 65% (315 articles) of the identified publications have been published in general interest outlets. With 85 publications, however, the Journal of Sports Economics is still the most popular outlet for empirical soccer economic research. Also, The International Journal of Sport Finance can be seen as a leading journal for empirical soccer economic research, with 25 publications, even though the journal was first ­published in only 2006. Next to these field  journals, Applied Economics, Applied Economics Letters and the Scottish Journal of Political Economy are among the top

five journals publishing research on empirical soccer economics, followed by journals such as Managerial and Decision Economics, Economic Inquiry and The Statistician. There are another 84 journals with fewer than four publications, including journals which are among the top 15 economic journals (Bornmann, Butz & Wohlrabe, 2017), such as the American Economic Review (3), Econometrica (2), Economic Journal (2), Journal of Political Economy (1), The Journal of Finance (1) or Journal of Economic Literature (1), and Review of Economics and Statistics (1).

Authors Sport economists strongly favor data on individual productivity. Hence, authors’ own contributions to the field of empirical soccer economics need to be examined. Figure 25.5 summarizes those authors (out of 615 different authors) with 10 or more publications. For anyone starting to work in the field of soccer economics the names of Rob Simmons, David Forrest, Carlos Pestana Barros, Stefan Szymanski, John Goddard, Bernd Frick, Egon Franck or Babatunde Buraimo should be familiar, since these are the researchers with the most publications up to June 2017.

251

THE ECONOMICS OF PROFESSIONAL SOCCER

h sc üe N

D

ob

so

n

as S.

Th

om

o im ra Bu

ck Fr

an

k ic

da od

Fr

rd

i G

an

sk

s Sz

ym

rro Ba

st rre Fo

Si

m

m

on

s

0

5

10

15

20

Number of Publications

25

30

Authors of Empirical Soccer Economic Publications

Figure 25.5  Authors with 10 or more empirical soccer economic publications In addition, Table 25.2 gives an overview of the researchers with the most publications per substream. While some substreams show a rather balanced distribution with regard to the number of researchers involved (e.g. team outcome, player, TV audience), other substreams are dominated by

a small fraction of researchers (e.g. soccer finance, sentiment, stadium attendance, mega events). ­ One explanation could be found in the existing entrance barriers to those research ­markets, which are either given by extraordinary data access or by special and complex data analyzing techniques.

Table 25.2  Authors with three or more publications per substream (in descending order according to the number of publications) Substream

Authors

League Teams and National Teams Efficiency Managers and Coaches Referee Player Financial Performance Soccer and Stock Exchange Stadium Attendance TV Audience Forecasting Sentiment Mega Events

Nüesch, Forrest, Frick, Pollard, Carmichael, Goddard, Schneemann, Scoppa, Simmons, Tena, Thomas, Wicker Barros, García-Cebrián, Espitia-Escuer, Carmichael, Haas, Garcia-del-Barrio, Leach, Thomas Dobson, Dawson Buraimo, Dawson, Simmons Frick, Simmons, Deutscher, Franck, Gerrard, Prinz, Schmidt, Thomas, Torgler Szymanski Berument, Ceylan Simmons, Forrest, Brandes, Buraimo, Franck, Andreff, Goddard, Peel, Scelles, Thomas Buraimo, Schmidt, Schreyer, Simmons, Torgler Dixon, Fenton, Goddard, Neil Forrest, Perez, Simmons Maennig

252

Leagues Considered

M

ul

En 0 g tip Ge lan N le L rma d at e n io ag y na u l t es In ea t. To S ms ur pa na in m en Ita t ly Sc US ot A la N Fr nd et a he nc rla e nd B s Po raz rtu il N ga or l Be wa lg y iu Af m r Tu ica rk Ja ey p Ire an Ar la ge nd U nt Sw ru ina itz gua er y Sw lan e d G den re Fi ece D nlan en d m a C rk hi Au na st ria

20

40

60

80

Number of Publications

100

120

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Figure 25.6  Leagues and nations considered in empirical soccer economic research

LEAGUES AND NATIONS Although most soccer data (especially information at the team level) is available to all researchers around the world, the distribution of research on the league of interest is highly skewed to the right (Figure 25.6). With a wide gap, the English Premier League and the German Bundesliga were of main interest to researchers. Only 30 (17) researchers considered the Spanish La Liga (the Italian Serie A) data, although these leagues are equally important in economic terms. Over time, data from international competitions such as the UEFA Champions League, UEFA Europa League or Copa Lipertadores as well as Major League Soccer is growing in importance. Only 43 studies used data from more than one league. Finally, there seems to be a strong correlation between the nationality of the researcher and the league focused on in the research.

CONCLUSION The existing research on soccer economics is rich and classical literature overviews on soccer

economics are already available. Thus, instead of giving another repetitive overview on the precise content of past literature, this chapter provides a literature structuring and a descriptive analysis of the past literature on empirical soccer economics research that is focused on actual outcomes. As a result, the area of empirical soccer economics is still of growing interest to researchers from both sports economics and general economists. 486 articles (from 615 different authors) were subdivided into six main streams of research, i.e. team productivity, soccer labor market and individual productivity, demand for soccer, betting and forecasting, soccer finance and penalties. It was shown that soccer data are suited for both sports economics journals and general interest journals, since sports data generally provide a good setting for investigations on certain phenomena of human and organizational behavior. Over time, the literature has increased as a whole, although interest in the substreams has fluctuated depending on ‘en vogue’ topics. A further analysis revealed those researchers with the most publications in each substream as well as the fact that the researchers’ nationality strongly determines her or his preference of data, leading to an overrepresentation of English and German soccer.

THE ECONOMICS OF PROFESSIONAL SOCCER

Based on this review, several recommendations for future research can be derived. First, empirical soccer economic research should focus on rising topics of interest, such as the Asian soccer leagues (Szymanski, 2016; Watanabe & Soebbing, 2017). Second, given the increasing importance to find solutions to crime and violence surrounding soccer events, it seems promising to focus on this topic in further empirical investigations. Third, despite its ­popularity in other sport economic settings, fantasy soccer is still a neglected topic, although it offers an interesting setting for soccer economic research (Gonçalves, 2013). Fourth, data from women’s soccer could be used to investigate gender-specific differences in soccer behavior. Fifth, there is also missing empirical evidence on the soft budget constraint assumption within European soccer (Storm & Nielsen, 2012). Sixth, due to its novelty, empirical investigation on the effectiveness of Financial Fair Play is still missing. In some years, ­sufficient data should be available to investigate the effect on competitive balance and financial performance of European ­soccer clubs. Seventh, empirical investigations on corruption also seem necessary (Forrest, 2012). Another recommendation (or rather hope) for future research is to base empirical analysis on different leagues rather than single leagues, if only to increase the external validity of the findings. Related to this, future research should step back from ‘salami’ publication techniques of using all available data in one article instead of publishing several ‘minor’ contributions. While such research strategies increase the list of publications, the marginal utility generated by every additional ‘salami paper’ is, however, often almost zero. Moreover, empirical soccer research should reduce the trend of ‘significosis’ (Antonakis, 2017, p. 7) by emphasizing the economic significance instead of purely looking at the statistical significance (Antonakis, 2017). Finally, while doing the literature structuring, it became evident that the fraction of repetitive studies is quite small. Researchers should not search for novelty effects only (Antonakis, 2017). Repetitive studies possibly generate more impact and benefit to the literature, especially if previous research c­ annot be confirmed (see examples of ­contradictory results in Klein et al., 2009; or Kocher, Lenz & Sutter, 2012).

Notes 1  Search queries on Google Scholar for the expression ‘Economic AND Team AND Club AND management AND Soccer’ OR ‘European football’ showed 5,050 hits for the years 2000– 2005, 13,300 hits for the years 2006–2011 and 16,100 for the years 2012–2017.

253

2  Neglecting studies using non-actual outcomes (survey data) is only motivated by objectively reducing the basic population to an achievable number. The restriction is not driven by any ­subjective preference for actual or survey outcome by the author. Both are seen to be identically important for sports economics research. 3  It is also not distinguished between studies primarily using sports data to investigate general economic phenomena and studies using economic theory to explain sports phenomena. However, most studies of the latter category are published in special sports journals, while the first category is only published in general economic journals. 4  ‘While the DEA focuses primarily on the overall assessment of clubs’ efficiency based on the analysis of sets of inputs and outputs that characterize the production function, the SFA use tools of deterministic correlation and regression analysis to determine the functional dependence of the variable considered in the model’ (Barros & Rossi, 2014, p. 2401). 5  In the ‘Pre-Bosman-Time’, the current holder of a player right could avoid any transfer of the player to an outside club until they found an agreement with the taking club. Afterwards, a club only has control on players’ movements as long as the contract is active. If a contract runs out, players now can move without restrictions (transfer fees) to another club (Feess & Mülheußer, 2002). 6  Rottenberg (1956) and Neale (1964) were the first emphasizing the potential relationship between ex-ante outcome uncertainty of sport competitions and the demand for sport. 7  Since, however, the quality of teams as driver of demand is closely related to uncertainty, most empirical studies on stadium attendance and TV demand incorporate also a measure of uncertainty (e.g. win probabilities derived from betting odds, the difference in league standings or differences in team market values).

REFERENCES Andreff, W., & Szymanski, S. (Eds.) (2006). Handbook on the Economics of Sport. ­Cheltenham, UK: Edward Elgar. Antonakis, J. (2017). On doing better science: From thrill of discovery to policy implications. The Leadership Quarterly, 28(1), 5–21. Barnsley, R. H., Thompson, A. H., & Barnsley, P.  E. (1985). Hockey success and birthdate: The relative age effect. Canadian Association for Health, Physical Education, and Recreation, 51, 23–28.

254

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Barros, C. P., & Rossi, G. (2014). A Bayesian stochastic frontier of Italian football. Applied Economics, 46(20), 2398–2407. Berument, M. H., & Ceylan, N. B. (2012). Effects of soccer on stock markets: The return–volatility relationship. The Social Science Journal, 49(3), 368–374. Bornmann, L., Butz, A., & Wohlrabe, K. (2017). What are the top five journals in economics? A new meta-ranking. Applied Economics. http://dx.doi. org/10.1080/00036846.2017.1332753 Budzinski, O., & Pawlowski, T. (2017). The behavioural economics of competitive ­ balance: Theories, findings and implications. International Journal of Sports Finance, 12(2), 109–122. Dobson, S., & Goddard, J. (2011). The Economics of Football. Cambridge: Cambridge University Press. Dohmen, T. J., & Sauermann, J. (2016). Referee bias. Journal of Economic Surveys, 30(4), 679–695. Ericson, T. (2000). The Bosman case: Effects of the abolition of the transfer fee. Journal of Sports Economics, 1(3), 203–218. Feess, E., & Mülheußer, G. (2002). Economic consequences of transfer fee regulations in European football. European Journal of Law and Economics, 13(3), 221–237. Forrest, D. (2012). The threat to football from betting-related corruption. International Journal of Sports Finance, 7(2), 99. Fort, R., & Maxcy, J. (2003). Competitive balance in sports leagues: An introduction. Journal of Sports Economics, 4(2), 154–160. Franck, E. (2014). Financial Fair Play in E­ uropean club football: What is it all about? International Journal of Sports Finance, 9(3), 193–217. Franck, E., & Lang, M. (2014). A theoretical analysis of the influence of money injections on risk taking in football clubs. Scottish Journal of Political Economy, 61(4), 430–454. Franck, E., & Müller, J. C. (2000). Problemstruktur, Eskalationsvoraussetzungen und eskalationsfördernde Bedingungen sogenannter Rattenrennen. Schmalenbachs Zeitschrift für betriebswirtschaftliche Forschung, 52(1), 3–26. Frick, B. (2007). The football players’ labor market: Empirical evidence from the major European leagues. Scottish Journal of Political Economy, 54(3), 422–446. Frick, B. (2009). Globalization and factor m ­ obility: The impact of the ‘Bosman-ruling’ on player migration in professional soccer. ­Journal of Sports Economics, 10(1), 88–106. Frick, B., & Prinz, J. (2006). Crisis? What crisis? Football in Germany. Journal of Sports ­Economics, 7(1), 60–75. Goddard, J., & Sloane, P. (Eds.) (2014). Handbook on the Economics of Professional F­ ootball. Cheltenham, UK: Edward Elgar.

Gonçalves, R. (2013). Empirical evidence on the impact of reserve prices in English auctions. The Journal of Industrial Economics, 61(1), 202–242. Gouguet, J. J., & Primault, D. (2006). The  French exception. Journal of Sports Economics, 7(1), 47–59. Hon, L. Y., & Parinduri, R. A. (2016). Does the threepoint rule make soccer more exciting? Evidence from a regression discontinuity design. Journal of Sports Economics, 17(4), 377–395. Hvattum, L. M., & Arntzen, H. (2010). Using ELO ratings for match result prediction in association football. International Journal of Forecasting, 26(3), 460–470. Jennett, N. (1984). Attendances, uncertainty of outcome and policy in Scottish league football. Scottish Journal of Political Economy, 31(2), 176–198. Késenne, S. (2006). The win maximization model reconsidered flexible talent supply and efficiency wages. Journal of Sports Economics, 7(4), 416–427. Klein, C., Zwergel, B., & Henning Fock, J. (2009). Reconsidering the impact of national soccer results on the FTSE 100. Applied Economics, 41(25), 3287–3294. Kocher, M. G., Lenz, M. V., & Sutter, M. (2012). Psychological pressure in competitive environments: New evidence from randomized natural experiments. Management Science, 58(8), 1585–1591. Koning, R. H. (2000). Balance in competition in Dutch soccer. The Statistician, 49(3), 419–431. Kuper, S., & Szymanski, S. (2009). Why ­England Lose. London: Harper Collins. Leach, S., & Szymanski, S. (2015). Making money out of football. Scottish Journal of Political Economy, 62(1), 25–50. Marcén, M. (2016). The Bosman ruling and the presence of native football players in their home league: The Spanish case. European Journal of Law and Economics, 42(2), 209–235. Moschini, G. (2010). Incentives and outcomes in a strategic setting: The 3-points-for-a-win system in soccer. Economic Inquiry, 48(1), 65. Neale, W. C. (1964). The peculiar economics of professional sports. The Quarterly Journal of Economics, 78(1), 1–14. Peel, D., & Thomas, D. (1988). Outcome uncertainty and the demand for football: An analysis of match attendances in the English football league. Scottish Journal of Political Economy, 35(3), 242–249. Peeters, T., & Szymanski, S. (2014). Financial Fair Play in European football. Economic Policy, 29(78), 343–390. Prinz, J. & Weimar, D. (2017). The golden generation: The personnel economics of youth recruitment in European professional soccer. Working Paper.

THE ECONOMICS OF PROFESSIONAL SOCCER

Radoman, M. (2015). Labor market implications of institutional changes in European football: The Bosman ruling and its effect on  productivity and career duration of ­ players. Journal of Sports Economics. First published 30 July. https://doi. org/10.1177/1527002515594555. Rosen, S. (1981). The economics of superstars. The American Economic Review, 71(5), 845–858. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Santos, J. M. S., & García, P. C. (2011). A bibliometric analysis of sports economics research. International Journal of Sports Finance, 6(3), 222. Scelles, N., Szymanski, S., & Dermit-Richard, N. (2016). Insolvency in French soccer: The case of payment failure. Journal of Sports Economics. First published 15 November. https://doi.org/10.1177/ 1527002516674510. Schausteck de Almeida, B., Bolsmann, C., Marchi Júnior, W., & de Souza, J. (2015). Rationales, rhetoric and realities: FIFA’s World Cup in South Africa 2010 and Brazil 2014. International Review for the Sociology of Sport, 50(3), 265–282. Simmons, R. (1997). Implications of the Bosman ruling for football transfer markets. Economic Affairs, 17(3), 13–18. Sloane, P. (1969). The labour market in professional football. British Journal of Industrial Relations, 7(2), 181–199. Sloane, P. J. (1971). The economics of professional football: The football club as a utility maximiser. Scottish Journal of Political Economy, 18(2), 121–146. Sloane, P. J. (1983). The economic crisis in professional football. Economic Affairs, 3(4), 273–275.

255

Sloane, P. J. (2015). The economics of professional football revisited. Scottish Journal of Political Economy, 62(1), 1–7. Storm, R. K., & Nielsen, K. (2012). Soft budget constraints in professional football. European Sport Management Quarterly, 12(2), 183–201. Sutherland, R. J. (1988). The labour market in professional football. Management Research News, 11(1/2), 5–6. Szymanski, S. (2016). Professional Asian football leagues and the global market. Asian Economic Policy Review, 11(1), 16–38. Szymanski, S. (2017). Entry into exit: Insolvency in English professional football. Scottish Journal of Political Economy. First published 23 May. doi: 10.1111/sjpe.12134. Szymanski, S., & Weimar, D. (2017). Insolvencies in professional sports: Evidence from German football. Working Paper. Villa, G., & Lozano, S. (2016). Assessing the scoring efficiency of a football match. ­European Journal of Operational Research, 255(2), 559–569. Watanabe, N., & Soebbing, B. (2017). Chinese Super League: Attendance, pricing, and team performance. Sport, Business and Management, 7(2), 157–174. Weimar, D., & Fox, A. (2012). Fananleihen als Finanzierungsmöglichkeit von Sportclubs? Eine Bestandsaufnahme am Beispiel der Fußballbundesliga. Corporate Finance Biz, 3(4), 181. Wilders, M. G. (1976). The football club ­manager: A precarious occupation? Journal of Management Studies, 13(2), 152–163. Williams, A. M., & Reilly, T. (2000). Talent i­ dentification and development in soccer. Journal of Sports Sciences, 18, 657–667.

26 The Economics of Cricket Ian Gregory-Smith, David Paton and Abhinav Sacheti1

INTRODUCTION Professional cricket is one of the most popular team sports in the world. It is the dominant sport in the Indian subcontinent and is also very popular in England, where the sport originated, and several other countries in the Commonwealth such as Australia, South Africa, Zimbabwe and a number of nations in the Carribean. In 2015, an estimated audience of a billion people watched a live broadcast of the match between India and Pakistan in the Cricket World Cup (Berry, 2015). Despite this, economists have generally paid less attention to cricket than to other sports, such as football, baseball, basketball or American football. This chapter argues that cricket has a number of features which make it conducive to economic analysis. The sport is structured in such a way that the public is offered differentiated products, from Test matches between countries which can last up to five days to domestic competitions comprising games lasting just a few hours. Further, these products have been subject to both product and technological innovation (Borooah, 2016; Shivakumar, 2018). At a micro level, playing the game involves a series of interesting decisions made by both officials and players under varying information sets and with uncertain outcomes. Many of these decisions are discrete in nature and outcomes depend not only on physical performance but also on chance, climatic and ground conditions as well as strategic

awareness. As a result, intellectual skills relating to judgement of risk and reward are at a premium. The rapid rate of innovation in the structure of cricket in recent years has also provided researchers with a number of naturalistic experiments within which aspects of decision-making can be analysed (Sacheti, Gregory-Smith, & Paton, 2015). Considering also the amount of data generated by cricket matches, cricket provides a rich context for conducting economic research (Bhaskar, 2009). The aim of this chapter is to highlight some of the more significant contributions made by economists to the analysis of cricket. In the next section, cricket’s historical and institutional background is sketched out. Then an overview of the cricket economics literature is provided. Next, the issue of bias among cricketing officials (umpires) is explored and some new data are used to analyse recent changes in decision-making technology. Finally, some implications for cricket administrators are identified as well as suggestions for future research by economists.

CRICKET: OVERVIEW AND DEVELOPMENTS Description of Cricket Cricket is a bat and ball game played between two teams of eleven players each. Each team alternates

THE ECONOMICS OF CRICKET

between ‘batting’ and ‘bowling’ (or ‘fielding’), with the decision to bat or bowl first made by means of a coin toss before a match begins. The game itself is built up by ‘overs’ of six balls each, in which one player from the bowling side bowls six legal balls in succession to the batting side. Although the earliest form of cricket is believed to have originated in southeast England in the thirteenth century, games played in a form recognizable to the current sport can be traced back to the eighteenth century. Birley (2003) provides a full account. At present, international cricket is played in three formats: the Test format, the One-Day International (ODI) format and the Twenty20 International (T20I) format. Each of these three formats has a counterpart in domestic cricket within the Test-playing nations. It is at the domestic level that most cricketers are first able to earn a living playing the sport. The Test format has been played since 1877 and is the earliest format of international cricket. It involves two batting ‘innings’ per team, and is peculiar in its length of playing time, which has varied between three days and ‘timeless’ over the history of the sport, but is now standardized at five days. Notably, despite the length of playing time, Test matches still frequently end up in a ‘drawn’ game with neither side victorious. The ODI format was introduced in 1971. ODI games are played only one innings a side, with each innings capped at a maximum number of overs (the cap has been fifty overs for over two decades) and are designed to be completed within a day (matches typically last no longer than seven to eight hours) and to produce a winner. Over three decades later, the T20I format was introduced in 2005. Played over a maximum of twenty overs per innings, the format can be described as a shorter version of ODI cricket, and typically lasts no longer than three to four hours.

Changes to the Structure and Regulations of Cricket The International Cricket Council (ICC) is the governing body of international cricket. The ICC is responsible for staging all global tournaments, such as the Cricket World Cup, and also selects match officials for all sanctioned international matches. The organization of domestic cricket and the national team is the responsibility of national cricket boards, as are all bilateral fixtures between cricket-playing countries. National cricket boards are the sole ‘buyers’ for professional cricketers who wish to ‘sell’ their services in the market for international cricket selection. This framework

257

has been challenged twice by unsanctioned cricket tournaments (the World Series Cricket (WSC) tournament, which begain in 1977 and the Indian Cricket League (ICL) tournament, which began in 2007 (Parker, Burns, & Natarajan, 2008)). In both cases, cricket administrators responded by banning participating players. As a result, both leagues were short-lived, but several of their innovations, such as day/night matches and coloured uniforms for players (rather than white clothing) now feature prominently in contemporary cricket. More recently, officially sanctioned domestic T20 leagues have begun to challenge the dominance of international cricket. Although T20 cricket started with an English county competition in 2003, a significant development took place in 2008 when the Board of Cricket Control in India (BCCI), the governing body of Indian cricket, allowed private ownership of franchisee-based teams in a domestic T20 league called the Indian Premier League (IPL). This is not the only novelty. Players are assigned to teams through an auction, and the resulting salaries can at times be in excess of those available from international cricket. The IPL’s success has spawned several other leagues with a similar business model, including the Big Bash League (BBL) in Australia, the Bangladesh Premier League (BPL), the Caribbean Premier League (CPL) and the Pakistan Super League (PSL). These leagues are threatening the historical primacy of international cricket, and several commentators have raised concerns that the significant salary differential between T20 leagues and Test cricket could eventually even lead to the demise of Test cricket (White, 2010; Fletcher, 2017). As well as innovations in the formats of the game, regulations affecting decision-making have also changed over the history of cricket. Prior to the 1990s, umpires (match officials) in all international cricket matches were from the hosting country. However, home umpires from a number of countries faced persistent questions regarding their objectivity and competence.2 In response, the ICC mandated one neutral umpire for Test matches held in the early 1990s and, by 2002, ruled that both umpires should be from neutral countries. ODI cricket continues to feature one home and one neutral umpire. In 2009, the ICC began formally including technology in the decision-making process through the Decision Review System (DRS), which allows teams to challenge a fixed number of umpire decisions in each match by seeking assistance from an off-field umpire (the third umpire, or TV umpire), who has access to television replays of the on-field activity. Other regulatory changes include the pre-match coin toss in which the captain winning the toss has the right to choose whether to bat or field first.

258

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The coin toss is thought to play an important role in the outcome of the match. Although the statistical evidence is somewhat mixed (Bhaskar, 2009; Sacheti, Gregory-Smith, & Paton, 2016b), the issue is considered significant enough to attract regulatory attention. For instance, in 2015, the England and Wales Cricket Board (ECB) ruled that the visiting team would be offered the chance to bowl first in English first-class cricket and a toss used only if this option is declined. Many of these decision-making features and innovations have been subject to economic analysis and we now turn to an overview of this academic literature.

THE CRICKET ECONOMICS LITERATURE Much of the work by economists on cricket can be conveniently divided into two lines of inquiry: 1 Demand, structure and organization of cricket 2 Decision-making in cricket Although both areas are surveyed, the focus in this chapter will be on the second. It is argued that it is within the study of decision-making where the empirical setting of cricket is particularly favourable relative to other sports. Discrete cricketing decisions are regularly made by key actors, such as leaders (cricket captains) and officials (cricket umpires), under reasonably controlled conditions. As such, it is possible to gain insights into potential biases that may be present in decisions, sometimes by exploiting exogenous rule changes which affect decisions. There are also other cricket-related fields of inquiry with a more limited literature which are briefly surveyed at the end of this section.

Demand, Structure and Organization While research on cricket has a long history,3 the literature on cricket demand has received relatively little attention. Out of nearly 50 studies on demand for sport reviewed by Borland and Macdonald (2003), only one is on cricket, this being Hynds and Smith (1994). To the authors’ knowledge, there are now in fact at least eight other published studies on the demand for international and domestic cricket – Schofield (1983); Chapman, Fisher & Maloney (1987); Bhattacharya & Smyth (2003); Blackham & Chapman (2004); Paton & Cooke (2005); Morley & Thomas (2007); Sacheti, Gregory-Smith & Paton (2014, 2016a).

Estimates of demand have typically employed Ordinary Least Squares (OLS) or Generalized Least Squares (GLS) regressions and have attempted to accommodate the various atypical aspects of cricket demand. First, watching live cricket carries a high opportunity cost in that spectators have to forgo considerable amounts of time, with the exception of T20I matches (Paton & Cooke, 2011). Second, prices for match day attendance can be endogenous to the quality of opposition. For example, cricket authorities in Australia set prices according to the strength of the opposition in order to maximize revenue (Bhattacharya and Smyth, 2003). Third, match day attendances can be censored due to venue capacity constraints, as is the case in England (Paton & Cooke, 2005), although this is far less likely in countries such as Australia due to the much bigger stadiums (Sacheti et al., 2014, 2016a). Like in other sporting settings, a stream of the cricket demand literature has considered the notion of competitive balance or uncertainty of outcome (e.g., Rocke, Ramkissoon, Iton & Khan, 2016). Debates on team strengths led the ICC to introduce an official ratings system, but consideration of outcome uncertainty has been limited to relatively short-run measures of uncertainty, such as match uncertainty and series uncertainty (Hynds & Smith, 1994; Bhattacharya & Smyth, 2003). Sacheti et al. (2014, 2016a) examine longer-run measures of uncertainty and find support for the uncertainty of outcome hypothesis in England but not in Australia.

Decision-making in Cricket Sporting settings such as the draft system in American football (Massey & Thaler, 2013) have been used to examine how actors depart from the axioms of rational decision-making. Decisionmaking in cricket is particularly well suited to economic analysis. The decisions are often discrete (out or not out, bat or field); governed by precise rules and repeated numerous times over the course of a season; the outcomes are publicly available and are feasible to assemble into a dataset. Further, decision-making in cricket offers analogies to behaviour in wider microeconomic contexts. For example, decision-making by cricket captains can parallel the behaviour of corporate leaders. In both settings, agents are making complex decisions under uncertainty on behalf of their organization and these decisions draw upon experience, skill, risk preferences and are influenced by the incentives (possibly financial) facing the agent. The impact of the coin toss and resulting decision by the captain to bat or field first has been analysed in each format of the game, typically by employing

THE ECONOMICS OF CRICKET

logit or probit regression analysis. Identification of impact on outcome is achieved by exploiting the random assignment of the decision following the coin toss. Overall, winning the toss appears to marginally increase the probability of winning a match, albeit results vary according to the specific empirical setting. Internationally, de Silva and Schwartz (1998), Bhaskar (2009) and Dawson et  al. (2009) analyse the impact of the toss in ODI cricket, with the last of these finding a stronger effect in day/ night matches. Allsopp and Clarke (2004) use data from both ODI and Test matches and find no effect of winning the toss on match outcome. Sacheti, Gregory-Smith and Paton (2016b) focus on T20I cricket and present evidence to suggest that captains do not always choose optimally after winning the toss when under pressure from external commentators. Domestically, Morley and Thomas (2005) and Forrest and Dorsey (2008) find winning the toss increases the probability of winning but to a lesser extent than other variables, such as home advantage, the weather and the strengths of the teams. The studies on umpire decision-making in Test cricket have focused on the Leg Before Wicket (LBW) dismissal, and whether this decision is being made optimally or exhibits bias. The LBW decision requires a high degree of judgement by the umpire in a very short period of time and can dramatically shift the momentum of a match. The early literature compared whether LBW decisions against home and away teams differ statistically significantly, using data from either one country e.g., Australia (Sumner & Mobley, 1981; Croucher, 1982; and Crowe & Middeldorp, 1996), several countries (Ringrose, 2006) or domestic cricket (Jones, Bray, & Bolton, 2001). Although these studies tend to find that batsmen of home teams are given out LBW less frequently than away teams, on average, it proved difficult to distinguish between umpire bias and pressure from home crowds as the cause of the difference. More recent evidence has exploited changes in umpiring regulations and decision-making technology to deepen our understanding of umpiring anomalies and the issue is considered in more detail below.

2

3

4

5

259

in rain-affected matches. The Duckworth-Lewis method was first used in international cricket in 1997 and remains in use for most one-day games. Other studies include Preston and Thomas (2000), who consider domestic limited overs cricket in England; Brooks, Faff and Sokulsky (2002), who predict Test cricket outcomes using an ordered response model; Norton and Phatarfod (2008), who look at ODI matches; and Akhtar and Scarf (2012), who use logistic regression models to forecast Test match outcomes by session. Studies that provide quantitative measures to assess cricketers’ performances. Barr and Kantor (2004) provide a criterion for comparing and selecting batsmen in ODI cricket, while Rohde (2011) applies the concepts of opportunity costs and supernormal profits to batting performance to produce a cardinal ranking system for players. Chedzoy (1997) examines the impact of umpiring decisions on player performance. Bullough et  al. (2016) look at the impact of central contracts on player performances. Studies that use cricket as a setting to examine wider economic issues. Mandle (1972) provides an overview of the cricket labour market in its early history while Aiyar and Ramcharan (2010) attempt to identify the role of fortune in labour markets by separating the impacts of luck and ability on an individual’s career progression. Mishra and Smyth (2010) look at the effect of the Indian cricket team’s performance in ODI matches on returns on the Indian stock market. Studies that develop models for the structure of international cricket. Preston, Ross and Szymanski (2000) presented a model for an international club championship. Studies that use T20 league auctions to draw insights on player valuations and on the economics of auction theory. Parker et al. (2008) assess the IPL auction in the wider setting of the economics of auctions. A small literature has looked at the determinants of player valuations, including Karnik (2010) and Lenten, Geerling and Kónya (2012).

OTHER THEMES IN CRICKET RESEARCH A number of other issues have been studied in the context of cricket. Among these are: 1 Studies that assess batting strategy and forecast match outcomes. Of particular note is Duckworth and Lewis (1998),4 who develop a method to reset targets and determine a winner

DECISION-MAKING ANOMALIES: THE LATEST EVIDENCE The sports economics literature has identified a tendency for sporting officials to favour home teams (Garicano, Palacios-Huerta, & Prendergast, 2005;

260

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Buraimo, Forrest, & Simmons, 2010; Dohmen & Sauermann, 2016). Cricket is a particularly interesting laboratory for studying the issue for two reasons. First, cricket umpires are able to apply a high level of subjective judgement in decisions that can be critical for match outcomes. Second, regulatory changes to the appointment and functioning of umpires over the last two decades provide natural experiments that can help to identify the source of any bias. In cricket, players first appeal to the umpire for a decision. Umpires have significant incentives to award the correct decision. Aside from the inherent satisfaction associated with making a correct decision, the more decisions umpires get correct, the more likely they are to be retained by the ICC. Nevertheless, umpires may exhibit a conscious or subconscious preference to favour home teams. In other sports, home support from large crowds has been shown to exert pressure on neutral officials (e.g. Buraimo et  al. (2010) find that home teams in the Bundesliga, the German football league, receive fewer yellow and red cards when crowds are closer to the field of play). To distinguish between these two sources of bias, Sacheti et  al. (2015) exploit a regulatory change which required first one and subsequently both umpires to be neutral (i.e. to be of a different nationality to both the playing teams) in Test cricket. Robust evidence is found that favouritism towards home teams was not associated with crowd size and, indeed, could be attributed almost exclusively to the presence of ‘home’ umpires. The implication is that umpires, on average, display favouritism to the team of their nationality, although it is impossible to know if this bias is conscious or subconscious. A more recent development in officiating allows further insights into anomalies in the decisionmaking process. As noted above, the use of technology supports umpire decision-making through the DRS. An obvious line of inquiry is whether or not the DRS corrects the bias of home umpires and, if so, how it has achieved this. Although Test matches are now officiated by two neutral umpires, ODIs continue to have one home and one neutral umpire. For this reason, presented below are some preliminary results on a new sample of data on umpiring decisions in ODIs.

The Decision Review System: Context and Literature International cricket has been one of the leaders in the use of technology in the umpire decisionmaking process (other sports, such as football, have adopted technology later). Following a trial of the system in 2008, the DRS has been gradually

introduced into the majority of international cricket matches. The system was welcomed at a relative early stage in its history by most Testplaying countries, the one exception being India, who refused to allow the use of DRS in their matches until late 2016 (Hoult, 2016). Some observers retain doubts about the quality and accuracy of the technology used, which includes computerized projections of the path of the ball and a device that detects noise to try to identify whether the ball hit the bat. Using a Bayesian approach, Borooah (2016) argues that rather than taking the DRS decision as infallible, umpires should recognize it as a tool for updating their prior belief as to whether the batsman was out or not out. Borooah concludes that in its current format, the marginal improvement to decision-making from DRS is not justified by the large set-up costs. Shivakumar (2018) analyses 912 LBW DRS referrals in Test matches between 2009 and 2014. The preliminary analysis below, although presently on a much smaller scale, advances Shivakumar’s analysis in two important directions. First, since Sacheti et al. (2015) provide evidence that decisionmaking varies between neutral and home umpires, data are collected separately for on-field home and neutral umpires. Second, Shivakumar restricts his analysis to on-field decisions that are challenged, while data on all appeals, including those that are not reviewed, are collected here.

Data We have data, collected by hand, on LBW decisions in ODIs played since 2015. Our sample comprises 20 matches where DRS was used and 20 where it was not. All matches are between Test-playing countries. For each match, we record the details of the match (e.g. venue, date, home and away teams, total runs scored and total overs bowled in each innings, etc.) and data relating to umpiring decisions. These include data on all 57 LBW wickets in this sample, at what stage each LBW decision occurred and which umpire made the decision along with their nationality. For the DRS matches, we also have data on every LBW appeal, 135 in total, as well as whether the appeal was given out or not out, whether this decision was reviewed and, finally, whether the review was upheld. In each case we also record the bowler and batsman involved.

Neutral Umpires and DRS Table 26.1 reports the data on LBW appeals in DRS matches along with whether the appeal was reviewed and, if so, whether it was upheld or

THE ECONOMICS OF CRICKET

overturned. Separate reports for decisions made by home and neutral umpires and also whether the decision concerned a home or away batsman are provided. In Table 26.2, data on LBW wickets broken down by DRS and non-DRS games are presented. Decisions are reported separately for home and neutral umpires and by whether a home or away batsman was given out. The first item of note in Table 26.1 is that the overwhelming majority of appeals are not reviewed. Of the 135 appeals, only 28 (21%) were reviewed. Further, nearly 80% of appeals (106 of 135) were given ‘not out’ by the on-field umpires and only 12 of these 106 (11%) were reviewed. In contrast, ‘out’ decisions are reviewed more frequently with 16 of 29 (55%) ‘out’ decisions reviewed. It is the bowling side that initiates the appeal and there is no limit on the number of appeals that can be made. As a result, many appeals will be of little merit and it is unsurprising that a relatively large proportion of the ‘not-out’ decisions go uncontested. Moving to review success rates, ‘out’ decisions are reviewed with much greater frequency than ‘not-out’ decisions. Ten of the 16 on-field ‘out’ decisions were overturned on review while only 1 of the 12 ‘not-out’ decisions was overturned. To some extent, this reflects the informational differences between a batting and bowling team at the moment when a decision to review is being made. For example, a batsman will be almost certain whether or not he has edged the ball prior to striking the pads, which automatically overturns an LBW. In contrast, reviews of ‘not-out’ decisions are more likely cast in hope than expectation. However, there could be other factors at play. Of particular interest is the high frequency of successful reviews in this small sample: 62.5% ‘out’ decisions are successful. This contrasts with Shivakumar’s (2018) finding that 28% (119 of 420) of decisions were overturned in Test matches. An important difference between this empirical setting of ODIs and Shivakumar’s is that on-field decisions are made by one home umpire and one neutral umpire in ODIs, whereas in Shivakumar’s sample of Tests, both umpires are neutral. Two additional variables are thus brought to the analysis: (1) whether the on-field decision was given by the home umpire and (2) whether the decision is in favour of the home team. An important question is whether this sample displays the same evidence of home favouritism as that found in previous work, such as Sacheti et al. (2015). To answer this, the differences between games where DRS operated and where it did not are considered (see Table 26.2). Home batsmen are given ‘out’ less often by on-field home umpires but only when DRS is not available. Away batsmen

261

are given ‘out’ more often by the on-field home umpire than the on-field neutral umpire, but the effect is much larger where DRS is not available. Together, this suggests that the presence of DRS could constrain the favouritism of home umpires towards home teams. However, we do find some evidence of favouritism even in DRS games. It appears that decisions by on-field home umpires are overturned with greater frequency than neutral umpires. If the post-DRS decision is the correct one (noting the objection by Borooah (2016) above), the original decisions of home umpires were correct 89% of the time (64 of 72) while neutral umpires were correct 95% (60 of 63) of the time. Furthermore, looking more closely at the times when home umpires ‘got it wrong’, three of these decisions were against home teams whereas five of these decisions were against away teams. In contrast, neutral umpires made a mistake against the away team only once. Of course, whether these differences hold up in a larger sample remains to be seen. To be clear, other explanations for the differences are not ruled out. It may be neutral umpires are simply better umpires. Neutral umpires are selected by the ICC from a pool of umpires known as the ‘Elite Panel of ICC Umpires’, which is considered the pinnacle of the profession for cricket umpires and hence it may be the case that they may make fewer mistakes than home umpires, on average.

Summary This preliminary exploration of ODIs by umpire nationality suggests that on-field home umpires have a tendency to give the away team out more frequently than on-field neutral umpires. This is consistent with home favouritism, as found in Sacheti et al. (2015) in the context of Test matches. When available, DRS appears to help correct this bias. In the final reckoning, the away team and the home team have the same number of dismissals by LBW. This provokes an interesting question. Why would home umpires favour home teams in a setting where they know they can be found out? Perhaps this suggests that the favouritism exhibited by home umpires occurs subconsciously. Another question that emerges from the analysis is whether the teams adjust their decision to review based on their beliefs as to whether a particular umpire is likely to be unbiased. Indeed, the preliminary results suggest that away batsmen review the decision of the home umpire more than they do that of the neutral umpire. Eleven out of 15 ‘out’ decisions given by the home umpire were reviewed. In eight out of these 11 cases, it was the away team

106 94 12 11 1

All given out Not reviewed Reviewed  Upheld  Overturned

42 36 6 5 1

12 7 5 1 4 (39.6%) (38.3%) (50.0%) (45.5%) (100.0%)

(41.4%) (53.8%) (31.3%) (16.7%) (40.0%)

Home batsmen

64 58 6 6 0

17 6 11 5 6 (60.4%) (61.7%) (50.0%) (54.5%) (0.0%)

(58.6%) (46.2%) (68.8%) (83.3%) (60.0%)

Away batsmen

57 48 9 8 1

15 4 11 4 7

Total

21 15 6 5 1

6 3 3 1 2 (36.8%) (31.3%) (66.7%) (62.5%) (100.0%)

(40.0%) (75.0%) (27.3%) (25.0%) (28.6%)

Home batsmen

36 33 3 3 0

9 1 8 3 5 (63.2%) (68.8%) (33.3%) (37.5%) (0.0%)

(60.0%) (25.0%) (72.7%) (75.0%) (71.4%)

Away batsmen

49 46 3 3 0

14 9 5 2 3

Total

21 21 0 0 0

6 4 2 0 2

(42.9%) (45.7%) (0.0%) (0.0%)

(42.9%) (44.4%) (40.0%) (0.0%) (66.7%)

Home batsmen

28 25 3 3 0

8 5 3 2 1

(57.1%) (54.3%) (100.0%) (100.0%)

(57.1%) (55.6%) (60.0%) (100.0%) (33.3%)

Away batsmen

Neutral umpires

Notes: (i) Data is derived from 20 ODIs that took place between January and February 2017. (ii) Source is match commentary and scorecards from www.espncricinfo.com with some supplementary information taken from www.cricbuzz.com, collected by the authors.

Given not out

Given out

29 13 16 6 10

All given out Not reviewed Reviewed  Upheld  Overturned

Total

Home umpires

.

All umpires

Table 26.1  Summary of LBW appeals in DRS matches

262 THE SAGE HANDBOOK OF SPORTS ECONOMICS

57 28 29

48 25 23

Out–all Final decision Out–home batsmen Out–away batsmen

24 11 13

30 12 18 (50.0%) (44.0%) (56.5%)

(52.6%) (42.9%) (62.1%)

Home umpire

24 14 10

27 16 11 (50.0%) (56.0%) (43.5%)

(47.4%) (57.1%) (37.9%)

Neutral umpire

20 9 11

29 12 17

Total

9 5 4

15 6 9 (45.0%) (55.6%) (36.4%)

(51.7%) (50.0%) (52.9%)

Home umpire

11 4 7

14 6 8 (55.0%) (44.4%) (63.6%)

(48.3%) (50.0%) (47.1%)

Neutral umpire

DRS matches

28 16 12

28 16 12

Total

15 6 9

15 6 9

(53.6%) (37.5%) (75.0%)

(53.6%) (37.5%) (75.0%)

Home umpire

13 10 3

13 10 3

(46.4%) (62.5%) (25.0%)

(46.4%) (62.5%) (25.0%)

Neutral umpire

Non-DRS matches

Notes: (i) Data is derived from 20 ODIs using DRS and 20 in which DRS was not used. The non-DRS matches took place between March 2015 and October 2016. The DRS matches took place between January and February 2017. (ii) Source is match commentary and scorecards from www.espncricinfo.com with some supplementary information taken from www.cricbuzz.com, collected by the authors.

Original decision

Out–all Out–home batsmen Out–away batsmen

Total

All matches

Table 26.2  Summary of LBW wickets in DRS and non-DRS matches THE ECONOMICS OF CRICKET 263

264

THE SAGE HANDBOOK OF SPORTS ECONOMICS

that initiated the review. This compares to a review of only five of 14 ‘out’ decisions given by the neutral umpire, of which only three were instigated by the away team. In other words, the away batsmen use their reviews to challenge the home umpire versus the neutral umpire at a ratio of 8:3. There are other interesting questions that could be explored. The decision to initiate a review has not been theorized in the literature. An unsuccessful review carries an opportunity cost since the number of reviews is finite, but the cost diminishes as the innings approaches an end (or every 80 overs in Test matches). Also, the value of overturning a decision is increasing in the expected score of the batsman. Both these factors may influence the original decision of the umpire but exactly how is currently unknown. It may be that umpires are more willing to give a star batsman out, now they are supported by DRS. Alternatively, they may be less willing to give a star batsman out, knowing that a ‘not-out’ is less likely to be reviewed and overturned than an ‘out’ decision.

POLICY IMPLICATIONS AND SUGGESTIONS FOR FUTURE RESEARCH The cricket economics literature offers several insights for the ICC, national cricket boards and other key stakeholders in the sport. For example, the away team has been given the choice of bowling first in English first-class cricket (to prevent home teams from deliberately preparing favourable pitches), in which case no coin toss occurs. However, the empirical evidence reviewed above suggests that such a policy change is unlikely to improve competitive balance substantially because winning the toss only marginally increases the probability of winning. On the demand side, the finding that interest in ODI cricket in England benefits from longer-run uncertainty of outcome suggests English administrators could prioritize matches against teams likely to provide a close contest. However, these findings need to be set against the argument that a more equitable structure of the sport requires presently weaker teams to receive exposure to better teams in order to develop in the future. This is of particular significance given protracted discussions between the ICC and national boards on the appropriate revenue-sharing model for international cricket as well as debates on the inclusion of non-Test-playing nations in global tournaments such as the World Cup and Champions Trophy, as well as the proposed inclusion of cricket in the Olympic Games.

Looking at decision-making technology, there are several interesting implications of the existing research for the ICC. For one, the use of home umpires in ODIs post-DRS could offer insight to inform the debate over whether or not to reintroduce home umpires in Test matches. The preliminary analysis of ODIs post-DRS in this chapter suggests that home teams receive an advantage when DRS is not in place, suggesting the ICC should not allow teams the choice of whether to use DRS. Finally, it is important for the ICC to carefully review the statistical evidence on the extent to which the DRS helps improve decisions, such as by reducing anomalies between home and neutral umpires, especially in light of the set-up costs for DRS. This can be used to assess whether the benefits of the DRS outweigh its costs. To conclude the chapter, some suggestions for areas of cricket research, which are likely to be fruitful for economists are offered. Recent upheavals in the revenue-sharing arrangements from shared television rights for international cricket matches (in the last few years, two significantly different revenue-sharing models have been put forward at the ICC:5 BBC, 2017; ESPNcricinfo, 2017; Gollapudi, 2017) highlight the need for an academic study that provides a comprehensive analysis of the drivers of revenues from international cricket and the various domestic T20 leagues. The literature on demand for cricket has focused predominantly on datasets recording attendance in England and Australia. This is at odds with the popularity of the sport in the Asian subcontinent. While data can sometimes be difficult to obtain in these countries, it is not reasonable to assume that past results generated with English and Australian data will automatically carry over to other countries. In particular, the IPL affords new opportunities for research on competitive balance, as it has now been played for twelve completed seasons and six different teams have won the IPL championship at the time of writing. The IPL uses a franchisee model with shared revenues from television rights, so there is also scope for research on optimal revenue-sharing arrangements, drawing on lessons learned from other sports. The IPL and other T20 leagues could also offer other interesting avenues for research in terms of their impact on product quality (e.g. how does participation in T20 leagues affect player performances in international cricket, particularly T20I cricket?) and product demand (e.g. have attendances and consumption of other media offering international cricket changed after the introduction of T20 leagues?). When measuring the demand for cricket, studies have limited themselves to attendance demand. Ideally, wider measures, such as television

THE ECONOMICS OF CRICKET

viewing figures, should be used. Viewership could even be examined in real time to help explore the extent to which demand varies according to the state of play in a match. Such analysis will become more feasible as streaming services of international cricket matches over the internet become commonplace. How do these different modes of consuming cricket interact? Are internet streaming services complements or substitutes to live attendance? Future work on decision-making would benefit from a better appreciation of the influence of live crowds and TV audiences on officials. In addition, there have been a number of recent instances of corruption in professional cricket, including several in T20 leagues. These have not yet been the subject of academic research and an obvious avenue to explore is any links between corruption and anomalies in decision-making by officials. There are already signs in the past few years of a growing interest in cricket among economists. Potential exists for economists to derive significant insights which will be of interest not only to sports economics but also to the more general field of the microeconomics of behaviour and decision-making.

Notes 1  We would like to thank Usha Sacheti for kindly assisting with data collection. 2  Perhaps the most notorious incident was during England’s 1987 tour of Pakistan between England captain Mike Gatting and Pakistani umpire Shakoor Rana (see Johnson, 1989). 3  Statistical analysis of cricket in the academic literature dates back to at least Elderton and Wood (1945) and Elderton (1945). Sloane (1976) included cricket as part of wider research on team sports. Work on the structure of the professional cricket industry can be traced back to Schofield (1982). An earlier overview of the literature is provided by Preston (2006). 4  Following Duckworth and Lewis’s retirements, Professor Steven Stern became the custodian of the method, and in November 2014, the method was renamed the Duckworth-Lewis-Stern method (or D/L/S method). 5  The first model was proposed in 2014. It involved distributing a significantly higher proportion of ICC revenues to India, England and Australia. In 2017, the ICC put forward an alternative approach which reduced the contributions to India and England in percentage terms, while making more provision for non-Test-playing countries.

265

REFERENCES Aiyar, S., & Ramcharan, R. (2010). What can international cricket teach us about the role of luck in labor markets? International Monetary Fund Working Papers, 10(225), 1–29. Akhtar, S., & Scarf, P. (2012). Forecasting test cricket match outcomes in play. International Journal of Forecasting, 28(3), 632–643. Allsopp, P.E., & Clarke, S.R. (2004). Rating teams and analysing outcomes in one-day and test cricket. Journal of the Royal Statistical Society Series A, 167(4), 657–667. Barr, G.D.I., & Kantor, B.S. (2004). A criterion for comparing and selecting batsmen in limited overs cricket. Journal of the Operational Research Society, 55(12), 1266–1274. BBC (2017). India’s share of ICC global revenues adjusted after initial vote. BBC, 22 June. Retrieved from: www.bbc.co.uk/sport/cricket/40374596. Berry, S. (2015). India beat Pakistan by 76 runs as estimated one billion viewers tune in to World Cup clash. The Telegraph, 15 February. Retrieved from: www.telegraph.co.uk/sport/cricket/cricket-world-cup/ 11413995/ India-beat-Pakistan-by-76-runs-asestimated-one-billion-viewers-tune-in-to-WorldCup-clash.html. Bhaskar, V. (2009). Rational adversaries? Evidence from randomised trials in one-day cricket. The Economic Journal, 119(534), 1–23. Bhattacharya, M., & Smyth, R. (2003). The game is not the same: The demand for test match cricket in Australia. Australian Economic Papers, 42(1), 77–90. Birley, D. (2003). A Social History of English Cricket. London: Aurum Press. Blackham, J., & Chapman, B.J. (2004). The value of Don Bradman: Additional revenue in Australian Ashes tests. Economic Papers: A Journal of Applied Economics and Policy, 23(4), 369–385. Borland, J., & Macdonald, R. (2003). Demand for sport. Oxford Review of Economic Policy, 19(4), 478–502. Borooah, V.K. (2016). Upstairs and downstairs: The imperfections of cricket’s Decision Review System. Journal of Sports Economics, 17(1), 64–85. Brooks, R.D., Faff, R.W., & Sokulsky, D. (2002). An ordered response model of test cricket performance. Applied Economics, 34(18), 2353–2365. Bullough, S., Ramchandani, G., Coleman, R., & Miller, R. (2016). Central contracts in Test cricket: A model of best practice? Managing Sport and Leisure, 21(5), 338–352. Buraimo, B., Forrest, D., & Simmons, R. (2010). The twelfth man? Refereeing bias in English and German soccer. Journal of the Royal Statistical Society Series A, 173(2), 431–449.

266

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Chapman, B., Fischer, A., & Maloney, R. (1987). An Analysis of the Determinants of English-Australian Test Cricket Crowds: Estimating the Value of Don Bradman. Mimeo, Centre for Economic Policy Research, Research School of Social Sciences, Australian National University. Chedzoy, O.B. (1997). The effect of umpiring errors in cricket. The Statistician, 46(4), 529–540. Croucher, J.S. (1982). Anglo-Australian test cricket dismissals 1946–80. Journal of Applied Statistics, 9(2), 179–193. Crowe, S.M., & Middeldorp, J. (1996). A comparison of leg before wicket rates between Australians and their visiting teams for Test cricket series played in Australia, 1977–94. The Statistician, 45(2), 255–262. Dawson, P., Morley, B., Paton, D., & Thomas, D. (2009). To bat or not to bat: An examination of match outcomes in day–night limited overs cricket. Journal of the Operational Research Society, 60(12), 1786–1793. de Silva, B.M., & Swartz, T.B. (1998). Winning the coin toss and the home advantage in one-day international cricket matches. Department of Statistics and Operations Research, Royal Melbourne Institute of Technology. Dohmen, T., & Sauermann, J. (2016). Referee bias. Journal of Economic Surveys, 30(4), 679–695. Duckworth, F., & Lewis, A. (1998). A fair method for resetting the target in interrupted one-day cricket matches. Journal of the Operational Research Society, 49(3), 220–227. Elderton, W.P. (1945). Cricket scores and some skew correlation distribution. Journal of the Royal Statistical Society Series A, 108(1–2), 1–11. Elderton, W., & Wood, G.H. (1945). Cricket scores and geometrical progression. Journal of the Royal Statistical Society Series A, 108(1–2), 12–40. ESPNcricinfo (2017). Revealed: The ICC’s new financial model. ESPNcricinfo, 9 February. Retrieved from: www.espncricinfo.com/story/_/id/18650019/ revealed-icc-new-financial-model. Fletcher, T. (2017). Cricket: More Twenty20 may sow seeds of demise for the quintessential English game. The Conversation, 6 April. Retrieved from: http:// theconversation.com/cricket-more-twenty20-­ may-sow-seeds-of-demise-for-the-quintessentialenglish-game-75520. Forrest, D.K., & Dorsey, R. (2008). Effect of weather and toss on County Cricket Championship outcomes. Journal of Sports Sciences, 26(1), 3–13. Garicano, L., Palacios-Huerta, I., & Prendergast, C. (2005). Favoritism under social pressure. The Review of Economics and Statistics, 87(2), 208–216. Gollapudi, N. (2017). BCCI, ICC broker peace over financial model. ESPNcricinfo, 22 June. Retrieved from: www.espncricinfo.com/story/_/id/19704972/ bcci,-icc-broker-peace-financial-model.

Hoult, N. (2016). India finally give in and agree to use DRS in England Test series. The Telegraph, 21 October. Retrieved from: www.telegraph.co.uk/ cricket/2016/10/21/india-finally-give-in-andagree-to-use-drs-in-england-test-series/. Hynds, M., & Smith, I. (1994). The demand for test match cricket. Applied Economics Letters, 1(7), 103–106. Johnson, M. (1989). Pakistan v England 1987. In S. Moss (Ed.), Wisden Anthology 1978–2006: Cricket’s Age of Revolution. Guildford, UK: John Wisden & Co. Jones, M.V., Bray, S.R., & Bolton, L. (2001). Game location and officiating bias in English club cricket. Perceptual and Motor Skills, 93(2), 359–362. Karnik, A. (2010). Valuing cricketers using hedonic price models. Journal of Sports Economics, 11(4), 456–469. Lenten, J.A., Geerling, W., & Kónya, L. (2012). A hedonic model of player wage determination from the Indian Premier League auction: Further evidence. Sport Management Review, 15(1), 60–71. Mandle, W.F. (1972). The professional cricketer in England in the nineteenth century. Labour History, 23, 1–16. Massey, C., & Thaler, R.H. (2013). The loser’s curse: Decision making and market efficiency in the National Football League draft. Management Science, 59(7), 1479–1495. Mishra, V., & Smyth, R. (2010). An examination of the impact of India’s performance in one-day cricket internationals on the Indian stock market. Pacific-Basin Finance Journal, 18(3), 319–334. Morley, B., & Thomas, D. (2005). An investigation of home advantage and other factors affecting outcomes in English one-day cricket matches. Journal of Sports Sciences, 23(3), 261–268. Morley, B., & Thomas, D. (2007). Attendance demand and core support: Evidence from limited-overs cricket. Applied Economics, 39(16), 2085–2097. Norton, P., & Phatarfod, R. (2008). Optimal strategies in one-day cricket. Asia-Pacific Journal of Operational Research, 25(4), 495–511. Parker, D., Burns, P., & Natarajan, H. (2008). Player valuations in the Indian Premier League. Frontier Economics, October, 1–17. Paton, D., & Cooke, A. (2005). Attendance at county cricket: An economic analysis. Journal of Sports Economics, 6(1), 24–45. Paton, D., & Cooke, A. (2011). The changing demands of leisure time: The emergence of Twenty20 cricket. In S. Cameron (Ed.), Handbook on the Economics of Leisure (Chapter 18, 602–639). Cheltenham: Edward Elgar. Preston, I. (2006). The economics of cricket. In W. Andreff and S. Szymanski (Eds.), Handbook on the Economics of Sport (Chapter 61, 585–593). Cheltenham: Edward Elgar.

THE ECONOMICS OF CRICKET

Preston, I., Ross, S.F., & Szymanski, S. (2000). Seizing the moment: A blueprint for reform of world cricket. ESRC Study Group, Economics of Sport, Arts and Leisure, 13. Preston, I., & Thomas, J. (2000). Batting strategy in limited overs cricket. The Statistician, 49(1), 95–106. Ringrose, T.J. (2006). Neutral umpires and leg before wicket decisions in test cricket. Journal of the Royal Statistical Society: Series A (Statistics in Society), 169(4), 903–911. Rocke, K.D., Ramkissoon, B.I., Iton, A., & Khan, M. (2016). The global trend of competitive balance of domestic Twenty20 cricket competitions: A survey of the 2010–2015 seasons. International Journal of Sports Studies, 6(2), 81–89. Rohde, N. (2011). An ‘economic’ ranking of batters in test cricket. Economic Papers: A Journal of Applied Economics and Policy, 30(4), 455–465. Sacheti, A., Gregory-Smith, I., & Paton, D. (2014). Uncertainty of outcome or strengths of teams: An economic analysis of attendance demand for international cricket. Applied Economics, 46(17), 2034–2046. Sacheti, A., Gregory-Smith, I., & Paton, D. (2015). Home bias in officiating: Evidence from international cricket. Journal of the Royal Statistical Society Series A (Statistics in Society), 178(3), 741–755. Sacheti, A., Paton, D., & Gregory-Smith, I. (2016a). An economic analysis of attendance demand for

267

One Day International cricket. Economic Record, 92(296), 121–136. Sacheti, A., Gregory-Smith, I., & Paton, D. (2016b). Managerial decision making under uncertainty: The case of Twenty20 cricket. Journal of Sports Economics, 17(1), 44–63. Schofield, J.A. (1982). The development of first-class cricket in England: An economic analysis. The Journal of Industrial Economics, 30(4), 337–360. Schofield, J.A. (1983). The demand for cricket: The case of the John Player League. Applied Economics, 15(3), 283–296. Shivakumar, R. (2018). What technology says about decision-making: Evidence from cricket’s Decision Review System (DRS). Journal of Sports Economics, 19(3), 315–331. Sloane, P.J. (1976). Restriction of competition in professional team sports. Bulletin of Economic Research, 28(1), 3–22. Sumner, J., & Mobley, M. (1981). Are cricket umpires biased? New Scientist, 91(1260), 29–31. White, J. (2010). Twenty20 will kill Test cricket within 20 years, says West Indian great Michael Holding. The Telegraph, 1 June. Retrieved from: www. telegraph.co.uk/sport/cricket/twenty20/7790638/ Twenty20-will-kill-Test-cricket-within-20-yearssays-West-Indian-great-Michael-Holding.html.

27 Rugby Union’s Late Conversion to Professionalism: An Economic Perspective Patrick Massey

INTRODUCTION

RUGBY – A BRIEF HISTORY

Rugby Union provides an interesting and somewhat unique sports economics case study. Unlike many other major team sports, Rugby Union remained an amateur sport until 1995. Rugby split into amateur (Union) and professional (League) codes in 1895, but the amateur Union code remained the more widely played of the two.1 Rugby Union has received relatively little attention in the sports economics literature compared to Rugby League (Hogan, Massey, & Massey, 2017). This has been attributed to Union’s amateur traditions (Williams, 2012) and the lack of reliable data for the amateur era (Jones, Schofield, & Giles, 2000). Rugby Union’s switch to professionalism resulted in significant structural changes, with clubs, leagues and national associations adopting a variety of different business models which potentially makes it particularly interesting for sports economists. The remainder of this chapter is structured as follows. The next section provides a brief history of rugby and the sport’s structure pre-1995 is then described. The transition to professionalism and the changes that followed are described next. Subsequent sections address league structures and governance, team business models and attendances. Some conclusions are then offered. The focus is largely confined to European rugby due to space constraints.

The Rugby Football Union (RFU) was established in London in 1871 and rules were drawn up based on those of Rugby School, where, according to tradition, the game originated.2 While rugby has expanded in recent years, only 10 countries are designated as Tier 1 nations by World Rugby (Hogan, Massey, & Massey, 2013). The issue of professionalism emerged in team sports once they began to attract large paying audiences (Szymanski, 2009). The RFU banned professionalism at a meeting in October 1886 in order to stem the flow of working-class men into the sport. Mr Harry Garnett, who became RFU President in 1889, told the meeting: ‘If working men desired to play [rugby] football, they should pay for it themselves, as they would have to do with any other pastime’ (Yorkshire Post, 5 October 1886, cited in Collins, 2015, p. 38). Rugby subsequently fell behind soccer in England in terms of popularity (Collins, 2015). In 1893 a compromise proposal by the Northern clubs that players should be compensated for wages lost as a result of playing matches on Saturdays3 was vigorously opposed by the RFU hierarchy and rejected. The RFU launched a determined campaign to stamp out illegal payments, which were widespread, and subsequently announced

RUGBY UNION’S LATE CONVERSION TO PROFESSIONALISM

that, from September 1895, all clubs and players accused of professionalism would be presumed guilty until proven innocent (Collins, 2015). In August 1895 the leading Northern clubs formed their own Northern Rugby Football Union which became the Rugby League in 1922. Prior to 1995 all but one of English Rugby League’s top-tier clubs was located in the North of England (Dobson, Goddard, & Wilson, 2001), which has few top-tier Union clubs. Rugby League has struggled to expand beyond its traditional heartland (Wilson, Plumley, & Barrett, 2015). Uniquely, in Australia, Union ranks well behind League in terms of popularity (Collins, 2015). Up to 1995, only five of the 617 England/Great Britain Rugby League international players had attended private schools whereas 75% of England Rugby Union internationals had attended such schools. ‘There could be no greater marker of how far apart the two types of rugby were in England’ (Collins, 2015, p. 275).

COMPETITIVE STRUCTURES Leagues are a particularly effective way for professional teams to generate revenue to pay their players (Szymanski, 2009). South Africa and France were unusual among major rugby-playing nations having national championships dating back to 1889 and 1892, respectively. The ‘four home unions’4 traditionally opposed organised competitions and viewed leagues as inevitably leading to professionalism (Davies, 2003; Collins, 2015). In 1973 the Scottish Rugby Union (SRU) established a national club league in order to improve its international team’s performances but the RFU rejected proposals for an English club league (Collins, 2015). The traditional focus in rugby was on international matches (Hogan et  al., 2013). An annual championship involving national teams from England, Ireland, Scotland and Wales began in 1883. France joined in 1910 but did not participate from 1932 to 1939 following alleged payments to players by French clubs (BBC Sport, 2002). Italy joined in 2000 when the competition changed from the Five to the Six Nations Championship. The European teams also played international matches against the main Southern Hemisphere countries. The first tour of Britain and Ireland by the New Zealand All-Blacks took place in 1905. The first major innovation in the rugby calendar in almost a century came with the launch of the Rugby World Cup in 1987 (Davies, 2003). By the 1980s there was a growing public demand in England for competitive rather than friendly

269

matches between clubs (Davies, 2003). English and Welsh newspapers began publishing unofficial merit tables (Williams, 2012). The establishment of a national league in England in 1987/88 was also prompted by concerns about poor international performances. Teams played each other once, with no set dates for fixtures, it being left to clubs to arrange dates. Fixed Saturday fixtures were introduced in 1988/89 and in 1993/94 the league adopted a double round-robin format.5 National club leagues were established in Ireland and Wales in 1990/91.6 An annual championship between the four provincial branches of the Irish Rugby Football Union (IRFU), involving a single round-robin series with home advantage alternating from one season to the next, began in 1947. It mainly served as a series of trial matches for the national team and attracted limited attendances. As one former player recalls: ‘In 1995 I played for Munster against Leinster in Thomond Park with maybe at best 1,000 supporters who were passionate and typically drawn from the clubs and families that provided the players on show’ (Toland, 2014). The SRU introduced a similar competition – the Inter-District Championship – involving four regional teams, in 1953. These regional structures would play an important role in the professional era in both countries. The European Rugby Cup (ERC) was launched in 1995, shortly before the removal of the professional ban.

A TURBULENT TRANSITION The RFU along with their Irish, Scottish and Welsh counterparts were vehemently opposed to professionalism (Williams, 2012). A 1986 history of Irish rugby stated: ‘Professionalism has no part to play in the game in Ireland and could not be sustained’ (Van Esbeck, 1986, p. 234). Things were set to change dramatically. The establishment of the Courage League in England led to increasing breaches of the amateur rules (Williams, 2012), confirming the RFU’s worst fears. A working group established to discuss the future of amateurism reported, incorrectly according to Collins (2015), that no written record existed of why the Northern clubs’ 1895 proposal for broken time payments had been considered undesirable. In November 1995 the International Rugby Board (IRB) voted to end the professional ban. McMillan (1997) claims a lack of leadership by the RFU reduced English rugby’s first professional season to a ‘shambles’. Williams (2012) also criticises the RFU’s failure to manage the transition to professionalism. Many English clubs were acquired by wealthy individuals, while some floated

270

THE SAGE HANDBOOK OF SPORTS ECONOMICS

on the stock exchange. The net result was a threeway war between the RFU, grass roots clubs and the professional teams (Collins, 2015). Bitter disputes erupted between the RFU and professional clubs over the release of players to play for the England team and the division of television revenues (McMillan, 1997). In 1997/98 the professional clubs established a new league – Premiership Rugby (EPR) – signalling a transition to a business model similar to soccer and Rugby League. Play-offs were introduced to decide the champions in 2002/03. Rampant wage inflation threatened EPR’s viability. Revenue sharing and a salary cap were introduced in 1999/2000 after two clubs went into administration the previous season (Williams, 2012). Such arrangements are common in many sports and are often justified as necessary to maintain competitive balance and ultimately supporter interest. EPR’s arrangements were designed to prevent further club collapses and ensure the league’s viability rather than promote competitive balance (Williams, 2012). A unique feature of team sports is that production is a joint process (Rottenberg, 1956). Teams have an interest in ensuring that their rivals do not go out of business. EPR nevertheless describes itself as ‘the world’s most competitive domestic rugby union competition’.7 Hogan and Massey (2018) report that revenue sharing and salary caps increased short-run competitive balance (uncertainty of individual match outcomes) in EPR. Professionalism saw the continuation of a trend, already underway, towards a reduction in the number of teams in the French championship and a shift from rugby’s traditional Southern French heartland to larger cities (Andreff, 2015). In 2004/05 a pyramidal league structure was introduced with promotion and relegation and end-of-season play-offs to decide

80

Player Gross Wages (ex Payroll Taxes)

the championship. The top division was reduced from 16 to 14 teams and renamed the Top 14 the following season. As in England, many clubs were acquired by wealthy individuals. Hogan and Massey (2018) found that professionalism reduced shortrun competitive balance in both EPR and Top14, a result at odds with Rottenberg’s (1956) invariance principle. Several Top14 clubs experienced financial difficulties and a salary cap was introduced in 2010/11, albeit at a much higher level than in EPR. The cap has had limited success in controlling Top14 payroll costs (see Figure 27.1). Demands to raise EPR’s salary cap to match the Top14 were rejected as many EPR clubs felt they could not afford such an increase (Williams, 2012). The fan base in the smaller European rugby nations was too small to sustain national professional leagues that could compete with EPR and Top14. Collins (2015) suggests that the IRFU and SRU managed the transition to professionalism better than their Welsh counterparts. Whether, in the Irish case this reflected a clearly thought out strategy or the IRFU simply ‘stumbled upon the current system of centrally contracted players attached to four conveniently placed provinces’ (Thornley, 2012) is a moot point. Following the introduction of professionalism, the IRFU and SRU adopted a vertically integrated structure with players on centralised contracts. When the ERC was launched in 1995, the IRFU entered its provincial representative teams rather than club teams. The SRU subsequently entered its regional teams in the ERC the following year.8 Initially, the Welsh Premier League (WPL) continued as before and WPL teams entered the ERC, but the WPL clubs struggled financially. Gross Wages and Payroll Taxes (All employees)

70 60 50 40 30 20 10 0

Figure 27.1  Top14 payroll as a percentage of revenue (2004/05–2016/17) Source: LNR/DNACG, various years.

RUGBY UNION’S LATE CONVERSION TO PROFESSIONALISM

The redevelopment of its national stadium, Murrayfield, in Edinburgh left the SRU heavily indebted, forcing it to close two of its four district professional teams in 1998/99. The two remaining teams joined the WPL, which was renamed the Welsh-Scottish League. In 2001/02 the IRFU, SRU and Welsh Rugby Union (WRU) agreed to establish a new league, known as the Celtic League, with 15 teams – the nine WPL teams, four Irish provincial teams and the two remaining Scottish districts – split into two sections. FIFA rules prohibit multi-national leagues in soccer9, although it has been suggested that mergers of smaller country leagues are necessary to address growing competitive imbalances at European level (Szymanski, 2009). In 2002/03 the SRU re-launched the Borders team, which it had closed three years earlier, bringing the league to 16 teams. In 2002 the WRU proposed a vertically integrated structure with four regional franchises similar to Ireland, following ‘a painful few years for Welsh rugby as it has been forced to adapt and change in an attempt to climb out of the financial mire’ (WRU, 2004, p. 16). This was flatly rejected by the clubs (Collins, 2015). Ultimately five regional teams were established but crucially these were club owned. Two clubs, Cardiff and Llanelli, were designated as regions. The other three regions involved joint ventures between six of the remaining WPL clubs.10 The regions subsequently opposed WRU proposals for centralised player contracts, a stand-off that was eventually resolved in 2014 by the introduction of dual contracts for Welsh international players (Verdier, 2014). Following Welsh rugby’s restructuring, the Celtic League switched to a single 12-team division with a double round-robin fixture schedule in 2003/04. The WRU closed one of the new Welsh regions – the Celtic Warriors – after one season, reducing the league to 11 teams and leaving the Welsh valleys without a top-tier professional team. The SRU unsuccessfully sought joint venture partners for its teams. It sold the Edinburgh franchise to a private consortium and, in 2006/07, closed the Borders franchise again, reducing the league to 10 teams.11 The SRU bought back the Edinburgh franchise following a legal dispute. The IRFU also considered closing one of its teams but backed down following strong public opposition (Thornley, 2017a). Play-offs were introduced to decide the championship in 2009/10. In 2010/11, the Celtic League became the Pro12 with the addition of two Italian teams. Ostensibly this was intended to raise standards in Italian rugby and enhance their ability and that of the Italian national team to compete at European and international level. The €3 million annual fee that the Italian federation agreed to pay for their inclusion and the additional two home

271

matches per season for the remaining teams was probably a consideration also. The Italian experiment has not been successful. A proposed Rome-based franchise failed to materialise (ESPN, 2009). One of the Italian teams collapsed after just two seasons and its replacement was taken over by the Italian federation in 2017/18 after it also experienced financial difficulties (Rees, 2017a). The Italian teams have consistently finished in the bottom three and have attracted poor attendances. The SRU and WRU are reportedly unenthusiastic about retaining the Italian teams (Thornley, 2017b). In 2017/18 the Pro12 became the Pro14 when it added two South African teams and adopted a twoconference structure. The Pro14 thus includes teams from half of the Tier 1 rugby nations (Rees, 2017a).

LEAGUE STRUCTURES AND GOVERNANCE12 The EPR and Top14 are open leagues with promotion and relegation, although the EPR has reportedly considered dropping promotion and relegation and adopting a closed league structure (BBC Sport, 2019). Both leagues were built on the existing club system and the business model is broadly similar to that of European soccer leagues, with elements from US sports, e.g. play-offs and salary caps. Wilson and Plumley (2017) argue that EPR (and Rugby League’s Super League) have suffered financially through having adopted some (but not all) elements of US sports league models. The Pro14 has evolved towards a US sports league model. It is a closed league without promotion and relegation, play-offs to decide the championship and a two-conference structure since 2017/18. In Ireland and Scotland, representative teams which only played three or four matches a season during the amateur era have been transformed into professional franchises, while completely new teams were created in Italy and Wales. The league was superimposed on the existing club leagues in all four countries and clubs in those leagues effectively became feeder clubs for Pro12 teams. Fort (2006) suggests that, in club-run leagues, a conflict of interests between the league and its member clubs leads to sub-optimal decision making from the viewpoint of the league overall. Similar conflicts may arise in the Pro14, which is jointly owned by the IRFU, SRU and WRU. The unilateral decisions by the SRU and WRU to each close one of their teams may have been sensible for both associations but arguably ignored externalities, i.e. reduced fixtures and revenue for the league’s remaining teams. It may be optimal for a league to

272

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 27.1  ERC/ERCC performances by league and country (1995/96–2017/18)

Top 14 EPR Pro12 of which Ireland Scotland Wales Italy

Champions

Quarter-finals

Group stage win %

8 8 7

65 53 58

56.4 57.7 49.0

7 0 0 0

35  3 20  0

62.3 33.5 44.6 10.8

Notes: There were no quarter-finals in 1995/96. Pro12 total excludes Italian teams as they only joined the league in 2010/11. Source: European Professional Club Rugby: https://www.epcrugby.com/champions-cup/matches/

include some teams that are not viable as they may enhance demand for the stronger teams such that the revenue-enhancing effect on the latter group exceeds the losses of the former (Sloane, 1971). A club-run league might have retained the Borders and Celtic Warriors rather than replace them with two Italian franchises. Pro14 experience contrasts with EPR where clubs, recognising their mutual interdependence, introduced revenue sharing and salary caps to ensure the viability of all the member teams. Revenue sharing within the Pro14 is complicated because all the teams depend on their national associations for financial support. The Pro14 has no salary cap, although the four Welsh teams introduced their own salary cap in 2012/13 (BBC Sport, 2011). League structure is an important issue for sports governing bodies and restructuring is often motivated by an expectation that it will increase attendances (Dobson et al., 2001). The introduction of play-offs in all three European rugby leagues is a clear example. Play-offs maintain fan interest over a longer period, as more teams remain in contention to win the championship for longer, and thus increase league profits (Fort & Quirk, 1995). Play-offs also introduce a degree of randomness into championship selection, reducing teams’ incentives to over-invest in players, thereby improving competitive balance. FIFA rules requiring teams to play in national leagues prevents inter-league competition in soccer (Szymanski, 2009). Che and Humphreys (2015) cite the absence of competing leagues in US sports where incumbent leagues pre-empted entry by leaving few viable franchise opportunities for rival leagues (Cave & Crandall, 2001). Super Rugby’s decision to drop two of its six South African franchises in 2017/18 saw them switch to an expanded Pro14. Other South African Super Rugby teams could seek to join European leagues (Dymock, 2017). EPR has proposed including South African teams in the Anglo-Welsh Cup, which currently involves EPR and Welsh Pro14 teams (BBC Sport,

2017). In the past the Welsh Pro14 teams had discussions with EPR about switching leagues (Jackson, 2014). In England and Australia, Union and League could be regarded as competing leagues. Rugby League sought to establish a Welsh franchise at the former home of the defunct Celtic Warriors (BBC Sport, 2005). European rugby, like soccer, involves a combination of domestic and multinational tournaments. It is suggested that soccer’s Champions’ League has reduced competitive balance both within and between national leagues (Hogan et  al., 2017). Table 27.1 summarises the performances of the various leagues in the ERC/ERCC between 1995/96 and 2017/18. The table indicates a high level of balance between the three leagues. Irish teams have outperformed their Scottish and Welsh counterparts. An unusual feature of many rugby competitions, including all three European leagues, is the awarding of bonus league points which provide an explicit reward for outcomes besides winning (or drawing) (Lenten & Winchester, 2015). Such arrangements, which are designed to make matches more attractive by encouraging exciting play, provide useful frameworks for analysing economic behaviour.13 Winchester (2014) analyses the effects of different bonus points systems used in rugby.14 Hogan and Massey (2017) analyse the effect of rule changes in rugby which increased the value of tries relative to penalties over time. Rule changes represent a unique natural experiment to test the consequences of theoretical pay-off structures (Banerjee, Swimmen, & Weersink, 2007).

REVENUE Teams’ business objectives have been the subject of intense debate in sports economics. The effects

RUGBY UNION’S LATE CONVERSION TO PROFESSIONALISM

Revenue

400 350 300 250 200 150 100 50 0

273

Expenditure

Figure 27.2  Top14 Clubs’ aggregate revenue and expenditure (2004/05–2016/17 € million) Source: LNR/DNACG, various years.

of revenue sharing and salary caps depend on teams’ objectives (Noll, 2006). US sports teams are generally regarded as profit maximisers. It is argued that most European soccer clubs are win maximisers (Frick, 2007). Total revenue and expenditure of Top14 clubs from 2004/05 to 2016/17 are illustrated in Figure 27.2. In revenue terms, rugby lags well behind the major European soccer and US sports leagues. Top14 is the richest of the three main European rugby leagues with total revenues in 2016/17 of almost €340 million ($380 million). Growth in expenditure has outstripped revenue over the period. Aggregate losses in 2016/17 amounted to €23.5 million ($26.5 million) compared

with €200,000 ($246,000) in 2005/06. A small number of Top14 clubs balance their budgets. Andreff (2015), however, criticises the financial indiscipline of the majority which pursue a winmaximising strategy financed by repeated bailouts from obliging shareholders. Barros, Bertrand, Botti, and Tainsky (2014) found significant differences in efficiency between Top14 clubs. Figure 27.3 gives a breakdown of Top14 teams’ revenue streams. Sponsorship accounted for 46% of Top14 revenue in 2016/17. Club owners are the main source of sponsorship for many Top14 clubs (Andreff, 2015). Gate receipts and league disbursements (mainly broadcast revenue) accounted for 14% and 20% of revenue respectively. Local Govt, 2.5

Merchandise, 3.7

Other, 13.9

Match Receipts, 14.4

League Disbursements, 20.1

Figure 27.3  Composition of Top14 team revenues in 2016/17 Source: LNR/DNACG (2018).

Sponsorship, 46.4

274

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Top14 broadcast revenue increased from approximately €12 million ($13.9 million) in 2002/03 to €90 million ($106 million) in 2016/17 following the entry of a second Pay-TV channel (O’Sullivan, 2014).15 Toulon club president Mourad Boudjellal has called for a merit-based distribution of Top14 broadcast revenue. Palomino and Sakovics (2004) suggest that when there is competition between leagues for players, as is the case in European rugby, performance-based revenue sharing represents an equilibrium outcome as the increase in league quality, as a result of recruiting players from other leagues, is likely to outweigh any decline in competitive balance and lead to higher broadcast income. Rejecting calls for limits on the number of non-French players in Top14, Monsieur Boudjellal argued: ‘The French league will get more than €70 million per season due to the foreign players who have come to improve the level of competition here and enhance its appeal to the general public’ (O’Sullivan, 2014). EPR clubs’ total revenue in 2016/17 amounted to €240-€250 million ($268-$279 million) (Conn & Rees, 2018)16 up from €148 million ($185 million) in 2005/06 (Wilson & Plumley, 2017). The increase in revenue was partly due to new entrant BT Sports outbidding Sky for the broadcast rights. Wages amounted to 64% of revenue in 2016/17 (Conn & Rees, 2018).17 As in Top14, a clear distinction can be drawn between EPR clubs which spend what they earn and win maximisers with rich backers with the majority in the latter category. Wilson and Plumley (2017) and Conn and Rees (2018) report that only two EPR clubs were profitable in 2014/15 and 2016/17 respectively. Total losses in 2016/17 amounted to €33.6 million ($37.6 million) (Conn & Rees, 2018) while debt levels have increased at most clubs (Wilson & Plumley, 2017). 2014/15 EPR champions Saracens reported accumulated debts of €62 million ($68.8 million). Club Chairman Nigel Wray indicated that the debt could be reduced by cutting player numbers and ‘finishing 11th’, ‘[b]ut that would send out a terrible message to players that we weren’t ambitious and would hardly attract sponsors’ (BBC Sport, 2016). In contrast the Exeter Chiefs’ CEO stated: ‘We run Sandy Park as a business, which means we run at a profit. If we can’t run at a profit then we shouldn’t be in business. It does feel a bit unfair that other people can buy players and run sporting businesses as a tax loss’ (ESPN, 2014).18 Changes introduced in EPR’s revenue-sharing arrangements in 2005/06 mean that newly promoted teams receive significantly less revenue than incumbents.19 Szymanski (2003) argues that promotion and relegation reduce the incentive for teams to share revenues, which may explain the

revised EPR arrangements. Williamson (2015), however, described them as a cartel-type arrangement designed to restrict entry. In 2016, the RFU agreed to pay EPR clubs €244 million ($270 million) for using players in international matches. The question of national associations compensating clubs for the use of players in international matches is a recurring issue in soccer. Pro12 broadcast income in 2016/17 was approximately €15 million ($16.8 million), well below EPR and Top 14, reflecting the smaller broadcast markets in Pro12 countries. The gap has closed in recent years as the addition of two South African teams in 2017/18 (Dymock, 2017) and a new broadcast deal in 2018/19 substantially increased broadcast revenue of the new Pro14. The possible addition of US-based franchises is seen as a way of further increasing broadcast revenue (Rees, 2017a). Financial data on Pro14 clubs is not published, but all depend on their national associations for financial support. In 2016/17, one of the Irish teams was unable to repay loans owed to the IRFU (Cummiskey, 2016). The WRU had to buy out its joint-venture partner in the Dragons franchise to prevent its closure while the other three Welsh teams also recorded financial losses (Rees, 2017b). The IRFU has altered its approach to external financing in recent years and now permits its teams to top-up player salaries through sponsorship deals in order to compete with EPR and Top14 clubs (Hogan et al., 2017). Nevertheless, the IRFU’s priority is the international team. The IRFU limits the number of Pro14 matches that Irish international players can play for their teams. Irish teams also require IRFU sanction to sign overseas players. SRU (2012) objectives for the period to 2016 provided that both Scottish Pro12 teams should ‘consistently’ reach the league play-offs and knockout stages of European competitions and achieve average attendances of 10,000, although the latter target remains an unfulfilled aspiration. In 2016, the SRU (2016) announced its intention to sell its two Pro14 teams. The ERC distributed approximately €45 million ($59.8 million) to participating teams in 2013 with 52% shared equally between the four Pro12 countries and 24% each for EPR and Top14 with some additional payments to teams based on progress to the knock-out stages (Hogan et al., 2017). In 2012, EPR and Top14 clubs announced their intention to withdraw from the ERC after the 2013/14 season and establish an alternative competition due to their dissatisfaction with the distribution of ERC revenues. Jones (2014) described the ERC arrangements as ‘preposterously slanted in favour of the Irish and Scottish, despite the fact that through their television and other commercial deals, the Irish provide very little for the community pot’. In 2014/15, the ERC was

RUGBY UNION’S LATE CONVERSION TO PROFESSIONALISM

16,000

EPR

Top14

275

Pro14 (ex. Italy)

14,000 12,000 10,000 8,000 6,000 4,000 2,000 0

Figure 27.4  Average attendances in the main European rugby leagues (2001/02–2017/18) Source: www.premiershiprugby.com; http://rd.pro14rugby.com/matchcentre and LNR/DNACG, various years.

replaced by the European Rugby Champions Cup (ERCC) since renamed the Heineken Champions Cup with revenue split equally between EPR, Pro12 and Top14, although the Pro12 countries received guarantees that they would not receive less than they had from the ERC (Hogan et al., 2017).

ATTENDANCES Owen and Weatherston (2004) argued that information on the determinants of rugby attendances was important because attendances represented a significant source of revenue. Apart from Top14, this remains the case. Ticket sales represent the main source of income for EPR clubs despite increased broadcast revenues (Wilson & Plumley, 2017). The same almost certainly applies to Pro14 clubs. Figure 27.4 illustrates attendance trends in the three main European rugby leagues since 2001/02.20 EPR attendances have increased steadily, averaging just over 14,400 in 2017/18, similar to tier-two English soccer clubs. Top14 attendances increased up to 2009/10, but have been relatively flat since then, although Andreff (2015) notes that ticket prices increased post 2009/10. Pro14 attendances have grown steadily with average attendances (excluding Italy and South Africa) reaching 9,800 in 2017/18. Figure 27.5 shows trends in Pro14 attendances by country over the same period. Attendances in Ireland averaged almost 13,000 in 2017/18. This was in line with several EPR and Top14 clubs and significantly higher than in the other Pro14 countries. This is even though rugby was traditionally the main sport in Wales (Collins, 2015), while it was very much a minority sport in

Ireland. Despite increases in recent years Scottish attendances in 2017/18 averaged around 7,500, which is well below the SRU (2012) target. Italian attendances have been poor. The 12 ERC group stage matches in the inaugural 1995/96 season attracted 58,000 spectators, an average of 4,800 per match. In 2017/18 the 60 ERCC group stage matches attracted over 750,000 spectators, an average of 12,561 per match. Hogan et al. (2013) found that home team quality was the key determinant of match attendance in all three European leagues. Whether or not the home team was in contention to win the league or reach the play-offs also had a significant positive effect. Home team quality, chances of reaching the play-offs and habit (the number of previous seasons a team had played in the competition) were the main determinants of attendances at ERC group stage matches (Hogan et al., 2017). Shortrun competitive balance did not affect attendances.

CONCLUSIONS This chapter has sought to provide an overview of the changes that have occurred in rugby since the removal of the professional ban in 1995. Space constraints precluded consideration of the Rugby World Cup, which has grown significantly since its launch in 1987 with increasingly strong competition between countries to host it – the literature on megasports events may be relevant here. Professionalism has resulted in rule changes designed to speed up the game and make it more attractive, particularly to television audiences. Concerns have grown about player welfare, particularly around concussion-type injuries.

276

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Ireland

Scotland

Wales

Italy

14,000 12,000 10,000 8,000 6,000 4,000 2,000 0

Figure 27.5  Average Pro14 attendances by country (2003/04–2017/18) Source: http://rd.pro14rugby.com/matchcentre

Following the move to professionalism, teams in all the main European rugby nations have experienced financial problems and several have folded. Many English and French clubs are dependent on rich benefactors, while teams in the smaller European rugby nations are financially dependent on their national associations. This begs the question of whether European rugby needs Financial Fair Play rules like soccer. An interesting outcome of European rugby’s move to professionalism is the adoption of different business models and league and governance structures in different countries, combining elements from European soccer and US sports leagues. Some smaller countries have adopted vertically integrated structures with mixed results. Unlike soccer and US sports, there is competition between leagues, and this may intensify as leagues look to expand geographically in order to increase revenue. Rugby traditionally received relatively little attention in the sports economics literature. The introduction of professionalism and the sport’s subsequent transformation make it a potentially rich source for future research.

Notes   1  Throughout this chapter, rugby refers to Rugby Union unless otherwise stated.   2  www.rfu.com/AboutTheRFU/History.aspx   3  Such payments became known as ‘broken time’ payments.   4  England, Ireland, Scotland and Wales.   5  Source: www.premiershiprugby.com/informtion/ history.php   6  There were regional leagues in Ireland prior to this.   7  www.premiershiprugby.com/history/

  8  Scottish (and English) teams did not participate in the ERC’s first season.   9  Pro12 is not unique. The Super Rugby Championship (formerly Super 12) includes teams from Argentina, Australia, Japan, New Zealand and South Africa. 10  The WRU had to step in at the last minute and acquire a 50% stake in one of the joint ventures when one of the proposed partners withdrew. 11  The borders region was the traditional heartland of Scottish rugby (Collins, 2015). 12  See Szymanski (2003) for an economic analysis of sports leagues. 13  See Frick (2003) for a review of contest theory in sport. 14  EPR, Pro14 and ERC/ERCC award bonus points to teams scoring four tries and for losing by seven points or less. Top14 awards a bonus point for scoring three tries more than the opposition and, since 2015/16, for losing by five (previously seven) points or less. Super Rugby switched from the EPR to the Top14 system in 2017. 15  French soccer’s broadcast income is approximately five times greater than that of Top14 (Andreff, 2015). 16  Conn & Rees (2018) include revenue data for 11 of the 12 EPR teams. Figures quoted here assume the remaining club, Bristol Bears, had revenue in line with the average of the 11 teams for which data is publicly available. 17  Based on results for 11 teams (see note 18). 18  Exeter succeeded Saracens as champions in 2016/17. 19  A promoted team would need to remain in EPR for more than six seasons before they would receive the same share as incumbents (Williamson, 2015). 20  Pro14 data excludes Italian teams which only joined in 2010/11.

RUGBY UNION’S LATE CONVERSION TO PROFESSIONALISM

REFERENCES Andreff, W. (2015). Analyse economique du rugby professionnel en France: equilibre compétitif et contrainte budgétaire. In P. Chaix (Ed.), Le Nouveau Visage du Rugby Professionnel Français: Argent, Succès & Dérives (pp. 157–189). Paris: L’Harmattan. Banerjee, A.N., Swimmen, J.F.M., & Weersink, A. (2007). Skating on thin ice: rule changes and team strategies in the NHL. Canadian Journal of Economics, 40(2), 493–514. Barros, C., Bertrand, G., Botti, L., & Tainsky, S. (2014). Cost efficiency of French rugby clubs. Applied Economics, 46(23), 2721–2732. BBC Sport (2002, January 28). Six Nations history. Retrieved December 8 2017 from: http://news.bbc. co.uk/sport2/hi/rugby_union/international/1776391. stm BBC Sport (2005, June 22). Welsh side joins pro league ranks. Retrieved December 8 2017 from: http:// news.bbc.co.uk/sport2/hi/rugby_league/4120700. stm BBC Sport (2011). Welsh rugby introduces salary cap at regions, Retrieved March 5 2019 from: https:// www.bbc.co.uk/sport/rugby-union/16267538 BBC Sport (2016, January 25). Saracens debt grows to £45.1m after £3.98m annual loss. Retrieved December 8 2017 from: https:www.bbc.com/ sport/rugby-union/35400355 BBC Sport (2017, April 25). Anglo-Welsh Cup: Premiership Rugby looks at inviting South African teams. Retrieved December 8 2017 from: www. bbc.com/sport/rugby-union/39713384 BBC sport (2019, January 28). Premiership ringfencing promotion and relation debate that will not go away. Retrieved March 26 2019 from: https:// www.bbc.com/sport/rugby-uniion/46876713 Cave, M., & Crandall, R. (2001). Sports rights and the broadcast industry. Economic Journal, 111, F4–F26. Che, X.G., & Humphreys, B.R. (2015). Rival sports league formation and competition. In P. Rodriguez, S. Késenne, & R. Koning (Eds.), The Economics of Competitive Sports (pp. 7–21). Cheltenham, UK: Edward Elgar. Conn, D. & Rees, P. (2018, August 28). Premiership rugby finances: The full club-by-club breakdown and verdict, the Guardian. Retrieved March 29 2019 from: https://theguardian.com/sport/2018/aug/28/ premiership-finances-the-full-club-by-clubbreakdown-and-verdict Collins, T. (2015). The Oval World: A Global History of Rugby. London: Bloomsbury. Cummiskey, G. (2016, July 16). IRFU Chief does not believe Munster can make debt repayments. Irish Times. Retrieved December 8 2017 from: www. irishtimes.com/sport/rugby/international/irfuchief-does-not-believe-munster-can-make-debtrepayments-1.2723985

277

Davies, G. (2003). The History of the Rugby World Cup. Bodmin, UK: MPG Books. Dobson, S., Goddard, J., & Wilson, J.O.S. (2001). League structure and match attendances in English rugby league. International Review of Applied Economics, 15(3), 335–351. Dymock, A. (2017 September). Introducing… The Guinness Pro14. Rugby World. Retrieved March 26 2019 from: www.rugbyworld.com/news/introducingguinness-pro14-86651 ESPN (2009, July 19). Aironi and Praetorians set for Magner’s League. Retrieved October 2 2018 from: http:/www/espn.co.uk/rugby/ story/0id/14293797/ aironi-praetorians-set-magners-league ESPN (2014, April 13). Club bosses cry foul on Financial Fair Play. Retrieved December 8 2017 from: http://en.espn.co.uk/premiership-2013-14/ rugby/story/221695.html Fort, R. (2006). Talent models in North American and world leagues. In P. Rodríguez, S. Késenne, & J. Garcia (Eds.), Sports Economics after Fifty Years: Essays in Honour of Simon Rottenberg (pp. 83–106). Oviedo: University of Oviedo. Fort, R., & Quirk, J. (1995). Cross subsidisation, incentives and outcomes in professional team sports leagues. Journal of Economic Literature, 33(3), 1265–1299. Frick, B. (2003). Contest theory and sport. Oxford Review of Economic Policy, 19(4), 512–529. Frick, B. (2007). The Football Players’ Labour Market: empirical evidence from the major European leagues. Scottish Journal of Political Economy, 54(3), 422–446. Hogan, V., & Massey, P. (2017). Teams’ responses to changed incentives: evidence from Rugby’s Six Nations Championship. International Journal of Sport Finance, 12(2), 140–159. Hogan, V., & Massey, P. (2018). Revenue sharing, salary caps and competitive balance: results of a natural experiment from Rugby Union. International Journal of Sport Finance, 13(1), 3–17. Hogan, V., Massey, P., & Massey, S. (2013). Competitive balance and match attendance in European Rugby Union Leagues. Economic and Social Review, 12(2), 425–446. Hogan, V., Massey, P., & Massey, S. (2017). Analysing match attendance in the European Rugby Cup: does uncertainty of outcome matter in a multinational tournament? European Sport Management Quarterly, 17(3), 312–330. Jackson, P. (2014, January 28). Deadline looms for Anglo-Welsh League plan. The Rugby Paper. Retrieved December 8 2017 from: www. therugbypaper.co.uk/featured-post/13860/deadlinelooms-for-anglo-welsh-league-plan/ Jones, J.C.H., Schofield, J.A., & Giles, D.E.A. (2000). Our fans in the North: the demand for British Rugby League. Applied Economics, 32(14), 1877–1887.

278

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Jones, S. (2014, March 2). Dragons undoubted fire doused by a competition that’s all over the place. Sunday Times. Retrieved December 8 2017 from: www.thetimes.co.uk/article/dragons-undoubtedfire-doused-by-a-competition-thats-all-over-theplace-l86p60r5qqm Lenten, L.J.A., & Winchester, N. (2015). Secondary behavioural incentives: bonus points and rugby professionals. Economic Record, 91(294), 386–398. Ligue Nationale de Rugby/Direction Nationale D’aide et de Contrôle De Gestion (DNACG). Rapport Economie du Rugby Professionnel Francaise Comptes des Clubs Professionnels. Paris: Fédération Francaise de Rugby, various years. McMillan, J. (1997). Rugby meets economics. New Zealand Economic Papers, 31(1), 93–114. Noll, R.G. (2006). Sports economics after fifty years. In P. Rodríguez, S. Késenne, & J. Garcia (Eds.), Sports Economics after Fifty Years: Essays in Honour of Simon Rottenberg (pp. 17–50). Oviedo: University of Oviedo. O’Sullivan, J. (2014, April 26). Irish Provinces caught in European power game. Irish Times. Retrieved December 8 2017 from: https://www.irishtimes. com/newspaper/archive/2014/0426/Pg026.html Owen, P.D., & Weatherston, C.R. (2004). Uncertainty of outcome and Super 12 Rugby Union attendance: application of a general-to-specific modelling strategy. Journal of Sports Economics, 5(4), 347–370. Palomino, F., & Sakovics, J. (2004). Inter-League competition for talent vs. competitive balance. International Journal of Industrial Organization, 22(6), 783–797. Rees, P. (2017a, August 30). Ambitious Pro14 looks at Germany, Spain and Canada for further expansion. The Guardian. Retrieved December 8 2017 from: www.theguardian.com/sport/2017/aug/30/ the-breakdown-pro-14-expansion-germany-spaincanada Rees, P. (2017b, March 22). Newport Gwent Dragons to be taken over by Welsh Rugby Union. The Guardian. Retrieved December 8 from: www. theguardian.com/sport/2017/mar/22/newportgwent-dragons-welsh-rugby-union Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Scottish Rugby Union (2012). Inspiring Scotland Through Rugby: The Journey to 2016. Retrieved December 8 2017 from: www.scottishrugby.org/ sites/default/files/editor/docs/book.swf Scottish Rugby Union (2016, October 28). Member Clubs vote unanimously for external investment at

SGM. Retrieved December 8 2017 from: www. scottishrugby.org/news/16/10/28/member-clubsvote-unanimously-external-investment-sgm Sloane, P.J. (1971). The economics of professional football: the club as a utility maximiser. Scottish Journal of Political Economy, 18(2), 121–146. Szymanski, S. (2003). The economic design of sporting contests. Journal of Economic Literature, 61(4), 1137–1187. Szymanski, S. (2009). Playbooks and Checkbooks: An Introduction to the Economics of Modern Sports. Princeton, NJ: Princeton University Press. Thornley, G. (2012, January 2012). Player guidelines move would mean no more Nacewas. Irish Times. Retrieved December 8 2017 from: https://www. irishtimes.com/newspaper/archive/2012/0103/ Pg019.html Thornley, G. (2017a). Front Up, Rise Up: The Official Story of Connacht Rugby. Dublin: Transworld Publishers Ireland. Thornley, G. (2017b, November 16). Scots and Welsh votes could have brought RWC to Ireland. Irish Times. Retrieved December 8 2017 from: www. irishtimes.com/sport/sport/scots-and-welsh-votescould-have-brought-rwc-to-ireland-1.3293388 Toland, L. (2014, April 18). Interaction with the clubs must be part of our new grand plan. Irish Times. Retrieved December 9 2017 from: https://www. irishtimes.com/newspaper/archive/2014/0418/ Pg021.html Van Esbeck, E. (1986). The Story of Irish Rugby. London: Stanley Paul. Verdier, N. (2014, April 28). The Welsh WAR is over – statement. The Rugby Paper. Retrieved December 8 2017 from: www.therugbypaper.co.uk/latestnews/18167/the-welsh-war-is-over-statement/ Welsh Rugby Union (2004). Annual Report 2003/04. Cardiff: WRU. Williams, P. (2012). Any given Saturday: competitive balance in elite English Rugby Union. Managing Leisure, 17(2–3), 88–105. Williamson, B. (2015). Premiership Rugby Union: through the antitrust looking glass. Competition Law Review, 11(1), 41–60. Wilson, R., & Plumley, D. (2017). Different shaped ball, same financial problems? A holistic performance of English Rugby Union (2006–2015). Sport, Business and Management, 7(2), 141–156. Wilson, R., Plumley, D., & Barrett, D. (2015). Staring into the abyss? The state of UK Rugby’s Super League. Managing Sport and Leisure, 20(6), 293–310. Winchester, N. (2014). Should bonus points be included in the Six Nations Championship. New Zealand Economic Papers, 48(1), 96–101.

28 ‘The Answer’ and the Economics of Basketball: Perceptions vs Production David Berri

We see Allen Iverson, over and over again, charge toward the basket, twisting and turning and writhing through a thicket of arms and legs of much taller and heavier men—and all we learn is to appreciate twisting and turning and writhing. We become dance critics, blind to Iverson’s dismal shooting percentage and his excessive turnovers, blind to the reality that the Philadelphia 76ers would be better off without him. (Gladwell, 2006, n.p.)

Malcolm Gladwell wrote these words in The New Yorker in 2006 about Allen ‘the Answer’ Iverson. When these words were published, Iverson had just completed his tenth season in the NBA. Across these ten seasons he won numerous awards and was paid more than $96 million (BasketballReference, n.d.). To suggest somehow the Philadelphia 76ers would be better off without him would seem completely ridiculous. But in December of 2006 the 76ers decided to see what would happen if Iverson was sent elsewhere. ‘The Answer’ was traded on December 19 to the Denver Nuggets. On that day the Sixers had the worst record in the NBA (five wins and eighteen losses). After the trade I analyzed what might happen for the Sixers for the rest of that season. My analysis from December of 2006 indicated that the Sixers could expect to win 30 more games in 2006–07 without Iverson (Berri, 2007). When the

season ended, the Sixers record was 35–47 and that means the Sixers did indeed win 30 games without Iverson (and were a bit better than a 0.500 team without ‘the Answer’). Or to quote Gladwell, ‘… the Philadelphia 76ers were better off without him.’ The story of Allen Iverson captures much of what we know about the economics of basketball. From this story we learn there has historically been a disconnect between perceptions of performance and the objective measurement of performance. But this story will also teach us how decision-makers have learned some things over time. And it also gives us a glimpse into the future of where basketball might be headed.

OBJECTIVELY MEASURING BASKETBALL PERFORMANCE So how good was Allen Iverson? In terms of awards and salary, people certainly thought he was very good. But when Iverson’s statistics are empirically evaluated, the story told about Iverson changes dramatically. Team sports tend to track statistics for the individual player. They do this – as Berri and Schmidt (2010) argue – in order to ‘…separate

280

THE SAGE HANDBOOK OF SPORTS ECONOMICS

a player from his team. We know at the end of a contest who won. What we don’t know is which players were responsible for a team’s success (or failure).’ Player statistics have been tracked in baseball since the 19th century. Over time a host of different measures have been offered to measure the productivity of baseball players. Not surprisingly, some of these numbers might be more useful than others. How do we decide which numbers are more ‘useful’? Bradbury (2007) argued there are two criteria to consider. First, we need to look at how a measure connects to current outcomes. Then we need to consider how consistent the measure is across time. Like baseball, the NBA also has a variety of statistics to track performance. In addition to a host of box score statistics, there are also a number of summary measures. Some of the more popular measures based on box score statistics are NBA Efficiency and Player Efficiency Rating.1 Berri (2015) reviews the problems with these measures. The first issue deals with ‘efficiency’. Although the word ‘efficiency’ appears in the title of these metrics, the measures actually reward inefficiency. To illustrate this point, Berri and Bradbury (2010) presented calculations of how well a player has to shoot to break even with respect to the metric (i.e. has the estimated benefit of a player’s shooting equaled the estimated cost of shooting). Specifically, according to NBA Efficiency a player breaks even when they hit 33% of their two-point shots and 25% of their threepoint attempts. If a player shoots this well – and virtually all NBA players who take these shots exceed these thresholds – then the more a player shoots the higher will be his NBA Efficiency score. The break-even points for the Player Efficiency Rating model are even lower. Berri and Bradbury (2010) note that these are 29.2% on two-point shots and 20.6% on three-point shots. Berri (2015) notes that an average player makes 49% of their two-point shots and 35% of their shots from three-point range. So it is very common for players to exceed the break-even points. This means that for most players the more they shoot the better they will look according to NBA Efficiency and the Player Efficiency Rating. Given the issue with shooting efficiency, it is not surprising that none of these measures actually explains wins. As Berri (2015b) notes, only 34% of the variation in team winning percentage can be explained by NBA Efficiency. And only 33% of the team wins can be explained by the simplified version of Player Efficiency Rating called Game Score.2 In sum, measures that argue inefficient scorers help a team win games do not actually explain wins very well.

Explaining wins is not the only issue when it comes to performance statistics. Bradbury (2007) also notes that consistency across time also matters. When we focus on that issue, two other approaches – Win Shares and plus-minus – appear to come up short. The former model was developed by Justin Kubatko of Basketball-Reference3 and is based on the work of Dean Oliver (2004). Win Shares begins in a similar place as the Wins Produced model. According to Oliver (2004), four factors determine wins: shooting efficiency, rebounds, turnovers, and free throws. As Berri (2012) notes, these four factors – for a team and the team’s opponent – explain 94% of wins. To move from this basic idea to the evaluation of players, the calculation of Win Shares involves a series of calculations that can best be described as convoluted and are most definitely not based on an empirical examination of the link between the individual player statistics and wins. Win Shares for the individual player is also not entirely consistent. As Berri (2012) notes, per-minute Win Shares has a season-to-season correlation for NBA players of 0.67 (and if a player changes team that correlation falls to 0.49). Although this is low relative to the Wins Produced model, which we will discuss momentarily4, it is much better than the plus-minus model. John Hollinger (2005) describes this approach as follows: Add all the points the team scores when a player is on the court, and subtract all the points the team allows when he is on the court. Subtract the latter from the former, and you end up with the player’s ‘plus-minus’—how many points better or worse (i.e., plus or minus) the team is with that man on the court.

The primary problem with plus-minus should be obvious. Player statistics should separate a player from his teammates. But a player’s plus-minus depends on his teammates. If you play with toplevel talent, you will tend to have a high plusminus. If your teammates are not very good, your measure will tend to be lower. To overcome this issue, people have created a model called adjusted plus-minus. The many difficulties with this measure are detailed in Berri (2012, 2018). For now, we will simply note that Berri (2012) reports that adjusted plus-minus only has a 0.26 correlation from season to season.5 So this measure is simply not very consistent, suggesting that it does not completely separate a player from his teammates. None of these metrics takes the most obvious and direct approach to building a performance measure. Instead of guessing the weights of the

‘The Answer’ and the Economics of Basketball: Perceptions vs Production

box score statistics (as is done with PER or Win Shares) or ignoring the box score statistics (as is done with plus-minus) one can simply regress team wins on the box score statistics and allow that regression to determine the weights. This is the approach taken in Berri (2008); a work that builds upon Berri, Brook, Schmidt (2007a). Like all regressions, one has to do some thinking before one does any estimation. In other words, one can’t simply regress wins on all the box score statistics. As Gerrard (2007) notes, basketball is a ‘complex invasion sport’ similar to soccer, hockey, and American football. For such sports, Gerrard (2007) cautions researchers not to employ one simple regression to connect statistics to outcomes. Gerrard’s work focused on modeling soccer. But his argument is consistent with the approach Berri (2008) takes in modeling basketball. This model begins by regressing team wins on each team’s offensive and defensive efficiency. Offense (Defense) efficiency is simply the points scored (opponent’s points) divided by possessions employed (possession acquired). Possession employed is a function of field goal attempts (FGA), free throw attempts (FTA), offensive rebounds (OREB), and turnovers (TOV), or what a team can do with the ball once it is acquired. Possessions acquired is determined by the factors that give a team possession of the ball. This list includes defensive rebounds (DREB), team rebounds (TMREB)6, opponent’s turnovers (Opp. TOV)7, opponent’s field goals made (Opp.FGM), and opponent’s free throws made (Opp.FTM).8 Additional regression analysis allows one to ascertain the impact of assists (AST), blocked shots (BLK), and the diminishing returns of DREB.9 This model can be applied to the NBA, the Women’s National Basketball Association (WNBA; Berri & Krautmann, 2013), the American Basketball Association (a rival league of NBA from 1967 to 1976), and both National Collegiate Athletic Association (NCAA) women’s (Darvin, Pegoraro, & Berri, 2018) and men’s basketball (Berri, 2016). As Table 28.1 details, the results – when one compares the values of all statistics to the reported value of a point scored – are remarkably consistent. Essentially for all sports, the following statistics have a similar impact (in absolute terms) on outcomes: points, field goal attempts, offensive rebounds, defensive rebounds, team rebounds, steals (opponent’s turnovers), and turnovers, opponent’s points, and opponent’s made shots. Measures associated with free throws tend to have about half the value of everything else. These results suggest that a player breaks even on two-point shooting when they make 50% of their shots and 33% of their shots beyond the arc.

281

In sum, if a player shoots well they are helping a team win. If they are below average shooters they are not helping. The steps needed to derive the value of each box score statistic is somewhat involved. However, the story this approach tells about basketball is quite simple. The primary steps a team must take to win – as detailed in Berri (2015b) – are: • gain possession of the ball without the other team scoring (i.e., grab defensive rebounds, force turnovers), • keep possession of the ball (i.e., avoid turnovers and grab offensive rebounds), and • ultimately turn possessions into points (shoot efficiently from the field, get to the line and hit free throws). If a basketball team in the ABA, NBA, WNBA, or NCAA follows these three steps, it will win. And that means players primarily produce wins by grabbing rebounds, forcing opponent turnovers and avoiding turnovers themselves, and shooting efficiently. If they do not do these things, the team is better off without them. With all that we know about why teams win in basketball, how good was Allen Iverson? For his career, Iverson was below average – relative to an average point guard – with respect to shooting efficiency from the field and rebounds. And he was above average with respect to turnovers (i.e., he turned the ball over too much). So it should not be surprising when we measure Iverson’s contribution to wins we see a player that was generally below average in his career. An average team will win about 0.500 games per 48 minutes. Given that five players play for a team at any given time, an average player will therefore produce 0.100 wins per 48 minutes. Across Iverson’s career, though, he only produced 34.47 wins and posted a 0.044 Wins Produced per 48 minutes (Box Score Geeks, n.d.). In sum, Iverson was a below average NBA player.

PERCEPTIONS OF PERFORMANCE IN BASKETBALL Given how few wins Iverson produced, why was he considered such a great player? The answer is not about measurements of performance but about perceptions of performance. And that story begins with the issue of race. When economists first began looking at the sport of basketball, the topic they primarily examined was race. The NBA is a league where the

282

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 28.1  Impact of various player and team factors on wins in National Collegiate Athletic Association Women’s and Men’s Basketball, the Women’s National Basketball Association, and National Basketball Association, and the American Basketball Association NBA 1987-88 to 2016-17 Player Factors PTS FGA FTA OREB TOV DREB STL Team Factors Opp.PTS Opp.FGM Opp.FTM Opp.TOVa TMREB

Marginal Value

WNBA 1997 to 2017

Value Relative to Points

0.033 −0.034 −0.015 0.034 −0.034 0.033 0.033

Marginal Value

PTS FGA FTA OREB TOV DREB STL Team Factors Opp.PTS Opp.FGM Opp.FTM Opp.TOVa TMREB

Marginal Value

Value Relative to Points

1.0 −1.0 −0.5 1.0 −1.0 1.0 1.0

0.033 −0.031 −0.014 0.031 −0.031 0.032 0.032

1.0 −0.9 −0.4 0.9 −0.9 1.0 1.0

0.032 −0.034 −0.011 0.034 −0.034 0.033 0.033

1.0 −1.1 −0.3 1.1 −1.1 1.0 1.0

−1.0 1.0 0.5 1.0 1.0

−0.033 0.032 0.014 0.032 0.032

−1.0 1.0 0.4 1.0 1.0

−0.031 0.033 0.011 0.033 0.033

−1.0 1.0 0.3 1.0 1.0

Marginal Value −0.032 0.033 0.015 0.033 0.033 NCAA Men’s 2016-17

Player Factors

Value Relative to Points

ABA 1968-69 to 1975-76

Marginal Value

NCAA Women’s 2016-17

Value Relative to Points

Marginal Value Value Relative to Points

0.025 −0.027 −0.012 0.027 −0.027 0.027

1.0 −1.0 −0.5 1.0 −1.0 1.0

0.021 −0.019 −0.009 0.019 −0.019 0.021

1.0 −0.9 −0.4 0.9 −0.9 1.0

0.027

1.0

0.021

1.0

Marginal Value −0.026 0.027 0.012 0.027 0.027

Marginal Value −1.0 1.0 0.5 1.0 1.0

−0.023 0.021 0.009 0.021 0.021

−1.1 1.0 0.5 1.0 1.0

a

Opp.TO includes steals. Steals are credited to the individual player. Opp.TO that are not steals are credited to the team.

majority of players are African-American. This led economists to wonder if there was any difference in how decision-makers, who often were white, treated white and black players. The research on this question is somewhat large and not entirely conclusive. But Berri (2006) reported an interesting result when one looked at what this research was saying beyond the topic of race. Studies of race typically look at how white

and black workers are treated, controlling for performance. In basketball, the controls for performance were box score statistics. In virtually all the studies Berri (2006) looked at, points scored was statistically significant and of the correct sign. No other box score statistic, though, was consistently significant. Berri, Brook, and Schmidt (2007b) decided to explore this issue further in a study of the NBA free

‘The Answer’ and the Economics of Basketball: Perceptions vs Production

agent market from 2001 to 2006. The dependent variable in their model was average salary across a free agent’s contract. This was then related to a player’s performance the two years prior to the contract being signed. In addition, a collection of non-performance measures was included. The results indicated that from 2001 to 2006 players who signed with the same team did receive more money. More money was also paid to starters, those who did not miss games, younger players, and players from winning teams. When we turn to the box score statistics, we only see four factors that were statistically significant: points scored, rebounds, assists, and personal fouls. These statistics are essentially all that was tracked for an NBA player from 1950–51 to 1973– 74.10 Obviously across these years these were the only box score statistics that could have impacted a player’s evaluation. But more than 25 years later, player evaluation was still driven by these statistics. These four statistics, though, did not have the same impact on free agent salary. Berri, Brook, and Schmidt (2007b) also looked at economic significance and found that points scored clearly dominated the free agent salary decision. It is important to emphasize what didn’t matter. Players in this time period were not paid to shoot

283

efficiently or avoid turnovers. Rebounds mattered, but again, not nearly as much as points scored. So looking back on what determines wins in basketball, it becomes clear, from 2001 to 2006 players, why Allen Iverson was paid so well despite not producing many wins. Iverson scored many points and he was not penalized for shooting inefficiently and committing turnovers. Although many thought Iverson was one of the league’s best players in 2006, his career after leaving Philadelphia was surprisingly short. He only ended up playing one complete season in Denver (the 2007–08 season) and was then traded to the Detroit Pistons on November 3, 2008. After just one season in Detroit, he then signed with the Memphis Grizzlies for the 2009–10 season. The Grizzlies, though, waived him during the season. Iverson then re-joined the 76ers, but they also waived him before the 2009–10 season ended. In 2016, Iverson was elected to the NBA Hall of Fame. But the way his career ended suggests that maybe player evaluation had changed. To test this, the free agent salary model estimated in Berri, Brook, and Schmidt (2007b) was reestimated with free agent data from 2007 to 2017. The results, reported in Table 28.2, suggests that player evaluation has indeed changed.

Table 28.2  Modeling NBA free agent salaries, 2007–2017 Independent variable Points*** Effective Field Goal Percentage*** Free Throw Percentage Turnover Percentage Total Rebounds Assists*** Steals** Blocked Shots** Personal Fouls*** Age*** Dummy, Signs with Same Team*** Ratio of Started Games/Games Played*** Ratio of Games Played/Games Possible*** Team Wins*** Dummy, Plays Center*** Dummy, Plays Power Forward Dummy, Plays Shooting Guard Dummy, Plays Point Guard Observations R-squared

Coefficient 0.0415 1.6332 0.3481 0.0043 0.0091 0.0568 0.1049 0.0817 −0.0928 −0.0280 0.1400 0.6477 1.0913 0.8677 0.2254 0.0669 0.0784 −0.0380 635 0.61

Model is semi-logged with robust standard errors *, **, and *** indicate significance at 10%, 5%, and 1% levels

Standard error 0.0064 0.5672 0.3216 0.0106 0.0128 0.0201 0.0523 0.0406 0.0274 0.0071 0.0501 0.0733 0.1704 0.1766 0.0837 0.0736 0.0751 0.0856

t-statistic 6.53 2.88 1.08 0.41 0.71 2.83 2.01 2.01 −3.39 −3.96 2.80 8.84 6.40 4.91 2.69 0.91 1.04 −0.44

p-value 0.0000 0.0040 0.2790 0.6820 0.4790 0.0050 0.0450 0.0450 0.0010 0.0000 0.0050 0.0000 0.0000 0.0000 0.0070 0.3640 0.2970 0.6570

284

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Across these latter 11 seasons, free agent salary is still statistically related to points, assists and personal fouls (but not rebounds!11). But it is also related to effective field goal percentage, blocked shots, and steals. When we turn to economic significance, reported in Table 28.3, we see that when we look at elasticity it is effective field goal percentage that now matters the most. That is not the same story we see when we look at how a one standard deviation change in each independent variable impacts salary. But nevertheless, it is clear in recent years that players are now being paid to shoot efficiently. That does not quite mean that player evaluation is perfect. Players are still rewarded for taking shots12 and scoring points. Such a result is consistent with other player evaluations. For example, Berri, Brook, and Fenn (2011) note how scoring dominates where a player is chosen in the NBA draft. Greer, Price, and Berri (2019) argue that scoring also is the factor that dominates whether a player is even in the draft pool in the first place. Berri, Van Gilder, and Fenn (2014) note that voting for the MVP award and All-Rookie team are also dominated by scoring. Finally, Berri, Deutscher, and Galletti (2015) note that scoring is the primary box score statistic that determines the allocation of minutes played in the NBA and Spanish Liga. Harris and Berri (2015,

2016) found that scoring also dominated decisionmaking with respect to the allocation of minutes and the draft in the WNBA. Of course, one might think that maybe these scorers really are necessary to team success. After all, don’t the best players ‘create’ shots? Once again, we return to the story of Iverson. As Berri (2018) notes, Iverson averaged 24.4 field goal attempts per game before he was traded from the Sixers in 2006. The team was averaging 78.1 field goal attempts per game with Iverson, so ‘the Answer’ took 31.4% of the team’s shots from the field. So did Iverson ‘create’ those shots? That seems unlikely since, as Berri (2018) notes, Philadelphia averaged 78.2 field goal attempts per game once Iverson left. One should note that the Sixers’ effective field goal percentage as a team also improved without Iverson. The story about Iverson can be generalized by looking at what determines a team’s field goal attempts. As Table 28.4 notes, there are seven factors that explain 98% of the variation in a team’s shots from the field.13 All of these factors relate to how a team acquires and keeps the ball. So if a team acquires and keeps the ball, they will take a shot. And that tells us that shots in the NBA are not ‘created’. Of course, better players should be able to take ‘better’ shots. We would measure that

Table 28.3  Economic significance of box score statistics, 2007–2017 Statistically significant box score statistic

Slope coefficient

Elasticity

Impact of one standard deviation change

Points Assists Effective Field Goal Percentage Blocked Shots

$343,691 $470,078 $13,522,579 $676,112

0.8034 0.2389 0.8228 0.0801

$1,764,313.15 $788,870.20 $657,271.73 $428,596.52

$868,792 −$768,275

0.1572 −0.3895

$424,907.43 −$861,861.08

Steals Personal Fouls

Table 28.4  Explaining team field goal attempts in the NBA, 1987–88 to 2016–17 Independent variable

Coefficient

Standard error

t-statistic

p-value

Defensive Rebounds*** Opponent’s Field Goals Made*** Opponent’s Free Throws Made*** Opponent’s Turnovers*** Offensive Rebounds*** Free Throw Attempts*** Turnovers*** Constant R-squared

0.9781 0.9891 0.4549 0.9683 1.0777 −0.4457 −0.9514 4.9397 0.982

0.0118 0.0073 0.0104 0.0164 0.0136 0.0079 0.0172 0.5372

82.95 134.62 43.91 59.14 79.00 −56.42 −55.23 9.19

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

*, **, and *** indicate significance at 10%, 5%, and 1% levels

‘The Answer’ and the Economics of Basketball: Perceptions vs Production

by looking at the ability of players to shoot efficiently. When they cannot shoot efficiently, we would conclude they are not taking ‘better’ shots. The examination of team shot attempts indicates that when a player takes shots, they are literally taking those shots from teammates. Without the player, those shots would still be taken. So one should not reward players for just scoring points but rather they should be rewarded for scoring efficiently. Again, it appears the NBA is learning this lesson. Players are increasingly paid to shoot efficiently. Nevertheless, players are also paid more for scoring points and playing on winning teams. As noted, player statistics are about separating a player from his teammates. Player evaluation requires that a decision-maker separate a player from his teammates. The continued importance of winning percentage suggests that this is still not happening in the NBA.

LOOKING TO THE FUTURE Let one imagine a future where the gap between perceptions and reality in player performance continues to grow. How will that impact the NBA? As Berri (2018) reports, the NBA’s level of competitive balance is persistently worse than what we see in Major League Baseball, the National Football League, and the National Hockey League. Berri, Brook, Fenn, Frick, and Vicente-Mayoral (2005) argued that this seems to be because the NBA suffers from a ‘short supply of tall people.’ Because basketball relies on extremely tall people, the supply of talent – relative to other sports – is relatively small. When the underlying population is restricted, some teams will be able to employ extremely good athletes while others will have to employ talent that is far less productive. Hence competitive balance in the sport will not be very good. Despite this, though, the NBA has grown dramatically in popularity over time. The NBA began playing in 1946. Twenty-one years later, average attendance was only 6,631 fans per game (Association for Professional Basketball Research, n.d.). To put that in perspective, the WNBA was drawing 7,716 fans per game in 2017, or in its 21st season (WNBA, 2017). The NBA would not pass the WNBA’s 2017 mark until the 1971–72 season and 40 years into its existence (i.e., the 1985–86 season) the league was still drawing less than 12,000 fans per contest. It is important to note that the NBA’s early struggles are not unusual. As Berri (2018) notes,

285

we can tell a similar story about Major League Baseball, the National Football League, and the National Hockey League. Attendance seems to build slowly in a sport league.14 But as it has in other sports, the NBA has clearly found an audience. In 2016–17 the NBA’s teams attracted nearly 22 million fans, or close to 18,000 fans per contest (NBA, 2017). The NBA has done this despite the fact it has relatively poor competitive balance. Nearly half of the NBA franchises have never won an NBA championship. And each year one can tell before the season is half over that most teams have no chance of winning a title that year (Berri, 2017). So how has the NBA grown? There certainly is evidence that competitive balance is not quite as important to a league’s demand as economists historically argued (see Berri (2018) for a summary). But perhaps another issue is how market size, team payroll, and team wins are related in the NBA. Berri (2018) reports that there is no statistical relationship between market size and team wins while team payroll only explains 11% of the variation in team wins. However, if we look at market size and team payroll from 1999–2000 to 2016–17, we see that market size does explain 25% of the variation in what teams spend on players. So there is some evidence that teams in larger markets are spending more on playing talent. But it is often not leading to many positive outcomes. For example, in recent years both the New York Knicks and Brooklyn Nets have been among the league leaders in payroll. Neither franchise, though, has had much oncourt success. What if this all changed? What if the Knicks and Nets learned how to evaluate players and it was these franchises that persistently won NBA titles? Berri and Schmidt (2010) note that basketball players – relative to what we see in baseball, football, and hockey – are much more consistent across time. Berri and Schmidt (2010) note that for box score statistics in the NBA, 47% of the variation in a player’s field goal percentage this season was explained by what his field goal percentage was last season. All other box score statistics were even more consistent, with rebounds being the most consistent (explanatory power is reported to be 90%). In contrast, no statistic tracked for quarterbacks or running backs in the NFL has an explanatory power in excess of 26% and for turnovers (fumbles and interceptions) explanatory power is 7% or less. In the NHL, shots on goal is consistent (explanatory power of 80%). But shooting percentage is only 39% and statistics tracked for goalies are far less consistent. Finally, Bradbury (2008) offers evidence that both hitters

286

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and pitchers in Major League Baseball are generally less consistent than NBA players. All this means that if perceptions of performance were better aligned with the reality of a player’s productivity, teams in larger markets who wish to spend more money on players might be able to more consistently build winning teams. And that means teams in New York would be much more likely to win titles. Such an outcome has been seen in North American sports before. Although teams in New York have not consistently won in the NFL and NHL, in Major League Baseball the New York Yankees did dominate the league from 1921 to 1964. Across these 44 seasons the Yankees appeared in the World Series 29 times and won the title in 20 years. Despite this dominance, average league attendance rose from 7,004 fans per game in 1921 to 13,088 fans per contest in 1964 (Baseball-Reference, n.d.). Again, it is not entirely clear that competitive balance drives attendance in sports. So it is not clear that dominance by teams from New York in basketball would significantly harm the league. In sum, better player evaluation may not help competitive balance. But that might not harm the economic prospects of the NBA.

CONCLUDING OBSERVATIONS The research on basketball indicates that there is a gap between the perceptions of performance and how much players actually contribute to team wins. Players have historically been primarily paid to take shots and score. But wins are primarily about shooting efficiently, rebounds, and turnovers. There is evidence that some learning has taken place. In the early part of the 21st century it seemed like player evaluation still focused on the basic box score statistics tracked in the 1950s. But in more recent years it appears player evaluation has evolved somewhat. This is especially seen with respect to shooting efficiency. It appears that NBA free agents – in recent years – are now being rewarded for shooting efficiently. Despite this move forward, the market still seems to have it flaws. One could argue that players are still being paid to take shots. In addition, if a player has better teammates they will also get more money. Finally, teams may be employing some ‘advanced’ statistics that are not wellconstructed measures of productivity. So the market for playing talent still can best be described in the same terms we would use to describe the process by which players are drafted, minutes are

allocated, and awards are given. In other words, all of these processes are somewhat flawed. Nevertheless, progress has been made. If this continues, it might be the case that in the future the teams with the most money might be able to employ more of the best talent. The research on competitive balance indicates that this probably will not significantly damage the prospects for the league. But it would likely make fans in the larger markets somewhat happier.

Notes   1  NBA Efficiency was historically reported at NBA. com and is quite similar to the TENDEX model reported by Heeran (1994). Player Efficiency Rating was developed by John Hollinger and is currently reported at ESPN.com and at BasketballReference. More recently, NBA.com has updated NBA Efficiency with a measure they call Player Impact Estimate (see https://stats.nba.com/help/ glossary/). This updated measure, though, suffers the same problems observed with NBA Efficiency and Player Efficiency Rating.  2  Berri and Bradbury (2010) note that Game Score per 48 minutes has a 0.99 correlation with Player Efficiency Rating (which is a per-minute evaluation).   3  see www.basketball-reference.com/about/ws.html   4  Berri (2012) notes that an NBA player’s Wins Produced per 48 minutes (unadjusted for position played) has a 0.83 correlation from season to season.   5  Galletti (2011) noted that one prominent version of adjusted plus-minus also explained less than 5% of outcomes in the NBA and involved some dubious statistical steps.  6  As Berri (2008) notes, TMREB refers to team rebounds that change possession. This factor has to be empirically measured.   7  Opponent’s turnovers include steals (STL).   8  Berri (2008) notes that from the impact of opponent’s free throws made we can determine the value of personal fouls (PF).   9  Berri and Schmidt (2010) detailed the additional regressions needed to ascertain the values of blocked shots, assists, and defensive rebounds. The regression for defensive rebounds does not impact the value these have for the team. However, there is evidence that some defensive rebounds a player accumulates are taken from teammates. Therefore, in evaluating individuals, this issue needs to be noted. 10  See www.basketball-reference.com. Total rebounds were added in 1950–51. Previously, only points, assists, personal fouls, and shot attempts were tracked. Steals, blocks, and turnovers were not

‘The Answer’ and the Economics of Basketball: Perceptions vs Production

added to the NBA’s standard box score for the team until the 1973–74 season. Turnovers were not tracked for individual players until the 1977– 78 season. 11  The insignificance of rebounds is somewhat surprising. One should note that some measures that people consider ‘advanced’ – such as Win Shares and box score plus-minus – argue that rebounds are not very important. There are significant statistical problems with these approaches. The issues with Win Shares have been noted. Box score plusminus (BPM) involves a regression of a player’s adjusted plus-minus on the box score statistics (see www.basketball-reference.com/about/bpm. html). In other words, BPM involves a regression of a measure with many problems on box score statistics. Or one could argue that BPM involves regressing ‘nonsense’ on player statistics. And oddly enough, the list of statistics includes minutes per game. So a player who plays more minutes is considered a better player. It should be obvious that this is not an approach that would yield a very good measure of a player’s performance. Nevertheless, it is possible that decision-makers are influenced by what these models argue. 12  The model reported in Table 28.3 can be reestimated with the following measure of shots attempts per 48 minutes: Shots per 48 minutes = FGA per 48 minutes + 0.44*FTA per 48 minutes. Substituting this measure for points scored per 48 minutes we see that shots per 48 minutes are statistically significant (and positive). Furthermore, our ability to explain free agents’ salary (measured via r-squared) doesn’t change. In other words, one can argue that players are being paid to take shots. 13  This model was reported in Berri and Schmidt (2010) and Berri (2018). 14  This is an important lesson to be learned when looking at women’s sports league. The WNBA is the oldest of these. It takes decades to build a following so it is not surprising that women’s sports leagues still have relatively low attendance. So low attendance in women’s sports is likely more about the age of the league, not the quality of the product.

REFERENCES Association for Professional Basketball Research (n.d.). NBA/ABA home attendance totals. Retrieved from: www.apbr.org/attendance.html Basketball-Reference (n.d.). Allen Iverson Stats. Retrieved from: www.basketball-reference.com/ players/i/iversal01.html

287

Baseball-Reference (n.d.). New York Yankees Team History and Encyclopedia. Retrieved from: www. baseball-reference.com/teams/NYY/ Berri, D. J. (2006). Economics and the National Basketball Association: Surveying the literature at the tip-off. In J. Fizel (Ed.), The handbook of sports economics research (pp. 21–48). New York: M.E. Sharpe. Berri, D. J. (2007). One correct prediction. The Wages of Wins Journal, 19 April. Retrieved from: http:// wagesofwins.com/2007/04/19/one-correctprediction/. Berri, D. J. (2008). A simple measure of worker productivity in the National Basketball Association. In B. R. Humphreys & D. Howard (Eds.), The business of sport (Vol. 3, pp. 1–40). Westport, CT: Praeger. Berri, D. J. (2012). Measuring performance in the National Basketball Association. In S. Shmanske & L. Kahane (Eds.), The handbook of sports economics (pp. 94–117). Oxford: Oxford University Press. Berri, D. J. (2015). Think you know basketball? You need to know the numbers to know the game. Sports & Entertainment Review, 1, 6–13. Berri, D. J. (2016). Paying NCAA athletes. Marquette Sports Law Review, 26, 479–492. Berri, D. J. (2017). For most NBA teams, the season is basically over when you wake up Christmas morning. Forbes.com, 24 December. Retrieved from: www.forbes.com/sites/davidberri/2017/12/24/ for-most-nba-teams-the-season-is-basically-overwhen-you-wake-up-christmas-morning/ #e92c3113b14b Berri, D. J. (2018). Sports economics. New York: Worth Publishers/Macmillan Education. Berri, D. J., & Bradbury, J. C. (2010). Working in the land of metricians. Journal of Sports Economics, 11, 29–47. Berri, D. J., Brook, S.L., & Fenn, A. (2011). From college to the pros: Predicting the NBA amateur player draft. Journal of Productivity Analysis, 35, 25–35. Berri, D. J., Brook, S. L., Fenn, A., Frick, B., & VicenteMayoral, R. (2005). The short supply of tall people: Explaining competitive imbalance in the National Basketball Association. Journal of Economics Issues, 39, 1029–1041. Berri, D. J., Brook, S. L., & Schmidt, M. B. (2007a). The wages of wins: Taking measure of the many myths in modern sport. Stanford, CA: Stanford University Press. Berri, D. J., Brook, S. L., & Schmidt, M. B. (2007b). Does one need to score to score? International Journal of Sports Finance, 2, 190–205. Berri, D. J., Deutscher, C., & Galletti, A. (2015). Born in the USA: National origin effects on time allocation in US and Spanish professional basketball. National Institute Economic Review, 232, R41–R50. Berri, D. J., & Krautmann, A. (2013). Understanding the WNBA on and off the court. In E. M. Leeds &

288

THE SAGE HANDBOOK OF SPORTS ECONOMICS

M. Leeds (Eds.), Handbook on the economics of women in sports (pp. 132–155). Cheltenham, UK: Edward Elgar. Berri, D. J., & Schmidt, M. B. (2010). Stumbling on wins: Two economists explore the pitfalls on the road to victory in professional sports. Princeton, NJ: Financial Times Press. Berri, D. J., & van Gilder, J., & Fenn, A. (2014). Is the sports media color-blind? International Journal of Sport Finance, 9, 130–148. Box Score Geeks (n.d.). Allen Iverson. Retrieved from: www.boxscoregeeks.com/players/544-allen-iverson Bradbury, J. C. (2007). The baseball economist: The real game exposed. New York: Dutton. Bradbury, J. C. (2008). Statistical performance analysis in sport. In B. R. Humphreys & D. Howard (Eds.), The business of sport (Vol. 3, pp. 41–56). Westport, CT: Praeger. Darvin, L., Pegoraro, A., & Berri, D. (2018). Are men better leaders? An investigation of head coaches’ gender and individual players’ performance in amateur and professional women’s basketball. Sex Roles, 78, 455–466. Galletti, A. (2011). Deconstructing a model. Arturo’s silly little stats v2.0, 4 March. Retrieved from: https://arturogalletti.wordpress.com/2011/03/04/ deconstructing-a-model/ Gerrard, B. (2007). Is the Moneyball Approach transferable to complex invasion team sports? International Journal of Sports Finance, 2, 214–228.

Gladwell, M. (2006). Game Theory. The New Yorker, 29 May. Retrieved from: www.newyorker.com/ magazine/2006/05/29/game-theory-2 Greer, T., Price, J. A., & Berri, D. J. (2019). Jumping in the pool: What determines which players the NBA considers in the draft? International Journal of Sport Finance, 14(1), 43–53. Harris, J., & Berri, D. (2015). Predicting the WNBA Draft: What matters most from college performance. International Journal of Sport Finance, 10, 299–309. Harris, J., & Berri, D. (2016). If you can’t pay them, play them: Fan preference and own-race bias in the WNBA. International Journal of Sport Finance, 11, 163–180. Heeren, D. (1994). Basketball Abstract 1994–95 Edition. Indianapolis, IN: Masters Press. Hollinger, J. (2005). Hockey stat, with a twist, useful in NBA, too. ESPN.com, 29 March. NBA (2017). NBA breaks attendance record for third straight season. NBA.com, 13 April. Retrieved from: www.nba.com/article/2017/04/13/nba-breaksall-time-attendance-record-third-straight-season#/ Oliver, D. (2004). Basketball on paper: Rules and tools for performance analysis. Washington, DC: Potomac Books. WNBA (2017). WNBA scores highest attendance in six years during record-breaking season. WNBA.com, 6 September. Retrieved from: www.wnba.com/news/ wnba-scores-highest-attendance-six-years-recordbreaking-season

29 Economics and the National Football League Benjamin Blemmings

INTRODUCTION The National Football League (NFL) creates a rich empirical environment for studying two economic topics. One, the focus of this chapter, is testing the assumption that firms maximize profit. The second investigates how labor markets value certain worker characteristics. Characteristics that have been investigated include age (Allen, 2015), criminal records (Weir & Wu, 2014), education (Böheim & Lackner, 2012), cognitive ability and race (Berri & Simmons, 2009; Pitts & Evans, 2018), and media exposure (Treme & Allen, 2011). These papers follow the broader suggestion of Kahn (2000) that sports represent an excellent environment to investigate labor issues. Firm maximization implies that firms can optimally trade off player characteristics, so I focus here on papers that examine broader NFL labor evidence which is connected more closely to how teams allocate compensation to employees, a topic closely related to testing whether NFL teams maximize. The NFL has also been fertile ground for examination of broader economic questions, particularly whether responses to shocks from NFL income conform to the permanent income hypothesis and whether teams play minimax strategies (Kovash & Levitt, 2009), a prediction from game theory. A recent application in the NFL context is a test of the permanent income hypothesis in terms of

predictions about how individuals optimally allocate lifetime consumption (Carlson et  al., 2015). The theory predicts that players should save most of a positive income shock from an NFL contract, recognizing that the average NFL career is short, but the earnings are very large, and they will likely never earn income that high again. Preliminary results indicate that 15.7% of players retiring from the NFL between 1996 and 2003 filed for bankruptcy, contrary to the predictions of the model. This chapter focuses on the analytic strategy used in Romer (2006) and subsequent papers that extended this line of inquiry on whether or not NFL teams maximize over specific variables and how economic research benefits from specific empirical tests of assumptions of standard economics models.

DO FIRMS MAXIMIZE? An essential building block of economic theory is the assumption of agent rationality. Rational agents are assumed to maximize utility or profit subject to a budget constraint (and parameters), presumably a useful approximation for human behavior. Modeling economic decisions is fraught with pitfalls for several reasons, including mathematical sophistication, computational tractability, institutional detail, and

290

THE SAGE HANDBOOK OF SPORTS ECONOMICS

omission of non-economic factors. Economists assess how well these assumptions hold up in realworld applications using newly available data and econometric methods. One difficulty with testing the assumptions that underlie economic models is that data on individual firm decision making are typically unavailable. Researchers cannot easily observe how many workers are assigned daily to work at McDonalds throughout the country or how much credit local businesses are willing to take on to expand. Even if one could observe these difficult-to-measure characteristics, how would one be able to calculate a benchmark optimal choice for comparison? Conditions are ripe for economic maximization in the NFL where decisions are repeatedly observed and there is a way to calculate optimal choices in certain strategic contexts. The incentives for all involved parties (players, coaches, team managers, etc.) seem purely centered on profit maximization. The league has high stakes, the highest of any professional league, with annual revenues over $30 billion. This creates intense competition among teams and it is plausible that a firm’s objective function in this context simplifies from profit maximization to win maximization. In addition to clear economic preconditions for maximization, the rules of American football result in a highly structured game which generates an excellent opportunity to observe repeated decisions in similar contexts. Additionally, high-quality information exists in this environment, including multiple measures of performance, opposing team behavior, and external information from scouting services and media coverage. As with any study, caveats about how results apply externally exist, but given the prevailing conditions in the NFL it would be especially noteworthy if teams regularly deviated from optimal decision making. The seminal paper on NFL game decision making is Romer (2006), which compares what teams do in a specific game situation to what the optimal decision would have been. After a team receives possession at a point on the field, they start with a first down. Subsequently they have four attempts (downs) to gain the 10 required yards to gain an additional set of downs or score a touchdown (worth on average 6.98 points in Romer’s (2006) data, it is a decimal due to the points for a touchdown being conditional on the next play). Typically, teams only use three downs to gain additional yardage, because not gaining the required yardage for a new set of downs or a touchdown on the fourth play results in the opposing team possessing the ball at the point on the field to where the offense advanced the ball. To avoid this risk, teams often use the fourth down to kick the ball in one of two ways. If teams are close enough to the opponent’s endzone, they

can kick a field goal, which is worth three points, or punt the ball, which almost always results in the opposing team’s offense beginning farther away than if the team would have failed to convert a fourth down attempt into a new set of first downs. Whether agents choose the optimal amount of risk is of central importance to economics because of what the results would imply for almost a century of economic models. This forms the cornerstone of the research question in Romer (2006). The analysis in Romer (2006) proceeds in several steps. First, the value of having the ball anywhere on the field is calculated. Second, the optimal decision is estimated using Bellman equations, because the decision yields both a current outcome and an alternative down/distance scenario. In addition to estimating the value of having a first down anywhere on the field, the value of kicking and ‘going for it’ are also estimated. With this information, the maximum number of yards-to-go for a first down, where teams should be indifferent between kicking and going for it, is determined along with the maximum number of yards-to-go where coaches choose to go for it at least as often as they choose to kick. Romer (2006) clearly shows that the maximum number of yards-to-go at which coaches go for it nearly as often as they kick falls systematically and significantly short of the optimal predicted decision. Coaches appear to make irrational decisions over almost the entire field of play, because actual attempted fourth down conversions only exceed the 2 standard error confidence band of the estimated optimal decision at 2–4 specific points on the field. Teams only go for it with the optimal frequency suggested by the point estimate of the gain from going for it at one point on the field. Over about half of the field, coaches never go for it on the fourth down and they are never close to going for it on the fourth down as much as the analysis suggests is optimal in terms of expected points scored. Statistical significance tests find z-statistics of between 3 and 7 for tests of the null hypothesis of no difference between optimal decisions and actual decisions on the fourth down. Given this striking empirical finding, it is appropriate to review the assumptions and conditions that generated this apparent refutation of rationality. First, three years of NFL data are analyzed which only contains 11,112 first quarter situations for analysis. Only first quarter situations are used to account for momentum effects, excessive point differentials, and modified behavior due to time remaining (second quarter begins play where first quarter ended) which seems appropriate to address this narrow question, but this leaves very few fourth down attempts for analysis. Due to the lack of fourth down attempts, third down attempts are used under the assumption that the decision

ECONOMICS AND THE NATIONAL FOOTBALL LEAGUE

on the third down closely resembles the decision on the fourth down. The estimated value of going for it, kicking, and of being in a certain position are all smoothed, which improves precision. By comparing expected points given a decision, agents are assumed to be risk-neutral over points scored. Romer (2006) explains deviations from rationality based on: rational risk aversion, third versus fourth down differences, asymmetric information, and momentum effects. In each case evidence is presented which suggests that the main result is not especially sensitive to these alternatives. One validation of Romer (2006) can be found in Yam and Lopez (2018), which finds remarkably similar impacts in terms of wins if teams were to go for it on the fourth down more often using matching methods as a causal inference approach. The importance and robustness of the results in Romer (2006) spawned an extensive literature which addresses issues related to firm maximization in the context of the NFL. For instance, if teams instead made the optimal decision by going for it more on the fourth down, how many more games would they win? Do teams make optimal in-game and draft strategy decision on other margins? Can the failure of strict economic maximization in the same environment be attributed to modern economic concerns with traditional maximization like psychological factors and behavioral economics? These are the main questions that the subsequent literature addressed. I classify papers in this section according to which explanation they advance as a potential theory of firms failing to maximize. Racism could be considered a failure of maximization, but racism is a tangential issue because economic theory could be used to make predictions about market conditions for the existence of racism. Instead of reviewing work on racism in the NFL, I refer readers to Berri and Simmons (2009), the seminal paper on this topic, for further details. I classify papers in this section as either based on psychological factors or economic factors such as heuristic decision making, agency problems, or risk aversion. More generally, Romer (2006) considers the possibility that the utility function of the firm may not simply reduce to win maximization, for example the case when utility depends on fans’ risk parameters (thus, the risk parameters of those who generate demand).

PSYCHOLOGICAL CONCERNS First, I consider evidence provided in studies which posit that the failure of strict maximization reflects psychological concerns. Psychological concerns

291

encompasses a wide variety of phenomena at odds with economic maximization. For example, the ‘hot hand’ effect hypothesizes that previous success will make future success more likely and implies that previous success makes a firm more likely to take risks. Standard economic theory would categorize the most recent event in a sports contest as a sunk cost and therefore not relevant for maximization in the next decision. This category contains the most direct test of the effect of momentum on risktaking. Additionally, an evaluation of the extent to which NFL teams maximize would not be complete without examining other margins, such as draft decisions and the examination of different rules which determine the payoffs of risky decisions.

In-game Decisions Basic economic models do not allow risk preferences to systematically vary with recent momentum, so why is momentum important? The variable focus of attention model (March & Shapira, 1992) provides an important framework for understanding why organizations are more likely to take risks which affects important economic variables like capital accumulation. In this model, attention focuses on reaching performance targets, experimenting with slack resources, or avoiding threats. Empirical tests focused on discovering conditions which trigger shifts towards survival or slack concerns. Survival concerns emerge when a firm operates far below aspiration level and is in danger of closing, while slack concerns emerge when organizations experiment with surplus resources. In Lehman and Hahn (2013), psychological insights about how firms maximize are combined with organizational frameworks to form a testable empirical theory about how firms behave, given that they may have gotten into a specific situation via different recent game situations. The paper makes two contributions: one pertains to the theory of how momentum affects risk-taking. The key relationships in this model are: (1) performance and risk-taking are negatively correlated, but the relationship is stronger for performance below the aspiration level than for performance above it. (2) For performance below the aspiration level, the negative correlation between performance and risk-taking is weaker when the organization is experiencing negative within-period momentum; for performance above the aspiration level, the negative correlation between performance and risk-taking is weaker when the organization is experiencing positive within-period momentum. (3) The relationship between across-period momentum and risk-taking

292

THE SAGE HANDBOOK OF SPORTS ECONOMICS

is negative but only when within-period feedback is not available. An empirical analysis using NFL data supports the prediction of this model. Lehman and Hahn (2013) define within-period positive momentum as two consecutive scores without an opponent scoring in between them and negative withinperiod momentum as the opposite. Across-period positive (negative) momentum occurs when a team wins (loses) two or more games in a row. Neither variable is dichotomous, both indicate how much momentum exists, not just an extensive margin, by reflecting the actual number of consecutive wins/ losses or points. Finally, the dependent variable is a binary variable which equals one if a team decided to go for it on the fourth down. The results are generated using a logit model with interactions of variables representing magnitude of momentums used to test the hypotheses. These tests support the hypotheses, with expected signs on regression parameters, and reasonable magnitudes. It is a compelling psychological/organizational theory, leading the authors to conclude that momentum is an important driver of risk-taking. Rothoff (2012) also analyzes in-game NFL decisions, but focuses on whether teams exhibit bankruptcy behavior. Bankruptcy behavior is when companies in financial distress have an incentive to take on high-risk/high-reward projects. His analyses use simple ordinary least squares (OLS) models which investigate the relationship between scoring in a quarter and in the previous quarter, the relationship between scoring in any previous quarter on scoring in the fourth quarter, and finally these relationships in subsamples of close or not close games. All models are also estimated separately for college and NFL teams, because NFL overtime begins with a coin toss determining which team gets the ball first, while both teams get equal chances to score in college football overtime. The results suggest that NFL teams exhibit bankruptcy behavior, because fourth quarter scoring is less closely related, and not related at all for close games, to other quarter scoring. This contrasts with college rules where scoring is always related to fourth quarter scoring. Romer (2006) finds that riskier decisions could lead to greater scoring and teams behind in the fourth quarter could conceivably take more risks to try to win, but the results do not suggest that this leads to more NFL scoring.

Draft Decisions Another paper that finds psychological concerns lead to NFL teams making systematic deviations from optimal strategy analyzes the NFL draft

(Massey & Thaler, 2013). The NFL draft is a good empirical environment because trades can be made during the draft that reflect how teams value players. Additionally, the data necessary to construct an optimal draft decision to benchmark actual draft decisions are available. By using how much one team pays their draft pick, how the draft pick performs, and where the player is drafted, Massey and Thaler (2013) test the economic condition for optimal draft day strategy. That equation is: (Mi / Mi+k) = ((E(Si)) / (E(Si+k)), showing that the relative value of any two picks that are traded should equal the relative expected surplus of the picks. This assumes a player’s value can be observed on the field and estimated from the labor market. I agree with their claim that assuming firms are profit-maximizing in these trade decisions is like assuming that they are maximizing winning percentage or chance of winning super bowl and that it is not a bad one to make. The authors test the alternative hypothesis that teams will overvalue the right to pick early, citing several psychological biases that unambiguously suggest a bias in that direction, although the separate biases’ effects are not decomposed in the empirical section. The biases are nonregressive predictions, overconfidence, the winners curse, false consensus from strong beliefs about a future player’s future productivity, and anticipated regret from missing out on a superstar. Nonregressive predictions hypothesize that intuitive predictions are insufficiently regressive, being more extreme and varied than is justified by the evidence on which they are based and overconfidence works similarly. The reader is left to explore these psychological concepts on their own if they are more curious. Importantly, the draft has been going on since 1936, teams have incentive to learn, and performance is observable, which should facilitate learning. Given these conditions, it would be glaringly inconsistent with standard economic theory if teams were not optimally trading picks. Massey and Thaler (2013) first estimate the value of draft picks as a function of draft order relative to the first pick in the draft using actual draft-day trades between 1983 and 2008 based on a two-parameter Weibull distribution. The model fits the data well and exhibits a steep decline in value relative to the number one pick as players are selected farther from number one. The decline only moderately flattened over time. A simple analysis suggests that when teams move up to get a player who they perceive is better than the player they could have drafted at the same position, the player they move up to get starts more games than that player only 52% of the time. Next, situations where a first-round pick was traded away for two later picks are examined. The average gain from trading down is 5.4 starts per

ECONOMICS AND THE NATIONAL FOOTBALL LEAGUE

season, while pro bowls are roughly equal. The paper concludes that in 74% of the cases, teams would have acquired players who started more games by trading down. The descriptive evidence strongly supports the notion that teams overvalue picking earlier in the draft, but more analysis is done to examine the magnitude of the effect and how it varies with draft position. The ‘cost-benefit’ analysis of draft-day trading relies on estimating the value that teams place on performance by looking at compensation of veteran players in the first stage. Next, these values are applied to drafted players and their surplus value is estimated by subtracting their compensation from these performance values. Veteran quarterbacks are paid more than 50% of the next-highest-paid position (defensive end), which is interesting when considered alongside a main result of Roach (2017), where very similar years of data (2000–2009) are used, which finds quarterbacks are underpaid relative to the value they provide measured by their injury and salary adjusted value. It appears that salary commitments are monotonically increasing, for a given position, as a function of average season utilization (no games, some games up to 14, more than 14 games, and pro bowl) which appears to be increasingly steep as utilization category increases. Massey and Thaler (2013) also notes that there is a steep relationship between draft position and player compensation, with a distinct discontinuity after pick 32, which is when the first round ends. These round discontinuities form the basis of identification strategies in the section on heuristic decision making. The next step is calculating surplus as the difference between the performance value estimated from the compensation model for a player’s position and actual performance. Finally, the surplus value is estimated on a linear spline of draft order, which is linear within round and knotted between rounds. The authors find strong statistical evidence that surplus decreases when draft order increases within the first round, which is demonstrated clearly in Massey and Thaler (2013, Figure 3). High picks performed better on average than lower picks, but compensation to lower picks within the first round fell faster relative to performance. The descriptive statistics support this regression result and when one compares the expected surplus to the actual trademark, the deviation is visible. Where the trade market projects that the value of players drops quickly after the first pick, the expected surplus measure projects the maximum surplus is in the second round. What is striking about Figure 3 in Massey and Thaler (2013) is that when the information on performance value and compensation are combined to create surplus value relative to first pick, then that is compared to the values implied by actual trades,

293

the shapes of the curves describing expected surplus and actual values are different. The surplus value is concave with a maximum around the beginning of the second round, while the trades are concave with significant convexity, especially in the first round where the compensation implies that the last player picked in the first round produces about 25% of the player selected first. All the round coefficients are significant, the first round is positive and rounds 2–5 are insignificant for ordinary least squares or quantile regressions. Even though teams can evaluate players well, as evidenced by the monotonically decreasing performance value, the evidence strongly suggests that teams are not good at using this to optimize their trading strategy. This is another result that economic theory would not suggest. Several checks demonstrate the robustness of the results. They are not sensitive to accounting for all-star players beyond pro bowls. Evidence that off-field utility (a player who increases demand for tickets and paraphernalia) drives the results is not found based on a subsample of offensive linemen, who are arguably the least likely to generate significant off-field utility. Finer performance measures do not alter the main conclusions for wide receivers, when performance is measured by receiving yards. One might be concerned about a team having a special need and/or think that they have particularly good information on a player. The performance of players that were traded up for were compared with all other NFL players and were not found to have a higher likelihood of playing in the NFL, playing or starting more games, or making the pro bowl more often. This suggests that the NFL players who are traded up for perform better than others. Finally, making smart trades produces an estimated 1.5 more wins per year, but this is likely to be positively correlated with other factors that also produce more wins, such as effective management, which leads to the estimated effect of smart drafting being biased upwards when those factors are not included as control variables. Similar results are shown in Hersch and Pelkowski (2016) where it is found that players who are traded for start more and have greater approximate value.

ASSESSMENT OF PSYCHOLOGICAL FACTORS AFFECTING MAXIMIZATION Each theory is plausible and not mutually exclusive, but the strength of the evidence which supports each varies. Massey and Thaler (2013) features the strongest empirical strategy, showing actual decisions substantially deviate from reasonably constructed benchmark optimal decisions like the

294

THE SAGE HANDBOOK OF SPORTS ECONOMICS

approach in Romer (2006). Empirical work (Allen, 2015) shows a difference between compensation of rookies and veterans, which Massey and Thaler (2013) implicitly assume away. Correcting the estimates in Massey and Thaler (2013) for this difference could alter the shape of the performance value curve, altering the expected surplus curve if the compensation difference between veterans and rookies varies systematically by draft round chosen. The starkly opposite shapes of the curves implies that this would have to have a very strong effect to alter the conclusions. Massey and Thaler (2013) also document the existence of a chart used by NFL teams that offers guidelines on the value of draft pick trades, which was created using somewhat similar analytic techniques in the late 1980s by a team executive. The chart has proliferated around the league and is often used in trades to value picks. So in addition to psychological biases, there is also a norm that tripled trades and reduced variance of valuations by 50% as the chart proliferated, which both explains how they fit the data so well and has interesting implications. The Massey and Thaler (2013) paper is titled the ‘loser’s curse’ because the first pick is the worst pick in the draft, measured by expected surplus, and limits to arbitrage prevent smarter competitors from taking advantage of this. There is no way to sell the first pick short and if it is traded away, then teams must find a buyer willing to take the standard overinflated price. This paper is unique in that it shows firm failures and how market characteristics will prevent surplus maximizing decisions that agree with predictions of the theory. The paper is a useful reminder that decision makers usually don’t appreciate how difficult decision making is and it expands the finding in Romer (2006) by showing that in-game decision making is not the only context where NFL teams showcase less than strict economic maximization. Momentum is a highly intuitive, plausible explanation for risk-taking in sports and the implications for other industries are important. The definitions of within and across period momentum are fair approximation of firm decision making, with minor caveats. Better descriptive statistics, which illustrate the effect of momentum, may have made the assumptions relying on regression more tenable, but endogeneity gives rise to serious questions about whether the coefficients can be interpreted as causal. Hesitation stems from endogeneity, which is introduced through reverse causation of risktaking and within-period momentum. A core result of Romer (2006) is that teams could score more by going for it and momentum is defined as consecutive scores so, by definition, the explanatory variables are endogenous. The question is

important still and a potential explanation for deviation, thus future research should focus on finding causal approaches to estimating momentum’s effect on risky decision making. Finally, the result that NFL teams are more likely to exhibit bankruptcy behavior in the fourth quarter (Rotthoff, 2012) should be interpreted cautiously as no fixed effects and correlation between the standard errors and the independent variables are accounted for. This means the coefficients are identified by cross-sectional variation under the assumption of homoskedasticity, which are potentially problematic. The papers reviewed found evidence that psychological factors are present and a potential impediment to strict economic maximization, which is often assumed in economic models. Evidence from Massey and Thaler (2013) is the strongest and shows that teams significantly overvalue picking early in the draft. The other two papers could use some additional tests of their robustness to updated methodological approaches, but offer promising theories for specific empirical tests of what may trigger additional risk tolerance, which would be a direction towards understanding why teams are not strict economic maximizers.

ECONOMIC FACTORS Behavioral economics offers a different explanation for why NFL teams are not maximizing. As behavioral concepts gained acceptance in economics, they have also been used in empirical tests of NFL decisions to see if outcomes are consistent with what we would expect from individuals who are maximizing simple profit-maximizing objective functions. The widespread proliferation of behavioral concepts into sports is covered in Coates and Humphreys (2018), which also points to the usefulness of behavioral economics to the larger study of industrial organization. I classify the papers in this section as behavioral because they focus primarily on heuristic decision making. Main topics focus on how a draft pick’s college team’s rank affects their draft pick number and how sunk costs affect decision making.

Behavioral Economics: Heuristic Decision Making Perhaps the simplest heuristic that could be used by NFL teams to evaluate players would be the quality of the undergraduate football program they played in. Past research did not include NFL draft combine performance measures (Hendricks, DeBrock, &

ECONOMICS AND THE NATIONAL FOOTBALL LEAGUE

Koenker, 2003) but results have been shown to hold when including them (Kitchens, 2015). Results indicate that the performance (in terms of final season poll ranking) of the college team of players selected in the NFL draft affects only the player’s draft selection number, not his NFL performance. If the draft was perfectly competitive and contained only profit-maximizing teams, players who played at higher-quality college football programs would only be picked higher if they had better performance in college. Teams may believe that the expected future productivity of a player from a higher-quality college football program is higher than a player with similar observables who played at a lower-quality college football program, because unobservable characteristics may be correlated with observables like football program quality. Thus, all other observables being equal, an NFL team would prefer a player from the higher-ranked college program. This is most related to the economics literature considered by Altonji (2005), where employers believe future productivity is a function of observables and unobservables which is used to understand statistical discrimination. From a theoretical perspective, employers receive an inefficient signal from potential employees and signals from one distinct group of employees is more inefficient relative to another. The consequences for group members could be positive or negative, conditional on whether the selection mechanism is one-shot or if true productivity can be learned over a probationary period. Statistical discrimination exists if signals received by employers are inefficient and the inefficient signal puts a certain group at a disadvantage, while option value predicts positive outcomes for the groups with more inefficient signals if some employers can profit from figuring out a way to capture rents with knowing their value. The analysis in Hendricks et al. (2003) examines different sections of the draft, while Kitchens (2015) focuses on estimating an average effect for the entire draft. The analysis in Kitchens (2015) shows that scout’s rank is not related to the player’s institution and that several NFL combine outcomes are able to explain variation in scout’s projected draft round for players, which establishes the importance of the combined outcomes, then shows that college team rank significantly affects draft pick number. College team rank also does not significantly affect a player’s career performance, measured by years in the league, games started, games played, and a metric which captures a player’s best three seasons, conditional on how far ahead of their predicted draft position they were selected. Controlling for how much a team reached (how much sooner they actually drafted him than they projected), players from higher-ranked institutions had similar careers, but the evidence also suggests that players from

295

higher-ranked institutions are chosen sooner. These results are shown to be robust to dropping players from teams in the top ten, interacting position dummies with specific combine variables, position interactions with college team’s rank, interactions of college team rank and year, and non-linear impacts of team rank. Teams statistically discriminate against players from lower-ranked institutions by picking players from higher-ranked institutions sooner, which contradicts economic maximization because a college’s rank is shown to have no effect on a player’s professional productivity. The analysis is repeated for just quarterbacks, with an emphasis on race, in Gill and Brajer (2012).

Heuristics and Sunk Costs Another violation of the predictions of standard economic theory occurs if teams equate the marginal benefit of an action to the marginal cost plus the sunk cost of that action. One empirical strategy for assessing the prevalence of such decisions is to exploit the end of a draft round, an arbitrary cutoff, as a discontinuity. This empirical strategy relies on the probability of being chosen in the second round being discontinuous along a running (x) variable, the draft selection number. There are 32 teams in the NFL, so each draft rounds contains 32 picks with each team drafting one player in each round. At pick number 33, the probability of being in the second round goes from 0 to 1. No evidence is found that players near this threshold, for the thresholds between rounds 1 and 2 and rounds 2 and 3, are significantly different in terms of productivity from players close to the other side of this threshold. Even though the players perform similarly, players picked at the end of round 1 are paid 36–38% more than those picked at the beginning of round 2 (Keefer, 2016, 2017a). These papers then estimate that a 10% increase in salary cap value yields an additional 2.7 games started by players, which suggests that NFL teams incorporate sunk costs in playing time decisions. Finally, since games started are a coarse productivity measure, the in-game usage of a running backs is regressed on their compensation, which is instrumented for using restricted and unrestricted free agency status dummies (Keefer, 2017b). These results also support the importance of sunk costs in team decisions.

Assessment of Heuristics The quality of the empirical evidence supporting the presence of heuristic decision making occurring in the NFL is strong. The ability of OLS estimates to

296

THE SAGE HANDBOOK OF SPORTS ECONOMICS

capture the causal effect of team rank depends on college team rank being randomly assigned, conditional on a rich set of covariates, which appears to be a reasonable assumption. Draft round-based discontinuities likely identify credible estimates of the local average treatment effect, because variation around the cutoff is effectively randomized and players clearly cannot sort onto either side of the discontinuity. In Keefer (2017b), the instruments do not pass the relevance benchmark of a first stage F statistic of at least 10 in the preferred specification, although other specifications reach the benchmark. Evidence from this line of research appears to be credibly causal. Analytically testing whether statistical discrimination arises from recruiting finds evidence that it occurs, although it does not appear to benefit teams. Additionally, an upside to analyzing statistical discrimination in the NFL is that many outcomes are observable, and information is relatively complete. This means that teams could select on observables but, all else being equal, also check for differences in football program rank. In the NFL, observable player characteristics are widely available and more detailed than in other settings, and those observables can be included in a regression model. Round discontinuities are a natural identification strategy and appear to offer reasonable evidence of supporting the sunk cost fallacy; teams are paying earlier round draft picks more and playing them longer. This is further evidence that teams are not maximizing along important margins. These heuristics have a quantifiable impact on players whose salaries are heavily influenced by early contracts due to short average career length, but no analysis has been carried out to assess the effect on other outcomes, such as team success. Studies in this section test for whether teams which are boundedly rational have economic effects. On initial inspections, the heuristics appear to not be very good for some players, but more work is needed to quantify the effects of heuristics on teams’ outcomes to better understand how far heuristics cause teams to deviate from optimal decisions.

Agency A final explanation for the discrepancy between actual and calculated optimal decisions in Romer (2006) is the existence of an agency problem, where the incentives of the coach do not align well with that of upper management. Owens and Roach (2018) show the amount of risk a coach is willing to accept on the fourth down to be a function of how likely that coach is to be fired or retained in the following season, using a sample of college football coach decisions. The literature would benefit from having

this approach applied to the NFL if this extended analysis would verify the effect of predicted firing/ promotion on risk-taking on the fourth down. Perhaps more information might be contained in the tweets or social media announcements of teams, coaches, players, and fans, making textual analysis a better predictor of promoted/fired outcomes. For an application of text analysis explaining outcome variance, see Kuziemko and Washington (2015).

Risk In the same spirit of thoroughly addressing the effect of risk on decision making, Goff and Locke (2018) reinvestigate Romer (2006) and find a similar divergence in decisions. They also take model uncertainty into account in the empirical analysis, which reduces the statistical significance of the difference between the optimal decision and the actual decision, but not below the 90% level. They conclude that a significant amount of the restraint in making risky decisions is due to coaches being more risk averse than originally attributed by Romer (2006). This result is intuitive, especially viewed in conjunction with the results in Owens and Roach (2018). These papers show that not only psychological factors undermine maximization, but also factors which have traditionally been in the domain of economics, such as risk aversion and agency, can lead to firms maximizing a more complicated objective function.

CONCLUSION The NFL provides an ideal context for analyzing firm decision making. Many theories have been advanced to explain the seminal result that NFL coaches do not maximize points scored, which is inconsistent with how economic agents in this competitive market would be expected to behave, including psychological, behavioral, agency, and risk aversion. The strongest evidence supports the hypothesis that behavioral heuristic decision making has important effects, but agency and risk aversion are being reconsidered in recent research. Possible future research should include testing the newer economic concepts that have recently been advanced, such as agency and risk aversion, to see if they can explain other decisions where teams do not maximize. Agency has only been tested in college and risk aversion has only recently been updated, so learning about how football teams maximize is still important and active. Economics can gain further understanding from continuing to test decision making in the NFL for micro- and macro-related questions.

ECONOMICS AND THE NATIONAL FOOTBALL LEAGUE

REFERENCES Allen, W. D. (2015). The demand for younger and older workers: patterns from NFL labor markets. Journal of Sports Economics, 16(2), 127–158. Altonji, J. G. (2005). Employer learning, statistical discrimination and occupational attainment. American Economic Review, 95(2), 112–117. Berri, D. J., & Simmons, R. (2009). Race and the evaluation of signal callers in the National Football League. Journal of Sports Economics, 10(1), 23–43. Böheim, R., & Lackner, M. (2012). Returns to education in professional football. Economics Letters, 114(3), 326–328. Carlson, K., Kim, J., Lusardi, A., & Camerer, C. F. (2015). Bankruptcy rates among NFL players with short-lived income spikes. American Economic Review, 105(5), 381–384. Coates, D., & Humphreys, B. R. (2018). Behavioral and sports economics. In V. J. Tremblay, E. Schroeder, & C. H. Tremblay (Eds.), Handbook of Behavioral Industrial Organization (Chapter 12). Cheltenham, UK: Edward Elgar. Gill, A., & Brajer, V. (2012). Wonderlic, race, and the NFL draft. Journal of Sports Economics, 13(6), 642–653. Goff, B., & Locke, S. L. (2018). Revisiting Romer: digging deeper into influences on NFL managerial decisions. Journal of Sports Economics, online first, September 13. https://doi.org/10.1177/1527002518798686 Hendricks, W., DeBrock, L., & Koenker, R. (2003). Uncertainty, hiring, and subsequent performance: the NFL draft. Journal of Labor Economics, 21(4), 857–886. Hersch, P. L., & Pelkowski, J. E. (2016). Are there too few trades during the NFL draft? Applied Economics Letters, 23(7), 516–519. Kahn, L. M. (2000). The sports business as a labor market laboratory. Journal of Economic Perspectives, 14(3), 75–94. Keefer, Q. A. (2016). Rank-based groupings and decision making: a regression discontinuity analysis of the NFL draft rounds and rookie compensation. Journal of Sports Economics, 17(7), 748–762. Keefer, Q. A. (2017a). Do sunk costs affect expert decision making? Evidence from the within-game usage of NFL running backs. Empirical Economics, online first, December 23, 1–28. http://dx.doi. org/10.1007/s00181- 017-1399-y Keefer, Q. A. (2017b). The sunk-cost fallacy in the National Football League: salary cap value and playing time. Journal of Sports Economics, 18(3), 282–297. Kitchens, C. T. (2015). Are winners promoted too often? Evidence from the Nfl draft 1999–2012. Economic Inquiry, 53(2), 1317–1330.

297

Kovash, K., & Levitt, S. D. (2009). Professionals do not play minimax: evidence from Major League Baseball and the National Football League (No. w15347). Cambridge, MA: National Bureau of Economic Research. Kuziemko, I., & Washington, E. (2015). Why did the Democrats lose the south? Bringing new data to an old debate (No. w21703). Cambridge, MA: National Bureau of Economic Research. Lehman, D. W., & Hahn, J. (2013). Momentum and organizational risk taking: Evidence from the National Football League. Management Science, 59(4), 852–868. March, J. G., & Shapira, Z. (1992). Variable risk preferences and the focus of attention. Psychological review, 99(1), 172. Massey, C., & Thaler, R. H. (2013). The loser’s curse: decision making and market efficiency in the National Football League draft. Management Science, 59(7), 1479–1495. Owens, M. F., & Roach, M. A. (2018). Decision-making on the hot seat and the short list: evidence from college football fourth down decisions. Journal of Economic Behavior & Organization, 148, 301–314. Pitts, J. D., & Evans, B. (2018). Evidence on the importance of cognitive ability tests for NFL quarterbacks: what are the relationships among Wonderlic scores, draft positions and NFL performance outcomes? Applied Economics, 50(27), 2957–2966. Online first, December 3, 2017. Roach, M. A. (2017). Testing labor market efficiency across position groups in the NFL. Journal of Sports Economics, online first, April 18. doi: 10.1177/1527002517704021. Romer, D. (2006). Do firms maximize? Evidence from professional football. Journal of Political Economy, 114(2), 340–365. Rotthoff, K. W. (2012). Bankruptcy behavior in the NFL: does the overtime structure change the strategy of the game? Journal of Economics and Finance, 36(3), 662–674. Treme, J., & Allen, S. K. (2011). Press pass: payoffs to media exposure among National Football League (NFL) wide receivers. Journal of Sports Economics, 12(3), 370–390. Weir, K., & Wu, S. (2014). Criminal records and the labor market for professional athletes: the case of the National Football League. Journal of Sports Economics, 15(6), 617–635. Yam, D., & Lopez, M. (2018). Quantifying the causal effects of conservative fourth down decision making in the national football league. SSRN Electronic Journal, January. http://dx.doi.org/10.2139/ ssrn.3114242

30 ‘The Baseball Players’ Labor Market’: An Update Anthony C. Krautmann

It is well known that the professional sports industry provides an excellent opportunity to examine the salary determination process as the industry generates easily observable data on both the marginal revenues (MR) of teams as well as the marginal products (MP) of players. Scully (1974) wrote the seminal paper on this topic by estimating the Marginal Revenue Product (MRP) of baseball players in the pre-Free Agency era of the 1960s by separate models of the team’s revenue and production functions. By looking at how team performance determines wins, and then calculating the revenue impact of one additional win, Scully imputed the marginal value of different inputs on the playing field. Scully found that the reserve clause in Major League Baseball (MLB) resulted in players receiving compensation that was only about 20 percent of their MRP. This approach has a long history in the literature (Zimbalist, 1992; Bradbury, 2007). An alternative approach to imputing a player’s marginal value is based on the assumption that in competitive labor markets firms are willing to pay up to the marginal value created by the worker. When a player sells his labor on competitive markets, as is the case in free agency, we would expect the final negotiated salary to lie somewhere between what he is expected to be worth to the team (i.e., his MRP) and his next best salary offer

(see Solow & Krautmann, 2011). As the number of demanders increases, one would expect the gap between these two values to narrow to the point where the observed salary approaches the player’s MRP on the team with the winning bid. Under these conditions, one can use this approach to approximate a Free Agent’s MRP by looking at his competitively negotiated salary – the process underlying the free market model of MRP (Krautmann, 1999, 2013).1 Monopsony in professional sports labor markets typically arises from institutional structures associated with the league’s collective bargaining agreement. When a player is drafted into a league, property rights to his services are typically assigned to the team. In the case of MLB, those players with less than three years of Major League service (EXP) are constrained under a reserve clause which indentures them to their team. Such players (‘Apprentices’) must accept whatever salary the team offers, meaning this segment of the labor market is operating under the greatest degree of monopsonistic exploitation. Apprentices in MLB are typically paid about the minimum salary (i.e., $535,000 in 2017) regardless of their MP, resulting in arbitration-ineligible baseball players being paid only about 20 percent of their free-market value (Zimbalist, 1992; Krautmann, Gustafson, & Hadley, 2000).

‘The Baseball Players’ Labor Market’: An Update

At some point in the player’s career, however, the assignment of property rights shifts from the team to the player. Such players (‘Free Agents’) are able to market themselves to the highest bidder and would, under sufficiently competitive conditions, receive a salary more commensurate with their marginal value. To the degree that a Free Agent’s salary aligns with his MRP, these players labor under the least degree of monopsonistic exploitation. Between Apprentices and Free Agents are players laboring under market conditions that exhibit varying degrees of monopsony exploitation. In MLB, the salaries of those players with between three and six years of EXP (‘Journeymen’) are affected by their eligibility to the arbitration process.2 Arbitration greatly reduces the ability of the team to unilaterally impose its will upon the player because the player is able to have his salary bid considered alongside the team’s offer. In final-offer arbitration, both the team and the player present potential salary bids to an impartial arbitration board, and this board must choose one or the other of the two bids.3 This type of arbitration is argued to produce ‘reasonable bids’ from both sides because both the team and player realize that the winning bid will be that which is closest to the (impartial) arbitrator’s estimate of the player’s MRP. In practice, the vast majority of arbitrationeligible baseball players settle with their teams without the case actually going in front of the arbitration board. But simply being eligible for arbitration greatly evens out the bargaining power on both sides of the negotiation. Past studies indicate that the salaries of such players range from about 50 percent to about 80 percent of their free-market values (Zimbalist, 1992; Burger & Walters, 2005; Krautmann, von Allmen, & Berri, 2009). The primary purpose of this chapter is to analyze the baseball labor market and provide an update of the degree to which player salaries reflect free-market values. Before moving on to updating these traditional cohorts, we need to recognize that contract extensions (CE) have become

an important mechanism for superstar-players to increase their salaries. Figure 30.1 shows that over the past decade the number of contract extensions in MLB has grown from about 10 per year to more than 40 per year.4 Since players signing CE tend to be superstars, teams may be pressured to negotiate in good faith so as to avoid the uncertainty of losing control of the player. While it is not clear whether this type of negotiation process results in these players being paid their full market value, we check the nature of these salaries alongside that of the Apprentices, Journeymen, and Free Agents.

EMPIRICAL ANALYSIS Free Agents Using the free agent market as our benchmark, to which all other cohorts are compared, we begin by regressing Free-Agent salaries on players’ Marginal Products and teams’ Marginal Revenues. From this regression, the free-market value of any type of player can be imputed (regardless of his contract status) by applying the regression coefficients to the player-specific inputs. The sample used here is 166 potential Free Agents who signed new contracts over the 2012 to 2014 seasons, consisting of 50 starting pitchers and 116 position players. Table 30.1 shows that the average Free Agent has almost nine years of Major League experience, is about 34 years old, and generates an extra 1.4 wins over and above his replacement (WARP). The productivity of the Free Agents in the sample range from about (−1.0) to nearly 8 extra wins (see Figure 30.2). Starting pitchers generate a slightly lower number of wins than position players (1.19 versus 1.42), yet they make about $2 million more per season. While the typical Free Agent signed a new contract worth about $21 million over two years, there were some

80 64

60 40

35 39

20 0

6

14 17 14

299

46 49

38

50 48

40 42

22

Figure 30.1  Contract extensions in Major League Baseball (2001–2014) Source: www.mlbtraderumors.com/extensiontracker, ‘Extension Tracker’

22 24

300

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 30.1  Summary statistics: Free Agents (2012–2014) Entire Sample

Pitchers Only

Position Only

Variable

Mean

Std Dev.

Mean

Std Dev.

Mean

Std Dev.

EXP Dummy for Pitcher1 Dummy for Mid-fielder1 Dummy for Corner Fielder1 Annual WARP Age POP (1000s) Nominal Annual Salary Real Annual Salary ($2014) Size of FA Contract Length of FA Contract TREND # observations

8.84 0.30 0.12 0.14 1.35 34.54 6185.1 $6,774,000 $6,852,000 $20,600,000 2.05 1.030 166

3.00 0.46 0.40 0.34 1.26 3.17 4838.9 $5,490,000 $5,551,000 $38,700,000 1.61 0.81

8.80 1.0 — — 1.19 34.12 6467.6 $8,157,000 $8,240,000 $20,390,000 1.96 1.14 50

3.22 0.0 — — 0.98 3.38 5241.9 $5,097,000 $5,148,000 $26,606,000 1.34 0.81

8.85 — 0.28 0.20 1.42 34.72 6063.4 $6,178,000 $6,254,000 $20,642,000 2.09 0.98 116

2.91 — 0.45 0.40 1.35 3.08 4673.2 $5,567,000 $5,632,000 $43,000,000 1.72 0.81

1 The pitcher dummy variable (DUMpit) includes just starting pitchers. The midfielder dummy variable (DUMmid) controls for the positions of catcher, shortstop, second-base, and centerfield. The corner infielder dummy variable (DUMcorner) controls for the positions of first- and third-base. Designated hitters and corner outfielders are the omitted positions.

Albert Pujos

8 7 6 5 4 3 2 1 0 -1 -2 Roberto Hernandez

Figure 30.2  Free Agents’ WARP (2012–2014) blockbuster contracts included in the sample. For example, Albert Pujols (who generated almost 8 extra wins per year) signed a 10-year, $250 million contract in 2012. To estimate the marginal impact of the different factors determining the Free-Agent negotiation process, we regressed these players’ real salaries

($2014) on a number of economic variables. This regression model is given by: ln(REALSal) = a + b1 WARP + b2 WARP2 + b3 POP + b4 DUMpit + b5 DUMmid + b6 DUMcorner + b7 EXP + b8 EXP2 + b9 TREND + e

301

‘The Baseball Players’ Labor Market’: An Update

Equation (1) assumes that salaries are determined by the Free Agent’s expected productivity (WARP),5 the size of the team’s market (POP), positional dummy variables (DUMpit, DUMmid, DUMcorner),6 the player’s Major League service (EXP) and a trend variable (TREND). Huber-White estimates of equation (1) are presented below in Table 30.2.7 According to Table 30.2, a Free Agent’s salary increases with productivity and experience (both at a diminishing rate). Furthermore, starting pitchers and corner infielders receive a salary premium, while largemarket teams pay a slightly higher salary. These regression results imply that a Free Agent receives his highest annual salary after about 9.7 years of Major League experience. To illustrate the economic impact of the productivity variable WARP, we evaluate the effect on fitted salaries of a onestandard deviation increase above the mean in this key right-hand side variable. Comparing a player whose productivity is approximately equal to the sample mean (e.g., David Ross) to a player whose productivity is about one standard deviation above the mean (e.g., Nelson Cruz) results in an increase in the (estimated) annual salary of $4.6 million – about a 70 percent increase above the mean salary! This chapter is interested in analyzing how various degrees of monopsony impact the difference between what a player is worth and what he is paid. The approach taken here is to separate the sample into the different types of players discussed above, and then compare their actual salaries (SACT) to their imputed, free-market salaries computed using the above Free-Agent S FA

( )

salary regression. It is worth noting that this approach is reasonable regardless of whether a Free Agent’s salary converges to his MRP. The primary assumption here is that free-market salaries exhibit the least degree of monopsony exploitation, establishing the benchmark to which the salaries of the other types of players are compared. We then impute the degree of monopsony power by looking at the ratio of the player’s actual salary (SACT) and what he would have received if he were a Free Agent S FA .

( )

Contract Extensions We begin with a sample containing 110 players who signed contract extensions (CE) with their teams between the 2012 and 2014 seasons – 37 starting pitchers and 73 position players. Figure 30.3 below illustrates the distribution of WARP across the CE players and shows player-productivity ranges from about (−1) to (+7). Table 30.3 summarizes a number of characteristics of these CE players. With an average WARP of 2.6, these players are some of the most productive players in the league – twice as productive, in fact, as the sample of Free Agents used to estimate equation (1). Over half of these players have between three and six years of experience and would have otherwise been eligible for arbitration. The size of a typical contract is about $47 million over a duration of four years. While the average player in this group receives a salary of about

Table 30.2  Huber-White estimates of Free Agent salaries (2012–2014) – Dependent variable: ln (REALSal) Variable

Coefficient

Constant WARP WARP2 POP DUMpit DUMmid DUMcorner EXP EXP2 TREND # observations R2 F-stat

12.789** 0.585** −0.033** 0.00002* 0.537** −0.024 0.397** 0.311** −0.016** 0.262**

** Significant at 5% level * Significant at 10% level

Robust Std Error 0.427 0.074 0.012 0.000001 0.114 0.129 0.143 0.084 0.004 0.057 166 0.58 32.7**

t-stat 30.0 7.94 −2.75 1.76 4.72 −0.19 2.77 3.71 −4.17 4.63

302

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Miguel Cabrera 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 -1.0 -2.0 Jeff Mathis

Figure 30.3  Contract extension players’ WARP (2012–2014) Table 30.3  Means of contract extension players (2012–2014) Entire Sample WARP Experience Years Remaining on Old Contract Length of CE Size of CE SACT – Before New CE SCE – After New CE  SFA

(S (S



ACT / SFA



CE / SFA

)

)

# obs. (% of total)

Apprentices (EXP ≤ 3 Years)

Journeymen (3 ≤ EXP ≤ 7)

Veterans (EXP > 6)

2.6 5.0 1.1 4.0 $47.1 mil. $4.9 mil. $9.7 mil. $8.0 mil.

2.47 1.9 1.0 4.9 $26.8 mil. $0.59 mil. $5.2 mil. $4.6 mil.

2.44 4.7 1.1 3.7 $47.8 mil. $4.0 mil. $9.6 mil. $7.2 mil.

2.89 8.4 1.2 3.8 $63.7 mil. $10.5 mil. $14.2 mil. $12.7 mil.

0.64

0.22

0.67

0.95

1.41

1.55

1.41

1.28

110 (100%)

25 (23%)

57 (52%)

28 (25%)

Notes: SACT = the player’s salary the season before signing a new contract extension SCE = the player’s salary associated with the new contract extension S FA = the player’s imputed, free-market salary

$10 million per year, there are some very lucrative signings included here (e.g., Miguel Cabrera signed a new eight-year, $248 million contract extension with the Detroit Tigers in 2014). Although all three cohorts of CE players have similar WARPs, the average CE salaries (SCE) rises as one goes from Apprentices to Journeymen to Veterans (i.e., those with EXP > 6). This positive

correlation between SCE and experience suggests that some consideration of a player’s next-best alternative plays an important role in the contract extension negotiation process. Most CEs are re-negotiated while the player is in his last year of an old contract (i.e., YEARS REMAINING), and extends this contract for an additional four years (i.e., LENGTH). Apprentices

‘The Baseball Players’ Labor Market’: An Update

tend to get somewhat longer extensions than Journeymen or Veterans. The average Apprentice with about two years of experience receives a five-year extension, meaning that by the time the Apprentice’s CE expires, the team will have locked up control of that player past his arbitrationeligible years. Given that the average Journeyman’s new CE lock him up beyond his sixth year of experience suggests that perhaps the CE negotiation process provides teams with a means to control their young superstars and avoid the uncertainty associated with the arbitration process. S  The ratio  ACT  implies that, before signing  S  FA  the new contract extension, the average CE player was only receiving about 64 percent of his freemarket value. Apprentices are particularly underpaid – on average, their prior salary is only about 22 percent of their free-market salary. Signing a new contract extension, however, is quite lucrative for these players. Comparing the player’s prior salary to his new CE salary shows that an Apprentice’s salary rises nearly tenfold (from about SACT = $590,000 to SCE = $5.2 million), while a Journeyman’s salary more than doubles (from $4 million to $9.6 million). Even the highly paid Veteran sees a 40 percent increase in his salary (from $10 million to $14 million). Comparing the new contract salary to the freeS  market salary,  CE  , we see that CE players  S FA  are paid about 40 percent more than the freemarket salary! Such a salary premium is consistent with our earlier supposition that perhaps teams use contract extensions as a risk-management strategy to maintain control of their superstars (see Walters, von Allmen, & Krautmann, 2017).

Journeymen The next cohort examined is the group of Journeymen who are in the arbitration-eligible range of experience, over the 2012 to 2014 seasons. Table 30.4 summarizes the sample of 339 Journeymen salaries – 117 starting pitchers and 222 position players. Figure 30.4 illustrates the distribution of WARP across the Journeymen, and shows that player-productivity ranges from (−1) to (+6.5). Since going to arbitration entails just a single-year commitment by the team, Journeymen on multi-year contracts are different from those on single-year contracts. As such, we separate the Journeymen sample into the 75 percent of the players on single-year contracts versus the 25 percent on multi-year contracts.

303

We have categorized the Journeymen sample according to the players’ years of Major League experience in the following manner: • the 21% of Journeymen with between five and six years of MLS (and one year away from free agency) are labeled as EXP = 6; • the 30% of Journeymen with between four and five years of MLS are labeled as EXP = 5; • the 37% of Journeymen with between three and four years of MLS are labeled as EXP = 4; • the 13% of Journeymen (called ‘Super-Twos’) with between two and three years of MLS are labeled as EXP = 3. In the last few years, teams are increasingly holding superstars back at the beginning of their rookie season to take advantage of the eligibility rules regarding Super-Twos. Holding back a player at the beginning of his rookie season means that a Super-Two would only be eligible for arbitration three (instead of four) times. This is particularly cost-effective to teams because arbitration salaries are significantly higher than pre-arbitration salaries. Furthermore, holding a player back for the first couple of weeks of the season secures the team an extra year of contract control over the player before he is eligible for free agency. For example, the Chicago Cubs were widely criticized for using this strategy in regards to holding MVP Kris Bryant back for the first couple of weeks of the 2015 season. While Journeymen as a whole receive about 65 percent of their free-market value, they capture more of their value as their tenure rises. For example, a Super-Two with EXP = 3 makes only 38 percent of his free-market value while a Journeyman with EXP = 6 captures nearly 90 percent of his value. Journeymen with multi-year contracts are 55 percent more productive, and receive salaries that are twice as high as those on single-year contracts. As a result, monopsony rents are lower for S  these elite players – the ratio  ACT  of multi-year   SFA  Journeymen is 85 percent versus only 58 percent for the single-year players. Furthermore, those multi-year Journeymen who are one year away from free agency (EXP = 6) receive 99 percent of their free-market value (i.e., they receive essentially the same salary as a Free Agent).

Apprentices The most severely exploited cohort are the Apprentices. They have little, to no, negotiating

304

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 30.4  Means of Journeymen (2012–2014) Single-year WARP

S ACT

(S ) FA

# obs. (% of total)

Multi-year

Entire sample

1.41 $3.35 mil.

2.19 $6.89 mil.

1.60 $4.24 mil.

$6.25 mil

$9.73 mil.

$7.30 mil.

254

85

339

(75%)

(25%)

(100%)

( )

 Ratio of (SACT) to SFA

EXP=3 EXP=4 EXP=5 EXP=6 TOTAL

WARP

Single-year

0.94 1.57 1.79 1.80 1.60

0.378 0.501 0.629 0.845 0.578

Multi-year na* 0.790 0.775 0.989 0.845

Entire sample 0.378 0.570 0.676 0.898 0.645

# obs. (%Total) 45 (13%) 121 (37%) 102 (30%) 71 (21%) 339

* The only Super-Twos that signed multi-year contracts during this time period were Buster Posey, Gio Gonzalez, and Starlin Castro. Given such a small sample, we omitted them from consideration.

Andrew McCutchen 7 6 5 4 3 2 1 0 -1 -2 Justin Smoak

Figure 30.4  Journeymen’s WARP (2012–2014) power and consequently must accept their teams’ salary offers (or sit out the season). We analyzed the salaries of 558 Apprentices who appeared in at least one game between 2012 and 2014, consisting of 172 starting pitchers and 386 position players. We separated Apprentices in terms of their years of MLS in the following manner:

• if MLS is between 2.0 and 2.99 (and not SuperTwos), then EXP = 3; • if MLS is between 1.0 and 1.99, then EXP = 2; • if MLS is between 0.26 and 0.99, then EXP = 1; and • if MLS is between 0.0 and 0.26, then EXP = 0 (‘Rookies’).8

305

‘The Baseball Players’ Labor Market’: An Update

The reserve clause in MLB allows teams to renew an Apprentice’s contract year-by-year (i.e., singleyear contracts), suggesting that those Apprentices with multi-year contracts must be different from those with single-year contracts. Apprentices with multi-year contracts are rare – they comprise just 4 percent of the sample. In any case, we separate those Apprentices with single-year contracts from those on multi-year contracts. To compute the player’s free-market salary S FA , we need a reliable measure of his expected productivity. Unfortunately, many of these Apprentices have very short professional careers upon which to base the expected WARP. While The Baseball Prospectus does provide forecasts of these players’ WARPs, one must use caution here as the errors on these estimates are likely quite large. In this regard, we proceed in this discussion recognizing that our estimates could come with a potentially large degree of uncertainty. Figure 30.5 illustrates the distribution of WARP across the Apprentices, and shows that player-productivity ranges from about (−1) to (+7), although this (+7) is an extreme outlier given that the next highest player has a WARP of only about (+4). One way to view the degree of monopsony exploitation on these young players is to look at how close the Apprentice’s salary (SACT) is to the Major League minimum salary (SMIN). If owners overwhelmingly pay their Apprentices the Major League minimum, then one might surmise that these players’ salaries are primarily determined

( )

by the rules of the collective bargaining agreement – and have little to nothing to do with the player’s productivity. Table 30.5 shows that for the entire Apprentice sample SACT exceeds SMIN by only about $27,000 (although multi-year Apprentices are paid about $200,000 above the minimum salary). But even by the time a player has accumulated three years of experience (i.e., his last year before becoming eligible for arbitration), his salary still only exceeds SMIN by about $57,000. Across the entire sample, Apprentices receive only about 31 percent of their market value. In contrast to Journeymen, however, monopsony rents increase (rather than decrease) with experience. That is, even though the productivity of an Apprentice doubles as he matures from EXP = 1 S  to EXP = 3, the ratio  ACT  falls from 40 percent   SFA  to 21 percent. As further evidence that teams extract the greatest monopsony rents from the most productive Apprentices, note the difference between those on single-year versus multi-year contracts. Multi-year Apprentices are twice as productive as single-year Apprentices – yet their actual salaries are only 36 percent higher. As a result, multi-year Apprentices receive (a slightly) lower percent of their market value than those on single-year contracts. Both the impact of experience and multi-year contracts suggest that teams extract the greatest rents from their most valuable Apprentices.

Mike Trout 8 7 6 5 4 3 2 1 0 -1 -2 Brett Marshall

Figure 30.5  Apprentices’ WARP (2012–2014)

306

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 30.5  Means of Apprentices (2012–2014) Single-year Multi-year WARP SACT

0.874 1.764 0.909 $509,000 $693,000 $516,000 $2,240,000 $3,768,000 $2,300,000

(S ) FA

SACT - (SMIN) SACT = (SMIN)

(S



ACT / SFA

)

$19,500 11% 0.314

# obs.(% of total) 536 (96%) WARP EXP=0* EXP=1 EXP=2 EXP=3 TOTAL

Entire sample

0.561 0.639 1.029 1.170 0.909

$204,000 4.5% 0.273

$26,700 10.8% 0.313

22 (4%)

558 (100%)

SACT

SACT - (SMIN) SACT = (SMIN)

$491,000 $495,000 $516,000 $547,000 $516,000

$1,100 $5,900 $26,700 $57,000 $26,700

58% 13% 3% 2% 10.8%

(S ) FA

$1,225,000 $1,554,000 $2,411,000 $3,311,000 $2,300,000

(S



ACT / SFA

0.50 0.40 0.27 0.21 0.313

)

# obs. (%) 52 (9%) 167 (30%) 180 (32%) 159 (28%) 558 (100%)

* Rookie-year status if ≤ 45 days on active roster. Since 1 year of Major League Service (MLS) is 172 days, then when a player’s MLS ≤ (45/172) = 0.26, he is considered a ‘rookie.’ Thus, those with less than 0.26 of MLS are assumed (roughly speaking) rookies and listed as EXP=0 in Table 30.5.

CONCLUDING REMARKS The segregated labor market for professional baseball players exhibits varying degrees of monopsony exploitation. The most underpaid are, not surprising, Apprentices, who make about $500,000 and receive only about 30 percent of their free-market value. This result is somewhat higher than that reported in previous studies comparing SACT to MRP (Zimbalist, 1992; Burger & Walters, 2005). Previous research suggests that this underpayment of Apprentices may be a mechanism by which owners recoup their investments in player development in the Minor Leagues (Krautmann & Oppenheimer, 1996; Krautmann et al., 2000; Bradbury, 2007). Next in line are the arbitration-eligible Journeymen who are paid a little over $4 million and receive about 65 percent of their value – similar to what is reported in previous studies (Zimbalist, 1992; Burger & Walters, 2005). This finding, however, is somewhat misleading as the Journeyman’s salary goes from 38 percent to 90 percent of market value as he approaches free agency. In fact, the most productive Journeymen (those with multi-year contracts and are one year away from free agency) essentially receive a Free Agent’s wage! However, as a player gains experience and becomes more productive, teams treat Apprentices quite differently than Journeymen. As a Journeyman

gains experience, and approaches free agency, he captures a greater proportion of his freemarket value. Yet exactly the opposite occurs with Apprentices – as they gain experience, and approach arbitration, they receive a smaller proportion of their market value. Finally, we find that those superstars signing contract extensions are the least exploited of all – in fact, CE salaries exceed free-market values by nearly 40 percent! Such players are exceptionally productive, and their new contracts pay them handsomely (about $10 million per year over four years). This overpayment may reflect a risk premium that teams are willing to pay to give them the ability to avoid the arbitration process, as well as avoid the uncertainty associated with the future supply of Free Agents (Walters et al., 2017).

Notes 1  One cannot explicitly test whether the Free Agent’s salary converges to his MRP, as this is an assumption of the model, but clearly, better players generate greater MRP and are paid higher salaries. 2  Zimbalist (1992) was the first to label arbitrationineligible players ‘Apprentices,’ and arbitrationeligible players ‘Journeymen.’ We continue that nomenclature here.

‘The Baseball Players’ Labor Market’: An Update

3  The arbitration board is restricted to considering only these two bids, without recourse to selecting an intermediate salary in between the two bids. 4  The recent decline in the number of reported extensions may be a result of the uncertainty associated with a new Collective Bargaining Agreement (which went into effect in 2016). 5  The productivity variable (WARP) is defined as the number of wins above replacement attributed to this player. Since contracts are negotiated on the basis of expected productivity, we use a proxy for this expectation. This proxy is the average WARP of the player over the prior two seasons. 6  The pitcher dummy variable (DUMpit) includes just starting pitchers. The midfielder dummy variable (DUMmid) controls for the positions of catcher, shortstop, second-base, and centerfield. The corner infielder dummy variable (DUMcorner) controls for the positions of first- and third-base. Designated hitters and corner outfielders are the omitted positions. 7  Huber-White estimates yield robust standard errors in the presence of heteroscedasticity. 8  Rookie-year status is roughly less than 45 days on an active roster. Since one year of Major League Service (MLS) is 172 days, then MLS ≤ (45/172) = 0.26 implies the player is a ‘rookie.’ Thus, those with less than 0.26 of EXP are assumed (roughly speaking) to be rookies and are listed as EXP=0.

REFERENCES Bradbury, J.C. (2007). What is a ballplayer worth? In J.C. Bradbury, The Baseball Economist: The Real Game Exposed (pp. 176–200). New York: Penguin.

307

Burger, J.D., & S.J.K. Walters (2005). Arbitrator bias and self-interest: lessons from the baseball labor market. Journal of Labor Research, 26(2), 267–280. Krautmann, A.C. (1999). What’s wrong with scullyestimates of a player’s MRP? Economic Inquiry, 37(April), 369–381. Krautmann, A.C. (2013). What’s right with scullyestimates of a player’s Marginal Revenue Product: reply. Journal of Sports Economics, 14(February), 97–105. Krautmann, A.C., E. Gustafson, & L. Hadley (2000). Who pays for minor league training costs? Contemporary Economic Policy, 18(1), 37–47. Krautmann, A.C., & M. Oppenheimer (1996). Training in Major League Baseball: are players exploited? In J. Fizel, E. Gustafson, & L. Hadley (Eds.), Baseball Economics: Current Issues. Westport, CT: Greenwood. Krautmann, A.C., P. von Allmen, & D. Berri (2009). The underpayment of restricted players in North American sports leagues. International Journal of Sport Finance, 4, 75–93. Scully, G.W. (1974). Pay and performance in Major League Baseball. American Economic Review, 64(6), 915–930. Solow, J.L., & A.C. Krautmann (2011). A Nash bargaining model of the salaries of elite Free Agents. Journal of Sports Economics, 12(3), 309–316. Walters, S.J., P. von Allmen, & A.C. Krautmann (2017). Risk aversion and wages: evidence from the baseball labor market. Atlantic Economic Journal, 45(3), September. Zimbalist, A. (1992). Salaries and performance: beyond the Scully model. In P. Sommers (Ed.), Diamonds are Forever: The Business of Baseball (pp. 109–133). Washington, DC: Brookings Institution.

31 Economic Issues of the National Hockey League: A Survey of the Literature D u a n e W . R o c k e r b i e a n d S t e p h e n T. E a s t o n

The National Hockey League (NHL) has undergone more changes to its business model than any of the other three major North American sports leagues in the last 15 years. This has resulted in a corresponding increase in published research devoted to NHL issues. Our purpose is to provide a useful starting point for new researchers who wish to explore the economic issues facing the NHL, as well as to suggest future research directions. We chose to survey the fields that best reflect the body of published work on the economic issues facing the NHL. These include the lost season of the 2004–05 player lockout that resulted in significant subsequent labour market developments, salary determination, discrimination and violence. We could not survey all of the research to date here without creating an extensively long chapter and we apologize if we excluded any relevant papers.1

THE 2004–05 LOCKOUT AND SUBSEQUENT DEVELOPMENTS Negotiations with the National Hockey League Players Association (NHLPA) failed to reach a new collective bargaining agreement (CBA) in 2004–05. Because talks were at an impasse, the NHL

cancelled its 2004–05 season of 1,230 regular season games. All of the four major North American professional sports leagues have suffered labor stoppages, but until 2004–05 no league had cancelled a season since the establishment of the National League in major league baseball (MLB) in 1876. MLB has had eight work stoppages, but only three involved lost games. Table 31.1 displays the stoppages and the lost games among the major leagues. Losing the entire NHL season and the subsequent reaction in the negotiated CBA has galvanized interest among economists who try to understand both the causes and the consequences of the extreme stoppage. Staudohar (2005) identifies two root causes of the 2004–05 lockout. Although the 1993–94 CBA tightened the rules regarding free agency, the average player salary had more than tripled ($558,000 to $1.83 million) by 2004–05, partly due to the more frequent inclusion of large bonuses in player contracts that were not subject to maximum salary rules in the CBA. Revenues expanded more modestly, although buoyed by expansion fees obtained from the addition of nine new teams over the period (around $500 million in total). However, US television revenues were cut in half in the 2002–03 season and the Levitt Report identified total losses of $273 million on revenues of $2 billion.2 Salaries ate up 75% of league revenues, much more than in the NFL (56.1%), NBA (57%) and MLB (55.4%).3

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

309

Table 31.1  Games lost due to work stoppages Season NHL NFL NBA MLB

2004–05 2012–13 1982 1987 1998–99 2011–12 1972 1981 1994 1995

Games lost

Games lost (%)

1,230 510 98 28 464 240 85 874 668 251

100 41 44 6.5 39 39 4.4 39.5 29.5 11.1

Source: www.hockey-reference.com

The 2004–05 NHL CBA was a bold and harsh step toward cost containment with few benefits for the players. Its main features for our purposes include: • A team payroll cap of $38 million with annual increases limited to 54% of anticipated league revenue. The maximum salary was capped at $7.8 million (20% of team payroll), while the minimum salary was increased from $175,000 to $450,000. The maximum rookie salary was capped at $850,000. Players took an immediate pay-cut of 24%. • More relaxed free agency. Players could become unrestricted free agents at age 27, instead of 31 in the 1993–94 CBA, although this was instituted over a number of seasons. • A revenue sharing system that required the top ten revenue teams to contribute 4.5% of league revenue into a fund to be redistributed to the bottom ten revenue teams. The salary cap is adjusted each season according to anticipated league revenues. The maximum and minimum team payrolls are set at $73 million and $54 million for the 2016–17 season. The revenuesharing contribution rate paid by the top ten revenue teams increased to 6% of league revenue in the 2013 CBA (which runs until the 2023 season). No significant changes in free agency rules have occurred since the 2005 CBA. There is a large literature on the theoretical effects of salary caps, free agency and revenue sharing with profit-maximizing team owners that we will not review here. Suffice it to say that salary caps reduce payrolls but are not profit-maximizing (Fort & Quirk, 1995); free agency reduces the league monopsony on talent and increases salaries (Krautmann, 1999); revenue sharing reduces the

demand for talent, thus reducing salaries and worsening competitive balance (Driskell & Vrooman, 2016; Késenne, 2015; Vrooman, 2009).4 There is a definite paucity of literature on whether these CBA innovations had the predicted effects using postlockout NHL data. This is important since here is a unique instance where a professional sports league adopted these three policies where almost none existed before, and there has been a full eleven seasons of data since. An exception is Buschemann and Deutscher (2011). These authors estimate the degree of economic efficiency of NHL teams by using revenue and team values as the measures of output in a frontier analysis approach. The sample covered the four seasons before and after the 2004–05 lockout. All teams became more efficient in achieving higher franchise values after the lockout, but most notably for the bottom ten performing teams in the league. A few papers have studied lockout-related issues. Winfree and Fort (2008) estimated an attendance regression model to test for substitution effects towards minor (NHL affiliates, players are paid) and junior (no NHL affiliation, amateur players) hockey teams during the lockout season. Both types of teams experienced a significant increase in attendance, although the minor league teams’ (mostly American Hockey League) increase was uniform whether the team was located in an NHL city or not. The increase was larger for junior league teams (mostly Canadian Hockey League) that operated in NHL cities. The authors conclude that the substitution effect could work the other way: when NHL teams move into the city of a minor or junior hockey league, their attendance declines. This is left untested. Jasina and Rotthoff (2016) found that the 2004–05 lockout had no effects on employment relative to trend in NHL county areas but was associated with a small decrease in employment income. These results suggest that the

310

THE SAGE HANDBOOK OF SPORTS ECONOMICS

240

Series: SALARY2003 Sample 1 747 Observations 747

200 160 120 80 40 0 5

2000005

4000005

6000005

8000005

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis

1768427. 1100000. 11000000 150000.0 1830582. 2.307934 8.817905

Jarque-Bera Probability

1716.676 0.000000

10000005

Figure 31.1  NHL salary distribution, 2003–04 season (2003 = 100) Source: www.nhlpa.com

possible job loss due to the lockout was countered by job increases in other industries. Poplawski and O’Hara (2012) estimated a gate revenue model for a revenue maximizing team owner using data for the 2005–09 seasons. They then assessed the probability of new NHL franchise locations to earn positive profits by inserting their market and performance statistics into the estimated model. Quebec City, Hamilton, Winnipeg, Portland and Houston were identified as good candidates (the Atlanta Thrashers have since moved to Winnipeg). Figures 31.1 and 31.2 provide histograms of salaries reported by the NHLPA for the 2003–04 and 2013–14 seasons respectively. Salaries for the

latter season are deflated using the US consumer price index (2003 = 100). The average real salary increased by only 9.69% in ten years. The coefficient of variation for the two seasons are 1.0446 and 0.8402, indicating that the relative variability in salaries is lower after the imposition of the salary cap. The decrease in the skewness coefficient indicates that there is much less salary in the right tail of the distribution. Only one player earned the minimum salary of $175,000 in the 2003–04 season, but by 2013–14, 29 players earned the minimum salary of $550,000. At the high salary range, 18 players earned over $8 million in 2003–04 and only four in 2013–14. The statistics indicate a substantial

300

Series: SALARY2013 Sample 1 747 Observations 747

250 200 150 100

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis

1939722. 1263823. 11058452 414691.9 1629766. 1.411086 5.510175

Jarque-Bera Probability

444.0178 0.000000

50 0 5

2000005

4000005

6000005

8000005

10000005

Figure 31.2  NHL salary distribution, 2013–14 season (2003 = 100) Source: www.nhlpa.com

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

degree of salary compression with very high salaries being less prominent (cap imposed maximum of $12.86 million equal to $10.15 million in 2003 dollars). The Gini coefficient provides a measure of salary equality that is bounded by zero and one with values closer to zero indicating greater equality. The Gini coefficient decreased from 0.4803 in the 2003–04 season to 0.4381 in the 2013–14 season. Overall, the salary cap introduced in the 2005 CBA has significantly reduced the growth in salaries and the minimum and maximum salary caps have made the salary structure more even. It makes sense that this should translate into lower salary dispersion for individual teams and that this might affect team performance (Bloom, 1999). Depken (2000) found that greater salary compression is associated with better team performance in MLB. Frick, Prinz and Winkelmann (2003) and Mondello and Maxcy (2009) found the same result for the NFL. A useful paper that combines a lengthy sample period of six NHL seasons and the use of the conditional team salary dispersion measure is Kahane (2012).5 The results were robust to different model specifications and suggest that team performance and team salary dispersion are negatively associated. How did the 2005 CBA affect league profitability? The salary cap and revenue sharing should have both improved profits and that seems to be the case based on estimated team profits from Forbes magazine.6 Forbes estimates a total loss of $96 million for the 2003–04 season with 17 out of 30 teams experiencing losses. The 2013–14 season saw an estimated total profit of $444.1 million ($350.8 million in 2003 dollars) with only ten teams suffering losses.7 The average loss decreased slightly from $9.74 million in 2003–04 to $8.2 million ($6.48 million in 2003 dollars), although the standard deviation of losses dropped from $8 million to $5.19 million ($4.1 million in 2003 dollars). The financial situation for all teams has improved, with the exceptions of the Florida Panthers and Columbus Blue Jackets, whose losses have increased since the 2005 CBA. The effect of the lockout experience proved sufficiently costly to both parties that the new CBA lasted for a decade with changes in many dimensions. It has proven to be an extreme example that neither party has appeared to want to repeat. Understanding the longer-term consequences of such shock therapy allows economists to begin to understand the implications of the major clauses of the agreement.

SALARY DETERMINATION The sports economics literature has considered two different approaches to player salary determination.

311

The first approach, developed by Scully (1974), assumes that talent is scarce, and its supply is upward-sloping in the market wage per unit of talent. This gives rise to a monopsony market in which players are paid less than their marginal revenue product (MRP) according to their bargaining power. The task is to determine econometrically how much players are exploited by team owners in different periods of player mobility. The method requires the estimation of a team revenue function that includes the team winning percentage in some form (quadratic or lagged), as well as estimating a winning percentage function that includes team and individual player performance variables. Differentiating each function allows for the calculation of MRP = MR × MP for each player. Papers that have employed the Scully (1974) method for NHL players have been sparse. While MLB adopted free agency extensively in its 1976 CBA, the NHL did not adopt a more restricted free agency until the 1993 CBA. However, the NHL did face the rival World Hockey Association that could have bid up salaries until the 1979 merger. Jones and Walsh (1987) estimate MRPs using the Scully method using data from the 1975–78 seasons and found that NHL player salaries were bid up to the point where they exceeded the MRP for many players. Richardson (2000) models the NHL player market as a bilateral monopoly whose solution is estimated best by the ultimatum game described by Telser (1995) in the reserve clause period.8 Although only 48 players qualified for free agency, the limited free agency rules incorporated in the 1993 CBA did move salaries closer to their MRPs in the 1993–94 season. The second approach to player salary determination, outlined first in the sports economics context by Krautmann (1999), begins with the premise that players are paid their MRPs and that players should be paid their MRP in a competitive bidding market when facing free agency.9 The method is simple: regress player salaries on team and individual performance statistics to determine what characteristics most contribute to salaries. The results can be used to predict a player’s salary but do not have the same ability to determine if the wage = MRP rule holds. Salary regressions are commonly used in the discrimination literature that we survey elsewhere in this chapter. Here we describe papers that deal with player mobility and differences in player attributes directly. Idsen and Kahane (2000) estimate a salary regression model using data for the 1990–92 NHL seasons to test whether free agency increases player earnings. They include a dummy variable equal to one if the player was a free agent at the end of the previous season and found that it was highly significant. Vincent and Eastman (2009) ask the same

312

THE SAGE HANDBOOK OF SPORTS ECONOMICS

question, but instead use lifetime player statistics for the 2003–04 season and a dummy variable equal to one if a player moved to a new team at any time during his career. Holding player productivity constant, the results suggested that moving to new teams does not increase salaries significantly, although moving to a new team may not be the result of free agency. Since the threat of moving can have the same effect as moving, looking only at those who moved is not a powerful test. Kahane (2001) tests for differences in rewarding players among NHL teams. The hierarchical linear model (HLM) approach he uses employs team revenues as the cause of the team differences.10 Kahane (2001) found significant differences in how teams compensate players of the same productivity (essentially the marginal revenue part of the MRP). Vincent and Eastman (2009) recognize that players at different positions in the salary distribution may be compensated differently for offensive and defensive performance.11 They estimate a quantile salary regression that includes performance measures, team revenue, player experience, height, weight and junior hockey experience using data from the 2003–04 NHL season. The model fit well for high-paid players, but poorly for low-paid players, suggesting that team owners look to other attributes when deciding what to pay less productive players. NHL players are eligible for salary arbitration at the end of a season when they are eligible for restricted free agency (which requires at least four seasons of NHL service).12 The goal of the

arbitrator is to award a salary that is an accurate representation of a player’s MRP. The decision of the arbitrator is binding. Lambrinos and Ashman (2007) test the efficiency of arbitrated salaries by first estimating a salary regression model that includes performance variables from the player’s previous season. Predicted salaries can then be compared to arbitrated salaries for players who go through arbitration. The data set included salaries for the 2001–02 season and performance statistics for the 2000–01 season, a rather limited sample that included only 17 arbitrated players. The results suggested that arbitrated salaries are no different than negotiated salaries. A related issue is whether coaching turnover affects team performance. If coaching ability enters the production function of team performance, it will affect the marginal product of the players and influence their salaries. High coaching turnover could be a signal of poor coaching ability and lower player salaries. Audas, Goddard and Rowe (2006) found that the coaching turnover rate had no effect on team performance by estimating a hazard function that predicts coaching departures using an extensive data set for the 1967–2002 NHL seasons. We have updated their table to include data to the end of the 2016–17 NHL season (in Table 31.2). Figure 31.3 provides a scatter diagram. The correlation coefficient of −0.18443 was not statistically significant at 95% confidence (t = −0.993), suggesting that coaching turnover rates are not associated with average winning percentages.13

Table 31.2  Coaching turnover and winning percentage, 1967–68 to 2016–17 Team Anaheim Arizona/Winnipeg Boston Buffalo Calgary Carolina/Hartford Chicago Colorado/QC Columbus Dallas/Minnesota Detroit Edmonton Florida

No. of coaching changes  7 16 19 19 16 15 19 14  6 26 22 15 10

1967–2016 seasons 21 35 47 44 42 35 47 35 14 47 47 35 21

Turnover rate 0.333 0.457 0.404 0.432 0.381 0.429 0.404 0.400 0.429 0.553 0.468 0.429 0.476

Average winning % 0.5411 0.4785 0.5998 0.5396 0.5335 0.4782 0.5412 0.5191 0.4795 0.5221 0.5367 0.5179 0.4923 (Continued)

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

313

Table 31.2  (Continued) Team

No. of coaching changes

Los Angeles Minnesota Montreal Nashville New Jersey/Col/KC New York I New York R Ottawa Philadelphia Pittsburgh San Jose St. Louis Tampa Bay Toronto Vancouver Washington Winnipeg/Atlanta

1967–2016 seasons

23  4 20  1 28 17 26  7 19 23  6 30  8 22 19 15  5

47 14 47 16 40 42 47 22 47 47 23 47 22 47 44 40 15

Turnover rate

Average winning %

0.489 0.286 0.426 0.063 0.700 0.405 0.553 0.318 0.404 0.489 0.261 0.638 0.364 0.468 0.432 0.375 0.333

0.4914 0.5433 0.6008 0.544 0.4916 0.5073 0.5459 0.5208 0.5757 0.5123 0.5324 0.5284 0.4773 0.4883 0.4886 0.5211 0.4719

Source: www.hockey-reference.com/ 0.61 0.59 0.57

Point %

0.55 0.53 0.51 0.49 0.47 0.45 0.000

0.100

0.200

0.300

0.400

0.500

0.600

0.700

0.800

Turnover Rate

Figure 31.3  Turnover rate versus winning percentage, 1967–68 to 2016–17

DISCRIMINATION Discrimination is a popular research topic in the NHL sports economics literature. While research into discrimination in MLB, NFL and the NBA has focused on racial divides among black and white players, researchers who study the NHL have focused

on gaps between French-Canadian, European and English-Canadian (including American) players. Longley (2012) gives a detailed review of the literature. European players were non-existent in the NHL in 1917 when the league was formed and did not feature in the league in significant numbers until the 1980 NHL draft that allowed non-North American

314

THE SAGE HANDBOOK OF SPORTS ECONOMICS

players to enter the amateur draft.14 French-Canadian players were abundant in the early years of the NHL and constituted a significant share of all players.15 The labour market literature has studied three outcomes of discrimination. Salary discrimination occurs when some workers are paid less than others on the basis of characteristics not directly related to productivity, such as race, ethnicity, sexual orientation, etc. Since accurate salary data for individual players of different skill levels are readily available, sports economists have focused on different aspects of salary discrimination. Segregation discrimination occurs when certain jobs are reserved for specific groups whose members may be paid the same wage, but differ in non-economic characteristics, such as race, gender, religion, or country of origin. In the context of sport, specific playing positions are reserved for stereotyped athletes (white versus black quarterbacks, for example). Entry discrimination occurs when firms restrict entry into a workplace by screening applicants and restricting their rights to be hired by the team that will pay them the highest salary. The first NHL amateur draft was held in 1963 on the argument that not all teams (there were only six at that time) had an equal ability to acquire talented players. The amateur draft also restricted the rights of amateur players since their NHL rights were held by the team that drafted them. Testing for salary discrimination in the NHL typically involves estimating a salary regression model that is developed from human capital theory that equates the player’s wage to his marginal revenue product (MRP).16 A sample of player salaries17 is collected for a season or two (yi for player i) and the log of these salaries is regressed on an ad hoc set of performance statistics (xi) and experience (ei, sometimes including an interaction effect between performance and experience) intended to represent his productivity. The marginal revenue part typically is handled by including a measure(s) of market size or team revenue as an independent variable. Discrimination, for example against FrenchCanadians, is modeled using a dummy variable (zi = 1 for French-Canadian, zi = 0 otherwise). If its estimated coefficient, d, is negative and statistically significant, this constitutes support for salary discrimination against French-Canadians.

lnyi = a + b´xi + cei + dzi(1)

This method overestimates the discrimination coefficient d if important omitted performance vari­ ables are correlated with the dummy variable (Goldberger, 1984), so it is important to include a good number of variables that are thought to be representative of what a general manager is using in salary negotiations. This is a serious problem in

the literature. Generally, those researchers who use few productivity variables find evidence for salary discrimination: Boucher (1984) and Longley (1995) used only one productivity variable (points) for the 1977–78 and 1989–90 NHL seasons respectively, Jones and Walsh (1988) used two (1977–78 season), Curme and Daugherty (2004) used three (1998–99 season). Other papers have found no evidence of salary discrimination (Bruggink & Williams, 2009; Jones, Nadeau, & Walsh, 1999; Lavoie, 2000; Lavoie & Grenier, 1992; McLean & Veall, 1992) for various seasons. It is hard to draw any firm conclusions from the literature except to say that salary discrimination appears to be less prevalent the more recent the NHL season. However, other modelling problems make even this result uncertain. For example, often independent variables are included that reflect the physical characteristics of the player: height, weight, age. It is likely that these variables do not affect a player’s salary directly, but instead affect the performance measures (goals, points, plusminus, penalty minutes, and so on) of the player. Given two players of equal ability, experience and cultural background, does a general manager pay the player who is ten pounds lighter any differently? This multicollinearity does not bias the coefficients but does increase their standard errors, making it difficult to assess their statistical significance. Some papers also ignore the effects of market size (Boucher, 1984; McLean & Veall, 1992), but it has become standard practice to summarize market size with an estimate of team revenue. The small number of French-Canadian players make it difficult to reach meaningful conclusions (Krashinsky & Krashinsky, 1997). Longley (1995) included only six French-Canadian players from the 1989–90 season.18 There were 41 Quebecborn players who played ten or more games in the 2015–16 season19 out of a total of 397 Canadianborn players and 867 players in total. It might be more fruitful to focus on salary discrimination toward European and American players who make up a much larger share (26.4% and 24.6% respectively) of NHL players today. Criticisms of the regressions begin with the selection criteria for players to include. The more games played, the better the precision. However, studies have included players with as few as ten games played. More generally, there is a time mismatch between the date at which a salary contract is signed and the season used for the performance measures. Salaries are presumably agreed upon based partly on past performance and the expected future performance of the player. Hence the negotiated salary never reflects only the performance of a single season in the duration of a player contract and may bias the coefficient estimates.

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

If salary discrimination is persistent in the NHL, many more than a single season are needed to estimate a salary regression. A pooled sample (with fixed effects) of ten or more seasons could reveal if salary discrimination has any long-run pervasiveness. Unfortunately, the 2005 collective bargaining agreement implemented a salary cap for the first time since the late 1940s, imposing a salary constraint that will affect any econometric estimates.20 With a more limited budget, teams may be more discriminating in the players they acquire. Using pre- and post-salary cap data (in separate regressions or pooled with dummy vari­ ables to break the sample) could provide interesting results into the salary cap effect on discrimination.

ENTRY DISCRIMINATION A small literature exists on entry discrimination in the NHL, particularly for French-Canadian players. Generally, two methods have been used to test for entry discrimination.21 The first uses simple statistical tests. Lavoie and Coulombe (1985) found that lifetime points for French-Canadian players was significantly greater than for EnglishCanadian players, suggesting that French-Canadian players need to be relatively better players to earn a spot in the NHL. They demonstrated that the same result held for black players in MLB and interpreted this as confirmation of discrimination. Lavoie, Grenier and Coulombe (1987) found the same result for French-Canadian and European players. Longley (2000) considered the representation of French-Canadian players on non-Quebec Canadian teams and US teams using frequency analysis. He found evidence that English-Canadian NHL teams discriminate against French-Canadian players, while US based teams do not, using a lengthy sample period (1943–98). American, European and French-Canadian players composed 24.5%, 25.2% and 5.3% of all NHL players in the 2015–16 season.22 The observed number of American-born players on Canadian teams was 57 out of 277 players; the observed number on US teams was 204 out of 788 players. Comparing observed and expected frequencies resulted in a chi-squared statistic equal to 2.36, less than the critical value of 5.99 with two degrees of freedom, suggesting that the frequencies of the two groups do not differ from the overall league frequency. Removing the Montreal Canadiens did not change the result. With a chi-squared equal to 1.469 and 0.611 respectively, we were unable to find evidence of entry discrimination for European or French-Canadian players.

315

The second method to test for entry discrimination uses a regression model where lifetime points for each player is regressed on draft position and dummy variables for playing position and ethnic origin. Lavoie, Grenier and Coulombe (1987) argue that French-Canadian players must work harder and demonstrate better performance than EnglishCanadian players due to entry barriers to the NHL. They test whether French-Canadian players demonstrate superior lifetime NHL points given the same draft position as English-Canadian players by regressing lifetime points on a quadratic function of draft position and ethnic background (using data from the 1983–84 and 1984–85 seasons). The results suggest that entry discrimination exists for French-Canadian players. However, Walsh (1992) found evidence to suggest that, rather than discriminating based on ethnic origin, NHL teams value defensive skills and size when evaluating players to draft and that French-Canadian players did not meet these criteria during most of the 1980s. In summary, salary and entry discrimination against French-Canadian players appeared to be likely through the 1990s, particularly during the two episodes of Quebec unrest (Longley, 2000), but there is little evidence of discrimination against European and American players. Of course, the most recent seasons are the most relevant for any policy recommendations to the NHL brass and little has been analyzed recently.

FIGHTING IN ICE HOCKEY Compared to other team sports, fighting during a game is a distinguishing feature of ice hockey, especially in the NHL. Rule books for the NFL, NBA and MLB specify that fighting is to be penalized with immediate disqualification from the game (rules 12.3, 12A and 6.04 respectively). Rule 46 in the NHL rulebook outlines the rules of engagement for fighting, as well as the penalties that can be imposed for ‘unsportsmanlike’ fighting (kicking, pulling off an opponent’s equipment, and so on).23 The speed of the game and the smaller ice surface in comparison to other sports results in frequent player contact, some of which can be quite violent. Players usually come to fighting as the result of one of three distinct situations. In the first, one player may believe that the physical contact initiated by another player is excessive and represents an intent to injure one of the players on his team. The second situation is less frequent. Two combatants agree to fight to settle a score from a fight or hit that may have occurred earlier in the game or in a previous game. A third situation is

316

THE SAGE HANDBOOK OF SPORTS ECONOMICS

when a team needs to be energized and a fight is one way to do so. Goldschmied and Espindola (2013) used data from the 2010–11 NHL season to examine whether fighting is just an impulsive action by players due to the physical contact of the game, or if fighting is a calculated action to gain an intimidating advantage over the opponent. Fights were found to be more likely to occur early in regular season games when the apparent costs are lower. Fighting was more frequent in pre-season games when the cost was nil, but almost non-existent in playoff games when the cost could be very large. The authors concluded this was evidence that players consider the benefits and costs of their actions when choosing to fight. They further suggest that raising the costs of fighting by increasing penalty time, fines and suspensions could be an effective deterrent to fighting. Sports fans who do not understand the unwritten code of fighting in ice hockey sometimes view fighting in hockey as a primitive method to injure an opponent and advocate that it should be banned. Hockey players do not share this position, despite suffering the possible injuries from fighting. In fact, in a 2011 poll of NHL players, 98% opposed a ban on fighting, arguing that fighting makes the game safer by deterring other types of injury-causing violence (hitting with the stick, hits to the head, violent boarding, and so on).24 Figure 31.4 plots the number of fights per game from the 1981–82 season through the 2013–14

season. The incidence of fighting began a steady increase in the 1967–68 season (expansion season from six to 12 teams), peaking at 1.31 fights per game in the 1987–88 season (21 teams), followed by a steady decline to 0.31 fights in the 2015–16 season. One explanation is that the rapid expansion of teams created a shortage of highly skilled players. Teams that managed to acquire these valuable assets needed to protect them on the ice by keeping specialized fighters on the roster to intimidate opponents. The influx of European and American players in the 1990s increased the overall quality of the league, reducing the need for their protection. More recently, the team salary cap (since the 2005–06 season) has limited the ability of teams to keep specialized fighters on their rosters. These factors have not been tested in the sports economics literature. Instead, research has focused on the effects of greater enforcement of the rules to limit physical play that may result in fighting, and the effects of fighting on game attendance and revenue.

GREATER ENFORCEMENT OF RULES The NHL added a third referee for the 1998–99 season with the intention of increasing the detection of penalties, thus increasing the probability of incurring the cost of a penalty, but not the cost of the

1.4 1.2 1 0.8 0.6 0.4

1960-61 1962-63 1964-65 1966-67 1968-69 1970-71 1972-73 1974-75 1976-77 1978-79 1980-81 1982-83 1984-85 1986-87 1988-89 1990-91 1992-93 1994-95* 1996-97 1998-99 2000-01 2002-03 2004-05* 2006-07 2008-09 2010-11 2012-13 2014-15

0.2

Figure 31.4  NHL fights per game, 1960–2015 Source: www.dropyourgloves.com

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

penalty itself. Using an economic model of crime, Allen (2002) found that occurrences of minor penalties decreased with the extra referee, but the occurrences of major penalties increased and appeared to be more random. Heckelman and Yates (2003) noted that two effects can result from adding more referees: the deterrent effect and the monitoring effect. The deterrent effect is observing a reduction in attempts by players to commit major penalties due to the greater likelihood of being caught. The monitoring effect arises from a greater ability by officials to detect penalties, resulting in more penalties called, without any increase in attempts to commit penalties. The authors found evidence that the monitoring effect was significant but that there was no deterrent effect, suggesting that increasing the costs of committing major penalties could be an effective deterrent. Depken and Wilson (2003) found that the addition of the second referee increased scoring and reduced fighting. Wilson (2005) found the same results. Allen (2005) extended his crime model with no qualitative changes in the results. By contrast, Levitt (2002) could find no evidence that the third referee reduced violent behavior. Fighting is easily detected and punished, unlike other types of major penalties (spearing, boarding, kicking, etc.), so it is not likely that the introduction of the third referee alone explains the dramatic decrease in fighting in Figure 31.4.

FIGHTING AND REVENUES Owners might encourage fighting as a way to improve their bottom lines. Evidence of the effects of fighting and rough play on attendance is mixed.25 Jones (1984) estimated an NHL attendance model that included independent variables that measured violent behavior. The results suggested that a team could increase ticket demand by fighting. Stewart, Ferguson and Jones (1992) construct a model of a profit-maximizing club owner who considers violence to affect attendance by the way it affects wins and through the ‘blood lust’ of the fans. The econometric results suggest that fighting is an effective strategy to increase attendance and revenue. Using data from the 1983–84 NHL season, Jones, Ferguson and Stewart (1993) found that American fans responded positively to violence, but Canadian fans responded negatively. Jones, Stewart and Sunderman (1996) repeated the estimation using a single attendance equation and data from the 1989– 90 NHL season and found similar results. Paul (2003) found that both American and Canadian fans preferred fighting, with American fans being somewhat more responsive from a game-by-game attendance model for the 1999–2000 NHL season.

317

Rockerbie (2012) extended the work of Jones et al. (1993) to include a much longer sample period (1997–2009) and found that doubling the number of fights per season (an historically large increase) decreased attendance by only 1.63%, suggesting that fans in the US or Canada were not particularly responsive to fighting. The attendance model was augmented in Rockerbie (2016) to account for limited arena capacities and exogenous ticket pricing, although fans were still found to be unresponsive to fighting.

CONCLUDING REMARKS The economic research on issues facing the NHL are constantly evolving due to the rather rapid changes in the business practices of the league in the last 15 years. Yet we believe that the pace of published work is still slow in comparison to the other major sports leagues, and even college sports, in North America. The 2004–05 lockout and subsequent CBA provides a useful laboratory to test the theoretical predictions of salary caps, revenue sharing, divisional realignments, rule changes, free agency and other key developments in the sports economics literature. The increasing number of post-lockout seasons should provide a rich data set for researchers. Useful areas for research using preand post-lockout data could include testing for a MRP-salary gap as evidence of monopsony; isolating the effects of the peculiar revenue-sharing model used by the NHL on team incentives; and analyzing the effects of coaching turnover on team performance. Historically, the NHL literature has focused on discrimination and violence. However, these fields may have run their course as a consequence of the multicultural integration of players and greater penalties for particularly violent rules offenses.

Notes 1  For example, there is an interesting literature on how changes to the NHL overtime rules and overtime points awarded, which began in the 1999–2000 season and changed again in the 2005–06 season, have created incentives to change effort during a game and allocating effort between conference and non-conference games. See Abrevaya (2004), Easton and Rockerbie (2005), Longley and Sankaran (2005), Shmanske and Lowenthal (2007) and Lopez (2015). 2  Levitt, A. (2004). Independent review of the combined financial results of the National Hockey League 2002–03 Season. Retrieved from: www2. nhl.com/images/levittreport.pdf

318

THE SAGE HANDBOOK OF SPORTS ECONOMICS

3  The NHL reported figures are not without criticism. Numerous expenses are deducted from revenue sources to arrive at a final revenue figure. For instance, team stores that operate outside the arena often report no net revenue since expenses incurred in operating the stores are deductible. See D. Ebner (2012). Accountant blows whistle on numbers. The Globe and Mail, September 21. Retrieved from: www.theglobeandmail.com/ sports/hockey/accountant-blows-whistle-onnumbers/article4560882/ 4  The typical form of revenue sharing in the theoretical assumes a simple gate sharing plan or a central pool plan with equal shares for all teams. The NHL system does not fit into these simple models even today, so making accurate predictions is somewhat difficult. An exception is Rockerbie and Easton (2017) for MLB. 5  The conditional salary dispersion measure is constructed using the standard deviation (or other dispersion statistic) of the residuals from a regression of player salaries on performance measures. In this way, differences in salaries due to differences in player productivities are held constant. 6  Taken from Rod Fort’s business of sports economics website. https://umich.app.box.com/s/ 41707f0b2619c0107b8b/0/320029263 7  It is not clear if the 2013–14 estimates are net of revenue sharing or not. 8  In this game, owners provide a fixed amount of money to be divided by the players each season. The most productive players will demand the lion’s share of the money, leaving little for the low productivity players, knowing that the low productivity players will accept any reasonable offer. The surplus to the owner is predicted to increase the lower the MRP of the player. 9  One shortcoming is that the Scully (1974) method estimates the MRP of a player by simply adding his or her contribution to the team performance statistics (goals, assists, home runs, batting average, etc.) without removing the contribution of the player being replaced. See Bradbury (2013) for a rebuttal and a response in Krautmann (2013). 10  The main advantage of the HLM approach is the use of fewer degrees of freedom in the estimation, resulting in more results of greater confidence. 11  Forwards are expected to obtain more points than defensemen, hence most studies separate players into different player categories. In fact, Chan, Cho, and Novati (2012) identify four forward types, four defensemen types and three goalie types that contribute to the team’s production using cluster analysis for the 2005–2010 NHL seasons. This might be too fine a level of disaggregation. 12  Restricted free agency requires signing teams to surrender draft picks to the team at the loss of the player. This can severely limit the bidding

for a restricted free agent so players often opt for salary arbitration until they can become an unrestricted free agent. 13  Winning percentage computed as total points divided by total maximum points in each season. 14  European players composed 26.8% of all NHL players at the start of the 2016–17 season. www. quanthockey.com/nhl/nationality-totals/nhlplayers-2016-17-stats.html 15  The Montreal Canadiens were granted the right of first refusal of French-Canadian players with the establishment of the National Hockey Association in 1909 (Ross, 2016). 16  The gap between the wage and the player’s MRP induced by monopsony power is typically ignored since the exercise is not to determine if players are actually paid their MRPs. Jones (1988) is an exception. Scully (1974) is the first attempt to measure a player’s MRP. When facing free-agency, it is more likely that a player is paid close to his MRP or even higher (Krautmann, 1999). Ultimately, the elasticity of the talent supply curve can be used to estimate the marginal resource cost and the wage gap at the existing level of talent. We have not seen this done. 17  The National Hockey League Players Association (NHLPA) has made individual player salaries publicly available since the 1989–90 season. www.nhlpa.com/the-players/team-compensation 18  Five of them played for one team – the Toronto Maple Leafs, whose owner (the authors show) underpaid all his players. Longley (1997) refutes the criticisms by including a dummy variable for Leafs players, but the results are mixed. Using data for more NHL seasons (which was available) could have provided a clearer result. 19  Only four of these players played for the Montreal Canadiens, Quebec’s only NHL team. 20  A salary cap was a standard feature of the NHL up to 1946. The salary cap for the 1925 season was set at $35,000 per team, increasing to $45,000 for the 1927 season. However, teams regularly exceeded the salary cap (with no penalty) due to competition for players. 21  Kahane (2005) uses a third method that utilizes frontier analysis. He shows that unusually high or low numbers of French-Canadian players reduces the technical efficiencies of NHL teams, suggesting that there exists an optimal number of French-Canadian players. 22  www.quanthockey.com/nhl/nationality-totals/ nhl-players-2015-16-stats.html 23  Rule 46 also specifies that players must cease fighting when they have been separated and are ordered to stop by an official. Failure to do so results in a major penalty (10 minutes) or expulsion from the game at the discretion of the official. Other penalties can be assessed for

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

removing a helmet before fighting, fighting off the ice surface, being the third player into a fight, fighting before play is started and wearing inappropriate equipment during a fight. 24  See S. Whyno (2013). NHL players bristle at fighting debate despite fan support for a ban. National Post, November 7. http://sports. nationalpost. com/2013/11/07/nhl-players-bristle-at-fightingdebate-despite-fan-support-for-ban/ 25  A small literature exists on NHL attendance models that do not include fighting or violence. Leadley and Zygmont (2006) found increased local attendance for five seasons after the opening of a new NHL arena. Coates and Humphreys (2012) tested the uncertainty of outcome hypothesis (UOH in the literature) for NHL teams and found an interesting asymmetric effect on attendance. Mills and Rosentraub (2014) found significant cross-border effects on attendance for the Buffalo Sabres and Toronto Maple Leafs.

REFERENCES Abrevaya, J. (2004). Fit to be tied: The incentive effects of overtime rules in professional hockey. Journal of Sports Economics, 5(3), 292–306. Allen, W. (2002). Crime, punishment and recidivism: Lessons from the National Hockey League. Journal of Sports Economics, 3, 39–60. Allen, W. (2005). Cultures of illegality in the National Hockey League. Southern Economic Journal, 71(3), 494–513. Audas, R., Goddard, J., & Rowe, W. (2006). Modelling employment durations of NHL head coaches: Turnover and post-succession performance. Managerial and Decision Economics, 27, 293–306. Bloom, M. (1999). The performance effects of pay dispersion on individuals and organizations. Academy of Management Journal, 42(1), 25–40. Boucher, M. (1984). Les Canadiens Français dans la Ligue Nationale de Hockey: une analyse statistique. L’Actualité économique, 60(3), 308–325. Bradbury, J. (2013). What is right with Scully estimates of a player’s marginal revenue product. Journal of Sports Economics, 14(1), 87–96. Bruggink, T., & Williams, D. (2009). Discrimination against Europeans in the National Hockey League: Are players getting their fair pay? American Economist, 54(2), 82–90. Buschemann, A., & Deutscher, C. (2011). Did the 2005 collective bargaining agreement really improve team efficiency in the NHL? International Journal of Sport Finance, 6(4), 298–306. Chan, T., Cho, J., & Novati, D. (2012). Quantifying the contribution of NHL player types to team performance. Interfaces, 42(2), 131–145.

319

Coates, D., & Humphreys, B. (2012). Game attendance and outcome uncertainty in the National Hockey League. Journal of Sports Economics, 13(4), 364–377. Curme, M., & Daugherty, G. (2004). Competition and pay for National Hockey League players born in Quebec. Journal of Sports Economics, 5(2), 186–205. Depken, C. (2000). Wage disparity and team productivity: Evidence from Major League Baseball. Economics Letters, 67(1), 87–92. Depken II, C., & Wilson, D. (2003). Wherein lies the benefit of the second referee in the NHL? Review of Industrial Organization, 24(1), 51–71. Driskill, R., & Vrooman, J. (2016). It’s not over till the fat lady sings: Game-theoretic analyses of sports leagues. Journal of Sports Economics, 17(4), 354–376. Easton, S., & Rockerbie, D. (2005). Overtime! Rules and incentives in the National Hockey League. Journal of Sports Economics, 6(2), 178–202. Fort, R., & Quirk, J. (1995). Cross-subsidization, incentives and outcomes in professional team sports leagues. Journal of Economic Literature, 33(3), 1265–1299. Frick, B., Prinz, J., & Winkelmann, K. (2003). Pay inequalities and team performance: Evidence from the North American major leagues. International Journal of Manpower, 24(3), 472–488. Goldberger, A. (1984). Reverse regression and salary discrimination. Journal of Human Resources, 19(3), 293–318. Goldschmied, N., & Espindola, S. (2013). Is professional hockey fighting calculated or impulsive? Sports Health: A Multidisciplinary Approach, 5(5), 458–462. Heckelman, J., & Yates, A. (2003). And a hockey game broke out: Crime and punishment in the NHL. Economic Inquiry, 41(4), 705–712. Idsen, T., & Kahane, L. (2000). Team effects on compensation: An application to salary determination in the National Hockey League. In A. Zimbalist (Ed.), The Economics of Sport, (Vol. 2). Cheltenham, U.K.: Elgar Publishing. Jasina, J., & Rothoff, K. (2016). The impact of the NHL lockout on county employment. International Journal of Sport Finance, 11(2), 114–123. Jones, J. (1984). Winners, losers and hosers: Demand and survival in the National Hockey League. Atlantic Economic Journal, 12(3), 54–63. Jones, J. (1988). Salary determination in the National Hockey League: The effects of skills, franchise characteristics and discrimination. Industrial and Labor Relations Review, 41(4), 592–604. Jones, J., Ferguson, D., & Stewart, K. (1993). Blood sports and cherry pie: Some economics of violence in the National Hockey League. American Journal of Economics and Sociology, 52(1), 63–78.

320

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Jones, J., Nadeau, S., & Walsh, W. (1999). Ethnicity, productivity and salary: Player compensation and discrimination in the National Hockey League. Applied Economics, 31(5), 593–608. Jones, J., Stewart, K., & Sunderman, R. (1996). From the arena into the streets: Hockey violence, economic incentives and public policy. American Journal of Economics and Sociology, 55(2), 231–243. Jones, J., & Walsh, W. (1987). The World Hockey Association and player exploitation in the National Hockey League. Quarterly Review of Economics and Business, 27(2), 87–101. Kahane, L. (2001). Team and player effects on NHL salaries: A hierarchical linear model approach. Applied Economics Letters, 8, 629–632. Kahane, L. (2005). Production efficiency and discriminatory hiring practices in the National Hockey League: A stochastic frontier approach. Review of Industrial Organization, 27, 47–71. Kahane, L. (2012). Salary dispersion and team production: Evidence from the National Hockey League. In L. Kahane & S. Shmanske (Eds.), The Economics of Sport (Vol. 1). Oxford: Oxford University Press. Késenne, S. (2015). Revenue sharing and absolute league quality, talent investment and talent allocation. Scottish Journal of Political Economy, 62(1), 51–58. Krashinsky, M., & Krashinsky, H. (1997). Do English Canadian hockey teams discriminate against French Canadian players? A comment. Canadian Public Policy, 23(2), 212–216. Krautmann, A.C. (1999). What’s wrong with Scullyestimates of a player’s marginal revenue product. Economic Inquiry, 37(2), 369–381. Krautmann, A.C. (2013). What is right with Scully estimates of a player’s marginal revenue product: Reply. Journal of Sports Economics, 14(1), 97–105. Lambrinos, J., & Ashman, T. (2007). Salary determination in the National Hockey League: Is arbitration efficient? Journal of Sports Economics, 8(2), 192–201. Lavoie, M. (2000). The location of pay discrimination in the National Hockey League. Journal of Sports Economics, 1(4), 401–11. Lavoie, M., & Coulombe, S. (1985). Discrimination à l’embauche et performance supérieure des francoquébécois dans la LNH: une mise au point. L’Actualité économique, 61(4), 527–530. Lavoie, M. & Grenier, G. (1992). Performance differentials in the National Hockey League: Discrimination versus style-of-play thesis. Canadian Public Policy, 18(4), 461–469. Lavoie, M., Grenier, G., & Coulombe, S. (1987). Discrimination and performance differentials in the National Hockey League. Canadian Public Policy, 13(4), 407–422. Leadley, J., & Zygmont, Z. (2006). When is the honeymoon over? National Hockey League

attendance, 1970–2003. Canadian Public Policy, 32(2), 213–232. Levitt, S. (2002). Testing the economic model of crime: The National Hockey League’s two-referee experiment. Contributions to Economic Analysis and Policy, 1(1), Article 2. Longley, N. (1995). Salary discrimination in the National Hockey League: The effects of team location. Canadian Public Policy, 21(4), 413–422. Longley, N. (1997). Do English Canadian hockey teams discriminate against French Canadian players? A reply. Canadian Public Policy, 23(2), 217–220. Longley, N. (2000). The underrepresentation of French Canadians on English Canadian NHL teams. Journal of Sports Economics, 1(3), 236–256. Longley, N. (2012). The economics of discrimination: Evidence from hockey. In L. Kahane & S. Shmanske (Eds.), The Economics of Sport (Vol. 2). Oxford: Oxford University Press. Longley, N., & Sankaran, S. (2007). The incentive effects of overtime rules in professional hockey: A comment and extension. Journal of Sports Economics, 8(4), 546–554. Lopez, M. (2015). Inefficiencies in the National Hockey League points system and the teams that take advantage. Journal of Sports Economics, 16(4), 410–424. McLean, R., & Veall, M. (1992). Performance and salary differentials in the National Hockey League. Canadian Public Policy, 18(4), 470–475. Mills, B., & Rosentraub, M. (2014). The National Hockey League and cross-border fandom: Fan substitution and international boundaries. Journal of Sports Economics, 15(5), 497–518. Mondello, M., & Maxcy, J. (2009). The impact of salary dispersion and performance bonuses in NFL organizations. Management Decision, 47(1), 110–123. Paul, R. (2003). Variations in NHL attendance: The impact of violence, scoring and regional rivalries. American Journal of Economics and Sociology, 62(2), 345–364. Poplawski, W., & O’Hara, M. (2012). The feasibility of potential NHL markets under the new collective bargaining agreement. Journal of Sports Economics, 15(1), 64–77. Richardson, D. (2000). Pay, performance and competitive balance in the National Hockey League. Eastern Economic Journal, 26(4), 393–417. Rockerbie, D. (2012). The demand for violence in hockey. In L. Kahane & S. Shmanske (Eds.), The Economics of Sport (Vol. 1). Oxford: Oxford University Press. Rockerbie, D. (2016). Fighting as a profit-maximizing strategy in the National Hockey League: More evidence. Applied Economics, 48(4), 292–299. Rockerbie, D., & Easton, S. (2017). Revenue sharing in professional sports leagues as a hedge for

ECONOMIC ISSUES OF THE NATIONAL HOCKEY LEAGUE: A SURVEY OF THE LITERATURE

exchange rate risk. In International Journal of Sport Finance. 12(4), 342–358. www.researchgate. net/profile/Duane_Rockerbie Ross, J. (2015). Joining the clubs: The business of the National Hockey League to 1945. Syracuse, N.Y.: Syracuse University Press. Scully, G. (1974). Pay and performance in Major League Baseball. American Economic Review, 64(6): 915–930. Shmanske, S., & Lowenthal, F. (2007). Overtime incentives in the National Hockey League: More evidence. Journal of Sports Economics, 8(4), 435–442. Staudohar, P. (2005). The hockey lockout of 2004–05. Monthly Labor Review, December, 23–29. Stewart, K., Ferguson, D., & Jones, J. (1992). On violence in professional team sport as the endogenous result of profit maximization. Atlantic Economic Journal, 20(4), 55–64.

321

Telser, L. (1995). The ultimatum game and the law of demand. Economic Journal, 105, 1519–1523. Vincent, C., & Eastman, B. (2009). Determinants of pay in the NHL: A quantile regression approach. Journal of Sports Economics, 10(3), 256–277. Vrooman, J. (2009). Theory of the perfect game: Competitive balance in monopoly sports leagues. Review of Industrial Organization, 34, 5–44. Walsh, W. (1992). The entry problem of Francophones in the National Hockey League: A systemic interpretation. Canadian Public Policy, 18(4), 443–460. Wilson, D.P. (2005). Additional law enforcement as a deterrent to criminal behavior: Empirical evidence from the National Hockey League. Journal of Socio-Economics, 34(2), 319–330. Winfree, J., & Fort, R. (2008). Fan substitution and the 2004–05 NHL lockout. Journal of Sports Economics, 9(4), 425–434.

32 The Economics of Australian Rules Football Ross Booth and Robert Brooks

INTRODUCTION: THE GAME, HISTORY, OWNERSHIP, OBJECTIVES AND GOVERNANCE Australian Rules football (or ‘AFL’ as it is now more commonly known) as a professional game is unique to Australia. Teams are comprised of 18 players (and four interchange) and play on an oval of preferably about 165 metres in length and 135 metres in width. Scoring consists of six points for a goal between the two goalposts and one point for a behind scored between the two goal posts and behind posts. Scoring is high, and scores with as many as 15 goals (with behinds totalling over 100 points) are not uncommon. The ball is oval-shaped and there is no off-side. Players can kick the ball, run and bounce the ball and mark (catch) the ball for a free possession. Players may also handpass the ball (punch with a clenched fist off the other open hand) but cannot throw the ball. Tackling is allowed below the shoulder and above the knee. Most of the sports economics literature on Australian Rules football focuses on the national league. It is the most popular football code in Australia with average home and away match attendance typically between 30,000 and 35,000. Formed in 1897, the Victorian Football League (VFL) expanded nationally to become the semi-­ professional Australian Football League (AFL)

and in 2018 comprises 18 predominantly memberowned, win-maximising clubs playing a unique brand of ‘Australian Rules’ football. The AFL is essentially owned by the football industry and is governed by a set of independent commissioners, who run not only the national league but are responsible for Australian football at all levels below the national league – state-based, suburban and country, senior and junior, male and female.

CONTRIBUTIONS AND FINDINGS This chapter discusses, in chronological order, the main theoretical contributions and research findings in the sports economics of Australian Rules football. We have restricted our survey to books and book chapters, and journal articles under a narrow definition of sports economics. Later in this chapter we cover some of the contributions from academics in other disciplines writing on sports economics. Perhaps the earliest contributions on the economics of Australian Rules football are by Dabscheck (1975a), which examines the VFL’s player transfer and zoning rules, and Dabscheck (1975b), which explains the wage determination process for VFL players. A pioneering work by

THE ECONOMICS OF AUSTRALIAN RULES FOOTBALL

Sandercock and Turner (1981) discusses the transformation of football in Victoria from amateur to semi-professional, including the media, the league, the clubs, the players, fans, and the future structure of the game. Another seminal work on the early business aspects of Australian football from the 1960s to early 1980s is that of Stewart (1983), which discusses the sources and distribution of revenue, player mobility regulations, player payments, and the increasing commercial and professional approach of the industry. After a prolonged period of financial difficulties and unevenness of competition, Stewart (1984) argues that the league has three courses of action available to it to try to ensure a viable competition. First, the league could redistribute playing talent from the stronger to the weaker clubs. Second, the league might minimise the growth in club costs and regulate player movements through the use of player payment controls and transfer and zoning regulations. Third, the VFL could redistribute league revenues in order to equalise the financial strength of clubs. Stewart (1984) argues that there is little evidence to show that the second approach adopted by the VFL of controls on player payments and transfer has been successful so far. Moreover, Stewart is sceptical that revenue sharing will be effective if non-football income is not shared, nor will a player draft work if clubs sell their players to overcome financial difficulties. Stewart is convinced that no single scheme will be able to ensure long-run league viability and concludes that some combination of revenue sharing and a player draft might be best able to ensure some semblance of parity. Borland (1987) conducted an econometric analysis of annual average attendance data of the demand for football in the VFL over the period 1950–1986. Borland found admission prices to have a significant negative effect and real income a positive effect on attendance. The author also found that variables for lagged attendance and uncertainty of outcome helped explain the demand for football attendance in the VFL. Dabscheck (1989) believes a player draft faces technical and legal problems, will be ineffective, and is unnecessary. He argues that the position in the draft selection order may be over-rated given the large number of players on a team, and suggests that if the league were serious, it should grant a club all its picks at once. Dabscheck argues that a player draft is not even necessary because a salary cap (set at some percentage of gross revenue) and a player list (roster) would be sufficient. Borland and Lye (1992) followed on the earlier work of Borland (1987) by looking at individual match attendance data for seasons 1981– 1986. The authors found that attendance is price

323

inelastic and an inferior good. Moreover, habitual attendance, stadium size, uncertainty of outcome and team success are also important. Dabscheck (1996) prefers the use of revenue-sharing devices rather than labour market controls such as a player draft or even a team salary cap. Borland and Lye (1996) examine matching effects as a determinant of mobility in the market for Australian Rules football coaches between 1931 and 1994. They find support for the existence of coach-team match-specific effects on team output, and that the probability of a coach-team separation decreases with the coach-team match effect on team performance and with years of coach tenure and coach experience. There is also some evidence that sorting of coaches by ability occurred between groups of coaches with only one spell and those with multiple spells. Fuller and Stewart (1996), in a study of attendance at VFL (1948–1994) and South Australian National Football League (SANFL, 1948–1993) home and away matches, found the main determinants of attendance to be the price of admission, the number of different days on which matches are played over a weekend, the construction of a bigger stand at the MCG (Melbourne Cricket Ground – used by up to four AFL clubs as their home ground over the period), and a competitiveness index. But spectators are creatures of habit, with these factors not having an immediate effect on attendance. Booth (1997) examines the player recruitment and transfer rules in the VFL/AFL, dividing 1897–1996 into five different periods: free agency (1987–1914); free agency and metropolitan zoning (1915–1929); free agency, metropolitan zoning and the Coulter Law (1930–1967); free agency, metropolitan and country zoning, player transfer fees, and various player payment schemes (1968–1984), and the salary cap and national player draft (1985–1996). Macdonald and Borland (2004) examine a number of Australian professional sports, including the AFL, under the headings of league history and structure, corporate governance, the labour market, the product market, attendance, television, and competitive balance. This chapter examines the levels of competitive balance in a league comprised of win-maximising clubs under a variety of labour market devices and revenue-sharing rules that make for interesting comparisons with competitive balance levels achieved in other professional sports leagues. In a series of articles, Booth (2004a, 2004b, 2005, 2006a, 2006b) examines the history and effectiveness of labour market devices and revenue-sharing rules the VFL/AFL has used to try to increase competitive balance. Booth (2004a) identifies six

324

THE SAGE HANDBOOK OF SPORTS ECONOMICS

different periods between 1897 and 2002 and then matches the different levels of within-season competitive balance against the devices and rules used in each period. It is suggested that the high levels of competitive balance achieved in the VFL/AFL in the most recent period result from the introduction of both a player draft and a team salary cap. The levels of competitive balance achieved in the VFL/ AFL are also compared with other major leagues in North America. Booth (2004b) extends the analysis of labour market intervention and revenue sharing in the VFL/AFL by incorporating two further years until 2004. The various devices used by the winmaximising clubs in the VFL/AFL are assessed in terms of their likely impact on competitive balance, with some significantly different theoretical predictions than under profit maximisation. Free agency results in a less equal distribution of player talent under win maximisation, cash sales of players and player trades will not always undermine the draft, and gate sharing, and league-revenue sharing tend to equalise playing strengths. The conclusion reached here is that a player draft, a team salary cap, and revenue sharing is the combination most likely to succeed in achieving higher levels of competitive balance. The evidence of competitive balance in the VFL/AFL is consistent with these predictions. Booth (2005) continues the investigation into whether the increase in competitive balance in the AFL post-draft and salary cap can be attributed to these labour market changes by comparing competitive balance outcomes and labour market changes in the AFL with two other Australian sports leagues, the National Basketball League (NBL) and the National Rugby League (NRL). The evidence suggests that since 1985 within-season competitive balance has increased slightly in all three leagues, and both pre- and post-1985 the NRL has been the most balanced and the NBL the least balanced. The distribution of championships/premierships is, in general, also more even in the post-1985 period in all three leagues. The most significant labour market change in both the NBL and the NRL post-1985 is their adoption of a team salary cap. Thus, the evidence on competitive balance is not inconsistent with the view that the introduction of a team salary cap (at least) in all three leagues has improved competitive balance since 1985. Booth (2006a) covers similar ground to Booth (2004a, 2004b, 2005) with the analysis ending in 2003. Discussion also includes the evolution of the game, the development of the league, club ownership and objectives, and league ownership, governance and objectives. Booth (2006b) details three main responses by AFL clubs to the season 2000 change from 50-50

gate-sharing, to the home team keeping the net gate. The first was a permanent move from small suburban grounds to larger stadia (Carlton and Collingwood) to cater for larger crowds and membership and sponsorship; second was a move of high attendance games to a larger venue (Sydney), and third was the move of low attendance games to interstate venues with the prospect of higher attendance, more membership and sponsorship, but lower venue costs. Stewart, Stavros and Mitchell (2007) use a similar approach to ‘Moneyball’ to identify potential inefficiencies in the recruitment of AFL players. Using match data, they use various types of regression models to identify and quantify the important player statistics in terms of their effect on match outcomes, and then use these results to decide which players to recruit. Stewart (2007) edits a book on the political economy of the four main football codes in Australia, including Australian football. Stewart introduces an Australian sport business model. Stewart and Dickson (2007) examine the development of Australian football leading to a national league – the AFL. Macdonald and Booth (2007), in a wide-ranging chapter, discuss a period of unprecedented reform, a time of significant competition restructuring and corporate reorganisation in all four codes – Australian football, rugby league, rugby union and soccer. The chapter addresses these reforms by surveying change in governance, regulatory, and competition structures, revenue, player salaries, attendance, participation, media exposure, competitive balance and national-team performance in each football code. The specific focus is the corporate era, comprising the 1990s and early years of the twenty-first century, when both crisis and deliberate strategy prompted reform of the governance structures and national competitions in each code. Borland, Chicu and Macdonald (2009) examine whether the player draft used in the Australian Football League (AFL) since 1986 has caused clubs to tank; that is, to seek to lose matches in order to obtain improved draft choices. The authors compare the change in performance between the pre-draft and draft eras of clubs that could have had an incentive to lose matches in the draft era, with the change in performance of clubs that would have had no difference in their incentive to lose matches between the pre-draft and draft eras. Analysing all AFL matches between 1968 and 2005, the authors find that clubs eliminated from the finals in the post-draft era are no more likely to lose matches in the last six rounds of the season than similarly performing clubs prior to the introduction of the draft. Moreover, clubs that were eligible for Special Assistance during the period 1997–2005 (where any team that won five

THE ECONOMICS OF AUSTRALIAN RULES FOOTBALL

matches or fewer would receive an extra priority choice in the player draft) were no more likely to lose matches in the last six rounds of the season than similarly performed clubs in other seasons. The main potential explanations for the absence of tanking in the AFL are the difficulty in identifying good talent, and the relatively small marginal effect of an extra or higher-rated drafted player on future club performance. Lenten (2009a) estimates a structural time-series model to investigate the relation between competitive balance, measured by the Actual Standard Deviation/Idealised Standard Deviation (ASD/ISD) ratio, and average match attendance in the Australian Football League from 1945 to 2005. He finds that increased competitive balance is associated with higher match attendances, at least on a season-toseason basis. Further, any shocks to competitive balance (from an unusually even or uneven season, for instance) tend to have quite persistent effects on attendance, which is consistent with the perception of AFL fans as habitual and loyal in nature. Lenten (2009b) proposes a new measure for competitive balance because within-season and between-season measures of competitive balance can produce differing results, thus making comparisons between leagues problematic. The new measure, based on each team’s win ratios from the current and previous seasons, overcomes one of the shortcomings of within-season measures, that is, the ability to pick up uncertainty of outcome from season to season, not just round to round. Unlike the ASD/ISD ratio for the AFL, according to this measure there appears to have been no large structural change in competitive balance in the AFL over the period 1898–2006. Lenten and Winchester (2010) consider a bonus point system for the AFL. For AFL data extending over seasons 1997–2008, the authors determine a bonus point system that is better at revealing strong teams than the current allocation of league points. For each match, the preferred league-scoring system awards six points: four points for a win, three points for a draw, two points for winning by 27 or more, and two points for losing by 26 or less. In addition to more accurately revealing strong teams, they suggest the inclusion of bonuses may increase spectator interest in matches where an obvious winner emerges prior to match completion. Lenten (2011) provides further evidence in the relationship between ASD/ISD ratios and attendances in the AFL, using the long-term trend components of the series. The sample period of 1945–2009 was a period in which there were profound changes in the demand for sport, driven by social and economic factors generally, but also by numerous league-specific factors, most notably competitive balance.

325

Stewart, Stavros and Mitchell (2011) examine whether the recruitment of Indigenous Australian footballers in the draft is based on rational expectations and market efficiency, by looking at the relationship between player performance and the expectations placed on these players when they were first drafted (using the player’s draft selection number) or traded (trade selection number). Results show that, on average, Indigenous Australian footballers outperform their expectations, which poses the question why AFL teams do not draft more Indigenous players. The authors suggest that this might arise because of the irrationality of recruiting staff, some form of statistical discrimination, or earlier discrimination at the junior level. Booth and Brooks (2011) examine violence in the AFL over the period 2000–2009, during which the AFL changed the tribunal system, most notably with the introduction of the Match Review Panel (MRP) in 2005, believing that the system would reduce the incidence of violence and injuries and help improve its public image as a sport for the young. The results suggest that the MRP has been effective in reducing the level of violent play and would appear to be related to three factors: changes in the rules of the game to reduce the likelihood of violent play; greater transparency in the penalty system; and the carry-over system that rewards previous good behaviour and sets higher penalties for repeat offenders. Interestingly, there does appear to be a greater attendance premium from aggressive play after the introduction of the MRP. Borland, Lee and Macdonald (2011) investigate escalation effects in the AFL. They use a sample of players selected in the AFL national player draft between 1986 and 2002, and test for escalation effects by examining whether a player’s draft order affects his subsequent utilisation by the club to which he was drafted. Limited evidence of an escalation effect is found. Any relation between a player’s draft order and his games played and tenure at the club to which he was drafted is concentrated in the early years (chiefly the second) of his career, and this apparent relation can be explained by the information about a player’s ability that is contained in the player’s draft order and to a lesser extent by incentives for clubs to provide greater playing experience to higher ability players. Escalation effects in the AFL competition are therefore much weaker than have been found in studies of the NBA. It is suggested that differences in the structure of the competitions may explain why the escalation effect in the AFL would be weaker than in the NBA – there are fewer players on an NBA team and in the AFL a greater disconnect between the drafting process and the coach. Dabscheck (2011) compares the share of revenue going to players in five Australian sports

326

THE SAGE HANDBOOK OF SPORTS ECONOMICS

(including the AFL) with a range of professional sports in Europe and the US and concludes that the percentage share going to Australian players (including the AFL) is much lower – in the 20s – than that of international players – above 50 and in some cases above 60. In the case of the AFL, the percentage share fell from the high 20s in 2001 to the low 20s in 2009. Booth, Brooks and Diamond (2012a) consider the AFL players share of revenue in the context of theoretical predictions of models of player salaries in both settings of profit-maximising and winmaximising clubs. The authors then show that the declining share of AFL players’ salaries as a proportion of revenue from 2001 to 2009 is consistent with the predictions from these theoretical models. This poses the question of what the league and the clubs do with the additional revenue if they are not paying it to the players. They suggest that the clubs’ strategy is to explore alternative talent investments (better coaching, improved facilities), while the league strategy is to spend on game development. Booth, Brooks and Diamond (2012b) provide a more detailed analysis of the finances of the AFL (club and league) which supports the general finding of Dabscheck (2011) that the players’ share of revenue in Australian sports is low by comparison with overseas sports. Macdonald (2012) comments on Dabscheck (2011), emphasising the centrality of competitive balance and fairness to the legal and economic dimensions of collective bargaining negotiations between the AFL and the AFLPA. Dabscheck (2012), in his response to both Booth, Brooks and Diamond (2012b) and to Macdonald (2012), is critical of both for ignoring the market structures in which professional sports in Australia, Europe and the US operate. Dabscheck argues that the major reason for the low player share of revenue in the AFL (and other Australian sports) is because of its ability to operate as a strong cartel. Lenten (2012) compares attendance and club membership (Hawthorn FC) as a measure of demand in the AFL. He finds that attendance is strongly correlated with win-ratio, as are club memberships, but with a one-year lag, since fans form expectations about the coming season based on the past season. Tuck and Whitten (2013) use a dynamic simulation model that captures the key components of a win-maximising sporting league (such as the AFL), and find reverse-order drafts can lead to some teams cycling between success and failure and to other teams being stuck in mid-ranking positions for extended periods of time. They also find an incentive for teams to tank exists, but it decreases as it becomes harder to determine player

quality, the fewer the number of teams, the smaller are list sizes, and as the number of teams tanking increases. Fujak and Frawley (2015) evaluate the impact of different strategies of the AFL and the NRL, by assessing the degree of exclusivity and geographic reach embedded in each broadcast agreement. In doing so, the research considered the impact of strategy in providing value to the broadcasters and teams, as well as utility to the fans of each league. It was found that the AFL has been the more successful in leveraging its broadcasters to achieve both financial and strategic outcomes. Fujak and Frawley (2016) analyse television ratings for the AFL and the NRL between 2007 and 2011 and conclude that there is significant variance in the coverage provided by the centralised broadcasting agreements and corresponding cumulative audience exposure of clubs within both leagues. This variance in television coverage between clubs may result in significant economic disparity – through sponsorship and the ability to engage fans. Notably, there was distinct favouritism shown towards those traditionally perceived as ‘powerhouse’ clubs. Dang, Booth, Brooks and Schnytzer (2015) use TV ratings to investigate the importance of match uncertainty in the TV viewing demand for AFL matches for the 2009–2011 home and away seasons. It is shown that greater television audience demand is strongly associated with higher match uncertainty, both expected (and actual) close contests. There is also evidence of substantially greater demand for matches shown on a weekday compared to a weekend. Tuck, Macdonald and Whitten (2015) build on the earlier work by Tuck and Whitten (2013) to show how dynamic non-equilibrium simulation represents a viable tool to aid league managers in the design of labour market regulations. Specifically, they simulate the expansion of a closed league, where a player draft is the primary player recruitment regulation, assessing the competitive balance implications of various concessional player draft selections to the expansion club. Stewart, Stavros, Mitchell, Barake and Phillips (2016) examine players on AFL player lists (2006– 2015) to quantify the effects on competitive balance of the father–son rule. The rule allows teams to prioritise the recruitment of the sons of former players in the player draft. The authors find that some teams have been able to access an on-field advantage via this rule, thereby compromising the AFL’s player draft, while also receiving significant marketing opportunities. More recently, the rule has been reformed, which has reduced this advantage. Frost, Lightbody, Carter and Halabi (2016) examine the business of ground sharing by cricket and

THE ECONOMICS OF AUSTRALIAN RULES FOOTBALL

Australian football at the Adelaide Oval prior to 1973. In 1973, the SANFL left the Adelaide Oval it had shared with cricket since 1877 and moved to its own stadium, ‘Football Park’, where it resided for 40 years, before a highly successful return to a redeveloped Adelaide Oval (shared with cricket) with new facilities and revenue streams, for the 2014 season. Using archival sources and financial reports, the authors construct counterfactual scenarios based on cricket and football continuing to share use of the Adelaide Oval after 1973 and projections of their outcomes. These are compared to the actual situation, based on the construction of Football Park. The authors suggest that the operation of separate grounds for football and cricket over this 40-year period was not economically efficient, because both cricket and football authorities placed higher values on non-monetary factors and had different objectives. Lenor, Lenten and McKenzie (2016) evaluate AFL scheduling policy (in which only some teams – key rivals – play each other twice) by estimating a model of match attendance with rivalry effects. Using attendance data from seasons 1997–2010, the authors find that rivalry effects are greatest for some large-market Melbourne teams, but there exists scope for increasing aggregate attendance if the AFL were to schedule other more well-supported matches a second time. They also observe some decline in interest of the second within-season meetings of the popular Melbourne teams but not in non-Victorian intrastate derbies. Wilson (2016) examines the relationship between crowd attendance and various measures of competitive balance in SANFL ‘home and away’ and finals matches, with some data sets covering a period as long as 1920 to 1983. The results are somewhat mixed, but the authors conclude that crowd attendance was not particularly sensitive to overall competitive balance. Wilson and Siegfried (2016) explore the issue of public funding of sports stadiums in Australia. Some benefit is captured by spectators as increased consumer surplus, yet the taxes which fund these facilities are mostly regressive. They find that those who buy tickets to sporting events have higher annual incomes and greater wealth than Australians who do not – the funding is not subsidising ‘working men’s’ recreation. The authors suggest that the funding of sporting arenas should be assessed on correcting market failures rather than on equity considerations. Frost, Borrowman and Halabi (2017) note that over the period 1920–1970, the VFL did not always use its largest and best equipped stadium for regular season games between its most popular teams, nor schedule those teams to play twice in a regular season. The authors calculate deadweight losses from the use of capital goods (stadiums)

327

and the effects of match scheduling. A counterfactual, based on the league scheduling a full set of games involving its five most popular teams each season (a scheduling effect) and using the MCG every weekend for these rivalry games (a venue effect), allows estimation of the dimensions of these inefficiencies in terms of lost revenue. Lenten (2017a) builds on the long-run structural Australian Football League (AFL) result obtained by Lenten (2009a) by using a structural time-series framework for AFL data between 1945 and 2010 to address the question of asymmetry in the sensitivity of demand (attendances) to relative team quality within-season in more- versus lessbalanced seasons. Strong evidence of asymmetry was identified, with attendance being more sensitive to variations in unexpectedly even seasons than uneven ones. However, there was no evidence to suggest asymmetry in levels, indicating that any decline in demand can be restored later without net loss of attendance revenue by a future reversal of the previous policy change. Lenten (2017b) investigated the possibility of racial discrimination by umpires on the basis of voting for the Brownlow Medal (a performancebased award), with individual-level data from 1998–2010 AFL matches. The results indicate that umpires are estimated to award significantly more votes to Indigenous footballers compared to their non-Indigenous contemporaries. The probability that they will award votes to Indigenous players is also significantly higher. Lenten argues that this result is quite a novel one when considered against the background of previous discrimination studies, including Stewart, Stavros and Mitchell (2011). Kendall and Lenten (2017), in a survey paper, look at specific examples where the rules of sports have led to unforeseen and/or unwanted consequences. The AFL is used as an example of a league where there have been a considerable number of rule changes, some designed to clean up and speed up the game, but others designed to overcome unintended consequences such as preventing the prevalence of the negative tactic of rushing behinds. One contentious rule which remains makes a ‘suspended’ player ineligible for the ‘fairest and best’ player in the AFL (the Brownlow medallist), and there have been instances of the leading vote-getter being ineligible. Lenten, Smith and Boys (2017) consider an alternative draft-pick order rule to the current AFL reverse order based on the end-of-season ladder, whereby the first pick is given to the team mathematically excluded first from making the finals (after fewest matches played), and then to other teams in order of elimination. Using data from home and away games between 1997 and 2009, the authors calculate estimates of improvement in

328

THE SAGE HANDBOOK OF SPORTS ECONOMICS

an already-eliminated team’s probability of winning late-season matches. Suggested benefits to the AFL would be an improvement in fans’ lateseason engagement, a fairer system of draft distributions that would increase competitive balance, while enhancing the integrity of the competition.

FUTURE HORIZONS FOR RESEARCH Having to suggest areas for future research is an onerous task. Aside from all the possible extensions of research outlined by the researchers discussed above, let us suggest some areas which might attract the attention of future academics and students interested in the sports economics of Australian football. One topic that is likely to receive considerable attention is women’s football as a professional competition. AFL Women’s (AFLW) is Australia’s national Australian Rules football league for female players. It is run by the AFL, which issued licences to eight AFL clubs for the first season, which began in February 2017. The rate of expansion has been a keen discussion point. The AFL has issued licences to two other clubs in 2019, and to a further four in 2020, and four clubs have yet to be awarded a licence. AFLW faces issues such as how to expand the season beyond seven rounds, when to fixture the season (currently in February–March in the AFL men’s pre-season competition time), where to play the fixtures, and what ticket prices to set (admission was free in the inaugural season). Apart from a few papers on the SANFL, there has not been a lot of sports economics research on aspects below the AFL – the national league. AFL Victoria imposes both a salary cap and a points system in some suburban and country leagues as a means of controlling player costs and increasing competitive balance. For example, a highly credentialed player, such as an ex-AFL player, attracts a lot of points, and so there is a limit on the number a club can employ (in addition to a team salary cap). Behavioural economics is very fashionable presently, and there must be a myriad of topics that will attract academic researchers. One obvious area of interest would be the trading strategies of clubs during the pre-draft trading period where there seems to be a flurry of completed trades towards the end of the trade period, no matter its length. The AFL clubs trialled variable ticket pricing in 2014 and dynamic ticket pricing in 2016, but neither has been well received by the football public. This is perhaps curious given the acceptance of both variable and dynamic ticket pricing in many sports in Europe and the US.

With 18 teams and a 22-round season, a longterm issue for the AFL has been an u­ neven fixture list – in the sense that not all teams play each other twice. This can result in some teams receiving ‘easy’ draws and some teams receiving ‘hard’ draws – notwithstanding that the AFL determines the fixture by giving the top six teams in the previous season the hardest draw and the bottom six teams the easiest draw. Many solutions have been proposed, including various conference systems. The problem of the fixture will be worse if the league conducts further expansion. Currently, two Melbourne-based clubs (Hawthorn and North Melbourne) play some of their home games in Tasmania, but the Tasmanian government is adamant the State could support an economically viable team. On the other hand, there is the issue of underperforming clubs. The focus at present is on the 2011 expansion club Gold Coast, which despite significant drafting and salary cap concessions has not been as successful as the 2012 expansion team Greater Western Sydney. Whether to expand or contract – that is the question! One of the issues creating concern are the current player drafting rules, particularly the limited free agency rule. There is a push from the players for free agency to be granted after a shorter qualifying period. This creates fears that the draft will be undermined as players seek to transfer to strong clubs in the prospect of enjoying more playing success. On the other hand, being able to move as a younger player might make some more attracted to the weaker teams as they will have more time to build up their playing stocks to achieve on-field success. Finally, one aspect of Australian Rules football which seems not to have had the attention it has in Europe and the US concerns the oft-maligned umpires and whether there is any bias of any kind in their decision-making, such as being influenced by the home team. In closing, it would be remiss of us not to make some additional observations, not on future horizons for research as such, but on research which has already been completed which might slip by unnoticed because it doesn’t neatly and obviously fit into the focus of this chapter. One of the frustrations of the academic literature is the common silo approach of experts in particular disciplines. One of the frustrations of the academic literature is the common silo approach of experts in particular disciplines. As authors, it is important to acknowledge that there have also been significant contributions by academics in other allied fields to the understanding of sports economics and Australian football, but which are not covered in this chapter. Moreover, a similar oversight can occur when we ignore not only contributions

THE ECONOMICS OF AUSTRALIAN RULES FOOTBALL

of a sports economic nature from other disciplinarians, but also of the multi-disciplinary field, which arguably are also necessary to consider in order to make sound policy decisions. In essence, what we are acknowledging is a significant contribution by academics in allied fields to the understanding of sports economics and Australian football, but which too often slips under the guard of we sports economists, and indeed in the coverage of this material in this chapter.

REFERENCES Booth, R. (1997). History of Player Recruitment, Transfer and Payment Rules in the Victorian and Australian Football League. Australian Society for Sports History Bulletin, 26, 13–33. Booth, R. (2004a). Labour Market Intervention, Revenue Sharing and Competitive Balance in the Australian Football League, 1897–2002. In R. Fort & J. Fizel (Eds.), International Sports Economics Comparisons (Chapter 19, pp. 319–336). Greenwich, CT: Praeger. Booth, R. (2004b). The Economics of Achieving Competitive Balance in the Australian Football League, 1897–2004. Economic Papers, 23(4), 325–344. Booth, R. (2005). Comparing Competitive Balance in Australian Sports Leagues: Does a Salary Cap and Player Draft Measure Up? Sport Management Review, 8(2), 119–143. Booth, R. (2006a). The Economic Development of the Australian Football League. In W. Andreff & S. Szymanski (Eds.), Handbook on the Economics of Sports (Chapter 58, pp. 552–564). Cheltenham, UK: Edward Elgar. Booth, R. (2006b). Some Economic Effects of Changes to Gate-Sharing Arrangements in the Australian Football League. In M. Nicholson, B. Stewart & R. Hess (Eds.), Football Fever: Moving the Goalposts (Chapter 9, pp. 115–132). Hawthorn, Victoria: Maribyrnong Press. Booth, R., & Brooks, R. (2011). Violence in the Australian Football League: Good or Bad? In R.T. Jewell (Ed.), Violence and Aggression in Sporting Contests: Economics, History, and Policy (Chapter 9, pp. 133–151). New York: Springer. Booth, R., Brooks, R., & Diamond, N. (2012a). Theory and Evidence on Player Salaries and Revenues in the AFL 2002–2009. Economic and Labour Relations Review, 23(2), 39–54. Booth, R., Brooks, R., & Diamond, N. (2012b). The Declining Player Share of AFL Clubs and League Revenue, 2001–2009: Where Has the Money Gone? Labour and Industry, 23(4), 433–445.

329

Borland, J. (1987). The Demand for Australian Rules Football. Economic Record, 63, 220–230. Borland, J., Chicu, M., & Macdonald, R.D. (2009). Do Teams Always Lose to Win? Performance Incentives and the Player Draft in the Australian Football League. Journal of Sports Economics, 10, 451–484. Borland, J., Lee, L., & Macdonald, R.D. (2011). Escalation Effects and the Player Draft in the AFL. Labour Economics, 18, 371–380. Borland, J., & Lye, J. (1992). Attendance at Australian Rules Football: A Panel Study. Applied Economics, 24, 1053–1058. Borland, J., & Lye, J. (1996). Matching and Mobility in the Market for Australian Rules Football Coaches. Industrial and Labor Relations Review, 50, 145–160. Dabscheck, B. (1975a). Sporting Equality: Labour Market versus Product Market Control. Journal of Industrial Relations, 17(2), 174–190. Dabscheck, B. (1975b). The Wage Determination Process for Sportsmen. Economic Record, 51(133), 52–65. Dabscheck, B. (1989). Abolishing Transfer Fees: The Victorian Football League’s New Employment Rules. Sporting Traditions, 6(1), 63–87. Dabscheck, B. (1996). Assault on Soccer’s Compensation System: Europe and Australia Compared. Sporting Traditions, 13(1), 81–107. Dabscheck, B. (2011). Player Shares of Revenue in Australia and Overseas Professional Team Sports. Labour and Industry, 22(1–2), 57–82. Dabscheck, B. (2012). Player Shares of Revenue in Australia and Overseas Professional Team Sports: A Response. Labour and Industry, 22(4), 465–475. Dang, T., Booth, R., Brooks, R., & Schnytzer, A. (2015). Do TV Viewers Value Uncertainty of Outcome? Evidence from the Australian Football League. Economic Record, 91(295), 523–535. Frost, L., Borrowman, L., & Halabi, A.K. (2017). Stadiums and Scheduling: Measuring Deadweight Losses in the Victorian Football League, 1920–70. Australian Economic History Review, online first 26 July. doi: 10.1111/aehr.12132 Frost, L., Lightbody, M., Carter, A., & Halabi, A.K. (2016). A Cricket Ground or a Football Stadium? The Business of Ground Sharing at the Adelaide Oval before 1973. Business History, 58(8), 1164–1182. Fujak, H., & Frawley, S. (2015). Evaluating Broadcast Strategy: The Case of Australian Football. International Journal of Sport Communication, 8, 431–451. Fujak, H., & Frawley, S. (2016). Broadcast Inequality in Australian Football. Communication and Sport, 4(2), 187–211. Fuller, P.J., & Stewart, M.F. (1996). Attendance Patterns at Victorian and South Australian Football Games. Economic Papers, 15(1), 83–93.

330

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Kendall, G., & Lenten, L.J.A. (2017). When Sports Rules Go Awry. European Journal of Operational Research, 257(2), 377–394. Lenor, S., Lenten, L.J.A., & McKenzie, J. (2016). Rivalry Effects and Unbalanced Schedule Optimisation in the Australian Football League. Review of Industrial Organization, 49(1), 43–69. Lenten, L.J.A. (2009a). Unobserved Components in Competitive Balance and Match Attendances in the Australian Football League, 1945–2005: Where is all the Action Happening? Economic Record, 85(269), 181–196. Lenten, L.J.A. (2009b). Towards a New Dynamic Measure of Competitive Balance: A Study Applied to Australia’s Two Major Professional ‘Football’ Leagues. Economic Analysis and Policy, 39(3), 407–428. Lenten, L.J.A. (2011). Long-run Trends and Factors in Attendance Patterns in Sport: Australian Football League, 1945–2009. In S. Cameron (Ed.), Handbook on the Economics of Leisure (pp. 360– 380). Cheltenham, UK: Edward Elgar. Lenten, L.J.A. (2012). Comparing Attendances and Memberships in the Australian Football League: The Case of Hawthorn. Economic and Labour Relations Review, 23(2), 23–38. Lenten, L.J.A. (2017a). A Formal Test for Asymmetry in the Uncertainty of Outcome Hypothesis. Journal of Sports Economics, 18(3), 253–270. Lenten, L.J.A. (2017b). Racial Discrimination in Umpire Voting: An (Arguably) Unexpected Result. Applied Economics, 49(37), 3751–3757. Lenten, L.J.A., Smith, A., & Boys, N. (2017). Evaluating an Alternative Draft Pick Allocation Policy to Reduce ‘Tanking’ in the Australian Football League. European Journal of Operational Research, 267(1), 315–320. Online first, 16 May. Lenten, L.J.A., & Winchester, N. (2010). Optimal Bonus Points in the Australian Football League. Economic Papers, 29(4), 407–420. Macdonald, R. (2012). ‘Player Shares of Revenue’, Collective Bargaining and the Australian Football League: A Comment on Dabscheck. Labour and Industry, 22(4), 447–464. Macdonald, R., & Booth, R. (2007). Around the Grounds: A Comparative Analysis of Football in Australia. In B. Stewart (Ed.), The Games Are Not the Same: The Political Economy of Football in Australia (Chapter 8, pp. 236–331). Carlton, Victoria: Melbourne University Press. Macdonald, R., & Borland, J. (2004). Professional Sports Competitions in Australia. In R. Fort & J. Fizel (Eds.), International Sports Economics Comparisons (Chapter 18, pp. 295–317). Greenwich, CT: Praeger.

Sandercock, L., & Turner, I. (1981). Up Where Cazaly? The Great Australian Game. London: Granada. Stewart, B. (1983). The Australian Football Business: A Spectator’s Guide to the VFL. Kenthurst, NSW: Kangaroo Press. Stewart, B. (1984). The Economic Development of the Victorian Football League. Sporting Traditions, 1(2), 2–26. Stewart, B. (Ed.) (2007). The Games Are Not the Same: The Political Economy of Football in Australia. Carlton, Victoria: Melbourne University Press. Stewart, B., & Dickson, G. (2007). Crossing the Barassi Line: The Rise and Rise of Australian Football. In B. Stewart (Ed.), The Games Are Not the Same: The Political Economy of Football in Australia (Chapter 6, pp. 71–113). Carlton, Victoria: Melbourne University Press. Stewart, M.F., Stavros, C., & Mitchell, H. (2007). ‘Moneyball’ Applied: Econometrics and the Identification and Recruitment of Elite Australian Footballers. International Journal of Sports Finance, 2(4), 231–248. Stewart, M.F., Stavros, C., & Mitchell, H. (2011). Does the AFL Draft Under-Value Indigenous Australian Footballers? Journal of Sports Economics, 12(1), 36–54. Stewart, M.F., Stavros, C., Mitchell, H., Barake, A.J., & Phillips, P. (2016). Like Father, Like Son: Analyzing Australian Football’s Unique Recruitment Process. Journal of Sport Management, 30(6), 672–688. Tuck, G.N., & Whitten, A.R. (2013). Lead Us Not into Tanktation: A Simulation Modelling Approach to Gain Insights into Incentives for Sporting Teams to Tank. PLoS ONE, 8(11), e80798. doi:10.1371/ journal.pone.0080798 Tuck, G.N., Macdonald, R., & Whitten, A.R. (2015). Management Reference Points for Sporting Leagues: Simulating League Expansion and the Effect of Alternative Player Drafting Regulations. In W. Andreff (Ed.), Disequilibrium Sports Economics: Competitive Imbalance and Budget Constraints (Chapter 3, pp. 50–103). Cheltenham, UK: Edward Elgar. Wilson, J.K. (2016). The Relationship between Crowd Attendance and Competitive Balance: Evidence from the SANFL – 1920–1983. In R. Pomfret & J.K. Wilson (Eds.), Sport Through the Lens of Economic History (Chapter 5, pp. 71–88). Cheltenham, UK: Edward Elgar. Wilson, J.K., & Siegfried, J. (2016). Who Sits in Australia’s Grandstands? Journal of Sports Economics, online first, 5 July. https://doi.org/ 10.1177/1527002516656728

33 The Economics of Major League Soccer from the NASL to MLS: A Brief History of North American Professional Soccer Nicholas Watanabe

To begin with, in order to understand the economics, organizational structure, and strategies of Major League Soccer (MLS), it is necessary to consider the development of professional soccer in North America over the last 50 years. The first professional league, the North American Soccer League (NASL) began play in 1968 with 17 teams, but many of the teams experienced financial troubles, causing 12 teams to withdraw before the 1969 season began. However, once the league managed to overcome this initial crisis, it quickly expanded to 15 teams in 1974, and 20 in 1975. Furthermore, in 1975 the New York Cosmos signed Brazilian superstar Pele, whose debut drew a national television audience of 10 million (Ludtke, 1976). This event changed the landscape of the NASL, and its teams embarked on a new strategy of expanding the league (which reached 24 teams in 1978), as well as signing aging stars from Europe and South America. However, as teams signed players with high wages, costs began to rise faster than revenues, and many franchises experienced financial trouble. Though the league worked with players to implement a salary cap in 1982, this move to cut costs came too late and by the end of the 1984 season the NASL shut down operations. Although the closure of the NASL was a major setback for professional soccer in North America, the lessons learned from

the operational successes and failures of the league would play an important role in the development of the MLS. In the aftermath of the death of the NASL, the United States Soccer Federation (USSF), which governs all soccer in the United States, began to work on a bid to host the 1996 FIFA World Cup in the United States. As part of this bid, the USSF promised it would re-establish a top-tier professional soccer league in the US, which led to the birth of a 10-team MLS in 1996 (Jewell & Molina, 2005). The most unique feature of the MLS, and the thing which distinguishes it from most other major professional sport leagues, is that it operated as a single-entity organization (Mathias, 1999). MLS was structured so that all the teams and players in the league are owned by a single corporation, giving the league ultimate control over which players will be signed by the league, player salaries, as well as providing the league some protection from antitrust cases (Mathias, 1999). Under the single-entity structure, MLS teams are not technically owned by any one individual (or group), but rather have investors who oversee the day-to-day operations of teams similar to owners in other professional sport leagues (Watanabe, 2006). The use of the single-entity model by MLS provided a number of benefits that helped the

332

THE SAGE HANDBOOK OF SPORTS ECONOMICS

league avoid many of the financial problems that plagued the NASL. The lack of financial control on player salaries in the NASL created competitive imbalance on the field, and also drove some teams into bankruptcy (Twomey & Monks, 2011). Considering this background, MLS placed a highly restrictive salary cap on every franchise in the league, with teams only allowed to spend $3.845 million in 2017 to fill 20 roster spots. Furthermore, as the league signs and allocates all players, it helps control not only how talent is distributed, but also to manage the pay scale for most players. Additionally, the single-entity model was beneficial in the early days of MLS, as there were not many investors who were interested in financing teams in a startup league. Thus, two main investors, Phil Anschutz and Lamar Hunt, ended up being the investor/owner for nine of the 12 franchises in the league from 1998 to 2001. Though this high concentration of ownership would be seen as a conflict of interest in most professional sport leagues, especially having two teams from a single owner competing against one another, the single-entity structure allowed these two investors to own multiple teams. As MLS continued to grow over the last two decades, the league began to encourage Anschutz and Hunt to begin divesting from the teams they were invested in, moving to a distributed club ownership model. This goal of divesting became a reality in 2015 when Anschutz Entertainment Group sold its remaining 50% stake in the Houston Dynamo, making it so all the teams in the league had their own unique investors and operators. As of 2017, Major League Soccer is 22 clubs spread across both the United States and Canada, with plans to expand to 26 clubs by 2020, and a target number of 28 clubs at a later date. Though the MLS is smaller in terms of the number of clubs when compared to other major North American professional sport leagues (which often have 30 or more franchises), the league does have more clubs than most top leagues in Europe, who usually range between 16 and 20 clubs in their top tier. Additionally, the growth of the MLS has not just been in terms of the numbers of clubs and the geographic reach of its fan base, but also in fan interest. In the early 2000s the league was averaging around 13,000 to 15,000 in average attendance at matches, but now boasts an average of over 20,000 attendees per match. Over the same time period the value of MLS franchises has also grown, with expansion teams such as Seattle having paid $30 million to enter the league in 2007 (Wilson, 2008), now being estimated to be worth around $285 million (Smith, 2016).

DESIGNATED PLAYER RULE After the first decade of operations, MLS began to explore ways in which they might be able to bring more high-profile players into the league to help draw more fan interest as part of an overall plan to continue expanding the league. However, the salary caps imposed on all teams by the league made this rather difficult, as the total amount investors were allowed to spend on salaries for a roster of around 20 players was less than $2 million in the mid-2000s (Twomey & Monks, 2011). In 2007, when the league was presented with the opportunity to sign David Beckham, perhaps one of the most famous English soccer stars of the modern era, it was clear that the league’s salary cap rules would not allow them to sign him. Thus, the league created what is known as the ‘Designated Player Rule,’ as it allowed them to bring Beckham and other stars to the league. The Designated Player Rule itself has evolved over time. In its first year, the rule allowed teams to have one Designated Player spot, which they could use or trade to another team, with all teams being allowed a maximum of two Designated Players. Additionally, any Designated Player would have the first $400,000 of their salary be part of the team salary cap of $2.1 million (in 2007), and then any remaining salary would not count towards the limit. Because of this, the Los Angeles Galaxy were able to sign David Beckham for a five-year contract that had a yearly salary of $6.5 million, with the majority of Beckham’s wages not counting against the league imposed cap. Following the creation of the original Designated Player Rule, MLS began to bring in a number of stars from other leagues and decided that there was a need to expand the rule. In 2010, the rule was updated so all teams were allowed to have two Designated Player spots on their roster, with only $335,000 of those players’ salaries charged to the salary cap. Additionally, teams were allowed to have a maximum of three Designated Players on their roster, of which they would have to pay a luxury tax of $250,000 for the third player. This tax on the third player allowed MLS to create what was known as ‘allocation money,’ or funds specifically earmarked for signing and paying top-level players. In 2015, the Designated Player Rule was altered to update the allocation money, with a new mechanism called ‘Targeted Allocation Money’ (TAM). The most recent change provided each team with $100,000 a year for the next five years that was specifically to be used in signing a high-profile player (Sung & Mills, 2018). All teams receive this money but are required to use the full $500,000 value within the next five years,

THE ECONOMICS OF MAJOR LEAGUE SOCCER FROM THE NASL TO MLS

trade it to other teams, or else they have to return the money to the league. Notably, the TAM funds were specifically designed to buyout the contracts of lower-level Designated Players (ones whose salary is close to that of the DP minimum), and thus be able to replace these players with individuals who are considered to be international superstars (Sung & Mills, 2018).

REVIEW OF THE LITERATURE To date, there exists limited research examining the economics of MLS. The sparsity of theoretical and empirical examinations of MLS can be partly attributed to the fact that the league has only completed 20 seasons of play, and thus there is a relative lack of data which can be produced in such a time period. Additionally, as the league operates under a single-entity model, it often makes it difficult to fully understand the complexities that exist in trying to sign players. That said, there does exist a handful of studies which have approached various aspects of MLS from an economic perspective. Specifically, the literature can mainly be divided into two main lineages, the first examining the wages of players and its relationship to factors such as team performance, and the second focusing on consumer demand for attending MLS matches. One thing that must be noted in reviewing the literature is that the majority of studies came about after the creation of the Designated Player Rule in 2007. The published research used this policy change as part of the motivation for conducting such studies. Whether these studies are focused on wage inequality, attendance, or team performance, they all theorize the value that superstar-level talent has for the league, and ways in which organizations may be able to maximize gains from these individuals. In the following, this chapter will first review the literature for the player salary literature, and then cover attendance demand research. In conclusion, the chapter will consider gaps that exist within the overall body of research, and how these may serve as an impetus for future sports economics studies of MLS.

PLAYER SALARIES, TEAM PERFORMANCE, AND THE SINGLE-ENTITY MODEL One of the earliest discussions regarding the economics of MLS was a paper by Mathias (1999) in

333

the University of Pennsylvania Law Review. His work analyzed both the antitrust and labor issues surrounding the single-entity structure of the league. Specifically, focus was placed on the case Fraser v. MLS, which was seen not only as a challenge by the players against the salary structure and policies of MLS but also as a ‘rite of passage’ (Mathias, 1999, p. 203) as the top North American professional sport leagues all have litigated antitrust court cases between the owners and players. However, what is most intriguing from an economic perspective about the work of Mathias (1999) is the detailed evaluation from a legal context of how the single-entity structure of MLS holds up against the Sherman Antitrust Act. Considering this structure, Mathias (1999) argued that through organizing the league as a single corporation in which there are a number of outposts (i.e., teams) run by investors, it allows the league to avoid section 1 lawsuits which are commonly raised against professional sport leagues. This corporate structure then creates the situation where if any groups do bring up an antitrust suit under the Sherman Act against MLS, it would have to do so under section 2. This change creates numerous advantages for any league that operates as a single-entity, including the ability to avoid the triple damages that can be awarded under section 1 cases, as well as the difficulty in being able to prove monopolistic behaviors per section 2 of the Sherman Act (Mathias, 1999). In continuing this discussion, it is argued that by organizing as a single-entity corporation and making it difficult for players to sue the league for antitrust violations in regards to contracts and salary, it provides MLS with a stronger bargaining position than most other professional sport leagues. Furthermore, Mathias (1999) also theorizes that by structuring the league in this manner, it also provides the league with greater control over the operations of individual clubs, which can prevent a range of behaviors, from violations of the existing salary structures to preventing teams from relocation or breaking away without consent from the league as a whole. From this, Mathias (1999) notes that this lawsuit and the formation of the MLS could have important implications for the economics and operational methods of all professional sport leagues in North America, including how reforming as single-entity organizations could help other leagues strengthen their legal standing and bargaining power. The first published economic examination of Major League Soccer likewise examined technical efficiency within the league (Haas, 2003). Specifically, the research by Haas (2003) provides the first empirical examination of how the wages

334

THE SAGE HANDBOOK OF SPORTS ECONOMICS

of players and coaches are related to organizational performance in the league. In order to conduct such an examination, Haas employs a Data Envelopment Analysis (DEA) using input and output measures in order to calculate whether teams are efficient in regards to their overall organizational performance. From this, the final results of the paper find that when using player and coaching salaries as inputs, and attendance and revenues as outputs, that MLS was remarkably efficient in the year 2000, with some teams achieving perfect or near-perfect efficiency. Overall, Haas’s (2003) research is important within the sports economics literature as being the first research to attempt to analyze the relationship between salaries and organizational performance in the league. However, this early study of the MLS does have a number of issues with regard to the data and the variables used to estimate and control for technical efficiency. First, because of the single-entity nature of MLS, the data for this study were collected partly through information published online, as well as through interviews with an individual who filled in the missing data on player and coaching salaries. Because player and coaching salaries were used as the two input variables for this research, it is possible that any misreporting by the individual who was interviewed could have resulted in biased or unreliable estimates of efficiency. Furthermore, the author notes the inability to control for the population size and other market-specific factors, which could have caused bias in the results by omitting these variables. Examining the results in more detail, Haas argued that four teams – the Chicago Fire, Tampa Bay Mutiny, Los Angeles Galaxy, and Kansas City Wizards (now known as Sporting Kansas City) – are ‘globally efficient’ (Haas, 2003, p. 209). The argument presented is that all of these teams were highly successful in all measurements of performance in relationship to salary. However, if one considers the off-field performance measures for these teams, the Wizards and Mutiny were near the bottom with regard to attendance and organizational revenues (only the Miami Fusion was worse in these two categories). Moreover, Haas (2003) argued that the single-entity structure of MLS may be what is helping to make the teams in the league remarkably efficient, especially as they are able to restrict salaries and player movements with greater ease than other leagues. However, Haas did not consider that his analysis of a startup league ignores the economic reality that many of the teams may have been efficient on and off the field, but were also in unstable financial situations. Notably, both Tampa Bay and Miami would be contracted from the league following the 2001

season due to having the worst attendance and revenue generation in the league. Considering this fact alongside the results from Haas’s research, it would seem to be the case that Major League Soccer was not concerned about efficiency, but rather on the profitability and overall attendance and revenue generation by clubs. Following the work of Haas (2003), it would be many years before another academic study continued the line of research examining player salaries in MLS. The main impetus for this came about because of two things. First, the Fraser v. MLS legal case discussed by Mathias (1999) ended up making it to the Court of Appeals in 2002, where it was ruled that there was no proof that MLS has restricted competition in the labor market for players (Stebbins, 2015). After losing this ruling, the players’ union entered into negotiations with the league, and managed to agree on the first Collective Bargaining Agreement in 2004. Following this, the players’ union began to publicly post the salaries of all players in the league, starting in 2005. The second change which came about was the creation of the aforementioned Designated Player Rule, which also created greater dispersion of salaries across teams and the league as a whole. Thus, because of the existence of an open data set and structural changes in how the league was able to pay players, researchers began to conduct more detailed analysis of the effect the Designated Player Rule had on salaries and team performance in MLS. The first paper to specifically examine the salaries using the new data provided by the league was conducted by Kuethe and Motamed (2010), who analyzed the effects that being a superstar player had on salaries in MLS. Similar to Haas (2003), Kuethe and Motamed (2010) only examined a single year of salary data. However, because of the new and complete data set, the later study was able to analyze the factors which helped to determine the salaries of 193 players who played in 2007. In this paper, the authors built one of the first comprehensive models to analyze player compensation MLS, including various factors such as age, experience, position, region the player was from, as well as if individuals were members of their national team or were Designated Players. Results from Ordinary Least Squares (OLS) and Quantile regressions find a superstar effect, where having Designated Player or All-star status within the league are significant determinants of salary. One of the more interesting findings of these models is that the relationship between age and salary is convex, which is different than most other professional sport leagues. From this, it is argued that MLS may be rewarding two types of players: young stars who have high potential to become

THE ECONOMICS OF MAJOR LEAGUE SOCCER FROM THE NASL TO MLS

great players, as well as older stars who have come to the MLS to finish out their career while still making a high wage. Where Kuethe and Motamed (2010) focused on determining individual player salaries, follow-up research by Twomey and Monks (2011) attempted to examine the link between monopsony, singleentity status, and salary suppression in MLS. Similar to Kuethe and Motamed, Twomey and Monks utilize the 2007 salaries of all players in MLS that were released by the players’ union. However, rather than look at determinants of wages, the authors consider how the structure and controls put into place by MLS may be allowing the league to suppress salaries and extract more rents from players than other leagues. Utilizing revenue data estimates released by Forbes, Twomey and Monks calculate that only 24.86% of league revenues are used to pay player salaries. When compared to other leagues, such as MLB, which pays 52%, or the NFL, which pays 59.5%, the authors argue that the single-entity nature of the MLS has allowed them to keep salaries low and allowed the investors in MLS to gain more profit per dollar of revenue than other leagues. Furthermore, Twomey and Monks argued that while the single-entity structure of the MLS has helped to keep the costs in check for the league, this type of minimal investment in human capital may also lead to a low growth rate for the league in its future. That is, they theorized that by spending such low amounts of money on player salaries, it will be the case that the American players who have the highest valuation and level of talent will be more likely to move to leagues in Europe where they are more likely to be paid better. While this theorization about the investment in player salaries in MLS is interesting, the timing of the study capturing only the first year of the Designated Player Rule likely influences this argument. On the one hand, if one were to examine the years before the existence of the Designated Player Rule, it is likely that there would be even higher salary suppression within MLS, as there were no loopholes to pay players outside of the imposed salary cap. On the other hand, further analysis of additional years of data after the Designated Player Rule was put into place would have seen the pay scale for the league drastically increase as the number of Designated Players increased due to more teams signing them, as well as rule changes to allow more Designated Player slots. In considering the overall body of literature examining player salaries in MLS, one of the main issues which exists is that the research has mostly utilized a single year of data in relatively small samples to conduct their analysis. In this manner, the research focused on player salary often

335

is only able to capture a snapshot of salaries and their relationship to other behaviors in the league. While much of this can be attributed to a lack of data in the early years of the league, the fact that there has been reliable information regarding player wages since 2005 means that the research has often not reached its full potential by examining multiple years. In response to this, the most recent study examining MLS player salaries by Coates, Frick, and Jewell (2016) provide the first multi-year analysis of player salary, by examining the data from 2005 through 2012 to determine whether team wages are important in determining on-field performance. Additionally, the Coates et  al. (2016) study advances the MLS literature by considering the total wages and dispersion of salaries within a team, through calculating the Gini Coefficient of wages for each team. Thus, this work makes major contributions to the literature not only in being the first to examine salaries in the league over a one-year time period, but by also considering how the Designated Player Rule could impact team performance. Notably, the management literature has long discussed how when wages in an organization are dispersed unfairly, it can lead to positive or negative performances by the group as a whole (Shaw, 2014). However, MLS provides an interesting backdrop through which to examine whether having high wage dispersion affects team performance, as the creation of the Designated Player Rule created the situation where a few players are often paid salaries that are higher than the wages for the rest of their team. Through using a Log-Log model where the dependent variable is the production of points accumulated through wins or ties in matches, the researchers find consistently that having higher dispersion of salaries (greater inequality) caused teams to earn significantly less points in a match (Coates et al., 2016). At the same time, the overall total salary for a team was positive and significant, indicating that teams that spent more on talent tended to perform better. Interestingly, in examining the coefficients from their results, Coates et  al. (2016) found a negative impact of salary dispersion and the positive impact of total wages for a team are of comparable magnitudes. Thus, it presents a conundrum for MLS teams in terms of how to maximize their performance while balancing the composition of their teams. On the one hand, if a franchise wants to increase the number of wins, the findings indicate that spending higher amounts on talent helps to improve on-field performance. Due to the salary cap structure put into place by the league, the only way to significantly increase the total wages for a team is through signing Designated Players who are able to be paid

336

THE SAGE HANDBOOK OF SPORTS ECONOMICS

outside of the salary cap rules. However, in doing so, this then increases the salary inequality within a team, and thus also lowers win production. Thus, all MLS teams are faced with the tradeoff of trying to spend more to increase team performance, but at the same time potentially decreasing their performance because of the wage gap between players.

ATTENDANCE DEMAND AND MAJOR LEAGUE SOCCER The second line of research which analyzed MLS focused on attendance demand for matches. Similar to the salary and team performance literature, the studies conducted primarily look at the attendance impact that has come about from adding star players and Designated Players to team rosters. This trend is one that would seem to be rather as expected, as the MLS presented a unique case where economists were able to look at the superstar effect as it has grown across the league. The initial study of MLS attendance by Jewell and Molina (2005) focused on the growth of the league in its early years. In this, the authors placed focus on examining the relationship between average team attendance and the size of Hispanic populations in regions with MLS franchises. In analyzing the first six years of the league, the authors suggest that having a higher population with roots from Latin American countries should provide a strong fan base for teams. The reason for this is that while the MLS and soccer in general were relatively new in the US at this time, Latin American countries have had a long tradition and history of supporting professional soccer leagues and clubs. Thus, the belief was that cities with higher Hispanic populations would be able to draw more attendees to matches. Following a traditional demand function similar to that laid out by Bird (1982) in his examination of professional soccer attendance in Europe, MLS attendance is modeled as being a function of the strength of teams, number of star players, and market characteristics. Included in this model are specific controls for the number of Hispanic and Black individuals in each Metropolitan Statistical Area (MSA). Curiously, the estimated results find that MLS franchises that played in regions with higher numbers of Hispanic individuals actually had significantly lower attendance. While the result does not match up with the initial expectations, the authors do provide explanations as to why cities with higher Hispanic populations had

lower live demand. First, in looking at markets with MLS franchises, it was found that matches were often played live on Spanish language television stations. Because of this, it is possible that many Hispanic fans were choosing to forgo attending in person, and instead would watch the matches on television in their native language. Furthermore, in examining the television schedules, it is the case that most MLS matches were traditionally played on weekend afternoons, which is also when Mexican League (and other Latin/ Central American league) games are aired. Thus, it is possible that Hispanic populations were not attending MLS matches because they were able to follow and watch the teams that they had traditionally supported on television. In this manner, it is possible that televised games of other leagues were serving as substitutes to live attendance at MLS matches. Following this, the next study to examine MLS demand focused on the specific context of the signing of 14-year-old soccer sensation Freddy Adu, and whether the league would be able to recoup the $500,000 a year salary they were paying him (DeSchriver, 2007). In order to conduct such an analysis, DeSchriver (2007) developed an attendance model following the prior work of Jewell and Molina (2005), to examine Adu’s impact on match-level attendance in his first full season in the league (2004). Using a fixed-effects regression estimated from 150 observations of match attendance, the results indicate that when Freddy Adu played on the road, the increase in attendance was estimated to be around 11,000. Conducting back-of-the-envelope calculations, DeSchriver (2007) estimated that in away matches alone, Adu produced around $155,000 of extra revenue per match, or about $2.3 million dollars in total. Thus, considering the amount of revenue Adu likely helped to produce in his first year in the league, it is the case that the investment in the teenage star was financially beneficial to the league. However, DeSchriver noted that while Adu was popular, he had limited production on-field. A player with the combination of marketability and a high-level of play on the field could potentially bring even bigger gains. Following the analysis of Adu, it is only natural that the next examination of the impact of a star player on attendance focused on the very first Designated Player, David Beckham. In this, Lawson, Sheehan, and Stephenson (2008) employed a similar approach to DeSchriver (2007) to try to estimate the effect that Beckham’s presence had on attendance in his first season. It is important to note a number of differences between DeSchriver (2007) and Lawson et  al.’s (2008) approach in estimating the effects of star players

THE ECONOMICS OF MAJOR LEAGUE SOCCER FROM THE NASL TO MLS

on attendance. First, where the previous paper focused on the effects that Adu had on road games, the latter study examined attendance across all matches in MLS. Furthermore, rather than using the number of attendees as a dependent variable, Lawson et  al. (2008) utilize the percentage of attendance at each game, and a Tobit regression. Results estimated by Lawson et al. (2008) find that Beckham’s impact on the league was rather substantial, with his presence on the roster increasing attendance by about 32%, and having him playing in a game further increased attendance by another 24%. In other words, it was the case that in his first season, David Beckham’s presence on the field could increase attendance at a match by over 50%. From this, a quick calculation finds that Beckham was able to generate an additional $400,000 per game. Although Beckham was paid a high salary to come to the US, Lawson et al. (2008) note that the amount the league was required to contribute to his wages was only $400,000 a year, meaning that his presence had potential for the league to make an additional $20 million a year. Most recently, Jewell (2017) examined the effects that Designated Players had on attendance in matches over a six-year period from 2007 through 2012. This research advances the literature by being the first to consider the effects of superstar players on the demand for MLS attendance beyond just examining a single season of play. In this manner, not only is Jewell (2017) able to provide a better picture of the effect that the Designated Player Rule has had on attendance across the league, but he is also able to offer a more nuanced examination of the impact that individual Designated Players had over this time period. First, through analyzing the change in average attendance based on the presence of each Designated Player over a six-year time period, the research shows that there are often large year-to-year fluctuations in average attendance based on marquee players appearing in a match. For example, while David Beckham is calculated to have had an impact on average attendance of about 45.9% in his first year (which is similar to the estimated results from Lawson et al. (2008)), his second year only witnessed a 0.6% increase in attendance. Furthermore, some players, like Thierry Henry (previously of Arsenal and Barcelona), actually moved from having a positive impact on attendance (17.1% in his first year) to having a negative effect (-7.8% and -1.6% in his second and third years, respectively). In continuing his analysis, Jewell (2017) estimated a censored regression to determine the year-to-year effects that three Designated Players (David Beckham, Blanco, and Marquez) had on average attendance at MLS matches. First, it

337

is the case that two of the three players did not have a significant impact on consumer demand for attendance in every year that they played in the league. While a player like David Beckham is found to have impacted attendance on average by 25% over his six years in the league, it was found that he did not have any effect on attendance in the league in 2010 and 2012. Additionally, it is found that while these three Designated Players did not all help to improve home attendance, they all significantly increased attendance at away matches in which they played. This finding may suggest a novelty effect in regards to Designated Players and their ability to increase demand. This belief is especially the case when you consider the yearly results for these players, and find that the highest impact on attendance usually comes in the first two years in which the player was appearing in MLS matches.

FUTURE RESEARCH HORIZONS Overall, the economic-based literature focused on examining Major League Soccer has touched on a number of topics of interest and produced intriguing theoretical and empirical results. In considering the overall lineage of research, the creation of the Designated Player Rule marks a major change in focus in the MLS literature, as almost all of the studies conducted after 2007 place their focus on the impact of the rule on wages, team performance, and consumer interest. At the same time, there are still a number of areas through which the economic examination of MLS can be advanced, all of which are covered in what follows.

Wages and Performance As was noted in the review of the literature, most of the studies focused on player salaries in MLS have used rather limited data sets and time frames through which to examine player and team compensation. However, considering that there is now well over a decade of salaries, it would seem that there is room for research studies to improve on the modeling provided in earlier studies because of the wealth of data now available to the public. When considering determinants of player compensation in MLS, the research has generally ignored controlling for player performance outside of controlling for their position, goals scored, or appearance in all-star games. While it is more difficult to model the performance of individual soccer players because of the nature of the sport,

338

THE SAGE HANDBOOK OF SPORTS ECONOMICS

MLS’s new partnership with data companies now has provided match-level and season-level data that can better help to quantify player performances. Thus, through combining salary and performance data sets, it is the case that researchers can attempt to empirically model the relationship between on-field performance of individual players in the league and the compensation they receive for their services. In this manner, it may be beneficial in helping to tease out a better understanding of whether Designated Players are being compensated for their on-field performance, their off-field marketability, or a combination of both.

DEMAND AND THE STADIUM SITUATION Next, considering the research focused on demand, there are still a number of gaps which exist within the literature. First, academics examining attendance demand for MLS matches have entirely done so through using live attendance at either the match or season levels. However, as the league has expanded and gained greater exposure in North America, it is certainly the case that there is a need to consider the television audience for MLS games. The examination of demand for television viewership provides a number of potential avenues for research. Primarily, these studies could possibly consider whether Designated Players impact television audience numbers, as well as if televised games from rival leagues (both the major North American professional sport leagues as well European soccer leagues) affects MLS television viewership. In the early years of MLS, it was the case that all of the teams in the league were the secondary tenant at the facility where they played. In most cases, this meant that the team was often playing in large cavernous stadiums built for National Football League contests which expected attendance of over 60,000 individuals per game. As of 2017, it is now the case that 14 of 22 MLS franchises have their own soccer-specific stadium. Despite the boom in building soccer-specific stadiums, there has been almost no research which has actually considered the impact of these specialized venues from a demand or revenue perspective. To date, the only study on soccer-specific stadiums (DeSchriver, Rascher, & Shapiro, 2016) concluded that these stadiums did cause a significant increase in attendance, but that this impact was minimized by the existence of the Seattle Sounders. With recent expansion teams such as Atlanta United FC also experiencing high attendance demand while

playing in a NFL stadium, the question remains whether soccer-specific stadiums will continue to play an important role for MLS teams.

LEAGUE EXPANSION Finally, the last subject which needs to be considered is the expansion of the league. The singleentity structure of the league was meant to protect it from previous mistakes that the NASL made, including the rapid expansion of the number of franchises and salaries paid to players. In the past decade, both of these have increased drastically within the MLS, and it is certainly the case that the league now has its focus on expanding to be bigger than the NASL was at its peak. As the expansion of the league continues, there is a need to consider whether the manner and speed in which the league has grown has been beneficial, or if it is possible that the league has attempted to put its product into too many markets. In this manner, continued expansion by MLS provides a number of avenues for academics to analyze a relatively new league that is attempting to grow and compete against both rival domestic professional sport leagues and overseas soccer leagues.

REFERENCES Bird, P. J. W. N. (1982). The demand for league football. Applied Economics, 14, 637–649. Coates, D., Frick, B., & Jewell, T. (2016). Superstar salaries and soccer success: The impact of designated players in Major League Soccer. Journal of Sports Economics, 17, 716–735. DeSchriver, T. D. (2007). Much adieu about Freddy: Freddy Adu and attendance in Major League Soccer. Journal of Sport Management, 21, 438–451. DeSchriver, T. D., Rascher, D. A., & Shapiro, S. L. (2016). If we build it, will they come? Examining the effect of expansion teams and soccer-specific stadiums on Major League Soccer attendance. Sport, Business and Management: An International Journal, 6(2), 205–227. Haas, D. J. (2003). Technical efficiency in the major league soccer. Journal of Sports Economics, 4, 203–215. Jewell, R. T. (2017). The effect of marquee players on sports demand: The case of US Major League Soccer. Journal of Sports Economics, 18(3), 239–252.

THE ECONOMICS OF MAJOR LEAGUE SOCCER FROM THE NASL TO MLS

Jewell, R. T., & Molina, D. J. (2005). An evaluation of the relationship between Hispanics and Major League Soccer. Journal of Sports Economics, 6, 160–177. Kuethe, T. H., & Motamed, M. (2010). Returns to stardom: Evidence from US Major League Soccer. Journal of Sports Economics, 11, 567–579. Lawson, R. A., Sheehan, K., & Stephenson, E. F. (2008). Vend it like Beckham: David Beckham’s effect on MLS ticket sales. International Journal of Sport Finance, 3, 189–195. Ludtke, M. (1976). Soccer is getting a toe hold. Sports Illustrated, 30 August. Accessed from: www.si.com/vault/1976/08/30/615444/soccer-isgetting-a-toehold Mathias, E. (1999). Big league Perestroika? The implications of Fraser v. Major League Soccer. University of Pennsylvania Law Review, 148, 203–237. Shaw, J. D. (2014). Pay dispersion. Annual Review of Organizational Psychology and Organizational Behavior, 1, 521–544. Smith, C. (2016). Major League Soccer’s most valuable teams 2016: New York, Orlando thrive in first seasons. Forbes Magazine, 7 September. Accessed from: www.forbes.com/sites/chrissmith/2016/

339

09/07/major-league-soccers-most-valuableteams-2016-new-york-orlando-thrive-in-firstseasons/#b741b91270df Stebbins, H. (2015). Blind Draw: How Major League Soccer’s Single Entity Structure and Unique Rules Have Impacted Soccer in the United States. Willamette Sports Law Journal, 13, 1–15. Sung, H., & Mills, B. M. (2018). Estimation of gamelevel attendance in major league soccer: Outcome uncertainty and absolute quality considerations. Sport Management Review, 21(5), 519–532. Online first, 14 December 2017. https://doi. org/10.1016/j.smr.2017.12.002 Twomey, J., & Monks, J. (2011). Monopsony and salary suppression: The case of Major League Soccer in the United States. The American Economist, 56, 20–28. Watanabe, N. M. (2006). Competitive balance and other variables effects on Major League Soccer attendance. Doctoral dissertation, University of Illinois at Urbana-Champaign. Wilson, D. (2008). MLS Team Valuations. Bleacher Report, 24 September. Accessed from: https:// bleacherreport.com/articles/61297-mls-teamvaluations

PART V

Sports Events

34 The Economic Impact Measurement of the Olympic Games Holger Preuss

In the run-up to the Games the responsible policy makers, organizers and their consultancy, for political reasons constantly estimate the economic impact of the Olympic Games. The International Olympic Committee (IOC) changed its bidding process following the Paris 2024 Games, and now demands long-term development plans, including an economic impact study (IOC, 2016). The size of the Olympic Games is growing in many dimensions even though the number of athletes, sports and participating nations stays stable. Over the past 20 years, the main cost drivers were the global audience (resulting in ever more media representatives and sponsor appearances) and politicians using the Games as an excuse for urban regeneration. The desire to participate live in the Olympic Games led to astronomic costs to provide additional accommodation, safety and local transport. National political forces, lobbying from the local construction industry, opportunistic profit maximization of decision makers and even corruption sky-rocketed the costs of the Games, as recently seen in Beijing 2008, Sochi 2014 and Rio de Janeiro 2016. On the other hand, the revenues also grew. The IOC (2017a, p. 24) reports that the last Olympiad (2013–2016) generated approximately US$4 billion in TV rights and $1 billion in worldwide sponsorship deals. This amount does not include

national sponsorship and ticket revenues by the Organizing Committee. For the Games in Los Angeles 2028 the IOC will pump $1.8 billion into California by supporting the organization (IOC, 2017b, p. 14). These examples give a good idea of why there is controversy in the debate about the size of the economic impacts that can be generated by the Olympic Games. To conduct a sound impact study, many premises (size of economy, time frame, etc.) must be taken into account. Impact studies without defining size, time or the welfare function generate different results even when the same Games is being investigated. Further, studies can differ by trying to consider the economic complexity while others only focus on a part of the economic impact. Some only calculate economic surplus by looking at IOC subventions and tourism expenditures that lead to economic stimulation. Others calculate all debts to construct new sport venues, which cannot be compensated by revenues from Olympic tourism and IOC subventions. There is no right or wrong, only more or less complete calculations. Positive studies are correct when claiming that the host city attracts autonomous money, causes accelerated city development and creates positive image effects. Their opponents are also right when criticizing that the Olympic Games force hasty planning and some unneeded or oversized

344

THE SAGE HANDBOOK OF SPORTS ECONOMICS

infrastructure investments. The fact is that many impact studies ‘have violated standard practices’ (Hudson, 2001, p. 21) by biased assumptions and methods. A literature review revealed that between 2004 and 2016 only 14 (out of 109) peer-reviewed papers were published about impact studies of mega-events. Most of them focus on soccer (the FIFA World Cup and European Championships), and many of them include only parts of an economic impact (e.g. employment, the hotel sector, tourism). Only two measured the entire economic impact (Ahlert, 2006; Kasimati & Dawson, 2009). The first Olympic economic impact study (Los Angeles 1984) was a direct result of the interest generated by the overwhelming deficit created by Montreal 1976 (Bohlmann & van Heerden, 2008, p. 384). However, besides peer-reviewed papers there are some other complete studies published by researchers (Preuss & Weiss, 2003; Preuss, Kurscheidt, & Schütte, 2009; Preuss, 2011b; CDES, 2016). Not many publications have measured seriously the economic impact of the Olympic Games. The purpose of this chapter is to assess the economic impact measurement, to recommend best practices and to point at typical mistakes. Many studies violate standard practices in impact analysis. The suggested structure has then to be customized to each host city as the conditions (economic strength, size, amount of investment, etc.) vary and therefore the economic impact is different for each Olympic Games.

IMPACT MEASUREMENT Other than cost–benefit analysis, an economic impact study is limited to the monetary flow in or out of a region for a given period. Opportunity costs, intangible effects, social or environmental changes are not considered but they would be in a more complete cost–benefit analysis (see Humphreys & Prokopowicz, 2007). Economic impact measurement is based on a simple model that refers to the Keynesian economic cycle (Figure 34.1). The basic idea is to identify the monetary streams that enter or leave a region and thus estimate the net primary economic impact. That impact is the increased demand that stimulates economic activity (output) and then leads to additional revenues. However, the money streams leading to a change in demand have to be differentiated (Figure 34.2). Figure 34.3 shows an economic region that can be a city, region or nation. First, we need a

systematic measurement of all monetary streams entering. These stimulate economic activity in a region because they cause demand. Most of the money entering an Olympic host region is related to tourism revenues and the IOC contribution. Depending on the size of the region, one can also count monetary streams from national sponsors and governmental financial support. The resources leaving the region due to the Olympic Games reduce demand. These are usually imports and tax payments. Finally, one also has to consider the monetary streams that are crowded-out, which are difficult to measure. To be considered are the money streams that would have entered without the Games (e.g. tourists who do not come due to high prices) or money that is no longer spent in the region due to increased demand (residents who leave the region to avoid the overcrowded city). Finally, we need to consider the monetary re-distribution. The most important step to calculate the true primary impact is the serious and valid collection of all data. Based on the primary net impact one can calculate the secondary effects, also called induced effects. The multiplier determines the induced impact, which is the additional change in consumption, investment and export spending that results from the initial change in spending due to the Olympics working its way through the economy. How the Olympics affects a region can be determined by one of two leading methods. One is using a computable general equilibrium (CGE) model (e.g. Madden, 2006; Dwyer, Forsyth & Spurr, 2006, 2007; Borowski et  al., 2013). CGE models are a class of economic models that use actual economic data to estimate how an economy might react to changes in policy, technology or other external factors. The other is to use a macroeconomic input-output model (e.g. Lee & Taylor, 2005; Ahlert, 2006). The advantage of input-output tables is that expenditures associated with the Olympic Games translate into information about income and employment and provide data about the industrial sector where additional activity occurs. Once the input-output table is constructed (which basically is done for each country but not for state or cities levels), multipliers for value added, income and employment can be generated via mathematical manipulation. Each multiplier measures the effect on total regional income of a unit change in some component of aggregate demand. For example, suppose that firms increase activities in construction for Olympic facilities. The result of the increase is a rise in regional income as the investment is paid out in the form of wages, salaries and profits to suppliers of factor services. The recipients of the income in turn purchase goods and services, which creates

The Economic Impact Measurement of the Olympic Games

345

Figure 34.1  Olympic Keynesian approach (CDES, 2016, p. 12; according to Preuss, 2000, p. 55)

Figure 34.2  Basic monetary streams in impact studies

Figure 34.3  Matrix of relevant monetary streams in impact studies

346

THE SAGE HANDBOOK OF SPORTS ECONOMICS

income, further expenditures and savings in the economy. In other words, it has multiplied by a factor such that the result is the initial investment plus the multiplied effects. However, care has to be taken that income multipliers are used and not output multipliers (see Crompton, 1995). An input-output table is usually only the center of a broader model that includes the changes of the economic situation, prices, taxes, interests, etc. As multipliers depend on the economic situation, the type of primary affected industries and how these spend their income, the income level of workers in those industries, import quotas of intermediate input, etc., they are different every year and for any region. The use of dynamic multipliers generates a better forecast of an economic impact than a static multiplier. For an economy close to full employment, additional economic activity is more likely to result in the crowding-out of existing economic activity (Mankiw, 2015, p. 486), resulting in less induced impact and a lower multiplier. In contrast, if many resources are underemployed and crowding-out does not occur, the multiplier will be much higher. So, the multiplier depends on the macroeconomic business cycle (how close the regional economy is to full employment) at the time of the highest event expenditures as well as the overall maturity, diversity and productivity of the economy and the productivity of the activities on which the money is spent (usually construction and service and tourism industries). That shows how difficult a good impact study is. For all impact studies, and particularly for ex-ante studies, it is more credible to work with two scenarios. Having a best-case and worst-case scenario, the investigator can anticipate a better and a worse economy as well as more or less imports and more or less tourists coming to attend the event, etc. The result will be then in between the span of best and worst case. Figure 34.4 illustrates a three-scenario study. For the Ile-de-France region a regional multiplier of 2.35 was used. To be conservative and not to overestimate the result, three scenarios corresponding to three different multiplier values were used. • Low scenario: Here the value of the multiplier is set at 1.1 to account for an unprecedented case where the Games would only produce a very limited number of effects beyond their primary impact (net inflow). It would require the convergence of extraordinary factors (major displacement of expenditure and tourists, etc.). A multiplier can never be below 1, as the money influx is 1 and if it immediately leads to imports, we reach an effect of 1.

Figure 34.4  Three-scenario ex-ante study on the Olympic Games Paris 2024 impact for Ile-de-France region (CDES, 2016, p. 56)

• Middle scenario: Here the value of the multiplier is set at 1.5 to reflect common results achieved based on multipliers aggregated at the level of the State or of large administrative regions. • High scenario: Here the value of the multiplier is set at 2.0 to take account of the fact that the result of the meta-analysis was higher than that derived from other case studies. In order to avoid criticism related to the value of the multiplier and, in turn, of the overestimation of the economic impact of the 2024 Games, a range from 1.1 to 2.0 was used.

THE IMPORTANCE OF GOOD DATA In principle, there are two ways to collect data (Table 34.1). The bottom-up approach gathers data during the Olympic Games or at the time when the impact occurs. The top-down approach

The Economic Impact Measurement of the Olympic Games

347

Table 34.1  Problems of data collection Top-down approach

Bottom-up approach

Causality problem: The ‘event case’ cannot be distinguished from Identification problem I: Impact interaction cannot be the ‘non-event case’. Other activities in the city intervene in considered. Gross data are collected and crowdingthe event impact out cannot be measured Visibility problem I: ‘White noise’ means the Olympic impact Identification problem II: Re-distribution of money disappears in statistics. Overall activities are bigger than the cannot be detected. Money spent is collected but Olympic impacts not where the money originates from Visibility problem II: ‘White noise’ intervenes in periods before Methodical problem: The measurement of intangibles and after the Games. Other activities are bigger than the is difficult: how to measure up-skilling, image, Olympic activities traffic improvements, etc. Missing data I: Regionalized statistics are often missing and are not available in short periods. Olympic Games are only two weeks’ duration but are often touching two months Missing data II: The statistics for intangibles are missing: how to measure up-skilling, image, traffic improvements, etc.

Figure 34.5  Factors influencing the net impact of the Olympic Games

uses secondary data from available statistics (e.g. Kasimati & Dawson, 2009, for Athens 2004; Tien, Lo, & Lin, 2011, for mega-events in general; and Mehrotra, 2012). Both methods have problems in finding all the data that are necessary for an impact study. Neither method of data collection satisfies. While the top-down approach solves the problem of getting close to the net impact (as the overall changes in an economy are displayed), the bottom-up method helps to identify effects that are caused by the Olympic Games.

All data entering and leaving the region related to the Olympic Games needs to be collected and calculated to determine the net impact. Figure 34.5 illustrates the complexity of making an accurate net impact. For each monetary stream caused by the Olympics, one has to decide if it is autonomous money or not. The decision depends on knowledge about whether the same money would have come to the region even without the Games. For example, the construction of a new metro line can be purely related to the Olympics or it can be a plan that existed before.

348

THE SAGE HANDBOOK OF SPORTS ECONOMICS

FRAMING AN IMPACT STUDY A reputable study has to define certain premises. Without a clear definition of time, region, welfare function (or perspective) and degree of completeness of the study, everyone can manipulate the result to produce any wanted outcome. The following examples will illustrate how different premises can produce different results.

Time During the life-cycle of the Olympic Games the monetary streams differ (Figure 34.6). First, the construction industry is stimulated and later the tourism service industry. Therefore, it is important to define for what period an impact study is conducted. A short period (the time of the Olympic Games only) would only consider the revenues from Olympic tourists and the consumption of the Organizing Committee during the event time. That would include all imports to prepare for the Games and investments in sport venues. A longer period includes the time after the Games, for example, the payment of interest for debts undertaken for construction and the transformation of venues into long-term

utilization. A study about the potential Olympic Games in Frankfurt 2012 showed a 30% larger impact when considering 17 years instead of two years (Preuss & Weiss, 2003, p. 213). Figure 34.6 illustrates the time that should be considered for an Olympic impact study. PreOlympic activities start seven years before the Olympic Games, when the Games are awarded to a city. However, for Los Angeles 2028 the decision was 11 years in advance. During this period operating expenditure (OCOG) and consumption from Olympic visitors who attend test events, visit Olympic conferences, prepare sponsor activities, and the visits of coaches and athletes to the Olympic facilities have to be considered. The largest amount of autonomous money entering the region in this period is for venue and general infrastructure construction. These often-massive investments are usually financed by national taxpayers’ money or by loans and stimulates the regional economy in the case of construction companies coming from that region. The Games stimulate the economy through operating expenditures and visitors coming from outside the host region (athletes, staff, facilitators, officials, tourists, media, Olympic family, etc.). Post-Olympic activities last for many years. However, the later the expenditures occur the more difficult it is to find the causal link to the Olympic Games. The monetary streams considered should be limited to payments for debts and related interest, additional visitor expenditures directly related to the Olympics and expenditures for transforming the venues into their permanent use. One should avoid mixing up economic impact with Olympic legacy. Legacy is any economic activity that is generated from the changed structure. For example, a tourist visiting the Olympic Games adds to the Olympic economic impact, whereas a tourist coming to the Olympic city due to better hotels or to visit the iconic Olympic stadium after the Games is a tourist attracted by the changed structure and therefore is counted as legacy. Figure 34.7 illustrates the three periods considered for Paris 2024. In the post-Games period, tourism attracted by the Olympic Games was considered.

Region

Figure 34.6  Time frame to measure the impact of the Olympic Games (CDES, 2016, p. 13; according to Preuss, 2000, p. 41)

The size of the region matters tremendously. It defines whether a monetary stream is coming from outside or whether any purchase is only a regional re-distribution or an import. The smaller the region under consideration, the more money entering the region (e.g. more people live outside the region and carry money in to attend the Games).

The Economic Impact Measurement of the Olympic Games

349

Figure 34.7  Three-scenario forecast of three periods for the Olympic Games in Paris 2024 (CDES, 2016, p. 57) On the other hand, the smaller the region the more imports are needed because not all necessary goods can be produced in that small region. Additionally, the multiplier is smaller for a smaller region because it reflects less economic activity. This leads to more imports and the induced effect is smaller (Figure 34.2). It is not trivial to determine whether a monetary stream counts for an economic impact or not. This will be illustrated by the Olympic visitors. The same consideration, which is shown in Figure 34.8, can be used for the decision if investment counts as economic impact. The example of the Los Angeles Games 2028 is used in Figure 34.8. Monetary streams must be differentiated by whether they are exogenous for California or not. All groups of visitors in Figure 34.8 are different if the size of the region changes. For example, a visitor from San Francisco will be in group K if we consider an impact for California, but the same person would be in group B if we only consider Los Angeles City as the region. The methodology used here needs to integrate the reason to visit the Games in 2028. To discover whether consumption should be included into the impact study or not depends on the person’s travel movement combined with how much the Olympics in Los Angeles 2028 were the reason for the person’s decision to travel. ‘Extenders’ (A), ‘Olympic visitors’ (B) and ‘Home stayers’ (C) are Olympic visitors bringing exogenous resources to California. The ‘Home stayers’ (C) are an example of ‘import substitution’ (Cobb & Weinberg, 1993) because those Californians who stay at home and spend their holiday budget in Los Angeles and not outside California do not carry money out of the state but keep it in.

A, B and C add money to the Californian economy. On the other hand, ‘Cancellers’ (E1) and the ‘Runaways’ (D) reduce this impact because they are crowded-out by the Olympic Games. The methodological challenge is to distinguish these two groups from those visitors who merely time switched their visit to California. For example, residents avoiding the Olympics are not necessarily ‘Runaways’ (D). They can also be ‘Changers’ (F), i.e. individuals who ‘time switch’ their holiday trip from another time of the year to the period of the Olympics. While ‘Runaways’ must be considered as crowded-out, the ‘Changers’ do not spend more money outside California than they would have done anyway. Similarly, while ‘Cancellers’ (E1) are crowded-out, ‘Pre/Post switchers’ (E2) are not. While ‘Cancellers’ (E1) would not visit California at another time, ‘Pre/Post switchers’ (E2) have just postponed their visit. To calculate the economic impact, ‘Changers’ (F) and ‘Pre/Post switchers’ (E2) must not be considered due to the time-shift in their consumption. However, the ‘Runaways’ (D) must be considered. This can be done either by a survey after the Olympics (a bottom-up approach) or by using statistics on individuals leaving the country (a topdown approach). The identification of ‘Cancellers’ (E1), who can be found only by using ex-post trend calculations, is the focus of this chapter (see also Preuss, 2011a). Figure 34.8 shows three other groups that would also have spent their money in California without the Olympics. Thus the ‘Casuals’ (G), ‘Time switchers’ (H) and ‘Citizens’ (K) are neutral in their contribution to the economic impact. However, all of them have presumably spent more money on their visits due to price increases and

350

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Figure 34.8  Importance of the size of the region when measuring an impact (Preuss, 2011a, p. 370) additional consumption than they would have done in their regular visit to the region. Event tourists spend more money on average than leisure time tourists (Preuss et al., 2009). Leeds (2008, p. 466) investigated displacement effects to be considered when the region is larger. That means that tourists stay away from visiting an Olympic host city, such as Salt Lake City in 2002, but go to other places instead, such as other ski resorts. Thus 16 ski resorts in Colorado gained a net impact of $160 million in 2002. However, Baade and Matheson (2004) and Teigland (1999) report an opposite effect. According to them, Sydney 2000 and Lillehammer 1994 attracted visitors who wanted to go to other Australian/Norwegian destinations. This fact opens an important discussion related to crowding-out. It is difficult to measure whether the Olympic Games have been the reason for crowding-out of visitors or residents. Crowding-out can occur due to high demand and therefore high prices of accommodation (Lamla, Straub, & Girsberger, 2014, p. 1702) and transport. However, crowding-out is only related to the group E1 and D as they do not just time switch. Table 34.2 shows the share of groups for nine major events measured towards a national

economy. One easily can see that the groups are different and that there is no standard pattern.

Welfare Function/perspective Many economic effects affect the population in a region differently. An impact study can focus on a particular group and therefore does not need to reflect the monetary effects for other groups.

Completeness Further, it is important to define how complete an impact study is. In a minimum case, only the monetary streams are considered. Extended studies can also consider the pecuniary effects (price level changes) or the motivation to invest and consume; in other words, if the consumption rate changes. A change in consumption rate would affect the multiplier, thus a dynamic model is needed. Finally, it is important to define how many effects are in the study. Most studies only calculate the primary impact, thus avoiding CGE models or multipliers. Others calculate solely the net effect

40.7  4.7  7.0 22.8  1.7 23.0

Residents (K) Home stayers (C) Extensioners (B) Event visitors (A) Casuals (G) Time switchers (H)

38.3  6.7  9.3 26.6  8.3 10.7

FIFA World Cup 2006 (Germany) 23.0  1.1  1.5 33.1 12.0 29.4

FIFA World Cup 2010 (South Africa) 27.0  5.0  2.0 37.0 23.1  5.9

EURO 2008 (Austria)

87.4  4.6  2.8  2.2  0.9  2.1

Handball World Cup 2007 (Germany)

Sources: Preuss, Kurscheidt, & Schütte, 2009; Preuss, 2011a; Preuss & Schütte, 2014; Kwiatkowski, 2016

FIFA World Cup 2014 (Brazil)

Group of event attendees

Table 34.2  Share of visitors at events based on economic importance

62.5  4.6  6.0 15.5  2.9  8.5

Hockey World Cup 2006 (Germany) 16.8 15.8 – 46.0 17.0  4.9

Commonwealth Games 2002 (City of Manchester)

 4.8  0  9.5 29.0 55.8  1.1

2012 PWA Windsurf World Cup (Sylt)

46.6  0  1.4  4.41 28.8 18.3

2012/2013 Ski Jumping World Cup (Oslo)

The Economic Impact Measurement of the Olympic Games 351

352

THE SAGE HANDBOOK OF SPORTS ECONOMICS

but no indirect effects. These are economic activities not directly related to the Olympic Games but also caused by them. Another explanation for the great differences in impact studies is a missing explanation of causality. Authors who try to make the Olympic Games appear expensive include expenditures that cannot be solely attributable to the Games. The important differentiation is between incremental and independent investments. Incremental investments are not paid by the Organizing Committee but often from taxpayers’ money and is for infrastructure that is only needed for the Olympic Games (e.g. a special sports venue). These must be included in an economic impact study. However, independent investments (e.g. airport, railway system or skiing resort) which are also needed regardless of the Games cannot be attributed as Games-related and therefore should not be included in the impact studies. The Olympic Games can cause an acceleration of these investments, in which case the interest to be paid for the money that is needed earlier is Olympicrelated. It is a question of definition regarding which infrastructure is only Games-related and which is needed regardless of the Games.

MOST FREQUENT MISTAKES The mistakes in economic impact assessments of the Olympic Games are manifold. Often there is

double counting, which is when the monetary streams entering a region are counted several times, leading to an overvaluation. For example, the OCOG expenditures have to be analyzed for their origin. If money stems from a state subsidy for infrastructure, the same subsidy has to be excluded from state investments. Or the national team expenditures paid to the OCOG cannot be counted as team expenditures to the region and the OCOG ticket revenues from spectators cannot be calculated as tourism expenditures. Another mistake often seen in impact stories is the non-inclusion of crowding-out effects regarding consumption and investment. These are difficult to estimate and even more difficult to measure (see Table 34.1). For example, consumers may cancel their stay for fear of overcrowding and price increases associated with the Games. Similar to crowding-out are the substitution effects. For this reason, it is good to not consider re-distributions in a region. However, a redistribution can be economically positive and negative and may be considered in the complete analysis. There is debate about whether the Olympic Games cause positive or negative re-distribution and also what their overall effects are (Matheson, 2008). Figure 34.9 displays two effects to consider, both the change of import rate and the change of multiplier. Both are based on the assumption that the Olympic Games stimulates more industries than would be the case without the Games. For simplicity, both effects are displayed separately.

Figure 34.9  Games versus non-Games effect by changes in import rates and different multipliers

The Economic Impact Measurement of the Olympic Games

On the left side, the first bar shows the economy without the Games. $100 million enter the region with an import rate of 30% on average. The Olympic Games may have a higher import rate (40%, right bar) but the higher primary impact due to the Games attracting more money overcompensates that effect. The right side of Figure 34.9 is showing the multiplier effect. The first two bars illustrate the economy without the Games with a multiplier of 1.25. When the Games are staged the same industries may not be demanded and re-distributions may additionally stimulate other industries (e.g. visitors go to see sports and not to visit the zoo). Thus, the economy may lose 20% by crowding-out and other industries are stimulated but not more than they would have been without the Games. Now more money goes into the construction and tourism industries. These may have a larger multiplier than the average multiplier without the Games because those industries are stronger in the region. For example, if Innsbruck (Austria) stages another Olympic Winter Games after it has already transformed its infrastructure towards sport and tourism after staging the Olympic Games in 1964 and 1976 and Youth Olympic Games in 2012, there would already be the necessary industry and infrastructure for any new sports event. Any publication that claims per se that the Olympic Games leads to crowdingout and therefore a lower overall impact may be wrong (Barclay, 2009, p. 64; see also Owen, 2005; Matheson, 2008). Another misinterpretation of economic impacts is the expectation that jobs are always created in the host regon only. However, logically the labour is increased there where the production of goods takes place. Event industries import much to the host city. Thus many jobs get created in regions that produce goods for the event. In impact studies the region that is considered is important to define because only then one can distinguish wether a job is created or not (see Feddersen & Maennig, 2013).

DIFFICULTIES IN GETTING THE FULL PICTURE There are several issues that make it difficult to get a full economic picture of the Olympic Games (see also Li & Jago, 2013). Irrespective of the aspects mentioned above, five more are listed below. First, the economic situation during the main time of construction is not known. In an economic recession the construction of sport venues for the Olympic Games adds demand and increases employment. However, if the economy is in a boom phase, the market overheats and prices and

353

imports increase. Then the same investments have a very different economic impact for the region. In other words, whether or not construction stimulates the industry is dependent on the overall economic situation in the industries that are stimulated by the Games. Second, any economic activity changes the structures of a host city. Those changes will cause follow-up economic activity (e.g. the construction of a stadium will lead to stadium tours). However, whether or not a legacy is positive depends on the post-Games utilization. A new venue can be a ‘white elephant’ if it is not used, or an engine to activate business for surrounding areas when it is properly used. It is extremely difficult to measure the value of the economic benefit because the Olympic Games just change the supply side (location factors) but the economic activity (demand) needs to follow. Third, the Olympic Games demand high capacities for tourism, sport venues and conference facilities. Often the supply increases more than the post-event demand for such capacity. Therefore, a post-event regulation is needed, and there are often additional costs to reduce the size of or dismantle venues. In some industries the reduction of overcapacity also causes bankruptcies (e.g. hotel industry) (Teigland, 1999; Humphreys & Prokowicz, 2007). Fourth, contrary economic effects mean that one group gains while another loses. The re-distribution of money in a region can be socially unjust. Economic impact studies do not distinguish who is gaining money. They just measure if there is a monetary gain (Pareto Optimum). Fifth, even though an economic impact study only measures the monetary impact, there are intangible effects related to economy. Intangibles are only considered in cost–benefit analyses. However, for a complete picture of the economic effects of the Olympic Games one should consider pecuniary changes, an increased interest in a destination, the upskilling of the workforce or better traffic regulation, which all lead to higher productivity. Other effects caused by the Olympic Games, such as an enhanced national identity, greater happiness and more civic pride, are valuable assets for a prospering city when it is competing for its position among many world cities and economic centers.

CONCLUSION Economic impact studies are still needed for political decision makers. They provide an

354

THE SAGE HANDBOOK OF SPORTS ECONOMICS

important view of the Olympic Games from a monetary perspective. The Olympic Games represent the highest performance in sport but, due to their immense cost (often for the taxpayer), it is important to consider the economic impact for the region. Moreover, recently, many residents have voted against staging the Olympic Games in their city and an often-mentioned reason is that the public money needed for the Games could be better spent. This chapter has showed that the measurement of the economic impact of the Olympic Games is very complex. Studies can be manipulated relatively easily to reach the politically desired result. This has led to a decrease in trust in these studies, and therefore it is ever more important to conduct good studies in future. There are four ways in which to manipulate a result. First, the quality of an economic impact study depends on the preciseness of the data collection to determine the primary impact. The best method is the bottom-up approach because only then the causality of monetary streams can be measured. However, any bottom-up data collection is costly and can only be done during the event. An ex-ante study is not possible. Second, any trustworthy impact study must define its parameters: the time, region, perspective and completeness of the report. If the parameters are not defined, the reader does not know what story of the impact is told thus, leading the reader into a false expectation. A third way to manipulate impact studies is by misusing one of the many methodological shortcomings, such as to measure crowding-out effects. The fact that regional input-output tables or CGE modelling is usually not available forces authors into making estimations. Therefore, a trustworthy study should work with various scenarios. By taking bigger multipliers or not considering imports or the re-distribution of purchasing power in different industries the calculated impact is different. Finally, any impact study always needs to forecast some data, such as the changing pattern of tourism due to the Olympic Games. Also difficult is the right attribution of investments to the Olympic Games. To do this effectively the future city development plans need to be known.

REFERENCES Ahlert, G. (2006). Hosting the FIFA World Cup Germany 2006: macroeconomic and regional economic impacts. Journal of Convention and Event Tourism, 8(2), 57–77.

Baade, R., & Matheson, V. (2004). The quest for the cup: assessing the economic impact of the World Cup. Regional Studies, 38, 343–354. Barclay, J. (2009). Predicting the Costs and Benefits of Mega-Sporting Events: Misjudgment of Olympic Proportions? Institute of Economic Affairs. Oxford: Blackwell. Bohlmann, H. R., & van Heerden, J. H. (2008). Predicting the economic impact of the 2010 FIFA World Cup on South Africa. International Journal Sport Management and Marketing, 3(4), 383–396. Borowski, J., Boratynski, J., Czerniak, A., Dykas, P., Plich, M., Rapacki, P., & Tokarski, T. (2013). Assessing the impact of the 2012 European Football Championships on the Polish economy. International Journal of Sport Management and Marketing, 13(1/2), 74–103. CDES (Centre de Droit et d’Economie du Sport). (2016). Paris 2024. Etude d’impact. Limoges. Retrieved from http://www.cdes.fr/sites/default/ files/files/Expertise/réf%20éco/Résumé%20étude %20finale%20JO2024.pdf (Accessed on 8/4/2019). Cobb, S., & Weinberg, D. (1993). The importance of import substitution in regional economic impact analysis: Empirical estimates from two Cincinnati area events. Economic Development Quarterly 7(3), 282–286. Crompton, J. L. (1995). Economic impact analysis of sports facilities and events: Eleven sources of misapplication. Journal of Sport Management 9(1), 14–35. Dwyer, L., Forsyth, P., & Spurr, R. (2006). Assessing the economic impacts of events: a computable general equilibrium approach. Journal of Travel Research, 45, 59–66. Dwyer, L., Forsyth, P., & Spurr, R. (2007). Contrasting the uses of TSAs and CGE models: measuring tourism yield and productivity. Tourism Economics, 13(4), 537–551. Feddersen, A., & Maennig, W. (2013). Mega-events and sectoral employment: the case of the 1996 Olympic Games. Contemporary Economic Policy, 31(3), 580–603. Hudson, I. (2001). The use and misuse of economic impact analyses. Journal of Sport & Social Issues, 25(1), 20–39. Humphreys, B. R., & Prokopowicz, S. (2007). Assessing the impact of sport mega-events in transition economies: Euro 2012 in Poland and Ukraine. International Journal of Sport Management and Marketing, 2(5/6), 496–509. IOC (2016). Report of the IOC 2024 Evaluation Commission. Lausanne: International Olympic Committee. IOC (2017a). IOC Annual Report 2016. Lausanne: International Olympic Committee.

The Economic Impact Measurement of the Olympic Games

IOC (2017b). Host City Contract. Principles. Lausanne: International Olympic Committee. Kasimati, E., & Dawson, P. (2009). Assessing the impact of the 2004 Olympic Games on the Greek economy: a small macroeconometric model. Economic Modelling, 26(1), 139–146. Kwiatkowski, G. (2016). Composition of event attendees: a comparison of three small-scale sporting events. International Journal of Sport Finance, 11(2), 163–180. Lamla, M. J., Straub, M., & Girsberger, E. M. (2014). On the economic impact of international sport events: microevidence from survey data at the EURO 2008. Applied Economics, 46, 1693–1703. Lee, C. K., & Taylor, T. (2005). Critical reflections on the economic impact assessment of a mega-event: the case of 2002 FIFA World Cup. Tourism Management, 26, 595–603. Leeds, M. A. (2008). Do good Olympic Games make good neighbors? Contemporary Economic Policy, 26(3), 460–467. Li, S., & Jago, L. (2013). Evaluating economic impacts of major sports events – a meta analysis of the key trends. Current Issues in Tourism, 16(6), 591–611. Madden, J. R. (2006). Economic and fiscal impacts of Mega sporting events: a general equilibrium assessment. Public Finance and Management, 6, 346–393. Mankiw, N. G. (2015). Macroeconomics, 9th Edition. Duffield: Worth Publishers. Matheson, V. A. (2008). The business of sports. In D. Howard & B. Humphreys (Eds.), Mega Events: The Effect of the World’s Biggest Sporting Events on Local, Regional, and National Economies (pp. 81–99). Westport, CT: Praeger.

355

Mehrotra, A. (2012). To host or not to host? A comparison study of the long-run impacts of the Olympic Games. Michigan Journal of Business, 2, 61–92. Owen, J. G. (2005). Estimating the costs and benefits of hosting Olympic Games: what can Beijing expect from its 2008 Games? The Industrial Geographer, 3(1), 1–18. Preuss, H. (2000). Economics of the Olympic Games: Hosting the Games 1972–2000. Sydney: Walla Walla Press. Preuss, H. (2011a). A method for calculating the crowding-out effect in sport mega-event impact studies: the 2010 FIFA World Cup. Development Southern Africa, 28(3), 367–385. Preuss, H. (2011b). Kosten und Nutzen Olympischer Winterspiele in Deutschland: Eine Analyse von München 2018. Wiesbaden: Springer Gabler Verlag. Preuss, H., & Schütte, N. (2014). Tourism at FIFA World Cups 2006, 2010 and 2014. First WASM Conference, Madrid. Preuss, H., Kurscheidt, M., & Schütte, N. (2009). Ökonomie des Tourismus durch Sportgroßveranstaltungen. Wiesbaden: Springer Gabler Verlag. Preuss, H., & Weiss, H. J. (2003). Torchholder Value Added. Eschborn: AWV-Verlag. Teigland, J. (1999). Mega events and impacts on tourism: the predictions and realities of the Lillehammer Olmypics. Impact Assessment and Project Appraisal, 17(4), 305–317. Tien, C., Lo, H., & Lin, H. (2011). The economic benefits of mega-events: a myth or reality? A longitudinal study on the Olympic Games. Journal of Sport Management, 25, 11–23.

35 Major Events: Economic Impact Wolfgang Maennig

INTRODUCTION Economic impact has developed into one of the most cited and used arguments to bid for major events. Ex-ante studies on the economic impact, which analyze regional income and employment impacts, are usually required by local chamberlains or regional financing authorities and occasionally are a basis for votes in the relevant assemblies, parliaments or public referenda. Occasionally, economic impact studies are required by law as proof of the efficient allocation of public resources.1 However, in most cases, economic impacts are part of public relations, attempting to convince the media and relevant stakeholders that the (planned) major event is beneficial for the local, regional or national economy and may, at least to a considerable degree, finance itself by the induced tax increases. There are different methods for estimating exante the potential benefits of hosting major events, including multiplier analysis, input-output calculations, surveys of decision makers in relevant sectors, and computable general equilibrium (CGE) models. Other rarely used methods are the social accounting matrix, the direct expenditure approach, or cost–benefit analysis (Davies, Coleman, & Ramchandani, 2013). The most common form of

multiplier and input-output analysis refers to the investment in stadiums and infrastructure as well as the spending of the organizing committee and the tourists as an expenditure shock, which in a Keynesian tradition, implies multiplying income effects. Additionally, most CGE models explicitly consider potential supply restrictions. Ex-post analysis focuses on impacts on macroor regional economic variables such as income and wages, employment and tax income, tourism, civic pride and wellbeing, and stock markets (Porter & Chin, 2012). This analysis occasionally occurs by analyzing a single time series, primarily on the basis of panel regression, and applying a difference-in-difference analysis that compares developments in the event location with the developments in comparable locations.

WORLD CUPS With no exception in recent times, World Cups have been the object of ex-ante impact analysis. For example, using an input-output model, the World Cup in Brazil in 2014 was suggested to ‘bring an additional R$112.79 billion (=US$50,56 billion at 2014/07/01 nominal exchange rate) to

Major Events: Economic Impact

the Brazilian economy, with indirect and induced effects being produced thereafter. In total, an additional R$142.39 billion will flow in the country from 2010 to 2014, generating 3.63 million jobs/ year’ (Ernst & Young Terco, 2011). The forecasted Brazil spending of R$22.46 billion (=US$10.12 billion) for infrastructure and organization would largely be financed by a positive impact on tax collection of R$18.13 billion. For the World Cup 2010 in South Africa, Grant Thornton (2004) calculated a net economic gain of R$21.3 billion (=US$2.0 billion at 2010/07/01 nominal exchange rate) for the South African economy based on 230,000 foreign tourists arriving for the tournament and construction costs totaling R$12.7 billion. Relative to current South African GDP, this amount would have corresponded to a 1.5% increase in GDP, ‘an equivalent of 159,000 annual jobs’. In an update, Grant Thornton (2008) increased its estimates to a GDP boost of US$6.0 billion, an additional employment of 381,000 jobs, and an additional tax income of US$2.1 billion. Grant Thornton estimated that at least 480,000 World Cup tourists would visit South Africa. A computable general equilibrium model-based forecast by Bohlmann and van Heerden (2005) assumed a 10% addition to the capital stock of the construction and transport industry, capital-­ augmenting technological changes and, ultimately, a positive impact of 0.94% of GDP (in the long term) and the creation of 50,000 jobs. For World Cup Germany 2006, a forecasting industry emerged, with the producers outdoing each other. One of the first scenario studies for Germany in 2010 (Rahmann et al., 1997) was commissioned by the German Football Association. They traced scenarios for a decade following the tournament based on the number of World Cup venues (which were undecided at the time) and the spending behavior of the World Cup tourists. Until immediately before the World Cup, the authors were quoting a positive economic impact of €1.5 billion as their ‘best guess’. The German Hotel and Catering Association proposed a figure of €3.4 billion based on the assumption of 3.3 million foreign visitors, spending an average of €150 to €200 per day (Unterreiner, 2006). The Postbank, a major sponsor of the World Cup, was more upbeat, predicting an overall effect of €9 to 10 billion (or 0.5% of German GDP) (e.g. Postbank, 2005). Ahlert (2001), building on an assumption of constant spending by foreign visitors of approximately €1.8 billion and modeled under various scenarios for the level of state investment, the type of financing and possible displacement effects, calculated a positive net effect of a maximum of €7.8 billion.

357

Surveys of relevant decision-makers in relevant businesses were also conducted in Germany. A private marketing research agency expected a volume of investment of €5.5 billion. In addition, the German Association of Chambers of Industry and Commerce reported that 15% of the member enterprises replied that they expected positive effects for their enterprise; 83% expected no net effects, and 2% expected negative effects. The enterprises that expected positive effects identified additional demand by consumers/tourists, public spending contracts, improved infrastructure or other aspects (e.g. a better image for Germany) as the reasons for their positive expectations (Maennig & du Plessis, 2007). In contrast to these estimates based on ex-ante studies, ex-post analyses of World Cups were less optimistic. There were reports that the arrival of tourists tripled in June 2014 in Brazil, increasing from 350,000 in 2013 to 1,020,000 in 2014, and a new high of 6.4 million tourists to Brazil during 2014 (Armstrong, 2015). It is unclear whether these tourists induced the expected increase in Brazil’s income: ‘An army of Argentine fans poured over the border for the tournament, with thousands partying by the beach in Rio. Some landed in tiny camper vans and toured the country to see their heroes, travelling hundreds of miles during the five weeks of competition’ (Armstrong, 2015). There is no statistical analysis of World Cup 2014 labor market effects available, but it can be assumed that the expected 3.63 million jobs would be difficult to confirm. For the 2010 South Africa World Cup, du Plessis and Maennig (2011), using data on additional international plane landings, observed no evidence of a net increase in World Cup-related overseas tourism beyond approximately 90,000 to 118,000 persons, equivalent to a short-term impact of the tournament of 0.1% of GDP. Matheson, Peeters and Szymanski (2012) estimated an increase of 123,000 to 202,000 of tourist arrivals above what would have been expected without the World Cup. Concerning World Cup 2006 in Germany, Hagn and Maennig (2009) showed that the 2006 FIFA World Cup had no short-term employment effects. Feddersen, Grötzinger and Maennig (2009) failed to identify a better economic development in the host cities in the run-up of the World Cup, potentially by the investment in stadiums and related infrastructure. However, Kurscheidt, Preuß and Schütte (2008), using poll data, estimated an impact of the 2006 FIFA World Cup in Germany through substitution-adjusted consumer spending of €3.2 billion. In addition, Allmers and Maennig (2009) identified 700,000 additional overnight stays by non-residents and €570 million in net

358

THE SAGE HANDBOOK OF SPORTS ECONOMICS

German tourism income. Feddersen and Maennig (2012), on the basis of data that are regionalized and sectoralized, found that only the hospitality sector, and exclusively in the second quarter of 2006, enjoyed an employment increase of approximately 4.2%. This effect can be translated into approximately 2,600 additional jobs, which is far from the five-digit employment effects predicted in most ex-ante studies. They rejected the hypothesis of a long-term and persistent employment boost caused by the 2006 World Cup. Sobering results also exist for other World Cups. Szymanski (2002) collected data on the 20 largest economies measured by current GDP over the last 30 years. Many of these countries have hosted the Olympic Games or the World Cup at least once in the past 30 years. He concluded that the growth of these countries was significantly lower during World Cup years.2 Sterken (2006) found that World Cups have a positive effect but that this effect is limited. Hagn and Maennig (2008) showed that the 1974 World Cup, which was held in Germany, did not generate significant short- or long-term employment effects in that country. Baade and Matheson (2004) showed that, as a result of the 1994 World Cup in the USA, nine of the 13 host cities suffered a reduction in growth. Overall, the 13 locations suffered losses of over US$9 billion.

OLYMPIC GAMES For the Olympic Games, an equivalent gap between boosting ex-ante calculations and sobering ex-post analysis exists. Baade and Matheson (2016) provide an overview. Certain additions are necessary: In a recent ex-ante projection for the 2020 Tokyo Games, Osada, Ojima, Kurachi, Takuji and Kawamoto (2016) analyze foreign tourism as well as construction investment. The planned investment budget for the Tokyo Olympics equals 0.1–0.2% of Japanese GDP, and the number of foreign visitors to Japan is assumed to increase from 10 million in 2014 to 33 million visitors in 2020, with a growing per visitor expenditure also currently in action in 2016/2017. Using input-output (I-O) analysis, Japan’s real GDP level in 2018 is suggested to be approximately 1% higher than would otherwise be the case. In absolute numbers, this percentage is equivalent to 5–6 trillion yen (= US$45–54 billion at the 2017/31/03 exchange rate). Employment is estimated to increase by as much as 730.000 full-employed persons in 2018. The Tokyo Metropolitan Government (2017)

analyses the 2013–2030 period and calculates that induced Japanese production will increase by 323 trillion yen, with an employment equivalent of 1.9 million person-years. Again, ex-post analysis is less optimistic. Feddersen and Maennig (2013b) in a sectoral analysis of the Atlanta 1996 Games using monthly data, suggested an increase of 29,000 jobs, exclusively for the Atlanta Olympic month, exclusively in Fulton County and exclusively in a few specific sectors. Examining the 1996 Atlanta Olympic Games with the same yearly data as Hotchkiss, Moore and Zobay (2003), Feddersen and Maennig (2013a) were unable to reject the hypothesis that the 1996 Olympics had no significant impact on the employment figures. Testing the effects of the Olympic Games in Seoul in 1988, Barcelona in 1992, Sydney in 2000, and Beijing in 2008 on tourism and foreign exchange earnings with an ARIMA-model, Mitchell and Steward (2015) exclusively found negative Olympic impacts for the host countries, with the exception of a positive level shift of tourist numbers for South Korea. Two ex-post econometric studies are notable. Analyzing the Olympic Games from 1960 to 2012, Rose and Spiegel (2011) suggested a permanent export boost of 39% in Olympic host countries. Brückner and Pappa (2015) find positive effects on GDP of 1.74, 2.60, and 1.41 percentage points at three, four and five years preceding the Games, respectively. The cumulative effect on output from ten years before the Games to seven years after the Games is approximately 15%. Both studies argued that the Olympic effect may be attributable to a signal of trade liberalization and increased openness a country sends when bidding for the Games rather than the act of actually hosting the Games. It should first be noted that with such an interpretation, the titles of the papers of Rose and Spiegel, ‘The Olympic Effect’, and of Brückner and Pappa, ‘Olympic Games and Their Macroeconomic Effects’, may be misleading. It must also be noted that significant portions of the growth effects calculated by the two studies occur after the nomination to host the Olympics, making it difficult to distinguish between any signal effects and effects induced by the Olympics themselves. Maennig and Richter (2012) and Langer, Maennig and Richter (2017) demonstrated that the empirical findings of Rose and Spiegel as well as of Brückner and Pappa suffer from selection bias. When comparing Olympic host nations to matching countries, no significant effects on exports or GDP prevail. Finally, if, as Rose and Spiegel as well as of Brückner and Pappa are implicitly claiming, the

Major Events: Economic Impact

causality is postulated to run from (strategies for) growth of GDP (or other economic variables) to Olympic bids or hosting, this causality direction could be tested directly. Maennig and Vierhaus (2017) did so and found that the Olympic hosts are indeed characterized by larger markets and higher medium-term growth rates.

OTHER MAJOR EVENTS There are also ex-post studies on the economic effects of other major events or sporting activities. A few studies have found significant positive effects from sports facilities and sports events expost. Baim (1994) found positive employment effects from Major League Baseball (MLB) and National Football League (NFL) teams for 15 cities in the USA. Tu (2005) found significant positive effects from the FedEx Field (Washington) on real estate prices in the surrounding neighborhood, as did Ahlfeldt and Maennig (2009, 2010a, 2010b) for three arenas in Berlin, Germany. Carlino and Coulson (2004) examined the 60 largest Metropolitan Statistical Areas (MSA) in the USA and found that having an NFL team allowed the cities to enjoy rents that were 8% higher; however, they did not enjoy higher wages.3 In contrast, other studies, particularly those by Coates and Humphreys (1999, 2000, 2001, 2003a, 2003b) and Teigland (1999), have indicated significant negative effects. Porter (1999), Coates and Humphreys (2002), Baade, Baumann and Matheson (2008), Coates (2006), and Matheson (2005) analyzed the economic impact of the Super Bowl and find negligible positive effects at the maximum, as did Allmers and Maennig (2009) for Soccer European Cups. This list of ex-post studies on the ‘core’ economic effects of major events on income, employment and taxation may not be complete, also because of the difficulty to define a ‘major event’ (Coates & Depken, 2011; Maennig & Zimbalist, 2012a).

INTANGIBLE EFFECTS AND ‘NON-ECONOMIC’ EFFECTS In addition to ‘core’ economic effects, major events may induce intangible benefits for the host cities, host countries and their citizens. First, the Olympics could induce or at least hasten policy changes on international relations and labor

359

markets. Note that this argument is reversing the causality proffered by Rose and Spiegel (2011) and Brückner and Pappa (2015). These two studies on Tokyo 2020 demonstrate the argument. The suggested (and to a large degree previously fulfilled) increase in international tourism is also a result of an easing of visa requirements for sightseeing. In addition, the calculated increase in employment induced by the Olympics in the Japanese economy, which is suffering from labor shortages, will only be realized if the labor participation of women and the elderly will be increased. For example, of the 730,000 additional jobs in 2018, 230,000 shall be created in the construction sector. Note that unemployment in the Japanese construction sector was no more than 10,000 persons in 2014 (Osada et  al., 2016). In order to facilitate any substantial impact effect, openingup reforms might be necessary in Japan. Second, since 1992, the Barcelona case of urban regeneration when hosting the Olympics served as a franchising model. After this event, policy-makers of cities all over the world no longer applied for the Olympics because they wanted to host the best athletes of the world; rather, they applied because they wanted to position themselves to blackmail their national governments for billions of dollars for investments in infrastructure that otherwise would never have been built or that would have been built much later. Typically, underused locations (i.e. Barcelona coast line, Sydney Homebush Bay, London east end and Rio Barra zone) were used to locate stadius and/or housing and connected to other areas with transport facilities. The Olympic Games are often misunderstood as exercises in urban ‘strategic planning’. Master plans and zoning, which normally take decades to be decided and implemented in modern, multilayered democratic societies, may be accelerated; the Olympics provide an exceptional imperative for bypassing established procedures in urban regeneration and fast-track decisions, breakingup the perceived sclerotic democratic and juridical processes in urban planning in democratic and transparent societies. There are attempts to delineate the general impact of stadiums on urban development (Ahlfeldt & Maennig, 2010b) and to measure the external effects of stadiums (Tu, 2005; Ahlfeldt & Maennig, 2009, 2010a). Note that there is also minimal doubt regarding the positive effects of improvement in transport infrastructure tied to Olympic projects. Nevertheless, thus far, no comprehensive calculations of the value of the urban development acceleration are available. This unavailability may be because such Olympicinduced urban accelerations have a cost. Concerns are growing about ‘white elephant’ projects in

360

THE SAGE HANDBOOK OF SPORTS ECONOMICS

stadiums that deliver minimal value after the event, and the occasionally reported urban impoverishment of removed parts of the population. In a more general view, doubts are being raised of the potential of urban master planning: Are districts in the world’s cities with the highest quality of life the result of master planning? Alternatively, is urbanity the result of development in which a multiplicity of different actors, interests and initiatives participate? The perceived sclerotic urban development with multi-layered democratic decisions, including bottom-up participation potentially challenged by juridical decisions, evolved because civilians believed it to be a suitable idea. Third, enhanced international awareness (‘putting a city on the map’), better recognition, effects from nation-building or ‘integration’ effects, or simply the joy of being a good host and to experience ‘live’ the best athletes of the world may induce a benefit which is primarily called a ‘feelgood effect’ or ‘civic pride’. Such effects may be relevant before, during and after the major event and can, by the very nature of apparently being intangible, not be directly measured by the usual economic statistical time series. These effects are important for economists. The analysis of ‘welfare’ and ‘utility’, which may now be termed ‘life satisfaction’, was at the center of economics from its very beginning. Several means of delimiting ‘intangible’ effects have evolved. Using contingent valuation surveys, Atkinson, Mourato, Szymanski and Ozdemiroglu (2008) (and similar: Walton, Longo & Dawson, 2008) identified a willingness to pay (WTP) for the 2012 London Olympics by people from London and throughout Great Britain of £22 and c. £12 per year (for 10 years), summing to £2 billion in total. For the World Cup 2010 in Germany, an (ex-post) WTP for the induced integration and feel-good of €1 billion has been identified. This feel-good effect was one of the largest effects of the World Cup. Note that before the event, the WTP was substantially lower (Süssmuth, Heyne & Maennig, 2010), indicating that a major event may have the characteristic of an ‘experience good’. With such a characteristic, people may underestimate the benefits associated with the major events, which may cast doubt on the efficiency of public referenda (see Chapter 36 by Maennig on ‘Olympic Games: Public Referenda, Public Opinion and Willingness to Pay’ in this volume). In a second research strand, life satisfaction statistics are used. Kavetsos and Szymanski (2010) found a positive ‘feel-good’ factor associated with hosting football events but not with the Olympics. Dolan et  al. (2016) found a positive impact on the happiness of Londoners during the London 2012 Olympics. The magnitude of the effects is

equivalent to moving from bottom to the fourth income decile, but the effects were gone within a year. A third research strand uses data from social media. Du Plessis and Maennig (2011), for the World Cup South Africa 2010, used the number of Google hits and Facebook group members and find the largest increases in awareness for the 2010 FIFA World Cup tournament itself, which may suggest that only part of the awareness of major events is directed towards the host country. Fourth, recall interviews have been conducted to measure the public awareness of past Olympic host sites in both Europe and North America (Ritchie & Smith, 1991). Based on several thousand telephone interviews conducted from 1986 to 1989, less than 10% of the North American residents surveyed and less than 30% of the Europeans could recall that the 1976 Winter Olympic Games had been held in Innsbruck, Austria. Only 28% of the North Americans and 24% of the Europeans surveyed remembered that the 1980 Winter Games occurred in Lake Placid, New York. Other research showed that recognition of Calgary having hosted the 1988 Winter Games had nearly entirely faded by 1991. There may be other tangible, but not easily monetized, sporting or athletic effects: in most cases, there are additional medals for the athletes of the hosting region or nation due to the home advantage (Anders & Rothoff, 2014). A positive long-run athletic effect may also apply if, for the major event, national high-performance sport structures are renovated. Finally, the new stadiums may have a positive effect on spectator demand. For example, in Germany, the ‘novelty effect’ of all stadium projects since 1963 was equivalent to a rise in spectator numbers of approximately 10% per match (Feddersen, Maennig & Borcherding, 2006). In addition, the average revenue per ticket increased due to the expansion of the area for VIP and business seating; therefore, the overall ticket proceeds may increase more than proportionally. These increased receipts improve the ability of a club to acquire top players in the international market, which, in the medium term, leads to increased national and international sporting competitiveness.

EXPLAINING THE GAP BETWEEN EX-ANTE AND EX-POST ANALYSIS The gap between the boosting ex-ante views on income and employment and the sobering ex-post analysis deserves explanations. First, the

Major Events: Economic Impact

incentives of the two groups of authors may differ. Ex-ante analysis, primarily performed by consultancy firms, banks or similar, are paid to serve a public relations function in the bids for major events. Currently, despite the many obvious cases of no relevant income effects of major events, this group of authors believes that people can be convinced that a major event ‘pays off’ in core economic terms such as employment and income (Késenne, 2005). The ex-post authors, primarily academics striving for publication in peer-reviewed journals, may suffer from a generally pessimistically biased view of the activities of policy-makers, which, in their perception, are tempted to differentiate between a short-term vote maximization and a long-term welfareenhancing policy. Second, impacts that appear impressive in absolute terms are small in most statistically reported spatial areas. For example, the above-mentioned study by Kurscheidt et al. (2008), who estimated that the impact of the 2006 World Cup in Germany of €3.2 billion equaled only a small relative impact of 0.14% of Germany’s 2006 GDP. Nearly any positive impact of a mega-event will thus be lost within the normal fluctuations in the economy and, from a statistical perspective, will disappear in the white noise. This effect will be stronger as the data become more aggregated. However, note that the above-mentioned ex-post studies barely find effects that are significantly different from zero using data disaggregation as follows: (1) on a regional scale, (2) on the scale of the target variable, (3) on an industry scale, and (4) on a time scale or a combination of it. A discussion on the quality of the data may be necessary. Data on income, exports and tourism are revised on a permanent basis, occasionally decades after the relevant point in time. Implicitly, ex-post studies may suffer from the (low) quality of the data, and the results may be biased in both directions. There are a few studies that note this aspect (Langer et al., 2017). Note that the 700,000 additional overnight stays by non-residents for the World Cup 2006 in Germany identified by Allmers and Maennig (2009) disappear if the 2016 data set is used (details are available on request). Third, the employment effects claimed by exante impact studies cannot strictly be rejected by testing for significant differences from zero. The effects’ rejection would be possible if the postulated values were tested directly. For example, Baade and Matheson (2006) tested hypotheses both against a zero impact and against the impact claimed by sports promoters. The researchers were able to reject, at a 5% significance level, any promoters’ claims of an economic impact of more than $300 million from the game.

361

Hagn and Maennig (2008) concluded that a hypothesis that the 1974 World Cup in Germany caused an employment increase of 40,923 jobs in the average employment levels in the host cities in the period between 1974 and 1988 could not be rejected. Note that non-testing the zero-hypothesis evades from the current ‘Popper’ world of science and is thus unusual. In a most general view, economists tend to believe in equilibriums (at least in a well-ordered competition and in medium terms). The notion of compensating differentials is one of the most used applications of this economic belief in equilibriums. For major events, it translates into a perception that any of their potential benefits will be compensated by other negative effects, namely, a kind of cost. An obvious transmission mechanism is at hand (Maennig & Zimbalist, 2012b, p. 571): The International Olympic Committee (IOC), for instance, auctions off the right to host the Winter and Summer Games in a multi-stage competition. Prospective host cities/countries bid against each other to purchase a unitary product. The competitive bidding, in turn, leads the would-be hosts to bid up to the point where the expected marginal social utility of benefit from the Games equals the expected marginal social cost; or worse, if the winner, due to imperfect information, is subjected to a curse and bids beyond the benefit.

Put differently and less abstractly, if there are positive effects of major events, such as additional sporting success or feel-good effects, it is barely imaginable in a competitive environment that they will also be accompanied by positive economic effects; goods have a positive price. However, why an ex-ante calculation is very far from ex-post realities deserves attention. Technically, a lower than expected impact might have two reasons: first, the direct impact of the event is smaller than suggested by the event expenditures. Second, multipliers may be particularly low for major events, especially when compared to alternative public spending. Concerning the former, net direct spending must be analyzed (instead of gross spending). Beginning with the public spending on infrastructure on the occasion of major events, most public budgets are restricted either by the capital markets or by legal issues. In a growing number of countries, there are debt brakes that occasionally forbid additional debts. In such cases, the economic notion of opportunity costs becomes most relevant. Any public dollar spent on major events cannot be spent on other issues, namely, education and health. In many cases, the main argument of positive economic impacts may well cease here;

362

THE SAGE HANDBOOK OF SPORTS ECONOMICS

the net additional public spending is zero. In the few nations with no such obvious public finance restrictions, all limits to the efficiency of fiscal spending, which may be summarized under the headline ‘crowding-out’, apply (Crompton, 1995; Diedering & Kwiatkowski, 2015). Concerning private expenditures, economists agree that the consumption of local/national visitors for major events should not be included; household budget restrictions will induce a reduced consumption elsewhere. For the remaining pillar of incremental spending, namely, the spending of tourists (and athletes as well as officials) attending the major events, a crowding-out argument may apply here as well, as normal tourism may be reduced due to perceived increased congestion, noise, violence and crime. In addition, a price crowding-out may be in effect, as du Plessis and Maennig (2011) showed for the World Cup in South Africa in 2010: the above-­mentioned ex-ante studies indicating as much as half a million of additional overseas visitors may have induced a pricing behavior that damped the potential increase in the number of tourists. For example, prices for flights to South Africa during the 2010 FIFA World Cup period were as much as three times higher than normal for bookings for flights between the end of January and the end of March 2010. A similar observation could be made for hotel prices and for price quotations for car rentals. This finding is not an argument against the appropriate use of ex-ante impact studies. In contrast, ex-ante quantitative forecasts or measurements play an important role in sport event economics as in other economic areas. However, this finding is a warning that ex-ante economic impact studies with overly optimistic claims may induce self-defeating expectations. Concerning the second factor, ex-ante studies tend to use excessively large multipliers. For example, the studies include using sales instead of household multipliers, using incremental instead of normal multiplier coefficients, and using borrowed multiplier coefficients that may not apply to spending on major events (Crompton, 1995). Specifically, in the long run, the efficiency of sport facilities may be lower than for alternative public investments, which may be particularly true in the case of an (over-)supply of poorly adapted sport facilities with oversize, ‘wrong’ architecture and no integration into the urban structures (‘white elephants’). Such wasteful public spending may particularly occur if major events relate to human resource and management deficits. Finally, the diversion for major events may decrease productivity. The comparison with alternative, potentially more efficient uses of public spending clarifies

why impact analysis in the form of multiplier ana­ lysis or I-O models are principally criticized. These analyses in general calculate increasingly more impacts when more public money is spent. Cost– benefit analysis could be a solution (Késenne, 2005). Those costs should calculate impact/public spending ratios, not only for the mega-events but also for other potential areas of public spending, such as education and health. Solely if a public investment mega-event is proven to have a ‘return to investment’ ratio larger than other investment opportunities, they should be undertaken (Abelson, 2011). Because of the involved effort of a full consideration of the socially most useful potential projects, no such holistic cost– benefit analysis exists (Maennig, 1998). A ‘bounded rationality’ (Braybrooke & Lindblom, 1963) would be accepted by most economists if at least a minimum quota, as prescribed in certain national regulations, can be achieved.4 Finally, a discussion of computable general equilibrium models is necessary. As inferred by their name, the models were developed to analyze the long-term effects, particularly of permanent shocks. Major events are a short-term, transitory shock, which may only be analyzed satisfactorily by largely adapting the existing models (Abelson, 2011). With rational (i.e. well-informed) voters, impact studies in all the forms noted above may be irrelevant; if major events, after balancing the individual costs and benefits, were an efficient means of spending public money for citizens, a majority of the citizens would support them. However, majorities in many public referenda do not support the hosting of major events in their home cities (see Chapter 36 in this volume). If, nevertheless, a social efficiency remains postulated, it would imply an unequal distribution of net benefits; a supporting minority may perceive large average net benefits per capita, while the majority may suffer net costs, although potentially smaller per capita than the per capita gains of the supporters.

ENHANCING THE IMPACTS FOR MAJOR EVENTS The potentially unequal distribution may form a starting point for measures to enhance the impacts of major events (Maennig & Zimbalist, 2012b; Maennig, 2016). Major events attract the desire of policy-makers to host them in all parts of the world. The sporting institutions that own the major events have top ambitions (and positions) in

Major Events: Economic Impact

athletic and financial areas that could and should be mirrored by similarly strong ambitions to serve humankind more generally. With few efforts and resources, these institutions could use their events to enforce standard requirements for good governance, for labor regulations and minority protection, which have been defined by UN institutions, by declaring them as a precondition for being eligible to bid for their events. Nations with ‘deficits’ in these fields would need to change their structures simply to be able to apply. Once this change is completed, IOC and FIFA would contribute to enforcing the internationally agreed standards of Good Governance, which other institutions are unable to enforce. It is time for a more ambitious, truly world-leading self-awareness and selfesteem within these sporting organizations. Second, referenda and public participation may be introduced as a formal prerequisite by the leading sporting federations (see Chapter 36 in this volume). Third, the IOC and FIFA could choose a ‘pool of future hosts’ (Maennig, 2016), granting four of five cities/nations to be host in the near future. Pool members could invest on a secured basis with less time pressure, with the final selection of the host with respect to the current status of preparation within a sufficient time frame. Finally, public participation should also be included in the process of finding the leadership for the bid-organizing team. Until now, in nearly all cases, the selection process was limited to a small circle of decision-makers in a non-transparent process. In an excessive number of cases, the selection process exposed the enthroning of politically but in many cases less successful persons. Furthermore, there are very few (if any) known cases of bids/organization processes of major sporting events where the leading person was not required to be removed after a short time. From a historical perspective and in general, the selection processes of leadership personnel for bidding for major sporting events cannot be regarded as successful. A selection that includes a public participation process may well increase the quality (and acceptance) of the leadership team.

Notes 1  For example, §6(2) of the German «Haushaltsgrundsätzegesetz» (Budgetary Principles Act) prescribes: ‘Für alle finanzwirksamen Maßnahmen sind angemessene Wirtschaftlichkeitsuntersuchungen durchzuführen.’ (Appropriate economic feasibility studies must be undertaken for all financially relevant activities.) 2  No significant effects at all were registered for the Olympic Games.

363

3  In a comment, Coates, Humphreys and Zimbalist (2006) showed that these results are not robust, for example, to the exclusion of extreme outliers. 4  In earlier versions of the German Bundesverkehrswegeplan (Plan for Federal Transport Routes), a minimum ratio of 3:1 was required.

REFERENCES Abelson, P. (2011). Evaluating major events and avoiding the mercantilist fallacy. Economic Papers, 30(1), 48–59. Ahlert, G. (2001). The economic effects of the soccer World Cup in Germany with regard to different financing. Economic Systems Research, 13(1), 109–127. Ahlfeldt, G., & Maennig, W. (2009). Arenas, arena architecture and the impact on location desirability: The case of ‘Olympic Arenas’ in Prenzlauer Berg, Berlin. Urban Studies, 46(7), 1343–1362. Ahlfeldt, G., & Maennig, W. (2010a). Impact of sports arenas on land values: Evidence from Berlin. The Annals of Regional Science, 44(2), 205–227. Ahlfeldt, G., & Maennig, W. (2010b). Stadium architecture and urban development from the perspective of urban economics. International Journal of Urban and Regional Research, 34(3), 629–646. Allmers, S., & Maennig, W. (2009). Economic impacts of the FIFA Soccer World Cups in France 1998, Germany 2006, and outlook for South Africa 2010. Eastern Economic Journal, 35, 500–519. Anders, A., & Rotthoff, K. W. (2014). Is home-field advantage driven by the fans? Evidence from across the ocean. Applied Economics Letters, 21(16), 1165–1168. Armstrong, J. (2015). FIFA World Cup 2014 leads to record number of foreign visitors to Brazil. The Mirror, 20 July. Retrieved from: www.mirror.co.uk/ news/world-news/fifa-world-cup-2014-leads6106072 Atkinson, G., Mourato, S., Szymanski, S., & Ozdemiroglu, E. (2008). Are we willing to pay enough to ‘back the bid’? Valuing the intangible impacts of London’s bid to host the 2012 Summer Olympic Games. Urban Studies, 45(2), 419–444. Baade, R. A., Baumann, R. W., & Matheson, V. A. (2008). Selling the game: Estimating the economic impact of professional sports through taxable sales. Southern Economic Journal, 74, 794–810. Baade, R. A., & Matheson, V. A. (2004). The quest for the Cup: Assessing the economic impact of the World Cup. Regional Studies, 38(4), 341–352. Baade, R. A., & Matheson, V. A. (2006). Padding required: Assessing the economic impact of the Super Bowl. European Sport Management Quarterly, 6(4), 353–374.

364

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Baade, R. A., & Matheson, V. (2016). Going for the gold: The economics of the Olympics. Journal of Economic Perspectives, 30(2), 201–218. Baim, D. V. (1994). The sports stadium as a municipal investment. Westport, CT: Greenwood Press. Bohlmann, H. R., & van Heerden, J. H. (2005). The impact of hosting a major sport event on the South African economy. University of Pretoria, Department of Economics Working Paper Series, No. 2005–09. Braybrooke, D., & Lindblom, C. E. (1963). A strategy of decision: Policy evaluation as a social process. New York: The Free Press. Brückner, M., & Pappa, E. (2015). News shocks in the data: Olympic Games and their macroeconomic effects. Journal of Money, Credit and Banking, 47(7), 1339–1367. Carlino, G., & Coulson, N. E. (2004). Compensating differentials and the social benefits of the NFL. Journal of Urban Economics, 56(1), 25–50. Coates, D. (2006). The tax benefits of hosting the Super Bowl and the MLB all-star game: The Houston experience. International Journal of Sport Finance, 1(4), 239–252. Coates, D., & Depken, C. (2011). Mega-events: Is the Texas Baylor Game to Waco what the Super Bowl is to Houston? Journal of Sports Economics, 12(6), 599–620. Coates, D., & Humphreys, B. R. (1999). The growth effects of sport franchises, stadia, and arenas. Journal of Policy Analysis and Management, 18(4), 601–624. Coates, D., & Humphreys, B. R. (2000). The stadium gambit and local economic development. The Cato Review of Business and Government, 23(2), 15–20. Coates, D., & Humphreys, B. R. (2001). The economic consequences of professional sports strikes and lockouts. Southern Economic Journal, 67(3), 737–747. Coates, D., & Humphreys, B. R. (2002). The economic impact of postseason play in professional sports. Journal of Sports Economics, 3(3), 291–299. Coates, D., & Humphreys, B. R. (2003a). The effect of professional sports on earnings and employment in the services and retail sectors in US cities. Regional Science and Urban Economics, 33(2), 175–198. Coates, D., & Humphreys, B. R. (2003b). Professional sports facilities, franchises and urban economic development. Public Finance and Management, 3(3), 335–357. Coates, D., Humphreys, B. R., & Zimbalist, A. (2006). Compensating differentials and the social benefits of the NFL: A comment. Journal of Urban Economics, 60(1), 124–131. Crompton, J. L. (1995). Economic impact analysis of sports facilities and events: Eleven sources of

misapplication. Journal of Sport Management, 9(1), 14–35. Davies, L., Coleman, R., & Ramchandani, G. (2013). Evaluating event economic impact: Rigour versus reality? International Journal of Event and Festival Management, 4(1), 31–42. Diedering, M., & Kwiatkowski, G. (2015). Economic impact of events and festivals on host regions – methods in practice and potential sources of bias. Polish Journal of Sport Tourism, 22, 247–252. Dolan, P., Kavetsos, G., Krekel, C., Mavridis, D., Metcalfe, R., Senik, C., Szymanski, S., & Ziebarth, N. R. (2016). The host with the most? The effects of the Olympic Games on happiness. CEP Discussion Paper, No. 1441. Centre for Economic Performance, London School of Economics and Political Science. du Plessis, S., & Maennig, W. (2011). The 2010 FIFA World Cup high-frequency data economics: Effects on international tourism and awareness for South Africa. Development Southern Africa, 28(3), 349–365. Ernst & Young Terco (2011). Sustainable Brazil: Social and economic impacts of the 2014 World Cup. Retrieved March 2017, from: www.ey.com/ Publication/vwLUAssets/Brazil_World_Cup_impacts_ 2014_TZ_priloha/$FILE/Sustainable_Brazil_2014 %20World%20Cup%20impactsTZ%20priloha_.pdf Feddersen, A., Grötzinger, A. L., & Maennig, W. (2009). Investment in stadia and regional economic development – evidence from FIFA World Cup 2006. International Journal of Sport Finance, 4(4), 221–239. Feddersen, A., & Maennig, W. (2012). Sectoral labour market effects of the 2006 FIFA World Cup. Labour Economics, 19(6), 860–869. Feddersen, A., & Maennig, W. (2013a). Employment effects of the Olympic Games in Atlanta 1996 reconsidered. International Journal of Sport Finance, 8(2), 95–111. Feddersen, A., & Maennig, W. (2013b). Mega-events and sectoral employment: The case of the 1996 Olympic Games. Contemporary Economic Policy, 31(3), 580–603. Feddersen, A., Maennig, W., & Borcherding, M. (2006). The novelty effect of new football stadia: The case of Germany. International Journal of Sport Finance, 1(3), 174–188. Grant Thornton. (2004). 2010 Soccer World Cup facts you should know. Retrieved June 2006 from: www.gauteng.net/research/ pdf/soccer1.pdf Grant Thornton. (2008). The economic impact of the 2010 Soccer World Cup. Paper provided by email by Stan du Plessis from Stellenbosch University, 8 May 2008. Hagn, F., & Maennig, W. (2008). Employment effects of the Football World Cup 1974 in Germany. Labour Economics, 15(5), 1062–1075.

Major Events: Economic Impact

Hagn, F., & Maennig, W. (2009). Large sport events and unemployment: The case of the 2006 soccer World Cup in Germany. Applied Economics, 41(25), 3295–3302. Hotchkiss, J. L., Moore, R. E., & Zobay, S. M. (2003). Impact of the 1996 Summer Olympic Games on employment and wages in Georgia. Southern Economic Journal, 69(3), 691–704. Kavetsos, G., & Szymanski, S. (2010). National wellbeing and international sports events. Journal of Economic Psychology, 31(2), 158–171. Késenne, S. (2005). Do we need an economic impact study or a cost–benefit analysis of a sports event? European Sport Management Quarterly, 5(2), 133–142. Kurscheidt, M., Preuß, H., & Schütte, N. (2008). Konsuminduzierter Impakt von Sportgroßevents am Beispiel der Fußball-WM 2006. Wirtschaftspolitische Blätter, 55(1), 79–94. Langer, V. C. E., Maennig, W., & Richter, F. J. (2017). The Olympic Games as a news shock: Macroeconomic implications. Journal of Sports Economics. Online first, 1 March. https://doi.org/10.1177/ 1527002517690788 Maennig, W. (1998). Möglichkeiten und Grenzen von Kosten-Nutzen-Analysen im Sport. Sportwissenschaft, 28(3), 311–327. Maennig, W. (2016). Preventing corruption in the planning of major sporting events: Open issues. In Transparency International (Ed.), Global corruption report sport (pp. 169–173). Abingdon, UK, and New York: Routledge. Maennig, W., & du Plessis, S. (2007). World Cup 2010: South African economic perspectives and policy challenges informed by the experience of Germany 2006. Contemporary Economic Policy, 25(4), 578–590. Maennig, W., & Richter, F. J. (2012). Exports and Olympic Games: Is there a signal effect? Journal of Sports Economics, 13(6), 635–641. Maennig, W., & Vierhaus, C. (2017). Winning the Olympic host city election: Key success factors. Applied Economics, 49(31), 3086–3099. Online first, 16 December 2016. Retrieved from http://dx.doi.org/10.1080/00036846.2016. 1254339 Maennig, W., & Zimbalist, A. (2012a). What is a mega sporting event? In W. Maennig & A. Zimbalist (Eds.), International handbook on the economics of mega sporting events (pp. 9–14). Cheltenham, UK, and Northampton, MA: Edward Elgar. Maennig, W., & Zimbalist, A. (2012b). Future challenges: Maximizing the benefits and minimizing the costs. In W. Maennig & A. Zimbalist (Eds.), International handbook on the economics of mega sporting events (pp. 571–586). Cheltenham, UK, and Northampton, MA: Edward Elgar.

365

Matheson, V. A. (2005). Contrary evidence on the economic effect of the Super Bowl on the victorious city. Journal of Sports Economics, 6(4), 420–428. Matheson, V. A., Peeters, T., & Szymanski, S. (2012). Tourism and the 2010 World Cup: Lessons for developing countries. Journal of African Economies, 23(2), 290–320. Mitchell, H., & Stewart, M. F. (2015). What should you pay to host a party? An economic analysis of hosting sports mega-events. Applied Economics, 47(15), 1550–1561. Osada, M., Ojima, M., Kurachi, Y., Takuji, K., & Kawamoto, M. (2016). Economic impact of the Tokyo 2020 Olympic Games. Research and Statistics Department, Bank of Japan Reports & Research Papers. Retrieved March 2017 from: www.boj.or.jp/en/research/brp/ron_2016/data/ ron160121b.pdf Porter, P. (1999). Mega-events as municipal investments: A critique of impact analysis. In J. Fizel, E. Gustafson, & L. Hadley (Eds.), Sports economics: Current research (pp. 61–67). Westport, CT: Praeger. Porter, P., & Chin, D. (2012). Economic impact of sports events. In A. Zimbalist & W. Maennig (Eds.), International handbook on the economics of mega sporting events (pp. 246–264). Cheltenham, UK, and Northampton, MA: Edward Elgar. Postbank (2005). FIFA Fussball-Weltmeisterschaft 2006 – Deutsche Wirtschaft steht als Gewinner bereits fest. Retrieved November 2006 from: www.postbank.de/Datei/RSpezialFebruar05. pdf Rahmann, B., Weber, W., Groening, Y., Kurscheidt, M., Napp, H.-G., & Pauli, M. (1997). Sozio-ökonomische Analyse der Fußball-Weltmeisterschaft 2006 in Deutschland. Köln. Ritchie, J. R. B., & Smith, B. H. (1991). The impact of a mega-event on host region awareness: A longitudinal study. Journal of Travel Research, 30(1), 3–10. Rose, A. K., & Spiegel, M. M. (2011). The Olympic effect. The Economic Journal, 121(553), 652–677. Sterken, E. (2006). Growth impact of major sporting events. European Sport Management Quarterly, 6(4), 375–389. Süssmuth, B., Heyne, M., & Maennig, W. (2010). Induced civic pride and integration. Oxford Bulletin of Economics and Statistics, 72(2), 202–220. Szymanski, S. (2002). The economic impact of the World Cup. World Economics, 3(1), 169–177. Teigland, J. (1999). Mega-events and impacts on tourism: The predictions and realities of the Lillehammer Olympics. Impact Assessment and Project Appraisal, 17(4), 305–317.

366

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Tokyo Metropolitan Government (2017). Economic benefits of the Tokyo 2020 Olympic and Paralympic Games. Study handed over by Mr Nakamura (TOKYO OCOG) on 29 March 2017. Tu, C. C. (2005). How does a new sport stadium affect housing values? The case of FedEx Field. Land Economics, 81(3), 379–395.

Unterreiner, F. P. (2006). Fußball-WM und Olympia: Profit und Prestige für den Gastgeber. Frankfurter Allgemeine Zeitung, 10 March, 10–11. Walton, H., Longo, A., & Dawson, P. (2008). A contingent valuation of the 2012 London Olympic Games: A regional perspective. Journal of Sports Economics, 9(3), 304–317.

36 Olympic Games: Public Referenda, Public Opinion and Willingness to Pay Wolfgang Maennig

HISTORY OF OLYMPIC REFERENDA AND DETERMINANTS OF THE VOTING OUTCOME Public referenda on Olympic bids have drawn considerable interest in the recent past, also because – with few exceptions – the voters decided against the Olympic ambitions of the local governments and/or local bid committees. For example, within the last five years before the finalization of this text, voters in Vienna and Hamburg decided against bids for Olympic Summer Games in 2028 and 2024, respectively. Ambitions for Olympic Winter Games have been ended by referenda in Graubünden 2026 (Switzerland), Munich 2022 (Germany), St. Moritz/ Davos/Graubünden 2022, and Krakow 2022 (Poland). Tables 36.1 and 36.2 present data of the history of Olympic referenda and clarify that Olympic referenda are by no means an innovation of recent times: the oldest reported referenda are from the 1960s. All public referenda were held on Olympic Winter Games, with the exception of the two Summer Games referenda for Vienna 2024 and Hamburg 2024 mentioned above. There was a balance between positive and negative referenda for Olympic Games bids up to the 1970s.

Since the 1980s, the number of negative referenda has clearly outperformed the number of positive referenda. With the exception of the two positive referenda of Salt Lake City 2002 and Vancouver 2010 as well as the negative referendum of Denver 1976, all referenda took place in central European locations. The case of Denver 1976 is outstanding in another respect: while all other referenda were concerning Olympic bid campaigns, the Denver referendum took place long after the International Olympic Committee (IOC) had awarded the Games to Denver city, and only a little more than three years before the planned opening of the 1976 Olympic Winter Games (Foster, 1976). The IOC decided to shift the Games to Innsbruck, which also had organized the Winter Games of 1964 and had most of the sporting facilities at hand. Beyond that, it is hard to draw systematic conclusions from the descriptive data in Tables 36.1 and 36.2. For example, a positive or negative attitude does not seem to depend on the size of the bidding city. In addition, the tightness of the referenda results does not seem to be connected with the size of the bidding city or to the year of the referendum or to the sign (yes/no) of the result. In the only multivariate econometric study concerning Olympic referenda known by the author, Coates and Wicker (2015), analyzing the

Winter Games 2018

Winter Games 2010

Winter Games 2006

Winter Games 1998 and 2002

Winter Games 1976

Winter Games 1976

Munich (GER)

Vancouver (CAN)

Sion (CH)

Salt Lake City (USA)

St. Moritz/Graubünden (CH)

Sion (CH)

13.07.1969

09.11.1969

09.11.1989

08.06.1997

22.02.2003

08.05.2011

YES: 70,1% NO: 29,9% YES: 68,7% NO: 31,3%

YES: 64% NO: 36% YES: 67% NO: 33% YES: 57% NO: 43%

YES: 53,6% NO: 46,4% YES: 58,07% NO: 41,93%

Winter Games 2022

Oslo (NOR)

10.09.2013

Summer/Winter Games Date of Referendum Result of Referendum

City

Table 36.1  Positive referenda

41.81%

54%

59,64% in GarmischPartenkirchen (20918) 46% (293263)

66%

Turnout (Eligibles)

Burgener (1972)

www.valais-wallis-digital.ch/ de/a/#!/explore/cards/161 http://articles.latimes.com/198911-09/sports/sp-1649_1_saltlake-city Burgener (1972)

www.ssb.no/en/valg/statistikker/ folkavs_kostra/aar/2014-06-03 www.merkur.de/lokales/ garmisch-partenkirchen/ olympia-buergerentscheidderabstimmung-1234301.html Hiller and Wanner (2011)

Data source

No, Denver

No, Denver

1998: No, Nagano 2002: Yes

No, Turin

Yes

No, PyeongChang

No, Beijing

Successful bid?

368 THE SAGE HANDBOOK OF SPORTS ECONOMICS

10.11.2013

03.03.2013

22.09.2002

Winter Games 2026

Summer Games 2024

Winter Games 2022

Winter Games 2022

Winter Games 2006

Hamburg (GER)

Krakow (POL)

Munich (GER)

St. Moritz/Davos/ Winter Games 2022 Graubünden (CH)

Winter Games 2010

Graubünden (CH)

Bern (CH)

Innsbruck (A)

09.03.1997

25.05.2014

28.11.2015

12.02.2017

12.03.2013

Summer Games 2028

Vienna (A)

Date of Referendum

Summer/Winter Games

City

Table 36.2  Negative referenda

YES: 21,2% NO: 78,8% YES: 47,3% NO: 52,7%

Munich YES: 47,9% NO: 52,1% Garmisch YES: 48,44% NO: 51,56% Traunstein YES: 40,33% NO: 59,67% Berchtesgadener Land: YES: 45,9% NO: 54,1% YES: 47,3% NO: 52,7%

YES: 30,3% NO: 69,7%

YES: 48,40 % NO: 51,60 %

YES: 39,9% NO: 60,1%

YES: 28,06% NO: 71,94%

Result of Referendum

35,7% (80246)

59.14%

Munich: 28,8% Garmisch: 58,8% Transtein: 39,98% Berchtesgadener Land: 38,25 %

36%

50,2% (1,3 Mio)

51%

31,54% (1,15 Mio)

Turnout (Eligibles)

(Continued )

www.srf.ch/news/schweiz/abstimmungen/ abstimmungen-gr/buendner-sagen-nein-zuolympischen-spielen www.nzz.ch/sport/olympische-spiele-in-derschweiz-ld.149399 www.innsbruck.gv.at/data.cfm?vpath=redaktion/ ma_i/allgemeine_servicedienste/statistik/ dokumente38/wahlen2/innsbruckervolksbefragungenpdf

https://kurier.at/chronik/wien/volks befragung-in-wien-klares-nein-zuolympia/4.695.888 www.nzz.ch/schweiz/abstimmung-graubuendendas-olympia-flaemmchen-isterloschen-ld.145131 www.hamburg.de/pressearchiv-fhh/4655260/201512-15-bis-pm-olympia-referendumendgueltiges-ergebnis/ www.zeit.de/sport /2014-05/olympische-winterspiele-krakaureferendum www.olympia-nein.ch/go/aktuelles/ meldungen/0zu4.php

Data source OLYMPIC GAMES: PUBLIC REFERENDA, PUBLIC OPINION AND WILLINGNESS TO PAY 369

Summer/Winter Games

Winter Games 2002

Winter Games 1998

Winter Games 1994

Winter Games 1988

Winter Games 1976

Winter Games 1976

Winter Games 1976

Winter Games 1968

City

Innsbruck (A)

Aosta Valley (I)

Lausanne (CH)

Chur/Graubünden (CH)

Denver (USA)

Zürich (CH)

Interlaken/Bern (CH)

Sion/Wallis (CH)

08.12.1963

26.10.1969

02.11.1969

07.11.1972

02.03.1980

26.06.1988

19.06.1992

17.10.1993

Date of Referendum

Table 36.2  Negative referenda (Continued)

YES: 22% NO: 78% YES: 48,4% NO: 51,6% YES: 49,4% NO: 50,6%

YES: 40,56% NO: 59,44%

YES: 23% NO: 77%

YES: 15,3% NO: 84,7% YES: 38% NO: 62%

YES: 26,6% NO: 73,4%

Result of Referendum

45.80%

60.70%

45.20%

Turnout (Eligibles)

Burgener et al. (1972)

Burgener (1972)

https://serval.unil.ch/resource/ serval:BIB_7E35973521A1.P001/REF; http://doc. rero.ch/record/110462/files/1988-06-27.pdf www.srf.ch/news/schweiz/abstimmungen/ abstimmungen-gr/buendner-sagen-nein-zuolympischen-spielen https://ballotpedia.org/Colorado_Wint er_Olympic_Games_Funding_and_Tax,_ Measure_8_(1972) Burgener (1972)

www.innsbruck.gv.at/data.cfm?vpath=redaktion/ ma_i/allgemeine_servicedienste/statistik/ dokumente38/wahlen2/innsbruckervolksbefragungenpdf www.storiavda.it/novecento-2.html

Data source

370 THE SAGE HANDBOOK OF SPORTS ECONOMICS

OLYMPIC GAMES: PUBLIC REFERENDA, PUBLIC OPINION AND WILLINGNESS TO PAY

Munich 2018 referendum, try to shed light on the determinants of the voting outcome. They find a significantly negative effect of the share of leftist voters and of the number of hotel beds per capita on the Yes votes, but a positive impact of unemployment. A positive impact of liberal voters as well as a negative impact of the share of green voters did not turn out to be robust. In a robustness test, farm property tax per capita (+) as well as real property tax per capita (-) turned out to be significant, which is interpreted as an indication of a more positive attitude toward the Munich Olympic bid in rural areas than in urban areas. Other variables, such as gender, conservative or social democratic voters, and age structure of the population (for which the share of 18–64 year-olds serves as a proxy) did not have a significant impact. Further insights may be derived from earlier econometric studies of referenda on major (sport) infrastructure. Again, in a case study of the city of Munich, Ahlfeldt and Maennig (2012) find in a spatial analysis at the precinct level of the 2001 referendum of the Allianz Arena that voters in proximity of the proposed site opposed the project. At the city level and in proximity of alternative sites, voters supported the sports arena, indicating that residents expected net costs in proximity to stadiums and engaged in the referendum in order to shift the stadium away from their neighborhood. Sport facilities may thus exhibit a NIMBY (Not In My Backyard) character. A similar result was found in the case of the Seahawk stadium in Seattle, WA, USA, where support for the stadium was highest at 10–30 miles driving

371

distance from the stadium, and beyond that distance, voting support fell off (Horn, Cantor, & Fort, 2015). By contrast, Coates and Humphreys (2006) found net benefits of proximity to stadiums in Wisconsin, Texas and Florida, USA. Proximity costs and benefits of sports facilities apparently may thus vary across sports and countries, and the selection of the suggested locations may influence the outcome of referenda. The case of the negative referendum on Hamburg 2024 may illustrate the case. Figure 36.1 depicts the outcomes of the Hamburg 2024 referendum. The dark dot represents the projected Hamburg Olympic Center at Kleiner Grasbrook. Kleiner Grasbrook seems to be centrally located, but in the perception of many inhabitants of Hamburg, their city is divided between the richer areas in the north of the river Elbe (flowing from east to west, more or less halving the city) and southern areas, characterized by a population with lower incomes and higher shares of unemployment. Indeed, the Hamburg Olympic concept was regarded as part of the attempt (many decades long and, up to now, less successful) to regenerate the southern areas (‘Sprung über die Elbe’ – ‘Jump over the river Elbe’). Rejection of the Olympic ambitions was most prevalent in the densely populated precincts neighboring Kleiner Grasbrook as well as in the south of Hamburg. Indeed, the 10 voting places with the highest shares of No votes, ranging from 71.7% to 83.1% No votes, were located at distances of a maximum of some 2.5 kilometers from Kleiner Grasbrook. This evidence is in contrast to the findings of Coates and

Figure 36.1  Results of the referendum on the bid for the Olympic Games in Hamburg 2024 (29 November 2015). Share of yes/no votes in the different districts (Statistik-Nord, 2015)

372

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Wicker (2015) of a positive effect of higher unemployment rates on the support of Olympic bids. The case of the more pronounced resistance in the poorer southern parts of Hamburg, which were supposed to particularly benefit from the re-urbanizing Olympic concept, draws attention to the potential role of gentrification. Voters, particularly in the proximity of new Olympic facilities, may expect rising rents and real estate prices. Indeed, rising land values and property prices induced by sporting facilities have been identified by Ahlfeldt and Kavetsos (2014), Ahlfeldt and Maennig (2009, 2010), Carlino and Coulson (2004), Dehring, Depken, and Ward (2007), Feng and Humphreys (2012) and Tu (2005). According to the home voter hypothesis (Fischel, 2005), homeowners should be in favor of such stadium projects. Several studies suggest that projected house price capitalization effects significantly influence the degree of support for public initiatives and projects (Ahlfeldt, 2011; Brunner & Sonstelie, 2003; Brunner, Sonstelie, & Thayer, 2001; Dehring, Depken, & Ward, 2008; Hilber & Mayer, 2009). More generally, the literature on the political economy of housing markets suggests that a strong link exists among the nature of the political process, the ownership of land, and patterns of development (e.g., DiPasquale & Glaeser, 1999; Solé-Ollé & Viladecans-Marsal, 2013). However, in renter-dominated communities, it is worth differentiating explicitly between home voter and lease voter behavior in a public referendum. Ahlfeldt and Maennig (2015), on the occasion of a referendum on the closure of Tempelhof airport in the city of Berlin, found home voters to be more likely to support or oppose initiatives that positively or negatively affect the amenity value of a neighborhood because some of the related benefits or costs of lease voters are neutralized by adjustments in market rents. By contrast, Horn et  al. (2015) on the above-mentioned case of the Seahawk stadium in Seattle, USA, find little effect of the proportion of renters relative to homeowners. Apart from spatial determinants of support or resistance, Ahlfeldt and Maennig (2012) find that people aged 25–35 exhibited a relatively lower share of Yes votes for the new Munich football stadium, whereas the share of Yes votes is increased in precincts with a higher proportion of people aged 18–25 or 60 and above, or precincts that are characterized by a large share of male or unemployed population, while the opposite is true for precincts with a particularly high purchasing power per capita. Horn et al. (2015) also find an increased support for the new stadium in older demographic precincts. An – at least – nonnegative impact of the elderly on the outcomes of

referenda concerning sporting issues is also confirmed in a meta-study by Ahlfeldt, Maennig, and Steenbeck (2016), who nevertheless warn that the generational shift in most western democracies may heavily impact the implementation of additional larger infrastructures (see also Kotlikoff & Burns, 2005). In addition, on the occasion of the referendum on the Allianz Arena, Ahlfeldt, Maennig, and Ölschläger (2014) find that the preference for the professional football stadium is characteristic of substrata or middle strata lifestyle groups with a limited modernity orientation. Compared to established economic variables such as income, the application of indicators of value and strata orientation outperform the traditional indicators of economic wealth, in terms of capturing the spatial distribution of support and opposition. The authors highlight the importance of accounting for lifestyle compatibility as a criterion in choosing locations for (public) facilities with local costs (and benefits).

PUBLIC SUPPORT IN THE BIDDING PROCESS Public support is a criterion in the decision process of the International Olympic Committee (IOC), but there is an open debate on the extent of its importance. Hiller and Wanner (2017) find that public support ‘is not valued highly in the final decision’ of the IOC, but Maennig and Vierhaus (2017), concluding their multivariate econometric study of the chances of winning Olympic bids, find that support of at least 67% is crucial. For example, they argue that bids of New York 2012 (support of 59% in the Candidate city phase) and Tokyo 2016 (56%) may have failed due to the relatively low support. The Tokyo 2020 bid had the support of 70% and easily won. There are hardly any (published) results of polls on Olympic bids commissioned by local bid committees, local authorities or local media worldwide that show support of less than 50%. Note that support is almost always lower according to polls commissioned by the IOC compared to support according to polls commissioned by local bid committees, etc. (Hiller & Wanner, 2017). In such polls, there seems to exist some type of a time dependency, if not time inconsistency. Ritchie and Lyons (1990) undertook yearly surveys on the occasion of the Winter Games of Calgary 1988 and demonstrated that support grew from 84.7% in 1983 to some 88% in 1987, although support did not grow consistently. Most bidding cities

OLYMPIC GAMES: PUBLIC REFERENDA, PUBLIC OPINION AND WILLINGNESS TO PAY

experience decreasing support in the years before the IOC decision (Hiller & Wanner, 2017), but support may increase during the Games (Hiller & Wanner, 2011). Comparing before and after the Games, negative opinions on Olympic Games are considerably muted after the event (Hiller & Wanner, 2017). Ritchie and Lyons (1990) even found 97.8% support some weeks after the Games of Calgary 1988. These comparative findings exante versus ex-post are in line with findings for the World Cup 2006 in Germany. In this case, willingness to pay was significantly higher ex-post, indicating that a major event may have the characteristic of an ‘experience good’ (Süssmuth, Heyne, & Maennig, 2010). The reported degrees of support found by polls stand in certain contrast to the results in referenda depicted in Tables 36.1 and 36.2, at least for more recent Olympic bids. One striking case of such a gap between polls and the referendum results occurred in the case of the Olympic bid of Hamburg 2024: on the day of the referendum, public support of some 56% was published (N.N., 2015). In the official referendum results, however, the share of Yes votes was at 48.4%.1 There are also indications that leaders for politics and media may have a biased estimation on people’s opinion (Trosien, Pfeffel, & Ratz, 2016). Having mentioned the limited parallels of Olympic polls and referenda, it may nevertheless be indicative to monitor analysis of polls of Olympic Games. In the following, we also include studies that attempt to measure the willingness to pay (WTP) in addition to support/no-support polls. Opposition to the Olympic bids of Los Angeles (1984), Atlanta (1996), and Salt Lake City (2002) has been identified in an attempt to deflect or mitigate the negative consequences of a particular development project, but no indication of a general anti-growth movement could be identified (Burbank, Heying, & Andranovich, 2000). Note that there is little to no empirical proof for any significant economic impact of major events such as the Olympic Games (see Chapter 35 by Maennig on ‘Major Events: Economic Impact’, in this volume). However, perceptions of the local populations may well differ from ex-post realities, and they may well do so because of the boosting of expectations usually disseminated by the local bid committees and other local authorities. Positive intangible effects are usually the (only) significant effect of mega-sporting events found in academic studies, and their expectation may increase WTP (Wicker, Whitehead, Johnson, & Mason, 2015). A positive effect of higher incomes is found for the support as well as the willingness to pay for Olympic Games (Atkinson, Mourato,

373

Szymanski, & Ozdemiroglu, 2008; Coates & Szymanski, 2015; Hiller & Wanner, 2011; Preuss & Werkmann, 2011; Walton, Longo, & Dawson, 2008; Wicker et  al., 2015). Age is generally found to have no significant effect on WTP (e.g., Atkinson et al., 2008; Preuss & Werkmann, 2011; for an exception, see Walton et al., 2008). Concerning other socio-economic variables, attending free unticketed events and supporting the Liberal party were found to be positively significant for supporting the Vancouver 2010 Games, whereas educational levels, gender and age were of minor or less robust influence (Hiller & Wanner, 2011). Full-time employees had a lower WTP in the case of the London 2012 Games, while homemakers had a higher WTP (Walton et al., 2008). Females have a lower WTP (Coates & Szymanski, 2015; for mixed evidence, see Wicker et  al., 2015). Being from one of the finalist US cities for the 2024 Games bid (San Francisco, Los Angeles, Boston, and Washington) had no impact on WTP for Olympic Games in the United States (Coates & Szymanski, 2015). A general interest in sports was not found to have a significant effect on WTP for Olympic Games in Germany, while active participation in a sport does (Wicker et al., 2015).2

POLITICAL IMPLICATIONS In sum, distinct types of households derive different net utilities from Olympic Games and their legacies, such as stadiums (Coates & Humphreys, 2006). The net utilities may depend on the members’ preferences for the consumption benefits of sports and proximity costs (or benefits). Ahlfeldt and Maennig (2015) argue that in the long run, a political process based on referenda will not necessarily lead to the allocation of local public goods according to their welfare impact. If the perceived costs and benefits also depend on experiencing the Olympic Games, people may systematically underestimate the net benefits associated with the major events, which may cast doubt on the efficiency of public referenda (and public polls as an indicator for political desirability). The potential political implications are of a wide range. Referenda may be regarded as inferior to decisions based on social cost–benefit analyses, particularly those based on revealed preference approaches (Osborne & Turner, 2010). In a more extreme view, the ‘rule of knowers’ may be called for (Brennan, 2016). In an opposing attitude, the IOC, the Fédération Internationale de Football Association (FIFA)

374

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and other sporting institutions could declare exante referenda or other participation processes, adapted to the nations’ usance, as a precondition for bidding. The quality of the bids might increase: Interested cities/nations will need to invest more resources in developing bidding concepts that convince their own populations (and, consequently, the deciding bodies in the sporting federations). Having the majority backing them, the organizing will afterwards be smoothened. It may occur that more effort will be saved during the biding and organizing period than is invested in the pre-bidding period. The people’s participation and inclusion of the views of more milieus on urban development may well increase the perceived quality of the concepts. Particularly in the case of urban development, the competences of the formal elites in politics and administration are not accepted anymore to be superior by a growing number of milieus. The ideas upcoming in such participation processes may well induce painful information on the weak points of major sporting events. For example, there may be requests to plan Olympic facilities in a way that makes them useable by sports for all afterwards. Compensating measures may not be asked exclusively in the ecological sphere but also in social respects, for example, in order to face the fear of increasing rents and real estate prices by many persons. There may even be demands for a downsizing of the Games, and there may be demands for a private financing of such major sporting events, not further stressing public finances (Maennig, 2016). This method of financing would bring the Games much closer to its roots as a sporting event rather than an occasion for urban regeneration. In the end, more participation may strengthen the bidding concepts for major sporting events.

Notes 1  Such striking gaps between poll data and voting results have occurred on many other occasions recently, with the most prominent cases being the BREXIT and the 2016 US presidential elections (e.g., Chalabi, 2016; Mercer, Deane & McGeeney, 2016). Possible causes for this gap may lie, for example, in the personal reluctance to speak up against the officially propagated activity that is at the same time often perceived as the majority of the public opinion (Rothschild & Malhotra, 2014). Further, the powerful tool of a vote as an expression of discontent with the current situation – not necessarily related to the voting topic – is not used very often. A further aspect may be a selection bias in the opinion

polls, as most of them try to represent all eligible voters in their results (Wang, Rothschild, Goel, & Gelman, 2015). However, there are crucial differences between eligible voters and actual voters in general (Freedman & Goldstein, 1996; Gelman & King, 1993; Petrocik, 1991; Rothschild & Malhotra, 2014), most of them due to age (e.g., Gutsche, Kapteyn, Meijer, & Weerman, 2014; Keeter, 2006), educational level (Reedy, Gastil, & Moy, 2016; Rosenstone & Hansen, 1993, among others), and employment status (Gutsche et  al., 2014). Homeowners tend to have higher voter participation (Brunner, Sonstelie, & Thayer, 2001). Further, referenda may be a powerful tool as an expression of discontent with the current political situation. 2  For more information about surveys on Olympic Games, see Deccio and Baloglu (2002), Guala (2009), Mihalik (2001, 2003), Preuss and Solberg (2006), and Scamuzzi (2006).

REFERENCES Ahlfeldt, G. (2011). Blessing or curse? Appreciation, amenities and resistance to urban renewal. Regional Science and Urban Economics, 41, 32–34. Ahlfeldt, G., & Kavetsos, G. (2014). Form or function? The effect of new sports stadia on property prices in London. Journal of the Royal Statistical Society Series A (Statistics in Society), 177(1), 169–190. Ahlfeldt, G., & Maennig, W. (2009). Arenas, arena architecture and the impact on location desirability: The case of ‘Olympic Arenas’ in Prenzlauer Berg, Berlin. Urban Studies, 46(7), 1343–1362. Ahlfeldt, G., & Maennig, W. (2010). Impact of sports arenas on land values: Evidence from Berlin. The Annals of Regional Science, 44(2), 205–227. Ahlfeldt, G., & Maennig, W. (2012). Voting on a NIMBY facility: Proximity cost of an ‘Iconic’ Stadium. Urban Affairs Review, 48(2), 205–237. Ahlfeldt, G., & Maennig, W. (2015). Homevoters vs. leasevoters: A spatial analysis of airport effects. Journal of Urban Economics, 87, 85–99. Ahlfeldt, G., Maennig, W., & Ölschläger, M. (2014). Measuring and quantifying lifestyles and their impact on public choices: The case of professional football in Munich. Journal of Economic and Social Measurement, 39(1–2), 59–86. Ahlfeldt, G., Maennig, W., & Steenbeck, M. (2016). Après Nous le Déluge? Direct democracy and intergenerational conflicts in ageing societies. CESIFO Working Paper No. 5779. ISSN 2364–1428. Atkinson, G., Mourato, S., Szymanski, S., & Ozdemiroglu, E. (2008). Are we willing to pay enough to

OLYMPIC GAMES: PUBLIC REFERENDA, PUBLIC OPINION AND WILLINGNESS TO PAY

‘back the bid’? Valuing the intangible impacts of London’s bid to host the 2012 Summer Olympic Games. Urban Studies, 45, 419–444. Brennan, J. (2016). Against democracy. Princeton, NJ: Princeton University Press. Brunner, E., & Sonstelie, J. (2003). Homeowners, property values, and the political economy of the school voucher. Journal of Urban Economics, 54(2), 239–257. Brunner, E., Sonstelie, J., & Thayer, M. (2001). Capitalization and the voucher: An analysis of precinct returns from California’s proposition 174. Journal of Urban Economics, 50(3), 517–536. Burbank, M. J., Heying, C. H., & Andranovich, G. (2000). Antigrowth politics or piecemeal resistance? Citizen opposition to Olympic-related economic growth. Urban Affairs Review, 35(3), 334–357. Burgener, L. (1972). Quatre référendums suisses sur les Jeux Olympiques. Revue économique et sociale: bulletin de la Société d’Etudes Economiques et Sociales, 30, 2. Carlino, G., & Coulson, N. E. (2004). Compensating differentials and the social benefits of the NFL. Journal of Urban Economics, 56, 25–50. Chalabi, M. (2016). Why were the election polls so wrong? How Donald Trump defied predictions. The Guardian, 9 November. Retrieved from: www. theguardian.com/us-news/2016/nov/09/donaldtrump-exit-polls-data-us-election on 29 May 2017. Coates, D., & Humphreys, B. R. (2006). Proximity benefits and voting on stadium and arena subsidies. Journal of Urban Economics, 59, 285–299. Coates, D., & Szymanski, S. (2015). Willingness to pay to host the Summer Olympic Games. Working Paper, University of Maryland, Baltimore, MD. Coates, D., & Wicker, P. (2015). Why were voters against the 2022 Munich Winter Olympics in a referendum? International Journal of Sport Finance, 10(3), 267–283. Deccio, C., & Baloglu, S. (2002). Nonhost community resident reactions to the 2002 Winter Olympics: The spillover impacts. Journal of Travel Research, 41(1), 46–56. Dehring, C. A., Depken, C. A., & Ward, M. R. (2007). The impact of stadium announcements on residential property values: Evidence from a natural experiment in Dallas-Fort Worth. Contemporary Economic Policy, 25(4), 627–638. Dehring, C. A., Depken, C. A., & Ward, M. R. (2008). A direct test of the homevoter hypothesis. Journal of Urban Economics, 64(1), 155–170. DiPasquale, D., & Glaeser, E. L. (1999). Incentives and social capital: Are homeowners better citizens? Journal of Urban Economics, 45(2), 354–384. Feng, X., & Humphreys, B. R. (2012). The impact of professional sports facilities on housing values: Evidence from census block group data. City, Culture and Society, 3, 189–200.

375

Fischel, W. A. (2005). The homevoter hypothesis. Cambridge, MA: Harvard University Press. Foster, M. (1976). Colorado’s defeat of the 1976 Winter Olympics. Colorado Magazine, 53, 163–186. Freedman, P., & Goldstein, K. (1996). Building a probable electorate from preelection polls. Public Opinion Quarterly, 60, 574–587. Gelman, A., & King, G. (1993). Why are American Presidential election campaign polls so variable when votes are so predictable? British Journal of Political Science, 23, 409–451. Guala, C. (2009). To bid or not to bid: Public opinion before and after the Games: The case of Turin. In J. Kennell, C. Bladon, & E. Booth (Eds.), The Olympic legacy: People, place, enterprise (pp. 21–30). London: University of Greenwich. Gutsche, T. L., Kapteyn, A., Meijer, E., & Weerman, B. (2014). The RAND continuous 2012 presidential election poll. Public Opinion Quarterly, 78, 233–254. Hilber, C. A. L., & Mayer, C. J. (2009). Why do households without children support local public schools? Journal of Urban Economics, 65(1), 74–90. Hiller, H. H., & Wanner, R. A. (2011). Public opinion in host Olympic cities: The case of the 2010 Vancouver Winter Games. Sociology, 45, 883–899. Hiller, H. H., & Wanner, R. A. (2017). Public opinion in Olympic cities: From bidding to retrospection. Urban Affairs Review, 1–32, DOI: 10.1177/ 1078087416684036 Horn, B. P., Cantor, M., & Fort, R. (2015). Proximity and voting for professional sporting stadiums: The pattern of support for the seahawk stadium referendum. Contemporary Economic Policy, 33(4), 678–688. Keeter, S. (2006). The impact of cell phone noncoverage bias on polling in the 2004 Presidential election. Public Opinion Quarterly, 70, 88–98. Kotlikoff, L. J., & Burns, S. (2005). The coming generational storm: What you need to know about America’s economic future. Cambridge, MA: MIT Press. Maennig, W. (2016). Preventing corruption in the planning of major sporting events: Open issues. In Transparency International (Ed.), Global Corruption Report Sport (pp. 169–173). London: Transparency International. Maennig, W., & Vierhaus, C. (2017). Winning the Olympic host city election: Key success factors. Applied Economics, 49(31), 3086–3099. Mercer, A., Deane, C., & McGeeney, K. (2016). Why 2016 election polls missed their mark. Hg. v. Pew Research Center. Retrieved from: www. pewresearch.org/fact-tank/2016/11/09/why2016-election-polls-missed-their-mark/ on 9 November 2016.

376

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Mihalik, B. (2001). Host population perceptions of the 1996 Atlanta Olympics: Attendance, support, benefits and liabilities. Tourism Analysis, 5(1), 49–53. Mihalik, B. (2003). Host population perceptions towards the 1996 Atlanta Olympics: Benefits and liabilities. In M. De Moragas, C. Kennett, & N. Puig (Eds.), The legacy of the Olympic Games 1984–2000 (pp. 339– 345). Lausanne: International Olympic Committee. N.N. (2015). ZDF-Prognose: 56 Prozent für Olympia in Hamburg [ZDF-Forecast: 56 percent in favor of Olympics in Hamburg]. Retrieved from: www. sueddeutsche.de/news/politik/sportpolitik-zdfprognose56-prozent-fuer-olympia-in-hamburg-dpa.urnnewsml-dpa-com-20090101-151129-99-01886 Osborne, M. J., & Turner, M. A. (2010). Cost benefit analyses versus referenda. Journal of Political Economy, 118(1), 156–187. Petrocik, J. R. (1991). An algorithm for estimating turnout as a guide to predicting elections. Public Opinion Quarterly, 55, 643–647. Preuss, H., & Solberg, H. (2006). Attracting major sporting events: The role of local residents. European Sport Management Quarterly, 6(4), 391–411. Preuss, H., & Werkmann, K. (2011). Erlebniswert Olympischer Winterspiele in München 2018. Sport und Gesellschaft, 8(2), 97–123. Reedy, J., Gastil, J., & Moy, P. (2016). From the secret ballot to the public vote: Examining voters’ experience of political discussion in vote-by-elections. Political Communication, 33, 39–58. Ritchie, J. R. B., & Lyons, M. (1990). Olympulse VI: A post-event assessment of resident reaction to the XV Olympic Winter Games. Journal of Travel Research, 28(3), 14–23. Rosenstone, S. J., & Hansen, J. M. (1993). Mobilization, participation, and democracy in America. New York: Macmillan. Rothschild, D., & Malhotra, N. (2014). Are public opinion polls self-fulfilling prophecies? Research and Politics, 1(2), 1–10.

Scamuzzi, S. (2006). Winter Olympic Games 2006 in Turin. In N. Müller, M. Messing, & H. Preuß (Eds.), Von Chamonix bis Turin. Kassel: Agon Sportverlag. Solé-Ollé, A., & Viladecans-Marsal, E. (2013). Do political parties matter for local land use policies? Journal of Urban Economics, 78, 42–56. Statistik-Nord (2015). Olympia-Referendum in Hamburg am 29. November 2015 – Ergebnisse. Statistisches Amt für Hamburg und Schleswig-Holstein. Retrieved November 2017 from: www.statistiknord.de/fileadmin/maps/election_2015_hh_ref2/ print.html?0|0,10000,0,9643.807,0|th0=1,th0-0=1, th0-1=0,th0-2=0,th1=0,th1-0=1,th1-1=0,th12=0,th1-3=0,th2=1,th2-0=1,th2-0-0=1|||th00xth0xxth1-0xth1 Süssmuth, B., Heyne, M., & Maennig, W. (2010). Induced civic pride and integration. Oxford Bulletin of Economics and Statistics, 72(2), 202–220. Trosien, G., Pfeffel, F., & Ratz, M. (2016). Governance and importance of cooperation of metropolitan regions: Learnings from the failed Olympic bid Hamburg 2024. Paper presented at SESM Conference, Berlin, 19 May. Tu, C. C. (2005). How does a new sports stadium affect housing values? The case of FedEx Field. Land Economics, 81, 379–395. Walton, H., Longo, A., & Dawson, P. (2008). A contingent valuation of the 2012 London Olympic Games: A regional perspective. Journal of Sports Economics, 9, 304–317. Wang, W., Rothschild, D., Goel, S., & Gelman, A. (2015). Forecasting elections with non-representative polls. International Journal of Forecasting, 31, 980–991. Wicker, P., Whitehead, J. C., Johnson, B. K., & Mason, D. S. (2015). Public support for hosting the Olympic Summer Games in Germany: The CVM approach. Department of Economics Working Paper No. 15–06, Appalachian State University, http://econ.appstate.edu/RePEc/pdf/wp1506.pdf.

37 Olympic Performance Eva Marikova Leeds

INTRODUCTION The Olympic Games are an enchanting spectacle. Despite the recent controversies surrounding them, the worldwide audience and the associated revenues remain strong. Viewers follow the Games for intrinsic and extrinsic reasons. They want to admire the beauty and wonder of the athletic performance as well as the resulting rankings. They want to see which records will be broken and which nation’s athletes will dominate. This chapter focuses on national performance measured by medal counts. The topic of Olympic national performance emerged in the economics literature decades after appearing in the sociology literature, very much in tandem with the development of econometric techniques. As new techniques arise, they offer researchers a new way to address old questions. Researchers start looking at the world through a new lens as they return to old questions and pose new ones: Why do some countries regularly collect so many medals? How does the political regime contribute to the bounty? Why do some countries fare relatively better in the Winter Olympics than the Summer Olympics? Why are they better in some disciplines than in others? What is the medal yield from hosting the Games? How many medals can a country expect to win at the next Olympics?

The next section provides a brief background and reviews papers that use regression analysis, but not stochastic frontier analysis. It presents their main findings, as they address the above questions. The papers apply different estimation techniques to different numbers of explanatory variables using different data sets. Over time, researchers have expanded their data sets and explanatory variables, and they started to use newer estimation methods. The rising ability to forecast the number of medals has led to a regional focus on Olympics performance. The last section summarizes the findings and suggests avenues for further research.

BACKGROUND AND LITERATURE REVIEW If we asked a layman how to predict the number of medals a nation would win at the Olympic Games, she would probably respond that one would have to know the size and quality of each nation’s delegation. The larger the delegation and the better its athletes, the more medals the country is likely to win. Until recently, the International Olympic Committee (IOC) had a two-tier standard for athletes to qualify for entry. The size of the delegation can be easily ascertained, and the

378

THE SAGE HANDBOOK OF SPORTS ECONOMICS

quality quickly gleaned from the proportion of athletes who qualify for the Olympics under the higher standard. Because the IOC wants all countries to send a delegation, some countries send athletes who qualify under a lower standard and have virtually no chance of winning a medal. Taking this method to an extreme, one would forecast the winners of each event, just like Sports Illustrated does, and add the resulting medals for each country. If this procedure were used statistically to generate forecasts, it would suffer from what is recognized as a dominant variable problem, where the dependent variable is effectively a transformation of the independent variable. Even using the size and quality of each delegation to predict the number of medals, does not generate perfect forecasts because of the presence of two confounding factors, team events and multiple events. Team sports garner medals very inefficiently because for ranking purposes they award a single medal to multiple athletes. In contrast, some athletic categories have many events in which a single athlete may win medals. This is most pronounced for swimming and gymnastics. Economists view the world differently, as they look for underlying determinants of success. Athletic performance is a function of dedicated resources. The honing of talented athletes’ skills is costly. This cost, and the accompanying expenditure, is borne by many parties, including the athletes themselves, the clubs to which they belong, and the nations of which they are citizens. However, it is virtually impossible to gather statistics on the personal resources that athletes devote to athletics. It is equally unrealistic to survey the resources extended by clubs. Finally, no researcher has tried to summarize the expenses of national organizations, such as National Olympics Committees. The beauty of the economics literature surveyed in this chapter is that it can sidestep both the ambiguous connection between the number of participants and medals as well as the unavailability of funding data. Instead, it explains Olympic performance based on a few easily accessible macroeconomic variables. The theoretical underpinning of the literature lies in a simple production function. Medals are produced using capital and labor, proxied by income and population; the regression approach applied below is implicitly based on the assumption that this function is universal. Total factor productivity, in turn, depends on additional variables, such as the political regime or climate of the country. Only some papers (e.g. Bernard & Busse, 2004; Leeds & Leeds, 2012; and Noland & Stahler, 2017) explicitly model this underlying production function. The lack of a more specific model allows different authors to use different

variables and introduces an element of arbitrariness to the enterprise. While researchers look for a stable link between income and population and other variables, institutional details ‘muddy the waters.’ To ensure broad international participation, the IOC allows countries without any qualifying athletes to bring a small delegation anyway. It also limits the number of qualifying athletes in each event to three per nation in the hope of spreading the medal bounty and raising international interest in the Games. More recently, globalization and the accompanying migration of athletes, noted by Lui and Suen (2008), allow countries to leapfrog local talent development by importing world-class athletes to earn medals irrespective of income and population. Researchers have been long aware that national output and population affect Olympic performance. Ball (1972), when examining the outcome of the 1964 Olympics, remarks that ‘Big and rich countries should be more successful than little and poor nations just because they are big and rich’ (p. 189). Grimes, Kelly, and Rubin (1974) add a political regime variable. They apply a tobit analysis to the results of the 1972 Olympics using per capita GNP, population, and population interacted with a Communist dummy variable. The results allow them to make a respectable in-sample prediction for the number of medals for each country and thus encourage further exploration. Baimbridge (1998) is the first Olympic performance paper to appear in an economics journal, but it does not focus on national performance per se. Baimbridge notes that everything about the Olympics has grown since their rebirth in 1896, but one variable stands in contrast to this broad trend. The proportion of countries that win any medals has decreased, as the number of countries that participate at the Olympics has grown to include effectively all countries. He interprets the lower aggregate proportion as the probability that a specific nation would win a medal and as greater uncertainty of outcome. Because each Summer Olympiad is the unit of observation, the sample of 23 is too small for any meaningful inference, but the ordinary least squares (OLS) regression indicates a negative correlation of the proportion of medal-winning countries with the number of participants per sporting event and a positive correlation with the number of participants per nation. The time trend has the statistically strongest, negative impact. Baimbridge also includes three dummies for Olympic boycotts, but only the 1984 dummy has an impact. It is negative because of the number of perennially medal-winning countries that boycotted the Los Angeles Games was very large.

Olympic Performance

After Baimbridge (1998), the literature turns back to examining national success. Hoffmann, Ging, and Ramasamy (2002), like Baimbridge, have a small sample, specifically of medal-winning countries at the 2000 Sydney Olympics. Using OLS, they bring back income and population, both of which have the expected positive impact on the number of medals won by each country. This study addresses many questions of interest in this literature. The regression contains a dummy for countries with former or current Communist regimes, which are assumed to have the power to marshal resources to desired ends; two dummies for whether a country hosted one or two Olympic Games, to indicate a proclivity for sports; and a dummy for the current host, to indicate homecourt advantage. All variables have a statistically significant impact. Two additional different regressions contain only climate variables to control for exogenous influence on training, which often starts informally outdoors. Tcha and Pershin (2003) ask a different question: Why do some countries dominate in certain sports? For each of six broad categories, such as swimming and gymnastics, they compute a nation’s revealed comparative advantage (RCA) as the ratio of the percentage of medals a country won in that sport to the percentage of all medals awarded in the sport. The ratios are constructed using data from the 1988, 1992, and 1996 Olympics for 66 countries. For each of the categories, Tcha and Pershin regress the RCAs on three sets of variables: one containing economic performance (GDP, population, and GDP per capita); one with geographic variables, for example, coastline length; and a third with three dummy variables for Asia, Africa, and a former Communist country. The authors transform the independent variables in two ways. Using a tobit regression, they consider the variables as proportions as well as their ranks, and they obtain very different results. For example, in the first version, only population affects the RCA for ballgames. In the second version, RCA is affected positively by landmass, GDP, and the African dummy and negatively by GDP per capita, coast length, and altitude. While the Communist dummy does not affect ballgames, it affects other RCAs, for example, gymnastics. Because it is hard to model the connection between a specific sport category and macroeconomic variables, this strand of the literature has not progressed much beyond this paper. In the second half of their paper, Noland and Stahler (2017) focus on individual events, referencing Otamendi and Doncel (2014). They calculate the HerfindahlHirschman Index (HHI) index of a country’s medal count for all individual sports, 26 summer categories for 1960–2012 and seven winter categories

379

for 1960–2010. They also report the top medalwinning country for each category, with Canada, for example, dominant in curling for both men and women. Because of data limitations, they explain only three categories (aquatics, athletics, and gymnastics) for the entire 1960–2012 period and 11 categories (including aquatics, athletics, and gymnastics) for 1996–2012 for the Summer Games. I highlight four of their many conclusions. In the Summer Games, population positively affects all sport categories. Second, the effect of hosting depends in part on whether an event is judged; for example, it is negative for fencing (see also Balmer, Nevill, & Williams, 2003). The authors, however, do not entertain the possibility that such a correlation may be spurious. They indirectly allow for spurious correlation between GDP per capita and gymnastics performance by noting that the Communist countries, which were poor, specialized in gymnastics. Finally, the effect of the Communist regime differs for men and women for many sports, including gymnastics. The Communist dummy has no effect for men, and for women it flips from positive for the entire period to negative for the shorter period. Contrasting their results with Balmer, Nevill, and Williams (2001), for the Winter Games they find that: Being from the communist bloc has aided female athletes more in skating, but male athletes more in skiing. Unlike the results from the Summer Games, where the home bias helps in judged competitions, there appears to be no host advantage in men’s and women’s skating. (p. 528)

Given the strong likelihood of spurious correlation, the sport-specific results should be interpreted with much caution. The two most prominent papers in this literature were written over the same time span. Bernard and Busse (2000) appeared as an NBER working paper and received national media attention, as it successfully predicted the Sydney Olympics medal count (see NPR, 2000) ‘for the 36 countries that won at least five medals in 1996’ (Bernard & Busse, 2004, p. 417). Johnson and Ali (2000) was not a formal working paper, but both papers were published four years later. Bernard and Busse (2004) is cited more frequently, although Johnson and Ali (2004) is more comprehensive. Bernard and Busse (2004) present their work as a pioneering effort and do not contrast their results with anyone else’s, despite acknowledging that the topic of country success has appeared in the literature. They use two basic economic variables, logs of GDP per capita and population, to explain the proportion of medals countries won between 1960 and 1996 using panel tobit regression.

380

THE SAGE HANDBOOK OF SPORTS ECONOMICS

In addition to this frugal formulation, Bernard and Busse (2004) use several fuller versions, in which the proportion of medals also depends on two Communist regime dummies (‘planned’ for ‘China, Albania, Yugoslavia (through 1988), and North Korea’ (p. 415) and ‘Soviet’ for former Warsaw-Pact countries), a dummy for hosting the Olympics and two boycott dummies. Soviet countries earn about 6% more medals and planned economies about 1.5% more. The estimated effect clearly depends on the definition of a regime. The effect of hosting ranges from 1% to 6%, depending on the specification. To make their predictions, however, Bernard and Busse add a lagged dependent variable; without it, the model predicts negative medals for most countries winning none. Johnson and Ali (2004) cast a wider net. They are the first to brave the world of the Winter Olympics in addition to the Summer Olympics, and they examine total participation in addition to the total medal count, but they do not use participation as an explanatory variable for success. Moreover, they examine female participation as well as the determination of the gold medal count, thus establishing all of the categories of future interest in this literature. (They also estimate individual sport regressions but do not report their results.) They use data from all Olympics between 1952 and 2000 and report their OLS results with country fixed effects. In addition to linear and squared income, linear and squared population, and regime variables, they specify a host and hostneighbor dummy and two climate variables, one for the share of land that experiences light frost and one for heavy frost. They explain participation and medal count for the Summer and Winter Olympics separately and together. The identical specification allows for a sideby-side comparison of the Summer and Winter Olympics. The most striking result is the different impact of hosting the Games, which adds 32 participants in the winter and 210 participants in the summer. While we expect cold countries to send relatively larger delegations to the Winter Olympics (they send 53 more athletes), it is surprising that even after controlling for income, heavy-frost countries have 33 more participants at the Summer Olympics. This is another example of spurious correlation. Female participation at the Summer Games is explained by the usual variables, but female participation at the Winter Games (which this is still the only paper to examine) is affected only by the host-country dummy. The squared GDP terms have no effect on medal counts for either Olympics, and population does not explain Winter Olympic performance. Home nation advantage holds for both seasons, as it adds three medals in the winter and 25 in the summer.

Heavy-frost countries not only send more athletes, but they also win more medals: four in the winter and 12 in the summer. Formerly Communist countries win 10 more medals in the Winter Games and 18 at the Summer Games. Lowen et  al. (2016) highlight the result that these countries do not have larger delegations, but they seem to use their athletes more efficiently. Finally, Johnson and Ali (2004) make forecasts for the Salt Lake City Olympics, and they report the expected medals for the top medal winners. The predicted and actual results for the top-five forecasts are 31 predicted for Germany (36 won), 21 predicted for Russia (13 won), 20 predicted for USA (34 won), 20 predicted for Norway (25 won), and 16 predicted for Austria (17 won). Lui and Suen (2008) use data from 1952– 2004 and confirm that large, wealthy countries win more medals. The dependent variable is the weighted number of medals (gold is multiplied by three and silver by two), and both income per person and population of each country appear as fractions of the US population. The only other variable is the host dummy. Based on the criterion of the best correlation between the predicted and actual dependent variable, they prefer the Poisson regression to tobit and, surprisingly, negative binomial. They note the persistence of country performance and run the tobit and Poisson models. The Poisson model includes a lagged dependent variable, which improves its fit. (This inclusion, however, violates the assumption of strict exogeneity and creates inconsistent estimates.) The authors calculate that, at the mean, the effect of hosting raises the number of medals by four medals in the Poisson regression and by 22 in the tobit regression. Bravely, Lui and Suen (2008) provide an outof-sample prediction that China would win 14% more medals in 2008 than in 2004. In contrast, the authors predict that Japan and Korea would win fewer medals. Since they wrote before the Olympics, they cannot check their predictions. China famously won 100 medals in Beijing, fully 37 more than in Sydney. This represents a 59% increase. While Japan won only 25 medals in 2008 as opposed to 37 in Sydney, Korea raised its medal count from 30 to 32. The lack of predictive power of this model is understandable given its sparse specification and the absence of the lagged dependent variable, which proved crucial both for Bernard and Busse (2000) and for Forrest, Sanz, and Tena (2010). Forrest et  al. (2010) go head-to-head with Bernard and Busse (2004) in their unabashed forecasting exercise for the medal count for the Beijing Games. But they also contribute to the analytical side of the literature by bringing in a new, important variable, which is a proxy for sport

Olympic Performance

expenditure. Several studies have noted the lack of comparable publicly available data for the amount of funds devoted to sport advancement in each country, which could be used to measure national dedication to Olympic performance. Forrest et al. (2010, p. 578) discovered that ‘The United Nations data on public expenditure by country breaks down spending into nine categories defined by the Classification of Functions of Government (United Nations, 2000), and one of these, “recreational, cultural and religious affairs”, includes sport.’ They report that where the breakdown of these components is available, the variable tracks the spending on sports rather closely. However, it is available only since 1990, thus limiting their sample to four Olympiads (starting in 1992) and creating an unbalanced panel as it becomes available across more countries. Forrest et al. (2010) use the Bernard and Busse (2004) specification of the share of medals won, including their two Communist variables, Soviet and Planned, the lagged dependent variable, as well as a host dummy and shares of world GDP, but not population, finding its effect insignificant. They add recreational spending and a dummy for upcoming host of the Olympics, arguing that the future host will already have some advantage in winning medals. Thus, the Japanese should have won a disproportionate fraction of medals in Rio. The effect of the prospective-host dummy is larger than the effect of the host dummy, but the effect of the lagged variable is by far the most significant. Even with this dominant variable, the effect of the new variable, expenditure, is also significant. Forecasts are typically based on many intuitive adjustments. Forrest et al. (2010) reduce the Soviet dummy coefficient, as they expect the advantage of being a formerly Communist country to fade over time. Without foreseeing the future Russian doping scandals, they correctly forecast a lower medal count for Russia based on this adjustment. Recognizing China as the first Communist country to host a non-boycotted Olympics, they double the impact of the Planned dummy, as well as the host-country dummy, which raises the forecast of medals for China to 90, just 10 shy of the eventual medal count. This success reflects the overall model improvement over Bernard and Busse (2004). However, the predictions for only 27 of the top 35 medal-winning countries are within two standard deviations of the predicted mean; the ratio is 35 out of 36 for Bernard and Busse. Leeds and Leeds (2012) bring two new aspects to the examination of Olympic performance. They estimate the number of total medals, gold medals and the number of silver and bronze medals separately for men and women, implicitly raising the possibility that the production functions for

381

the three medal counts differ for the two sexes. The gender breakdown motivates the use of new variables, including fertility, the date of obtaining women’s suffrage, and the ratio of male to female labor force participation rate. These variables can be seen as another aspect of the political and economic regime that affects total factor productivity. Results from the negative binomial estimation with random effects, using data from the 1996–2008 Olympics, reveal four surprises. First, the frequently used dummy for the host of current Olympics has a positive effect on all medal counts for men, but not on the silver and bronze medal count for women. Second, former Soviet republics and satellites earn more medals for men, but not for women. Third, the expected negative association between fertility and women’s medals also holds for all categories of the men’s medal count. The same is true for the date of women’s suffrage (except for the silver and bronze count for women). This implies that some aspects of socioeconomic development favoring women have positive spillovers for the whole country. Last, while Arab countries have a relatively low medal count, this study reveals that the Arab dummy does not have a separate effect after controlling for socioeconomic development. Lowen et  al. (2016) independently investigate the impact of the status of women on Olympic performance with the Gender Inequality Index (GII), a successor to the Gender Empowerment Measure and Gender Development Index. Like the Gini coefficient, zero represents complete equality and unity complete inequality. In a panel tobit regression with random effects, the authors use four dependent variables and three functional forms for the effect of GDP per capita and population: linear, squared, and logged. (They also estimate participation, which is not discussed here.) Like Leeds and Leeds (2012), they run separate equations for men and women. Using data from the 1996–2012 Olympics, they confirm the general finding of the importance of income and population on the number and proportion of medals won by a nation. The effect of GII is strong and statistically significant in explaining both the male and female versions of the dependent variables: treating women as full-fledged citizens is also good for men, as it raises the total factor productivity of the factors employed in producing medals. In addition, Lowen et al. (2016) show that the current host variable dummy matters for both men and women, increasing the number of medals won by six to 14 medals or by about 2%. Their regime variable (Polity2 score) is insignificant in both of these equations, but the Muslim variable, expressed as the percentage of Muslim population

382

THE SAGE HANDBOOK OF SPORTS ECONOMICS

as reported by the Pew foundation, has a negative effect. This contrast with Leeds and Leeds (2012) may result from the Leeds and Leeds Arab dummy being too crude and missing some Muslim countries. Alternatively, the Lowen et al. model may suffer from omitted variable bias, as Leeds and Leeds have more variables for the status of women. However, the effect of the Muslim variable on the share of medals won by women is strong and unambiguous: among countries that won at least one medal, the proportion falls by about 0.5% for every percentage point of population that is Muslim. In the first part of their paper, Noland and Stahler (2017) extend the preceding two papers (and Klein, 2004, and Noland & Stahler, 2016a) to explain the share of men’s and women’s medals. Moreover, for the first time since Johnson and Ali (2004) and Pfau (2006), they examine the Winter Olympics. They add several variables to the Bernard and Busse (2004) model, including education, as expressed by the average years of total schooling among the 15+ year-old population, and geographical distance in degrees of latitude from the equator. To examine the Winter Olympics medal shares, they also add a dummy variable for countries that experience heavy frost. They report results from tobit, tobit with lagged dependent variable, and random effects tobit regressions, and they provide OLS estimates in an appendix. For the Summer Games, using data from 1960–2012, population, being a host, and being a Communist country are significant with the expected signs, with the last variable coefficient greater for women than for men. Since their Communist dummy represents only currently Communist countries, its effect is smaller than in the earlier papers. GDP per capita is significant in only four equations. The authors suspect that this is caused by high correlation with the distance variable, but it can also be caused by correlation with the education variable, which is always significant. For the Winter Games, 1960–2010, the impact of GDP per capita is strongly significant, much more so than for the Summer Games. The effect of population is also significant, thus corroborating the usual results. While both the distance and the frost variables matter in all equations, education does not. Being a host affects only men, as does, generally, being a Communist country. The authors note the idiosyncratic nature of the Winter Games. Noland and Stahler (2016b) apply their previous framework to comprehensively examine the performance of Asian countries by adding a dummy for a Northeast Asian country (China, Japan and South Korea) and a dummy for 15 other Asian countries. Using data from the Summer

Games between 1960 and 2012, they find that all the variables have the expected, statistically significant impact on the share of total medals in a tobit regression without the lagged dependent variable. Both Asian dummies have a negative impact, indicating the relative underperformance of these countries. With the inclusion of the lagged dependent variable, the negative effect fully disappears for the large countries and almost disappears for the small ones. The effect of the dummies is similar in the regressions for the female share of medals. The model predicts the share of Asian medals to the authors’ satisfaction. They also examine the determinants of individual sport categories, confirming that ‘Northeast Asian countries excel in sports such as archery, badminton, gymnastics, judo, and table tennis’ (Noland & Stahler, 2016b, pp. 82–84). Noland (2016) uses this model to predict the Rio medal count. Recognizing that forecasts require some subjective adjustment, Noland adjusts the model for earlier Olympic boycotts and for Russian doping. He cautions that the forecast is subject to the uncertainty of the impact of the Zika virus; of home-court advantage, which is affected by the political turmoil in Brazil with the potential of unsettling the athletes; and of doping. With this proviso, Noland predicts that the US would win 106 medals, China 94, Russia 66, the UK 55, Germany 48, Japan 42 and that Brazil would place 9th with 35 medals. The final count was 121 medals for the US, 70 for China, 67 for the UK, 55 for Russia, 42 for Germany and 41 medals for Japan. Brazil had a disappointing performance and placed 13th with 19 medals.

CONCLUSION AND SUGGESTIONS FOR FURTHER RESEARCH Research on Olympic performance has come full circle. The recognition that Olympic success is a function of income per capita and population jumpstarted this literature. The cornerstone of the empirical investigation is a production function; the results confirm that income and population are the main determinants of Olympic performance. The literature has also examined the effect of other variables that matter for Olympic success, most frequently hosting the Games and the political regime. The best part of this literature is its ‘tongue-incheek’ quality. Without knowing anything about sports, sports funding, or athletic training, based on just a few macroeconomic variables, one can form a credible forecast for the number of medals

Olympic Performance

a country will win at the next Games. But how good is such a forecast? Is it better than a static forecast? Can it substitute for a sport-by-sport forecast that can be added for a total medal count? It is notable that every paper that makes a serious forecast includes the lagged dependent variable, thus implicitly acknowledging that the models we have are incomplete. One inherent weakness of this literature stems from the loose underlying models. In some ways, all the independent variables are proxies. For example, population can proxy both potential talent and the amount of labor devoted to sports production. GDP per capita can proxy capital devoted to the training of athletes. Geographic features can proxy proclivity and preconditions for different sports. The models surveyed here thus mix policy variables and exogenous control variables. They also mix direct and indirect policy variables, as the host dummy represents a direct policy and Title IX or women’s suffrage represent indirect policies that improve Olympic performance over time. Similarly, Muslim limits on athletic participation have an indirect negative effect on Olympic success. The models are too limited to give much guidance to policy makers. Another handicap stems from lack of financial data. Policy makers would be eager to know how subsidies to specific sports could affect the country’s medal count. The literature seeks to overcome this shortage of data, but the models it uses are not equipped to provide an answer. At the individual sport level, the literature lacks a clear identification strategy and thus cannot infer a causal link between aggregate variables and national success in individual sports. While there is some true correlation, like the success of Nordic countries in Nordic skiing, the distance from the equator, for example, strikes me as spuriously correlated with some (or all) of the Summer Olympic sport categories. The two variables that all papers in this literature contain, in addition to income and population, are dummy variables for hosting the Games and the political regime. The first one is directly relevant to policy makers considering hosting the Games. Prospective hosts of Olympic Games would love to know how to allocate funds to win medals and gain national prestige. The answers these papers offer differ based on the model and data used, but the model and data are not the only reason for the differences of the impacts. The host dummy proxies at least three separate influences. The advantage that local athletes enjoy by being in a familiar environment, the possible bias of judges in judged events, and the amount of additional resources that the host country devotes to the training of their athletes. The dummy for each

383

country captures different combinations of these three influences. As these three cannot be disaggregated, their effects remain unknown. Analogously, this literature shows that formerly and currently Communist countries win more medals, but the effect of the Communist dummy has several interpretations. While it could reflect the strength of ideology motivating athletes to greater performance, it could more realistically stand for better training and greater allocation of resources to athletics, including performanceenhancing drugs. Many athletes of these countries were professionals before professionals were officially allowed to compete in the Olympics. Thus, the interpretation of the effect of the Communist dummy remains unclear. Another concern about this literature, while perhaps common to much empirical literature, is that the structure of the model changes over time. Because the Olympics occur infrequently, we do not have enough observations to track the changes of the underlying regime. Noland and Stahler (2017) address this concern by interacting all their major variables with a time dummy. While the effect of the time-interaction term is either insignificant or inconsistent across specifications, the effect of the population variable has been decreasing for the Summer Olympics. This accords with the observation that institutional features, such as the training of foreign athletes in the US, weaken the link between population and Olympic success. One paper in this literature confirms some of the above concerns. Forrest et  al. (2010, p. 585) reflect on their effort: What purpose does all this serve? Of course, there is an immense media appetite for forecasts of high profile events. Efforts to satisfy that demand by statistical modelling can at least have the spin-off effect of aiding a more effective understanding of the structural determinants of national sporting success; and, by adding public spending to the model, we include something over which governments actually have influence.

As more sport financing data are collected and made available, this literature will start to analyze their impact. This will help to disentangle the effect of macroeconomic variables, including hosting the Games and the political regime. Until then, in many ways, this literature has run its course, especially after the recent comprehensive contributions of Noland and Stahler. In the future, the literature may turn to examining more regional differences, and it will make periodic medal forecasts, starting with PyeongChang 2018 and Tokyo 2020.

384

THE SAGE HANDBOOK OF SPORTS ECONOMICS

REFERENCES Baimbridge, M. (1998). Outcome Uncertainty in Sporting Competition: The Olympic Games 1896– 1996. Applied Economics Letters, 5, 161–164. Ball, Donald W. (1972). Olympic Games Competition: Structural Correlates of National Success. International Journal of Comparative Sociology, 15, 186–200. Balmer, Nigel J., Alan M. Nevill, & A. Mark Williams (2001). Home Advantage in the Winter Olympics (1908–1998). Journal of Sports Sciences, 19(2), 129–139. Balmer, Nigel J., Alan M. Nevill, & A. Mark Williams (2003). Modelling Home Advantage in the Summer Olympic Games. Journal of Sports Sciences, 21(6), 469–478. Bernard, Andrew B., & Meghan R. Busse (2000). Who Wins the Olympic Games: Economic Development and Medal Totals. National Bureau of Economic Research. Working Paper No. 7998. Cambridge, MA: NBER. Bernard, Andrew B., & Meghan R. Busse (2004). Who Wins the Olympic Games: Economic Resources and Medal Totals. The Review of Economics and Statistics, 86(1), 413–417. Forrest, D., I. Sanz, & J. D. Tena (2010). Forecasting National Team Medal Totals at the Summer Olympic Games. International Journal of Forecasting, 26, 576–588. Grimes, A. Ray, William J. Kelly, & Paul H. Rubin (1974). A Socioeconomic Model of National Olympic Performance. Social Science Quarterly, 55, 777–782. Hoffmann, Robert, Lee Chew Ging, & Bala Ramasamy (2002). Public Policy and Olympic Success. Applied Economics Letters, 9, 545–548. Johnson, Daniel K. N., & Ayfer Ali (2000). Coming to Play or Coming to Win: Participation and Success at the Olympic Games. Wellesley College. Mimeo. Johnson, Daniel K. N., & Ayfer Ali (2004). A Tale of Two Seasons: Participation and Medal Counts at the Summer and Winter Olympic Games. Social Science Quarterly, 85, 974–993. Klein, Michael (2004). Work and Play: International Evidence of Gender Equity in Employment and Sports. Journal of Sports Economics, 5(3), 227–242.

Leeds, Eva Marikova, & Michael A. Leeds (2012). Gold, Silver, and Bronze: Determining National Success in Men’s and Women’s Summer Olympic Events. Jahrbücher für Nationalökonomie und Statistik, 232(3), 279–292, www.npr.org/news/ specials/olympics2000/coverage.html Lowen, A., Deaner, R. O., & E. Schmitt (2016). Guys and Gals Going for Gold: The Role of Women’s Empowerment in Olympic Success. Journal of Sports Economics, 17(3), 260–285. Lui, H., & W. Suen (2008). Men, Money, and Medals: An Econometric Analysis of the Olympic Games. Pacific Economic Review, 13, 1–16. Noland, Marcus (2016). Converging on the Medal Stand: Rio 2016 Olympic Forecast. Peterson Institute for International Economics, Policy Brief 16–9. July. Noland, Marcus, & Kevin Stahler (2016a). What Goes into a Medal? Women’s Inclusion and Success at the Olympic Games. Social Science Quarterly, 97(2), 177–196. Noland, Marcus, & Kevin Stahler (2016b). Asian Participation and Performance at the Olympic Games. Asian Economic Policy Review, 11, 70–90. Noland, Marcus, & Kevin Stahler (2017). An Old Boys’ Club No More: Pluralism in Participation and Performance at the Olympic Games. Journal of Sports Economics, 18(5), 506–536. NPR (2000). Meting Out Medals. Sydney 2000: NPR Coverage of the Olympic Games. Retrieved from www.npr.org/news/specials/olympics2000/. Otamendi, Javier, & Luis M. Doncel (2014). Medal Shares in Winter Olympic Games by Sport: Socioeconomic Analysis after Vancouver 2010. Social Science Quarterly, 95(2), 598–614. Pfau, Wade D. (2006). Predicting Medal Wins by Country at the 2006 Winter Olympic Games: An Econometrics Approach. Korean Economic Review, 22(2), 233–247. Tcha, M., & V. Pershin (2003). Reconsidering Performance at the Summer Olympics and Revealed Comparative Advantage. Journal of Sports Economics, 4(3), 216–239. United Nations (2000). Classification of the Functions of Government (COFOG). M. No. 84. Statistics Division. New York: UN.

38 The Economics of Mega-Events: The Impact, Costs, and Benefits of the Olympic Games and the World Cup Candon Johnson

INTRODUCTION The Olympic Games and World Cup rank among the largest sports events in the world. The Summer Olympic Games, Winter Olympic Games, and World Cup all occur every four years and allow the world’s premier athletes to compete against one another. These are arguably the most important mega-events in the world and will be the focus of this chapter. The term mega-event(s) will be used interchangeably with the Olympic Games and the World Cup throughout. This chapter discusses the economics of hosting sports mega-events, the financing of these events, and the process through which the rights to host sports mega-events are assigned to cities and regions. The chapter documents the presence and persistence of large cost overruns associated with hosting mega-events and discusses their likely causes and reviews the potential economic impacts associated with these events. Finally, the existing scholarly evidence, based on retrospective econometric research is reviewed. Sports mega-events draw an incredibly large amount of spectators and generate billions in broadcast rights revenues. However, the host city or region also faces enormous costs, including opportunity costs. While many proponents of hosting sports

mega-events cite overly-optimistic projections of tangible economic benefits, a thorough assessment of all costs, and a close look at all available evidence on the past economic experience of host cities and regions casts serious doubts on the wisdom of hosting the Olympic Games or World Cup based solely on economic grounds.

HOSTING AND FINANCING MEGA-EVENTS Hosting the Olympics is not a short process. The host city is selected seven years in advance of the Games being held. A number of steps occur in the host city selection process. Take, for example, the election of Rio de Janeiro as the host for the 2016 Olympic Games1. The first phase is the application process. For the 2016 Games, seven cities were approved by their National Olympic Committees (NOCs) and applied to host the games: Chicago, Tokyo, Prague, Baku, Doha, Madrid, and Rio de Janeiro. Each country’s National Olympic Committee informs the International Olympic Committee (IOC) of the name of their Applicant City. The IOC reviews the

386

THE SAGE HANDBOOK OF SPORTS ECONOMICS

cities’ potential to organize the games. The Applicant City replies to an IOC questionnaire, and their answers are made into an Applicant File. This file is studied by an IOC-appointed committee to produce a report on each Applicant City. This report aides the IOC Executive Board in selecting the cities that will move onto the next phase as a Candidate City. After the applicant phase, the process moves on to the candidate phase. There were four candidate cities considered for the 2016 Olympics Games: Rio de Janeiro, Chicago, Madrid, and Tokyo. This phase requires cities to submit a Candidature File and guarantees to the IOC. This file details the city’s Olympic plan, forming the basis for the bid of each city. Following the file submission, the IOC Evaluation Commission visits each city. The visits for the 2016 Games occurred in April and May 2009, after which the Commission produces a report for each city. This report is a technical appraisal of each city’s bid. The 2016 Olympic application cycle introduced a technical briefing in Lausanne, Switzerland, allowing IOC members and cities to discuss technical aspects of each bid. Next was a full IOC session in Copenhagen in which each city had 45 minutes to present their Bid, followed by 15 minutes of questions. Following the presentation, the IOC voted to determine the host city. The voting consisted of three rounds. In each round, the city receiving the lowest number of votes was eliminated. In the first round, Chicago received the lowest amount of votes (18), followed by Tokyo (22), Rio de Janeiro (26), and Madrid (28). Round two eliminated Tokyo (20 votes), leaving Madrid (29 votes) and Rio de Janeiro (46 votes). The final round saw Rio de Janeiro win the bid, receiving 66 votes to Madrid’s 32. After winning the bid, the Organizing Committee for the Olympic Games (OCOG) is formed to organize the Games and is dissolved following the Games. The OCOG executive body is comprised of the IOC member(s) from that country, the president and secretary general of the country’s National Olympic Committee, and at least one member representing the host city.2 Brazil was also the host of the 2014 World Cup. The procedure for selecting the host country of the World Cup differs from that of the Olympic Games. The traditional approach used by FIFA (Fédération Internationale de Football Association) was to alternate hosts between Europe and the Americas. This changed with the selection of South Korea and Japan in a joint bid to host the 2002 World Cup, and further changed with the selection of South Africa in 2004 to host the 2010 World Cup. A new policy was implemented, rotating through the six continents represented by

FIFA. However, like the competition to host the 1984 Olympic Games, there was a lack of interest in hosting the 2014 World Cup when it was South America’s turn to host. When the 2014 World Cup host country was chosen, only one country submitted a bid, Brazil (Zimbalist, 2016, pp. 29). Without competition to host the World Cup, FIFA needed to further alter the hosting policy. The new policy allowed four of the six continents to be eligible in each cycle. The two ineligible continents are the continents that hosted the two most recent World Cups. This policy change seemed to spur more interest in hosting the World Cup. For the 2018 World Cup cycle, eleven countries bid to host compared to only Brazil for the 2014 World Cup. Hosts incur incredibly high costs of financing the Olympic Games and World Cup. For example, the 2008 Summer Olympic Games in Beijing cost US$40 billion, the 2014 Winter Games in Sochi cost $50 billion (Zimbalist, 2016, pp. 2). Brazil spent $20 billion hosting the 2014 World Cup and more than $13 billion hosting the 2016 Olympic Games held in Rio de Janeiro.3 The World Cup to be held in Qatar in 2022 has been projected to cost a staggering $200 billion (Zimbalist, 2016, pp. 2). Hosting the Olympics and World Cup do in fact generate some positive economic impact, but generally the revenues are not nearly the size of the costs. Consider Beijing as an example. The 2008 Summer Olympic Games in Beijing generated and estimated $3.6 billion in revenues. These games cost $40 billion, meaning that the revenues were less than 10% of the cost of hosting the Games. Host cities provide large subsidies for the OCOGs putting on the Games. The amount that the host city provides to the OCOGs depends on the amount of funding the OCOG receives from the IOC and the availability of private funding. The OCOGs’ budget mostly includes operating costs of the Games, while the host city is largely responsible for infrastructure. Revenues generated come largely from broadcasting rights fees, accounting for more than half of total revenues. Revenues are split between OCOGs, NOCs, the IOC, and International Federations (IFs). Overall, the IOC receives about 8% of revenues, while the rest is split between the other organizations (Humphreys & Howard, 2008). For the World Cup, FIFA recently changed the structure of costs and revenues associated with the tournament. The system introduced for the 2010 and 2014 World Cups included FIFA covering most of the operating expenses, but FIFA also retains all revenues. Meanwhile, the host country was responsible for infrastructure spending, including communications, transportation, and stadiums. It is important to note here that

THE ECONOMICS OF MEGA-EVENTS

there are distinct differences in hosting the World Cup and the Olympic Games. While the Olympic Games are generally hosted by, and held within, one city and the surrounding area, the World Cup is held throughout the host country in various host cities. For example, the United States hosted the World Cup in 1994. During this World Cup, matches were held in Chicago, Washington DC, Los Angeles, San Francisco, New York, Dallas, Boston, and Orlando. Considering that the World Cup matches are held in multiple cities throughout the host country, the infrastructure costs of the World Cup are substantial. The host country is required by FIFA to have eight modern stadiums with a minimum seating capacity of 40,000. In addition, of these eight stadiums there must be one stadium that seats 60,000 for the opening night match, and another that holds 80,000 for the finals (Zimbalist, 2016, pp. 31). This can be a large hurdle for developing countries like South Africa, South Korea, and Japan as they do not have adequate existing facilities. These host countries were required to build stadiums, spending billions of dollars in the process. In contrast, countries like Germany, France, and the United States are better equipped to host the World Cup and do not need to undertake as much new stadium construction. Hosting the World Cup and the Olympic Games is an incredibly costly venture for the host.

POTENTIAL IMPACTS OF MEGA-EVENTS Sports mega-events represent a major investment undertaken by a host city or country. Proponents of hosting mega-events frequently claim that the events will lead to positive outcomes, both socially and economically, in the host area. Opponents of hosting mega-events claim that there can be negative outcomes, and the positive impacts that do exist are not large enough to warrant the extreme cost of hosting these events. In this section, the potential impacts of hosting the Olympic Games and World Cup will be discussed. While this section will largely focus on the economic impacts of the sports mega-events, other potential impacts are to be considered as well. Pierre de Coubertin, founder of the IOC, believed the Olympic Games should bring nations together. In fact, the Olympic website notes: ‘The goal of the Olympic Movement is to contribute to building a peaceful and better world by educating youth through sport practised without discrimination of any kind, in a spirit of friendship, solidarity and fair play.’4 FIFA, the governing for the World Cup

387

has a similar statement on their website. FIFA’s World Cup vision is to ‘promote the game of football, protect its integrity and bring the game to all.’5 Clearly, both organizations, and the events associated with them, revolved around promoting worldwide solidarity and integrity. With this in mind, the Olympics and World Cup potentially affect the host in many different ways. In addition to economic impacts, the impact of these events can be felt socioculturally, psychologically, politically, and environmentally (Scandizzo & Piereloni, 2018). Sociocultural and psychological impacts represent an interesting area. Pride and prestige are associated with hosting mega-events, generating an uplifted mood in the host area. Hosting an event of this size can generate increased prestige for the host country, making the host a more desirable destination for tourists and immigrants. Hosting may also increase participation in sporting activities throughout the host country. This is an important potential positive spillover in that a more physically active society will arguably become a healthier society. However, negative impacts are possible as well. An important negative impact of hosting a mega-event is the potential increase in crime in the host area. This is not unique to mega-events, as studies have shown a rise in crime related to other sporting events (Card & Dahl, 2011). However, the IOC claims that the Olympic Games can reduce crime. Hosting mega-events can also impact local politics. Arguably, the most important positive impact stemming from megaevents is the host area becoming more recognizable globally, putting the host and the politicians associated with the mega-event at the forefront of world news. Local politicians also build valuable political and human capital through planning the Olympics and World Cup. These events are large-scale and require years of planning and coordination with the IOC or FIFA. A stronger relationship between the private and public sectors can be another positive impact of hosting megaevents. However, this relationship can be negative due to mega-events fostering corruption through self-interest. Hosting mega-events can lead to negative outcomes, particularly when hosting the events is not popular among local residents, causing political unrest. This was the case in Brazil prior to hosting the 2014 World Cup; widespread political unrest occurred in Brazil during the Confederations Cup. The Confederations Cup, an international soccer competition held the year prior to the World Cup (in 2013 in the case of Brazil), drew over a million Brazilian protesters to the streets. The protests were against high government spending $15–20 million for hosting the 2014 World Cup.

388

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The protests continued as the World Cup approached. Many cities also experienced strikes by police and teachers, among other workers, in the run-up to the World Cup. Hosting a mega-event requires large infrastructure investments which may have negative environmental and physical impacts. The Olympic Games and World Cup potentially draw substantial tourist activity, and hosts must be equipped to handle the increased inflow of visitors. In addition to the construction of new facilities for competitions, this also requires additional investment in the surrounding area for non-sport-related infrastructure. Of the more than $13 billion spent hosting the 2016 Summer Olympic Games in Rio de Janeiro, $8.2 billion was spent on non-sportrelated venues. These included a subway system and a doping laboratory. Additionally, Guanabara Bay in Rio was badly polluted prior to the Games, leading to part of the $8.2 billion being invested in cleaning up the bay and renovating the port.6 These investments in Rio de Janeiro may not have happened had it not been for the 2016 Olympic Games. Investments in infrastructure may not be positive. Numerous studies find no economic impact from the construction of sporting facilities. Furthermore, large expensive sport facilities for mega-events will typically be underutilized or unused following the events. In addition to the facilities being unlikely to generate positive impacts, there are other potentially negative impacts generated by hosting a mega-event. Negative externalities in the form of overcrowding and pollution can be generated by megaevents. Hosts areas may be ill-equipped to handle the large influx of foreign visitors, as well as increases in waste generated by crowds of visitors. Overbuilding is another potential negative impact. Hosts may build tourist facilities like hotels of excessive size to accommodate the large influx of mega-event-related tourism only to find them unoccupied after the event ends. Mega-events cost huge amounts of money, potentially generating large opportunity costs. The protests in Brazil provide an example of the opportunity costs of hosting mega-events. Brazilians were angry because they believed the billions of dollars the government spent hosting the Games should have been spent elsewhere. Money spent on hosting a mega-event could be used for a variety of other activities, such as education, infrastructure not associated with the megaevent, funding police and other emergency service departments, health care, etc. Another example of the opportunity costs of mega-events is New York City’s unsuccessful bid to host the 2012 Summer Olympic Games. The bid involved building a

new Olympic stadium on Manhattan Island that included first building a concrete pad estimated to cost $400 million alone. This stadium would have hosted the New York Jets of the NFL following the Games. Hosting an NFL team would mean the stadium would likely be used approximately twelve times a year, including eight home regular season games, two pre-season games, possibly a few post-season games, and occasional other events, such as concerts. In addition to the large costs of hosting the Olympic Games, this stadium would sit on one of the most valuable pieces of real estate in the world. This would have generated a large opportunity cost to hosting the Olympic Games in New York City. This land could be used for something much more productive instead of a stadium that would be used roughly a dozen times a year (Zimbalist, 2016, pp. x). The hosting of mega-events will generate important economic impacts in both the long run and short run. Potential short-run benefits include tourism, job creation, increased wages and real estate prices. Increased tourism is an obvious potential short-run benefit. However, it is not clear whether this benefit actually occurs when megaevents are hosted. While the Olympic Games draw substantial visitors, there are no clear effects on local foreign tourism. For example, China was visited by fewer foreign tourists in 2008, the year of the Beijing Olympic Games, than in 2007. A similar outcome occurred in the UK during the 2012 Olympic Games held in London. Furthermore, Beijing’s anticipated foreign tourists per night was 400,000 during the Games, but only 235,000 foreign visitors entered the country. Beijing is not alone in receiving fewer tourists than expected. Both Sydney and Athens received fewer foreign visitors than expected when hosting the Games. Sydney expected 132,000 visitors per night but received 97,000, while Athens expected 105,000 but received only 14,000. South Korea saw a decrease in foreign visitors compared to the previous year and overestimated the number of foreign visitors who would enter the country during the 2002 World Cup. Salt Lake City saw a decrease in the number of skier days during the Olympic Games. A few hosts received a modest increase in tourism during the Olympic Games. While Sydney overestimated the number of foreign tourists it would receive, there was a slight increase compared to the year before. The same occurred in British Columbia after Vancouver hosted the Winter Olympics (Zimbalist, 2016, pp. 50–52). The potential benefits of job creation from mega-events could stem from increased construction and tourism activities. However, the increase in construction and tourism may both be

THE ECONOMICS OF MEGA-EVENTS

temporary. Furthermore, tourism-related employment may not even experience an increase during a mega-event, as occurred in British Columbia during the Winter Games. The percentage increase in tourism-related employment in British Columbia was significantly lower in the year Vancouver hosted the Winter Games than in the years leading up to the Games. Increased construction can help the local economy if the economy is weak. An increase in construction can lower the unemployment rate, and the infrastructure investment associated with a mega-event could help boost output growth in the short run. Any increase in construction employment will be temporary, making this a shortrun impact. In contrast, if the local economy is strong, with an already low unemployment rate, an increase in construction will not increase local output, but instead reallocate scarce resources to mega-event-related activities. This could actually decrease local output by taking scarce inputs from productive uses in the local economy and reallocating them to hosting the mega-event, a potentially less productive activity. An important aspect of increased construction is that this construction will likely be largely publicly funded. To pay for this construction, local government may have to raise taxes, reduce public spending elsewhere, or borrow additional funds. A final potential negative outcome related to increased mega-event construction is a lack of sufficient capacity in the local labor pool. Megaevents require substantial construction, likely leading to a shortage in the local labor market, and potentially requiring construction labor to be imported. Imported labor can potentially be mistreated or underpaid. This has been an ongoing issue in Qatar, host of the 2022 World Cup. Qatar experienced a large number of deaths of migrant workers in 2012 and 2013 due to unsafe working conditions and poor living conditions on World Cup stadium construction sites. Construction workers in Qatar were reportedly paid as little as $0.76 per hour for work on these stadiums (Zimbalist, 2016, pp. 48–50). Proponents of hosting mega-events often argue that hosting provides local benefits through increased local real estate prices. However, this increase does not benefit all local residents. While local real estate prices do, in general, rise as a result of hosting mega-events, the benefits largely accrue to property owners and real estate agents. Low-income families can be hurt by increasing real estate prices, if they are unable to afford higher rent, causing these families to relocate. Understanding the short-run impact of hosting a mega-event requires a detailed understanding of the costs involved. From the beginning, bidding is

389

incredibly costly for potential host cities. The IOC and FIFA each require upfront payments from potential hosts during the bidding process. For the IOC, these payments include $150,000 at the Applicant stage and another payment of $500,000 at the final Candidate stage. In addition to these upfront costs, there are costs incurred putting together the bid. These include hiring consultants, advertising, employing executives, etc. Bidding for the rights to host the 2016 Olympic Games cost Chicago $100 million and Tokyo $150 million, and both bids failed (Zimbalist, 2016, pp. 41–42). Extravagant opening and closing ceremonies also lead to large costs that may not be obvious. It is clear that sporting and non-sport infrastructure represent expensive investment but holding opening and closing ceremonies can be expensive as well. The opening and closing ceremonies give hosts a stage to showcase their country in an effort to attract tourists from around the world. This leads to substantial spending on these ceremonies. For example, China spent nearly $350 million on their opening ceremony (Zimbalist, 2016, pp. 42). Additional costs of hosting a mega-event comes from business disruption due to the construction of facilities. Businesses are further disrupted by surrounding street closures and limited foot traffic, and often experience a decrease in sales during the construction period. This disruption of business operations may contradict expectations of an increase in business activity and increased investment, exacerbating negative effects. Security costs represent a final short-run cost to consider. Since the terrorist attacks of September 11, 2001, all security costs, including at megaevents, have risen substantially. The 2004 Summer Olympic Games held in Athens, Greece, were the first following the 9/11 attacks. Security for the Athens Games cost $1.5 billion; estimated costs in the bid, developed before 9/11, were only $400 million (Zimbalist, 2016, pp. 43). The IOC has consistently asserted that multiple long-term, or legacy, benefits will be generated by hosting the Olympic Games. Potential legacy benefits could apply to the World Cup as well. These claimed benefits include: sustained boosts in tourism for long periods of time; additional construction of sports venues, non-sport infrastructure, and other accommodations; increased trade and investment; increased education and public health; lower crime rates; and a long-run increase in real estate prices. The creation of jobs and increased wages are also frequently claimed long-run benefits. Tourism has been discussed as a potential shortrun impact, but a sustained increase in tourist

390

THE SAGE HANDBOOK OF SPORTS ECONOMICS

activity could provide a large long-run benefit to a host area. The Olympics and World Cup place host areas on the world stage, as the events are viewed by billions around the world; mega-events essentially act as an enormous advertisement for tourism to the host area. However, hosts may not experience a long-run increase in tourism after a mega-event ends. For example, to accommodate the expected increase in tourism during the 2000 Olympic Games, Sydney increased its hotel capacity by 30% (Zimbalist, 2016, pp. 59). However, the expected large tourism boost did not materialize and within four years, ten of the city’s largest hotels closed. Many other mega-event hosts saw little or no long-term increases in tourism after the events, and the increase did not last more than a few years. An exception to this was the 2004 Summer Olympic Games in Athens. Tourism in Greece was impacted positively. Leading up to the Games, Athens increased hotel capacity, but not as much as Sydney. Furthermore, a large portion of their spending on hosting went to upgrades to the airport, roads, and other public transportation, making it easier to travel to and in Greece. While the IOC cites higher real estate prices, construction, increased trade and investment, and lower crime rates as long-term benefits, it is not clear that these benefits materialize, or that these actually constitute benefits. Real estate price increases only benefit certain residents, as discussed earlier. Construction can be a negative if the hosts construct extravagant stadiums and venues that end up being unused after the mega-event ends. These over-large stadiums and venues, along with other overly ambitious plans, are referred to as ‘white elephants.’ Studies conducted on increased trade and investment following mega-events, and crime will be discussed in a later section. One clear long-run cost of hosting a mega-event is the debt incurred by the host city to finance construction and other hosting costs. Local government often goes into debt to pay for megaevent-related costs. This debt must be paid off, often requiring the cutting of other public services to cover the debt cost, as well as making it more difficult for hosts to invest in other projects, such as education, health care, or other infrastructure. This, again, points out the large opportunity costs of hosting a mega-event.

COST OVERRUNS Prospective mega-event hosts face a peculiar dilemma. Almost all mega-event hosts, especially Olympic Games hosts, experience cost overruns.

Cost overrun refers to the case when a planned project goes over budget, stemming from an underestimation of the total cost of the project. Not only do the Olympic Games experience cost overruns, the cost overruns are generally large. For example, the 2014 Olympic Games hosted by Sochi, Russia, were initially budgeted at $12 billion. However, the Games cost actually $50 billion to host, a cost overrun of over 400%. In fact, every Olympic Games since 1960 resulted in cost overruns, meaning the Olympic Games go over budget 100% of the time. This is highlighted by Flyvbjerg and Stewart (2012), who show that the average cost overrun associated with hosting the Olympic Games is 179%. Additionally, the maximum cost overrun during their sample period (1960–2012) occurred during the 1976 Summer Olympic Games in Montreal at 796% in real terms (Flyvbjerg & Stewart, 2012). There are numerous reasons why cost overruns occur. One stems from local self-interest, as bids are largely constructed by those who have the most to gain from hosting the mega-event. This incentivizes those submitting a bid to initially underestimate hosting costs, generating a bid including the bare minimum elements, and amending the bid to include extravagant add-ons later in the process. This intentional underestimating of the total costs to convince government bodies to support bidding efforts partially contributes to massive, persistent cost overruns. Other reasons behind cost overruns are not deliberate in nature. The bidding process is long and includes competition among potential hosts after the initial budget. Competition among potential hosts leads to them attempting to outdo one another in order to be awarded the hosting rights, increasing the costs. Inflation stemming from the length of the hosting process is also a contributing factor. To put the impact of inflation into context, without adjusting for inflation, the 1976 Olympics in Montreal experienced a cost overrun of 1266% in nominal terms. The final cause of cost overruns is also a claimed benefit of hosting the Olympic Games made by mega-event proponents and the IOC: increases in future local tourism, construction, and real estate prices. Increases of construction demand generated by hosting the games also increases the local price of all other construction. This can increase the costs of hosting mega-events throughout the period between the awarding of the hosting rights and the staging of the events. Increases in real estate prices, and an influx of tourists during the event can drive up local prices rapidly throughout the local economy. Cost overruns appear certain when hosting the Olympic Games; they also occur when hosting the

THE ECONOMICS OF MEGA-EVENTS

World Cup. The 2014 World Cup in Brazil experienced a cost overrun of over 450% just for the cost of stadium construction alone (Zimbalist, 2016, pp. 44). Local governments must keep this in mind when considering submitting a bid to host a megaevent. Flyvbjerg and Stewart (2012) note that the size of cost overruns had been decreasing in recent years, but the 2012 Summer Games in London did not follow this trend. London experienced a cost overrun of 133% in real terms following Beijing and Vancouver having cost overruns of ‘only’ 35% and 36% respectively. Furthermore, this change in the trend has persisted with Sochi’s cost overrun topping 400% in 2014, and the 2014 World Cup having an overrun of over 450%. However, the 2016 Summer Olympic Games in Rio de Janeiro had a cost overrun of ‘only’ 51%.7 This smaller percentage may be partially attributed to the fact the Brazil had already invested heavily to host the World Cup in 2014. Nonetheless, governments must consider this trend of massive cost overruns in the future when considering hosting a mega-event.

EVIDENCE OF THE ECONOMIC IMPACT OF THE OLYMPIC GAMES AND WORLD CUP Due to the large size and potential impacts of hosting mega-events, a large body of research assesses the impacts of hosting sports megaevents. A substantial portion of these studies focus on the Olympic Games, the largest sporting event in the world. Studies use both ex-ante and ex-post approaches. In other words, studies predict the expected hosting impacts and also, retrospectively, test for evidence of an impact.8 It is common for ex-ante studies predicting the impacts of sports mega-events to use regional input-output (I-O) models. I-O models formally describe relationships across different sectors of the local economy and generate a multiplier to determine how much a given input into the economy will be multiplied as it moves through different local economics sectors. These multipliers are generally large in reports promoting the hosting of a sports mega-event, ranging from 1.7 to 3.5. This multiplier indicates that an input or increased spending of $1,000 associated with a sports mega-event would generate a total local impact or increase in economic activity of between $1,700 and $3,500 (Zimbalist, 2016, pp. 34). However, the methodology underlying I-O models utilized by these ex-ante studies are rife with problems. They are based on inter-sectoral

391

relationships that are constantly evolving in response to local changes like hosting a sports mega-event and are highly aggregated in nature. I-O models include aggregated measures of local goods and services, such as textiles or food, instead of actual local goods and services, such as t-shirts or cheeseburgers. Trade represents another major issue for I-O model-based studies. For instance, a restaurant may use products from outside the local economy to produce a meal. They may import silverware, tables, and food products from abroad. I-O models do not fully account for such inflows. Further, the money spent by tourists attending a local mega-event may be spent at international businesses, such as an international hotel chain or restaurant (Zimbalist, 2016, pp. 35–36). These issues lead I-O models to generate overstated predictions of local economic impacts.9 Proponents of hosting mega-events claim a reduction in crime as a benefit. Baumann, Engelhardt and Matheson (2012) examined the impact of the Olympic Games. These authors use data at the Metropolitan Statistical Area (MSA) level from 1981 to 2006 in the United States. In addition to the Olympics, they include NHL, NBA, and MLB teams, the Super Bowl, and the World Cup. Their sample includes the 1984 Olympic Games in Los Angeles, 1996 in Atlanta, and 2002 in Salt Lake City. Additionally, the World Cup in 1994 included eight different MSAs that hosted games during the event. They find that an MSA hosting the Olympics Games was associated with a 10% increase in property crime rate. There are no significant results found for hosting the World Cup.10 This study provides evidence against the claim that hosting either the World Cup or Olympic Games reduces the crime rate, instead indicating an increase in crime associated with the Olympic Games. In terms of the impact of mega-events on trade and investment, the first study was conducted by Andrew K. Rose and Mark M. Spiegel (2011). Rose and Spiegel find a large and sustainable increase of exports of 20% for countries that host the Olympic Games, using an ordinary least squares (OLS) regression using data for 196 different locations and territories from 1950 to 2006. A similar effect is found for countries which the authors claim is a ‘signal’ effect, allowing a signal of openness to be sent from an Olympic bid. Both hosting the Olympics and unsuccessfully bidding to host the Games yield similar results, making the findings of this study questionable. Due to the questionable results in Rose and Spiegel (2011), Wolfgang Maennig and Felix Ritcher responded in 2012. Maennig and Ritcher (2012) challenge the results in Rose and Spiegel due to potential selection bias. Rose and Spiegel

392

THE SAGE HANDBOOK OF SPORTS ECONOMICS

(2011) compare countries that may be structurally different and include nonmatching groups of countries, due to the plausibility that hosts and potential host countries will be structurally different from countries that do not host or bid to host. This nonmatching issue is overcome using propensity score matching estimation, also using data from 1950 to 2006. After convincingly correcting for the methodological issues present in Rose and Spiegel’s (2011) paper, Maennig and Ritcher (2012) find that statistical significance disappears from the results, meaning there is no impact of hosting the Olympic Games on exports, refuting another benefit claimed by those who support hosting the Olympic Games. Trade is not the only area of sports mega-event research with disputed results. An early study of the effect of the 1996 Olympic Games in Atlanta on employment and wages provoked a paper criticizing the initial results. The authors of the original study responded to the response, providing an intriguing back and forth among economists. The original paper by Julie L. Hotchkiss, Robert E. Moore, and Stephanie M. Zobay (2003) found that the 1996 Olympic Games boosted employment by 17% in Georgia counties both affiliated with and close to the Olympic Games compared to other counties in Georgia. To obtain these results, the authors use a differences-in-differences method to test for pre- and post-event impacts. For robustness, a random-growth approach is used to support the results. The positive results of the 1996 Olympic Games were questioned by Arne Fedderson and Wolfgang Maennig (2013a). Maennig and Fedderson questioned the results in Hotchkiss et al. (2003) on multiple grounds. First, the earlier study does not adequately account for underlying trends. When correcting for trends, Maennig and Fedderson find no significant increase in employment associated with hosting the Games. They also run numerous nonparametric tests in lieu of the standard differences-in-differences model, again finding no effect. In response to Feddersen and Maennig (2013), Julie L. Hotchkiss, Robert E. Moore, and Fernando Rios-Avila (2015) revisit the topic of their original paper. They attempt to prove that the results they initially reported were robust to the alternative methods in Feddersen and Maennig (2013a). In their response, they again find that employment growth in Georgia counties that either hosted the Olympics or were near host counties outpaced employment gains in other Georgia counties by 11%. In correcting their approach, they still reported a positive effect from the Olympics, but not at the same magnitude of their original paper. The portion of this paper (Hotchkiss et al., 2015)

that may be more convincing is when they compare MSAs in Georgia that hosted the Olympics to similar MSAs throughout the southern United States. Results indicate that MSAs hosting the Olympics outpaced employment gain in other southern states by 5%. An additional study conducted by Maennig and Fedderson examined mega-events and sectoral employment, again using the 1996 Olympic Games held in Atlanta (Feddersen & Maennig, 2013b). They analyze monthly data for 16 different sectors using a nonparametric approach to isolate employment effects. Their results do show a slight boost in employment. However, they find no evidence of a persistent shift in employment growth. They find that the Games lead to an increase in 29,000 jobs in July 1996, when the Olympic Games were being held, but this effect was only seen in Fulton County and only in July. Additionally, the jobs increase was concentrated in only three sectors of the economy – the retail trade, accommodation and food services, and arts, entertainment, and recreation. To further test the merits of this claimed benefit, Robert Baumann, Bryan Engelhardt, and Victor A. Matheson (2012) study the impact of the 2002 Winter Olympics in Salt Lake City on employment in Utah. While their paper finds an increase in employment, promoters of the Olympics overestimated the impact substantially. While promoters estimated 35,000 job-years, the findings indicate an increase of 4,000–7,000 jobs using a control group of states adjacent to Utah. Furthermore, these jobs were focused in the leisure industry and the effect dissipated after a year. Considering these papers, there may be slight job growth associated with hosting the Olympics, but the growth is much less than the estimate ex-ante and dissipates quickly. Studies have also examined the impact of the Olympic Games on local consumption. Robert A. Baade, Robert A. Baumann, and Victor A. Matheson (2010) assess the impact of the 2002 Winter Olympic Games in Salt Lake City on local taxable sales. They examine quarterly taxable sales from 1982 through 2006 using an auto-regressive-moving-average (ARMA) model. Their analysis examines different sectors of the economy, as the Olympics will likely generate a differential impact on different sectors. Their findings indicate a net negative effect on taxable sales associated with hosting the Olympic Games. While there are gains experienced in hotels, and eating and drinking establishments, these gains are offset by losses elsewhere, leading to a net loss of $167.4 million. To further show the consumption losses from hosting the Olympic Games, James A. Giesecke and John R. Madden (2011) analyze the 2000

THE ECONOMICS OF MEGA-EVENTS

Olympic Games and find a net loss in Sydney, Australia, as well. In this study, the authors conduct simulations from 1997 to 2006 to assess Olympic-induced tourism demand. They also conduct simulations of a non-Sydney Olympics counterfactual, finding that the Olympics generates a loss in real consumption of AUS$2.1 billion. It is apparent from these consumption studies that hosting mega-events can have a large negative impact on consumption, further showing that the claimed benefits of hosting a mega-event do not occur. While the literature reviewed thus far focuses on the Olympics, Robert A. Baade and Victor A. Matheson (2004) examine the impacts of the 1994 World Cup in the United States. They use data from 1970 to 2000 to estimate the effect of hosting on income growth. They do so by comparing predicted growth to actual growth. Of the 13 cities that hosted World Cup games, nine experienced growth lower than the predicted value, indicating an economic loss from hosting the event. In total, in their ex-post analysis the cities hosting games experienced a cumulative loss of $9.26 billion compared to the ex-ante estimate $4 billion in benefits. Considering the numerous studies discussed here, it is difficult to conclude that the benefits claimed by those who wish to host a mega-event hold any validity. According to the literature, it appears that at most there would be a slight, temporary increase in local employment from hosting a mega-event. However, this temporary employment increase is likely accompanied by an increase in crime, a decrease in local consumption, and losses in income growth.

CONCLUSIONS This chapter discusses the history of hosting and financing both the Olympic Games and World Cup, the two largest sporting events in the world. Hosting these events are incredibly costly, but proponents of hosting mega-events often make claims of large benefits to be received from hosting the events. These benefits, including reduced crime, increased employment, income growth, increased trade, increased consumption, and many more, have been discussed in detail. Little evidence exists supporting any of the claimed benefits. Studies found increases in crime, reductions in local income growth, and reductions in local consumption. Furthermore, there is no clear evidence supporting an impact of mega-events on trade. There is evidence showing

393

slight employment growth associated with hosting the Olympic Games, although this growth is only seen in a few sectors, is temporary, and well under the expected increase. Before governments decide to submit a bid to host a mega-event they should fully consider all the costs and benefits of doing so. Particularly, given the large opportunity costs associated with hosting the Olympic Games and World Cup, the large amount of money spent on hosting these events might be better spent elsewhere in the local economy to spur growth or increase local welfare. Mega-events also utilize large amounts of land, which is a valuable, scarce resource. Considering the high costs, as well as the numerous studies refuting claimed benefits, hosting these megaevents may not be an economically valuable local investment.

Notes   1  This information comes from the Olympic website. https://www.olympic.org/2016-host-city-election   2  Details from the Olympic website: https://www. olympic.org/ioc-governance-national-olympiccommittees   3  h t t p s : / / w w w . u s a t o d a y . c o m / s t o r y / s p o r t s/ olympics/2017/06/14/ap-analysis-rio-de-janeiroolympics-cost-13-1-billion/102860310/   4  https://www.olympic.org/about-ioc-institution   5  https://www.fifa.com/about-fifa/index.html   6  https://www.usatoday.com/story/sports/olympics/2017/06/14/ap-analysis-rio-de-janeiroolympics-cost-13-1-billion/102860310/   7  https://www.bloomberg.com/news/articles/ 2016-07-06/rio-olympic-cost-overruns-reach1-6-billion-oxford-study-finds   8  Scandizzo and Pierleoni (2018) provide a thorough analysis of the literature on the economic impacts of the Olympic Games.   9  More detailed information on I-O models can be found in Zimbalist (2016), and Scandizzo and Pierleoni (2018). 10  Baumann et. al find that hosting the Super Bowl is associated with a 2.5% decrease in violent crime, but no results for any other sports.

REFERENCES Baade, Robert A., Robert Baumann, & Victor A. Matheson (2010). Slippery slope? Assessing the economic impact of the 2002 Winter Olympic Games in Salt Lake City, Utah. Région et Développement, 31, 81–91.

394

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Baade, Robert A., & Victor A. Matheson (2004). The quest for the cup: assessing the economic impact of the World Cup. Regional Studies, 38(4), 343–354. Baumann, R., Ciavarra, T., Englehardt, B., & Matheson, V. A. (2012a). Sports franchises, events, and city livability: An examination of spectator sports and crime rates. The Economics and Labour Relations Review, 23(2), 83–97. Baumann, Robert, Bryan Engelhardt, & Victor A. Matheson (2012b). Employment effects of the 2002 Winter Olympics in Salt Lake City, Utah. Jahrbcher fr Nationalkonomie und Statistik, 232(3), 308–317. Card, David, & Gordon B. Dahl (2011). Family violence and football: the effect of unexpected emotional cues on violent behavior. The Quarterly Journal of Economics, 126(1), 103–143. https:// doi.org/10.1093/qje/qjr001 Feddersen, Arne, & Wolfgang Maennig (2013a). Employment effects of the Olympic Games in Atlanta 1996 reconsidered. International Journal of Sport Finance, 8(2). Feddersen, A., & Maennig, W. (2013b). Mega-events and sectoral employment: The case of the 1996 Olympic Games. Contemporary Economic Policy, 31(3), 580–603. Flyvbjerg, Bent, & Allison Stewart (2012). Olympic proportions: cost and cost overrun at the Olympics 1960–2012. Said Business School Working Papers. Oxford: Oxford University.

Giesecke, James A., & John R. Madden (2011). Modelling the economic impacts of the Sydney Olympics in retrospect: game over for the bonanza story? Economic Papers: A Journal of Applied Economics and Policy, 30(2), 218–232. Hotchkiss, Julie L., Robert E. Moore, & Fernando RiosAvila (2015). Reevaluation of the employment impact of the 1996 Summer Olympic Games. Southern Economic Journal, 81(3), 619–632. Hotchkiss, Julie L., Robert E. Moore, & Stephanie M. Zobay (2003). Impact of the 1996 Summer Olympic Games on employment and wages in Georgia. Southern Economic Journal, 69(3), 691–704. Humphreys, Brad R., & Dennis R. Howard (2008). The Business of Sports. Vol. 1: Perspectives on the Sports Industry. Tunbridge Wells, UK: ABCCLIO. Maennig, Wolfgang, & Felix Richter (2012). Exports and Olympic Games: is there a signal effect? Journal of Sports Economics, 13(6), 635–641. Rose, Andrew K., & Mark M. Spiegel (2011). The Olympic effect. The Economic Journal, 121(553), 652–677. Scandizzo, P. L., & Pierleoni, M. R. (2018). Assessing the Olympic Games: The economic impact and beyond. Journal of Economic Surveys, 32(3), 649–682. Zimbalist, A. (2016). Circus Maximus: The economic gamble behind hosting the Olympics and the World Cup. Brookings Institution Press.

39 Economic Impact of Minor Sporting Events and Minor League Teams N o l a A g h a a n d M a r i j k e Ta k s

A considerable amount has been written on mega sporting events, large events, and professional sports, despite the fact that there are many more minor events and minor league teams (development teams) across the globe. To illustrate the scope of minor events, nearly every sizeable city in North America has a convention and visitor’s bureau (CVB), a destination management organization (DMO), or sport authority whose purpose is to bring these, mostly minor, sporting events or teams to the city. Since many more minor events and teams exist compared to professional teams and megaevents, one would expect a greater focus on these minor teams and events to better understand their economic effects. The tourism, leisure, and recreation literature is replete with studies in the context of traditional minor events such as soccer tournaments, hockey championships, and swim meets (e.g., Cela, Kowalski, & Lankford, 2006; Veltri, Miller, & Harris, 2009; Wilson, 2006). These studies are mostly ex ante using the direct expenditure approach (DEA) (Davies, Coleman, & Ramchandani, 2013) or input-output models, only look at benefits, and do not subtract out costs. On the other hand, research on minor events in the economics literature is rare and when it focuses on small stadiums and arenas (Colclough, Daellenbach, & Sherony, 1994; Hodur, Bangsund,

Leistritz, & Kaatz, 2006) and minor events (Ryan & Lockyer, 2001) it often takes the same DEA or ex ante, benefit-only approach. An exception to this approach is Taks, Késenne, Chalip, Green, and Martyn (2011), who contrasted and compared the outcomes of a DEA with a cost–benefit approach (CBA) for a medium-sized international sport event. For teams, the only exception is Agha (2013), who used an ex post regression approach to identify the effect of Minor League Baseball teams. The dearth of research on minor events and teams in the economic literature can be explained in various ways. To begin, there is less motivation to conduct the studies because minor sports are assumed to have smaller significance (Marsh, 1984) and generally lack status or prestige that draws attention and interest (Agha & Coates, 2015). In addition, economists generally prefer the ex post approach to economic impact as it more naturally controls for net effects (e.g., Agha, 2013) but data are harder to obtain for smaller events (Coates, 2012). Baade, Baumann, and Matheson (2008) pointed out that it is tough to find the effect of a large championship in a large economy due to normal fluctuations in the regional economic activity. By extension, it would be hard to find an effect of a small event in a small economy. Figure 39.1 illustrates not just the difficulty of identifying

396

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Large City

Small City

Large

Hard to find Multiple ex post attempts have been made

Likely can find

Small Event/ Team

Likely cannot find

Hard to find One ex post attempt has been made

Figure 39.1  Potential to find impact of events/teams using the ex post regression method impact in the ex post approach, but also helps to explain why so little research has been conducted on smaller events and teams. Despite these reasons to overlook inquiry on minor sports in the economic literature, there has been a long history suggesting that minor events are likely to be more beneficial than major sports (e.g., Coates, 2012; Daniels & Norman, 2003; Marsh, 1984; Matheson, 2006; Taks, 2013; Walo, Bull, & Breen, 1996). In part due to both increased data availability and a shift towards cost–benefit analysis (CBA) approaches (e.g., Késenne, 2005, 2012), recent efforts to develop theory specific to minor events (e.g., Agha & Rascher, 2016; Agha & Taks, 2015) and to conduct studies on minor events (Taks et al., 2011) and teams (Agha, 2013) using more robust approaches occur in the literature. Thus, we focus for the remainder of the chapter on the more recent theoretical and empirical contributions that take either the CBA or the ex post approach for both minor sport events and minor teams. We start by defining minor sport events and teams followed by an explanation of the different approaches to measuring impact. We then present theoretical models and provide empirical examples of economic impact of minor sport events and leagues. We end the chapter with recommendations for future research.

DEFINING MINOR SPORTING EVENTS AND TEAMS It is easy to distinguish minor league teams from their major league counterparts, simply based on their labeling (i.e., minor league versus major league). Similarly, there is consensus that the Summer Olympic Games and the FIFA World Cup are mega-events. Researchers, however, do not agree if other large events, such as the Commonwealth Games, EURO Football Cup,

and Winter Olympics are mega sport events (e.g., Getz, 2012) with them sometimes referred to as ‘second tier events’ (Grix, 2014). It is less clear what is understood by a minor sporting event and, thus far, the classification has generally been up to the researcher. Researchers refer to the following events as ‘small’ events: the Giro d’Italia, Windsurf World Cup, and Ski Jumping World Cup (Kwiatkowski, 2016); a North Dakota State University football game with 5,100 spectators (Hodur et  al., 2006); the Cooper River Bridge Run/Walk and the National Softball Association Girls Fastpitch World Series (Daniels & Norman, 2003). Several definitions of small events exist. Saayman and Saayman (2014) defined events based on seven variables, one of which is size. Higham (1999) suggested they could be ‘regular season sporting competitions (ice hockey, basketball, soccer, rugby leagues), international sporting fixtures, domestic competitions, Masters or dis­ abled sports, and the like’ (p. 87). This expansive definition spans from local to international events and from regular season games to one-off events. Gratton and Taylor (2000) proposed a typology of four different types of events (Types A, B, C and D) specifically for economic impact purposes; the smaller Type C and Type D events also include an extremely broad range of events. Wilson (2006) extended the model and suggested a smaller Type E category to cover regional and local events. Based on this, Ramchandani (2014) classified Type A events as ‘Mega’, Type B and C events as ‘Sub Mega’ (National and International), Type D and E as ‘Domestic’. Unfortunately, these typologies do little to help us understand what is minor and major because they focus on event outcomes (media coverage or assumed impact) instead of the event features. Taking a different approach, Daniels and Norman (2003) studied seven small events and found the unifying features of minor events were a lack of government investment and a reliance on existing infrastructure.

Economic Impact of Minor Sporting Events and Minor League Teams

Ultimately, none of these typologies and suggested definitions are based on economic theory. To address this problem, Agha and Taks (2015) developed a definition based on resource demands. They defined Event Resource Demand (ERD) as the continuum of resources needed to stage an event. Specifically, events require investments of three resource types: human (e.g., employees, volunteers), financial (e.g., private or government investments), and physical (e.g., venues, accommodation, transportation). As a result, events are operationalized by the multivariate combination of demanded resources where ‘large’ events are defined as those with high ERD and ‘small’ events as those with low ERD. There are an infinite number of events falling on the ERD continuum between the largest and smallest events. To put this definition into context, Lamla, Straub, and Girsberger (2014) investigated the UEFA EURO 2008 football competition which hosted over 1 million game attendees in four stadiums in Austria and four stadiums in Switzerland. For this pan-continental championship, no new stadiums were built and no new hotels were necessary to host the event (suggesting a small ERD in terms of physical resources). In addition, the host cities were common tourism destinations with available employees in the hospitality sector (suggesting a small ERD in terms of human resources). On the other hand, the financial resource demand was large, with Swiss taxpayers alone responsible for over US$130 million (SwissInfo, 2005). Given the ERD continuum, the EURO football championship was not the largest or smallest event, but somewhere in between (see Figure 39.2). While the prior discussion focused on events, it is worth noting that minor league teams have traditionally been defined very differently – simply as those teams not at the highest level of competition. It would be incredibly useful to apply the ERD definition to development teams to accurately capture the resource demands they place on host cities. While there are generally low human resource demands of minor league teams, a new minor league baseball stadium can have a large ERD

397

if it places large financial demands on the local government to produce the physical resources. For instance, Ramapo, NY, a small town of 127,000 inhabitants, issued $25 million in bonds to build a minor league baseball stadium in 2011 that ultimately cost the city $60 million (Klopott, 2013).

APPROACHES TO MEASURING IMPACT Minor sporting events create an impact if they are out of the ordinary (i.e., not part of the regular season), thus create a temporary ‘shock’ in the economy (Taks, Chalip, & Green, 2016). Generally, they require a whole range of resources, including products, services, and facilities for a very short period of time, thus increasing consumption in the local economy. As mentioned previously, a DEA measures this increased demand by focusing on the new money coming in to the local economy. While a DEA occasionally corrects for leakages (i.e., money not spent locally), it generally neglects the cost, and completely neglects the opportunity cost (e.g., diverting investments from other projects, crowding-out regular tourists, etc.). By only considering the positive impacts while ignoring the negative impacts, the standard DEA constantly overestimates the ‘economic benefits’ (Késenne, 2012). This overestimation is found by comparing ex post to DEA results (e.g., Baade & Matheson, 2001). However, no such ex post challenge of DEA has been conducted in minor events because no ex post research exists for minor events. Sport economists’ reliance on ex post regression began with Baade and Dye (1990) and is based on externalities. If events and teams produce positive production externalities in a local economy they will, theoretically, manifest themselves through pecuniary effects. These effects have been operationalized and observed through standard economic variables such as spending, sales tax collections, income, and employment.

Figure 39.2  Event Resource Demand (ERD) continuum

398

THE SAGE HANDBOOK OF SPORTS ECONOMICS

These pecuniary effects are then identifiable, measurable, and econometrically testable. The bulk of economic research to date relied on this approach but focused entirely on larger and medium-sized events, with the exception of Agha (2013) who studied minor league teams. One benefit of an ex post regression approach is it more easily captures net effects, especially since costs, leakages, and crowding-out are very difficult to identify and measure with a DEA. In part, this explains the stark differences between ex ante and ex post estimates of impact (e.g., Baade & Matheson, 2001). Ex post analysis also makes sense for major events and teams because they are the most likely to have production externalities. Smaller events have smaller production externalities and when organized in larger cities it becomes close to impossible to find the impact of smaller events ex post (e.g., a swim meet in a city of 1 million people – a tiny event, which will hardly make a difference in a large city). Thus, in general, ex post analysis makes little sense for minor events. More recently, calls for alternative methods of measuring economic impact of events, such as CBA (e.g., Késenne, 2005) or computable general equilibrium (CGE) (e.g., Dwyer, Forsyth & Dwyer, 2010), were made, which revealed more realistic (and often negative) outcomes (e.g., Taks et al., 2011). A CBA is concerned with net benefits for the local population (i.e., welfare economics), whereby the best option to improve the efficiency of resource allocation is the one in which the marginal social benefit exceeds the marginal social cost by the largest amount. CBA identifies which money flows in a standard economic impact study should be considered a cost and which are a benefit. It also identifies the value of intangible social benefits and social costs not reflected in market prices, such as consumer surplus (e.g., Falconieri & Palomino, 2004), public good value (e.g., Johnson & Whitehead, 2000), and opportunity costs (e.g., Késenne, 2012). Opportunity costs represent foregone earnings from spending public money on sport events. For example, money spent on sport facilities could alternatively have been used to build a school or a hospital. The return on investments (ROI) on these alternative projects could be higher than the ROI of the sport facility. It is very challenging to estimate the ROI of all possible alternatives. Thus, CBA estimates opportunity costs on the basis of crowding-out effects (e.g., regular tourists, local businesses) and all government expenditures related to the event (Taks et al., 2011). In criticizing the CBA approach to impact, Davies et  al. (2013) stated, ‘…CBA is arguably too data intensive from a practitioner perspective,

especially for medium-sized Type C and D events, and given increasing constraints with public sector funding across many countries, is unlikely to be adopted by event organisers and local governments as a regular tool for evaluation’ (p. 34). While data collection is a challenge for any attempt to measure impact (Wilton & Nickerson, 2006), we note that as soon as events do not require public subsidies, which is more likely the case with small events, there are, technically speaking, no opportunity costs. There still might be leakages and crowding-out effects if the local economy is at full capacity, and technically they should be identified, calculated, and subtracted from the benefits. In the absence of public subsidies, no data must be collected to calculate opportunity costs based on taxpayers’ dollars. Therefore, the overall economic impact is highly likely to be positive. In contrast to Davies et al. (2013), we argue that the CBA is a better approach than both DEA and regressionbased ex post analysis for small events with no government subsidies.

THEORETICAL IMPACT OF MINOR SPORTING EVENTS AND TEAMS A typical DEA or ex ante approach to economic impact requires hundreds or thousands of variables. Simplistic examples include the number of visitors, how long they stay, the amount spent in different industries, and an organizing committee’s budget. In taking a CBA approach, Agha and Taks (2015) noted that these variables could be simplified and categorized into 10 economic impact drivers: five that increase and five that decrease economic impact (see Figure 39.3). While a DEA generally focuses on B1 (and occasionally captures B2–B5 and C4) a CBA accounts for all 10 drivers. Note that some drivers are a function of the type of event (e.g., B1, New spending spent locally by visitors) while others are a function of city characteristics, such as normal tourism rates and the available hotel stock (e.g., C1, Crowding out other visitors). Seeing the necessity of defining cities in the same terms as events, Agha and Taks (2015) extended the idea of ERD by defining City Resource Supply (CRS) as the resources available in the host city to stage the event, including the human resources (supply of labor and volunteers), financial resources (public and private investments), and physical resources (infrastructure such as transportation, venues, and accommodation). Defining a city in terms of supply allows for local economic conditions and the reality that

Economic Impact of Minor Sporting Events and Minor League Teams

399

Figure 39.3  Economic impact drivers (Agha & Taks, 2015, p. 203) cities with excess capacity will benefit more than fully productive economies (Baade & Sanderson, 1997). Traditional city characteristics, such as population or GDP, are less relevant in this multi­ variate continuum of CRS because it is possible for a city with a smaller population to have better transportation, lodging, and venue options than one with a larger population (e.g., a small city which is a popular tourism destination). Similarly, some cities are well off financially and may have financial surpluses, regardless of their population size. In this conceptualization of cities, ‘large’ cities have high CRS and ‘small’ cities have low CRS in the context of events. An infinite number of cities fall on the CRS continuum between largest and smallest city. Economic impact is then a function of the interaction of the ERD and the CRS. In short, it is often the relative size of the event as a function of the city (or better, the city’s resources) that matters most. Wilson (2006) in the context of swim meets in the UK, Coates and Depken (2011) in terms of American college football games, and Coates and Agha (2015) in the context of Minor League Baseball support this point. We can see this idea of relativity expressed as the interaction of ERD and CRS in Figure 39.4. Whereas the demanded resources for event 1 (E1) exactly match the resource supply of city 1 (C1), city 2 (C2) actually has a surplus of resources to host event 1. On the other hand, Agha and Taks

(2015) introduced the idea of resource deficiency to illustrate that the lack of local resources often leads to a realization of zero or negative impact, as when event 2 (E2) is held in city 1 (C1). Because CRS is a multivariate measure, the deficiency (D1) could be too few hotel rooms for visitors, too few venues, or a lack of financial resources. Regardless of the specific deficiency, there will be a local cost to obtain them which decreases economic impact. Thus, only an equilibrium between ERD and CRS will lead to an optimal economic impact as in points O1 and O2 in Figure 39.4. High E2 ERD

E1

O2

D1 Deficiency CRSERD

Low

C1 CRS

C2

High

Figure 39.4  Optimum economic impact where Event Resource Demand (ERD) equals City Resource Supply (CRS) (adapted from Agha & Taks, 2015, p. 210)

400

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Using the concept of resource deficiency, and bringing local economic conditions into the analysis, Agha and Taks (2015) demonstrated that theoretically: (1) no city has the resources required to host a mega-event and will therefore never achieve the optimal economic impact; (2) smaller events have a higher potential for maximum optimal economic impact compared to larger events; and, (3) smaller events have positive impacts in many more cities than larger events. To see this interplay between ERD and CRS in action, we return to the example of EURO 2008. Given the definition of ERD, EURO 2008 demanded eight football stadiums, of which all eight were locally available. Given the strong tourism infrastructure in Europe, no new hotels were necessary although there was not necessarily slack in those establishments in the summer months leading to a resource deficiency, the result of which was crowding-out of other visitors which decreases impact. Lamla et al. (2014) found hotels and restaurants reported lower sales due to crowdingout. The financial ERD of over $130 million had real opportunity costs for the host cities. The deficiencies in some of the resources and the presence of clear cost drivers suggest a non-optimal economic impact. One could imagine EURO 2008 located at point D1 in Figure 39.4. Contrast the resource deficiency of EURO 2008 with the International Tennis Federation (ITF) women’s professional tennis tournament, the 2013 GDF-Suez Open in Seine-et-Marne, France. Despite its status as an international event, Schut and Pierre (2016) reported the use of an existing tennis facility and hotel complex, suggesting an optimal level of physical resources. Little to no crowding-out occurred because 91% of spectators lived in the local area and ‘stayed there for a few hours’ (Schut & Pierre, 2016, p. 77). In terms of financial demands, the Seine-et-Marne Department Council paid €50,000 to subsidize the event. With unemployment in France near 10%, there is supply to match the demand for human resources. Thus, the overall ERD seems likely to be below the CRS and closer to the optimal point (O1) than EURO 2008. Looking at the interaction of ERD and CRS in the context of minor league teams, we see a similar pattern whereby some minor teams can exceed the capacity of their cities even though they are small. Returning to the example of Ramapo, NY, the financial demand of the team for the stadium was in excess of the CRS. The per capita cost of $472 to build the stadium vastly exceeded the average per capita MLB stadium cost of $79 (Agha & Coates, 2015). The city of Ramapo is now fiscally stressed, and Moody’s issued a negative outlook on the debt (Klopott, 2013). This is illustrated by

point D1 in Figure 39.4. In contrast, the San Jose Giants are a minor league baseball team (Class A) located in San Jose, CA, a city with a population over 1 million inhabitants. The team employs 25 full-time personnel year-round, about 265 part-time seasonal, four full-time paid interns, and no unpaid volunteers (C. Seike, 2017, personal communication, May 10, 2017). The stadium requires approximately $100,000 in annual maintenance, and the few out-of-town visitors are easily accommodated in existing hotels. Given the CRS, the team is best represented by point S1 in Figure 39.4.

APPLICATION AND OUTCOMES FOR SMALL EVENTS All the theories and concepts discussed thus far relate equally to major and minor events including CBA, the economic impact drivers, and the ERD/CRS framework. Although the theories apply equally to all events, the impacts for minor events and teams are more consistently positive than those of major events and teams. There are a variety of reasons for this, some of which extend naturally from the ERD/CRS framework. Simply by the nature of their ERD and the number of cities with available CRS, minor events have a lower likelihood of exceeding local capacity (Agha & Taks, 2015), including a lower likelihood of public subsidies for infrastructure (Agha & Rascher, 2016; Higham, 1999), security (Matheson, 2006), and bidding costs (Higham, 1999) (driver C5). Available capacity also means minor events have a lower likelihood of crowdingout (Agha & Rascher, 2016; Matheson, 2006) (drivers C1, C2, and C3), a result that runs contrary to major events. Taks et  al. (2011) and Matheson (2006) state minor events are less likely to influence changes in normal business activity (including both positive changes in driver B2 and negative changes in driver C3) and thus are less likely to affect competing industries, multipliers, and exchange rates. Whereas the primary explanations for neutral and negative effects of major teams are substitution and leakages (e.g., Siegfried & Zimbalist, 2000), Agha and Rascher (2016) suggested that minor league teams have lower leakages (driver C4) and a higher propensity for new visitor spending to be captured locally (driver B1; what Ryan and Lockyer (2001) refer to as retained expenditures). Overall, the theoretical explanations for the differences are consistent with empirical findings of both minor teams and events.

Economic Impact of Minor Sporting Events and Minor League Teams

EMPIRICAL EXAMPLES OF MINOR SPORTING EVENTS AND TEAMS Despite Matheson’s (2006) call for more ex post analysis of ‘less prominent sporting events,’ (p. 194), little work has been done, in part because, as Figure 39.1 indicates, it is not an easy task to identify the effect. In this section we provide details of two empirical studies.

Minor Events Taks et  al. (2011) compared the outcomes of a standard economic impact analysis (EIA based on DEA and input-output modeling) with a CBA for the Pan American Junior Athletic Championships. While one-off, and international in nature, this event is considered a non-mega sport event. The 2005 edition was hosted in Windsor, Ontario, a medium-sized city in Canada of approximately 250,000 inhabitants. Thirty-five countries were represented by 443 athletes and 144 coaches. Most of the 600 volunteers were residents. The event attracted a substantial amount of local media attention, as if the Olympic Games were in town.

401

It drew 16,000 spectators to the stadium over the course of the four-day event. Most of the spectators were residents, while the competitors and participants were almost exclusively non-locals. This event is a rare example of a non-mega sport event for which a new $9.5 CND stadium was built (on University premises). Private funds covered 75% of the cost and the remainder through increased student fees. The results are presented in Tables 39.1 and 39.2. The DEA of the event indicated $11,023,162 in new spending (visitors: $971,759; capital: $9,506,883; and, operational spending: $544,521). After correcting for leakages (i.e., some of this money was re-spent locally, while the rest was re-spent outside the host region), the net increase in economic activity in Windsor was estimated to be $5,617,681. Furthermore, the event generated a total of 75.8 Full-Time Job Equivalents for the city. The total impact from wages and salaries was estimated to be $3,396,524. The example clearly demonstrates that even when standard EIA are corrected for leakages, the outcome is always positive, because it does not take into account the costs for hosting the event (Table 39.1). The benefit side of the CBA includes the nonlocal visitor spending, the revenue of the local

Table 39.1  Results from the Standard Economic Impact Analysis of the 2005 Pan American Junior Athletic Championships (adapted from Taks et al., 2011, p. 193): Economic Impact Summary – Combined Total (Visitor – Operational – Stadium) for the City of Windsor in $ CDN (results from the STEAM model; Canadian Sport Tourism Alliance, 2006) Initial expenditure:   Visitor spending  Organization  Construction Total initial expenditure Net increase in GDP Employment (full-year jobs) Wages and salaries

$ 971,759 $ 544,521 $9,506,883 $11,023,162 $  5,617,681   75.8 $   3,396,524

Table 39.2  Results from the cost–benefit analysis (in $ CDN) of the 2005 Pan American Junior Athletic Championships (adapted from Taks et al., 2011, p. 195) Benefits

Costs

Non-local Visitor Spending 971,759 LOC-Revenue 564,878 Consumer Surplus 39,944 Public Good Value 530,000 Total-Benefits (B)= 2,106,581 Net Benefit (B − C) = −2,421,676

Opportunity Cost of Labor Opportunity Cost of Borrowing Imports (indirect) Ticket Sales to Locals Total-Costs (C)=

0 2,500,000 1,948,368 79,889 4,528,257

402

THE SAGE HANDBOOK OF SPORTS ECONOMICS

organizing committee (LOC), the consumer surplus, and the public good value. The cost side consists of opportunity costs, which included: (1) the costs for building the stadium (labor costs, the cost of borrowing); (2) imports; and (3) ticket sales to locals. Money spent on building the new stadium, crowds-out other projects. The imports are considered a cost, as money flows out of the local economy because of the event (the numbers were retrieved from the standard EIA which provided numbers on imports). Ticket sales to locals crowd-out local businesses (e.g., movie theatre, bars, restaurants). This is particularly problematic in cases where the organizing committee takes its profits outside the host community (e.g., the IOC in the context of Olympic Games, or the FIFA in the context of the World Cup Soccer; Késenne, 2005). When subtracting the overall costs of approximately $4.5 million from the overall benefits of approximately $2.1 million, the outcome is a negative net benefit of $2.4 million (Table 39.2). What is important here is how the positive signs from the standard EIA revert into a negative sign when a CBA is executed for the same event, from an estimated net increase in economic activity in the City of $5,617,681, to a negative net benefit of $2.4 million – a discrepancy of about $7 million.

Minor Teams In order to investigate the effect of minor league baseball teams and stadiums on local per capita income, Agha (2013) relied on an ex post approach taken by Coates and Humphreys (1999) on major league teams and stadiums. To find the effect of smaller teams, she pooled data on 238 different metropolitan areas over 27 years. The bulk of the team and stadium effects for each classification were insignificant, which aligned with decades of results at the major league level. There were two important differences though. First, whereas there were known negative effects in major league results, there were no significant negative effects at the minor league level. Second, positive effects were found in four cases: teams at the AAA and A+ classifications and stadiums at the AA and rookie classifications. The results were particularly surprising because, as the first ex post investigation of any type of minor league team, the a priori expectations were that the results would be negative. Teams are small businesses that have shorter seasons, more frequent moves between cities, seasonal employees, and no national media exposure. Furthermore, the leagues in which the teams play have no national revenue sharing and they fold with much higher frequency than do

major leagues. Agha (2013) concluded, ‘Although these are undeniable features of minor league baseball, they are simply descriptive features of the product. It is faulty to assume they are sufficient to explain the relationship between the presence of a team and per capita income’ (p. 245). Instead, the reasons given for these positive results included little or no crowding-out, low leakages, and a higher likelihood of retained expenditures – explanations that align with the ERD/CRS framework.

CONCLUSIONS AND FUTURE RESEARCH When events do not exceed the resource capacity of their host cities there is greater potential for a host of other benefits as evidenced by the lengthy literature on the social benefits, quality of life, and network effects of minor events (e.g., Taks et al., 2016). As Walo et al. (1996) stated, ‘enhancement of the host population’s way of life, economy, and environment is possibly the most significant difference between local special events and largescale events’ (p. 104). If crowding-out is one reason why economic impact of large events is non-positive, then multiple smaller events will likely bring greater benefits than one large event (Matheson, 2006). This comment is consistent with the literature on optimizing event outcomes with strategic planning of an event portfolio (e.g., Ziakas & Costa, 2011). There is more to learn about minor events and teams. We encourage more comparative studies using a CBA approach rather than a DEA approach to better understand the features of events that increase their likelihood of benefiting an economy. CBA is especially recommended for minor events when there is no government investments, as this voids the need to calculate an important opportunity cost. More research is also necessary on minor league teams in sports beyond baseball (e.g., hockey, soccer, lacrosse) to capture the resource demands they place on host cities. This research area is increasingly important in the context of relativity as minor league facility costs can have major impacts on small cities. In addition, minor league teams affect thousands of cities across the globe compared to only a few hundred major league cities. Looking at the drivers of economic impact (Figure 39.3), there has been considerable attention paid to all the benefit drivers except for B2, increased spending (spent locally) by residents and businesses. Although we see claims that

Economic Impact of Minor Sporting Events and Minor League Teams

this occurs, the evidence thus far seems to suggest that increased local spending is simply time shifted (Agha & Taks, 2018). If major events do not affect B2, then it is even less likely for minor events, a point in alignment with Higham (1999) that minor events should have negligible impacts on residents. More inquiry is also necessary on the crowding-out effects (drivers C1, C2, and C3). In conclusion, calculating the economic impacts of events and teams remains a challenge and is often incomplete. In this contribution, we stressed the importance of going above and beyond direct expenditure by taking costs into account. Consistent with other sports economists, we strongly recommend performing CBA over DEA. Moreover, instead of defining event sizes in terms of outcomes, we defined events (and teams) in terms of resources needed (ERD) and combined this with the resource capacity of the host city (CRS) to better understand how events (or teams) can achieve optimal economic impact in the city where they are being hosted. There are no absolute sizes of events; instead it is the equilibrium of resources an event demands relative to the resources a city can supply that determines economic outcomes. Any event operating within the existing resource capacity of the host city will have low opportunity costs, higher community benefits, and more optimal economic impact. We demonstrated the greater likelihood for minor sporting events and minor league teams to operate within those parameters compared to their major counterparts.

REFERENCES Agha, N. (2013). The economic impact of stadiums and teams: The case of minor league baseball. Journal of Sports Economics, 14, 227–252. Agha, N., & Coates, D. (2015). A compensating differential approach to valuing the social benefit of minor league baseball. Contemporary Economic Policy, 33, 285–299. Agha, N., & Rascher, D. A. (2016). An explanation of economic impact: Why positive impacts can exist for smaller sports. Sport, Business and Management: An International Journal, 6, 182–204. Agha, N., & Taks, M. A. (2015). A theoretical comparison of the economic impact of large and small events. International Journal of Sport Finance, 10, 199–216. Agha, N., & Taks, M. (2018). Modeling resident spending behavior during sport events: Do residents contribute to economic impact? Journal of Sport Management, 32(5), 473–485.

403

Baade, R. A., Baumann, R., & Matheson, V. A. (2008). Selling the game: Estimating the economic impact of professional sports through taxable sales. Southern Economic Journal, 74, 794–810. Baade, R. A., & Dye, R. F. (1990). The impact of stadium and professional sports on metropolitan area development. Growth and Change, 21(2), 1–14. Baade, R. A., & Matheson, V. A. (2001). Home run or wild pitch? Assessing the economic impact of Major League Baseball’s All-Star Game. Journal of Sports Economics, 2, 307–327. Baade, R. A., & Sanderson, A. R. (1997). The employment effect of teams and sports facilities. In R. Noll & A. Zimbalist (Eds.), Sports, jobs and taxes: The economic impact of sports teams and stadiums (pp. 92–118). Washington, DC: Brookings Institution. Canadian Sport Tourism Alliance [CSTA]. (2006). Canadian Sport Tourism Alliance: Building business through sport. Retrieved from: www.canadiansporttourism.com Cela, A., Kowalski, C., & Lankford, S. (2006). Spectators’ characteristics and economic impact of local sports events: A case study of Cedar Valley Moonlight Classic Soccer Tournament. World Leisure Journal, 48(3), 45–53. Coates, D. (2012). Not-so-mega events. In W. Maennig & A. S. Zimbalist (Eds.), International handbook on the economics of mega sporting events (pp. 401–433). Cheltenham, UK: Edward Elgar. Coates, D., & Depken, C. A. (2011). Mega-events: Is Baylor football to Waco what the Super Bowl is to Houston? Journal of Sports Economics, 12, 599–620. Coates, D., & Humphreys, B. R. (1999). The growth effects of sport franchises, stadia, and arenas. Journal of Policy Analysis and Management, 18, 601–624. Colclough, W. G., Daellenbach, L. A., & Sherony, K. R. (1994). Estimating the economic impact of a minor league baseball stadium. Managerial and Decision Economics, 15, 497–502. Daniels, M. J., & Norman, W. C. (2003). Estimating the economic impacts of seven regular sport tourism events. Journal of Sport Tourism, 8, 214–222. Davies, L., Coleman, R., & Ramchandani, G. (2013). Evaluating event economic impact: Rigour versus reality? International Journal of Event and Festival Management, 4(1), 31–42. Dwyer, L., Forsyth, P., & Dwyer, W. (2010). Tourism economics and policy. Bristol, UK: Channel View. Falconieri, S., & Palomino, F. (2004). Collective versus individual sale of television rights in league sports. Journal of the European Economic Association, 2, 833–862. Getz, D. (2012). Event studies: Theory, research and policy for planned events. London: Routledge. Gratton, C., & Taylor, P. (2000). Economics of sport & recreation. London: E & FN Spon.

404

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Grix, J. (Ed.) (2014). Leveraging legacies from sports mega-events: Concepts and cases. New York: Palgrave Macmillan. Higham, J. (1999). Commentary – sport as an avenue of tourism development: An analysis of the positive and negative impacts of sport tourism. Current Issues in Tourism, 2(1), 82–90. Hodur, N. M., Bangsund, D. A., Leistritz, F. L., & Kaatz, J. (2006). Estimating the contribution of a multi-purpose event facility to the area economy. Tourism Economics, 12(2), 303–316. Johnson, B., & Whitehead, J. (2000). Value of public goods from sports stadiums: The CVM approach. Contemporary Economic Policy, 18, 48–58. Késenne, S. (2005). Do we need an economic impact study or a cost-benefit analysis of a sports event? European Sport Management Quarterly, 5, 133–142. Késenne, S. (2012). The economic impact, costs and benefits of the FIFA World Cup and the Olympic Games: Who wins, who loses? In W. Maennig & A. S. Zimbalist (Eds.), International handbook on the economics of mega sporting events (pp. 270–278). Cheltenham, UK: Edward Elgar. Klopott, F. (2013). FBI Ramapo probe shows risks of Minor-League stadium boom. BloombergMarkets. Retrieved from: www.bloomberg.com/news/ articles/2013-12-23/ramapo-fbi-probe-shows-risksof-minor-league-stadium-boom Kwiatkowski, G. (2016). Composition of event attendees: A comparison of three small-scale sporting events. International Journal of Sport Finance, 11, 163–180. Lamla, M. J., Straub, M., & Girsberger, E. M. (2014). On the economic impact of international sport events: Microevidence from survey data at the EURO 2008. Applied Economics, 46, 1693–1703. Marsh, J. S. (1984). The economic impact of a small annual sporting event: An initial case study of the Peterborough church league atom hockey tournament. Recreation Research Review, 11(1), 48–55. Matheson, V. A. (2006). Is smaller better? A comment on ‘comparative economic impact analyses’ by Michael Mondello and Patrick Rishe. Economic Development Quarterly, 20, 192–195. Ramchandani, G. (2014). Economic, sport development and elite performance consequences of sports events. Doctoral dissertation, Sheffield Hallam University, Sheffield, UK. http://shura.shu. ac.uk/9165/

Ryan, C., & Lockyer, T. (2001). An economic impact case study: The South Pacific Masters’ Games. Tourism Economics, 7, 267–275. Saayman, M., & Saayman, A. (2014). Appraisal of measuring economic impact of sport events. South African Journal for Research in Sport, Physical Education and Recreation, 36, 151–181. Schut, P. O., & Pierre, J. (2016). The economic impact of a women’s professional tennis tournament: The example of the GDF-Suez Open of Seine-et-Marne, France. Journal of Policy Research in Tourism, Leisure and Events, 8, 71–86. Siegfried, J., & Zimbalist, A. (2000). The economics of sports facilities and their communities. The Journal of Economic Perspectives, 14, 95–114. SwissInfo (2005). Euro 2008 to cost much more than expected. SwissInfo, 2 October. Retrieved from: www.swissinfo.ch/eng/euro-2008-to-cost-muchmore-than-expected/4764528 Taks, M. (2013). Social sustainability of non-mega sport events in a global world. European Journal for Sport and Society, 10, 121–141. Taks, M., Chalip, L., & Green, B. C. (Eds.) (2016). Impacts and strategic outcomes from non-mega sport events for local communities. London: Routledge. Taks, M., Késenne, S., Chalip, L., Green, B. C., & Martyn, S. (2011). Economic impact study versus cost-benefit analysis: An empirical example of a medium sized international sporting event. International Journal of Sport Finance, 6, 187–203. Veltri, F. R., Miller, J. J., & Harris, A. (2009). Club sport national tournament: Economic impact of a small event on a mid-size community. Recreational Sports Journal, 33(2), 119–128. Walo, M., Bull, A., & Breen, H. (1996). Achieving economic benefits at local events: A case study of a local sports event. Festival Management and Event Tourism, 4(3–1), 95–106. Wilson, R. (2006). The economic impact of local sport events: Significant, limited or otherwise? A case study of four swimming events. Managing Leisure, 11, 57–70. Wilton, J. J., & Nickerson, N. P. (2006). Collecting and using visitor spending data. Journal of Travel Research, 45, 17–25. Ziakas, V., & Costa, C. A. (2011). Event portfolio and multi-purpose development: Establishing the conceptual grounds. Sport Management Review, 14, 409–423.

40 Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans’? Peter Dawson

‘Quite simply the Games are the biggest opportunity sport in this country has ever had. It is one that we must not squander’. (Lord Sebastian Coe, Chairman of the London Organising Committee of the Olympic and Paralympic Games)

Do elite athletes and sport events provide inspiration to the general public? How does this manifest itself? What form does it take? The focus of this chapter is on participation in sport and physical activity and the extent to which elite athletes and elite sport provide inspiration to enact behaviour change amongst the general population.

BACKGROUND A substantial amount of attention has been devoted towards the legacy effects of major sporting events. Whereas urban planners and policy-makers extol the virtues of major sporting events, such as the Olympics and soccer World Cups, as catalysts for urban renewal, economic growth and job creation, the academic literature has, in general, been sceptical over such claims, typically drawing attention to the questionable modelling and assumptions associated with the overall economic impact. As such, attention has increasingly turned towards analysing

the softer, intangible, impacts of these types of events. Central among these appear to be associated with participation in sport and physical activity. It is well established, for example, that physical activity, including sports participation, can improve health through improved respiration and cardiovascular performance as well as psychological improvements in mood and well-being. And there are several studies that have analysed these relationships using subjective health and/or subjective well-being data (e.g. Huang & Humphreys, 2012; Humphreys, McLeod & Ruseski, 2014; Downward & Dawson, 2016). The need to increase physical activity and reduce sedentary behaviour of individuals to improve health and well-being is undeniable. Drawing upon evidence from the UK, successive governments from the early 2000s onwards have used sport as a vehicle to encourage such change. This began with the publication of ‘Game Plan’ in 2002 (DCMS/Strategy Unit, 2002), which emphasised a twin-track approach of growing mass sports participation both directly and indirectly through hosting and being successful at major sporting events. This culminated in the successful strategy of bidding for and hosting the London 2012 Olympic and Paralympic Games. Since then the UK has gone on to host the Commonwealth Games (2013), Rugby League World Cup (2013), Rugby

406

THE SAGE HANDBOOK OF SPORTS ECONOMICS

World Cup (2015), the IAAF World Athletics and World Para Athletics Championships (2017) and, at the time of writing, is due to host the Cricket World Cup (2019), leading to a number of commentators describing this as a ‘golden decade’ for British sport. It was claimed that these events would build on, or at least maintain, the feel-good factor and legacy from the London 2012 Games. A central, if perhaps implicit, assumption embedded within this and many subsequent policy documents, in both the UK and elsewhere, is the conjecture that the general population will be inspired by the performances of elite athletes which will then motivate them to engage in sport and physical activity. This has been described as a ‘demonstration’ effect or ‘trickle down’ effect and refers to the process by which mass sports participation is stimulated by public exposure to the performances of elite athletes. Sport England’s most recent strategy document, Towards an Active Nation (Sport England, 2016), builds on the UK Government’s Sporting Future strategy (HM Government, 2015), which aims to get people from every background to take part in sport and physical activity and suggests that major sporting events can provide an inspirational effect: Since London 2012, England has continued to witness some of the greatest sporting events being contested across the world. Hosting major events not only provides home advantage to our athletes and wide-ranging economic impacts, but our Major Events Engagement Fund has shown that – with careful planning – they can also inspire people to engage in sport through taking part, coaching, volunteering and spectating (Sport England, 2016: 42). Olympic and Paralympic success can provide the inspiration to get people involved in sport and the positive wellbeing and

social development benefits from shared national success. (HM Government, 2015: 44)

It is undeniable that a positive relationship between funding of sport and sporting success exists. As Figure 40.1 reveals, increases in UK Sport funding of sports and their athletes has resulted in significant returns in terms of total medals, including gold medals, at successive Summer Olympic Games over the period 2000 (Sydney Olympics) to 2016 (Rio Olympics). What is less well understood is whether, and to what extent, this success has inspired the general public to participate in sport and physical activity. The remainder of the chapter is structured as follows. The next section outlines a number of analytical frameworks advocated to understand the determinants of demand for participation. Here both economic and broader social science perspectives are considered. A central feature of this section relates to theories of behavioural change, with a particular focus on approaches that have been developed within the physical activity and sport management and sport policy literatures. In the following section, a discussion of the empirical analysis of demonstration and trickle-down effects, which is considered in the context of both survey (subjective) and objectively based data as well as avenues for future research, are considered. The final section offers some concluding remarks.

THEORETICAL AND ANALYTICAL FRAMEWORKS Economic Framework The economic framework is based on the assumption that individuals aim to maximise utility subject to

Figure 40.1  UK performance funding and (Summer) Olympic medals won 2000–2016

Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans’? 407

time and money constraints. The constrained optimisation problem, as applied to the demand for leisure, was initially empirically estimated as a residual from work. As detailed, large-scale data have become available, the optimisation problem has been developed within an explicit choicebased framework that is well established in the economics literature (see Gratton & Taylor, 2000; and Downward, Dawson & Dejonghe, 2009, for detailed reviews). This neoclassical perspective typically draws upon Becker’s (1965) theory of time allocation as well as household production models and has been extended in a number of ways. For example, Cawley (2004) developed a model categorising all 24 hours of a day into five domains – Sleep, Leisure, Occupation, Transportation and Home (SLOTH) – that has been adopted and adapted in a number of studies (e.g. Humphreys & Ruseski, 2011).

Heterodox Perspective As detailed by Downward (2007), a heterodox approach considers the decision to undertake leisure activities, including participation in sport, from a broader social science context. Drawing upon the disciplines of psychology (e.g. Scitovsky, 1976) and sociology (Bourdieu, 1988), more emphasis is placed on how social networks and the environment influence behaviour. As discussed in Gratton and Taylor (2000), Scitovsky’s view is that the demand for sport is driven by concepts such as ‘arousal’ and ‘sensation seeking’ and that individuals seek to obtain an optimal level of stimulation. Scitovsky argues that stimulation is the main driver of leisure demand and participation in sport and physical activity. Of course, stimulation can be incorporated within an economic framework, particularly since stimulation can impact utility (both positively and negatively). However, unlike the economic framework, these broader approaches are considered better at developing an understanding of how preferences are formed in the first place rather than simply treating them as exogenous, as is typically the case in the economic framework. There is now a well-developed international literature examining the determinants of sports participation which have incorporated either or both of these perspectives (Ruseski & Maresova, 2014; Cabane & Lechner, 2016). Broadly speaking, the literature identifies that males are more likely to participate in sport and also do so more frequently than females. Increasing age reduces sports participation but its frequency in some activities can rise with age. Increasing income and higher levels

of socio-economic status also increase participation and its frequency. However, higher levels of education can increase the likelihood of participating in sport but reduce its frequency. Studies also find that household structure can affect sports participation. Typically, being married or a couple, and having children, reduces sports participation, though there is heterogeneity across specific activities by gender. Belonging to an ethnic minority group or being a recent migrant to a country is associated with lower participation. The research also suggests a strong socio-­economic inertia in sports participation.

Theories of Behaviour Change While the economic perspective is useful for understanding the influences on sport and physical activity and heterodox approaches can be helpful in understanding how preferences are formed, neither provides a wholly adequate understanding for the way preferences may change. To consider this, a third strand of theories has been developed. These are collectively referred to as theories of behaviour change. The theories and models that have been developed include social cognitive theory, theory of reasoned action, theory of planned behaviour and the stages of change model, more commonly referred to as the trans­ theoretical model. Here, attention is restricted to the transtheoretical model (TTM) since it is the most widely adopted within the sport and exercise science and sport management and policy literatures (Boardley, 2013). The TTM was developed by Prochaska, DiClemente and Norcross (1992). While it was originally developed to understand the stages and processes of negative health behaviours, such as smoking cessation, it has become more broadly applied for understanding change across a variety of behaviours, including sport and physical activity (see, for example, Marshall & Biddle, 2001). TTM is considered a comprehensive model of behaviour change and includes the following elements: stages of change, processes of change, decision balance, self-efficacy and temptation. The stages of change provide a framework for understanding when change occurs whereas the processes of change provide an understanding of how change occurs (Ramchandani et al., 2015). Figure 40.2 outlines the five stages of change. The first stage is precontemplation. This stage is representative of individuals who have no intention of change (or have not even considered behaviour change). In the context of sport and physical activity, this would include those

408

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Maintenance Action Preparation Contemplation

Precontemplation

Figure 40.2  Stages of change within the transtheoretical model (TTM) (adapted from Prochaska, DiClemente, & Norcross, 1992; Ramchandani et al., 2015) groups identified in the previous section as physically inactive. The second stage is contemplation, whereby the individual has begun to think about a change in behaviour. Preparation is the third stage. In this stage the individual has immediate intentions to change behaviour. The fourth stage, the action stage, is where the individual has initiated a change, but such a change has only occurred recently and not been sustained over a period of time. Only when behaviour change has occurred and been maintained for a prolonged period is the fifth, and final, stage of maintenance reached. Ramchandani et  al. (2015) argue that the first three stages can be categorised as attitudinal and inspirational with the final two stages providing the behaviour change through to actual participation.

TTM and Sport Participation The TTM suggests 10 processes which enact such change. An important process identified at the initial stage of contemplation is consciousness raising. This is particularly pertinent in the context of sport and physical activity since the media exposure associated with events such as the Olympics, for example, increase people’s awareness of sport, the multitude of sports available and how to participate in them. It has been claimed that they may also engender a knowledge of the benefits of participation.

Progression beyond the contemplation stage to the final stages of action and maintenance does, however, hinge upon the attitudes of respondents being able to emulate the performances of elite athletes, and is closely related to the ideas of role models and imitation that has been applied across a range of different areas in economics (Mutter & Pawlowski, 2014). Vicarious experiences leading to positive outcomes associated with watching elite athletes rely on the perceived similarities between the observed (elite athlete) and the observer (viewer) and, it has been argued, will generally only apply to a small minority of people watching, typically those who already have or had (in the case of lapsed participation) some level of participation in sport and physical activity. These shared characteristics might include coming from a same locality or having a similar type of up-bringing and education. For others, observing elite athletes could lead to demotivation since it might strengthen beliefs that positive outcomes from sport are confined to just those performing at or near the elite level. As well as the impact of role models, social interactions and peer effects are also likely to be important. There is evidence that peers can (positively) influence an individual’s sport and physical activity behaviour (Downward & Rasciute, 2016). This may manifest itself through demonstration effects, where the appropriate trigger may be an alternative form of vicarious experience or more directly through social pressure. This is particularly apparent in the context of participation of

Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans’? 409

parents and their children, where the motivation may be driven by the parent (i.e. the child being influenced by the past (or present) behaviour of the parent), by the child (i.e. the child influencing the parent to engage (or re-engage) with the activity), or both. For example, Downward, Hallman and Pawlowski (2014) find a causal link between male parental participation when growing up and participation of their male child, whereas no such affect is observed for female children. What matters, it appears, is that demonstration effects, whether it be through elite athletes or peers, is that a sense of shared characteristics and identity is crucial. Another important construct within the TTM is decisional balance. Decisional balance reflects the individual’s beliefs of the pros and cons of changing their behaviour. In the context of the discussions about emulation, it may be, for example, that any perceived health benefits associated with participation are outweighed by the costs associated with achieving a similar standard of performance. Effective demand may also not be realised due to a number of other (personal) barriers. These are likely to include a lack of free time and the direct and indirect costs associated with the activity. An important and necessary condition for many activities is the availability and accessibility of facilities (Hallman et al., 2012). Theories of behaviour change, including TTM, share many facets with behavioural economics. A particularly relevant concept here is the role played by ‘nudges’. Nudge theory refers to ways to encourage people to make good decisions rather than bad ones. Within the framework of TTM, moving from the attitudinal (pre-contemplation – contemplation) stages to behaviour change (action – maintenance) may be influenced not only by the awareness and inspiration drawn from elite athletes and elite sport, but also by ‘nudges’, which might be triggered by social pressure and social norms, incentives to encourage healthy lifestyles through reward or loyalty schemes or mechanisms which induce behaviour change in order to avoid losses. Behavioural patterns of, for example, gym and health club membership (e.g. DellaVigna & Malmendier, 2006), which are built on models of dynamic inconsistency that include habit formation and projection bias, have recently been extended to the physical activity literature (Humphreys, Ruseski & Zhou, 2015). It would appear, therefore, that simply relying on the event alone to increase engagement in sport and physical activity is not enough. As documented by Misener et al. (2015), leveraging of events through, for example, existing initiatives and education programmes and campaigns which raise awareness of the benefits of participation in

sport and physical activity, may also encourage behaviour change.

DEMONSTRATION EFFECTS AND BEHAVIOUR CHANGE: EMPIRICAL EVIDENCE A ‘demonstration effect’ or ‘trickle-down effect’ occurs whereby performances and achievements of elite athletes inspire the general public to increase their involvement in sport and physical activity. This section explores the empirical evidence associated with these effects in the context of major, non-major and single-sport events from the perspective of both the general population and specific groups. A distinction is also made between subjective studies which draw upon data from surveys and objective data.

Studies Using Subjectively Generated (Survey) Data In the context of major sporting events such as the Olympics, a number of systematic reviews have been carried out. Mahtani et al. (2013) conducted a review of systematic reviews, highlighting the studies of Weed et al. (2009) and McCartney et al. (2010). Both studies indicate little evidence of an increase in sporting activity following the event. A more recent systematic review by Weed et  al. (2015) came to similar conclusions but did suggest that during the pre-Games period there is the potential to increase participation and to reengage lapsed participants. This positive effect associated with the pre-Games period has also been highlighted elsewhere. For example, Dawson (2012), using data from the British Household Panel Survey (BHPS), finds a positive and statistically significant increase in participation in the UK around the time when London was chosen to host the 2012 Olympics. One limitation of these studies is that the analysis is typically conducted at an aggregated, general-population level, which masks potentially important heterogeneity across sub-populations. Analysis at a more disaggregated level, focusing on specific sub-groups (such as age, gender, ethnicity, etc.), can reveal important differences. Differences can arise between the number participating and the frequency and intensity of such participation. It is entirely possible, for example, that a demonstration effect leads to increases in participation frequency but leaves the number of

410

THE SAGE HANDBOOK OF SPORTS ECONOMICS

participants unchanged. This is consistent with the view that those who are already engaged in participation increase their levels of frequency and/or intensity whereas non-participants remain inactive. Opportunities to distinguish between number of participants, frequency and intensity of participation have increased in the UK through the introduction of the Active People Survey (renamed the Active Lives Survey from 2015) and the Taking Part Survey in 2005. The Taking Part Survey is a repeated cross-sectional survey across England of activities associated with sport, culture, heritage and media. It was first undertaken in 2005 and has continued annually since. The Taking Part Survey asks questions about the frequency of participation during the last four weeks in the form of days, hours and minutes for 68 different sports. Some studies have exploited the fact that the primary sampling units are surveyed on a rolling basis over the year and the monthly allocations of sampling units take place following a string pattern that is based on a random starting point. This means that events such as the 2012 Olympics and Paralympic Games can be examined from an event-study, time-series perspective. The socio-economic and demographic variables that are also included in the Taking Part Survey provide opportunities to disaggregate the data by sub-groups (gender, age, ethnicity, region and disability). In the case of population-level perspectives, Downward, Dawson and Mills (2013), find some evidence that the intensity of participation in England increased during the post-Beiing, pre-­ London Games period (2008–2012). Increases in intensity of participation are also found for the 16–25 age group, ethnic minorities and male respondents. On the other hand, in the post-games period, Downward, Dawson and Mills (2016), utilising predictive failure tests for a breakpoint at August 2012, find no evidence that the 2012 Olympic Games shifted behaviour either at the population level or by gender and age-gender groups. The Taking Part Survey also contains questions about attitudes towards the Olympics in the form of sentiment and motivation. In the case of sentiment, Dawson, Downward and Mills (2014) find that the good news associated with Team GB’s largely unexpected medal success in the 2008 Beijing Olympics positively affected sentiment more than ‘bad’ news. This was particularly the case for younger respondents, females and those living in regions outside London. For males and those of middle age, there was a reduction in support in the context of the ‘bad’ news associated with the dramatic increase in costs of staging the Games.

Attitudes of respondents with respect to participation were initially asked via the question, ‘Do you think the UK winning the bid to host the 2012 Olympics has motivated you to do more sport or recreational physical activity?’ changing to ‘Do you think that the UK hosting the 2012 Olympics has motivated you to do more sport or recreational physical activity?’ in the post-Games period. The data reveal a decline in motivation from 7.68% in 2005/06 to 4.33% in 2007/08. This was followed by an increase to 7.52% in 2008/09 and remained stable in the subsequent two periods (2010/11 and 2011/12). A significant spike was observed in 2012, coinciding with the London 2012 Olympics, and this trend has remained stable, with a marginal increase in the post-event period. Despite the increase in motivation in the Games and post-Games period, levels of participation in sport in England have remained largely unchanged. Downward, Dawson and Mills (2013) perform (Granger) causality tests to explore the temporal ordering between motivation and levels/intensity of participation. While the evidence generally suggests that some of the temporal ordering is based on pre-disposition (i.e. those who are already participating in sport and physical activity are more likely to be motivated by the Olympics), some evidence that the Olympics had a positive impact on participation of typically less active participants (females, ethnic minorities and those with long-standing illness or disability) was found. Regarding specific sports, Sportscotland (2004) find that curling participation increased after Great Britain’s Winter Olympic success in 2002. Veal, Toohey and Frawley (2012) analysed participation trends associated with the 2003 Rugby World Cup held in Australia and found that participation rates steadily increased over the pre-event and post-event periods. Considering rugby union and cricket, Weed (2009) suggests a demonstration effect is most likely to occur among current participants, lapsed participants and participant switchers (i.e. the switching of participation between activities among current participants). This view that exposure to an event facilitates increased participation is confined to already active individuals is consistent with Ramchandani et  al. (2015) in their analysis of nine events held in England between 2010 and 2012. An important limitation of these and other studies is that they are typically based on relatively short time-frames, often within one or two years of the event taking place. Even where longer time horizons have been considered, the data seldom track the same individuals over time. The availability of longitudinal data would potentially overcome some of these limitations but many

Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans’? 411

nationally representative surveys in the UK, such as the BHPS, included only limited information on sport and physical activity (Dawson, 2012; Downward & Rasciute, 2016). The Understanding Society Survey – the successor to BHPS – has incorporated more information on sport and physical activity but to date has only done so within two waves (Waves 2 and 5) and, rather disappointingly, there are no immediate plans to incorporate details of sport and physical activity frequency and intensity into future Waves. Despite this, and since 2011/12, the Taking Part Survey has included a longitudinal component in which a sub-sample of adults are re-interviewed annually. Currently three years of data are available, encompassing 4,637 respondents. As documented in Dawson and Downward (2016), while the data reveals there were 12% new participants between the first interview and third interview, 14% were described as former participants (respondents who stipulated that they had participated in the first and/or second interview but no longer participate – this is consistent with the notion of lapsed participants described above). In a similar manner, while 20% of respondents reported increased sports participation, 21% reported less frequent participation. Evidence suggests that younger (16–24 age group) males were more likely to consistently and regularly participate in sport. They are also more likely to live in the least deprived/wealthy achiever areas. Interestingly, new participants were more likely to be from London and to be black or minority ethnic. This is consistent with some of the ideas mentioned earlier that those in closer proximity to the event and exposure to role models were more likely to be inspired. The view that watching elite athletes compete fails to inspire participation among the inactive is supported in a number of qualitative studies. For example, Carter and Lorenc (2013) found that among the respondents from the general public that were interviewed, none expressed a positive identification with elite athletes and elite sport. However, what they did find is that respondents were more receptive in their view that elite athletes could inspire children to engage with sport. Perhaps due to data availability, only a small number of studies have investigated whether demonstration effects are prevalent among children and adolescents. In a study of 12–19 year-olds in Canada, Potwarka and Leatherdale (2016) find no effect at the national level for both male and females but find positive effects for females living in the Vancouver region, host of 2010 Winter Olympics. From a UK-perspective, a Child Taking Part Survey is available and has been administered to

5–15 year-olds in England. Despite the modest sample sizes, Downward and Dawson (2016) find a 4 percentage-point and 1 percentage-point fall in those taking part in sport outside school among 5–10 year-olds and 11–15 year-olds, respectively. Shibli, Kokolakakis and Davies (2014) find a positive association between watching sport on television and participation for the 11–15 year-olds. For the younger group, attending the events in person was more important. In what appears to be the first attempt to track a cohort over the life course, Aizawa et al. (2018) analyse participation rates and trends of people in Japan following the 1964 Tokyo Olympic Games. Specifically, they are interested in whether those respondents aged between 11–19 in 1964 participated more in sport and physical activity compared to older age groups. They also considered the impact of similar age cohorts when exposed to other, arguably lesser, sporting events that took place in Japan over the sample period (e.g. the 1972 Sapporo Winter Olympics, the 1991 World Athletics Championship and the 1998 Nagano Winter Olympics). Controlling for the usual socio-economic and demographic factors, the study finds those aged 11–19 in 1964 participated in sport and physical activity more than other groups and other generations.

Studies Using ObjectivelyGenerated Data A feature of the studies addressed in the previous section is that they are all based on surveys where the data relating to participation in sport and physical activity rely on respondents’ recall ability. The advantage of such surveys is that they are nationally representative and thus generalisable to the overall population. A disadvantage is that the responses about sport participation may suffer from recall bias as well as other cognitive problems associated with, for example, the ordering of questions and scales which introduces measurement error. A few studies circumvent the recall bias issue associated with large, nationally representative surveys by using more objectively measured data in analysing demonstration and trickle-down effects. One such example is data on sport club membership. Membership data have been analysed with respect to football (Frick & Wicker, 2016; Wicker & Frick, 2016), tennis (Feddersen, Jacobson & Maennig, 2009) as well as studies that include membership across a range of, largely Olympic, sports (Weimar, Wicker & Prinz, 2015). In these studies, a dynamic

412

THE SAGE HANDBOOK OF SPORTS ECONOMICS

regression analysis framework is used, with total membership acting as the dependent variable and control variables that include economic variables such as GDP. Factors associated with hosting and success in sporting events as well as the presence of sport stars are typically captured through the use of lagged variables to allow for the fact that effects on membership are not instantaneous. There is some support for demonstration effects in this literature but, consistent with what was found in the section above, the impact varies across sub-groups such as males and females. For example, Wicker and Frick (2016) find that the success of the male national team has a significant positive effect on the number of female club memberships, but success of the female national team has no effect either on male or female membership. This may be due to the relative lack of exposure of women’s sport compared to men’s sport if media exposure and increased opportunities to watch the events are more likely to generate awareness and ultimately inspiration. On the other hand, and consistent with the demotivation effect, Feddersen, Jacobson and Maennig (2009) find a negative effect on tennis membership levels in Germany during the period of success of the German tennis stars Boris Becker, Steffi Graf and Michael Stich. However, and what appears to contradict the demotivation effect, these negative effects persist once these stars have retired. Of course, membership is only an indirect measure of physical activity. Another source of objective data that provide a more direct measure of physical activity are pedometers. Craig and Bauman (2014) use pedometer data to evaluate demonstration effects in the context of the 2010 Vancouver Winter Olympics. Analysing pre-event, event and post-event periods, they find no discernible increase in ambulatory physical activity across children and adolescents in any of these three periods either at national or local level, including the region (British Columbia) within which the Games were held. Another publicly available and under-utilised resource is data from the mass community event Parkrun. Parkrun operates a series of weekly, timed 5-kilometre runs in areas of open space, operating within a safe and supportive environment. It is open to all ages and abilities and is inexpensive. Established in London in 2004, it has grown nationally and internationally. As of July 2018, there were 546 locations in the UK and over 500 locations internationally in Australia, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Malaysia, Norway, New Zealand, Poland,

Russia, Singapore, South Africa and the United States, among others. Parkrun uses electronic timing and barcode technology to generate the results of each event. The finish times and athlete numbers (with their finish position) are uploaded to a server which automatically generates the results tables and statistics on the Parkrun website. The use of the unique runner number allows the website to collate historical data, including personal bests, overall performance and total number of runs. The few studies that have utilised data from Parkrun have tended to do so from a public health intervention perspective, but there is clearly scope to undertake an analysis of the impact of major sporting events, for example, as a further means of using objective data to test for demonstration or trickle-down effects. To the extent that a number of these Parkrun events are well established, so that a significant number of observations can be made, there is scope to undertake analysis which captures both the pre- and post-event periods. The individual longitudinal element also provides opportunities to investigate lapsed and non-regular participants and changes in the performance of regular participants.

CONCLUDING REMARKS This chapter has considered the role of demonstration or trickle-down effects in relation to sport and physical activity. Outlining a number of theories and frameworks and drawing upon a range of empirical studies, using both subjectively- and objectively-driven data, the evidence presented here suggests a lack of support for demonstration effects – the success of elite athletes or hosting elite sporting events does not appear to provide sufficient inspiration to effect population-level sport participation. There is, however, some limited evidence of demonstration effects on motivation and actual participation among typically active participants (males and younger adults), as well as typically less active individuals (females and ethnic minorities). However, more work needs to be undertaken to explore whether these effects last over longer periods of time. Studies should also be conducted to further untangle the mechanisms by which these changes occur and are maintained. Greater consideration should be applied to an analysis of children and adolescents, and longitudinal studies that fully incorporate and track detailed participation data covering both frequency and intensity should also be conducted.

Participation and Demonstration Effects: ‘Couch Potatoes to Runner Beans’? 413

REFERENCES Aizawa, K., Wu, J., Inoue, Y., & Sato, M. (2018). Long-term impact of the Tokyo 1964 Olympic Games on sport participation: A cohort analysis. Sport Management Review, 21(1), 86–97. Becker, G. (1965). A theory of the allocation of time. The Economic Journal, 75, 493–517. Boardley, I.D. (2013). Can viewing London 2012 influence sport participation? A viewpoint based on relevant theory. International Journal of Sport Policy and Politics, 5(2), 245–256. Bourdieu, P. (1988). Program for a sociology of sport. Sociology of Sport Journal, 5, 153–161. Cabane, C., & Lechner, M. (2016). Physical activity of adults: A survey of correlates, determinants, and effects. Jahrbücher für Nationalökonomie und Statistik, 235(4–5), 376–402. Carter, R.V., & Lorenc, T. (2013). A qualitative study into the development of a physical activity legacy from the London 2012 Olympic Games. Health Promotion International, 30(3), 793–802. Cawley, J. (2004). An economic framework for understanding physical activity and eating behaviors. American Journal of Preventive Medicine, 27, 117–125. Craig, C.L., & Bauman, A.E. (2014). The impact of the Vancouver Winter Olympics on population level physical activity and sport participation among Canadian children and adolescents: Populationbased study. International Journal of Behavioral Nutrition and Physical Activity, 11(1), 107. Dawson, P. (2012). Economics of the Olympics. In L. Kahane & S. Shmanske (Eds.), The Oxford Handbook of Sports Economics. Vol. 1: The Economics of Sports. Oxford: Oxford University Press. Dawson, P., & Downward, P. (2016). London 2012: A sporting legacy? In Global Sports Impact Report 2016. London: Sportcal. Dawson, P., Downward, P., & Mills, T. (2014). Olympic news and attitudes towards the Olympics: A compositional time-series analysis of how sentiment is affected by events. Journal of Applied Statistics, 41(6), 1307–1314. DCMS/Strategy Unit (2002). Game Plan: A Strategy for Delivering Government’s Sport and Physical Activity Objectives. London: Department for Culture, Media and Sport. DellaVigna, S., & Malmendier, U. (2006). Paying not to go to the gym. American Economic Review, 96(3), 694–719. Downward, P. (2007). Exploring the economic choice to participate in sport: Results from the 2002 General Household Survey. International Review of Applied Economics, 21(5), 633–653. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports Economics: Theory, Evidence and Policy. London: Elsevier.

Downward, P., & Dawson, P. (2016). Is it pleasure or health from leisure that we benefit from most? An analysis of well-being alternatives and implications for policy. Social Indicators Research, 126(1), 443–465. Downward, P., Dawson, P., & Mills, T. (2013). The impact of the Olympic Games on sports participation, motivation, health and well-being. In 2012 Games mega-evaluation: Report 5 (post-games evaluation) sports evidence base. London: ­Department for Culture, Media and Sport. Avail­ able at: www.gov.uk/government/uploads/system/ uploads/attachment_data/file/224144/Report_5_ Sport_Evidence_Base_FINAL.pdf Downward, P., Dawson, P., & Mills, T.C. (2016). Sports participation as an investment in (subjective) health: A time series analysis of the life course. Journal of Public Health, 38(4), e504–e510. Downward, P., Hallmann, K., & Pawlowski, T. (2014). Assessing parental impact on the sports participation of children: A socio-economic analysis of the UK. European Journal of Sport Science, 14(1), 84–90. Downward, P., & Rasciute, S. (2016). ‘No man is an island entire of itself’: The hidden effect of peers on physical activity. Social Science & Medicine, 169, 149–156. Feddersen, A., Jacobsen, S., & Maennig, W. (2009). Sport heroes and mass participation: The (double) paradox of the ‘German tennis boom’. Working Paper, No. 29. Hamburg. Frick, B., & Wicker, P. (2016). The trickle-down effect: how elite sporting success affects amateur participation in German football. Applied Economics Letters, 23(4), 259–263. Gratton, C., & Taylor, P. (2000). Economics of Sport and Recreation. London: Spon Press. Hallmann, K., Wicker, P., Breuer, C., & Schönherr, L. (2012). Understanding the importance of sport infrastructure for participation in different sports: Findings from multi-level modeling. European Sport Management Quarterly, 12(5), 525–544. HM Government (2015). Sporting Future: A New Strategy for an Active Nation. London: Cabinet Office. Huang, H., & Humphreys, B.R. (2012). Sports participation and happiness: Evidence from US microdata. Journal of Economic Psychology, 33(4), 776–793. Humphreys, B.R., McLeod, L., & Ruseski, J.E. (2014). Physical activity and health outcomes: Evidence from Canada. Health Economics, 23(1), 33–54. Humphreys, B.R., & Ruseski, J.E. (2011). An economic analysis of participation and time spent in physical activity. B.E. Journal of Economic Analysis and Policy, 11(1), Article 47. Humphreys, B.R., Ruseski, J.E., & Zhou, L. (2015). Physical activity, present bias and habit formation: Theory and evidence from longitudinal data.

414

THE SAGE HANDBOOK OF SPORTS ECONOMICS

University of Alberta Department of Economics Working Paper No. 2015–06. Mahtani, K.R., Protheroe, J., Slight, S.P., et al. (2013). Can the London 2012 Olympics ‘inspire a generation’ to do more physical or sporting activities? An overview of systematic reviews. BMJ Open, 3, e002058. Marshall, S.J., & Biddle, S.J.H. (2001). The transtheoretical model of behavior change: A meta-analysis of applications to physical activity and exercise. Annals of Behavioral Medicine, 23, 229–246. McCartney, G., Thomas, S., Thomson, H., Scott, J., Hamilton, V., Hanlon, P. et  al. (2010). The health and socioeconomic impacts of major multi-sport events: A systematic review. British Medical Journal, 340, c2369. Misener, L., Taks, M., Chalip, L., & Green, B.C. (2015). The elusive ‘trickle-down effect’ of sport events: Assumptions and missed opportunities. Managing Sport and Leisure, 20(2), 135–156. Mutter, F., & Pawlowski, T. (2014). The causal effect of professional sports on amateur sport participation: An instrumental variable approach. International Journal of Sport Finance, 9, 172–188. Potwarka, L.R., & Leatherdale, S.T. (2016). The Vancouver 2010 Olympics and leisure-time physical activity rates among youth in Canada: Any evidence of a trickle-down effect? Leisure Studies, 35(2), 241–257. Prochaska, J.O., DiClemente, C.C., & Norcross, J.C. (1992). In search of how people change: Applications to addictive behaviors. American Psychologist, 47, 1102–1114. Ramchandani, G., Davies, L.E., Coleman, R., Shibli, S., & Bingham, J. (2015). Limited or lasting legacy? The effect of non-mega sport event attendance on participation. European Sport Management Quarterly, 15(1), 93–110. Ruseski, J., & Maresova, K. (2014). Economic freedom, sport policy and individual participation in physical activity: An international comparison. Contemporary Economic Policy, 32(1), 42–55. Scitovsky, T. (1976). The Joyless Economy. New York: Oxford University Press.

Shibli, S., Kokolakakis, T., & Davies, L. (2014). Child taking part survey: Multivariate analysis of the determinants of child participation in arts, sport, heritage, museums and libraries. Sport Industry Research Centre, Sheffield Hallam University, UK. Sport England (2016). Sport England: Towards an Active Nation Strategy 2016–2021. London: Sport England. Sportscotland (2004). Curling Success and its Impact on Participation. Research Report No. 92. Edinburgh: Sportscotland. Available at: www. sportscotland.org.uk/sportscotland/Documents/ Resources/CurlingStudyResearchReportNo92.pdf Veal, A.J., Toohey, K., & Frawley, S. (2012). The sport participation legacy of the Sydney 2000 Olympic Games and other international sporting events hosted in Australia. Journal of Policy Research in Tourism, Leisure and Events, 4(2), 155–184. Weed, M. (2009). The potential of the demonstration effect to grow and sustain participation in sport. Review Paper for Sport England. Centre for Sport, Physical Education and Activity Research, Canterbury Christ Church University. Weed, M., Coren, E., Fiore, J., Mansfield, L., Wellard, I., Chatziefstathiou, D., & Dowse, S. (2009). A Systematic Review of the Evidence Base for Developing a Physical Activity and Health Legacy from the London 2012 Olympic and Paralympic Games. London: Department of Health. Weed, M., Coren, E., Fiore, J., Wellard, I., Chatziefstathiou, D., Mansfield, L., & Dowse, S. (2015). The Olympic Games and raising sport participation: A systematic review of evidence and an interrogation of policy for a demonstration effect. European Sport Management Quarterly, 15(2), 195–226. Weimar, D., Wicker, P., & Prinz, J. (2015). Membership in sport clubs: A dynamic panel analysis of external organizational factors. Nonprofit and Voluntary Sector Quarterly, 44(3), 417–436. Wicker, P., & Frick, B. (2016). The inspirational effect of sporting achievement and potential role models in football: A gender-specific analysis. Managing Sport and Leisure, 21(5), 265–282.

41 Willingness to Pay in Sports Johannes Orlowski and Pamela Wicker

INTRODUCTION Discussions on public investments in sport facilities, sport events, and elite sport policies have turned out to be a lively back and forth of the costs and benefits of the respective investment. With public investments being subject to increased scrutiny by taxpayers, proponents of investments have established several legitimation strategies, including the provision of rational arguments, value-based arguments, and authority-based arguments (Sant & Mason, 2018). For example, rational arguments encompass the spin-off of local businesses, downtown revitalization, and enhanced reputation as a result of hosting an event, while value-based arguments include enhanced entertainment and leisure opportunities, community building, quality of life, and civic pride (Sant & Mason, 2018). These examples visualize that legitimation strategies are not only based on anticipated tangible economic impacts, such as the development of local businesses, but also on intangible effects, like enhanced reputation, pride, quality of life, and civic pride. Therefore, these intangible effects should be included in estimates of the benefits and costs of sport-related investments to provide a more holistic picture.

The inclusion of intangible effects in economic analyses is associated with at least two challenges. The first challenge is the non-market nature of the underlying goods or services. For example, goods like reputation and pride do not have prices and are not sold on the market. Therefore, several valuation methods have emerged to assign a monetary value to these non-market goods and services, such as the hedonic pricing approach, the travel cost method, the compensating variation approach, and the contingent valuation method (CVM) (for an overview of valuation approaches, see Orlowski & Wicker, 2018). Though differing in their underlying assumptions and required data, all approaches yield willingness to pay (WTP) estimates, i.e., the monetary value individuals, usually taxpayers, assign to the respective facility, event, or policy. The second challenge is that value estimates have to be available ex ante to inform decisionmaking at the policy level. However, many valuation approaches, including hedonic pricing, compensating variation, and the travel cost method, are only able to provide ex post estimates as they rely on revealed preference data. Particularly, the possibility to obtain monetary values ex ante, i.e., before a facility was built, an event took place, or a policy was implemented,

416

THE SAGE HANDBOOK OF SPORTS ECONOMICS

drove the rapidly increasing interest of researchers in CVM as a valuation approach (Walker & Mondello, 2007). CVM is a survey-based method where respondents are presented with a hypothetical, yet realistic scenario that describes positive (or negative) changes in a specific good or service. Then they are asked to state their WTP for the scenario to occur (or to be avoided) (Carson, 2000). Another advantage of CVM is that it allows integrating non-use (passive-use) values into an economic analysis. While use values solely represent the value of utility directly related to the consumption of a good or service, such as attending a sport event or a league match, non-use values also take into account this type of utility, which is potentially gained through a passive, non-physical consumption of a good or service (Carson, 2000). This means that residents can also benefit from sport facilities, sport teams, or sport events when they do not attend competitions or games, for example, by talking to other people about the team or event (Johnson, Groothuis, & Whitehead, 2001), deriving happiness and pride from sporting success (e.g., Wicker, Hallmann, Breuer, & Feiler, 2012a), and enhanced prestige and reputation (Wicker, Whitehead, Mason, & Johnson, 2017). The purpose of this chapter is to provide an overview of existing WTP studies in sports economics applying CVM. This overview includes a discussion of theoretical underpinnings, applied scenarios, methodological specifications within CVM (payment vehicle and period), and key findings (mean WTP, significant determinants of WTP). Having said this, it extends existing reviews of CVM in sport that have focused on the conceptual understanding of CVM and its appropriateness to inform the sport management discourse (Johnson, 2008; Walker & Mondello, 2007) and an overview of (dis)advantages and applications in terms of country, research context, database, sample size, and estimator (Orlowski & Wicker, 2018). The focus of this review is on specific methodological choices within CVM and concrete findings. An updated review of CVM studies is warranted as the popularity of CVM applications in sports has rapidly increased since these earlier reviews (Johnson, 2008; Walker & Mondello, 2007). Entering ‘contingent valuation method’ and related terms into common search engines and the keyword search of leading journals in sport economics (Journal of Sports Economics, International Journal of Sport Finance) and sport management (Sport Management Review, Journal of Sport Management, European Sport Management Quarterly) yielded over 40 articles in the English language which were published between 2000 and August 2018. While 15 of these

studies have been published between 2000 and 2010, almost twice as many have been published since 2011, indicating the increasing popularity of CVM in sport.

THEORETICAL FOUNDATIONS Different theoretical approaches were used to frame CVM studies, to explain the level of individuals’ WTP, and to conceptualize determinants of WTP. The assessment of WTP is rooted in the discussion of the concept of utility. Frey (2008) distinguished between decision utility and experienced utility, with the latter being more relevant to consuming sports and the public goods and externalities created by sport (e.g., pride, happiness, reputation). Existing research has relied on the concept of experienced utility (e.g., Frick & Wicker, 2018a, 2018b; Wicker, Kiefer, & Dilger, 2015) and supported the notion that in particular sport events are experience goods as people have difficulty in adequately valuing these goods when they have not experienced them (Humphreys, Johnson, Mason, & Whitehead, 2018; Süssmuth, Heyne, & Maennig, 2010). Conceptually, the idea is that the concrete level of utility is reflected in individuals’ statements of WTP for a certain good or service (Becker, Degroot, & Marschak, 1964). The level of experienced utility differs among consumers as the value derived from experiencing sports is subjective. As postulated by the random utility maximization model (Hanemann, 1984), the utility function is only known by the individual and typically includes the commodity under analysis, individual income, price, and some random terms. Individual utility increases with increasing income and the commodity under analysis (e.g., hosting a sport event, sporting success, relocation of a team) and decreases with increasing prices. Several studies have used this framework as theoretical underpinning (e.g., Barros, 2006; Johnson et  al., 2001; Johnson, Mondello, & Whitehead, 2006; 2007a; Johnson, Whitehead, Mason, & Walker 2007b; Johnson, Whitehead, Mason, & Walker, 2012). Given the subjective nature of value assessments, existing studies have relied on further theories to explain why some respondents are expected to report higher or lower WTP values than others and what factors are expected to be relevant in this regard. The most frequently applied theory is consumption capital theory (e.g., Frick & Wicker, 2018b; Wicker et al., 2012a; Wicker, Prinz, & von Hanau, 2012b). This theory holds that individuals

WILLINGNESS TO PAY IN SPORTS

develop consumption capital through the repetitive consumption of similar goods, with higher levels of consumption capital yielding more (experienced) utility (Stigler & Becker, 1977) and, accordingly, also higher WTP. Less frequently applied theoretical approaches include other forms of capital (e.g., human capital; Wicker et al., 2012a), rational choice theory (Castellanos, García, & Sánchez, 2011), the theory of planned behavior (Funahashi & Mano, 2015; Johnson et al., 2007b), equity theory (Kiefer, 2015), organizational capacity (Swierzy, Wicker, & Breuer, 2018), and public choice theory (Groothuis, Johnson, & Whitehead, 2004). Notably, approximately half of CVM studies have not stated any theoretical foundation to conceptualize possible determinants of WTP.

LITERATURE REVIEW This section reviews existing studies applying CVM to estimate WTP in the field of sport. This review of literature is organized around four main aspects: (1) research context and presented scenario, (2) payment vehicle and payment period, (3) mean WTP, and (4) significant determinants of WTP where applicable (see Table 41.1 for an overview).

Research Contexts and Scenarios The second column in Table 41.1 summarizes the contexts and scenarios of existing studies. CVM has been applied to estimate WTP in several research contexts. The most frequently applied context was the hosting of various sport events, including the Olympic Games (e.g., Walton et al., 2008), World or European Championships (e.g., Barros, 2002, 2006; Süssmuth et  al., 2010), or smaller one-day events such as a stage for a cycling race (e.g., Vekeman et  al., 2015). The second most frequent application was in the context of national (e.g., Humphreys et  al., 2018; Wicker et  al., 2012a, 2012b, 2015) or regional sporting success (e.g., Barlow & Forrest, 2015; Frick & Wicker, 2018a; Wicker et al., 2016). While studies on regional sporting success of sports teams were typically conducted in Europe, North American studies focused on team relocation, i.e., the WTP to attract a professional sports team to a city or keep an existing team in the city (e.g., Groothuis et al., 2004; Owen, 2006; Santo, 2007). An important decision for team relocation is the construction of a new stadium. Therefore, many studies were conducted to assess the WTP

417

of taxpayers to build new sport facilities or stadiums (e.g., Fenn & Crooker, 2009; Harter, 2015; Johnson et  al., 2001, 2012), including the first application of CVM to sports (Johnson & Whitehead, 2000). CVM has also been applied in a variety of further research contexts. In Europe, non-profit sport clubs have repeatedly been the object of CVM research. Existing studies have estimated the WTP for membership fees (Swierzy et al., 2018; Wicker, 2011), overall club quality improvements (Kiefer, 2015), and the provision of training sessions by voluntary coaches (Orlowski & Wicker, 2016). Furthermore, previous research has estimated the WTP for sport policies at the elite (e.g., Funahashi & Mano, 2015; Morgan & Whitehead, 2018) and grassroots level (Johnson et  al., 2007b). Within sports leagues, existing studies have assessed the WTP for tickets for football games (Nalbantis, Pawloski, & Coates, 2017), VIP seating at a basketball game (Drayer & Shapiro, 2011), and competitive balance in football (Pawlowski & Budzinski, 2013). More remote topics include the WTP for increased security in backcountry skiing (Leiter & Rheinberger, 2016) or a college football tradition (Interis & Taylor, 2017). These diverse applications support the flexibility of CVM, which can be adjusted to fit many research contexts and problems in an effort to provide knowledge about the valuation of specific goods or services ex ante. Although the estimation of WTP values ex ante represents a core advantage of CVM, the notion of experienced utility suggests that investigating the impact of the consumption or experience of the valued good or service on individuals’ WTP is important. Therefore, a handful of studies applied a longitudinal research design where the same respondents were surveyed ex ante and ex post a sport event. Interestingly, no clear tendency in terms of estimated WTP values with respect to ex ante/ex post estimates could be observed. Two studies reported higher ex ante estimates (Andersson et  al., 2004; Morgan & Whitehead, 2018), while three studies documented higher ex post values (De Boer et  al., 2019; Humphreys et al., 2018; Süssmuth et al., 2010).

Payment Vehicles and Periods The third and fourth columns of Table 41.1 summarize the payment vehicles and periods, respectively. They show that various payment vehicles have been applied in previous research, with the most frequently used vehicle being some form of tax. These forms encompass changes in an already existing city tax (e.g., Groothuis et  al., 2004),

WTP for new university basketball arena and minor league baseball stadium WTP for new arena to keep Pittsburgh Penguins (NHL) in city WTP to support Portugal in hosting the UEFA Euro 2004

WTP for hosting the World Skiing Championship in Trondheim, Norway WTP to keep NFL team (Pittsburgh Penguins) in Pittsburgh WTP of Lisbon citizens to host 2004 UEFA Euro in Portugal WTP to keep NFL team in Jacksonville and WTP to attract a new NBA team WTP to keep professional sports team (MLB, NFL, NBA, NHL) in Minnesota and Michigan

Johnson & Whitehead, 2000

Andersson et al., 2004

Owen, 2006

Johnson et al., 2006

Barros, 2006

Groothuis et al., 2004

Barros, 2002

Johnson et al., 2001

Scenario

Author, year

Not specified

Tax increase

Not specified

City taxes

Individual earmarked tax

Individual earmarked tax

City tax increase

Tax increase

Payment vehicle

Annual payment (period not specified) Lump-sum payment Annual payment over 5, 10, and 20 years Annual payment

Lump-sum payment

Annual payment (period not specified) Annual payment (period not specified) Lump-sum payment

Payment period

Michigan: MLB: $15.72; NFL: $19.84; NBA: $13.79 NHL: $56.05 Minnesota: MLB: $27.29; NFL: $83.95 NBA: $25.40

NBA: $60.15 NFL: $161.04

Nonsupporters: $9.00; Supporters: $30.76 €1.75

Ex ante: $143; Ex post: $135

€0.18

Basketball arena: $6.36; Baseball stadium: $6.17 $5.57

Mean WTP

Michigan: +Interest as fan; +No. of games attended; −Region Minnesota: +Interest as fan; +No. of games attended; +Income

+Average no. of games attended per season; −Income; +Being a soccer fan; −Expected prestige from Euro 2004; +Education NBA & NFL: +Starting point in bid; +Long payment period; +No. of games attended/expected to attend; +Income

−Age; +Education; −Income; +Male; +No. of games attended; +No. games watched on TV; +Consuming newspaper football coverage; +Member of football club +Optimism towards crowdedness; +Optimism towards traffic conditions; +Optimism towards tax spending; +Confidence that city will become a more popular tourist destination +Income; −Education; +Hockey attendance; +Penguin civic pride

+Games attended in old arena; +Expected to attend more games; +Public good consumption; +Expected to attend future games; +No. of games expected to attend in the future +Games attended; +Public good consumption; +Attended or watched Stanley Cup games

Significant determinants of WTPa

Table 41.1  Overview of CVM studies estimating WTP (in chronological, then alphabetical order)

418 THE SAGE HANDBOOK OF SPORTS ECONOMICS

Castellanos et al., 2011

Süssmuth et al., 2010

Fenn & Crooker, 2009

Ex ante / Ex post WTP for hosting the 2006 FIFA World Cup in Germany WTP to keep a First Division football club in the city of A Coruña, Spain

WTP of residents in Bath, UK, for hosting the 2012 Olympic Games in London WTP to finance new stadium for a NFL team (Minnesota Vikings)

Walton et al., 2008

Voluntary contribution

Voluntary contribution

Voluntary contribution

London: Council tax increase; Manchester, Glasgow: Donation National tax increase

WTP of London, Manchester, and Glasgow residents for hosting the 2012 Olympic Games in London

Atkinson et al., 2008

Santo, 2007

WTP for enhanced sport Provincial income and recreation programs to tax increase increase sport participation by 2% or 10 % in Alberta, Canada WTP for consumption benefits Tax increase associated with hosting a MLB team in Portland, Oregon

Johnson et al., 2007b

Tax increase

WTP to keep NFL team in Jacksonville and WTP to attract a new NBA team

Johnson et al., 2007a

Annual payment (period not specified)

Lump-sum payment

Lump-sum payment

One year

Period not specified; Aggregate WTP over 30 years Not specified; 10 years

€10.77

Ex ante: €4.26; Ex post: €10.07

$312.52

£70.11

London: £21.95; Manchester: £12.40; Glasgow: £10.87

$14.35

Annual payment NBA: $81.63 (total over 5, 10, and 20 over 10 years) years NFL: $148.36 (total over 20 years) Annual (period $18.32 not specified)

(Continued)

−Bid presented; +Public goods consumption; +Money spent on tickets, merchandise, and travel cost; +New stadium brings more prestige to area; +New stadium will help Vikings to win Super Bowl; +Vikings will leave without new stadium; +Support Twins over Vikings for new stadium; +Support joint stadium with University stadium Ex ante: +Believes Germany benefits from World Cup; +Education; −Age Ex post: +Believes Germany benefits from World Cup; +Male; −Age +Income; +No. of games attended; +No. of games watched on TV; +Male; +Public goods consumption

+Male; +Homemaker; +Income; +Intention to attend live events; +Participates in active exercise; −Age; −Full-time employed

+Income; +Expects London to be better or as well organized as previous host cities; +Intention to attend live events; +Support for London 2012 bid; +Intangibles perceived more important than tangibles; −More than 10 year payment period

+No. of MLB games expected to attend; +Anticipated public consumption benefits; +Income; +Believe that economy will benefit; −Tax aversion; −Believe in other priorities

NBA: +First scenario presented; +Long payment period; No. of games attended/expected to attend; −Age; +Income NFL: +Long payment period; +No. of games attended/expected to attend; −Age; +Public goods consumption; +Jaguars make Jacksonville major city; +Team improves race relations −Respondent does not live in Calgary or Edmonton

WILLINGNESS TO PAY IN SPORTS 419

WTP for premium seat tickets to a NBA game against the LA Lakers WTP for membership fees in German non-profit sports clubs WTP for new NHL arenas in Edmonton and Calgary, Canada

WTP for first rank in medal table and for gold medal in track and field at 2012 London Games WTP of the German population for winning the 2010 Football World Cup WTP for competitive balance in German, Dutch and Danish first division football leagues WTP to keep a First Division football club in the city of A Coruña, Spain WTP to guarantee professional league status for Bury FC/ Luton Town, UK WTP for Japanese elite sport policy, i.e., achieving rank 5 in medal tally of Olympic Summer and rank 10 in Winter Games

Drayer & Shapiro, 2011

Wicker et al., 2012a

Funahashi & Mano, 2015

Barlow & Forrest, 2015

Castellanos et al., 2014

Pawlowski & Budzinski, 2013

Wicker et al., 2012b

Johnson et al., 2012

Wicker, 2011

Scenario

Author, year

Annual payment for 5 years

Annual payment

Lump-sum payment

Payment period

Tax increase

Council tax increase

Voluntary contribution

Ticket price surcharge

Voluntary contribution

Annual payment (period not specified) Monthly payment (period not specified) Annual payment over 10 years

Period not specified

Lump-sum payment

Elite sport support Lump-sum program payment

Property tax increase

Membership fee increase

Ticket price

Payment vehicle

+Income; +No. of games attended; +No. of games watched on TV; +Public good consumption; +Quality of life; −Age

+Football participation; +Games watched; +Identification with country; +Identification with national team; +Personal importance; +Importance for country; +Female; +Income n.a.

Edmonton: +Existence of casino; +Income; +Attending home games; +Lively downtown improves quality of life; +Lively downtown is necessary for a city to be great Calgary: −Taxes; +Distance from downtown; +Lives in downtown; +Lively downtown improves quality of life +Pride; +Male; +Age 54-65; +Income

+Current membership fee; +Income; +Education; +Years of participation; +Level of performance

+Team interest; +Importance of future team success

Significant determinants of WTPa

JPY 1802 (€14.59)

+Perceived social benefits; +Perceived personal benefits; +Regarding elite athletes as role models; +Age; +Income; −Perception of risks

Bury: £0.22-£1.79 n.a. Luton: £0.20-£1.68

DE: €3.09 NL: €2.98 DK: €3.10 2003: €14.48; 2012: €10.18

€25.79

Medal table: €6.13; Gold medal T&F: €5.21

Edmonton: $17.58($12.73) Calgary: $18.01($13.35)

€148.07

$55.39

Mean WTP

Table 41.1  Overview of CVM studies estimating WTP (in chronological, then alphabetical order) (Continued)

420 THE SAGE HANDBOOK OF SPORTS ECONOMICS

Five years

Fan bonds

Orlowski & Wicker, WTP for coaching sessions Voluntary 2016 provided by volunteer coaches contribution in German non-profit sports clubs

Wicker et al., 2016 WTP of football fans to avoid negative (relegation) or achieve positive outcome (promotion, success)

Annual payment (period not specified) Lump-sum payment

Annual payment (period not specified) Ongoing

Voluntary contribution

Income tax increase

Annual payment (period not specified) Increase in annual Ongoing membership fee

Special tax

Leiter & WTP for enhancements in Subscription Rheinberger, 2016 avalanche forecasting services

WTP to keep a high school gymnasium open in Anderson, Indiana Kiefer, 2015 WTP of riding club members for overall quality improvement Vekeman et al., WTP of Flemish population for 2015 hosting a cycling event (Tour of Flanders) Wicker et al., 2015 WTP for winning the 2012 UEFA European Championships and first place in medal tally of 2012 London Games

Harter, 2015

Open format: €67.26 Dichotomous choice: €17.51 Payment ladder: €18.46 Negative €735.76 Positive: €765.23

CHF 78.00

UEFA European Championships: € 40.74; London Olympics: €46.47

(Continued)

Negative (attendees): +Play; +Identify; −happy; +Sad; −Without bonds; +With bonds; +Likelihood; −Not redeem; +Male, +Age; +Schooling; +Income; −Third scenario Positive (attendees): −Interest; +Play; +Discuss; +Without bonds; −With bonds; +Likelihood; −Not redeem; +Age; −Tenure city; +Schooling; +Income

+Male; +Age; +Active cyclist; +Intensive user; +Probability to be a spectator −Living in arrival, departure or village of the tour UEFA: −Expectations towards German success; −Identification with German national football team; −Personal importance of football team performance; +Age; −Age2; −Income London Olympics: −General interest in sport; +Interest in Summer Olympics; +Watching the Olympic Games; +Expectations of Germany in final medal rank; +Identification with German Olympic team; − Male; −Education; +No. of days before Olympic Games +Perceived risk; −Deliberate risk taking; +Forecast usage; +Defensive rider; +Expenditure on equipment; +Income; +Swiss; −Skills Open format: +Training group size; +Income. Dichotomous choice: +Social interaction in club; +Being a volunteer; +Training frequency; −Awareness of coach’s qualifications; +Attitude towards coach; +Training group size; +Income. Payment ladder: +Training group size; +University degree

+Membership fee; +Income; +Social ties; −Satisfaction with club; −Membership fee expensiveness

€117.39

€10.3

+Interest in sports; −White; +Household member is a teacher; +Attended Anderson high school

$9.33

WILLINGNESS TO PAY IN SPORTS 421

Frick & Wicker, 2018a Frick & Wicker, 2018b

De Boer et al., 2019

WTP for hosting the big start of the 2016 Giro d’Italia in The Netherlands WTP for ‘first division status’ of Football Bundesliga teams WTP of Austrian population for broadcasting of 2017 World Championships in Alpine Skiing; WTP for athlete support

WTP for a ticket to the first division football match of VfB Stuttgart against Borussia Mönchengladbach Wicker et al., 2017 WTP for hosting Olympic Summer Games in Germany

WTP for banning a college sport tradition (ringing cowbells) at Mississippi State University football games

Interis & Taylor, 2017

Nalbantis et al., 2017

Scenario

Author, year

Fee; tax

Lump-sum payment Lump-sum payment; Monthly payment over 5 years

Lump-sum payment

Monthly payment over 5 years

Income tax increase

Donation to the organizer of the event Voluntary fund

Lump-sum payment

Ongoing

Payment period

Ticket

Ticket price increase

Payment vehicle

Broadcasting: €24.80; Athlete support: €8.12

€110.49

Ex ante: €3.58; Ex post: €4.45

€ 51.38

Student/student season ticket: $2; Student/singlegame ticket: $9; Non-student/ general public season ticket: $8 €25.10

Mean WTP

Broadcasting: +Interest; +No. of competitions planned to watch; − Medal expectations; +Public goods; +Higher education. Athlete support: +Interest; +High income; +Male; −Age

+Home games; +Income; +FC Schalke 04

−Scenario order; −Regularly practicing sport; −Identification with Germany; +Happy when German athletes are successful; −Success considered important for Germany’s reputation; +Athletes are considered role models; +Results of survey are expected to affect sport; +All information in survey understood; −Confidence in German sport policy; +Income; +Male; −Age; +Tenure in city; +Education +Visit; +Media; −Age; +Age²; +Income

+Suspense; −Suspense²; +Bundesliga suspense; +Income; +Travel distance; −Travel distance²

+Male; −Prior games attended; +Survey provided enough info; +Cowbells are annoying; +Cowbells are/could be annoying; +Cowbells project negative image

Significant determinants of WTPa

Table 41.1  Overview of CVM studies estimating WTP (in chronological, then alphabetical order) (Continued)

422 THE SAGE HANDBOOK OF SPORTS ECONOMICS

WTP for memberships in Increase German nonprofit sports clubs in monthly membership fee

WTP for hosting the 2024 Olympic Summer Games in Hamburg, Germany

Swierzy et al., 2018

Wicker & Coates, 2018

Annual payment over 5 years

Ongoing

Annual payment over 4 years

Annual payment over 3 years

€25

€10.50

Ex ante: $35.02 Ex post: $28.95

Ex ante: $38.19 Ex post: $89.70

Note: a plus (+) denotes positive association with WTP, minus (−) indicates negative association; n.a. = not analyzed.

City government, Federal government; German Olympic Sports Confederation; Sport foundation; Account in trust

American households’ WTP for Household tax soccer player development in the United States

Morgan & Whitehead, 2018

Tax increase

WTP for medals won by Team Canada in the 2010 Winter Olympic Games in Vancouver, Canada

Humphreys et al., 2018

Ex ante: +Expected gold medals; + Expected other medals; +Income; +Perception of prestige; +Proud to host games; +Proud if Canada wins medal table; +Province dummies Ex post: +Expected gold medals; +Expected other medals; +Income; +Perception of prestige; +Proud if Canada wins medal table; +Province dummies Ex ante: −Taxes; +Identification with team; +Interest in soccer; +Age. Ex post: −Taxes; +Identification with team; +Interest in soccer; − Income; −Age −Number of paid staff; +Club breaks at least even; +Bad financial situation; +Club engages in youth work; −Club should stay as it is; −Age; +Education; +Current fee; +Identification with club; +Satisfaction with club; +Utilization of club facilities +Informed; +Role models; +Age; +Income

WILLINGNESS TO PAY IN SPORTS 423

424

THE SAGE HANDBOOK OF SPORTS ECONOMICS

income tax (e.g., Wicker et al., 2017b), or a specifically implemented earmarked tax (Andersson et  al., 2004). Other prominent payment vehicles include voluntary contributions (e.g., Castellanos et al., 2011; Fenn & Crooker, 2009) and markups to existing prices (e.g., Drayer & Shapiro, 2011; Pawlowski & Budzinski, 2013) or membership fees (e.g., Kiefer, 2015; Wicker, 2011). More uniquely, fan bonds (Wicker et al., 2016) or subscriptions (Leiter & Rheinberger, 2016) have served as payment vehicles and one study has invited respondents to select their payment vehicle among five suggested options (Wicker & Coates, 2018). The variety of payment vehicles supports the flexibility of adapting CVM surveys to the specific demands of each research context. However, the choice of payment vehicles other than taxes can be associated with compromises in terms of incentive compatibility, meaning that some payment vehicles do not induce respondents to reveal their true preferences (Wicker et al., 2017; Wiser, 2007). Different types of payment schemes were identified. Most researchers have presented scenarios with (annually or monthly) recurring payments (e.g., Johnson et  al., 2001) or lump-sum payments (e.g., Barros, 2002) to survey participants. With respect to recurring payments, a clear definition of the overall payment period is required in order to obtain meaningful WTP values (Johnson et  al., 2006). In previous research, these periods varied from one year up to more than 20 years. However, a substantial number of studies do not provide detailed information on the duration of the funding scheme within the scenarios, suggesting potential biases from temporal embedding effects (Johnson et al., 2006).

WTP Estimates and their Determinants The last two columns in Table 41.1 display mean WTP and significant determinants of WTP. Since obtained WTP estimates are contingent on the provided scenario and research context, utilized payment vehicle, and payment period, absolute measures are difficult to compare. Also, different years, countries, and currencies make comparisons between studies difficult. Having said this, it is difficult to observe any patterns in WTP estimates. Hence, the observation by Johnson (2008) that WTP seems to be higher for active sport consumption than for spectator sports cannot be supported a decade later. The comparison of determinants of WTP is also difficult as existing studies have included different

factors. Nevertheless, some overarching observations can be made. The majority of research has identified some measure of consumption capital as a significant and positive determinant of individuals WTP. Applied measures of consumption capital include, for example, the number of games attended (Groothuis et  al., 2004; Owen, 2006) or watched on television (Barros, 2002). Furthermore, in the context of non-profit sports clubs, research has found a significant positive relationship between social capital measures and WTP (Kiefer, 2015; Orlowski & Wicker, 2016; Swierzy et al., 2018). Several studies examined the proximity to a respective facility or event and found significant effects for included regional or provincial controls (Humphreys et  al., 2018; Johnson et  al., 2007b; Owen, 2006), indicating that WTP might depend on geographic factors. More specifically, Vekeman et  al. (2015) documented that the proximity to a sport event had a significant negative effect on individuals’ WTP, while Nalbantis et  al. (2017) suggested that the relationship between event distance and WTP might be an inverse u-shape. Attitudes towards the non-market good in question were very frequently included in models analyzing WTP. Estimated coefficients across various research contexts are predominantly positive and significant, indicating that more positive attitudes towards the respective context in general or the non-market good in particular, the higher the WTP. For example, a higher general interest in sport is positively and significantly associated with WTP for hosting sport events (e.g., Walton et al., 2008), new sport facilities (e.g., Harter, 2015), or national sporting success (e.g., Wicker et al., 2015). Socio-demographic characteristics of respondents were included in many studies, with some general tendencies being noteworthy. First, individuals’ age is included in nearly all CVM studies. However, the effect is statistically significant in merely one-third of studies. Interestingly, the age effect seems to be predominantly negative with regard to the WTP for hosting major sport events and prevent professional sports team allocation (e.g., Barros, 2002; Süssmuth et al., 2010; Walton et  al., 2008). On the contrary, the coefficient is typically positive in studies estimating the WTP for sport policies and national sporting success (e.g., Funahashi & Mano, 2015; Morgan & Whitehead, 2018; Wicker et al., 2012a). Second, gender was routinely included as a control variable in WTP equations but was insignificant in many studies. If significant, men tended to have a higher WTP than women (e.g., Interis & Taylor, 2017; Vekeman et  al., 2015), with the exception of two German studies on national sporting success (Wicker et  al., 2012b;

WILLINGNESS TO PAY IN SPORTS

Wicker et  al., 2015). Third, education was infrequently found to be a significant determinant of WTP. While some studies documented a negative association between educational level and WTP (Groothuis et al., 2004; Wicker et al., 2015), others found a positive relationship (Barros, 2002, 2006; Süssmuth et  al., 2010; Wicker, 2011). Lastly, income seems to have a consistent, positive effect on individuals’ WTP across various studies and research contexts, which is in line with the random utility maximization model.

CONCLUSIONS This chapter reviewed existing studies applying CVM in sports. The review indicates that the body of research has developed over time and that CVM has become an increasingly popular method to assign a monetary value to non-market goods and services ex ante. Looking at the existing body of CVM studies, at least two concluding remarks can be made that also outline perspectives for future research. First, the theoretical underpinning of studies is dominated by discussions of utility and different forms of capital (e.g., consumption capital, human capital, social capital). Moreover, several studies have neglected a theoretical discussion as to how and why which factors are expected to influence individuals’ WTP. Future studies should pay more attention to the theoretical framework that guides the research. Given the national or regional character of many CVM studies, a worthwhile endeavor might be to consider not only individual factors, but also higher-level factors, such as characteristics of the district or community, in explaining the level of WTP. Second, scholars should reflect on their research designs in several ways. For example, the review has shown that many studies have not specified the payment period which might affect WTP estimates and the credibility of CVM studies. Also, more efforts should be made to establish longitudinal research designs that allow comparing ex ante and ex post WTP values, having in mind that such designs compromise the anonymity of surveys. Nevertheless, they facilitate establishing a better link between applied concepts (experienced utility) and the empirical analysis. Moreover, the selected scenarios and determinants indicate that the valuation of positive externalities has attracted the most research interest. Only a few studies included negative externalities, such as risk, crowdedness, or congestion. Future studies should consider the aforementioned and

425

further negative externalities of sport to a greater extent, such as crime and negative environmental outcomes, and provide a more balanced picture of potential (intangible) benefits and costs.

REFERENCES Andersson, T. D., Rustad, A., & Solberg, H. A. (2004). Local residents’ monetary evaluation of sports events. Managing Leisure, 9(3), 145–158. Atkinson, G., Mourato, S., Szymanski, S., & Ozdemiroglu, E. (2008). Are we willing to pay enough to ‘back the bid’? Valuing the intangible impacts of London’s bid to host the 2012 Summer Olympic Games. Urban Studies, 45(2), 419–444. Barlow, A., & Forrest, D. (2015). Benefits to their communities from small town professional football clubs. National Institute Economic Review, 232(1), 18–29. Barros, C. (2002). Evaluating the regulatory procedure of host-country selections for the UEFA European Championship: A case study of Euro 2004. European Sport Management Quarterly, 2(4), 321–349. Barros, C. (2006). Evaluating sport events at European level: The Euro 2004. Sport Management and Marketing, 1(4), 400–410. Becker, G. M., Degroot, M., & Marschak, J. (1964). Measuring utility by a single-response sequential method. Behavioral Science, 9, 226–232. Carson, R. T. (2000). Contingent valuation: A user’s guide. Environmental Science & Technology, 34(8), 1413–1418. Castellanos, P., García, J., & Sánchez, J. (2011). The willingness to pay to keep a football club in a city: How important are the methodological issues? Journal of Sports Economics, 12(4), 464–486. Castellanos, P., García Villar, J., & Sanchez, J. M. (2014). Economic crisis, sport success and willingness to pay: The case of a football club. Sports, Business and Management: An International Journal, 4(3), 237–249. De Boer, W. I. J., Koning, R. H., & Mierau, J. O. (2019). Ex ante and ex post willingness to pay for hosting a large international sport event. Journal of Sports Economics, 20(2), 159–176 (in press). DOI: 10.1177/1527002518771445 Drayer, J., & Shapiro, S. L. (2011). An examination into the factors that influence consumers’ perceptions of value. Sport Management Review, 14(4), 389–398. Fenn, A. J., & Crooker, J. R. (2009). Estimating local welfare generated by an NFL team under credible threat of relocation. Southern Economic Journal, 76(1), 198–223.

426

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Frey, B. S. (2008). Happiness: A revolution in economics. Cambridge, MA: MIT Press. Frick, B., & Wicker, P. (2018a). The monetary value of having a first division Bundesliga team to local residents. Schmalenbach Business Review, 70, 63–103. Frick, B., & Wicker, P. (2018b). The value of alpine skiing to the Austrian population: A CVM study of the 2017 World Championships. Managing Sport and Leisure (in press). DOI: 10.1080/23750472. 2018.1510742 Funahashi, H., & Mano, Y. (2015). Socio-psychological factors associated with the public’s willingness to pay for elite sport policy. Managing Sport and Leisure, 20(2), 77–99. Groothuis, P. A., Johnson, B. K., & Whitehead, J. C. (2004). Public funding of professional sports stadiums: Public choice or civic pride? Eastern Economic Journal, 30(4), 515–526. Hanemann, W. M. (1984). Welfare evaluations in contingent valuation experiments with discrete responses. American Journal of Agricultural Economics, 66, 331–355. Harter, J. F. R. (2015). Is the Wigwam worth it? A contingent valuation method estimate for a smallcity sports arena. Contemporary Economic Policy, 33(2), 279–284. Humphreys, B. R., Johnson, B. K., Mason, D. S., & Whitehead, J. C. (2018). Estimating the value of medal success in the Olympic Games. Journal of Sports Economics, 19(3), 398–416. Interis, M. G., & Taylor, N. J. (2017). Estimating the non-market value of a college sports tradition. International Journal of Sport Finance, 12(3). Johnson, B. K. (2008). The valuation of nonmarket benefits in sport. In B. R. Humphreys & D. R. Howard (Eds.), The business of sports (pp. 207–233). Westport, CT: Praeger. Johnson, B. K., Groothuis, P. A., & Whitehead, J. C. (2001). The value of public goods generated by a major league sports team: The CVM approach. Journal of Sports Economics, 2(1), 6–21. Johnson, B. K., Mondello, M. J., & Whitehead, J. C. (2006). Contingent valuation of sports: Temporal embedding and ordering effects. Journal of Sports Economics, 7(3), 267–288. Johnson, B. K., Mondello, M. J., & Whitehead, J. C. (2007a). The value of public goods generated by a National Football League team. Journal of Sport Management, 21(1), 123–136. Johnson, B. K., & Whitehead, J. C. (2000). Value of public goods from sports stadiums: the CVM approach. Contemporary Economic Policy, 18(1), 48–58. Johnson, B. K., Whitehead, J. C., Mason, D. S., & Walker, G. J. (2007b). Willingness to pay for amateur sport and recreation programs. Contemporary Economic Policy, 25(4), 553–564.

Johnson, B. K., Whitehead, J. C., Mason, D. S., & Walker, G. J. (2012). Willingness to pay for downtown public goods generated by large, sports-anchored development projects: The CVM approach. City, Culture and Society, 3(3), 201–208. Kiefer, S. (2015). Are riding club members willing to pay or work for overall quality improvement? Managing Sport and Leisure, 20(2), 100–116. Leiter, A. M., & Rheinberger, C. M. (2016). Risky sports and the value of safety information. Journal of Economic Behavior and Organization, 131, 328–345. Morgan, O. A., & Whitehead, J. C. (2018). Willingness to pay for soccer player development in the United States. Journal of Sports Economics, 19(2), 279–296. Nalbantis, G., Pawlowski, T., & Coates, D. (2017). The fans’ perception of competitive balance and its impact on willingness-to-pay for a single game. Journal of Sports Economics, 18(5), 479–505. Orlowski, J., & Wicker, P. (2016). The monetary value of voluntary coaching: An output-based approach. International Journal of Sport Finance, 11(4), 310–326. Orlowski, J., & Wicker, P. (2018). Monetary valuation of non-market goods and services: A review of conceptual approaches and empirical applications in sports. European Sport Management Quarterly (in press). Owen, J. G. (2006). The intangible benefits of sports teams. Public Finance and Management, 6(3), 321–345. Pawlowski, T., & Budzinski, O. (2013). The monetary value of competitive balance for sport consumers: A stated preference approach to European professional football. International Journal of Sport Finance, 8(2), 112–123. Sant, S., & Mason, D. S. (2018). Rhetorical legitimation strategies and sport and entertainment facilities in smaller Canadian cities. European Sport Management Quarterly, 2(1), 6–21 (in press). DOI: 10.1080/16184742.2018.1499789 Santo, C. A. (2007). Beyond the economic catalyst debate: Can public consumption benefits justify a municipal stadium investment? Journal of Urban Affairs, 29(5), 455–479. Stigler, G., & Becker, G. S. (1977). De gustibus non est disputandum. The American Economic Review, 67(2), 76–90. Süssmuth, B., Heyne, M., & Maennig, W. (2010). Induced civic pride and integration. Oxford Bulletin of Economics and Statistics, 72(2), 202–220. Swierzy, P., Wicker, P., & Breuer, C. (2018). Willingnessto-pay for memberships in nonprofit sports clubs: The role of organizational capacity. International Journal of Sport Finance, 13(3), 261–278.

WILLINGNESS TO PAY IN SPORTS

Vekeman, A., Meulders, M., Praet, A., Colpaert, J., & van Puyenbroeck, T. (2015). Contingent valuation of a classic cycling race. Journal of Sports Economics, 16(3), 268–294. Walker, M., & Mondello, M. J. (2007). Moving beyond economic impact: A closer look at the contingent valuation method. International Journal of Sport Finance, 2(3), 149–160. Walton, H., Longo, A., & Dawson, P. (2008). A contingent valuation of the 2012 London Olympic Games: A regional perspective. Journal of Sports Economics, 9(3), 304–317. Wicker, P. (2011). Willingness-to-pay in non-profit sports clubs. International Journal of Sport Finance, 6(2), 155–169. Wicker, P., & Coates, D. (2018). Flame goes out: Determinants of individual support at the 2024 Hamburg Games referendum. Contemporary Economic Policy, 36(2), 302–317. Wicker, P., Hallmann, K., Breuer, C., & Feiler, S. (2012a). The value of Olympic success and the intangible effects of sport events – a contingent

427

valuation approach in Germany. European Sport Management Quarterly, 12(4), 337–355. Wicker, P., Kiefer, S., & Dilger, A. (2015). The value of sporting success to Germans: Comparing the 2012 UEFA Championships with the 2012 Olympics. Journal of Business Economics, 85(8), 897–919. Wicker, P., Prinz, J., & von Hanau, T. (2012b). Estimating the value of national sporting success. Sport Management Review, 15(2), 200–210. Wicker, P., Whitehead, J. C., Johnson, B. K., & Mason, D. S. (2016). Willingness-to-pay for sporting success of football Bundesliga teams. Contemporary Economic Policy, 34(3), 446–462. Wicker, P., Whitehead, J. C., Mason, D. S., & Johnson, B. K. (2017). Public support for hosting the Olympic Summer Games in Germany: The CVM approach. Urban Studies, 54(15), 3597–3614. Wiser, R. H. (2007). Using contingent valuation to explore willingness to pay for renewable energy: A comparison of collective and voluntary payment vehicles. Ecological Economics, 62(3–4), 419–432.

42 Positive and Negative Externalities of Sport Events: From Well-Being, Pride, and Social Capital to Traffic and Crime Pamela Wicker and Paul Downward

INTRODUCTION Hosting major sport events has become an integral part of national sport policies. One major aim is to develop sport infrastructure and engagement with sport. For example, sport events are often used as a catalyst to improve sporting and transport infrastructure. The sport plan of the Gold Coast, Australia, host of the 2018 Commonwealth Games, states that one of its objectives is to ‘create a sustainable sporting community with quality fit for purpose facilities that provide optimal sporting programs and activities to promote health and well being’ (City of Gold Coast, 2015, p. 15). Other objectives are to establish international connections through sport and to ‘enhance the reputation of the Gold Coast as a leading sports event city’ (City of Gold Coast, 2015, p. 23). Such objectives are also identified in the legacy plans for the London 2012 Olympic and Paralympic Games which sought to increase grass-roots participation is sport in order to encourage the whole population to be more physically active; to promote community engagement across all groups in society; to exploit opportunities for economic growth; and to regenerate East London (Department of Culture, Media and Sport [DCMS], 2010).

It is clear that as well as seeking to enhance sporting engagement and infrastructure, and to elicit economic impacts, policy now targets outcomes like health and well-being, which is indicative of a desire to achieve less tangible economic benefits. This desire is now a widespread general policy aspiration. For example, the Sporting Future Strategy of the United Kingdom emphasises the relevance of experiencing sport events arguing that ‘people who regularly turn up and experience live sport, in particular when they support a specific team or athlete, can enjoy improved wellbeing or greater community engagement’ (HM Government, 2015, p. 39). Likewise, such benefits are ascribed to sporting success: ‘it provides significant wellbeing, social and economic benefits to the nation. Put simply, the more our teams win, the better the nation feels’ (HM Government, 2015, p. 43). Within sports economics, research has long focused on critically assessing the purported tangible economic impacts of sport events, such as tourism and employment effects. However, most studies have documented only small and shortterm impacts at best (Kasimati, 2003; Noll & Zimbalist, 2011). Given the difficulty of legitimising the spending of taxpayer money based on expected tangible benefits, research efforts

POSITIVE AND NEGATIVE EXTERNALITIES OF SPORT EVENTS

have shifted towards assessing the claimed benefits of a more intangible nature, such as health (e.g., Downward, Dawson, & Mills, 2016), wellbeing (e.g., Pawlowski, Downward, & Rasciute, 2014), national and civic pride (e.g., Humphreys, Johnson, Mason, & Whitehead, 2018), social capital (Schulenkorf, Thomson, & Schlenker, 2011), and image (Grix, 2012; Rowe & McGuirk, 1999). Valuations of such benefits have been found to be substantial, with estimates of their monetary value even exceeding estimated event costs and tangible event impacts in some studies (e.g., Atkinson, Mourato, Szymanski, & Ozdemiroglu, 2008; Wicker, Whitehead, Mason, & Johnson, 2017). The purpose of this chapter, therefore, is to reflect upon existing research studying these intangible impacts of sport events through the lens of sports economics. Specifically, the aim is to discuss various aspects of previous studies, including research contexts and main findings. The next section explores the theoretical foundation of the intangible impacts. Then, the empirical research that has investigated both the claimed positive outcomes of policy is reviewed as well as negative impacts that are less discussed in policy. Consequently, the impacts on both health and well-being, pride, and social capital are explored as well as traffic, congestion, and pollution, and different types of criminal behaviour. The reflection finishes by identifying lines of enquiry for future research.

THEORY OF VALUE It is well known in economics that the evaluation of policy alternatives can be couched in terms of their relative impacts on social welfare. This evaluation can take place based upon the set of Pareto efficient consumption and production possibilities coupled with assumptions about the ethic that is desired in society, or whose implications are to be evaluated, as expressed in a social welfare function (Bator, 1957). The Pareto efficient alternatives over which society chooses reflect allocations of resources in which the ratio of marginal productivities of factors of production are equal to the ratio of their factor costs, and the ratio of marginal utilities of consumption goods reflect the ratio of their market prices. Significantly, it can be argued that if individual consumers are able to exercise free choices to allocate their income and time resources to activities that maximise their utility, free competitive markets will ensure that Pareto efficient resource allocation occurs which can maximise social welfare (Downward, Dawson, & Dejonghe, 2009).

429

This implies that market prices reveal the underlying subjective value of resources to both individuals and to society and why, for example, the level and distribution of gross domestic product (GDP) has historically been emphasised in policy discussion (Stiglitz, Sen, & Fitoussi, 2010). A precondition for Pareto efficiency is the ability of markets to allocate resources which, in turn, requires that property rights are fully established. These are: …the rules (whether formal and legal or informal custom) which specify which individuals are allowed to do what with resources and the outputs of those resources. Property rights define which of the technologically feasible economic decisions individuals are permitted to make. (Gravelle & Rees, 2004, p. 9, italics in original)

If they are not fully established, then markets can fail. As Downward et  al. (2009) note, there are two main sources of market failure: inefficiency and inequity. The former arises from externalities in which property rights are not fully allocated, meaning that the market prices established in exchanges of resources do not represent the true resource costs and benefits to individuals or society. The latter reflects an explicit value judgement of social welfare that the distribution of resources provided by markets is inequitable. Such arguments are typically used to justify public sector investment in events and their infrastructure to regenerate neighbourhoods, as is implied in the introduction. In this chapter the focus is upon the values associated with sport events that are not fully captured in market prices and which manifest themselves as externalities. Figure 42.1 provides a clear outline of the potential form that the externalities might take. It shows that the total economic value of the event can be understood from both use and non-use values. The former is where market exchange reveals the value of the event. However, while one component of this value is the actual monetary exchanges that take place, for example in the form of entrance fees, it remains that additional value is not accounted for through consumer surplus, i.e., the utility obtained that exceeds that revealed in the price. Two other sources of value are also possible. One is the potential use values that could emanate from the option to experience an event, should it recur, and the other reflects the endowment of the event and infrastructure to future generations and the impacts that this can have on a community. In a similar way, non-use values reflect the impacts on society and the economy simply from the intrinsic nature of the event and this is even

430

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Total economic value

Use value

Actual use value

Entrance fees, household and business expenses

Non-use value

Potential use value

Optional value

Legacy value

Future attendance

Future profits

(if possible)

Consumer surplus

Community memory

Social cohesion

‘willingness-to-pay’

Intrinsic value of the event Moral values Satisfaction from others attending Youth education Reduction in criminality

Pride/identity

Reduction in social costs Socio-economic values Improvement in entrepreneurial system Economic growth

Figure 42.1  The total economic value of sport events (modified from Barget & Gouguet, 2007, p. 169) though beneficiaries might not engage in the event at all. Enjoyment from others’ pleasure in the event, the building of community relationships and reductions in social cost, as well as stimulating economic initiatives could be examples of this value. In this way it must be recognised that the utility or value associated with sport events does not necessarily conform to utility as represented in the axiomatic approach to consumer decision making in textbooks accounts of economics, but also the hedonic experience of engaging with an

event that could also include altruistic feelings towards others’ pleasure and enjoyment through social interactions in consumption. Consequently, Frey (2008) makes a distinction between decision utility and experienced utility. The latter nests externality effects within it. Experienced utility may be, for example, captured in relational goods (Rasciute, Downward, & Greene, 2017) and is connected with the formation of social capital. More generally, it is recognising these theoretical issues that has led,

POSITIVE AND NEGATIVE EXTERNALITIES OF SPORT EVENTS

following Stiglitz et  al. (2010), to an attempt to measure social welfare directly by asking about an individual’s utility expressed as subjective well-being (SWB) (Frey, 2008). As a result, in the United Kingdom (UK), SWB is now considered to be an important concept to measure in evaluating alternative policy outcomes and investment (HM Treasury, 2011). These developments bring with them recognition that SWB is multifaceted and can comprise physical and mental health, and in part affect, and also to be influenced by the personal, social and economic development of individuals that might stem from policies, such as those connected with the promotion of sport events.

REVIEW OF EVIDENCE This section reviews existing research exploring the externalities of sport events. The above discussion indicates that the policy narrative tends to focus on positive outcomes. The discussion below also includes some consideration of negative externalities, which tends to draw upon the sports league context, though it is clear that parallels for future research might be made. The assignment of externalities to positive or negative is made based on their claimed effect in policy document; hence health is listed among positive externalities. The discussion includes both hosting and attending sport events as well as sporting success. Indicative references are provided throughout.

POSITIVE EXTERNALITIES Health and Subjective Well-being Following the World Health Organisation (WHO, 2010), a person’s individual health status consists of two components, physical health and mental health. Physical health is typically assessed with the absence of various diseases, such as obesity, cancer, hypertension, osteoporosis, cardiovascular heart diseases, or type 2 diabetes (e.g., Humphreys, McLeod, & Ruseski, 2014; WHO, 2010) or with an individual’s subjective perception of his/her health status (e.g., Downward et al., 2016). Mental (or psychological) health is regarded as the absence of mental disorders, like anxieties and depression, or cognitive decline (WHO, 2010), and is captured in epidemiological studies using respective measures (e.g., Fergusson et al., 2015). Sports economists have mainly studied SWB – ‘a

431

person’s cognitive and affective evaluations of his or her life’ (Diener, Oishi, & Lucas, 2002, p. 53) – using self-reported measures for individual life satisfaction (e.g., Kavetsos & Szymanski, 2010) or happiness (e.g., Pawlowski et al., 2014). These measures can also be regarded as early indicators of potential mental health issues (KoivumaaHonkanen, Kaprio, Honkanen, Viinamaki, & Koskenvuo, 2004). Considering the perspective of sports economics and the terminology within the field, this chapter refers to physical health as health and uses the terms SWB, happiness, and life satisfaction interchangeably, reflecting the psychological dimension of health. Regarding health and SWB, the discussion focuses on research studying direct links. Studies examining indirect externalities, such as inspirational effects of sport events on sport participation levels, which can also yield health and well-being outcomes, have been discussed elsewhere (for an overview, see Breuer & Wicker, 2015; and Chapter 40 by Dawson in this volume). Few studies have examined direct health effects of sport events. The promotion of healthy living was considered to be an intangible benefit by residents of London preceding the 2012 Olympic Games (Atkinson et al., 2008). However, research conducted after the Games has found no empirical evidence of the event on the subjective health of the population (Downward et al., 2016). Notably, some studies have pointed towards negative health outcomes of sport events, such as asthma and illicit drug use (for an overview, see McCartney et al., 2010) as well as unhealthy habits and practices, including excessive alcohol consumption and unhealthy food (Inoue, Berg, & Chelladurai, 2015). Overall, existing reviews have documented a lack of supporting evidence for public health benefits of major sport events as claimed by policy makers (Inoue et al., 2015; McCartney et  al., 2010; Murphy & Bauman, 2007). Hence, the empirical evidence of health benefits can be considered ­limited and unconvincing. The majority of existing studies have examined the relationship between sport events and the SWB of the population. Starting with hosting sport events, it has been argued that major sport events, such as the 2012 Olympic Games, can create a feel-good factor for local residents (Atkinson et  al., 2008). Importantly, such expectations of intangible benefits were argued not to be limited to the Olympic host city but were also evident for residents outside London (Walton, Longo, & Dawson, 2008). Indeed, after the Games, more respondents agreed to the statement that the London Olympics had a positive effect on the overall mood of the general public compared to the pre-Games period (Hiller & Wanner, 2015),

432

THE SAGE HANDBOOK OF SPORTS ECONOMICS

supporting the notion that sport events are experience goods and individuals can benefit from experienced utility. The latter study also pointed to differences between population groups with regard to the SWB derived from sport events: women, older people, those with children, individuals working full-time or part-time, being retired, and with medium incomes were significantly more likely to be happy from hosting the London Olympics (Hiller & Wanner, 2015). On the contrary, a study comparing the effects of several Olympic Games and major events in football reported that the feel-good factor was only significant for hosting football events, not for the Olympic Games (Kavetsos & Szymanski, 2010). Some studies have examined specific FIFA World Cups. For example, Cornelissen and Maennig (2010) concluded that the feel-good effect was the largest and most obvious outcome of the 2006 FIFA World Cup in Germany. Likewise, the 2014 tournament in Brazil was found to yield positive SWB. Specifically, the effect of hosting the World Cup on the SWB of residents in Rio de Janeiro was larger during the event than before the event (Schlegel, Pfitzner, & Königstorfer, 2017). This also supports the view that experiencing an event can be important for generating wellbeing benefits. Experiencing can imply attending (Pawlowski et al., 2014), being aware of an event (Taks, Littlejohn, Snelgrove, & Wood, 2016), or enjoying the celebrative atmosphere in the city (Schlegel et al., 2017) – all aspects were found to be positively associated with resident SWB. Turning to sporting success, the empirical evidence is inconclusive. On the one hand, existing research has found no significant effect on wellbeing in terms of life satisfaction (Kavetsos & Szymanski, 2010). Likewise, pride from sporting success has been shown not to be related to individual happiness (Pawlowski et al., 2014) and happiness from sporting success not to be associated with individual support for bidding for the 2024 Olympic Summer Games in Hamburg, Germany (Wicker & Coates, 2018). On the other hand, happiness from sporting success has been found to have a significant positive effect on public support of hosting any Olympic Summer Games in Germany (Wicker et  al., 2017) and on financial support of Olympic athlete development (Wicker, Hallmann, Breuer, & Feiler, 2012). Happiness from sporting success is not equally distributed among the population: women, individuals with a migration background, and those who participate themselves showed higher levels of happiness from sporting success (Hallmann, Breuer, & Kühnreich, 2013). Finally, some studies have examined the role of the outcome of specific games on SWB. A

smartphone-based study during the 2014 football World Cup revealed that SWB was higher among supporters and spectators after the matches of the German team, with effects augmenting with increasing goal differential (Stieger, Götz, & Gehrig, 2015). In the National Football League (NFL), unexpected wins had a positive effect on life satisfaction (Janhuba, 2017). In an experimental study, fans were found to be significantly happier after a boring win game than after an exciting loss game (Jang, Wann, & Ko, 2018).

Pride Several studies have examined national and civic pride resulting from hosting sport events or sporting success, with many of them also investigating closely related outcomes, such as national identity and national identification. Starting with hosting events, residents tend to anticipate that hosting major sport events, such as the London Olympic Games, can increase national pride and unite people (Atkinson et al., 2008). Notably, research looking at the bid of the Netherlands for the 2018 football World Cup indicates that national pride, solidarity, and identity also come at a price (De Nooij, van den Berg, & Koopmans, 2013). Hence, pride is not automatically related to individuals supporting the hosting of a sport event: for example, being more proud of the city hosting the 1997 World Championships in Skiing did not significantly affect individual support of Trondheim (Norway) residents (Andersson, Rustad, & Solberg, 2004). Importantly, the above studies examined preevent perceptions of pride from hosting events, neglecting the notion that sport events can be experience goods and perceptions might be different after having experienced the event. For example, a Canadian study revealed that the share of respondents being proud of Vancouver hosting the 2010 Winter Olympics was higher after the event than before the event (Humphreys et  al., 2018). Similarly, a German study comparing ex-ante and ex-post perceptions indicated that civic pride can be fostered by collective experiences during the 2006 football World Cup (Süssmuth, Heyne, & Maennig, 2010). Likewise, event-related pride was significantly higher after the 2010 football World Cup in South Africa among residents of five host cities, especially among males and black South Africans (Gibson et al., 2014). Accordingly, Cornelissen and Maennig (2010) concluded that national pride and cohesion were the most important and most durable outcomes of the 2010 FIFA World Cup.

POSITIVE AND NEGATIVE EXTERNALITIES OF SPORT EVENTS

With regard to sporting success, existing research indicates that the prestige of hosting a major sport event is mainly driven by pride from sporting success (Wicker et al., 2017) rather than by the mere staging of an event (Barros, 2006). Accordingly, pride from sporting success had a positive effect on individual support related to bidding for any Olympic Summer Games in Germany (Wicker et  al., 2017). In Japan, perceived social and personal benefits of elite sporting success, including national pride and national identity, were found to positively affect the support of the population for financing the development of Olympic athletes (Funahashi & Mano, 2015). Likewise, pride from sporting success had a positive effect on financial support of Olympic athlete development in Germany (Wicker, Hallmann, et al., 2012). Prior research indicates that pride from sporting success is not equally distributed among the population. For example, a study in the Netherlands revealed that the effects of sporting success on national pride are particularly pronounced for males, athletes, and non-immigrants (Elling, van Hilvoorde, & van den Dool, 2014). In Germany, national pride from sporting success was found to be significantly higher for women and sport participants, but significantly lower for older people and those with higher education (Hallmann et al., 2013). Some studies looked at the time frame of the relationship between sporting success and national pride. For example, a Canadian study revealed that respondents scored higher on different facets of national pride, such as pride from Canadians winning a gold medal and Canada winning more medals than the United States, after the 2010 Vancouver Olympics than before these Games (Humphreys et al., 2018). A Dutch study indicated that the effects of international sporting success on national pride were only small and short-term in nature (Elling et al., 2014). The ability of sporting success to affect national pride has been challenged, however. For instance, a Dutch study suggested that a sense of belonging would be more of a precondition for national pride from sporting success rather than a result of it (van Hilvoorde, Elling, & Stokvis, 2010). Scholars have also concluded that national pride could also be a stable characteristic of national identification, which is difficult to change by sporting success (Elling et al., 2014; van Hilvoorde et al., 2010). Similarly, a German study suggested that sport competitions may not be ‘serious’ enough to raise ‘real’ national pride (Haut, Prohl, & Emrich, 2014). Another German study showing no significant relationship between national identity and the perceived value of sporting success supports this

433

notion (Wicker, Kiefer, & Dilger, 2015). On the contrary, a different German study found a positive relationship between national identity and the value of sporting success (Wicker, Prinz, & von Hanau, 2012). Moreover, Topic and Coakley (2010) pointed to differences between established nation-states and developing nations, such as postSoviet nations, where sport and sport events can facilitate establishing and maintaining national identity.

Social Capital Moving from the level of national identity and pride, research has studied social capital and wider social outcomes of sport events, such as local unity. For example, expectations of local unity were regarded as one potential personal benefit which positively affects the support of the population for financing the development of Olympic athletes in Japan (Funahashi & Mano, 2015). The empirical evidence with regard to social capital and other social outcomes of sport events is mixed. Starting with positive effects, research studying the 2006 football World Cup among Munich residents confirmed beneficial social impacts, particularly with regard to positive fan behaviour and the atmosphere surrounding the event (Ohmann, Jones, & Wilkes, 2006). Another German study found that the hosting of this World Cup contributed to social cohesion and integration in a reunified country (Süssmuth et al., 2010). In South Korea, comparisons between ex-ante and ex-post perceptions of residents regarding the impact of the 2002 football World Cup showed that social problems were perceived as significantly smaller after the event (Kim, Gursey, & Lee, 2006). Moreover, in the sport for development literature, existing research examining the impact of a sport event, the International Run for Peace, on intergroup relations and social capital stocks in Sri Lanka, found evidence for the generation of social capital. Specifically, there was evidence for socio-cultural experiences resulting from the event, including socialising, trust, reciprocity and solidarity, contacts and networks, and learning and development, which also aided in increasing social capital levels and reducing social barriers (Schulenkorf et al., 2011). On the contrary, some studies do not report positive social outcomes. For example, 33.2% of Munich residents anticipated that socially deprived people would not benefit from hosting the 2018 Winter Games (Preuss & Werkmann,

434

THE SAGE HANDBOOK OF SPORTS ECONOMICS

2011). Also, hosting the 2002 football World Cup did not fulfil South Korean residents’ expectations regarding benefits from cultural exchange (Kim et  al., 2006). A study examining social capital before and after the 2010 football World Cup even revealed decreasing levels among three dimensions of social capital (i.e., collective action, social connections, tolerance of diversity), while two dimensions (i.e., value of life, trust and safety) remained unchanged (Gibson et  al., 2014). This study also suggested that not all population groups benefit equally from hosting a major sport event: social capital levels were found to be higher for black and younger South Africans (Gibson et al., 2014). Taken together, the empirical evidence suggests that sport events can be considered as catalysts for enhancing social capital levels among residents (Schulenkorf et al., 2011). Having said this, they cannot ‘fix’ social problems, but can assist in community development by exploiting the community’s strength (Misener & Schulenkorf, 2016).

NEGATIVE EXTERNALITIES As noted earlier, policy tends to focus on the positive benefits from hosting sport events and this is even implied in Figure 42.1. However, it is well known in the literature that sport events can have negative impacts on individuals and communities. Some of these negative impacts are discussed below.

Traffic, Congestion, and Pollution Sport events can affect the daily routines of residents, for example through increased travel by event participants and spectators traveling to, during, and from the sport event. This eventrelated travel causes carbon dioxide emissions and pollution (e.g., Jones, 2008), which might be perceived to be a cost imposed on local residents. Indeed, possible crowding and transport issues were listed among the expected intangible costs of the 2012 Games by people before the event – both by London residents (Atkinson et al., 2008) and people living in Weymouth and Portland (Ritchie, Shipway, & Cleeve, 2009). In Munich, 37.6% of residents anticipated environmental damages as a result of the city hosting the 2018 Winter Games (Preuss & Werkmann, 2011). Expectations of traffic and crowding can also affect support levels for an event. For example, Trondheim (Norway) residents considering

crowding a problem showed lower support for hosting the 1997 World Championships in Skiing, while traffic conditions and noise had no effect on individual support (Andersson et al., 2004). In South Korea, resident perceptions about negative impacts of traffic congestion and pollution associated with hosting the 2002 football World Cup turned out to be less severe than expected before the event (Kim et al., 2006).

Criminal Behaviour Examining the link between sport events and criminal behaviour, including domestic violence (e.g., Kirby, Francis, & O’Flaherty, 2014), has attracted increasing research interest within sports economics in recent years. The review of eventrelated research is enriched by studies examining league games in professional team sports. Starting with perceptions before the event took place, London residents stated increased safety and security risks as well as risk of theft as a major intangible cost of hosting the 2012 Summer Games (Atkinson et al., 2008). After the event, only a few Munich residents perceived increased crime and prostitution as an issue resulting from Germany hosting the 2006 football World Cup (Ohmann et al., 2006). A few studies have examined the link between major sport events and criminal behaviour by assessing official records after the event took place. For example, a North American study found that the Olympic Games were associated with a 10% increase in property crime, while the Super Bowl yielded a decrease in violent crime by 2.5% (Baumann, Ciavarra, Englehardt, & Matheson, 2012). Likewise, research examining the 2002 Winter Olympics showed that, despite the presence of social control agents, routine crime increased significantly in Salt Lake City, measured by calls for police service, crime reports, and police arrests (Decker, Varano, & Greene, 2007). In the UK, Kirby et al. (2014) identified that a rise in domestic violence followed the national football team playing in successive tournaments, with more violence arising if England lost, and a rising trend over time. The majority of research has studied sport leagues, with the empirical evidence being inconclusive, as suggested in the review by Inoue et al. (2015). For example, Baumann et al. (2012), studying teams in the four major professional sports leagues, did not find any significant effects of league games on property crime or violent crime. On the contrary, a study looking only at the NFL reported that home games were associated with an increase in total crimes by 2.6%, with larceny

POSITIVE AND NEGATIVE EXTERNALITIES OF SPORT EVENTS

increasing by 4.1% and theft by 6.7% (Kalist & Lee, 2016). Likewise, prior studies documented an increase in larceny and assaults on home game days of the Detroit Lions (NFL) (Pyun & Hall, 2016) and the Washington Nationals (Major League Baseball) (Pyun, 2018), respectively. In college football, alcohol-related arrests (Merlo, Hong, & Cottler, 2010) and other crimes, such as assaults, vandalism, and arrests for disorderly conduct (Rees & Schnepel, 2009), were found to be significantly higher on game days than on nongame days. Turning to European soccer, property crimes were found to increase in communities hosting matches of London teams, presumably as a result of police displacement during home games (Marie, 2016). A study looking specifically at crime in the area that surrounds the Wembley Stadium in the UK documented a higher rate of crime per-unit time when the stadium was used compared to when not (Kurland, Johnson, & Tilley, 2014). Prior research has also studied the relationship between game outcome and criminal behaviour. Looking at North American professional sports teams in 30 metropolitan areas, Fernquist (2000) reported that making the playoffs is significantly associated with a decline in suicide and homicide rates, while winning championships is only related to declines in suicide rates. A study looking specifically at professional football found that upset losses were associated with a 10% increase in domestic violence, with the rise in violence being concentrated at the end of games and larger in size for more important games (Card & Dahl, 2011). Likewise, in college football, upsets were related to larger increases in various offences on game days (Rees & Schnepel, 2009). Within soccer, frustration resulting from an unexpected loss yielded a spike in violent crime, while euphoria resulting from an unexpected win was associated with a reduction (Munyo & Rossi, 2013). These findings support the notion that not only the mere hosting of an event or league game, but also the actual sporting outcome can yield criminal behaviour.

CONCLUSIONS AND FUTURE RESEARCH This chapter has set out to reflect upon both the positive and negative externalities of sport events through the lens of sports economics. Existing research has shown that hosting or being successful at sports events can have quite mixed results on the social welfare of society. For example, while the literature above has identified positive impacts

435

on SWB and social capital, the results are much less robust for health and pride.1 It is also true that policy tends to neglect the negative impacts that sport events can bring, such as congestion and pollution, which can of course also affect health, as well as crime and antisocial behaviour. The latter impacts are much more researched in the geographical and environmental literatures and the sociological literature. It is clear that future research in sports economics has an agenda here in seeking to address these issues further and to integrate them into a more holistic assessment of the value associated with sport events.

Note 1  Existing research also stresses that positive event outcomes might not occur because an event takes place, but that strategic processes are required to achieve social leverage of sport events (Chalip, 2018).

REFERENCES Andersson, T., Rustad, A., & Solberg, H. A. (2004). Local residents’ monetary evaluation of sports events. Managing Leisure, 9(3), 145–158. Atkinson, G., Mourato, S., Szymanski, S., & Ozdemiroglu, E. (2008). Are we willing to pay enough to ‘back the bid’? Valuing the intangible impacts of London’s bid to host the 2012 Summer Olympic Games. Urban Studies, 45, 419–444. Barget, E., & Gouguet, J. J. (2007). The total economic value of sporting events theory and practice. Journal of Sports Economics, 8(2), 165–182. Barros, C. P. (2006). Evaluating sport events at European level: The Euro 2004. International Journal of Sport Management and Marketing, 1(4), 400–410. Bator, F. M. (1957). The simple analytics of welfare maximization. The American Economic Review, 47(1), 22–59. Baumann, R., Ciavarra, T., Englehardt, B., & Matheson, V. A. (2012). Sports franchises, events, and city livability: An examination of spectator sports and crime rates. The Economic and Labour Relations Review, 23(2), 83–98. Breuer, C., & Wicker, P. (2015). Arguments and evidence of bridging elite performance and mass participation sports from an economic perspective. In R. Bailey & M. Talbot (Eds.), Elite sport and sportfor-all: Bridging the two cultures? (pp. 30–40). London: Routledge.

436

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Card, D., & Dahl, G. B. (2011). Family violence and football: The effect of unexpected emotional cues on violent behavior. The Quarterly Journal of Economics, 126(1), 103–143. Chalip, L. (2018). Trading legacy for leverage. In I. Brittain, J. Bocarro, T. Byers, & K. Swart (Eds.), Legacies and mega events: Fact or fairy tales? (pp. 25–41). London: Routledge. City of Gold Coast. (2015). Gold Coast sport plan – 2013–2023. City of Gold Coast. Cornelissen, S., & Maennig, W. (2010). On the political economy of ‘feel good’ effects at sport megaevents: Experiences from FIFA Germany 2006 and prospects for South Africa 2010. Alternation, 17(2), 96–120. De Nooij, M., van den Berg, M., & Koopmans, C. (2013). Bread or Games? A social cost-benefit analysis of the World Cup bid of the Netherlands and the winning Russian bid. Journal of Sports Economics, 14(5), 521–545. Decker, S. H., Varano, S. P., & Greene, J. R. (2007). Routine crime in exceptional times: The impact of the 2002 Winter Olympics on citizen demand for police services. Journal of Criminal Justice, 35(1), 89–101. Department of Culture, Media and Sport (DCMS). (2010). Plans for the legacy from the 2012 Olympic and Paralympic Games. London: DCMS. Retrieved December 13, 2017 from: www. gov.uk/government/uploads/system/uploads/ attachment_data/file/78105/201210_Legacy_ Publication.pdf Diener, E., Oishi, S., & Lucas, R. E. (2002). Subjective well-being: The science of happiness and life satisfaction. In C. R. Snyder & S. J. Lopez (Eds.), The handbook of positive psychology (pp. 63–73). New York: Oxford University Press. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sport economics: Theory, evidence and policy. Oxford: Butterworth-Heinemann. Downward, P., Dawson, P., & Mills, T. (2016). Sports participation as an investment in (subjective) health: A time series analysis of the life course. Journal of Public Health, 38(4), e504–e510. Elling, A., van Hilvoorde, I., & van den Dool, R. (2014). Creating or awakening national pride through sporting success: A longitudinal study on macro effects in the Netherlands. International Review for the Sociology of Sport, 49(2), 129–151. Fergusson, D. M., McLeod, G. F. H., Horwood, L. J., Swain, N. R., & Poulton, R. (2015). Life satisfaction and mental health problems (18 to 35 years). Psychological Medicine, 45, 2427–2436. Fernquist, R. M. (2000). An aggregate analysis of professional sports, suicide, and homicide rates: 30 U.S. metropolitan areas, 1971–1990. Aggression and Violent Behavior, 5(4), 329–341.

Frey, B. (2008). Happiness: A revolution in economics (Munich lectures in economics). Cambridge, MA: MIT Press. Funahashi, H., & Mano, Y. (2015). Socio-psychological factors associated with the public’s willingness to pay for elite sport policy: Does risk perception matter? Managing Sport and Leisure, 20(2), 77–99. Gibson, H. J., Walker, M., Thapa, B., Kaplanidou, K., Geldenhuys, S., & Coetzee, W. (2014). Psychic income and social capital among host nation residents: A pre-post analysis of the 2010 FIFA World Cup in South Africa. Tourism Management, 44, 113–122. Gravelle, H., & Rees, R. (2004). Microeconomics (3rd ed.). Harlow, UK: Prentice Hall. Grix, J. (2012). ‘Image’ leveraging and sports megaevents: Germany and the 2006 FIFA World Cup. Journal of Sport & Tourism, 17(4), 289–312. Hallmann, K., Breuer, C., & Kühnreich, B. (2013). Happiness, pride and elite sporting success: What population segments gain most from national athletic achievements? Sport Management Review, 16(2), 226–235. Haut, J., Prohl, R., & Emrich, E. (2014). Nothing but medals? Attitudes towards the importance of Olympic success. International Review for the Sociology of Sport, 51(3), 332–348. Hiller, H. H., & Wanner, R. A. (2015). The psychosocial impact of the Olympics as urban festival: A leisure perspective. Leisure Studies, 34(6), 672–688. HM Government. (2015). Sporting future: A new strategy for an active nation. London: HM Government. Retrieved April 15, 2016 from: www. gov.uk/government/uploads/system/uploads/ attachment_data/file/486622/Sporting_Future_ ACCESSIBLE.pdf HM Treasury. (2011). Valuation techniques for social cost-benefit analysis: Stated preference, revealed preference and subjective well-being approaches. A discussion of the current issues. London: HM Treasury. Retrieved December 14, 2017 from: www.gov.uk/government/uploads/system/ uploads/attachment_data/file/209107/greenbook_ valuationtechniques.pdf Humphreys, B. R., Johnson, B. K., Mason, D. S., & Whitehead, J. C. (2018). Estimating the value of medal success in the Olympic Games. Journal of Sports Economics, 19(3), 398–416. Humphreys, B. R., McLeod, L., & Ruseski, J. E. (2014). Physical activity and health outcomes: Evidence from Canada. Health Economics, 23(1), 33–54. Inoue, Y., Berg, B. K., & Chelladurai, P. (2015). Spectator sport and population health: A scoping study. Journal of Sport Management, 29(6), 705–725. Jang, W., Wann, D. L., & Ko, Y. J. (2018). Influence of team identification, game outcome, and game

POSITIVE AND NEGATIVE EXTERNALITIES OF SPORT EVENTS

process on sport consumers’ happiness. Sport Management Review, 21(1), 63–71. https://doi. org/10.1016/j.smr.2017.03.002 Janhuba, R. (2017). Do victories and losses matter? Effects of football on life satisfaction. CERGE-EI Working Paper Series No. 579. Retrieved December 08, 2017 from: https://ssrn.com/abstract=2886008 Jones, C. (2008). Assessing the impact of a major sporting event: The role of environmental accounting. Tourism Economics, 14(2), 343–360. Kalist, D. E., & Lee, D. Y. (2016). The National Football League: Does crime increase on game day? Journal of Sports Economics, 17(8), 863–882. Kasimati, E. (2003). Economic aspects and the Summer Olympics: A review of related research. International Journal of Tourism Research, 5, 433–444. Kavetsos, G., & Szymanski, S. (2010). National wellbeing and international sports events. Journal of Economic Psychology, 31(2), 158–171. Kim, H. J., Gursey, D., & Lee, S. (2006). The impact of the 2002 World Cup on South Korea: Comparisons of pre- and post-games. Tourism Management, 27(1), 86–96. Kirby, S., Francis, B., & O’Flaherty, R. (2014). Can the FIFA world cup football (soccer) tournament be associated with an increase in domestic abuse? Journal of Research in Crime and Delinquency, 51(3), 259–276. Koivumaa-Honkanen, H., Kaprio, J., Honkanen, R., Viinamaki, H., & Koskenvuo, M. (2004). Life satisfaction and depression in a 15-year follow-up of healthy adults. Social Psychiatry and Psychiatric Epidemiology, 39, 994–999. Kurland, J., Johnson, S. D., & Tilley, N. (2014). Offenses around stadiums: A natural experiment on crime attraction and generation. Journal of Research in Crime and Delinquency, 51(1), 5–28. Marie, O. (2016). Police and thieves in the stadium: Measuring the (multiple) effects of football matches on crime. Journal of the Royal Statistical Society: Series A, 179(1), 273–292. McCartney, G., Thomas, S., Thomson, H., Scott, J., Hamilton, V., Halon, P., Morrison, D. S., & Bond, L. (2010). The health and socioeconomic impacts of major multi-sport events: Systematic review (1978–2008). British Medical Journal, 340, 1–9. Merlo, L. J., Hong, J., & Cottler, L. B. (2010). The association between alcohol-related arrests and college football game days. Drug and Alcohol Dependence, 106(1), 69–71. Misener, L., & Schulenkorf, N. (2016). Rethinking the social value of sport events through an Asset-Based Community Development (ABCD) perspective. Journal of Sport Management, 30(3), 329–340. Munyo, I., & Rossi, M. A. (2013). Frustration, euphoria, and violent crime. Journal of Economic Behavior & Organizations, 89, 136–142.

437

Murphy, N. M., & Bauman, A. (2007). Mass sporting and physical activity events – are they ‘bread and circuses’ or public health interventions to increase population levels of physical activity? Journal of Physical Activity and Health, 4(2), 193–202. Noll, R. G., & Zimbalist, A. (Eds.) (2011). Sports, jobs, and taxes: The economic impact of sports teams and stadiums. Washington, DC: Brookings Institution Press. Ohmann, S., Jones, I., & Wilkens, K. (2006). The perceived social impacts of the 2006 Football World Cup on Munich residents. Journal of Sport & Tourism, 11(2), 129–152. Pawlowski, T., Downward, P., & Rasciute, S. (2014). Does national pride from international sporting success contribute to well-being? An international investigation. Sport Management Review, 17, 121–132. Preuss, H., & Werkmann, K. (2011). Experiential value of hosting the 2018 Winter Olympics in Munich. Sport and Society, 8(2), 97–123. Pyun, H. (2018). Exploring causal relationship between Major League Baseball games and crime: A synthetic control analysis. Empirical Economics, online first, May 26, 2018, pp. 1–19. doi: 10.1007/ s00181-018-1440-9 Pyun, H., & Hall, J. C. (2016). Does the presence of professional football cause crime in a city? Evidence from Pontiac, Michigan. Working Paper No. 16-02, College of Business and Economics. Retrieved December 08, 2017 from: http://be.wvu. edu/phd_economics/working-papers.htm Rasciute, S., Downward, P., & Greene, W. H. (2017). Do relational goods raise well-being? An econometric analysis. Eastern Economic Journal, 43(4), 563–579. Rees, D. I., & Schnepel, K. T. (2009). College football games and crime. Journal of Sports Economics, 10(1), 68–87. Ritchie, B. W., Shipway, R., & Cleeve, B. (2009). Resident perceptions of mega-sporting events: A non-host city perspective of the 2012 London Olympic Games. Journal of Sport & Tourism, 14(2–3), 143–167. Rowe, D., & McGuirk, P. (1999). Drunk for three weeks: Sporting success and city image. International Review for the Sociology of Sport, 34(2), 125–141. Schlegel, A., Pfitzner, R., & Königstorfer, J. (2017). The impact of atmosphere in the city on subjective well-being of Rio de Janeiro residents during (vs. before) the 2014 FIFA World Cup. Journal of Sport Management, 31(6), 605–619. Schulenkorf, N., Thomson, A., & Schlenker, K. (2011). Intercommunity sport events: Vehicles and catalysts for social capital in divided societies. Event Management, 15(2), 105–119.

438

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Stieger, S., Götz, F. M., & Gehrig, F. (2015). Soccer results affect subjective well-being, but only briefly: a smartphone study during the 2014 FIFA World Cup. Frontiers in Psychology, 6, 1–8. Stiglitz, J. E., Sen, A., & Fitoussi, J. P. (2010). Mismeasuring our lives: Why GDP doesn’t add up. New York: The New Press. Süssmuth, B., Heyne, M., & Maennig, W. (2010). Induced civic pride and integration. Oxford Bulletin of Economics and Statistics, 72(2), 202–220. Taks, M., Littlejohn, M., Snelgrove, R., & Wood, L. (2016). Sport events and residential happiness: The case of two non-mega sport events. Journal of Global Sport Management, 1(3–4), 90–109. Topic, M. D., & Coakley, J. (2010). Complicating the relationship between sport and national identity: The case of post-socialist Slovenia. Sociology of Sport Journal, 27(4), 371–389. Van Hilvoorde, I., Elling, A., & Stokvis, R. (2010). How to influence national pride? The Olympic medal index as a unifying narrative. International Review for the Sociology of Sport, 45(1), 87–102. Walton, H., Longo, A., & Dawson, P. (2008). A contingent valuation of the 2012 London Olympic Games: A regional perspective. Journal of Sports Economics, 9, 304–317.

WHO. (2010). Global recommendations on physical activity for health. Geneva: World Health Organisation. Retrieved 16 January, 2015 from: ­ http://whqlibdoc.who.int/publications/2010/ 9789241599979_eng.pdf?ua=1 Wicker, P., & Coates, D. (2018). Flame goes out: Determinants of individual support at the 2024 Hamburg Games referendum. Contemporary Economic Policy, 36(2), 302–317. Wicker, P., Hallmann, K., Breuer, C., & Feiler, S. (2012). The value of Olympic success and the intangible effects of sport events: A contingent valuation approach in Germany. European Sport Management Quarterly, 12(4), 337–355. Wicker, P., Kiefer, S., & Dilger, A. (2015). The value of sporting success to Germans: Comparing the 2012 UEFA Championships with the 2012 Olympics. Journal of Business Economics, 85(8), 897–919. Wicker, P., Prinz, J., & von Hanau, T. (2012). Estimating the value of national sporting success. Sport Management Review, 15(2), 200–210. Wicker, P., Whitehead, J. C., Mason, D. S., & Johnson, B. K. (2017). Public support for hosting the Olympic Summer Games in Germany: The CVM approach. Urban Studies, 54(15), 3597–3614.

PART VI

Individual Sports

43 The Economics of Running Bernd Frick, Katharina Moser and Katrin Scharfenkamp

RUNNING: ANCIENT LEGEND AND CONTEMPORARY SITUATION In 490 BC, an army from Persia landed in Marathon, about 25 miles from Athens, with the intention of capturing and enslaving the city. While the Athenians prepared for the battle, they sent a messenger named Philippides to Sparta to ask the Spartans to support them in the upcoming fight. In the meantime, back in Marathon the decision was made not to wait for the Spartans. Surprisingly, the Athenian army defeated the much larger Persian forces and sent a runner – most likely not Philippides again – to Athens to carry the news of the victory. According to legend, the messenger reached the city, spread the news and then died of exhaustion. When Pierre de Coubertin, a French aristocrat with a keen interest in athletics, revived the ancient Olympic Games in 1896, the footrace from Marathon to Athens was considered the greatest of all events, because it linked the legend of Philippides to modern Greece. Until today, winning the Olympic Marathon remains the ultimate accomplishment for any elite long-distance runner. Running in general and marathon running in particular has become a business in recent years.

While a large number of specialized travel agencies offer their services to thousands of runners who prepare for the big city marathons, such as New York, London, Berlin, Boston, and Chicago (to name just a few), the organizers of those events compete for a small number of top athletes, who – due to long recovery periods – can participate in a limited number of races only (usually two full marathons per year). The athletes, in turn, compete for monetary rewards that vary across races and change over time. In this chapter, we first describe the major events in professional middle- and long-distance running before turning to the incentive and selection effects of prize money. We then go on to review the available evidence on the changing composition of the world’s marathon elite and the gender difference in competitiveness. The second half of the chapter is devoted to recreational running with a particular emphasis on issues of self-selection and (over)confidence, the impact of training, pacing and socio-demographic characteristics on recreational runners’ marathon performance and the impact of chronological age on the performance of male and female runners.

442

THE SAGE HANDBOOK OF SPORTS ECONOMICS

THE ECONOMICS OF PROFESSIONAL MIDDLE- AND LONG-DISTANCE RUNNING Major Events in Middle- and Long-distance Running Apart from the quadrennial Olympic Summer Games and the biannual IAAF World Championships, elite runners compete for titles and prize money in continental and national championships and in especially designed contests, such as city marathons, the ‘World Marathon Majors’ and the ‘International Association of Athletics Federation (IAAF) Diamond League’. Table 43.1 shows that the prize purses offered by the organizers of the most prestigious city marathons differ considerably in terms of size and in the way the money is distributed across the top finishers. In Chicago, only the top five are entitled to receive prize money, whereas in Boston the top 15 are paid an ex ante announced share of the pie. Thus, different marathons have different strategies for their prize purse. Some simply provide payouts to the top athletes based on the order in which they finish, while others set the prize purse based on the order of finish, but then supplement that with time bonuses to encourage the athletes to run fast as well. London – focusing on time – goes to the extreme, where the finish order prizes are rather low (compared to the other races displayed in Table 43.1), but many of the top runners earn

more money through the time bonuses than they will from the placement prizes. The time bonuses to be earned in London are quite significant (see Table 43.2 and Table 43.3) and are thought to ensure a fast race. On top of that, it is well known, that many of the invited elite athletes have their own contracts with the London organization in which their appearance fees are tied to the achievement of specific time goals. The ‘World Marathon Majors (WMM)’ (since 2015 the Abbott World Marathon Majors) is a championship style competition for marathon runners that started in 2006. It comprises six annual races in the cities of Tokyo (starting in 2013), Boston, London, Berlin, Chicago and New York City (except 2012 when the race had to be cancelled due to weather conditions), a biennial race (the IAAF World Championships Marathon), and a quadrennial race (the Olympic Games Marathon). Each WMM series originally spanned two calendar years where the second year of a series overlapped with the first year of the next. Starting in 2015, each series began with a city race and ended with next year’s race in the same city; Series IX started in February 2015 at the Tokyo Marathon and ended there in February 2016; Series X started and finished with the Boston Marathon 2016/2017; Series XI started with the London Marathon in 2017 and ended there in April 2018. At the end of each WMM series, the leading man and woman each win US$500,000, while the contenders receive no prize money at all, i.e. the competition is a ‘winner-takes-all’ contest.

Table 43.1  Prize purses, 2017, in USD Rank

London

Chicago

New York

Boston

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

55,000 30,000 22,500 15,000 10,000 7,500 5,000 4,000 3,000 2,000 1,500 1,000 — — —

100,000 50,000 25,000 15,000 10,000 — — — — — — — — — —

100,000 60,000 40,000 25,000 15,000 10,000 7,500 5,000 2,500 2,000 — — — — —

150,000 75,000 40,000 25,000 15,000 12,000 9,000 7,400 5,700 4,200 2,600 2,100 1,800 1,700 1,500

Sources: MarathonGuide.com (2016a); Chicagomarathon (2016a); Tcsnycmarathon (2016a); Bostonmarathonmediaguide (2016a)

443

THE ECONOMICS OF RUNNING

Table 43.2  Time bonuses, men 2017, in USD Time

London

Chicago

New York

Boston

World Record Course Record Sub 2:05:00 Sub 2:05:30 Sub 2:06:00 Sub 2:06:30 Sub 2:07:00 Sub 2:07:30 Sub 2:08:00 Sub 2:08:30 Sub 2:09:00 Sub 2:09:30 Sub 2:10:00 Sub 2:11:00

125,000 25,000 100,000 — 75,000 — 50,000 — 25,000 15,000 10,000 5,000 3,000 1,000

— 50,000 — — 40,000 30,000 20,000 — 10,000 — — — — —

— — — 50,000 45,000 40,000 35,000 30,000 25,000 — 15,000 — 10,000 —

50,000 25,000 — — — — — — — — — — — —

Sources: MarathonGuide.com (2016b); Chicagomarathon (2016b); Tcsnycmarathon (2016b); Bostonmarathonmediaguide (2016b)

Table 43.3  Time bonuses, women 2017, in USD Time

London

Chicago

New York

Boston

World Record Course Record Sub 2:18:00 Sub 2:20:00 Sub 2:21:00 Sub 2:22:00 Sub 2:22:30 Sub 2:23:00 Sub 2:23:30 Sub 2:24:00 Sub 2:24:30 Sub 2:25:00 Sub 2:26:00 Sub 2:27:00 Sub 2:28:00

125,000 25,000 100,000 75,000 — 50,000 — 25,000 — 15,000 — 10,000 5,000 3,000 1,000

— 50,000 — 40,000 30,000 20,000 —

— — — — — — 50,000 45,000 40,000 35,000 30,000 25,000 15,000 10,000 —

50,000 25,000 — — — — — — — — — — — — —

— 10,000 — — — — —

Sources: MarathonGuide.com (2016c); Chicagomarathon (2016c); Tcsnycmarathon (2016c); Bostonmarathonmediaguide (2016c)

Athletes originally received points for finishing in any of the top five places (1st: 25 points; 2nd: 15 points; 3rd: 10 points; 4th: 5 points, 5th: 1 point). Their four highest ranks over the two-year period were counted; if an athlete scored points in more races, that person’s four best races were scored. To be eligible for the jackpot, an athlete had to compete in at least one qualifying race in each calendar year of the series. In 2015, the scoring

was revised to 1st place 25 points, then 16 (2nd), 9 (3rd), 4 (4th) and 1 (5th). The two highest ranks during the scoring period were counted, with only the best two if more than that number. The ‘IAAF Diamond League’ has also implemented a championship style model with the finalists competing for a prize pool of US$3.2 million. Athletes earn points in the first 12 IAAF Diamond League meetings to qualify for two final meetings

444

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 43.4  Prize money (in USD) in middle- and long-distance running events, 2017 Rank

1 2 3 4 5 6 7 8

IAAF Diamond League Qualification Meeting

Final

10,000 6,000 4,000 3,000 2,500 2,000 1,500 1,000

50,000 20,000 10,000 6,000 5,000 4,000 3,000 2,000

IAAF Track & Field World Championship

IAAF World Championship Cross Country & Half Marathon

60,000 30,000 20,000 15,000 10,000 6,000 5,000 4,000

30,000 15,000 10,000 7,000 5,000 3,000 — —

Sources: MarathonGuide.com (2016d); Athletics (2017); IAAF (2017)

where US$100,000 is at stake in each of the 32 Diamond Disciplines, including US$50,000 for each winner. Thus, the season is a race to reach the finals with the winners crowned as ‘IAAF Diamond League Champions’. As in a championship, the performance of athletes in the final will only determine who the champion will be and the prize money won. At each of the qualification meetings, the winner of each Diamond Discipline will receive US$30,000 prize money. Each qualification meeting will stage either 13 or 14 Diamond Disciplines, meaning the total prize money awarded at a single meeting will total either US$390,000 or US$420,000. Each of the disciplines is staged either four or six times before the final. At each of the 12 qualification meetings, athletes are awarded 8 to 1 point(s) for ranking 1st to 8th respectively. The top eight or 12 athletes (depending on the discipline) will be awarded a start at the final. In case of a tie, the best legal performance of the qualification phase wins. The winner at the final of each Diamond Discipline will become ‘IAAF Diamond League Champion’ and be awarded a Diamond Trophy, US$50,000 prize money and a wild card for the IAAF World Championships. In 2017, the Diamond League included the 800 meters, 1,500 meters, 3,000 meters, 5,000 meters and 3,000 meters steeplechase events for both, men and women. At the IAAF World Championships, the traditional middle- and long-distance events are the 800 meters, 1,500 meters, 5,000 meters, 10,000 meters, marathon, and 3,000 meters steeplechase. Moreover, the IAAF World Champion­ ships in cross-country and half-marathon now also offer prize money to the top six finishers.

Incentives and Performance Based on the theoretical approaches of Lazear and Rosen (1981) and Rosen (1988),1 race organizers

can be assumed to adjust the size and the distribution of the purse to maximize athletic performance which, in turn, leads to an increase of the organizer’s reputation. A higher reputation secures or even increases media attention for next year’s race and, at the same time, makes the race an attractive venue for potential sponsors. Income-maximizing athletes who are usually familiar with the abilities of their competitors know that winning a sizeable piece of the prize money ‘pie’ requires effort that they increase until the marginal utility of effort equals marginal costs. The assumption that athletes constantly compare the costs and benefits of (an additional unit of) effort can be confirmed by looking at individual behavior during a race: Runners who still belong to the leading group beyond the 30 kilometer mark will as a rule quit the race only in case of an injury. On the other hand, indisposed athletes typically drop out of the race prior to the 25 kilometer mark. These seemingly incompatible observations can be reconciled easily: the regeneration period after a marathon usually takes several weeks before the body is fully recovered. The length of the recuperation period increases over-proportionally with every kilometer completed between the 25 and the 35 kilometer mark, while increasing only slightly thereafter. Athletes trying to maximize their income therefore not only consider the prize money yields of the respective race, but also calculate the opportunity costs in the form of foregone prize money in the event that an unsuccessful finish of the race deters their participation in races for the weeks to come. In light of this dilemma, the choice of one of the two alternatives – quitting a race early versus finishing it – can be interpreted as an attempt to allocate physical resources in a way that is compatible with the idea of income maximization. An elite athlete who drops out of a race after 32 kilometers faces a long recovery period similar to that faced by one who finishes

THE ECONOMICS OF RUNNING

the race, whereas the one who abandons after 25 kilometers at the latest can go to the start again after a relatively short period of 10–14 days. Thus, the parameters of the prize money structure that are at the disposal of the race organizers to maximize the performance of the athletes must be investigated separately as to their incentive, selection, and cooperation effects. Apart from the obvious incentive and (self-) selection effects of tournaments, race organizers must also consider the peculiarities of most endurance events when designing the prize money structure. First, since athletes in a marathon must be encouraged to cooperate (even elite runners need the support of a number of equally able competitors), a ‘winnertakes-all’ prize structure will be detrimental to the performance of the field. Second, since elite athletes who ‘have caught a bad day’ are most likely to leave the race rather early, a higher number of money ranks may be a suitable instrument to keep these runners on the road by increasing the opportunity costs of relinquishing. Thus, the parameters to be considered are the total purse, additional bonuses that reward absolute performance, the distribution of prize money to the individual ranks, the prize money differences between two adjacent ranks, and the number of ‘in-the-money-ranks.’ Maloney and McCormick (2000) use data from 115 races, ranging in distance from one mile to a full marathon, that were held in the southeastern United States between 1987 and 1991. They found that both the average prize paid and the prize spread had the predicted negative and statistically significant influence on finish times. Doubling the average prize leads to a fall in average times of about 2% and doubling the prize spread leads to a fall in average times of about 4%. Unfortunately, the data used cover predominantly provincial races that attract small numbers of elite runners only. With an average finishing time for 10 km races of 34:40 and an average prize of, at most, US$450 (in the open class for men), the races covered seem to be of a rather low quality. The maximum purse paid was US$23,900; this is a small fraction of the sums paid to the top finishers in the big city marathons. Moreover, pooling data from races as short as one mile and as long as a full marathon is likely to produce findings that can hardly be interpreted, because the tactics and strategies of long-distance runners are significantly different from the ones employed by runners in middle distances. Using data from 135 different races ranging in distance from 5 kilometers to a full marathon held between 1993 and 1995, Lynch and Zax (2000) also support the hypothesis that times are faster in races offering higher prize money. However, they argue that this is not due to the incentive effects of financial rewards but to sorting: when runner

445

ability is controlled through fixed effects or ranking points, the incentive effects decrease or completely disappear. While information on the total purses of the races under consideration is not provided by the authors, the relatively small prize differences between two adjacent places reported in Table 2 of the paper indicate that, as in Maloney and McCormick (2000), the data set is likely to contain many races with rather low purses and, therefore, presumably a rather low quality of the field. Using data from 13 different city marathons held in the United States, Britain and Germany between 1983 and 2001 (n=57 different races), Frick and Prinz (2007) show that doubling the average prize leads to a reduction in finishing times by more than 1.5%, or about 120 seconds. This is attributed to the selection or participation effect – a higher purse attracts better runners. The coefficient of the prize spread indicates that if the spread doubles, average times fall by almost 7%. This is attributed to the individual responses of runners to the payoffs they face during the competition. As predicted, smaller prize money differences between two adjacent ranks lead to a slower race, while doubling the bonus payments (the rewards for absolute performance) leads to a reduction of finish times by 3%. However, controlling for differences in runner abilities, it appears that both the coefficient of the total prize purse and the coefficient of the available time bonuses, are now statistically insignificant, suggesting that predominantly differences in runner abilities explain the variation in finishing times. When looking at the determinants of the ‘tightness’ or ‘excitement’ of the races, a different picture emerges. First, a larger prize money ‘pie’ produces tighter races, while the same implication can be inferred from the coefficients of the prize structure variables. In particular, the negative sign of the concentration of the prize money variable supports a basic argument proposed by tournament theory: a higher degree of inequality induces effort, i.e. since dispersion of the prize money varies considerably across the marathons analyzed, the authors expect – and indeed find – that events with an ‘unequal’ distribution of the prize money are more ‘entertaining’ in the sense that the top runners finish closer to each other. Using data from 368 races with a distance of at least 5 km that were listed in the 2004/2005 edition of the ‘Road Race Management Directory’, Azmat and Möller (2009) find that contests awarding multiple prizes attract more participants than contests offering their entire prize budget to the winner. However, a steeper prize structure increases the participation of top runners, which in turn, leads to better winning times. According

446

THE SAGE HANDBOOK OF SPORTS ECONOMICS

to the formal model developed in the paper, the optimal prize structure becomes steeper as the race distance increases, because in longer races winning probabilities are less affected by changes in runners’ efforts. This is due to the fact that, first, the impact of external factors beyond the athlete’s control (such as weather conditions and course profile, for example) increase with the distance of the race and, second, longer races exhibit a higher level of randomness than shorter races. Consistent with the model, the empirical evidence shows that the prize structure of a marathon is 3.7% more concentrated toward the first prize than the prize structure of a 5 km race. When moving from 5 km to marathon, the prize awarded to the winner increases by about 8% while the difference between first and second prize increases by 48%.

The Changing Composition of the World’s Marathon Elite The world’s marathon elite today is dominated by East African runners, primarily from Kenya and Ethiopia. However, in the mid-1970s, the female elite consisted almost exclusively of runners from the USA (70%) and Europe (23%) and the male

elite of runners from Europe (36%), Asia (37%), and the USA (14%). Since then, the composition of the elite has changed dramatically. Today, 66% of women and 94% of men come from Africa, and Europeans and North Americans have nearly disappeared (see Figure 43.1).2 Most likely, the percentage of African female runners making it into the ‘bel étage’ of professional road running will increase even further in the near future as the successful runners are considered ‘role models’ by many young girls not only in Kenya and Ethiopia, but also in, for example, Eritrea and Uganda. Closer inspection of the data reveals that the increasing percentage of Africans is associated with significantly faster finish times among both genders and a lower performance dispersion (i.e. a smaller standard deviation) especially among women, suggesting that the performance of the top athletes is increasing with increasing size of the talent pool. Moreover, since the introduction of the WMM series, African athletes have won 109 out of the 126 city marathons that are part of this championship (70 of the winners came from Kenya, 38 from Ethiopia and one from Eritrea) and 16 of the 18 gold medals at the World Championships and the Summer Olympics (10 of these went to Kenya, three to Ethiopia, two to Uganda and one to Eritrea).3

1

.8

.6

.4

.2

0 1970

1980

1990 Female African Female European

2000

2010

2020

Male African Male European

Figure 43.1  Percentage of Africans and Europeans in the world’s marathon elite, 1973–2015 (Source: Own compilation)

THE ECONOMICS OF RUNNING

Gender and Competitiveness Among Elite Runners Frick and Klären (1997) and Frick (1998) find that female marathon runners seem to respond more to an increase of the prize purse and to changes in its distribution than males. On the other hand, women seem not to respond to bonus payments. The explanation for the different behaviors of men and women is quite simple: with regard to their performances, the male marathon elite was – and still is – more homogeneous than the female elite. In 2015, for example, the difference between the fastest runner of the year and no. 50 on that list was 3:44 minutes in the men’s field and 5:19 minutes in the women’s. Given the same number of lucrative races for men and women, this implies that members of the female elite can (and indeed do) avoid competing against each other. Such a behavior is not possible for men, who (due to the homogeneity of the competition) will always face other runners of similar strength. Thus, the observable gender differences in competitiveness in the 1980s and early 1990s were mostly due to differences in the presence of equally talented contestants. This, however, has changed considerably with the increasing number of African females entering this highly competitive environment (see Figure 43.1). Using data from ultramarathon running over the period 2005–2009, Frick (2011a) finds that the pool of female runners is far more heterogeneous than the pool of male runners, as indicated by the larger standard deviation of finishing times. However, this performance differential between men and women declines rapidly as time progresses and the decline is larger on the 100 km, i.e. the race where incentives to perform well are particularly high (due to, for example, the existence of a World Cup) than on the 50 km. Moreover, the rank corrected time difference between men and women has declined considerably over the last 40 years on the ‘traditional’ long-distances (3,000 m, 5,000 m, 10,000 m, marathon) too and continues to narrow. This suggests that the women are catching up rapidly and that the gender gap in competitiveness will have disappeared soon in this (admittedly idiosyncratic) context because the male–female differential in competitiveness declines by about 2% per year (Frick, 2011b). Moreover, it appears that the rank corrected percentage difference in finishing times between men and women is particularly small on the 3,000 meters outdoor track, the 5,000 meters outdoor track, the 10,000 meters outdoor track. These particular track races are equally attractive to men and women because they are part of the IAAF World Championships and/or the annual ‘Diamond League’ meetings (the most

447

lucrative events in track and field athletics apart from city marathons). Marathons, in turn, are also equally attractive to male and female athletes as the prize purses have increased considerably in the recent past and are equal for men and women. Thus, it appears that in the race types where prize money and/or prestige is particularly high, the gender difference in competitiveness is significantly lower than in races that are – for various reasons – considered less attractive, suggesting that today men and women respond similarly to (changes in) financial incentives.

THE ECONOMICS OF RECREATIONAL RUNNING Training, Pacing, and Marathon Performance Using data from an online survey with some 2,300 recreational runners, Vickers and Vertosick (2016) find that men running intervals during their regular training have marathon times that are on average 4:46 minutes faster than those of men not doing intervals. For women the respective difference is just 3:07 minutes. Not surprisingly, the weekly mileage is far more important for marathon finish times than (occasional) speed work: increasing weekly mileage from 30 to 50 is associated with 16 minutes better finish times for men and 22 minutes for women. This suggests that among recreational runners women benefit comparably more from extensive training while men benefit comparably more from intensive training. March, Vanderburgh, Titlebaum, and Hoops (2011) use a data set including some 320 male and female runners finishing a marathon in the Midwest of the United States from 2005 to 2007 to analyze recreational athletes’ ability to maintain a constant pace, which is typically considered important to successfully finish this kind of race. The course is a 1.6 km loop with pace markers throughout, facilitating the measurement of runners’ pacing behavior electronically by a chip attached to the individuals’ shoes. Pacing is defined as mean velocity during the last 9.7 km divided by that of the first 32.5 km. It appears that men have a significantly larger drop in velocity over the last 10 km than women and that older runners show less of a velocity decrease later in the race than younger runners. Moreover, more able runners, i.e. those finishing in better times, are also better in running at a constant pace. The major problem with these – and comparable – studies is that they only report associations, but

448

THE SAGE HANDBOOK OF SPORTS ECONOMICS

not causal effects. Most likely, people running 50 instead of 30 miles per week are physically stronger and are therefore able to train more, which in turn leads to better race results. Moreover, older runners who are found to be better at pacing may simply be more experienced and, therefore, better able to maintain their pace. Thus, the identification of causal effects remains a challenge for this kind of research.

Chronological Age and Marathon Performance Frick and Schmitz (2017) use a large data set with the finish times of 264,647 German men completing 662,251 marathons and 77,630 German women completing 169,649 international marathons (yielding a total sample of 831,902 finishes) in the years 2002–2013 to analyze the impact of age on the individuals’ performance. Controlling for unobserved heterogeneity and selection effects, Frick and Schmitz (2017) find that men and women get faster until their early fifties when their performance starts to deteriorate (see Figure 43.2). Sixty year-old men and women are still as fast as 20 year-old men and women. Moreover, young men seem to improve their performance faster than young women, while older women seem to slow down faster than older men.4 The differences for men and women of the same age are statistically significant for all

but one age group, as are the differences between two adjacent age groups for all but two female cohorts. The question that the authors fail to answer convincingly is why the performance of female runners deteriorates (much) faster than that of male runners. There are (at least) two competing hypotheses that can potentially explain the observed pattern. First, it may well be that women’s physical fitness declines more rapidly than men’s (‘health hypothesis’). Second, women may be more likely to ‘run for fun’ and continue entering races even when their fitness declines, while men stop entering competitions once they discover that their deteriorating fitness results in poorer performance (‘selection hypothesis’). Moreover, the authors need to better control for sample attrition. It may be that the fitter persons continue to run while those with more injuries and a lower level of fitness stop running at an earlier age. Thus, even the fixed-effects estimates presented in the paper may be biased and either over- or underestimate the impact of age on (running) performance. Given the massive demographic changes that most industrialized countries are confronted with over the next decades, the impact of age on physical and/or mental performance needs to be addressed again to design, for example, adequate pension policies because ‘policies on aging should take into account physical deterioration rates since it may be that societies have been too pessimistic about losses from aging for individuals who stay

7000 6000 5000 4000 3000

Men

2000

Women

1000 0 -1000 -2000

Figure 43.2  Age and marathon performance of recreational runners by gender (Source: Own compilation)

THE ECONOMICS OF RUNNING

healthy and fit. Societies may have passed laws dealing with old people under incorrect assumptions’ (Fair, 1994, p. 117).

Reference Points, Self-selection and (Over-)confidence Using a large data set of marathon finishing times with nearly 10 million observations, Allen et  al. (2017) test for reference dependence, i.e. check whether individuals evaluate outcomes as gains or losses relative to a neutral reference point. The authors provide evidence of bunching of finish times at round numbers (e.g. a four-hour marathon) and interpret this as evidence that reference points are important. While these round numbers may be goals that individuals set to motivate themselves, goals are not always rational expectations. In a small sample of 350 participants of the Warsaw Marathon, Krawczyk and Wilamowski (2017) find that most recreational runners are indeed competent forecasters, i.e. their actual finish time is (nearly) identical with the finish time predicted the day before the race. Closer inspection of the socio-demographic characteristics of those giving grossly inaccurate predictions reveals that men and middle-aged runners are overrepresented, suggesting that female or younger and older runners are better at forecasting their individual performance. The latter findings seem to support the evidence presented in an already large body of experimental literature finding statistically significant gender differences in competitiveness, confidence and risk behavior. However, using data from two different footraces, two recent studies (Nekby, Skogman-Thoursie, & Vahtrik, 2008; Garratt, Weinberger, & Johnson, 2013) question these findings. Nekby et al. (2008) use information on a sub-sample of runners who participated in both the 2005 and 2006 ‘Midnight Run’ (Stockholm), where runners have the opportunity to self-select into either a ‘competitive’ or a ‘recreational’ race. Controlling for the individuals’ performance in 2005, it appears that female runners in 2006 are 6% more likely than male runners to be overconfident and self-select into the competitive race. Using data from a popular road race in California, the ‘State Street Mile’, where male and female participants also have a choice between two different levels of competition, Garratt et al. (2013) find large, systematic age and gender differences in the relationship between true ability (measured via the performance in previous years) and the decision to enter the more competitive race. Overall, qualified women and older runners are less likely than

449

qualified young men to enter a competitive race with prizes. However, the fastest young women unanimously enter the competitive race. Finally, using data from three marathons offering the opportunity to qualify for the Boston Marathon, Burdina, Hiller, and Metz (2017) find that runner performance improves as a goal becomes more attainable, if past performance is properly controlled for. The main finding is that when qualifying standards get easier, because a runner moves up one age group (e.g. from 40–44 to 45–49 years), this has a positive effect on performance. In cases in which the standard became more challenging (due to changes in the qualification rules), a negative relationship with performance is found, suggesting that the setting of a challenging yet achievable goal increases motivation and, as a result, improves performance. The authors conclude that in order to boost performance, goals must be specific, challenging, and realistic. Unrealistic and unreasonable goals will result in failure, which will lead to a drop in motivation and lower overall performance.

DISCUSSION In this chapter, we summarize the available research by economists on professional as well as recreational (long-distance) running. We first provide some information on the historical background and then discuss the current situation by describing the major events, their incentive systems, and their outcomes. Moreover, we document the changing composition of the World’s Marathon elite by gender. We then present selected findings on the economics of recreational running by focusing on (1) training and pacing, (2) chronological age and performance, and (3) reference points, self-selection and (over)confidence. Irrespective of the evidence presented in this chapter, economic research on running is still in its infancy, leaving a couple of questions for future research. First, while there is compelling evidence on the impact of (different forms of) training on the performance of recreational as well professional marathon runners, little to nothing is known about the impact of (different forms of) doping, that is now apparently widespread even among recreational runners. Second, based on the pioneering studies by, for example, Emerson and Hill (2017) and Boyd and Boyd (1995), the strategic behavior of middle- and long-distance-runners trying to maximize their sporting success and/or their prize money needs to be investigated in more detail. Finally, nothing is known so far about

450

THE SAGE HANDBOOK OF SPORTS ECONOMICS

individual runners’ endorsement contracts and their (potential) impact on performance and the likelihood of using illegal performance-enhancing substances.

Notes 1  For a more extensive review of the literature on the incentive effects of sports tournaments see, e.g., Frick and Simmons (2008a, 2008b). 2  The descriptive and econometric findings presented in this and the next section are based on pooled cross-section data including information on the performance of the Top 200 male and female professional distance runners during the period 1973 through 2015. The data was retrieved from the homepage of the Association of Road Running Statisticians (www.arrs.net/). 3  Emerson and Hill (2017) argue that the recent dominance of East African runners is due to the superior race strategies they use, such as, for example, running the second half of the race faster than the first and running faster from 30 km to 35 km than the previous 5 km split. Since the choice of these strategies requires superior physical abilities, they are most likely not available to either Europeans or North Americans. 4  Similar results – all pointing towards a drop in performance around the age of 50–55 – were found by Connick, Beckman, and Tweedy (2015), Lara, Salinero, and Del Coso (2014), Lepers and Cattagni (2012), Leyk et  al. (2009, 2010), and Van Ours (2009).

REFERENCES Allen, E.J., Dechow, P.M., Pope, D.G., & Wu, G. (2017). Reference-dependent preferences: evidence from marathon runners. Management Science, 63, 1657–1672. Athletics (2017). Prize money (in USD) in middle- and long-distance running events, 2017. Retrieved from: http://athletics.com.au/Portals/56/High%20 Performance/Documents/WCH/Prize%20Money. pdf (November 21, 2017. Azmat, G., & Möller, M. (2009). Competition among contests. RAND Journal of Economics, 40, 743–768. Bostonmarathonmediaguide (2016a). Prize purses, 2017, in USD. Retrieved from: www.boston marathonmediaguide.com/prize-structure/ (November 21, 2017). Bostonmarathonmediaguide (2016b). Time bonuses, men 2017, in USD. Retrieved from: www.

bostonmarathonmediaguide.com/prize-structure/ (November 21, 2017). Bostonmarathonmediaguide (2016c). Time bonuses, women 2017, in USD. Retrieved from: www. bostonmarathonmediaguide.com/prize-structure/ (November 21, 2017). Boyd, D.W., & Boyd, L.A. (1995). Strategic behaviour in contests: evidence from the 1992 Barcelona Olympic Games. Applied Economics, 27, 1037–1043. Burdina, M., Hiller R.S., & Metz, N.E. (2017). Goal attainability and performance: evidence from Boston marathon qualifying standards. Journal of Economic Psychology, 58, 77–88. Chicagomarathon (2016a). Prize purses, 2017, in USD. Retrieved from: http://assets.chicagomarathon. com/wp-content/uploads/2012/05/1_MediaInformation-and-Fast-Facts.pdf (November 21, 2017). Chicagomarathon (2016b). Time bonuses, men 2017, in USD. Retrieved from: http://assets. chicagomarathon.com/wp-content/uploads/ 2012/05/1_Media-Information-and-Fast-Facts.pdf (November 21, 2017). Chicagomarathon (2016c). Time bonuses, women 2017, in USD. Retrieved from: http://assets. chicagomarathon.com/wp-content/uploads/ 2012/05/1_Media-Information-and-Fast-Facts.pdf (November 21, 2017). Connick, M.J., Beckman E.M., & Tweedy, S.M. (2015). Relative age affects marathon performance in male and female athletes. Journal of Sports Science and Medicine, 14, 669–674. Emerson, J., & Hill, B. (2017). Elite marathon runners: do East Africans utilize different strategies than the rest of the world? Economics Bulletin, 37(3), 1851–1860. Fair, R.C. (1994). How fast do old men slow down? Review of Economics and Statistics, 76, 103–118. Frick, B. (1998). Lohn und Leistung im professionellen Sport: Das Beispiel Stadt-Marathon. Konjunkturpolitik, 44(2), 114–140. Frick, B. (2011a). Gender differences in competitive orientations: empirical evidence from ultramarathon running. Journal of Sports Economics, 12, 317–340. Frick, B. (2011b). Gender differences in competitiveness: empirical evidence from professional distance running. Labour Economics, 18, 389–398. Frick, B., & Klären, R. (1997). Die Anreizwirkungen leistungsabhängiger Entgelte: Theoretische Überlegungen und empirische Befunde aus dem Bereich des professionellen Sports. Zeitschrift für Betriebswirtschaft, 67(12), 1117–1138. Frick, B., & Prinz, J. (2007). Pay and performance in professional road running: the case of city marathons. International Journal of Sport Finance, 2, 25–35.

THE ECONOMICS OF RUNNING

Frick, B., & Schmitz, H. (2017). Age-related differences in the performance of recreational marathon runners. Mimeo, Management Department, Paderborn University. Frick, B., & Simmons, R. (2008a). Pay and performance of players in sports leagues: inter-national comparisons. In B.R. Humphreys & D.R. Howard (Eds.), The business of sports (pp. 153–180). Westport, CT: Praeger. Frick, B., & Simmons, R. (2008b). The allocation of rewards in athletic contests. In B.R. Humphreys & D.R. Howard (Eds.), The business of sports (pp. 1–25). Westport, CT: Praeger. Garratt, R.J., Weinberger, C., & Johnson, N. (2013). The State Street Mile: age and gender differences in competition aversion in the field. Economic Inquiry, 51, 806–815. IAAF (2017). Prize money (in USD) in middle- and long-distance running events, 2017. Retrieved from: www.iaaf.org/news/press-release/prize-moneyiaaf-world-cross-country-champion (November 21, 2017). Krawczyk, M., & Wilamowski, M. (2017). Are we all overconfident in the long run? Evidence from one million marathon participants. Journals of Behavioral Decision Making, 30, 719–730. Lara, B., Salinero, J.J., & Del Coso, J. (2014). The relationship between age and running time in elite marathoners is U-shaped. Age, 36, 1003–1008. Lazear, E.P., & Rosen, S. (1981). Rank-order tournaments as optimum labor contracts. Journal of Political Economy, 89, 841–864. Lepers, R., & Cattagni, T. (2012). Do older athletes reach limits in their performance during marathon running? Age, 34, 773–781. Leyk, D., Erley, O., Gorges, W., Ridder, D., Rüther, T., Wunderlich, M., Sievert, A., Essfeld, D., Piekarski, C., & Erren, T. (2009). Performance, training and lifestyle parameters of marathon runners aged 20–80 years: results of the PACE-study. International Journal of Sports Medicine, 30, 360–365. Leyk, D., Rüther, T., Wunderlich, M., Sievert, A., Eßfeld, D., Witzki, A., Erley, O., Küchmeister, G., Piekarski, C., & Löllgen, H. (2010). Physical performance in middle age and old age: good news for our sedentary and aging society. Deutsches Ärzteblatt International, 107, 809–816. Lynch, D., & Zax, J.S. (2000). The rewards to running: prize structure and performance in professional road racing. Journal of Sports Economics, 1, 323–340. Maloney, M.T., & McCormick, R.E. (2000). The response of workers to wages in tournaments:

451

evidence from foot races. Journal of Sports Economics, 1(2), 99–123. MarathonGuide.com (2016a). Prize purses, 2017, in USD. Retrieved from: www.marathonguide.com/ news/exclusives/LondonMarathon_2016/LondonMarathonPrizeMoneyAndStarterList.cfm (November 21, 2017). MarathonGuide.com (2016b). Time bonuses, men 2017, in USD. Retrieved from: www.marathonguide. com/news/exclusives/LondonMarathon_2016/ LondonMarathonPrizeMoneyAndStarterList.cfm (November 21, 2017). MarathonGuide.com (2016c). Time bonuses, women 2017, in USD. Retrieved from: www.marathonguide. com/news/exclusives/LondonMarathon_2016/ LondonMarathonPrizeMoneyAndStarterList.cfm (November 21, 2017). MarathonGuide.com (2016d). Prize money (in USD) in middle- and long-distance running events, 2017. Retrieved from: www.diamondleague.com/ rules/ (November 21, 2017). March, D.S., Vanderburgh, P.M., Titlebaum, P.J., & Hoops, M.L. (2011). Age, sex, and finish times as determinants of pacing in the marathon. Journal of Strength and Conditioning Research, 25, 386–391. Nekby, L., Skogman-Thoursie, P., & Vahtrik, L. (2008). Gender and self-selection into a competitive environment: are women more overconfident than men? Economics Letters, 100, 405–407. Rosen, S. (1988). Promotions, elections and other contests. Journal of Institutional and Theoretical Economics, 144, 73–90. Tcsnycmarathon (2016a). Prize purses, 2017, in USD. Retrieved from: www.tcsnycmarathon.org/aboutthe-race/prize-money-and-bonus-awards (November 21, 2017). Tcsnycmarathon (2016b). Time bonuses, men 2017, in USD. Retrieved from: www.tcsnycmarathon.org/ about-the-race/prize-money-and-bonus-awards (November 21, 2017). Tcsnycmarathon (2016c). Time bonuses, women 2017, in USD. Retrieved from: www.tcsnycmarathon.org/ about-the-race/prize-money-and-bonus-awards (November 21, 2017). Van Ours, J.C. (2009). Will you still need me: when I’m 64? De Economist, 157, 441–460. Vickers, A.J., & Vertosick, E.A. (2016). An empirical study of race times in recreational endurance runners. BMC Sports Science, Medicine and Rehabilitation, 8, 8–26.

44 Hitting the Ball Forward: The Economics of Racquet Sports Julio del Corral and Carlos Gomez-Gonzalez

INTRODUCTION As a global-impact sport, tennis has received much attention in the academic literature and, therefore, it is the axis around which this chapter will revolve. However, the existence and importance of other racket sports do not allow us to dismiss the opportunity to discuss their implications and future research opportunities. Table tennis, badminton, squash and padel are important sports in so many countries that still remain in the shadow of tennis in the literature on economics. The impact of the ‘other racket sports’ on the global sport context also differs. While table tennis and badminton have been Olympic sports since 1988 and 1992 respectively, squash and padel are still seeking their positions. Since 1926, the International Table Tennis Federation (ITTF) has governed this sport, which is particularly popular in East Asia and Europe, with the support of 222 member associations in all five continents (ITTF, 2017). The World Table Tennis Championships, the Olympics and the ITTF World Tour are the most important competitions. The importance of this sport is not only based on its professional development, but also on popular participation. In the US, more than 16 million people were reported to have played table tennis at least occasionally in 2016 (SFIA, 2016a). Similarly, the Badminton

World Federation (BWF) has managed the organization of professional badminton since 1934, with five confederation partners located in each of the continents. With a great presence in Asia, this sport is slowly arriving in other parts of the world; for example, in the US, almost 7.2 million people were reported to practice this sport in 2016 (SFIA, 2016b). The recent development of the other two racket sports, squash and padel, determines their currently limited impact. The World Squash Federation (WSF), founded in 1967, is the main governing body and is supported by 149 national associations on all continents. US Squash (2017) reported more than 20 million players worldwide, which is an important number in terms of total participation and acceptance. This support justifies the intense debate about including squash as an Olympic sport in countries with an older tradition and higher participation numbers, such as the United Kingdom. On the other hand, padel has limited representation. The International Padel Federation (FIP), created in 1991, only has 28 associate members and the World Padel Tour (WPT) is the main competition (FIP, 2017). Official participation numbers are difficult to obtain in any country due to its recent development. In any event, the number of associations and players indicates the importance of these racket sports.

Hitting the Ball Forward: The Economics of Racquet Sports

Although the number of associations and players provide evidence of the importance of the ‘other racket sports’, tennis is without doubt the most influential racket sport in the world in terms of participants and media attention. In the US, the Tennis Industry Association (TIA) reported more than 18 million players and an approximate impact of $5.94 billion on the economy (TIA, 2017). With regard to broadcast coverage worldwide, the Association of Tennis Professionals (ATP) events attained close to 1,000 million cumulative viewers, and the Women’s Tennis Association (WTA) events had an audience of 395 million in 2015. This number of followers worldwide justifies the investments of media companies to obtain television rights to tennis competitions; for example, the beIN media group invested $525 million (2017–2026) in WTA rights. Barget (2006) provided a detailed explanation of the structure and organization of tennis in the professional context. To understand the organization of professional tennis, it is necessary to acknowledge some singularities. An important characteristic is the ‘fight’ to manage professional tennis activities between the International Tennis Federation (ITF), which organizes the four Grand Slam tournaments and team competitions (for example, the Davis Cup) and the players’ trade union (ATP), which manages the World Tour tournaments (the ATP World Tour Finals, the ATP World Tour Master 1000, the ATP World Tour 500 series, the ATP World Tour 250 series, and the ATP Challenger Tour) and provides players with supportive services. The WTA also has a strong influence on the management of professional women’s tennis. Thus, the aim of this chapter is to analyze the main academic contributions to the economics of tennis and other racket sports, and to provide recommendations for future research. In recent years, the literature on Sports Economics has allowed further understanding of tennis in the following areas: demand, behavior, betting market, and performance modeling. The next section provides a detailed compendium of the literature, and the chapter concludes with a discussion of potential research lines concerning tennis and other racket sports.

LITERATURE REVIEW Kahn (2000) considered sports to be a natural laboratory in which to study the labor market because of the amount of data available and the observable outcomes. Due to its social impact and development, tennis has offered researchers a tremendous opportunity to obtain information and to analyze organizations, formats, results, players,

453

wages, and awards, for different purposes. Moreover, women’s tennis has more economic and social impact than any other women’s sport, which has allowed a more complete picture of the research. This literature review section is organized as follows: demand, forecast and betting, behavior, and other topics.

Demand Since the path-breaking articles by Rottenberg (1956) and Neale (1964), demand has been one of the main focuses in the literature on Sports Economics. However, this is not true of tennis, for which demand articles are scarce. Bizzozero, Flepp, and Franck (2016) and Konjer, Meier, and Wedeking (2017) are the only papers, to the best of our knowledge, which have analyzed the demand for tennis on television. Specifically, Bizzozero et  al. (2016) analyzed 80 Wimbledon men’s matches that were broadcast on Swiss television in order to test the influence of both surprise and suspense on television audiences. They found that both suspense and surprise were drivers of media entertainment demand. On the other hand, Konjer et  al. (2017) analyzed the demand for live broadcasts on German television during the period 1999–2010. They analyzed 6,939 single matches and found that consumers preferred prestigious competitions, relevant matches, and highly ranked and German players. In Sports Economics, demand and competitive balance are often two interrelated issues. The literature on competitive balance in tennis is also scarce. Two papers from the previous decade took the first steps forward in the literature. Du Bois and Heyndels (2007) were pioneers in this strand of the literature. They proposed a measure of interseasonal competitive balance that considered two aspects: newcomers in the top 10 and the volatility of the remaining players. They found that men’s tennis was competitive. An analysis of long-term uncertainty reinforced this result. Subsequently, del Corral (2009) developed a measure to compute the competitive balance in individual sports in order to analyze the effect of the change in the number of seeded players from 16 to 32 in Grand Slam tournaments in 2001. The results indicated that the outcome of the change in the seeding system differed according to gender. In particular, competitive balance significantly decreased for men. For women, no significant effect on competitive balance was observed as a result of the change. Moreover, the competitive balance measured by the performance of seeded players was much higher for men than it was for women.

454

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Krumer, Rosenboim, and Shapir (2016) adopted a different approach. They analyzed gender differences in competitiveness (tightness) with regard to the final score. They found that men’s matches were tighter than were women’s matches. However, if physical characteristics (i.e., height and weight) are incorporated into the analysis, the gender difference disappears. Hence, these authors suggested that it might be worthwhile for the tennis authorities to consider changing the rules of tennis on a gender basis in order to equalize the physical conditions.

Forecast and Betting Market In tennis, both ATP and WTA produce player rankings that have been found to be a good predictor of player performance in several papers. Boulier and Stekler (1999) found that rankings are useful predictors by using data from Grand Slam tournaments between 1986 and 1995. Subsequently, Clarke and Dyte (2000) used information provided by the rankings to calculate the probability of winning the tournament for each player. In so doing, they compiled information from 1997 ATP tennis, and they calibrated a model in order to determine the probability of winning the set given the players’ ranks. By using this information in Wimbledon 1998, they were able to calculate the probability of winning for each player once the draw had taken place by simulation analysis with 10,000 replications. Moreover, the model was updated on a daily basis. Del Corral and Prieto-Rodríguez (2010) estimated separate probit models for men and women using Grand Slam tennis match data from 2005 to 2008 in order to test whether the differences in rankings between individual players were good predictors of Grand Slam tennis outcomes. They found that the most relevant explanatory variable appeared to be the difference in the ATP or WTA rankings, the effect of which was not statistically different for men and women. They also found that age difference had an important impact on the probability of winning, which decreased monotonically for men, while it is U-shaped for women. McHale and Norton (2011) compared the accuracy of prediction between two kinds of models: on the one hand, logistic regression that used both the classification and points from tennis rankings and, on the other hand, a model from Bradley and Terry (1952), which uses the information from a player’s past result and surface of the contest. They found that the latter model predictions outperformed the predictions of those based on ranking alone. Moreover, they found a strategy of positive returns

in the betting market. Lastly, Ovaska and Sumell (2014) used an extensive database of ATP matches to test several hypotheses and reached the conclusions that top players do their best when it truly matters, tennis is a game of momentum, and being Spanish in clay games is very important, among other findings. Sport betting markets have been shown as a fruitful environment to test the efficient market hypothesis (Sauer, 1998), and to analyze some biases in wagering markets (Forrest & Simmons, 2008; Humphreys, Paul, & Weinbach, 2013). Forrest and McHale (2007) considered that ‘tennis is an especially valuable source of evidence because of the very wide range of odds available on singles bets’ (p. 752). Some papers have addressed the longshot bias, which states that betting houses provide odds in such a way that betting on underdogs has a lower return than does betting on favorites. Forrest and McHale (2007) found evidence of this favorite longshot bias in ATP tennis by using data from 2002 to 2005 from bet365, although similar results were obtained from other bookmakers. Lahvicˇka (2014) extended the work of Forrest and McHale (2007) by incorporating more years into the analysis and data from WTA matches, and also found evidence of a longshot bias in tennis. In addition, the results showed that a favorite-longshot bias was much stronger in matches between lower-ranked players, in laterround matches and in matches in high-profile tournaments, while it is practically nonexistent in the other matches. Both papers collected the data from www.tennis-data.co.uk, which is a very complete data source. To further study this issue, Abinzano, Muga and Santamaría (2016) tested the existence of a longshot bias using data from betting exchanges instead of from bookmakers. They found similar results to those reported by Lahvicˇka (2014), which suggested that bookmakers’ adjustments to respond to informed betting are not the main driver of the favorite longshot bias. Brown (2012), who also used data from the betting exchange market, tried to shed some light on whether ‘informed’ traders have an advantage due to access to private, inside information, or due to a superior ability to process public information. Betting companies provide odds in such a way that the sum of the inverse of the odds is higher than one. This margin of betting companies, best known as the over-round, is a measure of transaction costs for bettors, which can be viewed as the price in the betting market. The lack of published works analyzing the determinants of the over-round is surprising. Lyócsa and Fedorko (2016) aimed to fill this gap. They determined that there is an overall tendency towards a reduction in

Hitting the Ball Forward: The Economics of Racquet Sports

margins, which is in line with the results obtained by Gómez-Roso (2013) and by Feddersen, Che, and Humphreys (2013) regarding European football. Moreover, also in line with Gómez-Roso (2013), Lyócsa and Fedorko (2016) found that the more prestigious the match, measured as the importance of the tournament and the round, the lower the over-round.

Behavior As pointed out by Palacios-Huerta (2014), ‘any type of data about human activity is potentially useful to evaluate economic theories. In fact, sports are in many ways the perfect laboratory for testing economic theories for a number of reasons. There is an abundance of readily available data, the goals of the participants are often uncomplicated (score, win, enforce the rules), and the outcomes are extremely clear. The stakes are typically high, and the subjects are professionals with experience’ (p. 3). Thus, it is not surprising to find a great number of papers using tennis for this purpose. However, the literature that studies behavioral issues in tennis is very heterogeneous. The characteristics of this sport provide the opportunity to test strategic decisions and to further develop the literature on game theory. With regard to the distribution of effort, Sunde (2009) analyzed the last two rounds of professional tournaments to show that higher ex-ante heterogeneity (ranking differences) reduced the effort exertion for favorites, which moderates the positive effect of ability and helps underdogs’ performances. Gilsdorf and Sukhatme (2007) showed that larger prize differentials were an incentive for favorite male tennis players, whose probability of winning the match significantly increased. The same positive effect of prize differentials on winning probabilities for favorites was found in WTA tournaments (Gilsdorf & Sukhatme, 2008; Lallemand, Plasman, & Rycx, 2008). In this line, Sunde (2003) also found a positive effect of prizes on performance (number of games won) and effort (number of games played) on players in the last two rounds of Grand Slams and Masters tournaments. Rotthoff, Zanzalari, and Jasina (2011) studied the probabilities for a ‘true outcome’ in tennis and demonstrated that increasing the number of sets did not significantly help higher-ranked players to win. To understand the timing of the strategic effort of tennis players, Malueg and Yates (2010) used betting data to identify equally skilled opponents and found that the winner of the first set put more effort into the second, which is eventually equalized in the third.

455

Tennis is often considered to be a laboratory in the field of game theory because of the limited number of outcomes between the two contestants. Walker and Wooders (2001) first used tennis serve-and-return play to test the mixed strategy equilibrium, which is consistent except for the random choices to switch serves from right to left. Hsu, Huang, and Tang (2007) extended this dataset using women’s and juniors’ games to further confirm support for the hypotheses of equal probabilities of winning serve directions and serial independence. Paserman (2007) also presented a game-theoretic model in which the options of aggressive play and unforced errors were considered. In sports such as tennis, in order to analyze the determinants of winning probabilities, points are often assumed to be independent and identically distributed. However, Klaassen and Magnus (2001) demonstrated that points in tennis are neither independent nor identically distributed using data from Wimbledon matches. These authors found that winning the previous point had a significant positive influence on the current point, while it is more difficult for the server to win important points than less important points. To define the dependency of results, Jackson and Mosurski (1997) explained that ‘If winning a trial increases the probability of winning the next trial, then that kind of dependency structure is quite properly called psychological momentum’ (p. 27). They demonstrated that winning the first set increased the probability of winning the second, and the same for the third. Jetter and Walker (2015) also recently found evidence of this ‘hot hand’ effect in professional tennis. Moreover, since Morris (1977) introduced the importance of points in tennis, several works have used this concept to understand the behavior of players and the probability of winning at different moments and scorelines, for example, tournament rounds (Dietl & Nesseler, 2017), break points (Knight & O’Donoghue, 2012), the final set (Magnus & Klaassen, 1999a), when serving first (Magnus & Klaassen, 1999b), or at all possible scores (Croucher, 1986; O’Donoghue, 2001). With regard to career success, Geyer (2010) found that higher income (prize money) and performance (ranking, games, and titles won) delayed players’ decisions to quit professional tennis, which is consistent with previous studies on team sports. Thus, career success has been found to be significantly correlated with players’ ability to recognize the important points and to adapt performance optimally (González-Díaz, Gossner, & Rogers, 2012). The recognition of important points also seems crucial for players to use the hawk-eye, i.e., the technological system to help referee’s decisions, due to the limited attempts

456

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and decisive influence on the score. Similarly, Abramitzky, Einav, Kolkowitz, and Mill (2012) quantitatively confirmed that players’ decisions are close to optimal because calls are more likely to happen when the reward is important, and the option value is low, as theory predicts. Moreover, the utilization ratio and success of line-call challenges are similar for men and women, although women are significantly more accurate with their calls at tiebreaks than are men (Anbarci, Lee, & Ulker, 2016). Home advantage, which has been widely studied in team sports competitions (Varela-Quintana, del Corral, & Prieto-Rodríguez, 2015) remains quite overlooked in the literature on tennis. Koning (2011) made an effort to determine the influence of home advantage, which had a positive effect for men (particularly in matches between high-ranked players) and non-effect for women. The type of surface and the crowd’s support were argued to be influential. In addition, Holder and Nevill (1997) found little evidence of home advantage when using data from the 1993 Grand Slam tournaments. The great impact of women’s tennis (which is not the case for the majority of women’s sports) has provided researchers with sufficient data to analyze the behavior of professional female tennis players and to test differences by gender. In this line, several works have contributed to the understanding of prospective theory issues during recent years. Anbarci, Arin, Okten, and Zenker (2017) analyzed the serve speed of women and men in the Dubai 2013 tournament in different scenarios (ahead, behind, and at tie score), and concluded that female players were more risk averse when serving ahead than were men. Paserman (2007) also found that females serve less aggressively and more conservatively when plying important points. This study showed that women made more unforced errors in important situations than did men when analyzing point-by-point data, but this significant difference disappeared when using data at a set level. The differences in performance between men and women under competitive pressure in professional tennis have attracted the attention of several researchers. Jetter and Walker (2015) demonstrated that professional male and female tennis players had similar performances in important matches in Grand Slam events. Similarly, Cohen-Zada, Krumer, Rosenboim, and Shapir (2017) found that women were better under competitive pressure (different measures were used to define pressure) in Grand Slam tournaments. Coate and Robbins (2001), motivated by previous studies on gender comparisons of competitiveness and aggressiveness in the labor market, used professional tennis

to demonstrate that females, despite lower earnings, played as many tournaments as males with similar career longevity during the period 1979– 1994. More recently, Wozniak (2012) showed that previous positive results (short-term) had a positive effect on tournament participation for both female and male tennis players, although this effect lasts slightly longer for males.

Others Topics Two interesting strands of the literature that cannot be embedded in a full subsection are described in this final part of the literature review. Linn-Brit Bakkenbüll analyzed the effect of attractiveness on performance in two different papers (Bakkenbüll & Kiefer, 2015; Bakkenbüll, 2017). In both papers they found evidence that attractiveness and performance have a positive correlation. However, the reasons for such a relationship are yet to be identified. Future research can help to find the determinants of this result. In Economics, an important strand of the literature is related to the efficiency and productivity of Decision-Making Units. In tennis, however, we have only found Ramón, Ruiz, and Sirvent (2012) and Ruiz, Pastor, and Pastor (2013), in which Data Envelopment Analysis is used to complement the information provided by official rankings. In this regard, Baker and McHale (2014) performed a dynamic paired comparison in order to determine the best male tennis player in history. The authors found that Roger Federer appeared to be the most likely candidate.

RESEARCH OPPORTUNITIES In tennis, there are various topics that have not received much attention in the literature on economics. Moreover, given the scarcity of papers on other racket sports, the literature could focus on several economic aspects that remain overlooked. Tennis is one of the most frequently televised and watched sports in many countries. Thus, it is really surprising that we have found only two papers that analyze the television broadcast demand for tennis (Bizzozero et al., 2016; Konjer et  al., 2017). The data from television, such as ratings, could be used to further understand several issues. The prizes in tournaments are a very interesting topic as, some years ago, Grand Slam tournaments started to concede the same prizes to men and women. However, is this fair? Analyzing the demand for tennis on television can help to

Hitting the Ball Forward: The Economics of Racquet Sports

answer this question. There are several papers that have tested the uncertainty of outcome hypothesis using data from television in different sports: football (Alavy, Gaskell, Leach, & Szymanski, 2010), cycling (Van Reeth, 2013), Formula One (Schreyer & Torgler, 2018), American football (Paul & Weinbach, 2007), Australian football (Dang, Booth, Brooks, & Schnytzer, 2015), and college basketball (Grimshaw, Sabin, & Willes, 2013). Bizzozero et al.’s (2016) paper is the only one analyzing this issue in tennis. However, this contribution only uses data from the Wimbledon men’s matches on Swiss television; thus, more research is needed in this area. Also related to the demand, and to the best of our knowledge, there is no paper that analyzes attendance at professional tennis matches. The sold-out matches that major tournaments often report in important games is a factor that may have limited the interest of researchers in this issue. However, this is not the case for most matches in the early rounds of these tournaments. These data can also contribute to the debate on the differences in prizes for men and women. Moreover, as Kemper and Breuer (2016) pointed out, San Francisco Giants introduced dynamic pricing to the sporting industry in 2009. This possibility should be of interest for both researchers and tournament managers. Rodenberg, Sackmann, and Groer (2016) indicated that tennis integrity is a critical governance issue, and that economics should play a central role in both analyzing facts and in advising policymakers. The literature on this issue that focuses specifically on tennis is scarce. Rodenberg and Feustel (2014) found a way to predict fixed matches before the match, whereas Paulden (2016) analyzed how to identify a fixed match once it had been played. Both contributions have opened a pathway to further analyze this problem, which affects professional tennis at different levels. Some important tournaments can be considered ‘franchises’, and their events could be moved to other cities. Hence, it is very important for policymakers and tournament managers to be aware of the economic impact of such tournaments on the host cities. At present, most of this work is performed by private companies, but academic analyses under the peer-review processes can help to make these studies more transparent. Similar studies can also be very important for cities bidding for the Davis Cup, as indicated by Barget and Gouguet (2007). There is an ongoing debate about the Davis Cup. On the one hand, most players are not entirely happy about the current system, particularly with regard to the calendar and prizes for players. On the other hand, the ITF is not particularly reluctant to consider

457

a change in its current form. Thus, economics literature can play a role in analyzing the consequences of different formats. The same applies to other possible rule changes, such as golden points at the deuce, or same number of sets played by females and males, or even a change in the height of the net, as proposed by Krumer et al. (2016). Du Bois and Heyndels (2007) and del Corral (2009) wrote the first papers that analyzed the competitive balance in tennis. However, since these path-breaking papers, to the best of our knowledge, only Krumer et al. (2016) have shown some interest in this topic. It is true that the analysis of the competitive balance in individual sports is not as common in the literature when compared with team sports. However, it seems that this literature could be expanded by either updating the results based on the methodologies of Du Bois and Heyndels (2007) and del Corral (2009), or by proposing new methodologies that are able to address the individual nature of tennis. A completely neglected issue in the economics of tennis is coaches. The literature on coaches in team sports is extensive. Some popular topics are efficiency and the determinants (e.g., del Corral, Maroto & Gallardo, 2017), and the effect of dismissal on performance (e.g., de Dios Tena & Forrest, 2007). Tennis is a sport in which top players often change their coaches; hence, interesting data could be collected to analyze several research questions. The first is obvious; namely, who are the most efficient and less efficient coaches within ATP and WTA. A subsequent research question would aim to analyze the determinants of becoming a good coach, which raises two important questions. One is whether having been a top tennis player plays a positive role, and the other pertains to the differences between men and women. Amelie Mauresmo and Conchita Martínez are two examples of female coaches training top male players. Another research idea would be to investigate the role that the change of coach plays in the performance of tennis players. The inclusion of technology to assist referees’ decisions, i.e., the hawk-eye in professional tennis matches, is a recent development of the game. This is the most influential non-game related factor that has been included in professional tennis in its history, and players needed to incorporate it in the playing strategy. Abramitzky et  al. (2012) demonstrated that its use by players is in line with the predictions of decision theory, and Anbarci et  al. (2017) found a similar use among female and male tennis players, albeit with some differences in success rates. However, more research is needed to study the effect that negative and positive calls have on players’ performances at different levels (game, set, and/or match).

458

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Finally, tennis has offered a fruitful environment for testing game and behavioral theories thus far. The data from tennis are becoming the so-called big data, with trillions of data being generated in each tournament. This increase in the quality of the data could be used to learn more about the behavior of tennis players. With regard to the other racket sports, table tennis offers several research opportunities. Tainsky, Xu, and Yang (2017) analyzed the competitive (im)balance of the table tennis Chinese League. However, other leagues and the dynasty of Chinese players in recent years remain uninvestigated in the literature. In addition, ITTF has made several rule changes, such as the change of the balls, and the scoring system, for which an analysis of its effect on the game should be relevant. Badminton, among other research possibilities, offers the opportunity to test the importance of a rising star in a country in which the popularity indexes for this sport are somewhat low. Since Carolina Marin, the Spanish badminton female player, emerged as an important figure in international badminton competitions, the popularity of badminton has increased greatly in Spain, but to what degree? What is its value? These are interesting questions in order to understand the importance of investing in sports that have lower levels of participation. Although odds on most international badminton, table tennis, and tennis matches are offered by most bookmakers, only data from tennis have been used in the literature. Therefore, analyzing the betting market for table tennis and badminton provides future research opportunities. In summary, many researchers have contributed to the understanding of economic issues using data from racket sports, particularly tennis due to its great impact. Moreover, these efforts have provided evidence that has also helped the organization and evolution of this sport. In the future, data from other racket sports, which are played worldwide yet remain unaddressed in the literature, can also be used for similar purposes.

REFERENCES Abinzano, I., Muga, L., & Santamaria, R. (2016). Game, set and match: The favourite-long shot bias in tennis betting exchanges. Applied Economics Letters, 23(8), 605–608. Abramitzky, R., Einav, L., Kolkowitz, S., & Mill, R. (2012). On the optimality of line call challenges in professional tennis. International Economic Review, 53(3), 939–964.

Alavy, K., Gaskell, A., Leach, S., & Szymanski, S. (2010). On the edge of your seat: Demand for football on television and the uncertainty of outcome hypothesis. International Journal of Sport Finance, 5(2), 75–95. Anbarci, N., Arin, K. P., Okten, C., & Zenker, C. (2017). Is Roger Federer more loss averse than Serena Williams? Applied Economics, 49(35), 3546–3559. Anbarci, N., Lee, J., & Ulker, A. (2016). Win at all costs or lose gracefully in high-stakes competition? Gender differences in professional tennis. Journal of Sports Economics, 17(4), 323–353. Baker, R. D., & McHale, I. G. (2014). A dynamic paired comparisons model: Who is the greatest tennis player? European Journal of Operational Research, 236(2), 677–684. Bakkenbüll, L. B. (2017). The impact of attractiveness on athletic performance of tennis players. International Journal of Social Science Studies, 5(3), 12–20. Bakkenbüll, L. B., & Kiefer, S. (2015). Are attractive female tennis players more successful? An empirical analysis. Kyklos, 68(4), 443–458. Barget, E. (2006). The economics of tennis. In W. Andreff, & S. Szymanski (Eds.), Handbook on the Economics of Sport (pp. 418–431). Cheltenham: Edward Elgar. Barget, E., & Gouguet, J. J. (2007). The total economic value of sporting events theory and practice. Journal of Sports Economics, 8(2), 165–182. Bizzozero, P., Flepp, R., & Franck, E. (2016). The importance of suspense and surprise in entertainment demand: Evidence from Wimbledon. Journal of Economic Behavior & Organization, 130, 47–63. Boulier, B. L., & Stekler, H. O. (1999). Are sports seedings good predictors? An evaluation. International Journal of Forecasting, 15(1), 83–91. Bradley, R. A., & Terry, M. E. (1952). Rank analysis of incomplete block designs I: The method of paired comparisons. Biometrika, 39, 324–345. Brown, A. (2012). Evidence of in-play insider trading on a UK betting exchange. Applied Economics, 44(9), 1169–1175. Du Bois, C., & Heyndels, B. (2007). It’s a different game you go to watch: Competitive balance in men’s and women’s tennis. European Sport Management Quarterly, 7(2), 167–185. Clarke, S. R., & Dyte, D. (2000). Using official ratings to simulate major tennis tournaments. International Transactions in Operational Research, 7(6), 585–594. Coate, D., & Robbins, D. (2001). The tournament careers of top-ranked men and women tennis professionals: Are the gentlemen more committed than the ladies? Journal of Labor Research, 22(1), 185–193.

Hitting the Ball Forward: The Economics of Racquet Sports

Cohen-Zada, D., Krumer, A., Rosenboim, M., & Shapir, O. M. (2017). Choking under pressure and gender. Journal of Economic Psychology, 61, 176–190. del Corral, J. (2009). Competitive balance and match uncertainty in Grand-Slam Tennis: Effects of seeding system, gender, and court surface. Journal of Sports Economics, 10(6), 563–581. del Corral, J., Maroto, A., & Gallardo, A. (2017). Are former professional athletes and native better coaches? Evidence from Spanish basketball. Journal of Sports Economics, 18(7), 698–719. del Corral, J., & Prieto-Rodríguez, J. (2010). Are differences in ranks good predictors for Grand Slam tennis matches? International Journal of Forecasting, 26(3), 551–563. Croucher, J. S. (1986). The conditional probability of winning games of tennis. Research Quarterly for Exercise and Sport, 57(1), 23–26. Dang, T. M., Booth, R., Brooks, R., & Schnytzer, A. (2015). Do TV viewers value uncertainty of outcome? Evidence from the Australian Football League. Economic Record, 91, 523–535. Dietl, H., & Nesseler, C. (2017). Momentum in tennis: Controlling the match. International Journal of Sports Psychology, 48(4), 459–471. de Dios Tena, J., & Forrest, D. (2007). Within-season dismissal of football coaches: Statistical analysis of causes and consequences. European Journal of Operational Research, 181(1), 362–373. Feddersen, A., Che, X., & Humphreys, B. R. (2013). Price setting in fixed odds betting markets. Paper presented at VIII Gijón Conference on Sports Economics. FIP (International Padel Federation) (2017). Federaciones-Federations. Retrieved from: www. padelfip.com/federaciones Forrest, D., & McHale, I. (2007). Anyone for tennis (betting)? The European Journal of Finance, 13(8), 751–768. Forrest, D., & Simmons, R. (2008). Sentiment in the betting market on Spanish football. Applied Economics, 40(1), 119–126. Geyer, H. (2010). Quit behavior of professional tennis players. Journal of Sports Economics, 11(1), 89–99. Gilsdorf, K. F., & Sukhatme, V. A. (2007). Testing Rosen’s sequential elimination tournament model: Incentives and player performance in professional tennis. Journal of Sports Economics, 9(3), 287–303. Gilsdorf, K. F., & Sukhatme, V. A. (2008). Tournament incentives and match outcomes in women’s professional tennis. Applied Economics, 40(18), 2405–2412. Gómez-Roso, J. M. (2013). El uso de cuotas de apuestas deportivas para analizar la competencia en el sector del juego online. Undergraduate thesis. Retrieved from Julio del Corral webpage.

459

González-Díaz, J., Gossner, O., & Rogers, B. W. (2012). Performing best when it matters most: Evidence from professional tennis. Journal of Economic Behavior & Organization, 84(3), 767–781. Grimshaw, S. D., Sabin, R. P., & Willes, K. M. (2013). Analysis of the NCAA men’s Final Four TV audience. Journal of Quantitative Analysis in Sports, 9(2), 115–126. Holder, R. L., & Nevill, A. M. (1997). Modelling performance at international tennis and golf tournaments: Is there a home advantage? Journal of the Royal Statistical Society: Series D (The Statistician), 46(4), 551–559. Hsu, S. H., Huang, C. Y., & Tang, C. T. (2007). Minimax play at Wimbledon: Comment. The American Economic Review, 97(1), 517–523. Humphreys, B. R., Paul, R. J., & Weinbach, A. P. (2013). Bettor biases and the home-underdog bias in the NFL. International Journal of Sport Finance, 8(4), 294–312. ITTF (International Table Tennis Federation) (2017). Member associations. Retrieved from: www.ittf. com/directory/member-associations/ Jackson, D., & Mosurski, K. (1997). Heavy defeats in tennis: Psychological momentum or random effect? Chance, 10(2), 27–34. Jetter, M., & Walker, J. K. (2015). Game, set, and match: Do women and men perform differently in competitive situations? Journal of Economic Behavior & Organization, 119, 96–108. Kahn, L. M. (2000). The sports business as a labor market laboratory. The Journal of Economic Perspectives, 14(3), 75–94. Kemper, C., & Breuer, C. (2016). How efficient is dynamic pricing for sport events? Designing a dynamic pricing model for Bayern Munich. International Journal of Sport Finance, 11(1), 4–25. Klaassen, F. J., & Magnus, J. R. (2001). Are points in tennis independent and identically distributed? Evidence from a dynamic binary panel data model. Journal of the American Statistical Association, 96(454), 500–509. Knight, G., & O’Donoghue, P. (2012). The probability of winning break points in Grand Slam men’s singles tennis. European Journal of Sport Science, 12(6), 462–468. Koning, R. H. (2011). Home advantage in professional tennis. Journal of Sports Sciences, 29(1), 19–27. Konjer, M., Meier, H. E., & Wedeking, K. (2017). Consumer demand for telecasts of tennis matches in Germany. Journal of Sports Economics, 18(4), 351–375. Krumer, A., Rosenboim, M., & Shapir, O. M. (2016). Gender, competitiveness, and physical characteristics: Evidence from professional tennis. Journal of Sports Economics, 17(3), 234–259.

460

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Lahvička, J. (2014). What causes the favouritelongshot bias? Further evidence from tennis. Applied Economics Letters, 21(2), 90–92. Lallemand, T., Plasman, R., & Rycx, F. (2008). Women and competition in elimination tournaments: Evidence from professional tennis data. Journal of Sports Economics, 9(1), 3–19. Lyócsa, Š., & Fedorko, I. (2016). What drives intermediation costs? A case of tennis betting market. Applied Economics, 48(22), 2037–2053. Magnus, J. R., & Klaassen, F. J. (1999a). The final set in a tennis match: Four years at Wimbledon. Journal of Applied Statistics, 26(4), 461–468. Magnus, J. R., & Klaassen, F. J. (1999b). On the advantage of serving first in a tennis set: Four years at Wimbledon. Journal of the Royal Statistical Society: Series D (The Statistician), 48(2), 247–256. Malueg, D. A., & Yates, A. J. (2010). Testing contest theory: Evidence from best-of-three tennis matches. The Review of Economics and Statistics, 92(3), 689–692. McHale, I., & Morton, A. (2011). A Bradley-Terry type model for forecasting tennis match results. International Journal of Forecasting, 27(2), 619–630. Morris, C. (1977). The most important points in tennis. In S. P. Kadany & R. E. Machol (Eds.), Optimal Strategies in Sports (pp. 131–140). Amsterdam: North-Holland. Neale, W. C. (1964). The peculiar economics of professional sports. The Quarterly Journal of Economics, 78(1), 1–14. O’Donoghue, P. G. (2001). The most important points in grand slam singles tennis. Research Quarterly for Exercise and Sport, 72(2), 125–131. Ovaska, T., & Sumell, A. J. (2014). Who has the advantage? An economic exploration of winning in men’s professional tennis. The American Economist, 59(1), 34–51. Palacios-Huerta, I. (2014). Beautiful Game Theory: How Soccer Can Help Economics. Princeton, NJ: Princeton University Press. Paserman, M. D. (2007). Gender differences in performance in competitive environments: Evidence from professional tennis players. IZA Discussion Paper 2834. Bonn, Germany: Institute for the Study of Labor. Paul, R. J., & Weinbach, A. P. (2007). The uncertainty of outcome and scoring effects on Nielsen ratings for Monday Night Football. Journal of Economics and Business, 59(3), 199–211. Paulden, T. (2016). Smashing the racket. Significance, 13(3), 16–21. Ramón, N., Ruiz, J. L., & Sirvent, I. (2012). Common sets of weights as summaries of DEA profiles of weights: With an application to the ranking of professional tennis players. Expert Systems with Applications, 39(5), 4882–4889.

Rodenberg, R. M., & Feustel, E. D. (2014). Forensic sports analytics: Detecting and predicting matchfixing in tennis. Journal of prediction markets, 8(1), 77–95. Rodenberg, R. M., Sackmann, J., & Groer, C. (2016). Tennis integrity: A sports law analytics review. The International Sports Law Journal, 16(1–2), 67–81. Rottenberg, S. (1956). The baseball players’ labor market. Journal of Political Economy, 64(3), 242–258. Rotthoff, K., Zanzalari, D., & Jasina, J. (2011). What are the odds? A measure of the small sample problems. Applied Economics Letters, 18(12), 1139–1143. Ruiz, J. L., Pastor, D., & Pastor, J. T. (2013). Assessing professional tennis players using data envelopment analysis (DEA). Journal of Sports Economics, 14(3), 276–302. Sauer, R. D. (1998). The economics of wagering markets. Journal of Economic Literature, 36(4), 2021–2064. Schreyer, D., & Torgler, B. (2018). On the role of race outcome uncertainty in the TV demand for Formula 1 Grands Prix. Journal of Sports Economics, 19(2), 211–229. SFIA (Sports and Fitness Industry Association) (2016a). Table Tennis Participation Report 2016. Retrieved from: www.sfia.org/reports/452_Table-TennisParticipation-Report-2016 SFIA (2016b). Badminton Participation Report 2016. Retrieved from: www.sfia.org/reports/469_ Badminton-Participation-Report-2016 Sunde, U. (2003). Potential, prizes, and performance: Testing tournament theory with professional tennis data. IZA Discussion Paper No. 947. Bonn, Germany: Institute for the Study of Labor. Sunde, U. (2009). Heterogeneity and performance in tournaments: A test for incentive effects using professional tennis data. Applied Economics, 41(25), 3199–3208. Tainsky, S., Xu, J., & Yang, Q. (2017). Competitive balance and the participation–spectatorship gap in Chinese table tennis. Applied Economics, 49(3), 263–272. TIA (Tennis Industry Association) (2017). Tennis Talking Points. Retrieved from: www.tennisindustry. org/pdfs/Industry_Talking_Points.pdf US Squash (2017). Squash Facts. Retrieved from: www.ussquash.com/squash-facts/ Van Reeth, D. (2013). TV demand for the Tour de France: The importance of stage characteristics versus outcome uncertainty, patriotism, and doping. International Journal of Sport Finance, 8(1), 39–60. Varela-Quintana, C., del Corral, J., & PrietoRodríguez, J. (2015). Where to play first (away or home) in a best-of-two tournament? An analysis from UEFA competitions. In P. Rodriguez,

Hitting the Ball Forward: The Economics of Racquet Sports

S. Késenne & R. Koning (Eds.), The Economics of Competitive Sports (pp. 56–73). Cheltenham: Edward Elgar. Walker, M., & Wooders, J. (2001). Minimax play at Wimbledon. The American Economic Review, 91(5), 1521–1538.

461

Wozniak, D. (2012). Gender differences in a market with relative performance feedback: Professional tennis players. Journal of Economic Behavior and Organization, 83(1), 158–171.

45 The Economics of Road Cycling D a a m Va n R e e t h

INTRODUCTION Professional road cycling is one of the oldest professional sports. The first road cycling races date back to the late 19th century and some of cycling’s most famous events are well over 100 years old. Paris-Roubaix was first organized in 1896, the Tour de France was created in 1903 and the first edition of the Tour of Flanders took place in 1913. Already in 1893, a first World Championships race was organized. Since the sport was practiced professionally by the end of the late 19th century, road cycling required well-established cycling federations right from the start. The International Cyclists’ Association, founded in 1892, was in fact one of the first international sports federations in history. It was replaced in 1900 by the UCI, the International Cycling Union, which is still the governing body of the sport today (Mignot, 2015a, p. 16). Road cycling has steadily grown throughout the 20th century and it gathered a stable following in core European countries like France, Belgium, Italy and Spain. Despite the problems professional road cycling currently faces, such as financial distress, internal conflicts of interest, and doping, the sport continues to expand in the 21st century with increasing media attention across the

world, especially in Anglo-Saxon and Asian markets. Cycling’s biggest events, the ‘Grand Tours’ (the Tour de France, the Giro d’Italia and the Vuelta a España) and the major classics called ‘the monuments of cycling’ (Milan–Sanremo, Tour of Flanders, Paris-Roubaix, Liège–Bastogne–Liège and Tour of Lombardy), attract tens of millions of fans who watch the spectacle from the roadside or in front of their television sets (Van Reeth & Larson, 2015, pp. 1–2). But although road cycling is hugely popular in certain European countries and breaking ground in other parts of the world, from a worldwide point of view it is still a relatively small and commercially underdeveloped sport, especially when compared to football, tennis, Formula One or basketball. Consequently, road cycling is one of the least-appreciated and undervalued potential investments in the global sporting market. This chapter offers an insight into the overall state of sports economists’ understanding of professional road cycling. We first illustrate the particular nature of the sport. With this in mind, we next discuss the current institutional setting and league structure of professional road cycling, and we present some anecdotal data on its economic value. In the penultimate section we explain the main challenges road cycling currently faces. An overview of some further research opportunities concludes.

THE ECONOMICS OF ROAD CYCLING

THE PARTICULAR NATURE OF ROAD CYCLING There are a number of critical differences between cycling and almost all other sports. Road cycling is a team sport that is won by individuals, its competitions are heterogeneous, and the sport is practiced on public roads with teams without a home base. These differences strongly determine the particular business model of the sport and make the use of classic sports economics concepts such as competitive balance and outcome uncertainty less straightforward.

A Team Sport Won by Individuals Professional road cycling shows many characteristics of a team sport. On the organizational level, the participants in cycling competitions are the cycling teams. Individual riders can only compete if selected as members of the team to whom they belong (Freeburn, 2013a, p. 202). Furthermore, the salary a rider gets from the team is his main source of income. This has strong repercussions on his behavior given the fact that most contracts in professional cycling are relatively short term and seldom exceed two years. At the competitive level, it is virtually impossible for a rider to win a race without a strong team support. Team strategies protect the leader, reducing his workload and making sure that an opposing team’s leader must do as much work as possible. Even the best riders thus need the team’s strength and tactics to win a ‘Grand Tour’ or a classics race. Few other sports see athletes completely sacrificing themselves to help another team member win. But in contrast to classical team sports, cycling races are won by individual riders. It is the only Olympic sport where the team contributes to the victory of an individual, with a medal only for the winning athlete (Jutel, 2002, p. 197). Therefore, as a sport, professional road cycling is probably best compared with Formula One or MotoGP racing, where the results of individual pilots are also highly dependent on the technical support of the team’s engineers and the team strategy. However, unlike what is the case in motorized sports, cycling teams are not given much credit for their contribution to the result. In cycling, the individual winners receive all the glory and the prize money, although the latter is usually shared between all team members, including staff. Consequently, professional road cycling is neither a typical individual sport, like golf or tennis, nor a pure team sport, like football or basketball (Benijts, Lagae, & Vanclooster, 2011, p. 606). Professional road cycling could,

463

in fact, be described as the most individual of all team sports. Candelon and Dupuy (2015) are the first to model this particular feature of road cycling in great detail. The authors investigate the relation between hierarchical organization and performance inequality and propose an equilibrium theory for an implicit labor market, organized in hierarchies with one team leader (the buyer) at the top and several helpers (the sellers) below. Using game theory, Mignot (2015b) explains that the team aspect of road cycling offers opportunities for strategic interactions, both between and within teams. Scelles, Mignot, Cabaud and François (2017) investigate empirically the determinants of breakaway success in the Tour de France and illustrate how the likelihood of breakaway success depends on strategic moves such as attack timing and the percentage of riders with a teammate in the breakaway. Larson and Maxcy (2014) show how within teams, the diffusion of two-way radio communication between the team director and his riders has increased the importance of the team directors’ strategic decisions on the dynamics of bicycle finals since the 1990s. The effect of being a team leader on a cyclist’s performance is analyzed in Rodriguez-Gutiérrez (2014). It is concluded that the most decisive feature in enhancing rider efficiency is team status.

A Sport with Heterogeneous Competitions Many sports have homogeneous events. All football games are played on similar sized pitches in games of equal duration and universal rules, and a 100-meter sprinter performs the same act throughout his entire career (Rebeggiani & Tondani, 2008, p. 24). In some sports, however, competitions are to a certain extent heterogeneous. Tennis games are played on grass, gravel or hardcourt and alpine skiing courses or Formula One racing tracks are different at every event. The differences between individual events are even more prominent in road cycling competitions. Not only is there a difference between stage races, which can run from a couple of days up to three weeks, and one-day races, but all competitions also have their own identity and race trajectory. Since climbing a steep mountain pass requires totally different skills than riding on cobblestones, certain top riders (like Chris Froome) will never participate in one-day classics (such as Paris–Roubaix) that do not very well match their strengths. The heterogeneity in races also makes it difficult to determine the best overall rider. The rider who wins the most races (usually a sprinter type of rider) is unlikely

464

THE SAGE HANDBOOK OF SPORTS ECONOMICS

to be the winner of the most important cycling competition (the Tour de France) (Van Reeth & Lagae, 2015, p. 317). In addition, there is also a heterogeneity in prizes. In stage races, multiple prizes are at stake at the same time. Thus, competitors in the same competition may have very different objectives which do not necessarily correspond with trying to win the race. For instance, in the Tour de France, next to the competition for the overall win in the general time classification (yellow jersey), during each stage there are at least four other competitions for secondary prizes: the general points classification (green jersey), the king of the mountain classification (polka dot jersey), the best young rider classification (white jersey) and the team classification. Since there is also a winner in each of the 21 individual stages, there are at least 26 major prizes to be won in a single Tour de France. Performance studies of professional road cycling are relatively rare in the economics of sports literature and have been focused exclusively on cycling’s hallmark event, the Tour de France. The heterogeneous nature of its competitions has indeed made analyses of performance in road cycling less evident (Cabaud, Scelles, François, & Morrow, 2015, p. 25). In fact, both Torgler (2007) and Prinz and Wicker (2012) entirely ignore the heterogeneity in prizes and focus exclusively on the overall time classification. Both studies explain how factors like strength, experience, bodyweight, team composition and nationality determine the overall ranking of a rider. But by neglecting the multi-prize nature of the Tour de France and by using a linear relative ranking measure, which does not discriminate between, on the one hand, the highly relevant difference between the winner and the runner-up and, on the other hand, the totally irrelevant difference between the rider finishing in 110th position and the rider finishing in 111th position, both studies do oversimplify the context. In contrast, Cherchye and Vermeulen (2006) present a multidimensional approach. They develop a new methodology to aggregate information from six different performance indicators (yellow jersey, second and third place in GC, stage wins, green jersey and polka dot jersey) into a single ranking and thus embed the multi-prize context into the analysis. Still, by focusing on individuals they neglect an important real element of strategic positioning that is largely at a team manager’s discretion. Rogge, Van Reeth, and Van Puyenbroeck (2013) are the first to analyze overall team performance rather than individual rider results. They also account for the multiple ways in which a cycling team can be successful by differentiating between ranking teams, sprint teams and mixed teams, based on the team’s major goals

in the Tour de France. Using Data Envelopment Analysis, for each team two efficiency scores are determined: a score that measures the team’s performance relative to similarly classified teams, and a score that is the consequence of the team type. Generally, ranking teams are more efficient than other types of cycling teams.

A Sport Practiced on Public Roads with Teams without a Home Base Most sports are played in stadiums or take place on closed tracks. This allows organizers to charge admission fees, creating a revenue stream that supports the sport. Since road cycling takes place on public roads, this source of revenue is unavailable to race organizers. Consequently, although the Tour de France attracts millions of spectators along the streets every year, this does not lead to any revenues for the organizers (Rebeggiani & Tondani, 2008, p. 24). The road nature of the sport also prevents cycling teams from directly collecting revenue from their fans. The absence of a home base makes it harder to create long-term fan loyalty, especially given the fact that cycling teams regularly change sponsors and thus change names. Instead of following a team, cycling fans rather identify with riders. While a fan of Barcelona will usually dislike a former Barcelona player the moment he is transferred to a rival team, a fan of Peter Sagan will remain a fan no matter what team he is in. In general, most professional cycling teams lack the regional fandom that is typical in many other sports. Consequently, cycling has always wrestled with massive financial challenges. Today, cycling teams still almost exclusively depend upon sponsorship for their financial viability (Van Reeth, 2018, p. 199). The use of public roads also has other important consequences. First, it implies that cycling competitions cause negative externalities to society. Main roads and sometimes entire city centers are blocked for hours to the discomfort of the local inhabitants and causing traffic congestion in the wider surroundings of the race trajectory. Second, the use of public infrastructure also imposes direct costs to society. To warrant safety within the race as well as along the race trajectory, a significant amount of public money must be spent on, for example, municipal staff and policing. Traditionally borne voluntarily by local and regional entities, these costs are now increasingly passed on to the race organizers. Third, safety investments in road infrastructure pose a significant problem to the riders. It is a paradox that the introduction of measures to increase road safety, like speed

THE ECONOMICS OF ROAD CYCLING

humps, roundabouts, or separate bike paths, have adversely increased the risk of crashes within the peloton. Finally, as a sport run on the road, TV coverage of cycling is complex and expensive. It requires a huge array of specialized equipment and personnel, such as helicopters, motorcycles, mobile high-definition cameras, satellite uplink trucks, a TV relay aircraft, and so on. Especially when compared to ‘on site’ sports, the production costs of a cycling race quickly become relatively high. This is one of the reasons why, except for a small number of top races like the Tour de France, TV revenues are virtually non-existent in professional road cycling and race organizers often pay broadcasters to cover their events, rather than the other way around (Van Reeth, 2015a, p. 77).

PROFESSIONAL ROAD CYCLING TODAY Professional road cycling today is discussed from three points of view. We first describe the institutional setting. This is followed by a discussion of the league structure and we conclude with some information on the economic value of the sport.

Institutional Setting Professional road cycling is organized around four main types of agents: the governing bodies, the race organizers, the cycling teams and the riders. As explained above, the UCI acts as cycling’s governing body, and is recognized as such by the International Olympic Committee. It groups together the close to 200 member federations and operates as a traditional regulatory body, licensing races and riders, providing referees and enforcing rules. The UCI not only overlooks male professional road cycling, it is also responsible for junior and women’s road cycling, as well as for seven other cycling disciplines, such as track racing and BMX (Rebeggiani, 2015, p. 39). The UCI has often been described as a weak institution (Long, 2012, p. 361), especially in comparison to powerful organizations such as the FIFA in football or the FIA in Formula One. Freeburn (2012) even claims that there are significant issues of representative democracy and nationality discrimination within the UCI that are so fundamental that they affect the credibility of the UCI and pose real questions as to its legitimacy as the governing body of the sport. In another article that uses UCI’s anti-doping policy as a case study, Freeburn (2013b) concludes that the UCI requires structural reforms and suggests a number of measures that could be adopted to

465

improve the accountability and management of the UCI. The main reason for UCI’s weakness is the missing market power since the most prestigious and commercially most interesting races are controlled by private race organizers. Race organizers have a prominent role in road cycling because of their ability to control access to their events unregulated by ex-ante defined rules (Rebeggiani, 2015, p. 41). Most race organizers are jointly organized in the AIOCC (Association Internationale des Organisateurs de Courses Cyclistes) representing over 100 competitions. The most influential member is without a doubt Tour de France organizer ASO (Amaury Sports Organisation), with Tour de France director Christian Prudhomme also being the president of the AIOCC. The basic organizational unit of professional road cycling is the cycling team. These teams are supported by one or two principal commercial sponsors and are referred to by these sponsor names (Brewer, 2002, p. 281). The interests of the cycling teams are, at least theoretically, protected by the AIGCP (Association Internationale des Groupes Cyclistes Professionnels). But in contrast to the AIOCC, the AIGCP is a poorly organized and internally divided association with many feuds, resulting in teams constantly leaving and rejoining the organization. Therefore, some cycling teams started to create their own interest groups. In 2007 a number of primarily French teams founded the MPCC (Mouvement Pour un Cyclisme Crédible) to defend the idea of clean and healthy cycling (Rebeggiani, 2015, p. 44). In 2014 about a dozen teams created the Velon Group, a joint-venture company that aims for a business model that ensures a sustainable future for the teams (Van Reeth, 2015a, p. 79). The fact that only a couple of teams joined both projects clearly exposes the continuing schism between the cycling teams. The most important organization representing the riders is the CPA (Cyclistes Professionnels Associés). The general aim of the CPA is to defend the riders’ interests before the UCI, the race organizers and the teams (Rebeggiani, 2015, p. 46). Major problems with the CPA, however, are that it only represents the riders from a dozen countries, excluding cyclists from, for example, Australia, Colombia or Russia, and that it is led by former cyclists. There is almost no engagement from current professional riders. Therefore, although the CPA occupies seats in several UCI commissions, overall the riders are still a rather powerless group in cycling, especially when compared to athletes in other sports. Cycling’s labour market has not yet been researched intensively. In fact, the aforementioned study by Candelon and Dupuy (2015)

466

THE SAGE HANDBOOK OF SPORTS ECONOMICS

is the only one that explicitly focuses on the role of cyclists in cycling’s labor market. Some studies deal with aspects that are closely related to cycling’s labor market. Larson and Maxcy (2011, 2013) explore the impact of cycling coaches while Brocard (2015) analyzes rider agents in professional road cycling, a role that has only been regulated by the UCI since 2012.

League Structure The UCI WorldTour is the top league of the pyramidal three-tier competition structure of professional road cycling. In 2019, it included 38 major cycling events, 18 cycling teams and 478 riders (Table 45.1). Below the WorldTour level are the ProContinental level (25 teams, 481 professional riders) and the Continental level (175 teams, over 2,000 semi-professional or amateur riders). There are 314 races at the Continental level, divided over five Continental circuits: the America Tour (20 races), the Asia Tour (28 races), the Africa Tour (14 races), the Oceania Tour (3 races) and the Europe Tour (249 races). The barriers between the different levels are not strict, though. ProContinental teams can be invited to participate in WorldTour races, and WorldTour teams as well as Continental teams participate in Continental races. One of the main goals of the UCI WorldTour is to globalize professional road cycling. Yet, Table 45.1 illustrates that professional road cycling at the top level is still very much a European sport. Although the expansion of the WordTour from 28 races to 38 in 2017 increased the part of the non-European calendar from 10% to 20%, close to 80% of the riders and the races in the WorldTour are still European. The observation that seven out of 18 WorldTour teams are non-European hides the fact that many of these teams still have a majority of European riders and are often guided by European staff. The current league structure, the UCI ProTour as it was initially called, was introduced in 2005 and has been widely debated in academic

literature since then. At the time it was promoted as the Champions League of cycling. Freeburn (2007) was the first to raise reasonable concern about the reform. He stated that the ProTour could be a ‘train wreck waiting to happen’ if all of the parties involved do not show the necessary amount of solidarity. Morrow and Idle (2008) examine the 2005 reform as a case study, using stakeholder theory and network theory, with a focus on power relationships between different stakeholders. A similar analysis, with a specific focus on the implications of the reform on the position of the race organizers, is carried out by Desbordes (2008). Rebeggiani and Tondani (2008) use an oligopolistic model to demonstrate why the new competition fails to reach one of its major goals, namely to increase overall competition. They claim the pattern of non-competitive behavior displayed by cycling teams results from a poorly designed licensing assignment procedure and recommend the introduction of a promotion and relegation system. Deubert (2010) states that road cycling’s business model could benefit from a more American approach. Because too much power is in the hands of the race organizers, he claims a lucrative partnership between cycling teams and cyclists through a collective bargaining process would make cycling a more stable sport. Based on a market analysis, a rather positive attitude of team sponsors towards the 2005 reform was reported by Benijts, Lagae, and Vanclooster (2011). The authors find that the introduction of the UCI WorldTour had a positive influence on the financial value of the teams’ business-to-business market but did not result in a change in the business demographics of corporate sponsors. In a later study, Benijts and Lagae (2012) nevertheless conclude that overall the reform was unsuccessful, mainly because of problems related to conflicts between cycling teams and race organizers, and the continuing Eurocentric nature of the sport. Rebeggiani (2015) concludes the failure of the WorldTour can be attributed to the UCI paying too little attention to the preferences of the race organizers and not working hard enough to emphasize the importance of cooperation for creating a better product beneficial to all stakeholders.­

Table 45.1  The competition structure of professional road cycling in 2019 WorldTour level

Teams Riders Races

Total

Europe

Rest of the world

18 478 38

11 (61%) 365 (76%) 30 (79%)

7 (39%) 113 (24%) 8 (21%)

Sources: www.cqranking.com and www.uci.ch

ProContinental level

25 481

Continental level

175 over 2000 +/−350

467

THE ECONOMICS OF ROAD CYCLING

The overall consensus in the literature on the failure of the 2005 reform is confirmed by Van Reeth and Lagae (2015). They conclude that the WorldTour is a poorly managed and badly promoted product, illustrated by the fact that in 10 years of existence, a sponsor for the competition was never found.

Economic Value Not much information on the finances of cycling teams, cycling races or rider contracts is publicly available. It is therefore hard to determine the economic value of professional road cycling. Nevertheless, in Table 45.2 some key financial data about professional road cycling are presented for the 1990–2018 period. The data on the team budgets and the broadcasting rights are estimates based on a variety of sources; the information on the prize money is obtained from the official Tour de France website. Team budgets kept increasing at a steady rate in the past quarter of a century. In nominal terms, the average budgets of the best 10 cycling teams more than quadrupled from about €4 million in 1990 to about €17.5 million in 2018. This represents an annual increase of over 5%, which is significantly higher than the long-term inflation rate of 2–3% in western countries in recent decades (Van Reeth, 2015a, p. 59). The budget increase since 2010 is especially worrying since it was mainly the result of state-funded conglomerates, wealthy individuals and oligarch business owners entering professional road cycling. In 2019 only nine out of 18 WorldTour teams are still financed exclusively by classic commercial sponsors: Ag2R La Mondiale, Bora-Hansgrohe, CCC, Movistar, Dimension Data, EF Education First, Jumbo-Visma, Sky and Sunweb. The bulk of the team sponsors of the other teams are either bike manufacturers (Merida, Scott, Trek), national lotteries (FDJ, Lotto), state-funded conglomerates (Astana, Bahrain, Katusha, United Arab Emirates) or powerful benefactors (Zdenek Bakala – Quick Step Floors, Gerry Ryan − Orica-Scott, Igor Makarov – Katusha). The result is a rat race

between the deep-pocketed sponsors promising ever-increasing wages to a select group of top riders (Lagae & Van Reeth, 2015). French television began to pay for the right to cover the Tour de France in the 1960s. The real explosion in TV broadcasting rights started from the 1980s on. In the early 1980s, Tour de France TV broadcasting rights were valued at €250,000 only (Mignot, 2014, p. 28). Thirty-eight years later, they are estimated at about €65 million, of which €26 million is paid by France Télévisions. This remarkable growth is, of course, not that much different from what happened to the value of the broadcasting rights in many other sports in the recent past. Yet, the situation for the Tour de France is exceptional in professional road cycling. Although, as in other sports, television is how 99% of fans follow the competitions, in cycling, outside a handful of prominent events, TV revenues are virtually non-existent. Table 45.2 also illustrates how in cycling the commercial benefits generated by a race like the Tour de France are hardly redistributed to the main actors in the sport, i.e. the riders or the cycling teams. While broadcasting rights (in nominal terms) more or less increased by a factor of 25 between 1990 and 2018, total Tour de France prize money, which also includes the participation fees to the teams, only increased by a factor of 2.5 from just over €1.5 million in 1990 to €4 million in 2018. Since total Tour de France revenue (broadcasting rights, sponsorship, subventions, and fees by hosting cities) is estimated to be over €150 million (Van Reeth, 2015a, p. 77), only 2–3% of the revenue generated by the event is currently redistributed to riders and teams. Moreover, the prize awarded to the overall Tour de France winner only increased from €305,000 to €500,000 over a 28-year period, which does not even compensate for inflation. In fact, the winning prize was set at €450,000 in 2006 and remained unchanged for almost 10 years until 2016, when it was raised to its current level of €500,000. From a tournament theory point of view, it is also remarkable that the share of the winner in the total amount of prize money decreased over the years from 20% in 1990 to only 12.5% in 2016.

Table 45.2  Some key financial data for professional road cycling (1990–2018, in nominal euro)

Average budget top 10 teams Broadcasting rights Tour de France Prize money Tour de France winner Total Tour de France prize money

1990

2000

2010

2018

4,000,000 2,500,000 305,000 1,535,000

7,000,000 18,000,000 335,000 2,363,000

10,000,000 50,000,000 450,000 3,377,000

17,500,000 65,000,000 500,000 4,000,000

Sources: www.letour.fr, Van Reeth (2015a) and Mignot (2014)

468

THE SAGE HANDBOOK OF SPORTS ECONOMICS

THE CHALLENGES FOR PROFESSIONAL ROAD CYCLING Professional road cycling faces numerous challenges that threaten the long-term viability of the sport. Below, we discuss four of the most significant threats: the fragile business model, the dominance of ASO, the decreasing TV audiences, and the doping image.

The Fragile Business Model The Tour de France is not only the greatest cycling event, it is probably also the biggest free professional sporting event in the world (Wille, 2003, p. 128). From a fan’s perspective, professional road cycling is indeed among the cheapest spectator sports. Apart from travel costs, watching a race from the roadside is completely free, and in many European countries most major cycling races are broadcast on subscription-free TV channels (Van Reeth & Lagae, 2015, p. 338). But there is no such thing as a free lunch. The free nature of the sport comes at a cost, namely the high dependence of teams and races on sponsor revenue. Since cycling teams are funded for over 90% by sponsorship money (Van Reeth, 2015a, p. 71), and few companies are willing to commit themselves for more than a couple of years, the future of a cycling team is always insecure. But also race organizers face insecurity. In most sports, television revenues are a major source of revenue but in cycling, as explained before, they are limited. The fact that there is no gate revenue either makes that race organizers also mainly depend on sponsorship money to cover the organizational costs of the races. One of the main challenges for professional road cycling is how to adjust the sponsorship-driven business model, i.e. how to create better financial stability through a diversification of revenue.

of, among others, motor sports, athletics, equestrian sports, golf events and cycling. Next to the Tour, ASO also owns other well-known cycling races in France and Belgium, such as Paris–Nice, Paris– Roubaix, Paris–Tours, la Flèche Wallonne and Liège–Bastogne–Liège, it holds a majority stake in the Vuelta a España, and it is involved in the organization of many more smaller cycling races around the world. The company is without a doubt road cycling’s most powerful actor and its dominant position is a mixed blessing for professional road cycling (Van Reeth & Lagae, 2018, p. 197). The commercial success of the Tour de France allows ASO to invest in less profitable races, to support new races logistically, and to award a significantly higher amount of prize money than other races. By doing so, ASO indirectly subsidizes professional cycling. But its dominant position is also a threat to the further development of the sport. As a private company, above all it defends the commercial value of the Tour de France and it uses its market power to block any reform plan that is likely to threaten its unique position. Consequently, while theoretically the UCI is responsible for managing world cycling, in practice it is ASO that determines the future of the sport.

The Decreasing TV Audiences TV audiences for road cycling are much smaller than most of the stakeholders in cycling like to believe. For example, while ASO claims a multibillion TV audience for the Tour de France, the real per-stage average audience is about 20 to 25 million only (Van Reeth, 2019, p. 810). Furthermore, cycling is confronted with an ageing fan base. The average Tour de France TV viewer in the Netherlands is 57 years old and more than half of the French Tour de France TV audience is over 60 (Van Reeth, 2015b, p. 121). Although cycling fans in Anglo-Saxon countries are, on average, younger than their Continental counterparts, professional road cycling seems to lose touch with the younger generation, largely because of the lengthy broadcasts in which regularly little of interest happens until the last couple of minutes.

The Dominance of ASO Cycling’s most prestigious event, the Tour de France, is absorbing most of the media attention for professional road cycling and in many countries it is the only cycling race that is widely covered. The commercial and sportive importance of the race, with a market share of two-thirds of all television revenue and sponsorship income generated by other WorldTour organizers, threatens to virtually reduce the cycling season to just a single event. The Tour de France is organized by ASO, a for-profit company that specializes in the organization and promotion

The Doping Image Professional road cycling has a long association with doping. Vandeweghe (2015) presents an extensive overview on how doping has been with the sport from its inception in the late 19th century, and how practices changed throughout the 20th and 21st century. Brewer (2002) explains how the commercial development of cycling resulted in extra pressure on team managers and riders to resort to doping practices. Although the introduction of,

THE ECONOMICS OF ROAD CYCLING

respectively, the whereabouts system, a blood passport and a no-needle policy has improved the situation significantly in the past decade, the sport continues to find it very difficult to deal with this historic legacy. There is some evidence on the public’s reaction to doping in cycling. Van Reeth and Lagae (2014) show that in Flanders most spectators do prefer a clean performance over exceptional results due to doping. Yet, this belief does not really affect whether they watch cycling on TV since newly revealed doping cases during the Tour de France appear to have practically no lasting impact on average Flemish TV audiences (Van Reeth, 2013, p. 57). Still, doping remains a threat to the further long-term development of road cycling because sponsors and broadcasters are much more sensitive to new doping cases than the general public (Lagae & Van Reeth, 2015).

469

Tour of Flanders is €10.3 million and arrive at a total economic value for the Tour of Flanders of about €25 million (Vekeman et al., 2015, p. 286). Women’s competitions are largely neglected in academic research on sports, and road cycling is no different in this respect. Since, to our knowledge, at the moment there exists no published academic research on the economics of women’s cycling, a third research opportunity lies in closing this gender gap in the road cycling literature. A better understanding of competitive balance and outcome uncertainty, of the monetization of fan loyalty, and of the development of women’s cycling, can support decisions on the vital changes that the sport desperately needs.

REFERENCES FURTHER RESEARCH OPPORTUNITIES Academic research on professional road cycling has developed substantially in the last decade. Still, there remain numerous interesting research opportunities. Below we identify three of the more significant ones. First, outcome uncertainty and competitive balance, key concepts in the sports economics literature, have received little attention so far. Van Reeth (2013), Rodriguez, Peréz, Puente, and Rodriguez (2015), Andreff (2015), and Rodriguez-Gutierrez and Fernandez-Blanco (2016) develop indices for outcome uncertainty, but only to be used as variables in an empirical analysis. The only academic publication specifically dedicated to competitive balance and outcome uncertainty in professional road cycling is a book chapter (Cabaud et  al., 2015) in which the authors present a ‘competitive intensity’ model that continuously measures outcome uncertainty as the race develops. Because of the sponsorship-driven business model of road cycling, research on improving cycling’s finances is a second relevant research opportunity. There has been some unpublished and mostly non-academic research on the economic impact of staging the World Championships or organizing the start of the Tour de France or the Giro d’Italia, but this is revenue that does not directly benefit cycling. Furthermore, with a few exceptions, the results of most of these studies are questionable. In fact, Vekeman, Meulders, Praet, Colpaert, and Van Puyenbroeck (2015) is the only academic paper so far to specifically deal with willingness-to-pay for cycling races. The authors estimate the average willingness-to-pay for the

Andreff, W. (2015). The Tour de France: a success story in spite of competitive imbalance and doping. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 233–355). London: Springer. Benijts, T., & Lagae, W. (2012). Using program theory to evaluate league reforms: the case of professional road cycling. European Sport Management Quarterly, 12(1), 83–109. Benijts, T., Lagae, W., & Vanclooster, B. (2011). The influence of sport leagues on business-to-business marketing of teams: the case of professional road cycling. Journal of Business & Industrial Marketing, 26(8), 602–613. Brewer, B. (2002). Commercialization in professional cycling 1950–2001: institutional transformations and the rationalization of doping. Sociology of Sport Journal, 19(3), 276–301. Brocard, J.F. (2015). Agents in professional road cycling. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 147– 163). London: Springer. Cabaud, B., Scelles, N., François, A., & Morrow, S. (2015). Modeling performances and competitive balance in professional road cycling. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 257–283). London: Springer. Candelon, B., & Dupuy, A. (2015). Hierarchical organization and performance inequality: evidence from professional cycling. International Economic Review, 56(4), 1207–1235. Cherchye, L., & Vermeulen, F. (2006). Robust rankings of multidimensional performances: an application to Tour de France racing cyclists. Journal of Sports Economics, 7(4), 359–373. Desbordes, M. (2008). The future of the Tour de France: from an independent style of organization to ‘a

470

THE SAGE HANDBOOK OF SPORTS ECONOMICS

formula one model’. In D. Arthur & S. Chadwick (Eds.), International cases in the business of sport (pp. 358–371). Oxford: Butterworth-Heinemann. Deubert, C. (2010). Putting shoulder pads on Schleck: how the business of professional cycling could be improved through a more American structure. Brooklyn Journal of International Law, 37(1), 65–115. Freeburn, L. (2007). The UCI ProTour: an enduring reform or a ‘train wreck’ waiting to happen? Australian and New Zealand Sports Law Journal, 2(1), 15–58. Freeburn, L. (2012). The unrepresentative and discriminatory governance structure of cycling. The International Sports Law Journal, 12(1–2), 27–36. Freeburn, L. (2013a). A breakaway league in professional cycling: issues for the governance and organisation of the sport. The International Sports Law Journal, 13(3–4), 193–210. Freeburn, L. (2013b). The Union Cycliste Internationale: a study in the failure of organisational governance of an international federation. The International Sports Law Journal, 13(1–2), 71–81. Jutel, A. (2002). Olympic road cycling and national identity: where is Germany? Journal of Sport and Social Issues, 26(2), 195–208. Lagae, W., & Van Reeth, D. (2015). Paradoxes in professional road cycling: a plea for a new cycling industry. In Y. Vanden Auweele, E. Cook, & J. Parry (Eds.), Ethics and governance in sport: the future of sport imagined (pp. 121–128). Abingdon, UK: Routledge. Larson, D., & Maxcy, J. (2011). Why the master? Human capital development for practicing U.S. cycling coaches. International Journal of Sport Finance, 6(3), 265–280. Larson, D., & Maxcy, J. (2013). The industrial organization of sport coaches: road cycling as a distinguished case. Journal of Sport Management, 27(4), 363–378. Larson, D., & Maxcy, J. (2014). Uncertainty of outcome and radio policy in professional road cycling. Journal of Sport Management, 28(3), 311–323. Long, J.G. (2012). Tour de France: a taxpayer bargain among mega sporting events? In W. Maennig & A. Zimbalist (Eds.), International handbook on the economics of mega sporting events (pp. 357–385). Cheltenham, UK: Edward Elgar. Mignot, J.F. (2014). Histoire du Tour de France. Paris: La Découverte. Mignot, J.F. (2015a). The history of professional road cycling. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 7–31). London: Springer. Mignot, J.F. (2015b). Strategic behavior in road cycling competitions. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 207–232). London: Springer.

Morrow, S., & Idle, C. (2008). Understanding change in professional road cycling. European Sport Management Quarterly, 8(4), 315–335. Prinz, J., & Wicker, P. (2012). Team and individual performance in the Tour de France. Team Performance Management, 18(7–8), 418–432. Rebeggiani, L. (2015). The organizational structure of professional road cycling. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 33–54). London: Springer. Rebeggiani, L., & Tondani, D. (2008). Organizational forms in professional road cycling – efficiency issues of the UCI ProTour. International Journal of Sport Finance, 3(1), 19–41. Rodriguez, C., Peréz, L., Puente, V., & Rodriguez, P. (2015). The determinants of television audience for professional cycling: the case of Spain. Journal of Sports Economics, 16(1), 26–58. Rodriguez-Gutiérrez, C. (2014). Leadership and efficiency in professional cycling. International Journal of Sport Finance, 9(4), 305–330. Rodriguez-Gutiérrez, C., & Fernandez-Blanco, V. (2016). Continuous TV demand in road cycling: the 2015 Vuelta a España. European Sport Management Quarterly, 17(3), 1–21. Rogge, N., Van Reeth, D., & Van Puyenbroeck, T. (2013). Performance evaluation of Tour de France cycling teams using data envelopment analysis. International Journal of Sport Finance, 8(3), 236–257. Scelles, N., Mignot, J.F., Cabaud, B., & François, A. (2017). Temporary organizational forms and cooperation in cycling. What makes a breakaway successful in the Tour de France? Team Performance Management: An International Journal, 24(3/4), 122–134. Torgler, B. (2007). La Grande Boucle: determinants of success at the Tour de France. Journal of Sports Economics, 8(3), 317–331. Vandeweghe, H. (2015). Doping in cycling: past and present. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 285– 311). London: Springer. Van Reeth, D. (2013). TV demand for the Tour de France: the importance of stage characteristics versus outcome uncertainty, patriotism and doping. International Journal of Sport Finance, 8(1), 39–60. Van Reeth, D. (2015a). The finances of professional cycling teams. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 55–82). London: Springer. Van Reeth, D. (2015b). TV viewing of road cycling races. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 100– 128). London: Springer. Van Reeth, D. (2019). Forecasting Tour de France TV audiences: A multi-country analysis. International Journal of Forecasting, 35, 810–821.

THE ECONOMICS OF ROAD CYCLING

Van Reeth, D., & Lagae, W. (2014). Public opinion on doping in cycling: differences among population groups. In O. Budzinski & A. Feddersen (Eds.), Contemporary research in sports economics (Proceedings of the 5th ESEA Conference) (pp. 247–268). Frankfurt am Main: Peter Lang. Van Reeth, D., & Lagae, W. (2015). The future of professional road cycling. In D. Van Reeth & D. Larson (Eds.), The economics of professional road cycling (pp. 313–341). London: Springer. Van Reeth, D., & Lagae, W. (2018). A blueprint for the future of professional cycling. Sport, Business

471

and Management: An International Journal, 8(2), 195–210. Van Reeth, D., & Larson, D. (Eds.) (2015). The economics of professional road cycling. London: Springer. Vekeman, A., Meulders, M., Praet, A., Colpaert, J., & Van Puyenbroeck, T. (2015). Contingent valuation of a classic cycling race. Journal of Sports Economics, 16(3), 268–294. Wille, F. (2003). The Tour de France as an agent of change in media production. The International Journal of the History of Sport, 20(2), 128–146.

46 The Economics of Golf Stephen Shmanske

INTRODUCTION As a concentrated subject area in the economics of sport, golf was a latecomer to the field. Nevertheless, many of the same themes and economic questions that were first considered in the context of team sports have exact or very similar counterparts in the sport of golf. Direct counterparts exist for the area of production functions or earnings functions in which the skills that are required for sporting success, for salaries, or for earnings are examined. The much looked at issue of sporting competition versus economic competition in league sports is addressed by a cartel of team owners who have a type of exemption from antitrust prosecution. This issue has a parallel in golf in terms of professional golfer’s associations, which govern the officially sanctioned tournaments worldwide.1 There are also parallels between golf and team sports in terms of differing strategies that may become appropriate at different situations in a competition. Going for it on fourth down presents a risk/reward situation that has similarities to attempting a risky golf shot over or near a hazard to a tight pin placement. The question of the economic impact of staging a championship or other mega-event also has direct counterparts in both the traditional league sports and golf.

There are, however, important areas of difference between golf and league sports that support a different variety of research questions. Consider discrimination. Issues of salary discrimination in league sports are absent for golf, in which the contestants are individuals playing for shares of a prize fund rather than employees working for a salary. However, issues of gender discrimination across competitions have been examined. The nature of golfing competitions lends itself very naturally to the consideration of the degree to which the participants supply effort over the course of a tournament. The effort decision also comes up in the context of going up against a superstar in a competition with a tournament compensation scheme for the prize fund. Golf also presents an entry or participation dimension, in which golfers pick and choose which events to enter, which is relatively absent in the usual league sport setting. Additionally, and in homage to Yogi Berra, who reportedly quipped that 90% of this game (baseball) is half mental, studies of the mental aspect of many sports have been rare but mental decision making and performance in golf has been analyzed in the context of propositions stemming from behavioral economics. Perhaps the biggest commonalities in all of sports economics stem from the abundant, highly reliable, and detailed data that exist over

The Economics of Golf

long periods of time with respect to each participant’s skills and earnings. For the most part, the data are publicly available and easy to work with. Additionally, and not to be forgotten, for your typical sports enthusiast the data are interesting in and of themselves. For me personally, on a rainy day or long cold winter when playing golf was out of the question, working with, or should I say playing with, the golf statistics was the next best thing. The remainder of this chapter is organized by sub-topics within the area of golf. The first section tells the story of the development of the golf production or earnings functions from early multiple regression models through multiple equation models with more advanced econometric techniques and better and better data. While refinements can still be made to these functions and the data, economic hypotheses can be illustrated or tested with whatever production functions exist at the time. Some of these clever applications of the production or earnings functions are summarized in the following section. The next section looks at gender and golf, and then I consider the supply decision by golfers of which tournaments to enter. A different type of supply, namely the supply of effort, is considered next. The new data and golf’s application of behavioral economics is then examined. Miscellaneous topics including those motivated by economic examination of the golf course industry are introduced in the final section.

THE GOLF PRODUCTION OR EARNINGS FUNCTION The topic receiving the most attention is the relationship between a multidimensional set of golfing skills and the sporting performances or earnings that the skills create. The earliest such paper in the economics tradition (Shmanske, 1992) performed multiple regression analysis on the top 60 1986 Professional Golfers’ Association (PGA) TOUR golfers with measures of the level of earnings as the dependent variable and a number of right-hand side variables to capture the skills.2 In 1992 there were five skills that were directly measured by the PGA TOUR: driving distance, driving accuracy, approach shot accuracy, putting skill, and the ability to recover from sand-filled hazards surrounding the green. With yearly earnings as the dependent variable, the coefficient estimates in such an equation are interpretable as values of marginal products (VMPs) of the individual skills. For example, in this paper, an additional yard of driving distance is worth $6,775 per year; an additional percentage point in accuracy in hitting greens in regulation is

473

worth $16,037; an increase in putting skill that would save one stroke per tournament is worth $62,500 per year, and so forth. These point estimates are both statistically and economically significant. Other control variables captured golfer experience and a measure of short game ability that was calculated from other raw data.3 A number of other researchers carried out similar analyses using different years and sources for the data, different functional forms, and different lists of control variables.4 Sommers (1994) used the full set of 183 PGA TOUR golfers for 1992 while correcting for heteroskedasticity by using the logarithmic transformation of the dependent variable, earnings per tournament. Not to be outdone, Moy and Liaw (1998) develop separate estimates of the VMP’s for three separate tours: the PGA TOUR, the Champions Tour, and the Ladies Professional Golf Association (LPGA) Tour. Extending the data crunching even further, Alexander and Kern (2005) use 10 years (1992– 2001) of PGA TOUR data in an unbalanced panel to track changes in the VMPs over time. An interesting result from this paper suggests that the game of golf may not have changed much over this period. The results roughly show a tripling of VMPs over the period, tracking an approximate tripling of purse sizes during the 10 years. Most of the attempts to examine the effect of skills on earnings have concentrated on the PGA TOUR, which has the most developed data. For example, in measuring the skill of putting, both the PGA TOUR and the LPGA tracked the total number of putts per round. This does not control for the length of the putts and, therefore, measures the skill with error. Early on, the PGA TOUR also tracked the number of putts taken only on greens reached in regulation, which takes out the effect of many short tap-in putts which are left after short chip shots. Having a lot of these short putts on greens that are not reached in regulation lowers the total number of putts but has less to do with the putting skill than with the short game skill (and perhaps also a degree of inaccuracy with approach shots). More recently, the PGA TOUR keeps track of the total length of made putts but there may also be problems with this measure. For example, making three 20-foot putts for birdies is better (score wise and probably shows more skill than making one (lucky?) 58-foot putt for birdie and two onefooters for pars, but both will count as 60 feet of made putts on the three holes. Even more recently, the PGA TOUR has developed a method of comparing the success rate on similar putts that will be discussed later in this chapter. Notwithstanding the data difficulties, some researchers have focused on the LPGA Tour. Shmanske (2000, 2004) examined the putting

474

THE SAGE HANDBOOK OF SPORTS ECONOMICS

measurement difficulty and showed that the PGA TOUR method was better but that the LPGA statistics were still okay. Fried, Lambrinos, and Tyner (2004) and Pfitzner and Rishel (2005) examined the LPGA earnings function for the years 1998 and 2004 respectively. Most of these papers use ordinary least squares (OLS) and the results differ only slightly. Almost universally the signs of the coefficients of the skills variables are as theoretically proposed. The statistical significance of the measures varies but a consistent story emerges. Putting skill and approach shot accuracy are almost universally important, both statistically and quantitatively. Sometimes driving distance is significant and sometimes driving accuracy seems more important. Occasionally the sand save percentage is statistically significant but rarely is its VMP large enough to matter too much. The overall percentages of explained variation run from 30% to over 80% but, of course, these are not comparable across studies using different data sets or different dependent variables. Techniques besides multiple regression have also been employed. Fried, Lambrinos, and Tyner (2004) first developed the use of Data Envelopment Analysis and a frontier production function method to relate golfer earnings to their skills. They show how two golfers with almost identical statistical skill levels (measured as yearlong averages) can end up with a huge disparity in earnings over the year. They attribute the disparity to the ability or lack thereof to perform up to the golfer’s skill levels in pressure situations. This ability to perform under pressure is generally unmeasured and uncaptured in the literature, and arguably is quantitatively important. Subsequently, Fried and Tauer (2011, 2012) used the technique to examine the age/productivity profiles of PGA TOUR and LPGA golfers respectively. Another variation in statistical procedure was undertaken by Kahane (2010), who uses quantile regression on PGA TOUR data from 2004 to 2007. As an example of the results obtained in this procedure, consider the VMP of driving distance. Whereas an OLS regression indicates a VMP of $926 per event per yard of distance, the quantile regression reports a VMP of only $407 per yard at the median and a larger $1,602 per yard at the 90th percentile. There is still room for growth using these techniques, especially in conjunction with the data improvements introduced below and towards the end of this chapter. Chronologically, the next development in this literature was to recognize that regressing earnings on skills was essentially a reduced-form examination of a multiple step process. Scully (2002) was the first to argue that the skills, rather than producing earnings directly, actually produced scores in

competitions and that this physical performance produced earnings in a second step. Callan and Thomas (2007) took up and extended this suggestion with 2002 data by estimating a three-equation system in which skills produced scores, scores produced tournament ranks, and tournament ranks produced earnings. With this method, the VMPs are calculated from the coefficients of three different equations and can be compared to the OLS VMPs, which are calculated for comparison sake. The major differences in the results are a doubling of the effect of approach shot accuracy as measured by the percentage of greens hit in regulation, and a halving of the VMP of driving distance. The game of golf can be disaggregated even further. Berry (1999) has argued that each of the statistics measured by the PGA TOUR actually depend upon the other skills in identifiable ways. For example, recovery from sand bunkers depends upon how good the sand bunker shot is and how good a putter the player is. So, the sand save percentage is a combination of two skills. Along the same lines, hitting greens in regulation is easier if one is a longer and straighter driver, so the greens in regulation percentage is a combination of three skills. Similarly, the number of putts per green reached in regulation depends upon the accuracy of the approach shot, which depends on the length and accuracy of the drive. More careful painstaking data crunching could be used in the examination of the golf statistics. The payoff of such research may be highest to the golfers themselves. Knowing the true, independent, effect of each of the skills could influence their practice routines and their on-course strategy in much the same way that statistical analysis influenced Major League Baseball, as suggested in Moneyball (Lewis, 2004) or that the so-called ‘analytics’ has transformed the NBA. A younger, more energetic version of myself would seriously be considering this as a career option. The next major improvement was undertaken in Shmanske (2008). Up until this point all of the variables were measured as year-end averages. As such, each of the skills is measured with error because the individual, tournament-by-tournament components of each skill are measured under different conditions, and because not all players enter each tournament. So, for example, if a player enters a tournament at high altitude but not one at sea level, his driving distance will be overstated compared to a golfer who does the opposite. Fortunately, by gathering the underlying data for each tournament, corrections can be made to the measurement of each skill. Only year-to-date data were available for my 2008 paper, from which it was possible to painstakingly back out the new data from each tournament from the change in the

The Economics of Golf

year-to-date data. Fortunately, the PGA TOUR now allows access to all its underlying data for bonafide academic researchers. A secondary improvement becomes available by having the whole year-long distribution of data instead of just the year-end average. Namely, other aspects of the distribution, such as the variance and skewness, become available for examination and hypothesis testing. Indeed, it is not the yearend average measure of any of the skills that will win a particular tournament. Perhaps instead, it is exceptional performance in one or a few tournaments rather than year-long consistency that pays handsomely. The results of the study indicate that the mean, variance, and skewness of the scoring distributions affect earnings in the predicted directions. However, only the mean of the scoring distribution is significantly affected by the parameters of each of the distributions of the skills. That is to say, variance and skewness of the scoring distributions are remunerative but are essentially random with respect to the skills themselves. Perhaps this is not surprising. There are both very good shots that hit the flagstick and carom into a nearby hazard, and what would otherwise be poor shots (at least in speed if not direction) that luckily hit the flagstick and fall into the hole. In concluding this section, I suggest that future analysis of the production or earnings functions will be somewhat different. In the discussion so far, the literature that started out with year-end data progressed to tournament-by-tournament data. Most recently, the PGA TOUR and an army of volunteers now track the beginning and end of every shot, by every player, in every tournament. Future analysis will be done with shot-by-shot data providing less measurement error and more finely developed testable hypotheses. I will return to this issue below.

APPLICATIONS OF THE GOLF PRODUCTION OR EARNINGS FUNCTION The typical error term in a regression equation includes the effects of all unmeasured influences whether random or not. Therefore, it is possible to conjecture that an unmeasured influence is systematically important and to try to get a handle on it. This is what Harold Fried with two sets of colleagues has done in a series of papers.5 These papers estimate a multidimensional frontier production function which identifies the lowest amount of talent required to achieve any given level of earnings. If a golfer with this level of earnings actually has more talent, as statistically

475

captured in year-end averages, then that golfer is not using his or her talent efficiently, and a numerical measure of the extent of the inefficiency can be calculated. In their explanation, the golfer may be failing under pressure in the last round or on the last few holes when in contention for a first place or other top-level finish. So, while the yearlong statistics might indicate high skill, if the skill disappears when it really counts, the earnings will be lower than the skill would indicate. The numerical level of inefficiency is only measurable ex-post, after the frontier production function has been calculated, so it cannot be used as an explanatory variable in developing a production function. However, if it can be meaningfully related to other data, its credibility will be enhanced. Fried and his colleagues relate the level of inefficiency to experience and to an ageearnings profile for PGA TOUR and LPGA golfers with some success. Further extensions along this line hold forth promise, especially with the data on hole-by-hole scoring that is now available. As the work stands, it is a nice example of the use of sports statistics to illustrate and enhance understanding of a significant field of study, namely, the measurement of productive efficiency and productivity growth. Another area where golf earnings functions can be applied is the issue of discrimination. By estimating the same earnings function separately for men and women6 or for young and old7 and then using the earnings decomposition method of Oaxaca (1973), one can determine how much of any earnings differential is due to lower input levels and how much is due to lower implicit prices paid for the inputs. If women earn less than men because of lower driving distance, it cannot be attributed to discrimination. But if the implicit price (the VMP from an earnings equation) of an extra yard of driving distance is lower for women, a claim of discrimination is appropriate. In more general industrial or managerial settings, it is next to impossible to get good measures of the multidimensional inputs involved. This is where sports economics pays off. The earnings functions developed for professional golf coherently explain the majority of the variation in earnings across individuals so the earnings decomposition method has the ability to tell most of the story. The results from my papers (Shmanske, 2000, 2012b) looking at 1998 and 2008 data are compelling. Although women playing in LPGA events earn less than men playing in PGA TOUR events, they are being fairly (actually more than fairly) paid given the skill levels they exhibit. If anything, there is discrimination in favor of women. Upon reflection this makes perfect sense. Women could enter and compete for PGA TOUR prize funds which are open to both

476

THE SAGE HANDBOOK OF SPORTS ECONOMICS

genders but find it more rewarding to compete in LPGA events from which men are barred. Perhaps the most straightforward (least imaginative?) application of a golf production function is to assess the truth of the old age, ‘drive for show, putt for dough.’ Shmanske (2009), using the tournament-by-tournament data collected for the 48 official PGA TOUR tournaments in 2006, suggests that the answer could differ depending upon the characteristics of the tournament venue. I estimated a scoring function for each tournament using standardized regression.8 The results show that putting was the most important skill in over half of the tournaments, hitting greens in regulation was the most important in 20 of the tournaments, and driving distance was never the most important skill and was statistically significant in only 10 of the tournaments. The old adage gets statistical support from this application of the golf production function.

GOLF AND GENDER In addition to the earnings decomposition that addresses the issue of discrimination there have been other papers looking at the economics of women’s golf or using women’s golf to highlight other economic or societal issues. In Shmanske (2013), I address recent trends in the driving distance statistic to see if women are catching up to men in the area of the game where there is the starkest gender difference. Although women made some modest gains in the early 1990s, driving distance has leveled off for all professionals, with men outdistancing women by between 35 and 40 yards. This paper also employs a Granger (1969) causality test on time series of purse size and driving distance to examine whether larger purses are encouraging professional golfers to develop greater distance, or whether the prodigious distances the professional golfers drive the golf ball are increasing fan support and prize funds. The results show that for men, increases in driving distance do seem to bring about larger purses but not vice versa. For women, the inferences are reversed. Increases in purse size in LPGA events (but not in PGA TOUR events) do lead to increases in driving distance. The time series examined in this paper (Shmankse, 2013) ends in 2010 and this issue should be revisited as more data become available. Ahn and Lee (2014) use LPGA golfers to study the relationship between beauty and earnings and discover a link between them. For golfers with above median attractiveness there is a positive correlation between tournament earnings and

attractiveness. Why should this be so? The authors argue that champion golfers can earn extra income in commercial endorsement possibilities, and that this avenue is more lucrative the more attractive one is. Essentially, the total income prize (purse plus endorsements) is larger the more attractive one is. Thus, there is a greater return to developing golf skills the more attractive one is. As is common in sports economics, we have another example of how careful, detailed sports statistics illuminate economic propositions stemming from other fields, in this case the issues of customer discrimination and the return to beauty in labor markets.9 Along similar lines, Lee, Park, Kang, and Lee (2013) find evidence to support a standard labor migration model (Borjas, 1987, 1994) based on earnings differentials between two regions. There has been a marked influx of female Korean golfers to play in LPGA events based in North America. The migration is greater for Korean females than for Japanese females and greater for Korean females than Korean males. This pattern is what would be expected based upon the prize funds and the level of competition in men’s and women’s tours in Korea, Japan, and North America. Gender differences with respect to responses to incentives are explored by Gilsdorf and Sukhatme (2013). Using 2009 data, they find that there are some differences between men and women. However, they report several findings that higher prize funds tend to increase, that is worsen, the actual scores of the participants. This is contrary to both theoretical expectation and some past results from similar formulations. As such, this line of research cries out for further examination.

SUPPLY: THE TOURNAMENT ENTRY DECISION The top professional golfers can achieve and maintain ‘exempt’ status, which means that they can pick and choose which, and how many, tournaments to enter during the year. Non-exempt golfers get to fill in any available slots after the exempt golfers have made their choices. The examination of the exempt golfer’s choices has been the topic of a handful of papers. An early example (Gilley & Chopin, 2000) is motivated by the possibility of documenting a so-called ‘backward-bending’ supply curve, in which the number of tournaments entered is negatively related to the prize funds. Matthew Hood (2006) looks at the entry decision with the tournament as the unit of analysis and with entry measured by the percentage of the top 30 golfers who enter a particular tournament.

The Economics of Golf

By looking at multiple years, he is able to determine that high purses do encourage entry of the top golfers, while, simultaneously, the entry of top golfers increases the purse in the future. He also establishes the importance of the golfer’s performances in previous years at the tournament and the dynamic effects of which tournaments are scheduled immediately before and after the tournament in question. Using a similar analysis on players from the European PGA TOUR, Hood (2012) confirms these results and is also able to control for and show that there is a home country effect in which golfers select tournaments in their home country with greater frequency than otherwise. Rhoads (2007) uses golfers as the unit of analysis and looks at several years of data, 1995 to 2002, during which a change in the rules regarding exemption status took place. He is able to address questions about the elasticity of supply as well as questions about how the issue of maintaining exempt status changes the supply decision. Shmanske (2009) takes the number of tournaments that a golfer enters as a given and attempts to determine which tournaments the golfer will enter. In this paper, I rely on the fact that different golfers have different skill sets and that tournaments reward skill sets differentially in that some require long drives, some require straight drives, etc. Evidence from a combinatorial analysis shows that about half of the top professional golfers clearly systematically choose the set of tournaments to enter based upon their expected relative performances on the courses in question. In a regression setting, some dynamic influences are added to the analysis. Maintaining exempt status or obtaining entry into special invited field tournaments by crossing certain income thresholds are also shown to be important influences.

SUPPLY: THE SUPPLY OF EFFORT Economists have cleverly examined the supply of effort by professional golfers along two separate lines of inquiry. First, there is the connection between the incentives implied by the prize fund and how much effort is expended. At issue is not only the total amount of the prize fund, but also one’s position in the rank order and the nonlinear nature of the tournament prize structure. Second, and not unrelated, is the issue of competition against a superstar, which might be seen as a lowering of the incentives as the expectation of prevailing against a superstar diminishes. In two of the earliest papers applying economics to golf, Ehrenberg and Bognanno (1990a,

477

1990b) explain how the nonlinear tournament prize structure supplies different amounts of monetary reward incentives depending on the golfer’s position on the leaderboard. For those at the top of the prize structure, one shot might be worth one half million dollars or more, whereas in the middle of the pack a one-stroke difference might only be $10,000 or so. At the bottom of the list, the difference between 69th and 70th place might be less than $100. Exploiting this fact, and looking at final round scores, the authors show that those with more money on the line play better than those who are out of serious contention. Following up this line of research, Orszag (1994), using a different year of data, is not able to replicate the incentive effects of greater purse size on overall scores, but does not look at final round scores, and suggests the desirability of controlling for weather differences. Matthews, Sommers, and Peschiera (2007) also look at overall scores and fail to discover a positive effect of the purse size on quality of play in LPGA events in 2000. Additionally, Gilsdorf and Sukhatme (2013) look both at overall scores and final round scores from the PGA TOUR and the LPGA in the year 2009, while mostly contradicting the results that money matters. Given the contradictions in the foregoing, this area is ready for further research, especially using new more detailed data, as described in the next section. Matthews, Sommers, and Peschiera (2007) also look at the possibility of a superstar effect in the 2000 LPGA data.10 Karrie Webb won six of the tournaments in their sample and was in the top 10 over half the time she competed. The authors look at differences in the coefficients from splitting the sample based on whether Webb entered. While some of the coefficients show a greater sensitivity of score to purse when Webb does not compete, the general impression is a lack of support for a superstar effect on effort and performance of the rest of the field. However, differences in golfer performance also depend on course difficulty, weather, and other factors which are difficult to control for with one year of data. Brown (2011) cleverly uses differences in playing schedules over several years (1999–2006) of PGA TOUR competition to isolate the effects of course difficulty from the absence or presence of Tiger Woods in the field. Among many results, Brown reports that during Tiger’s hot streaks the scores of the other members of the top 20 golfers are over two strokes higher than otherwise would be the case. Meanwhile, there is no appreciable effect during Tiger’s relatively cool periods when other golfers seemed to think they had more of a chance. Thus, there is some evidence in support of a superstar effect on effort.

478

THE SAGE HANDBOOK OF SPORTS ECONOMICS

THE NEW DATA, STRATEGY, AND BEHAVIORAL ECONOMICS Economic analysis of golf started with data on yearly averages and moved to tournament-bytournament data, and also to round-by-round data. More recently, hole-by-hole data and shot-by-shot data have become available and lend themselves to refinements of old research questions and the introduction of wholly new testable implications. For example, whereas a few of the above papers tested propositions related to final round scores, Oettinger and Bronars (2014) are able to focus on the back nine and even on the last few holes to examine the relationship between performance and the actual money at stake, confirming the results that money matters. It is also possible to look at the variance of hole scores relative to par (or relative to the average score on that hole). One can determine an individual golfer’s style along a continuum of playit-safe conservative to go-for-broke risk taking. The tournament promoters can look at a particular hole’s variance to predict score swings and generate excitement for the golf fans. More importantly for sports economics, one can also test propositions about on-course strategy, as explained below. Looking at a round of golf as a set of independent trials, a golfer would be expected to adopt whatever style (safe or risky) was consistent with achieving the lowest average score per hole, in order to achieve the lowest expected score per round. However, there are situations in which a golfer would be willing to accept a higher average score by playing it safer than usual in order to avoid the possibility of a really bad score. This would be the case if the golfer had a lead going into the last few holes or had a score that was above but near the cut line after the second round. The opposite is also possible. If the golfer is just below the cut line or just behind the leader, the golfer would want to increase the probability of making a birdie by adopting a riskier strategy on a shot or a hole. This would increase the expected score on the hole but would be okay because a bad score would not hurt and only an exceptional score would help the golfer make the cut or catch the leader.11 Oettinger and Bronars (2014) look at precisely this issue in their clever research. With shot-by-shot data, even more clever applications are possible. The PGA TOUR now uses an army of paid and volunteer workers and global positioning to track the beginning and end of every shot by every golfer in every tournament. From this data, a ‘shots gained’ statistic is developed for each golfer for each type of shot. Consider a 200yard shot from which the other golfers take three

strokes to finish the hole, an approach shot to 20 feet, a putt to within a few inches, and a tap in. Now consider golfer A who holes out in two shots. If he hits the 200-yard shot to within tap-in range, then he gets a stroke gained (that is one stroke less than the average) on the 200-yard shot and a zero strokes gained on the putt. Alternatively, if he hits the 200-yard shot to 20 feet like everyone else but makes the putt, then he gets zero strokes gained on the approach shot and one stroke gained on putting. Other combinations are also possible. Suppose the 200-yard shot ends up 10 feet from the hole in a position from which the other competitors make the putt half the time. In this case, the golfer would have gained one-half stroke on 200-yard shots and one-half stroke putting. The papers by Stockl, Lamb, and Lames (2011) and Fearing, Acimovic, and Graves (2011) carefully explain the data and concentrate on the examination of the strokes-gained on putting statistics. This new measurement statistic promises to further refine the production and earnings functions described earlier. The new statistics have also allowed a test of the loss aversion hypothesis that stems from behavioral economics. Pope and Schweitzer (2011) looked at over 2.5 million putts and, using par as a reference point, determine that par putts are hit more aggressively and made more often than similar birdie putts. Their results are statistically and economically significant and withstand numerous robustness checks and alternative explanations, such as the number of previous putts the golfer is able to view before hitting his own. The detailed statistics from golf have allowed arguably the best test of behavioral economics where real-world seasoned professionals with a significant amount at stake seemingly make biased decisions. If golfers take the conclusion in this paper to heart, their behavior may change, and the loss aversion bias may disappear in the future. I am looking forward to follow-up research on this important issue in economic theory.

MISCELLANEOUS The additional variety in the application of economics to golf is almost endless and will receive brief mention in this final section. Continuing with studies related to professional golf, Shmanske (2005) uses insights from finance to look at informational and transaction efficiency in posted odds gambling markets for PGA TOUR events. Agrawal, Grimm, and Fung (2013) and Knittel and Stango (2014) used event studies to look at

The Economics of Golf

the value of celebrity endorsement for the case of Tiger Woods. And Shmanske (2012a) looks at economic activity in a rich and varied time series of major golf tournaments to assess the impact of hosting a mega-event. Many other studies have used data from recreational golf and the golf course industry to test or illustrate economic propositions. For example, looking at a large variety of pricing schemes used by golf courses, Shmanske (1998) shows that price discrimination is profitable and that the overwhelming majority of demand studies that only capture price in a one-dimensional vector all suffer from an omitted variable bias. In Shmanske (1999) I looked at the effects of golf course design and maintenance on demand. Do and Grudnitski (1995) use hedonic pricing to assess the additional value of real estate adjacent to a golf course. Shmanske (2004) looks at slow play on golf courses through the lens of Just-in-Time production smoothing arguments. I have also shown how government subsidized pricing on municipal golf courses actually depresses golf course construction and raises the full price (monetary and waiting cost) of a round of golf (Shmanske, 1996). Finally, I simply mention two other sources which have not been intensively mined by professional economists. First, the National Golf Foundation does numerous studies of the recreational golf industry for reports that they sell on their website, (www.ngf.org) but typically do not share their data. And second, the World Scientific Congress of Golf first met in St Andrews, Scotland, in 1990 and has met occasionally since (by my count, the seventh meeting was in 2016). The focus is primarily on equipment technology, golf course management, agronomy, and kinesiology, but economics is often not far below the surface, and some papers are specifically economic in scope. Conference abstracts and proceedings are usually published (for an example, see Cochran & Farrally, 1994).

Notes  1  See Cottle (1981), who uses insights from standard cartel theory in the earliest application of economic analysis to golf. See also Shmanske (2004).  2  An earlier paper from the exercise and sport literature (Nix & Koslow, 1991) does not report statistics from a defensible statistical model. For those unfamiliar with the sport of golf or the typical list of skills, see Shmanske (1992) for an overview of the play of the game and its required skills.  3  In the only attempt of its kind, this paper also examined the production and maintenance of the

479

skills themselves by correlating them to the practice routines of a small subset of golfers.  4  A nice listing of 16 such studies appears in Shmanske (2013), which highlighted driving distance and reports the VMPs and the elasticities of earnings with respect to it.  5  See Fried, Lambrinos, and Tyner (2004) and Fried and Tauer (2011, 2012).  6  See Shmanske (2000, 2012b, 2015).  7  See Rishe (2001).  8  In standardized regression the data are transformed into z-scores and the coefficients measure the effect of being one standard deviation above the mean of the data. In this sense the results show which skill provides the biggest gain from being one standard deviation better than the mean of your peers.  9  See Hamermesh and Biddle (1994). 10  In a different type of superstar effect, Gooding and Stephenson (2016) show how television ratings are influenced by the presence of certain golf superstars, notably Tiger Woods. 11  With respect to missing the cut, the bad score costs nothing. With respect to one’s final rank, the bad score costs less than the possible gain from a good score in the typical tournament prize scheme.

REFERENCES Agrawal, J., Grimm, P., & Fung, S. (2013). Benefits and costs of hiring and firing Tiger Woods. California State University, East Bay, Working Paper, December. Ahn, S. C., & Lee, Y. H. (2014). Beauty and productivity: the case of the Ladies Professional Golf Association. Contemporary Economic Policy, 32(1) 155–168. Alexander, D. L., & Kern, W. (2005). Drive for show and putt for dough? An analysis of the earnings of PGA TOUR golfers. Journal of Sports Economics, 6(1), 46–60. Berry, S. M. (1999). Drive for show and putt for dough. Chance, 12(4), 50–55. Borjas, G. J. (1987). Self-selection and the earnings of immigrants. American Economic Review, 77(4), 531–553. Borjas, G. J. (1994). The economics of immigration. Journal of Economic Literature, 32(4), 1667–1717. Brown, J. (2011). Quitters never win: the (adverse) incentive effects of competing with superstars. Journal of Political Economy, 119(5), 982–1013. Callan, S. J., & Thomas, J. M. (2007). Modeling the determinants of a professional golfer’s tournament earnings: a multiequation approach. Journal of Sports Economics, 8(4), 394–411.

480

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Cochran, A. J., & Farrally, M. (Eds.) (1994). Science and Golf II: Proceedings of the World Scientific Congress of Golf. London: E & FN Spon. Cottle, R. L. (1981). Economics of the Professional Golfers’ Association Tour. Social Science Quarterly, 62, 721–734. Do, A. Q., & Grudnitski, G. (1995). Golf courses and residential house prices: an empirical examination. The Journal of Real Estate Finance and Economics, 10(3), 261–270. Ehrenberg, R. G., & Bognanno, M. L. (1990a). Do tournaments have incentive effects? Journal of Political Economy, 98(6), 307–324. Ehrenberg, R. G., & Bognanno, M. L. (1990b). The incentive effects of tournaments revisited: evidence from the European PGA TOUR. Industrial and Labor Relations Review, 43, 74S–88S. Fearing, D., Acimovic, J., & Graves, S. C. (2011). How to catch a tiger: understanding putting performance on the PGA TOUR. Journal of Quantitative Analysis in Sports, 7(1) article 5. Fried, H. O., Lambrinos, J., & Tyner, J. (2004). Evaluating the performance of professional golfers on the PGA, LPGA, and SPGA Tours. European Journal of Operational Research, 154(2), 548–561. Fried, H. O., & Tauer, L. W. (2011). The impact of age on the ability to perform under pressure: golfers on the PGA TOUR. Journal of Productivity Analysis, 35, 75–84. Fried, H. O., & Tauer, L. W. (2012). Age and performance under pressure: golfers on the LPGA Tour. In S. Shmanske & L. Kahane (Eds.), The Oxford Handbook of Sports Economics. Vol. 2: Economics through Sports (pp. 135–152). New York: Oxford University Press. Gilley, O. W., & Chopin, M. C. (2000). Professional golf: labor or leisure. Managerial Finance, 26(7), 33–45. Gilsdorf, K. F., & Sukhatme, V. A. (2013). Gender differences in responses to incentives in sports: some new results from golf. In E. M. Leeds & M. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 92–114). Northhampton, MA: Edward Elgar. Gooding, C., & Stephenson, E. F. (2016). Superstars, uncertainty of outcome, and PGA TOUR television ratings. Journal of Sports Economics, online first, 21 April. doi: 1527002516637649. Granger, C. W. J. (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37(3), 424–438. Hamermesh, D. S., & Biddle, J. E. (1994). Beauty and the labor market. American Economic Review, 84(5), 1174–1194. Hood, M. (2006). The purse is not enough: modeling professional golfers entry decision. Journal of Sports Economics, 7(3), 289–308. Hood, M. (2012). Remembering three economic studies on professional golf. In L. Kahane & S.

Shmanske (Eds.), The Oxford Handbook of Sports Economics. Vol. 1: The Economics of Sports (pp. 301–317). New York: Oxford University Press. Kahane, L. H. (2010). Returns to skills in professional golf: a quantile regression approach. International Journal of Sport Finance, 5(3), 167–180. Knittel, C. R., & Stango, V. (2014). Celebrity endorsements, firm value, and reputation risk: evidence from the Tiger Woods scandal. Management Science, 60(1), 1–17. Online first, 16 September 2013. Lee, Y. H., Park, I., Kang, J.-H., & Lee, Y. (2013). An economic analysis of the sudden influx of Korean female golfers into the LPGA. In E. M. Leeds & M. A. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 388–409). Northampton, MA: Edward Elgar. Lewis, M. (2004). Moneyball: The Art of Winning an Unfair Game. New York: W. W. Norton. Matthews, P. H., Sommers, P. M., & Peschiera, F. (2007). Incentives and superstars on the LPGA Tour. Applied Economics, 39(1–3), 87–94. Moy, R. L., & Liaw, T. (1998). Determinants of professional golf tournament earnings. The American Economist, 42(1), 65–70. Nix, C. L., & Koslow, R. (1991). Physical skill factors contributing to success on the professional golf tour. Perceptual and Motor Skills, 72, 1272–1274. Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International Economic Review, 14, 693–709. Oettinger, G. S., & Bronars, S. (2014). Giving 110% and going for broke: the effort and risk-taking responses of professional golfers to tournament incentives. Unpublished manuscript. Orszag, J. M. (1994). A new look at incentive effects in tournaments. Economics Letters, 46(1), 77–88. Pfitzner, C. B., & Rishel, T. D. (2005). Performance and compensation on the LPGA Tour: a statistical analysis. International Journal of Performance Analysis in Sport, 5(3) 29–39. Pope, D. G., & Schweitzer, M. E. (2011). Is Tiger Woods loss averse? Persistent bias in the face of experience, competition, and high stakes. American Economic Review, 101(1), 129–157. Rhoads, T. A. (2007). Labor supply on the PGA TOUR: the effect of higher expected earnings and stricter exemption status on annual entry decisions. Journal of Sports Economics, 8(1), 83–98. Rishe, P. J. (2001). Differing rates of return to performance: a comparison of the PGA and senior golf tours. Journal of Sports Economics, 2(3), 285–296. Scully, G. W. (2002). The distribution of performance and earnings in a prize economy. Journal of Sports Economics, 3(3), 235–245. Shmanske, S. (1992). Human capital formation in professional sports: evidence from the PGA TOUR. Atlantic Economic Journal, 20(3), 66–80.

The Economics of Golf

Shmanske, S. (1996). Contestability, queues, and governmental entry deterrence. Public Choice, 86, 1–15. Shmanske, S. (1998). Price discrimination at the links. Contemporary Economic Policy, 16(3), 368–378. Shmanske, S. (1999). The economics of golf course condition and beauty. Atlantic Economic Journal, 27(3), 301–313. Shmanske, S. (2000). Gender, skill, and earnings in professional golf. Journal of Sports Economics, 1(4), 385–400. Shmanske, S. (2004). Golfonomics. River Edge, NJ: World Scientific Publishing. Shmanske, S. (2005). Odds-setting efficiency in gambling markets: evidence from the PGA TOUR. Journal of Economics and Finance, 29(3), 391–402. Shmanske, S. (2008). Skills, performance, and earnings in the tournament compensation model: evidence from PGA TOUR microdata. Journal of Sports Economics, 9(6) 644–662. Shmanske, S. (2009). Golf match: the choice by PGA TOUR golfers of which tournaments to enter. International Journal of Sports Finance, 4(2), 114–135.

481

Shmanske, S. (2012a). The economic impact of the golf majors. In W. Maennig & A. Zimbalist (Eds.), Handbook on the Economics of Mega Sporting Events (pp. 449–460). Northampton, MA: Edward Elgar. Shmanske, S. (2012b). Gender and discrimination in professional golf. In S. Shmanske & L. Kahane (Eds.), The Oxford Handbook of Sports Economics. Vol. 2: Economics through Sports (pp. 39–54). New York: Oxford University Press. Shmanske, S. (2013). Gender and skill convergence in professional golf. In E. M. Leeds & M. A. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 73–91). Northampton, MA: Edward Elgar. Shmanske, S. (2015). Super Golfonomics. Singapore: World Scientific Publishing. Sommers, P. M. (1994). A bread and putter model. Atlantic Economic Journal, 22, 77. Stockl, M., Lamb, P. F., & Lames, M. (2011). The ISOPAR method: a new approach to performance analysis in golf. Journal of Quantitative Analysis in Sports, 7(1) article 10.

47 IRON(O)MICS: The Market for Long-Distance Triathlon Joachim Prinz

INTRODUCTION The past two years saw the triathlon market turn its best performance since its inaugural race in Hawaii in 1978. Over the course of the 2015/2016 season, output grew in terms of new and additional triathlon races as well as finishers all over the world. Different than running events (i.e. city marathons) short- and long-distance triathlon participation rose by 15–20% pa. According to a recently conducted triathlon market survey performed by the International Triathlon Union (ITU) in 2013, more than 2 million recreational athletes competed in almost 13,000 contests around the world. In this context, as expected, the number of German triathlon club members increased by more than 50% between 2010 (35,000 athletes) and 2015 (55,000 athletes), for an average rate of 4,000 ‘rookies’ per year alone in Germany. Those who are determined to finish a race do so at considerable cost. Apart from the (extreme) physical and mental investments necessary to accomplish a race, direct costs and expenses for equipment, travelling and starting licences (Wicker, Prinz & Weimar, 2013) accumulated worldwide to US$4 billion – approximately the sum that the NBA economized in 2015. Given the current hype, triathlon, and especially the long-distance market, has become a booming mass-participation sport,

challenging and partly substituting traditional city marathon events. Since ‘triathloning’ is regarded as a lifestyle development particularly lived by recreational athletes, spending huge amounts of dollars, triathlon organizers, and especially professional athletes, have mostly benefited from the commercialization and globalization of triathlon. Among the competition organizers, the long-haul market, typically known as the IRONMAN distance, stays in focus by recreational athletes, and the media. Here, the IRONMAN™ organizer is the most prestigious leader, supplying triathlons worldwide (Adler, 2017). Very recently, the Chinese Wanda group has paid $650 million to the former IRONMAN owner, the American World Triathlon Corporation (WTC), in an effort to spread triathlon popularity in the Far East, and especially to the huge Chinese market. In this new ‘money business context’, two world-class triathletes cashed in the highest paychecks in triathlon’s history and in their individual careers, receiving quite considerable prize money: while Jan Frodeno won $120,000 for winning the IRONMAN World Championship in Hawaii, Daniela Ryf won the so-called ‘Triple Crown’ tournament by being the first athlete ever winning a series of three ex-ante specified races in 2015 (IRONMAN 70.3 Dubai; IRONMAN 70.3 World Championship Zell am

483

IRON(O)MICS: The Market for Long-Distance Triathlon

See, and IRONMAN 70.3 Bahrein), receiving a purse of US$1 million. A simple look at these figures implies that triathlon has become a big industry. However, compared to individual athletes showing up on the Forbes ‘World’s Highest Paid Athletes’ list, it illustrates that professional triathletes are marginally remunerated. But even compared to two other endurance sports, such as marathon running and swimming, professional triathletes earn considerably less. Table 47.1 illustrates the purse that the top three athletes in their selected sport have earned in 2015. Apart from Jan Frodeno, the money list leader in professional triathlon, the runner-ups do not earn similar stakes than swimmers or marathon athletes. The exception is certainly Daniela Ryf, who has won the $1 million prize from the special triple crown event, and Jan Frodeno, who cashed in $120,000 for winning the World Championship in Hawaii in 2015. Compared to other big market sports, such as tennis, golf or Formula One, individual endurance sports are poorly paid. Nevertheless, the displayed earnings of the top money makers in Table 47.1 do not represent athletes’ real annual income. Much more, their personal income is largely driven by sponsorships, grants and other endorsement money. This is particularly true for professional triathletes where only one event (the Championship in Hawaii) determines the probability of receiving lucrative sponsorships. Therefore, the triathlon money market is heavily skewed to the right, in a sense that only the Hawaii winner takes it all. The remainder of the chapter is structured as follows: the first section provides a historical explanation about triathlon in general, but more so on long-distance races, especially the IRONMAN

race. It describes the organization of the events, illustrates their location and gives some insights about the triathlon market. The subsequent section theorizes the determinants of elite athletes’ output and the interest of the triathlon organizer. Specifically, the next section analyzes the demand function of the long-distance triathlon market. Finally, the chapter concludes and raises some questions of improvement.

TRIATHLON-EVOLUTION AND THE ORGANIZATION OF THE TRIATHLON MARKET For some, triathlon is regarded as one of the world’s most admired and exhausting sports (Leper, 2008). This is especially true for the longdistance race, the so-called IRONMAN. It is particularly long and grueling and combines swimming, cycling and running. It is one of the most extreme sports and requires a specific level of physical and psychological fitness of different skills (Knechtle, Wirth & Rosemann, 2010). On average, the IRONMAN-triathlete faces energy costs between 8,000 and 10,000 kcal during a race (Laursen & Rhodes, 2001).1 The triathlon idea was born in Honolulu, Hawaii, in 1978. Three navy officers argued whether swimming, biking or running is supposed to be the world’s most endurable sport (Sowell & Mounts, 2005). To settle the discussion, Commander John Collins issued the challenge to combine three long-distance events on one day: the ‘3.8k Waikiki Roughwater Swim’, the 180k ‘Around Oahu Bike Race’ and the ‘Honolulu Marathon’. This race took place on October 18,

Table 47.1  Top three annual prize money (US$) makers in selected sports, 2015 Money Rank

1 2 3

Marathon

Triathlon#

Swimming##

Male athletes

Male athletes

Male athletes

Name

Purse ($) Name

Purse ($) Name

Purse ($)

Desisa Benti Eliud Kipchoge Berhanu Hayle

275.000 226.570 219.135

213.000 77.750 75.500

262.200 105.500 92.250

Female athletes

1 2 3

Jan Frodeno Andreas Raelert Andy Potts Female athletes

Name

Purse ($) Name

Mergia Medessa Keitang Chepkosgei Dibaba Harssa

307.000 244.650 195.000

Daniela Ryf Meredith Kessler Litz Blatchford

Cameron van der Burgh Chad le Clos Mitch Larkin Female athletes

Purse ($) Name 1.223.000 Katinka Hosszu 86.000 Emily Seebohm 79.750 Zsuzsanna Jakobs

Purse ($) 325.000 188.000 99.500

Notes: # includes prize money from long-distance and half-long-distance races in 2015; ## includes prize money from the FINA World Cup Tour 2015, different lengths.

484

THE SAGE HANDBOOK OF SPORTS ECONOMICS

1978, on Oahu island. Only 12 American men competed in this ‘opener’. ‘Whoever finishes first, we’ll call him the Iron Man’ was the slogan of these early years of triathlon evolution. Gordan Haller, a marathon runner, won the challenge in 11:46 hours. In 1981, this competition was relocated from Waikkiki to Kailua Kona on Big Island. Since then, the long-haul triathlon, also called IRONMAN triathlon, gained great popularity and attention all over the world. In many other countries, much shorter triathlons (i.e. the San Francisco triathlon ‘Escape from Alcatraz’) were established and actually the first swim and run took place in San Diego in 1974. In 1983 the World Triathlon Cooperation (WTC) was founded. The WTC is the owner of the IRONMAN™ and operates by licensing IRONMAN Hawaii qualification contests and slots. This business model was the origin of the IRONMAN series consisting of qualifying races in New Zealand (1985), Japan (1985), Australia (1985), Canada (1987), Roth (1987), Nice (1988) and Lanzarote in 1992. From the beginning of the IRONMAN family in 1985, the series has grown to 39 races worldwide and most of them are sold out (capacity between 2,000 and 3,000 athletes) in a couple of a days or even minutes (McCarville, 2007).2 Table 47.2 portrays the roster of the current 39 IRONMAN (IM) locations worldwide. The table includes average values of 150 events between 2011 and 2015. With the exception of the race on Hawaii, all races are qualifying competitions for the most popular and one of the toughest long-distance triathlons, the annual World Championship in Hawaii. As always, participants need to cover a 3.8k swim, a 180k bike ride and a final marathon run within 15 hours.3 While all races cover the same distance, they vary in terms of course severity, where topography and weather impairs the level of effort. For example, while IM Lanzarote is well known for its heavy winds and elevated bike split, IM Wales is notorious for its cold water temperature, while athletes in Malaysia face great humidity and temperatures of more than 35 degrees Celsius. On the other hand, the good results from IM Barcelona are mainly determined by the fast-track cycle course (flat out) and IM Chattanooga provides athletes with a strong current towards the swim finish. Depending on the level of difficulty (elevation, winds, water temperature, humidity, ocean swimming, etc.) finishing times, and the number of starters and drop-outs vary considerably. Moreover, these differences lead to different levels of prestige within the triathlon community. Similar to many other sports, triathlon is separated between professional and recreational athletes. However, in triathlon, hobby athletes

(‘age-groupers’) compete in the same race as professional athletes do, but only the ‘pros’ will receive the prize money. In order to enter the World Championship on Hawaii, both groups (pros and age-groupers) need to qualify for a slot in Kona. Every athlete needs to qualify for the IRONMAN Hawaii by matching specific thresholds in one of the remaining 38 IRONMAN races worldwide. Similar to the ATP tennis ranking, the pros collect points for the Kona Pro Ranking (KPR) while doing qualifying races. The top 50 professional athletes are then eligible to participate on Big Island, if, and only if, they have collected a specific threshold of Kona points and have finished at least one of the season’s 38 qualification races. The winner in Hawaii is the annual World Champion. Recreational athletes follow a different mechanism. As before, these athletes have to meet the qualifying times in one of the qualifying races (Table 47.2). However, recreational athletes only compete with peers from their same age and gender group. Depending on the number of athletes running in the respective group, a predetermined number of (Hawaii) slots is awarded to the fastest athletes. In 2006 the WTC undertook some product differentiation. Targeting new customers, it spun-off its IRONMAN 70.3 series. IRONMAN 70.3 is half the distance of a full long-distance race (140.6 miles). Currently, professional and age-grouping athletes have the opportunity to participate in more than 100 IRONMAN 70.3 races. Apart from the chance to qualify for the IM 70.3 World Championships, some selected 70.3 events also allow qualification to Hawaii. The IRONMAN brand is the flagship of the WTC and the market leader in organizing fulland half-distance triathlon events. It accounts for a 91% global market share of long-haul competitions (www.active.com/triathlon). As a result, this marketplace is determined by quasi-monopoly power, giving marginal room for only two other (marginal) players, the German-based ChallengeFamily (Roth, Germany) and the Ican series from Gandia, Spain. Both series expand the number of half- and full-distance triathlons around the world. The Challenge family adds 10 races, and the Ican sets three more races. The Challenge corporation is the owner of the biggest (in terms of participants) and fastest full-distance race – the competition in Roth, Germany. With the exception of Hawaii, it is the most traditional triathlon event in the world and was the first race on European ground, formerly known as the IRONMAN Europe. In 2001 the licensed family lost its IRONMAN brand name status and the IRONMAN Europe was relocated to Frankfurt, Germany, where the European Championships have been regularly held since then. Although the long-distance ‘market’ and

485

IRON(O)MICS: The Market for Long-Distance Triathlon

Table 47.2  Worldwide IRONMAN race characteristics (2011–2015, daily averages) Race

Air-Temp Humidity WindSpeed

WaterTemp

Salt Elevation Elevation Bike Run

Money

Starters Finishers

Finish Time in Sec.

Lanzarote France Austria Frankfurt UK Switzerland Maastricht Sweden Copenhagen Vichy Wales Mallorca Barcelona Kapstadt Taiwan Malaysia New Zealand Australia Melbourne Cairns W Australia Texas Lake Placid Canada Boudler Mont-Tremb Japan

21,8 22,4 19,8 21,6 17,2 20,6 18 16,5 17,3 27 14 20 19 18,8 18 28 14 17,2 16 21,5 22,2 26 23,2 15 23,5 17 19

60 69 61 64 75 75 55 79 77 33 77 70 77 74 86 84 69 71 57 84 50 82 68 65 41 85 84

26,4 9,4 5,8 12,4 13,4 7,8 5 11,5 18,3 10 14,4 5,5 10,5 16,4 18 3 11,6 13,8 17 14,3 19,8 23 12,2 1,67 3,5 2,5 16,7

18,4 24 23 22 16 22,2 16 19 19 21 16 24,5 21 20 24 30 18 21 18 22 21 25 20,6 19 23 21 21,3

1 1 0 0 0 0 0 1 1 0 1 1 1 1 1 1 0 0 1 1 1 0 0 0 0 0 0

2668 2053 1771 1144 1877 1550 800 530 633 1098 2105 2105 653 1800 1735 1397 927 1255 928 1050 227 756 1681 1889 1254 1882 2355

286 113 149 372 329 212 120 312 173 132 499 499 124 107 250 88 237 205 109 87 121 104 298 343 141 341 283

25000 75000 70000 125000 25000 50000 25000 68750 50000 25000 25000 25000 50000 90000 25000 40000 55000 30000 131250 125000 50000 100000 20000 75000 12500 118750 25000

1824 2682 2812 2946 1654 2151 877 2033 2557 1610 1724 1724 2522 1912 1023 927 1401 1498 2004 1109 1450 2651 2780 2009 2764 2379 1494

1533 2220 2328 2289 1380 1728 714 1732 2201 1186 1361 1361 2023 1573 815 766 1310 1300 1717 910 1238 2101 2277 1724 2031 2131 1197

32903 31750 30068 29972 32923 32076 31699 30787 30559 31370 34213 34213 29663 31286 31403 33320 31481 32180 29147 31450 30780 30497 33124 32462 32965 31810 34504

Coeur d’Alene Wisconsin Chattanooga Maryland Hawaii Louisville Los Cabos Florida Arizona Cozumel Brazil Fortaleza

18

58

8,2

17

0

1701

286

4000

2479

1947

31983

17,6 22 16 27,2 24,6 23,3 18,2 18,4 21,4 17,6 27,5

71 79 75 66 63 46 74 50 79 78 68

7,6 2,5 4 12 7,6 10 11,4 9,4 8,2 6,6 22

22 22 22 26 23,6 28 21 18 25 18 27

0 0 0 1 0 1 1 0 1 1 1

1530 1280 309 1171 1374 1578 359 490 107 777 960

208 438 59 311 75 157 48 156 62 205 126

2000 75000 12500 636000 20000 16666,67 20000 80000 80000 90000 75000

2797 2512 2054 2117 2629 1042 2985 2949 2266 1884 894

2326 2156 1371 1953 2216 757 2406 2392 1699 1540 687

32738 30080 32740 30188 32939 32643 29946 29589 30842 30312 33204

the IRONMAN Championship in Hawaii is the ‘holy grail’ within the triathlon social world, there are many smaller triathlon races in the world.

According to the active.com registration portal (www.active.com/triathlon), in the US alone more than 480,000 people participated in shorter events.

486

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Most typically, these races cover the Sprint- and the Olympic distance. Mark Allen, the IRONMAN Hawaii winner from 1986 and 1987, was the first winner (1989) of the short-distance World Championships established by the world triathlon governing body, the ITU (International Triathlon Union). For professional athletes, the ITU operates an annual tour of (short) triathlon events, the World Triathlon Series (WTS). The tour consists of nine (global) events, culminating with a grand final race. The athlete with the highest number of points accomplished with his respective rank is the ITU world champion. Athletes usually do the Olympic distance (OD) of 1500m open water swimming, a draft-legal bike ride (40k) and a 10k running leg.4 Tables 47.3 and 47.4 summarize the triathlon market. Table 47.3 highlights the most important

races and triathlon series by incorporating the respective total prize purse, athletes’ productivity and the number of finishers for the extracted 2015 season. Table 47.4 shows some time series of the development of the IRONMAN market between 2008 and 2016.

THE ECONOMICS OF IRONMAN TRIATHLON: A THEORETICAL FRAMEWORK Being a (professional) IRONMAN triathlete is a risky choice. Even for elite athletes, completing the number of kilometers in a contest is a difficult adventure. Sometimes athletes face temperatures of more than 40 degrees Celsius, making events highly demanding both physically and mentally. Moreover, athletes incur considerable direct costs,

Table 47.3  Triathlon-portfolio 2015 (professional athletes) Type

Description

Kona, Hawaii IM World Champs IM 38 IM ‘Qualifiers’ IM70.3 70.3 World Champs IM 70.3 107 IM 70.3 ‘Qualifiers’ Challenge 8 full distance races ITU (OD) 9 races Sum

164

Total purse Prize money (US$, 2015) for winner (US$, 2015)

No. of athletes in prize money rank (2015)

Example(s)

Course record (location, year)

650.000

120.000

20 (m/f)

Kona, Hawaii

(8:04h) (Kona, 1996)

2.315.000

10.000–50.000

318 (m/f)

7:44h (Tempe, 2016)

250.000

60.000

20 (m/f)

2.177.500

5000–10.000

400 (m/f)

360.000

6000

83 (m/f)

2.385.000

18.000

36 (m/f)

Frankfurt, Florida, Cairns Zell am See, Mooloolaba Wiesbaden, Dubai Roth, Wanaka (NZ) Hamburg, Gold Coast

n.a. 3.34h 7.35h (World Record, Roth, 2016)

6.959.500

Table 47.4  Development of the IRONMAN market 2008–2016 Year

No. of IM races

2008 2009 2010 2011 2012 2013 2014 2015 2016

22 23 24 24 25 29 35 39 39

Prize purse (US$) / / / 1.550.000 2.125.000 2.300.000 2.615.000 2.965.000 2.965.000

#Starter / / / 45.799 53.214 63.273 78.842 79.960 80.152

#Finisher / / / 37.452 43.439 53.639 65.894 63.380 68.587

Annual best time (male) 7:59:55h 7:55:53h 7:52:05h 7:41:33h 7:57:21h 7:53:12h 7:48:43h 7:48:45h 7:44:39h

∅ Age Group No. of annual IM races (elite) Hawaii / 11:39h 11:14h 11:26h 11:32h 11:06h / /

1,7 1,7 1,8 1,9 2,0 2,0 2,0 / /

IRON(O)MICS: The Market for Long-Distance Triathlon

i.e. for equipment and travelling expenditures around the world (Wicker, Hallmann, Prinz & Weimar, 2012; Wicker, Maxcy & Prinz, 2019). While input costs are high, output in terms of measurable returns (i.e. prize money) is low. As in marathon running, elite athletes do not have many opportunities to race competitively during the season since full regeneration from a past race and necessary qualifying for an IRONMAN race is comparably long (Frick & Klaeren, 1997; Frick & Prinz, 2007). Based on the results of 230 IRONMAN races between 2000 and 2011, Adler (2017) shows that the average number of performed long-distance races is about two a year, which limits the athlete’s opportunities for making (prize) money. Consequently, triathletes need to select races in order to be successful. Traditionally, the highest return on investment is achieved at the World Championships in Hawaii. Hawaii is the athlete’s marquee event. Nevertheless, even for pros, especially when suffering, participating and finishing is certainly their priority. As stated above, the WTC controls almost 40 long-distance races around the world. Different than the ITU, the WTC is a private race organizer and its main objective is to increase its profit. Apart from merchandizing and sponsoring money, revenues are mainly generated by registration fees as well as the overall reputational gains from conducting IRONMAN competitions. Although the new Chinese owner has installed new events under the umbrella of the WTC, the number of athletes triathloning exceeds by far the slots available at major long-distance races (Britt, 2014), while, simultaneously, entry fees have risen to US$1,000 per race (IM Switzerland, Zurich, 2017). The main reason for this disequilibrium (demand exceeds supply) is a bottleneck on the supply side, imposed by the nature of the infrastructure (bike and run course) necessary to conduct such a long race (15 hours is the maximum length for athletes to finish the race). Moreover, safety concerns related to the first discipline is a crucial element affecting the organizer’s and the local organizing committee’s (town, region) reputation. Typically, the swim leg opens the competition in a mass start, with 3,000 athletes heating up the water. However, congested triathlon swims invoke significant contact (McCarville, 2007) between triathletes (kicks to the head), who are already confronting human-generated and natural waves, which makes it difficult not to hyperventilate. Since the director is well aware of this problem, the number of active athletes is artificially constrained somewhere between 2,500 and 3,000 participants in order to save the organizer’s reputation and to avoid deaths by drowning. This capacity limitation leads to the fact that many major IRONMAN races are sold out immediately

487

after registration has begun. Consequently, the organizer has discretion to raise the starting fee for recreational athletes, which opens the door to prizes being doled out for professional athletes. In this sense, age-group athletes cross-subsidize elite athletes, while a higher prize purse might attract better athletes, generating faster and more interesting races (Frick & Klaeren, 1997). High registration fees and the (limited) number of age-group athletes participating in a race determine the profit function of the organizer. From this point of view, the organizer has an incentive to sell as many slots as possible, while not overreaching the race-specific limitation. As a result, the organizer needs to find out the relevant determinants influencing the demand of age-group athletes. Most interesting in this context seems to be the number of starters and the number of finishers. After all, finishing an endurance race that covers 226 kilometers in 15 hours is the priority of most age-group athletes. Therefore, starting, and especially finishing, the race is influenced by the difficulty of the course environment (the topography of the bike and run course), swim conditions (lake or ocean swim) as well as other externalities, such as wind and heat. Given the point that the organizer tries to maximize the demand of athletes and that athletes are interested in finishing the race while expending as little effort as possible, ‘easier’ races are expected to be ‘sold out’ more frequently than more advanced IRONMAN competitions.

THE ECONOMICS OF IRONMAN TRIATHLON: LITERATURE REVIEW The majority of existing triathlon papers come from sports scientists and have been published in medical and physiology journals. Generally, these papers investigate the productivity development of athletes, distinguishing between gender and different age-groups. Particularly, IRONMAN data are used in applied research, because they offer a lot of relevant information (output, age, gender, etc.) that enable such investigations (Kahn, 2000). Picking this up, Sowell and Mounts (2005) were the first economists to use the individual data of 111 men and 91 women participating in the 1998 and 1999 IRONMAN Triathlon World Championships in Kailua-Kona, Hawaii. Expanding the model proposed by Fair (1994), Sowell and Mounts analyzed the human capital model and tested the relationship between an athlete’s age and his/her finishing time. As in Fair (1994), the objective of the study was to estimate the rate of output decline over years. Since athletes are typically recorded in age

488

THE SAGE HANDBOOK OF SPORTS ECONOMICS

groups (in five-year increments, male/female) it was possible to use the stochastic frontier estimation technique to determine the most efficient (the fittest) triathlete per age group. Different than Fair (1994), who used data from track and field (world records), Sowell and Mounts (2005) find that mens’ output declines faster than other studies implicated but that men’s rate of aging is slower than the rates computed for women. Using data from elite IRONMAN racers in Hawaii (the top 10, male/female), Leper (2008) analyzed the performance of 400 prize money-ranked athletes between 1988 and 2008. Specifically, he examined whether male and female athletes made output improvements (i.e. faster finishing times) and whether the magnitude of the gender difference became smaller among the three disciplines (3.8k swim, 180k cycling and 42k run). He used ordinary least squares (OLS) regression and computed the percentage difference in time between female and male elite athletes for the three endurance legs separately and collectively. His findings indicate that both sexes improved their overall productivity during the 1980s but stayed relatively flat within the 1990s until 2007. While both genders increased bike performance, women reduced the running gap (compared to men), while the swim difference among the genders remained more or less small. The latter finding is due to females’ higher share of body fat. Performing an observational field study of nonprofessional male IRONMAN athletes (n=83) in Zurich, 2009 (IRONMAN SWITZERLAND 2009), Knechtle, Wirth and Rosemann (2010) tried to estimate the most relevant determinants of these athletes’ output. The aim of their study was to find the most important factor in predicting an age-grouper’s performance. The authors distinguished between three theories explaining an athlete’s productivity: physical characteristics (body fat, body mass, height), training volume and intensity in swimming, cycling and running (volume, speed) and pre-race experience and personal best times in marathon running. All things being equal, the authors displayed that high-intensity runs during training sessions as well as individual marathon (best) times best explained participants’ finishing times (dependent variable). Using data from eight IRONMAN races around the world between 1994 and 1999, Prinz (1999) analyzed the incentive effects of prize money provided by tournament theory. Different than Frick and Klaeren (1997), who used data from worldwide city marathons, the prize purse distribution among the top 10 triathletes is more equal and lower in IRONMAN than in marathon races. Nevertheless, results suggest that triathletes put more effort into events with a higher purse. In a

similar vein, Wies (2016) collects prize money data from 62 IRONMAN events between January 2014 and October 2015 (including the Hawaii race). Overall, 462 elite athletes performed in the prize money ranks. Wies extended the list of the traditional independent variables (prize money and prize spread) by controlling idiosyncratic factors such as courses (slow and fast track) and wetsuit bans. His findings are still robust compared with the results by Prinz (1999) and Frick and Klaeren (1997). Obviously, the course influences the finishing times, but money still does motivate athletes to put in (marginally) more effort. With the help of prospect theory (Kahnemann & Tversky, 1979), Wicker, Maxcy and Prinz (2016) analyzed the motives of recreational triathletes participating in a long-distance race. Given high opportunity costs, equipment expenditures and especially the pain associated with finishing an IRONMAN race (McCarville, 2007), the authors scrutinized whether such a torture can be explained rationally. Using an online survey, they collected information from 206 non-elite IRONMAN athletes in 2015 and asked about their rewards for the ‘torture’. Specifically, respondents were interviewed with respect to their feeling when crossing the finish line and then how they felt some weeks after ‘being an IRONMAN’. The authors documented that mental torture during the race harms the feeling of happiness but that suffering during the race and then finishing the torture increases happiness. The following studies are not concerned with the determinants of athletes’ output. These papers examine the economic impact attached to organizing an IRONMAN competition. Jose Raya (2012) examined the length of stay of triathletes during the former Challenge-Family event in Calella/ Barcelona 2009 (since 2014 IRONMAN is the new organizer). Using data from a survey issued by the local county government, Raya estimates Cox hazard rate models in order to find out the most relevant covariates explaining IRONMAN participants’ duration in Calella. The results show some socio-economic relevance. On average, triathletes stay 4.34 days in the neighborhoods of Calella. The higher the number of companions brought to the event, the higher the expenditures spent and only Spanish athletes will stay longer. Adversely, the higher the hotel is ranked (i.e. more stars), the shorter the period of stay, and age does not have a significant effect. Using daily airplane arrival data made available by Hawaii’s Department of Business and Tourism, Baumann, Matheson and Muroi (2009) determined the net change in tourism for three events being held in Hawaii between 2004 and 2008 – the NFL Pro Bowl, the Honolulu Marathon and the IRONMAN Triathlon, Kona). Similar to other authors, they

IRON(O)MICS: The Market for Long-Distance Triathlon

conduct some ARIMA models to assess the impact of incoming and outgoing people to/from Hawaii. They found that all three events generate positive net impact effects. However, while the Pro Bowl reaches the highest number of extra tourists (~6,000), the IRONMAN Triathlon only generates between 1,900 and 3,500 additional tourists (mainly age-grouping participants). Nevertheless, the IRONMAN figure of 1,900–3,500 people is highly lucrative for the state of Hawaii because many participants arrive well in advance (seven days prior to the race) and thus spend (holiday) money. Moreover, unlike the Pro Bowl, the IRONMAN Hawaii is not subsidized by the government.

DETERMINING THE DEMAND SIDE OF THE LONG-DISTANCE MARKET: AN EMPIRICAL ANALYSIS Assuming that the organizer is interested in maximizing his demand and that athletes are more interested in finishing the race, advanced races are supposed to be less demanding than tougher races. As illustrated in Table 47.2, the 39 races vary considerably regarding their level of difficulty, prize money and participants. Therefore, the number of starters, the number of finishers as well as the sold-out rate should be affected by the idio­ syncrasies of the respective 39 IRONMAN competitions. The following analysis is based on a data set that covers 150 IRONMAN races between 2011 and 2015 (see Table 47.2). Overall, more than 300,000 athletes participated in these races but not all races were sold out. While the average number of athletes is around 2,000, the race in Frankfurt and the races in Florida, Barcelona and Copenhagen are usually sold out within a couple of days, comprising approximately 3,000 runners. On the other hand, other prestigious contests (i.e. Lanzarote and Malaysia) are never in high demand since both races are among the most difficult races in the IRONMAN circus. Typically, triathletes in Lanzarote face strong headwinds and very hilly terrain on the bike ride (see Table 47.2), while the race in Malaysia is very hot and humidity is extremely high. To see whether a race-specific environment (location) determines the demand side of the IRONMAN market, (daily) average values of these 150 races are investigated. Specifically, two OLS models (STARTER and FINISHER) as well as one LOGIT model (SOLD-OUT) were estimated. Overall, the findings from this exercise does not reveal a clear picture. Not many independent variables influence the demand side of the

489

long-distance market. Since the number of registered (paid!) athletes is presumably the organizer’s most important income function, model one is the preferred estimation. Since athletes have exact information regarding course topography and swim location, they face uncertainty concerning humidity, air temperature, water temperature and wind conditions. The results in Table 47.5 illustrate that the ‘customers’ select themselves into easier races but that these choices are rather made randomly. For example, the number of starters is lower if wind conditions are expected to be bad (i.e. Copenhagen, Lanzarote). The same effect (also not significant) can be seen from elevation parameters (bike and run). Both variables are negatively sloped but insignificant. Athletes refrain from registration if the race location shows more climbing meters to overcome. This finding is particularly pronounced in model three. The logit (sold-out = 1/not sold-out = 0) displays that customers shy away from mountainous bike stages. Additionally, demand is decreased if athletes know they face ocean swimming since choppy water, current and shark attacks impede an athlete’s IRONMAN choice.5 SALT is the variable controlling for open-water ocean swims. This coefficient highlights that the demand is reduced by approximately 400 athletes if the location of the event is surrounded by the sea. Further, if the race is ex-ante known as a wetsuit illegal event6 (i.e. Malaysia, Cozumel, Hawaii), starters are not that prone to participate there (with the exception of Hawaii, of course!)

MANAGEMENT IMPLICATIONS Especially in the past years, triathlon longdistances races have become very popular around the world. Demand and supply increased and the IRONMAN organizer is constantly adding new events all over the world. However, since the number of participants is a constraint to a certain limit, the organizer needs to find and select the best spots. One strategic implication from the results documented in this chapter is to find new locations that increase the probability of finishing the race and to finish the race as fast as possible. Finishing, and finishing well, is in the interest of the age-group athlete, so consequently race directors might select a flat and wind-sheltered bike course because cycling is the most relevant discipline. Rival leagues, such as the Challenge family or the Ican series, might take this aspect into consideration when allocating new events. Since both rivals (Challenge and Ican) are less capitalized

490

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Table 47.5  The demand of IRONMAN races Variable

Model (1) OLS (STARTER)

Model (2) OLS (FINISHER)

Air-Temp

5.04 (0.32)+ −1.92 (−0.47)+ −10.00 (−1.31)+ 18.1 (0.56)+ −0.071 (−0.79)+ −0.46 (−0.83)+ 0.000 (1.60)+ −420.1 (−3.76)*** −276.4 (−1.36)+ 2.306 (3.46)*** 2.57* 0.15 / 150

0.002 (0.95)+ −0.000 (−1.25)+ −0.001 (−1.28)+ 0.002 (0.72)+ −0.000 (−0.65)+ 0.000 (−1.07)+ −0.000 (−6.71)*** 0.017 (1.23)+ 0.239 (0.95)+ 0.129 (1.75)* 7.50*** 0.18 / 150

Humidity Wind-Speed Water-Temp Elevation_Bike Elevation_Run Preis Salt Wet-Ban Const. F-Value R2 McFadden R2 N of cases

compared to the market leader (IRONMAN), this strategy is helpful in acquiring more age-group athletes and thus staying competitive.

Notes 1  Neumann et al. (2010) report that elite triathletes spent up to 40 hours of training a week. This results in 30k of swimming, 700k of cycling and 100k of running a week (tri.mag, No. 139, pp. 23). 2  For example, the IRONMAN race for Klagenfurt 2018 was already sold out within two days after the 2017 edition on July 2. The race in Roth is traditionally sold out within a couple of hours. 3  Generally, long-distance races always cover 3.8k (swim), 180k (ride) and 42.2 (run). But only the IRONMAN organizer is allowed to sell its product as an IRONMAN. Other races from rival series are long-distance races. 4  Triathlon’s Olympic inauguration took place at the Summer Games in Sydney, 2000. 5  Hyperventilation and panic is often recognized when running into the water (swim start). Kicks

Model (3) LOGIT (SOLD-OUT) 0.068 (0.94)+ −0.006 (−0.38)+ 0.011 (0.34)+ 0.024 (0.22)+ −0.000 (−2.48)** 0.001 (0.61)+ 0.000 (1.06)+ −0.077 (−0.19)+ −0.854 (−1.09)+ 0.634 (0.24)+ / / 0.056*** 150

are the rule not the exception. This is especially true for first-time athletes. Approximately 25% of race participants are rookies. In order to alleviate this problem, the IRONMAN organizer introduced the rolling swim start. Athletes self-select according to their estimated swim time and enter the water in a group of four individuals within an interval of five seconds. 6  Buoyancy wetsuits allow athletes to swim significantly faster and more securely. Because of overheating, wetsuits are banned if the water temperature exceeds 24.5 degrees. A day before the event starts, water temperature is measured by the organizer and the result is made public a few hours before the (swim) start.

REFERENCES Adler, K. (2017). Rattenrennen im Individualsport – eine Untersuchung in der Sportart Triathlon. Eingereichte Dissertation an der Sportwissenschaftlichen Fakultät der Universität Leipzig.

IRON(O)MICS: The Market for Long-Distance Triathlon

Baumann, R., Matheson, V., & Muroi, C. (2009). Bowling in Hawaii: examining the effectiveness of sports-based tourism strategies. Journal of Sports Economics, 10, 107–123. Britt, R. (2014). Sold out Ironman races: How many choose not to race? Retrieved September 19, 2015, from http://runtri.com/2011/01sold-outironman-races-how-many-chooose.html. Fair, R. (1994). How fast do old men slow down? Review of Economics and Statistics, 75, 103–118. Frick, B., & Klaeren, R. (1997). Die Anreizwirkungen leistungsabhängiger Entgelte: Theoretische Überlegungen und empirische Befunde aus dem Bereich des professionellen Sports. Zeitschrift für Betriebswirtschaft, 67, 1117–1138. Frick, B., & Prinz, J. (2007). Pay and performance in professional road running: the case of city marathons. International Journal of Sport Finance, 2, 25–35. Kahn, L. (2000). The sport business as a labor market laboratory. Journal of Economic Perspectives, 14, 75–94. Kahneman, D., & Tversky, A. (1979). Prospect theory: an analysis of decision under risk. Econometrica, 47(2), 263–291. Knechtle, B., Wirth, A., & Rosemann, T. (2010). Predictors of race time in male Ironman triathletes: physical characteristics, training or prerace experience? Perceptual and Motor Skill, 111(2), 437–446. Laursen, P., & Rhodes, E. (2001). Factors affecting performance in an Ultraendurance Triathlon. Sports Medicine, 31(3), 195–209. Leper, R. (2008). Analysis of Hawaii Ironman performances in elite triathletes from 1981 to 2007. Medicine and Science in Sports and Exercise, 40(10), 1828–1834. McCarville, R. (2007). From a fall in the mall to a run in the sun: one journey to Ironman triathlon. Leisure Science, 29, 159–173.

491

Neumann, G., Pfützner, A., & Hottenrott, K. (2010). Das große Buch vom Triathlon. Aachen: Meyer & Meyer. Prinz, J. (1999). Performance-pay and incentive effects in tournaments. Copenhagen: Copenhagen Business School. Prinz, J. (2017). Whoever finishes first, we’ll call him the Ironman. In Kirmße, S. & Schüller, S. (Eds.), Standortattraktivität und Nachfrage von Ironman – Rennen. Managementreihe des zeb. Frankfurt am Main: Fritz Knapp. Raya, J.M. (2012). Length of stay for triathlon participants in the Challenge Maresme-Barcelona: a survival approach. Journal of Sport and Social Issues, 36(1), 88–104. Sowell, C., & Mounts, S. (2005). Ability, age and performance: conclusions from the Ironman Triathlon World Championship. Journal of Sports Economics, 6(1), 78–97. Wicker, P., Hallman, K., Prinz, J., & Weimar, D. (2012). Who takes part in triathlon events? Application of lifestyle segmentation to triathlon participants. International Journal of Sport Management and Marketing. Special Issue: Consumer Behaviour in Sports, 12, 1–24. Wicker, P., Maxcy, J., & Prinz, J. (2019). Happiness as a reward for torture: is participation in a long distance race a rational choice? Manuscript Sporthochschule Cologne. Wicker, P., Prinz, J., & Weimar, D. (2013). Big spenders in a booming sport: consumption capital as a key driver of triathletes’s sport-related expenditure. Managing Leisure, 18, 286–299. Wies, S. (2016). Preisgeld als Anreizeffekt im Profisport Triathlon. Eine empirische Untersuchung der Preisgeldstruktur bei Ironmanveranstaltungen. Duisburg, University of Duisburg-Essen.

48 NASCAR Economics Peter von Allmen

Compared to other areas of sports economics, the literature related to auto racing, and more specifically NASCAR, is relatively new. While one can easily find publications that number into the hundreds in most major North American sports and European football that date back well over 50 years, most research on NASCAR has occurred in the last 20 years, with a significantly smaller total volume of work. Given the wealth of opportunities to test hypotheses in the areas of labor supply and incentives, economic impact, demand and others, as well as the prominence of NASCAR on the American sports landscape, this is surprising. Though less-so elsewhere, NASCAR is very popular in the United States. This is particularly true in specific regions (e.g. the southeast), where some local economies are driven (no pun intended) by the auto racing industry. While the relative scarcity of published work may lead to frustration among those in search of well-defined questions with settled answers, it represents a fertile area for new research. This chapter places the economics of NASCAR racing in the larger context of sports economics, discusses examples of the existing literature in those areas in which substantive work has been published and notes those areas in which the greatest opportunities lie for future work. Due to space limitations, no attempt is made to discuss all published work

related to NASCAR economics. Unless otherwise noted, NASCAR refers to the highest level of NASCAR stock car racing competition, currently known by the title sponsor name Monster Energy NASCAR Cup Series or simply the NASCAR Cup Series. The next section contains a brief overview of the structure of the industry and the nature of demand, providing context for the remainder of the chapter.

OVERVIEW Though, as the saying goes, auto racing began the day that the second car was built, NASCAR racing began in 1948. Since that time, NASCAR has expanded to include a wide variety of racing types at many levels. Thus, while in the common vernacular, the highest level of stock car racing is simply known as NASCAR, its full designation is the Monster Energy NASCAR Cup Series.1 Other prominent series include the Xfinity Series and the Camping World Truck Series. It is not uncommon for races in all three series to be held on the same weekend at the same track. Both the Xfinity and Camping World Series serve as training grounds (not unlike minor leagues) for drivers

NASCAR Economics

aspiring to compete in the NASCAR Cup Series, but unlike minor leagues in sports such as baseball, NASCAR Cup series drivers sometimes compete in both Xfinity Series races and Cup races on the same weekend. Unless otherwise noted, this chapter is limited to discussion of Cup Series competition and is simply referred to as NASCAR or the NASCAR Cup Series. Interestingly, and likely not coincidentally, NASCAR was founded around the same time that other professional sports began to stabilize into well-formed leagues in the US. The Daytona 500, the first 500-mile race and still the signature event in the NASCAR season, was first run in 1959. Since then, there have been a number of changes to the length of season, number of cars permitted per race and contest rules that lead to a seasonending champion that are discussed in the following section. Briefly, the season currently consists of 36 races (not including special events such as all-star weekend and the two Daytona qualifying races). The first 26 of these races are referred to as the regular season and the last 10 races constitute the playoffs (previously known as ‘the Chase’), during which the 16 cars in playoff contention are serially eliminated until four cars remain for the final competition at the last race of the season. Unlike all other major professional sports in the US, participation in the playoff races is not limited to those teams (cars) still in contention. Rather, NASCAR fields a full complement of cars for each race. NASCAR crowns both a regular season and a playoff season champion. In addition to monetary prizes earned by final position at the races’ end, drivers compete for both single race and season-long rewards by accumulating points at each race. Drivers are employed by teams that in turn are run by larger motorsports organizations that may enter multiple cars in any given race. As of 2016, the number of cars per race was reduced from 43 to 40, with guaranteed entry to each of the 36 chartered teams (similar to a franchise) with four spots left for open teams.2 NASCAR events are held across the country, though a majority of races are held in the southeastern US. Teams do not have ‘home tracks’ and as such there is no parallel to the home team, as in other sports, as it relates to the teams or to the fans. The majority of tracks used to host races are owned by two corporations: International Speedway Corporation (ISC) and Speedway Motorsports International (SMI). NASCAR events have traditionally boasted very strong attendance numbers across all races (though official figures are no longer released). In recent years there are reports that attendance has dropped markedly. In response to the live attendance decline, a number of tracks have reduced seating capacity by

493

removing sections of the grandstands. For example, attendance at the Brickyard 400 has reportedly dropped from over 200,000 to estimates of as low as 50,000 in 2016.3 According to the Associated Press, television viewership has also suffered a steep decline, falling by over 40% between 2004 and 2016.4 Competing in the NASCAR Cup Series is an extremely expensive proposition, prohibitive for a single, independently-owned entrant. Instead, drivers are employed by teams, many of which have multiple entrants in a given race. For example, Hendrick Motorsports maintains four Cuplevel teams. As noted in the final section, the influence of multi-car teams on competitiveness remains a significantly underexplored area of the sport.5 As with other North American sports, most elements of NASCAR racing are privately held, making detailed analysis of team finances, the influence of sponsorship revenues and labor contracts difficult. The remainder of the chapter is organized by area of analysis. This includes contest design, demand studies, safety, economic impact and the labor market. The final section offers suggestions for future research and concludes.

CONTEST DESIGN As noted in the overview, NASCAR Cup races are both stand-alone contests and part of a seasonlong series. Each race represents a rank-order tournament, with the winner and subsequent positions determined by finish order. Drivers are awarded both cash prizes and points that are used to determine the season-long champion. Beginning in 2017, NASCAR, in an attempt to increase intrarace excitement, substantially revamped the prize structure. As data become available, these changes will serve as an important natural experiment as it relates not only to demand but also to driver and team behavior. As with all rank-order tournaments, race organizers attempt to construct a set of incentives that motivate participants to elicit an efficient level of effort such that the contest generates demand.6 Becker and Huselid (1992) were the first to analyze the impact of the NASCAR payoff scheme and found that drivers did respond to increases in the marginal prize.7 They also found weak evidence of increases in accidents as the prize spread increases. Von Allmen (2001) presented a detailed overview of the prize profiles in racing and their potential impact on behavior, noting that marginal

494

THE SAGE HANDBOOK OF SPORTS ECONOMICS

prize increments (both in terms of dollars and championship points) have historically been considerably smaller than in other high-level tournaments, such as professional golf and tennis.8 In that paper, I introduce three reasons why such small increments may be optimal in auto racing: the ‘sponsorship effect’, the ‘sabotage effect’, and cost implications. The first two in particular are related to the increased likelihood of a crash eliminating the car (and possibly others) from the race. Briefly, the sponsorship effect is the desire for cars to remain in the race because it increases on camera exposures, providing value for sponsors (whose name and logos appear on the cars). The sabotage effect is based on Lazear’s (1989) description of industrial sabotage, in which competitors may take actions that impact opponent’s chances of winning. In the case of racing, winnertake-all type prize profiles create the incentive for drivers to drive at the absolute limit of control and perhaps beyond, including pursuing reckless strategies that may cause other cars to crash. The cost-based argument is that highly non-linear profiles may lead to unbalanced competition as winners are rewarded with resources that provide a competitive advantage in future races. Von Allmen (2001) does not include empirical tests of these hypotheses beyond showing that crash rates were significantly higher in a small sample of winnertake-all reward races. Subsequently, a series of papers sought to test various elements of the arguments in von Allmen (2001) and whether or not increasing marginal prize profiles elicit increased absolute levels of performance in NASCAR racing. Depken and Wilson (2004) were the first to explicitly investigate the sabotage and the cost hypothesis. Their goal was not to test the sabotage hypothesis using race-level data but instead to discover if the conditions under which the sabotage hypothesis would logically occur are present in season-long performance and winnings data. They were able to show, using HHI, CR4 and CR8 measures of concentration for performance and winning, that the necessary (but not sufficient) conditions for the sabotage hypothesis to be supported are present. That is, the one-unit increase in the concentration of performance points results in a positive but less than proportional increase in season winnings. They did not, however, find much support for the cost hypothesis as their tests of whether increased concentration in winnings (Granger, 1969) cause increased concentration of future races were largely unsupported. Schwartz, Isaacs and Carilli (2007) also investigate the driver response to rewards but provide a richer theoretical and empirical construct to do so by assuming that drivers fall into two categories:

skilled and unskilled. Because, they posit, the unskilled driver will lose and receive the lower payout in the absence of high-risk behavior (that is not required of the high-skill driver) in order to earn the higher wage, they must engage in highrisk maneuvers (e.g. more dangerous passes). In contrast, high-skill drivers can increase effort with less impact (sabotage) on other drivers because they ‘…impose less cost on themselves, and on others, given the same level of effort’ (Schwartz et al., 2007, p. 636). Their results support the sabotage hypothesis as less-skilled drivers are involved in more accidents. Using a much larger data set than previous authors, Frick and Humphreys (2011) test whether the existing prize profiles generate incentive effects consistent with Rank-Order-Tournament (ROT) theory. Using data from over 1,000 races over a period of 34 years, they show that, particularly in the top five positions, average winnings are strongly non-linear, with the winner earning on average 53% more than the second-place driver, who in turn earns on average 24% more than the third-place driver, though after the first few positions, the profile begins to flatten considerably. Using both the inter-quartile range and the standard deviation, they find that the greater the dispersion of prize money earned, the faster the average speed of the winning car. Per-race rewards, as Frick and Humphreys (2011) clearly show, are non-linear, at least at the top of the distribution. Yet the non-linearity is much less pronounced than in golf, wherein the first-place finisher receives 66% more than the second-place finisher, who receives about 58% more than the third-place finisher, and much less pronounced than professional tennis, wherein prize money at the US Open, for example, falls by half for the runner up, in half again for the semi-finalists and nearly in half again for quarterfinalists. The results of Frick and Humphreys’ (2011) empirical model indicate that drivers respond to larger prize spreads by driving faster. Thus, their contribution is important as they show that contest organizers can increase speeds with larger prize spreads. Yet this does not fully answer the question of whether the distributions are set optimally, i.e. why would organizers not offer prize distributions that are more non-linear, as in golf and tennis? Based on the significantly higher crash rates from the nearly all-or-nothing races, known as ‘The Winston’, noted in von Allmen (2001), it may be that drivers do respond to the relatively modest rewards with faster driving and that a more highly skewed reward distribution would lead to negative consequences related to cost, sponsorship problems and sabotage. It could also be that, as Groothuis, Groothuis and Rotthoff

NASCAR Economics

(2011) suggest, the structure of the reward profile is considerably more non-linear than individual race data suggest once one includes the value of sponsorship and that race organizers do not see wisdom in making rewards even more non-linear through the per-race prize awards. Fortunately for economists interested in the relationship between rewards and performance, NASCAR continues to evolve and make changes to the contest design. As discussed in the next section, such changes may generate the natural experiments that help to answer the question.

CHANGES TO CONTEST DESIGN OF THE SEASON-LONG CHAMPIONSHIP In addition to the per-race reward structure, the design of the season-long contest for the overall championship has also drawn the attention of economists, in part due to several major changes in contest design. Prior to 2004, NASCAR drivers were able to score well in the final season standings simply by completing a large number of laps (as opposed to winning many races). The existence of this apparent flaw in the contest design was brought sharply into focus in 2003 when Matt Kenseth won the overall points championship despite winning only one race (Rishel, Baker & Pfitzner, 2015). In an attempt to increase the incentive to win races, NASCAR instituted ‘The Chase’, which later became known as the Chase I era as the rules were again modified in 2014. In the Chase I era, the first 26 races of the season were known as the regular season and the final 10 races were dubbed the playoffs. After the first 26 races, only the top 10 drivers (with minor modifications, including expanding the field to 12 drivers over the next several years) would be eligible to win the season championship and points were reset to reduce the difference between these finalists to increase the premium on winning. Pfitzner, Glazebrook and Rishel (2014) test the impact of the modifications within the initial Chase format and find that they did not alter driver incentives. The subsequent study by Rishel, Baker and Pfitzner (2015) indicates that the Chase I era design did not produce the desired effect, as consistency still had the largest impact on points accumulation. O’Roark, Wood and Demblowski (2012) found evidence of additional problems with the Chase format. Their results show that during the Chase period (the final 10 races) the number of accidents increases among both the Chase and non-Chase cars. With greater stakes available to winners, this can be interpreted as a

495

negative consequence of a steeper reward profile (and support of the sabotage hypothesis). Perhaps in recognition of the need to increase the incentives to win, NASCAR again made significant changes to the championship format in 2014, creating an elimination playoff system in which 12 of the 16 drivers who qualified for the Chase were systematically eliminated over the ensuing nine races, leaving four to compete in the final race with only the order of finish determining the championship (i.e. no laps led or other bonuses awarded). To date, no known empirical studies exist of the impact of the 2014–2016 contest rules. In 2017, NASCAR made even more significant changes to the structure of the season-long points competition designed to increase effort throughout the course of each race. Each race now has three distinct stages that drivers will compete to ‘win’. Each of the first two stages will comprise approximately 30% of the race with the final stage making up the remainder. In each of the first two stages, the top 10 drivers will earn bonus points based on their order of finish and the winner receives one postseason bonus point. The winner of the final stage (the conclusion of the race) wins the race and receives 40 points. Points are awarded in descending fashion from second (35 points) through 40th in one-point increments. Race winners are also made eligible for the postseason (the final segment of races). Points are no longer awarded for leading laps. At the conclusion of the regular season drivers in the top 10 receive bonus points in descending fashion (15, 10, 8…1) towards the postseason.9 Once the postseason begins, drivers are eliminated, as in the previous Chase format, until the final four in the last race. The new format puts a premium on driver effort throughout the race and over the course of the season. The 2014–2016 Chase system as well as new system for 2017 and beyond will be well worth empirical investigation once the data become available. Perhaps most important will be to consider the effect of the new system on fan demand given that the ultimate test of a contest design in a forprofit environment is whether fans find it appealing.

DEMAND STUDIES The demand for NASCAR Cup races is frequently purported to be among the highest of all North American sports, perhaps eclipsed only by the NFL.10 As noted in the introduction, however, there is some evidence that NASCAR’s grip on the interest of the American sports audience may be slipping. Cheney (2017) reports that between 2007 and 2014, 11 different tracks reduced their

496

THE SAGE HANDBOOK OF SPORTS ECONOMICS

seating capacity by as much as 72,000. As such, there is likely much to be learned from additional study of consumer demand. To date, there have been just two published empirical investigations of the demand for live attendance or television viewership: von Allmen and Solow (2011), who studied television demand, and Berkowitz, Depken and Wilson (2011), who studied both television and live audience demand. Von Allmen and Solow (2011) used Nielsen data from 267 races between 2001 and 2009 to investigate the often-asked question of whether fans tune in to watch the racing or to see the crashes. Their results across three different specifications that span the entire time period consistently show that crashes increase television demand. They also find that NASCAR and NFL broadcasts are substitutes but could find no evidence that broadcasts of other sporting events, such as triple crown races, PGA Major events or playoff baseball games, had any impact on the NASCAR audience. They also found that larger purses attract larger audiences but that uncertainty of outcome was not a significant determinant of ratings. However, when the sample was restricted to the Chase era, uncertainty of outcome was a significant determinant of viewership as was the speed of the pole winner (i.e. higher anticipated race speed). Unsurprisingly, ratings were significantly lower for weekday races (held due to rainouts of the weekend date) in all specifications. Berkowitz, Depken and Wilson (2011) model both live attendance and television audience using data from three seasons in the Chase I era (2007– 2009).11 Their results for both television ratings and viewership are largely consistent with the von Allmen and Solow (2011) specification restricted to the Chase era. They find that uncertainty of outcome increases ratings and viewership and that races held on Sundays or evenings and superspeedway events draw larger audiences. They also find that televised races are sensitive to substitute broadcast sporting events, though the impact of substitution declines over the course of the season. In the live attendance model, Berkowitz et  al. (2011) find that greater uncertainty of outcome and superspeedway events increase attendance while, not surprisingly, events at road courses where fans can only see a small portion of the track from a single vantage point, suffer significantly lower attendance. The sample periods used by both von Allmen and Solow (2011) and Berkowitz et al. (2011) end in 2009. Since that time, there have been several changes to the structure of the in-race and season-long contest (described in the previous section) as well as the reported but not empirically verified decrease in attendance. As such, additional research on the level and determinants of demand would be particularly valuable (discussed further in the conclusion).

SAFETY As described above, an auto race is structured similarly to many other forms of racing (such as foot races) in which all competitors compete simultaneously against the entire field. Unlike a foot race, however, the ever-expanding envelope of engine, tire and chassis technology has meant systematic increases in the speed of the contest and attending risk of catastrophic accidents. While NASCAR makes substantial efforts to safeguard drivers (discussed in detail below), it is not possible to entirely mitigate risk, particularly if drivers respond to increased safety measures with more aggressive driving. As is the case with changes to the contest and the inherent driver incentives described in the previous section, changes to contest rules represent natural experiments and provide opportunities for economists to study the response of drivers and teams. There are two fundamental questions regarding issues of safety in auto racing. The first is whether drivers respond to increases in safety measures with what Peltzman (1975) described as offsetting behavior (now known as the Peltzman effect), wherein drivers increase risk-taking as the odds of injury from a crash decrease. The second is whether the risk of accidents is an important component of demand. These are, of course, inter-related questions. From the driver and team standpoint, accidents that lead to elimination or, worse yet, injury or death, are clearly undesirable. However, if fans derive utility from seeing accidents and drivers are more willing to engage in high-risk tactics as safety improves, then increases in the safety of the cars ultimately increase the demand for the sport. In the first published study to address safetydriven changes in contest design, O’Roark and Wood (2004) evaluated the determinants of racetrack safety following the introduction of restrictor plates at Talladega and Daytona – the two ‘superspeedways’ – in 1987. Restrictor plates reduce the flow of air to the engine, reducing the top speed. Prior to the introduction of restrictor plates, cars would routinely reach speeds in excess of 200 mph, with significant risk to both drivers and spectators. As O’Roark and Wood (2004) point out, however, while the plates did decrease top speeds, which increases fan safety, it also decreased the variance in speed. Reduced variance in speed may increase safety on public roads (see O’Roark & Wood, 2004, for a discussion of this literature), but decreased variance in a race on a closed track may have just the opposite effect due to the resulting congestion. Their empirical models test for both changes in the incidence of wrecks and injuries. They find that restrictor plate races, wherein the

NASCAR Economics

variance in speed is reduced, have significantly more accidents. Thus it is congestion rather than speed that causes accidents. In an interesting contrast, they find that injuries are unrelated to the use of restrictor plates, supporting the hypothesis that it is the speed that results in injuries rather than congestion. In a separate analysis, O’Roark (2011) investigated the impact of mixing the skill level of drivers and also tested for the impact of restrictor plates and again found that restrictor plates increase crash rates. Anecdotally, both this result and the controversy surrounding it remain. In the 2016 race at Talladega, 35 of the 40 cars were involved in crashes and drivers seem either resigned to the plates or openly disdainful.12 Sobel and Nesbit (2007) construct a model designed to more directly test the Peltzman effect. They test the relationship between the number of accidents and the probability of an injury given that there is an accident. Using two different samples and a variety of specifications, they find significant support for the Peltzman effect with an elasticity of approximately .2 across all specifications: ‘Thus, at 10% improvement in automobile safety results in approximately a 2% increase in reckless driving’ (p. 79). The result that increased safety increases accident rates is reinforced by the work of Pope and Tollison (2010), who found that mandating the use of the head and neck restraint system (HANS) led to a significant increase in laps run under caution (caution laps are essentially a pause in the race where all cars drive in a highly controlled and reduced speed formation while accident debris is cleared). It is worth reiterating here that an auto race is a very different sort of contest than, say, a golf tournament, where no defensive actions are permitted and the chances of injuring another player are almost non-existent. In auto racing, drivers may make both offensive and defensive strategic maneuvers that lead to accidents. Accidents are almost by definition events wherein a driver loses control of her/his vehicle with significant potential for extensive negative externalities that involve other cars and drivers, as described in von Allmen (2001) and O’Roark et  al. (2012). Thus, if the Peltzman effect is real in the contest of NASCAR (and evidence seems to unanimously support this conclusion), the costs of the resulting aggressive driving are likely to be widespread. If future research in this area is focused on the external costs of aggressive driving, it may help to more fully understand the cost of the behavior. As noted in the introduction to this section, aggressive driving and the resulting crashes have two inter-related effects. To the extent that advances in safety measures increase aggressive driving and cause accidents and injuries,

497

NASCAR should actively promote policies to reduce this behavior. If, on the other hand, fans view crashes as part of the allure of watching racing, they enhance demand and reducing their occurrence will leave fans less satisfied. As noted above, von Allmen and Solow (2011) indicate that in the pre-Chase era, aggressive driving is positively related to the size of the television audience with an additional crash per race increasing viewership by about 6%. Interestingly, they find that the effect disappears in the Chase era, indicating that the change to the structure of the tournament had a significant impact on driver behavior. Given the very substantial changes in the design of the championship tournament as well as individual races, it would be well worth revisiting this conclusion.

ECONOMIC IMPACT The economic impact of sporting events on local communities has long been the subject of study. When considering economic analysis of the impact of sports, it is useful to divide the analysis (and analysts) along two dimensions. The first is to separate out what Coates and Humphreys (2008) refer to the ‘promotional literature’ (sometimes also called ex-ante or ‘booster’ studies), which are usually sponsored by members of a local community with a vested interest in a positive outcome, from ex-post analytical work, which are usually done by academic economists with no particular interest in a conclusion that either supports or does not support the merits of the activity.13 The second useful distinction in economic impact studies is to separate those that consider the economic impact of continuous hosting of a team in an organized league, e.g. the benefit to Los Angeles of luring the National Football League’s (NFL) Chargers to move there from San Diego, from so-called mega-events. Mega-events are onetime or infrequently held events, such as the NFL Super Bowl, the World Cup or the Olympics. Such events are fundamentally different than a league in that, to be worthy of support from local taxes on economic development grounds, they must generate enough benefit from a single event or short series of events that they cover the contribution. Setting aside non-academic studies, the question is, are NASCAR events mega-events or long-term hosting arrangements? The answer is – something in between. Most tracks that host NASCAR events do so annually over the course of many years. In that sense, they are long-term arrangements. Yet from an economic impact standpoint, one or two weekends per year make the races much like a

498

THE SAGE HANDBOOK OF SPORTS ECONOMICS

mega-event. As stated in the introduction, all drivers (and many fans) travel from race to race each week of the season. At any given event, all are visitors and, as such, there is no ‘home team’. As Baade and Matheson (2000) note, that so many attendees come from outside the local community mitigates one of the most common problems with impact studies: substitution spending. Substitution occurs as consumers in a local economy simply attend one local event instead of another, such that actual net increased spending is significantly smaller than gross spending associated with the event. At the same time, it may exacerbate two well-known problems – the direct costs and negative externalities associated with as many as 100,000 or more visitors to a local community lacking the infrastructure to accommodate them and crowding-out, as otherwise visitors to an area stay away to avoid the crowds associated with race weekend traffic. Finally, unlike any other major team sport in North America, NASCAR as a product is ‘produced’ in part through research and development and construction of the vehicles. Much of this activity occurs in the Charlotte, North Carolina area, and is subsequently exported to events in the form of completed cars and ancillary equipment. Connaughton and Madsen (2007) find that the benefits to Charlotte of this activity was nearly $6 billion in 2005 alone. Perhaps in part because much of the product is produced in this single region, large-scale benefits related to individual tracks or races are more difficult to be shown conclusively. Baade and Matheson (2000) analyze the impact of the Daytona 500, which many regard as the equivalent to the Super Bowl for motor sports. Though the event appears to increase taxable sales increases of about $42 million per year, after accounting for reasonable estimates of substitution and crowding-out as well as leakages from the local multiplier, the impact would be considerably smaller. While the authors do not explicitly consider costs to the local community, they are surely significant. After paying expenses such as local police overtime, event security, waste removal and the like, the net impact could be negative. Bernthal and Regan (2004) use survey data to estimate the impact of two NASCAR race weekends at Darlington in the Pee Dee region of South Carolina and find a total impact of about $46 million and just over 900 jobs, though they acknowledge that by counting local spending as net new, they may overstate the impact. In addition, it appears that all hotel spending is counted as local and air travel was included in the total travel expenditures. To the extent that these dollars do not actually flow into the local economy, it would again overstate the impact and, as with Baade and Matheson (2000), costs are not considered. These concerns aside,

given the rural nature of the area surrounding the Darlington track, which would likely reduce substitution spending and increase net exports, racing as a mega-event in this context likely generates significant benefits for the local economy. Coates and Gearhart (2008) use a panel data set that includes 52 different race tracks to test the public goods nature of tracks, noting that they may be negative or positive. Rather than attempting to measure expenditures as described above, they instead consider the impact of the presence of a track as well as specific race events on rental housing prices. Interestingly, their results indicate that tracks and events are jointly significant (positive) for urban areas, though the impact appears to be driven at least in part by non-race events (e.g. on-track driving schools). In non-urban areas, they generally find either no impact or negative impacts. Given that this result stands in contrast to Bernthal and Regan (2004), who argue that Darlington’s rural location magnifies the impact, future impact analyses should carefully include accounting for the costs imposed by the events as well as the benefits.

LABOR MARKET Racing labor markets (i.e. the supply of and demand for drivers) have received the least attention in the broader NASCAR literature. This is somewhat surprising given the many areas of potential interest and available data. For example, on the supply side, auto racing is a sport that requires extensive human capital gained from experience, yet not a single study exists that investigates how human capital is acquired and the mechanisms by which it leads to promotion from one series to another, more prestigious one. On the demand side, labor demand is, as always, derived from output demand and so ultimately by consumers. As noted above, auto racing depends heavily on sponsorship funding. Such funding is likely to be easier to obtain when racing teams have popular drivers who can serve as the ‘face of the franchise’. Rotthof, Depken and Groothuis (2014) use value of time on camera (VTOC) data to investigate the role of performance and celebrity status. They find that winning, leading laps, previous wins, finishing races and being the son of a NASCAR driver all contribute to VTOC. Groothuis and Groothuis (2008) note that nepotism creates the potential to exacerbate the influence of consumer discrimination in hiring decisions. While they find no significant evidence of nepotism in career duration, they do

NASCAR Economics

find intergenerational effects that may reflect consumer discrimination as branding (i.e. driver name recognition). Beyond the scope of their study, consumer discrimination may be evident through the usual transmission mechanism of consumer (fan) preferences for specific worker types. Racially, the NASCAR fan base is quite homogeneous. A Nielsen study indicated that the NASCAR audience is 94% white and only 2% black.14 The racial composition of the labor force is similar, if not even more extreme. Only three African-American drivers have ever competed in a Cup Series race. It is reasonable to assume that these proportions are interdependent, though economists have yet to explore the direction of causality. Interestingly, the parallel between fans and drivers is not consistent across all characteristics. About 40% of NASCAR fans are women, though female drivers are nearly as scarce as blacks.15 Only 16 women have ever run a NASCAR Cup race (i.e. raced at the highest level) and only two since 2000. Other than Danica Patrick, who has started more than 165 races and is still active, only one other woman has started more than five regular series Cup races since 1950.16 NASCAR is clearly aware of the need to explore and address diversity issues. In 2004, NASCAR instituted the ‘Drive for Diversity’ program to encourage women and minorities to enter and develop skills in the sport of racing. They also engaged the services of well-known diversity advocate Richard Lapchick (Director of the Institute for Diversity in Sport) to deliver diversity training programming.17 Given the differences in the fan base across race and gender lines that stand in contrast to the remarkably homogeneous characteristics of the drivers, the Drive for Diversity and internal training programs deserve economists’ attention as they represent opportunities to test hypotheses related to the efficacy of such programs to influence the labor market.

CONCLUSION AND SUGGESTIONS FOR FUTURE RESEARCH Research in sports economics as applied micro­ economics can serve at least two important purposes. First, as first noted by Kahn (2000), it represents an opportunity to test hypotheses from the larger discipline that are difficult to validate elsewhere due to data limitations. Second, empirical investigations of various sports help economists and practitioners to better understand the nature of the industry. Within these areas, the economics of NASCAR is fertile ground for new

499

research, particularly given the dramatic changes to the contest structure that we do not observe in other sports. There are three areas in particular in which new research is most needed. First, we still have much to learn about how drivers (and teams) respond to incentives. Existing research indicates that while the per-race prize profile is flatter than other tournament sports, such as golf and tennis, drivers do appear to respond to those differences. What we do not yet fully understand are the influences of season-long rewards on individual race behavior. The as yet un-studied second era of the Chase (2014–2016) and the new 2017 rule changes provide a series of natural experiments that can help to untangle the various influences of cash prizes, championship points and sponsorship revenue as well as driver risk preferences. Second, there is a clear need for additional work to explain the demand for NASCAR. Changes to the contest structure (presumably with fan interest in mind) along with the reported decrease in fan interest, call for a reconsideration of television and live attendance demand determinants. Finally, racing labor markets have gone almost completely unexamined. There is much to learn about how drivers accumulate human capital and move from one series to another. The same is true for the relationship between fan preferences (i.e. consumer discrimination) and driver characteristics and the impact of NASCAR incentives to increase diversity.

Notes  1  The name of this series has changed several times as the title sponsorship changes. Previous names include Winston Cup, Nextel Cup and Sprint Cup.  2  George Diaz. NASCAR charter system will cut Cup field from 43 to 40 cars. Orlando Sentinal, February 9, 2016. At www.orlandosentinel. com/sports/nascar/os-nascar-sprint-cup-charterstarters-0210-20160209-story.html (accessed May 1, 2017).  3  Gregg Doyel. 2016 Brickyard 400 – a race no one saw. IndyStar, July 24, 2016. At www. indystar.com/story/sports/columnists/greggdoyel/2016/07/24/doyel-2016-brickyard-400-raceno-one-saw/87507680/ (accessed May 1, 2017).  4  Michael Casey. New museum tells history of New England car racing history. Associated Press, June 7, 2017. At https://apnews.com/322fb66d108d4 41094b49ea24d1c2ec/ (accessed June 7, 2017).  5  A notable exception is Depken and Mackey (2009).  6  See Sherwin Rosen (1986) for more on the purpose and effect of non-linear rewards, and O’Keefe, Viscusi and Zeckhauser (1984, p. 29), who describe four distinct circumstances under which

500

THE SAGE HANDBOOK OF SPORTS ECONOMICS

tournament style rewards are appropriate, including: the utility of the contest itself, indivisible rewards, high monitoring costs and risk reduction. 7  Though not discussed here, Becker and Huselid (1992) also investigated incentives and performance in IMSA racing. 8  The small marginal increase in points received for moving up one position has led to speculation that drivers and their teams prefer conservative driving in order to stay in the race simply to earn points through completed laps, an assertion supported by the work of Rishel and Pfitzner (2006), who found that season points were strongly influenced by laps completed independent of performance. 9  Point descriptions are from NASCAR.com, FAQ for NASCAR’s 2017 race format enhancements (at February 27, 2017). 10  Groothuis, Groothuis and Rotthoff (2011) cite several sources stating that the demand for NASCAR compares favorably with all North American sports and at least at one time had higher television ratings than all but the National Football League. 11  Berkowitz, Depken and Wilson (2011) note that attendance figures are taken directly from NASCAR as opposed to track owners and may be unreliable as they are always rounded to the nearest thousand. 12  Jordan Bianchi. Talladega the latest evidence that NASCAR restrictor-plate insanity needs to stop. SBNation, May 1, 2016. At www.sbnation. com/nascar/2016/5/1/11554082/2016-nascartalladega-results-recap-restrictor-plates (accessed May 2, 2017). 13  No attempt is made to review the promotional or ‘booster’ literature in this chapter. 14  Jonathan Jones. Bubba Wallace is best hope for fulltime black driver in NASCAR Cup series. Charlotte Observer.com, May 22, 2015. At www.thoughtco. com/bubba-wallace-biography-2471774 (accessed May 31, 2017). 15  Tori Petry. NASCAR not just for the boys. ESPNW. com, August 20, 2012. At www.espn.com/espnw/ news-commentary/article/8284577/espnw-nascarnot-just-boys (accessed May 31, 2017). 16  Female NASCAR Drivers across the Three National Series. ESPN.com, May 15, 2017. At www.espn. com/jayski/stats/story/_/id/19054896/femalenascar-drivers (accessed May 31, 2017). 17  Jones (2015), in note 14.

REFERENCES Baade, R.A., & Matheson, V. (2000). High octane? Grading the economic impact of the Daytona 500. Marquette Sports Law Review, 10(2), 401–415.

Becker, B.E., & Huselid, M.A. (1992). The incentive effects of tournament compensation schemes. Administrative Science Quarterly, 37, 336–350. Berkowitz, J.P., Depken, C.A., & Wilson, D.P. (2011). When going in circles is going backward: outcome uncertainty in NASCAR. Journal of Sports Economics, 12(3), 253–283. Bernthal, M.J., & Regan, T.H. (2004). The economic impact of a NASCAR racetrack on a rural community and region. Sport Marketing Quarterly, 13, 26–34. Cheney, P. (2017). The rise and fall of NASCAR: Why racetracks are removing hundreds of thousands of seats. The Globe and Mail. January 5, 2017. At https://www.theglobeandmail.com/globe-drive/ adventure/red-line/the-rise-and-fall-of-nascarwhy-tracks-are-removingseats/article27211911/. Coates, D., & Gearhart, D. (2008). NASCAR as a public good. International Journal of Sport Finance, 3(1), 42–57. Coates, D., & Humphreys, B.R. (2008). Do economists reach a conclusion for sports franchises, stadiums, and mega-events? Economic Journal Watch, 5(3), 294–315. Connaughton, J.E., & Madsen, R.A. (2007). The economic impacts of the North Carolina motorsports industry. Economic Development Quarterly, 21(2), 185–197. Depken, C.A., & Mackey, L. (2009). Driver success in the NASCAR Sprint Cup series: the impact of multi-car teams. North American Association of Sports Economists Working Paper Series 09-15. Depken, C.A., & Wilson, D.P. (2004). The efficiency of the NASCAR reward system: initial empirical evidence. Journal of Sports Economics, 5(4), 371–386. Frick, B., & Humphreys, B.R. (2011). Prize structure and performance: evidence from NASCAR. University of Alberta Working Paper 2011–12. Granger, C.W.J. (1969). Investigating causal relations by econometric models and cross spectral methods, Econometrica 37, 424–438. Groothuis, P.A., & Groothuis, J.D. (2008). Nepotism or family tradition? A study of NASCAR drivers. Journal of Sports Economics, 9(3), 250–265. Groothuis, P.A., Groothuis, J.D., & Rotthoff, K.W. (2011). Time on Camera: An additional explanation of NASCAR tournaments. Journal of Sports Economics, 12(5), 561–570. Kahn, L.M. (2000). The sports business as a labor market laboratory. Journal of Economic Perspectives, 14(3), 75–94. Lazear, E.P. (1989). Pay equality and industrial politics. Journal of Political Economy, 97(3), 561–580. O’Keefe, M., Viscusi, W.K., & Zeckhauser, R.J. (1984). Economic contests: comparative reward schemes. Journal of Labor Economics, 2(1), 27–56. O’Roark, J.B. (2011). Buschwhacking in NASCAR: mixing skill on the road in NASCAR’s junior circuit. Virginia Economic Journal, 16, 1–11.

NASCAR Economics

O’Roark, J.B., & Wood, W.C. (2004). Safety at the racetrack: results of restrictor plates in superspeedway competition. Southern Economic Journal, 71(1), 118–129. O’Roark, J.B., Wood, W.C., & Demblowski, B. (2012). Tournament chasing NASCAR style: driver incentives in stock car racing’s playoff season. Eastern Economic Journal, 38(1), 1–17. Peltzman, S. (1975). The effects of automobile safety regulation. Journal of Political Economy, 83(4), 677–725. Pfitzner, C.B., Glazebrook, T. & Rishel, T.D. (2014). Success in NASCAR: the statistical determination of points. SEDSI 2014 Conference Proceedings (pp. 115–124). Wilmington, NC. Pope, A.T., & Tollison, R.D. (2010). ‘Rubbin’ is racin’: evidence of the Peltzman effect from NASCAR. Public Choice, 142(3–4), 507–513. Rishel, T.D., Baker, E.W., & Pfitzner, C.B. (2015). Finishing or winning? The variables that impacted the NASCAR championship in the Chase I format. The Coastal Business Journal, 14(1), 26–41. Rishel, T.D., & Pfitzner, C.B. (2006). Success in NASCAR: a preliminary look at points and

501

money winnings. Virginia Economic Journal, 11, 29–39. Rosen, S. (1986). Prizes and incentives in elimination tournaments. American Economic Review. 76 (September 1986), 701–715. Rotthoff, K.W., Depken, C.A., & Groothuis, P.A. (2014). Influences on sponsorship deals in NASCAR: indirect evidence from time on camera. Applied Economics. 46(19) 2277–2289. Schwartz, J.T., Isaacs, J.P., & Carilli, A.M. (2007). To race or to place? An empirical investigation of the efficiency of the NASCAR points competition. Journal of Sports Economics, 8(6), 633–641. Sobel, R.S., & Nesbit, T.M. (2007). Automobile safety regulation and the incentive to drive recklessly: evidence from NASCAR. Southern Economic Journal, 74(1), 71–84. von Allmen, P. (2001). Is the reward system in NASCAR efficient? Journal of Sports Economics, 2(1), 62–79. von Allmen, P., & Solow, J. (2011). The demand for aggressive behavior in American stock car racing. In R.T. Jewell (Ed.), Violence and Aggression in Sporting Contests (pp. 79–95). New York: Springer.

PART VII

Future Research

49 Behavioral Economics and Sport Yu l i a C h i k i s h a n d B r a d R . H u m p h r e y s

INTRODUCTION Most economic theories employ simple, powerful models in which individual agents make choices to maximize some objective like utility or profit using all pertinent information available in an efficient manner. These models assume that agents have identical, standard, well-behaved, time-­ consistent preferences that are unaffected by trivial factors like the framing of the choice, perceptions of others, or information not relevant to the choice. These models have been widely applied to outcomes in sports markets, as well as other settings. A substantial body of evidence, from experiments and the field, indicates that outcomes often fail to match predictions because agents’ behavior differs in important ways from the assumptions underlying these basic economic models. Individuals frequently make decisions based on reference points or that reflect framing, exhibit time inconsistency, over-­estimate their own abilities, use heuristics rather than optimal approaches, and appear to be influenced by factors omitted from the basic economic models. Many decisions appear to reflect preferences that differ substantially from those assumed by the basic models. This chapter reviews research in sports economics that takes a behavioral economic perspective. The behavioral economic perspective provides new

insight into old problems in sports economics and also helps to explain many outcomes that cannot be easily explained using standard economic models. We critically review research that explicitly incorporates reference dependence and investigates the role played by biases reflecting non-standard preferences for athletes, referees, teams, and firms. The review is selective in that we do not attempt to comprehensively assess all areas of behavioral sports economics.

REFERENCE–DEPENDENT PREFERENCES AND LOSS AVERSION Prospect theory (Kahneman & Tversky, 1979) represents an important early development in behavioral economics. Prospect theory posits that individuals interpret losses and gains differently. The idea of asymmetric effects of expected gains and losses received considerable attention from empirical researchers in many contexts, including sports. Reference-dependent preferences (RDP) – the notion that people do not simply value outcomes but rather compare outcomes with some benchmark value called a reference point in the valuation process – are an important component of prospect theory. If an outcome exceeds the

506

THE SAGE HANDBOOK OF SPORTS ECONOMICS

reference point, an individual interprets it as a gain; outcomes not meeting the reference point are treated like losses. The second key aspect of prospect theory is loss aversion, which means that the utility function is steeper in the loss domain than in the gain domain, or that the absolute value of marginal utility from losses exceeds the marginal utility of gains. Loss aversion motivates the presence of the endowment effect – the fact that people are willing to pay less to acquire some good than they are willing to accept to part with the same good – which has been empirically verified in a number of settings. Reference dependence and loss aversion have been shown to influence a large number of decisions and outcome in sport. Pope and Schweitzer (2011) test whether professional golfers exhibit loss aversion in performance in competitions. They consider putts – the final shots players take to complete a hole – and how putting outcomes vary with a specific reference point. Their reference point is putts attempted for a score of par.1 A score of par is good, scores below par are very good, and scores above par are not good, making par a natural reference point. Pope and Schweitzer find that when golfers are under par (in the gain domain) they are significantly less accurate than when they are over par (in the loss domain). Golfers try harder to avoid a loss (score over par) than to obtain a gain of the same magnitude which is consistent with loss aversion. Allen et al. (2017) analyze 9.5 million marathon finishing times to test the reference-dependent nature of marathon runners’ preferences. The data contains evidence of significant clustering of race finishing times near round numbers (3 hours, 4 hours, etc.). Clustering cannot be explained by explicit rewards such as qualifying for elite competitions like the Boston Marathon, peer effects, or institutional settings, which leaves the most probable explanation that round numbers act as reference points in running. Reference dependence and loss aversion play a critical role in fans’ decisions to attend sporting events. Attending a sporting event involves uncertainty, as the outcome of the game is unknown when the fan decides to attend. The sports economics literature assumes that fans prefer games with an uncertain outcome to games with a certain outcome, an idea called the uncertainty of outcome hypothesis (UOH). Coates, Humphreys and Zhou (2014) develop a consumer choice model to explain the role of the UOH in game attendance decisions. In this model, consumers have RDP about game outcomes, standard consumption utility from attendance, and ‘gain–loss’ utility that reflects differences between expectations and actual game outcomes as well as loss aversion. The UOH emerges as a special case with no RDP, but reference dependence and loss

aversion appear to explain more observed variation in game attendance than the UOH in many settings. This model also simplifies to a case where fans only care about home team wins, a version with considerable appeal. RDP and loss aversion appear to be important components in fans’ decisions to attend games. Several studies assume that fans have rational expectations (proxied by betting market data) about their home team’s expected game outcomes (wins and losses) and explore the relationship between game outcomes that violate these expectations and violent behavior. Card and Dahl (2011) investigate the relationship between family violence and the outcomes of US professional football games. Their key assumption is that wins and losses by the home team invoke emotional cues associated with gain–loss utility around a rationally set reference point. They find that ‘upset losses’ (games where the home team is defeated though betting market predicted it would win) increase family violence by 10% in the city home to the losing team. Losses when a close game was expected, and victories when losses were expected, do not have a significant effect on violent crime. Munyo and Rossi (2013) also study the effect of violations of expectations in a game on violent crime. They match data on robberies in Montevideo, Uruguay, with results of soccer matches played by the two most prominent Uruguayan teams. Munyo and Rossi improve on Card’s and Dahls (2011) methodology by isolating the impact of violations of expectations from the impact of game outcomes. They define two kinds of emotions generated by expectation violations: frustration (losses) and euphoria (wins) and use variation in these expectations while holding game outcomes constant. Their findings suggest that frustration is followed by sharp increases in violent crime, while euphoria is followed by reductions in violent crime. Dickson, Jennings and Koop (2016) further analyze the relationship between soccer match outcomes relative to expectations and domestic violence. They use data on the number of domestic violence incidents in the Strathclyde region of Scotland from January 2003 to October 2011 and data on Scottish Premier League football matches that involved two prominent Scottish teams, Celtic and Rangers, during that period (both teams are from Glasgow and are long-time rivals). The results suggest that derby matches, matches with both teams playing against each other, lead to increases in domestic violence of about 36% regardless of the outcome. Disappointing results (a loss when victory was expected) are not related to increases of domestic violence, except for matches played at the end of the season. Reference dependence and loss aversion (LA) represent important factors that explain a number of

BEHAVIORAL ECONOMICS AND SPORT

outcomes in sports, including the behavior of fans. Recent research also shows that RDP/LA affect decisions made by local governments when subsidizing the construction of new sports facilities (Humphreys & Zhou, 2015). RDP/LA models likely apply to many other sports decisions and outcomes, including team decisions about acquiring and paying players, and retaining coaches. Application of RDP/LA models to other sports outcomes appears to be a fruitful avenue for future research.

PRESENT BIAS Economic decisions with inter-temporal aspects, especially those involving current costs and future benefits, may be influenced by non-standard preferences over current and future outcomes. Present bias, or hyperbolic discounting, refers to preferences that give stronger weight to outcomes in the present or near future and weaker weight to outcomes in the distant future. Standard economic models assume identical discount rate between any two subsequent periods of time, no matter how far in the future those two periods are. A growing body of research examines the economic determinants of individuals’ participation in physical activity, which may reflect present bias. Walking, running, and home exercise constitute the most common types of individual exercise. The decision to exercise can be analyzed using standard economic approaches (Humphreys & Ruseski, 2011). Another common form of exercise involves fitness center or health club membership. This membership decision involves a contractual relationship between a business and individuals interested in exercising, an activity subject to selfcontrol problems like present bias. Fitness club membership and use decisions represent an ideal type of consumer choice to study in the context of self-control problems. Exercise involves current monetary and non-monetary costs (discomfort from exertion, muscle soreness) that generate disutility from exercise and future health benefits only realized after prolonged participation. Consumers must accurately forecast future health benefits and form expectations of future participation levels when making current exercise decisions, making current and future self-control crucial. A growing literature finds that fitness club membership contracts systematically exploit consumer self-control problems. DellaVigna and Malmendier (2006) analyze contracts and attendance decisions made by 7,752 members of three health clubs. The clubs offered multiple membership options, including automatically renewable monthly

507

contracts, paid-in-advance annual contracts expiring after one year, and pay-per-visit 10-visit passes. Actual attendance patterns reveal that choosing the monthly option cost members 70% more than the pay-as-you-go option, and 80% of members who chose monthly contracts would have spent less using the pay-per-visit pass. Many possible explanations exist for these decisions. DellaVigna and Malmendier (2006) posit that overconfidence, in the form of consumer naiveté about future visits that will actually be made, leads members to over-forecast future attendance, leading to suboptimal contract choices. Alternatively, club members could have time inconsistent preferences leading to systematic differences between expected future benefits and current costs that reduce the expected value of future benefits. For two consumption bundles, x and y, received at time t, an individual has time inconsistent preferences when she evaluates them differently at time t − 1 and at time t − k, where k > 1. This reflects a preference reversal, where x is preferred at time t − 1 but y is preferred at time t − k. Present biased individuals with time inconsistent preferences and partial naiveté also exhibit overconfidence about future fitness club attendance. Since exercise has current monetary and non-monetary costs and future benefits, consumers with such preferences would choose more costly monthly contracts over pay-as-you go options with high per-visit fees. This study represents an important initial step in research on the role of present bias in exercise decisions. Health clubs design contracts to exploit members’ present bias and overconfidence. DellaVigna and Malmendier (2006) use data from three health clubs in a single US region. Thousands of health clubs and recreation centers exist; participation rates in exercise vary systematically across regions. Also, many exercise decisions involve going for a walk or a run, which require only time and a safe place to exercise. Much more research is needed before a complete picture of the decision to be physically active can be formed. In addition, many other decisions in sports markets involve current costs and future benefits; for example, new talent acquisition by sports teams. The extent to which present bias affects other decisions in sports markets is largely unexplored.

JUDGING BIASES Officials play an important role in sporting events. During sporting events, officials make almost instantaneous decisions to enforce the rules of sport. Because these decisions sometimes involve a high

508

THE SAGE HANDBOOK OF SPORTS ECONOMICS

degree of subjectivity, officials should preside over games or competitions from a neutral point of view and not affect outcomes. Referees, judges and umpires are highly trained professionals and their actions are scrutinized by the public and league officials. And yet, the existing literature identifies many factors, such as the difficulty of an athlete’s program in a competition, athletes’ nationality, and their affiliation and reputation, that can influence the objectivity of officials’ decisions. We next turn to a review of the large literature on judging bias in sport.

Difficulty Bias Difficulty bias refers to athletes attempting more difficult programs receiving higher execution scores during competition, even when the execution score is not supposed to reflect the difficulty of the program. Morgan and Rotthoff (2014) found evidence of difficulty bias in elite-level gymnastics competitions using data from the 2009 World Artistic Gymnastic Championship. The paper finds that gymnasts’ overall score increases with their routines’ difficulty. Increasing difficulty by one standard deviation increases execution scores by 0.21 standard deviations while attempting less difficult routines reduces the overall score. The authors ensure that the results do not reflect the presence of famous athletes by controlling for the presence of gymnasts from countries with a reputation of producing great gymnasts. Despite this robustness check, the evidence is not completely convincing. The authors’ claim that the execution score should be unrelated to the difficulty score appears to be strong. The quality of execution is likely to be higher for more gifted athletes, and these athletes are more likely to attempt difficult routines in the first place. Therefore, without some control for ability it is hard to attribute the higher performance scores of gymnasts with complicated routines to difficulty bias alone.

Home Team Bias Courneya and Carron (1992) describe home advantage as the consistent finding that home teams win games significantly more often than visiting teams. Carron, Hausenblas and Eys (1998) note that home advantage is present in professional and amateur sports, and in individual and team sports for athletes of both genders. Courneya and Carron (1992) and Carron, Loughhead and Bray (2005) provide an overview of home advantage. Both identify subjective decisions made by officials in favor of home teams and players as a possible contributing factor.

Early empirical research documented evidence that officials make more subjective decisions against visiting teams. For example, analyzing data from the 1973–1974 season of the Belgian National Soccer League, Lefebrve and Passer (1974) find that visiting teams received more yellow cards and penalties than home teams. Varca (1980) examines men’s college basketball games from the 1977–1978 Southeastern Conference (SEC) and concludes that away teams were called for significantly more fouls than home teams. Greer (1983) uses data from games played by Kansas State University and the University of Illinois basketball teams to find that after the crowd protested a call during a game, referees were less likely to call a violation on the home team and more likely to call a violation on the visiting team. Despite this evidence, it is hard to attribute those findings solely to officials’ preferential treatment of home teams. Carron et  al. (2005) observe that disproportional fouls on visiting teams might be a reasonable reaction of referees to more aggressive behavior of visitors’ trying to compensate for home advantage. Lehman and Reifman (1987) address the issue by comparing the number of fouls awarded to star and non-star NBA players home and away. They argue that a cheering crowd is the main mechanism affecting referees’ decisions in favor of the home team, and that the effect will be more pronounced for star players, since calling a foul on the home team star provokes a more strenuous fan reaction. Analyzing 82 games from the 1984–1985 NBA season involving the Los Angeles Lakers, Lehman and Reifman (1987) find that significantly fewer fouls were called on star players at home but there were no differences in fouls for non-star players. These findings suggest referee bias for home teams only under the assumption that star and non-star players adjust their aggressiveness between home and away games symmetrically. Some studies generate evidence supporting home advantage from outcomes in contests with high levels of subjectivity. Balmer, Nevill and Williams (2001) construct indices of home advantage based on medals won for each Winter Olympic event from 1908 to 1998 and found evidence of home advantage in subjectively judged events like figure skating and freestyle skiing, but not for objectively judged (timed) events. Balmer, Nevill and Williams (2003) reach similar conclusion about Summer Olympic Games medals. Using data from 1896 to 1996, they find a significant home advantage in subjectively judged events (boxing and gymnastics) as well as events involving some subjective decisions (team sports), but no home advantage for objectively judged events (athletics and weightlifting). Finally, Balmer, Nevill and Lane (2005) analyze data from

BEHAVIORAL ECONOMICS AND SPORT

European boxing championship bouts from March 1910 to June 2002 across all weight divisions and find that bouts ending in points decisions (relatively more subjectively judged) have a significantly higher proportion of home wins than those decided by objectively determined knockouts. Analysis of soccer referees’ bias towards home teams received significant attention in the literature after the pioneering work of Garicano, PalaciosHuerta and Prendergast (2005), in which they use injury time as a measure of bias. Injury time is added at the end of a game to compensate for time lost due to injuries and time wasting. Referees have full discretion over the amount of time added. Using data from 1994/1995 and 1998/1999 seasons of Spanish Primera Division football, Garicano et  al. (2005) show that referees use injury time to create an advantage for the home team. When a game is close, referees reduce the amount of extra time when the home team is winning and increase the amount of extra time when the home team is losing. When a game is not close, there is no evidence of such bias. Similar evidence is found by Sutter and Kocher (2004) for the German Bundesliga in the 2000/2001 season; Dohmen (2008) for Bundesliga 1992/1993–2002/2003; Rickman and Witt (2008) for in the English Premier League (EPL) 1999/2000 and 2002/2003 seasons; Scoppa (2008) for Italian Serie A 2003/2004 and 2004/2005 seasons; Lucey and Power (2009) for Serie A 2002/2003 season and Major League Soccer 2003 season; Rocha et al. (2013) for the Brazilian Football Championship, 2004–2008; Mendoza and Roses (2013) for the Colombian Professional League, 2004–2008; and Watanabe, Wicker and Reuter (2015) for the EPL 2012–2013 and 2013–2014 seasons. Using German Bundesliga data (from 2000/2001 to 2010/2011 seasons), Reidl et  al. (2015) conclude that referees’ decisions on injury time do not contribute to home advantage because there is no significant effect on goals scored. Some studies also find that referees demonstrate bias towards home teams while making decisions to award goals and penalty kicks, and showing yellow and red cards (Boyko, Boyko and Boyko, 2007; Dohmen, 2008; Buraimo, Forrest and Simmons, 2010; Buraimo, Simmons and Maciaszczyk, 2012). Dohmen and Sauermann (2016) provide a detailed literature review. Given evidence of referees treating home teams more favorably, the next issue is identifying what factors make referees more or less biased. Garicano et al. (2005) argue that the most probable source is social pressure put on the referees by the crowd. They find that higher game attendance increases home team bias. But when attendance is unusually high (measured by the attendance to capacity ratio), home team bias decreases substantially. Using German soccer data, Dohmen

509

(2008) also finds that crowd size and composition matters. But even more strikingly, distance from the crowd to the field (measured by the presence or absence of running tracks in stadiums) has a negative effect on home team bias. Buraimo et al. (2010) and Buraimo et al. (2012) find smaller bias (the home team is more likely to be awarded a yellow/red card) for stadiums with running tracks. Nevill, Balmer and Williams (2002) conduct a laboratory study showing videotaped tackles from an EPL match to professional referees and asking them to classify a tackle as regular or irregular. Referees were divided in two groups: one group could hear the crowd’s reaction and the second group could not. Referees in the first group were significantly less likely to classify home team tackles as irregular than referees in the second group. The decisions of the actual referees in the matches resembled that of the first group, suggesting that referees are affected by social pressure from the crowd. Page and Page (2010a) show that some referees are more prone to be influenced by the crowd than others using data from football matches in 15 different leagues. Based on safety concerns after violent clashes between soccer hooligans in Italy in 2007, local governments temporarily banned all spectators from matches conducted in stadiums with insufficient safety standards around the country. Pettersson-Lidbom and Priks (2010) use this variation to study referee decisions with and without a crowd present. They find home teams are treated significantly more favorably when spectators are present compared to games in empty stadiums. Deutscher (2015) compares fouls called by NBA referees during close games (teams are within five points with two or less minutes to play) with expost judgments of these fouls by the league. Based on data from 113 close games in the 2014/2015 regular season (1,229 calls), he does not find any evidence of referee bias towards the home team.

Nationalistic Bias Nationalistic bias refers to officials treating compatriot athletes or teams more favorably than athletes or teams from other countries. There are relatively few empirical studies of nationalistic referee bias in team sports, mainly because referees are not usually allowed to officiate matches involving a team of their own nationality. Mohr and Larsen (1998) compare the number of free kicks awarded by instate umpires to instate and outstate football teams in Australian Football League (AFL) from 1992 to 1995, controlling for the final score differential. Teams from the same state as the central umpires received 11% more free kicks than outstate teams.

510

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Home bias magnified the effect – the instate advantage on home grounds (15%) was greater than on outstate home grounds (5%). Using data from two rugby leagues, Page and Page (2010b) find that rugby referees favor teams from their own nationality. The favoritism is stronger in critical decisions such as awarding yellow and red cards, but dissolves when the level of scrutiny increases. The empirical evidence of nationalistic bias in individual sports is vast, especially in sports where officials are allowed some degree of subjectivity when evaluating performance. Ste-Marie (1996) finds that gymnastics judges award higher scores to competitors from the same country compared to other judges on the same panel. She assumes that nationalistic bias might be at least partially explained by the perception fluency effect – judges might have previously seen the routine of gymnasts from their country in domestic competitions, and multiple exposures to the same routine might lead to the perception fluency which, in turn, might be misattributed to increased appreciation of the routine. The data provide no empirical support for this hypothesis. Using data from the 2003 World Championships, Myers et  al. (2006) find evidence of nationalistic bias in MuayThai (Thai Kickboxing). Judges who share nationality with a boxer were more likely to award him more points. Nationalistic bias of figure skating judges received substantial attention. Seltzer and Glass (1991) analyze data from all skating events in the Winter Olympics between 1968 and 1988 and find that judges were more likely to award higher scores to their compatriots or to skaters belonging to the same political bloc. Campbell and Galbraith (1996) find strong evidence of nationalistic bias in the judging of figure skating events during the 1976, 1988, 1992 and 1994 Olympic Games. The effects are more pronounced for the medal contenders. Sala, Scott and Spriggs (2007) develop evidence supporting this conclusion by demonstrating the existence of ‘patriotic bias’ in a broader set of data from Winter Olympic Games 1948–2002. Using data from the 2012 Winter Olympics and other major competitions, Zitzewitz (2006) finds evidence of significant nationalistic bias in figure skating and ski jumping – judges were more likely to give their compatriots higher scores. Moreover, a skater with a judge from the same country on the panel was more likely to receive higher scores from all judges, which might reflect vote trading. After the 2002 Winter Olympics vote trading scandal, the International Skating Union (ISU) introduced a new set of rules which included judge anonymity. It was supposed to reduce opportunities for voting collusion and nationalistic bias, but, according Zitzewitz (2014), this policy actually

decreased transparency and did not reduce favoritism and vote trading. One potential problem with the results of these studies is that this evidence might not reflect the strategic behavior of judges trying to boost compatriot athletes higher in the rankings, but rather manifest some specific preferences that judges from different countries might have over athletes’ performance styles. There are at least two studies that attempt to address this problem. Emerson, Seltzer and Lin (2009) find strong evidence of nationalistic bias in Olympic diving competitions in 2000. The authors claim that the magnitude of the bias could have affected the medal standings. To ensure that the results reflect nationalistic bias and not style preferences, they regress residuals (the difference between the actual score of a judge and the predicted score which contains the preferences of a judge for individual athletes’ styles) on dummies for individual divers. Only one judge demonstrated significant differences in preferences. Sampaio (2012) analyzes data from an elite-level surfing competition, the Association of Surfing Professionals (ASP) World Tour, which is very well suited for overcoming difficulties with estimation of nationalistic bias. The competition design allows observation of scores awarded to the same pair of surfers multiple times by the same judging panel. If a surfer has compatriots in the judging panel, his scores are observed when he is losing and when he is winning, distinguishing the desire of a judge to help a fellow countryman (by strategically inflating his points or deflating the points of the competitors) from preferences for surfing style. Sampaio finds that surfing judges neither overscore nor underscore athletes from the same country; rather, they significantly underscore their opponents, especially when their compatriots are losing. The magnitude of the bias is large enough to affect the final standings in the tournament.

Reputation/expectation Bias Judges or referees might rely on information based on prior behavior or outcomes that should be irrelevant to the evaluation of current performance when making decisions. This tendency is called reputation or expectation bias. One of the most wellknown examples of reputation bias is the placement effect in gymnastics. In gymnastics competitions, coaches control the order in which athletes will perform and judges expect coaches to put their best athletes later on in the competition. Therefore, judges expect the quality of performances to improve with team order. Scheer and Ansorge (1975), Ansorge et  al. (1978) and Plessner (1999) conduct experimental studies that show identical

BEHAVIORAL ECONOMICS AND SPORT

videotaped gymnastics routines to professional judges, manipulating the order of the routines shown within teams. All three find that gymnasts appearing in the final team position received significantly higher scores than those appearing earlier. Jones, Paull and Erskine (2002) conduct an experiment that randomly assigned 38 football referees to two groups and showed each group identical video clips of football plays involving a generic ‘blue team.’ Both groups were asked to indicate the calls they would have made if refereeing the game. The treatment group received additional information: the blue team was known for their aggressive style of play. The results indicate that referees exhibited reputation bias – the treated group awarded significantly more yellow and red cards to the blue team even though both groups judged the same video clips. More recent works use field data. Analyzing data from Major League Baseball, Mills (2014) finds evidence of umpire bias toward players with higher status. Umpires are significantly more likely to call a strike on pitches thrown by older pitchers and pitchers with higher career WAR (Wins Above Replacement), which summarizes a player’s total contributions to his team’s success. Using data provided by a new monitoring technology, Kim and King (2014) directly measured the accuracy of strike calls and analyzed whether pitcher status affects umpires’ decisions. The results suggest that umpires treat high-status pitchers more favorably – they are more likely to over-recognize quality for high-status pitchers. Lago-Peñas and Gómez-López (2016) analyze football data from La Liga in Spain for the 2014–2015 season and find that referees favored higher ranked teams by awarding more injury time in games when they were behind and less time when they were ahead.

Representativeness Bias Behavioral biases also affect fans and athletes. The hot hand – outcomes where athletes or teams experience periods where outcomes appear certain to be successful – has long attracted the attention of researchers. The hot hand reflects representativeness bias (Tversky & Kahneman, 1974), which applies to probabilistic decisions, such as ‘what is the probability outcome X is generated by process Y?’, and captures the difficulty people have in interpreting random sequences of events. Hot hand implications for economic decisionmaking stem from the ubiquity of economic decisions made under uncertainty. Consumer decisions about purchases and time allocation decisions about education and occupation have uncertain outcomes. Financial market participants repeatedly

511

face decisions involving uncertain returns. In sports, fans make economic decisions about attending games with uncertain outcomes, and teams offer players contracts before observing uncertain player performance. If agents routinely misinterpret random events and believe that observed sequences reflect deterministic hot hand outcomes, then these decisions could be biased, leading to poor decisions and suboptimal resource allocation. Gilovich, Vallone and Tversky (1985), in the seminal hot hand study, investigate the hot hand based on data on shots taken by individual basketball players. Their probabilistic research question, ‘what is the probability that an observed sequence of shots by a basketball player come from a random process with known mean?’, is subject to interpretation in different ways. Objectively, if an observed sequence of basketball shots exceeds some expected value, based on past outcomes or perceived ability, then a basketball player is said to have a hot hand. Any observed shot outcome sequence represents a random variable and individual perceptions of these and other random variables can reflect behavioral biases. Gilovich et  al. (1985) analyze sequences of basketball shot outcomes, including field goal attempts and free throw attempts, made by National Basketball Association (NBA) players in the 1980–1981 season, interview NBA coaches and players from the Philadelphia 76ers about outcome sequences, survey NBA fans about perceptions of sequences of shot outcomes, and conduct controlled experiments on men’s and women’s basketball players at Cornell University. Both participants and observers uniformly believed that the hot hand existed in all settings, including experimental evidence based on Cornell basketball players and a teammate observer who placed small bets on shot outcomes based on previous shots. In the experiment, the shooter and observer placed bets that were consistent with a belief in the hot hand but all statistical analyses of shot outcome sequences did not support hot hand effects in the data. Perception differed from reality, and behavioral biases known to affect decision-making under uncertainty can explain these differences. Since Gilovich et al. (1985), many papers tested for the presence of hot hand effects in multiple settings.2 A substantial body of hot hand research has been published in economics journals. As a result, the hot hand literature is too extensive for a comprehensive review. In general, results appear to be mixed. For example, Bar-Eli, Avugos and Raab (2006) survey ‘important’ hot hand research and find 13 studies supporting the presence of hot hand effects and 11 studies not supporting it. Green and Zwiebel (2017) contains a discussion of the recent literature.

512

THE SAGE HANDBOOK OF SPORTS ECONOMICS

The classic exchange between Camerer (1989) and Brown and Sauer (1993) illustrates the lack of agreement in the hot hand literature and the economic consequences of mis-interpreting random sequences. Camerer (1989) develops evidence supporting team-level hot hand outcomes in NBA games over the 1983–1986 seasons using point spreads, which reflect both pre-game bettor and bookmaker perceptions and past game outcomes. The point spread represents an estimate of the number of points by which a stronger team is expected to beat a weaker team; point spreads are known to be accurate predictors of actual game outcomes (Sauer, 1998). Camerer (1989) undertakes a statistical analysis of point spreads and game outcomes on NBA games conditional on consecutive past games won or lost by each team. Teams on winning streaks may reflect hot hand outcomes and differences between point spreads and game outcomes can be interpreted as forecast errors. When the point spread exceeds the game scoring differential, the forecast error is positive and the betting market over-forecast the performance of the stronger team. Camerer (1989) develops evidence of positive forecast errors for NBA teams on long (2-8 game) winning streaks relative to teams not on winning streaks, suggesting that performance perceptions of teams with the hot hand exceeded their actual performance. Brown and Sauer (1993) challenge the results in Camerer (1989) for failing to distinguish between perceived ‘hot hand’ effects in betting markets and actual team performance consistent with hot hand effects affecting point spreads. In other words, Camerer (1989) assumes that the hot hand represents a pure behavioral bias even though actual game outcomes exhibit streaky performance. Brown and Sauer (1993) test for both streaks in performance and for the effects of performance streaks on point spreads, and develop evidence consistent with the presence of both hot hand effects in performance and betting markets containing hot hand effects. These results suggest a nuanced story about hot hand effects. Rather than reflecting simple widespread behavioral biases on the part of betting market participants, some betting market participants may make mistakes attributable to hot hand biases, but the presence of persistent team success is consistent with the presence of bettor preferences is more complex than the predictions of standard economic models.

PEER EFFECTS Peers might affect other agent’s decisions in many contexts – educational attainment, workplace productivity, and anti-social behavior to name a few.

Peer effects refer to externalities generated by spillovers from one individual to other members of some group. The basic economic model assumes that no spillovers across group members exist. Studying peer effects in the field involves challenges. First, precisely identifying appropriate peer groups and collecting data on outcomes for all individuals and their peers. Second, Manski (1993) establishes several identification problems in estimating causal impacts of peer effects; many can be resolved with random peer assignment, which occurs infrequently in the field. Analyzing peer effects in the context of sports allows researchers to avoid these problems. Sports outcome data can be easily obtained, peer groups are often defined by rules, regulations and institutions in each particular sport. Moreover, the rules assigning competitors into pairs or groups often involve randomization. One hotly debated question in the peer-effects literature is the impact of heterogeneity in competitors’ abilities. This issue has been addressed using data from many settings. Brown (2011) uses data from Professional Golfers’ Association of America (PGA) tournaments from 1999 to 2010 to explore the relationship between the presence of a superstar in the competition and effort exerted by other participants. She finds that Tiger Woods’ participation in PGA events adversely affects other golfers. The effect is more pronounced for higher-skilled competitors and stronger when Woods was performing well. Brown does not find any evidence that reduced performance stems from the adoption of risky strategies; she explains reduced performance as a result of a decrease in non-stars’ incentives to win in the presence of a superstar golfer due to lower expected payoffs. Employing data from the Japan Golf Tour from 1994 to 1996 and using the presence of Masashi ‘Jumbo’ Ozaki to generate superstar effects, Tanaka and Ishino (2012) also find that the presence of a superstar player has an adverse effect on other players’ scores. Guryan, Kroft and Notowidigdo (2009) extend the analysis to the effects of the ability of a golfer’s playing partner on his performance. Using data from PGA tournaments from 1999 to 2006, they find no evidence of peer effects. The results are robust to employing a number of ability measures, including average ability, maximum ability, and minimum ability, focusing only on Tiger Woods’ partners’ performance, employing a measure of heterogeneity of partners’ abilities and examining the performance of golfers paired with partners in top 10%, top 25%, bottom 25% and bottom 10%. The findings of Guryan et  al. (2009) are consistent with those of Brown (2011). The former analyze pair-level data while the latter analyzes t­ournament-level data. But Connolly and

BEHAVIORAL ECONOMICS AND SPORT

Rendleman (2014) argue that Brown’s (2011) study contains major methodological issues. They discover that, after altering the regression model specification and adjusting the estimated standard errors, Tiger Woods’ presence in a tournament actually has no significant effect on other golfers’ performance. Rather than outcomes, Ozbeklik and Smith (2017) concentrate their attention on golfers’ risktaking behavior in match play competitions where each hole is won/lost by the player with the lowest/ highest number of strokes. Risk taking is defined as the standard deviation of players’ score to par on types of holes (par 3s, par 4s, and par 5s). The results suggest that trailing players take more risks and take increasingly more risk the further behind they are in a competition. Also, lower-ranked players adopt riskier strategies as the relative strength of their opponent increases. This implies that when playing against a superstar, non-superstar golfers do not give up; they are taking more risks. Several studies estimate peer effects exploiting institutional design factors inherent in specific competitions. Using the data from the preliminary round of the 2009 World Gymnastic Championship, where each participating country is randomly assigned to one of three performance spots, Rotthoff (2015) finds that the score of a gymnast does not depend on the performance of the athlete performing immediately before him or her. This result holds for gymnasts following superstar athletes, indicating no superstar peer effects in this setting. Yamane and Hayashi (2015) use data on swimming competitions and find that swimmers’ performance is affected by that of a competitor in an adjacent lane. In particular, swimmers swim faster when they are ahead of the competitor in the adjacent lane compared to when swimming alone, but they swim slower when a competitor in an adjacent lane is ahead of them. The authors demonstrate that observability is a key determinant of peer effects, finding no effect of competitors’ performance in adjacent lanes in backstroke competitions where swimmers are not able to see other swimmers. Booth and Yamamura (2016) use the unique institutional design of Japanese speedboat racing to analyze peer effects in sporting events. In Japan, female speedboat racers compete under the same conditions as males, and participants are randomly assigned into groups for each race, creating variation in the gender composition of groups of competitors. Booth and Yamamura (2016) find that women’s race times are slower in mixed-gender races than in all-women races, while men’s race times are faster in mixed-gender races than in all-male races. They also find that men with slower race times in exhibition races tend to exert less effort during the actual race, but women’s performance appears to be unaffected by their exhibition time.

513

Evidence of peer effects from sports markets appears mixed, even using data from the same events. Since peer effects could affect a large number of important economic variables, such as human capital investment and worker productivity, developing strong evidence supporting or rejecting the presence of peer effects represents an important goal in empirical economic research. Sports represent a rich setting for empirical research on peer effects, and many examples of outcomes based on quasi-random peer assignment in sports remain unexamined.

Notes 1  Each hole on a golf course is assigned a specific number of strokes that a typical golfer should make to complete it, which is called par on that hole. 2  Gilovich et  al. (1985) has been cited more than 1,000 times.

REFERENCES Allen, E. J., Dechow, P. M., Pope, D. G., & Wu, G. (2017). Reference-dependent preferences: Evidence from marathon runners. Management Science, 63(6), 1657–1672. Online first, April 20, 2016, https://doi.org/ 10.1287/mnsc.2015.2417 Ansorge, C. J., Scheer, J. K., Laub, J., & Howard, J. (1978). Bias in judging women’s gymnastics induced by expectations of within-team order. Research Quarterly: American Alliance for Health, Physical Education and Recreation, 49(4), 399–405. Balmer, N. J., Nevill, A. M., & Lane, A. M. (2005). Do judges enhance home advantage in European championship boxing? Journal of Sports Sciences, 23(4), 409–416. Balmer, N. J., Nevill, A. M., & Williams, A. M. (2001). Home advantage in the Winter Olympics (1908– 1998). Journal of Sports Sciences, 19(2), 129–139. Balmer, N. J., Nevill, A. M., & Williams, A. M. (2003). Modelling home advantage in the Summer Olympic Games. Journal of Sports Sciences, 21(6), 469–478. Bar-Eli, M., Avugos, S., & Raab, M. (2006). Twenty years of ‘hot hand’ research: Review and critique. Psychology of Sport and Exercise, 7(6), 525–553. Booth, A. L., & Yamamura, E. (2016). Performance in mixed-sex and single-sex tournaments: What we can learn from speedboat races in Japan. IZA Discussion paper 10384. Boyko, R. H., Boyko, A. R., & Boyko, M. G. (2007). Referee bias contributes to home advantage in English Premiership football. Journal of Sports Sciences, 25(11), 1185–1194.

514

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Brown, J. (2011). Quitters never win: The (adverse) incentive effects of competing with superstars. Journal of Political Economy, 119(5), 982–1013. Brown, W. O., & Sauer, R. D. (1993). Does the basketball market believe in the hot hand? Comment. The American Economic Review, 83(5), 1377–1386. Buraimo, B., Forrest, D., & Simmons, R. (2010). The 12th man? Refereeing bias in English and German soccer. Journal of the Royal Statistical Society: Series A (Statistics in Society), 173(2), 431–449. Buraimo, B., Simmons, R., & Maciaszczyk, M. (2012). Favoritism and referee bias in European soccer: Evidence from the Spanish League and the UEFA Champions League. Contemporary Economic Policy, 30(3), 329–343. Camerer, C. F. (1989). Does the basketball market believe in the ‘hot hand’? The American Economic Review, 79(5), 1257–1261. Campbell, B., & Galbraith, J. W. (1996). Nonparametric tests of the unbiasedness of Olympic figureskating judgments. Journal of the Royal Statistical Society Series D (The Statistician), 45(4), 521–526. http://dx.doi.org/10.2307/2988550 Card, D., & Dahl, G. B. (2011). Family violence and football: The effect of unexpected emotional cues on violent behavior. The Quarterly Journal of Economics, 126(1), 103–143. Carron, A. V., Hausenblas, H. A., & Eys, M. A. (1998). Group dynamics in sport. Morgantown WV: Fitness Information Technology. Carron, A. V., Loughhead, T. M., & Bray, S. R. (2005). The home advantage in sport competitions: Courneya and Carron’s (1992) conceptual framework a decade later. Journal of Sports Sciences, 23(4), 395–407. Coates, D., Humphreys, B. R., & Zhou, L. (2014). Reference-dependent preferences, loss aversion and live game attendance. Economic Inquiry, 52(3), 959–973. Connolly, R. A., & Rendleman, R. J. (2014). The (adverse) incentive effects of competing with superstars: A reexamination of the evidence. SSRN Papers, December 5. Courneya, K. S., & Carron, A. V. (1992). The home advantage in sport competitions: A literature review. Journal of Sport and Exercise Psychology, 14(1), 13–27. DellaVigna, S., & Malmendier, U. (2006). Paying not to go to the gym. American Economic Review, 96(3), 694–719. Deutscher, C. (2015). No referee bias in the NBA: New evidence with leagues assessment data. Journal of Sports Analytics, 1(2), 91–96. Dickson, A., Jennings, C., & Koop, G. (2016). Domestic violence and football in Glasgow: Are reference points relevant? Oxford Bulletin of Economics and Statistics, 78(1), 1–21. Dohmen, T. (2008). The influence of social forces: Evidence from the behavior of football referees. Economic Inquiry, 46(3), 411–424.

Dohmen, T., & Sauermann, J. (2016). Referee bias. Journal of Economic Surveys, 30(4), 679–695. Emerson, J. W., Seltzer, M., & Lin, D. (2009). Assessing judging bias: An example from the 2000 Olympic Games. The American Statistician, 63(2), 124–131. Garicano, L., Palacios-Huerta, I., & Prendergast, C. (2005). Favoritism under social pressure. The Review of Economics and Statistics, 87(2), 208–216. Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball: On the misperception of random sequences. Cognitive Psychology, 17(3), 295–314. Green, B., & Zwiebel, J. (2017). The hot-hand fallacy: Cognitive mistakes or equilibrium adjustments? Evidence from Major League Baseball. Management Science, September 15. https://doi.org/10.1287/ mnsc.2017.2804 Greer, D. L. (1983). Spectator booing and the home advantage: A study of social influence in the basketball arena. Social Psychology Quarterly, 46, 252–261. Guryan, J., Kroft, K., & Notowidigdo, M. J. (2009). Peer effects in the workplace: Evidence from random groupings in professional golf tournaments. American Economic Journal: Applied Economics, 1(4), 34–68. Humphreys, B. R., & Ruseski, J. E. (2011). An economic analysis of participation and time spent in physical activity. The BE Journal of Economic Analysis & Policy, 11(1), 1–32. Humphreys, B. R., & Zhou, L. (2015). Referencedependent preferences, team relocations, and major league expansion. Journal of Economic Behavior & Organization, 109, 10–25. Jones, M. V., Paull, G. C., & Erskine, J. (2002). The impact of a team’s aggressive reputation on the decisions of association football referees. Journal of Sports Sciences, 20(12), 991–1000. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–292. Kim, J. W., & King, B. G. (2014). Seeing stars: Matthew effects and status bias in major league baseball umpiring. Management Science, 60(11), 2619–2644. Lago-Peñas, C., & Gómez-López, M. (2016). The influence of referee bias on extra time in elite soccer matches. Perceptual and Motor Skills, 122(2), 666–677. Lefebvre, L., & Passer, M. (1974). The effects of game location and importance on aggression in team sport. International Journal of Sport Psychology, 5, 102–110. Lehman, D. R., & Reifman, A. (1987). Spectator influence on basketball officiating. Journal of Social Psychology, 127(6), 673–675. Lucey, B. M., & Power, D. (2009). Do soccer referees display home bias? Working Paper, SSRN. Manski, C. F. (1993). Identification of endogenous social effects: The reflection problem. The Review of Economic Studies, 60(3), 531–542.

BEHAVIORAL ECONOMICS AND SPORT

Mendoza, J., & Rosas, A. (2013). Referee bias in professional soccer: Evidence from Colombia. Working Paper 011059, Universidad del Pacífico. Mills, B. M. (2014). Social pressure at the plate: Inequality aversion, status, and mere exposure. Managerial and Decision Economics, 35(6), 387–403. Mohr, P. B., & Larsen, K. (1998). Ingroup favoritism in umpiring decisions in Australian Football. The Journal of Social Psychology, 138(4), 495–504. Morgan, H. N., & Rotthoff, K. W. (2014). The harder the task, the higher the score: Findings of a difficulty bias. Economic Inquiry, 52(3), 1014–1026. Munyo, I., & Rossi, M. A. (2013). Frustration, euphoria, and violent crime. Journal of Economic Behavior & Organization, 89, 136–142. Myers, T. D., Balmer, N. J., Nevill, A. M., & Al Nakeeb, Y. (2006). Evidence of nationalistic bias in MuayThai. Journal of Sports Science & Medicine, 5(CSSI), 21. Nevill, A. M., Balmer, N. J., & Williams, A. M. (2002). The influence of crowd noise and experience upon refereeing decisions in football. Psychology of Sport and Exercise, 3(4), 261–272. Ozbeklik, S., & Smith, J. K. (2017). Risk taking in competition: Evidence from match play golf tournaments. Journal of Corporate Finance, 44, 506–523. Page, K., & Page, L. (2010a). Alone against the crowd: Individual differences in referees’ ability to cope under pressure. Journal of Economic Psychology, 31, 192–199. Page, L., & Page, K. (2010b). Evidence of referees’ national favouritism in rugby. Working Paper 62, National Centre for Econometric Research Working Paper Series. Pettersson-Lidbom, P., & Priks, M. (2010). Behavior under social pressure: Empty Italian stadiums and referee bias. Economics Letters, 108(2), 212–214. Plessner, H. (1999). Expectation biases in gymnastics judging. Journal of Sport and Exercise Psychology, 21(2), 131–144. Pope, D. G., & Schweitzer, M. E. (2011). Is Tiger Woods loss averse? Persistent bias in the face of experience, competition, and high stakes. The American Economic Review, 101(1), 129–157. Rickman, N., & Witt, R. (2008). Favouritism and financial incentives: A natural experiment. Economica, 75(298), 296–309. Riedl, D., Strauss, B., Heuer, A., & Rubner, O. (2015). Finale furioso: Referee-biased injury times and their effects on home advantage in football. Journal of Sports Sciences, 33(4), 327–336. Rocha, B., Sanches, F., Souza, I., & Carlos Domingos da Silva, J. (2013). Does monitoring affect corruption? Career concerns and home bias in football refereeing. Applied Economics Letters, 20(8), 728–731.

515

Rotthoff, K. W. (2015). (Not finding a) sequential order bias in elite level gymnastics. Southern Economic Journal, 81(3), 724–741. Sala, B. R., Scott, J. T., & Spriggs, J. F. (2007). The cold war on ice: Constructivism and the politics of Olympic figure skating judging. Perspectives on Politics, 5(1), 17–29. Sampaio, B. (2012). For love of country: Nationalistic bias in professional surfing. In 34° Meeting of the Brazilian Econometric Society. Sauer, R. D. (1998). The economics of wagering markets. Journal of Economic Literature, 36(4), 2021–2064. Scheer, J. K., & Ansorge, C. J. (1975). Effects of naturally induced judges’ expectations on the ratings of physical performances. Research Quarterly. American Alliance for Health, Physical Education and Recreation, 46(4), 463–470. Scoppa, V. (2008). Are subjective evaluations biased by social factors or connections? An econometric analysis of soccer referee decisions. Empirical Economics, 35(1), 123–140. Seltzer, R., & Glass, W. (1991). International politics and judging in Olympic skating events: 1968– 1988. Journal of Sport Behavior, 14(3), 189. Ste-Marie, D. M. (1996). International bias in gymnastic judging: Conscious or unconscious influences? Perceptual and Motor Skills, 83(3), 963–975. Sutter, M. and Kocher, M. G. (2004). Favoritism of agents–the case of referees’ home bias. Journal of Economic Psychology, 25(4), 461–469. Tanaka, R., & Ishino, K. (2012). Testing the incentive effects in tournaments with a superstar. Journal of the Japanese and International Economies, 26(3), 393–404. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124–1131. Varca, P. E. (1980). An analysis of home and away game performance of male college basketball teams. Journal of Sport Psychology, 2(3), 245–257. Watanabe, N. M., Wicker, P., & Reuter, J. C. (2015). Determinants of stoppage time awarded to teams in the English premier league. International Journal of Sport Finance, 10(4), 310. Yamane, S., & Hayashi, R. (2015). Peer effects among swimmers. The Scandinavian Journal of Economics, 117(4), 1230–1255. Zitzewitz, E. (2006). Nationalism in winter sports judging and its lessons for organizational decision making. Journal of Economics & Management Strategy, 15(1), 67–99. Zitzewitz, E. (2014). Does transparency reduce favoritism and corruption? Evidence from the reform of figure skating judging. Journal of Sports Economics, 15(1), 3–30.

50 Is There a Gender Difference in the Response to Competitive Settings? Michael A. Leeds

INTRODUCTION Almost a half-century after the start of the ‘women’s movement’ in the United States, women remain a long way from gender equality in the workplace. According to the American Community Survey, the median annual earnings of women in the US ($40,905) were 80% of men’s earnings ($51,388).1 Just as women in general struggle to achieve equality with men, women’s sports have generally played second-fiddle to men’s sports. While few doubt that sex discrimination continues to play a role in sport and society, recent research suggests that women may face insurmountable barriers to achieving parity with men, even in the absence of discrimination. (For a good summary of this literature, see Croson & Gneezy, 2009). A sizable experimental literature resulted in three disturbing findings regarding how women respond to high-stakes contests. First, studies find that, while men perform better in the winner-takeall settings associated with rank-order tournaments, women typically perform worse the higher the stakes (e.g., Gneezy, Niederle, & Rustichini, 2003). Second, they find women become more readily discouraged than men in such contests (e.g., Booth & Nolen, 2012). Finally, they find that, unlike men, women try to avoid high-stakes

situations (e.g., Gil & Prowse, 2014).2 These findings have implications for women in sports and in the economy as a whole. If women respond to pressure-filled situations with conservative or sloppy play, the competitions will be less appealing to fans and sponsors. The lack of such sponsorship has, for example, been used to explain why the highly successful US Women’s National Soccer Team has been paid much less than their less-successful men’s counterparts (see Das, 2016, for a review of this example). More generally, because rewards in the workplace are frequently discontinuous, bringing large rewards to the ‘winners,’ these behaviors could put women at a perpetual disadvantage. If this observation is true, only major changes in the structure of the labor market will enable women to achieve true equality in the workplace. One problem with the research on gender preferences and their impact on the relative pay of women and men is the lack of data to test the hypothesis that women respond differently to incentives than men do. Except for a literature on gender differences in response to high-stakes academic exams (e.g., Attali, Neeman, & Schlosser, 2011; Jurajda & Münich, 2011; Ors, Palomino, & Peyrache, 2013), most of the evidence regarding gender differences in competitive settings comes from experiments rather than real-world data.

Is There a Gender Difference in the Response to Competitive Settings?

It has become increasingly apparent, however, that economic experiments are not the panacea they were once considered to be by the academic community (see Deaton & Cartwright, 2018, for an overview).3 Economists have used sports as an empirical setting to test hypotheses about how economic contests affect incentives since the work of Ehrenberg and Bognanno (1990a, 1990b) on professional golf. They studied the impact of golf’s steep earnings gradient, in which the difference between prizes increases as one moves up the leaderboard, on performance. They showed that, all else being equal, golfers who are closer to the lead when entering the final round of a tournament – and hence stand to increase their reward by a greater amount – show greater improvement than other golfers. Academic studies have applied Ehrenberg and Bognanno’s work (1990a, 1990b) to other sports. Lynch and Zax (2000) found larger payouts resulted in faster times in road races. However, this result disappeared when they controlled for the quality of the runners involved. They concluded that higher rewards improve times by attracting better runners rather than spurring the runners to perform better. Economists also use sports to test for differences in how women and men respond to economic contests. Evidence from these studies indicates, contrary to the implications of the experimental studies, that the impact of economic contests on women is often no different from the impact on men (e.g., Banko, Leeds, & Leeds, 2016; Gilsdorf & Sukhatme, 2008a, 2008b; Krumer, Rosenboim, & Shapir, 2016). This chapter shows how economists have used a variety of sports to test whether contests induce different behavior from women and men, and hence whether high-stakes settings form an insuperable barrier to gender equality. The next two sections of this chapter briefly introduce the reader to rank-order tournaments and how they affect incentives. First, I motivate and summarize the basic model and briefly point out how sports economists have contributed to the literature. Next, I show how rank-order tournaments might create a barrier for women. I then turn to the main body of the chapter, the contribution of sports economists to the literature.

RANK-ORDER TOURNAMENTS Economists’ fondness for freely-functioning competitive markets stems from the fact that

517

competitive markets provide buyers and sellers with incentives that induce them to act in the best interest of the whole. As Adam Smith (1974) noted: It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest. We address ourselves, not to their humanity but to their self-love, and never talk to them of our necessities but of their advantages. (p. 45)

Smith’s insight into the product market holds equally well for the labor market. In a competitive labor market, workers are paid a wage equal to their marginal revenue product (MRP). Their MRP is the product of the workers’ marginal physical product (the extra output produced by an additional worker) and the firm’s marginal revenue (the extra revenue brought in by an additional worker). Thus, a slightly more productive worker receives a wage that is slightly higher than a less capable or diligent coworker. In the real world, however, minor differences in performance often lead to major differences in reward. It took Roger Federer five sets and almost four hours to defeat Rafael Nadal in the 2017 Australian Open (Garber, 2017), but he received almost twice Nadal’s prize money ($2.7 million versus $1.4 million; Jones, 2017). In the corporate world as well, a slight difference in performance is often all that separates a CEO from one of many vice presidents (e.g., Bognanno, 2001). Employers generally do not base pay on a worker’s marginal product because, except for a few special cases, such as car sales, it is difficult to measure exactly how much a worker produces and reward her accordingly. Constantly monitoring the workforce is prohibitively expensive, and periodically observing a worker is unlikely to yield a reliable measure of worker productivity. Rather than try to measure the absolute level of performance, firms often rely on the workers’ relative performance. Relative performance is easier to observe and is less subject to random fluctuations. This observation led Lazear and Rosen (1981) to propose the rank-order tournament as an alternative compensation mechanism. They show that, in a world with identical, risk-neutral workers and perfectly competitive firms, one can establish a tournament that provides the same incentives as a system in which one can identify workers’ output and reward them accordingly. Specifically, Lazear and Rosen (1981) show that, rather than attempt to measure precisely how much a worker produces, firms can pre-announce a contest, a rank-order tournament, in which the winner receives a high wage (relative to MRP),

518

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and the loser receives a low wage. By increasing or decreasing the spread between the two, firms can induce workers to exert greater or less effort. O’Keeffe, Viscusi, and Zeckhauser (1984) generalized this basic model to include heterogeneous agents. They showed that rank-order tournaments do not motivate effort when one contestant dominates the field. Brown (2011) later verified the impact of uneven competition by analyzing performance on the Professional Golfers’ Association (PGA) Tour from 1999 to 2010, when Tiger Woods dominated men’s golf. Brown (2011) also found other golfers’ performances improved when Woods was absent from a tournament. Moreover, this improvement mostly occurred among Woods’s closest competitors. Lazear (1989) noted pay inequality created by rank-order tournaments can cause friction among workers. The friction, in turn, can erode firm performance. This friction is particularly acute when work is done in teams rather than individually. Numerous studies of Major League Baseball teams show team performance declines as pay inequality rises, all else being equal (e.g., Annala & Winfree, 2011; Depken, 2000; Jewell & Molina, 2004). Interestingly, Kamada and Katayama’s (2014) study of pay inequality among professional baseball teams in Japan, where pay is notably more equal than in the US, found a team’s performance improved with greater dispersion in pay.

RANK-ORDER TOURNAMENTS AND GENDER The literature on gender preferences and tournaments is nicely summarized in Croson and Gneezy (2009). I present a few of the more salient studies here. Their findings clearly indicate that women respond less positively to tournaments than men do, that women are more readily discouraged, and that women seek to avoid the winner-take-all system inherent in tournaments. All these results work to the detriment of women in the labor market. In one of the first experimental studies of gender preferences and tournaments, Gneezy et al. (2003) had students at the Technion – Israel Institute of Technology – work on a series of mazes. The students were randomly divided into two groups. The first group was told that there was a fixed prize for each correctly solved maze. The second group was told that there would be only one (large) prize for the person who solved the most mazes. Gneezy et al. (2003) found that women and men in the first

group solved the same number of puzzles but that men in the second group solved significantly more puzzles than the women did. Gneezy and Rustichini (2004) obtained comparable results for a much younger group of students. They had Israeli fourth-graders run individual 40-meter sprints and then repeated the process in a series of head-to-head competitions. They found that the boys’ performances improved when they ran competitively, while the girls’ performances worsened. These results were robust to different forms of competition (mixed-sex, single-sex, fast runner and slow runner, two slow runners, and two fast runners). This result is also notable because it appears that competition itself, and not the prospect of a prize, affects performance. The experimental literature also found ‘women report more intense nervousness and fear than men in anticipation of negative outcomes’ (Croson & Gneezy, 2009, p. 452). Perhaps for this reason, women respond worse to negative feedback than men do. Discouragement, in turn, can depress performance, as it reduces the perceived impact that one’s effort can have on the contest’s outcome. Gill and Prowse (2014) test this second hypothesis by randomly pairing individuals who competed without knowing their opponent’s identity. Each participant used a mouse to place a series of computer cursors in the center of a computer screen over a two-minute period. Each contestant was graded according to the number of cursors centered and how well placed each cursor was. The two participants’ grades, however, did not by themselves determine who won the prize. Instead, their scores set their relative probabilities of winning a lottery. The winner of the lottery was then awarded a prize. This process was then repeated in subsequent rounds. Gill and Prowse (2014) found two noteworthy results. First, women responded more profoundly to losing than men did. Second, women’s response to a ‘bad draw’ in the lottery was particularly pronounced relative to that of men, so much so that the ‘differential responses to luck account for about half of the gender performance gap that we observe in our experiment’ (Gill & Prowse, 2014, p. 351). Booth and Nolen (2012) asked whether women and men have different attitudes towards tournaments. They randomly assigned a group of English high school students into singlesex and mixed-sex groups. Each group was given a sequence of mazes to solve. In the first round, both groups were rewarded in the piece rate system used by Gneezy et al. (2003). In the second round, they were confronted by a winner-take-all tournament. After each group had experienced both reward mechanisms, all the participants were allowed to choose whichever mechanism

Is There a Gender Difference in the Response to Competitive Settings?

they wished for the third round. Booth and Nolen (2012) found the girls were significantly more likely than boys to choose the piece-rate setting. This finding was less pronounced for girls who attended all-female schools. All three of the above conclusions paint a bleak picture of women’s labor market prospects.

WHAT SPORTS ECONOMICS TELLS US ABOUT GENDER PREFERENCES The majority of the sports economics research on gender differences in the response to rank-order tournaments takes place in the context of individual sports. The use of individual sports as an empirical setting for this research area occurs largely for two reasons. First, there are relatively few professional team sports that have both men’s leagues and women’s leagues. Second, the few sports that have both men’s and women’s leagues are marked by vastly different pay scales. It is much less difficult to hold all else constant in individual sports. Men and women participate in many sports whose rules and parameters closely resemble each other, such as tennis, golf, skiing, and many track and field events. In addition, the pecuniary and, where relevant, non-pecuniary rewards for men’s and women’s sports are much closer in these sports than in most team sports. Hence, the bulk of this section is devoted to individual sports.

Evidence from Tennis Tennis has received the most attention from scholars studying gender differences and incentives in sports. Tennis is a popular research setting largely because it is an individual sport and because the rules and rewards that women face resemble those rules and rewards faced by men more closely than in other sports. The evidence from tennis generally does not support the conclusion that women respond worse to incentives than men do, though there is some evidence that women become discouraged more easily. Gilsdorf and Sukhatme (2008a, 2008b) found no difference in how women and men respond to the incentives provided by elimination tournaments – contests in which a player must win to advance to the next round (and a higher payout) – in tennis. In separate studies for women and men, they found tournaments with higher payoffs have fewer upsets. This finding corresponds to Rosen’s

519

(1986) conjecture that higher rewards in elimination tournaments elicit greater effort. When all participants play closer to their best, the more talented player is more likely to win, resulting in fewer upsets. Paserman (2010) used dynamic programming to analyze stroke-level data from men’s and women’s Grand Slam tournaments, allowing him to evaluate whether women and men respond differently to crucial points in the match. In general, he concluded both men and women play more conservatively at key moments in a match. More specifically, Paserman (2010) noticed that women make more unforced errors at crucial moments4 than at other times, while men do not. However, he did not conclude that women respond to pressure worse than men do, as men’s play also deteriorates. While men do not make more unforced errors, they fail to hit as many ‘winners’ – shots that cannot be returned by their opponent – at crucial points in the match. Paserman (2010) concludes that these two effects roughly offset each other. Thus, game outcomes do not differ by sex. Several papers found that women’s tennis matches are far more likely to end in blowouts than are men’s matches. Magnus and Klaassen (1999) found that women’s matches at Wimbledon last fewer games compared to men’s matches, even when accounting for the fact that women play best-of-three sets while men play best-of-five sets. Krumer, Rosenboim, and Shapir’s (2016) study broadly supported the findings of Magnus and Klaassen’s (1999) earlier research. Neither of these studies, however, linked their findings to gender preferences. In fact, Krumer et al. (2016) claimed the cause is physical rather than psychological differences between women and men. They found gender differences in match length disappear when accounting for the physical attributes of the contestants. Banko, Leeds, and Leeds (2016) addressed the question of whether women are more easily discouraged than men by testing whether women are more likely than men to lose a match in straight sets.5 Using data from the Association of Tennis Professionals (men’s) and Women’s Tennis Association (women’s) tours for 2011, they found that, controlling for player ability, women who lose the first set of a match are no more or less likely than men to lose the second as well. They also found women who lose the second set are no more or less likely than men to lose the third set. Banko et al. (2016), however, did find one significant difference between men and women. Women who do lose in straight sets play fewer games in the second set than men who lose in straight sets. Thus, women may be no more likely to lose in straight sets than men are, but – similar to Magnus

520

THE SAGE HANDBOOK OF SPORTS ECONOMICS

and Klaassen (1999) and Krumer et  al. (2016) – they found women who do lose in straight sets are more likely than men to lose the second set badly.

Evidence from Golf One problem with using tennis as a setting to understand gender differences is that, although it is an individual sport, the players interact directly with one another. This element of player interaction can make it hard to isolate the performance of one player from that of another. For example, a ‘winner’ may be set up by a weak shot by one’s opponent. In golf, the players compete only by seeing who can complete the course in the fewest strokes. This element allows researchers to compare performances without worrying about whether a given outcome is the result of bad play by one competitor or good play by the other. The golf studies extend the earlier work of Ehrenberg and Bognanno (1990a, 1990b) on the men’s golf tour, the PGA, by contrasting the behavior of men with their results for the women’s golf tour, the Ladies Professional Golf Association (LPGA). The golf studies, however, reach very different conclusions, as one set of studies strongly suggests that women and men respond similarly to incentives, while another finds just the opposite. Gilsdorf and Sukhatme (2013) asked whether women improve their performance when faced with the size of the purse (the total prize money for the tournament) and the steepness of the prize gradient (how much of the total purse goes to the golfers who finish first, second, and so on). They hypothesized that women would respond differently from men to tournament settings. Using data they compiled from the 2009 PGA and LPGA professional golf tours, Gilsdorf and Sukhatme (2013) found total scores attained by women and men respond similarly to incentives. Looking specifically at final round scores, they found female golfers respond more positively to incentives than male golfers do. This result directly contradicts an earlier finding by Matthews, Sommers, and Peschiera (2007). They use data from the 2000 LPGA tour and find that higher purses lead to higher, not lower, scores for women. Matthews et  al. (2007) concluded that women do not respond well to the added pressure. Shmanske (2013) took a different tack. Rather than analyze overall scores, he examined how rewards affect driving distance, a key aspect of men’s and women’s performances. Shmanske applied Granger’s (1969) causality tests to men’s

and women’s data from the PGA and LPGA tours from 1992 to 2010. He found that an increase in the purse ‘Granger causes’ greater driving distance for men but that the opposite is true for women. He explained the latter finding by saying that longer drives and generally greater athleticism by men has attracted larger crowds and increased sponsorship.

Evidence from Track and Field Track and field events attracted several economic studies that evaluate gender differences. These studies have reached several different conclusions. Frick (2011) found gender convergence in distance races, while Frick and Scheel (2013) found divergence in sprints. Böheim and Lackner (2015) examined high jump and pole vault competitions and concluded that women avoid high pressure events more than men do. Frick (2011) observed that the relatively small number of women who participate in distance races results in a steeper drop-off in performance among women as one moves down the rankings. As a result, the top female distance runners can avoid head-to-head competition. Frick (2011) showed that this drop-off exists by noting the coefficient of variation in personal-best times is significantly greater in women’s races than in men’s races. Facing a greater depth-of-field, men have not been able to divvy up the competitions. The lack of elite competition, in turn, allowed the winning times of women’s races to lag behind those of men. Over time, however, the overall quality of women runners has improved, forcing elite women to compete with one another more frequently. The result has been a steady diminution in the gap between the winning times in distance races involving women and men. Frick and Scheel (2013) performed a similar analysis of 100-meter races and reached the opposite conclusion of Frick (2011). As with the distance races, Frick and Scheel (2013) found much greater heterogeneity among women sprinters than among men, which has allowed the best female runners to avoid competing with one another more than has been possible for men. Unlike distance running, however, the gender gap in sprint times has been rising rather than falling. Rather than attribute the growing difference to gender preferences, Frick and Scheel (2013) asserted that the cause is likely to be performanceenhancing drugs (PEDs). Since steroids are more detectable in women, PED-use has been more prevalent among male sprinters compared to female sprinters. This difference in PED detection

Is There a Gender Difference in the Response to Competitive Settings?

among sprinters is not the case in distance running, where the PED of choice, EPO, is not more detectable among women and hence is an ‘equal opportunity’ drug. Böheim and Lackner (2015) asked whether women avoid risky situations more than men do by comparing how likely the two are to pass in high jump and pole vault competitions. Unlike most other track and field events, competitors in these two events can choose strategically when to compete. A jumper can ‘pass’ at a given height, conserving energy for higher attempts and avoiding failed attempts at lower heights. Misses at lower heights can harm a jumper, as misses at lower heights serve as a tie-breaker when jumpers are tied at the final height. Passing also involves a risk, as the probability of failure increases with the height of the jump. A jumper with three failed jumps at a given height is disqualified. Using data from 305 indoor and outdoor competitions, Böheim and Lackner (2015) estimated linear probability models of passing for women and men, holding constant such factors as the current height; the number of previous jumps, competitors, and previous failures; and the nature of the competition (e.g., whether it was during the Olympics). They found that women were 7% less likely to pass in the second round of a high jump competition, a difference that steadily declines until it becomes insignificant in round nine. Women were 12% less likely to pass in the pole vault, a difference that disappears in the fifth round. They explained the quicker drop-off in the pole vault by noting that most pole vault competitions do not last more than six rounds, while high jump competitions typically last much longer. In addition, Böheim and Lackner (2015) found jumpers were more likely to succeed in a round immediately following a pass and that this difference is greater for women than it is for men. Hence, jumpers – particularly women jumpers – would increase their likelihood of winning a competition by passing more than they currently do. They attributed the reluctance of women to pass to their greater risk aversion.

Evidence from Winter Sports Generally, winter sports received relatively little attention from sports economists. Thus, it is no surprise that only a few papers look at women’s responses to incentives. Che and Humphreys (2013) found some evidence that female skiers respond to incentives, though they have no comparable results for men. Leeds and Leeds (2013)

521

provided unambiguous evidence that women respond more strongly than men to incentives in figure skating competition and that women do not avoid highly-contested events. The evidence from winter sports therefore contradicts the experimental findings noted earlier in this chapter. Che and Humphreys (2013) studied the performance of women in the downhill and giant slalom competitions in the Fédération Internationale de Ski’s Alpine World Cup Tour. In the downhill, the skier makes relatively few turns over a course that has a steep vertical drop. In the giant slalom, the skier must make many more turns between ‘gates’ over a much flatter course. They applied pooled ordinary least squares (POLS) to an unbalanced panel, with data covering the 2001–2002 through 2010–2011 seasons. Overall, Che and Humphreys (2013) found greater differences in prizes led to faster times, though their result does not apply to all specifications of the prize gradient. They also revealed the size of the purse is positively related to finish times in the downhill events, but not in giant slalom events. While the different effect of prize money could reflect systematic differences in the preferences of the skiers, enough skiers participated in both events to make this conclusion unlikely. It is more probable that the different outcomes stem from the natures of the events themselves. Because the downhill event is inherently a ‘speed’ event, a higher prize places a greater premium on getting down the hill quickly. The greater technical demands of the giant slalom event mean the incentive to go more quickly may be offset by an increase in the desire to be certain that one passes through the gates that mark the curves. Leeds and Leeds (2013) analyzed several aspects of incentives in figure skating. Because the most important figure skating events – national, continental or world championships, and the Olympics – do not offer an explicit prize, one cannot directly measure how monetary incentives affect performance. However, one can compare the performances of skaters who are higher up the leaderboard with those skaters who are farther down. While there is no explicit prize gradient, a better finish, particularly in high-profile events, can significantly increase the pecuniary and nonpecuniary rewards. Leeds and Leeds (2013) also compared performance in high-stakes competitions with competitions that are less prominent. They use data from the 2009–2010 figure skating season to regress skaters’ scores in the free skate on a series of control variables, including dummy variables indicating whether the competition is the World Championship or the Olympics. The women’s regressions show that, all else being

522

THE SAGE HANDBOOK OF SPORTS ECONOMICS

equal, women who are closer to the leader after the Short Program – and hence have more to gain from performing well – score more highly in the Free Skate. In addition, women score better in the high-profile Olympics and World Championships than in the other tour events. Estimating the same regression model for men yields very different results compared to women, as male skaters do not respond positively to high pressure situations. Men who are higher up the leaderboard entering the Free Skate do not perform better. In fact, male skaters respond negatively to the pressure of being among the final six skaters – those skaters who have the best chance of finishing with a gold, silver, or bronze medal. Also, unlike women, men show no improvement in performance when they perform in highstakes events, such as the Olympics or World Championships. Finally, unlike Frick’s (2011) findings for distance runners, Leeds and Leeds (2013) found female figure skaters do not enter events strategically to avoid competing with one another. Specifically, the coefficient of variation in skaters’ standing in the previous season (as measured by International Skate Union points in 2008–2009) is not consistently higher for women than for men. This finding suggests one of two things. First, female figure skaters might behave differently from the female distance runners studied by Frick (2011). The second possibility is female figure skaters do not have enough events for them to be able to avoid other elite competitors.

Evidence from Basketball As noted earlier, virtually all the evidence regarding gender differences in response to incentives comes from individual sports. One notable exception to this pattern is Böheim, Freudenthaler, and Lackner’s (2016) study comparing the behavior of men in the National Basketball Association (NBA) and women in the Women’s National Basketball Association (WNBA). They found that men engage in significantly riskier behavior than women do in the final minutes of close games. There is, however, reason to doubt whether this behavior reflects the reluctance of women to engage in risky behavior. Böheim et al. (2016) used data from post-season (playoff) games in the NBA and WNBA for the 2002–2003 through 2013–2014 seasons. Playoff games are particularly relevant because of the high stakes associated with individual games in a short playoff series. Success in these games, and in the playoffs overall, can lead to higher payments for

winning teams, higher future salaries for players on winning teams, as well as the personal satisfaction from advancing to the next round. Like Paserman (2010), Böheim et  al. (2016) focused on particularly crucial moments in the game. They restricted their attention to the last minute of play when neither team had more than a two-point lead. They compared the likelihood of men and women to attempt a three-point field goal in this situation. Because a three-point shot takes place farther from the basket, it is generally a riskier play. However, it comes with a higher reward because, if successful, scoring three points makes it harder for the other team to respond. Böheim et al. (2016) found striking differences in the behavior of men and women. Players on a men’s team that is trailing in the closing seconds of a close match were more likely to take the riskier three-point shot than at other points in the game. Women, however, were less likely to attempt a three-point shot at these crucial points of the match. The authors naturally concluded that this result shows female basketball players to be more risk averse than their male counterparts. This difference in behavior, however, could result from sources other than differences in risk aversion. As was the case with tennis, it is difficult to separate the actions of one team from the actions of the other team. The failure of a women’s team to take three-point shots could reflect superior play by the defensive team in preventing three-point shots. In addition, coaches often have a more direct influence on team play than on individual play. Once a tennis match, golf round, or skating routine begins, a player’s coach is typically reduced to being a spectator. Plays in basketball games, particularly those in the closing minutes of a close game, are often set pieces that are carefully choreographed by the coach, with only limited room for improvisation by the players. Differences in play could therefore reflect differences in attitudes of the coaches rather than the players.

CONCLUSIONS Economic experiments almost uniformly conclude that (1) women respond negatively to competitive situations, while men respond positively, and (2) women are more likely to become discouraged by poor performance or even random luck. Finally, given the choice, women in experiments avoid competitive situations. These findings paint a dire picture of what faces women in the labor market. The lumpiness of

Is There a Gender Difference in the Response to Competitive Settings?

rewards in the labor market – small differences in performance can lead one worker to be promoted with another left behind – mean that workers often face the kind of rank-order tournament that leads to the differences in behavior described above. These differences, in turn, would lead to significant gender differences in pay and position, even in a world without discrimination. This chapter shows that non-experimental results from a variety of sports do not support the experimental findings. The individual sports provide no clear evidence that women and men behave differently in the face of high-stakes competition. Studies of golf and tennis show no clear pattern, with some studies finding gender differences and others finding none. Analyses of track and field return different results for specific events. Distance runs show convergence in behavior while sprints and jumping show clear differences. The winter sports show that women respond positively to incentives. In the case of figure skating, they respond more positively than men do. In the one example from team sports, women appear to avoid risky behavior while men embrace it. However, it is harder to identify player behavior in basketball than in the individual sports. Shot selection could reflect play by the defending team as much as it does play by the offense. It could also reflect the dictates of the coach rather than choices by the players. In sum, the strong findings of the experimental literature do not uniformly extend to sports. This overall conclusion has implications for professional and amateur sports, as one cannot conclude that reward structures elicit different behavior by men and women. To the degree that sport settings provide a strong empirical setting that can be generalized to the overall labor market (e.g., Kahn, 2000), it also indicates that policy makers should be cautious about extending the experimental results to the labor market in general.

Notes I thank Elinor Dittes for her excellent research assistance. 1  While African American women earned 90% of what African American men did ($34,426 versus $38,243), this is largely because African American men earned far less than white men. 2  Whether these preferences are biologically or socially determined is beyond the scope of economic research. 3  A general debate on experiments in social science appear in the August 2018 issue of Social Science & Medicine.

523

4  Paserman applies dynamic programming to strokeby-stroke data to create an ‘importance’ variable. To do this, he calculates the impact of every point in every match on the eventual outcome of each match. 5  Aside from Grand Slam events, women and men both play best-of-three sets.

REFERENCES Annala, C., & Winfree, J. (2011). Salary distribution and team performance in Major League Baseball. Sport Management Review, 14, 167–175. Attali, Y., Neeman, Z., & Schlosser, A. (2011). Rise to the challenge or not give a damn: Differential performance in high vs. low-stakes tests. IZA Discussion Paper No. 5693. Bonn: Institute for the Study of Labor (IZA). Retrieved from: http://ftp.iza.org/ dp5693.pdf Banko, L., Leeds, E.M., & Leeds, M.A. (2016). Gender differences in response to setbacks: Evidence from professional tennis. Social Science Quarterly, 97, 161–176. Bognanno, M. (2001). Corporate tournaments. Journal of Labor Economics, 19, 290–315. Böheim, R., Freudenthaler, C., & Lackner, M. (2016). Gender differences in risk-taking: Evidence from professional basketball. IZA Discussion Paper No. 10011. Bonn: Institute for the Study of Labor (IZA). Böheim, R., & Lackner, M. (2015). Gender and risktaking: Evidence from jumping competitions. Journal of the Royal Statistical Society, 178, 883–902. Booth, A., & Nolen, P. (2012). Choosing to compete: How different are girls and boys? Journal of Economic Behavior and Organization, 81, 542–555. Brown, J. (2011). Quitters never win: The (adverse) incentive effects of competing with superstars. Journal of Political Economy, 119, 982–1013. Che, X., & Humphreys, B. (2013). Earnings and performance in women’s skiing. In E.M. Leeds & M. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 115–131). Northampton, MA: Edward Elgar. Croson, R., & Gneezy, U. (2009). Gender differences in preferences. Journal of Economic Literature, 47, 1–27. Das, A. (2016). Pay disparity in US Soccer? It’s complicated. New York Times, April 21. Retrieved from: www.nytimes.com/2016/04/22/sports/soccer/ usmnt-uswnt-soccer-equalpay.html?em_pos=large& emc=edit_sp_20160421&nl=sports&nlid=222373 1&ref=sports&ref=headline&te=1&_r=1. Deaton, A., & Cartwright, N. (2018). Understanding and misunderstanding randomized control trials. Social Science & Medicine, 210, 2–21.

524

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Depken, C. (2000). Wage disparity and team productivity: Evidence from Major League Baseball. Economics Letters, 61, 87–92. Ehrenberg, R.G., & Bognanno, M.L. (1990a). Do tournaments have incentive effects? Journal of Political Economy, 98, 1307–1324. Ehrenberg, R.G., & Bognanno, M.L. (1990b). The incentive effects of tournaments revisited: Evidence from the European PGA Tour. Industrial and Labor Relations Review, 43, 74–88. Frick, B. (2011). Gender differences in competitiveness: Empirical evidence from professional distance running. Labour Economics, 18, 389–398. Frick, B., & Scheel, F. (2013). Gender differences in competitiveness: Evidence from 100m races. In E.M. Leeds & M. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 293–318). Northampton, MA: Edward Elgar. Garber, G. (2017). Roger Federer beats longtime rival Rafael Nadal, wins 18th Grand Slam. ESPN.com, January 29. Retrieved from: www.espn.com/tennis/ story/_/id/18577366/roger-federer-defeats-rafaelnadal-win-australian-open-title Gil, D., & Prowse, V. (2014). Gender differences and dynamics: The role of luck. Quantitative Economics, 5, 351–376. Gilsdorf, K., & Sukhatme, V. (2008a). Tournament incentives and match outcomes in women’s professional tennis. Applied Economics, 40, 2405–2412. Gilsdorf, K., & Sukhatme, V. (2008b). Testing Rosen’s Sequential Elimination Tournament Model: Incentives and player performance in professional tennis. Journal of Sports Economics, 9, 287–303. Gilsdorf, K., & Sukhatme, V. (2013). Gender differences in response to incentives: Some new results from golf. In E.M. Leeds & M. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 92–114). Northampton, MA: Edward Elgar. Gneezy, U., Niederele, M., & Rustichini, A. (2003). Performance in competitive environments: Gender differences. Quarterly Journal of Economics, 118, 1049–1074. Gneezy, U., & Rustichini, A. (2004). Gender and competition at a young age. American Economic Review, 94, 377–381. Granger, C.W.J. (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37(3), 424–438. Jewell, R.T., & Molina, D. (2004). Productive efficiency and salary distribution: The case of U.S. Major League Baseball. Scottish Journal of Political Economy, 51, 127–142. Jones, M. (2017). Australian Open 2017 prize money: Complete purse and earnings from Melbourne. Bleacher Report, January 15. Retrieved from: https://bleacherreport.com/articles/2687031-

australian-open-2017-prize-money-completepurse-and-earnings-from-melbourne Jurajda, Š., & Münich, D. (2011). Gender gap in performance under competitive pressure: Admission to Czech universities. American Economic Review, 101, 514–518. Kahn, L.M. (2000). The sports business as a labor market laboratory. Journal of Economic Perspectives, 14, 75–94. Kamada, T., & Katayama, H. (2014). Team performance and within-team salary disparity: An analysis of Nippon Professional Baseball. Economics Bulletin, 34, 144–151. Krumer, A., Rosenboim, M., & Shapir, O.M. (2016). Gender, competitiveness, and physical characteristics: Evidence from professional tennis. Journal of Sports Economics, 17, 234–259. Lazear, E.P. (1989). Pay equality and industrial politics. Journal of Political Economy, 97, 561–580. Lazear, E.P., & Rosen, S. (1981). Rank order tournaments as optimal labor contracts. Journal of Political Economy, 89, 841–864. Leeds, E.M., & Leeds, M.A. (2013). Do men and women respond differently to pressure? Evidence from figure skating. In E.M. Leeds & M. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 319–344). Northampton, MA: Edward Elgar. Lynch, J.G., & Zax, J.S. (2000). The rewards to running: Prize structure and performance in professional road racing. Journal of Sports Economics, 1, 323–340. Magnus, J., & Klaassen, F. (1999). On the advantage of serving first in a tennis tournament: Four years at Wimbledon. Journal of the Royal Statistical Society Series D (The Statistician), 48, 247–256. Matthews, P.H., Sommers, P.M., & Peschiera, F.J. (2007). Incentives and superstars on the LPGA Tour. Applied Economics, 39, 87–94. O’Keeffe, M., Viscusi, W.K., & Zeckhauser, R.J. (1984). Economic contests: Comparative reward schemes. Journal of Labor Economics, 2, 27–56. Ors, E., Palomino, F., & Peyrache, E. (2013). Performance gender gap: Does competition matter? Journal of Labor Economics, 31, 443–499. Paserman, M.D. (2010). Gender differences in performance? Evidence from professional tennis players. Unpublished manuscript. Rosen, S. (1986). Prizes and incentives in elimination tournaments. American Economic Review, 76, 701–715. Shmanske, S. (2013). Gender and skill convergence in professional golf. In E.M. Leeds & M. Leeds (Eds.), Handbook on the Economics of Women’s Sports (pp. 73–91). Northampton, MA: Edward Elgar. Smith, A. (1974). The Wealth of Nations. New York: Sheba Blake.

51 Dynamic Pricing in Sports Rodney J. Paul

Ticket pricing for sporting events was forever changed by the introduction of dynamic pricing in Major League Baseball (MLB) by the San Francisco Giants for the 2009 season (Branch, 2008). The practice quickly spread from the early adopters (e.g., San Francisco Giants) to others across the baseball landscape, and teams in other sports leagues were soon to follow. Dynamic pricing, as it is used in sports, is like pricing systems used in the airline and hotel industries (Paul & Weinbach, 2013). Prices of tickets by teams who use dynamic pricing are generally allowed, with some internal restrictions, to move freely based upon consumer demand and the remaining supply of tickets at any moment in time. Although the adoption of this form of ticket pricing was reported as revolutionary, many consumers in the marketplace were not overwhelmed by its introduction as they implicitly understood demandbased pricing from the secondary market for sport tickets from resellers on websites such as StubHub. In an interview with Athletic Business, Russ Stanley, managing vice president of ticket services and client relations of the San Francisco Giants, was quoted as it relates to sports tickets being viewed as a commodity: ‘I think people understand. They get it.’ Dynamic pricing of sports tickets is in many ways a natural progression from variable pricing,

a system virtually all North American professional sports teams used (particularly after the National Football League (NFL) eventually allows for its usage in 2014 (Kaplan, 2015)) in some form with varying levels of sophistication. Variable pricing involves setting different prices for different games on the schedule when tickets are offered to the public, typically before the start of the season (Rascher & Schwarz, 2012). Once released, under variable pricing, these prices do not fluctuate at the primary point of sale from the teams. Differences in ticket prices in variable pricing are based upon the opponent, day of the week, time of the season, and other factors easily recognizable in advance by the teams (e.g., Rascher et al., 2007). The difference between dynamic pricing and variable pricing is that, under dynamic pricing, the prices will move throughout the season up to the point of the start of the game. Whereas variable pricing remains fixed, dynamic pricing leads to ticket prices changing based upon real-time information. In some cases, this information could not have been known in advance and would not have been incorporated into ticket prices in a variable pricing system. A key issue as it relates to the adoption and use of dynamic pricing in sports stems from the advantages and disadvantages of dynamic pricing

526

THE SAGE HANDBOOK OF SPORTS ECONOMICS

compared to variable pricing. The major differences in the two systems are based upon expectations and strategic design. Potential advantages of dynamic over variable pricing stem from gains in revenues that will occur if a team performs above expectations or other favorable conditions exist which increase ticket prices due to high demand. In addition, dynamic pricing allows for the discounting of tickets in the unexpected case of a team performing poorly (or other adverse factors), which will lead to higher revenues than would have been generated if prices did not change. Dynamic pricing also offers the potential advantage that if fans realize ticket prices could rise in the future, it may encourage more advanced purchases and sales of partial and full season ticket plans. The downside of dynamic pricing is that the team incurs transactions costs as this service is outsourced and can result in costs that are considerably above those of setting prices in a variable pricing system. In addition, the strategy could confuse or alienate certain fans, which may negatively influence current and future sales. Another disadvantage is that care needs to be taken to not decrease prices in adverse conditions below the season ticket price. If prices fall below this level, it risks angering the season ticket base, who may not choose to renew their season tickets. Therefore, setting a price floor in this case is likely the optimal long-run strategy. In an interview with CIO.com, members of the Giants front office reviewed the challenges of implementing dynamic pricing. Fears of alienating season ticket holders were discussed. In the article, it is explained how the strategy of the San Francisco Giants is to initially set prices low and then allow them to rise with demand. Russ Stanley noted: ‘We don’t want season-ticket holders to feel like they made a bad investment.’ This chapter gives an overview of dynamic pricing as it is used by sports teams. A very brief background is given on dynamic pricing in nonsports markets, but this brief background is mainly in the form of key articles and literature reviews of the subject. This section is followed by a discussion of studies that examine dynamic pricing in secondary markets for tickets and is subsequently followed by an overview of the articles directly related to the selling of dynamically-priced tickets directly from teams. A conclusion and suggestions for future research close the chapter.

BRIEF BACKGROUND ON DYNAMIC PRICING OUTSIDE OF SPORTS Although sports are a relatively recent adopter of dynamic pricing, the practice has existed for many

years in other industries where a good is perish­ able and the setting exhibits a fixed capacity. In general, Bremaud (1980) and Gallego and van Ryzin (1994) modeled the continuous-pricing problem as it relates to dynamic pricing. They found the greater the duration remaining in which to sell the product, the higher the optimal price. In addition, they found the optimal price is lower as inventory of the product increases. As noted in the introduction, this pricing strategy is most commonly associated with the airline and hotel industries. Examples of research on the use of dynamic pricing in the airline industry was conducted by Belobaba (1987, 1989), Burger and Fuchs (2004), and Escobari (2009), while research on the hotel industry has been performed by Choi and Mattila (2004). In addition to these two commonly associated industries, previous research on dynamic pricing took place in the restaurant setting (Kimes & Robson, 2004; Kimes & Wirtz, 2002) along with entertainment markets like amusement parks and golf courses (Heo & Lee, 2009; Kimes & Wirtz, 2003). Given the focus of this chapter is on the use of dynamic pricing in the sports industry, the theory and research on the topic is deferred to excellent literature reviews about dynamic pricing. Elmaghraby and Keskinocak (2003) provided an overview of the literature where they distinguish between posted-price and price-discovery market research. Gonsch, Klein, Neugebauer, and Steinhardt (2013) offered a nice summary of the growing body of literature on dynamic pricing with strategic customers. This literature review not only notes the theme and industry of the various papers on the subject matter, but also usefully classifies the literature into categories such as pricing policy, demand arrival process, information setting, and others, which serves as a valuable resource to researchers in the field.

DYNAMIC PRICING IN THE SECONDARY MARKET FOR SPORTS Although the focus of this chapter is on the use of dynamic pricing by teams in professional sports leagues, a variety of insights into the nature of dynamic pricing from secondary market sellers was revealed in Sweeting (2012), who gathered 2007 MLB ticket prices of tickets from both eBay and StubHub. A feature of the model of dynamic pricing of a perishable good, such as a ticket to a baseball game, is declining prices as the event approaches.

Dynamic Pricing in Sports

The price of a perishable good should continue to decline if demand does not become less elastic during this time segment. In this case, the day and time of the game is the end point for analysis as the ticket does not have value after the game has been played. Sweeting (2012) studied the quantity of sales for each game in the sample as it related to ticket price for a range of seating options which could be separated into expensive and inexpensive seats for the game. In the model, team performance variables were included in addition to the number of competing listings of tickets, and other factors. The key test for the study was if prices decline as the game approaches and if there is a difference in the stated declines between the classifications of expensive and inexpensive seats. The study revealed that prices did decline over time as game day approached. In addition, Sweeting (2012) found that both expensive and inexpensive seats had their prices cut in a similar fashion by about the same amount. An important point brought to light in this research, particularly as the author notes on the side of the expensive tickets, is that prices will fall even further if the owner of the ticket has little to no value in attending the game. For fans who do enjoy going to the game, when prices fall below a certain level, they obtain more utility by attending the game than by selling the ticket. Two other key points revealed in Sweeting’s (2012) work are important to the understanding of the secondary market for tickets and other markets using dynamic pricing. One, the results revealed that sellers face demand that was time invariant. The second point was that current prices of the tickets had no effect on the value of trying to resell the tickets in the future. Another paper with key insights for the selling of sports tickets in the secondary market was conducted by Drayer, Rascher, and McEvoy (2012). In their study, they investigated the secondary market for NFL tickets for the 2007 season. Drayer et al.’s (2012) research aimed to determine the answers to four questions about the secondary market for sports tickets. The four research questions related to the determining of factors important in predicting secondary market prices, quantity of sales in the secondary market, the consumer surplus generated in purchasing tickets in the primary market, and the level of excess demand in the secondary market. Using data from an undisclosed secondary market firm, they could calculate the average sale price per ticket and the number of tickets sold for NFL games. These two variables became the dependent variables in their model using twostage least squares as their estimation technique. In the first stage, they estimated transaction quantity and then used the results to perform the second

527

stage with the transaction price as the dependent variable in their model (Drayer et al., 2012). Drayer et  al. (2012) used a wide variety of explanatory variables to address their research questions and estimate their model. Independent variables included city factors such as population and income, the week of the season, the day of the week of the game, geographic distance between the teams, and team performance variables, including the home and visiting team win percentages and power rankings of the teams. Stadium effects were also included using factors such as a new stadium dummy, the capacity of the stadium, and the percentage capacity of the stadium already sold for the game. In addition, the average face value of the game ticket from the primary market was included along with expectations for the game taken from the betting market point spread. Statistically significant results from the first stage revealed the average face value of the ticket and the local population had positive impacts on the quantity of tickets sold on the secondary market. The percentage of capacity sold for the game had a negative effect on quantity, while two variables representing team quality, the power ranking and the point spread were shown to have negative and significant effects on the quantity of tickets sold. These results revealed that more reselling occurs with higher team quality and greater expectations of uncertainty of outcome. The second stage of the estimation process, with the average price of the ticket as the dependent variable, revealed several important determinants of demand for secondary market NFL tickets. Greater uncertainty of outcome and team quality, estimated by the point spreads and power rankings, were shown to have a negative and significant effect on the price of tickets. Home and visiting team quality were shown to have positive and significant effects. Additional positive effects on secondary ticket prices were found for the average face ticket value, the dummy variable for a new stadium, and the percentage of capacity of tickets sold for the game (Drayer et al., 2012). Beyond these discoveries, Drayer et al. (2012) used their results to estimate the answers to their other research questions on consumer surplus and excess demand. Given the findings from their data set, they estimated about $260,000 per game was generated as consumer surplus above the face value of tickets for the game. In addition, they estimated excess demand of about 20,000 fans based upon face value prices posted in the primary market for NFL tickets. Another important paper in this area is the study by Diehl, Maxcy, and Drayer (2015), who investigates price elasticity in the NFL. Previous

528

THE SAGE HANDBOOK OF SPORTS ECONOMICS

ticket studies using the primary ticket market finds tickets to be inelastic (e.g., Fort, 2004). Diehl et al. (2015) estimate price elasticity using data from a prominent secondary market firm for the stadium as a whole, upper and lower levels of the stadium, and four distinct seat locations. The dependent variable in their model was the quantity of tickets sold and the key variable of interest in the study was the per ticket average price to estimate elasticity. Control variables include a wide range of factors about the stadium, city, teams, and game itself. In terms of the ticket itself, independent variables in the regression model included the face value of the ticket and the concessions price per ticket. Population and per capita income were included to control for the city where the game takes place. Home and visiting team quality were included, based upon factors such as win percentage, probability of a home team win, and probability of a Super Bowl win. Other independent variables include the distance between the home and visiting team cities, seat location, day of the week of the game, team age, and other factors (Diehl et al., 2015). The results of their study indicate the demand for tickets in the secondary market for NFL games was price elastic. This finding was confirmed for the overall sample and for the different subcategories. The elasticity did vary across seat quality, with the best seats being significantly more elastic. In addition to these primary results, the study revealed that concession prices were not statistically significant, which could signify that they do not enter the ticket demand function by consumers of NFL tickets, a potential interesting avenue for future research.

DYNAMIC PRICING IN THE PRIMARY MARKET IN SPORTS As MLB teams began to use dynamic pricing, some aspects of the data became available to analyze for researchers. Paul and Weinbach (2013) captured individual game data from four teams that used dynamic pricing for the 2011 season: San Francisco Giants, St. Louis Cardinals, Chicago White Sox, Houston Astros. The data, obtained from the team websites, consisted of pricing data for each of the seating areas in the local stadium. Since quantity data were not available, only the determinants of pricing could be studied using the information embedded in dynamic ticket prices for these teams. These data were supplemented with information from the team schedule, team performance, daily starting

pitchers, weather, and promotional information for each team. The dependent variable used in the study was the average ticket price on game day for each of the four Major League Baseball teams. The independent variables focused on dummy variables for the day of the week and months of the season, the team win percentage entering the game, the opponent, game-day promotions, starting pitcher, and weather-related variables. The days of the week and months of the season variables revealed a significant premium for Friday and Saturday games, with Sunday games having a premium in some cities. The lowest prices for games were shown to be early in the week, specifically on Monday, Tuesday, and Wednesday. For teams in the pennant race, there was a late season premium in ticket prices. The win percentage, calculated as a moving average throughout the season, was found to have a positive and significant effect on ticket prices as fans were willing to pay a premium for teams that were successful. Key opponents on the schedule led to the highest premiums seen in the sample, with key games such as the Cubs at the White Sox and any games involving the Red Sox or Yankees as the visiting team leading to significantly higher prices. In terms of events and promotions, some premiums on ticket prices were seen for these games. Opening day, merchandise giveaways, fireworks nights, post-game events, and kids’ group nights were all found to increase ticket prices for the teams using dynamic pricing. On the other hand, starting pitchers’ impact were shown to be relatively small, with only a few pitchers having a significant impact on pricing. This finding either could have been a function of the teams in the sample or may reflect that fans are not willing to pay a premium to see the top pitchers perform. Weather-related variables revealed a premium for games played on sunny days, a discount for games played on windy days, and a positive effect on pricing as it relates to game-day temperature for fans in San Francisco. These results showed some potential advantages of the use of dynamic pricing. Premiums for key opponents, weather-related factors, and promotions may not fully be known at the start of the season and could increase revenues when using dynamic pricing. Other results, such as the limited impact of the starting pitcher on price, helps to reveal what fans may truly demand and not only has ramifications for dynamic pricing, but adds to the positives of variable pricing as an alternative system. Shapiro and Drayer (2014) studied the market for dynamic ticket pricing and secondary ticket

Dynamic Pricing in Sports

pricing simultaneously for MLB. Using data from the 2012 season for 12 home games for the San Francisco Giants, the authors compiled data from the team and StubHub to create a data set of ticket prices for these dozen games. Ticket prices were established under different conditions at varying periods before the games were played. The data were arranged so that prices in both the primary and secondary market were compared for similar seats in the stadium. The goal of their research was to identify the significant determinants of price in the dynamic (primary) and secondary ticket marketplaces. Most of the significant determinants of dynamic ticket prices directly purchased from the Giants were also significant as it related to the secondary market on StubHub, although some differences in terms of both magnitude and statistical significance existed (Shapiro & Drayer, 2014). With ticket price in either the primary or secondary market as the dependent variable, the authors created a regression model with explanatory variables that encompassed a wide range of potential determinants. They broke their independent variables into categories that included time (part of season, day of week, time of game, days before the game), game-related variables (which game of series, interleague, division, promotion, and national TV game dummies), environment (temperature, precipitation), team performance (winning percentage, games back, prior playoffs, etc.), individual performance (pitcher and hitter variables, all stars on roster, etc.), ticket-related variables (seat location, game sellout, season ticket price), and market-based variables (tickets available and secondary market price). For both the primary and secondary market, regression model results were shown with and without the season ticket prices included in the model. For the dynamic ticket pricing model in the primary market for the Giants, the key results illustrated that seat location and secondary market prices were important determinants. Team and individual performance variables also were shown to significantly impact ticket prices along with game-related variables such as division games and interleague opponents. Time-related variables were also shown to be important, as the number of days before the game, the start time, and part of the season variables were also statistically significant. In the model without season ticket price, variables such as seat location, time of game, and playoff participation by the opponent took on added importance. The results for the secondary market revealed ticket-related variables, team and individual performance variables, and time-related variables had significant impacts on price. Ticket-related

529

variables included season ticket price, seat location, number of seats available, and sellouts. Team and individual performance variables consisted of current team performance, prior season playoff appearance, divisional opponent, and number of all stars on opponent’s roster. The time-related variables were time of game and part of season. When season ticket price was not included in the secondary market price model, seat location and opponent playoff appearance showed much greater importance. Paul and Weinbach (2016a) expanded upon these previous studies by looking at dynamic ticket pricing in the 2012 MLB, which saw another seven teams adding dynamic pricing as their method of selling tickets. In addition to modeling the closing price, this study focused on the price change between market open and close with data gathered from each team’s website every night following the start of the day’s game. The focus on the movement in ticket prices from open to close deepens the study of the topic by illustrating some of the factors that may be unanticipated at the time of initial sale, which could be captured by dynamic pricing, but would not be feasible in variable pricing. It also illuminates team strategies of setting prices upon release, compared to the price on the day of the game. The movement in price from open to close was the dependent variable in the model, with explanatory variables consisting of the day of the week, month of the season, key opponents, start time, weather, promotions, team quality, and uncertainty of outcome. The key results found that Friday and Saturday games showed an increase in price of around $5–6 from open to close and Mondays were shown to increase by approximately $2–3, all being statistically significant. Results in relation to opponents that were found to be statistically significant included increases in price for games against the Red Sox, Yankees, and Cubs ($5–$10). Division opponents also showed a significant increase ($2) in the sample (Paul & Weinbach, 2016a). Other independent variables with positive and significant increases were afternoon games, cloudy weather, and fireworks nights. Afternoon games saw an increase of about $2–3 on the average, while cloudy weather on game day increased prices by about $2. Fireworks saw the largest change in price from open to close with an average increase of $2–3 per ticket. In relation to individual game effects, fans appeared to be willing to pay for games with a higher level of uncertainty of outcome as the home team win probability, calculated through the betting market odds on the game, was shown to have a negative and significant effect on the change in

530

THE SAGE HANDBOOK OF SPORTS ECONOMICS

ticket prices. They also preferred both the home and road teams to have higher win percentages. Through the focus on price movements from open to close, some key results related to unexpected versus expected results were revealed. In terms of factors that likely could have been anticipated in advance, results for key opponents, division opponents, days of the week, afternoon games, and fireworks nights could all reasonably have been known. This may be corrected over time as dynamic pricing was new to some of these teams. An alternative explanation is the teams were strategically offering an early buyer discount. This strategy is like offering discounts on full and partial season ticket plans or teams charging a premium for day-of-game sales. If this choice is strategic by teams, this strategy would not be expected to change in the future. The study also revealed some factors that could not be known in advance which highlights the advantages of dynamic pricing. Price increases due to weather factors, team success, and uncertainty of outcome are unlikely to be known before the season starts. Each variable was shown to significantly change prices when these conditions were favorable and contribute to additional revenues earned using dynamic pricing. Further highlighting the 2012 MLB season, Paul and Weinbach (2016b) focused on differences in determinants of prices for the highest and lowest dynamically-priced tickets for games. Differences between fans who purchase these tickets could be based upon income levels and opportunity costs of time. The log of the highest and lowest priced ticket were the dependent variables in two models. Some independent variables had a positive and significant effect on both ticket prices. These independent variables were key opponents, temperature, opening day, and weekend games. The key result in relation to these variables, however, was that they did not influence both groups equally. The lowest priced tickets showed a percentage change in price that was 2–3 times greater than the highest priced tickets. Fans purchasing the least expensive tickets to the game were shown to be much more responsive to these factors, compared to the fans purchasing the most expensive tickets. Considerable differences were also revealed between those purchasing the highest priced and lowest priced tickets as it relates to team success and uncertainty of outcome. Those individuals paying the highest ticket prices for the game were shown to prefer more uncertainty of outcome, while fans purchasing the lowest priced tickets to the game were more influenced by the sum of the win percentages of the teams, less by individual game uncertainty.

Some independent variables were shown to only impact the highest priced or lowest priced tickets. Division games were shown to be sold at a premium for the highest priced tickets, signaling a potential preference for known opponents and rivals. Variables that increased prices at the lowest priced levels were afternoon games and some midweek days, while the variables shown to decrease ticket prices were clear days, games with a closed roof, and games in September. One other key factor that influenced the consumers who purchased the lowest priced tickets to the games was fireworks. Fireworks nights were shown to have a negative and significant effect on the lowest priced tickets to these games. It is doubtful those individuals purchasing the lowest priced tickets are not fans of fireworks, but this result could represent strategic behavior on the part of the teams to increase attendance at the stadium to meet a quota or generally please the sponsor of the fireworks shows. Dynamic pricing spread to the sport of ice hockey, as National Hockey League (NHL) teams adopted strategies like baseball teams in pricing tickets. Paul and Weinbach (2017) studied the use of dynamic pricing by three teams in the NHL: Anaheim Ducks, Minnesota Wild, Ottawa Senators. The data on ticket prices was gathered from the team websites and the aim of the study was to differentiate between dynamic and variable pricing. This research began by examining the setting of opening prices by the three teams in the sample to determine if opening prices in a dynamic marketplace are set in the same fashion as variable pricing methods. The study then examined the change in prices from open to close to identify the variables in the marketplace which are anticipated and unanticipated from the perspective of the team. In addition, the study investigates if the team over- or underestimated the demand for certain game attributes and if there was any apparent strategic behavior involving discounts in the setting of opening prices by the teams. The regression model used for opening prices set by the Ducks, Wild, and Senators organizations utilized three specifications with the dependent variable being either the average ticket price, lowest ticket price, or highest ticket price to each home game. The explanatory variables consisted of the opponent, the day of the week, month of the season, and other factors that would only be known at the time of initial sale. The results presented a stark contrast in terms of opening price strategy across the three teams studied. Anaheim had little variation in their opening prices. Minnesota, on the other hand, had a very aggressive pricing strategy with large differences in prices across the schedule. Ottawa staggered the release of their

Dynamic Pricing in Sports

tickets, but also adopted a system of aggressively differentiating prices across games on the schedule (Paul & Weinbach, 2017). The opening price strategy of Anaheim only had slight differences in prices, mostly for certain key opponents (e.g., Los Angeles Kings), and slightly lower prices for games in December and March. Minnesota had large differences in opening prices by opponent and the days of the week, charging premiums for weekend games. The opening strategy of Ottawa appeared to mainly focus on the opponent, with substantial premiums for opponents from other cities in Canada and notable US teams, such as Pittsburgh and Chicago (Paul & Weinbach, 2017). In changing the focus to the change in price from open to close, additional explanatory variables were added to Paul and Weinbach’s (2017) model to incorporate factors which would be very difficult to know before the start of the season. These variables included team success, fights per game, and individual game forecasts for uncertainty of outcome and expected scoring. In terms of the independent variables included in the opening price model, if the impact of these variables were forecasted in the correct manner, they should not see any significant changes from the time of market open to close, unless the team is engaging in some strategic pricing behavior. The results provided different insights across the three teams in the sample. For Anaheim, there was a general upward movement in price for all games, prices were higher for games involving a higher level of uncertainty of outcome, while key opponents and weekend games saw even greater price increases. In Minnesota, the only significant increases in price were seen at the end of the season (March/April), while the only significant decrease was for their game against Pittsburgh. The end-of-the-season results were driven by the team’s successful playoff push following the trade deadline and the Pittsburgh results were likely driven by key injuries to Penguins players. These significant results were from factors that would have been difficult, if not impossible, to predict before the start of the season, clearly illustrating the advantages of dynamic pricing. Ottawa had statistically significant increases in price for some opponents. There were also statistically significant decreases in price for early season games, which were likely influenced by poor team performance to start the season. The key opponents might have been able to be forecasted in advance or could just be a function of unanticipated performances by the opposing teams heading into the home games in Ottawa. The early season woes of the team were likely unanticipated. The use of dynamic pricing allowed the team to

531

discount prices to help with gate revenues compared to static ticket prices. Overall the study of NHL teams illustrated two key points. First, unanticipated factors, good or bad, can be incorporated using dynamic pricing to offer an advantage to the team over variable pricing. The extent of the advantage depends upon the level of unanticipated events throughout the season. Second, teams have a choice if they would like to aggressively price from the start or strategically offer discounts for purchasing early and then allow for prices to increase through the season. While the majority of the research on dynamic pricing looks at North American sports, some recent research looks at dynamic ticket pricing in European football. Kemper and Breuer (2016) looked at Derby County in English football. Their findings, as it relates to time before the sporting event, were consistent with the work of Shapiro and Drayer (2014) but different from findings in other industries. In addition, various control variables presented in previous work, such as seat category and opponent quality, were consistent with early work.

CONCLUSIONS This chapter illustrated some of the early empirical findings on the use of dynamic pricing in professional sports. The adoption of a pricing strategy that was well known in the airline and hotel industry spread to the MLB and moved quickly across professional sports. The key advantages of dynamic pricing are that prices can move in real time, capturing changes in demand based on factors such as team success, upcoming opponent success, star player movement across teams, player or team milestones, the weather, and other factors coupled with the remaining supply inventory of seats for a given game. This pricing system offers upside on revenues in the case of unanticipated changes in demand above and beyond its competing system of pricing in sports, variable pricing. While the additional costs, including explicit payments and transactions costs, are a negative compared to setting prices under a variable pricing system, the rush of teams across leagues to use dynamic pricing gives some evidence that the anticipated gains outweigh the additional costs. Moving forward, research on dynamic pricing has many fronts on which it can estimate the costs and benefits of this system, in addition to deepening our understanding of how teams are using this pricing method. Also, future research can

532

THE SAGE HANDBOOK OF SPORTS ECONOMICS

help to isolate where dynamic pricing in sports has similarities and differences compared to other industries that use this technique. An obvious piece of research that is needed is the estimation of the elasticity of demand and overall revenues associated with dynamic pricing in the primary marketplace. In addition, with data on remaining supply at any point in time, it can be determined if pricing in the primary market fits the characteristics that define dynamic pricing and sees similar results to those discussed in the secondary market, as described by Sweeting (2012). Analyzing these factors may be difficult, however, without a team or teams being willing to share their ticket quantity data associated with the prices in which they were sold. If achieved, however, the findings would go a long way to quantifying the advantages of dynamic pricing over variable pricing. Other issues on which research may focus are determining the systems and strategies that teams use when selling their tickets in a dynamic fashion. How are initial prices set? Do initial prices mimic differences across the schedule in pricing seen previously for the team when they used variable pricing? Are starting prices set in the same manner that other teams set prices under variable pricing? Is there evidence of early purchase discounts? How do these discounts compare to those given to full and partial season ticket holders? Continuing with season tickets, how do season ticket purchases compare pre- and post-adoption of a dynamic pricing system? Another key area to pursue relates to deepening our understanding of fan demand for sports by using dynamic pricing data to determine if fans are homogeneous in terms of their game attribute preferences. Do fans sitting closer to the action value certain attributes, such as uncertainty of outcome, presence of star players, scoring, etc., differently than those sitting further from the action? This research could also be coupled with comparisons to findings from the secondary market to determine if buyers from teams have similar preferences to those who buy from StubHub or other sellers. As teams adopt more and different ways to use dynamic pricing, such as parking, concessions, and merchandise prices, it would be informative to learn best practices in these markets. In addition, determining the consequences of the interaction of the use of dynamic pricing across these different markets within the sports organization would be informative. The future of dynamic pricing in sports would also serve as a rich research area as we as a profession try to decipher how far this pricing phenomenon will spread. How likely is this practice to spread to other areas of the world? How far into the lower divisions or minor leagues

is dynamic pricing beneficial? If the rapid adoption of dynamic pricing across sports and teams is any indication of the spread of this practice in the future, there will be many opportunities for researchers to shed additional light on this important subject to both sports teams and fans.

REFERENCES Belobaba, P. P. (1987). Airline yield management: An overview of seat inventory control. Transportation Science, 21, 63–73. Belobaba, P. P. (1989). Application of a probabilistic decision model to airline seat inventory control. Operations Research, 37, 183–197. Branch, A., Jr (2008). San Francisco Giants to introduce dynamic pricing for some 2009 tickets. Ticket News, December 5. Retrieved from: www.ticketnews. com/2008/12/san-rancisco-giants-to-introducedynamic-pricing-for-some-2009-tickets/ Bremaud, P. (1980). Point Processes and Queues: Martingale Dynamics. New York: Springer. Burger, B., & Fuchs, M. (2004). Dynamic pricing – a future airline business model. Journal of Revenue and Pricing Management, 4, 39–53. Choi, S., & Mattila, A. S. (2004). Hotel revenue management and its impact on customer fairness perceptions. Journal of Revenue and Pricing Management, 2, 303–314. Diehl, M. A., Maxcy, J. G., & Drayer, J. (2015). Price elasticity of demand in the secondary market: Evidence from the National Football League. Journal of Sports Economics, 16, 557–575. Drayer, J., Rascher, D. A., & McEvoy, C. D. (2012). An examination of underlying consumer demand and sport pricing using secondary market data. Sport Management Review, 15, 448–460. Elmaghraby, W., & Keskinocak, P. (2003). Dynamic pricing in the presence of inventory considerations: Research overview, current practices, and future directions. Management Science, 49, 1287–1309. Escobari, D. (2009). Systematic peak-load pricing, congestion premia and demand diverting: Empirical evidence. Economics Letters, 103, 59–61. Fort, R. (2004). Inelastic sports pricing. Managerial and Decision Economics, 25, 87–94. Gallego, G., & van Ryzin, G. (1994). Optimal dynamic pricing of inventories with stochastic demand over finite horizon. Management Science, 40, 999–1020. Gonsch, J., Klein, R., Neugebauer, M., & Steinhardt, C. (2013). Dynamic pricing with strategic customers. Journal of Business Economics, 83, 505–549. Heo, C.Y., & Lee, S. (2009). Application of revenue management practices to the theme park industry.

Dynamic Pricing in Sports

International Journal of Hospitality Management, 28, 446–453. Kaplan, D. (2015). Dynamic ticket pricing makes successful debut in NFL. Sports Business Journal, October 26. Retrieved from: www. sportsbusinessdaily.com/Journal/Issues/2015/1 0 /26 /Leagu es -a nd- Gov e r ning- Bodie s / NF L dynamic.aspx Kemper, C., & Breuer, C. (2016). Dynamic ticket pricing and impact of time: An analysis of price paths of the English soccer club Derby County. European Sport Management Quarterly, 16, 233–253. Kimes, S. E., & Robson, S. K. A. (2004). The impact of restaurant table characteristics on meal duration and spending. Cornell Hotel and Restaurant Administration Quarterly, 45, 333–346. Kimes, S. E., & Wirtz, J. (2002). Perceived fairness of demand-based pricing for restaurants. Cornell Hotel and Restaurant Administration Quarterly, 43, 31–37. Kimes, S. E., & Wirtz, J. (2003). Perceived fairness of revenue management in the golf industry. Journal of Revenue and Pricing Management, 2(1), 332–344. Paul, R. J., & Weinbach, A. P. (2013). Determinants of dynamic pricing premiums in Major League Baseball. Sport Management Quarterly, 22, 152–165. Paul, R. J., & Weinbach, A. P. (2016a). Price adjustments with dynamic pricing in Major League

533

Baseball. Journal of Law and Economics of Sports (Rassegna di Diritto ed Economia dello Sport), 2, 214–237. Paul, R. J., & Weinbach, A. P. (2016b). Using prediction market prices to differentiate factors that influence the highest and lowest priced tickets in dynamic pricing for Major League Baseball. Journal of Prediction Markets, 9, 43–63. Paul, R. J., & Weinbach, A. P. (2017). An exploration of dynamic pricing in the National Hockey League. In B. Frick (Ed.), Breaking the Ice: The Economics of Hockey (pp. 177–197). Cham, Switzerland: Springer. Rascher, D., McEvoy, C., Nagel, M., & Brown, M. (2007). Variable ticket pricing in Major League Baseball. Journal of Sport Management, 21(3), 407–437. Rascher, D. A., & Schwarz, A. D. (2012). Illustrations of price discrimination in baseball. In S. Shmanske & L. H. Kahane (Eds.), The Oxford Handbook of Sports Economics (Vol. 2, pp. 380–399). Oxford: Oxford University Press. Shapiro, S. L., & Drayer, J. (2014). An examination of dynamic ticket pricing and secondary market price determinants in Major League Baseball. Sport Management Review, 17, 145–159. Sweeting, A. (2012). Dynamic pricing behavior in perishable goods markets: Evidence from secondary markets for Major League Baseball tickets. Journal of Political Economy, 120, 1133–1172.

52 Sports Betting David Forrest

HISTORY There has been sports betting for as long as there has been organised sport. Indeed, the emergence of a professional sports sector in the eighteenth century owed much to betting interest. From his study of the development of rules in sport, Vamplew (2007) concluded that, in this early period, ‘primacy in the formation of rules can be attributed to gambling’. Munting (1996) noted that the rules of golf and cricket, two of the most popular modern sports, were in fact written down for the first time by betting interests, which needed rules to make bettors confident in the fairness of contests and to prevent possible ambiguity in terms of which were winning bets.1 But rules, once set down and accepted, also fostered the development of sport because, for example, clubs from different parts of a country could then compete against each other within a common framework of how the game would be conducted whereas, previously, different forms of the sport had evolved in different regions.2 The mutual dependence between sports and betting was therefore evident in the early history of sport. As sports matured, this relationship seems to have become less important and the focus of interest came to be mainly on the potential for manipulation

of sport by those looking for gains in the betting market, achieved by bribing players to underperform. Occasionally this phenomenon blighted even the highest level of sport, as in baseball’s World Series (1919) and in a notorious cricket test match between South Africa and England (2000). Through the twentieth century, many sports were actively hostile to the betting sector, whether from concern over integrity or as part of a moralistic stance. In a short-lived experiment in the 1920s, English football even went so far as to keep fixtures secret until it was too late for bettors to send their wagers to postal bookmakers operating offshore in Belgium at a time when betting was illegal in Great Britain (Munting, 1996). And in America, lobbying by professional and college sports led to the Professional and Amateur Sports Protection Act (1993), which forbade states from legalising sports betting.3 In the twenty-first century, whether for better or worse, the relationship between sports and betting appears to have become much closer again. On the one hand, as will be documented below, many prominent sports have dropped their overt hostility to betting and have sought to exploit it to provide a new revenue stream. On the other hand, reported cases of manipulation for betting gain have become much more frequent and affect

Sports Betting

a wide range of sports.4 Both developments may be related to the great increase in the scale of the sports betting sector since the millennium.

THE SCALE OF SPORTS BETTING ACTIVITY It is hard to assess the scale of sports betting because it operates illegally in many parts of the world, including most of North America and Asia. Nevertheless attempts are made. IRIS (2017) offers recent estimates from CK Consulting, based on reviews of countries accounting for more than 80% of the world population. Its figures appear conservative relative to some from other sources, which it attributes to its care to avoid double-counting. In Asia and increasingly in North America, betting takes place within an agency structure where illegal local bookmakers pass on bets to licensed multinational online operators based in offshore jurisdictions. Inflated estimates of the size of the betting sector will be obtained if the same wager is counted at both the point where it is initially placed and at the final destination (in betting hubs such as Cagayan (Philippines) and Curaçao). According to IRIS estimates, amounts wagered in sports betting5 in 2016 totalled about €475 billion, of which only about 18% was in the legal sector. But this dizzying figure does not represent the economic value of the betting sector because most of it is returned to bettors as winnings. Gross gaming revenue (GGR), the amount operators win from their clients, is the appropriate figure for consumer expenditure on betting products because it is the amount customers leave behind after engagement in the activity (making GGR equivalent to, for example, box office revenue for cinemas). Global GGR was estimated in IRIS (2017) as €30 billion, almost equally divided between legal and illegal sectors.6 The largest national markets were those of China (€8.4 billion) and the USA (€2.2 billion). It might be noted that, while these figures are impressively high, the betting sector which has developed around sport appears to be worth less than some other sports-related markets. For example, Nielsen forecasts7 for the global value of sports sponsorship and media rights in 2017 were $62 billion and $45 billion respectively.8 However, we shall note below that these figures themselves may be so high because of the high level of betting interest in sport.

DEVELOPMENTS IN SPORTS BETTING What is perhaps most remarkable about the size of the sports betting market is its rapid rate of

535

growth. Sport Accord (2011) quotes estimates of GGR from the same source as employed in IRIS (2017). Whereas the IRIS estimate for 2016 was €30 billion, estimates for 2000 and 2010 had been only €6 billion and €19 billion respectively. It is worth noting again that caution must be exercised when accepting such estimates given the difficulty of measuring activity in illegal and unregulated markets. Nevertheless, changes in estimates over the last two decades clearly point to the current scale of sports betting being several orders of magnitude greater than at the millennium. This indeed explains why sport is now much more active in seeking to create revenue streams from betting and why the threat from betting-related manipulation is greater than before. Some of the extraordinary growth may be related to ‘routine’ demand factors, such as the rapid increase in income levels in East Asia, where the taste for sports betting seems to be particularly high. But the change in the scale of activity is plausibly just part of the transformation of the betting industry by the internet. In a prescient paper, Borenstein and Saloner (2001) predicted that online technology would cause the biggest transformation in industries where there is no physical product, such as insurance selling, stockbroking and travel agency services. In such sectors, traditional suppliers are not protected by transport costs, and the ability of consumers to purchase services online from anywhere in the world would create competitive pressures that would drive down prices. Much activity would move from retail outlets to online since online suppliers enjoy lower costs than those which use bricks-and-mortar outlets. Obviously, in betting, there is no physical product. So, in sports betting, one would have expected value-for-money to improve substantially for consumers as online supply emerged, the more so because local provision had often been limited to a state-owned monopolist or (as in Great Britain) to a single permitted betting shop in any one neighbourhood. Indeed, the price of betting fell. Forrest (2012a) calculated the expected loss from random betting (home win/ draw/away win), at a single British bookmaker, on English Premier League football matches in each season from 2000–2001 to 2010–2011. In the first season, the loss was 10.7%, which may be interpreted as the price of betting. By the last season, expected loss had fallen to 6.1%. And, had the bettor placed each random bet at the best odds available for that particular bet from an odds comparison website, his loss would have been only 0.7%. Even given modest demand elasticity, one would have expected such changes in the implicit bookmaker commission to have yielded

536

THE SAGE HANDBOOK OF SPORTS ECONOMICS

substantial market growth even with no change in product specification. But the betting product did change because online technology also enabled the emergence of in-play betting as a new gambling activity. The speed with which bets can be placed through personal computers and mobile telephones, together with the development of algorithms which change odds automatically as data from the venue capture developments in the match, enables the betting operator to offer consumers an interactive experience. For example a customer can respond to events in a game shown on television or streamed to his device as they evolve, placing bets on a variety of propositions (not just the final result of a tennis match, for example, but also on items such as who will win the next game or how many aces will be served). This development represents essentially a new product and it has proven particularly popular in tennis and football. The online betting platform Betfair, a betting exchange which enables peer-to-peer betting,9 reports that 68% of its turnover was in-play rather than pre-play.10 As with the increase in the volume of betting, the shift within it towards in-play presents new opportunities for sport and also elevated threats from match fixing.

COMPLEMENTARITY BETWEEN THE SPORT AND BETTING PRODUCTS In a general sense, it has likely always been the case that the existence of betting markets increases the demand for sport, if only because betting on an event provides extra interest when watching it. To that extent, the remarkable growth in betting should have had some positive spill­ overs on sport. But it is plausible that these spill­ overs will be greater in the contemporary era than before because of changes in the way spectator sport is consumed (with most consumers viewing on television or through streaming services rather than in the stadium) and because of changes in the nature of the betting product (towards interactivity). Ely, Frankel and Kamenica (2015) famously argued that, across a range of entertainments, such as casino gaming, soap operas, mystery novels and indeed sport, consumers’ utility is essentially dependent on the attributes of surprise and suspense. For example, their interest in a drama series is enhanced by the cliff-hanger, such as where an episode ends in the hero facing some mortal peril. This is an example of ‘suspense’: the viewer is eager to know what will happen next. But for this

to be stimulating, he has to care about what will happen. Hence the successful writer will end the episode with the hero rather than the villain facing the mortal peril. In a sports match too, the spectator watches a narrative unfold. In a ‘good’ and closely fought match, the traditional partisan spectator in the home stadium will experience suspense because of uncertainty over what will happen next and he will care because he has an emotional affinity with one of the teams. One could say that fandom involves training oneself to care about a team because then one can extract more utility from a match since one will care what happens next. In much of contemporary sport, the potential revenue from spectators outside the stadium exceeds that from the traditional local audience inside the stadium. Keeping their interest in viewing games is therefore important because the willingness of media companies to pay for rights is dependent on the size of audience. But most of the potential viewers in a national or international market will not have affinity with either team contesting a particular match. They are not stakeholders and will therefore find it hard to manufacture suspense as a means of enhancing their enjoyment of the game. Betting, however, provides a means for any viewer to become a stakeholder by taking a financial interest in a given outcome, making the experience sufficiently attractive for him to join the audience. In this way, betting and watching the event become strongly complementary activities and heavy engagement with betting should promote demand for the sport. There is casual empirical evidence for this notion that an active betting sector produces positive spillovers for sport. For example, the value of the English Premier League (EPL) broadcasting rights per unit of population is highest in Hong Kong and Singapore, both known as strong betting cities.11 More formal evidence of betting interest promoting the demand for televised sport is provided by Salaga and Tainsky (2015), who analysed the within-match evolution of Nielsen Ratings for college football games over five seasons. They found a strong tendency for one-sided games to retain the television audience when the score was close to that indicated by the betting spread (i.e. the game was dead as a sporting contest but the betting outcome was not yet decided). Similarly, ratings were lower for games when the number of points scored had gone over the number quoted by betting houses in the over/under market so that the outcomes of all bets in that market had been decided. That such statistically significant patterns are identified suggests that a sizeable part of the viewership was interested in the betting outcome (even in a country where

Sports Betting

sports betting was illegal for most of the population) and that many lost the motivation to continue viewing when they no longer had a stake in what happened next. As in-play betting extends its reach, it becomes possible for viewers to continue to enhance their sporting experience by betting even when the outcomes of traditional bets made pre-play have been determined. For example, if there is an unexpectedly large number of goals early in a football game, bets placed in the over/under 2.5 total goals market may be settled even though a lot of time remains in the match; but betting-interested viewers could then place fresh bets in-play in, say, the over/under 4.5 goals market. Thus interest in continuing to watch the telecast can be maintained and this is just part of an overall trend towards consumption of a joint sports-betting product which should benefit sport.

EXTERNALITIES Such positive spillovers from betting to sport are not, of course, part of the scenario traditionally presented by sports organisers. Rather, sports lobbyists represent sport as sending a positive externality to the betting sector (consumer surplus for bettors, profit for operators) and receiving a negative externality from the betting sector (integrity risk with its attendant costs in the form of defensive measures and potential for reputational damage). This simpler scenario, which lends natural support to lobbying for a sports right where bookmakers would have to pay a fee for permission to use a particular sports event as the subject for bets, is implicitly accepted in some academic writing as an adequate representation of the relationship between sport and betting. Thus, Dietl and Weingärtner (2014) argue for property rights to be granted to sports organisers as ‘producing institutions’, whereas they characterise betting houses as ‘exploiters’. With such property rights in place (as is the case in Australia), they would then depend on Coasian bargaining to produce a socially optimal level and pattern of use of sporting fixtures for betting purposes. For example, sports might not be willing to sell in-play betting rights for a price which bookmakers were willing to pay because they judge integrity risks to be too high. An alternative mentioned by Dietl and Weingärtner (2014) would be to consider a statutory levy on GGR, such as is applied in British horse racing, where sums raised are allocated to the governing body of the sport.12

537

Implementation of a sports right would be difficult in practice. Sports betting is provided mainly in illegal and unregulated markets and, for example, the bulk of stakes on most sports events taking place in Europe are wagered in Asia. There would be no feasible mechanism for sports organisers to exert property rights over non-licensed betting providers and, for example, if licensed operators withdrew from in-play markets, the demand might shift to the illegal sector, actually exacerbating integrity risk. In any case, the argument in this chapter is that the relationship between sport and betting is more complex. Rather than regarding the situation as one in which sport is exploited by the betting sector, it is argued to be more realistic to view the industries as mutually dependent, the more so given technological change which increasingly facilitates joint consumption of the two products. To be sure, property rights over the use of sports fixtures implicitly lie with the betting industry since, in most jurisdictions, it has no legal obligation to pay for use of fixtures. But emphatically this does not mean that sport cannot capture some of the profits made by betting operators and it has shown itself keen to do so.

DATA AND STREAMING RIGHTS Just as betting can enhance the enjoyment of sport, so following the event as it unfolds can make a wager more interesting for the bettor. Thus it is well understood that showing a match on mainstream television significantly increases the volume of betting on the fixture (for evidence from American basketball and ice hockey, see Paul & Weinbach, 2010). But, while the ability of the bettor to follow the game made even traditional betting products more attractive, the new model of interactive betting in-play positively demands that, at least at some level, the bettor consumes the sport simultaneously with the betting. Consequently, there must be fast transmission of data (or better still, live pictures) from the stadium to the bettor’s device if he is to be able to participate in the in-play market. As in-play moves towards dominating the market, this puts sports organisers in a more powerful position vis-à-vis the betting sector. They may have no ‘betting right’ to sell but they do have effective monopoly power over the supply of data and audio-visual content to betting operators to be used on the platforms where they offer in-play betting on the event.13 This will allow them to capture a share in the revenue that the betting operators derive from the new in-play product.

538

THE SAGE HANDBOOK OF SPORTS ECONOMICS

A number of specialist firms have emerged which acquire media and data rights which then allow them to sell live coverage and up-to-date data directly from the stadium to operators’ betting platforms. This content is fundamental to the operators’ ability to provide in-play betting opportunities. News of such agreements are reported regularly in the trade press. I inspected a sample of such reports from September–October 2016. New rights sales were recorded not only for economically powerful leagues (such as the National Basketball Association – NBA), but also for relatively minor sports, such as bowls and Gaelic football. To illustrate, consider three transactions involving basketball: 1 Sportradar purchased the right to distribute NBA data and audio-visual content to gaming operators and its CEO undertook ‘to place NBA content at the heart of its betting client provisions’.14 2 Genius Sports ‘was selected as exclusive rights holder to distribute the ACB’s Data for betting purposes’ [the ACB is Spain’s basketball league].15 3 Perform Group acquired the right to ‘manage the broadcast and digital rights to all BBL [British Basketball League] games including exclusive streaming and data rights for Perform’s Watch&Bet and Running Ball products’. All 242 BBL games were to be streamed Worldwide on Watch&Bet.16 To the NBA, the reported value of the agreement with Sportradar could be regarded as relatively marginal given its very high income from conventional television. However, for the BBL, a relatively low-level league, the news report described the agreement as a ‘game changer’. For the first time, it could hope to earn money from overseas markets which would accept it as a betting product even if the level of competition was too low for it to sell mainstream television rights. Since the volume of in-play sports betting is increasing and since the online bookmaker market is intensely competitive, such that operators need to offer a comprehensive range of markets, the value of data and streaming rights can be expected to grow. Given that in-play betting involves the joint consumption of sports and betting products, it can be seen as part of a process of convergence between entertainment industries (Lopez-Gonzalez & Griffiths, 2017). This may make the notion that a betting right is necessary to correct the ‘externality’ of integrity risk seem somewhat anachronistic.17 Big (including all Major League sport in

the USA, where, until very recently, nearly all single-event sports betting has been supplied illegally) and small leagues alike have chosen to sell a facility which is necessary to the development of the fastest-growing sports betting product and one which seems to be implicated in perhaps the majority of sports fixes, certainly in football (Van Rompuy, 2015). Evidently, they have revealed that their share of the revenue from in-play betting is adequate to compensate for added integrity risk.18 On the other hand, a major report (Lewis et  al., 2018) commissioned by governing bodies in tennis has challenged the sport to withdraw from data supply agreements on lower-level tournaments because they facilitate betting markets used for widespread manipulation of matches.19

MATCH FIXING Thus far the chapter has focused on the potential benefits for the sports industry from the growth of betting and the emergence of new betting models. Now it considers in more detail the cost: the increase in the prevalence of manipulation of sports competitions for betting gain. As with other crimes, fixing cases can range from petty to major. An example of petty crime is where players themselves manipulate an event to win (often modest) bets placed by themselves or their families. An example was the fixing of an end-of-season French handball match20 where players for the club, which had already secured the Championship, placed bets against their team at local retail outlets, reportedly to get some extra money for their forthcoming holidays. This was an unsophisticated crime where bets were placed in a heavily regulated and monitored market where a spatially concentrated increase in betting volume, most against the local team, was readily picked up and investigated. Examples at the other end of the scale, where multiple matches are manipulated by international organised crime, are the Bochum case (tried in Germany with the defendants found guilty of fixing more than 300 matches in 13 countries) and the Calcioscommesse case in Italy (which involved players from more than 20 clubs across all tiers of Italian football). In each of these cases, nearly all associated betting took place in Asian markets. And the sums of money involved were very substantial. For example, the Italian prosecutor found that the result of one of the many matches implicated had been ‘purchased’ for €600,000, enabling the instigator to win approximately €8 million in the Asian over/ under market on that match (IRIS, 2017).

Sports Betting

The importance of ‘petty’ fixes cannot be discounted because, regardless of the amount of money involved, revelation that a result has been manufactured may still have the potential to lower consumer confidence in the sporting product, resulting in economic damage. Forrest (2012b) notes that widespread fixing led to the collapse of a number of Asian leagues when revelations of manipulation triggered falls in attendance and withdrawal of sponsorship.21 However, player fixing related to betting on their own account is not a new phenomenon (nor is fixing instigated by local criminals). What is new in the contemporary era is international organised crime, working to fix matches on an industrial scale, across national boundaries and sports. This can be linked to the growth of the betting market, described above. Liquidity in markets on even low-level competitions can now reach extraordinary levels, allowing very large wagers to be placed without attracting undue attention. Moreover, the high liquidity sits in barely regulated markets where the source of funds is not traceable even if suspicions about a match arise. High potential profits and low chances of being caught have encouraged organised crime to add fixing to its portfolio of illicit activities (Forrest, 2012b, 2018).

WHERE IS RISK GREATEST? There have been sufficient criminal cases, and cases where, on the basis of suspicious odds movements, monitoring agencies have reported matches as very likely to have been fixed, for some general conclusions to be drawn about where (which competitions, which betting markets) integrity risk is greatest. 1 Although top-level competition (Italian first division football, Indian Premier League cricket) has not been entirely immune, most cases in team sports have been observed where the level of player remuneration is low but the level of liquidity in the betting market is high. In European football, this applies to low-status national leagues (e.g. Sweden, Ireland) and lower-level leagues in big countries (e.g. fifth- and sixth-division football in England, third-tier in Spain). Outside football, English county cricket (which attracts a low domestic audience but a high level of betting in India) and American college basketball (Wolfers, 2006) are further examples where corruption has come to light in environments where there is a

539

lack of congruence between the level of betting activity and the rewards for players. 2 In individual sports, there have been many fixing cases in sports (tennis, badminton, snooker) organised as international tours where a large majority of players struggle to cover the costs of travelling between venues. With little prospect that they can advance to stages of the tournament which deliver significant prize money, many players have proven susceptible to approaches to fix early round matches. Tennis is linked to a large worldwide betting market and badminton and snooker attract particular betting interest in East Asia.22 3 In many football leagues, financial governance is weak and players are often not paid according to contractual commitments. Such players are disproportionately likely to be approached by fixers (FifPro, 2016). Poor governance may also allow owners to enter the industry with the specific intention of manipulating matches through the control they have over players and coaches. 4 In a high proportion of cases, bets associated with fixing are placed in Asian markets. The bets are frequently in-play and, in football, are made almost exclusively in markets related to the final result of the match or the total number of goals. Across sports, spot fixing is largely a myth. Where a player is paid for manipulation of a part of a match, for example provoking the issue of a red card, this will normally be for an action that has a substantial effect on the likely final outcome in terms of result or goals. Only these markets have sufficient liquidity to generate worthwhile profits for professional criminals. For example, it would not be possible to place large bets on the number of red cards.

POLICY Given that risk appears greatest in betting markets on lower-tier competition and in in-play markets, a natural policy to consider is prohibition of betting on minor sports events and of in-play betting. Such measures have been introduced in Australia and France. But this approach seems unpromising in terms of likely benefits and costly in terms of lost consumer surplus (given that, where it is permitted, in-play betting is what bettors increasingly prefer to do). It is unpromising because local fixes with regulated operators are relatively easy to detect anyway (given appropriate monitoring,

540

THE SAGE HANDBOOK OF SPORTS ECONOMICS

such as practised in Great Britain and France), and major fixes are carried out with bets placed in illegal or unregulated markets. Prohibition of what bets can be made, and on what events, in regulated markets even runs the risk of inducing bettors to transfer to unregulated markets, increasing liquidity and the potential gain to fixing. Since most large-scale fixing involves bets placed in effectively unregulated offshore markets serving Asia and America, the greatest gain in protecting sport would come were large jurisdictions, such as China, India and the USA, to legalise and regulate betting. On a micro scale, the benefits of changing to a legal market were illustrated when Great Britain legalised sports betting in 1961. Because the new betting shops offered value-formoney to consumers, they chose to use the new legal sector and the illegal sector then withered away as its liquidity dried up. Fixers could not use the new outlets because activity was scrutinised and operators had their licences to protect. In English football, the once endemic corruption seemed to disappear (Hill, 2010). But the prospects for important markets in East and South Asia opting for legal, regulated sports betting in the near future are limited. And, while a decision by the US Supreme Court has now authorised individual states to permit sports betting, early indications are that many states will treat betting as just a source of tax revenue and they will end up with a product offering insufficient value-for-money to shift significant liquidity out of the illegal sector. Meanwhile, public policy, where there is legal betting, needs to ensure that it is kept free of crime by strong regulation, such as implementation of know-your-customer obligations on operators. But, in essence, there are no regulatory levers to modify the demand for fixes which comes from the existence of unsupervised liquid betting markets in other jurisdictions. It follows that policy must, of necessity, focus on reducing the willingness of players and other sports insiders to supply fixes. This means that the burden of addressing threats of manipulation must fall on sport itself, although public policy has an essential supplementary role to play (e.g. ensuring that Law Enforcement is ready and willing to investigate and prosecute cases). Reducing the willingness of athletes to supply fixes requires addressing incentives. The individual player’s decision on whether to commit the offence will, if he is risk-neutral, depend on the expected value of the loss incurred were he to be caught and punished. This in turn is the product of the probability of detection and the size of the penalty which would be incurred. However,

it would not seem unreasonable to suppose that the typical sports person is risk-loving rather than risk-neutral. Entry into the professional ranks of any sport involves costly investment, usually during adolescence when education as well as leisure time may have to be sacrificed to arduous hours of training. Even then, only a low proportion of those seeking to enter the sport are successful. Having taken that risk is itself suggestive of risk-love. So is the known high prevalence of gambling (and problem gambling) among professional players.23 According to Becker’s (1968) classic model of criminal behaviour, the chance of being detected bulks more highly in the calculus of whether to offend than the size of penalty imposed if individuals are risk-loving. Given that, it appears to be a reasonable working hypothesis that athletes tend towards risk-love, and sport is therefore correct to respond to elevated integrity risk by investing in the infrastructure of detection. Investment in the infrastructure of detection may include sports setting up specialist intelligence units (such as the Tennis Integrity Unit), taking powers to scrutinise players’ bank and mobile telephone accounts,24 for example, and establishing whistle-blower lines. In European football, and increasingly in other sports, there is also costly monitoring of betting markets to detect patterns in betting (most commonly sharp movements in odds during a match that are not reconcilable with the state of the game) that are suggestive of a fixing operation. UEFA funds an external contractor to monitor hundreds of betting websites around the world before and during matches in the top two divisions of each member country’s national league as well as in its own competitions. Initial screening is by the use of algorithms and then betting experts scrutinise events in the betting market and on the field in all matches flagged by the automated system. Finally, up to 1% of matches in any one year are determined as likely to have been manipulated. This process identifies cases to be investigated by the sport. In an independent audit of the system, Forrest and McHale (2015) were unable to reach a conclusion on the sensitivity of the detection system but presented reasons why it was likely to deliver high specificity, i.e. cases labelled as fixes by this application of forensic statistics were indeed very likely to be true cases. Their finding was subsequently accepted by the Court of Arbitration for Sport in a Judgement,25 which saw the exclusion of a club from the 2016–2017 European Champions League on the grounds of involvement in match fixing. While the chance of being caught is important, sport should also consider whether it could

Sports Betting

reduce incentives to accept bribes by working on the loss experienced by athletes found to have engaged in manipulation. Typically, such players face lengthy or permanent exclusion from the sport, with consequent financial loss.26 But this prospective loss might not be great for veteran players, who collectively have accounted for a large share of sportsmen accused of match fixing. The risk of corruption from late-career employees is faced in many other sectors, such as the police and armed forces where the policy response has often been to increase the proportion of remuneration accounted for by deferred pay (Lazear, 1979). In many countries, police officers receive a high pension, which is lost if service ends with a dishonourable discharge rather than retirement. Sport might consider the potential for equivalent payment mechanisms, although there is less scope for this in minor leagues where pay is already low. An extensive literature in sports economics documents that players in individual sports tend to be sensitive to incentives when deciding how much effort to supply. For example, Gilsdorf and Sukhatme (2008) test contest theory (Rosen, 1986) in the context of men’s tennis and report evidence consistent with the hypothesis that players exert more effort the greater the expected gain from progression to the next round. This will include the known prize for winning the present match plus the expected prize revenue from winning in subsequent matches in the tournament. Of course, a fixer offers a reward for not progressing. For players who would have a low probability of advancing far into the tournament, the prize for winning the current match may be too low to justify turning down a fix. At Grand Slam tournaments, the prize for winning in the first round is $30,000–$60,000 (Jetter & Walker, 2017), but will be just a fraction of that in lesser tournaments where betting markets are often still highly liquid. Tennis might review both the structure of prizes in individual tournaments and the overall distribution of income between elite players and the rest. This might test the commitment of organisers to addressing manipulation since there may be a conflict with commercial goals (e.g. there may be priority for attracting at least one very high-ranked player to take part and this will involve using funds for prizes late in the tournament or for paying appearance money to a star). In general, the literature on manipulation of sports competitions has not yet fully explored ideas from the wider literature on the economics of corruption. Seeking to apply ideas from the general field to the specific context of sport represents one potential avenue for future research.

541

Notes   1  Rules could also be related to their impact on the demand for betting. For example, handicapping promotes outcome uncertainty and this might increase not only sporting but also betting interest.   2  The same phenomenon may be observed in the emergence of e-sports today. Betting houses have taken an active role in regularising the sports, with governing bodies established and rules set down to try to safeguard integrity.   3  A grandfather clause allowed betting to continue in Nevada and existing pools-style products to continue in a few other states.   4  It would not be an exaggeration to say that there is now a continuing torrent of match fixes. Readers may, for example, consult the archive of the bi-weekly Interpol Integrity in Sport Bulletin (www.interpol.int/Crime-areas/Crimes-in-sport/ Integrity-in-sport), which reprints news stories about fixing cases around the world.   5  Here, and in official figures in most jurisdictions, ‘sports betting’ is defined to exclude wagering on horse racing. While horse racing is unquestionably a sport, betting on it is nearly always regulated differently from that on other activities (often it is the only betting legally permitted) and this has led to data being collected in different ways and by different agencies.   6  The much higher share of the legal sector in GGR compared with in total stakes reflects that legal betting typically offers much less attractive odds than illegal bookmakers. Many legal operators are state monopolies. Further, they sometimes, as in China, offer only pools betting rather than single-event betting and, as with other long-odds gambling products, consumers appear willing to pay higher operator commissions.   7  www.nielsen.com/uk/en/insights/reports/2017/ commercial-trends-in-sports-2017.html   8  On the other hand, the size of the associated betting market dwarfs the size of the sports market itself at the level of many individual competitions. In the Republic of Georgia, second-tier football matches can attract fewer than 50 spectators but the Interpol Integrity in Sport Bulletin (April 18–May 1, 2017) reported individual bets of the order of $100,000 having been placed. That bookmakers accepted such bets is indicative of the high liquidity in betting markets on even minor European leagues.   9  Itself another innovation among many which have stimulated interest in sports betting. 10  www.betangel.com/blog_wp/2017/02/04/thegrowth-and-growth-of-inplay/ 11  According to estimates in IRIS (2017), annual per capita GGR from sports betting was €187.40 in

542

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Hong Kong, which was more than three times as great as in the next highest jurisdiction (Greece, €58.30). 12  Other jurisdictions have followed other models for correcting for externalities. For example, in Hong Kong and the United States, merger is the favoured solution with the racing authorities given sole rights to offer betting. A severe limitation of this approach is that it creates integrity risk through conflicts of interest. This is the likely explanation of why these jurisdictions allow only pari-mutuel betting where the betting provider has no interest in the outcome of the race. Pari-mutuel betting may be less attractive to consumers than fixed odds betting. 13  Courtsiders (agents of the betting industry who use mobile telephones to provide live information on the progress of a match) are imperfect substitutes, e.g. it is impractical for them to supply high-quality pictures and they bring their own betting-market-integrity risk: they could delay news so that co-conspirators could trade at odds which do not reflect, e.g. a goal that has just been scored. 14  https://sbcnews.co.uk/featurenews/2016/09/22/ sportradar-nets-nba-real-time-data-statisticsglobal-distribution-partnership/ 15  https://sbcnews.co.uk/retail/2016/10/13/spainacb-unlocks-data-value-genius-sports/ 16  https://sbcnews.co.uk/Europe/uk/2016/09/ 30perform-secures-british-basketball-league-bblmedia-rights-distributor-deal/ 17  Even in British racing, where there is a statutory levy on bookmaker GGR, the value of media rights (primarily related to relaying pictures to domestic betting shops and foreign-based operators) was more than twice the revenue inflow from the 10.75% levy in 2014 (Frontier Economics, 2016). 18  There could still be inefficiency from a social perspective if society’s concern for sports integrity were not fully reflected when sports governing bodies chose to make agreements which facilitate betting. 19  Growth in sports betting has also allowed sports to gain sponsorship and advertising income from bookmakers. However, this sort of income is subject to regulatory risk, e.g. in 2017 Australia announced restrictions on betting advertisements during sports events because of concern that association with sport would encourage minors to gamble. 20  http://en.rfi.fr/sports/20150711-french-handballplayer-found-guilty-match-fixing 21  There appears to have been no formal econometric analysis of the effect of betting-related fixing on the demand for sport. However, Buraimo, Migali and Simmons (2016) modelled attendance at Italian football matches and found significant falls

at matches involving clubs which had been sanctioned for involvement in attempts to fix matches to gain advantage in the Championship (rather than to enable gains on the betting market). 22  At the time of writing, table tennis is the latest sport to have generated fixing stories. 23  See, for example, evidence from a large sample of players, across sports and across countries, in Grall-Bronnec et al. (2016). 24  As with measures to detect doping, the exercise of such powers may be judged by some to be a disproportionate intrusion into athletes’ privacy. 25  Available at https://jurisprudence.tas-cas.org/ Shared%20Documents/4650.pdf 26  Sometimes they may, as in cases in English cricket and Australian football, face criminal charges and a prison sentence. However, this is less common because the standard of proof in criminal proceedings is higher than in disciplinary proceedings. Investigation to the depth that would provide evidence to satisfy a criminal court may be very costly in resources and Law Enforcement may not give priority to these cases.

REFERENCES Becker, G. S. (1968). Crime and punishment: An economic approach. Journal of Political Economy, 76(2), 169–217. https://doi.org/10.1086/259394 Borenstein, S., & Saloner, G. (2001). Economics and electronic commerce. Journal of Economic Perspectives, 15(1), 3–12. https:/doi.org/10.1257/ jep.15.1.3 Buraimo, B., Migali, G., & Simmons, R. (2016). An analysis of consumer response to corruption: Italy’s Calciopoli Scandal. Oxford Bulletin of Economics and Statistics, 78(1), 22–41. https://doi.org/10.1111/ obes.12094 Dietl, H., & Weingärtner, C. (2014). Betting scandals and attenuated property rights: How bettingrelated match-fixing can be prevented in future. International Sports Law Journal, 14(1–2), 128–137. https://doi.org/10.1007/s40318-014-0040-x Ely, J., Frankel, A., & Kamenica, E. (2015). Surprise and suspense. Journal of Political Economy, 123(1), 215–260. https://doi.org/10.1086/677350 FifPro (2016). 2016 FifPro Global Employment Report. Retrieved from: www.fifpro.org/news/2016global-professional-football-players-survey/en/ Forrest, D. (2012a). Online gambling: An economics perspective. In R. J. Williams, R. T. Wood, & J. Parke (Eds.), Routledge International Handbook of Internet Gambling (pp. 29–45). Abingdon, UK: Routledge. Forrest, D. (2012b). The threat to football from bettingrelated corruption. International Journal of Sport Finance, 7(2), 99–116.

Sports Betting

Forrest, D. (2018). Match-fixing. In C. Breuer & D. Forrest (Eds.), The Palgrave Handbook on the Economics of Manipulation in Sport (pp. 91–114). Cham, Switzerland: Springer International. Forrest, D., & McHale, I. G. (2015). An Evaluation of Sportradar’s Fraud Detection System. Retrieved from: https://integrity.sportradar.com/wpcontent/ uploads/sites/15/2016/03/Sportradar-Security-Services_ Univsersity-of-Liverpool_An-Evaluation-of-theFDS.pdf Frontier Economics (2016). An Economic Analysis of the Funding of Horseracing. A Report Prepared for the Department for Culture, Media and Sport. London: DCMS. Retrieved from: https://assets. publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/586305/ Frontier_Economics_An_economic_analysis_of_ the_funding_of_horseracing.pdf Gilsdorf, K. F., & Sukhatme, V. A. (2008). Testing Rosen’s sequential elimination tournament model: Incentives and player performance in professional tennis. Journal of Sports Economics, 9(3), 287–303. Grall-Bronnec, M., Caillon, J., Humeau, E., Perrot, B., Remaud, M., Guilleux, A., Rocher, B., Sauvaget, A., & Bouju, G. (2016). Gambling among European professional athletes: Prevalence and associated factors. Journal of Addictive Diseases, 35(4), 278–290. https://doi.org/doi/10.1080/10550887. 2016.1177807 Hill, D. (2010). A critical mass of corruption: Why some football leagues have more match-fixing than others. International Journal of Sports Marketing and Sponsorship, 11(3), 221–235. https:// doi.org/10.1108/IJSMS-11-03-2010-B005 IRIS (2017). Preventing Criminal Risks Linked to the Sports Betting Market. Retrieved from: www. iris-france.org/wpcontent/uploads/2017/06/ PRECRIMBET_2017_FINAL.pdf Jetter, M., & Walker, J. K. (2017). Good girl, bad boy? Evidence consistent with collusion in professional tennis. Southern Economic Journal, 84(1), 155–180. https://doi.org/10.1002/soej.12213

543

Lazear, E. P. (1979). Why is there mandatory retirement? Journal of Political Economy, 87(6), 1261–1284. https://doi.org/10.1086/260835 Lopez-Gonzalez, H., & Griffiths, M. D. (2017). Understanding the convergence of markets in online sports betting. International Review for the Sociology of Sport, 53(7), 807–823. https://doi.org/10.1177/ 1012690216680602. Lewis, A., Wilkinson, B., & Henzelin, M. (2018). Independent Review of Integrity in Tennis: Final Report. Available at http://www.tennisintegrityunit.com/ storage/app/media/Independent%20Reviews/ Final%20Report_191218.pdf Munting, R. (1996). An Economic and Social History of Gambling in Britain and the USA. Manchester, UK: Manchester University Press. Paul, R., & Weinbach, A. (2010). The determinants of betting volume for sports in North America: Evidence of sports betting as consumption in the NBA and NHL. International Journal of Sport Finance, 5(2), 128–140. Rosen, S. (1986). Prizes and incentives in elimination tournaments. American Economic Review, 76(4), 701–715. Retrieved from: www.jstor.org/stable/ 1806068 Salaga, S., & Tainsky, S. (2015). Betting lines and college football television ratings. Economics Letters, 132(July), 112–116. https://doi.org/10.1016/ j.econlet.2015.04.032 Sport Accord (2011). Integrity in Sport: Understanding and Predicting Match Fixing. Moudon, Switzerland: Sport Accord. Vamplew, W. (2007). Playing with the rules: Influences on the development of regulation in sport. International Journal of the History of Sport, 24(7), 843–871. https://doi.org/10.1080/09523360701311745 Van Rompuy, B. (2015). The odds of match fixing: Facts and figures on the integrity risk of certain sports bets. Retrieved from: www.asser.nl/media/ 2623/the-odds-of-matchfixing-report-2015.pdf Wolfers, J. (2006). Point shaving: Corruption in NCAA basketball. American Economic Review, 96(2), 279– 283. https://doi.org/10.1257/000282806777211757

53 The Economics of Doping in Sports: A Special Case of Corruption Eugen Dimant and Christian Deutscher

INTRODUCTION Corruption occurs in various forms and in every area of sports, as the case of the FIFA scandal has taught us. Embezzlement, match fixing or any kind of manipulation of the final outcome are corresponding examples, and, as we will also argue, doping (see Maennig, 2008; Masters, 2015). Indeed, professional sports attract not only individuals’ interest, but also significant money, and thus corruption in sports can create societal and economic burdens. An athlete’s incentives to win are shaped not only by the (expected) inflow of prize money, but also by the increasing prestige and self-fulfillment. Such incentives give rise to crossing legal boundaries in order to gain an edge. In professional sports, doping is a ubiquitous problem. Cases of doping in athletic sports – such as the exposures of Asafa Powell, Tyson Gay, Veronica Campbell-Brown and Sherone Simpson – or revelations of past and current doping cases in cycling undermine the sport’s reputation. In 1991, coaches conceded that steroids propelled the success of East German women’s swimming teams over a period of two decades during the so-called ‘Golden Period’ between the late 1960s and the late 1980s. The East German women’s swimming teams excelled in nearly every contest during this period, culminating in crushing the competition by winning

10 out of 14 gold medals and setting eight world records at the first world swimming championship in 1973 (Janofsky, 1991). Doping introduces frictions into the system through the creation of unwarranted disadvantage against opponents. It also imposes negative externalities on third parties by distorting the level playing field of fair competition and weakens the public trust in the institutions involved. Following this reasoning, doping shares fundamental features with general corruption, as it generates personal gains to the detriment of others. However, corruption is mainly driven by rent seeking and rent extraction behavior, whereas doping is not (see Maennig, 2002; and for a general discussion of the various definitions of corruption as well as the multiple causes and effects of corruption, see Dimant & Schulte, 2016; Dimant, 2019; Dimant & Tosato, 2018). In fact, Masters (2015, p. 113) argued that ‘corruption in sports equates to the deviation from public expectations that sport will be played and administered in an honest manner’, thus validating our approach. This chapter provides an overview of the economic sports corruption literature, with a particular focus on doping. It is worth mentioning that although general corruption is highly relevant to sports, in this chapter we are particularly interested in the antecedents and effects of doping. Our motivation to shed light exclusively on doping is

THE ECONOMICS OF DOPING IN SPORTS: A SPECIAL CASE OF CORRUPTION

that doping strongly affects fair competition, distorts incentives, and causes substantial monetary and non-monetary burdens (Lessig, 2013). Along these lines, we argue that the mechanism of doping entails all the ingredients to distort fair competition and trust in the game, and consequently corrupts the entire sports system. Our aim is to bring together economic theories from both the rational and behavioral spheres to analyze the athlete’s inclination towards doping. We add to the existing discussion by providing some behavioral explanations to doping beyond rational decisionmaking. This assumption makes it particularly difficult to fight doping most effectively and sustainably. We conclude that the fight against doping can only succeed with strong regulatory bodies in place. We make the case that, although doping has an inwardlooking purpose, its negative impacts go beyond the individual level and instead impose detrimental outcomes on the aggregate. We also discuss possible countermeasures that have the potential to reduce the incentives for doping from both the individual and institutional perspectives. In the next section we briefly discuss the main aspects and history of corruption as well as the detrimental effects of doping on the individual and aggregate levels. Then we discuss the rational and behavioral approaches explaining the individual decision-making process with respect to doping. In the following section we discuss the characteristics and detrimental effects of corruption in sports and offer possible countermeasures, and finally, we conclude in the last section.

CORRUPTION IN SPORTS – AN OVERVIEW History and Magnitude People lean towards the belief that corruption in sports has only recently become a problem. However, research indicates that the general attitudes towards doping differ and are culturally rooted (Wagner & Ziebarth, 2014). Cases of general corruption in sports, like those of the former soccer referee Robert Hoyzer, the former professional basketball referee Tim Donaghy, the Italian soccer betting scandal, or the case of Lance Armstrong, are hyped by the media and thus bias our perceptions. A wide range of corruption cases in modern sports have taken place over time, which are comprehensively covered in Maennig (2008). Academics exposed corruption in various sports (for NCAA Basketball, see Wolfers, 2006; for Sumo Wrestling, see Duggan & Levitt, 2002; for soccer, see Deutscher, Dimant, & Humphreys,

545

2017), resulting in policy recommendations that were picked up by decision-makers (Dietl, Lang, & Werner, 2010). However, in some cases, excessive media attention leads to perverted outcomes (for more detail, see Ahrens, 2014). Doping distorts fair competition, dilutes incentives, and harms the reputation of the respective sport. The loss of public and media interest in professional cycling due to uncovered doping cases peaked in the German television stations abandoning the sport almost completely from their program in 2007 (Dilger, Frick, & Tolsdorf, 2007). It is alarming that we don’t have data of sufficient quality to draw sophisticated inferences about the spread of doping (ab)use in sports. Some have even gone so far as to argue that no reliable estimate of the prevalence of doping in elite sports has been published so far (Sottas et al., 2011). Exemplarily, according to official World Anti-Doping Agency (WADA) doping tests, roughly 1% of all professional athletes dope (Tsivou et  al., 2006). This, however, is most likely a vastly biased underestimation, as the number of doping cases is a function of, among other things, the frequency of conducted doping tests and their quality. Consequently, official instances of doping as the result of (infrequent) doping tests do not allow clean inferences to be drawn on the actual pervasiveness of doping. Since empirical evidence is rare, survey results and anecdotal evidence indicate that the percentage of dopers is set to increase at the level of competition in which they compete, at best (Pitsch & Emrich, 2012). Still, some effort has been made to approach this question empirically. Exemplarily, Sottas et al. (2011) indicate that, on average, 14% of the tested athletes used blood doping. The data also suggest a strong heterogeneity of doping among athletes, ranging from 1% to 48% for subsamples stratified according to nationality, sex and type of sport (endurance versus non-endurance). Furthermore, a group of German scholars in 2013 evaluated 2,997 triathletes who participated in various German Ironman races in a triathlondoping study. Results showed that 13% admitted to physical doping (such as steroids, EPO, human growth hormone), 15% admitted to cognitive doping (for example, antidepressants, beta blockers, Modafinil, methylphenidate), and 10% admitted to both physical and cognitive doping (Dietz et al., 2013). It is conceivable that this is a lower-bound result given that professional athletes face higher incentives and have the required skill levels to boost their performance at a margin that makes the difference between mediocrity and superstardom. The results from other (anonymous) surveys draw a different and more heterogeneous picture. For example, the abuse of performance-enhancing drugs for German top-level athletes is estimated

546

THE SAGE HANDBOOK OF SPORTS ECONOMICS

at around 25–50%, whereas the doping prevalence for elite-level sport adolescents is estimated at 7% (Pitsch, Emrich, & Klein, 2007; Striegel, Ulrich, & Simon, 2010).

that athletes are well aware of the fact that doping is dishonest and entails negative side effects (PerettiWatel et al., 2004).

The aggregate perspective Detrimental Effects of Doping In this section, we discuss a range of adverse effects associated with the (mis-)use of doping in competitive sports from both an individual and aggregate perspective. Such a subdivision is necessary since doping imposes different negative externalities at the individual level (mainly a health perspective) and the aggregate level (mainly a cost perspective).

The individual perspective

According to Preston and Szymanski (2003), there are four basic reasons why doping can be harmful to sports in general and athletes in particular: 1 2 3 4

It damages the health of athletes. It gives doped athletes an unfair advantage. It undermines interest in the sport. It undermines the reputation of a sport.

In this section, we will focus on points (1) and (2). The US Anti-Doping Agency (USADA) lists a number of effects associated with performanceenhancing drugs (PEDs) on their website (www. usada.org). For their purpose, the USADA subdivides the PEDs into 11 categories. The negative effects are substantial and vary for different age groups and genders. The side effects range from liver damage and impotence to generally higher risks for strokes and committing suicide. In their intent to improve sporting performance, athletes rely on performance-enhancing supplements to protect health (Nieper, 2005) and improve performance by, for example, reducing recovery time (Erdman et  al., 2007). The magnitude of the impact of PEDs on the athlete’s performance crucially depends on the type of sport and the individual (Alaranta et al., 2006). In particular, the demand for PEDs differs between endurance athletes and athletes involved in activities that demand motor skills. Consequently, one has to be careful in trying to measure universally the magnitude of unfair competitive advantage caused by the usage of PEDs. It is hard to find a direct link between PED usage and performance, because talent, training, technique, and nutrition factor in (de Hon, Kuipers, & Bottenburg, 2015). Hence, field studies do not serve as a feasible approach to test the effectiveness of PEDs on professional sporting performance and published studies predominantly rely on self-reported attitudes and beliefs stated by the athletes. Research indicates

An intertwined reputation system might lead to a perverted equilibrium in which clean athletes are unable to signal their fair sportsmanship (meaning their refusal to dope). Consequently, one might start doping not purely because of their individual selfmaximizing calculus but based on a more sophisticated set of incentives, including peer pressure or behavioral contagion. Consequently, the whole sport might accelerate into deviant behavior that is conditional on both other athletes’ deviant behavior and the (false) public perception. From a cost perspective, performing doping tests entails enormous annual costs to society. The World Anti-Doping Agency (WADA) and Maennig (2014) estimates the costs in 2013 to range between US$229 million and $500 million in order to cover 270,000 doping tests. Given the testing results, every exposed case of violation against the rules costs about $70,000. For an assumed sensitivity of doping tests of about 40% (Hermann & Henneberg, 2013) combined with a very short window of detectability implies (a) high social costs in attempts to convict athletes who dope and, despite these high costs, (b) a residual uncertainty that remains with respect to detecting offenders. Increasing testing frequencies and improved testing methods are feasible measures to detect doping at a higher rate, but this is again costly (Hanstad & Loland, 2009). In addition, the repetitive occurrence of doping has the power to cause a sport to lose its credibility. The best-known case is professional cycling. During the period 1940–2013, more than 600 riders were detected to be cheating (www.cycling4fans. com). In 2007, media attention peaked and led the German Telekom to terminate the sponsorship of its T-Mobile-Team. Television audiences decreased in most European countries (Dilger et  al., 2007) and one year later German television stations abandoned their broadcasting of the Tour de France as a reaction to the doping information being published.

THE DECISION TO DOPE: EXPLAINING BEHAVIOR In what follows, we shed light on an athlete’s decision-making process – before choosing to engage in doping – from various perspectives. First, the individual approach will serve as a starting point to explain deviant behavior, highlighting the

THE ECONOMICS OF DOPING IN SPORTS: A SPECIAL CASE OF CORRUPTION

athlete’s decision as the result of (bounded) rational assessments. Second, the aggregate approach highlights the fact that individual decision-making is subject to the peer group’s decisions and reputation mechanisms, consequently being correlative with the surrounding social environment.

The Individual Perspective Monetary and nonmonetary incentives play a decisive role in an athlete’s calculus to even consider the intake of performance-enhancing drugs that will have harmful consequences in the future. Research confirms the existence of such incentive effects, for example, in golf (see Ehrenberg & Bognanno, 1990a, 1990b). At the same time, evaluating relative performance makes sporting contests prone to sabotaging behavior (see Garicano & Palacios-Huerta, 2005; Deutscher et al., 2013; Deutscher & Schneemann, 2017) and potentially doping (Kräkel, 2007). We analyze an athlete’s doping decision by providing both purely rational and behavioral perspectives that are conducive to the understanding of doping behavior.

A rational approach

If anything, corruption in sports has widely been discussed from a delineative perspective in the literature of sports economics. While the ultimate practice is well understood, the underlying motivations and decision-making processes lack a behavioral perspective. In what follows, we will review contributions that help us understand individual decisions to dope from a rational perspective. In their seminal work, Becker and Murphy (1988) develop a general theory of rational addiction that can easily be adapted to explain individual doping decisions. Hereby, addiction might be the result of either a physical or a social dependency (for example, one’s need to achieve recognition and approval). Approaching this topic from a rational perspective, they argue that rational addiction implies the presence of an active calculus that would consistently maximize utility over time. Their theoretical results suggest that even strong addictions are driven by rational decisions and involve a forward-looking maximization of stable preferences. These restrictions have been eased in subsequent research involving quasi-hyperbolic preferences, leading to an addict’s time inconsistent decisions. Hereby, individuals with high discount rates for future events and thus a high preference for the present are more likely to become addicted. Using a similar approach, Maennig (2008) explicitly models the individual’s decision to engage in crime in a rational risk-assessment style.

547

He transforms Becker’s (1968) general model of crime exertion into the context of corruption in sports and argues that athletes are able to carry out proper risk assessments in order to weigh their expected benefits against the expected costs. This neoclassical approach can also be used to shed light on the athlete’s decision to dope. In line with this cost–benefit perspective, it is important to analyze the impact of rising stakes on the probability of using PEDs. In the more recent past, one can observe a substantial increase in absolute wage disparity for both team and individual sports. Acknowledging an increase of rewards, especially for the top performers, let’s assume the presence of incentives to dope. Prize money development observed throughout the last two decades increased marginal returns to improved performance (and hence doping), especially for top athletes competing for the highest rewards. We argue that the incentive’s strength in manipulating an athlete’s decision as to whether or not to take performance-enhancing drugs is a function of his own age that most likely exhibits a U-shape characteristic. A competitive athlete in his younger years has both the physical condition and sufficient upward leeway to allow for a skill boost large enough to create an edge that makes the difference between mediocrity and superstardom. Under these circumstances, the expected monetary and non-monetary benefits might very well outweigh the risks accompanied by taking PEDs. While this advantage vanishes with the athlete aging, this flattening off is substituted and the initial decline is likely to be overcompensated by what is known as the ‘endgame effect’ at the end of his active career. Here, existing punishment mechanisms, such as exclusion from participation in tournaments, have no credible sanctioning effect on an old athlete who is close to his retirement. Doping decisions are often arranged in line with game theoretical logic. Simply put, athletes are in a prisoner’s dilemma type situation. Both of them would be better off not engaging in doping in the first place. However, as nobody can trust the other, both end up taking drugs in order to enhance their chances to win (at least in the one-shot consideration without allowing for trust and reputation to build up). An extension to this approach is the so-called ‘inspection game’, in which the relationship between athletes and organizations in charge of doping tests is modeled (Berentsen, 2002; Kirstein, 2014). Extending the existing approaches, Büchel, Emrich, and Pohlkamp (2014) introduce the customer as an additional player to the game. The underlying motivation is straightforward as customers play a decisive role in making sports profitable. Once customers lose interest, professional sports might experience a downfall, especially in monetary terms

548

THE SAGE HANDBOOK OF SPORTS ECONOMICS

(cycling, for example). In this vein, the possible threat of customers turning away from sports is included in the athlete’s decision-making sphere. Here, the basic assumption is a trigger strategy on the side of the customers: they provide support for sports until a doping scandal occurs, which will then lead to the withdrawal of support. Under mild assumptions, the unique equilibrium is that athletes dope while organizers underinvest in testing them. The reasoning is that the customer’s threat to withdraw their support leads to a situation where the organizers tolerate (undiscovered) doping rather than run the risk of losing support due to discovered cases of doping. To sum up, the athlete’s opportunity costs of not being able to earn (prize) money, his increasing loss of value as he advances in age, and his exposure to a prisoner’s type of dilemma work in the same direction and serve as an incentive to take performanceenhancing drugs. As the evidence suggests, the combination of these incentives is strong enough to outweigh the disincentives of punishment.

A behavioral approach

In this section, we shed light on individual doping decisions that are driven by behavioral aspects beyond clear-cut rational decision-making. Consequently, some well-known behavioral approaches that are conducive to understanding doping decisions will be discussed. Bounded rationality, an individual’s cognitive limitation due to the presence of depletive resources, unleashes peculiar behavioral patterns. The human brain relies on fundamental patterns, simplifying and accelerating processing, often leading to more intuitive decisions, to which people frequently refer to as ‘gut feelings.’ By any means, humans simultaneously and sequentially implement various techniques: they make mental accounts, organize choices in a lexicographical style, and enforce selective perception. As a result, choices are pondered less deliberatively, potentially inducing inadvertent behavior. One particular example of a person’s cognitive bias is the concept of self-control. As research indicates, self-control is an integral part of continuous decision-making processes, allowing a more deliberate assessment of each situation and facilitating the individual’s capability to resist temptations. Self-control is treated ‘as the capacity of one “more rational” self to override the decisions of a more impulsive one (or several)’ (Achtziger, AlósFerrer, & Wagner, 2015). The underlying idea is that the resources needed to exhibit self-control are the same as those that are used for controlling and restraining thoughts and impulses and persisting in cognitive tasks. As is evidently true, these resources are limited and using them for one task leaves only

a reduced amount of self-control (if any) for subsequent tasks. Using up cognitive resources necessary for self-control induces a state of ego depletion. Along these lines, it is reasonable to assume that being exposed to constant physical and psychic pressure, certain physiological conditions resulting from, in particular, ego depletion, lower the athlete’s self-control and consequently the intrinsic threshold to withstand the temptation of doping. Consequently, athletes who are exposed to extensive physical and psychic strain are more prone to the abuse of performance-enhancing drugs. From a crime perspective, a lack of self-control is perceived to be the driving factor behind deviant behavior (see Hirschi & Gottfredson, 1990; Muraven, Pogarsky, & Shmueli, 2006). However, for self-control to be effective, sufficient mental resources are needed. At short sight, these resources are finite and once they are depleted, the decision-maker ‘gives in’ and resorts to using heuristics rather than a deliberative cost–benefit analysis. In a sports context, this could mean that if the athlete engages in tasks that eat up cognitive resources needed for self-control, subsequent decisions are taken less appraisingly, possibly leading to more inconsiderate outcomes, for example in the form of heuristics. One can easily imagine that professional athletes who are consistently under physical and psychological pressure of various kinds (frequent and intense workout, strict nutritional protocols, restraints and constrictions of various kinds) quite frequently deplete their resources, which might influence their intrinsic inclination towards taking performance-enhancing substances. Experimental evidence points to the idea that impairing cognitive resources leading to ego depletion has a substantial and a prolonging effect on changes in behavior and shows an intensifying character as more decisions are made (Vohs et al., 2008). In such a mental state, an individual resorts to the use of heuristics. One particular heuristic approach that is involved in the individual decision-making process is ‘win-stay, lose-shift’ and is closely connected to habits and standard operating procedures (Nowak & Sigmund, 1993). This concept is particularly helpful to explain why athletes might stick to their previous choice to dope. Such a strategy explains the evolution of certain behavioral patterns and can, in particular, be easily applied to any type of repeated decision problems. In more detail, once a decision in favor of doping was made, as long as the outcome of the last ‘round’ was a success, the ‘player’ will stick to his previous decision. For our purposes, the outcome of the last round may be represented by the outcome of last blood test or contest; the player is the athlete and the decision sphere is denoted by the athlete’s decision about whether or not to dope in a previous contest.

THE ECONOMICS OF DOPING IN SPORTS: A SPECIAL CASE OF CORRUPTION

The Aggregate Perspective

Peer effects: behavioral spillovers and social contagion

In this section, focus lies on the role of conditionality in the individual’s decision-making process with respect to using PEDs. This approach represents a combination of imitating and following other people’s behavior based on complying with possible existing norms. As will be argued, this approach suggests that an individual’s decision to engage in doping might be the result of spillover effects resulting from the peer’s doping decision observed by the individual. Research indicates that crime has severe contagion effects and that behavioral contagion is subject to the social context (for a comprehensive discussion, see Dimant, 2019). Along these lines, an individual’s inclination to behave unethically is a function of the peer group’s unethical behavior (Glaeser, Sacerdote, & Scheinkman, 1996). Theory suggests that individuals are likely to engage in herding behavior by following the (bad) example of their peers. In the context of sports, one might think of a sports team where the usage of doping spills over from one athlete to the next. Ichniowski and Preston (2014) find evidence for the existence of peer effects and spillovers of skill among soccer players. Especially in team sports, spillover effects are likely to appear within members of the same team. In the context of sports, when a sportsperson is exposed to his/her peers’ deviant behavior, such as doping, even an intrinsically motivated honest person might be inclined to conform to the (perceived) social norms in order to avoid being the odd one out. In retrospect, this observation seems to be true of cycling. More often, not only were single professionals convicted of doping, but doping was also found to be institutionalized among the teams, exerting extensive pressure (with respect to both performance and social belonging) on those who do not comply.

Threat of reputation loss

Reputation is an essential and immanent feature of everyday life. In a social context, reputation determines one’s own trustworthiness, and once that is undermined, it is hard to rehabilitate into society. Along these lines, the impact of possible reputation loss might deter individuals from doping in the first place. However, an individual’s reputation depends not only on one’s own behavior but also on the behavior of peers and the group dynamics (Tirole, 1996). In the context of sports, such an approach sheds light on how past group members’ decisions on, for example, a cycling team impact a current team member’s decision with respect to

549

engaging in doping. One of the main results is that under particular conditions, current members of a group are ‘locked-in’ to deviant behavior as a result of past group members’ deviant behavior. Here, the group’s reputation has been damaged in the past and engaging in deviant behavior in the present becomes the best response. While the loss of reputation functions as a threat to the individual and thus should represent a deterrent factor, once the other group members smear reputation, a short-run and long-run corruption steady-state might be reached in which it is not worthwhile to remain faithful and abstain from using PEDs. In consequence, using this approach, one can comprehend that it is not only the individual’s inclination towards misbehavior that is important. Rather, an athlete’s decision to dope is driven by the economic and social context as well as the peers’ behavior.

COUNTERMEASURES TO FIGHT DOPING As argued throughout the chapter, the reasons to dope are diverse and hence must also be the approaches to fight doping. Existing research indicates that a variety of leverage points are conceivable. From a classical cost–benefit perspective, raising the (expected) costs for doping might do the trick, as this, all else being equal, can be expected to reduce the incentives for such deviant behavior in the first place. This can be implemented through both pecuniary penalties in the form of fines and an extended ban from the federation or from any form of competition events on corrupt athletes. The possible loss of reputation represents a strong costdriving factor. If the media sticks together and carries out extensive media coverage, the concomitant costs would rise significantly. One feasible approach is to extend the (randomized) testing of professional athletes for PEDs. In particular, given that athletes depending on their skill level and age exhibit different incentives to dope, such results could give rise to more selective additional testing (for example, of athletes who have just rehabilitated from a severe injury). Preston and Szymanski (2003) argue that although randomized testing would increase the chance of exposing doped athletes and thus increase the (expected) costs as well, most professional US sports organizations have managed to reach agreements, through the player unions, that put a ban on randomized testing. The concomitant costs can be partly internalized through reinforcing incentives for self-reporting, blowing the whistle, and asymmetric punishment (see Innes, 1999; Basu, Basu, &

550

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Cordella, 2014; Maennig, 2014). While increasing the fines to reach an incentive-compatible level to deter athletes from using PEDs might seem to represent a feasible approach in general, excessive fines like the ones imposed on Petar Korda might not bring about the desired outcome (see Maennig, 2014). Further attempts to fight the doping issue involve harsher measures, such as temporarily excluding the tainted sports disciplines from the Olympic program, banning the television broadcast of sports or shifting the cost burden to official institutions of the respective sports (Maennig, 2014). In accordance with the deferred compensation model developed by Lazear (1979), Maennig (2002) proposes a mechanism by which professional athletes would deposit part of their prize money into a fund. This money will be safely stored and paid out after they retire if, in retrospect, they remained clean over the course of their sports career. This in fact might counterbalance the athlete’s incentive to dope induced by the earlier discussed endgame effect. However, it remains questionable whether athletes are willing to take the real loss in income brought about by deferring payments to the future. Since athletes (just as regular people) discount future income and thus prefer consumption now over consumption later, such an approach might distort incentives. In the short term, this is especially true for athletes who are active at the time of the introduction of such measures due to a shift in their reference point. Having been used to a system where they were in charge of their whole prize money, the idea of giving up this prestige might provoke broad rejection. However, in the medium and long term, and especially for young (amateur) athletes, such a one-time cut can be expected to cause a less pronounced rejection. One first step to create an impartial judge has been taken by the German Federal Cabinet by approving an anti-doping law, clearly discerning between delicts of anti-drugs and anti-doping violations. Punishment and hence costs of doping will increase dramatically in the future as competition bans will be accompanied by prison sentences or fines (Zeit.de, 2015). This will also affect postsporting career chances for athletes since misdemeanor will lead to them having a criminal record.

FINAL REMARKS Doping in particular is a persistent trait of professional sports. This chapter brings together economic theories from both the rational and behavioral spheres to analyze the athlete’s inclination

towards doping. The implications suggest that both approaches are useful in explaining doping decisions and that athletes are driven by complex bundles of cost–benefit calculations, incentives, reputation concerns, spillover effects, and social contagion. Understanding the mechanism of doping decisions is a pivotal element for designing institutions capable of curbing its persistence. So far, the institutions in charge have had trouble implementing the right mix of rules and leeway to allow for a clean and competitive sport. One fundamental problem is the lack of quality data. Given the excessively high monetary and nonmonetary stakes involved in (professional) sports, this deficiency is worrisome.

REFERENCES Achtziger, A., Alós-Ferrer, C., & Wagner, A. K. (2015). Money, depletion, and prosociality in the dictator game. Journal of Neuroscience, Psychology, and Economics, 8, 1–14. Ahrens, P. (2014). Lance Armstrong im US-Fernsehen: ‘Perfekt vermarkteter Betrug’. Spiegel Online, August 20, 2014, www.spiegel.de/sport/sonst/ lance-armstrong-doping- luegen-beichte-a-987105. html Alaranta, A., Alaranta, H., Heliövaara, M., Airaksinen, M., & Helenius, I. (2006). Ample use of physicianprescribed medications in Finnish elite athletes. International Journal of Sports Medicine, 27(11), 919–925. Basu, K., Basu, K., & Cordella, T. (2014). Asymmetric punishment as an instrument of corruption control. The World Bank: Policy Research Working Paper 6933. Becker, G. (1968). Crime and punishment: An economic approach. Journal of Political Economy, 76, 169–217. Becker, G., & Murphy, K. (1988). A theory of rational addiction. Journal of Political Economy, 96, 675–700. Berentsen, A. (2002). The economics of doping. European Journal of Political Economy, 18, 109–127. Büchel, B., Emrich, E., & Pohlkamp, S. (2014). Nobody’s innocent: The role of customers in the doping dilemma. Journal of Sports Economics, online first, October 7. doi:10.1177/1527002514551475 de Hon, O., Kuipers, H., & van Bottenburg, M. (2015). Prevalence of doping use in elite sports: A review of numbers and methods. Sports Medicine, 45(1), 57–69. Deutscher, C., Dimant, E., & Humphreys, B. (2017). Match Fixing and Sports Betting in Football:

THE ECONOMICS OF DOPING IN SPORTS: A SPECIAL CASE OF CORRUPTION

Empirical Evidence from the German Bundesliga (No. 0008). Philosophy, Politics and Economics, University of Pennsylvania. Deutscher, C., Frick, B., Gürtler, O., & Prinz, J. (2013). Sabotage in tournaments with heterogeneous contestants: Empirical evidence from the soccer pitch. The Scandinavian Journal of Economics, 115, 1138–1157. Deutscher, C., & Schneemann, S. (2017). The impact of intermediate information on sabotage in tournaments with heterogeneous contestants. Managerial and Decision Economics, 38(2), 222–237. Dietl, H. M., Lang, M., & Werner, S. (2010). Corruption in professional sumo: An update on the study of Duggan and Levitt. Journal of Sports Economics, 11(4), 383–396. Dietz, P., Ulrich, R., Dalaker, R., Striegel, H., Franke, A., Lieb, K., & Simon, P. (2013). Associations between physical and cognitive doping: A cross-sectional study in 2,997 triathletes. PLoS ONE, 8, 11. Dilger, A., Frick, B., & Tolsdorf, F. (2007). Are athletes doped? Some theoretical arguments and empirical evidence. Contemporary Economic Policy, 25, 604–615. Dimant, E. (forthcoming). Contagion of pro- and anti-social behavior among peers and the role of social proximity. Journal of Economic Psychology. Dimant, E., & Schulte, T. (2016). The nature of corruption: An interdisciplinary perspective. German LJ, 17, 53. Dimant, E., & Tosato, G. (2018). Causes and effects of corruption: What has past decade’s empirical research taught us? A survey. Journal of Economic Surveys, 32(2), 335–356. Duggan, M., & Levitt, S. (2002). Winning isn’t everything: Corruption in Sumo Wrestling. The American Economic Review, 92, 1594–1605. Ehrenberg, R., & Bognanno, M. (1990a). Do tournaments have incentive effects? Journal of Political Economy, 98, 1307–1324. Ehrenberg, R., & Bognanno, M. (1990b). The incentive effects of tournaments revisited: Evidence from the European PGA Tour. Industrial and Labor Relations Review, 43, 74–88. Erdman, K. A., Fung, T. S., Doyle-Baker, P. K., Verhoef, M. J., & Reimer, R. A. (2007). Dietary supplementation of high-performance Canadian athletes by age and gender. Clinical Journal of Sport Medicine, 17(6), 458–464. Garicano, L., & Palacios-Huerta, I. (2005). Sabotage in tournaments: Making the beautiful game a bit less beautiful. London School of Economics: Discussion Paper. Glaeser, E., Sacerdote, B., & Scheinkman, J. (1996). Crime and social interactions. Quarterly Journal of Economics, 11, 507–548. Hanstad, D., & Loland, S. (2009). Elite athlete’s duty to provide information on their whereabouts:

551

Justifiable anti-doping work or an indefensible surveillance regime? European Journal of Sport Science, 9, 1. Hermann, A., & Henneberg, M. (2013). Anti-doping systems in sports are doomed to fail: A probability and cost analysis. University of Adelaide. Hirschi, T., & Gottfredson, M. (1990). A General Theory of Crime. Stanford, CA: Stanford University Press. Ichniowski, C., & Preston, A. (2014). Do Star Performers Produce More Stars? Peer Effects and Learning in Elite Teams. NBER Working Paper No. 20478. Cambridge, MA: National Bureau of Economic Research. Innes, R. (1999). Remediation and self-reporting in optimal law enforcement. Journal of Public Economics, 72, 379–393. Janofsky, M. (1991). Olympics: Coaches concede that steroids fueled East Germany’s success in swimming. The New York Times, December 3, www.nytimes.com/1991/12/03/sports/ olympics-coaches-concede-that-steroids-fueledeast-germany-s-success-in-swimming.html Kirstein, R. (2014). Doping, the inspection game, and Bayesian enforcement. Journal of Sports Economics, 15, 385–409. Kräkel, M. (2007). Doping and cheating in contestlike situations. IZA Discussion Papers No. 2059. Bonn, Germany: Institute for the Study of Labor. Lazear, E. (1979). Why is there mandatory retirement? Journal of Political Economy, 87, 1261–1284. Lessig, L. (2013). Institutional corruption’ defined. Journal of Law, Medicine and Ethics, 41, 3. Maennig, W. (2002). On the economics of doping and corruption in international sports. Journal of Sports Economics, 3, 61–89. Maennig, W. (2008). Corruption in international sports and how it may be combatted. International Association of Sports Economists: Working Paper Series No. 08–13. Maennig, W. (2014). Inefficiency of the anti-doping system: Cost reduction proposals. Substance Use & Misuse, 49, 1201–1205. Masters, A. (2015). Corruption in sport: From the playing field to the field of policy. Policy and Society, 34(2), 111–123. Muraven, M., Pogarsky, G., & Shmueli, D. (2006). Selfcontrol depletion and the general theory of crime. Journal of Quantitative Criminology, 22, 263–277. Nieper, A. (2005). Nutritional supplement practices in UK junior national track and field athletes. British Journal of Sports Medicine, 39(9), 645–649. Nowak, M., & Sigmund, K. (1993). A strategy of winstay, lose-shift that outperforms tit-for-tat in the Prisoner’s Dilemma game. Nature, 364(6432), 56–58. Peretti-Watel, P., Guagliardo, V., Verger, P., Mignon, P., Pruvost, J., & Obadia, Y. (2004). Attitudes toward doping and recreational drug use among

552

THE SAGE HANDBOOK OF SPORTS ECONOMICS

French elite student-athletes. Sociology of Sport Journal, 21(1), 1–17. Pitsch, W., & Emrich, W. (2012). The frequency of doping in elite sport: Results of a replication study. International Review for the Sociology of Sport, 47, 559–580. Pitsch, W., Emrich, W., & Klein, M. (2007). Doping in elite sports in Germany: Results of a www survey. European Journal of Sport and Society, 4(2), 89–102. Preston, I., & Szymanski, S. (2003). Cheating in contests. Oxford Review of Economic Policy, 19(4), 612–624. Sottas, P.-E., Robinson, N., Fischetto, G., Dollé, G, Alonso, J., & Saugy, M. (2011). Prevalence of blood doping in samples collected from elite track and field athletes. Clinical Chemistry, 57(5), 762–769. Striegel, H., Ulrich, R., & Simon, P. (2010). Randomized response estimates for doping and illicit drug use in elite athletes. Drug and Alcohol Dependence, 106(2–3), 230–232. Tirole, J. (1996). A theory of collective reputations (with applications to the persistence of corruption and to firm quality). The Review of Economic Studies, 63, 1–22.

Tsivou, M., Kioukia-Fougia, N., Lyrics, E., Aggelis, Y., Fragkaki, A., Kiousi, X., Simitsek, Ph., Dimopoulou, H., Leontiou, I.-P., Stamou, M., Spyridaki, M.-H., & Georgakopoulos, C. (2006). An overview of the doping control analysis during the Olympic Games of 2004 in Athens, Greece. Analytica Chimica Acta, 555(1), 1–13. Vohs, K., Baumeister, R., Schmeichel, B., Twenge, J., Nelson, N., & Tice, D. (2008). Making choices impairs subsequent self-control: A limited-resource account of decision making, self-regulation, and active initiative. Journal of Personality and Social Psychology, 94(5), 883–898. Wagner, G., & Ziebarth, N. (2014). The shaping of attitudes toward doping: Evidence from divided Germany. Mimeo. Wolfers, J. (2006). Point shaving: Corruption in NCAA Basketball. The American Economic Review, 96(2), 270–283. Zeit.de (2015). Kabinett beschließt Haftstrafen für gedopte Sportler. Zeit.de Online. Retrieved from: w w w . z e i t . d e / s p o r t / 2 0 1 5 - 0 3 / a n t i - d o p i n ggesetz-bundeskabinett-beschluss

54 Performance Analytics Bill Gerrard

INTRODUCTION Analytics has been one of the buzzwords of the 21st century. A Google search on ‘analytics’ for the calendar year 2000 yields 143,000 hits but five years later this had increased to 2.92 million. By 2010 there were 5.81 million hits on Google but thereafter the growth has been exponential with 122 million hits in 2016. The average annual growth rate between 2000 and 2016 was 52.5% but in the most recent five-year period, 2011–2016, the annual growth rate was 68.1%. But although data analytics is now widely recognised as playing an important role in the management of performance in business organisations, it is only just beginning to become a subject of academic study in the business disciplines (e.g. Angrave et al., 2016). Sotiriadou (2013, p. 2) has commented that performance analysis and data mining are taking high-performance sport to new heights. The use of data analytics to support the player recruitment process in professional team sports has been popularised by Moneyball (Lewis, 2003) but there has been little theoretical framing of the topic. This chapter addresses this gap in the sport economics and sport management literature and draws on Davenport’s taxonomy of five stages that organisations need to progress through to become analytical competitors with data analytics

as their principal source of a sustainable competitive advantage (Davenport, 2006; Davenport & Harris, 2007). Davenport’s taxonomy provides a useful way of categorising the extent to which an organisation has developed an analytical capability. The chapter begins with a discussion of the role of data analytics as a management tool in the business sector. The following section focuses on the development of performance analytics in professional team sports, showing how the applications of data analytics to player recruitment is underpinned by research on player valuation and player rating in sports economics. The Moneyball story is then considered as a popularised case study of how data analytics has actually been used to influence player recruitment decisions. In the next section it is argued that performance analytics can be applied to a wide range of coaching decisions as part of a thoroughgoing evidence-based approach to coaching. The chapter ends with a discussion of the managerial implications and suggestions for future research.

DATA ANALYTICS AS A MANAGEMENT TOOL Definitions of analytics abound. A recent discussion of the phenomenon of analytics in the area of

554

THE SAGE HANDBOOK OF SPORTS ECONOMICS

human resource management starts with the following definition of analytics: Analytics is the discipline which has developed at the intersection of engineering, computer science, decision-making and quantitative methods to organise, analyse and make sense of the increasing amounts of data being generated by contemporary societies. (Angrave et al., 2016, p. 1)

While this definition recognises that analytics draws on a number of disciplines – one could further break down quantitative methods into statistics, econometrics, management science and operational research – it fails to emphasise that analytics is above all a practice-led area of endeavour. It is analysis with purpose, a form of practical wisdom, or what Aristotle called ‘phronesis’ (Aristotle, 1955; Flyvbjerg, 2006). Davenport and Harris (2007) make the importance of practice very clear in their definition of analytics: ‘…the extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and fact-based management to drive decisions and actions’ (p. 7). Analytics produces actionable insight (MayerSchönberger & Cukier, 2013). It can be summarised by the three Ds: Data, Domain and Decision. Analytics involves the application of data analysis in specific domains to support organisational decision making. Analytics is descriptive (what happened?), diagnostic (why did it happen?), predictive (what could happen?) and prescriptive (what should be done?). The recognition that the objective of analytics is to support organisational decision making necessarily leads to more ‘negative’ definitions in terms of what analytics is not, i.e. non-analytical approaches to decision making. For example, Davenport, Harris and Shapiro (2010) state that ‘analytics takes the guesswork out of fresh management approaches’ (p. 55). Referring specifically to talent analytics, they go on to describe the mind-set associated with an analytical approach: ‘Leaders who believe that human-capital insights should be used to solve business problems must constantly press for decisions and analyses based on facts and data rather than on tradition, hearsay, or supposition’ (Davenport, Harris & Shapiro, 2010, p. 58). In this respect analytics can be seen as part of the broader movement towards evidencebased practice which started in medicine in the early 1990s and has spread across many practicefocused disciples, including management. The analytical mind-set described by Davenport, Harris and Shapiro (2010) has much in common with Pfeffer and Sutton’s (2006) description of the ‘attitude of wisdom’ associated with evidence-based management: ‘Evidence-based

management is conducted best not by know-italls but by managers who profoundly appreciate how much they do not know’ (p. 73). Davenport (2006) classifies organisations based on the extent to which they utilise data analytics. Davenport reports the results of a study of 32 organisations that are committed to an evidence-based approach involving quantitative data analysis. Davenport (2006) argues that only 11 of these organisations are analytical competitors. These organisations share three key attributes: 1 The widespread use of modelling and optimisation. 2 An enterprise-wide approach to the use of data analytics. 3 Senior executives as advocates for the use of data analytics. Davenport extended his argument in the book, Competing on Analytics: The New Science of Winning (Davenport & Harris, 2007), in which a five-stage hierarchy of organisations in their use of data analytics is proposed. Organisations at Stage 1 are analytically impaired, with little or no analytical capabilities and a knowledge-allergic culture in which managers pride themselves on gut-based decisions. Organisations at Stage 2 have localised analytics. Data analytics is used on an ad hoc basis in different departments to improve performance within different functional activities. Organisations at this stage have no coherent plan to adopt a more analytical approach so that analytical initiatives are small-scale and undertaken in isolation. Organisations at Stage 3 have analytical aspirations. They have made a strategic commitment to developing a distinctive analytical capability as a source of competitive advantage and have started to integrate their analytical efforts in a more coordinated, organisation-wide approach. It is often at this stage that organisations create a centralised database and may bring together analysts into a single department responsible for providing data analytics to the rest of the organisation. Organisations at Stage 4 are designated as analytical companies/ organisations with an organisation-wide analytical capability. At this stage analytics is seen as contributing significantly to organisational effectiveness, with the extensive use of data analytics setting Stage 4 organisations apart from other organisations in their sector. But, crucially, data analytics is not yet seen as the primary source of sustainable long-term competitive advantage. Organisations at Stage 5 are analytical competitors, with analytics considered to be the principal driver of organisational performance. Stage 5 organisations are

Performance Analytics

555

Stage 5 Analytical competitors Stage 4 Analytical organisations

Stage 3 Analytical aspirations

Stage 2 Localised analytics

Stage 1 Analytically impaired

Figure 54.1  The five stages of analytical competition (adapted from Davenport & Harris, 2007) committed to a continual search for analytics-led innovations to maintain their competitive success. The process of moving up through the analytical hierarchy is described by the DELTA model (Davenport, Harris, & Morison, 2010), in which the move from one stage to the next is described in terms of five dimensions – Data, Enterprise, Leadership, Targets and Analysts (forming the acronym, DELTA). Progress towards becoming an analytical competitor requires developments along all five dimensions. Data must become more comprehensive, more integrated and better quality. The analytical effort must be enterprise-wide, with passionate and committed leadership by senior executives. Analytical targets should initially aim for depth before broadening out to embrace organisational targets. And the organisation must be able to attract top analysts who are able to work effectively with managers. In setting out a roadmap to analytics success, Cuddeford-Jones (2015) offers similar insights to the DELTA model, stressing the importance of building a strategic focus around analytics with leadership commitment and a data-conscious culture.

PERFORMANCE ANALYTICS Alamar (2013) defines sports analytics as ‘the management of structured historical data, the application of predictive analytic models that

utilize that data, and the use of information systems to inform decision makers and enable them to help their organizations in gaining a competitive advantage on the field of play’ (p. 4). This is a comprehensive definition that captures the essential characteristic of analytics, namely the use of data analysis to support decision makers. In the case of high-performance sport, the decision makers are the coaches and support staff tasked with the responsibility of facilitating winning performances by athletes/players. In some senses data analytics is nothing new in high-performance sport since at the elite level of many sports it is a well-established practice to collect physical performance data both during training and in competition. It is therefore common in high-performance sport for the coaching and support staff to include sports scientists, typically with degree-level qualifications, who, as scientists, are very familiar with the use of statistical methods to analyse quantitative data. Recent developments, particularly the advent of Global Positioning System (GPS) monitors and other wearable technologies, has massively increased the amount of physical performance data collected, particularly on distance covered and speed, and this has put ever greater demands on sports scientists in high-performance sport. But although the quantity and quality of physical performance data has increased, this type of data analysis has not fundamentally changed the role of analytics in high-performance sport.

556

THE SAGE HANDBOOK OF SPORTS ECONOMICS

What is new in high-performance sport is the application of data analysis to tactical performance data in the invasion-territorial team sports. Tactical performance data are defined as data on the actions, decisions and techniques of athletes/players during competitive performance. The invasion-territorial team sports, such as the various codes of football, hockey and basketball, are all characterised by the need for coordinated teamwork to move a ball (or puck) into defended territory to score (i.e. attacking play) while preventing the opposing team from entering and scoring in the team’s own defended territory (i.e. defensive play). Until the last 20 years or so there has been little tactical performance data available for the invasion-territorial team sports for individual players beyond data on their appearances, scoring and discipline. The wide range of skill-activities involved in these sports as well as the continuity and speed of play rendered it very difficult to collate meaningful data during games using paper-and-pencil methods. As a consequence, invasion-territorial team sports have relied almost exclusively on video analysis to explore the actions, decisions and techniques of individual players. In contrast, the striking-and-fielding team sports, such as baseball and cricket, with the central focus on the one-to-one contest between the hitter/batter and pitcher/bowler, have been much more amenable to paper-and-pencil methods to capture key performance data during games at all levels of competition, not just high-performance sport. The key performance data for individual players at the elite level in baseball and cricket have been published in the national newspapers as box scores and scoring cards almost since the inception of these sports as professional competitions, with the consequence that there is a long history of statistical analysis of these sports, and especially in baseball. In the case of the invasion-territorial team sports, technology has been a key enabler in the collection of tactical performance data and the subsequent development of performance analytics. Previously, tactical performance data could be collected post-match from game videos using paperand-pencil methods and the pause and rewind buttons on video machines, but this was very time-consuming. The development of notational systems quickened the data extraction process and also allowed some tactical performance data to be collected in real time during games. But the real breakthrough came in the 1980s and 1990s with the advent of digital technology and the development of computer software to facilitate the coding of games. Commercial companies specialising in the supply of tactical performance data in highperformance sport started to emerge. For example,

in England, two companies, Opta and Prozone, were formed in the mid/late 1990s to supply tactical performance data on association football. The availability of tactical performance data and the computer processing power provided the opportunity for professional sports teams to develop a performance analytics capability. But in order to exploit this opportunity and progress through the various stages of analytical competition required teams to identify potential applications and develop the appropriate analytical tools. In many cases this involved adapting existing analytical methods for specific practical uses. One particular source of analytical methods has been sports economics, which has utilised a whole array of statistical methods to analyse sports data in order to test economic hypotheses. Performance analytics has taken some of the analytical methods of sports economics and adapted them to provide an evidential base for decision making by coaches and team executives, as exemplified by developments in the valuation of playing talent. The core business process in professional team sports is transforming financial expenditure on playing talent into sporting performance. Quite simply, the business of professional team sports is turning wages into wins. Playing talent is one of the key resources for any professional sports team so that an ability to value playing talent accurately in order to optimise the allocation of the available financial budget to acquire and retain playing talent is a critical strategic component of organisational effectiveness. From the analytics perspective, the valuation of playing talent can be seen as an application of asset valuation techniques. These fall into two broad categories – comparative valuation techniques and fundamental valuation techniques. Comparative valuation is an anchorand-adjustment method of asset valuation in which observed market values of recently traded assets in the same asset class (or reference set) act as the ‘anchor’ value to be adjusted to take account of the specific characteristics of the asset to be traded. Fundamental valuation is an expectedreturns method, in which an asset value represents the present value of the expected future net cash flows adjusted to include any non-financial returns. Both types of asset valuation technique have been employed in sports economics in order to test economic hypotheses. There are two main types of comparative valuation technique. In the case of assets in which future performance is essentially bivariate, valuation ratios can capture the direct proportionality between asset value and the single characteristic of the asset. Valuation ratios, such as the priceearnings ratio and the market-to-book ratio, are widely used in corporate valuation. However,

Performance Analytics

in the case of playing talent, particularly in the invasion-territorial sports, the valuation problem is multivariate and requires the use of some form of hedonic-pricing technique, in which statistical techniques, such as multiple regression analysis, are used to estimate the imputed (or hedonic) prices for the set of individual asset characteristics associated with future performance. The hedonicpricing technique has been frequently applied in the sports economics literature, although often not recognised as such since the academic purpose has usually been to test a specific economic hypothesis and not to value playing talent per se. For example, there were a number of published studies in the 1990s that analysed football transfer fees in English professional football. Carmichael and Thomas (1993) used transfer fees to examine bargaining theory, while Reilly and Witt (1995) were concerned with racial discrimination, and Gerrard and Dobson (2000) focused on the degree of monopsony in the labour market. But in order to test economic hypotheses, all three studies had to control for player quality and in order to do so they identified a composite set of proxy variables for player quality that jointly modelled around 70–80% of the variation in individual transfer fees. These variables can be categorised into four groups of player quality indicators: • Player characteristics (e.g. age, career experience, current appearance rate, goals scored and national team appearances) • Selling-club characteristics (e.g. economic size, divisional status and current league performance) • Buying-club characteristics (e.g. economic size, divisional status and current league performance) • Market conditions (e.g. inflation rates and transfer deadlines). Note that these studies all pre-dated the general availability of tactical performance data so that the only performance data that could be included in these regression models were appearances and goals scored. The estimated coefficients in these published econometric studies of transfer fees can be interpreted as the hedonic prices for each of the player quality indicators. This is the approach adopted by Gerrard (2004), who adapted the estimated regression model in Dobson and Gerrard (1999) to create the SOCCER TRANSFERS player valuation system to provide transfer valuations of football players. The adaption involved three main elements. First, the estimated regression model was rearranged to create an algorithm with seven aggregate player quality indicators – age/experience, current squad status, goal scoring, international

557

experience, current club, potential buying club and transfer market inflation. Second, the player valuation algorithm needed to be extended to deal with the changing structure of the players’ labour market. In particular, the original econometric studies of football transfer fee all used data from the preBosman period when the players’ labour market in professional football was largely organised on a national basis and highly monopsonistic. All of that changed following the Bosman ruling by the European Court of Justice in September 1995, which led to the introduction of free agency across European professional team sports and ultimately to the internationalisation of players’ labour markets, especially in football. In order to accommodate this radical restructuring of the players’ labour market, the remaining time until the contract expiry date had to be introduced as a key driver of transfer values since Bosman free agency implied that a player had zero transfer value when their contract expired. The internationalisation of the players’ labour market also required that the player valuation algorithm be extended to apply to players and clubs outside England by ensuring that their data are consistent with the data for players and clubs in England. The SOCCER TRANSFERS player valuation system was one of the first applications of performance analytics in football, albeit that the tactical performance data used were limited to appearances and goals scored. But it broke new ground by using data analysis to support realworld decision making. Initially, it was used to provide player and squad valuations to guide teams in player transfer transactions. It was also used to determine the appropriate level of player insurance cover. But it was also used by financial institutions either to assist in the valuation of the equity of football clubs in acquisition transactions or to assess the security offered by the asset value of players in debt-financing arrangements. Player valuations also played a role in the resolution of various legal and tax disputes (Gerrard, 2014). The basic methodology for the fundamental valuation of players has also been developed in the academic economics literature. Scully (1974) was interested in testing a basic hypothesis of labour economics that the degree of economic exploitation depends on the competitive structure of labour markets. Economic exploitation occurs whenever the wage rate is less than the marginal revenue product (MRP) of labour. Scully used data from the players’ labour market in Major League Baseball (MLB) in the late 1960s which at that point was highly monopsonistic due to the Reserve Clause giving exclusive rights to the teams to retain their out-of-contract players. In order to calculate the degree of exploitation,

558

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Scully developed a two-stage regression procedure to estimate a player’s MRP. This first stage involved estimating the relationship between team performance (as measured by the win percentage) and the team totals for the player performance metrics, with various additional controls included in the regression model. The second stage was the estimation of the relationship between team revenue and team performance, again with various controls included. Taken together, the two estimated models provided the means of converting performance metrics into financial value in terms of incremental revenues. Since the performance metrics represented the aggregate contributions of the individual players, the incremental revenues could be easily disaggregated and attributed to individual players. Scully found that professional baseball players received only around 15–20% of their MRP. Scully’s methodology has been subsequently refined and applied in sports economics to test hypotheses on the impact of the introduction of free agency on the pay–performance relationship in the MLB (e.g. Zimbalist, 1992). Scully’s two-stage methodology for estimating player MRP can be adapted to provide player valuations. There are several practical limitations to Scully’s original work, particularly the focus on (1) annual salary costs only, thus ignoring both the multi-year nature of player contracts and acquisition costs such as transfer fees; and (2) on-thefield performance only, thus ignoring the image value of players for merchandising and sponsorship. But the most important practical limitation is that Scully used data from a striking-and-fielding sport in which player performance is effectively unidimensional. Scully measured batter performance with the slugging average and measured pitcher performance with the strikeout-to-walk ratio. However, as soon as you turn to the invasionterritorial team sports, player performance is multidimensional, requiring the construction of player rating systems (Gerrard, 2007). The difficulty is analogous to the problem faced with comparative valuation and the need to use multivariate hedonicpricing methods rather than simple valuation ratios. One approach is to use regression analysis to estimate the relationship between team performance and an appropriate set of performance metrics and use the estimated coefficients as weightings in the construction of a player performance rating. However, the regression approach to player ratings is often bedevilled by a number of estimation problems due to multicollinearity, the sequential game structure and the difficulty in controlling for opposition performance. As a consequence, regression-based player rating systems tend to involve a combination of statistical estimates and expert judgement as, for example, in

the EA Sports Player Performance Index (McHale, Scarf, & Folker, 2012) and my own STARS player rating system (Gerrard, 2014). An alternative nonregression approach to player ratings is to convert the raw player performance metrics into Z-scores or rankings and combine these standardised metrics into a single rating using equal weights.

PERFORMANCE ANALYTICS IN ACTION: THE MONEYBALL STORY Some professional sports teams have led the way in the use of data analytics. Davenport (2006, 2014) identifies three major league teams as exemplars of the use of human capital analytics in the recruitment process – Oakland Athletics and Boston Red Sox in the MLB, and New England Patriots in the National Football League. These three teams are all categorised by Davenport as having progressed to be either Stage 4 analytical organisations or Stage 5 analytical competitors. The experience of Oakland Athletics in using performance analytics has received international attention following the publication of the best-selling book, Moneyball: The Art of Winning an Unfair Game (Lewis, 2003) and the subsequent Hollywood movie starring Brad Pitt. Moneyball has been a real game-changer in the perceptions of the possibilities for performance analytics in high-performance sport. Oakland Athletics are a small-market team lacking the financial resources to compete with big-market teams such as the New York Yankees. Led by Billy Beane, a former player appointed as General Manager in late 1997, Oakland began to use performance analytics to identify players who were undervalued by other MLB teams. In particular, Oakland started to tap into the emerging field of sabermetrics, which is the statistical analysis of baseball pioneered in modern times by Bill James, who began to publish the Baseball Abstract in 1977. One of the key findings of sabermetrics is that the conventional performance metrics for batters, namely, the batting and slugging averages, were not the optimal predictors of team win percentages because they did not take into account being walked to base. James and other sabermetricians had long known that on-base percentage (OBP) is a much better predictor of team win percentage. Indeed, James advocated on-base plus slugging (OPS), calculated as the sum of OBP and the slugging average, as providing the best single performance metric for batters. Oakland realised that other MLB teams were ignoring the insights of sabermetrics so that batter salaries were determined principally using batting

Performance Analytics

and slugging averages that only allowed for players hitting themselves to base. This reflected conventional industry wisdom that walks represented pitcher error. But walks partly reflect a batter’s skill in swing selection, so to be ignored in salary valuation represented what economists call a ‘free lunch’, a valuable commodity with a zero market price. Oakland started to recruit players using OBP and were very successful in identifying players with high OBP relative to their market salaries. Gerrard (2007) estimated that over the first nine seasons of Beane’s tenure, 1998–2006, Oakland achieved an efficiency gain of 59.3% compared to the league average. Moneyball mainly focuses on the 2001 and 2002 seasons. In both of these seasons, Oakland ranked as the second-best team in the MLB in terms of regular-season win percentage, yet had one of the lowest salary budgets in the MLB. Hakes and Sauer (2006) explicitly interpret Moneyball as an example of the efficient market hypothesis. Using MLB data for the five-year period, 1999–2003, they found very strong evidence supporting the sabermetric proposition that OBP is the best statistical predictor of team win percentage. Hakes and Sauer then estimated a salary equation for batters using both OBP and the slugging average as predictors, as well as including control variables such as arbitration eligibility. They found that up until 2003 OBP had no significant impact on batter salaries, providing strong evidence of informational inefficiency in the market valuation of batters, just as Oakland had identified and exploited to their competitive advantage. The efficient market hypothesis predicts that when an informational inefficiency is successfully exploited for profit, and this becomes observable, the market will very quickly adjust as other traders seek to exploit the inefficiency. And that is exactly what Hakes and Sauer found in the MLB players’ labour market. They show that the market inefficiency was corrected in 2004 immediately after the publication of Moneyball, which alerted other MLB teams to the importance of using OBP in their salary valuations of batters. So much so, in fact, that OBP is the most significant performance predictor of batter salaries in the batter salary equation that Hakes and Sauer estimated for 2004. But the impact of Moneyball has gone way beyond baseball. It highlighted that performance analytics could provide a source of competitive advantage, a possible ‘David’ strategy for resource-constrained teams attempting to compete effectively against resource-richer ‘Goliath’ rivals in any professional team sport. And, as Moneyball has moved beyond baseball, it has led teams to consider other applications of performance analytics, beyond player recruitment, as part of a more thoroughgoing evidence-based approach to coaching.

559

BEYOND PLAYER RECRUITMENT: PERFORMANCE ANALYTICS AS AN EVIDENCE-BASED APPROACH TO COACHING The Moneyball story focuses primarily on the way in which Oakland Athletics have used performance analytics to inform their player recruitment decisions. There is some indication that Oakland also use performance analytics to guide decisions on game tactics, such as restricting the number of attempts to steal bases. But how extensively Oakland use data analytics elsewhere in the sporting operation and in the business side of the team is unclear and never discussed by Lewis (2003) in Moneyball. Hence, from the perspective of Davenport’s five-stages model (2006), it is probably most appropriate to categorise Oakland as a Stage 4 analytical organisation, in which the extensive use of performance analytics to assist player draft and player trading decisions sets them apart from other MLB teams, although the competitive advantage has narrowed over the years as other MLB teams have adopted and refined their use of performance analytics. Compared to baseball and other striking-andfield team sports, the invasion-territorial team sports offer much greater opportunities for the operational use of performance analytics to support coaching decisions beyond player recruitment. One example of an elite rugby union team that has followed a path similar to Oakland and transformed itself into at least a Stage 4 analytical organisation is Saracens (Slot, 2014; University of Leeds, 2014). Saracens are an elite rugby union team based in north London, competing in the English Premiership, the top tier of professional rugby in England, as well as competing in the top European team tournament, the Champions Cup (previously known as the Heineken Cup). Saracens have a long history, dating back to the late 19th century, but had won only one domestic knockout tournament in the professional era. Following the acquisition of a controlling interest in Saracens by a South African consortium, Saracens embarked in the 2008/2009 season on a comprehensive renewal of the organisation. This was led in the rugby operation by a new Director of Rugby, Brendan Venter, an experienced South Africa ex-international player and coach who is also a qualified medical practitioner. Venter initiated a process of cultural change that emphasised a people-centred, evidence-based approach to coaching and resulted in Saracens progressing rapidly from an organisation that made limited use of data analysis to becoming at least a Stage 4 analytical organisation

560

THE SAGE HANDBOOK OF SPORTS ECONOMICS

within five years. The coaching staff describe their use of performance analytics as ‘a unique point of difference that has made Saracens one of the leading clubs in European rugby’, ‘a game changing impact on conduct’ and ‘an integral part of the sustained success’ (University of Leeds, 2014). The step-by-step process by which Saracens developed their analytical capability started with the creation of expert, team-specific data by the coaches on own-team performance. The main form of tactical analysis in elite team sports is video analysis, although this is increasingly supported by summary game statistics supplied by commercial sports data providers. Under Venter’s leadership, the coaches started to systematically record their observations on own-team performance from the game video. A basic data collection system emerged consisting of tally counts of different types of activity based on team-specific definitions as well as expert evaluation of tactical decisions and skill technique. As this data accumulated during the 2009/2010 season, it was recognised that it could be used to investigate patterns and trends across games. I was appointed as their data analyst in March 2010 and that is when performance analytics really began. The focus initially was on reviewing individual games supplemented by quarterly reviews throughout the season. A reporting system was constructed for player and team performance with Key Performance Indicators (KPIs) identified and a traffic-lights system of categorising performance levels. The performance reports impacted on team selection and game tactics as well as being used in planning training schedules. The first full season in the implementation of the analytics system coincided with Saracens winning the Premiership title for the first time in May 2010. This major sporting success helped reinforce the belief among players, coaches, support staff and the owners in the efficacy of the new culture being created at Saracens, including the evidence-based approach to coaching. The application of performance analytics expanded after May 2010. In particular, it was used as part of a strategic benchmarking process to determine how Saracens could become competitive at the European level. In addition, at the initiative of the coaching staff, performance analytics was applied to opposition analysis. Gradually performance analytics became established as one of the principal sources of competitive advantage at Saracens (Slot, 2014). Their credentials as a Stage 4 analytical organisation were further reinforced by the involvement of Deloitte, one of the leading accounting and consulting firms internationally, who were appointed in 2014 to support the further development of the analytical capability, including the construction of a centralised database and

reporting facility. Saracens were European champions in 2016 and 2017, and their success is likely to provide encouragement to both their rivals as well as teams in other sports to develop their use of performance analytics. There are a number of cutting-edge developments in performance analytics. One of these is the analysis of tracking data on the location of players during games using video systems or GPS wearables. To date, most teams have used tracking data solely to assess the physical performance of players in terms of distance covered and speed. But there is a huge amount of tactical information contained in tracking data. One of the team sports leading the way in using tracking data for tactical purposes is basketball (Shea & Baker, 2013; Shea, 2014). For example, it is now possible to quantify the impact of ‘floor spacers’, those players whose presence on the court tends to stretch the defensive formation of the opposing team and thereby create space for teammates to shoot. The control and exploitation of space is crucial to success in the invasion-territorial team sports so being able to measure and assess the tactical positioning of players in and out of possession has enormous potential importance to coaches.

SUMMARY AND CONCLUSIONS Performance analytics is the application of statistical analysis and other related methods to investigate tactical performance data. Crucially, performance analytics is purpose-led. The data analysis is undertaken to provide an evidential base to support both operational and strategic decisions by the coaching staff. There is a long history of statistical analysis of performance data in the striking-and-fielding sports such as baseball, where the key data on individual player performance can be easily recorded by paper-and-pencil methods and at the elite level is widely reported in the media. What differentiates the development of performance analytics in the last 20 years is the adoption of data analysis by teams themselves, and the increasing uptake in the invasion-territorial sports as technological developments in video tracking systems and wearables have allowed the efficient collection of comprehensive performance data in these sports. The game changer for performance analytics undoubtedly has been Moneyball. Oakland Athletics represent (at least) a Stage 4 analytical organisation in Davenport’s taxonomy and provide an exemplar of how performance analytics can be used strategically to create and maintain a

Performance Analytics

sustainable competitive advantage. In the case of Oakland, performance analytics underpinned a David strategy in player recruitment to compete with resource-richer rivals by identifying undervalued players. But performance analytics is opening up new opportunities in the invasion-territorial team sports, as exemplified by Saracens in rugby union and basketball teams in the NBA. Performance analytics is increasingly establishing itself in the invasion-territorial team sports as a coaching tool that complements the more traditional coaching tools of video analysis, scouting and reporting. Sports economics and related disciplines have an important role in the development of performance analytics. Analytical methods developed to test hypotheses in discipline-directed research can be adapted to be used as analytical tools to support coaching decisions, as demonstrated by the development of player rating systems and player valuation algorithms. In adopting performance analytics, elite sports teams are merely following a similar path to organisations in other sectors utilising data analytics as part of a more thoroughgoing evidence-based approach to management. Ultimately, performance analytics is about replacing assertion with analysis, using the available data to help inform decision making rather than relying only on experience and intuition.

REFERENCES Alamar, B.C. (2013). Sports Analytics: A Guide for Coaches, Managers and Other Decision Makers. New York: Columbia University Press. Angrave, D., Charlwood, A., Kirkpatrick, I., Lawrence, M., & Stuart, M. (2016). HR and analytics: why HR is set to fail the Big Data challenge. Human Resource Management Journal, 26(1), 1–11. Aristotle (1955). Nicomachean Ethics. London: Penguin. Carmichael, F., & Thomas, D. (1993). Bargaining in the transfer market: theory and evidence. Applied Economics, 25, 1467–1476. Cuddeford-Jones, M. (2015). The Roadmap to Analytics Success: How to Create a Strong Foundation for Analytics Investment and Implementation. London: FC Business Intelligence. Davenport, T. H. (2006). Competing on analytics. Harvard Business Review, January, 99–107. Davenport, T. H. (2014). Analytics in Sport: The New Science of Winning. Portland, OR: International Institute for Analytics. Davenport, T. H., & Harris, J. G. (2007). Competing on Analytics: The New Science of Winning. Boston, MA: Harvard Business School Press.

561

Davenport, T. H., Harris, J. G., & Morison, R. (2010). Analytics at Work: Smarter Decisions, Better Results. Boston, MA: Harvard Business School Press. Davenport, T. H., Harris, J. G., & Shapiro, J. (2010). Competing on talent analytics. Harvard Business Review, October, 52–58. Flyvbjerg, B. (2006). Making organisation research matter: power, values and phronesis. In S. C. Clegg, C. Hardy, T. B. Lawrence & W. R. Nord (Eds.), The Sage Handbook of Organisation Studies (2nd ed., pp. 370–387). London: Sage. Gerrard, B. (2004). The measurement and valuation of player quality in association football. In J. J. Gouguet (Ed.), Professional Sport after the Bosman Case: An International Economic Analysis (pp. 143–164). Limoges, France: IASE/Pulim. Gerrard, B. (2007). Is the Moneyball approach transferable to complex invasion team sports? International Journal of Sport Finance, 2, 214–230. Gerrard, B. (2014). Achieving transactional efficiency in professional team sports: the theory and practice of player valuation. In J. Goddard & P. J. Sloane (Eds.), Handbook on the Economics of Football (pp. 189–202). Cheltenham, UK: Edward Elgar. Gerrard, B., & Dobson, S. (2000). Testing for monopoly rents in the market for playing talent. Journal of Economic Studies, 27, 142–164. Hakes, J. K., & Sauer, R. D. (2006). An economic evaluation of the Moneyball hypothesis. Journal of Economic Perspectives, 20, 173–185. James, B. (1977). Baseball Abstract: Featuring 18 Categories of Statistical Information That You Can’t Just Find Anywhere Else. Mimeo. Lewis, M. (2003). Moneyball: The Art of Winning an Unfair Game. New York: W.W. Norton. Mayer-Schönberger, V., & Cukier, K. (2013). Big Data: A Revolution That Will Transform How We Live, Work and Think. London: John Murray. McHale, I. G., Scarf, P. A., & Folker, D. E. (2012). On the development of a soccer player performance rating system for the English Premier League. Interfaces, 42, 339–351. Pfeffer, J., & Sutton, R. I. (2006). Evidence-based management. Harvard Business Review, January, 62–74. Reilly, B., & Witt, R. (1995). English league transfer prices: is there a racial dimension? Applied Economic Letters, 2, 220–222. Scully, G. W. (1974). Pay and performance in Major League Baseball. American Economic Review, 64, 915–930. Shea, S. (2014). Basketball Analytics: Spatial Tracking. Marston Gate: Amazon. Shea, S., & Baker, C. E. (2013). Basketball Analytics: Objective and Efficient Strategies for Understanding How Teams Win. Lake St Louis, MO: Advanced Metrics.

562

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Slot, O. (2014). Moneyball helps Saracens to play hard ball with Aviva Premiership rivals. The Times, 12 September, www.thetimes.co.uk/article/ moneyball-helps-saracens-to-play-hard-ball-withaviva-premiership-rivals-j0h7g2ndj9p (accessed 27 April 2017). Sotiriadou, P. (2013). The roles of high performance directors within national sporting organisations. In P. Sotiriadou & V. De Bosscher (Eds.), Managing High Performance Sport (pp. xi–xii). Abingdon, UK: Routledge.

University of Leeds (2014). Using statistical analysis to support decision making by managers and coaches in elite team sports. REF2014 Impact Case Studies. Available at: http://impact.ref.ac.uk/CaseStudies/ CaseStudy.aspx?Id=20050 (accessed 27 April 2017). Zimbalist, A. (1992). Salaries and performance: beyond the Scully model. In P. M. Sommers (Ed.), Diamonds are Forever: The Business of Baseball. Washington, DC: Brookings Institution.

Index Page numbers in bold indicate tables and in italic indicate figures. Abinzano, I., 454 Abramitzky, R., 456, 457 academic performance, sports participation and, 65–6 Achtziger, A., 548 Active Lives/People Surveys, UK, 27, 35, 36, 67, 410 Adachi, P.J., 40 Adelaide Oval, 327 Adler, K., 487 Adu, Freddy, 336–7 AFL see Australian Rules Football age marathon performance and, 448–9, 448 money donations and, 105–6 sports attendance and, 77 sports participation and, 37, 76, 407 volunteering and, 105–6 willingness to pay and, 424 Agha, Nola, 395–403 Agrawal, J., 478–9 Ahlert, G., 357 Ahlfeldt, G.M., 234, 235, 359, 371, 372, 373 Ahn, S.C., 65, 476 AIGCP (Association Internationale des Groupes Cyclistes Professionnels), 465 AIOCC (Association Internationale des Organisateurs de Courses Cyclistes), 465 Aiyar, S., 259 Aizawa, K., 411 Ajzen, I., 104 Akhtar, S., 259 Alamar, B.C., 555 Alameda Seven, 49 Alexander, Donald L., 113, 473 Ali, Ayfer, 379, 380 Allen, E.J., 449, 506 Allen, Mark, 486 Allen, W., 317 Allison, M., 85 Allmen, Peter von, 492–9 Allmers, S., 357–8, 359, 361 Allsopp, P.E., 259 Alston, Shawne, 119 Altonji, J.G., 295 altruistic behaviour, 102–8 multi-level model of, 105–7, 105, 108 theoretical framework, 103–4 see also volunteering American Basketball Association (ABA), 281, 282 American football, 289–96

behavioral economics, 294–6 competitive balance, 158 criminal behaviour and, 434–5 draft decisions, 292–3, 294–6 in-game decisions, 291–2 happiness and wins, 432 heuristic decision making, 294–6 home advantage, 222 intercollegiate, 114–16, 117–18, 119 player consistency, 285 psychological factors affecting maximization, 291–4 secondary ticket markets, 194–5, 195, 198, 201 in-stadium attendance, 164, 165 team relocations, 231, 232 ticket pricing, 184, 185, 527–8 work stoppages, 309 see also National Football League (NFL), US Analysis of Competitive Balance (ACB), 154, 156, 157 analytics see performance analytics Anbarci, N., 456, 457 Anders, C., 155, 185–6 Andersson, T.D., 418 Andreff, Wladimir, 8–16, 18, 139, 273, 275 Andreoni, J., 103 Angrave, D., 554 Angrist, J., 47–8 Anheier, H.K., 83 Ansorge, C.J., 510–11 antitrust legislation competitive balance defense, 146, 149–50 Major League Soccer and, 333, 334 professional baseball exemption, 146 US college sports and, 115, 116, 120 apprentices, baseball, 298, 302, 303–5, 305, 306 arbitration see salary arbitration Armstrong, J., 357 Arntzen, H., 246 Ashman, T., 312 ASO (Amaury Sports Organisation), 465, 468 Association of Surfing Professional (ASP) World Tour, 510 Association of Tennis Professionals (ATP), 453 Athens Olympics, Greece, 388, 389, 390 Atkinson, G., 419 Atkinson, S.E., 129 Atlanta Olympics, USA, 147, 358, 373, 391, 392 attendance see sports attendance Audas, R., 312

564

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Australia household sports consumption, 73 Olympic Games, 350, 358, 379, 380, 388, 390, 393 rugby, 269 sports participation, 36 see also Australian Rules Football Australian Bureau of Statistics, 36 Australian Football League Women’s (AFLW), 328 Australian Rules Football, 322–9 audience demand, 322, 323, 324, 325, 326, 327 competitive balance, 323–8 labour market for talent, 323–8 officials and bias, 327, 328, 509–10 stadiums, 326–7 Austria, 18, 23, 24, 73, 85 auto racing see NASCAR racing Azmat, G., 445–6 Baade, Robert A., 350, 358, 361, 392, 393, 395, 397, 498 badminton, 452, 458 Badminton World Federation (BWF), 452 Baim, D.V., 359 Baimbridge, M., 378 Bairam, E.A., 213 Baker, R.D., 456 Bakkenbüll, L.B., 456 Balia, S., 49 Balish, S., 95 Ball, Donald W., 378 Balmer, Nigel J., 379 Balmer, N.J., 508–9 Bandura, A., 104 Banko, L., 519–20 Barcelona Olympics, Spain, 358, 359 Bar-Eli, M., 511 Barget, E., 453, 457 Barlow, A., 420 Barr, G.D.I., 259 Barron, J.M., 66 Barros, Carlos Pestana, 215, 216, 251, 273, 418 Barsky, S.F., 222, 224 Bartolomeo, G.D., 58, 60 baseball antitrust exemption, 146 competitive balance, 146, 158 congestion and, 237 home advantage, 222, 224, 226 judging biases, 511 marginal revenue product (MRP) of players, 211–12, 298, 557–8 player consistency, 285–6 player productivity (WARP), 299–301, 300, 300, 301, 302, 302, 303, 304, 304, 305, 305, 306 secondary ticket markets, 193, 195, 195, 196, 198, 201 in-stadium attendance, 163, 164–5 team relocations, 232 ticket pricing, 183, 184, 185, 193, 457, 525, 526–7, 528–30

work stoppages, 308, 309 see also baseball labor market; Major League Baseball (MLB) baseball labor market, 9, 14, 15, 298–304 apprentices, 298, 302, 303–5, 305, 306 arbitration, 14, 299, 303, 306 contract extensions, 299, 299, 301–3, 302, 302 free agents, 299–301, 300, 300, 301, 302, 311 journeymen, 299, 302, 303, 304, 304 Moneyball story, 558–9 reserve clauses, 298, 305, 557–8 sabermetrics, 558–9 basketball, 279–86 coaches, 522 competitive balance, 158, 285, 286 data and streaming rights, 538 European leagues, 136 gender and, 522 home advantage, 222, 224, 508, 509 hot hand effects, 511, 512 objective measurement of performance, 279–81, 282 perceptions of performance, 281–5, 283, 284 in-play betting, 538 player consistency, 285–6 player salaries, 282–4, 283 secondary ticket markets, 195, 195, 202 in-stadium attendance, 165, 285, 286 team relocations, 232 ticket pricing, 184, 185 work stoppages, 309 see also National Basketball Association (NBA), US Bauman, A.E., 412 Baumann, Robert, 237, 391, 392, 434–5, 488–9 Baumeister, R.F., 226, 227 Baumol, W.J., 130 Beane, Billy, 558, 559 Becker, B.E., 493 Becker, G.S., 27, 34–5, 35, 37, 46, 74, 76, 94, 103, 407, 540, 547 Becker, S., 56–9, 57 Beckham, David, 332, 336–7 behavioral contagion, individual doping decisions and, 546, 549 behavioral economics, 505–13 American football, 294–6 difficulty bias, 508 golf, 478, 506, 512–13 heuristic decision making, 294–6, 548 home team bias, 224, 226, 227–8, 259–64, 262, 263, 508–9 hot hand effects, 291, 511–12 individual doping decisions, 548 judging biases, 507–12 loss aversion, 478, 506–7 nationalistic bias, 509–10 peer effects, 512–13 present bias, 507 reference dependent preferences, 449, 505–7 representativeness bias, 511–12

INDEX

reputation/expectation bias, 510–11 tennis, 455–6 Behavioral Risk Factor Surveillance Survey, US, 35–6, 50 behaviour change theories, 407–9, 408 Beijing Olympics, China, 343, 358, 380, 381, 386, 388, 389, 391, 410 Belgium, 18, 24, 73, 85 Belloc, N., 48–9 Belobaba, P.P., 526 Benijts, T., 466 Berkowitz, J.P., 496 Bernard, Andrew B., 379–80, 381 Bernthal, M.J., 498 Berri, David J., 158, 185, 279–86 Berry, S.M., 474 betting see sports betting Bhaskar, V., 259 Big Five personality traits, sports participation and, 66 Bird, P.J.W.N., 185, 336 bivariate probit models, 48, 49 Bizzozero, P., 158, 183, 453, 457 Blanco, Serge, 140 Blemmings, Benjamin, 289–96 Board of Cricket Control in India (BCCI), 257 body mass index (BMI), 49 Bognanno, M.L., 477, 517, 520 Böheim, R., 520, 521, 522 Bohlmann, H.R., 357 bonding social capital, 56 Booth, A.L., 513, 518–19 Booth, Ross, 322–9 Borenstein, S., 535 Borestein, S., 184 Borland, Jeff, 166, 167, 181–2, 186, 258, 323, 324, 325 Borooah, V.K., 260, 261 Bosca, J.E., 216 Bosman ruling, 13, 14, 139, 203, 204, 208, 244–5, 557 Boston Marathon, USA, 442, 442, 443, 449 Boucher, M., 315 Boudjellal, Mourad, 274 Boulier, B.L., 454 Bourdieu, Pierre, 55, 74, 96 Boyd, D.W., 449 Boyd, L.A., 449 Boyko, R.H., 228 Bradbury, J.C., 280, 285–6 Bradbury, S., 96 Bradley, R.A., 454 Brandes, L., 159 Brazil Olympic Games, 147, 343, 382, 385–6, 388, 391 World Cup, 351, 356–7, 386, 387–8, 391, 432 breast cancer, 46 Brehm, J., 60 Bremaud, P., 526 Breslow, L., 48–9 Brettschneider, W.-D., 57, 59 Breuer, Christoph, 75, 76, 77, 78, 82–8, 457, 531 Brewer, B., 468

bridging social capital, 56 British Basketball League (BBL), 538 British Household Panel Survey (BHPS), 36, 409 broadcasting rights Australian Rules Football, 326 betting and, 536 cricket, 264 English Premier League, 171–4, 173 Olympic Games, 343 professional team sports, 139, 148–51, 163 road cycling, 465, 467, 467 rugby union, 274 US college sports, 114–16, 117, 118 Brocard, Jean-François, 135–41, 466 Bronars, S., 478 Brooks, R.D., 259 Brooks, Robert, 322–9 Brown, A., 454 Brown, J., 477, 512–13, 518 Brown, W.O., 512 Brownlow Medal, 327 Brückner, M., 358 BSkyB, 173, 173, 177–8 Buchanan, J.M., 83 Büchel, B., 547–8 Budzinski, Oliver, 144–51, 155, 420 Bulgaria, 24 Bullough, S., 259 Buraimo, Babatunde, 171–9, 226, 251, 509 Burdina, M., 449 Burger, B., 526 Burgham, M., 93, 94, 95, 105 Buschemann, A., 309 Busse, Meghan R., 379–80, 381 Butzen, P., 76, 77 Buzzacchi, L., 157 Cabane, C., 67–8 Cain, L.P., 137 Cairns, J., 136, 137, 166, 167, 185 Calcagno, P.T., 233–4 Callan, S.J., 474 Camerer, C.F., 512 Campbell, B., 510 Canada employment and sports participation, 68 health and physical activity, 49 national pride and sport, 432, 433 Olympic Games, 344, 390, 432, 433 physical activity guidelines, 46 social capital formation, 58, 60 sports participation, 36 volunteering, 96 see also Major League Soccer (MLS) Canadian Community Health Survey, 36, 49 Candelon, B., 463, 465–6 Cappellari, L., 106 Card, D., 168, 506 Carlino, G., 234, 359

565

566

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Carmichael, F., 213–15, 216, 557 Carroll, K.A., 128 Carron, A.V, 508 cartel arrangements European sports leagues, 136–7 professional team sports, 148–51 US college sports, 115–20 see also sports systems Carter, R.V., 411 Caruso, R., 177 Case, A., 67 Castellanos, P., 419, 420 Celtic League, 271 championship uncertainty measure, 155, 167 Che, X.G., 272, 521 Chedzoy, O.B., 259 Chelladurai, P., 19 Chen, C.-A., 107 Cheney, P., 495–6 Cherchye, L., 464 Chester, D.N., 138 Chicago Marathon, USA, 442, 442, 443 Chikish, Yulia, 236, 505–13 Child Taking Part Survey, UK, 411 China household sports consumption, 73 Olympic Games, 343, 358, 380, 381, 386, 388, 389, 391, 410 Choi, S., 526 Chopin, M.C., 476 Chumacero, R.A., 225 circular flow of income, 20, 20 City Resource Supply (CRS), 398–400, 399 civic engagement, 55, 57–8, 59–60 Clark, A.E., 67 Clarke, S.R., 259, 454 Classification of Products by Activity (CPA) system, 21 clubs see non-profit sports clubs coaching, evidence-based approach to, 559–60 coaching turnover, ice hockey, 312, 312–13 Coakley, J., 433 Coate, Douglas, 456 Coates, Dennis, 87, 168, 185, 186, 231–7, 335–6, 359, 399, 423, 497, 498, 506 Cobb-Douglas production function, 212 cognitive bias, individual doping decisions and, 548 cognitive doping, 545 cognitive skills, sports participation and, 65–6 Cohen-Zada, D., 456 Coleman, James, 55 collective bargaining agreements (CBAs) baseball, 14, 298, 305 ice hockey, 308–9, 311, 315 Major League Soccer, 334 College Football Association (CFA), US, 115 college sports, United States, 112–20, 145 broadcasting rights, 114–16, 117, 118 criminal behaviour and, 435

player compensation, 114, 116, 117, 118–20 reasons for university sponsorship of, 112–14 revenue maximization, 130–1 utility maximization, 128 Collins, T., 269, 270 colon cancer, 46 Communist regimes, Olympic performance and, 379, 380, 381, 382, 383 comparative valuation of players, 556–7 competition intensity measure, 155 competition policy professional team sports, 144–51 US college sports, 115–20 competitive balance (CB), 154–9, 157, 163–4 as antitrust defense, 146, 149–50 Australian Rules Football, 323–8 cricket, 258, 264 football (soccer), 141, 245, 246 game uncertainty measures, 154–5, 157–8, 167 inter-seasonal uncertainty measures, 155–6, 158–9, 167 Major League Baseball (MLB), 146, 158 National Basketball Association (NBA), US, 158, 285, 286 road cycling, 469 seasonal uncertainty measures, 155, 158, 167 tennis, 453 transfer market and, 204–6, 205 uncertainty of outcome hypothesis, 154, 156–9, 163, 165, 166, 167, 210, 258, 506 competitive balance ratio (CBR), 156, 158 competitive intensity (CI), 141 computable general equilibrium (CGE) models, 344, 356, 357, 362, 398 concentration ratios, 156 congestion, traffic, 236, 237, 434 Conn, D., 274 Connaughton, J.E., 498 Connolly, R.A., 512–13 consecutive-season uncertainty measure, 167 consumer choice model of sports attendance, 168 consumer expenditure sports betting, 535 see also household sports expenditure consumption capital theory, 416–17, 424 contingent valuation method (CVM), 415–25, 418–23 Contoyannis, P., 49 cooperative norms, 55–6, 58, 60 Cornelissen, S., 432 Cornelissen, T., 67, 68 coronary heart disease, 46, 49 corruption match fixing, 457, 538–41 see also doping cost–benefit approach (CBA), 293, 344, 353, 356, 362, 395, 396, 398, 401–2, 401, 403 Coubertin, Pierre de, 387, 441 Coulombe, S., 315

INDEX

Coulson, N.E., 234, 359 County Championship, England and Wales, 135, 137, 213 County Health Rankings, US, 45 Courneya, K.S., 508 Court of Arbitration for Sport, 540 Courty, Pascal, 187, 190–9 CPA (Cyclistes Professionnels Associés), 465 Craig, C.L., 412 cricket, 135, 136, 137, 256–65 audience demand, 258, 264–5 coin toss, 226, 257–9, 264 ground sharing, 326–7 home advantage, 219, 222–8, 223, 225, 226, 227, 259–64, 262, 263 leg before wicket (LBW) decisions, 228, 259, 260–4, 262, 263 neutral umpires, 257, 260–4, 262, 263 officials and bias, 227–8, 257, 259–64, 262, 263 team performance, 212–13 Cricket Archive website, 222 criminal behaviour domestic violence, 434, 435, 506 match fixing, 457, 538–41 sports events and, 434–5 stadiums and, 236–7 Croatia, 24 Crompton, J.L., 184 Crooker, J.R., 419 Croson, R., 518 crowding-out effects, 346, 350, 352, 352, 353, 362, 398, 399, 400, 402, 498 Cuddeford-Jones, M., 555 Cui, Y., 197 cultural capital, 74 Curme, M., 315 Cushion, C., 216 Cuskelly, G., 95, 106 cycling see road cycling Cyprus, 18, 24, 73 Czech Republic, 24 Dabscheck, B., 322, 323, 325–6 Dahl, G.B., 168, 506 Dana, J.D.Jr., 184 Dang, T., 326 Daniels, M.J., 396 Daugherty, G., 315 Davenport, T.H., 553, 554, 558, 559 Davies, L.E., 78, 97, 397 Davis Cup, 226, 227, 457 Dawson, Peter, 40, 50–1, 94, 95, 227, 259, 405–12 Daytona 500, 493, 498 De Boer, W.I.J., 422 de Silva, B.M., 259 DeAntonio, D., 236 Decision Review System (DRS), cricket, 257, 260–4, 262, 263 decisiveness of a game measure, 155

567

Decker, O., 105 Deelen, I., 41 del Corral, Julio, 157, 452–8 Delaney, L., 57, 58, 60 Deliège, Christelle, 146–7 DellaVigna, S., 507 Demmert, Henry, 137, 165 demonstration and trickle-down effects, 405–6, 409–12, 428 Deneckere, R., 197 Denmark, 24, 73, 88 Depken, C.A., 156, 183, 317, 399, 494 Desbordes, M., 466 DeSchriver, T.D., 336–7 Designated Player Rule, Major League Soccer, 332–3, 334, 335–7, 338 Deubert, C., 466 Deutscher, Christian, 309, 509, 544–50 diabetes, 46, 49 Dickson, A., 506 Dickson, G., 324 Diehl, M.A, 198, 527–8 Dietl, H.M., 131, 537 difficulty bias, 508 Dimant, Eugen, 544–50 direct expenditure approach (DEA), 356, 395, 397, 398, 401, 403 discrimination in Australian Rules Football, 327 in golf, 472, 475–6 in ice hockey, 313–15 open and closed leagues, 15 distance running, 441–50 age and performance, 448–9, 448 composition of marathon elite, 446, 446 gender and, 446, 447, 449, 520–1 incentives and performance, 444–6, 447 Parkrun, 412 prize structure, 442–6, 442, 443, 444 professional, 442–7, 442, 443, 444, 446 recreational, 447–9, 448 reference dependent preferences, 449, 506 diving, judging biases, 510 Do, A.Q., 479 Dobson, S., 227, 251, 557 Dohmen, T.J., 244, 509 Dolan, P., 360 domestic violence, 434, 435, 506 donations, money, 102–8, 105 Doncel, Luis M., 379 doping, 544–50 countermeasures, 549–50 detrimental effects of, 546 gender and, 520–1 individual decision-making, 546–9 peer effects, 546, 549 prevalence of, 545–6 road cycling and, 468–9, 545, 546

568

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Dorsey, R., 259 Dote, L., 106 Double Hurdle model, 37, 39, 75 Downward, Paul, 33–41, 50–1, 58, 60, 67, 68, 74, 93, 94, 95, 105, 407, 409, 410, 411, 428–35 draft see player draft Drayer, J., 193, 194, 196, 198, 420, 527, 528–9, 531 Drewes, M., 15 drugs see doping Du Bois, C., 453, 457 du Plessis, S., 357, 360, 363 Duckworth, F., 259 Dupuy, A., 463, 465–6 Dury, S., 107 Dutch Football League, 140 Dwyer, B., 198 Dye, R.F., 397 dynamic ticket pricing, 187, 328, 457, 525–32 primary ticket markets, 192, 193, 196, 528–31 secondary ticket markets, 197–8, 526–8, 529 Dyte, D., 454 Eakins, J., 76, 77 Eastman, B., 311–12 Easton, Stephen T., 308–17 ecological models, 75 econometric modelling sports participation, 36–7, 38–9, 47–8, 75 transfer market, 557 economic impacts of sports events crowding-out effects, 346, 350, 352, 352, 353, 362, 398, 399, 400, 402, 498 enhancing, 362–3 gap between ex-ante and ex-post analysis, 360–2 minor events and minor league teams, 395–403, 396, 397, 399, 401 NASCAR racing, 497–8 Olympic Games, 343–54, 345, 346, 347, 347, 348, 349, 350, 352, 358–9, 360, 387–90, 391–3 visitor groupings, 349–50, 350, 351 World Cup, 351, 356–8, 360, 361, 362, 387–90, 391–3 see also intangible effects of sports events economic impacts of sports facilities, 233–6, 359–60 economic objective functions, 125–32 profit maximization, 125, 128–30 revenue maximization, 125, 130–1 utility maximization, 125–8 economic value of sport, 18–28, 429–31, 430 employment, 19, 21, 23–4, 23, 24, 25, 25, 26 Gross Value Added (GVA), 19, 21, 23–4, 23, 24 multipliers, 19, 25–7, 26 National Income Accounting (NIA), 19, 20–1, 20 Sport Satellite Accounts (SSAs), 18, 21–3, 27–8 Vilnius definition of sport, 21–3, 22 volunteering, 27

education sports attendance and, 77 sports participation and, 37, 65–6, 76, 407 volunteering and, 94 willingness to pay and, 425 Ehrenberg, R.G., 477, 517, 520 Eide, E.R., 66–7 Eime, R.M., 41 Eisenberg, J., 165 Eisenberg, N., 103–4 El Hodiri, M., 126–7, 128–9, 164, 165 Elfenbein, D.W., 194 Elmaghraby, W., 526 Elo rankings, 222, 223, 223, 224, 225 Ely, Jeffrey, 155, 158, 168–9, 536 Emerson, J.W., 449, 510 employment economic value of sport, 19, 21, 23–4, 23, 24, 25, 25, 26 major events and, 357, 358, 359, 361, 388–9, 392 minor events and, 400, 401, 401 physical activity and, 64, 66–9 stadiums and, 236 volunteering and, 94, 98 see also baseball labor market; labour markets for talent; transfer market, football English Football League, 9, 135, 136, 138, 139 English Premier League, 139–40, 144, 150, 171–9 betting and, 535, 536–7 broadcasting rights, 171–4, 173 team performance, 214, 215, 216 television exposure and viewing demand, 171–9, 173, 175, 176, 177 transfer market, 204–5 English Premiership Rugby (EPR), 270, 271, 272, 272, 274, 275, 275 enterprise zones, 233 entrepreneurship theories of non-profit organizations, 84 entry discrimination, ice hockey, 314, 315 environmental influences, on sports participation, 40–1, 75, 77 Ernst & Young Terco, 356–7 Escobari, D., 526 ESEA see European Sport Economics Association (ESEA) Espindola, S., 316 Espitia-Escuer, M., 216 Esser, H., 94 Esteban-Cornejo, I., 65 Estonia, 24 EURO Football Cup, 215, 351, 396, 397, 397, 400 Eurobarometer Survey, 36, 50, 88 European Competition Authority, 173–4 European Court of Justice (ECJ), 147, 203, 244–5, 557 European Judo Union (EJU), 147 European Labour Force Survey, 25 European Rugby Cup (ERC), 269, 270, 272, 274–5 European Sport Charter, 33

INDEX

European Sport Economics Association (ESEA), 3–4 European Sport Management Quarterly, 3–4 European sports leagues Financial Fair Play (FFP) regulations, 139, 140, 245 labour market for talent, 10–11, 14–15, 138, 139 organizational features, 9, 10–11, 11–12, 136–41 origins, 9, 135–6 see also transfer market, football European Super League, 141 European Union Bosman ruling, 13, 14, 139, 203, 204, 208, 244–5, 557 competition policy cases, 148, 149–51 economic value of sport, 18, 21–8, 22, 23, 24, 25, 26 labour market for talent, 14–15 physical activity guidelines, 46 Sport Satellite Accounts (SSAs), 18, 21–3, 27–8 Vilnius definition of sport, 21–3, 22 Working Group on Sport and Economics, 19, 21 Eurostat, 25 Event Resource Demand (ERD), 397, 397, 398–400, 399 evidence-based approach to coaching, 559–60 evidence-based management, 554 expectation bias, 510–11 externalities of sport events, 353, 359–60, 428–35, 430 criminal behaviour, 434–5 demonstration and trickle-down effects, 405–6, 409–12, 428 domestic violence, 434, 435, 506 pride, 432–3 social capital formation, 433–4 subjective well-being, 360, 431–2 traffic, congestion, and pollution, 236, 237, 434 Fair, R.C., 448–9, 487–8 family structure sports attendance and, 77–8 sports participation and, 40, 76–7 Fan Cost Index, 185, 186 father–son rule, Australian Rules Football, 326 Fearing, D., 478 Feddersen, Arne, 358, 392, 412, 455 Federer, Roger, 456, 517 Fedewa, A.L., 65 Fedorko, I., 454–5 feel-good factor, 360, 406, 431–2 Feiler, Svenja, 82–8, 106, 107 Felfe, C., 57, 59, 65, 66 Feng, X., 234–5 Fenn, A.J., 419 Fernquist, R.M., 435 Feustel, E.D., 457 FIFA (Fédération Internationale de Football ­Association), 14, 146, 203, 363, 386–7, 389 see also World Cup fighting, in ice hockey, 315–17, 316

569

figure skating gender and, 521–2 judging biases, 510 final offer arbitration (FOA), baseball, 14 Financial Fair Play (FFP) regulations, 139, 140, 245 Finland, 24, 68, 73, 85, 86, 87 fitness club membership, 507 Flores, R., 157 Florida Suncoast Dome, 231 Flyvbjerg, Bent, 390, 391 football (soccer) betting, 246, 534, 535, 536–7, 539, 540 Bosman ruling, 13, 14, 139, 203, 204, 208, 244–5, 557 broadcasting rights, 171–4, 173 competitive balance, 141, 245, 246 criminal behaviour and, 237, 434, 435 domestic violence and, 434, 435, 506 economic impacts of minor events, 396, 397, 397, 400 economic impacts of World Cup, 351, 356–8, 360, 361, 362, 387–90, 391–3 economics literature, 243–53, 247, 248, 249, 250, 251, 251, 252 finance, 245–6 Financial Fair Play (FFP) regulations, 139, 140, 245 home advantage, 223, 224, 225, 226, 227, 228, 508, 509 judging biases, 508, 509, 511 labour market for talent, 14, 138, 139, 244–5, 248, 332–6 match fixing, 539, 540 origins of leagues, 135–6 penalties, 246 stadium attendance, 246 stock market floatation, 11, 245–6 team performance, 214–16 television exposure and viewing demand, 171–9, 173, 175, 176, 177, 246, 336, 338 ticket pricing, 183, 185–6, 531 see also European sports leagues; Major League Soccer (MLS); transfer market, football football, American see American football football, Australian Rules see Australian Rules Football football, rugby see rugby league; rugby union Football Association (FA), England, 9 Forbes magazine, 311 Forrest, David, 50, 174, 177, 185, 221, 251, 259, 380–1, 383, 420, 454, 534–41 Fort, Rodney, 12, 125–32, 154, 155, 156, 157, 158, 163, 167, 185, 271, 309 France economic value of sport, 24, 26 features of sports leagues, 137, 138, 139, 140 Olympic Games, 346, 346, 349 origins of sports leagues, 135, 136 rugby, 135, 138, 139, 140, 269, 270 sport employment, 26

570

THE SAGE HANDBOOK OF SPORTS ECONOMICS

franchise relocations see team relocations Franck, Egon, 140, 159 Frankel, Alexander, 155, 158, 168–9 Fraser v. MLS case, 333, 334 Frawley, S., 326 free agents baseball, 299–301, 300, 300, 301, 302, 311 basketball, 282–4, 283 ice hockey, 309, 311 Freeburn, L., 465, 466 Freeman, R.B., 92 French Football Amateur Federation, 139 French Football Federation, 140 French Rugby Championship, 135 Frey, B., 430 Frick, Bernd, 41, 50, 51, 97, 215, 251, 311, 412, 422, 441–50, 488, 494, 520–1 Fricke, H., 65–6 Fried, Harold O., 474, 475 Frisco, M.L., 57, 60 Frost, L., 326–7 Fuchs, B., 66 Fuchs, M., 526 Fujak, H., 326 Fuller, P.J., 323 Funahashi, H., 420 fundamental valuation of players, 556, 557–8 Galbraith, J.W., 510 Gallego, G., 526 Gallo, E., 227 gambling see sports betting game theory, 13, 140, 246, 289, 455, 463, 547 game uncertainty measures, 154–5, 157–8, 167 García, J., 37, 40, 41 Garcia-Cebrián, L.I., 216 Garcia-del-Barrio, P., 204–5, 215 Garcia-Villar, J., 166, 167 Garicano, L., 228, 509 Garnett, Harry, 268 Garratt, R.J., 449 Gärtner, M., 165 Gayton, W.F., 226, 227 Gearhart, D., 498 Geenens, G., 157 gender basketball and, 522 distance running and, 446, 447, 449, 520–1 doping and, 520–1 golf and, 472, 475–6, 520 human capital accumulation and sport, 66 money donations and, 105 NASCAR racing and, 499 Olympic performance and, 381–2 rank-order tournaments and, 518–23 response to competitive settings and, 447, 449, 453, 456, 516–23 sports attendance and, 77

sports participation and, 37, 40, 76, 407 tennis and, 453, 456, 519–20 track and field and, 520–1 volunteering and, 105 willingness to pay and, 424 winter sports and, 521–2 Gender Inequality Index (GII), 381 Generalised Method of Moments (GMM), 50–1 Gerlach, E., 57, 59 German Bundesliga, 140, 150, 185–6, 215, 216, 509 German Telekom, 546 Germany economic value of sport, 18, 23, 24, 26 employment and sports participation, 67–8 health and physical activity, 50 household sports consumption, 73, 73, 76 human capital accumulation, 65, 66 national pride and sport, 432, 433 negative impacts of sports events, 434 non-profit sports clubs, 82, 85, 86, 87 Olympic Games, 432, 433, 434 Olympic referenda, 371–2, 371 prevalence of doping, 545–6 social capital formation, 56–9, 57, 58 sport employment, 26 sports participation, 36 ticket pricing, 183, 185–6 volunteering, 97–8, 102 World Cup, 351, 357–8, 360, 361, 432, 433, 434 Gerrard, Bill, 211, 212, 281, 553–61 Geyer, H., 455 Giesecke, James A., 392–3 Gil, D., 518 Gilley, O.W., 476 Gilovich, T., 511 Gilsdorf, K.F., 455, 476, 477, 519, 520, 541 Gini coefficients, 156, 311, 335 Gladwell, Malcolm, 279 Glaeser, E.L., 233 Glass, W., 510 Gneezy, U., 518 Goddard, John, 251 Goff, B., 296 Gold Coast, Australia, 428 Goldschmied, N., 316 golf, 472–9 behavioral economics, 478, 506, 512–13 gender and, 472, 475–6, 520 golf course industry, 479 hole-by-hole and shot-by-shot data, 478 loss aversion, 478, 506 prize structure, 477, 494 production or earnings functions, 473–6 supply of effort, 472, 477, 512–13, 518 tournament entry decisions, 472, 476–7 Gómez, M.A., 222–3 Gomez-Gonzalez, Carlos, 452–8 Gómez-López, M., 511

INDEX

Gómez-Roso, J.M., 455 Gonsch, J., 526 Gottlieb, J.D., 233 Gouguet, J.J., 457 governmental failure theory of non-profit organizations, 83–4 Granger, C.W.J., 476, 520 Grant, D., 183 Grant Thornton, 357 Gratton, Chris, 18–28, 33–4, 74, 77, 139, 396, 407 Gravelle, H., 429 Greece, 24, 73, 85, 388, 389, 390 Green, B., 511 Greenbaum, R.T., 233 Greer, D.L., 508 Greer, T., 284 Gregory-Smith, Ian, 256–65 Grimes, A.Ray, 378 Groball, Guenther, 18–28 Groot, L., 157 Groothuis, J.D., 498–9 Groothuis, P.A., 418, 494–5, 498–9 gross gaming revenue (GGR), 535 Gross Value Added (GVA), 19, 21, 23–4, 23, 24 Grossman, M., 46–7 ground sharing, 326–7, 338 group-selection, 104 Grudnitski, G., 479 Guerrero, P.R., 181, 182, 185, 186 Guryan, J., 512 gymnastics judging biases, 508, 510–11 peer effects, 513 Haas, D.J., 215, 216, 333–4 habitus, 74 Haddock, D.D., 137 Hadley, L., 158 Hagn, F., 357, 358, 361 Hahn, J., 291–2 Hakes, J.K., 559 Hall, J.C., 236–7 Haller, Gordan, 484 Hallmann, Kirstin, 94, 95, 102–8 Hansmann, H., 84 happiness sports events and, 360, 431–2 sports participation and, 49–51 volunteering and, 96–7 Harger, K., 236 Häring, A., 56–9, 57 Harrington, D.E., 198 Harris, J., 284 Harris, J.G., 554 Hart, R.A., 165 Harter, J.F.R., 421 Hartmann, B., 105 Harvey, J., 96

571

Haugen, K.K., 157 Hayashi, R., 513 health sports events and, 360, 431–2 sports participation and, 45–51 volunteering and, 96–7 Health and Lifestyle Survey, UK, 49 health production model, 46–7, 49 Healy, K., 106–7 heart disease, 46, 49 Hecht, Rodolfo, 141 Heckelman, J., 317 Heckman model, 37, 39, 75, 76 hedonic pricing, 415, 479, 557 Hefner, F., 233–4 Heilmann, R., 165 Heinemann, K., 83 Hendricks, W., 295 Henley Centre for Forecasting, 18, 19 Henry, Thierry, 337 herding behavior, individual doping decisions and, 549 Herfindahl-Hirschman Index (HHI), 155–6, 379 Herzog, W., 57, 59 heuristic decision making, 294–6, 548 Heyndels, B., 453, 457 high jump, gender and, 521 Higham, J., 396 Hill, B., 449 Hiller, H.H., 372 Hinshaw, C.E., 165 Hirschman, A.O., 155 Hoehn, T., 141 Hoekman, R., 85 Hoffmann, Robert, 379 Hogan, V., 270, 272, 275 Holder, R.L., 221, 223, 224, 227, 456 Hollinger, John, 280 Hollingsworth, B., 106 home advantage, 219–28, 221 in cricket, 219, 222–8, 223, 225, 226, 227, 259–64, 262, 263 crowds and, 224, 226, 509 familiarity and, 224–5, 225 home team bias, 224, 226, 227–8, 259–64, 262, 263, 508–9 officials and, 219, 224, 226, 227–8, 259–64, 262, 263, 508–9 in Olympic Games, 379, 380, 381, 383, 508 rules and, 224, 225–8 in tennis, 226, 227, 456 Hood, Matthew, 476–7 Hopkins, J., 65, 66 Horch, H.-D., 83, 85 horizontal relationships, 56 Horn, B.P., 372 hot hand effects, 291, 511–12 Hotchkiss, Julie L., 358, 392 Household Panel Survey, UK, 36, 409

572

THE SAGE HANDBOOK OF SPORTS ECONOMICS

household production model, 27, 46, 74, 75, 76 household sports expenditure, 72–8 economic importance of, 72–3, 73 on sports attendance, 77–8 on sports participation, 73, 75–7 theoretical approaches, 74–5 household structure sports attendance and, 77–8 sports participation and, 40, 76–7 housing values and rents, 234–6, 359, 372, 389 Howard, D.R., 184 Hsu, S.H., 455 Huang, H., 50 human capital sports participation and, 37, 64–6, 74, 76 volunteering and, 94, 98, 103 Humphreys, Brad R., 19, 35, 40, 49, 50, 51, 113, 128, 155, 156, 157, 158, 168, 183, 184, 185, 186, 234–5, 236, 237, 272, 359, 423, 494, 497, 505–13, 521 Hungary, 24, 73, 85, 87 Hunt, J., 165 Huselid, M.A., 493 Hvattum, L.M., 246 Hynds, M., 258 hyperbolic discounting, 507 hypertension, 46, 49 Hyytinen, A., 68 IASE see International Association of Sports Economists (IASE) Ibsen, B., 85, 88 ICC see International Cricket Council (ICC) ice hockey, 308–17 coaching turnover, 312, 312–13 collective bargaining agreement (CBA), 308–9, 311, 315 competitive balance, 158 discrimination, 313–15 entry discrimination, 314, 315 fighting in, 315–17, 316 home advantage, 222, 224 player consistency, 285 player salaries, 308–9, 310–12, 310 salary discrimination, 314–15 secondary ticket markets, 195, 195, 201 in-stadium attendance, 165, 317 team relocations, 232 ticket pricing, 184, 185, 530–1 work stoppages, 308–11, 309 see also National Hockey League (NHL), US Ichniowski, C., 549 Idle, C., 466 Idsen, T., 311 income circular flow of, 20, 20 money donations and, 106 sports attendance and, 77, 164–5, 167

sports participation and, 40, 67, 68, 74, 76 volunteering and, 94, 106 willingness to pay and, 425 income–leisure trade-off model, 34–5, 35 Indian Premier League (IPL), 257, 259, 264 Indigenous Australian footballers, 325, 327 in-play betting, 536, 537–8, 539 input-output (I-O) models, 22, 344–6, 356, 358, 391, 395 in-stadium attendance see sports attendance instrumental variables models, 47–8, 49–50, 66–7 intangible effects of sports events, 353, 359–60, 428–35, 430 criminal behaviour, 434–5 demonstration and trickle-down effects, 405–6, 409–12, 428 domestic violence, 434, 435, 506 pride, 432–3 social capital formation, 433–4 subjective well-being, 360, 431–2 traffic, congestion, and pollution, 236, 237, 434 Intercollegiate Athletic Association of the United States (IAAUS), 114 intercollegiate sports see college sports, United States interdependence theory of non-profit organizations, 84 Interis, M.G., 422 International Association of Athletics Federation (IAAF) Diamond League, 442, 443–4, 444 International Association of Sports Economists (IASE), 3 International Cricket Council (ICC), 257, 258, 259, 260, 261, 264 International Cycling Union (UCI), 462, 465, 466, 468 International Judo Federation (IJF), 146, 147 International Olympic Committee (IOC), 343, 361, 363, 367, 372, 377–8, 385–6, 387, 389 International Padel Federation (FIP), 452 International Run for Peace, 433 International Skating Union (ISU), 510 International Social Survey Programme (ISSP), 50, 60 International Speedway Corporation (ISC), 493 International Table Tennis Federation (ITTF), 452 International Tennis Federation (ITF), 400, 453 International Triathlon Union (ITU), 482, 486 inter-seasonal uncertainty measures, 155–6, 158–9, 167 invariance proposition, 163, 204, 270 IOC see International Olympic Committee (IOC) IPL see Indian Premier League (IPL) Ireland, 24, 73, 76 Irish Rugby Football Union (IRFU), 269, 270, 271, 272, 274, 275, 276 Ironman triathlon, 482–90, 485, 486 demand, 489, 490 doping, 545 prize structure, 483, 483, 486, 488 Ishino, K., 512 Italian Football League, 214–15, 216 Italy economic value of sport, 24, 26 non-profit sports clubs, 87

INDEX rugby union, 271, 272, 275, 276 social capital formation, 58 sport employment, 26 ITU see International Triathlon Union (ITU) Iverson, Allen, 279, 281, 283, 284 Jackson, D., 455 Jackson, E.N.J., 106 James, Bill, 558 James, M., 85 Jamil, Mikael, 210–16 Janssens, P., 155 Japan economic value of sport, 18 national pride and sport, 433 Olympic Games, 358, 359, 411, 433 speedboat racing, 513 Jasina, J., 236, 309 Jennett, N., 155, 158, 246 Jetter, M., 455, 456 Jewell, R.T., 336, 337 Johansson, Lennart, 140 Johnson, B.K., 418, 419, 420, 424 Johnson, Candon, 385–93 Johnson, Daniel K.N., 379, 380 Jones, A., 49 Jones, H., 18 Jones, J., 311, 317 Jones, J.C.H., 126 Jones, J.H.C., 165 Jones, M., 165 Jones, M.V., 511 Joo, S., 106 Journal of Sport Management, The, 3–4 journeymen, baseball, 299, 302, 303, 304, 304 Judde, C., 157 judging biases, 507–12 difficulty bias, 508 home team bias, 224, 226, 227–8, 259–64, 262, 263, 508–9 nationalistic bias, 509–10 representativeness bias, 511–12 reputation/expectation bias, 510–11 judo, 146–7 Kääriäinen, J., 107 Kahane, L.H., 311, 312, 474 Kahn, L.M., 289, 453, 499 Kahneman, Daniel, 168 Kalist, D.E., 236 Kamada, T., 518 Kamenica, Emir, 155, 158, 168–9 Kantor, B.S., 259 Karnik, A., 259 Karp, L., 197 Katayama, H., 518 Kavetsos, G., 235, 360 Kay, T., 96

573

Keaney, E., 57, 58, 60 Keefer, Q.A., 296 Kemper, C., 457, 531 Kendall, G., 327 Kenkel, D., 49 Kenseth, Matt, 495 Kent, R.A., 157 Kern, M., 215 Kern, William, 113, 473 Késenne, Stefan, 13, 76, 77, 130, 155, 203–9 Keskinocak, P., 526 Kessler, Jeffrey, 119 Kiefer, S., 96, 421 Kim, J.W., 511 Kim, M., 106 King, B.G., 511 kin-selection, 104 Kirby, S., 434 Kitchens, C.T., 295 Klaassen, F.J., 455, 519 Klaeren, R., 488 Klären, R., 447 Knaus, M.C., 65, 66 Knechtle, B., 488 Knittel, C.R., 478–9 Kocher, M.G., 228 Kokolakakis, Themis, 18–28, 40 Koning, R.H., 156, 227, 246, 456 Konjer, M., 453 Koski, P., 86 Koutrou, N., 94, 95 Koyama, M., 223, 228 Krautmann, Anthony C., 155, 158, 185, 298–304, 311 Krawczyk, M., 449 Krueger, A., 47–8 Krumer, A., 454, 457, 519 Kubatko, Justin, 280 Kuethe, T.H., 334–5 Künemund, H., 105 Kuo, T., 40 Kurscheidt, M., 357, 361 labour market outcomes sports participation and, 64, 66–9 volunteering and, 98 labour markets for talent, 10–11, 14–15 Australian Rules Football, 323–8 European sports leagues, 10–11, 14–15, 138, 139 football (soccer), 14, 138, 139, 244–5, 248, 332–6 ice hockey, 311–12 Major League Soccer, 332–6 NASCAR racing, 498–9 performance analytics in, 558–9 player valuation, 556–8 road cycling, 465–6 see also baseball labor market; player salaries; transfer market, football

574

THE SAGE HANDBOOK OF SPORTS ECONOMICS

labour stoppages, 308–11, 309 Lackner, M., 520, 521 Ladany, S., 165 Ladies Professional Golf Association (LPGA) Tour, 473–4, 475–6, 477, 520 Lagae, W., 466, 467, 469 Lago-Peñas, C., 511 Lahtonen, J., 68 Lahvička, J., 454 Lambrinos, J., 312 Lamla, M.J., 397, 400 Landers, J., 233 Lang, M., 131, 140 Langer, V.C.E., 358 Lapchick, Richard, 499 Larsen, K., 509 Larson, D., 463, 466 Latvia, 24 Laurence, J., 107 Lavoie, M., 315 Lawson, R.A., 336–7 Lazear, E.P., 444, 494, 517–18, 550 Leach, S., 215, 216 league standing effect measure, 155 Leatherdale, S.T., 411 Lechner, Michael, 49, 64–9 Lee, D.Y., 236 Lee, Y.H., 155, 158, 215, 476 Lee, Y.-j., 107 Leeds, Eva Marikova, 377–83, 521–2 Leeds, Michael A., 350, 381, 382, 516–23 Lees, C., 65, 66 Lefebvre, L., 508 leg before wicket (LBW) decisions, cricket, 228, 259, 260–4, 262, 263 Lehman, D.R., 508 Lehman, D.W., 291–2 Lehtonen, H., 107 Leisure Industries Research Centre, 18 Leiter, A.M., 421 Lenor, S., 327 Lenten, L.J.A., 157, 221, 259, 325, 326, 327–8 Leper, R., 488 Lera-López, Fernando, 72–8 Leslie, P., 192 Levitt, S., 317 Levitt Report, 308 Lewis, A., 259 Lewis, K., 165 Lewis, M., 211, 558 Li, J., 198 Liaw, T., 473 life satisfaction see subjective well-being (SWB) lifestyle choices health and, 45, 46–7, 48–9 sports participation and, 37–40 Lilley, A., 105 Lim, C., 107

Lipscomb, S., 65 Lithuania, 18, 24 Locke, S.L., 296 Logit/Probit estimators, 36, 38 London Marathon, UK, 442, 442, 443 London Olympics, UK, 360, 373, 388, 391, 405–6, 409, 410, 428, 431–2, 434 Long, J.G., 231 Longley, N., 313, 314, 315 Lopez, M., 291 Lorenc, T., 411 Los Angeles Olympics, USA, 343, 344, 348, 349–50, 350, 373 loss aversion, 478, 506–7 Louis–Schmeling paradox, 210 Lowen, A., 380, 381–2 Loyland, K., 76, 77 Lui, H., 378, 380 Luxembourg, 24 Lye, J., 323 Lynch, D., 445 Lynch, J.G., 517 Lyócsa, Š., 454–5 Lyons, M., 372–3 McCartney, G., 409 McCormick, R.E., 445 Macdonald, Robert, 166, 167, 181–2, 186, 258, 323, 324, 326 McHale, I.G., 50, 454, 456, 540 Machol, R., 165 Mackenzie, R., 216 McMillan, J., 269 McSharry, P.E., 225 Madden, John R., 392–3 Madsen, R.A., 498 Maennig, Wolfgang, 235, 356–63, 367–72, 391–2, 432, 545, 546, 547, 550 Magnus, J.R., 455, 519 Mahtani, K.R., 409 Major League Baseball (MLB), 9, 12 antitrust exemption, 146 competitive balance, 146, 158 congestion and, 237 economic impacts of, 359 judging biases, 511 marginal revenue product (MRP) of players, 211–12, 298, 557–8 Moneyball story, 558–9 player consistency, 285–6 sabermetrics, 558–9 secondary ticket markets, 193, 195, 195, 196, 198, 201 team relocations, 232 ticket pricing, 183, 184, 185, 193, 457, 525, 526–7, 528–30 work stoppages, 308, 309 see also baseball labor market

INDEX

Major League Soccer (MLS), 12, 216, 331–8 audience demand, 336–7, 338 Designated Player Rule, 332–3, 334, 335–7, 338 player salaries and team performance, 333–6, 337–8 single-entity structure, 331–2, 333, 334, 335 stadiums, 338 major sports events crowding-out effects, 346, 350, 352, 352, 353, 362, 498 demonstration and trickle-down effects, 405–6, 409–12, 428 enhancing impacts of, 362–3 gap between ex-ante and ex-post impact analysis, 360–2 national pride and, 432–3 social capital formation and, 433–4 visitor groupings, 349–50, 350, 351 see also intangible effects of sports events; Olympic Games; World Cup Makarova, E., 57, 59 Malmendier, U., 507 Maloney, M.T., 445 Malta, 24 Malueg, D.A., 455 Mandle, W.F, 259 Mano, Y., 420 Manski, C.F., 512 marathon running, 441–50 age and performance, 448–9, 448 composition of marathon elite, 446, 446 gender and, 446, 447, 449 incentives and performance, 444–6, 447 prize structure, 442–6, 442, 443, 444 professional, 442–7, 442, 443, 444, 446 recreational, 447–9, 448 reference dependent preferences, 449, 506 Marburger, D.R., 185 March, D.S., 447 marginal revenue product (MRP) of players, 517 baseball, 211–12, 298, 557–8 ice hockey, 311–12, 314 Marie, O., 237 Marin, Carolina, 458 marital status, sports participation and, 40, 76, 407 market failure theory of non-profit organizations, 84 Marris, R., 126 Martínez, Conchita, 457 Massey, C., 292–4 Massey, Patrick, 268–76 Masters, A., 544 Mastromarco, C., 157 match fixing, 457, 538–41 match relevance measure, 155 Match Review Panel (MRP), Australian Rules Football, 325 Matheson, Victor A., 350, 357, 358, 361, 392, 393, 400, 401, 498

575

Mathias, E., 333, 334 Matthews, P.H., 477, 520 Mattila, A.S., 526 Mauresmo, Amelie, 457 Maxcy, J., 154, 156, 311, 463, 466 Medoff, M., 212 mega-events see major sports events Mehra, S.K., 149 mental health sports events and, 360, 431–2 sports participation and, 49–51 volunteering and, 96–7 Mignot, J.F., 463 Milano, M., 19 Miller, P.A., 103–4 Mills, B.M., 158, 511 Mincer, J.A., 220 minor league teams, 395–403, 396, 397, 399 minor sports events, 395–403, 396, 397, 399, 401 Mishra, V., 259 Mitchell, H., 358 MLB see Major League Baseball (MLB) MLS see Major League Soccer (MLS) Mohr, P.B., 509 Molina, D.J., 336 Möller, M., 445–6 Mondello, M., 186, 311 money donations, 102–8, 105 Moneyball story, 558–9 Monks, J., 335 monopsony labour markets, 9, 11, 14–15, 206, 557 ice hockey, 311–12 Major League Soccer, 332–6 see also baseball labor market; transfer market, football Montolio, D., 237 Montreal Olympics, Canada, 344, 390 Moore, D., 194 Morgan, H.N., 508 Morgan, O.A., 423 Morley, B., 259 Morris, C., 455 Morrow, S., 466 Morton, A., 454 Moser, Katharina, 441–50 Mosurski, K., 455 Motamed, M., 334–5 motorsports see NASCAR racing Mounts, S., 487–8 Moy, R.L., 473 MPCC (Mouvement Pour un Cyclisme Crédible), 465 Multiregional Input-Output Model (MRIO), 22 multi-tiered ticket pricing, 184 multivariate probit models, 48 Muñiz Rui, Cristina, 33–41 Munting, R., 534 Munyo, I., 506 Murphy, K., 547

576

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Musgrave, R.A., 83 Musick, M., 98 Mutter, F., 41 Myers, T.D., 510 NAASE see North American Association of Sports Economists (NAASE) Nadal, Rafael, 517 Nagel, S., 85, 94, 95, 107 Nalbantis, Georgios, 154–9, 157, 422, 424 NASCAR racing, 492–9 audience demand, 493, 495–6, 497 contest design, 493–5 diversity issues, 499 economic impact of, 497–8 labor market, 498–9 Peltzman effect, 496–7 prize structure, 493–5 safety, 496–7 season championship, 495 Nash equilibrium model, 13–14 NASL see North American Soccer League (NASL) National Association of Baseball Players, US, 9 National Basketball Association (NBA), US, 9, 116, 279–86 competitive balance, 158, 285, 286 data and streaming rights, 538 gender and, 522 home advantage, 508, 509 hot hand effects, 511, 512 objective measurement of performance, 279–81, 282 perceptions of performance, 281–5, 283, 284 in-play betting, 538 player consistency, 285–6 player salaries, 282–4, 283 secondary ticket markets, 195, 195, 202 team relocations, 232 ticket pricing, 184, 185 work stoppages, 309 National Basketball League (NBL), Australia, 324 National Collegiate Athletic Association (NCAA), 112–20, 281, 282 broadcasting rights, 114–16, 117, 118 challenges to cartel operations, 116–18 court cases, 115, 118, 119 player compensation, 114, 116, 117, 118–20 ticket pricing, 193 National Football League (NFL), US, 9, 11, 15, 116, 144, 289–96 behavioral economics, 294–6 competitive balance, 158 criminal behaviour and, 434–5 draft decisions, 292–3, 294–6 economic impacts of, 234, 235, 359 in-game decisions, 291–2 happiness and wins, 432 heuristic decision making, 294–6

player consistency, 285 psychological factors affecting maximization, 291–4 secondary ticket markets, 194–5, 195, 198, 201 team relocations, 231, 232 ticket pricing, 184, 185, 527–8 work stoppages, 309 National Hockey League (NHL), US, 9, 165, 308–17 2004–05 player lockout, 308–11, 309 coaching turnover, 312, 312–13 collective bargaining agreement (CBA), 308–9, 311, 315 competitive balance, 158 discrimination, 313–15 entry discrimination, 314, 315 fighting, 315–17, 316 player consistency, 285 player salaries, 308–9, 310–12, 310 salary discrimination, 314–15 secondary ticket markets, 195, 195, 201 team relocations, 232 ticket pricing, 184, 185, 530–1 National Hockey League Players Association (NHLPA), US, 308 National Income Accounting (NIA), 19, 20–1, 20 National Labor Relations Act 1935, US, 116 National Population Health Survey (NPHS), Canada, 36, 49, 68 national pride, sports events and, 432–3 National Rugby League (NRL), Australia, 324, 326 National Union of Professional Players (UNFP), France, 139 nationalistic bias, 509–10 NBA see National Basketball Association (NBA), US NCAA see National Collegiate Athletic Association (NCAA) Neale, Walter C., 9, 154, 164, 210 neighborhood revitalization, stadiums and, 233–6 Nekby, L., 449 nepotism, in auto racing, 498–9 Nesbit, T.M., 497 Netherlands economic value of sport, 18, 24 features of sports leagues, 140 household sports consumption, 73 national pride and sport, 432, 433 non-profit sports clubs, 85, 87 World Cup, 432 Nevill, A.M., 221, 223, 224, 226, 227, 228, 456, 509 New York Marathon, USA, 442, 442, 443 New Zealand, 269 NFL see National Football League (NFL), US NHL see National Hockey League (NHL), US NIA see National Income Accounting (NIA) Nichols, G., 85 NIMBY (Not In My Backyard) attitude, 371 Nobel Prize for Economics, 168 Noland, Marcus, 379, 382, 383 Nolen, P., 518–19

INDEX

Noll, Roger, 9, 136, 137, 138, 164–5 non-cognitive skills, sports participation and, 66 non-profit organizations, theories of, 83–4 non-profit sports clubs functions of, 83 funding, 86–7 goals of, 85–6, 87, 92 organization of, 82–8 role identity in, 82, 95–6 volunteering in, 84, 86, 88, 92–9, 93, 102 Norman, W.C., 396 North American Association of Sports Economists (NAASE), 3–4 North American Soccer League (NASL), 331, 332, 338 Northern Rugby Football Union, England, 135 Norton, P., 259 Norway human capital accumulation, 65 national pride and sport, 432 negative impacts of sports events, 434 non-profit sports clubs, 85, 87, 88 social capital formation, 57, 58, 59, 60 Nowak, A., 235 Oakland Athletics, 558–9 Oaxaca, R., 475 O’Bannon, Ed, 119 obesity, 45, 46, 49 objective functions see economic objective functions O’Brien, W., 95 Oettinger, G.S., 478 officials and bias in Australian Rules Football, 327, 328, 509–10 in cricket, 227–8, 257, 259–64, 262, 263 home advantage and, 219, 224, 226, 227–8, 259–64, 262, 263, 508–9 judging biases, 224, 226, 227–8, 259–64, 262, 263, 507–11 nationalistic bias, 509–10 reputation/expectation bias, 510–11 O’Hara, M., 310 O’Keeffe, M., 518 Olympic Games, 385–93 cost overruns, 390–1 criminal behaviour and, 434 demonstration and trickle-down effects, 405, 406, 409, 410–11, 412, 428 economic impacts of, 343–54, 345, 346, 347, 347, 348, 349, 350, 352, 358–9, 360, 387–90, 391–3 feel-good factor, 360, 406, 431–2 home advantage in, 379, 380, 381, 383, 508 host city selection process, 385–6 hosting costs, 386–7, 389, 390–1 intangible effects, 353, 359–60, 431–2, 433, 434 judging biases, 510 national performance, 377–83, 406, 406 national pride and, 432, 433 opportunity costs of, 388

577

public referenda, 362, 363, 367–72, 368–70, 371 public willingness to pay for, 360, 367–72, 368–70, 371 subjective well-being and, 431–2 opportunity costs of hosting sports events, 388, 398, 399, 400, 401, 402 of sports attendance, 167–8, 258 of sports participation, 35, 40 of volunteering, 97, 98 organizational capacity theory of non-profit organizations, 84 Organizing Committees for the Olympic Games (OCOGs), 386 Orlowski, Johannes, 97, 415–25, 421 O’Roark, J.B., 495, 496–7 Orszag, J.M., 477 Osikominu, A., 66 osteoporosis, 46 Otamendi, Javier, 379 outcome uncertainty, 210 cricket, 258, 264 dynamic ticket pricing and, 527, 529–30, 531 game uncertainty measures, 154–5, 157–8, 167 inter-seasonal uncertainty measures, 155–6, 158–9, 167 road cycling, 469 seasonal uncertainty measures, 155, 158, 167 sporting prizes and, 141 sports attendance and, 156–9, 163, 165, 166, 167, 210, 258, 264, 506 television viewing demand and, 158–9 Ovaska, T., 454 Owen, J.G., 418 Owen, P.D., 275 Owens, M.F., 296 Ozbeklik, S., 513 padel, 452 Page, K., 509, 510 Page, L., 509, 510 Pagliero, M., 196 Palacios-Huerta, I., 455 Palomino, F., 274 Pan, X., 184 Pan American Junior Athletic Championships, 401–2, 401 Papa, S., 58, 60 Pappa, E., 358 Pareto efficiency, 429 Paris Olympics, France, 346, 346, 349 Parker, D., 259 Parkrun, 412 Paserman, M.D., 455, 456, 519 Passer, M., 508 Paton, David, 256–65 Patrick, Danica, 499 Paul, Rodney J., 187, 317, 525–32 Paulden, T., 457

578

THE SAGE HANDBOOK OF SPORTS ECONOMICS

Pawlowski, Tim, 41, 50, 54–61, 57, 65, 66, 75, 76, 77, 78, 154–9, 157, 185–6, 420 Paxson, C., 67 Peck, J., 197 peer effects, 512–13 doping and, 546, 549 Peeters, T., 140, 157 Pele, 331 Pelnar, G., 184–5 Peltzman, S., 496 Peltzman effect, 496–7 performance analytics, 553–61 data analytics as management tool, 553–5, 555 evidence-based approach to coaching, 559–60 Moneyball story, 558–9 in player recruitment, 558–9 player valuation, 556–8 tactical performance data, 556, 557, 560 performance-enhancing drugs (PEDs) see doping Perks, T., 58, 60 Perloff, J.M., 197 permanent income hypothesis, 289 Pershin, V., 379 personal relationships, 55, 56–9, 57 personality traits, sports participation and, 66 Peru, 57, 59, 61, 65 Pettersson-Lidbom, P., 226, 228, 509 Pfeffer, J., 554 Pfeifer, C., 67, 68 Pfitzner, C.B., 474, 495 Phatarfod, R., 259 physical activity economic determinants of participation in, 507 employment and, 64, 66–9 health and, 45–51 see also sports participation Pierre, J., 400 Pigou, Arthur, 182 Planells-Struse, S., 237 planned behaviour theory, 104 player compensation US college sports, 114, 116, 117, 118–20 see also player salaries player contracts see labour markets for talent; transfer market, football player draft Australian Rules Football, 323, 324–5, 326, 327–8 National Basketball Association (NBA), US, 284 National Football League (NFL), US, 292–3, 294–6 National Hockey League (NHL), US, 313–14, 315 player mobility, 10–11, 14–15 see also transfer market, football player salaries Australian Rules Football, 323, 324, 325–6, 328 basketball, 282–4, 283 ice hockey, 308–9, 310–12, 310 Major League Soccer, 332–6, 337–8

pay inequality and team performance, 311, 335–6, 518 rugby union, 270, 270, 272, 274 transfer market and, 204, 206–7, 206, 207 see also baseball labor market; labour markets for talent player unions, 14, 15, 116, 139, 308, 334, 549 player valuation, 556–8 see also marginal revenue product (MRP) of players playoff uncertainty measure, 155, 167 Plessner, H., 510–11 Plumley, D., 274 plus-minus model of performance, 280 Poland, 18, 24, 73, 87 pole vault, gender and, 521 political regime, Olympic performance and, 378, 379, 380, 381, 382, 383 Pollard, G., 219, 221, 222, 223, 224 Pollard, R., 219, 221, 222, 223, 224 Pommerehne, W., 165 Pope, A.T., 497 Pope, D.G., 478, 506 Poplawski, W., 310 Portugal, 18, 24, 73 Potwarka, L.R., 411 present bias, 507 Preston, A., 549 Preston, I., 259, 546, 549 Preuss, Holger, 140, 343–54 pride, sports events and, 432–3 Prieto-Rodríguez, J., 454 Priks, M., 226, 228, 509 Prinz, Joachim, 445, 464, 482–90 prisoner’s dilemma, 547 private consumption model, 103 prize structure distance running, 442–6, 442, 443, 444 golf, 477, 494 Ironman triathlon, 483, 483, 486, 488 NASCAR racing, 493–5 road cycling, 464, 467, 467 tennis, 455, 494, 517 Pro12/Pro14 (rugby union), 271, 272, 272, 274, 275, 275, 276 Prochaska, J.O., 407, 408 production frontiers, 215–16 production functions golf, 473–6 Olympic performance, 378, 381, 382 team performance, 211–15 Professional and Amateur Sports Protection 1993, USAct, 534 Professional Footballers Association (PFA), England, 139 profit maximization, 125, 128–30, 273 transfer market and, 204, 205, 205, 206, 206 property values and rents, 234–6, 359, 372, 389

INDEX

Propheter, G., 236 prospect theory, 505–6 Prowse, V., 518 Prudhomme, Christian, 465 psychological approaches altruistic behaviour, 103–4 sports participation, 74–5, 407 psychological continuum model, 75 public goods theory, 83–4, 103 public investments in sport stadium subsidies, 231–4, 237, 327, 417 willingness to pay, 360, 367–72, 368–70, 371, 415–25, 418–23 public referenda, 362, 363, 367–72, 368–70, 371 Pujols, Albert, 300, 300 Putnam, Robert, 54, 55, 56, 60, 96 Pyun, H., 236–7 Qatar World Cup, 386, 389 Quirk, J., 126–7, 128–9, 131, 157, 164, 165 racial discrimination, in Australian Rules Football, 327 racquet sports, 452, 458 see also tennis Rahn, W., 60 Ramchandani, G., 396, 408 Ramcharan, R., 259 Ramón, N., 456 random utility maximization model, 416, 425 rank-order tournaments, 517–18 gender and, 518–23 Rank-Order-Tournament (ROT) theory, 494 Rapún-Gárate, M., 77 Rascher, D.A., 127, 183, 184, 187, 400 Rasciute, S., 40, 50 ratio of standard deviations (RSD), 155 rational addiction theory, 547 rational-choice framework, 74, 93–4 Raya, Jose, 488 Reade, J.James, 219–28 Rebeggiani, L., 466 Rees, D.I., 65, 236 Rees, P., 274 Rees, R., 429 referees see officials and bias reference dependent preferences (RDP), 449, 505–7 Regan, T.H., 498 Reifman, A., 508 Reilly, B., 557 relational goods, 97 Rendleman, R.J., 512–13 replacement cost approach to volunteering, 97, 98 representativeness bias, 511–12 reputation bias, 510–11 Resaland, G.K., 65 reserve clauses, baseball, 298, 305, 557–8 resource dependence theory, 84 revenue maximization, 125, 130–1

579

revenue sharing, 11, 129, 130, 131 Australian Rules Football, 323–4 cricket, 264 ice hockey, 309, 311 rugby union, 270, 272, 273, 274–5 youth training compensation system, 208–9 Rheinberger, C.M., 421 Rhoads, T.A., 477 Richardson, D., 311 Richter, F.J., 358 Riedl, D., 509 Ringrose, T.J., 228 Ringstad, V., 76, 77 Rio de Janeiro Olympics, Brazil, 147, 343, 382, 385–6, 388, 391 Riordan, J., 35 Rishe, P., 186, 193 Rishel, T.D., 474, 495 Ritcher, Felix, 391–2 Ritchie, J.R.B., 372–3 Roach, M.A., 296 road cycling, 462–9 challenges, 468–9 competition structure, 463–4, 466–7, 466 doping, 468–9, 545, 546 economic value, 467, 467 institutional setting, 465–6 particular nature of, 463–5 public road use, 464–5 sponsorship, 464, 465, 466, 467, 468, 469 television rights and demand, 465, 467, 467, 468, 546 Robbins, D., 456 Rockerbie, Duane W., 308–17 Rodenberg, R.M., 457 Rodriguez, Placido, 163–9 Rodríguez-Guerrero, P., 166, 167 Rodriguez-Gutiérrez, C., 463 Rogge, N., 464 Rohde, N., 259 Romania, 24 Romer, D., 289, 290–1, 292, 294, 296 Ronan, N., 66–7 Roosevelt, Theodore, 114 Rooth, D.O., 67 Rose, Andrew K., 358, 359, 391–2 Rose, N.L., 184 Rosen, S., 182, 444, 517–18, 519 Rosenfield, A.M., 182 Rosentraub, M.S., 233, 234 Rossi, Lea, 102–8 Rossi, M.A., 506 Rotolo, T., 104 Rottenberg, Simon, 126, 129, 138, 154, 163, 181, 204, 210, 211, 243, 270 Rotthoff, K.W., 236, 292, 309, 455, 498, 508, 513 round-robin competitions, 9, 135, 138–9 Rugby Football Union (RFU), 268–70, 274 rugby league, 213–14, 269

580

THE SAGE HANDBOOK OF SPORTS ECONOMICS

rugby union, 135, 268–76 attendances, 275, 275, 276 competitive structures, 269 evidence-based approach to coaching, 559–60 league structures and governance, 271–2, 272 player salaries, 270, 270, 272, 274 revenue, 272–5, 273 transition to professionalism, 269–71, 275–6 Ruiz, J.L., 456 Runkel, M., 157 running gender and, 520–1 Parkrun, 412 see also distance running Ruseski, Jane E., 19, 35, 40, 45–51 Russel, S., 22 Russia, 343, 386, 390, 391 Rustichini, A., 518 Ryan, M.E., 221 Saayman, A., 396 Saayman, M., 396 sabermetrics, 558–9 Sabia, J.J., 65 Sacheti, Abhinav, 228, 256–65 safety Ironman triathlon, 487 NASCAR racing, 496–7 Sakovics, J., 274 Sala, B.R., 510 Salaga, S., 536–7 Salamon, L.M., 84 salaries, player Australian Rules Football, 323, 324, 325–6, 328 basketball, 282–4, 283 ice hockey, 308–9, 310–12, 310 Major League Soccer, 332–6, 337–8 pay inequality and team performance, 311, 335–6, 518 rugby union, 270, 270, 272, 274 transfer market and, 204, 206–7, 206, 207 see also baseball labor market; labour markets for talent salary arbitration baseball, 14, 299, 303, 306 ice hockey, 312 salary caps, 11, 14 Australian Rules Football, 323–4, 328 ice hockey, 309, 310–11, 315, 316 Major League Soccer, 332, 335–6 rugby union, 270, 272, 273 salary discrimination in golf, 472, 475 in ice hockey, 314–15 open and closed leagues, 15 Saloner, G., 535 Salt Lake City Olympics, USA, 350, 367, 373, 380, 388, 391, 392, 434

Sampaio, B., 510 Samuelson, P.A., 83 San Francisco Giants, 195, 196, 198, 231, 457, 525, 526, 528, 529 Sandercock, L., 323 Sanderson, Allen R., 112–20 Santo, C.A., 419 Saracens, 274, 559–60 Sari, N., 49, 68 Sarma, S., 49, 51 Sauer, R.D., 512, 559 Sauermann, J., 244 Scarf, P., 259 Scelles, Nicolas, 135–41, 155, 463 Scharfenkamp, Katrin, 441–50 Scheel, F., 520–1 Scheer, J.K., 510–11 Schlesinger, T., 94, 95, 107 Schmidt, M.B., 158, 279–80, 285 Schmitz, H., 448 Schnepel, K.T., 236 Schofield, J.A., 136, 137, 212–13 Schreyer, D., 183, 187 Schupp, J., 105 Schut, P.O., 400 Schüttoff, Ute, 54–61, 57, 58 Schwartz, B., 222, 224 Schwartz, J.T., 494 Schwarz, A.D., 183, 184, 187 Schweitzer, M.E., 478, 506 Scitovsky, T., 407 Scottish Rugby Union (SRU), 269, 270, 271, 272, 274, 275, 276 Scrivens, K., 55–6 Scully, G.W., 211–12, 298, 311, 474, 557–8 season tickets, 183, 186, 187, 193, 194, 196, 198, 526, 529, 532 seasonal uncertainty measures, 155, 158, 167 secondary ticket markets, 190–9 dynamic pricing, 197–8, 526–8, 529 fairness and legitimacy, 196 price elasticity, 527–8 reasons for, 192–3 resale laws and restrictions, 194 season tickets and, 193, 194, 196, 198 size of, 192 sponsored resale marketplaces, 194–7, 195, 201–2 traditional and modern ticketing models, 191–2, 191 segregation discrimination, 314 Seippel, Ø., 58, 60, 88 self-control fitness club membership and, 507 individual doping decisions and, 548 self-determination theory, 75, 104 self-esteem, physical activity and, 66 Seltzer, R., 510 Shapiro, S.L., 196, 198, 420, 528–9, 531

INDEX

Sherman Antitrust Act, US, 115, 333 Shibli, S., 411 Shivakumar, R., 260, 261 Shmanske, Stephen, 472–9, 520 Siegfried, J., 165, 327 Siegfried, John J, 112–20 Simmons, R., 177, 183, 185, 187, 251 Singh, A., 65 single-entity model, Major League Soccer, 331–2, 333, 334, 335 single-point-of-sale efficiency, 149, 150 skiing, gender and, 521 Sloane, Peter J., 125–6, 127, 136, 137, 138, 139, 165, 245 Slonim, R., 105 Slovakia, 24 Slovenia, 24, 73 SMAANZ see Sport Management Association of Australia and New Zealand (SMAANZ) Smith, Adam, 517 Smith, C., 55–6 Smith, I., 258 Smith, J.J., 66 Smith, J.K., 513 Smyth, R., 259 Sobel, R.S., 221, 497 soccer see football (soccer) Sochi Olympics, Russia, 343, 386, 390, 391 social accounting matrix, 356 social capital formation, 54–61, 57–8 civic engagement, 55, 57–8, 59–60 personal relationships, 55, 56–9, 57 social network support, 55, 57, 59 sports events and, 433–4 trust and cooperative norms, 55–6, 58, 60 volunteering and, 56, 58, 59, 96, 106, 107 social contagion, individual doping decisions and, 546, 549 social integration theory, 104, 106 social learning theory, 104 social network support, 55, 57, 59 societal benefits approach to volunteering, 97, 98 socio-biological approaches, altruistic behaviour, 104 socio-demographic characteristics money donations and, 105–6 sports attendance and, 77 sports participation and, 37–40, 76, 407 volunteering and, 105–6 willingness to pay and, 424–5 Socio-Economic Panel, Germany, 36, 66, 67–8 sociological approaches altruistic behaviour, 104 sports participation, 74–5, 407 Soebbing, Brian P, 181–8 Solberg, H.A., 139 Solow, J., 496, 497 Sommers, P.M., 473 Sorensen, A., 192

581

Sottas, P.-E., 545 South Africa national pride and sport, 432 rugby, 269, 272 World Cup, 351, 357, 360, 362, 432 South Australian National Football League (SANFL), 323, 327, 328 South Korea, 388, 433, 434 Sowell, C., 487–8 Spain economic value of sport, 24, 26 household sports consumption, 73, 73 non-profit sports clubs, 85, 87 Olympic Games, 358, 359 sport employment, 26 sports participation, 37 Spanish La Liga, 148, 204–5, 215, 216, 511 speedboat racing, 513 Speedway Motorsports International (SMI), 493 Spiegel, Mark M., 358, 359, 391–2 sponsored ticket resale marketplaces, 194–7, 195, 201–2 sponsorship global value of, 535 Ironman triathlon, 483 NASCAR racing, 494, 495, 498 non-profit sports clubs, 87 Olympic Games, 343 road cycling, 464, 465, 466, 467, 468, 469 rugby union, 273, 273, 274 Sport England, 27, 406 Sport Industry Research Centre (SIRC), 19, 27 Sport Management Association of Australia and New Zealand (SMAANZ), 4 Sport Management Review, 3–4 Sport Satellite Accounts (SSAs), 18, 21–3, 27–8 Sporting Future strategy, UK, 18–19, 406, 428 Sportradar, 538 sports attendance, 163–9 American football, 164, 165 Australian Rules Football, 322, 323, 324, 325, 326, 327 baseball, 163, 164–5 basketball, 165, 285, 286 cricket, 258, 264–5 football (soccer), 246 household expenditure on, 77–8 ice hockey, 165, 317 Major League Soccer, 336–7, 338 NASCAR racing, 493, 495–6 opportunity costs of, 167–8, 258 outcome uncertainty and, 156–9, 163, 165, 166, 167, 210, 258, 264, 506 reference dependence and loss aversion and, 506 rugby union, 275, 275, 276 suspense and surprise and, 155, 158, 168–9 ticket pricing and, 181–2, 184, 186, 187 violent play and, 317, 325

582

THE SAGE HANDBOOK OF SPORTS ECONOMICS

sports betting, 534–41 data and streaming rights, 537–8 demand for sport and, 536–7 football (soccer), 246, 534, 535, 536–7, 539, 540 match fixing, 457, 538–41 in-play betting, 536, 537–8, 539 scale and growth of, 535–6 sports rights, 537 tennis, 454–5, 536, 538, 539 sports clubs see non-profit sports clubs sports events demonstration and trickle-down effects, 405–6, 409–12, 428 health and, 360, 431–2 household expenditure on, 77–8 mental health and, 360, 431–2 national pride and, 432–3 opportunity costs of hosting, 388, 398, 399, 400, 401, 402 public referenda, 362, 363, 367–72, 368–70, 371 public willingness to pay for, 360, 367–72, 368–70, 371, 415–25, 418–23 social capital formation and, 433–4 sports participation and, 41 see also economic impacts of sports events; intangible effects of sports events; secondary ticket markets; sports attendance; ticket pricing sports facilities Australian Rules Football, 326–7 congestion and, 236, 237 criminal behaviour and, 236–7 ground sharing, 326–7, 338 impacts on employment and business, 236, 359 impacts on property values and rents, 234–6, 359, 372 impacts on urban development, 359–60 Major League Soccer, 338 neighborhood revitalization and, 233–6 NIMBY (Not In My Backyard) attitude, 371 opportunity costs of, 388 public subsidies, 231–4, 237, 327, 417 public willingness to pay for, 367–72, 368–70, 371, 415–25, 418–23 sports participation, 33–41, 34, 405–12 behaviour change theories and, 407–9, 408 demonstration and trickle-down effects of sports events, 405–6, 409–12, 428 econometric modelling, 36–7, 38–9, 47–8, 75 environmental influences, 40–1, 75, 77 health and, 45–51 household expenditure, 73, 75–7 household structure and, 40, 76–7 human capital and, 37, 64–6, 74, 76 income and time effects, 40, 74, 76 income–leisure trade-off model, 34–5, 35 labour market outcomes and, 64, 66–9 measurement of, 35–6 mental health and, 49–51 social capital formation and, 54–61, 57–8

socio-demographic characteristics and, 37–40, 76, 407 time allocation model, 34–5, 35, 37, 40, 74, 407 transtheoretical model (TTM), 407–9, 408 volunteering and, 95–6 sports systems, 8–16 features of open and closed, 8–12, 10–11 labour market for talent, 10–11, 14–15 modeling, 12–14 see also cartel arrangements squash, 452 Sri Lanka, 433 SSAs see Sport Satellite Accounts (SSAs) stadium attendance see sports attendance stadiums and arenas Australian Rules Football, 326–7 congestion and, 236, 237 criminal behaviour and, 236–7 ground sharing, 326–7, 338 impacts on employment and business, 236, 359 impacts on property values and rents, 234–6, 359, 372 impacts on urban development, 359–60 Major League Soccer, 338 neighborhood revitalization and, 233–6 NIMBY (Not In My Backyard) attitude, 371 opportunity costs of, 388 public subsidies, 231–4, 237, 327, 417 public willingness to pay for, 367–72, 368–70, 371, 415–25, 418–23 Stahl, S., 85 Stahler, Kevin, 379, 382, 383 stakeholder theories of non-profit organizations, 84 Stango, V., 478–9 Stanley, Russ, 525 Staudohar, P., 139, 308 Steckenleiter, Carina, 64–9 Steinhilber, A., 226, 227 Stekler, H.O., 454 Ste-Marie, D.M., 510 Sterken, E., 358 Stevenson, B., 67 Stewart, Allison, 390, 391 Stewart, B., 323, 324 Stewart, K., 317 Stewart, M.F., 323, 324, 325, 326, 358 Stigler, George, 117 Stiglitz, J.E., 431 stock market floatation, 11, 245–6 Stockl, M., 478 Stone, Richard, 20 Stratton, M., 40 Strigas, A.D., 106 stroke, 46 StubHub, 190, 192, 195, 196, 198, 201–2, 525, 529 subjective well-being (SWB) sports events and, 360, 431–2 sports participation and, 49–51 volunteering and, 96–7

INDEX

Suen, W., 378, 380 suicide, 435 Sukhatme, V.A., 455, 476, 477, 519, 520, 541 Sumell, A.J., 454 Sunde, U., 455 superstar effects on attendance, 164 in football (soccer), 244 on supply of effort in golf, 477, 512–13 Supreme Court, US, 115, 118, 119, 146 surfing, judging biases, 510 Süssmuth, B., 215, 419 Sutherland, R.J., 244 Sutter, M., 228 Sutton, R.I., 554 Swartz, T.B., 259 Sweden, 24, 67, 73, 85, 87 Sweeney, K., 193, 197–8 Sweeting, A., 193, 197–8, 526–7, 532 Swierzy, Philipp, 82–8, 423 swimming doping, 544 peer effects, 513 Switzerland, 18, 57, 59, 65–6, 86, 87 Sydney Olympics, Australia, 350, 358, 379, 380, 388, 390, 393 Szymanski, Stefan, 8, 13, 139, 140, 141, 154, 157, 167, 203, 204–5, 251, 274, 358, 360, 546, 549 table tennis, 452, 458 tactical performance data, 556, 557, 560 Tainsky, S., 458, 536–7 Taking Part Survey, UK, 35, 36, 50–1, 410 Taks, Marijke, 76, 395–403 Tanaka, R., 512 Tang, H.L., 40 Tauer, L.W., 474 tax increment financing (TIF), 233 Taylor, N.J., 422 Taylor, P., 33–4, 74, 77, 396, 407 Tcha, M., 379 Team Marketing Report, 185, 186 team performance, 210–16 coaching turnover and, 312, 312–13 production frontier approach, 215–16 production function approach, 211–15 salary dispersion and, 311, 335–6, 518 see also performance analytics team relocations, 10, 12 stadium subsidies and, 231–4, 417 Teigland, J., 350, 359 television broadcasting rights Australian Rules Football, 326 betting and, 536 cricket, 264 English Premier League, 171–4, 173 Olympic Games, 343 professional team sports, 139, 148–51, 163

583

road cycling, 465, 467, 467 rugby union, 274 US college sports, 114–16, 117, 118 television viewing demand Australian Rules Football, 326 betting and, 536–7 cricket, 264–5 English Premier League, 171–9, 173, 175, 176, 177 football (soccer), 171–9, 173, 175, 176, 177, 246, 336, 338 Major League Soccer, 336, 338 NASCAR racing, 493, 496, 497 outcome uncertainty and, 158–9 road cycling, 468, 546 tennis, 453, 456–7 Telser, L., 311 tennis, 452–8 audience demand, 453–4 behavioral economics, 455–6 betting, 454–5, 536, 538, 539 coaches, 457 demonstration effects, 412 economic impacts of minor event, 400 gender and, 453, 456, 519–20 home advantage, 226, 227, 456 match fixing, 457, 539, 541 player rankings, 454 prize structure, 455, 494, 517 television viewing demand, 453, 456–7 Tennis Industry Association (TIA), 453 Terry, M.E., 454 Thaler, Richard, 168, 292–4 Theil, H., 154 Theil measure, 154, 174, 176, 177, 177 theory of participation, 75 Thibaut, E., 74, 75, 76 third-party government theory of non-profit ­organizations, 84 Thomas, D., 213–14, 215, 216, 251, 259, 557 Thomas, J., 259 Thomas, J.M., 474 ticket pricing, 181–8 dynamic pricing, 187, 192, 193, 196, 197–8, 328, 457, 525–32 multi-tiered pricing, 184 price discrimination, 182–5, 187 price dispersion, 184, 196 price elasticity, 185–6, 187, 527–8 season tickets, 183, 186, 187, 193, 194, 196, 198, 526, 529, 532 traditional and modern ticketing models, 191–2, 191 two-part tariffs, 183 underpricing, 193 variable pricing, 184–5, 193, 328, 525–6 see also secondary ticket markets Ticketmaster, 190, 192, 194, 195, 201–2 time allocation model, 34–5, 35, 37, 40, 74, 94, 407

584

THE SAGE HANDBOOK OF SPORTS ECONOMICS

time availability sports participation and, 40, 74, 76 volunteering and, 94 Title IX regulation, US, 67, 128 Tobit model, 36–7, 38–9, 75, 76 Tokyo Olympics, Japan, 358, 359, 411 Tollison, R.D., 497 Tondani, D., 466 Top 14 (rugby union), 270, 270, 271, 272, 273–4, 273, 275, 275 Topic, M.D., 433 Torgler, B., 464 Tour de France, 462, 463, 464, 465, 467, 467, 468, 469, 546 Tour of Flanders, 462, 469 track and field economic impacts of minor event, 401–2, 401 gender and, 520–1 see also distance running traffic, congestion, and pollution, 236, 237, 434 transaction costs, 149, 150, 197, 454, 526 transfer market, football, 139, 203–9, 244–5 alternatives to, 207–9 average player salary and, 206–7, 206, 207 Bosman ruling, 13, 14, 139, 203, 204, 208, 244–5, 557 competitive balance and, 204–6, 205 player valuation, 557 salary distribution and, 207 training compensation as alternative to, 208–9 youth training and, 208–9, 245 transtheoretical model (TTM), 407–9, 408 Treber, J., 198 triathlon, 482–90, 485, 486 demand, 489, 490 doping, 545 prize structure, 483, 483, 486, 488 trust, 55–6, 58, 60 trust-related theory of non-profit organizations, 84 Tu, C.C., 235 Tuck, G.N., 326 Turner, I., 323 Tversky, Amos, 168 Twomey, J., 335 two-stage least squares (2SLS) model, 47–8 UCI see International Cycling Union (UCI) UEFA (Union des Associations Européennes de ­Football), 139, 540 Champions League, 140, 141, 150, 155, 540 EURO Football Cup, 215, 351, 396, 397, 397, 400 Europa League, 140 Financial Fair Play (FFP) regulations, 139, 140, 245 Ulseth, A.-L.B., 57, 59 umpires see officials and bias uncertainty of outcome hypothesis (UOH), 154, 156–9, 163, 165, 166, 167, 210, 258, 506 see also outcome uncertainty Understanding Society Survey, UK, 36, 411 Unger, L.S., 106

United Kingdom demonstration and trickle-down effects of sports events, 406, 409, 410–11, 428 domestic violence, 434, 506 economic value of sport, 18–19, 20–1, 23, 24, 25–6, 26, 26, 27 features of sports leagues, 136, 137, 138, 139 health and physical activity, 49, 50 household sports consumption, 73, 73 London Marathon, 442, 442, 443 national pride and sport, 432 negative impacts of sports events, 434, 435 non-profit sports clubs, 85, 86, 88 Olympic Games, 360, 373, 388, 391, 405–6, 406, 409, 410, 428, 431–2, 434 open sports system, 9 origins of sports leagues, 9, 135, 136 performance funding, 406, 406 social capital formation, 57, 58, 60 sport employment, 26 Sporting Future strategy, 18–19, 406, 428 sports betting, 534, 535, 540 sports participation, 35, 41 subjective well-being and, 431–2 ticket pricing, 185 volunteering, 27, 96 see also English Premier League; rugby union United Nations Office on Sport for Development and Peace, 54 United States closed sports system, 8, 9–11, 12, 14, 15 economic value of sport, 19 employment and sports participation, 67 health and physical activity, 45, 49, 50 human capital accumulation, 65 marathons, 442, 442, 443, 449 negative impacts of sports events, 434–5 Olympic Games, 147, 343, 344, 348, 349–50, 350, 358, 367, 373, 380, 388, 391, 392, 434 physical activity guidelines, 46 social capital formation, 57, 60 sports betting, 534 sports participation, 35–6, 41 ticket pricing, 183, 184–5, 186 ticket resale laws and restrictions, 194 World Cup, 331, 358, 387, 393 see also college sports, United States; Major League Baseball (MLB); Major League Soccer (MLS); National Basketball Association (NBA), US; National Football League (NFL), US; National Hockey League (NHL), US United States Soccer Federation (USSF), 331 University of Oklahoma, 115, 118 University of Pennsylvania, 114–15 UOH see uncertainty of outcome hypothesis (UOH) US Anti-Doping Agency (USADA), 546 Uslaner, E., 56 utility maximization, 103, 125–8, 165, 406–7, 416, 425

INDEX

Vamplew, W., 534 van Heerden, J.H., 357 Van Reeth, Daam, 462–9 van Ryzin, G., 526 van Schie, S., 107 Vandeweghe, H., 468 Varca, P.E., 508 variable ticket pricing, 184–5, 193, 328, 525–6 Veal, A.J., 410 Vekeman, A., 421, 424, 469 Venter, Brendan, 559–60 Vermeulen, F., 464 vertical relationships, 56 Vertosick, E.A., 447 Vickers, A.J., 447 Victorian Football League (VFL), 322–4, 327 Vierhaus, C., 359, 372 Villar, J.G., 181, 182, 185, 186 Vilnius definition of sport, 21–3, 22 Vincent, C., 311–12 violence, domestic, 434, 435, 506 violent play in Australian Rules Football, 325 in ice hockey, 315–17, 316 volunteering, 102–8 economic importance of, 27, 97–8 mental health and, 96–7 multi-level model of, 105–7, 105, 108 in non-profit sports clubs, 84, 86, 88, 92–9, 93, 102 role identity and, 95–6 social capital formation, 56, 58, 59, 96, 106, 107 sports participation and, 95–6 subjective well-being and, 96–7 theoretical framework, 103–4 Vrooman, J., 128, 129–30 Walker, J.K., 455, 456 Walker, M., 455 Walo, M., 402 Walras equilibrium model, 12–13 Walsh, W., 311, 315 Walton, H., 419 Wanner, R.A., 372 Warburton, D., 45 WARP (wins above replacement), baseball, 299–301, 300, 300, 301, 302, 302, 303, 304, 304, 305, 305, 306 Watanabe, Nicholas M., 184, 186, 196, 331–8 Weatherston, C.R., 275 Webb, Karrie, 477 Weed, M., 409, 410 Weimar, Daniel, 243–53 Weinbach, A.P., 528, 529–31 Weingärtner, C., 537 Weisbrod, B.A., 83–4 Welsh Premier League (WPL), 270, 271 Welsh Rugby Union (WRU), 271, 272, 274, 275, 276 Wendling, W., 165

585

Werding, M., 105 Whitehead, J.C., 418, 423 Whitten, A.R., 326 Wicker, Pamela, 41, 50, 51, 73, 75, 85, 86, 87, 92–9, 105, 186, 412, 415–25, 420, 421, 422, 423, 428–35, 464, 488 Wies, S., 488 Wilamowski, M., 449 Wilde, J., 48 Wilders, M.G., 244 Wildman, J., 106 Wilken, Claudia, 119 Williams, P., 269 Williamson, B., 274 Williamson, O.E., 126 willingness to pay (WTP) dynamic ticket pricing and, 528, 529–30 for Olympic Games, 360, 367–72, 368–70, 371 public investments in sport, 360, 367–72, 368–70, 371, 415–25, 418–23 for road cycling events, 469 Willoughby, T., 40 Wilson, D.P., 317, 494 Wilson, J., 98 Wilson, J.K., 327 Wilson, R., 274, 396 win maximization, 273 Australian Rules Football, 324, 326 transfer market and, 204–6, 205, 207 Winchester, N., 272, 325 Winfree, J., 309 winter sports gender and, 521–2 judging biases, 510 Witt, R., 557 WNBA see Women’s National Basketball Association (WNBA) Women’s National Basketball Association (WNBA), 281, 282, 284, 522 Women’s Tennis Association (WTA), 453 Woo, B., 75 Wood, W.C., 496–7 Wooders, J., 455 Woods, Tiger, 477, 479, 512, 513, 518 World Anti-Doping Agency (WADA), 545, 546 World Cup, 246, 385–93 cost overruns, 390–1 criminal behaviour and, 434 economic impacts of, 351, 356–8, 360, 361, 362, 387–90, 391–3 feel-good factor, 432 host city selection process, 331, 386 hosting costs, 386–7, 389, 390–1 intangible effects, 432, 433, 434 national pride and, 432 World Health Organization (WHO), 46, 51, 431 World Hockey Association, 311 World Scientific Congress of Golf, 479

586

THE SAGE HANDBOOK OF SPORTS ECONOMICS

World Squash Federation (WSF), 452 World Triathlon Cooperation (WTC), 484, 487 Wozniak, D., 456 Wray, Nigel, 274 Xu, J.J., 193 Yam, D., 291 Yamamura, E., 513 Yamane, S., 513 Yates, A., 317 Yates, A.J., 455 youth training, football, 208–9, 245

Zarnowitz, V., 220 Zax, J.S., 445, 517 Zech, C.E., 212 Zero-Inflated Negative Binomial (ZINB) model, 36, 38 Zero-Inflated Ordered Probit (ZIOP) model, 36, 38 Zhou, L., 155, 157, 168, 236 Zhu, J.D., 196 Zimbalist, A., 231, 361 Zitzewitz, E., 510 Zuercher, T.J., 149 Zwiebel, J., 511