Handbook of Behavioural Economics and Smart Decision-Making: Rational Decision-Making Within the Bounds of Reason 1782549579, 9781782549574

Citation preview

HANDBOOK OF BEHAVIOURAL ECONOMICS AND SMART DECISION-MAKING

M4225-ALTMAN_9781782549574_t.indd i

03/05/2017 08:20

M4225-ALTMAN_9781782549574_t.indd ii

03/05/2017 08:20

Handbook of Behavioural Economics and Smart Decision-Making Rational Decision-Making within the Bounds of Reason

Edited by

Morris Altman Professor of Behavioural and Institutional Economics and Dean and Head, Newcastle Business School, University of Newcastle, Australia

Cheltenham, UK • Northampton, MA, USA

M4225-ALTMAN_9781782549574_t.indd iii

03/05/2017 08:20

© Morris Altman 2017 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2016957242 This book is available electronically in the Economics subject collection DOI 10.4337/9781782549598

ISBN 978 1 78254 957 4 (cased) ISBN 978 1 78254 959 8 (eBook)

02

Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

M4225-ALTMAN_9781782549574_t.indd iv

03/05/2017 08:20

Contents

List of contributors Foreword by Vernon L. Smith Acknowledgements 1

Introduction to smart decision-making Morris Altman

PART I

2

ix xix xxi 1

SMART DECISION-MAKERS, DIFFERENT TYPES OF RATIONALITY AND OUTCOMES

Rational inefficiency: smart thinking, bounded rationality and the scientific basis for economic failure and success Morris Altman

11

3

Rational mistakes that make us smart Nathan Berg

43

4

Rational choice as if the choosers were human Peter J. Boettke and Rosolino A. Candela

68

5

Smart predictions from wrong data: the case of ecological correlations Florian Kutzner and Tobias Vogel

86

6

Heuristics: fast, frugal, and smart Shabnam Mousavi, Björn Meder, Hansjörg Neth and Reza Kheirandish

101

7

The beauty of simplicity? (Simple) heuristics and the opportunities yet to be realized Andreas Ortmann and Leonidas Spiliopoulos

119

Smart persons and human development: the missing ingredient in behavioral economics John F. Tomer

137

8

PART II 9

10

ASPECTS OF SMART DECISION-MAKING

Behavioral strategy at the frontline: insights and inspirations from the US Marine Corps Mie Augier Feminist economics for smart behavioral economics Siobhan Austen

157 173

v

M4225-ALTMAN_9781782549574_t.indd v

03/05/2017 08:20

vi

Handbook of behavioural economics and smart decision-making

11

How regret moves individual and collective choices towards rationality Sacha Bourgeois-Gironde

188

12

Is it rational to be in love? Paul Frijters and Gigi Foster

205

13

Behavioral economic anthropology Giuseppe Danese and Luigi Mittone

233

PART III 14

DEVELOPMENT AND GOVERNANCE

Do changes in farmers’ seed traits align with climate change? A case study of maize in Chiapas, Mexico C. Leigh Anderson, Andrew Cronholm and Pierre Biscaye

251

15

Rationality, globalization, and X-efficiency among financial institutions Roger Frantz

275

16

The evolution of governance structures in a polycentric system Edward McPhail and Vlad Tarko

290

PART IV

TAX BEHAVIOUR

17

Taxation and nudging Simon James

317

18

Income tax compliance Erich Kirchler, Barbara Hartl and Katharina Gangl

331

PART V

SMART MACROECONOMICS AND FINANCE

19

Financial decisions in the household Bernadette Kamleitner, Till Mengay and Erich Kirchler

349

20

Employing priming to shed light on financial decision-making processes Doron Kliger

366

21

Experimental asset markets: behavior and bubbles Owen Powell and Natalia Shestakova

375

22

To consume or to save: are we maximizing or what? Tobias F. Rötheli

392

PART VI

DIMENSIONS OF HEALTH

23

Time orientation effects on health behavior Jannette van Beek, Michel J.J. Handgraaf and Gerrit Antonides

413

24

Behavioral aspects of obesity Odelia Rosin

429

M4225-ALTMAN_9781782549574_t.indd vi

03/05/2017 08:20

Contents 25

26

Time inconsistent preferences in intertemporal choices for physical activity and weight loss: evidence from Canadian health surveys Nazmi Sari Suicide among smart people Bijou Yang and David Lester

PART VII

27

28

29

30

449 464

SOCIOLOGICAL DIMENSIONS OF SMART DECISIONMAKING

Seeing and knowing others: the impact of social ties on economic interactions Astrid Hopfensitz Weakness of will and stiffness of will: how far are shirking, slackening, favoritism, spoiling of children, and pornography from obsessivecompulsive behavior? Elias L. Khalil The role of identity, personal and social capital in community crime prevention Ambrose Leung and Brandon Harrison Norms, culture, and cognition Shinji Teraji

PART VIII

vii

479

492

515 526

MORALS AND ETHICS

31

Rational choice in public and private spheres Herbert Gintis

543

32

Ethics and simple games Mark Pingle

557

Index

M4225-ALTMAN_9781782549574_t.indd vii

573

03/05/2017 08:20

M4225-ALTMAN_9781782549574_t.indd viii

03/05/2017 08:20

Contributors

Morris Altman is Dean and Head of the Newcastle Business School and is Professor of Behavioural and Institutional Economics at the University of Newcastle, Australia. He is also Professor Emeritus at the University of Saskatchewan, Canada. Morris was the Head of the School of Economics and Finance and Professor at Victoria University of Wellington, New Zealand. He earned his PhD in economics from McGill University, Montreal, Canada in 1984. A former visiting scholar at Cambridge (Elected Visiting Fellow), Canterbury (Erkine Professor), Cornell, Duke, Hebrew, Stirling and Stanford Universities, he served as Editor of the Journal of Socio-Economics for ten years and is currently the co-founder and Associate Editor of the Review of Behavioral Economics. He is also past President of the Society for the Advancement of Behavioral Economics and of the Association for Social Economics. Morris has published over one hundred refereed papers and given over 150 international academic presentations on behavioural economics, x-inefficiency theory, institutional change, economics of cooperatives, economic history, methodology and empirical macroeconomics and has published eight books including: Handbook of Contemporary Behavioral Economics, Behavioral Economics for Dummies, Economic Growth and the High Wage Economy and Real-World Decision Making: An Encyclopedia of Behavioral Economics. Morris is on the International Co-operative Alliance (ICA) international committee on research as well as that for the Asian-Pacific region. C. Leigh Anderson is the Marc Lindenberg Professor for Humanitarian Action, International Development and Global Citizenship at the University of Washington’s Evans School of Public Policy and Governance, USA. Anderson’s research focuses on how individual and household decision-making is affected by economic and attitudinal factors including poverty, rural isolation, agricultural livelihoods, and preferences over risk, time and social standing. Of interest is how policy and programmatic interventions can be best designed and delivered to improve the lives of the poor and food insecure. Gerrit Antonides is an Emeritus Professor of Economics of Consumers and Households at Wageningen University, the Netherlands. He has published in the areas of behavioural economics, economic psychology and consumer behaviour. He has been an editor of the Journal of Economic Psychology and has authored and co-authored several textbooks on consumer behaviour and economic psychology. The behavioural aspects of consumer decision-making concerning issues of finance, household, environment and health are an important part of his current research activities. Mie Augier is Associate Professor at the Naval Postgraduate School, USA. Her scholarly and academic research interests include strategy, organizations, innovation, interdisciplinary social science, how organizations cultivate innovation capability (including the role of strategic organizational design), the influence of culture and globalization on strategic decision-making, and the past and future of management education and business schools. Her research has been published in more than 50 articles and book chapters in outlets ix

M4225-ALTMAN_9781782549574_t.indd ix

03/05/2017 08:20

x

Handbook of behavioural economics and smart decision-making

such as Organization Science, Industrial and Corporate Change, Journal of Management Inquiry, Management International Review, Organization Studies; Research Policy and California Management Review, among others. With collaborators she has published on topics such as the history of business schools (including her 2011 book with James March, The Roots, Rituals, and Rhetorics of Change, Stanford University Press) and the organizational mechanisms leading to the rise (and decline) of novelty and innovation in organizations (‘The flaring of intellectual outliers’, 2015, Organization Science). Active research interests include: (1) organizational and strategic analysis of the US Marine Corps as an organization, how they have evolved and organized for innovation, and their strategic decision-making; (2) the evolution of the teaching of ethics and values within the history of business schools and management education; and (3) behavioural strategy as a field. Siobhan Austen is Professor of Economics and Director of Women in Social and Economic Research (WiSER) at Curtin University Perth, Western Australia. She works on feminist and institutional economics, with a particular focus on the circumstances and experiences of women in labour markets. Nathan Berg is Associate Professor of Economics at University of Otago in Dunedin, New Zealand. Berg’s work appears in Journal of Economic Behavior and Organization, Psychological Review, Social Choice and Welfare and Review of Behavioral Economics. Berg was a Fulbright Scholar in 2003 and Visiting Research Scientist at the Max Planck Institute-Berlin in the 2000s. His research has been cited in the Financial Times, Business Week, Canada’s National Post, The Village Voice, The Advocate, Science News, Slate and the Atlantic Monthly. Pierre Biscaye is the Research Coordinator for the Evans School Policy Analysis and Research Group (EPAR) at the University of Washington, USA. He manages and supports research looking at issues across agricultural development, poverty reduction, financial inclusion, global health, and development policy. Pierre received a Master of Public Administration (MPA) from the University of Washington’s Evans School of Public Policy and Governance. For his capstone, he developed a monitoring and evaluation system and implementation plan for a small non-profit organization supporting education projects in Sierra Leone. Peter J. Boettke is a University Professor of Economics and Philosophy at George Mason University, a BB&T Professor for the Study of Capitalism, and the Director of the F.A. Hayek Program for Advanced Study in Philosophy, Politics, and Economics at the Mercatus Center at George Mason University, USA. He is Co-Editor-in-Chief of The Review of Austrian Economics and President of the Southern Economic Association. Sacha Bourgeois-Gironde is Professor of Economics at Université Paris II and research faculty member of Institut Jean-Nicod at Ecole Normale Supérieure, France. His work lies at the interface between decision-theory and cognitive sciences. The first aim is to understand how recent developments in formal decision-theory can supply new testable psychological insights on our use (or non-use) of probabilities, indeterminacies of our beliefs and values, and long-term rational purposes in life. The second aim, using interdisciplinary approaches (computer science, neuroscience and experimental psychology), is to

M4225-ALTMAN_9781782549574_t.indd x

03/05/2017 08:20

Contributors

xi

probe how bounded cognitive systems can adapt to complex decisional environments and how the interaction between the two brings about the emergence of particular institutions. Rosolino A. Candela is a PhD candidate in Economics at George Mason University and a Graduate Research Fellow in the F.A. Hayek Program for Advanced Study in Philosophy, Politics, and Economics at the Mercatus Center at George Mason University, USA. He holds a BA in History from St John’s University and an MA in Economics and International Political Economy from Fordham University. Previously, he was also a visiting PhD student in the Department of Political and Social Sciences at the European University Institute and a Charles G. Koch PhD Fellow at Suffolk University, where he was also a Koch Summer Fellow at the Beacon Hill Institute. Andrew Cronholm is an analyst with the King County Office of Performance, Strategy and Budget in Washington State, USA, where he provides policy, finance, and budgeting expertise. Andrew previously held analytical roles supporting the US Environmental Protection Agency and the City of Seattle’s Department of Transportation. Originally hailing from Massachusetts, Andrew received a Bachelor of Arts in Political Science from Drew University and obtained his Master of Public Administration (MPA) and Certificate in Environmental Management from the University of Washington’s Evans School of Public Policy and Governance. He currently resides in Seattle, Washington. Giuseppe Danese is a Fellow at CEGE, the research centre of Católica Porto Business School in Porto, Portugal. He holds a PhD in Economics from Simon Fraser University. His research interests are social norms, organizations, property rights, and the psychophysiological roots of decision-making. Gigi Foster is an Associate Professor with the School of Economics at the University of New South Wales, Australia. She works in many literatures, including education, social influence, corruption, laboratory experiments and time use. With support from the Australian Research Council and other bodies, she published a holistic behavioural economics treatise with Cambridge University Press (An Economic Theory of Greed, Love, Groups, and Networks, jointly with Paul Frijters) in 2013 and has authored over 25 academic papers published in a range of outlets such as the Journal of Public Economics, Quantitative Economics, Human Relations and Journal of Economic Psychology. Roger Frantz is Professor of Economics at San Diego State University, USA and Founding Editor of the Journal of Behavioral Economics for Policy. He has edited the Handbook of Behavioral Economics (London: Routledge, 2016). He is Editor of Renaissance in Behavioral Economics. He is co-editor, with Leslie Marsh, of Minds, Models, and Milieux: Commemorating the Centennial of the Birth of Herbert Simon and, with Robert Leeson, of Frederick Hayek and Behavioral Economics. He has also authored Two Minds: Intuition and Analysis in the History of Economic Thought and X-Efficiency: Theory, Evidence, and Applications. His work has been published in many journals, including the Journal of Socio-Economics, Journal of Economic Psychology, American Economic Review, Papers & Proceedings, Economics and Philosophy, Public Choice, Journal of Post Keynesian Economics, Journal of Behavioral Economics and the Southern Economic Journal. Paul Frijters is a Professorial Research Fellow at the Wellbeing Program within the Centre for Economic Performance at the London School of Economics, United Kingdom and Project

M4225-ALTMAN_9781782549574_t.indd xi

03/05/2017 08:20

xii

Handbook of behavioural economics and smart decision-making

Director of the LSE’s World Wellbeing Panel. Paul holds a PhD in welfare and well-being in Russia from the University of Amsterdam and has a wide range of research interests, having published over 70 papers and books in fields including happiness, labour markets, health economics, behavioural economics and econometrics. Before joining the LSE, he was the Research Director of the Rumici Project, an international project into the migration from the countryside to the cities in China and Indonesia, sponsored by ministries, the World Bank, the Ford Foundation and many others, tracking 20 000 individuals for many years. In 2009 Paul was awarded the Economic Society of Australia’s Young Economist Award (best economist under 40 in Australia). He regularly comments on economic issues in the national and international media, including the New York Times and the BBC. Katharina Gangl is Assistant Professor at the University of Göttingen, Germany, as the Chair of Economic and Social Psychology. She received her Diploma and PhD in Economic Psychology at the University of Vienna, Austria, and was a visiting scholar at the Queensland University of Technology, Australia. Her main research areas are ethical decision-making in organizations and tax behaviour. Herbert Gintis is External Professor at the Santa Fe Institute, Santa Fe, New Mexico, USA. His recent books include Game Theory in Action (with Stephen Schechter) (Princeton University Press 2016), A Cooperative Species (with Samuel Bowles) (Princeton University Press 2011), The Bounds of Reason (Princeton University Press 2009), Game Theory Evolving (Princeton University Press 2009), and Moral Sentiments and Material Interests (MIT Press 2005). His most recent book is Individuality and Entanglement: The Moral and Material Bases of Social Life (Princeton University 2016). His recent work on market dynamics includes: ‘The stability of general equilibrium with decentralized prices’ Journal of Mathematical Economics (with Antoine Mandel, 2016); ‘Stochastic stability in the Scarf economy’, Mathematical Social Sciences (with Antoine Mandel, 2014); and ‘The dynamics of general equilibrium’, Economic Journal (2007). His work on the unification of the behavioural sciences includes: ‘Zoon politikon: the evolutionary origins of human political systems’ (with Carel van Schaik and Christopher Boehm), Current Anthropology (2015); ‘Inclusive fitness and the sociobiology of the genome’, Biology & Philosophy (2014), ‘Homo socialis: an analytical core for sociological theory’ (with Dirk Helbing), Review of Behavioral Economics (2015); ‘The biology of cultural evolution’, Quarterly Review of Biology (2013), and ‘The evolutionary roots of property rights’, in Kim Sterelny et al. (eds), Cooperation and its Evolution (MIT Press 2013). Professor Gintis is a top reviewer of scientific books at Amazon.com and was recently cited as a gold star reviewer for Nature. Michel J.J. Handgraaf received his PhD in Social Psychology from Leiden University. Since 2011 he has been an Associate Professor at the Economics of Consumers and Households Group of Wageningen University, the Netherlands. Most of his research uses (field) experimental methods and surveys, can be described as ‘behavioural economics’ and mainly deals with differences between what rational economic theories would predict and the psychology behind deviations from such predictions. Besides research on fairness and ethics, Handgraaf’s current research focuses on decisions in the environmental domain. These decisions typically feature uncertainty, temporal trade-offs and social trade-offs.

M4225-ALTMAN_9781782549574_t.indd xii

03/05/2017 08:20

Contributors xiii Brandon Harrison obtained a Bachelor of Arts Criminal Justice (Honours) from Mount Royal University in Calgary, Alberta, Canada. Brandon is currently enrolled in the Faculty of Law at Thompson Rivers University in Kamloops, British Columbia, Canada. Brandon is interested in criminal law and energy law. Barbara Hartl holds a post-doctoral position at the Institute of Organization and Global Management Education at the Johannes Kepler University, Linz, Austria. Her research interests include cooperation in social dilemma, sustainable consumption and psychology of tax behaviour. She follows a multi-method approach, including laboratory experiments, survey studies, interviews and free associations. Astrid Hopfensitz is currently a Lecturer in Economics at the Toulouse School of Economics (TSE) in France. Her main research interest concerns the influence of emotions and psychological dimensions on economic decision-making and behaviour. In her work she employs economic experiments in combination with psychological methods of measuring emotions and character traits. Since 2012 she has also been affiliated with the Institute of Advanced Study in Toulouse (IAST). Simon James is an Associate Professor of Economics, Department of Organisation Studies, University of Exeter Business School, United Kingdom. He has held visiting positions at six universities overseas, published many research papers and his 16 books include a four-volume edited collection of papers Taxation: Critical Perspectives on the World Economy, 2002, A Dictionary of Taxation, 2nd edition 2012, The Economics of Taxation: Principles, Policy and Practice (with Christopher Nobes) 16th edition 2016 and, jointly edited with Adrian Sawyer and Tamer Budak, The Complexity of Tax Simplification: Experiences from around the World, 2016. Bernadette Kamleitner is Professor of Marketing at WU Vienna University of Economics and Business, Austria. She is head of the Institute for Marketing and Consumer Research at WU and President of the Austrian Forum Marketing. Her internationally published research is positioned at the cross-section of psychology, marketing and economics. Her particular research interests comprise the psychological underpinnings and consequences of experiences of ownership and financial decision-making. Elias L. Khalil is an Associate Professor of Economics at Monash University, Australia. He received his PhD in Economics from the New School for Social Research in 1990. His research focus is building a unified theory of human action, a theory that is based on rationality, virtue and an expanded notion of the self. His papers appeared in journals such as Economic Inquiry, Behavioral and Brain Sciences, Biology and Philosophy, Biological Theory, Theory and Decision, Journal of Economic Behavior and Organization,  Cambridge Journal of Economics, Journal of Evolutionary Economics, International Negotiation, Theoria, Philosophy, Economic Modelling and Economics and Philosophy. Reza Kheirandish is Associate Professor of Economics at the College of Business, Clayton State University, Morrow, Georgia, USA, and is affiliated with the Center for Adaptive Behavior and Cognition, Max Planck Institute for Human Development, Berlin, Germany. He received his PhD in economics from Virginia Tech and his BSc degree in electrical engineering from Sharif University of Technology. Reza has been a (summer)

M4225-ALTMAN_9781782549574_t.indd xiii

03/05/2017 08:20

xiv

Handbook of behavioural economics and smart decision-making

Visiting Researcher at the Max Planck Institute for Human Development in Berlin since 2014. He has served as the Director of the Center for Research in Economic Sustainability and Trends (CREST) at CSU (2010–15), the Program Co-Chair for the SABE/IAREP (2013) and the President, Vice-President, Treasurer/Secretary and Program Chair of SEINFORMS. He is the 2017 Program Chair for the SEDSI and has been a board member and webmaster of SABE since 2010. Erich Kirchler is Professor of Economic Psychology at the University of Vienna, Faculty of Psychology, and Guest Professor at WU Vienna University of Economics and Business, Austria. He is head of the Department of Applied Psychology: Work, Education, Economy at the Faculty of Psychology and past President of the International Association of Applied Psychology (IAAP), Division 9 (Economic Psychology) and the Austrian Psychological Society. His research is positioned at the cross-section of psychology and behavioural economics. His particular research interests comprise financial decisions in the household and psychology of tax behaviour. Doron Kliger is the Chair of the Economics Department at the University of Haifa, Israel, specializing in finance and behavioural economics. While in the US, he has been affiliated with the Wharton School, University of Pennsylvania. Kliger has published in a wide range of journals in finance, economics, insurance and probability, on topics including asset pricing, behavioural economics and finance, bond rating, decision-making, industrial organization and insurance pricing. He is a co-author of the book Event Studies for Financial Research, helping readers to obtain valuable hands-on experience with event study tools and to gain required technical skills for conducting their own studies. Florian Kutzner received his doctoral degree in psychology from the University of Heidelberg. After a research stay at the Warwick Business School he is currently affiliated with the Department of Cognitive Research in Social Psychology (CRISP) at the University of Heidelberg, Germany. His research focuses on the descriptive models of decision-making and learning in the context of social stereotypes and sustainable behaviour. David Lester has doctoral degrees from Cambridge University, United Kingdom, in social and political science and Brandeis University, USA, in psychology. He is Emeritus Professor of Psychology at Stockton University in Galloway, New Jersey, USA. He is a former President of the International Association for Suicide Prevention. He has published extensively on economic issues and suicide, including Suicide and the Economy (Nova Science, 1997) and ‘Calculating the economic cost of suicide’ (Death Studies, 2007, 31, 351–61). Ambrose Leung is Associate Professor at the Department of Economics, Justice, and Policy Studies, Mount Royal University in Calgary, Alberta, Canada. Ambrose received his PhD in Economics from Carleton University in Ottawa, Ontario, Canada. His main fields of research include socioeconomics, economics of crime, economic psychology and economics education. Ambrose has also acted as a consultant for the Department of Justice Canada. Edward McPhail is Professor of Economics at Dickinson College in Pennsylvania, USA. He has published papers in the American Journal of Economics and Sociology, Challenge, Eastern Economic Journal, European Journal of Political Economy, History of Economics

M4225-ALTMAN_9781782549574_t.indd xiv

03/05/2017 08:20

Contributors

xv

Review, History of Economic Ideas, Historical Journal, Journal of Economic Behavior and Organization, Review of Political Economy and others. Björn Meder is a research scientist at the Center for Adaptive Behavior and Cognition (ABC) at the Max Planck Institute for Human Development in Berlin, Germany. His research interests include judgement and decision-making, causal inference, information search and cognitive modelling. Björn holds a PhD in psychology from the University of Göttingen, Germany. Till Mengay was Research Assistant at the Institute for Marketing and Consumer Research, WU Vienna University of Economics and Business, Austria. Currently he is working at the Federal Ministry of Education, Department for Adult Education, Austria. His particular research interests are sustainable consumption and sharing within groups. Luigi Mittone (PhD, Bristol) is Full Professor of Economics at the University of Trento, Italy. At the University of Trento he is Director of the Doctoral School in Social Sciences, Director of the Cognitive and Experimental Economics Laboratory and coordinator of the International Master in Economics (MEC). He is also coordinator of the research project in ‘Experimental economics and nudging’ at the Bruno Kessler Research Centre. He has published extensively on fiscal evasion, consumer behaviour, mental modelling of uncertain events, intertemporal choices, cooperation, and fiscal system dynamics with heterogeneous agents. Shabnam Mousavi’s research revolves around making sense of the ways in which people make their decisions. She holds a PhD in economics and one in statistics from Virginia Tech, USA, serves on the Faculty of Business at the Johns Hopkins University, USA, and is a researcher at the Max Planck Institute for Human Development, Berlin, Germany. At the moment she is writing her first book, Fast-and-Frugal Decision Making. Hansjörg Neth is Lecturer in Social Psychology and Decision Sciences at the University of Konstanz, Germany and an associate member of the Max Planck Institute of Human Development, Berlin, Germany. His theoretical and experimental research focuses on the analysis of adaptive behaviour, interactive cognition and ecological rationality, as well as applied aspects of choice, and heuristic decision-making under uncertainty. He has served as acting Chair of Cognition, Emotion, and Communication at the University of Freiburg, taught cognitive and decision sciences at the University of Göttingen, and was Research Assistant Professor in Cognitive Science at the Rensselaer Polytechnic Institute. He holds a PhD in psychology from Cardiff University, UK. Andreas Ortmann is Professor of Experimental and Behavioural Economics in the School of Economics at the UNSW Australia Business School, Sydney, Australia. He was formerly Professor at CERGE-EI, Prague, Czech Republic, and Researcher at the Center for Adaptive Behavior and Cognition. Mark Pingle has been a member of the University of Nevada, Reno Department of Economics, USA, since 1990. He was appointed Conjoint Professor of Economics to the Department of Economics, University of Newcastle in 2016. He received his PhD in Economics in 1988 from University of Southern California. Professor Pingle has

M4225-ALTMAN_9781782549574_t.indd xv

03/05/2017 08:20

xvi

Handbook of behavioural economics and smart decision-making

published in the areas of behavioural economics, experimental economics and macroeconomics. He is Book Editor for the Journal of Behavioral and Experimental Economics, Associate Editor for the Review of Behavioral Economics and is a past President of the Society for the Advancement of Behavioral Economics. Owen Powell is an Assistant Professor in the Department of Economics at the University of Vienna, Austria. He holds a PhD in economics from the University of Tilburg. His research interests include experimental finance, growth and growth accounting, and computational economics. His work has been published in the Journal of Econometrics, the Review of Finance and the Journal of Behavioral and Experimental Finance. Odelia Rosin is a health economist. She holds a PhD in economics from Bar-Ilan University, Israel. Her doctoral dissertation dealt with obesity, its behavioral economic aspects and related public policy. Her research interests are health economics, behavioral economics, nutrition and public health. Odelia is a Lecturer in the College of Management – Academic Studies (COMAS) in Israel. She also serves as the academic head of one of COMAS’s campuses. Tobias F. Rötheli is Professor of Macroeconomics at the Department of Economics of the University of Erfurt in Germany. He holds a doctorate and a Venia Legendi in Economics from the University of Bern. His research focuses on behavioural models of expectations. Much of this work is built on the concept of pattern recognition and combines experimental methods and applied econometrics. A further area of research is the modelling of boundedly rational agents and their role in financial boom–bust dynamics. Finally, in historical studies Rötheli investigates the coevolution of quantitative methods in academic economics and in business practice. Nazmi Sari is Professor in the University of Saskatchewan, Department of Economics, Canada. In addition to his primary appointment at the university, he is a faculty associate with the Canadian Center for Health Economics, University of Toronto, and an adjunct scientist at the Health Quality of Council. His research interests are economics of physical activity and smoking, quality and efficiency in hospital markets, provider reimbursements and healthcare financing reforms. He has published articles in health economics, public health, and health policy journals. Natalia Shestakova is an Assistant Professor in the Department of Economics at the University of Vienna, Austria. She holds a PhD in Economics from the Center for Economic Research and Graduate Education – Economics Institute (CERGE-EI). Her research interests include behavioural economics, contract theory and experimental economics. Leonidas Spiliopoulos received a BA in Economics from Yale University in 1997, an MSc from the Athens University of Economics and Business in 2003 and a PhD in Economics from the University of Sydney in 2008. He is currently a Visiting Research Fellow at the Max Planck Institute for Human Development (Center for Adaptive Rationality), Berlin, Germany, where he also served as an Alexander von Humboldt Experienced Research Fellow. He previously held a Vice-Chancellor’s Postdoctoral Research Fellowship at the University of New South Wales, an Endeavour Cheung Kong Research Fellowship at the Hong Kong University of Science and Technology, and lectured at the University of

M4225-ALTMAN_9781782549574_t.indd xvi

03/05/2017 08:20

Contributors

xvii

Athens. His work focuses on how economics, game theory, cognitive psychology/neuroscience and artificial intelligence can inform models of decision-making and learning. Vlad Tarko is Assistant Professor of Economics at Dickinson College in Pennsylvania, USA. He is the author of Elinor Ostrom: An Intellectual Biography (Rowman & Littlefield International, 2017) and co-author with Paul Aligica of Capitalist Alternatives: Models, Taxonomies, Scenarios (Routledge, 2015). He has published papers in the American Political Science Review, Academy of Management Papers and Proceedings, Comparative Economic Studies, Kyklos, Constitutional Political Economy, Review of Austrian Economics and others. Shinji Teraji is Professor of Economics, Yamaguchi University, Japan. His research is mainly concerned with a synthesis of behavioural and institutional economics. He is the author of Evolving Norms: Cognitive Perspectives in Economics (2016). John F. Tomer is Emeritus Professor of Economics at Manhattan College, USA. He was born in 1942 and grew up in New Jersey. He has a PhD in Economics (1973) from Rutgers University, New Brunswick, New Jersey. He is a founder and past President of the Society for the Advancement of Behavioral Economics. His research areas are behavioural economics and human capital. He has written four books and 50 articles. His recent research integrates human capital with human development. Jannette van Beek completed a dissertation on time orientation in relation to both eating and exercising behaviour. The main aim of this dissertation is to provide insight into the relationship between time orientation and both eating and exercising behaviour in order to better understand individuals’ intertemporal decision-making in the health domain and ultimately stimulate healthy eating and exercising behaviour. Currently, Jannette works as a Lecturer at both the Economics of Consumers and Households Group and the Marketing and Consumer Behaviour Group of Wageningen University, the Netherlands. Tobias Vogel received his doctoral degree in psychology from the University of Heidelberg. After research stays at the Universities of Louvain-la-Neuve, Belgium, Basel, Switzerland, and San Diego, USA, he is currently affiliated with the Department of Consumer and Economic Psychology at the University of Mannheim, Germany. His research focuses on the psychology of evaluative judgements, with an emphasis on the cognitive processes underlying attitude acquisition and construction. He is author of the book Attitudes and Attitude Change (Vogel and Wänke, 2016). Bijou Yang has BA and MA degrees in economics from the National Taiwan University and MA and PhD degrees in economics from the University of Pennsylvania. Her dissertation was on econometric forecasting of the world economy, and she worked for Wharton Econometric Forecasting Associates (WEFA) before returning to academia. She joined Drexel University, USA, in 1987. Her research has focused on contingent employment, e-commerce and the behavioural economic approach to suicide and criminal behaviour. She served as Treasurer of the Society for the Advancement of Behavioral Economics (SABE) from 1986 to 2010 and as President from 2006 to 2008. She has published two books and some 190 scholarly articles and notes. Her research has appeared in Applied Economics and the Journal of Socio-Economics, most recently ‘Personality traits and economic activity’ in Applied Economics, 2016, 48, 653–57.

M4225-ALTMAN_9781782549574_t.indd xvii

03/05/2017 08:20

M4225-ALTMAN_9781782549574_t.indd xviii

03/05/2017 08:20

Foreword

In the study of decision-making by people in the world, the laboratory, in surveys, or in all of the above, many scholars have derided our decisions as irrational, uninformed, biased or vulnerable to illusions, if not delusions, that steer us off track. You won’t find that simplistic reduction in this book. You will find plenty of cases of error, sometimes random, sometimes systematic, and sometimes in the models that are alleged to specify rational behaviour. You will also find penetrating analyses of institutions and other social systems that have made us smart, or smart enough to muddle through in an uncertain world. For me this shift in methodology from the search for anomalies that prove that the standard model is wrong – a search that was assured of finding what was sought – to a closer examination of the circumstances that make for success or failure is particularly welcome. Both experimental economics and the anomalies literature grew out of a welcome new wave of empirical investigation that can only be understood against the intellectual backdrop of a hundred-odd years of equilibrium theory development. That development had been jump-started by the marginal utility revolution of the 1870s, devolving into powerful new theory by the mid-twentieth century. Theoretical insights into topics ranging from individual decision-making and two-person games to the determinants of prices in markets invited new experimental investigations by psychologists and economists in the 1950s and 1960s. Both verifying and falsifying evidence surfaced as part of these investigations. When you are looking to verify the predictions of a theory and get glaring contrary evidence proving your beliefs are wrong, it changes the way you think about the phenomenon. In retrospect, neoclassical economic theory provided insights so powerful and influential that it displaced rather than supplemented the classical economic perspective. Under the influence of neoclassical theory, my first supply and demand experiments were designed to show that complete information – a pillar of theory at the time – was necessary to observe efficient competitive outcomes. However, the experiments demonstrated the opposite. We were wrong. Inadvertently, I was rediscovering that process is what matters. ‘The propensity to truck barter and exchange’ is a process; empower people with a trading institution, and they will use it to discover rich forms of specialization that otherwise did not exist. Neoclassical equilibrium in outcomes displaced rather than supplemented socioeconomic interactive processes of change, prominent in the writings of David Hume and Adam Smith. I see this book as a return to that perspective, but driven by new and exciting ways of modelling and thinking about the great issues that have created the modern world. Vernon L. Smith, Professor of Economics and Law George L. Argyros Endowed Chair in Finance and Economics 2002 Nobel Laureate in Economics Chapman University, USA xix

M4225-ALTMAN_9781782549574_t.indd xix

03/05/2017 08:20

M4225-ALTMAN_9781782549574_t.indd xx

03/05/2017 08:20

Acknowledgements

This was a huge project and one that breaks from various conventional perspectives on economic theory and behavioural economics. I must thank Matthew Pitman, our publishing editor, for supporting this project and for his helpful advice. Of course, it goes without saying that all contributors devoted so much time and effort towards constructing some enriching and excellent chapters. Thanks for your contributions and support. The meticulous work from the whole Edward Elgar team was invaluable. I also express gratitude to life partner and wife, Louise Lamontagne, for her comments and suggestions. Many thanks as well to our daughter, Hannah Altman, now blossoming into a first-rate economist and health scientist.

xxi

M4225-ALTMAN_9781782549574_t.indd xxi

03/05/2017 08:20

To the late Professor Harold (H.R.C.) Wright, brilliant and modest teacher-scholar, my Master’s and PHD supervisor, teacher, mentor and friend at McGill University.

M4225-ALTMAN_9781782549574_t.indd xxii

03/05/2017 08:20

1

Introduction to smart decision-making Morris Altman

This is an original contribution of essays on behavioural economics, which builds upon the research of Herbert Simon and, more generally, the Carnegie-Mellon school of behavioural economics. This perspective can be referred to as the bounded rationality methodological approach to behavioural economics (Altman 1999, 2005, 2015, 2017). In this perspective, the prior assumption is that decision-makers are relatively rational, intelligent and smart (satisficing, boundedly rational and evolutionarily rational). As one of the intellectual leaders of the Carnegie-Mellon school, James March (1978, p. 589), stated, it is of primary importance to determine if we can explain human behaviour in terms of rationality, broadly defined, even if at first glance such behaviour does not appear rational and might even appear to be error-prone or ‘biased’. More generally, I refer to this methodological approach as smart decision-making, which encompasses bounded rationality, procedural rationality, fast and frugal heuristics, the brain as a scarce resource (following the insights of Friedrich Hayek) and the institutional, sociological and psychologicalneurological determinants of decision-making. This is counter-posed to the world view of conventional or neoclassical rationality as well as the heuristics and biases perspective on behavioural economics, pioneered by Kahneman and Tversky (Kahneman 2003, 2011), that dominates contemporary behavioural economics. Smart decision-making encompasses intelligent or smart decision-makers or agents, who develop or adopt decision-making processes and make decisions given their cognitive limitations, the decision-making mechanism of the brain, individuals (or economic agents) decision-making capabilities, decision-making experience, environmental factors, which include institutional and legal parameters, culture and norms, relative power in the decision-making process and related sociological factors. It is also recognized that cognitive limitations are affected by technology (computers and calculators, for example), the capabilities to effectively use new or improved technologies and the learning processes that affect how the brain is hardwired (neuroplasticity). Smart decision-makers or agents do the best they can, given the pertinent circumstances that affect the decision-making process and related outcomes. Herbert Simon refers to the act of doing the best we can as satisficing behaviour. Satisficing need not result in the best possible or optimal outcomes for the firm, household, society or individual; but it can, depending on circumstances. Deviations from optimality do not imply that decision-makers are not smart and, in this sense, irrational. Nor does establishing that decision-makers are smart imply that decision-making outcomes are optimal. Here rationality, broadly defined, relates to the choices people make and the decision-making processes adopted by individuals given their various constraints and opportunities as well as their decision-making environment. Optimality in production and consumption at an individual, firm, household or social level need not necessarily flow from smart decision-making. Smart decision-making, however, would often be a necessary but not a sufficient condition for optimality to be 1

M4225-ALTMAN_9781782549574_t.indd 1

03/05/2017 08:20

2

Handbook of behavioural economics and smart decision-making

obtained. What these sufficient conditions might be are critically important to research that stems from the smart agent or smart decision-making perspective. Inadequate decision-making environments, for example, would preclude smart agents from achieving optimal results from their own and from society’s perspective (where externalities exist). For example, you might wish to increase your savings for retirement, but you invest in high-risk high-return financial paper because of the false or misleading financial information provided to you, resulting in you losing much of your savings. Women might want to have one child, but they end up giving birth to four or five, because they are not empowered to realize their preferences. A firm’s productivity might not be maximized because decision-makers are maximizing a complex utility function that includes managerial slack and short-term returns. None of the above is a product of irrationality. They are a product of preferences, decision-making capabilities, experience and the overarching decision-making environment. Conventional theory’s point of focus is on very generalized concepts related to how humans should behave and are expected to behave to generate optimal outcomes. As long as the analytical prediction is correct, all is well. This is effectively the correlation-based analysis promoted by Friedman (1953). If you get the prediction correct, you can assume for reasons of simplicity that humans behave as if they are maximizing profits, minimizing costs and maximizing utility (which is often assumed to be identical wealth maximization, controlling for risk). The realism of the simplifying assumptions we make about decisionmakers, the decision-making processes and the decision-making environment are not of importance from this perspective. We can simply assume that individuals behave as if they are maximizing profits or utility, as long as the analytical prediction is the correct prediction. The assumption here is that individuals ideally should behave ‘neoclassically’, if they are rational, which they are assumed to be. Rationality is defined in terms of neoclassical rationality. Apart from this, what transpires in the decision-making process is not of substantive interest. We simply abide methodologically with neoclassical simplifying assumptions of how individuals behave within the firm and in the household. Moreover, it is further assumed that the decision-making environment allows for the realization of optimal outcomes, given neoclassical rationality, for the individual, the household and the firm. The analytical focus, therefore, is on correlation as opposed to true causation, where the latter relates to determining what particular behaviours and decision-making environments generate particular outcomes. Modelling true causation would address issues of spurious correlation, omitted variables and the possibility of alternative behaviours, yielding similar sustainable outcomes. What is key is the determination of what specific behaviours, decision-making processes and institutional and sociological variables yield specific outcomes. This deeper modelling agenda is part of the bounded rationality approach to behavioural economics. The bounded rationality tradition in behavioural economics plays particular attention to identifying the actual decision-making process that generates particular outcomes. It ventures into the black box of the firm, the household and the individual. Only by understanding how individuals actually behave, how they make decisions, can we determine if these decisions are smart and in this broad sense rational. Hence, rationality here is contextualized. Benchmarks for what is rational are, therefore, not constructed by some imagined ideal unrelated to the decision-making capabilities and environments of the individual, household or firm.

M4225-ALTMAN_9781782549574_t.indd 2

03/05/2017 08:20

Introduction to smart decision-making 3 For this reason, a core attribute of the approach taken in this book is, following from Simon, the overall importance of reasonable, reality-based, simplifying modelling assumptions for robust economic analysis. Related to this is the significance of situating our definition of rationality and smart decision-making in context. Simon writes (1986, p. S209): The judgment that certain behavior is ‘rational’ or ‘reasonable’ can be reached only by viewing the behavior in the context of a set of premises or ‘givens.’ These givens include the situation in which the behavior takes place, the goals it is aimed at realizing, and the computational means available for determining how the goals can be attained. In the course of this conference, many participants referred to the context of behavior as its ‘frame,’ a label that I will also use from time to time. Notice that the frame must be comprehensive enough to encompass goals, the definition of the situation, and computational resources.

The smart agent, smart decision-making approach to decision-making and behavioural economics not only stands in contrast to what we find in much of conventional economics, it also stands in contrast, as mentioned above, to a theme running through much of contemporary behavioural economics where much of the typical individual’s behaviour is deemed irrational and error-prone. This is the heuristics and biases approach pioneered by Kahneman and Tversky (Kahneman 2003, 2011). A common thread running through this approach and conventional economics is adopting neoclassical benchmarks for rationality and, flowing from this, benchmarks for optimal outcomes in the domain of consumption and production (although the latter is not a point of focus in the heuristics and biases approach). In the heuristics and biases approach, as in conventional economics, these various benchmarks are not empirically derived. Rather, they are taken for granted. As in the conventional approach, causal analysis is not the point of focus, and it appears that analytical prediction (correlation analysis) is of greatest significance. However, in the heuristics and biases approach psychological factors are introduced into the modelling framework to supplement or replace economic variables. Typically such new variables are said to generate deviations from neoclassical optimality and, therefore, errors in decision-making. This is often derived from assumed, but not proven, hardwired biases in the human decision-maker. However, in terms of the derivation and introduction of psychological variables, these are often not predicated upon an assessment of how individuals behave within the household and the firm. Rather, they are generalized descriptors of human behaviour introduced into the modelling framework to produce improved analytical predictions or predictions that are as robust as those generated in conventional models, but now contain more realistic behavioural assumptions. To reiterate, the realism of these new assumptions is typically not tested against how individuals actually behave in the real world of decision-making. A point of commonality between the bounded rationality approach, the broader smart agent approach and the heuristics and biases approach is recognizing that real-world decision-makers typically do not behave like the individuals in the traditional economic models. We should note that Gary Becker (1996), for example, makes a similar point with regard to neoclassical models ignoring sociological variables to their analytical and scientific peril. He argues that neoclassical predictions are often wrong because they systematically ignore how social context impacts the decisions of rational agents. Douglass North (1971) makes a similar point with respect to neoclassical economics systematically ignoring the importance of institutional variables to decision-making by rational agents.

M4225-ALTMAN_9781782549574_t.indd 3

03/05/2017 08:20

4

Handbook of behavioural economics and smart decision-making

Especially with respect to the heuristic and biases approach, a large scholarly industry has developed documenting the extent to which actual human behaviour deviates from predicted neoclassical behaviour. More generally, experimental economics, often done in classroom settings, has documented significant deviations from neoclassical norms. The fact that individuals tend not to behave neoclassically is no big surprise, even to many neoclassical economists. The latter simply assume that individuals behave as if they make decisions and choices based on neoclassical norms, not that they actually behave in this fantastical manner. Still this research remains important as it disabuses economists (theoretical and applied), model users and various types of practitioners, including policy-makers, from the notion that humans behave neoclassically. The big question is what does this actually means for analysis and policy? Experiments suggest that, on average, individuals engage in a wide array of behaviours that are contrary to what conventional economics assumes. For example: ● ● ● ● ● ● ● ● ● ● ●

Individuals weigh losses more than gains. Emotions and intuition drive much of decision-making. Individuals are willing to self-sacrifice to punish those who they deem are treating them unfairly. Individuals are willing to punish or hurt those they don’t like. Individuals are willing to self-sacrifice for those towards whom they feel sympathy. Ethical concerns play a role in economics decision-making. Wealth maximization, even when controlled for risk, finds many exceptions. Framing affects choices. Relative positioning often matters more than absolute levels of income or wealth. Sentiment or animal spirits often matter more to decision-making than ‘real’ economic variables. Individuals often follow the leader when making decisions (herding).

Are these ‘average’ human traits a sign of hardwired cognitive biases, yielding suboptimal choices, as the heuristics and biases approach intimates? Or, are these characteristics of smart agents given their capabilities, experience and decision-making environment, even when some of their decisions are wrong, at least in the first instance (a one-shot game)? This is where the smart agent or smart decision-making approach and bounded rationality approach part company with the heuristics and biases approach. From the smart agent approach, deviations from neoclassical norms typically imply that rational decisionmakers do not abide by these norms for good rational reasons that need to be identified and understood to better engage in robust causal analysis. From the heuristics and biases approach, deviations from the neoclassical norms imply systemic biases and errors in decision-making, typically a function of how the brain is hardwired. Humans do not and typically cannot behave the way they should behave to obtain optimal outcomes. Free will in decision-making can result in perverse socio-economic outcomes that can sometimes be corrected by experts nudging individuals to behave in the appropriate fashion as defined and articulated by the expert (referred to in the literature as the choice architect) (Thaler and Sunstein 2008). From the smart decision-making perspective, errors in decision-making can and do

M4225-ALTMAN_9781782549574_t.indd 4

03/05/2017 08:20

Introduction to smart decision-making 5 occur. There can be biases in decision-making, individuals can make decisions that are not in their own self-interest or they can make decisions in their self-interest but not in the interest of their group, organization or society, and preferences can be inconsistent across individuals and within an individual across historical time. All such non-traditional behaviours can be consistent with the hypothesis that economic agents are smart and, broadly speaking, rational. Moreover, these smart agents need not generate choices that are in any sense efficient. This is in stark contrast to the conventional approach wherein being ‘rational’ implies efficient outcomes. However, rationality need not imply efficiency or optimality in either consumption or production. Not conforming to neoclassical behavioural norms need not be symptomatic of irrationality, and free will in choice behaviour in and of itself need not result, therefore, in perverse socio-economic outcomes. Errors and biases and suboptimal socio-economic outcomes, for example, can be the product of inadequacies in decision-making capabilities, suboptimal decision-making environments and lack of experience. In this sense rationality does not mean perfection in actual behaviour or outcomes. Of critical importance is the determination of the conditions under which decisions and the decisionmaking processes can be improved upon; under what circumstances can rational or smart decision-making result in efficiency or optimality in either consumption or production? Identifying these circumstances is a critically important research agenda. Also, non-neoclassical behaviours can generate superior outcomes to those that flow from traditional neoclassical norms, such as narrowly maximizing behaviour. In other words, conforming to neoclassical behavioural norms can generate suboptimal outcomes and might therefore even signal irrationality in behaviour or at least serious biases and errors in decision-making. Gerd Gigerenzer (2007) and his colleagues have articulated this perspective in their fast and frugal heuristics narrative. Heuristics (decision-making short  cuts), often considered to be biased and error-prone in the heuristics and biases narrative, is argued to exemplify superior decision-making processes in the fast and frugal modelling of decision-making. From this perspective individuals have evolved decisionmaking processes that are partially derived from the fact that the brain is a scarce resource, has a particular processing capability and processes information within a particular decision-making environment. A prior assumption here is that individuals are broadly speaking rational. Hence, it is important to investigate whether, and the extent to which, non-neoclassical behavioural norms (such as fast and frugal heuristics) yield superior outcomes, and under what circumstances. At one extreme it could be argued that not only are individuals always rational, but their decision-making processes and decisions are always optimal as well. This perspective is derived from Hayek and his notion of ecological rationality (Hayek 1948; Smith 2003; Gigerenzer 2007). But it is critical to determine benchmarks for smart or broadly rational behaviour and, moreover, contextualized benchmarks for efficiency and optimality in decision-making outcomes. Smart decision-making is not a necessary and sufficient condition for efficiency and optimality. Kahneman (2011) has articulated a categorization of different types of decisionmaking, which he refers to as slow and fast thinking. He is basically looking at when particular thought processes yield better outcomes. Sometimes these might be fast thinking; very often these would be slow thinking. Some would argue that individuals do not know which type of thinking best serves their own self-interest and that of their organization

M4225-ALTMAN_9781782549574_t.indd 5

03/05/2017 08:20

6

Handbook of behavioural economics and smart decision-making

and, all too often, individuals make the wrong choices as to which thinking decisionmaking platform to use. This would be contrary to the fast and frugal approach that maintains that typically individuals make the right choices with regard to decision-making platforms. From the smart decision-making perspective, it is a testable hypothesis as to which thinking platform would be best. This hypothesis needs to be contextualized by the capabilities and experience of individuals and their decision-making environment. A critical point here is that the thinking platform the individual should adopt is not determined a priori by the expert or by theory. It is context dependent. The smart decision-making approach has differential implications for policy and approaches for structuring decision-making. The conventional wisdom is, in its extreme, very ‘hands off’ on policy, both in terms of government and even on suggestions of what can be done inside the firm and household to improve decision-making processes and decision-making outcomes. The prior assumption is that ‘free’ markets plus rational agents would generate optimal results. So, government could intervene to make markets ‘freer’ and perhaps to better secure property rights. If individuals are hardwired to be error-prone and biased (the heuristics and biases approach) then intervention must be much more proactive, nudging or more forcefully driving individuals to make what are deemed optimal or at least better decisions. With the smart decision-making or smart agent approach, it is assumed, at least as an analytical starting-point, that individuals are rational. Hence, we need to address issues of capabilities, decision-making environments, experience and externalities to determine what is required to facilitate best practice, but also context informed, decision-making processes. Barring externalities, it becomes critically important to construct decisionmaking capabilities and environments to facilitate and nurture informed decisions, based on the free choice of decision-makers. Therefore, it also becomes important to understand the circumstances under which individuals lose the capacity (or this capacity is severely reduced) to make informed choices, such as possibly severe addictions and mental illness, and perhaps even more importantly the power and even the legal rights to make informed choices. These methodological differences between the smart agent–smart decision-making approach to behavioural economics (related to the concept of bounded rationality), the heuristics and biases approach to behavioural economics and conventional economics are illustrated in Figure 1.1. The smart decision-making approach incorporates and is informed by bounded rationality, process rationality and institutional design. These are informed by a variety of variables, inclusive of human capital, mental models, preferences, information, power and learning. Smart decision-making can result in either optimal or suboptimal outcomes depending on the above economic, sociological and institutional variables. Both these outcomes can be ‘rational’ from the perspective of the individual, but they can generate socially inefficient outcomes. We can have what I refer to as rational inefficiency, but this can be corrected (more often than not) by changing some of the key variables mentioned above. However, benchmarks for what yields optimal outcomes is largely unrelated to neoclassical behavioural norms. Rather, it is reality based. In contrast, the heuristics and biases approach predicts that what is often hardwired behaviour yields deviations from conventional norms for optimal behaviour, that is, from neoclassical rationality. The latter is retained as the gold standard for achieving optimality for the individual, the household, the firm and society. Deviations from the neoclas-

M4225-ALTMAN_9781782549574_t.indd 6

03/05/2017 08:20

Introduction to smart decision-making 7 Suboptimality

Optimality

Persistent errors Neoclassical rationality Smart decision-making

Bounded rationality

Process rationality

Conventional methodology

Deviations from neoclassical rationality

Heuristics and biases

Institutional design

Hardwired Human capital

Mental models

Preferences Nudging

Information

Figure 1.1

Power

Learning multi-shot games

Not predicted Information

Decision-making models

sical rationality yield persistent errors in decision-making, hence suboptimal outcomes. This can be corrected by nudging (which can involve varying degrees of paternalism) and, sometimes, by correcting for failures in institutional design. The latter includes improvements to information. Also, the latter as well as institutional design are critically important to the smart decision-maker approach to behavioural economics. Neoclassical models predict neoclassical rationality and optimal outcomes. They do not predict persistent deviations from neoclassical rationality which have been well documented in the literature. This book covers a wide range of themes from micro to macro, sub-disciplines within economics, economic psychology, heuristics, fast and slow thinking, experimental economics, the capabilities approach, institutional and sociological dimensions, methodology, nudging, ethics and morals, and public policy. The book is divided into a number of parts: ‘Smart decision-makers, different types of rationality and outcomes’; ‘Aspects of smart decision-making’; ‘Development and governance’; ‘Tax behaviour’; ‘Smart macroeconomics and finance’; ‘Dimensions of health’; ‘Sociological dimensions of smart decision-making’; and ‘Morals and ethics’. The authors critically explore the modelling, methodological and policy implications of a smart decision-making or smart agent approach to behavioural economics. This alternative approach to behavioural economics, rooted in the tradition established through the research of Herbert Simon and his colleagues, holds much promise, incorporating learning from the bounded rationality approach, the heuristics and biases approach as well as important insights from other disciplines, such as psychological, institutional and sociological analyses and neuroscience.

M4225-ALTMAN_9781782549574_t.indd 7

03/05/2017 08:20

8

Handbook of behavioural economics and smart decision-making

REFERENCES Altman, M. (1999), ‘The methodology of economics and the survivor principle revisited and revised: some welfare and public policy implications of modeling the economic agent’, Review of Social Economics, 57 (4), 427–49. Altman, M. (2005), ‘Behavioral economics, power, rational inefficiencies, fuzzy sets, and public policy’, Journal of Economic Issues, 39 (3), 683–706. Altman, M. (2015), ‘Introduction’, in M. Altman (ed.), Real-World Decision Making: An Encyclopedia of Behavioral Economics, Santa Barbara, CA: Greenwood, ABC-CLIO. Altman, M. (2017), ‘A bounded rationality assessment of the new behavioral economics’, in R. Frantz, S.-H. Chen, K. Dopfer, F. Heukelom and S. Mousavi (eds), Routledge Handbook of Behavioral Economics, New York: Routledge, pp. 179–94. Becker, G.S. (1996), Accounting for Tastes, Cambridge, MA: Harvard University Press. Friedman, M. (1953), ‘The methodology of positive economics’, in M. Friedman (ed.), Essays in Positive Economics, Chicago, IL: University of Chicago Press, pp. 3–43. Gigerenzer, G. (2007), Gut Feelings: The Intelligence of the Unconscious, New York: Viking. Hayek, F.A. (1948), Individualism and the Economic Order, Chicago, IL: University of Chicago Press. Kahneman, D. (2003), ‘Maps of bounded rationality: psychology for behavioral economics’, American Economic Review, 93 (5), 1449–75. Kahneman, D. (2011), Thinking, Fast and Slow, New York: Farrar, Straus and Giroux. March, J.G. (1978), ‘Bounded rationality, ambiguity, and the engineering of choice’, Bell Journal of Economics, 9 (2), 587–608. North, D.C. (1971), ‘Institutional change and economic growth’, Journal of Economic History, 31 (1), 118–25. Simon, H.A. (1986), ‘Rationality in psychology and economics’, Journal of Business, 59 (4), S209–24. Smith, V.L. (2003), ‘Constructivist and ecological rationality in economics’, American Economic Review, 93 (3), 465–508. Thaler, R.H. and C. Sunstein (2008), Nudge: Improving Decisions about Health, Wealth, and Happiness, New Haven, CT and London: Yale University Press.

M4225-ALTMAN_9781782549574_t.indd 8

03/05/2017 08:20

PART I SMART DECISION-MAKERS, DIFFERENT TYPES OF RATIONALITY AND OUTCOMES

M4225-ALTMAN_9781782549574_t.indd 9

03/05/2017 08:20

M4225-ALTMAN_9781782549574_t.indd 10

03/05/2017 08:20

2

Rational inefficiency: smart thinking, bounded rationality and the scientific basis for economic failure and success Morris Altman

INTRODUCTION The core argument of this chapter is that individuals (economic agents) can behave inefficiently in a number of domains, at both the micro or macro (social) level. But this behaviour can be considered to be rational in the sense that such inefficiency can be a product of smart or considered choice behaviour. Smart people can be efficient or inefficient. From a smart-rationality assumption, we cannot necessarily derive choices that will have efficient outcomes. Moreover, what might appear to be irrational and, therefore, inefficient behaviour from the perspective of conventional economics might very well be, and often is, rational, smart, intelligent, considered and even purposeful behaviour from a smart agent perspective. Rational or, more generally speaking, smart behaviour should also not be confused with socially rational behaviour. What is rational from the individual’s perspective might very well be irrational from the social perspective as preferences across individuals and social groups might, and typically do, differ dramatically. Maximizing the preferences of a chief executive officer (CEO) need not be consistent with the long-term viability of the firm. Maximizing the well-being of the male partner in a relationship can be inconsistent with maximizing the well-being of the female partner. In addition, it is important to differentiate rational individual choice behaviour from behaviour that is error-free or decisions that are not subject to regret. Making mistakes and regretting these errors in decision-making can be consistent with rational or smart behaviour. Much depends on the decisionmaking capabilities of the individuals and the relevant decision-making environment. This chapter presents a modelling narrative on rational choice behaviour from a bounded rationality perspective. This builds on the pioneering work of Herbert Simon (1959, 1978, 1986, 1987) integrating the concepts of bounded and procedural rationality, and overlaps and is informed by the research and research orientations of Gerd Gigerenzer (2007), Friedrich Hayek (1944, 1945, 1948), Harvey Leibenstein (1957, 1966, 1979) and Vernon Smith (2003, 2005), as well as my own research on the subject (Altman 1999, 2005, 2006a, 2010, 2011, 2012, 2015, 2017). It is also informed by the research of Daniel Kahneman (2003, 2011; Kahneman and Tversky 1979; see also Tversky and Kahneman 1981); the heuristics and biases perspective. However, the smart decision-making approach generates results and orientations that contravene the heuristics and biases approach to decision-making and behavioural economics, which maintains conventional economic benchmarks for rationality and efficiency.

11

M4225-ALTMAN_9781782549574_t.indd 11

03/05/2017 08:20

12

Handbook of behavioural economics and smart decision-making

INTRODUCING RATIONALITY AND RATIONAL INEFFICIENCY To proceed in this narrative we have to clarify what is meant by rationality and by efficiency and what benchmarks we have to meet to be deemed rational and efficient. This remains a gap in the literature critical of conventional economics and, indeed, of the literature critical of the heuristics and biases approach and of the nudging approach to behavioural economics. Conventional economics is relatively clear on what is meant by rationality, what rational behaviours are, and what the expectations are for rational decision-making and its relationship to choices and outcomes. Conventional economics not only has reasonably clear benchmarks for rationality (discussed below), it also predicts rationality in human decision-making and hence in the outcomes emanating from these decisions. However, we should acknowledge that market failures remain a theoretical possibility even within the domain of conventional or neoclassical rationality when negative or positive externalities are present and not internalized by the decision-maker. Overall, conventional economics hypothesizes that rational inefficiency should not occur. This is predicated on the assumption that decision-makers are neoclassically rational (related to conventional economics definitions of rationality, discussed below). Moreover, given the prior assumption of the neoclassical rationality of decision-makers, it is assumed or predicted that the choices made by such rational agents will be efficient (assuming, for simplicity, no externalities exist). Given this overarching assumption of neoclassical rationality, we can predict that an individual’s choices yield optimal or ‘best’ outcomes given the constraints faced by the individual. By the prior assumption of rationality and de facto optimality on the part of individual decision-makers, we end up predicting that outcomes must be efficient and optimal. That which exists is presumed to be efficient by assumption as opposed to determining the extent of efficiency by empirical analyses. In this modelling scenario, it becomes possible to presume efficiency and optimality even when they actually do not exist. This can detract scholars from actively pursuing an analysis of the actual state of affairs, be it efficient or not, and its specific determinants. Note that in this approach, rationality is defined such that rational choice behaviour yields efficient outcomes. A core argument in this chapter is that smart decision-making is rational, but not necessarily neoclassically so. Rational behaviour can be inconsistent with ‘neoclassical’ behaviour and need not yield efficient outcomes. The conventional methodological approach fits nicely with what is referred to as the Humean fallacy, articulated in the A Treatise of Human Nature (Hume 1738 [2014], p. 576). Hume raises the problem of individuals deducing from what is, what ought to be (efficiency) and then attributing particular causes to the assumed efficiency, in this case a particular type of rationality. Since these deductions are not empirically based, they represent fallacies according to Hume. In fact, assuming that outcomes are necessarily efficient when choices are neoclassically rational is merely a testable hypothesis. This Humean fallacy is rooted in the dominant methodology in economics best articulated by Milton Friedman (1953, pp. 21–3) in his classic work on the praxis of positive economics. He argues that economically efficient outcomes are invariably the product of neoclassically rational behaviour. Hence only particular types of choices yield efficient outcomes. The derivative of this is that we do not have to investigate how actual humans behave or the process by which choices are made. It is good enough to assume that indi-

M4225-ALTMAN_9781782549574_t.indd 12

03/05/2017 08:20

Rational inefficiency 13 viduals behave as if they are neoclassically rational. Why? Because only efficient agents can survive on the market. If they survive they must be efficient. This efficiency can only be a product of neoclassically rational behaviour inside the firm. ‘Natural selection’, additionally, forces neoclassical rationality (in this case, what Friedman refers to as maximization-of-returns consistent behaviour) to dominate the behaviour of firm members and more specifically the decision-makers inside the firm. Therefore, the evidence in favour of rationality, joined with efficiency, is revealed by the survival of existing firms. Moreover, the persistence and dominance of the maximization-of-returns cum rationality assumption in the literature, both scholarly and popular, aided and abetted by no ‘credible’ alternative hypothesis explaining firm survival is, according to Friedman, further evidence of the scientific validity of the maximization-of-returns assumption. This type of argument is also developed in Alchian (1950) who argues that market forces create an environment wherein efficient choices are imposed on decision-makers, at least those with a preference for surviving on the market. There is no need for individuals to attempt to explicitly or carefully maximize profit of utility. Firms that survive are relatively efficient, because they survive. Hence, the behaviour of firm decision-makers must be consistent with such outcomes. This line of argument can be situated within the analytical domain of a Humean fallacy; any behaviour consistent with survival is considered to be acceptable and appropriate. For Alchian (1950), survival of the firm is evidence of relative efficiency, but there is no theory of human choice behaviour to benchmark which type of behaviours can yield, or should yield, relatively efficient outcomes. As with Friedman, there is little interest in how individuals actually behave. What is of concern is that any such behaviour yields economically efficient outcomes, at least in the long run. Alchian argues (1950, p. 213): Realized positive profits, not maximum profits, are the mark of success and viability. It does not matter through what process of reasoning or motivation such success was achieved. The fact of its accomplishment is sufficient. This is the criterion by which the economic system selects survivors: those who realize positive profits are the survivors; those who suffer losses disappear. The pertinent requirement – positive profits through relative efficiency – is weaker than ‘maximized profits,’ with which, unfortunately, it has been confused. Positive profits accrue to those who are better than their actual competitors, even if the participants are ignorant, intelligent, skilful, etc.

The conventional world view (and there are variations of this) is that individuals either behave in a fashion consistent with neoclassical rationality or they behave as if they are so doing. Ultimately it is expected that the outcomes will be economically efficient or utility maximizing either because market forces will guarantee this outcome or because individuals are hardwired to behave in this manner. The latter stronger assumption is all too often made. Typically, this is done implicitly. The end result is that a dominant prior assumption in the conventional wisdom is that outcomes will be economically efficient or utility maximizing. Moreover, it is assumed that because individual neoclassical rationality results in micro-level economic efficiency, this morphs into macro-level or social economic efficiency. It is then no longer analytically important to determine how individuals make their choices and which choices are made, or whether or not outcomes are in some identifiable sense efficient. Even institutional design and policy lose their importance if neoclassical rationality is predicted to yield economic efficiency in and of

M4225-ALTMAN_9781782549574_t.indd 13

03/05/2017 08:20

14

Handbook of behavioural economics and smart decision-making

itself – appropriate institutional design and policy are assumed to evolve naturally in an accommodating manner. Mancur Olson elaborates on this point with regards to social or macro-level economic efficiency derived from assumptions pertaining to micro-level neoclassical rationality and economic efficiency. Olson (1996, pp. 4–5) writes: The idea that the economies we observe are socially efficient, at least to an approximation, is not only espoused by economists who follow their logic as far as it will go, but is also a staple assumption behind much of the best-known empirical work. In the familiar aggregate production function or growth accounting empirical studies, it is assumed that economies are on the frontiers of their aggregate production functions. . . If the ideas evoked here are largely true, then the rational parties in the economy and the polity ensure that the economy cannot be that far from its potential, and the policy advice of economists cannot be especially valuable.

As evidenced above, critical to this neoclassical or conventional rational economic efficiency perspective is the assumption that the survival of economic entities is proof of economic efficiency and, correlated to this, economic efficiency is demonstrated by survival being proof of rationality. This survival principle builds upon the assumption that only efficient economic entities, which also happen to be rational, can survive in the market. To the extent that inefficient economic entities can survive in a moment in time (cross-sectionally) and over time, then survival can no longer serve as proof of efficiency or neoclassical rationality – critical to the conventional efficiency-rationality narrative. Survival would imply neither efficiency nor rationality. Moreover, if neoclassical rationality is not necessary to economic efficiency, then economic efficiency is not proof of economic agents being neoclassically rational. Another point that is important to note, and which will be elaborated on further below, is that the rationality–inefficiency–efficiency narrative can be applied to the realm of consumption or consumer behaviour. Conventional economics assumes that the revealed preferences of consumers through their choices off and on the market coincide with their true preferences – their wants and desires. In the realm of consumption this assumption represents an important aspect of consumption efficiency. Moreover, it is assumed that the process by which choices are made are consistent with the carefully calculating and prescient behaviour of consumers assumed in conventional economics. Hence consumption efficiency presumes the identity between revealed preferences and true preferences and these preferences being actualized within the parameters of neoclassical behavioural processes. However, even if we assume neoclassical processes, this is not sufficient to guarantee that revealed preferences are identical to the true preferences of decision-makers. This prior assumes that neoclassical processes are indeed the most effective means for achieving preferred ends. This is typically not the case in the real world of complex, costly and asymmetric information. The assumption equating revealed and true preferences builds implicitly upon very strong and unrealistic assumptions about the necessary conditions required for this equality to hold. Institutional parameters are critical in determining the extent to which revealed preferences are below optimal preferences. Moreover, unlike the reference to firm rationality and efficiency, market forces cannot guarantee that individuals should adhere to neoclassical behavioural decision-making protocols. There is no so-called survival of the fittest in the domain of consumption, even if we accept the

M4225-ALTMAN_9781782549574_t.indd 14

03/05/2017 08:20

Rational inefficiency 15 assumption that competitive forces can drive such neoclassical outcomes in the domain of production (Altman 2010). From the perspective of the heuristics and biases approach (Kahneman and Tversky 1979), there would be such suboptimal consumer behaviour, but this would be a product of the hardwired cognitive limitations of decision-makers, not the ‘inefficiency’ of institutional parameters, for example. Also, we expect suboptimal behaviour to be the rule, not the exception. This is most clearly elaborated and expressed in the Nudge approach to decision-making, which is well articulated in Thaler and Sunstein (2008). The more Simon-related behavioural economists situate consumer decision-making (as they do all types of decision-making) in the ‘environmental’ space, broadly speaking, within which decision-making takes place. Given this space, outcomes are considered to be optimal even though the realization of such outcomes does not follow neoclassical decision-making processes. This is now referred to as ecological rationality. Different decision-making processes (non-neoclassical processes) are expected to generate these optimal outcomes in the realm of consumption. So, what might appear to be irrational or error-prone and biased behaviour from the perspective of both the conventional wisdom and from heuristics and biases approaches could very well be both rational and optimal given the decision-making environment. This particular approach championed by Gerd Gigerenzer (2007), referred to as fast and frugal heuristics and ecological rationality, has roots in the work of Herbert Simon. Still, it remains an empirical question if, when and where particular heuristics yield optimal outcomes from the perspective of the individual or society. However, clearly, the benchmarks for what is consumer efficiency and rationality and what are the expected outcomes of the decision-making process are quite different in the conventional wisdom compared to what would be the case from the various perspectives in behavioural economics. Also, of significance to the smart agent approach to decision-making is an understanding of how sociological variables impact on choice behaviour, affecting the constraints and opportunities that frame the decision-making process and the choices made by economic agents. What is rational choice behaviour must also be contextualized by sociological variables such as peers, families, social norms and culture, for example (Becker 1996; Akerlof and Kranton 2010). Changing the social context of the choice environment impacts on what choices a smart individual will make. More generally, the conventional world view as well as most other methodological perspectives in economics, inclusive of the ‘new’ behavioural economics (heuristics and biases and related nudging approaches) and heterodox modelling, often also tend to rely, analytically, on the typical agent, household and firm where these are supposed to be the equivalent of representative agents. Where it is assumed that all economic agents are neoclassically rational and economically efficient, it is the typical agent, household and firm that is assumed to be so. This typical agent more often than not implicitly or explicitly refers to the average behaviour of economic agents, households and firms. However, the average cannot represent typical behaviour unless most individual behaviour is identical to the average. This assumption is unlikely to be true and cannot be assumed to be true without empirical validation. A critical focus on the typical or average is clearly articulated by Alchian (1950) – attention is placed on accurately predicting average outcomes and then imputing economic efficiency from these outcomes. However, if the objective is to determine the extent of economic efficiency and its

M4225-ALTMAN_9781782549574_t.indd 15

03/05/2017 08:20

16

Handbook of behavioural economics and smart decision-making

determinants (very much a function of the choices made by agents given their constraints, opportunities and capabilities), we have to go beyond the average and analyse the various empirical slices that comprise the average. There might be slices that comprise the average or typical that are efficient and others that are not; and there might be different means by which economic efficiency is achieved, which are smart but not neoclassically rational. Moreover, within each analytical slice, agents might be facing different opportunities and constraints and might possess different capabilities. These will affect what we mean by rational or smart decisions. The latter must be contextualized by the decision-makers’ overarching decision-making environment.

INSTITUTIONS, EFFICIENCY AND RATIONALITY Institutional frames are vitally important to a discussion of rationality and efficiency at both the micro and the macro (social) level. Whether or not decision-making is rational, and the extent to which efficiency is achieved, can only be determined if the decisionmaking process and the choices that flow from this process are contextualized by the institutional environment within which decisions are made. This is typically not done by conventional economists or even by economists off the mainstream. Not only must decision-making be institutionally contextualized, the framing must be empirically based. This entails that the framing must be derived from the actual institutional parameters within which the decision-making process takes place. We cannot assume that optimal institutional parameters are in place or will evolve willy-nilly. This is a point addressed by Simon (1987), articulating the importance of institutional parameters for decisionmaking processes and outcomes. Douglass North, one of the founding ‘fathers’ of what is often referred to as the new institutional economics, critiques conventional economics for paying no attention to the role institutions play in affecting choice behaviour and thereby economic outcomes, especially when decision-making is a dynamic process taking place through historical time. This is exactly the type of environment within which decision-making is embedded. Conventional theory is more concerned with stipulating equilibrium conditions given a particular institutional environment; often conditional upon an assumed institutional design that yields optimal economic outcomes. North (1994, p. 359) remarks: Neoclassical theory is simply an inappropriate tool to analyze and prescribe policies that will induce development. It is concerned with the operation of markets, not with how markets develop. How can one prescribe policies when one doesn’t understand how economies develop? . . . The very methods employed by neoclassical economists have dictated the subject matter and militated against such a development. In the analysis of economic performance through time it contained two erroneous assumptions: (i) that institutions do not matter and (ii) that time does not matter.

North (1991, p. 97) provides one possible definition of institutions and it would be this framework that North argues is ignored or assumed to be analytically irrelevant to cogent economic analysis: ‘Institutions are the humanly devised constraints that structure political, economic and social interaction. They consist of both informal constraints (sanctions, taboos, customs, traditions and codes of conduct) and formal rules (constitutions, laws, property rights).’

M4225-ALTMAN_9781782549574_t.indd 16

03/05/2017 08:20

Rational inefficiency 17 The reason why institutions are of analytical importance, argues North, is because they affect the incentive environment within which decision are made. For example, relative prices and relative opportunity costs of various types are conditional upon institutional parameters. North writes (1991, p. 97): Institutions and the effectiveness of enforcement (together with the technology employed) determine the cost of transacting. Effective institutions raise the benefits of cooperative solutions or the costs of defection, to use game theoretic terms. In transaction cost terms, institutions reduce transaction and production costs per exchange so that the potential gains from trade are realizeable [sic]. Both political and economic institutions are essential parts of an effective institutional matrix.

North continues that optimality in outcomes is not inevitable even if one assumes neoclassically rational agents. The types of institutions that are constructed, monitored and enforced determine outcomes. North remarks (1991, p. 110): When economies do evolve, therefore, nothing about that process assures economic growth. It has commonly been the case that the incentive structure provided by the basic institutional framework creates opportunities for the consequent organizations to evolve, but the direction of their development has not been to promote productivity-raising activities. Rather, private profitability has been enhanced by creating monopolies, by restricting entry and factor mobility, and by political organizations that established property rights that redistributed rather than increased income.

In North’s take on institutional economics, institutional design plays a pre-eminent role in determining the choices taken by decision-makers. Institutions, therefore, play a critical role in determining whether the outcomes of these institutionally derived choices are economically efficient. It is important to reiterate that North’s decision-makers can be neoclassically rational. However, such rationality need not generate economic efficiency at a micro level or at a social level, but given the institutional parameters imposed by a particular institutional design, such rationality would be utility maximizing, at least broadly speaking, from the perspective of the decision maker. North makes the case for utility maximizing rational inefficiency, contingent on whether or not institutional design incentivizes such inefficiency. Another pre-eminent economist, often associated with the conventional world view, also makes the case for rational inefficiency, contingent upon institutional design. Mancur Olson argues that the evidence is overwhelming that there are trillions of dollars lying on the sidewalk – something that should not occur in a world of rational wealth cum utility maximizing agents (1996, 19): The evidence from the national borders that delineate different institutions and economic policies not only contradicts the view that societies produce as much as their resource endowments permit, but also directly suggests that a country’s institutions and economic policies are decisive for its economic performance.

A key point made by Olsen is that even given the assumption of individual-based neoclassical rationality, societies can be socially inefficient and socially ‘irrational’, and they are socially irrational because they are socially inefficient. Institutional design determines the extent to which neoclassically rational agents generate socially inefficient economic

M4225-ALTMAN_9781782549574_t.indd 17

03/05/2017 08:20

18

Handbook of behavioural economics and smart decision-making

outcomes. He guestimates that such rationally inefficient outcomes are more the rule than the exception. Olson writes (1996, p. 23): ‘Some important trends in economic thinking, useful as they are, should not blind us to a sad and all-too-general reality: as the literature on collective action demonstrates . . . individual rationality is very far indeed from being sufficient for social rationality.’ At a very basic level the new institutional economics is incompatible with the core modelling assumptions of conventional economics. It makes the point that economic inefficiency can be rational and that economic efficiency requires particular institutional parameters to be in place. We can have billions, if not trillions, of dollars lying on the sidewalk even in the long term without rational or smart people picking these up. The incentive environment need not be appropriate for socially optimal behaviour to take place among utility maximizing individuals. Private utility maximization, which can take the form of rent-seeking behaviour, for example, can be consistent with social inefficiency. An economic agent, a decision-maker, might be maximizing utility, operating along her or his utility function, while the economy is operating in the interior of the economy’s production possibility frontier. Given a person’s utility function and given the institutional environment, it makes sense for the individual to maximize utility and profits through redistributing wealth as opposed to wealth creation. Taking this argument further, derivative of the ‘old’ institutional economics, I argue that such suboptimal (inefficient) outcomes could easily and predictably take place even with more appropriate (lower) transaction costs and more secure private property rights where agents are relatively secure that their legal gains from trade or their assets will not be arbitrarily confiscated by the state or by private agents. The capabilities of individuals, the preferences of decision-makers and the power relationship between decision-makers and potential decision-makers can impact the efficiency of economic outcomes, even given effective property rights and competitive market structures being in place. For example, inefficiency producing preferences, which would be a consequence of a preference for managerial slack (firm inefficiencies) or rent-seeking (social inefficiencies), can dominate even across different institutional parameters (Altman 2005). This social and power perspective to institutional economics is also absent not only from the new institutional economics, but also from the current and various perspective in behavioural economics.

DIFFERENT TYPES OF RATIONALITY It is important to have an understanding of what conventional economists tend to agree are the behavioural norms for optimal behaviour; that is, behaviour that yields efficient economic outcomes. Not everyone would completely agree on what these norms are. However, there are certain core assumptions that are often made reference to by both neoclassical and conventional economists, and by behavioural economists. What I outline below is not a straw ‘man’ that’s easy to attack and shoot down, and there are variations and modifications to this narrative. However, many would agree that the following is representative of the assumptions underlying much of conventional economic modelling: 1.

Individuals can and do make consistent choices across all possible bundles of goods and services and through time.

M4225-ALTMAN_9781782549574_t.indd 18

03/05/2017 08:20

Rational inefficiency 19 2.

3.

4. 5. 6.

7.

It is assumed that all individuals have a thorough knowledge of all relevant available options at any given point in time and they all have the means to process and understand this information in a timely manner – the brain is assumed not to be a scarce resource and individuals’ computational ability is assumed to be unlimited with respect to the decision-making process and problem in hand. Individuals can forecast the implications of their decisions through time and hence calculate, at least in a measurable probabilistic sense, the consequences of their choices. Individuals are assumed to make choices across alternatives that maximize utility or well-being, hence choices should not be subject to regret. It is typically assumed either explicitly or implicitly that, controlling for risk, utility maximization is consistent with wealth or income maximization. It is assumed that individuals are effective and efficient calculating machines, or at least they behave as if they are, irrespective of age, experience, education or social context. It is assumed that all individuals independent of context should behave in the same calculating manner (following conventional behavioural norms) to maximize utility or efficiency.

Herbert Simon rejects this neoclassical or conventional economics definition of rationality in favour of what he refers to as bounded rationality and, related to this, satisficing. Simon considers the conventional definition to be completely unrealistic and therefore useless with respect to constructing models that can speak to causation (as opposed to correlation) and to actual normative requirements to achieve optimal decisions and choices. Bounded rationality refers to smart or considered choice behaviour in the context of the choice environment and the decision-making capabilities of the decision maker. Hence, for Simon there is no unequivocal benchmark for rationality. It is context dependent and recognizes that decision-making capabilities differ across individuals, firms, households, ethnicities, cultures, religions, regions and nations. Satisficing is doing the best possible with what means are realistically available given the reality of bounded rationality. A key message here is that simply because decision-makers are not behaving neoclassically in their decision-making processes, this does not imply that they are irrational or inefficient. Indeed, behaving neoclassically might very well be irrational given the decision-making environment, yielding suboptimal outcomes (Simon 1978, 1986, 1987). This point is clearly articulated by James March, a close colleague of Simon. Rationality cannot be defined and modelled outside the context of the decision-making environment and the decision-making capabilities of decision makers (March 1978, p. 589): Engineers of artificial intelligence have modified their perceptions of efficient problem solving procedures by studying the actual behavior of human problem solvers. Engineers of organizational decision making have modified their models of rationality on the basis of studies of actual organizational behavior . . . Modern students of human choice behavior frequently assume, at least implicitly, that actual human choice behavior in some way or other is likely to make sense. It can be understood as being the behavior of an intelligent being or group of intelligent beings.

Vernon Smith, a pioneer of experimental economics, makes a related point, basing his understanding of rational behaviour on what works in effectively generating the preferred

M4225-ALTMAN_9781782549574_t.indd 19

03/05/2017 08:20

20

Handbook of behavioural economics and smart decision-making

outcomes of decision-makers. Such non-neoclassical behaviour, which we might refer to as satisficing, would be the most rational course of action for smart agents and should form the basis for constructing general norms for best practice behaviour or decisionmaking processes. Smith (2005, pp. 149–50; see also Smith 2003) writes: It is shown that the investor who chooses to maximize expected profit (discounted total withdrawals) fails in finite time. Moreover, there exist a variety of nonprofit-maximizing behaviors that have a positive probability of never failing. In fact it is shown that firms that maximize profits are the least likely to be the market survivors. My point is simple: when experimental results are contrary to standard concepts of rationality, assume not just that people are irrational, but that you may not have the right model of rational behavior. Listen to what your subjects may be trying to tell you. Think of it this way. If you could choose your ancestors, would you want them to be survivalists or to be expected wealth maximizers?

We can also refer to Gerd Gigerenzer (2007), who developed the concept of fast and frugal decision-making. The latter refers to decision-making processes that appear to be efficient in spite of being inconsistent with neoclassical processes. Fundamentally, the argument presented here is that decision-making must be contextualized and evaluated in terms of the decision-making environment and the decision-making capabilities of the individual (Gigerenzer refers specifically to Simon’s conceptualization of bounded rationality). Todd and Gigerenzer (2003, pp. 147–8) argue: ‘[B]ounded rationality can be seen as emerging from the joint effect of two interlocking components: the internal limitations of the (human) mind, and the structure of the external environments in which the mind operates. This fit between the internal cognitive structure and the external information structure underlies the perspective of bounded rationality as ecological rationality – making good (enough) decisions by exploiting the structure of the environment . . . Heuristics that are matched to particular environments allow agents to be ecologically rational, making adaptive decisions that combine accuracy with speed and frugality. (We call the heuristics ‘fast and frugal’ because they process information in a relatively simple way, and they search for little information.) The study of ecological rationality thus involves analyzing the structure of environments, the structure of heuristics, and the match between them.

The foundational behavioural economists, led by Simon, made a point of emphasizing that they do not dispute that human beings acting in the economic sphere (economic agents) are rational. They do not dispute this assumption of conventional economics, but they disagree on how conventional economics defines rationality. On rationality, Simon writes (1986, p. S210): I emphasize this point of agreement at the outset-that people have reasons for what they dobecause it appears that economics sometimes feels called on to defend the thesis that human beings are rational. Psychology has no quarrel at all with this thesis. If there are differences in viewpoint, they must lie in conceptions of what constitutes rationality, not in the fact of rationality itself. The judgment that certain behavior is ‘rational’ or ‘reasonable’ can be reached only by viewing the behavior in the context of a set of premises or ‘givens.’ These givens include the situation in which the behavior takes place, the goals it is aimed at realizing, and the computational means available for determining how the goals can be attained.

Simon further elaborates on rationality with regard to other social sciences, emphasizing that the conventional economics definition of rationality is a significant outlier in the social sciences (Simon 1986, p. S210):

M4225-ALTMAN_9781782549574_t.indd 20

03/05/2017 08:20

Rational inefficiency 21 In its treatment of rationality, neoclassical economics differs from the other social sciences in three main respects: (a) in its silence about the content of goals and values; (b) in its postulating global consistency of behavior; and (c) in its postulating ‘one world’ that behavior is objectively rational in relation to its total environment, including both present and future environment as the actor moves through time.

In defining rationality relative to decision-making Simon (1986, p. S211) points out that: The rational person of neoclassical economics always reaches the decision that is objectively, or substantively, best in terms of the given utility function. The rational person of cognitive psychology goes about making his or her decisions in a way that is procedurally reasonable in the light of the available knowledge and means of computation.

Simon elaborates on his concept of bounded rationality, making it more specific and nuanced. This brings him to a discussion of process rationality, which refers to the process of and the procedures used in arriving at a decision given the decision-making environment, the capabilities of the decision-maker and the objectives of the decision-maker. Moreover, process rationality takes into consideration that decision-makers’ understanding of what is best practice or optimal might be misconstrued or flat out wrong, but they rationally act upon such a misperception. Simon (1986, p. S211) argues that: if we accept the proposition that knowledge and the computational power of the decision maker are severely limited, then we must distinguish between the real world and the actor’s perception of it and reasoning about it . . . we must construct a theory (and test it empirically) of the processes of decision. Our theory must include not only the reasoning processes but also the processes that generate the actor’s subjective representation of the decision problem, his or her frame . . . The rational person of neoclassical economics always reaches the decision that is objectively, or substantively, best in terms of the given utility function. The rational person of cognitive psychology goes about making his or her decisions in a way that is procedurally reasonable in the light of the available knowledge and means of computation [it is context dependent].

Bounded rationality, satisficing and process rationality, all fit into a modelling paradigm that has as its core assumption that decision-makers are fundamentally smart. There can be exceptions to this rule. However, of critical importance is that we need to begin the analysis with a premise of smart agents doing the best they can, given their circumstances, their preferences, their understanding of available choices and their understanding of the best or optimal means of achieving their objectives. Deviations from neoclassical behavioural norms should not imply irrationality or inefficiency. More nuanced contextdependent norms need to be constructed for rational behaviour and what this implies for economic efficiency. This also implies a better understanding of how social context, social relationships, social norms and cultural factors, most of which can be reconfigured, impact on the rational choices that individuals make (Becker 1996; Akerlof and Kranton 2010). The ‘new’ behavioural economics, emanating from the initial research outcomes and initiatives of Kahneman and Tversky (1979; Kahneman 2003, 2011; see also Tversky and Kahneman 1981), sets out to develop theories that are better able to describe human behaviour, where often such behaviour is related to economic issues. This heuristics and biases approach rejects the neoclassical prediction that decision-makers will behave in a manner that will generate predicted ‘optimal’ choices. In this vein, for example, they

M4225-ALTMAN_9781782549574_t.indd 21

03/05/2017 08:20

22

Handbook of behavioural economics and smart decision-making

developed prospect theory as an alternative to subjective expected utility theory. Certainly, Kahneman and Tversky view their scientific project as bearing down on better describing choice behaviour than conventional economic theory. In the Kahneman and Tversky approach, such descriptive theories are typically related to the behaviour of the average individual. The focus on the average has also been a mainstay of conventional economics. This implicitly assumes that the average is the most appropriate point of reference for descriptive and analytical purposes. This ‘new’ behavioural economics also interprets the ‘average’ individual’s deviations from the conventional economic norms for optimal decision-making to be error-prone and biased, and typically persistently so. On the one hand, this perspective on behavioural economics maintains and adheres to a fundamental premise of conventional economics, that there is particular way of behaving in the economic realm resulting in a particular set of choices and therefore outcomes that are optimal (most, effective, efficient and unbiased). However, it represents a big break with conventional economics in that individuals tend not to behave optimally in a large array of choice scenarios. It is argued that individuals tend to engage in biased and error-prone behaviours; but they do so because they do not conform to conventional or neoclassical behavioural norms. Hence, the heuristics and biases approach retains neoclassical normative benchmarks for efficient and rational behaviour (although little mention is made of the term rational) (Berg and Gigerenzer 2010; Berg 2014). In the bounded rationality or smart agent approach to behavioural economics, errors and biases are not hardwired. There are those individuals with mental disabilities who engage in hardwired-biased behaviour – but these are clearly the exception to the rule. Overall, there are rational reasons that would explain most such biased and error-prone behaviour. At least this is the starting point of the smart agent perspective to economic modelling. What is meant by rational and even by efficiency (at least in the domain of consumption) would be different from that specified by conventional wisdom and by the heuristic and biases approach to behavioural economics (Altman 2017). Two key points need to be made and developed further. One is that it is important to specify or to think through (or model) the conditions under which rational decisionmakers generate either persistent local or social inefficiencies. It is important to also specify the extent to which such rational inefficiencies are a product of preferences of decision-makers, gaps in their capabilities, and/or biases or problems with institutional design. This is true for both the production and the household and consumer space. Modelling the necessary conditions for rational inefficiencies is the mirror image of modelling the necessary conditions for rational efficiencies. The focus of most conventional and behavioural economists has been on the process of achieving efficiencies often based on unrealistic assumptions of rationality, often decontextualized from pertinent institutional parameters. The second key point is the importance of better articulating the benchmarks for rationality and efficiency. For behavioural economists, following on from the Simon or bounded rationality perspective, this is a much more nuanced and complex narrative from what one finds in conventional economics or from the heuristics and biases approach, which rely largely on conventional benchmarks (Altman 2017).

M4225-ALTMAN_9781782549574_t.indd 22

03/05/2017 08:20

Rational inefficiency 23

PRODUCTION INEFFICIENCY In production, inefficiency can be defined as not making the best use of resources that are available for the task at hand. Hence, we would be operating inside the production possibility frontier. Or, we would be operating along a production isoquant that is further removed from the origin than it need be. In the latter case we would be using more inputs than required to generate a given level of output. This is also referred to as x-inefficiency in production (following the researches of Leibenstein 1966, 1979) as opposed to allocative inefficiency. The latter is a function of a distortion to relative prices, typically caused by oligopolistic market structures and presumed government distortions of the price mechanism. This leads to the misallocation of resources and hence to lower levels of productivity below that which would be the case when market prices are not distorted. However, it appears that allocative inefficiency is only of marginal importance as compared to x-inefficiency (Frantz 1997; Leibenstein 1966). Leibenstein considers x-inefficiency to be a product of irrational behaviour largely because decision-makers deviate from the norms of rational neoclassical behaviour (see also Cyert and March 1963). Leibenstein maintains that x-inefficient firms are a product of decision-makers, such as managers, not maximizing profits or minimizing costs as they should and would if they behaved in accordance with conventional economic norms. However, Leibenstein’s definition of rationality, although consistent with the overarching perspective of the heuristics and biases approach (using neoclassical behavioural benchmarks), is not at all related to whether decision-makers are making smart decisions given their constraints and opportunities and their preferences. Rationality is narrowly defined as it is in the conventional approach and in the heuristics and biases perspective. More importantly, Leibenstein creates an analytical space for persistent economic inefficiency by modelling x-inefficiency as a product of the preference function of decision-makers, where there is a preference for leisure as opposed to maximizing profits and minimizing costs. Here we have a preference function embodying managerial slack, yielding x-inefficiency in production. Decision-makers are, broadly speaking, maximizing their utility which, given their preferences, yield x-inefficiency. An important assumption in the conventional model is that preference functions of decision-makers are consistent with there being x-efficiency in production – firms using the fewest inputs possible to produce a given level of output. In reality, preferences of decision-makers are all too often not consistent with x-efficiency in production. This conventional benchmark for x-efficiency, minimizing inputs per unit of output, is a reasonable one, unlike the assumption of agents being super-calculators with prescience and perfect knowledge (in the relevant decision-making domain). I have argued that preferences inconsistent with x-efficiency (minimizing inputs per unit of output) are consistent with rational or smart behaviour. Agents can be purposeful, deliberative and even calculating, whilst still making choices that yield economic inefficiency (Altman 1999, 2005, 2006a, 2015). Leibenstein introduces the concept of effort discretion into the modelling of economic agents, something that runs contrary to conventional wisdom’s typical exposé of the economic agent. Effort should be, at a minimum, constant or even constant at some maximum according to the conventional wisdom. However, when managers organize the firm such that effort inputs diminish, then productivity falls and, ceteris paribus, average

M4225-ALTMAN_9781782549574_t.indd 23

03/05/2017 08:20

24

Handbook of behavioural economics and smart decision-making

cost increases. The firm is better off if it is x-efficient but, in this case, x-inefficiency in production is consistent with the preferences of decision-makers and, hence, with these agents maximizing their utility. Rational or smart agents attempt to ‘maximize’ their utility even if this results in suboptimal outcomes for the firm and society at large. This point can be illustrated in the equation 2.1, representing a simple economy with labour as the only factor input. Fundamental results do not change as we add other factor inputs to the production function. Ac 5

w Q a b L

(2.1)

AC is average cost; w is the wage rate or, more generally, the unit cost of inputs; (Q/L) is the average product of labour; Q is total output; and L is labour input measured in terms of hours worked. Reducing productivity by, for example, increasing managerial slack will, ceteris paribus, increase average cost (AC). Leibenstein assumes that w remains constant in the face of changes to productivity and average cost. Another way to visualize this argument is as follows: De 1 D(Q/L) 1 DAC

(2.2)

Going to the basic point, changes in effort input (e) yield changes in labour productivity (Q/L) which, in turn, yield changes in average costs (AC). In this model, maximizing effort inputs maximizes average product and, thereby, minimizes average cost. This would be consistent with x-efficiency in production. Such effort maximization is possible when the preferences of decision makers are consistent with this particular objective. I argue that effort maximization is rational or smart only under certain circumstances. Hence, economic efficiency (maximum x-efficiency), even among rational agents, should not be assumed as the natural state of things, given that economic inefficiency can be consistent with the preferences of decision-makers. And such preferences cannot be assumed to be irrational simply because they are not consistent with effort maximization. Leibenstein maintains that, given that decision-makers prefer the easy way out (managerial slack), unless product markets are highly competitive, x-inefficiency will persist. Since most markets are not highly competitive, he argues, x-inefficiency should be expected to dominate at different rates in different sectors, with a predicted strong positive causal relationship between more competition and more x-efficiency. However, Leibenstein argues that the political economy of market economies (which includes lobbying) would preclude product markets from being competitive enough for economic efficiency to be achieved. Within the context of imperfect product markets, managerial preferences play a key causal role in determining the extent of x-inefficiency. We can take this one step further. Smart agents and their preferences have a critical role in determining the extent of economic inefficiency because less than maximum levels of effort need not yield higher average costs, hence potentially threatening such firms’ survival on the market. The point made in Altman (2005, 2006a, 2010, 2011, 2012, 2015) is that managerial decisions (the extent of managerial slack, for example) affect the quality

M4225-ALTMAN_9781782549574_t.indd 24

03/05/2017 08:20

Rational inefficiency 25 and quantity of effort inputs among the ‘community’ of economic agents that comprise the firm as a unit of production. However, changes to effort levels are a costly process, affecting the levels of compensation to economic agents as well as investments in the quality of the work environment. Moreover, fixed costs are incurred if the system of management is transformed to change the level of productivity. This being said, if effort levels decrease this can be accompanied by lower wages and deteriorating working conditions and we would anticipate higher wages and improvements to working conditions when effort levels increase. From equation 2.1, we would expect w to be positively related to changes in productivity (Q/L). We would anticipate cost offset changes in effort levels. Rational or smart agents, therefore, have significant discretion as to how efficient firms end up being in long-run equilibrium since even with highly competitive product markets, inefficient firms can remain competitive and efficient firms need not have a cost advantage over less efficient firms. In this scenario, even competitive market forces cannot enforce economic efficiency on economic agents where this is incompatible with the preferences of the firm’s decisionmakers. Simply introducing more competitive product markets need not generate optimal economic efficiency. We can end up, as the evidence suggests we do, with firms ranging from highly inefficient to highly efficient even when the highly efficient outcomes are feasible and viable given current institutional parameters. We have multiple equilibrium in outcomes that flow from a multiple equilibrium in preferences (Altman 2016). In this narrative the preferences of members of the firm, whose preferences dominate the decision-making process, become of primary importance. This argument is illustrated in Figure 2.1. In this figure, aLCM represents our cost curve for the conventional firm if wages increase (effort levels are held constant) and for the Leibenstein model when effort levels are unrelated to changes in wages. In both cases the level of x-efficiency is held constant. And, average cost would increase in both scenarios. For the firm to survive, they would need protection, at a maximum of PLPL*.

Protection and multiple equilibrium

P’

P

LCM Extent of protection

d

Average cost

n del a

PL l mo iona t n e onv model ity) st: c variabil e co enstein g a r (effort b t e i c u e v d o L A e pr inal Averag orig PL* a

BM’

BMT

c

BM

Average cost (behavioural model with cost offsets)

0

d

X-efficiency

b

Induced technical change g

Wages and different levels of x-efficiency

Figure 2.1

Multiple equilibrium in production

M4225-ALTMAN_9781782549574_t.indd 25

03/05/2017 08:20

26

Handbook of behavioural economics and smart decision-making

Leibenstein’s x-inefficient firms can only survive when such firms are protected either through government policy or through imperfect product markets; but such protection is often afforded to inefficient firms. Alternatively, x-inefficient firms can survive by offsetting lower productivity related higher costs, by reducing labour costs. Along aBM, x-inefficient firms at different levels of x-inefficiency all produce at the same average cost as the x-efficient firm, given at point b. Cost offsets allow for multiple equilibrium with respect to x-inefficient and x-efficient firms. Market forces need not eliminate x-inefficient firms, even in the absence of product market imperfections and government protection, and even in the long run. The other side of the coin is that higher wages and improved working conditions need not generate higher average costs if compensated for by higher effort inputs which, in turn, yield compensating higher levels of labour productivity, here given by aBM’. In this scenario, higher levels of x-efficiency are consistent with higher labour costs. Indeed, the latter might be the cause of the former, forcing a reduction in the level of managerial slack, for example. Such higher labour cost firms can generate further cost offsets if higher labour costs induce technological change, which is illustrated by a shift in the average cost curve from aBM to ABMT (Altman 2009). This modelling narrative is consistent with what is articulated in the traditional prisoner dilemma (PD) model wherein particular ‘common knowledge’ assumptions yield social outcomes that represent a worst case scenario, even given the assumption of neoclassical rationality. In the realm of production, the worst case social outcome is one where productivity or output is at some minimum – the PD solution. It occurs when each participant in the game believes (common knowledge) that the other invests the least possible amount of time and effort in the process of production; maximizes her or his gains. This is consistent with narrowly self-interested maximizing behaviour (neoclassical rationality). In this narrative, we can increase (and maximize) our own individualized benefits by behaving in very narrowly self-interested fashion, if the other party actually contributes more than the anticipated minimum to the process of production. This is the case even though pie size is less than it might be otherwise (x-inefficiency in production). On the other hand, if we choose to behave in a manner that increases the size of the economic pie we risk a reduction in individualized benefits if the other party acts in a narrowly self-interested fashion. If the common knowledge is that the other party will act in a narrowly self-interested fashion, it would be rational to do the same, for only in this way can we minimize any potential losses to ourselves. Only if we change the common knowledge of the other’s behaviour will it be rational to behave in a fashion consistent with the common or social good, increasing the size of the economic pie. With increased pie size, each player of the economic game could see her or his real income increase – everyone is a potential beneficiary. This would be a cooperative solution to the economic problem, in direct contrast with the PD solution. Non-cooperative solutions are possible, as discussed above, when non-cooperative firms are protected from market forces or when they are able to trade-off low productivity with low wages and poor working conditions. Both PD and cooperative outcomes are sustainable and rational, given the preferences of decision-makers and the decision-making environment within which their decisions are made. It also needs mention that the constraints on decision-making within the firm are set by members of the firm hierarchy in the traditional investor-owned firm. If joint preferences of

M4225-ALTMAN_9781782549574_t.indd 26

03/05/2017 08:20

Rational inefficiency 27 the firm hierarchy are of the non-cooperative type wherein utility is maximized, the PD solution is inevitable. If a cooperative solution is what maximizes joint utility, then a cooperative solution would follow. In cooperative (worker-owned) firms the joint preferences would veer towards the cooperative solution. Also, power dynamics within the firm can effect which solution dominates. More bargaining power in the hands of workers can, but does not guarantee, a more cooperative solution as members of the firm hierarchy must find the means to increase productivity to offset the increasing direct costs of production that often follow when the bargaining power of employees is enhanced. Also, for firms where employees have a more substantive say on managerial and corporate decisions (a mixed hierarchical model), a cooperative solution is more likely. The same would be the case if owners and managers have a joint preference in favour of more cooperative outcomes (Altman 2002). Further related to the neoclassical assumption of what comprises rational behaviour within the firm, in behavioural-type models of the firm, simplistic formulations of profit maximization or cost minimization, especially in their mathematical presentation, tell us little about what is required for firms to be economically efficient. Being efficient is not a matter of equating marginal revenue to marginal cost. For example, even within the framework of a very simple model, assuming that firm decision-makers can actually and effectively do this calculation in a dynamic fashion, we can equate marginal cost and marginal revenue without effort being maximized. For any given level of effort input we can do this calculation. Hence, the firm could be economically inefficient even when marginal cost equals marginal benefit. The relevant marginal cost and marginal revenue functions would simply be different from what they would be if effort input was maximized. Also, in this type of modelling, the decision-makers would be maximizing their utility at different levels of effort input. The utility maximizing level of effort input is given by the preferences of the decisionmakers. Hence, any model that is scientifically robust must incorporate the conditions under which effort levels inside the firm are higher or lower, since these conditions are critically important for any determination of why and how rational or smart agents generate a particular level of effort input and, therefore, a particular level of productivity. The details of what transpires inside of the ‘black box’ of the firm becomes critically important because it is in the black box that we can deconstruct the methods adopted by decision-makers to achieve their chosen ends. There may also be alternative means to achieve efficiency, all of which might be consistent with the generic and often vacuous normative directive that efficiency is achieved when economic agents equate or behave as if they equate marginal costs to marginal benefits. There are those who argue that a heavily monitored and punitive environment where labour costs are minimized (such as wages and quality of the work environment) serves to maximize labour productivity. However, there is strong evidence to suggest that a more collaborative work environment based on teamwork, trust and reciprocity is better able to achieve economic efficiency. Here there is a more equitable (but not equal) distribution of power and income inside the firm. Both organizational structures and related processes could be rational from the perspective of the dominant decision makers (and their preferences), even though neither adheres to the behavioural processes that fit into the simplistic marginal cost equals marginal benefit narrative of conventional economics (Altman 2002). Related to the conventional prediction that rational behaviour should yield economic

M4225-ALTMAN_9781782549574_t.indd 27

03/05/2017 08:20

28

Handbook of behavioural economics and smart decision-making

efficiency, there is the commentary of Richard Posner (2009) on the (2007–08) global economic crisis. Posner was a leading proponent of the efficient market hypothesis and neoclassical rationality as the best way of modelling the economy and the relationship between law and the economy. He shifted theoretical ground to overlap with the Simon or bounded rationality modelling perspective. His perspective also overlaps with the view that rationality should not be interpreted as neoclassical rationality. Among the critical points made by Posner is that decision-makers’ rational behaviour in terms of efforts to maximize income need not take the form of neoclassical processes (they could involve emotion, intuition and herding). All these rational behaviours, however, can cause longrun harm to the firm, even while generating significant short- and even long-run benefits to the individuals engaging in such rational behaviours. Posner (2009, p. 111) elaborates: In sum, rational maximization by businessmen and consumers, all pursuing their self-interest more or less intelligently within a framework of property rights and contract rights, can set the stage for an economic catastrophe. There is no need to bring cognitive quirks, emotional forces, or character flaws into the causal analysis. This is important both in simplifying analysis and in avoiding a search, likely to be futile, for means by which government can alter the mentality or character of businessmen and consumers.

Posner argues that to prevent an economic meltdown, or at least to reduce the probability of one, we should not attempt to re-wire decision-makers so that they behave more neoclassically. Neither should we attempt to re-wire them so that they become less greedy or less narrowly self-interested – which Posner argues is very difficult to operationalize with substantive effect on the economy. To prevent or minimize the probability of narrowly rational income, wealth maximizing or ‘greedy’ individuals (those attempting to maximize their private income or wealth) causing social harm, which incorporates reducing longrun firm real income, wealth and/or productivity, governments must change the institutional environment. This goes beyond simplistic references to improvements to property rights and reducing transaction costs, which is often the focus of the new institutional economics. Also of importance would be providing decision-makers with improved information sets, improved information processing and analytical capabilities, and better understanding of viable organization options (low wage versus high wage, for example), and internalizing externalities to the firm and individual decision-makers (hence reducing the probability of moral hazard). Shiller (2008, 2012) argues for the improvements in the legally enforceable and regulated provision of transparent, accurate and understandable information to be important to a well functioned and socially efficient market economy. In summary, rational inefficiency is a very reasonable outcome given the preferences of dominant decision-makers and the institutional environment within which they are embedded. By acknowledging the possibility of rational inefficiency and its underlying determinants, we can suggest means of achieving more efficient outcomes. Moreover, if we are able to better model the conditions underlying rational inefficiencies, we can better identify when and where they exist as opposed to assuming ex ante that decision-makers make choices that yield economically efficient outcomes. Thus far, I have discussed rationality and efficiency in terms of productive sectors as opposed to rent-seeking sectors of the economy. However, it is important to note that even if all agents behave in an economically efficient fashion, this does not preclude this result-

M4225-ALTMAN_9781782549574_t.indd 28

03/05/2017 08:20

Rational inefficiency 29 ing in x-inefficiency in production in the economy as whole. One can have x-efficiency in sectors of the economy that are of a rent-seeking nature, wherein the firm’s wealth is a product of transferring resources from one sector to another or from one individual to another. Here, the non-productive sectors are x-efficient, but the economy as a whole is operating below its production possibility frontier as a consequence of the efficiency in the rent seeking sectors. What is important to note also is that criminal behaviour, lobbying, corruption and war machines that engage in income transfers to the conquering population can be run in an economically efficient manner. Institutional parameters can make such organizational forms more attractive (profitable) to economic agents. Economic efficiency, even when agents are neoclassically rational at the organizational level, in no way necessarily translates into social efficiency. At the extreme, we can have a rent-seeking based society, run by rational agents, that is efficient at the level of the organization but which is socially inefficient. Rationality implies neither efficiency nor inefficiency in production. Rationality also does not imply neoclassical behavioural norms in the realm of production. Smart decision-makers can deliver firm and socially efficient outcomes in production, contingent on the preferences of decision-makers, decision-making and organizational capabilities, and the overall incentive environment; but these conditions all too often do not prevail. From a smart agent perspective, it is a critically important scientific task to identify those conditions conducive to economic efficiency at both the firm and social levels. Smart agents can generate x-efficiency at both the firm and social levels given the appropriate circumstances.

CONSUMPTION INEFFICIENCIES Although I have devoted considerable attention to the rationality–efficiency–inefficiency narrative in the domain of production, contemporary behavioural economics devotes considerable energies to rationality–efficiency–inefficiency in the realm of consumptionrelated behaviour. A fundamental prior in contemporary microeconomic theory is that the revealed preferences of individuals represent their true and, related to this, utility maximizing preferences. Moreover it is assumed that these true preferences can be realized through the choices a person makes given her or his income and given relative prices. I have referred to this as choice x-efficiency (Altman 2010). Conventional economics assumes that choice x-efficiency is the rule in any given society and at any given point in historical time. In brief, true preferences represent those preferences of an individual that are formed in an environment wherein he or she has access to relatively complete and truthful information pertaining to pertinent choice decisions, has the capabilities to process and understand such information, and where this person’s preference formation and choices are not constrained by coercive circumstances. These assumptions are layered over the assumption that individuals are rational in their decision-making process. All these assumptions must hold for choice x-efficiency to prevail. Harsanyi (1982), one of the pioneers of choice theory, makes similar points with regard to the necessary conditions for revealed preferences to equal what we might refer to as true preferences (see also, Altman 2010).

M4225-ALTMAN_9781782549574_t.indd 29

03/05/2017 08:20

30

Handbook of behavioural economics and smart decision-making

I have argued that rational or smart choice behaviour requires the prevalence of the above preference formation and choice environment. However, to be realized, such rational choice behaviour only requires boundedly rational behaviour as opposed to the unreasonable and unobtainable prescient and super-calculating behaviour of conventional neoclassical economics. Even smart decision-makers cannot realize choice x-efficiency unless the appropriate preference formation and choice environment prevails. Hence, we can end up, under very reasonable circumstances, with rational inefficiencies (what I refer to as choice x-inefficiencies) in the realm of choice. Even if we can form ‘true’ preferences, rational individuals may not have the power to translate these preferences into choices or revealed preferences. In this scenario individuals are not free to choose. For example, a women may want to have one child, but may be forced into having six, or a parent might want her daughter to learn to read and write but is not empowered to do so, given social norms and legal parameters. Building on a bounded rationality platform, we can model conditions wherein choice x-inefficiencies/x-efficiencies can be obtained. Only under particular institutional/ environmental/social circumstances can choice x-efficiency be realized. Hence we should be able to identify the circumstances under which choice x-inefficiencies (with smart decision-makers) exist and how such circumstances need to be changed for revealed preferences to converge to an individual’s true preferences. It is important to note that true preferences, these utility maximizing preferences, irrespective of how ‘rational’ they might be, need not be socially rational. Choices that cause harm to others can be rational and reveal the true preferences of the individual decisionmaker. This socially suboptimal behaviour represents a form of market failure wherein externalities are not internalized by the individual decision-maker. Such market failures are not part of the conventional narrative even when we can legitimately assume that revealed preferences equal true preferences. This is the case even though the ‘forefathers’ of preference theory recognized this very real possibility (Harsanyi 1982). But market failure of this type can be easily incorporated into a modelling of preference formation and choice realization, as articulated above. In my modelling of choice x-inefficiencies and choice x-efficiencies, as in the conventional wisdom, there is the possibility of revealed preferences being identical to true preferences; but there is also the possibility that this equality need not hold – there is an analytical space for rational choice inefficiencies, irrespective of whether or not one models agency from a neoclassical or boundedly rational perspective. Moreover, there is a possibility that individuals do not have the capabilities and are not in a decision-making environment for true preferences to be formed. There is also the possibility of market failure in the domain of choice. This modelling narrative, therefore, does not accept as a prior working assumption that choice efficiency at an individual and a social level prevails everywhere and always. Its existence and prevalence is an empirical question, very much contingent upon the necessary institutional parameters (inclusive of appropriate power relationships) and individual decision-making capabilities being in place. Freedom of choice philosophically underpins the conventional wisdom’s normative preference for the revealed preference-utility maximizing modelling of decision-making. The individual’s preferences determine choice that, in turn, allows the individual to maximize her or his utility. Here freedom of choice is valued as core to the ability of individuals to ‘maximize’ their utility, their level of well-being. Modelling preference for-

M4225-ALTMAN_9781782549574_t.indd 30

03/05/2017 08:20

Rational inefficiency 31 mation and choice from a critical and bounded rationality perspective does not obviate a normative focus on the critical importance of individual freedom for utility maximization (accept when the latter causes harm to others). This overlaps with the critical approach taken by Sen and Nussbaum (Sen 1985; Nussbaum 2011) on this matter applying their capabilities analytical framework. Here, too, freedom of choice is critically important. The problem is that this freedom only exists if the institutional and individual decisionmaking capabilities are present. Once these conditions are met then, in this modelling framework, as in the conventional wisdom, we would predict that individuals’ choices should be ‘maximizing’ their utility or level of satisfaction. However, in the bounded rationality-choice x-efficiency approach, public policy would be required to assure that conditions for choice efficiencies and hence for freedom of choice are met. In the conventional approach it is typically assumed, ex ante, that such conditions are present everywhere and always. This normative approach has been challenged by the stream of behavioural economics linked with the heuristics and biases analytical framework developed by Kahneman and Tversky. A key point made here is that individuals are hardwired to be error-prone in decision-making. The capacity to form and execute our true preferences, therefore, will not preclude persistent errors and biases in decision-making. If anything, such freedom (even assuming that there are no choice x-inefficiencies – true preferences can be realized) can predictably cause more harm than good. This perspective has been most forcefully and poignantly developed and articulated by Thaler and Sunstein (2008) in their nudge approach to behavioural economics and public policy. They argue that there are clear objective benchmarks for what it means for individuals to be better off or maximizing their utility. These benchmarks appear to be universal, running across individuals, but it is argued that these universal benchmarks for utility maximizing, ‘best-practice’ behaviour cannot be realized by the typical individual exercising free choice. This is in part because individuals are not properly hardwired to do so – hence the persistent biases in choice behaviour, resulting in individuals’ choices yielding suboptimal outcomes for the individual decision-maker and society at large. An important assumption in this modelling is that preferences are the same across individuals – homogeneous preferences. Hence what is good for all individuals is based upon what is deemed to be good from the perspective of the expert. Individual preferences do not inform the content of what is ‘good’. The baseline for what is good is largely based on ‘neoclassical’ benchmarks and a depth of knowledge and emotionless understanding beyond the limit of typical human decision-makers. However, it is assumed that the expert has the capabilities, knowledge and understanding to identify what is truly utility or welfare maximizing and the means to achieve this in a most efficient and effective manner. Thaler and Sunstein (2008, p. 176) maintain: We intend ‘better off’ to be measured as objectively as possible, and we clearly do not always equate revealed preference with welfare. That is, we emphasize the possibility that in some cases individuals make inferior choices, choices that they would change if they had complete information, unlimited cognitive abilities, and no lack of willpower.

Critical to this interpretation of what is good for the individual and what is the baseline for the good is choice architecture and the choice architect. An important feature of choice architecture is reconfiguring the choice environment in a manner that induces

M4225-ALTMAN_9781782549574_t.indd 31

03/05/2017 08:20

32

Handbook of behavioural economics and smart decision-making

or, in more extreme circumstances, forces the individual to make choices that the expert deems to be in the individual’s best interest. Note that each individual does not have her or his own specific choice architecture. The latter is generic, as all individuals are assumed to be homogeneous in preferences. Actual differences in preferences across individuals are not recognized here (a type of simplifying assumption). The choice architect is the expert who designs the choice environment nudging the individual to make choices that will make her or him better off (higher level of utility, satisfaction or well-being) from the perspective of the choice architect or expert. This would be the case even if the affected decision-maker did not believe that her or his nudged choices increases her or his level of utility or satisfaction – making this person better off. The expert – the choice architect – knows best. A fundamental policy implication of the heuristics and biases approach is that people opposing choice architecture do so because they make the assumption that each individual knows what is in her or his best interest. This assumption is fundamentally flawed from the heuristics and biases approach. That is, this approach contests a fundamental world view of conventional neoclassical economics as well as that of the boundedly rational-smart decision-maker perspective articulated here. In the conventional perspective revealed preferences always (or almost always) reveal the true preferences of the individual, which equates with behaviour that maximizes an individual’s level of utility or satisfaction or well-being. Smart decision-makers would do what is in their best interest if they have the capabilities to form and then to realize their true preferences. However, from the heuristics and biases approach, true preferences are expected to be inconsistent with what is in the best interest of the decision-maker (Thaler and Sunstein 2009, p. 6): [A]lmost all people, almost all of the time, make choices that are in their best interest or at the very least are better than the choices that would be made by someone else. We claim that this assumption is false. In fact, we do not think that anyone believes this on reflection.

It is important to recognize that the nudging perspective contains many elements, some of which are paternalist, and others which are consistent with creating the conditions for the formation and realization of true preferences. However, the focus has been on the paternalist component, inducing or forcing individuals to make choices consistent with the expert’s preferences. An important component of the nudging approach is framing options such that individuals make expert-consistent choices. This could involve forcing organizations to re-frame options available to consumers so that consumers make expertconsistent choices. A fundamental argument put forth by Thaler and Sunstein is that choice options are always framed and that there is always someone who constructs the frame. Little analytical attention has been paid to framing because conventional economics assumes that framing does not affect the choices made by decision-makers. The implicit assumption in the nudging approach is that different frames contain no new information pertinent to a particular decision. Hence, individuals are easily manipulated by changing the framing of a choice option even when the revised frame is not substantively different from the prior frame. A classic example given is that of the framing of pension options. If the default option is not to invest in a pension, then most employees will not invest. However, if the default is to invest, most people will invest. The frame, in this case, is the default option. Simply changing the frame appears to have a huge impact on whether or not individuals

M4225-ALTMAN_9781782549574_t.indd 32

03/05/2017 08:20

Rational inefficiency 33 invest in a pension. The assumption is that individuals are indifferent in terms of utility between the two different frames and simply make their choices based on the different frames while their utility remains constant. However, it is further assumed that one of the frames yields choices that make the individual and society better off (they are both individually and socially optimal and welfare improving). Hence the positive view of the interventionist role of the expert, of the choice architect. This apparently clear-cut example appears to demonstrate the case for soft and, even, hard paternalism. However, at a minimum, in a world of complex, asymmetric and even misleading information, and the limited decision-making capabilities of decision-makers (partially based on learning deficiencies), defaults can represent signals to decision-makers as to which choice or choices have the highest probability of making them better off. When the default is not to invest in pensions (especially in a particular pension plan), this signals that experts deem this not to be the best idea, and the opposite if the default is to invest. Hence, the decision-maker is relying on the integrity of those setting the default to inform the decision-maker on which choice might be the best choice. By changing the default from non-investing to investing, the ‘expert’ must assure that the investor knows what he or she is getting into, such as various opportunity costs and risks. Re-framing is not simply changing the frame from one to another wherein no substantive information is being changed. Re-framing typically involves making substantive changes to the information affording to the decision-maker. This is why it is rational for decision-makers to change their behaviour when frames are changed. When the default is not investing, the onus is on the decision-maker to determine what is in her or his best interest, as the expert is not signalling to invest in this instance. Even here, framing becomes important, but not in the sense emphasized by the heuristics and biases or nudging perspective (Altman 2011, 2012; Gigerenzer 2007). How choices are framed is important because frames contain information fundamental to the determination of choice. Hence, for choices to be ‘optimal’ requires that the frame provides the individual with truthful, comprehensible and accessible information so that the individual can make the best possible decision given the choice-set available. This alternative, smart agent approach to framing focuses on providing decision-makers with an environment wherein they can better form their preferred choices and exercise these choices. This approach would not apply to situations when an individual’s optimal choice causes harm to others. Examples of this would be smoking in public spaces, taking heroin while pregnant, being abusive to your spouse and children, closing factories and asset stripping to maximize short-term gain for major shareholders, and cheating and deceiving customers. The smart agent approach focuses on improving the preference formation and decisionmaking environment, while accounting for and incorporating negative and positive externalities in this endeavour. Although there is some overlap between this and the nudging approach, the prior working assumption of the smart agent approach is that individuals’ preferences should, for the most part, be respected and that when choices are suboptimal even for smart decision-makers, they tend to be so for reasons of institutional design, for environmental reasons or for reasons of capabilities. Hence the focus is on institutional design, capabilities development and empowerment of decision-makers. This approach also pays attention to the enforcement of rules and regulations that can contribute towards an improved decision-making environment. From this perspective errors in

M4225-ALTMAN_9781782549574_t.indd 33

03/05/2017 08:20

34

Handbook of behavioural economics and smart decision-making

decision-making can and are often made. However, this is related to environmental and capabilities issues as opposed to the hardwiring of the human brain. The expert plays a role in this analytical construct by contributing to improvements in the decision-making environment as opposed to determining which decisions individuals should make. This bounded rationality smart agent approach is libertarian in orientation, but recognizes the importance of various levels of government and expert intervention to improve the overall decision-making environment and the decision-making capabilities of decision-makers as well as developing an incentive environment that accounts for negative and positive externalities. Martha Nussbaum, the co-developer, along with Amartya Sen, of the capabilities approach, makes a similar point. She argues (1999, p. 49): ‘Government is not directed to push citizens into acting in certain valued ways; instead, it is directed to make sure that all human beings have the necessary resources and conditions for acting in those ways. By making opportunities available, government enhances, and does not remove, choice.’ Related to my narrative on choice x-efficiency and smart decision-making, what Nussbaum is referring to is the creation of optimal preference formation and decisionmaking environments, as opposed to experts determining the choices that people should make.

MACROECONOMIC CHOICES AND RATIONAL BEHAVIOUR As with microeconomic behaviour, in the macroeconomic domain, conventional economics makes the case that decision-makers must be neoclassically rational. The evidence suggests this is not how individuals behave and this has had some major repercussions in the construction of macroeconomic theory and for finance theory (Akerlof 2002; Akerlof and Shiller 2009). However, these reconstructions are rejected outright by those who remain strict adherents to the conventional assumptions of rationality combined with assumptions related to flexible factor prices and the capacity of micro decisions having direct and immediate impact in the macroeconomic domain. The latter ‘school of thought’, in their pre-Keynesian incarnation, has been dubbed the classical school, whereas their modern equivalents have been referred to as the new classical school of macroeconomics. Many of the underlying revisions to classical macroeconomic theory were made decades ago by Keynes in his articulation of business cycle theory, more specifically, his theoretical narrative on the making of deep recessions and the mechanism involved in the economy transitioning from a deep recession or depression to recovery. Keynes’s narrative is largely based on the assumption of smart agents making decisions in a world of complex and asymmetric information with asymmetric power relationships across decision-makers. Keynes introduces the notion of ‘animal spirits’ as important to the determination of the timing and depth of recessions and upturns. His narrative suggests that animal spirits as a determinant of decision-making are rational in the sense that decision-makers are doing their best, given their decision-making environment. Hence, Keynes recognizes the importance that non-economic variables can play in determining economic (macroeconomic) outcomes (Keynes 1936). In the conventional wisdom, such non-economic variables are assumed away. Keynes

M4225-ALTMAN_9781782549574_t.indd 34

03/05/2017 08:20

Rational inefficiency 35 also recognizes the importance of sticky prices as being a possible and possibly important determinant of recession/depression, given negative demand shocks. However, the relative importance of sticky prices in determining economic downturns, especially severe downturns, is subject to heated debate among those writing in the Keynesian tradition, but there is no denying the empirical significance of sticky prices. Most recently Akerlof (2002), in theory, and Bewley (1999), empirically, have made the case that sticky prices in the face of a negative demand shock are rationally determined. This is based on what is referred to as efficiency wage theory, first modelled by Harvey Leibenstein (1957). Smart agents make local (within the firm) utility and profitmaximizing decisions that have negative macroeconomic consequences, such as persistent unemployment. Firms do not cut real wages for fear that workers will retaliate by cutting effort inputs, thereby reducing productivity. Here effort is a variable in the production function. Employers are also concerned that their best workers will quit, given the opportunity, for what are perceived to be fairer firms, also damaging firm productivity; but workers maximize their utility by taking such action, which is common knowledge to employers. Akerlof considers sticky-price related unemployment to be involuntary. The employed do not want to lose their jobs or keep others unemployed even though their locally rational decisions have this effect. It is important to note that classical economists, old and new, pay no attention to this efficiency wage modelling of unemployment, but they could interpret such behaviour as a reflection of labour’s preference for leisure, at least on the margin. There would be the assumption that in the face of negative aggregate demand shocks, employment would be restored by cutting real wage below where it was prior to a particular negative demand shock. The assumption is also made that workers can determine their real wages as opposed to simply their nominal wages. In terms of the narrative of this chapter, what is critical for causal analysis and policy is whether or not decisions-makers are rational, and the implications of this for analysis and policy. The pre-Keynesian and new classical economics perspectives assume that rational agents would endeavour to clear all markets (prices are flexible) and behave as if prices are flexible. Hence, if unemployment exists or if it increases, this is related to the rational decision to keep real wages too high, for example. Here, unemployment or increases in unemployment are voluntary. There can no substantive demand-side problem, especially in the longer run. The assumption here is that increases in the unemployment rate are a product of changing preferences of workers in favour of more leisure or non-labour market activities or government interventions that make labour markets less flexible and/ or increase the real wage above what would be generated in a ‘pure’ market economy. The increased real wage is predicted to increase the rate of unemployment. Such institutional interventions (such as minimum wages and unions) increase the structural rate of unemployment. Here, too, the demand side is not important. A popular rendition of this perspective was put forward in Friedman’s classic 1968 article, making a case for supply-side determinants of macro outcomes. He focuses on what he refers to as the natural rate of unemployment, which is determined by the structure of real wages. Not much attention is paid to severe negative demand shocks. Ultimately, if workers wanted more employment they should and would cut their real wages. It is assumed that this would not have a knock-on effect of reducing aggregate demand and therefore further increasing the rate of unemployment. Moreover, as unemployment

M4225-ALTMAN_9781782549574_t.indd 35

03/05/2017 08:20

36

Handbook of behavioural economics and smart decision-making

increases, even dramatically so, it can be attributed to changing preferences of workers in favour of more leisure time or changing government policy that permanently increases real wages to higher levels – facilitating workers’ preferences for more leisure time. Notice that among the old and new classical economists, and among many Keynesian economists, there is a prior assumption that decision-makers are rational, but the understanding of rationality differs across schools of thought, with significant implications for policy. Across the board, Keynesians regard spikes in unemployment yielding substantive increases in the unemployment rate to be involuntary. These increases in unemployment would be impossible for the market to deal with quickly and efficiently, that is, in the real world of complex and asymmetric information, limited foresight, inflexible prices and the consequential reliance (to a lesser or greater extent) on decision-making heuristics, such as herding. There is no evidence that markets naturally clear swiftly after a severe demandside shock. However, the classical economists assume that this reflects the preferences of decision-makers (there is very little modelling attention paid to different preferences and different power relationships across agents). This adds weight to the argument that our definition of rationality, what it means to be a smart decision-maker, and the realism of our modelling of the decision-making process, is vitally important for causal analysis and, in the macro domain, for public policy. A core Keynesian argument is that increasing demand either through monetary or fiscal policy will restore the economy to full employment in a relatively quick and efficient manner. Hence, the excessive demand-side related unemployment would be eliminated and the economy restored to the prior and lower natural rate of unemployment. The higher unemployment rate that is realized during a depression or deep recession is not the natural rate of unemployment – which is the claim of classical economists, old and new. A critical assumption made by Keynesian economists is that for involuntary unemployment to be eliminated, workers must accept lower real wages, as increasing employment requires the formerly employed less-productive workers (lower marginal product) to accept lower real wages. It is assumed here that a downward sloping marginal product of labour curve, over its relevant portion, characterizes the representative firm, which is a very big short-run assumption indeed. The decreased real wage must coincide with adequate increases in aggregate demand. Classical economists argue that accepting lower real wages would not be the rational response of the typical worker. Hence, increasing aggregate demand can have no real effect on the economy, measured by increased employment. However, the side-effect of such activist demand-side policy would be increased prices or increasing the rate of inflation. Akerlof has attempted to provide a scientific quasi-rational basis for government policy to restore employment towards its pre-recession levels (Akerlof 2002). He maintains that workers in some sense suffer from money illusion (quasi-rationality) and will therefore not pay attention to reductions in real wages that are a function of low rates of inflation. Basically, the transaction costs of computing the impact of low rates of inflation on real wages are not worth the benefits. Hence, increasing aggregate demand to increase employment should be effective as long as we buy into the realism of this transaction-cost based money illusion argument. Decades earlier, Keynes rejected any presumption of money illusion on the part of

M4225-ALTMAN_9781782549574_t.indd 36

03/05/2017 08:20

Rational inefficiency 37 workers, although he accepted the assumption that real wages need to be decreased for pre-recession or depression rates of unemployed to be restored. Workers would accept cuts to real wages that were generalized across sectors and occupations, as these would be seen as fair especially when accompanied by increased employment. This could be achieved through aggregate demand-side induced inflation. Workers, themselves, could not orchestrate such a cut in real wages. This would have to be effected through macroeconomic government policy. There is no money illusion here at all. Moreover, Keynes theorizes that self-imposed cuts to money wages would simply reduce aggregate demand, further dampening animal spirits and, thereby, further increasing unemployment. Keynes (1936, pp. 14–15) argues: [T]hey [workers] do not resist reductions of real wages, which are associated with increases in aggregate employment and leave relative money-wages unchanged, unless the reduction proceeds so far as to threaten a reduction of the real wage below the marginal disutility of the existing volume of employment. Every trade union will put up some resistance to a cut in moneywages, however, small. But since no trade union would dream of striking on every occasion of a rise in the cost of living, they do not raise the obstacle to any increase in aggregate employment which is attributed to them by the classical school.

Simply because nominal wages are sticky in no way implies that real wages are not flexible enough in a world of rational (smart) agents, for employment to be restored to prerecession levels through monetary and fiscal policy. Increased longer-term unemployment need not be a product of workers suddenly shifting their preferences towards more leisure but, rather, of misconstrued macro policy that equates sticky nominal prices (especially wages) with sticky real wages. On a related note, given the empirics and theory underlying x-efficiency theory, even if real wages increase as aggregate demand increases, if this is accompanied by compensating increases in labour productivity (a rational response by economic agents), increasing real wages would not impede the employment of more workers as aggregate demand increases. In this case, increasing real wages will not affect the economic capacity of the firm to hire more workers on the margin. The marginal product of the labour curve shifts to the right as real wages increase (Altman 2006b). Here, too, by assuming rational individuals, we cannot logically deduce that increasing unemployment is a function of workers’ preference for more leisure. Rather, a large reduction of aggregate demand requires a compensating increase in aggregate demand, given that rational or smart workers pose no fundamental obstacle to restoring employment to its pre-recession levels. This x-efficiency perspective strengthens the rational worker approach presented by Keynes in his narrative on workers accepting generalized fair cuts to real wages, given the expectation that employment will increase as a consequence. In this instance, rational inefficiency becomes a product of government not pursuing a policy that restores aggregate demand, in the face of rational decision-making at the firm level. The latter is a product of the belief by government decision-makers in the capacity of markets to self-correct and that the ultimate source of the persistence in the increased level of unemployment following a severe economic downturn is the unwillingness of workers to reduce their real wages. This belief in the classical model might be rational given the information set of decision-makers. However, they yield economic inefficiencies

M4225-ALTMAN_9781782549574_t.indd 37

03/05/2017 08:20

38

Handbook of behavioural economics and smart decision-making

at the macroeconomic level, keeping unemployment rates unnecessarily high and output well below what it might otherwise be.

CONCLUSION A key argument presented in this chapter is that smart individuals (economic agents) can make decisions that are economically inefficient in the realm of production and consumption, and at both the micro and macro level. Being smart and being rational, from this boundedly rational perspective, does preclude outcomes being inefficient and suboptimal. In the conventional wisdom, rational efficiencies are assumed away at the micro level. Rational agents behaving in accordance with the dictates of neoclassical theory should produce results that are both economically efficient in production and utility maximizing, reflecting the true preferences of decision-makers in the firm and the household in both the realm of production and consumption. However, if individuals were to deviate from neoclassical behavioural norms we would expect inefficiencies in both production and consumption, as they would be behaving irrationally at least from the perspective of conventional wisdom. However, the more empirically based smart agent approach redefines rationality more broadly in terms of smart decision-making. This builds upon the contributions of Simon and the bounded rationality/procedural rationality modelling platform that he developed. Here, a rational baseline for decision-making is predicated on the capabilities of the individual and the decision-making environment. This introduces a different set of norms for what is rational and even for what is efficient. Moreover, in the narrative presented in this chapter, economic inefficiencies can flow from rational or smart behaviour in both the realm of production and consumption. Such inefficiencies can be a function of the preferences of decision-makers and the decisionmaking capabilities of smart individuals and their decision-making environment. In fact, even given optimal decision-making capabilities and optimal decision-making environments, inefficiencies can arise given the preferences of smart decision-makers. Economic efficiency cannot be achieved simply by constructing appropriate decision-making capabilities and environments. The latter two serve as the necessary but not sufficient conditions for economic efficiency. Overall, the different approaches to behavioural economics empirically unmask the fact that individuals typically do not behave as predicted and as is normatively preferred by conventional economics. However, the heuristics and biases approach to behavioural economics, which feeds into and overlaps with the nudging approach, regards such deviations from conventional norms as indicators of suboptimal behaviour, typically hardwired into the human brain. From this perspective, modelling choice behaviour requires investigating and documenting deviations from the conventional norms and determining means of inducing decision-makers to behave in accordance with these norms for optimal behaviour to be achieved. Hence, the heuristics and biases approach, although critical of the conventional assumption that individuals behave ‘rationally’, retain the conventional economic benchmarks for rationality. Building upon the evidence, the argument presented in this chapter is that although smart people do not behave in accordance with conventional economic norms, this should not imply that such behaviour and the choices flowing from it are irrational, suboptimal or

M4225-ALTMAN_9781782549574_t.indd 38

03/05/2017 08:20

Rational inefficiency 39 inefficient, given individuals’ capabilities and their decision-making environment. Smart people make boundedly rational decisions. Benchmarks for what is rational, smart and intelligent need to be based upon what makes sense given the decision-makers capabilities and their decision-making environment. There are no specific optimal decision-making norms that apply across time, space and individuals, although there might be general behavioural normative rules of thumb. The approach taken in this chapter, and implicit in Simon’s notion of bounded and procedural rationality, is that individuals can make mistakes and can even be biased, but this is not part of the human condition – hardwired in the human brain. Environmental factors and decision-making capabilities, which can be altered, play a determining role. We can determine the conditions under which optimal decisions can be achieved by individuals, households and firms. Herein lies a critical role for societal (from community to state to international) interventions in economy and society; to facilitate the provision of improved decision-making environments and capabilities. Also, it is important to correct for externalities, positive and negative, many of which are related to information imperfections and coordination failures as well as preferences that, if realized, cause harm to others. Some of the differences and similarities of the different approaches to rationality and their implications for understanding the source and determinants of the relative inefficiencies in production and consumption are illustrated in Figure 2.2. Conventional economics presumes that narrowly defined rationality best explains human behaviour, and yields substantive predictions of production and consumption efficiencies across time and space. Public policy is of limited importance apart from assuring competitive markets and secure property rights. The heuristics and biases approach, while retaining conventional normative benchmarks for optimal behaviour and efficient choice outcomes, documents the persistent deviations from conventional norms. Hence, we have persistent inefficiencies (errors and biases in decision-making), typically a function of behaviours hardwired into the human brain. This yields policy prescriptions designed to nudge individuals towards what experts (choice architects) deem to be in the best interest of the decision-maker. The smart agent approach, building upon the bounded rationality contributions to the decision-making literature, rejects many of the conventional norms for optimal behaviour, while agreeing with the heuristics and biases proponents that humans typically do not behave in accordance with these norms. However, here rationality is defined relative to the capabilities of the decision-makers, the decision-making environment, preferences and power relationships, as well as recognizing differences in these variables across agents and across time and space. In this smart agent modelling, rational agents can make errors and be biased in their decisions, and generate inefficiencies in the domain of production and consumption. But these suboptimal outcomes can be affected by, for example, changes to individual capabilities and the overall decision-making environment. This underlines the significance of public policy in facilitating choices that yield more efficient outcomes while increasing the ability of agents to form and realize their true preferences.

M4225-ALTMAN_9781782549574_t.indd 39

03/05/2017 08:20

40

M4225-ALTMAN_9781782549574_t.indd 40

03/05/2017 08:20

Figure 2.2

Capabilities and environmental based benchmarks

Capabilities decision-making environment

Deviations from neoclassical norms

Rationality and inefficiency

▪ Improve capabilities and decision-making environments ▪ Improve markets and property rights ▪ Internalize externalities

Policy

Extent of production and consumption inefficiencies

Fast and frugal heuristics

Bounded and procedural rationality

Smart agents

▪ Nudging and reframing ▪ Hard and soft paternalism

Policy

Hardwired errors and biases

Neoclassical behavioural benchmarks

Heuristics and biases

Libertarian: ▪ Minimize government Less libertarian: ▪ More competitive markets ▪ Development and protection of property rights ▪ Internalize externalities

Policy

Production and consumption efficiencies

Assumed neoclassical behaviour

Neoclassical rationality

Rational inefficiency 41

REFERENCES Akerlof, G.A. (2002), ‘Behavioral macroeconomics and macroeconomic behavior’, American Economic Review, 92 (3), 411–33. Akerlof, G.A. and R.E. Kranton (2010), Identity Economics: How Our Identities Shape Our Work, Wages, and Well-Being, Princeton, NJ: Princeton University Press. Akerlof, G.A. and R.J. Shiller (2009), Animal Spirits: How Human Psychology Drives the Economy, and Why It Matters for Global Capitalism, Princeton, NJ: Princeton University Press. Alchian, A.A. (1950), ‘Uncertainty, evolution, and economic theory’, Journal of Political Economy, 58 (3), 211–21. Altman, M. (1999), ‘The methodology of economics and the survivor principle revisited and revised: some welfare and public policy implications of modeling the economic agent’, Review of Social Economy, 57 (4), 427–49. Altman, M. (2002), ‘Economic theory, public policy and the challenge of innovative work practices’, Economic and Industrial Democracy: An International Journal, 23 (2), 271–90. Altman, M. (2005), ‘Behavioral economics, power, rational inefficiencies, fuzzy sets, and public policy’, Journal of Economic Issues, 39 (3), 683–706. Altman, M. (2006a), ‘What a difference an assumption makes: effort discretion, economic theory, and public policy’, in M. Altman (ed.), Handbook of Contemporary Behavioral Economics: Foundations and Developments, Armonk, NY: M.E. Sharpe, pp. 125–64. Altman, M. (2006b), ‘Involuntary unemployment, macroeconomic policy, and a behavioral model of the firm: why high real wages need not cause high unemployment’, Research in Economics, 60 (2), 97–111. Altman, M. (2009), ‘A behavioral-institutional model of endogenous growth and induced technical change’, Journal of Economic Issues, 63 (1), 685–713. Altman, M. (2010), ‘A behavioral and institutional foundation of preference and choice behavior: freedom to choose and choice x-inefficiencies’, Review of Social Economy, 68 (4), 395–411. Altman, M. (2011), ‘Behavioural economics, ethics, and public policy: paving the road to freedom or serfdom?’, in J. Boston (ed.), Ethics and Public Policy: Contemporary Issues, Wellington: Victoria University Press, pp. 23–48. Altman, M. (2012), Behavioral Economics for Dummies, Mississauga, Ontario: Wiley. Altman, M. (2015), ‘Introduction’, in M. Altman (ed.), Real-World Decision Making: An Encyclopedia of Behavioral Economics, Santa Barbara, CA: Greenwood, ABC-CLIO, pp. xv–xxxi. Altman, M. (2016), ‘Multiple equilibria, bounded rationality, and the indeterminacy of economic outcomes: closing the system with institutional parameters’, in R. Frantz and L. Roger (eds), Minds, Models and Milieux: Commemorating the Centennial of the Birth of Herbert Simon, Basingstoke: Palgrave Macmillan, pp. 167–85. Altman, M. (2017), ‘A bounded rationality assessment of the new behavioral economics’, in R. Frantz, S.-H. Chen, K. Dopfer, F. Heukelom and S. Mousavi (eds), Routledge Handbook of Behavioral Economics, New York: Routledge, pp. 179–94. Becker, G.S. (1996), Accounting for Tastes, Cambridge, MA: Harvard University Press. Berg, N. (2014), ‘The consistency and ecological rationality approaches to normative bounded rationality’, Journal of Economic Methodology, 21 (4), 375–95. Berg, N. and G. Gigerenzer (2010), ‘As-if behavioral economics: neoclassical economics in disguise?’, History of Economic Ideas, 18 (1), 133–66. Bewley, T.F. (1999), Why Wages Don’t Fall During a Recession, Cambridge, MA, and London: Harvard University Press. Cyert, R.M. and J.C. March (1963), A Behavioral Theory of the Firm, Englewood Cliffs, NJ: Prentice-Hall. Frantz, R.S. (1997), X-Efficiency Theory, Evidence and Applications, Boston, MA, Dordrecht and London: Kluwer Academic. Friedman, M. (1953), ‘The methodology of positive economics’, in M. Friedman (ed.), Essays in Positive Economics, Chicago, IL: University of Chicago Press, pp. 3–43. Friedman, M. (1968), ‘The role of monetary policy’, American Economic Review, 58 (1), 1–17. Gigerenzer, G. (2007), Gut Feelings: The Intelligence of the Unconscious, New York: Viking. Hayek, F.A. (1944), The Road to Serfdom, Chicago, IL: University of Chicago Press. Hayek, F.A. (1945), ‘The use of knowledge in society’, American Economic Review, 35 (4), 519–30. Hayek, F.A. (1948), Individualism and the Economic Order, Chicago, IL: University of Chicago Press. Harsanyi, J. (1982), ‘Morality and the theory of rational behavior’, in A. Sen and B. Williams (eds), Utilitarianism and Beyond, Cambridge: Cambridge University Press, pp. 39–62. Hume, D. (1738), A Treatise on Human Nature, London, reprinted 2014, Some Good Press Kindle file. Kahneman, D. (2003), ‘Maps of bounded rationality: psychology for behavioral economics’, American Economic Review, 93 (5), 1449–75.

M4225-ALTMAN_9781782549574_t.indd 41

03/05/2017 08:20

42

Handbook of behavioural economics and smart decision-making

Kahneman, D. (2011), Thinking, Fast and Slow, New York: Farrar, Straus and Giroux. Kahneman, D. and A. Tversky (1979), ‘Prospect theory: an analysis of decision under risk’, Econometrica, 47 (2), 263–91. Keynes, J.M. (1936), The General Theory of Employment, Interest, and Money, New York: Harcourt, Brace and Company. Leibenstein, H. (1957), Economic Backwardness and Economic Growth, New York: John Wiley and Sons. Leibenstein, H. (1966), ‘Allocative efficiency vs “x-efficiency”’, American Economic Review, 56 (3), 392–415. Leibenstein, H. (1979), ‘A branch of economics is missing: micro-micro theory’, Journal of Economic Literature, 17 (2), 477–502. March, J.G. (1978), ‘Bounded rationality, ambiguity, and the engineering of choice’, Bell Journal of Economics, 9 (2), 587–608. North, D.C. (1991), ‘Institutions’, Journal of Economic Perspectives, 5 (1), 97–112. North, D.C. (1994), ‘Economic performance through time’, American Economic Review, 84 (3), 359–68. Nussbaum, M. (2011), Creating Capabilities: The Human Development Approach, Cambridge, MA: Harvard University Press. Nussbaum, M.C. (1999), Sex and Social Justice, Oxford and New York: Oxford University Press. Olson, M. (1996), ‘Distinguished lecture on economics in government: big bills left on the sidewalk: why some nations are rich, and others poor’, Journal of Economic Perspectives, 10 (2), 3–24. Posner, R.A. (2009), A Failure of Capitalism: The Crisis of ’08 and the Descent into Depression, Cambridge, MA, and London: Harvard University Press. Sen, A. (1985), Commodities and Capabilities, Amsterdam: North-Holland. Shiller, R.J. (2008), The Subprime Solution: How Today’s Global Financial Crisis Happened, and What to Do about It, Princeton, NJ: Princeton University Press. Shiller, R.J. (2012), Finance and the Good Society, Princeton, NJ: Princeton University Press. Simon, H.A. (1959), ‘Theories of decision making in economics and behavioral science’, American Economic Review, 49 (3), 252–83. Simon, H.A. (1978), ‘Rationality as process and as product of thought’, American Economic Review, 68 (2), 1–16. Simon, H.A. (1986), ‘Rationality in psychology and economics’, Journal of Business, 59 (4), S209–24. Simon, H.A. (1987), ‘Behavioral economics’, in J. Eatwell, M. Millgate and P. Newman (eds), The New Palgrave: A Dictionary of Economics, London: Macmillan, pp. 221–5. Smith, V.L. (2003), ‘Constructivist and ecological rationality in economics’, American Economic Review, 93 (3), 465–508. Smith, V.L. (2005), ‘Behavioral economics research and the foundations of economics’, Journal of SocioEconomics, 34 (2), 135–50. Thaler, R.H. and C. Sunstein (2008), Nudge: Improving Decisions about Health, Wealth, and Happiness, New Haven, CT, and London: Yale University Press. Todd, P.M. and G. Gigerenzer (2003), ‘Bounding rationality to the world’, Journal of Economic Psychology, 24 (2), 143–65. Tversky, A. and D. Kahneman (1981), ‘The framing of decisions and the psychology of choice’, Science, 211 (4481), 453–58.

M4225-ALTMAN_9781782549574_t.indd 42

03/05/2017 08:20

3

Rational mistakes that make us smart Nathan Berg

[E]very intelligent system makes good errors; otherwise it would not be intelligent. The reason is that the outside world is uncertain, and the system has to make intelligent inferences based on assumed ecological structures. Going beyond the information given by making inferences will produce systematic errors. Not making these errors would destroy intelligence. (Gerd Gigerenzer 2005, p. 199)

I describe theoretical and empirical examples of errors – both in games against nature and in strategic settings – that confer individual-level and, in some cases, Pareto-improving benefits to an entire economy or social system. My goal is to demonstrate the wide range of mechanisms by which we individually and collectively benefit from behaviors that many behavioral economists have been too quick to label as mistakes, simply because those behaviors do not conform to the orthodox rational-choice standard of rationality based on internal logical consistency. I want to invite you to reconsider the interpretations that can and should be attached to behavioral patterns that are commonly described by many behavioral economists as decision-making errors. I argue that mistakes are vital for strengthening and maintaining valuable relationships and enabling perceptual, inferential, social and financial success. Making mistakes ex ante, that is, as part of our planned behavior, is an expected and regular characteristic of smart behavior. Smart people must (that is, descriptively, as a logical consequence of the requirements of success) and should (that is, prescriptively) make mistakes. How could success require mistakes? Are such notions of ‘mistakes’ merely a semantic parlor trick that disappears once proper definitions of success are introduced? Does the claim that smart people must make mistakes sound like a bad joke? In fact, the mistake of telling a ‘bad joke’ (that is, an ill-chosen attempt at humor that unintendedly winds up annoying or offending someone you care about and want to feel good) illustrates the point precisely that smart people must err (cf., the examples and arguments in Gigerenzer 2005).

BAD JOKE Consider what happens if I tell my girlfriend a story that I heard from an entertaining and rough-talking (read ‘severely politically incorrect’) friend of mine. It turns out that my girlfriend does not like the joke and finds it deeply offensive. Normally, she would update her beliefs about any person (for example, me) heard to have spoken those specific words aloud. My girlfriend might, in fact, be interpreted here as a Bayesian updater whose subjective belief that I am a high-quality, worthy person (after conditioning on the historical sequence of speech acts by me that she has observed). Normally, her conditional assessment of my value would decline sharply, conditional on my telling of the bad joke. The joke was so bad that, conditional on observing me tell it, her updating function abruptly 43

M4225-ALTMAN_9781782549574_t.indd 43

03/05/2017 08:20

44

Handbook of behavioural economics and smart decision-making

downgrades the level of subjective esteem associated with the speaker (given the finite number of virtuous acts that were previously observed in our shared history). I am assuming that she also uses a threshold condition to accept men worthy of dating (that is, satisficing on mate choice while updating beliefs according to Bayes’ law). Her updated level of esteem for me, conditional on the bad joke, now falls strictly below her minimum threshold required for mate acceptance. She would normally reject me out of hand as a partner based on her usual belief-updating system. The bad joke I told would therefore normally exclude me from her consideration set and lead to a breakup of the relationship already in progress. What should I expect comes next? Instead of breaking up, she decides to forgive me. She says, ‘I didn’t like the words I heard you say, but I forgive you. Please don’t say it again.’ Just like that, something outside the normative performance metrics introduced in the model so far newly enters the analysis. Like bones that heal stronger than in their previously unbroken state or an immune system that heals stronger, apparently relationships, too, can grow deeper, richer, more valuable and stronger – in love, business, science, and friendships of many kinds – thanks to the event of a mistake followed by forgiveness (or other means of relationship repair). I consider several modeling strategies with the goal of representing the mechanism of relationships whose depth or robustness benefits from mistakes and shared adversity in their shared history. Among my reasons for raising this example are the subtleties it raises regarding the methodological conventions of constrained optimization and game-theoretic reasoning that behavioral economists typically use as the benchmark of perfect rationality in relation to which deviations are thought to measure irrationality, a-rationality, or various normative gradations of irrationalities (Berg 2003, 2014a; Berg and Gigerenzer 2006, 2010). Next I will present contrasting representations of this interaction corresponding to different views of other players’ action sets, whether those action sets include the possibility of intentionally versus randomly bad jokes (as a result of ‘nature’s move’); and whether the continuation value of the relationship itself is included explicitly, possibly strengthened as a result of withstanding a threatening event and then recovering thanks to forgiveness and repair.

SMALL WORLD WITH NO INTENTIONAL TELLING OF BAD JOKES AND NO INDIVIDUAL CONTROL OVER PROBABILITY, P Figure 3.1 shows a simple, small world with no possibility of intentionally telling a bad joke. Bad jokes are modeled in Figure 3.1 as a random event detached from any other variable under the joke teller’s control such as effort. Note, too, that there is no explicit model of risk preferences, cautionary motives, pro-social affections or anti-social motives such as spite. The only decision variable that the joke teller, referred to throughout as Agent 1, has to make in Figure 3.1 is whether to tell a joke or not. The second player whose payoffs are represented in Figure 3.1 is Agent 2, the receiver of Agent 1’s joke (for example, my girlfriend in the discussion above). Without loss of generality, both players’ payoffs are normalized to zero at the left-most no-joke node, so that the payoffs associated with the other two terminal nodes, corresponding to bad-

M4225-ALTMAN_9781782549574_t.indd 44

03/05/2017 08:20

Rational mistakes that make us smart 45 1

Joke

No joke Nature

Prob (bad joke) = p (b1, b2)

(0, 0)

Figure 3.1

Prob (good joke) = 1 – p (g1, g2)

Small-world event tree with no possibility of intentionally telling a bad or offensive joke, with bad jokes occurring as an act of nature with probability p, 0 0. Assuming from now on that b1 < 0 and b2 < 0, should we then agree with conventional wisdom in behavioral economics that the bad-joke outcome is always best avoided if possible? In Figure 3.1, there is nothing in either player’s choice set that enables him or her to control the probability of the bad-joke outcome. However, if there were, would rationality then trivially require (that is, by definition) that agents avoid making the mistake of telling a bad joke? The next model gives Agent 1 clairvoyance to focus on the question of whether he would ever rationally choose a bad joke he has perfect control to avoid.

M4225-ALTMAN_9781782549574_t.indd 45

03/05/2017 08:20

46

Handbook of behavioural economics and smart decision-making

AGENT 1 HAS CLAIRVOYANCE AND THEREFORE PERFECT ABILITY TO AVOID TELLING THE BAD JOKE In Figure 3.2, Agent 1 is clairvoyant. An equivalent assumption is that nature moves first and in a manner that is visible to both players, which determines the quality of the joke before Agent 1 decides whether to tell it. Therefore, Agent 1 knows in advance if the joke will land in Agent 2’s ears as a bad or good joke. An own-payoff-maximizing Agent 1 will never choose to tell a bad joke in the model depicted in Figure 3.2. If a bad joke occurs, then, because the joke teller is clairvoyant, Agent 2 knows that Agent 1 actually intended harm or offence. The bad-joke outcome in Figure 3.2 can never be accidental. Therefore, the possibility of spite or malevolence is now an unavoidable consideration for Agent 2 upon observing the bad joke. The unnatural abstraction of the payoffs from the context of the agents’ relationship shows up starkly and reflects a razor-edge view of what can be rational in Figures 3.1 and 3.2. The missing context of relationship remains in the next representation, which returns to the setup in Figure 3.1 but endows Agent 1 (in Figure 3.3) with the capacity to choose cautionary effort x in a way that reduces the bad-joke probability p(x). Nature

Prob (bad joke) = p

Prob (good joke) = 1 – p 1

1

Figure 3.2

No joke

Joke

No joke

Joke

(0, 0)

(b1, b2)

(0, 0)

(g1, g2)

Nature moves first (deciding whether Agent 1’s joke will turn out to be good or bad) or, equivalently, Agent 1 is clairvoyant

1 No joke

(0, 0)

Figure 3.3

x1 Joke x2 Agent 1 chooses bad-jokeavoidance effort x xi xn Nature Prob (bad joke) = p(x) Prob (good joke) = 1 – p(x) (b1, b2)

(g1, g2)

Joke teller chooses a continuously-valued cautionary effort variable, x P [0, ∞), such that the bad-joke probability is p(x) 5 e–ax

M4225-ALTMAN_9781782549574_t.indd 46

03/05/2017 08:20

Rational mistakes that make us smart 47

JOKE TELLER CHOOSES CAUTIONARY EFFORT X SUCH THAT BAD-JOKE PROBABILITY P(X) IS DECREASING IN X Figure 3.3 is an extension of Figure 3.1 (returning from now on to the original assumption of no clairvoyance) in which the joke teller is endowed with a continuously-valued cautionary effort variable, x P [0, ∞), that effectively reduces the probability of a bad joke. Therefore, p(x) is assumed to be a decreasing function of x. For simplicity, the specific functional form p(x) 5 e-ax is introduced as a specific case drawn from the more general family of decreasing (that is, controllable) bad-joke risk functions. Assuming that the unit cost of cautionary effort is measured by parameter k, then Agent 1’s expected payoff objective (conditional on telling a joke) can be written as follows: p(x) 5 p(x)b1 + (1 − p(x))g1 − kx,

(3.2)

which Agent 1 seeks to maximize by choosing x such that telling a joke is better than no joke (that is, p > 0) and that p(x) 5 e-ax. The constrained maximization problem just described has a global maximum at x* 5 [ln(a) + ln(g1 − b1) − ln(k)]/a (in the dense subset of the parameter space of b1, g1 and k satisfying the conditions that x* > 0 and p(x*) > 0). In the models of Figure 3.1 and Figure 3.3, we have not considered Agent 1’s reasoning about Agent 2’s conditions for continuing the relationship or any inferences she makes about Agent 1’s intentionality. The interactions so far represented are one-offs unless the payoff parameters are interpreted as depending on both agents’ valuations of continuing the relationship in future rounds of interaction. Before proceeding fully toward the fundamental issue of modeling the intentionality of the joke teller, I now model Agent 2’s decision to either break off the relationship (that is, not continue) versus continue. Introducing Agent 2’s continuation decision turns out to be enough to generate the possibility of the relationship increasing in value following forgiveness in the bad-joke outcome.

AGENT 2 CHOOSES ‘NO’ OR ‘YES’ TO CONTINUING THE RELATIONSHIP Figure 3.4 shows an extension of the basic model in Figure 3.1 (without a cautionary effort choice x that influences bad-joke probability p), which now includes the possibility of forgiveness and relationship repair in addition to the possibility that Agent 2 chooses to end the relationship by choosing ‘no’. New notation introduced in Figure 3.4 includes each agent’s valuation of the relationship itself, ex payoffs from the joke-telling interaction. The agents’ continuation values whenever Agent 2 chooses ‘yes’ to continue are denoted r1 and r2, respectively. Down the event branch in which a bad-joke outcome occurs and Agent 2 decides ‘yes’ to continue nevertheless, then both agents’ valuations of their relationship increase to R1 and R2, respectively. In keeping with the previous three figures where payoffs represent changes relative to the no-joke state, the continuation value of the relationship does not show up in this representation along the continuation path but instead as a lost continuation value whenever Agent 2 chooses ‘no’.

M4225-ALTMAN_9781782549574_t.indd 47

03/05/2017 08:20

48

Handbook of behavioural economics and smart decision-making Value to R1 > r1 and R2 > r2, respectively 1 No joke

Joke

Nature

Prob (bad joke) = p

Prob (good joke) = 1 – p

2

2

2 No

(–r1, –r2)

Figure 3.4

Yes

(0, 0)

No

Yes

No

Yes

(b1 – r1, b2 – r2) (b1 + R1 – r1, b2 + R2 – r2) (g1 – r1, g2 – r2)

(g1, g2)

Same as Figure 3.1, but Agent 2 now chooses whether to continue the relationship (‘no’ not continue or ‘yes’ continue), depending on whether her relationship valuation r2 remains positive, with forgiveness of bad jokes having the effect of increasing both agents’ relationship

The loss of the relationship value shows up at nodes where Agent 2 chooses not to continue. At the left-most node, for example, the payoffs are now written as (−r1, −r2) if Agent 2 chooses to not continue and (0, 0) if Agent 2 chooses to continue following the no-joke outcome. If, for whatever reason, Agent 2 perceives negative continuation value, then it is rational for her to discontinue because discarding the negative value achieves a positive payoff relative to continuation (r2 < 0 implies −r2 > 0). At the two nodes following the good-joke branch, the relationship value is assumed not to change and therefore does not show up in the payoffs (g1, g2) but is instead deducted from the payoffs as the lost value of continuing the relationship in the payoffs (g1 − r1, g2 − r2). Presumably, Agent 2 never has any reason to rationally choose ‘no’ following a good-joke outcome so long as r2 > 0. The real innovation and main point of focus in the analysis of payoffs in Figure 3.4 are those that follow the bad-joke outcome and Agent 2’s decision ‘yes’ continuing with the relationship nevertheless. In this case, in addition to the bad-joke payoffs (b1, b2), each agent sees something new about their respective assessments of the value of continuing the relationship. The changes in their relationship values, R1 − r1 and R2 − r2, respectively, are added to the bad-joke payoffs. Note here there is a distinct new possibility that the agents who are most well-off are those who endured the bad joke and recovered from it – or have simply chosen to continue the relationship, thereby revealing to each other greater continuation values than would otherwise have been observable to either player without the failure or mistake. The performance advantage referred to here as ‘most well-off’ could be interpreted as individuals who enjoy the strongest, most durable relationships founded on joint awareness that they are mutually valuable to each other – enough so to withstand a large range of negative-payoff events and nevertheless retain positive relationship value. The ideas here draw on the pioneering work by Rapoport and Chammah (1965), Rapoport (1984) and Axelrod (1984).

M4225-ALTMAN_9781782549574_t.indd 48

03/05/2017 08:20

Rational mistakes that make us smart 49 Result 1 In the interaction represented in Figure 3.4, if Ri − ri > bi − gi, then Agent 1 is better off after a bad joke is mistakenly told and Agent 2 chooses ‘yes’ to continue the relationship than Agent 1 would have been without Agent 1 having made the mistake. It follows from Result 1 that, if the inequality holds for both agents, then the mistake causes a Pareto improvement (that is, mistakes can unambiguously increase the size of the economic pie by revealing otherwise latent information about the strength of social ties). The possibility that mistakes can strengthen social ties through such a transformatively positive (that is, relationship-strengthening) act of forgiveness modeled along the bad-joke-‘yes’ path in Figure 3.4 brings with it profound implications. Note, too, that instead of forgiveness, the transformative event that occurs can be interpreted as information revelation – that the mistake simply reveals otherwise latent (that is, unobservable) information about others’ subjective valuations of their relationships with us. From this observation, a large set of new mechanisms that map mistakes into aggregate-value expanding outcomes emerges. For example, if I fail to deliver on a contractual commitment to a key business partner in a repeated interaction, and that partner expresses understanding and agrees to continue even though I see that my mistake imposed large costs on the partner, then I may be willing to take joint risks with that partner that I would not have otherwise. The reason for the shift in willingness to undertake value-generating risk may be that the process of dealing with my past failure and the hardships it caused both of us transformed the relationship or focused our attention on jointly observing the value of our collaboration. Or it could have simply revealed otherwise unobservable information about my partner’s willingness to endure joint losses and remain committed to continuing together, which, in turn, triggers my own willingness to take on new projects where our joint actions expose each other to new risks. I want to make the case that the interaction in Figure 3.4 and its scope for generating welfare-enhancing mistakes can be rather broadly interpreted. Telling a bad joke; failing to deliver on a contractual obligation; or being very late in delivering a promised book chapter to an editor whom I respect greatly, whose friendship is dear to me and whose book project I feel great passion for – these examples are illustrative of smart people’s rational mistakes. I discuss additional examples below. Before considering more examples, I want to begin addressing the as yet unexamined question of intentionality and why the mechanism of welfare-enhancing mistakes generally breaks down if this mechanism is deliberately exploited.

MODEL IN WHICH AGENT 1 CAN CHOOSE TO DELIBERATELY TELL A BAD JOKE If a prototypical Agent 1 looks at the payoffs in Figure 3.4 and perceives that the highest possible outcome is indeed the path along which a bad joke occurs and Agent 2 forgives, then could Agent 1 rationally pursue this outcome as his goal? Certainly if Agent 2 knew that Agent 1 were hurting her intentionally in order to coax her into forgiving and revealing her high latent value for continuing with him, then she would likely modify her assessed continuation value of the relationship downward. If Agent 1 is not sociopathic,

M4225-ALTMAN_9781782549574_t.indd 49

03/05/2017 08:20

50

Handbook of behavioural economics and smart decision-making

then he likely experiences some guilt (denoted g > 0, representing Agent 1’s psychic cost of deliberately telling the bad joke or otherwise intentionally hurting Agent 2). To represent intentionality, I introduce the notation w P{B, G} to code Agent 1’s intention to tell a bad or good joke. Figure 3.5 depicts, by now, the rather elaborate joke-telling interaction featuring two distinct bad-joke branches which correspond to intentional versus accidental bad jokes. The dotted oval represents Agent 2’s uncertainty: when she observes the bad-joke outcome, she does not know whether Agent 1’s intention was to tell a bad one or not – intentionally offending and hurting her, or, alternatively, intending to tell a good joke that led accidentally to causing offense or hurt. Because Agent 2’s valuations of continuing the relationship now depend on, and vary with, intentionality type w (through the functions r(w) and R(w)) while holding constant the bad-joke outcome, the model therefore becomes non-consequentialist. That is, agents’ subjective rankings depend not only on the final outcome but also on the process that led to that outcome. Figure 3.5 expresses payoffs corresponding to each of the two bad-joke outcomes that differ only in Agent 1’s intention w. But Agent 2 is not clairvoyant and does not know Agent 1’s intention (or intentionality type w) with perfect certainty. In the second dotted oval below the main payoff nodes, Figure 3.5 also provides Agent 2’s expected payoffs (in her state of uncertainty about w), which depend on Agent 2’s probabilistic belief b that Agent 1 is a bad-intention type. The function R(w) represents Agent 2’s assessment of the value of continuing the relationship with Agent 1 following a bad-joke outcome as a function of Agent 1’s type. The function r(w) represents Agent 2’s assessment of the value of continuing the relationship with Agent 1 along any other discontinuation or continuation path that does not involve the bad-joke outcome and forgiveness. Note that when Agent 1 is a bad-intention type, both continuation values are assumed to take on very negative and nearly equal payoff values: R(B) 5 r(B) r(G) > 0 and R(G) − r(G) > 0 (when Agent 1 is a good-intention type). In Agent 2’s state of uncertainty about Agent 1’s type, w, and having observed the bad-joke outcome, we can compute the difference between Agent 2’s expected payoff from choosing ‘yes’ (to continue) minus her expected payoff from choosing ‘no’ (to not continue), which I denote Dyes−no|bad-joke: Dyes−no|bad-joke 5 (1 − b)R(G) + br(B).

(3.3)

Agent 2 (assumed to be an expected payoff maximizer with belief b that measures her subjective probability that 1’s type is bad) chooses to continue if, and only if, Dyes−no|bad-joke ≥ 0 (assuming continuation whenever ‘no’ and ‘yes’ have equal expected payoffs) and discontinue otherwise. That is, Agent 2’s continuation decision in the face of being exposed to the bad joke and uncertainty about Agent 1’s intentionality type turns on the upwardly revised relationship value and Agent 1’s intentionality type being good, R(G), weighted by 2’s belief that 1 is in fact a good type − and then comparing this positive expected continuation value to the negative expected value if 1 were a bad type, r(B), weighted by 2’s belief that 1 is in fact a bad type. Under what circumstances will 1 deliberately cause 2 harm under the expectation that 2 will forgive and continue, thereby yielding a greater player-1 payoff than by trying to tell a good joke: b1 + R1 − r1 − g > p(b1 + R1 − r1) + (1 − p)g1? Recall that 1 effectively chooses his intentionality type w. Figure 3.5 assumes that, if Agent 1 chooses w 5 B, then the

M4225-ALTMAN_9781782549574_t.indd 50

03/05/2017 08:20

51

M4225-ALTMAN_9781782549574_t.indd 51

03/05/2017 08:20

Figure 3.5

(–r1, –r2)

No

Yes

(b1 + R1 – r1 – γ, b2)

Yes

(b1 – r1, b2 – r(G))

No

2

b2 – r(B) – (1 – )r(G)

No

b2 + (1 – )(R(G) – r(G))

Yes

Yes

Nature

2

No

Yes

(g1 – r1, g2 – r(G))

Prob (good joke) = 1 – p

(b1 + R1 – r1, b2 + R(G) – r(G))

Prob (bad joke) = p

Intend to tell good joke ( = G)

Agent 2’s expected payoffs after observing a bad joke but facing uncertainty about Agent 1’s intentionality type with belief Prob( = B) =

(b1 – r1 – γ, b2 – r(B))

No

2

Agent 1 can deliberately tell a bad joke

(0, 0)

2

bad joke

Intend to tell

No joke

1

(g1, g2)

52

Handbook of behavioural economics and smart decision-making

joke will turn out bad with probability 1. If 1 chooses w 5 G, however, then we are back in the non-degenerate probabilistic world where p measures the probability of the badjoke outcome. A further condition is required if 1 is to believe that 2 will indeed choose to continue. That is, 1 must believe that Dyes−no|bad-joke ≥ 0. The issue arises as to whether intentions are ever conclusively observable. The model here reflects an attempt to model real-world scenarios where distinct sets of intentions are in fact observable. If an agent deliberately misrepresents his or her intentions in a repeated interaction setting, then the modeling exercise here rests on the assumption that such misrepresentation is eventually discoverable (for example, my partner overhearing me tell my friend that I deliberately told a bad joke or showed up late to test the extent of her forgiveness). Agent 2’s preferences are non-consequentialist, a fact made explicit through the dependence of the functions R(w) and r(w) on w. Agent 2’s view of her own payoffs is not invariant with respect to w holding the bad-joke outcome fixed. Agent 1’s intention – to deliberately tell the bad joke (w 5 B) or to at least try to tell a good joke (w 5 G) – matters quite explicitly to Agent 2. The next section applies a slightly different interpretation to the payoff schemes in the figures above to illustrate the general nature of the phenomenon described above in the figures and Result 1, namely, that individually and collectively welfare-improving mistakes are commonplace and broadly distributed throughout the decision environments people face. The integrity of the mistakes, as distinguished in Figure 3.5, matters. There are honest mistakes and fraudulent or deliberate mistakes. Smart agents should, in general, be adept at detecting fraudulent mistakes, although doing so is not necessarily easy in practice. The broader issue is that errors can be welfare improving when they elicit new information or provide opportunities for others to reveal more about the objectives they are pursuing.

YOU’RE LATE! One way I can learn how much you value my work, or my contribution to a joint venture, or our relationship, is by observing your willingness to forgive. I sometimes show up late (or as it turns out, deliver work or other outputs much later than originally promised, thereby testing the patience – completely unintentionally – of dear colleagues, people about whom I care deeply and hold in truly great esteem, for example, by delivering late a chapter for an edited volume on Rational Decision-Making within the Bounds of Reason). In response to my lateness, some colleagues may classify me as unreliable and choose not to engage with me on future projects; others may classify me as (once again) unreliable but nevertheless choose to continue engaging with me, effectively revealing that they forgive me for being late, or that they value my contributions highly enough to offset the substantial costs I unintentionally imposed on them with my lateness, regardless of whether forgiveness is formally expressed. In game-theoretic terms, the act of my colleague effectively forgiving my lateness sends an important signal about their implicit valuation of engaging with me relative to the respective costs that my lateness imposed on them (again, completely unintentionally on my part). Why do I emphasize my a priori intention to not be late followed by ex post lateness (that is, unintended lateness)? Consider re-labeling ‘bad joke’ outcomes in Figures 3.1–3.5

M4225-ALTMAN_9781782549574_t.indd 52

03/05/2017 08:20

Rational mistakes that make us smart 53 with ‘late’. The interactions represented in Figures 3.4 and 3.5, for example, can then be reinterpreted as: as long as I am not late, then my payoffs and those of my colleagues correspond to good-joke payoffs, (g1, g2), which represents the normal state of affairs based on the productivity of our relationship with no change in trust, no disappointments and no forgiveness. However, as soon as I violate my colleague’s expectation, the colleague’s decision about continuing can then be interpreted as a signal of forgiveness (or otherwise revealing additional mutual value in continuing). Then something new happens. There is an objective loss to both players: bi < gi for i 5 1, 2. My colleague bears the cost of my lateness equal to g2 − b2. I pay a cost g1 − b1 based on embarrassment, loss of reputation for punctuality, and perhaps stress over future opportunities now at risk. However, our aggregate payoffs are now mutually recognized as being greater – for both of us – if we continue, thanks to the synergistic interaction of both individuals (assuming the condition in Result 1 holds). Acknowledging that these costs of lateness can be, and often are, substantial, what then justifies Result 1 and its possibility of greater payoffs – for both of us – corresponding to the action profile of (late, forgive) with associated payoffs of (b1 + R1 − r1, b2 + R2(G) − r2(G)) > (g1, g2)? The answer must be the existence of an offsetting or compensating deepening of the value of our working relationship, where a signal has now been transmitted showing the intention to collaborate cooperatively into the future within a larger-than-expected space of perturbations in the form of missed expectations of various kinds. Another possibility is more direct: the incremental increase in well-being that follows from an expression of (relatively) unconditional acceptance. What have I learned by being seriously late and then receiving implicit forgiveness? I may have learned that my colleague enjoys interacting with me or benefits to a sufficient degree that he or she is willing to incur higher costs than I had perhaps previously realized to keep the working relationship alive. Given the benefit discovered by lateness and subsequent forgiveness, might I then pursue intentional lateness as a mechanism to force colleagues to reveal signals about their willingness to forgive transgressions and maintain working relationships? No. None of this works if lateness (or the bad joke, or any other setback, mistake, disappointment or missed expectation) is intentional. Suppose I am considering a sequence of lateness decisions, coded as binary for simplicity, with a new person in my life with whom a potentially valuable relationship might unfold. I would like to know how this other person regards me or, more crassly, assesses the potential value of our relationship. In other words, I have positive willingness to pay for a costly-to-fake signal of affection, esteem or some form of perceived value from continued engagement. The other person would also like such a signal from me. Could I test the other person by deliberately being late, or deliberately telling an offensive joke, for example, to get a live observation of the other person’s willingness to forgive? Intentional mistakes are no longer mistakes, however, and this strategy is unlikely to work. Problems include the high risk of being discovered and my guilt or embarrassment (g), not to mention the new risk of being discovered, the possibility of an extremely negative payoff that the other person would perceive if I were outed as a perpetrator of intentional lateness, deliberately offensive joke telling or some equivalently dis-pleasurable breach of the other person’s expectations. In case of rare sociopath types who apparently feel no remorse (that is, g 5 0), the model should be interpreted as representing a world

M4225-ALTMAN_9781782549574_t.indd 53

03/05/2017 08:20

54

Handbook of behavioural economics and smart decision-making

where others engaged in repeated interaction should eventually be able to see the sociopath’s intention to deceive, leading non-sociopaths to break off the relationship and look for a normal human being with whom to continue interacting.

GAMES AGAINST (OR IN ACCORDANCE, COOPERATION, HARMONY WITH) NATURE Gigerenzer (2005) discusses physicist Feynman’s arguments in favor of violating invariance with respect to logically equivalent re-descriptions of the same problem. Feynman sought out scientifically useful framing effects by which different intuitions about the laws governing a set of variables became more readily apparent using different frames or logically equivalent re-descriptions. He wrote that these logically equivalent re-descriptions are valuable because ‘psychologically they are different’ (quoted in Gigerenzer 2005, p. 207). In contrast, behavioral economists largely adopt the opposite normative view: that framing effects and other patterns of making different inferences or taking different actions in response to logically identical re-descriptions of the ‘same’ decision problem constitute evidence of irrationality. Gigerenzer argues that the mind’s perceptual system similarly makes smart bets; the intelligence of those bets depends necessarily on making mistakes. For example, in making three-dimensional inferences based on two-dimensional visual input, the mind bets that there is only one source of light that is located above, implying that objects with dark shading below are likely to be ‘sticking out’ toward the observer. From this, Gigerenzer observes that the perceptual system correctly assumes that the world (that is, its threedimensional structure) is fundamentally uncertain (that is, we face the challenge of missing information about three-dimensional structure in our environments) and therefore use associational rules to make reasonable guesses. If instead the mind proceeded as an agnostic Bayesian and waited for irrefutable evidence before logically deducing the correct three-dimensional structure, it would be paralyzed. Similarly, if it had access to a veridical copy of all information required to produce an objectively accurate model of all relevant detail in its environment, the mind and its perceptual system would be overwhelmed. The functionality of the simple bivariate-association rule, ‘objects with dark shading below are sticking out toward me’, depends on its partiality and imperfection with respect to veridical descriptive accuracy. A hypothetically perfect (that is, veridically accurate) perceptual mechanism would still be too little, leaving perceptual holes when facing new or unknown environments (that is, situations where a quick action based on a snap perceptual bet is required without inputting the vast amount of information that a perfect perceptual mechanism would require). This perfectly veridical perceptual representation of the world would also be overwhelmingly too much, presenting the mind with paralysingly large volumes of spatial information. There are different representations of the truth with varying degrees of detail. Representations with more information, even when the additional information is perfectly valid, may induce strictly inferior judgments and decisions compared to less complete representations which enable minds to make quicker and more accurate inferences. The same goes for memory. Is more better? Not necessarily (for example, Schooler and Hertwig 2005, show that forgetting is beneficial in inference tasks). Gigerenzer

M4225-ALTMAN_9781782549574_t.indd 54

03/05/2017 08:20

Rational mistakes that make us smart 55 (2005) discusses individuals with unusually large recall memory that suffer, as a result of their special pneumonic endowment, with acute inability on tasks requiring abstraction. Perhaps having more recall memory means less practice at efficiently coding the gist of what is taking place, abstracting and forming equivalence classes in memory. Are larger consideration sets better than smaller ones? Among the successful entrepreneurs from whom data were collected in Berg (2014b), very small consideration sets with only three potential locations for a high-stakes investment decision were the rule rather than the exception. Larger consideration sets were associated with below-average investment performance. Less (that is, consideration of fewer feasible choices) was more (that is, greater than expected financial return). When choosing where to stand to catch a ball, three observations about how professional baseball players do it are worth noting in that they deviate from how robots would be programmed to do it using a veridical causal model based on initial velocity, wind speed, rotation and so on. Many researchers believe that veridical causal models stand unquestioningly as the gold standard for rational choice. In that view, deviations from how an idealized robot would do it are automatically labeled as mistakes. This view forces the interpretation that the deviations of professional baseball players – who are the best in the world at what they do – are prima facie evidence of irrationality rather than intelligence and high functionality. If the mind were essentially solving the physics problem of where to stand to catch the ball based on initial velocity, wind speed and rotation, then players who can reliably catch the ball in this way should be able to point to and predict the landing point without actually running to catch the ball. They cannot (see references in Gigerenzer 2005, and Berg and Gigerenzer 2010 on as-if behavioral economics). If players’ minds were evolved to approximate the veridical causal mechanism, then they should also run straight to the ball’s landing spot and do so as fast as they can to leave time for last-minute adjustments. Instead, they use a gaze heuristic that requires no causally relevant information at all and no precisely optimal angle (but, rather, allows for a large and forgiving range of angles that function just fine) at which to fix their gaze. The gaze heuristic is a process model: fix the angle of one’s gaze to the ball, start running and maintain the angle. It requires no causally relevant information. And it works.

BIAS–VARIANCE TRADE-OFF The bias–variance trade-off well known in statistics, machine learning and, more recently, psychology, implies that deliberate bias is, in general, a requirement of virtually any wellperforming statistical procedure that fits unknown parameters on a training set and then measures performance in generalization tasks requiring out-of-sample prediction. This trade-off forces the conclusion that insisting on zero bias will lead inexorably to maximal variance which, in any application with a single, finite dataset, violates most notions of ‘well performing’.

M4225-ALTMAN_9781782549574_t.indd 55

03/05/2017 08:20

56

Handbook of behavioural economics and smart decision-making

LEXICOGRAPHIC ORDER OF ASYMPTOTIC CONSISTENCY OVER VARIANCE REDUCTION IN ECONOMETRICS Classical econometrics is still taught in many if not most economics PhD programs as if there is a unanimous tacit agreement that the normative criteria of an estimator being unbiased or consistent (asymptotically converging to the correct value with probability 1) is infinitely more important than variance (not to mention performance in out-ofsample prediction). Orthodox econometric pedagogy advances a lexicographic order over biasedness and variance, ranking any estimation technique that is biased (or inconsistent) as strictly inferior, no matter what compensating characteristics (for example, speed, accuracy, low information requirements, lower variance, and so on) it may offer. Therefore, orthodox economic methodology applies lexicographic preferences over the methodological variables that economists choose when doing their work, while, in contrast, assuming that the consumers and firms in their models (with utility and profit functions satisfying the usual smoothness conditions) are never lexicographic in the way they reason about high-stakes decisions they face. This odd juxtaposition of methodological norms seems worth noting. Conditional mean functions are specified as flexible but always compensatory functions of the vector of conditioning variables. Utility functions are used that assume preferences cannot be lexicographic. In contrast, in econometrics, economists work under the assumption that lexicographic preferences over the characteristics of estimators are reasonable (that is, unbiasedness and consistency trump any comparisons of variance).

MORE INFORMATION NOT NECESSARILY BETTER EVEN IN GAMES AGAINST NATURE How much information should we pay attention to? When is it rational to ignore relevant information even when facing no cognitive constraints or costs of conditioning information? Berg and Hoffrage (2008) provide a formal definition of an economic or psychological environment and the matching concept of ecological rationality. They demonstrate that there are dense sets of environments in which, because payoffs and probabilities cancel out under the expected payoff operation, a non-redundant predictor or decision cue X that is veridically correlated with future payoffs may nevertheless drop out of optimal action rules, giving rise to the phenomenon of rational ignoring environments. Berg et al. (forthcoming) present data collected from economists that measure both individual-level belief consistency with respect to Bayes’ rule and belief accuracy with respect to published point estimates for disease frequencies in the medical literature. Which economists had the most objectively accurate beliefs about prostate cancer risks? It was not the economists whose conditional beliefs were perfectly Bayesian. Formal analysis of the analytic measures of belief consistency and belief accuracy, as well as the empirical data, show that performing well by one of these two distinct criteria does not imply good performance on the other. In many settings the multiple normative criteria that are observable in choice data may be negatively correlated. Perfect time consistency may arise mostly as a result of consistently impatient behavior (so that time consistency and the present value of lifetime wealth or laboratory earnings are negatively correlated).

M4225-ALTMAN_9781782549574_t.indd 56

03/05/2017 08:20

Rational mistakes that make us smart 57 Perfect conformity with the Savage Axioms may arise primarily as the result of consistently risk-averse choices with far-below average mean earnings. Perfect conformity with transitivity may result primarily as a very clear orientation toward leisure over money, implying that transitive types are, on average, less wealthy, less entrepreneurial and lower earning in laboratory experiments. When might the ‘mistake’ of failing to maximize expected utility and satisficing instead lead to social welfare improvements? Berg and Gigerenzer (2007) demonstrate just such an environment. Their model provides a thought experiment about a benevolent social planner who wants to achieve the greatest possible individual and aggregate payoffs for all stakeholders in her society. Now suppose she were able to choose whether the agents were expected utility maximizers or satisficers. Would the benevolent social planner follow behavioral economists’ preference for constrained optimization and advocate that individual members of society strive to be expected utility maximizers as opposed to satisficers? Berg and Gigerenzer (2007) show that the society of satisficers is unambiguously better off according to the same social welfare function. The satisficers achieve higher social welfare and require far less paternalistic intervention when compared from the vantage point of a Benthamite social-welfare metric.

STRATEGIC GAMES AGAINST SELF-INTERESTED COMPETITORS Mistakes can make an agent’s behavior less predictable and therefore thwart exploitative attacks. Like Columbo’s feigned ineptitude and lack of cleverness, agents that adopt decision styles that allow for and plan on committing errors can induce their adversaries into less cautious play, less aggressive best-response functions and greater revelation of information. To clarify: the errors considered in this chapter so far have nothing to do with strategically portraying anyone as stupid. However, it is worth including feigned irrationality in this list of examples that illustrate the breadth of mechanisms through which mistakes confer genuine value added. If others are convinced that I am stupid, then I may have more freedom to discover information or trade in markets without others strategizing against me. Inflated or wrong beliefs can make me a stronger negotiating partner. Mistakes lead to discoveries when the environment (for example, the reward- or payoffgenerating process) is changing (Bookstaber and Langsam 1985).

MARKETS AND SOCIAL SYSTEMS THAT BENEFIT FROM LOGICAL INCONSISTENCY AND OTHER ALLEGED ERRORS At the species level, suboptimal individual decisions may be rewarded by what is effectively a species-level portfolio diversification effect. There are some individuals failing to maximize in today’s environment, which may seem like a suboptimal waste. In the event that the payoff environment is buffeted by shocks so that previously optimal behaviors can no longer survive, however, then the currently suboptimal individuals may come into their own. Suppose the energy yield from grazing on the north side of the lake is 80 but only 20 on

M4225-ALTMAN_9781782549574_t.indd 57

03/05/2017 08:20

58

Handbook of behavioural economics and smart decision-making

the south side. What is individually rational is, of course, to graze at the north side. At the group level, however, when attacks, pests or poisons can appear on one side or the other, it is adaptive for some individuals to graze on the low-energy-yielding south side. This individual-level mistake averts group-wide cataclysm had all individuals chosen north and an unexpected attack takes place on the north. Market liquidity itself depends on noise or liquidity traders. Behavioral, belief and preference heterogeneity are primary reasons underlying why trade (that is, exchange itself) creates economic value. Berg and Lien’s (2005) model of Pareto-improving overconfidence in the precision of information possessed by insiders (in beliefs among the uninformed) shows that overconfidence in experts, while sometimes damaging, can generate surprising liquidity benefits in financial markets. These positive externalities in the form of lowered transactions costs more than offset the individual costs of having wrong beliefs. The model features informed experts and uninformed non-experts who may be overconfident in experts’ expertise (that is, the precision of experts’ information about future payoffs). If these agents’ (otherwise typical) payoff functions were alternatively interpreted as representing evolutionary fitness functions, then a striking conclusion emerges: there is no sense in which rational expectations (that is, objectively accurate subjective probabilistic beliefs) are adaptive; overconfident belief profiles support equilibria that Pareto dominate the rational expectations equilibrium. Sampling to learn about a changing environment is another benefit of making mistakes. That may explain why experimental subjects who switch their responses (perhaps randomly) to the very same decision tasks at different experimental sessions have been observed to earn more, on average, than do consistently impatient and consistently riskaverse individuals. The consistent types’ behavior passes the rationality test according to the norm of internal logical consistency, which is the sole claimant to rationality in rational choice orthodoxy. These consistent individuals earn significantly less, however (Berg et al. 2010b). Such contrasts, once again, highlight the multiple normative standards that economists employ, whether tacitly or explicitly (Berg, 2014), in characterizing the rationality of observed choice data. Randomization may confer other surprising benefits. For example, in social systems that offer opportunities for random face-to-face encounters, Berg et al. (2010a) show that agents who use a simple lexicographic heuristic for judging the acceptability of potential neighbors based on face recognition are capable of achieving stable multi-ethnic neighborhoods and preventing Schelling-type location-choice dynamics that tend toward absolute segregation. In public goods games, behavioral economists alternate in their interpretation of what constitutes mistaken behavior. Usually, failing to free ride, as required by the Nashequilibrium strategy (under the assumption that all players maximize standard rational choice own-payoff objective functions), is cast as an alleged mistake and serves as one of the main outcome variables that behavioral economists focus on. Kameda et al (2011) report evidence of strikingly intelligent behavior in the nonlinear public goods games they study. The ‘error’ in this case would be choosing an action that is different than the actions that everyone else in one’s group chooses even though all group members face exactly the same payoff functions and resource endowments. (That is, the game is completely symmetric, but Nash equilibrium requires an asymmetric profile of actions that theorists view as being very difficult to achieve). The equilibrium in the symmetric public goods games

M4225-ALTMAN_9781782549574_t.indd 58

03/05/2017 08:20

Rational mistakes that make us smart 59 they study requires asymmetric action profiles. Therefore, some means of coordinating or deciding which group member will volunteer to be the sole contributor and agree to be free-ridden upon is required. Rather than widespread pathology, their data reveal a wide range of individual and group intelligence. Regulation to prevent overuse of a commons is another longstanding question in public economics. The less-is-more principle underlying individual intelligence in Gigerenzer’s heuristics reappears, once again, as relevant to regulatory policy across multiple settings. For example, Berg and Kim (2015) show that permissive regulation that places fewer restrictions on the use of a commons (for example, road transportation networks, fisheries, or bandwidth for stock-price quotation networks that are exploited by high-frequency trading algorithms) can, counterintuitively, be more effective at mitigating overuse than stricter restrictions would have been, given imperfect enforcement of the regulation. A similar surprise regarding what looks mistaken through one lens of benefit–cost calculus becoming rational when viewed from another such lens shows up in models of social dynamics that include positive payoffs for coordinating with like types (as well as potentially negative externalities possibly resulting from extreme racial and religious segregation). The Kahneman-inspired normative position of much of behavioral economics condemning human judgment and decision making as generally pathological can be turned on its head once again: rather than widespread pathology as the default normative assessment of behavior that deviates from simple rational choice models and their assumed consistency criteria, there is as yet much intelligence that can be observed in apparently mistaken behavior. Take, for example, the money sacrificed on religious products for which there is an intrinsically equal-value substitute available at substantially lower price. Such behavior can, even without intrinsic benefit, provide socially valuable signaling and coordination functions (for example, Berg and Kim 2014, show that paying a higher premium for Islamic banking services can provide a signaling service that makes it worth paying for among highly pious types).

SINGULAR VERSUS PLURAL NORMS USED IN DEFINING RATIONALITY? It sounds paradoxical and unbelievable to many behavioral economists, which makes it worthwhile to reiterate: rational choice orthodoxy underlies much of behavioral economics, and the two share a methodological commitment to there being a single normative standard of rationality that does not depend on context or domain but instead is decided based solely on internal logical consistency (Berg 2003, 2014a; Berg and Gigerenzer 2010). The Kahneman-inspired biases literatures within behavioral economics and the field of judgment and decision making typically focus on deviations from some standard of logical consistency. Behavioral economists working in this vein are generally interested in the observational phenomenon of deviations from such a standard of internal logical consistency. Rather than question whether this normative standard used to define bias and deviations, the normative validity of the rational choice benchmark remains largely unquestioned among both what appears to be most behavioral economists and proponents of the rational choice orthodoxy. Their shared singular normative standard defines the

M4225-ALTMAN_9781782549574_t.indd 59

03/05/2017 08:20

60

Handbook of behavioural economics and smart decision-making

deviations that comprise the main outcome variables of interest to many behavioral economists. Such standards of rationality based solely on logical consistency include: logical invariance as the rationality standard against which framing effects become interestingly pathological; transitivity as the core component of the rational preference standard against which studies of intransitive and incomplete preference gain traction; Bayes’ rule in papers about non-Bayesian beliefs; the logic of set theory in investigations of the conjunction fallacy; and, even, Nash equilibrium as a benchmark of rationality in hundreds of studies by behavioral economists that report non-Nash play frequencies as the main dependent variable without ever comparing dollar payoffs (or comparisons by other normative metrics) among Nash versus non-Nash subsamples. In this chapter, I am considering the rationality of mistakes and errors of the kinds described above. To do so automatically implies that a newly pluralistic set of normative concepts are required. Ecological rationality is explicitly pluralistic by requiring good-enough (that is, satisficing levels) of match between a decision procedure and the environment in which it is used. This standard asks that, in a well-specified set of task environments, the decision procedure performs to a functional and pragmatic standard such that, despite and sometimes thanks to making mistakes, the procedure is readily seen as sensible, purposeful and, yes, rational! In the ecological rationality framework, a particular decision procedure or heuristic is, in itself, neither rational nor irrational. Unlike the rational choice and behavioral economics standard in which a single pair of intransitive choices or violation of logical invariance earns the universal assessment of irrationality, a choice procedure in the ecological rationality framework has performance characteristics that are alternatively rational and irrational depending on the external environment in which it is considered. It is only once the decision procedure is embedded in a particular environment that Herb Simon’s two blades of the ecological rationality scissors (decision procedure and external environment that jointly determine reasonable performance metrics for defining what is good enough to achieve success) can do their work at identifying boundaries that circumscribe the set of task environments in which a particular decision procedure achieves ecological rationality. It appears that any normative framework integrating the possibility of beneficial mistakes, as categorized above, necessarily implies that pluralistic normative metrics and the adaptive toolkit approach to defining what rationality means are in play.

WHICH ORGAN IN THE HUMAN BODY IS BEST? Does it make any sense to ask which organ in the human body is the best or most valuable? Using the massively interdependent body as an analogy, the behavioral phenomena of interest to social scientists will generally require multiple normative metrics akin to separately measuring and considering kidney function, liver function, cholesterol, triglycerides, blood glucose levels, and so on. Would it make sense to integrate all known organ-specific performance metrics or results from standard blood panels into a single, scalar-valued assessment, perhaps using a label such as generalized aggregate physiological (GAP) score? We would be hard pressed to think of any application where such aggregated summaries that compress the body’s multiple interdependent systems into a single

M4225-ALTMAN_9781782549574_t.indd 60

03/05/2017 08:20

Rational mistakes that make us smart 61 scalar-valued metric would be more informative or pragmatically useful than the disaggregated components considered as a fundamentally multivariate normative outcome. By analogy, when we ask the normative question using experimental choice data or theoretical models whether an observed set of behavioral patterns could be rationalized as if it were maximizing some scalar-valued objective function with newly exotic preference parameters to more flexibly mop up variation in the data, we are most likely asking a similarly wrong question. The standard analysis of a scalar-valued normative metric asks us to rely on the optimal choice function (that is, the program that maps exogenous parameters into an endogenous inference or action maximizing the narrowly defined objective). This method leads to the erection of a dug-in methodological phalanx that severely limits behavioral economics to persisting in egregious repetition of what statistician John Tukey called a type-III error: providing the right answer to the wrong question. In a massively interactive and interdependent biological or social system, the right way to behave depends on context. Rationality norms must be pluralistic and thoughtfully well-matched to a specific (that is, explicitly delimited) class of decision problems where a particular (that is, explicitly defined, possibly multivariate) standard of rationality makes sense (cf. Simon’s, 1976, notion of procedural rationality).

IS ECONOMICS THE ONLY DISCIPLINE WITH A COMMITMENT TO MONO-METHODOLOGICAL SINGULARISM? Yes.

INFLUENCE BY AND PARALLELS WITH THE AXIOMATIZATION PROGRAM IN MATHEMATICS? The axiomatization program in economics was in part inspired by the axiomatization program of mathematicians such as David Hilbert, Whitehead and Russell, and the Bourbaki group, which overlaps with the consistency school of normative bounded rationality (Berg 2014b). This axiomatization program profoundly influences (that is, restricts) economists’ normative analysis (that is, the normative questions that can be asked) in subtle ways that go mostly unnoticed in methodological treatises on the realworld applicability of behavioral economics and bounded rationality. Economic studies of bounded rationality would benefit by noticing the waning trajectory of this axiomatization program in mathematics and, like many in mathematics have, choose instead to pursue applied problems and the informal mathematics described in Backhouse (1998). I define the axiomatization program in economics as the body of economic theory that seeks a short list of axioms (perhaps minimal in some sense) that exhaustively characterizes the rationality of: preference orderings; sets of observed choices or demanded bundles (the extensive literature on revealed preference typically associated with Paul Samuelson); or orderings on choice sets. This axiomatization program can be narrowed further to investigations that pursue the question of postulating maximally general axioms (that is, the weakest possible) that can ‘rationalize’ observed choice behaviour. The

M4225-ALTMAN_9781782549574_t.indd 61

03/05/2017 08:20

62

Handbook of behavioural economics and smart decision-making

methodological priority of (topological) generality that characterized much of Hilbert’s program peaked in the latter half of the twentieth century. Since then, the dominance of the axiomatic program in mathematics has waned, whereas its methodological force in economics appears to have remained relatively undiminished. The history of the axiomatization program in economics reflects numerous borrowings and inspirations from mathematicians: David Hilbert, Bertrand Russell and the Bourbaki group all sought to rid mathematics of the possibility of inconsistencies. Bertrand’s paradox provides a primary motivation for early twentieth-century mathematicians’ program of eliminating inconsistency. That well-known paradox posits a collection of all sets that do not belong to themselves. The contradiction turns on ambiguity in the definition of the aforementioned ‘collection’ enjoying the status of set. By restricting the definition of a set to exclude some otherwise well-defined collections of mathematical objects, Frege, Whitehead and Bertrand, and Fraenkel introduced a new formalism into mathematics to resolve such paradoxes, most often beginning with axiomatization. There are even earlier links in the works of mathematicians such as Georg Cantor in the late 1800s (and axiomatization programs in set theory which followed) to the later axiomatization program in economics, based on the goal of providing a minimal list of conditions to ‘rationalize’ choice data. The ‘characterization’ of rationality and the ‘rationalization’ strand of the axiomatization program in economics can be thought of as beginning with a set of axioms and a universe of observable patterns of behavior and then projecting the graph that characterizes all allowable patterns of behavior that satisfy the axioms, which is a strict subset of the larger universe of possible patterns of behavior. This can be backward engineered as follows: Given the observed set of choices or behavior patterns, what axioms must this set of choice data satisfy in order to (1) recover a preference ordering that could have generated the choice data, and (2) assuming a preference ordering exists for ranking vector-valued bundles or payoff distributions in the case of risky choice, what axioms must the data satisfy for the rankings of multidimensional objects to be representable as scalar-valued utility or value scores? Note that this rationalization subset of the axiomatization program in economics contains, for example, Tversky and Kahneman’s (1992) loss-averse cumulative prospect theory, specifically, versions of it that attempted to rationalize the choice data generated in Allais’ paradox (which are interpreted as anomalous with respect to expected utility theory). Rationalizing anomalous choice data is described by Gerd Gigerenzer, Werner Güth and Reinhardt Selten as a repair program. The goal is to take choice data (from binary choices over pairs of risky gambles in the case of prospect theory as a resolution to Allais’ Paradox) that cannot be represented with an expected utility function, and then show that those data could have been generated by prospect theory, for some unspecified but theoretically possible parameters that determine the shape of the value-function and the nonlinear function mapping objective probabilities into decision weights. Note that this rationalization project, or repair program, bears some similarity to the fallacy of ranking regression models according to their R-squared. Finding a list of axioms that ‘can explain’ choice data is analogous to a regression model with more right-hand-side variables fitting a dataset better. As econometric textbooks correctly caution, a model that fits the data better may not necessarily make more accurate out-of-sample predictions. Fit can always be made to reach 100 percent if enough free parameters are added to the model specification, one for each observation in the fitting or training sample. Arguments

M4225-ALTMAN_9781782549574_t.indd 62

03/05/2017 08:20

Rational mistakes that make us smart 63 in favor of prospect theory that use different parameterizations to fit each new sample, for example, do not typically discipline that model to the risks of out-of-sample prediction. Instead, they use the flexibility of the additional parameters in prospect theory (compared with standard expected-utility theory) to increase ‘fit’ in a manner entirely analogous to increasing R-squared by trivially increasing the number of free parameters used in a regression model. Cantor proved – more than century ago – that if a binary relation is linearly ordered, then it is also embeddable as an isomorphism in the real numbers. Technically, this is almost identical to the intellectual work of writing down axioms (that is, restrictions on the preference ordering) that guarantee representability with utility, expected utility or prospect-theory value-function scores. Ragnar Frisch is credited as the first economist to define preferences as binary relations. Contemporary graduate textbooks use very different notation (deleting Frisch’s more broad-ranging ‘choice field’ formulation, which distinguishes commodity space from what Frisch referred to as the decision maker’s problem space). Frisch played a leading role in the founding of the Econometric Society and the journal Econometrica, advocating formalism and math modeling as a primary source of ‘rigor’ needed to put economics on a ‘scientific’ footing (Bjerkholt and Dupont 2010). Despite his view that mathematizing economics was needed to displace the ‘verbal’ approaches of institutionalists, his sophisticated appreciation of the fact that the decisions modeled as constrained utility maximization (exhaustively searching through a feasible set in commodity space) are embedded in a larger problem space that includes problems perhaps not best handled by the techniques of constrained optimization is striking. This notion of a larger ‘problem space’ foreshadows the notion of ‘environment’ used by writers such as Gigerenzer and Vernon Smith in advocating ecological rationality. Frisch’s concept of constrained maximization in commodity space as only one decision domain embedded in a larger problem space notably does not appear in most contemporary PhD textbooks, which instead emphasize the flexibility and universality of preference maximization devoid of context specificity. As is well known to economic methodologists and historians, early representation theorems in utility theory sought to address debates in economics between those who interpreted utility as a potentially measurable psychological metric of hedonic satisfaction and those influenced by logical positivism wanting to remove psychological notions (Bruni and Sugden 2007). Early representation theorems establishing utility as a purely ordinal concept devoid of cardinal meaning led to representation theorems in expected utility theory, axiomatizations of Bayesian updating as rational belief functions, and, more recently, weaker axiomatizations that can account for (as bounded rational) some well-known anomalies with respect to rational choice theory. It is this last subliterature of economists writing on rationality axioms in behavioral economics and making reference to Herbert Simon’s phrase ‘bounded rationality’ that is relevant to this chapter’s focus on bounded rationality and smart people’s rational mistakes. It is instructive to recall that the central motivation of Hilbert and Whitehead and Russell’s (1927 [2009]) axiomatization program was to formalize mathematics and philosophy with the explicit goal of eliminating inconsistency. Hilbert and Russell undertook this program and advocated that others join them to rid mathematics – and science – of the possibility of generating inconsistent statements, whether those statements be abstract or detailed descriptions of the world.

M4225-ALTMAN_9781782549574_t.indd 63

03/05/2017 08:20

64

Handbook of behavioural economics and smart decision-making

While Hilbert’s move toward formalism profoundly influenced mathematics (and at the same time attracted well-established critics), it eventually waned as new subfields in mathematics applying methods outside the Hilbert program grew up and gained acceptance as making substantial contributions to mathematics. Applied problem solving, combinatorics, category theory and subfields of mathematics overlapping with computer science achieved influence and prominence, while other theorists working in the constructivist and intuitionist traditions similarly produced new knowledge that followed distinct methodological priorities. The methodological influence of formalism and the axiomatization program in economics followed an arguably equal if not more profound influence in economics (see Backhouse 1998, regarding formalism in economics versus informal mathematics). One minor parallel between the trajectories of formalism in mathematics and economics was the desire to shed old interpretations (for example, the interpretation of points, lines and planes in geometry and the psychological or hedonistic interpretation of utility in nineteenth-century economics in favor of utility as a purely ordinal device). Another speculative parallel that can be seen in the restrictions that choice axioms placed on what had been a previously more libertarian view of consumer sovereignty is to see them echoing the restrictions that Frege, Whitehead and Russell, Fraenkel and Hilbert applied to the definition of a set (in order to avoid paradoxes such as Russell’s). Beyond these similarities, however, the differences in the historical trajectories of axiomatization programs in mathematics and that of economics are many. Formalism in economics (until very recently) did not have a long struggle with concepts such as the definition of a set as its core methodological problem, syntactical formalism, the incompleteness theorems of Gödel, and many others. The mathematical issues in the development of economists’ formalism were, by and large, far simpler mathematically, and focused on applying topological formalisms already established in mathematics to preferences and representations of preferences. In the axiomatization program in economics, the role of interpretation and motivation of axioms were the primary objects of notable theoretical economists’ writing. Critiques and crises over the roles of an axiomatization program (and the ‘interpretation-free’ view of mathematics as a content-free set of primitives and a formulaic set of statements based on definitions of operations juxtaposed or concatenated to generate all permutations allowed by the axioms) did not surface or echo in economics, at least in obvious ways. These differences, however, serve to cast into sharp relief the one overriding similarity between the axiomatization programs of math and economics: internal logical consistency as the pre-eminent normative value.

BEHAVIORAL ECONOMICS IS NORMAL SCIENCE PORTENDING NO PARADIGM SHIFT IN NORMATIVE ANALYSIS Some argue that behavioral economics should be interpreted as a paradigm shift or otherwise momentous contestation in reaction to the axiomatization program in economics. Behavioral economists’ work could, if such an interpretation were granted, be seen as echoing earlier methodological shifts in mathematics following the rise and decline of

M4225-ALTMAN_9781782549574_t.indd 64

03/05/2017 08:20

Rational mistakes that make us smart 65 the Bourbaki group’s influence in mathematics in the twentieth century. I argue against this methodological view of behavioral economics as a paradigm shift and instead demonstrate that its overriding normative value remains firmly rooted in the axiomatization program’s normative view, namely, that the central concern is, and should be, internal logical consistency. See Berg (2014a) or Berg and Gigerenzer (2010) for further detail distinguishing different camps within behavioral economics, and Berg (2003) on Thaler’s and Kahneman’s changing positions regarding the normative implications of implicit and explicit methodological challenges that behavioral economics put to orthodox rational choice models. Those who see behavioral economics and modelers of bounded rationality acting as an ensemble to ‘expand’ or ‘loosen’ the methodological strictures of rational choice theory miss a crucial difference in the normative views of the consistency and ecological rationality schools as I have defined them in earlier work (Berg 2014a). Behavioral economists in the consistency school propose radically narrow normative definitions of rationality and use the label ‘bounded rationality’ (in a manner that would seem to contradict Herbert Simon’s normative view). The result is to harden the methodological commitment to internal consistency as the sole criterion that economists are expected to use in characterizing what it means to make rational decisions – and in prescriptive policy proposals that paternalistically intervene, aiming to induce people’s private actions to more closely conform to axiomatic models of rationality. Backhouse (1998) reminds us that axiomatization, mathematicization and formalization are distinct. Gigerenzer and Selten’s (2001) ecological rationality program provides a clear example of normative decision analysis that draws on quantitative data to produce theories that can be expressed in the language of mathematics, yet have nothing to do with axiomatization. Backhouse notices (as many other writers on mathematics and philosophy, and the history of mathematics have) that mathematics itself can be either formal or informal. In the development of proofs of Euler’s theorem, for example, which relates the numbers of vertices, faces and edges of a polyhedron, Backhouse (1998, p. 1848) describes different authors’ proofs as somewhere ‘between formalism and irrationalism. . ..There is more to mathematics than driving the properties of formal systems’. The implication would seem to be that applied economics, welfare economics and prescriptive policy analysis cannot be entirely about deductive logic (Berg 2007). Indeed, the proper role of deductive logic led to animated and productive debates about mathematical methodology and philosophy regarding the Hilbert program among constructivists, intuitionists (including Hilbert’s students Brouwer and Weyl), subsequent work in proof theory, category theory and those inspired by Turing on computability. Given these prolific bodies of work by mathematicians that raised questions about consistency as the core methodological concern in mathematics, it would seem wrong for economists to draw the lesson from mathematics – in the name of ‘providing rigor’ or ‘putting economics on a more scientific basis’ – to insist on applying consistency alone as the ultimate methodological value. What are we to make of the long tradition among neoclassical economists – and now behavioral economists – who seem to follow Hilbert’s singular normative premise in pursuit of logical consistency? I think we can note the positions of economists like Debreu and Binmore as playing a role similar to Hilbert’s role in mathematics. Their staunch position in favor of consistency as a singular methodological and normative-prescriptive

M4225-ALTMAN_9781782549574_t.indd 65

03/05/2017 08:20

66

Handbook of behavioural economics and smart decision-making

value is simply one among multiple, competing normative claims within economics. Heterogeneity of methodological priorities is a positive symptom of productive scientific investigation. In light of the productivity generated by those who raised questions and took positions against Hilbert’s consistency program in mathematics, however, economists might also notice that competing normative claims are likely to play a similarly productive role in economics. Such methodological debate is no small side issue but rather a substantial object of core investigation in normative economics.

CONCLUDING REMARKS Casual empiricism and the theoretical economics, biological sciences and biostatistics literatures provide a rich collection of source material from which one finds a broad range of mechanisms by which smart people make rational mistakes. Also, economies that generate value added and nurture richly multidimensional measures of well-being generate numerous opportunities by which aggregate performance is enhanced thanks to systematic deviations from standard rationality criteria based solely on internal logical consistency. I have provided examples that hopefully give a sense of the technical, substantive and historical range of context-specific mechanisms in which alternative normative criteria that allow for welfare-enhancing deviations from logically consistent axiomatic rationality can be given even-minded consideration. May further study of this important phenomenon bloom forth and melt away the methodological strictures unnecessarily limiting behavioral economists’ evaluations of rationality.

REFERENCES Axelrod, R. (1984), The Evolution of Cooperation, New York: Basic Books. Backhouse, R.E. (1998), ‘If mathematics is informal, then perhaps we should accept that economics must be informal too’, Economic Journal, 108 (451), 1848–58. Berg, N. (2003), ‘Normative behavioral economics’, Journal of Socio-Economics, 32 (4), 411–27. Berg, N. (2007), ‘Behavioural economics, business decision making and applied policy analysis’, Global Business and Economics Review, 9 (2–3), 123–5. Berg, N. (2014a), ‘The consistency and ecological rationality schools of normative economics: singular versus plural metrics for assessing bounded rationality’, Journal of Economic Methodology, 21 (4), 375–95. Berg, N. (2014b), ‘Success from satisficing and imitation: entrepreneurs’ location choice and implications of heuristics for local economic development’, Journal of Business Research, 67 (8), 1700–1709. Berg, N. and G. Gigerenzer (2006), ‘Peacemaking among inconsistent rationalities?’, in C. Engel and L. Daston (eds), Is There Value in Inconsistency?, Baden-Baden: Nomos, pp. 421–33. Berg, N. and G. Gigerenzer (2007), ‘Psychology implies paternalism? Bounded rationality may reduce the rationale to regulate risk-taking’, Social Choice and Welfare, 28 (2), 337–59. Berg, N. and G. Gigerenzer (2010), ‘As-if behavioral economics: neoclassical economics in disguise?’, History of Economic Ideas, 18 (1), 133–66. Berg, N. and U. Hoffrage (2008), ‘Rational ignoring with unbounded cognitive capacity’, Journal of Economic Psychology, 29 (6), 792–809. Berg, N. and J.Y. Kim (2014), ‘Prohibition of riba and gharar: a signaling and screening explanation?’, Journal of Economic Behavior and Organization, 103 (July), 146–59. Berg, N. and J.Y. Kim (2015), ‘Quantity restrictions with imperfect enforcement in an over-used commons: permissive regulation to reduce over-use?’, Journal of Institutional and Theoretical Economics, 171 (2), 308–29. Berg, N. and D. Lien (2005), ‘Does society benefit from investor overconfidence in the ability of financial market experts?’, Journal of Economic Behavior and Organization, 58 (1), 95–116.

M4225-ALTMAN_9781782549574_t.indd 66

03/05/2017 08:20

Rational mistakes that make us smart 67 Berg, N., G. Biele and G. Gigerenzer (forthcoming), ‘Consistent Bayesians are no more accurate than nonBayesians: economists surveyed about PSA’, Review of Behavioral Economics (ROBE). Berg, N., C. Eckel and C. Johnson (2010), ‘Inconsistency pays? Time-inconsistent subjects and EU violators earn more’, working paper, University of Texas-Dallas, Dallas, TX. Berg, N., U. Hoffrage and K. Abramczuk (2010a), ‘Fast acceptance by common experience: FACE-recognition in Schelling’s model of neighborhood segregation’, Judgment and Decision Making, 5 (5), 391–410. Bjerkholt, O. and A. Dupont (2010), ‘Ragnar Frisch’s conception of econometrics’, History of Political Economy, 42 (1), 21–73. Bookstaber, R. and J. Langsam (1985), ‘On the optimality of coarse behavior rules’, Journal of Theoretical Biology, 116 (2), 161–93. Bruni, L. and R. Sugden (2007), ‘The road not taken: how psychology was removed from economics, and how it might be brought back’, Economic Journal, 117 (516), 146–73. Gigerenzer, G. (2005), ‘I think therefore I err’, Social Research, 72 (1), 195–218. Gigerenzer, G. and R. Selten (2001), Bounded Rationality: The Adaptive Toolbox, Cambridge, MA: MIT Press. Kameda, T., T. Tsukasaki, R. Hastie and N. Berg (2011), ‘Democracy under uncertainty: the wisdom of crowds and the free-rider problem in group decision making’, Psychological Review, 118 (1), 76–96. Rapoport, A. (1984), ‘Game theory without rationality’, Behavioral and Brain Sciences, 7 (1), 114–15. Rapoport, A. and A.M. Chammah (1965), Prisoner’s Dilemma, Ann Arbor, MI: University of Michigan Press. Schooler, L.J. and R. Hertwig (2005), ‘How forgetting aids heuristic inference’, Psychological Review, 112 (3), 610–28. Simon, H.A. (1976), ‘From substantive to procedural rationality’, in S.J. Latsis (ed.), Method and Appraisal in Economics, Cambridge: Cambridge University Press, pp. 129–48. Tversky, A. and D. Kahneman (1992), ‘Advances in prospect theory: cumulative representation of uncertainty’, Journal of Risk and Uncertainty, 5 (4), 297–323. Whitehead, A.N. and B. Russell (1927), Principia Mathematica, 3 vols, Cambridge: Cambridge University Press; 2nd edn, 1925 (vol. 1), 1927 (vols 2, 3), vols 1, 2 and 3 originally published in 1910, 1912 and 1913; 1st edn reprinted 2009 by Merchant Books, USA.

M4225-ALTMAN_9781782549574_t.indd 67

03/05/2017 08:20

4

Rational choice as if the choosers were human Peter J. Boettke and Rosolino A. Candela

1

INTRODUCTION

In a recent paper entitled ‘Principles of (behavioral) economics’, economists David Laibson and John List claim that ‘behavioral economics is a series of amendments to, not a rejection of, traditional economics’ (2015, p. 385), which studies ‘how people try to pick the best feasible option, including the cases in which people, despite their best efforts, make mistakes’ (2015, p. 389, original emphasis). For a classroom of undergraduates they would summarize the principles of (behavioral) economics in this way: If you want to boil behavioral economics down for a classroom summary you might say that most people are located somewhere between Mr. Spock and Mr. Simpson (aka Homer). Like Mr. Spock, Mr. Simpson is also an optimizer – he tries to choose the best feasible option. He’s just not good at it. We need to study and model all optimizers: the good, the bad, and those in between. (Laibson and List 2015, p. 389)

Oddly enough, these statements made by Laibson and List about human decision-making parallel strongly with the following statement made by Ludwig von Mises, one of the strongest proponents of rationality in the history of economic thought: It is a fact that human reason is not infallible and that man very often errs in selecting and applying the means. An action unsuited to the end sought falls short of expectation. It is contrary to purpose, but it is rational, i.e., the outcome of a reasonable – although faulty – deliberation and an attempt – although an ineffectual attempt – to attain a definite goal. The doctors who a hundred years ago employed certain methods for the treatment of cancer which our contemporary doctors reject were – from the point of view of present-day pathology – badly instructed and therefore inefficient. But they did not act irrationally; they did their best. (Mises 1949 [2007], p. 44)

From the quotes stated above, both Laibson and List and Mises seem to be depicting rational choosers as if they were human. However, if traditional economics does indeed need to be amended by behavioral economics, as Laibson and List argue, then the question is what notion of man has occupied ‘traditional’ economics? Implicitly, it would seem that the notion of man they have in mind for traditional economic analysis is none other than Homo economicus. The concept of economic man, or Homo economicus, has been under assault throughout much of the history of the discipline. It has often been a criticism intimately tied to an effort to discount the lessons that can be learned from economics for the practical understanding of public policy. Since economics as a discipline stresses scarcity and thus choice within constraints, the debate often turns on how competent people are in making choices, and how binding those constraints are. In an idealized world, the argument goes, individuals are fully informed and perfectly rational in making their decisions, and the constraints 68

M4225-ALTMAN_9781782549574_t.indd 68

03/05/2017 08:20

Rational choice as if the choosers were human 69 they face are hard and unyielding. Thus, correct decisions will not just ‘tend’ to be made, but will inevitably be made. Since this applies equally to all in the economy, the equilibrium that results will exhibit exchange efficiency, production efficiency, and product-mix efficiency. In short, all the gains from trade will be exhausted, all the gains from technological improvement will be incorporated into production, and the array of products that buyers are willing to pay for would be available on the market. Perfectly rational actors interacting freely in a frictionless environment produce an efficient outcome. With that narrative in the background, then consider the argumentative strategy of those who wanted to critique the free market system – they can go after the notion of the rational actor, they can go after the frictionless environment, and they can challenge the ethical status of the efficiency standard. Throughout the history of the discipline, all three intellectual strategies have been pursued. The easiest target has been the bogeyman of Homo economicus. Economic man is a bogeyman in two ways – the claim that the concept implies that economic ends, or monetary motives, enter the decision calculus, and that the model implies that the decision makers are imbued with omniscience with respect to all the relevant factors to the decision. ‘The hedonistic conception of man’, Veblen wrote, ‘is that of a lightning calculator of pleasures and pain, who oscillates like a homogenous globule of desire of happiness under the impulse of stimuli that shift him about the area, but leave him intact’ (1898, p. 389). Or consider how Keynes in his rhetorical brilliance was able to link both perfect rationality and perfect markets together and dismiss both claims in ‘The end of laissez faire.’ As he put it: Let us clear from the ground the metaphysical or general principles upon which, from time to time, laissez-faire has been founded. . ..The world is not so governed from above that private and social interests always coincide. It is not so managed here below that in practice they coincide. It is not a correct deduction from the principles of economics that enlightened self-interest always operates in the public interest. Nor is it true that self-interest generally is enlightened; more often individuals acting separately to promote their own ends are too ignorant or too weak to attain even these. (Keynes 1926 [1978], pp. 277–8, original emphases)

Human actors to Keynes are not rational as the ‘model’ presumes, and the market system does not function as smoothly as the ‘model’ suggests assuring that private and public interests align. The choice Keynes provides is binary – either perfect actors and perfect markets, and thus laissez-faire, or imperfect actors and imperfect markets, and thus activist government policy as a corrective. There simply is no way in his intellectual schematic that imperfect actors operating in an imperfect world could be stumbling upon coping mechanisms for the complex reality in which they find themselves, enabling them to realize the productive gains from specialization and peaceful cooperation without the activist hand of enlightened government. Keynes brilliantly identified the ‘dark forces of time and ignorance’ (1936 [1964], p. 155) in The General Theory, but in his depiction human actors are unable to navigate in that world. The debates over individual rationality and system-level efficiency have proceeded along these lines ever since. By modelling the actor as a close-ended decision maker and the economy as exhibiting a single exit, our result is the deterministic model of rational ‘choice’ depicted in a standard economics textbook, in which the human actor is devoid of any cognitive ability. If we introduce some form of imperfection, either with the actor (say, informational asymmetries) or in the market structure (say, monopolistic competition),

M4225-ALTMAN_9781782549574_t.indd 69

03/05/2017 08:20

70

Handbook of behavioural economics and smart decision-making

then the welfare conclusions derived from the determinant solution shifts. The government as a corrective either at the actor level to provide the requisite information, or at the structural level to provide the requisite regulatory measures, seems to follow naturally from the model. However, the strict binary intellectual choice that a Veblen or Keynes imposed on the economic conversation need not be followed. This is true for the contemporary discussion of behavioral economics, renewed calls for paternalism, and the entire practice of nudges. In this chapter, we contribute to the theme of this book by evaluating how rational human choosers are in fact ‘smart’ in the decision-making process once we have taken into account the particular institutional context within which they are evaluating the costs and benefits of their choices. As Morris Altman states, ‘conventional economics assumes that people’s choices are made in a vacuum’ (2012, p. 4), which is not only institutional, but also historical and cultural in nature as well. What we hope to demonstrate is that there has been from Adam Smith to Vernon Smith a tradition of economic scholarship that is grounded in the decision calculus of individuals, but requires neither the heroic assumptions of omniscience, nor that the individuals are interacting with others in frictionless environments. Instead, they see man as pursuing their varied purposes and caught, as they often are, between alluring hopes and haunting fears, and interacting in institutional environments that are constituted by vaguely and imperfectly understood rules that are often poorly enforced and path-dependent on the imprint of culture and history. Yet the filtering mechanisms of this institutional environment are guided to act in ways that coordinate their activities with those of others to realize the mutual gains from social cooperation. It is precisely because these scholars emphasize the open-endedness of choice that they can identify the role that even imperfect institutions play in coordinating economic affairs through time. Section 2 provides a survey of economic thinkers who rejected the caricature of Homo economicus that critics claimed was in fact held by classical and neoclassical economists, but who nevertheless defend the rational choice perspective in the social sciences, and economics in particular. These will include figures such as Frank Knight, Ludwig von Mises, F.A. Hayek, James Buchanan, Douglass North, Vernon Smith and Elinor Ostrom. Section 3 focuses on how Hayek, Buchanan and Ostrom develop the argument to move to the rules level of analysis in human decision making and human interaction. Section 4 discusses the concept of path dependency and imperfect institutions as developed by North and Ostrom. Section 5 concludes.

2

METHODOLOGICAL INDIVIDUALISTS WHO REJECT HOMO ECONOMICUS BUT EMBRACE RATIONAL CHOICE

In his Epistemological Problems of Economics, Ludwig von Mises (1933 [1960]), writing before the rise of the Keynesian revolution in macroeconomics and the growing emphasis on mathematical formalism and equilibrium analysis in microeconomics, claimed that there had been a consolidation of certain core propositions from different strands of economic thought that had emerged from the Marginal Revolution of 1871. These developments in neoclassical economics, according to Coase (1992, p. 713), were rooted in filling ‘the gaps in Adam Smith’s system, to correct his errors, and to make his analysis vastly

M4225-ALTMAN_9781782549574_t.indd 70

03/05/2017 08:20

Rational choice as if the choosers were human 71 more exact’. Choice within constraints had been a staple of economic analysis since at least the eighteenth century, but the Marginal Revolution led to a deeper understanding of the subjective nature of utility, the unit of analysis being the individual, and the choice calculation on the margin of decision. With the intellectual revolution in value theory, any Ricardian notion that long run costs of production determined value and price was to be jettisoned, and fallible yet competent human decision makers became the focal point of economic analysis. As Mises argued: Within modern subjectivist economics it has become customary to distinguish several schools. We usually speak of the Austrian and the Anglo-American Schools and the School of Lausanne . . . these three schools of thought differ only in their mode of expressing the same fundamental idea and that they are divided more by their terminology and by peculiarities of presentation than by the substance of their teachings. . .Today we have only one theory for the solution of the problems of catallactics, even if it makes use of several forms of expression and appears in different guises. (1933 [1960], pp. 214–15)

The ‘one theory’ for the analysis of catallactics that Mises had emphasized constituted a set of positive propositions that led to the further development of ‘mainline’ economics (see Boettke 2012). These propositions, which were held in common by economists from classical political economists such as Adam Smith to modern experimental economists such as Vernon Smith, explained the emergence of social order based on invisible hand theorizing that reconciled a broad notion of self-interest (that is, purposive action) with the public interest via institutional analysis. Figure 4.1 illustrates this point. Rational individuals, though imperfect in their cognitive capabilities, yet guided by the institutional prerequisites of private property, money prices, and profit/loss, will nonetheless coordinate their subjective plans through the unintended design of the invisible hand, yielding a social order. These mainline economists did not explain social order or disorder by collapsing selfinterest on to the public interest or by assuming the super-human cognitive capabilities, or lack thereof, upon individuals. However, in textbook presentations of mid-twentieth century microeconomics (and unfortunately true till this day) the argument is that social order results if, and only if, actors are fully informed and perfectly rational, and the market structure is perfectly competitive. Otherwise, decision making and system-wide efficiency will be lacking, and in need of correction.1 These two views of what has become ‘mainstream’ economics are illustrated in Figure 4.2. Rather than utilizing a behaviorally contingent explanation, their analysis was based on an institutionally contingent process Self-interest

Figure 4.1

Institutional filter

Social order (i.e. public interest)

Sequence of causation in mainline economics

Perfectly rational self-interest Irrational self-interest

Figure 4.2

Invisible hand

Perfect market structure Imperfect market structure

Social order Social disorder

Sequence of causation in mainstream economics

M4225-ALTMAN_9781782549574_t.indd 71

03/05/2017 08:20

72

Handbook of behavioural economics and smart decision-making

of reconciliation via exchange between fallible but capable individuals within a context of private property, freedom of contract, and the rule of law. All the developments that we are talking about are, as we quoted Coase above as saying, seen as filling in the gaps of Adam Smith’s scientific system. However, as the mainstream of economics deviated significantly from the mainline of economics as developed by the classical political economists and early neoclassical economists, acts of scientific entrepreneurship were initiated to try to place the individual once again at the center of economic analysis, and to resurrect institutional analysis as critical in explaining observed patterns of social order (disorder). It is in these acts of scientific entrepreneurship that we see ‘schools of thought’ playing out their function – in our narrative this includes the ‘Austrian school’, the ‘property rights school’, the ‘public choice school’, and the ‘new institutional school’ of contemporary economic thought. For Mises and the Austrians of the 1930s, the major opponents of this mainline notion of economic theorizing were perceived ‘not as being the followers of Walras or of Marshall, but as being the historical and institutionalist writers’ (Kirzner 1988, p. 9) who had criticized mainline reasoning by presuming that catallactics was behaviorally dependent on a notion of Homo economicus. For example, Institutionalist economist Thorstein Veblen criticized neoclassical economists for basing economic theory upon ‘a faulty conception of human nature’, which he rejected as a ‘hedonistic’ conception of man as a lightning calculator of pleasure and pains, namely because ‘under hedonism the economic interest is not conceived in terms of action’ (1898, p. 394). Remarking on such renditions made by Institutionalists and Historicists during the Methodenstreit, Mises asserted that: It was a fundamental mistake of the Historical School . . . in Germany and of Institutionalism in America to interpret economics as the characterization of the behavior of an ideal type, the Homo oeconomicus . . . Such a being does not have and never did have a counterpart in reality; it is a phantom of a spurious armchair philosophy. No man is exclusively motivated by the desire to become as rich as possible; many are not at all influenced by this lean craving. It is vain to refer to such an illusory homunculus in dealing with life and history. Even if this really were the meaning of classical economics, the Homo oeconomicus would certainly not be an ideal type. The ideal type is not an embodiment of one side or aspect of man’s various aims and desires. It is always the representation of complex phenomena of reality, either men, of institutions, or of ideologies. The classical economists sought to explain the formation of prices. They were fully aware of the fact that prices are not a product of the activities of a special group of people, but the result of an interplay of all members of the market society. (Mises 1949 [2007], p. 62, original emphasis)

Just like the classical economists, neoclassical economists of the twentieth century also were preoccupied with explaining the formation of prices, but they were also increasingly occupied with conceptualizing price determination along Walrasian and Marshallian lines, both of which take cost curves to be objective and therefore measurable. It is from this backdrop that a debate emerged over the use of marginal analysis, in which the criticisms of institutionalist economists against the principles of neoclassical economics would resurface. As Robert Prasch has argued, ‘this episode, now remembered as the “Marginal Cost Controversy”, presents us with something of an American Methodenstreit’ (2007, p. 815). During the 1940s, economist Richard Lester challenged the empirical reality of economic actors engaging in marginal decision making. According to Lester, survey data of labor markets demonstrated that actors had no clue about weighing marginal benefits and

M4225-ALTMAN_9781782549574_t.indd 72

03/05/2017 08:20

Rational choice as if the choosers were human 73 marginal costs. For Lester, like the Institutionalists, economic generalizations could be inferred without theory from generalizations that could be verified empirically. Moreover, when ‘analytical concepts, including the competitiveness of the market, the nature of economic rationality, or the structure of a firm’s costs, are assumed or asserted without reference to widely understood and accepted facts, then that theory lacked genuinely scientific foundation’ (Prasch 2007, p. 814). Although Lester was rejecting marginalist principles, the premise of Lester’s argument rested implicitly on the notion that cost curves were objective in the sense that they were measurable by an outside observer. In this respect, Austrian economist Fritz Machlup responded that cost curves were subjective, and therefore his conclusions were invalid. Machlup’s position is consistent with that of Hayek’s presentation in ‘Economics and knowledge’ (1937), where the marginal conditions are not assumptions going into an analysis, but by-products that emerge out of decision making, ‘discovered anew’ within the process of market competition itself. Too often Machlup’s contribution is captured under the heading of ‘as if’ modeling. While Machlup often used the instrumentalist language of his day to try to communicate his point, a careful reader will note that he always makes subtle shifts in language, which was understood by those at the time as qualifiers, but which have failed to travel through time with him. Such a classic case is his shift in the debate over verification in economic science where he switches the claim about ‘predictability’ to one focused on ‘intelligibility’ (Machlup 1955; see also Boettke 2015; Zanotti and Cachanosky 2015). A similar subtle switch occurs when Machlup in the science wars argues that economics is as scientific as the natural sciences are, but it is just that in the sciences of man ‘matter can talk’, changing the epistemological problems that must be confronted in the practicing of the science. A comprehensive review of the marginal cost controversy is beyond the scope of this chapter (see also Lavoie 1990). What is important for our analysis of rational choice, in which economic actors are fallible, but capable, human beings and neither mechanistic automatons nor hopelessly disoriented actors, is that many textbook presentations of the Lester–Machlup debate present Machlup as the winner, but present his argument in the ‘as if’ tradition of thinking championed by Milton Friedman (1953). Individuals act ‘as if’ they balance marginal benefits and costs even if they do not explicitly do so in their own minds. However, this misses the point in the sense that the debate has been understood in terms of behavioral assumptions of whether or not individuals are profit maximizing or not. What was lost in the exchange was an analysis of the institutional conditions within which individuals are enabled to or inhibited from pursuing maximum profits, not only pecuniary but also non-pecuniary. The marginal conditions have little or nothing to do with how individuals actually make decisions. Rather, the marginal conditions emerge as a by-product of the market process within an institutional context of private property, prices, and profit/loss accounting. We do not have direct access to motivations of individuals. What we can study is the systemic effect of different institutional arrangements on the incentives that actors face. However, we cannot detail what motivates individuals. As the renowned Chicago economist Frank Knight has argued: We really know very little about human motives, and still oversimplify them disastrously in nearly all discussion . . . The larger problem is to arrange things so that people will find their

M4225-ALTMAN_9781782549574_t.indd 73

03/05/2017 08:20

74

Handbook of behavioural economics and smart decision-making lives interesting and will grow into such personalities that they can respect themselves, admire others and enjoy their society, appreciate thought and beauty, and in general look upon creation and call it good. (1919, p. 806)

Among those institutional arrangements, Knight not only emphasized the formal constraints such as private property, freedom of contract, and the rule of law, but also understood that such formal institutions are dependent on informal norms that are conducive to self-interest ‘rightly understood’ as the harmony or dovetailing of individual ends among members of society: From the falsity of the atomistic-individualistic view of human nature and human desires it is an easy inference that any mechanical theory of social organization is subject to narrow limitations. The most potent agency of social control, even today, in spite of all the obstacles thrown in its way by an antiquated and wooden system of association, is the moral control of the individual’s sense of decency and the pressure of the opinions of his fellows. (Knight 1920 [2011], p. 87)

Moreover, market interactions are not defined solely by monetary exchanges, but also encompass and depend upon a realm of voluntary, non-monetary associations, which Coase recognized are prohibitively costly to effect through monetary exchange because of the costs of defining separate contracts (Coase 1937). Because of these transactions costs, not only firms but also other institutions such as marriage and families emerge to avoid the costliness of pricing non-monetary attributes, such as love amongst marriage partners and parental devotion towards children: The great advantage of the market is that it is able to use the strength of self-interest to offset the weakness and partiality of benevolence, so that those who are unknown, unattractive, or unimportant, will have their wants served. But this should not lead us to ignore the part which benevolence and moral sentiments do play in making possible a market system. Consider, for example, the care and training of the young, largely carried out within the family and sustained by parental devotion. If love were absent and the task of training the young was therefore placed on other institutions, run presumably by people following their own self-interest, it seems likely that this task, on which the successful working of human societies depends, would be worse performed. (Coase 1976, p. 544)

Returning to Veblen’s critique of neoclassical economics, rational choice does not depend on ‘mechanical’ or ‘hedonistic’ responses to objective cost and profit functions. To counter Veblen, economics is in fact an evolutionary science, but one that is firmly grounded in an open-ended notion of rational choice that embraces both discovery under uncertainty. Alluding to the marginal cost controversy described above, Armen Alchian states the following in his classic article ‘Uncertainty, evolution, and economic theory’: While it is true that the economist can define a profit maximization behavior by assuming specific cost and revenue conditions, is there any assurance that the conditions and conclusions so derivable are not too perfect and absolute? If profit maximization (certainty) is not ascertainable, the confidence about the predicted effects of changes, e.g., higher taxes or minimum wages, will be dependent upon how close the formerly existing arrangement was to the formerly ‘optimal’ (certainty) situation. What really counts is the various actions actually tried, for it is from these that ‘success’ is selected, not from some set of perfect actions. The economist may be pushing his luck too far in arguing that actions in response to changes in environment and changes in satisfaction with the existing state of affairs will converge as a result of adaptation or adop-

M4225-ALTMAN_9781782549574_t.indd 74

03/05/2017 08:20

Rational choice as if the choosers were human 75 tion toward the optimum action that should have been selected, if foresight had been perfect. (Alchian 1950, p. 220)

What Alchian is arguing is that neither the behavioral assumption of profit maximization nor perfect foresight is necessarily required ex ante for human rationality. What is sufficient is awareness of the institutional conditions within which human rationality manifests itself ex post: Even if each and every individual acted in a haphazard and nonmotivated manner, it is possible that the variety of actions would be so great that the resulting collective set would contain actions that are best, in the sense of perfect foresight . . . The essential point is that individual motivation and foresight, while sufficient, are not necessary. Of course, it is not argued here that therefore it is absent. All that is needed by economists is their own awareness of the survival conditions and criteria of the economic system and a group of participants who submit various combinations and organizations for the system’s selection and adoption. (Alchian 1950, pp. 215–17)

Regardless of the behavioral assumptions, given the ubiquitous presence of scarcity, rational human action (that is, the continuous application of discovered means to individual aims) will generate the contextual knowledge, manifested through the price system, for the pursuit of maximum profits given the particular institutional context (Boettke and Candela 2015). The science of economics analyzes how fallible, but capable individuals do their best under particular institutional constraints. The art of economics, however, is uncovering those institutional constraints for understanding how in each particular case individuals attempt to do their best: Even if environmental conditions cannot be forecast, the economist can compare for given alternative potential situations the types of behavior that would have higher probability of viability or adoption. If explanation of past results rather than prediction is the task, the economist can diagnose the particular attributes which were critical in facilitating survival, even though individual participants were not aware of them. (Alchian 1950, p. 216)

The outside observer of human behavior who assesses some particular action as ‘irrational’ makes his or her evaluation based on either a value judgment of the ends pursued or narrowly defining the actor’s utility function to fit a close-ended model of choice. Criticisms of Homo economicus have been based both on the former, namely, by challenging efficiency as an ethical benchmark, as well as the latter by subjecting the model to narrowly defined monetary motives. As Elinor Ostrom states, this ‘thin model of rationality needs to be viewed . . . as the limiting case of bounded or incomplete rationality’ (1998, p. 9), that emerges only after all the gains from trade and specialization have been exhausted. However, ‘as we move away from these conditions we must explore not only the immediate consequences in terms of choices but particularly the kinds of institutions that will evolve in such contexts to structure human interaction’ (North 1993, p. 161). Rational action understood among mainline economists refers to the discovery of the means applied towards the fulfillment of a particular end; it does not necessarily depend on all our preferences and means being given or specified in one’s utility function. ‘Consistent with all models of rational choice is a general theory of human behavior that views all humans as complex, fallible learners who seek to do as well as they can given the

M4225-ALTMAN_9781782549574_t.indd 75

03/05/2017 08:20

76

Handbook of behavioural economics and smart decision-making

constraints that they face and who are able to learn heuristics, norms, rules, and how to craft rules to improve achieved outcomes’ (Ostrom 1998, p. 9). It encompasses learning from our errors through time, without which institutions would not matter (North 1994). Thus, the innumerable manifestations of rationality depend on the institutional context within which learning through time takes place. Therefore, as Vernon Smith explains, what seems to be ‘irrational’ to the outside observer of human behavior is no more than a misunderstanding of the institutional context: Thus, if people in certain contexts make choices that contradict our formal theory of rationality, rather than conclude that they are irrational, some ask why, reexamine maintained hypotheses including all aspects of the experiments – procedures, payoffs, context, instructions, etc. – and inquire as to what new concepts and experimental designs can help us to better understand the behavior. (Smith 2003, p. 471)

Moreover, institutional analysis does not imply that rules will always be perfectly specified or that individuals respond passively to the institutional reward structure. Rather, because of the cost of defining all of the possible actions that may be prohibited or sanctioned by the institutional framework, entrepreneurial discovery by individuals will generate endogenous institutional solutions to problems that are institutional in nature, resulting in gradual changes to the institutional framework. In the next section, we elaborate on the insights of Hayek, Buchanan and Ostrom in shifting to the rule level of analysis in analyzing human decision making and human interaction.

3

DEVELOPMENT OF THESE INSIGHTS TO THE RULE LEVEL OF ANALYSIS

We have previously argued that exposition of economic phenomena in terms of competitive equilibrium as a description of reality rather than using equilibrium analysis as a heuristic tool had rendered institutional analysis of little concern to economists. By extending the pure logic of rational choice to closed-form solutions, real-world markets act ‘as if’ they were in competitive equilibrium, not only purging the analysis of institutional derivations of the logic of choice, but also resulting in economic analysis becoming increasingly reliant on behavioral assumptions characterized as Homo economicus, around which advocates and critics of the market order would base their arguments.2 By collapsing the gap between economic models and economic reality, the behavioral intentions of economic actors correspond one-to-one with the outcomes ‘predicted’ within the model (Boettke and Candela 2014). What Buchanan, Hayek, and Ostrom acknowledged was that a richer notion of economic theory included institutional analysis and that incorporating institutional analysis enabled economic analysis to explain a broader notion of rational choice as if choosers were human. Moreover, what distinguished them from earlier classical as well as neoclassical economists was their application of rational choice to the rule level of analysis as well. Unlike the behavioral and physical laws of nature, which they took as given, what they explicitly drew attention to was that ‘rules are interesting variables precisely because they are potentially subject to change. That rules can be changed by humans is one of their key characteristics’ (Ostrom 1986, p. 5).

M4225-ALTMAN_9781782549574_t.indd 76

03/05/2017 08:20

Rational choice as if the choosers were human 77 Hayek as early as 1937 in ‘Economics and knowledge’ recognized that rational choice analysis, or what he referred to as the pure logic of choice, was a necessary, though not a sufficient, condition for equilibrium analysis. What was sufficient for the derivation of equilibrating tendencies within the market order, however, was a shift to the rule level of analysis, or comparative institutional analysis. Fundamentally, the importance of rules to Hayek was to provide a framework of predictable guidelines within which individuals could adapt to unforeseen circumstances. As he states: We can produce the conditions for the formation of an order in society, but we cannot arrange the manner in which its elements will order themselves under appropriate conditions. In this sense the task of the lawgiver is not to set up a particular order but merely to create conditions in which an orderly arrangement can establish and ever renew itself. As in nature, to induce the establishment of such an order does not require that we be able to predict the behavior of the individual atom – that will depend on the unknown particular circumstances in which it finds itself. All that is required is a limited regularity in its behavior; and the purpose of the human laws we enforce is to secure such limited regularity as will make the formation of an order possible. (Hayek 1960, p. 161)

Ostrom also acknowledged that rules ‘are the result of implicit or explicit efforts by a set of individuals to achieve order and predictability within defined situations’ (1986, p. 5). More so than Buchanan and Ostrom, Hayek emphasized that rules emerged from a spontaneous order based on human action, though not of human design. However, like Buchanan and Ostrom, he also acknowledged that rules that have evolved spontaneously can also be improved upon by marginal deliberative choices on the level of rules to facilitate different patterns of interactions within those rules: At the moment our concern must be to make clear that while the rules on which spontaneous order rests may also be of spontaneous origin, this need not always be the case. Although undoubtedly an order originally formed itself spontaneously because the individuals followed rules which had not been deliberately made but had arisen spontaneously, people gradually learned to improve those rules; and it is at least conceivable that the formation of a spontaneous order relies entirely on rules that were deliberately made. (Hayek 1973, p. 45)

The rule level of analysis requires neither that rational agents are homogenous in their objectives, nor does it imply that they share only pecuniary aims, such as that attributed to Homo economicus. As Buchanan states: The central rationality precept states only that the individual choose more rather than less of goods, and less rather than more of bads. There is no requirement that rationality dictates choice in accordance with the individual’s economic interest, as this might be measured by some outside observer of behavior. The individualistic postulate allows the interests or preferences of individuals to differ, one from another. And the rationality postulate does not restrict these interests beyond the classificatory step noted. Homo economicus, the individual who populates the models of empirical economics may, but need not, describe the individual whose choice calculus is analyzed in constitutional political economy. (Buchanan 1990, pp. 14–15)

Buchanan argued that economists could analyze the derivation of that framework separately through the tools of rational choice political philosophy, namely, social contract theory, but differently from Hayek and Ostrom, argued that institutions were provided

M4225-ALTMAN_9781782549574_t.indd 77

03/05/2017 08:20

78

Handbook of behavioural economics and smart decision-making

exogenously in the first place. Ostrom, building more from Hayek in this regard, saw the framework itself as an endogenous set of rules that emerge from the bottom up, through the individual and group striving to minimize conflicts and realize gains from cooperation. It is the notion of ‘constitutional craftsmanship’ that is foundational to the work of Ostrom that provides the conciliatory link between the dual spontaneous order analysis argued for here and the restriction of spontaneous order analysis to the market process, argued notably by Buchanan. Yet the common thread uniting their shift to the rule level of analysis was that the ability for individuals to coordinate their actions fell within a paradigm of exchange behavior to achieve a more preferred arrangement of rules in order to facilitate outcomes conducive to cooperation through a division of labor. What they did not share was a vision of political economy through an ‘allocation paradigm’ (Buchanan 1964), in which rational choosers were purged of any human deliberation as well as confined to perfectly defined constraints not subject to change and improvements by human rational choosers themselves.

4

THE INSTITUTIONAL IMPRINT, RATIONAL CHOICE, AND PATH DEPENDENCY

The fundamental task that has plagued economists since Adam Smith, both mainline and mainstream, has been to inquire into the nature and causes of the wealth of nations. Particularly since the collapse of communist regimes in Central and Eastern Europe after 1989, this inquiry has been increasingly marked by a dovetailing of the mainline and mainstream through its emphasis on comparative institutional analysis and institutional path dependency in understanding the lag in economic development not only among countries emerging from communism, but also in Asia, Africa, and Latin America. The plain fact is that the ultimate source of poor economic performance in third-world countries is the polity that specifies and enforces the economic rules of the game. As yet the study of third-world polities is as underdeveloped as their polities themselves. But one thing is for sure: not much progress is going to be made in modeling such polities without taking into account the limits of rational choice and the importance of ideologies. (North 1993, pp. 160–61)

The point that Douglass North makes in this quote is that not only do the formal rules of the game matter for the economic performance of a country, but also that informal constraints provide path dependency in cultural attitudes towards trade and exchange. As North argues, ‘once a development path is set on a particular course, the network externalities, the learning process of organizations, and the historically derived subjective modeling of the issues reinforce the course’ (North 1990, p. 99). In this respect, neoclassical economics could not underpin the reform of centrally planned economies for two particular reasons. First, the theoretical model of perfect competition operates as a behavioral filter of the limiting conditions that apply to individuals after all the gains from trade and specialization have been exploited. It illustrates an idealized world in which individuals are fully informed and perfectly rational in making their decisions, and that the constraints they face, such as prices and income, are taken as given. However, lacking any institutional filter of the structure of incentives that gener-

M4225-ALTMAN_9781782549574_t.indd 78

03/05/2017 08:20

Rational choice as if the choosers were human 79 ates tendencies towards such limiting conditions, deviations from this behavioral model can only lead the economist to conclude that individuals are behaving ‘irrationally’ and that the economy is prone to ‘market failure,’ characterized by the prevalence of asymmetric information, externalities, and monopoly power. As Ronald Coase argued, ‘These ex-communist countries are advised to move to a market economy, and their leaders wish to do so, but without the appropriate institutions no market economy of any significance is possible. If we knew more about our own economy, we would be in a better position to advise them’ (1992, p. 714). Secondly, as Coase alluded to above, without understanding the ‘tacit presuppositions’ (Buchanan 1997) that are embodied in the underlying norms, customs, and behavioral attitudes of individuals that reinforce the prevailing institutions of society, it would be unclear whether the institutional designs of economic reformers would have the intended effect on the economic performance of a country. James Buchanan clearly makes this point: In Western regimes, markets work tolerably well, within the political-legal framework of widely dispersed property rights, when the workings of ordinary politics do not interfere excessively. They do so because they have evolved through a long history which has been interpreted and understood by experts in such fashion as to reinforce the behavioural attitudes necessary to make such institutions function. In failed socialist regimes, markets have neither the history nor the interpretation-understanding that informs behavioural attitudes. It seems naïve in the extreme to assume that the market order is ‘natural’ to the extent that it can emerge full blown without history, without institutional construction and, most importantly, without understanding. (1997, p. 106)

It is not just that institutions matter, but history and ideas matter for understanding the feasible institutional opportunity set within which the economist is able to propose reforms. Furthermore, Buchanan makes a distinction between an ‘exchange culture’ and a ‘command culture’ to distinguish the underlying behavioral attitudes in Western and post-socialist economies, respectively. To understand this point, Buchanan is neither denying that individuals are choosing rationally, nor is relying on any notion of Homo economicus. Rather, it is the underlying informal norms prevalent throughout members of society that motivate the degree of toleration and acknowledgement of the mutually beneficial nature of trade under anonymity, which fundamentally extends the limit of the market, and widens the scope for rational self-interest to encompass activity beyond the behavioral confines of Homo economicus. In Western countries, Buchanan argues the following: The attitude of reciprocity in the market relationships remains relatively pervasive in Western economies, even in those settings where, in a behavioural sense, there remains little or no rational foundations for such attitude. The salesclerk in the Sheraton Hotel in Houston, Texas, offers me a postcard as if she, personally, has an interest in my purchase, even when both of us know that her wage or position depends only in some extremely remote sense on her behaviour in our exchange relationship. The exchange relationship tends to foster the attitude of reciprocity, even in as if settings, and behavior reflecting such an attitude tends in itself to promote a mutuality of expectations that becomes reinforcing. (1997, p. 97, original emphases)

In those countries that have failed to emerge successfully, in terms of economic growth, from the failures of socialism, the underlying informal norms of society are conducive to

M4225-ALTMAN_9781782549574_t.indd 79

03/05/2017 08:20

80

Handbook of behavioural economics and smart decision-making

a command culture, one that ‘describes an idealization of collective reality, as this reality is interpreted by those who experience it’. It reinforces the idea that the ‘exertion of effort creates no claim to share in product. Effort is directed toward common purposes, and linkage between effort and reward becomes the source of envy rather than emulation’ (Buchanan 1997, p. 101, original emphasis). From this cultural context, it would therefore seem rational for individuals to regard the market order with suspicion, especially when viewed in the zero-sum terms of a command culture. Extending Buchanan’s point, Pejovich elaborates: In many parts of the region, gains from trade are seen as a redistribution of income rather than as rewards to innovators for creating new wealth. State authorities are more likely to impose price controls on producers and/or merchants who earn large profits than to seek ways to create incentives for others to emulate such individuals in competitive markets. The cultural heritage in [Central and Eastern Europe] supports an activist state. Historical development and nationalism are major reasons for cultural differences within the region . . . By feeding on the conviction that the community’s common good transcends the private ends of its members, nationalism in many [Central and Eastern European] countries has reinforced the culture of collectivism. (2003, p. 351)

Although our discussion thus far has emphasized a comparative institutional analysis between Western economies and the economies of Central and Eastern Europe, our observations have broader implications for income differences across time as well. Not only do the formal institutions matter for economic growth, but, perhaps more importantly, the fact that customary practice dictates the legitimacy of formal institutions is because informal rules ‘are not a policy variable’ (Pejovich 2003, p. 348) and, therefore, formal institutions must be crafted to be congruent, or ‘stick’ to the underlying informal rules of society. Although economic reformers may succeed in designing institutions that exhibit ‘institutional stickiness’ (Boettke et al. 2008) to informal institutions, this is by no means sufficient for generating economic growth: There is no guarantee that beliefs and institutions that evolve through time will produce economic growth . . . In fact, most societies throughout history got ‘stuck’ in an institutional matrix that did not evolve into the impersonal exchange essential to capturing the productivity gains that came from the specialization and the division of labor that have produced the Wealth of Nations. (North 1994, pp. 363–4)

With respect to the logic of rational choice, societies that exhibit path dependency towards economic stagnation does not imply irrationality on the part of economic actors within that society. Rationality must be understood as entirely subjective and forward looking; an individual’s perception of costs and benefits are shaped by the institutional payoffs: In every system of exchange, economic actors have an incentive to invest their time, resources, and energy in knowledge and skills that will improve their material status. But in some primitive institutional settings, the kind of knowledge and skills that will pay off will not result in institutional evolution towards more productive economies. (North 1991, p. 102)

As North elaborates on how institutional path dependence can be self-sustaining: In the case of economic growth, an adaptively efficient path . . . allows for a maximum of choices under uncertainty, for the pursuit of various trial methods of undertaking activities, and for an

M4225-ALTMAN_9781782549574_t.indd 80

03/05/2017 08:20

Rational choice as if the choosers were human 81 efficient feedback mechanism to identify choices that are relatively inefficient and to eliminate them . . . But so, too, can unproductive paths persist. The increasing returns characteristic of an initial set of institutions that provide disincentives to productive activity will create organizations and interest groups with a stake in the existing constraints. They will shape the polity in their interests. Such institutions provide incentives that may encourage military domination of the polity and economy, religious fanaticism, or plain, simple redistributive organizations, but they provide few rewards from increases in the stock and dissemination of economically useful knowledge. The subjective mental constructs of the participants will evolve an ideology that not only rationalizes the society’s structure but accounts for its poor performance. As a result the economy will evolve policies that reinforce the existing incentives and organizations. (North 1990, p. 99)

However, the observation that certain societies are locked into an institutional path unconducive to economic growth does not necessarily imply that intervention from reformers external to the particular institutional context will resolve the situation, namely, by establishing private property rights or transplanting other formal institutions that evolved within the historical context of Western economic development. As Ostrom has argued: When analysts perceive the human beings they model as being trapped inside perverse situations, they then assume that other human beings external to those involved – scholars and public officials – are able to analyze the situation, ascertain why counterproductive outcomes are reached, and posit what changes in the rules-in-use will enable participants to improve outcomes. Then, external officials are expected to impose an optimal set of rules on those individuals involved. It is assumed that the momentum for change must come from outside the situation rather than from the self-reflection and creativity of those within a situation to restructure their own patterns of interaction. (2010, p. 648)

Ostrom recognized that when the definition and enforcement of property rights are devised from the bottom up rather than from the top down, individuals will have a greater incentive to conserve on resources used in the process than when that process is imposed exogenously, not only because they exhibit greater residual claimancy over their actions, but also because they are able to utilize their contextual knowledge not often available to external reformers. The ability of individuals to craft rules that are effective within their own communities hinges upon the mutually agreed-upon rules of governance that then establish reliable expectations among the community. Elinor Ostrom emphasized the legitimacy of rules as essential to minimizing the enforcement and monitoring costs of rules (1990, p. 205). If rules are developed internally, by actors with local legitimacy and knowledge of the community’s history, then monitoring can be a ‘natural by-product’ of the system of rules (Ostrom 1990, p. 96). In addition, because of the path-dependent nature of bottom-up institutional solutions, formal enforcement of rules cannot run contrary to how individuals perceive and understand them: Without individuals viewing rules as appropriate mechanisms to enhance reciprocal relationships, no police force and court system on earth can monitor and enforce all the needed rules on its own. Nor would most of us want to live in a society in which police were really the thin blue line enforcing all rules. (Ostrom 1998, p. 16)

A summary of the core argument of this section can be restated as follows (Boettke 2001, pp. 251–9):

M4225-ALTMAN_9781782549574_t.indd 81

03/05/2017 08:20

82

Handbook of behavioural economics and smart decision-making

1. 2. 3.

People respond rationally to incentives as they perceive them. Incentives and therefore economic performance are a function of the rules of the game, both formal and informal. Rules are only rules if customary practice dictates.

5

CONCLUSION

In this chapter, we have argued that the caricature of economic man as Homo economicus has been an invalid and unwarranted characterization of the individual in their interactions with other individuals within the market order. From Adam Smith to Vernon Smith there has been a common thread that has united economic thought on man’s epistemic and behavioral capacity, one that rests on institutional analysis and the emergence of customs, norms, and monetary prices to guide social interaction towards social order. Hayek best summarizes this common lineage in mainline economic thought: Perhaps the best illustration of the current misperceptions of the individualism of Adam Smith and his group is the common belief that they had invented the bogey of ‘economic man’ and that their conclusions are vitiated by their assumption of a strictly rational behavior or generally by a false rationalistic psychology. They were, of course, very far from assuming anything of the kind. It would be nearer to the truth to say that in their view man was by nature lazy and indolent, improvident and wasteful, and that it was only by the force of circumstances that he could be made to behave economically or carefully to adjust his means to his ends . . . The main point about which there can be little doubt is that Smith’s chief concern was not so much with what man might occasionally achieve when he was at his best but that he should have as little opportunity as possible to do harm when he was at his worst. It would be scarcely too much to claim that the main merit of the individualism which he and his contemporaries advocated is that it is a system under which bad men can do least harm. It is a social system which does not depend for its functioning on our finding good men for running it, or on all men becoming better than they now are, but which makes use of men in all their given variety and complexity, sometimes good and sometimes bad, sometimes intelligent and more often stupid. (1948, pp. 11–12)

The excessive preoccupation with the behavioral characteristics of man in economic analysis from the late nineteenth century to the mid-twentieth century resulted not only from misunderstandings about the role of theory and history (Mises 1957) in economic analysis, but also from depicting economic phenomena in terms of competitive equilibrium. Because facts are theory laden, the purpose of economy theory is to engage in historical explanation of facts. To argue that man is rational – that is, that he or she evaluates the marginal costs and benefits of undertaking an activity towards the fulfillment of a particular goal – does grant that individual infallibility or omniscience. This is the realm of price theory, consistent with the understanding of mainline economists discussed in this chapter. Rather, such decision making and the manifestations of rationality must be evaluated within its particular institutional context. This is the realm of history. The modern analytical narrative approach employed by Bates et al. (1998) embodies this distinction between theory and history that Mises specified: The process of deciding the appropriate individuals, their preferences, and the structure of the environment – that is, the right game to use – is an inductive process much like that used in modern comparative politics, by historical institutionalists, and by historians. Once that induc-

M4225-ALTMAN_9781782549574_t.indd 82

03/05/2017 08:20

Rational choice as if the choosers were human 83 tion is complete, we can use the deductive methods to study behavior within the context of the game. (Bates et al. 2000, p. 697).

However, when economic theory becomes dependent on behavioral assumptions, not only does this tend to crowd out institutional analysis, collapsing human intentions on to outcomes, but it also leads to misleading characterizations of human ‘irrationality’ when historical events, such as the Great Depression or more recently the Great Recession, are not ‘predicted’ by theory or do not correspond to some particular efficiency benchmark. As a result, arguments about the success of comparative economic systems, even those in defense of the market order, will hinge upon the behavioral capabilities of man. The predictive power of mainstream theorizing in macroeconomics, both Keynesian as well as new classical, no less depend on whether individuals are hopelessly irrational or perfectly rational, respectively. With the resulting disconnect of theory from history, new emphasis was drawn to the rule level of analysis, which had not only been emphasized by Adam Smith and Frank Knight, but was reincarnated as new insights manifesting themselves as the public choice of James Buchanan, new institutional economics of Armen Alchian, Ronald Coase and Elinor Ostrom, the market process of modern Austrian economics developed by Mises and Hayek, the experimental economics of Vernon Smith, and the new economic history of Douglass North. Each of these scholars, while rejecting the notion of Homo economicus, did not throw rational choice by the wayside either. Instead, their contributions to economics were built upon ‘Adam Smith’s view of man’ as ‘he actually is – dominated, it is true, by self-love but not without some concern for others, able to reason but not necessarily in such a way as to reach the right conclusion, seeing the outcomes of his actions but through a veil of self-delusion’ (Coase 1976, pp. 545–6). While we do not deny that ‘Adam Smith frequently wrote about the psychology of decision-making’ (Laibson and List 2015, p. 385), accepting this view of man, who is a fallible but capable individual, will draw the economist’s attention, whether mainline or mainstream, behavioral or traditional, to the realization that man’s capacity to foster social order and capture the gains from exchange and innovation depends not on his reason or lack of reason per se, but on rules that are discovered and crafted to marshal individual reasoning dispersed among individuals throughout society.

NOTES 1. Economist Mark Thoma, in a 2015 article, goes so far as to say ‘I believe in markets as much as anyone. But for markets to work well the conditions for perfect competition must be approximated’. The notion that markets only ‘work’ after all the gains from trade and exchange of goods and services (including information) have been exploited (that is, perfect competition) is a description of the end result of market competition, not an analysis of the economic forces at work. Moreover, this characterization of the market ‘working well’ not only commits the nirvana fallacy (see Demsetz 1969) of comparing imperfect markets to a perfectly efficient benchmark, but more importantly, it also lacks comparative institutional analysis of market forces under different conditions. 2. During the socialist calculation debate between the 1920s and 1940s, market socialists, arguing against Mises’s claim that rational calculation was impossible under socialism, utilized neoclassical equilibrium analysis to establish the invalidity of Mises’s claim. Mises’s as well as Hayek’s fundamental argument during the debate held that, absent the institutional prerequisites of private property, central planners would be precluded from the prices and contextual knowledge required to engage in economic calculation. However,

M4225-ALTMAN_9781782549574_t.indd 83

03/05/2017 08:20

84

Handbook of behavioural economics and smart decision-making presuming the conditions of equilibrium and that perfect knowledge was available to central planners, market socialist Oskar Lange argued that central planners would be able to calculate the opportunity costs of resources, just as the market supposedly does, through a trial and error process. Surprisingly, Frank Knight and Joseph Schumpeter, both market-oriented economists of a Marshallian and Walrasian stripe, respectively, also agreed with Lange’s analytical assessment. Knight’s argument was that there was no economic problem of socialism because economic science is limited to applying marginalist principles to economic decision-making in circumstances of perfect knowledge and perfect competition (see Boettke and Vaughn 2002, p. 159). Schumpeter’s argument was that, assuming the conditions of general competitive equilibrium, the valuation of the factors of production can be logically imputed ‘ipso facto’ directly from the valuation of consumers’ goods (see Hayek 1945, pp. 529–30).

REFERENCES Alchian, A. (1950), ‘Uncertainty, evolution, and economic theory’, Journal of Political Economy, 58 (3), 211–21. Altman, M. (2012), Behavioral Economics for Dummies, Mississauga: John Wiley & Sons Canada. Bates, R.H., A. Greif, M. Levi, J.-L. Rosenthal and B. Weingast (1998), Analytical Narratives, Princeton, NJ: Princeton University Press. Bates, R.H., A. Greif, M. Levi, J. Rosenthal and B. Weingast (2000), ‘Review: the Analytical Narrative Project’, American Political Science Review, 94 (3), 696–702. Boettke, P.J. (2001), ‘Why culture matters: economics, politics, and the imprint of history’, in P.J. Boettke (ed.), Calculation and Coordination: Essays on Socialism and Transitional Political Economy, New York: Routledge, pp. 248–65. Boettke, P.J. (2012), Living Economics: Yesterday, Today, and Tomorrow, Oakland, CA: Independent Institute. Boettke, P. (2015), ‘The methodology of Austrian economics as a sophisticated, rather than naive, philosophy of economics’, Journal of the History of Economic Thought, 37 (1), 79–85. Boettke, P. and R. Candela (2014), ‘Alchian, Buchanan, and Coase: a neglected branch of Chicago price theory’, Man and the Economy, 1 (2), 189–208. Boettke, P. and R. Candela (2015), ‘What is old should be new again: methodological individualism, institutional analysis and spontaneous order’, Sociologia, 49 (2), 5–14. Boettke, P. and K. Vaughn (2002), ‘Knight and the Austrians on capital and the problem of socialism’, History of Political Economy, 34 (1), 153–74. Boettke, P., C. Coyne and P. Leeson (2008), ‘Institutional stickiness and the new development economics’, American Journal of Economics and Sociology, 67 (2), 331–58. Buchanan, J. (1964), ‘What should economists do?’, Southern Economic Journal, 30 (3), 213–22. Buchanan, J. (1990), ‘The domain of constitutional economics’, Constitutional Political Economy, 1 (1), 1–18. Buchanan, J.M. (1997), Post-Socialist Political Economy: Selected Essays, Cheltenham, UK and Lyme, NH, USA: Edward Elgar. Coase, R. (1937), ‘The nature of the firm’, Economica, 4 (16), 386–405. Coase, R. (1976), ‘Adam Smith’s view of man’, Journal of Law and Economics, 19 (3), 529–46. Coase, R. (1992), ‘The institutional structure of production’, American Economic Review, 82 (4), 713–19. Demsetz, H. (1969), ‘Information and efficiency: another viewpoint’, Journal of Law and Economics, 12 (1), 1–22. Friedman, M. (1953), ‘The methodology of positive economics’, in M. Friedman, Essays in Positive Economics, Chicago, IL: University of Chicago Press, pp. 3–43. Hayek, F.A. (1937), ‘Economics and knowledge’, Economica, 4 (13), 33–54. Hayek, F.A. (1945), ‘The use of knowledge in society’, American Economic Review, 35 (4), 519–30. Hayek, F.A. (1948), Individualism and Economic Order, Chicago, IL: University of Chicago Press. Hayek, F.A. (1960), The Constitution of Liberty, Chicago, IL: University of Chicago Press. Hayek, F.A. (1973), Law, Legislation, and Liberty, Volume I: Rules and Order, Chicago, IL: University of Chicago Press. Keynes, J.M. (1926), ‘The end of laissez faire’, reprinted 1978 in E. Johnson and D. Mogridge (eds), The Collected Writings of John Maynard Keynes, Volume IX: Essays in Persuasion, New York: Cambridge University Press. Keynes, J.M. (1936), The General Theory of Employment, Interest, and Money, reprinted 1964, New York: Harcourt, Brace & World. Kirzner, I. (1988), ‘The economic calculation debate: lessons for Austrians’, Review of Austrian Economics, 2 (1), 1–18. Knight, F. (1919), ‘Review of Co-operation and the Future of Industry by Leonard W. Woolf’, Journal of Political Economy, 27 (9), 805–6.

M4225-ALTMAN_9781782549574_t.indd 84

03/05/2017 08:20

Rational choice as if the choosers were human 85 Knight, F. (1920), ‘Social organization: a survey of its problems and forms from the standpoint of the present crisis’, Research in the History and Methodology of Economic Thought, vol. 29-B, 65–88, reprinted 2011, Bradford: Emerald Publishing. Laibson, D. and J. List (2015), ‘Principles of (behavioral) economics’, American Economic Review, 105 (5), 385–90. Lavoie, D. (1990), ‘Hermeneutics, subjectivity, and the Lester/Machlup debate: toward a more anthropological approach to empirical economics’, in W.J. Samuels (ed.), Economics as Discourse: An Analysis of the Language of Economists, Dordrecht: Springer, pp. 167–87. Machlup, F. (1955), ‘The problem of verification in economics’, Southern Economic Journal, 22 (1), 1–21. Mises, L. von (1933), Epistemological Problems of Economics, reprinted 1960, Princeton, NJ: Van Nostrand. Mises, L. von (1949), Human Action: A Treatise on Economics, reprinted 2007, Indianapolis, IN: Liberty Fund. Mises, L. von (1957), Theory and History: An Interpretation of Social and Economic Evolution, New Haven, CT: Yale University Press. North, D. (1990), Institutions, Institutional Change, and Economic Performance, New York: Cambridge University Press. North, D. (1991), ‘Institutions’, Journal of Economic Perspectives, 5 (1), 97–112. North, D. (1993), ‘What do we mean by rationality?’, Public Choice, 77 (1), 159–62. North, D. (1994), ‘Economic performance through time’, American Economic Review, 84 (3), 359–68. Ostrom, E. (1986), ‘An agenda for the study of institutions’, Public Choice, 48 (1), 3–25. Ostrom, E. (1990), Governing the Commons: The Evolution of Institutions for Collective Action, New York: Cambridge University Press. Ostrom, E. (1998), ‘A behavioral approach to the rational choice of collective action’, American Political Science Review, 92 (1), 1–22. Ostrom, E. (2010), ‘Beyond markets and states: polycentric governance of complex economic systems’, American Economic Review, 100 (3), 641–72. Pejovich, S. (2003), ‘Understanding the transaction costs of transition: it’s the culture, stupid’, Review of Austrian Economics, 16 (4), 347–61. Prasch, R. (2007), ‘Professor Lester and the neoclassicals: the “marginalist controversy” and the postwar academic debate over minimum wage legislation: 1945–1950’, Journal of Economic Issues, 41 (3), 809–25. Smith, V. (2003), ‘Constructivist and ecological rationality in economics’, American Economic Review, 93 (3), 465–508. Thoma, M. (2015), ‘The problem with completely free markets’, accessed 2 July 2015 at http://www.thefiscal times.com/columns/2015/06/30/problem-completely-free- markets?onswipe_redirect5no&oswrr51. Veblen, T. (1898), ‘Why is economics not an evolutionary science?’, Quarterly Journal of Economics, 12 (4), 373–97. Zanotti, G. and N. Cachanosky (2015), ‘Implications of Machlup’s interpretation of Mises’s epistemology’, Journal of the History of Economic Thought, 37 (1), 111–38.

M4225-ALTMAN_9781782549574_t.indd 85

03/05/2017 08:20

5

Smart predictions from wrong data: the case of ecological correlations* Florian Kutzner and Tobias Vogel

INTRODUCTION With the economic crisis that hit the European continent in 2007, many were interested in the consequences for those affected, from economic hardship, to emotional well-being to family planning. Take the relationship between the severity of the crisis and family planning as an illustration. Across 28 European countries, it was found that an increase in unemployment rates from 2007 to 2011 correlated with a decrease in birth rates (Goldstein et al., 2013). To illustrate, we plotted changes in unemployment rates against the respective changes in birth rates (see Figure 5.1). The correlation of –.34 indicates that those countries with larger increases in unemployment rates are marked by larger decreases in birth rates. It seems there exists ‘a strong relationship between economic conditions and fertility’ (Goldstein et al. 2013, p. 85) and that ‘the extent of joblessness . . . does in fact have an effect on birth rates’ (BBC 2013). Can we conclude, however, that people who lost their job delayed having children? In this chapter we show that people respond ‘Yes’ to this question, that this answer can be wrong at times, and yet this response can still be smart or rational on average. The conclusion that the correlation between changes in unemployment and birth rates across countries also holds for individual people, that those becoming unemployed are

Change in birth rate

0.20 0.10 0.00 –0.10 –0.20 –0.30 –5%

0%

5%

10%

15%

Change in unemployment rate Note: Dashed line 5 regression line. Source:

Data from Eurostat.

Figure 5.1

Changes in unemployment rates (2011 minus 2007) in 28 European countries plotted against respective changes in birth rates 86

M4225-ALTMAN_9781782549574_t.indd 86

03/05/2017 08:20

Smart predictions from wrong data 87 more likely to delay having children, can be severely wrong. Yet, when considering real-life constraints on decision making, being ignorant about different levels of data aggregation offers a smart strategy. In many cases, using aggregated data is the only option owing to a lack of more appropriate data or a lack of resources for appropriate analyses. Also, even though some conclusions will be wrong, an illustrative demonstration lends credence to the fact that correlations across countries (or other ecologies) have some predictive value for correlations across individuals.

ECOLOGICAL CORRELATIONS AND AGGREGATION BIAS The term ‘ecological’ in ecological correlations is applicable to any grouping of observations. Talking about people, it is easy to conceive ecologies as macro variables such as geographical entities. Yet, many other ecological groupings of variables are almost equally plausible and prominent, including social characteristics such as income or age groups. An ecological correlation results when a correlation between two variables is computed using the variables’ mean values for different ecologies. In Figure 5.1 the ecologies are countries that are assigned mean values for two variables: change in unemployment rates and change in birth rates. Across 28 of these ecologies a correlation is computed. That is, statistically the correlation is based on 28 observations. At the same time, thousands of individuals are behind the mean values, each of which has either become unemployed or not, and has become a parent or not. This individual information, however, is lost when aggregated into a country index. Which inferences about individuals are still valid given such aggregated data has puzzled statisticians, epidemiologists and sociologists for decades (Hannan 1971; Hammond 1973; King 2013). With aggregation comes greater reliability for estimating mean values. Asking 30 individuals about their change in employment status will be a less reliable estimate of the mean tendency than asking 1000 individuals. At the same time, correlations between these mean values across a handful of ecologies do not become more reliable. Even more dramatic than losing the reliability benefit, the size and even sign of correlations can differ from before to after aggregation. This sets the stage for ‘aggregation biases’ or ‘ecological fallacies’ (Hammond 1973). An early demonstration of diverging correlations given individual and aggregated data can be found in Robinson’s work (1950). Across nine US districts Robinson showed that the correlation between the averages of African-Americans and illiterates living in these districts was about r ≈ .95. The districts with higher rates of African-Americans had almost certainly higher illiteracy rates. At the same time, using individual data showed that the likelihood of being illiterate was barely higher for African-Americans than for non-African-Americans, only resulting in a correlation of r ≈ 0.2. An analogous divergence can be found with the effect of the economic crisis on birth rates. An analysis based on the changes in job and parental status of individual people, across a comparable time span and set of countries, reveals that becoming unemployed does not, on average, affect becoming a parent (Schmitt 2012). There are several reasons for a possible divergence (Hannan 1971; Hammond 1973). Perhaps the most obvious is the presence of an unobserved third variable that influences both measured variables. For example, allocation of people to districts might be

M4225-ALTMAN_9781782549574_t.indd 87

03/05/2017 08:20

88

Handbook of behavioural economics and smart decision-making

driven in part by socio-economic status. If status is correlated with ethnicity as well as with literacy, the percentage of African-Americans and the percentage of illiterates are likely to be correlated across districts. Within every district and in the whole population African-Americans can be as likely to be literate as their counterparts. In a similar vein, a third-variable – yet, with a different function – could explain the unemployment–birth rate correlation. If the economic downturn does increase subjective uncertainty, then everybody, not just those losing their jobs might delay having children. Yet, while many variables might cause a divergence between aggregated or ecological and individual-level correlations, it will be hard to know for a given question whether any of these are at work, especially for the incidental user of aggregated data. In the next sections of this chapter we elaborate on the use of ecological correlations from a descriptive and a normative perspective. We first report evidence to demonstrate that lay people are insensitive as to whether correlations are based on aggregated or individual-level data. Then, we will illustrate that such ecological inferences are potentially smart, and though error-prone, provide a good guess about individual-level correlations.

THE USE OF ECOLOGICAL CORRELATIONS IN LAY PEOPLE The technical term ecological correlations might obscure how prevalent they are as part of everyday life, especially in the media. The BBC coverage of the change in birth rates is but one example. Another widely covered article (Preis et al. 2012) found that the countrywide tendencies to Google for dates in the future, as compared to dates in the past, are positively related to a country’s gross domestic product (GDP). Conclusions such as ‘a focus on the future supports economic success’ (Preis et al., 2012, p. 2) were readily recited in the media (Wall Street Journal 2012; Guardian, 2013). But, again, what do readers conclude from such coverage? Ecological Inferences when Individual-level Data are Unavailable Before reviewing a host of pertinent laboratory research, let us present a brief and straightforward survey of what incidental consumers of the BBC news coverage conclude from the birth rate article. Shortly after the article appeared under the headline ‘Europe birth rates “have fallen” since economic crisis’, we had UK citizens (N 5 159) recruited online read either the headline or the entire article. The article included a figure similar to the one presented in Figure 5.1. Subsequently, participants were asked to answer the following question: ‘Who is more likely to delay having children?’ Respondents had three options: ‘The employed’, ‘The unemployed’ or ‘Don’t know’. A majority of readers of the headline (see the white bars in Figure 5.2) concluded that the unemployed would be more likely to delay having children. Only a few concluded that this was true for the employed or that they did not know. After only reading the headline, though, this pattern might results from a shared stereotype of what influences family planning. More interestingly, for those participants who were confronted with the whole article, including the ecological correlation, this trend intensified, c2 (2) 5 8.25, p 5 .016. That is, even though the correlation presented was not based on individuals, it made readers more certain that their conclusions about the correlation between employment

M4225-ALTMAN_9781782549574_t.indd 88

03/05/2017 08:20

Smart predictions from wrong data 89 ‘Who is more likely to delay having children?’ 100% BBC article

80%

BBC headline 60% 40% 20% 0% Unemployed

Employed

Don't know

Note: BBC headline 5 participants just receiving a headline stating ‘Europe birth rates “have fallen” since economic crisis’.

Figure 5.2

Responses to being confronted with an ecological correlation across countries (BBC article) between the change in unemployment rate and the change in birth rate

and parental status for individuals were correct. Of course, the conclusion is intuitively plausible and the argument well presented. Thus, the more general question is whether people draw similar conclusions in more controlled settings or when incentivized to be accurate. More systematic evidence for the use of ecological correlations by lay people stems from a series of controlled laboratory experiments. In one experiment (Vogel et al. 2013), we presented participants with information about a fictitious city. The graphical display drew on online newspaper formats. Participants saw a schematic map of the city, which was separated into nine districts. For each of the districts, there was information about the percentages of citizens belonging to an ethnic majority or an ethnic minority group, as well as about the percentage of people satisfied (versus dissatisfied) with their lives. As our critical manipulation, we varied the ecological correlation across districts. For half of the participants, the percentage of satisfied people increased with the number of majority members across districts. For the other half of participants, the percentage of satisfied citizens decreased with an increase of majority members. Asked about individual citizens, participants in the latter group judged majority members as less satisfied with their lives than minority members. This finding reversed for participants in the former group, who had been exposed to a positive ecological correlation between majority group and life satisfaction. A second experiment, using an interactive map corroborated this notion. The interactive format required participants to request the relevant information by clicking on the respective areas. Only participants who were motivated to compare different areas with regard to the relevant information learned the ecological correlation and transferred it to the individual judgments. Hence, using ecological correlations do not seem to reflect cognitive laziness. Instead, using ecological correlations requires attention to statistical regularities, which are taken to represent individuals as well.

M4225-ALTMAN_9781782549574_t.indd 89

03/05/2017 08:20

90

Handbook of behavioural economics and smart decision-making

The media often provides us with aggregate rather than individualized data. As is evident from the experiments reported above, consumers of such content readily use the provided ecological correlations to predict individual behavior or attributes. Yet, what happens when individual-level correlations are available? Ecological Inferences in Spite of Individual-level Zero Correlations We might conjecture that the use of ecological correlations as a proxy for individual data is a necessary evil owing to missing individual data. Indeed, in the case of missing information, lay people as well as scientists use the ecological correlation as a best guess. However, would we expect lay people to continue relying on ecological correlations in the presence of individual data? This is what a plethora of experimental learning studies suggests (for a review, see Fiedler et al. 2009). In one of the first studies to test the use of ecological correlations, Fiedler and Freytag (2004) presented participants with information about psychiatric patients. Participants learned about each patient’s test score on two personality tests, called type X and type Y. Across individuals, high scores in one test implied neither an increased or a decreased chance of a high score on the other test. In technical terms, the correlation between the test scores across individuals was zero. Yet, there was also an ecological correlation. Patients were additionally classified as belonging to one of two groups or ecologies: those for which previous psychotherapy had been successful and those for whom it had been unsuccessful. Among the successfully treated patients, high levels of both personality traits were three times as frequent as for patients with unsuccessful prior treatment. This created a perfect ecological correlation. The mean levels of both personality traits varied hand in hand across groups. The critical finding was that participants expected individuals with a high score on personality trait X to have a high score on trait Y. Thus, even with individual-level data easily accessible, a zero correlation is ignored and substituted with a positive ecological correlation. Ecological Inferences in Spite of Conflicting Individual-level Correlations Thus far, we have considered evidence for ecological correlations being used to infer correlations across individuals. We have implied that this correlation is the same within the different ecologies. For example, there was a zero individual-level correlation between the personality traits within both groups, for patients with successful and for patients with unsuccessful prior therapy. Yet, this might not always be the case. Another related yet different type of ecological inference fallacy has been demonstrated when the individual-level correlation for an entire population is different from the individual-level correlations within ecologies of this population. This constellation, in which different or even reverse individual-level correlations exist within ecologies, is known as Simpson’s paradox (Simpson 1951). Studies typically found that participants act as if they do not take the ecological variable into account reproducing the overall correlation across the entire population (for example, Schaller 1994). This neglect of an ecological variable seems at odds with claiming there is a strong influence of ecological correlations. Intriguingly, evidence that has been used to support the idea that the ecological variable is neglected in a Simpson’s paradox can equally well

M4225-ALTMAN_9781782549574_t.indd 90

03/05/2017 08:20

Smart predictions from wrong data 91 be reconciled with the more recent idea that ecological correlations, not overall individuallevel correlations, are used. Meiser and Hewstone (2004) provided evidence that favors this ecological correlation explanation. In a setup typical for a Simpson’s paradox, their participants were presented with a series of positive and negative behavioral descriptions involving members of two groups, A and B, distributed over two towns, X and Y. While the individual-level correlation was positive across all individuals, say, being from group A was predictive of behaving positively, it was negative within both towns. Such a setup, however, implies that in one town group A members and in the other group B members form the majority. At the same time, one town is characterized by many and the other by few positive behaviors. Thus, in creating the Simpson’s paradox, an ecological correlation results between being a member of group A and positive behavior. This is because the average ‘group-A-ness’ for a town is perfectly predictive of the average positivity. When asked to make predictions about individuals belonging to either of the groups and residing in either of the towns, participants predicted group A members to be more likely to behave in a positive way, contrary to the correlation actually observed within the towns, but in line with both the ecological correlation and the individual-level correlation neglecting the town variable. Thus, either participants neglected the ecological town variable and made judgments based on an individual-level correlation or they relied on the ecological correlation. Speaking in favor of the ecological correlation explanation, it was particularly those participants who had accurately learned the ecological correlation, that is, which town was characterized by a majority of group A members and a prevalence of positive behaviors, who exhibited this tendency. Thus, the failure to ‘solve’ the rather complex Simpson’s paradox might not reflect an attempt to simplify the task by neglecting a confounding ecological variable. It might instead reflect a genuine attempt to use the ecological correlation as a smart and parsimonious proxy for individual-level correlations. Ecological Inferences in Spite of Monetary Rewards Converging evidence for the use of ecological correlations stems from even simpler operant conditioning experiments where accurate responses were rewarded with money (Kutzner et al. 2008). In these experiments, only one ‘ecology’ with two variables was encountered. Participants were repeatedly asked to predict whether a ‘left’ or a ‘right’ response was the correct reaction after one of two signals, a high- and a low-pitch sound. The sounds were not predictive of which was the correct reaction, that is, on the individual trial level there was a zero correlation between sound and correct reaction. At the same time, the whole setting or ecology that participants encountered was special. One of the sounds, say the high-pitch sound, preceded responses clearly more frequently than the low-pitch sound and, irrespective of the sound presented, one of the responses, say ‘left’, was rewarded clearly more frequently than the other. This strangely ‘skewed’ ecology contrasts with an ‘ignorant prior’ ecology where signals and correct reactions can be expected to be evenly distributed. As compared to this ignorant prior ecology, there is again a perfect ecological correlation. Knowing the mean value of the high-pitch sound in one ecology is perfectly predictive of mean value of ‘left’ being the correct response. The choice pattern in this operant conditioning scenario is in line with the use of an

M4225-ALTMAN_9781782549574_t.indd 91

03/05/2017 08:20

92

Handbook of behavioural economics and smart decision-making

ecological correlation (see Fiedler et al. 2013 for a discussion of ecological inferences in single ecologies). Even though choices tended towards the more frequently rewarded option, this trend was less pronounced for the less frequently presented sound, significantly reducing participants’ payoff (Kutzner et al. 2008). Many more experiments testify to the use of ecological correlations that did not use monetary incentives but social incentives, calling on participants to form accurate and non-discriminatory impressions about others (McGarty et al. 1993; Eder et al. 2011; Kutzner et al. 2011). Ecological Inferences in Both Directions Given the robustness of inferences from ecological to individual correlations, the question is whether people are actually insensitive to the level of data aggregation or whether they are simply not able to assess individual-level correlations. In the end, ecological correlations require assessing less information saving time and processing costs. To answer this question we conducted two experiments (Vogel et al. 2014). We provided participants with information about the level of product demand in two supermarkets. Different ecologies were created in terms of eight different product segments (for example, cheese, fruit). A first study replicated evidence for the use of an ecological correlation. Participants acted as if high demand for a specific product in one supermarket was predictive of high demand for that product in the other supermarket. Yet, demand on product level was not predictive. Across the product categories, however, high average category demand in one supermarket predicted high average category demand in the other supermarket. In the second study, the reverse was true. Here, demand across product categories did not correlate between supermarkets but demand for individual products did.1 In that case, participants correctly identified the individual product-wise correlations. Additionally, participants also acted as if there was a correlation across ecologies, predicting higher average demand for product categories that had been highly demanded in the other supermarket. These experiments suggest that people substitute ecological correlations for individual-level ones just as readily as they substitute individual-level correlations for ecological ones, committing the so-called atomistic fallacy (Diez-Roux 1998). In that participants correctly identify existing individual-level correlations but readily and, in this case, incorrectly generalize them to a higher level of data aggregation, the results demonstrate a genuine insensitivity to the level of data aggregation. In sum, examining results across a variety of paradigms, inferences about correlations have revealed that people are insensitive as to whether aggregated or individual-level data provides the input to their judgments. Further studies have extended this evidence to content domains including scholastic achievement (Fiedler et al. 2007) and political attitudes (Vogel et al. 2013). Together, these studies demonstrate the robustness of the use of ecological correlations as a proxy for individual-level correlations by lay people. It appears that whenever there is an ecological variable (for example, countries) allowing for the grouping of individual observations, lay people will consider the ecological correlation to learn about the individual-level relationship.

M4225-ALTMAN_9781782549574_t.indd 92

03/05/2017 08:20

Smart predictions from wrong data 93

THE SMART POTENTIAL BEHIND ECOLOGICAL CORRELATIONS Given the resilience of the use of ecological correlations or, more generally, the indifference between aggregation levels, a pressing question is whether this indifference is potentially smart or, broadly speaking, rational? Parsimony Using ecological correlations to infer individual correlations is parsimonious. To make inferences, only the base rates of variable (how relatively frequently things happen) have to be assessed. This assessment can be based on experience gathered on different occasions. It is not necessary to wait for or remember combined observations of, for example, people becoming unemployed and becoming a parent. Such a complete data matrix, as ideally assumed in statistics books, is hardly ever available to the lay decision maker in real life. To illustrate, consider trying to assess which of four dichotomous variables are predictive of happiness. The four zero-order correlations already result in combining, recording, and remembering 16 types of observations (four for each correlation). Trying to avoid a Simpson’s paradox, taking into account each of these variables as moderating the others’ influence, creates 24 correlations (96 types of observations) to be handled (four variables related on each level of the other three variables). Beyond complexity that might prevent accurate processing, the data might not be available at the time of judgment. Trying to assess the correlation for a novel variable, say, engaging in volunteer work, might simply be impossible because the relevant observations have not (yet) been gathered. However, the base rate of people that volunteer in a given ecology should be recorded and recalled quite automatically (Hasher and Zacks 1984). Using these readily available base rates, ecological inferences seem designed to enable inferences about novel correlations. In sum, ecological correlations pose relatively low demands on either cognitive resources or the amount of data that has to be available. Thus using ecological correlations to infer individual correlations satisfies one of the components of being a smart inference strategy: they are feasible and efficient. The Validity of Ecological Correlations: The Case of Happiness Feasibility and efficiency alone, however, do not render an inference strategy smart. Only if the strategy used has some degree of validity can it be justified. Confounding variables, as in the case of the Simpson’s paradox, threaten the validity of taking ecological correlations for individual level correlations. Yet, checking for the presence of confounding variables usually amounts to making sure there is no needle in the haystack. A pragmatic workaround to the elusive analytical answer to the validity of ecological correlations in general is to quantify their validity for a specific question. Such an analysis is presented below. This analysis is but an example of how the validity of ecological correlations can be quantified. It makes no claim to their validity in general. Instead, it could inspire the classification of questions into those for which ecological inferences are valid and those for which they are not.

M4225-ALTMAN_9781782549574_t.indd 93

03/05/2017 08:20

94

Handbook of behavioural economics and smart decision-making

The strategy put forward to assess the validity of using ecological correlations is simple. In essence, it compares the correlation of a criterion with a meaningful predictor assessed across individuals with the same correlation assessed across ecologies. This comparison is repeated for multiple predictors and multiple ways to form ecologies in the same data. Ecological correlations are deemed valid to the degree that the comparisons show correspondence between ecological and individual-level correlations in sign and size. To do so for a given question, we had to select (1) a criterion for which individual-level data are available, (2) variables used to predict or correlate with this criterion and (3) ecologies or ecological variables across which ecological correlations are computed. All of these selections will possibly change the validity of ecological correlations. Therefore, the resulting selection should be as representative as possible of what lay decision makers would do to make inferences important to them. As a criterion we selected happiness. Happiness studies have enjoyed a privileged status among scientists, resulting in representative large-scale surveys optimal to assess individual-level correlations.2 Prominent philosophers from Socrates, Meister Eckart (Meister Eckart and Davies 1995) and Mill (1863) to the Dalai Lama (Dalai Lama and Cutler 1998), but also economists (for example, Anielski 2007), lawyers (Bronsteen et al. 2015) and psychologists (for example, Maslow 1962; Seligman 2002) have dedicated their work to the identification of the very predictors of this arguably ultimate goal. Lay people, pursuing hedonic or utilitarian motives, are also likely to reason about the correlates of happiness to finally infer what causes it. At the operational level, we used happiness as represented in the sixth European Social Survey’s (ESS) question C1 ‘Taking all things together, how happy would you say you are?’ ranging from ‘00 Extremely unhappy’ to ‘10 Extremely happy’. To select predictors of happiness we relied on variables prominent in the scientific psychological literature. We selected health, social integration, income, religiousness, personal freedom, educational achievement and gender. Note that some of the documented correlations seem robust and strong, such as those for health and social integration, whereas others are feebler, such as those with income or gender (for a review, see Diener et al. 1999). Selecting ecological variables, we tried to reflect social categories that people might spontaneously think of when thinking about happiness. We included social categories that are generally considered readily available when thinking about the social world, such as nationality, gender and age (Gavanski and Hui 1992). We also tried to capture ‘warmth’ and ‘competence’, variables that are usually considered fundamental dimensions in social perception (Fiske et al. 2002). As proxy for warmth we used what a person’s main occupation is, including housework, work, being in school or retired. As proxy for competence we used a person’s income group. Our selection resulted in eight correlations with happiness, which we calculated five times, once at the individual level and four times for the four ecological variables. These ecological correlations include, for example, people thinking about, or being confronted with, data indicating that middle-aged people are, on average, least free and least happy. Would it be smart for decision makers to infer that high personal freedom is likely to go along with high personal happiness? In Table 5.1 we present the results for one nation, Slovenia. We focus on evidence within one nation because we assume that people’s information will be biased towards

M4225-ALTMAN_9781782549574_t.indd 94

03/05/2017 08:20

Smart predictions from wrong data 95 Table 5.1

Correlation table for predicting high levels of happiness in Slovenia

Units of observations Predictors Good health High income Good social integration High freedom High edu. achievement High religiousness Being female Higher age

Individuals (N 5 1257) 0.69 0.56 0.54 0.50 0.25 0.16 0.04 −0.38

Income Nationality Main Age Gender (N 5 5) (N 5 29) occupation (N 5 5) (N 5 2) (N 5 5)

Mean validity

1.00 − 0.98 0.80 1.00 −0.86 −0.97 −0.98

0.70 0.34 0.74 0.62 −0.05 −0.19 – −0.19

0.80 0.91 0.96 −0.24 0.29 −0.61 0.08 −0.75

0.93 0.84 0.97 −0.76 0.57 −0.63 −0.64 –

−1.00 −1.00 −1.00 1.00 −1.00 1.00 – 1.00

Validity for Slovenia % correct sign Rank order agreement

71 0.83

63 0.83

75 0.78

50 0.65

25 −0.61

57 0.49

Validity for all 29 countries % correct sign Rank order agreement

75 0.72

62 0.84

77 0.71

56 0.51

48 0.02

64 0.56

Note: Ecologies with N 5 5 are based on quintile ranks, except for ‘Main occupation’ which represents the groups’ paid work, education, housework, unemployed and retired. Rank order agreement 5 correlation with individual correlation coefficients.

their local social circle (Galesic et al. 2012). We focus on Slovenia because it represents a modal pattern of results for the validity of ecological correlations among the 29 countries we analyzed. As visible in the first column of Table 5.1, results largely replicate research on what correlates with happiness. Strong individual-level correlations can be found with health, income and social integration, whereas small correlations result for gender and even a negative one for age. These results also provide insights into the validity of ecological correlations. Consider, first, the ecological correlations with happiness across income ecologies. Even though treating an income group as ‘ecology’ might sound strange, contrasting rich and poor people when thinking about happiness might not. For example, we might conclude that health is lowest for those with little income and highest for those with high income. We might observe that happiness is also lowest for those with little income and highest for those with high income. In this case we have an ecological correlation between happiness and health across income groups. When computing exactly this ecological correlation for Slovenia, a perfect correlation of r 5 +1 results across the five income quintiles. As visible in Figure 5.3(a), every move upward in income goes along with a joint move upward in health and happiness. As compared to the r 5 .69 individual-level correlation between health and happiness (see the first column in Table 5.1), this ecological correlation is inflated but has the same sign. Yet, we also found evidence for the ecological and individual-level correlations to also diverge in sign. As visible in the first column of Table 5.1, being religious is correlated at r 5 +.16

M4225-ALTMAN_9781782549574_t.indd 95

03/05/2017 08:20

96

Handbook of behavioural economics and smart decision-making

Mean happiness (0–10)

(a) Health-happiness correlation across income quintiles 8.0

Mid high

High

Average

7.5 7.0 Mid low

Low

6.5

3.4

3.6

3.8

4.0

Mean health (1–5)

Mean happiness (0–10)

(b) Religiousness-happiness correlation across income quintiles High

8.0 7.5

Mid high

Average

7.0

Mid low Low

6.5 2.6

2.7

2.8

2.9

3.0

Mean religiousness (1–5) Source:

Data from the sixth ESS survey for Slovenia.

Figure 5.3

(a) Ecological correlation across income groups (quintiles) between health and happiness; (b) ecological correlation across income groups (quintiles) between religiousness and happiness

with happiness in Slovenia. When analyzed across income quintiles, a lower mean level of religiousness almost always goes along with a higher mean level in happiness, r 5 –.86. Overall, however, five of seven predictors,3 71 percent, correlate in the same direction with happiness across income groups and across individuals. Additionally, the relative importance of the different predictors of happiness at the individual level should optimally also reflect in the ecological correlations. Comparing the ranks of the correlation coefficients in the first and second columns in Table 5.1 seems to support this. If a predictor’s coefficient has a high rank when computed at the individual level it tends to have a high rank among the ecological correlations as well. The correlation of correlation coefficients amounts to r 5 +.83 (see rank order agreement in Table 5.1). Thus, at least across income groups as ecologies, not only the sign but also the relative importance of the predictors tends to be preserved when going from individual level to ecological correlations. Similar levels of validity can be found when considering different ecological variables such as nationality, main occupation, and age (see third to fifth columns in Table 5.1). The only exception is gender where the validity in terms of detecting the correct sign or rela-

M4225-ALTMAN_9781782549574_t.indd 96

03/05/2017 08:20

Smart predictions from wrong data 97 tive importance is systematically off. There might be various reasons for this. Foremost, the difference in happiness between genders is, if present in the population, very small in the data set. This might render any ecological correlation computed across two gender categories an unsystematic chance gamble between +1 and –1. More substantially, for some correlations gender is likely to be a confounding variable. For example, as being female tends to be related to lower income but females are slightly happier in the data set, a perfect negative correlation between income and happiness across gender groups mainly reflects the correlation between income and happiness. In summary, for our representative case of Slovenia relying on ecological correlations to infer individual-level correlations has a high degree of validity for inferring which variables correlate with happiness and in what direction. Inferring the sign allows for 57 percent of correct predictions overall. Rank ordering the size of the correlations amounts to a validity of r 5 +.49. To generalize these results to the other countries in the data set, we averaged the validity indices separately for the five ecological variables across all 29 countries included in the ESS survey. As visible in the bottom lines of Table 5.1, on average the validities are even higher, allowing for 64 percent of correct sign inferences and a rank ordering with a validity of r 5 +.56. Only ecological differences across gender, possibly for the reasons discussed above, are on average uninformative for individual-level correlations. Further, for every country there is an overall positive validity in terms of the percentage of correct sign inferences and the rank ordering. The percentages and rank correlations range from 52 percent in Sweden and .00 in Iceland, to 88 percent and +.91 in the Czech Republic. Where does this leave us – is the glass half empty or half full? As is obvious from the imperfect rank correlations, decision makers using ecological correlations will sometimes mis-estimate the importance of a given predictor, and as evident from the percentage of sign inferences, ecological correlations also leave space for sign errors. However, in our data, ecological correlations are well above chance performance when inferring the sign and relative importance of individual-level correlations. Thus, to infer which variables correlate and how they correlate with happiness on an individual level, ecological correlations seem to be a valid inference tool overall. Of course, this analysis cannot be more than exemplifying evidence. Yet, it illustrates the potential validity of ecological inferences for variables meaningful to the passing consumer of information. These arguments add to the claim, though cautiously, that using ecological correlations represents a smart strategy. More systematic analyses of the circumstances under which they are smart are an endeavor for future research.

CONCLUSIONS This chapter deals with the use of ecological correlations as a proxy for individual-level correlations as a smart or, broadly speaking, rational inference strategy. In the first part we demonstrate that people, from passing consumers of newspaper articles to monetarily incentivized participants in laboratory studies, use correlations that are computed across ecologies, such as nations, supermarkets or social groups, to infer correlations at the individual level. This might be owing to the relative parsimony of ecological correlations in terms of the processing demands and the demands on the available data. In the second

M4225-ALTMAN_9781782549574_t.indd 97

03/05/2017 08:20

98

Handbook of behavioural economics and smart decision-making

part, we show that in the example of inferring what variables correlate with happiness, using ecological inferences seem to be a valid strategy. Even though the case of ecological correlations might seem technical at first, our current social and informational world is ripe with aggregated information inviting their usage. Many different internet-based services, such as Google and Twitter allow for access to their aggregated user data. For example, using Google Trends (http://www.google.com/ trends/), everybody can, within minutes, visualize on a map how the propensity to search for, say, ‘sushi’ differs between regions in the US, and how this compares to the propensity to search for, say, ‘democrats’. If similar colors are generated on a map, an easily accessible form of ecological correlation is born. Most government-based statistical services, such as Eurostat or the European Social Survey, and government agencies now provide similar access. For example, the London police (http://maps.met.police.uk/) offer results for recent crime statistics mapped out across different suburbs, ready to be ‘ecologically’ correlated with, for example, knowledge about the population composition of these suburbs. Concluding, using ecological correlations to infer individual-level correlations appears to be a valid inference strategy in the absence of adequate data or processing abilities. Yet, it is to some degree error-prone. Whether this probabilistic inference strategy is smart, ultimately depends on the costs, social or material, incurred by inferring wrong correlations.

NOTES *

The research underlying this paper was supported by a grant from the Deutsche Forschungsgemeinschaft (KU – 3059/2-1). 1. The distribution used in Vogel et al. (2014) is compatible with an ecological variable moderating the correlation between two variables, that is, individual correlations actually differ between ecologies. There are many examples for this kind of aggregation bias in the economist literature (for example, Jaworski and Kohli 1993; Grewal et al. 2013). For the sake of simplicity, in this chapter, we only consider divergence between ecological correlations and individual contingencies pooled across ecologies, but omit discussing heterogeneity regarding within ecology correlations. 2. For the subsequent analyses, we rely on the sixth round of the European Social Survey conducted in 2012 with 54 637 respondents in 29 countries (including Switzerland, Russia and Israel). 3. We removed the correlation of income across income groups from the analysis. Even though possible to compute, an ecological correlation of average income across income groups seems implausible. Also, resulting high correlations seemed to artificially inflate the validity scores. The same was true for age and gender.

REFERENCES Anielski, M. (2007), The Economics of Happiness: Building Genuine Wealth, Gabriola Island, BC: New Society. BBC (2013), ‘Europe birth rates “have fallen” since economic crisis’, BBC News website, 10 July, accessed 2 March 2015 at http://www.bbc.com/news/world-europe-23259127. Bronsteen, J., C. Buccafusco and J.S. Masur (2015), Happiness and the Law, Chicago, IL: University of Chicago Press. Dalai Lama and H. Cutler (1998), The Art of Happiness: A Handbook for Living, New York: Riverhead Books. Diener, E., E.M. Suh, R.E. Lucas and H.L. Smith (1999), ‘Subjective well-being: three decades of progress’, Psychological Bulletin, 125 (2), 276–302, doi:http://dx.doi.org/10.1037/0033-2909.125.2.276. Diez-Roux, A.V. (1998), ‘Bringing context back into epidemiology: variables and fallacies in multilevel analysis’, American Journal of Public Health, 88 (2), 216–22, doi:10.2105/AJPH.88.2.216. Eder, A.B., K. Fiedler and S. Hamm-Eder (2011), ‘Illusory correlations revisited: the role of pseudocontingencies and working-memory capacity’, Quarterly Journal of Experimental Psychology, 64 (3), 517–32, doi:10.1 080/17470218.2010.509917.

M4225-ALTMAN_9781782549574_t.indd 98

03/05/2017 08:20

Smart predictions from wrong data 99 Fiedler, K. and P. Freytag (2004), ‘Pseudocontingencies’, Journal of Personality and Social Psychology, 87 (4), 453–67, doi:10.1037/0022-3514.87.4.453. Fiedler, K., P. Freytag and T. Meiser (2009), ‘Pseudocontingencies: an integrative account of an intriguing cognitive illusion’, Psychological Review, 116 (1), 187–206, doi:10.1037/a0014480. Fiedler, K., P. Freytag and C. Unkelbach (2007), ‘Pseudocontingencies in a simulated classroom’, Journal of Personality and Social Psychology, 92 (4), 665–77, doi:10.1037/0022-3514.92.4.665. Fiedler, K., F. Kutzner and T. Vogel (2013), ‘Pseudocontingencies – logically unwarranted but smart inferences’, Current Directions in Psychological Science, 22 (4), 324–9. Fiske, S.T., A.J. Cuddy, P. Glick and J. Xu (2002), ‘A model of (often mixed) stereotype content: competence and warmth respectively follow from perceived status and competition’, Journal of Personality and Social Psychology, 82 (6), 878–902, doi:10.1037/0022-3514.82.6.878. Galesic, M., H. Olsson and J. Rieskamp (2012), ‘Social sampling explains apparent biases in judgments of social environments’, Psychological Science, 23 (12), 1515–23, doi:10.1177/0956797612445313. Gavanski, I. and C. Hui (1992), ‘Natural sample spaces and uncertain belief’, Journal of Personality and Social Psychology, 63 (5), 766–80, doi:10.1037/0022-3514.63.5.766. Goldstein, J., D. Karaman Örsal, M. Kreyenfeld and A. Jasilioniene (2013), ‘Fertility reactions to the “great recession” in Europe’, Demographic Research, 29 (4), 85–104, doi:10.4054/DemRes.2013.29.4. Grewal, R., M. Chandrashekaran, J.L. Johnson and G. Mallapragada (2013), ‘Environments, unobserved heterogeneity, and the effect of market orientation on outcomes for high-tech firms’, Journal of the Academy of Marketing Science, 41 (2), 206–33, doi:10.1007/s11747-011-0295-9. Guardian (2013), ‘Which countries are the most forward thinking? See it visualised’, Guardian, 8 February, accessed 5 March 2015 at http://www.theguardian.com/news/datablog/2013/feb/08/countriesmost-forward-thinking-visualise. Hammond, J. (1973), ‘Two sources of error in ecological correlations’, American Sociological Review, 38 (6), 764–78. Hannan, M.T. (1971), ‘Problems of aggregation’, in H. Blalock (ed.), Causal Models in the Social Sciences, Chicago, IL: Aldine, pp. 473–508. Hasher, L. and R. Zacks (1984), ‘Automatic processing of fundamental information: the case of frequency of occurrence’, American Psychologist, 39 (12), 1372–88, doi:dx.doi.org/10.1037/0003-066X.39.12.1372. Jaworski, B.J. and A.K. Kohli (1993), ‘Market orientation: antecedents and consequences’, Journal of Marketing, 57 (3), 53–70, doi:10.2307/1251854. King, G. (2013), A Solution to the Ecological Inference Problem: Reconstructing Individual Behavior from Aggregate Data, Princeton, NJ: Princeton University Press. Kutzner, F., P. Freytag, T. Vogel and K. Fiedler (2008), ‘Base-rate neglect as a function of base rates in probabilistic contingency learning’, Journal of the Experimental Analysis of Behavior, 90 (1), 23–32, doi:10.1901/ jeab.2008.90-23. Kutzner, F., T. Vogel, P. Freytag and K. Fiedler (2011), ‘A robust classic: illusory correlations are maintained under extended operant learning’, Experimental Psychology, 58 (6), 443–53, doi:10.1027/1618-3169/ a000112. Maslow, A.H. (1962), ‘Introduction: toward a psychology of health’, in A.H. Maslow, Toward a Psychology of Being, Princeton, NJ: D Van Nostrand, pp. 3–8. McGarty, C., S. Haslam, J. Turner and P. Oakes (1993), ‘Illusory correlation as accentuation of actual intercategory difference: evidence for the effect with minimal stimulus information’, European Journal of Social Psychology, 23 (4), 391–410, doi:10.1002/ejsp.2420230406. Meiser, T. and M. Hewstone (2004), ‘Cognitive processes in stereotype formation: the role of correct contingency learning for biased group judgments’, Journal of Personality and Social Psychology, 87 (5), 599–614, doi:10.1037/0022-3514.87.5.599. Meister Eckart and O. Davies (1995), Meister Eckart – Selected Writings, Harmondsworth: Penguin Classics. Mill, J.S. (1863), Utilitarianism, London: Parker, Son, and Bourn. Preis, T., H.S. Moat, H.E. Stanley and S.R. Bishop (2012), ‘Quantifying the advantage of looking forward’, Scientific Reports, 2, art. 350, doi:10.1038/srep00350. Robinson, W. (1950), ‘Ecological correlations and the behavior of individuals’, reprinted 2009, American Sociological Review, 15 (3), 351–7, doi:10.1093/ije/dyn357. Schaller, M. (1994), ‘The role of statistical reasoning in the formation, preservation and prevention of group stereotypes’, British Journal of Social Psychology, 33 (1), 47–61, doi:10.1111/j.2044-8309.1994.tb01010.x. Schmitt, C. (2012), ‘A cross-national perspective on unemployment and first births’, European Journal of Population/Revue Européenne De Démographie, 28 (3), 303–35, doi:10.1007/s10680-012-9262-5. Seligman, M.P. (2002), Authentic Happiness: Using the New Positive Psychology to Realize Your Potential for Lasting Fulfillment, New York: Free Press. Simpson, E.H. (1951), ‘The interpretation of interaction in contingency tables’, Journal of the Royal Statistical Society. Series B (Methodological), 13 (2), 238–41.

M4225-ALTMAN_9781782549574_t.indd 99

03/05/2017 08:20

100

Handbook of behavioural economics and smart decision-making

Vogel, T., F. Kutzner, K. Fiedler and P. Freytag (2013), ‘How majority members become associated with rare attributes: ecological correlations in stereotype formation’, Social Cognition, 31 (4), 427–42, doi:10.1521/ soco_2012_1002. Vogel, T., F. Kutzner, P. Freytag and K. Fiedler (2014), ‘Inferring correlations: from exemplars to categories’, Psychonomic Bulletin & Review, 21 (5), 1316–22, doi:10.3758/s13423-014-0586-5. Wall Street Journal (2012), ‘Europe takes digital lead; divide persists’, The Wall Street Journal, 6 April, accessed 5 March 2015 at http://blogs.wsj.com/tech-europe/2012/04/06/europe-takes-digital-lead-divide-persists/.

M4225-ALTMAN_9781782549574_t.indd 100

03/05/2017 08:20

6

Heuristics: fast, frugal, and smart Shabnam Mousavi, Björn Meder, Hansjörg Neth and Reza Kheirandish

Recent years have seen important new explorations along the boundaries between economics and psychology. For the economist, the immediate question about these developments is whether they include new advances in psychology that can fruitfully be applied to economics. (Simon 1959, p. 253)

INTRODUCTION Individuals often make smart decisions despite the inherent limitations of cognitive and material resources. Whereas mainstream economics has focused mainly on the allocation mechanisms of material resources by cognitively unbounded (fully rational) agents, behavioral economics aims to include allocation of cognitive resources by using the insights from the heuristics and biases program in psychology (Kahneman et al. 1982). In this chapter, we introduce another psychological program with a more optimistic perspective inspired by Simon’s view of bounded rationality and developed systematically in the study of fast-and-frugal heuristics (Gigerenzer et al. 1999).1 We make a distinction between human decision making in two situations: under uncertainty, in which case we reason that simple heuristics are successful strategies, and under risk, in which case we discuss the enhancing role of risk literacy and statistical thinking (for the ways in which information is processed and knowledge created under risk versus under uncertainty, see Mousavi and Gigerenzer 2014, Table 1; and for the realms of rationality see Neth and Gigerenzer 2015, Table 1). Our goal is to make sense of smart and efficient decisionmaking processes demonstrated by individuals who use their evolutionary developed and learned capacities. Which recent advances in psychology are important to economic theory and behavioral economics? This chapter has emerged from a series of dialogues between two psychologists and two economists exchanging views on the study of fast-and-frugal heuristics as it pertains to the methods of understanding economic behavior and decision making. Our discussions developed before a backdrop of what we view as the thrust of our fields as well as their overlaps in relation to formal treatments of human behavior. What economists now practice and profess as the basis and criterion for rigorous study of economic behavior traces back to the normative interpretation of subjective utility.2 Psychologists, by contrast, have often searched for systematic patterns of behavior in laboratory and case studies, often formulating models without subscribing to or aiming for accordance with universal maxims of behavior. The heuristics and biases research program (Tversky and Kahneman 1974) commenced with an inquiry into uncovering general cognitive mechanisms. Mainstream behavioral economics has combined the findings of this psychology program with the maxims of economics. The resulting body of work has been valuable in 101

M4225-ALTMAN_9781782549574_t.indd 101

03/05/2017 08:20

102

Handbook of behavioural economics and smart decision-making

crossing disciplinary boundaries, inspiring scientific inquiry into untapped domains and generating potential for further discoveries. At the same time, like any other field of study, behavioral economics has carried baggage from its mother discipline, which generated byproducts and implications for methods of research. Our assertion is that the study of the ecological rationality of fast-and-frugal heuristics (Gigerenzer et al. 1999) can provide important insights and tools for alternative and complementary analyses of behavior. The framework of fast-and-frugal heuristics can be distinguished from the heuristics and biases approach by the following characteristics and standpoints. In our view, heuristics are indispensable strategies for successfully dealing with uncertain situations in the real world. Notably, most real-world situations do not allow identification of all alternatives, consequences, and probabilities, even subjectively, as required for finding the optimal solution. Moreover, the best solution from a social perspective does not necessarily accord with rational choice based on self-interest (for example, public goods). For these reasons, smart decision makers regularly develop and use heuristics, relying on the wisdom and experience that simple heuristics can outperform supposedly optimizing strategies in uncertain situations. More often than not, satisficing with respect to a good enough aspiration level turns out to be both rational and smart for boundedly rational agents. It is thus not irrational but intelligent to be less than fully rational, in the neoclassical sense, in many decision-making situations. A two-way influence and exchange between psychology and economics can build upon shared notions such as ecological rationality of simple heuristic strategies. This is the core around which we have structured this chapter. We start with juxtaposing economic and psychological views of ecological rationality (based on interviews with Vernon Smith and Gerd Gigerenzer, both leading researchers in their respective fields), pointing out the overlaps between the two views, and then extend questions pertaining to behavioral economics as a set of suggestions for advancing the dialogue between economics and psychology. Viewing heuristics as adaptive tools for decision making is discussed next. Although heuristic strategies can be used both under uncertainty and under risk, the simplicity of heuristics makes them particularly successful under the irreducible uncertainty of many decision situations. This point is illustrated by connecting the Knightian distinction between risk and uncertainty to inferential rules, amended by heuristics. The practical success of simple heuristics is then illustrated in the domains of financial investments and business decision-making. Next, we consider potential implications of ecological rationality in two applied scenarios: the current debate on nudging and the use of natural frequencies in risk communication. We close by providing a brief summary and extending our collaborative challenge to economists.

WHERE ECONOMIC RATIONALITY MEETS PSYCHOLOGY In his Nobel Prize lecture, Vernon Smith (2002) focused on two forms of rationality in economics and their functions with respect to the understanding of human behavior. The first form is constructivism, which is rooted in Hume’s and British empiricism; here the study of human behavior starts with observing an outcome and then reconstructing the steps with which such an outcome can be generated through a deliberate reasoning process. This reconstruction provides a variety of possibilities and options to choose

M4225-ALTMAN_9781782549574_t.indd 102

03/05/2017 08:20

Heuristics: fast, frugal, and smart 103 from, which are not sufficient for the realization of action. For that reason, Smith points out that ‘constructivism alone leads nowhere; its roots must find ultimate nourishment outside of [such] reason. Outside means knowledge derived from experience, from social interactions, and from unconscious sources and processes – the nexus that I have called ecological rationality’ (Smith 2008, p. 287). Interestingly, this second form of rationality, the notion of ecological rationality in economics, is shared with the psychological study of fast-and-frugal heuristic decision-making: The term ‘ecological rationality’ has been used fittingly by Gigerenzer et al. (1999) for application to important discoveries captured in the concept of ‘fast and frugal decision making’ by individuals: ‘A heuristic is ecologically rational to the degree that it is adapted to the structure of an environment.’ (p. 13). My application of the term is concerned with adaptations that occur within institutions, markets, management, social, and other associations governed by informal or formal rule systems – in fact, any of these terms might be used in place of ‘heuristic’ and this definition works for me. (Smith 2008, p. xix)

The similarities and some specific connections between the research traditions established by Vernon Smith in economics and Gerd Gigerenzer in psychology that evolve around this shared notion of ecological rationality and lead to a functional view of heuristics are juxtaposed in Table 6.1 (for a juxtaposition of Smith’s and Kahneman’s approach to the theory and modeling of human behavior, see Altman 2004). Moreover, Table 6.1 provides two items under each connected notion. The first item illustrates the overlap between the two views, and the second outlines research questions that relate the particular preceding topic to the core of behavioral economics inquiry. In the study of human action, Smith calls for supplanting the traditional constructivist framework of rationality with the ecological one. In a similar vein, Gigerenzer calls for ‘a better understanding of human rationality than that relative to content-blind norms’ (2008, p. 19). Constructivist rationality derives normative benchmarks from formal frameworks such as logics and probability theory, where the situation in which a choice is made is abstracted from its content. Thus, these norms are blind to the content of the decisionmaking situation. Regrettably, ‘these were of little relevance for Homo sapiens, who had to adapt to a social and physical world, not to systems with artificial syntax, such as the laws of logic’ (Smith 2008, p. 19). In cognitive science, the study of error has fallen prey to a major error by maintaining norms of logic and statistics, which despite their coherent and consistent elegance, and at the price of preserving this elegance, could lack meaningful association to evaluation of human decision-making behavior.3 Pointing out that this is an unjustified extension from the study of perceptual errors to the cognitive domain, Mousavi and Gigerenzer (2011; see also Gigerenzer 1991, 1996) argue for adopting and developing content-sensitive norms for the study of human cognition and behavior. For example, when the famous Wason selection task (Wason 1966) is given content by assigning two roles of employee and employers to the players who both are tasked with cheating detection, one group’s correct strategy aligns with the logical truth table associated with the conditional, but the other does not. Thus, logic appears to capture one part of the content and miss the other part. In this case, if judgment is evaluated based on logical truth, one group appears to have wrong judgment, whereas their judgment is completely correct with respect to the role (content) that they are assigned (Gigerenzer and Hug 1992). Thus, what counts as human rationality depends on the content and domain,

M4225-ALTMAN_9781782549574_t.indd 103

03/05/2017 08:20

104

Handbook of behavioural economics and smart decision-making

Table 6.1

Ecological rationality and heuristics à la Smith (economics) and Gigerenzer (psychology)

Fast-and-frugal heuristics program

Constructivist versus ecological rationality in economics

A heuristic is ecologically rational to the degree that it is adapted to the structure of an environment Humans have to adapt to a social and physical world, not to systems with artificial syntax, such as logic

Ecological rationality is concerned with adaptations that occur within institutions, markets, management, and social and other associations governed by informal or formal rule systems

Overlap between psychology and economics: same definition of ecological rationality, when heuristic can be replaced by markets, management, or other rule systems Research questions for behavioral economics: what is the relationship between rule systems and heuristics? Do they overlap, or is one nested in the other? What can be learned from establishing such characterizations? Unbounded rationality can generate optimal solutions for simple situations, e.g., tic-tactoe; omniscience and omnipotence can also be used for theoretical examination of human behavior, but applying them as universal standard of rationality is a scientific error

Constructivism or reason provides a variety of ideas to try out but often no relevant selection criteria, whereas ecological process selects the norms and institutions that serve the fitness needs of societies

Overlap between psychology and economics: Norms produced by unbounded or constructivist rationality are not useful as selection criteria in complex situations; the ultimate evaluation comes from the real world, not from theoretical sophistication Research questions for behavioral economics: In the study of human behavior where does realism matter, and where does it not? If norms are chosen conditional to the situation, how can we judge across situations? Can ecological fitness be formalized? Experimental games are bound to study social behavior as rule-obeying behavior and not as rule-negotiating or rule-changing behavior

Observing how people actually behave reveals unanticipated system rules, e.g., hubs emerged unexpectedly (like an equilibrium) when airlines were deregulated

Overlap between psychology and economics: rules are to be discovered as they emerge from social behavior. Formal models can be used to provide a possible description of what was observed Research questions for behavioral economics: to what extent can field experiments improve the relevance of solution concepts used for the study of human behavior and specify their limitations? Fast-and-frugal heuristics are strategies triggered by environmental situations and enabled by evolved or learned capacities

Heuristics are a kind of cognitive capacity that we can access, although we are not completely aware of our access to it

Overlap between psychology and economics: The choice of heuristic strategy is often not fully deliberate. This does not exclude the possibility of training or altering the trigger conditions Research questions for behavioral economics: When is a heuristic successful, and when does it fail? In real-world situations, when is it not rational to be ‘rational’? Source: Based on interviews in Mousavi and Kheirandish (2014).

M4225-ALTMAN_9781782549574_t.indd 104

03/05/2017 08:20

Heuristics: fast, frugal, and smart 105 whereas logic is content-free. Extending on this point, Mousavi and Gigerenzer (2011, p. 102) argue that ‘cognitive scientists studied judgment errors in order to discover rules that govern our minds, just as visual errors were studied to unravel the laws of perception. This practice has generated a long list of so-called cognitive biases, with disappointingly little insight into how the human mind works’. Alongside Smith (2008, p. 31), we find that ‘[t]he failed objective of this constructivist adventure is cause for joy, not despair’. Behavioral economics is concerned with making sense of human behavior, with the goal of developing a framework for analyzing decision-making in real-world situations, and evaluating and predicting human choice and actions therein. Vernon Smith’s following remark refers to these tasks directly: ‘Whatever it is that people do, it is evident that they do not think about the problem the way an economist does, nor do they model it that way’ (Mousavi and Kheirandish 2014, p. 1784). As Gigerenzer elaborates, ‘the question is not whether it is good or bad to ignore information but what ignoring information does psychologically . . . “Why a certain strategy?” and “When does it work?” rather than assuming “It maximizes something,” and that something may be psychological’ (Mousavi and Kheirandish 2014, p. 1784). This view, which emphasizes the interplay between heuristics and environments, and relies on the notion of ecological rationality to evaluate the rationality of human behavior, provides an alternative way for understanding our adaptive minds where constructivist rationality reaches its limits.

HEURISTICS AS ADAPTIVE TOOLS FOR DECISION MAKING UNDER UNCERTAINTY A prevalent view in both psychology and behavioral economics, the heuristics and biases program (Tversky and Kahneman 1974), presumes that heuristics result from a trade-off between accuracy and effort and lead to flawed and biased thinking. Typically, the benchmarks used to corroborate these claims are formal frameworks such as logic, probability theory, and expected utility theory. These are presumed to provide normatively correct solutions, and deviations in human decision making constitute errors. We advocate a different view based on formal models of heuristics. Within the framework of fast and frugal heuristics (Gigerenzer et al. 1999), heuristics are adaptive tools that ignore information to make fast and frugal decisions that are accurate and robust under conditions of uncertainty (Neth and Gigerenzer 2015). Heuristics are successful when they exploit an ecologically rational match to the structure of information in the environment. In the previous section, we discussed the role of ecological rationality in understanding the success of simple individual and organizational strategies in the juncture of psychology and economics. Here, we turn our focus to two central concepts in the theory of decision-making, namely, uncertainty and knowledge. The path towards making a decision starts with a disequilibrium that triggers a search for solutions through processing information to create the knowledge we need (Dewey 1938 [1986]). Traditionally, we model this procedure in two forms, deductive and inductive, depending on the structural properties of the situation to be resolved (Goldman 1988). Also, we acknowledge the unknowns of the situation by specifying alternatives, consequences, and their probabilities, which leads to a characterization of the risk associated with the problematic situation. The way in which a problematic situation is

M4225-ALTMAN_9781782549574_t.indd 105

03/05/2017 08:20

106

Handbook of behavioural economics and smart decision-making

Table 6.2

Decisions under risk versus uncertainty

Nature of unknown

Knightian probability

Decision process

Method

Generated knowledge

Risk

A priori (design; propensity)

Deductive

Risk

Statistical (frequencies in the long run)

Inductive (statistical inference)

Use probability theory to model the underlying structure; optimization Use statistical inference; optimization

Uncertainty

Estimate; conduct based on opinion; not fully reasoned

Heuristic

Deterministic knowledge (as in lotteries); e.g., objective odds Stochastic knowledge; e.g., estimates of correlations Satisficing solutions when optimizing is not feasible; intuition (as in entrepreneurship)

Select a heuristic that is ecologically rational for a task; exploratory data analysis

Source: Adapted from Mousavi and Gigerenzer (2014) with permission.

characterized, in turn, shapes and limits the type of solution that can be produced because it dictates the form of knowledge generated from the processing of information. This is illustrated in Table 6.2, wherein, in addition to deductive and inductive processes, a third heuristic process is proposed, which involves a less than exhaustive search for or consideration of information and leads to knowledge that is simply good enough for making a successful decision, but by no means exhausts information or optimizes across conditions. Note that the idea is not that these decision processes are mutually exclusive categories. Rather, the current categorization is meant to shed light on the nature of knowledge used and created in the process of decision making. For resolving disequilibrium, boundedly rational agents tend to use different types of strategies compared to fully rational agents (Simon 1955; Selten 1998). They restore the equilibrium by finding satisficing answers to their problems in situations with irreducible uncertainty, wherein exhaustive search is often unhelpful or even impossible. Heuristic decision-making, based on good-enough reasons to act, characterizes the observed behavior primarily with respect to a functional (rather than mirror image) match between the mind of the decision maker, the particular strategy employed, and properties of the task environment. A large number of situations involving unknowns are characterized by what we refer to as fundamental uncertainty that cannot be reduced to risk calculations. This fundamental uncertainty includes what Knight refers to as an ‘estimate’ and extends to situations where some options, outcomes, or probabilities are fundamentally unknown (Meder et al. 2013). Heuristics are then to be viewed as less than fully reasoned strategies to deal with complexities of such uncertain unknowns by not trying to assign a probability to (including zero for ignoring) every unknown, but just forming an opinion that allows an action, what Knight calls an estimate: Suppose we are allowed to look into the urn containing a large number of black and red balls before making a wager, but are not allowed to count the balls: this would give rise to

M4225-ALTMAN_9781782549574_t.indd 106

03/05/2017 08:20

Heuristics: fast, frugal, and smart 107 an estimate of probability in the correct sense; it is something very different from either the mere consciousness or ignorance on which we act if we know only that there are balls of both colors without any knowledge or opinion as to the numbers or the exact knowledge of real probability attained by an accurate counting of the balls. In the second place, we must admit that the actual basis of action in a large proportion of real cases is an estimate. Neither of these interpretations, however, justifies identifying probability with an estimate. . ..The exact science of inference has little place in forming the opinions upon which decisions of conduct are based, and that this is true whether the implicit logic of the case is prediction on the ground of exhaustive analysis or a probability judgment, a priori or statistical. We act upon estimates rather than inferences, upon ‘judgment’ or ‘intuition’, not reasoning, for the most part. (Knight 1921, p. 23)

Note that heuristics can be applied to a variety of situations. For instance, the priority heuristic (Brandstätter et al. 2006) provides a lexicographic strategy to choose among lotteries, the classical paradigm for decision-making under risk. The priority heuristic chooses between lotteries by comparing their probabilities and outcomes (gains or losses) lexicographically (that is, one at a time) instead of combining probabilities and gains in a weighted sum. Surprisingly, this simple model logically implies long-lasting anomalies of human choice behavior, such as the Allais paradox, the fourfold pattern of risk, and the certainty effect (Katsikopoulos and Gigerenzer 2008). Also, heuristic methods may be applied to a wide range of other problems, such as catching a ball. One way of solving the problem would be to compute the trajectory of the ball and move towards the inferred landing point, but owing to the number of causally relevant variables (for example, velocity and wind resistance) and the associated uncertainties, this is difficult to impossible. However, the problem can be tackled by a relatively simple algorithm according to which the catcher does not compute the landing point, but focuses on the ball and keeps a constant angle of elevation of gaze while running in the direction (McLeod and Dienes 1996). This example also illustrates the tight connection between heuristics and evolutionary or learned capacities. Applying the gaze heuristic requires certain capacities (that is, to fixate a moving object with your eyes, locomotion, and so on) that are necessary for using the strategy, which are far from trivial and cannot be reduced to merely computing the solution. The simplicity of heuristics is a feature, rather than a flaw. Heuristics are successful because of their simplicity, which involves a beneficial degree of ignoring information, not despite it – something that may puzzle many economists, when trying to make sense of the observed behavior through the lens of constructivist rationality, but is practiced regularly by laypeople. Whether the benefits of heuristics come at a prohibitive cost is not a matter of opinion but should be understood as an empirical question. In the following we turn to the world of business and finance as an example of an uncertain environment in which the successful use of heuristic strategies accords with the ecological notion of rationality.

SUCCESSFUL HEURISTICS IN FINANCE AND BUSINESS DECISION-MAKING The previous sections have provided theoretical arguments for a shift towards the norm of ecological rationality and proposed that heuristics are appropriate tools to tackle

M4225-ALTMAN_9781782549574_t.indd 107

03/05/2017 08:20

108

Handbook of behavioural economics and smart decision-making

complex problems under conditions of uncertainty. Given that practitioners care more about measurable results than about abstract beauty or consistency with axioms, it is not surprising that some of the strongest examples for successful use of heuristics stem from the world of finance and business decision making. Any form of resource allocation faces two fundamental problems: (1) how should we distribute our assets over all available options, and (2) when should we switch from one option to another? Theoretically, the asset allocation problem is solved by the Nobel prize-winning mean-variance model of Markowitz (1952), which provides the optimal investment portfolio by maximizing profit for a given level of risk. By contrast, a dominant strategy employed by many people is a simple 1/N heuristic that allocates resources equally across all considered assets. When contributing to retirement savings plans, 1/N has been called ‘naive diversification’ and is believed to incur substantial costs to investors (Benartzi and Thaler 2001). However, when DeMiguel et al. (2009) compared Markowitz’s solution and its modern variants with 1/N, the heuristic performed at least as strongly as the mean-variance model. One reason for the surprising success of the simple 1/N heuristic lies in the so-called bias-variance dilemma (Geman et al. 1992), pertaining to minimizing the prediction error. The prediction error has two contributing components: bias and variance. The error due to bias has been at the center of behavioral economics, and has led to enlisting several debiasing techniques. The error due to variance, however, has not been receiving much attention. 1/N exemplifies a simple allocation mechanism, which is highly biased but has no variance, and overall generates less prediction error under certain circumstances. 1/N can be viewed as a special case of the Markovitz model, which implies that the flexibility of the Markowitz model comes at the potential cost of an increased estimation error (Neth et al. 2014; see Gigerenzer and Brighton 2009, for a general discussion of heuristics and the bias-variance dilemma). As the benefits of 1/N have been shown to generalize to investments in international stock markets and different asset classes (Jacob et al. 2013) it seems smart of Markowitz to have used 1/N himself, rather than his own method of portfolio optimization (Benartzi and Thaler 2001, p. 80). The 1/N heuristic is an instance of a more general equality rule (Messick 2008) that is also applied in parental investments (Hertwig et al. 2002). Regarding the switching problem (that is, when and how to switch between different options), biologists and psychologists have examined simple, yet highly effective stopping rules in animal foraging theory (Green 1984; Stephens and Krebs 1986) and research on human multitasking behavior (for example, Payne et al. 2007). An applied instance of a simple and successful temporal threshold rule is the hiatus heuristic (Wübben and von Wangenheim 2008), which allows directing marketing efforts by abandoning customers who have not purchased anything for a certain amount of time, say, a number of months. This period of time that sets the threshold is called the hiatus. Interestingly, heuristic models combine explanatory parsimony with higher predictive power for situations of uncertainty.4 This is in direct contrast with the prevalent method used by mainstream behavioral economics of adding flexible parameters to Bernoulli utility functions in order to incorporate psychological factors of observed behavior, which in turn adds to the complexity of the model but often costs predictive power (Berg and Gigerenzer, 2010). The abundance and ubiquity of successful heuristics in applied contexts raises the question whether existing heuristics can be used to create new or improve existing strategies.

M4225-ALTMAN_9781782549574_t.indd 108

03/05/2017 08:20

Heuristics: fast, frugal, and smart 109 One aspect of ecological rationality – as a research program – aims at teasing out the elements of successful strategies to adapt and refine them to novel situations. In addition, understanding how, when, and why heuristics work well can guide the design of intuitive decision systems that fit the strategies that people naturally use. For instance, highly transparent and teachable fast-and-frugal trees (Martignon et al. 2003) have been designed for coronary care unit allocations (Green and Mehr 1997), for diagnosing patients with clinical depression (Jenny et al. 2013), and for identifying vulnerable banks in financial regulation (Aikman et al. 2014; Neth et al. 2014). Thus, successful heuristics are not only discovered, but can also be specifically designed to create efficient and effective tools. Next, we demonstrate this transformative potential of ecological rationality in the context of public policy decisions and the communication of medical risks.

APPLIED LESSONS FROM THE STUDY OF ECOLOGICAL RATIONALITY In the following, we extend our discussion of ecological rationality to applied issues. First, we critically evaluate the idea of nudging, a policy-making tool rooted in the behavioral economics approach. Subsequently, we discuss probabilistic reasoning and risk literacy as an example of how successful decision engineering can be guided by psychological research that takes the match between cognitive processes and the information structure of the environment seriously. This approach is based on the idea of making people risk literate to help them make better, more informed decisions, rather than merely nudging them towards an externally specified goal. The Risk of Using Nudges in an Uncertain World How to conceptualize human rationality is not only an academic issue, but has strong implications for policy making and the question of how to help people make better decisions. A prominent example is the so-called ‘nudge’ approach, which (in the tradition of the heuristics and biases program) assumes that people frequently make inferior decisions because their thinking is fundamentally biased and error-prone (Thaler and Sunstein 2008). The proposed remedy is to structure the choice situation so that people are more likely to make better decisions, while retaining freedom of choice (Grüne-Yanoff and Hertwig 2015). Examples include nudging people towards healthier dietary choices by arranging food items (for example, in a canteen) such that healthier items are more readily available, or setting default options in retirement saving plans in a way that people automatically enroll in higher saving contributions, unless they deliberately opt out (for an alternative approach to public policy that advocates financial literacy, see Altman 2012). However, what are the implications of nudges in an uncertain world, where it may not always be clear what it means to make better decisions? Consider nudges in the health domain. In the past decades, several countries have set up screening programs (for example, for breast cancer and prostate cancer), with the long-term goal of reducing cancer-related mortality rates. The idea behind these programs is to detect cancer in early stages, in order to treat people earlier and more effectively (or at least more cost-efficiently). A key question is how to provide information to the target group to increase participation; one way

M4225-ALTMAN_9781782549574_t.indd 109

03/05/2017 08:20

110

Handbook of behavioural economics and smart decision-making

of doing so is to resort to the nudge approach. For instance, different nudges have been used in the Danish breast cancer screening program to increase participation rates (Ploug et al. 2012; see Gøtzsche and Jørgensen 2013, for a related analysis of the British NHS breast cancer screening programme). Women in the target group received an invitation to participate, along with an information leaflet. The default was a pre-booked appointment, so that women needed to actively opt out. The leaflet also stated that after evaluating the pros and cons the Danish National Board of Health recommends participating in screening. These strategies aim at nudging people towards a goal defined by experts and policy makers, based on the assumption that it is in people’s interest to participate. It could be argued that people should be nudged to participate in screening – after all, is it not to their own benefit to participate if such a program can reduce the risk of dying from cancer (or at least lead to better treatment with less side effects)? What remains unclear, however, is whether participating in screening always serves people’s interests, given that there are different benefits and costs associated. In the case of breast cancer screening, the (currently) available data show that over a period of ten years, eight out of 2000 women who do not participate in screening die from breast cancer, compared with seven out of 2000 women who do participate (Gøtzsche and Jørgensen 2011). At the same time, however, screening entails potential and harms, such as overtreatment resulting from false positive test results (for example, unnecessary removal of the breast). Also, the overall mortality rate (that is, total number of women dying from all causes) does not vary between women participating and not participating in breast cancer screening (Gøtzsche and Jørgensen 2013). Yet this crucial information was omitted from the leaflet of the Danish breast-screening program, thereby undermining the possibility to make an informed decision based on considering and evaluating the potential benefits and harms (see Gigerenzer 2014a; Gigerenzer and Edwards 2003; Gigerenzer et al. 2007). For other screening programs, such as PSA-based screening for prostate cancer, the current evidence indicates that the potential harms actually outweigh the potential benefits. Consequently, the US Preventive Services Task Force explicitly recommends against prostate-specific antigen (PSA)-based screening for prostate cancer (Moyer 2012). This recommendation was issued after a period of uncertainty in which not enough evidence was available to determine whether PSA-based screening would be beneficial or not. These examples highlight critical issues in the foundation and application of nudges. An important precondition for the nudge approach is the possibility to determine – from the perspective of the choice architect – which decision is in the best interest of the decision maker. It may be self-evident that an apple is a healthier choice for a snack than a chocolate bar, but in other domains, such as medical treatments, determining which choice is in the decision maker’s best interest may be highly uncertain and dependent on individual preferences. A woman provided with the currently available evidence on breast cancer screening may decide that the benefits outweigh potential harms and therefore participate. However, she may also conclude that the potential harms outweigh the potential benefits and therefore decide that she would be better off by not participating. In each case, the decision will depend on how she values the associated benefits and costs. This, in turn, highlights the importance of providing people with the necessary information to make informed decisions, relative to their goals, values, and individual preferences, and not merely nudging them towards an externally specified goal. Importantly, not all informa-

M4225-ALTMAN_9781782549574_t.indd 110

03/05/2017 08:20

Heuristics: fast, frugal, and smart 111 tion is created equal, and identifying or designing transparent and intuitive information presentation formats is crucial for both psychology and behavioral economics (see below). In sum, the nudge approach rests on the assumption that there is one right way to make decisions, which applies to everybody, and that the choice architect knows what is in the best interest of the decision maker and can therefore enforce it. While this may well be true in some cases, we advise against an uncritical application of the approach to policy making. Nudges may be an effective tool in some circumstances, but like any tool (fastand-frugal heuristics included), they can cut both ways and need to be handled with care. In an uncertain and changing world, nudges may lead to adverse outcomes that are not in the best interest of decision makers. In our view, rather than precluding the possibility that people can make good decisions, the goal should be to develop means for communicating the relevant information in a way that facilitates people’s understanding of it, so that they can make better, more informed decisions. From the perspective of the nudge program, educating and informing people to make them risk literate (Gigerenzer et al. 2007; Gigerenzer and Muir Gray 2011) is likely to be ineffective, because the assumption (rooted in the heuristics and biases program) is that human thinking and decision making are fundamentally flawed. This view, however, neglects recent research that demonstrates how people can be helped to make better inferences (for example, inferring posterior probabilities, such as the probability of breast cancer given a positive mammogram) by conveying the relevant information in a transparent and intuitive way, without being patronizing (Gigerenzer and Hoffrage 1995; Sedlmeier and Gigerenzer 2001; Meder and Gigerenzer 2014). Similar views have been presented in other domains. In the equilibrium analysis of financial markets, although an equilibrium state is always Pareto optimal, this optimality does not necessarily coincide with the best ‘wanted’ outcome for all agents. This is illustrated in phishing equilibria, where ‘phools,’ who do not act according to what they want or is good for them, are systematically ‘phished’. Akerlof and Shiller (2015) argue that when information can be used systematically in forms that would deceive the consumers, the very structure of free markets provides opportunity for exploitation, a point overlooked by behavioral economists, [C]uriously, to the best of our knowledge, they [behavioral economists] have never interpreted their results in the context of Adam Smith’s fundamental idea regarding the invisible hand . . . It’s a major reason why just letting people be Free to Choose – which Milton and Rose Friedman, for example, consider the sine qua non of good public policy – leads to serious economic problems. (Akerlof and Shiller 2015, p. 6)

As we discuss next, the use of accessible and intuitive representative formats such as natural frequencies can improve decision making by enhancing their probabilistic reasoning abilities. Moving Beyond Nudges by Making People Risk Literate Understanding of and reasoning with probabilistic and statistical information is crucial for making good decisions. For instance, an informed decision on whether to participate in a screening program requires understanding of the relevant evidence regarding potential benefits and harms and the implications of medical test results. The nudge approach

M4225-ALTMAN_9781782549574_t.indd 111

03/05/2017 08:20

112

Handbook of behavioural economics and smart decision-making

and the heuristics and biases program that provides its conceptual foundation presume that people lack this capacity. Since humans are assumed to be biased and error-prone when it comes to probabilistic thinking, the suggested remedy is to nudge people into making better decisions. However, is nudging the only way to help people make better decisions? Also, what does the psychological literature have to say about people’s capacity to reason with probabilistic and statistical information? In fact, the psychological literature to date has given very different answers to these questions. In the 1950s and 1960s, researchers began investigating experimentally whether people’s inferences correspond (approximately) to probability theory in general, and to Bayes’ rule in particular. For instance, Phillips and Edwards (1966) used (incentivized) bookbag and poker chip scenarios, in which they presented subjects with a sequence of draws that came from either a bag with more red than blue chips or a bag with more blue than red chips. The question of interest was whether subjects would update their beliefs regarding the bag in accordance with Bayes’ rule, given the observed data. This and other studies indicated that the human mind is able to deal with probabilistic inferences, although it was frequently observed that the amount of belief revision was not as extensive as prescribed by Bayes’ rule (a phenomenon referred to as conservatism; Edwards 1968). Peterson and Beach (1967) coined the term ‘man as intuitive statistician,’ mirroring the Enlightenment view that the laws of probability are also the laws of the mind (Daston 1988). This view stands in stark contrast to the conclusions drawn from later research in the heuristics and biases tradition: ‘In making predictions and judgments under uncertainty, people do not appear to follow the calculus of chance or the statistical theory of prediction’ (Kahneman and Tversky 1973, p. 237). A key empirical finding used to corroborate this claim was that people often do not seem to appreciate base rate information (prior probabilities) when making Bayesian inferences. A prominent example is the so-called ‘mammography problem’ (Eddy 1982; Gigerenzer and Hoffrage 1995). Figure 6.1 (left-hand side) gives an example of the task in which the goal is to derive the posterior probability of a woman having breast cancer given a positive mammogram, based on information about the base rate (prior probability) of cancer, the probability of obtaining a positive test result for a woman having the disease, and the probability of obtaining a positive test for women without cancer. The probability tree (Figure 6.1, middle left) visualizes the given information, which consists of a set of unconditional and conditional probabilities. The posterior probability can be inferred using Bayes’ rule (Figure 6.1, bottom left), according to which the probability of cancer given a positive test result is about 8 percent. Yet many people give much higher estimates in this particular scenario, which has been interpreted as neglect of the base rate. These and similar findings have led to the view that people’s probabilistic reasoning is fundamentally flawed (but see Koehler 1996 for a critical review). More recently, however, psychologists have begun to identify the conditions under which people are able to make sound probabilistic inferences. This is a case in point for successfully exploiting the ecological rationality of designed tools. Instead of focusing on human errors, the focus is shifted to human engineering: What can be done to help people with probabilistic reasoning? A key insight from this line of research is the power of presentation formats: The extent to which people are able to make sound probabilistic inferences crucially depends on the ways in which the relevant information is conveyed.

M4225-ALTMAN_9781782549574_t.indd 112

03/05/2017 08:20

Heuristics: fast, frugal, and smart 113 Task The probability of breast cancer is 1 percent for a woman at the age of 40 who participates in routine screening. If a woman has breast cancer, the probability is 80 percent that she will get a positive mammography. If a woman does not have breast cancer, the probability is 9.6 percent that she will also get a positive mammography. A woman in this age group had a positive mammography in a routine screening. What is the probability that she actually has breast cancer? ___ percent

Ten out of every 1000 women at the age of 40 who participate in routine screening have breast cancer. Eight of every 10 women with breast cancer will get a positive mammography. Ninety-five out of every 990 women without breast cancer will also get a positive mammography. Here is a new representative sample of women at the age of 40 who got a positive mammography in routine screening. How many of these women do you expect to actually have breast cancer? ___ out of ___

Representation Conditional Probability Tree

Natural Frequency Tree

1 woman

1000 women

Cancer

No cancer

1%

Cancer

No cancer

10

99%

990

80%

20%

9.6%

90.4%

8

2

95

895

Test positive

Test negative

Test positive

Test negative

Test positive

Test negative

Test positive

Test negative

Inference P(cancer | test positive) =

=

P(test positive | cancer) × P(cancer) P(test positive)

P(cancer | test positive) =

0.8 × 0.01 ≈ 0.08 0.08 × 0.01 + 0.096 × 0.99

=

N(test positive cancer) N(test positive) 8 ≈ 0.08 (8 + 95)

Note: The middle panel shows two types of task representations, a conditional probability tree (left) and a natural frequency tree (right). The bottom row shows two (mathematically equivalent) ways of deriving the posterior probability of having cancer given a positive mammogram, P(cancer|test positive), either by using Bayes’ rule (left) or by deriving it from the natural frequency information. Source: Task descriptions (top row) are taken from Gigerenzer and Hoffrage (1995).

Figure 6.1

Example of a simple probabilistic reasoning task

M4225-ALTMAN_9781782549574_t.indd 113

03/05/2017 08:20

114

Handbook of behavioural economics and smart decision-making

Particular frequency formats, presented verbally or graphically, have been shown to foster people’s inferences, in the laboratory and outside of it. Consider the variant of the mammography problem shown in Figure 6.1 (top right), adapted from Gigerenzer and Hoffrage (1995). Here, instead of using conditional probabilities, information is presented in terms of natural frequencies. The key difference to conveying information in terms of conditional probabilities is that natural frequencies preserve base rate information. The natural frequency tree (Figure 6.1, middle right) illustrates this. This tree represents information as it would result from natural sampling (Kleiter 1994), providing a joint frequency distribution over the two variables (cancer and test result) that reflects the base rate of cancer in the sample (as opposed to systematic sampling, which fixes base rates a priori). Several studies have shown that presenting information this way strongly improves the accuracy of people’s inferences (for a review, see Meder and Gigerenzer 2014). One reason is that the provided information makes it easier to calculate the desired quantity, namely that of 103 women who receive a positive mammogram (95 + 8), only eight actually have breast cancer (Figure 6.1, bottom right). This echoes Simon (1978; see also Larkin and Simon 1987), who noted that two representations are informationally equivalent if one representation can be translated into the other without losing information, but that this does not imply that they are computationally equivalent. Importantly, these findings have guided the development of efficient tools and teaching methods to help people deal with statistical information. Key examples include the use of natural frequencies for understanding the implications of diagnostic tests (Hoffrage and Gigerenzer 1998; Labarge et al. 2003) and forensic evidence (Lindsey et al. 2003), as well as the use of so-called fact boxes to convey medical information in a concise and easily understandable format (Schwartz et al. 2009; Gigerenzer 2014b). Research also shows that training people to use the power of presentation formats is more sustainable than merely teaching them the application of Bayes’ rule (Sedlmeier and Gigerenzer 2001). Over the past decade, different ways have been explored for the intuitive and transparent communication of health information, as well as for the development of graphical presentation formats that help people make sense of health statistics (for reviews see Akl et al. 2011; Gigerenzer et al. 2007). The upshot is that the human mind is not necessarily doomed when it comes to probabilistic thinking. Whereas many researchers endorse the view that people inevitably fall prey to ‘cognitive illusions’, harnessing the power of presentation formats offers a means to help people make sound probabilistic inferences. This, in turn, can provide a foundation for helping people make better decisions without nudging them towards an externally specified goal.

CONCLUSION A functional match between mind and the task environment leads to successful decision making. Fast-and-frugal heuristics are ecologically rational when used under conditions that satisfy such functional matches. Thus, boundedly rational agents make smart decisions by exploiting the ecological rationality of heuristics. Heuristics can capitalize on learned and evolved capacities, or can be designed to create efficient and effective tools

M4225-ALTMAN_9781782549574_t.indd 114

03/05/2017 08:20

Heuristics: fast, frugal, and smart 115 for decision making. A heuristic is neither good nor bad per se. Rather, the effectiveness of a heuristic strategy can only be gauged with respect to the structure of information in the environment within which it is used. As such, errors can be scrutinized as informative where they are indicating a mismatch between the environment, strategies, or evolved and learned capacities. Specifying proper matches and teasing out the mismatches between heuristics and their task environment constitutes the study of ecological rationality of heuristics. Intelligent behavior, when appearing less than neoclassically rational, can be understood by breaking free of the restrictive benchmarks imposed by constructivist rationality. The answer is to be found in the ecological rationality of intelligent behavior, because in many real-world situations it is simply not rational to be rational. Through the complementary frame of ecological rationality, intelligence can be understood to be beyond the agents’ mind and without requiring complete comprehension of rules. Also, smart behavior emerges where proper evolved or learned capacities are triggered in reaction to the structure of the task environment. This chapter primarily focused on recent theoretical developments regarding the ecological nature of humans’ bounded rationality that ought to be of interest to behavioral economists. We invite economists and psychologists to join our dialogue and dig into less explored insights from psychology that promise informing behavioral economics on realworld, practical, and smart decision making. Beyond encouraging economists to adopt psychological insights into their models and theories, Simon (1959, p. 253) also challenges economists to communicate their ideas and findings to psychologists: the psychologist will also raise the converse question – whether there are developments in economic theory and observation that have implications for the central core of psychology . . . Influence will run both ways.

NOTES 1. Fast-and-frugal heuristics are interchangeably used in this text with simple and smart heuristics. 2. Savage (1954) drew significantly on the Theory of Games and Economic Behavior (von Neumann and Morgenstern 1947) and proposed a normative reading of the subjective expected utility theoretical framework. 3. McCloskey (1991) spells out the pragmatic significance of this point in ‘Economic science: a search through the hyperspace of assumptions?’ where she portrays the practice of axiomatic economics as a mathematical practice faithful to math departments’ ideal of consistency and coherence, but incapable of grasping and providing solutions to real-world problems. 4. A series of papers (Journal of Business Research, vol. 67, 2014) on the effectiveness of fast-and-frugal heuristics in business decision-making demonstrate this point.

REFERENCES Aikman, D., M. Galesic, G. Gigerenzer, S. Kapadia, K.V. Katsikopoulos, A. Kothiyal et al. (2014), ‘Taking uncertainty seriously: simplicity versus complexity in financial regulation’, Financial Stability Paper No. 28, May, Bank of England, London. Akerlof, G.A. and R.J. Shiller (2015), Phishing for Phools: The Economics of Manipulation and Deception, Princeton, NJ: Princeton University Press.

M4225-ALTMAN_9781782549574_t.indd 115

03/05/2017 08:20

116

Handbook of behavioural economics and smart decision-making

Akl, E.A., A.D. Oxman, J. Herrin, G.E. Vist, I. Terrenato, F. Sperati et al. (2011), ‘Using alternative statistical formats for presenting risks and risk reductions’, Cochrane Database of Systematic Reviews, 16 March, CD006776, doi:10.1002/14651858.CD006776.pub2. Altman, M. (2004), ‘The Nobel Prize in behavioral and experimental economics: a contextual and critical appraisal of the contributions of Daniel Kahneman and Vernon Smith’, Review of Political Economy, 16 (1) 3–41. Altman, M. (2012), ‘Implications of behavioural economics for financial literacy and public policy’, Journal of Socio-Economics, 41 (5), 677–90. Benartzi, S. and R.H. Thaler (2001), ‘Naive diversification strategies in defined contribution saving plans’, American Economic Review, 91 (1), 79–98. Berg, N. and G. Gigerenzer (2010), ‘As-if behavioral economics: neoclassical economics in disguise?’, History of Economic Ideas, 18 (1), 133–66. Brandstätter, E., G. Gigerenzer and R. Hertwig (2006), ‘Priority heuristic: making choices without trade-offs’, Psychological Review, 113 (2), 409–32. Daston, L.J. (1988), Classical Probability in the Enlightenment, Princeton, NJ: Princeton University Press. DeMiguel, V., L. Garlappi and R. Uppal (2009), ‘Optimal versus naive diversification: how inefficient is the 1/N portfolio strategy?’, Review of Financial Studies, 22 (5), 1915–53. Dewey, J. (1938), ‘Logic: the theory of inquiry’, reprinted in J.A. Boydston (ed.) (1986), John Dewey: The Later Works, vol. 12, Carbondale, IL: Southern Illinois University Press. Eddy, D.M. (1982), ‘Probabilistic reasoning in clinical medicine: problems and opportunities’, in D. Kahneman, P. Slovic and A. Tversky (eds), Judgment under Uncertainty: Heuristics and Biases, Cambridge: Cambridge University Press, pp. 249–67. Edwards, W. (1968), ‘Conservatism in human information processing’, in B. Kleinmuntz (ed.), Formal Representation of Human Judgment, New York: Wiley, pp. 17–52. Geman, S., E. Bienenstock and R. Doursat (1992), ‘Neural networks and the bias/variance dilemma’, Neural Computation, 4 (1), 1–58. Gigerenzer, G. (1991), ‘How to make cognitive illusions disappear: beyond “heuristics and biases”’, European Review of Social Psychology, 2 (1), 83–115. Gigerenzer, G. (1996), ‘On narrow norms and vague heuristics: a reply to Kahneman and Tversky (1996)’, Psychological Review, 103(3), 592–6. Gigerenzer, G. (2008), Rationality for Mortals: How People Cope with Uncertainty, New York: Oxford University Press. Gigerenzer, G. (2014a), ‘Breast cancer screening pamphlets mislead women’, BMJ, 348 (25 April), g2636–g2636, doi:10.1136/bmj.g2636. Gigerenzer, G. (2014b), Risk Savvy: How to Make Good Decisions, New York: Viking. Gigerenzer, G. and H.J. Brighton (2009), ‘Homo heuristicus: why biased minds make better inferences’, Topics in Cognitive Science, 1 (1), 107–43, doi:10.1111/j.1756-8765.2008.01006.x. Gigerenzer, G. and A. Edwards (2003), ‘Simple tools for understanding risks: from innumeracy to insight’, BMJ, 327 (7417), 741–4, doi:10.1136/bmj.327.7417.741. Gigerenzer, G. and U. Hoffrage (1995), ‘How to improve Bayesian reasoning without instruction: frequency formats’, Psychological Review, 102 (4), 684–704, doi:10.1037/0033-295X.102.4.684. Gigerenzer, G. and K. Hug (1992), ‘Domain-specific reasoning: social contracts, cheating, and perspective change’, Cognition, 43 (2), 127–71. Gigerenzer, G. and J.A. Muir Gray (eds) (2011), Better Doctors, Better Patients, Better Decisions: Envisioning Health Care 2020, Cambridge, MA: MIT Press. Gigerenzer, G., W. Gaissmaier, E. Kurz-Milcke, L.M. Schwartz and S. Woloshin (2007), ‘Helping doctors and patients make sense of health statistics’, Psychological Science in the Public Interest, 8 (2), 53–96, doi:10.1111/j.1539-6053.2008.00033.x. Gigerenzer, G., P.M. Todd and the ABC Research Group (1999), Simple Heuristics that Make Us Smart, New York: Oxford University Press. Goldman, A. (1988), Empirical Knowledge, Berkeley, CA: University of California Press. Gøtzsche, P.C. and K.J. Jørgensen (2011), ‘The breast screening programme and misinforming the public’, Journal of the Royal Society of Medicine, 104 (9), 361–9, doi:10.1258/jrsm.2011.110078. Gøtzsche, P.C. and K.J. Jørgensen (2013), ‘Screening for breast cancer with mammography’, Cochrane Database of Systematic Reviews, (6), 4 June, CD001877, doi:10.1002/14651858.CD001877.pub5. Green, R.F. (1984), ‘Stopping rules for optimal foragers’, American Naturalist, 123 (1), 30–43. Green, L. and D.R. Mehr (1997), ‘What alters physicians’ decisions to admit to the coronary care unit?’, Journal of Family Practice, 45 (3), 219–26. Grüne-Yanoff, T. and R. Hertwig (2015), ‘Nudge versus boost: how coherent are policy and theory?’, Minds and Machines, 25 (1–2), 1–35, doi:10.1007/s11023-015-9367-9.

M4225-ALTMAN_9781782549574_t.indd 116

03/05/2017 08:20

Heuristics: fast, frugal, and smart 117 Hertwig, R., J.N. Davis and F.J. Sulloway (2002), ‘Parental investment: how an equity motive can produce inequality’, Psychological Bulletin, 128 (5), 728–45. Hoffrage, U. and G. Gigerenzer (1998), ‘Using natural frequencies to improve diagnostic inferences’, Academic Medicine: Journal of the Association of American Medical Colleges, 73 (5), 538–40. Jacobs, H., S. Müller and M. Weber (2013), ‘How should individual investors diversify? An empirical evaluation of alternative asset allocation policies’, Journal of Financial Markets, 19 (June), 62–85. Jenny, M.A., T. Pachur, S. Lloyd Williams, E. Becker and J. Margraf (2013), ‘Simple rules for detecting depression’, Journal of Applied Research in Memory and Cognition, 2 (3), 149–57. Kahneman, D. and A. Tversky (1973), ‘On the psychology of prediction’, Psychological Review, 80 (4), 237–51, doi:10.1037/h0034747. Kahneman, D., P. Slovic and A. Tversky (eds) (1982), Judgment under Uncertainty: Heuristics and Biases, Cambridge: Cambridge University Press. Katsikopoulos, K.V. and G. Gigerenzer (2008), ‘One-reason decision-making: modeling violations of expected utility theory’, Journal of Risk and Uncertainty, 37 (1), 35–56, doi:10.1007/s11166-008-9042-0. Kleiter, G.D. (1994), ‘Natural sampling: rationality without base rates’, in Contributions to Mathematical Psychology, Psychometrics, and Methodology, New York: Springer, pp. 375–88. Knight, F.H. (1921), Risk, Uncertainty and Profit, New York: Hart, Schaffner and Marx. Koehler, J.J. (1996), ‘The base rate fallacy reconsidered: descriptive, normative and methodological challenges’, Behavioral and Brain Sciences, 19 (1), 1–54. Labarge, A.S., R.J. McCaffrey and T.A. Brown (2003), ‘Neuropsychologists’ abilities to determine the predictive value of diagnostic tests’, Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists, 18 (2), 165–75. Larkin, J.H. and H.A. Simon (1987), ‘Why a diagram is (sometimes) worth ten thousand words’, Cognitive Science, 11 (1), 65-100, doi:10.1111/j.1551-6708.1987.tb00863.x. Lindsey, S., R. Hertwig and G. Gigerenzer (2003), ‘Communicating statistical DNA evidence’, Jurimetrics, 43 (2), 147–63. Markowitz, H.M. (1952), ‘Portfolio selection’, Journal of Finance, 7 (1), 77–91. Martignon, L., O. Vitouch, M. Takezawa and M.R. Forster (2003), ‘Naïve and yet enlightened: from natural frequencies to fast and frugal decision trees’, in D. Hardman and L. Macchi (eds), Thinking: Psychological Perspectives on Reasoning, Judgment and Decision Making, Chichester: John Wiley & Sons, pp. 189–211. McCloskey, D. (1991), ‘Economic science: a search through the hyperspace of assumptions?’, Methodus, 3 (1), 6–16. McLeod, P. and Z. Dienes (1996), ‘Do fielders know where to go to catch the ball, or only how to get there?’, Journal of Experimental Psychology: Human Perception and Performance, 22 (3), 531–43, doi:10.1037/0096-1523.22.3.531. Meder, B. and G. Gigerenzer (2014), ‘Statistical thinking: no one left behind’, in E.J. Chernoff and B. Sriraman (eds), Probabilistic Thinking, Dordrecht: Springer, pp. 127–48. Meder, B., F. Le Lec and M. Osman (2013), ‘Decision making in uncertain times: what can cognitive and decision sciences say about or learn from economic crises?’, Trends in Cognitive Sciences, 17 (6), 257–60, doi. org/10.1016/j.tics.2013.04.008. Messick, D.M. (2008), ‘Equality as a decision heuristic’, in B.A. Mellers and J. Baron (eds), Psychological Perspectives on Justice: Theory and Applications, Cambridge, MA: Cambridge University Press, pp. 11–31. Mousavi, S. and G. Gigerenzer (2011), ‘Revisiting the “error” in studies of cognitive errors’, in D.A. Hofmann and M. Frese (eds), Errors in Organizations, New York: Routledge, pp. 97–112. Mousavi, S. and G. Gigerenzer (2014), ‘Risk, uncertainty, and heuristics’, Journal of Business Research, 67 (8), 1671–8. Mousavi, S. and R. Kheirandish (2014), ‘Behind and beyond a shared definition of ecological rationality: a functional view of heuristics’, Journal of Business Research, 67 (8), 1780–85. Moyer, V.A. (2012), ‘Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement’, Annals of Internal Medicine, 157 (2), 120–34, doi:10.7326/0003-4819-157-2-201207170-00459. Neth, H. and G. Gigerenzer (2015), ‘Heuristics: tools for an uncertain world’, in R. Scott and S. Kosslyn (eds), Emerging Trends in the Social and Behavioral Sciences: An Interdisciplinary, Searchable, and Linkable Resource, New York: Wiley. pp. 1–18. Neth, H., B. Meder, A. Kothiyal and G. Gigerenzer (2014), ‘Homo heuristicus in the financial world: from risk management to managing uncertainty’, Journal of Risk Management in Financial Institutions, 7 (2), 134–44. Payne, S., G. Duggan and H. Neth (2007), ‘Discretionary task interleaving: heuristics for time allocation in cognitive foraging’, Journal of Experimental Psychology: General, 136 (3), 370–80. Peterson, C.R. and L.R. Beach (1967), ‘Man as an intuitive statistician’, Psychological Bulletin, 68 (1), 29–46. Phillips, L.D. and W. Edwards (1966), ‘Conservatism in a simple probability inference task’, Journal of Experimental Psychology, 72 (3), 346–54.

M4225-ALTMAN_9781782549574_t.indd 117

03/05/2017 08:20

118

Handbook of behavioural economics and smart decision-making

Ploug, T., S. Holm and J. Brodersen (2012), ‘To nudge or not to nudge: cancer screening programmes and the limits of libertarian paternalism’, Journal of Community Health, 66 (12), 1193–6. Savage, L.J. (1954), The Foundations of Statistics, 2nd edn, New York: Dover. Schwartz, L.M., S. Woloshin and H.G. Welch (2009), ‘Using a drug facts box to communicate drug benefits and harms: two randomized trials’, Annals of Internal Medicine, 150 (8), 516–27. Sedlmeier, P. and G. Gigerenzer (2001), ‘Teaching Bayesian reasoning in less than two hours’, Journal of Experimental Psychology: General, 130 (3), 380–400. Selten, R. (1998), ‘Aspiration adaptation theory’, Journal of Mathematical Psychology, 42 (2), 191–214. Simon, H.A. (1955), ‘A behavioral model of rational choice’, Quarterly Journal of Economics, 69 (1), 99–118. Simon, H.A. (1959), ‘Theories of decision-making in economics and behavioral science’, American Economic Review, 49 (3), 253–83. Simon, H.A. (1978), ‘On the forms of mental representation’, in C.W. Savage (ed.), Perception and Cognition. Issues in the Foundations of Psychology, vol. 9, Minneapolis, MN: University of Minnesota Press, pp. 3–18. Smith, V.L. (2002), ‘Constructivist and ecological rationality in economics’, accessed 20 February 2015, at http:// www.nobelprize.org/nobel_prizes/economic-sciences/laureates/2002/smith-lecture.html. Smith, V.L. (2008), Rationality in Economics: Constructivist and Ecological Forms, New York: Cambridge University Press. Stephens, D.W. and J.R. Krebs (1986), Foraging Theory, Princeton, NJ: Princeton University Press. Thaler, R.H. and C.R. Sunstein (2008), Nudge: Improving Decisions about Health, Wealth, and Happiness, New Haven, CT: Yale University Press. Tversky, A. and D. Kahneman (1974), ‘Judgment under uncertainty: heuristics and biases’, Science, 185 (4157), 1124–31. Von Neumann, J. and O. Morgenstern (1947), Theory of Games and Economic Behavior, 2nd edn, Princeton, NJ: Princeton University Press. Wason, P.C. (1966), ‘Reasoning’, in B.M. Foss (ed.), New Horizons in Psychology, vol. 1, Harmondsworth: Penguin Books. Wübben, M. and F. von Wangenheim (2008), ‘Instant customer base analysis: managerial heuristics often “get it right’’’, Journal of Marketing, 72 (3), 82–93.

M4225-ALTMAN_9781782549574_t.indd 118

03/05/2017 08:20

7

The beauty of simplicity? (Simple) heuristics and the opportunities yet to be realized* Andreas Ortmann and Leonidas Spiliopoulos

INTRODUCTION In this chapter we focus on the history of fast and frugal heuristics, as sketched out comprehensively in Gigerenzer et al. (1999) and scores of follow-up books (for example, Gigerenzer et al. 2011; Todd et al. 2012; Hertwig et al. 2013) and articles. What we consider must-read papers are listed in the further reading section at the end of the chapter. A recurring theme of this edited volume is that individuals can be smart and procedurally rational despite displaying errors in decisions. Such an argument implicitly assumes that there is an effort–accuracy tradeoff. Consider the feasible set of combinations of effort and accuracy as being constrained by the decision maker’s cognitive processes and features of the decision environment – this is the basis of prominent theories of bounded rationality. In this view, decision errors can be rationalized by arguing that regardless of which specific combination of effort and accuracy is chosen, as long as it is on the efficient frontier, a choice resulting in said decision error cannot automatically be classified as irrational per se. Errors are thus uncoupled from the notion of rationality, in contrast to neoclassical economics where errors are synonymous with irrational behavior. The notion of fast and frugal heuristics goes a step further than this argument, and its proponents contend that there exist decision environments – found with sufficiently high frequency in the real world – which can be exploited by appropriately adapted heuristics in a way that transcends the effort–accuracy tradeoff. Under such circumstances, normative models such as expected utility may be dominated by simple heuristics in both the accuracy and effort dimensions. We contextualize the emergence of this so-called ‘Ecological-Rationality’ (ER from here on) program as an explicit counterpoint to the ‘Heuristics-and-Biases’ (HandB from here on) program initiated by Kahneman and Tversky (for example, Tversky and Kahneman 1974; Kahneman and Tversky 1979; Kahneman 2003a, 2003b, 2011) that informed and inspired scores of early behavioral economists. Simple heuristics are here understood to be fast and frugal rules of thumb because they ignore information that is available and hence can shorten decision-making time. Also, they ought to reflect cognitive processes (and hence be able to predict) rather than be as-if modelling exercises that explain ex post. Our focus seems warranted by the fact that the HandB program has invaded economics, and other social sciences, to the extent that it is now by many measures thoroughly mainstream (for example, Camerer et al. 2004, 2011; Heukelom 2015; Thaler 2015). While in the past few years increasingly critical questions have been asked about the HandB program (for example, Ortmann 2015a, 2015b and references therein), the predominance 119

M4225-ALTMAN_9781782549574_t.indd 119

03/05/2017 08:20

120

Handbook of behavioural economics and smart decision-making

of it has largely overshadowed the ER program, in our view to the detriment of both economics and the ER program. It has not helped that those in favor of an ER program have not done as much out-reach to economics as might have been desirable. Sketching with a very broad brush, we argue that the HandB program suggested that various bounds on rationality, and the make-up of human judgment- and decisionmaking facilities, induced humans to make rash decisions that produced systematic biases, or cognitive illusions. Cognitive illusions were rationalized with reference to optical illusions whose reality was well established. The heuristics that people were said to use, such as representativeness, availability, and anchoring and adjustment, were motivated by appeal to the principles also underlying optical illusions. An implicit – and increasingly explicit claim (for example, Thaler 1980, p. 40) – was that cognitive illusions were as robust as optical illusions (see also Kahneman 2003a, 2003b). Heuristics were considered to be problematic and decision makers as fallible, even gullible, and in dire need of all the help that they could get to improve on their decision-making skills. As Cochrane (2015) has noted, not inappropriately, this view represents for the HandB program proponents a considerable moral hazard problem. It is worth noting that the assessment of people’s performance as being severely wanting was quite a departure from the prevailing view in the 1950s, 1960s and early 1970s (for example, Edwards 1956; Peterson and Beach 1967; see also Ortmann 2015a). Even Tversky and Kahneman (1974), in the article that started it all, did not make the kind of sweeping claims that were made in the following decades. Drawing on arguments by Herb Simon (1947, 1955, 1956) and his insight that rationality cannot be defined through cognitive and emotional processes alone, Gigerenzer and the ABC Research group showed that many of the demonstrations of the HandB program were highly problematic. The main criticism was directed at the design and implementation of the experiments used to produce supporting evidence (for example, prominently Gigerenzer 1991), and that indeed heuristics could have surprising performance properties, particularly so as environments became more uncertain (Gigerenzer and Gaissmaier 2011). We first review in more detail how this battle of programs unfolded, then lay out what we consider the considerable accomplishments of the ER program and point out some overlooked connections between the ER program and economics, and finally, enumerate what we consider to be open questions and challenges. In the interest of full disclosure, we note that both authors spent time at the Max Planck Institute for Human Development, which now houses the ABC and ARC groups (both of which contribute to the ER program; more about this below), Ortmann for one year each in 1996–97 and 1999–2000 with the ABC group, and Spiliopoulos having visited the ABC group twice (for a couple of weeks each) and since mid-2014 being first a Humboldt Experienced Research Fellow with the ARC group and then a senior researcher.

M4225-ALTMAN_9781782549574_t.indd 120

03/05/2017 08:20

The beauty of simplicity? 121

HOW THE BATTLE OF THE HANDB PROGRAM AND ER PROGRAM UNFOLDED First, the Heuristics and Biases Program (HandBP) Calling Richard Cyert, James March, Herbert Simon the ‘old behavioral economists, who focused on bounded rationality, satisficing, and simulations’ (Sent 2004, p. 740), historian of economics Esther-Mirjam Sent explained the transition from old to new behavioral economics (ibid., pp. 742–7), thus: ‘The roots of new behavioral economics may be traced to the 1970s and the work of especially Amos Tversky and Daniel Kahneman’ (ibid., p. 742). She identifies the ‘Behavioral foundations of economic theory’ conference held at the University of Chicago in October 1985 as a key event. In the preface to their book that drew on the conference, Hogarth and Reder (1987, p. vii) argued that there was ‘a growing body of evidence – mainly of an experimental nature – that has documented systematic departures from the dictates of rational economic behaviour.’ In his review of the book, Smith (1991, p. 878) dismissed such a claim: ‘(experimental economics) documents a growing body of evidence that is consistent with the implications of rational models’. Acknowledging that Simon’s work on bounded rationality had influenced them, too, Kahneman (2003a, p. 1449) identified three separate lines of research. The first explored the heuristics that people use and the biases to which they are prone in various tasks of judgment under uncertainty, including predictions and evaluations of evidence . . . The second was concerned with prospect theory, a model of choice under risk . . . and with loss aversion in riskless choice . . . The third line of research dealt with framing effects and with their implications for rational-agent models . . .

and Our research attempted to obtain a map of bounded rationality, by exploring the systematic biases that separate the beliefs that people have and the choices they make from the optimal beliefs and choices assumed in rational-agent models. The rational-agent model was our starting point and the main source of our null hypotheses, but Tversky and I viewed our research primarily as a contribution to psychology, with a possible contribution to economics as a secondary benefit. We were drawn into the interdisciplinary conversation by economists who hoped that psychology could be a useful source of assumptions for economic theorizing, and indirectly a source of hypotheses for economic research (Richard H. Thaler, 1980). (Kahneman 2003a, p. 1449)

Kahneman and Tversky’s HandBP was based on the idea that thinking was typically fast and rarely slow, and very fundamentally about accessibility or intuition. The argument was that, since our thinking was typically fast, it had to rely on rules of thumb (heuristics) which led to systematic divergences (biases) from normative behavior as described by standard economic theories (Tversky and Kahneman 1974; Kahneman and Tversky 1996; Kahneman 2003a, 2003b). People were increasingly conceptualized as bumbling fools and this theme was the general drift taken up by those starting the movement that later became behavioral economics. Thaler (1980, p. 40), for example, exclaimed that ‘Research on judgment and decision making under uncertainty, especially by Daniel Kahneman and Amos Tversky (1974; Tversky and Kahneman 1979) has shown that . . .

M4225-ALTMAN_9781782549574_t.indd 121

03/05/2017 08:20

122

Handbook of behavioural economics and smart decision-making

mental illusions should be considered the rule rather than the exception. Systematic, predictable differences between normative models of behavior and actual behavior occur . . .’. Importantly, the cognitive illusions were explicitly constructed (for example, Kahneman 2003a, 2003b) in parallel to optical illusions whose reality and robustness had been reasonably well established. It is striking that the optical-illusion analogy was not taken to its logical conclusion, namely, that the documented illusions either never occur in the environment or, in the few instances when they do, they rarely impose any real cost on the organism. We have argued elsewhere (Spiliopoulos and Ortmann 2014) that specific diagnostic tasks, that is, specific parameterizations of tasks where competing models make starkly different predictions, should not be used to infer the rationality of agents. Rationality can only be assessed on a wide range of parameterizations that must include those found in the real environment (on this, see Hogarth and Karelaia 2005, 2006, 2007; Erev et al. 2017). There were some obvious problems with the HandB approach, and two decades ago they were the subject of a highly visible dispute between Kahneman and Tversky (1996) and Gigerenzer (1996) about the reality of cognitive illusions. From the critics’, and the present authors’, view the HandBP was characterized by a lack of process models (key concepts such representativeness, anchoring and adjustment, and availability being hardly more than labels), too much story-telling, un-incentivized scenario studies, polysemy, often deception, and experimenter demand effects, to name a few. There was, in Nobel Prize laureate Vernon L. Smith’s sarcastic but brilliant observation, too much fishing in the tail ends of human behavior (Smith 2003, p. 467, fn. 8). No surprise then that many anomalies were found that were taken as proof of people’s limited rationality. The interpretation of that evidence as being indicative of humans’ typically underwhelming performance has been contested ever since it was proposed, by the ER program and many others working in the neoclassical tradition (for example, Smith 1991). Second, the Ecological-Rationality Program (ERP) The ABC research program (see also Lopes 1992) was constructed in contrast to the HandBP. Gigerenzer (1991), for example, successfully deconstructed some key findings of Kahneman and Tversky who eventually found themselves prompted to respond to Gigerenzer’s critique (Kahneman and Tversky 1996; Gigerenzer 1996). ABC also developed a fundamentally different view of heuristics and did so by formulating cognitive process models that could be tested. It is interesting to note that many of the process models were also based on a frequentist view of the world, with ABC researchers taking broadly an evolutionary-psychology perspective, which conceptualized humans as intuitive statisticians that were almost naturally good at navigating environs that were familiar to them. It was also demonstrated persuasively that an important moderator of these findings is the way statistical information is presented (Sedlmeier and Gigerenzer 2001; see Hertwig and Ortmann 2004 for a summary). To the extent that the HandBP was gobbled up entirely by the initial waves of behavioral economics/finance, ABC remained an outsider of sorts although its influence has grown, as recently evidenced by a 20-year celebration that was attended by more than 100 participants. Part of the problem is that ABC rarely engaged with modern economics and focused its critiques on normative economic models of deductive reasoning. We

M4225-ALTMAN_9781782549574_t.indd 122

03/05/2017 08:20

The beauty of simplicity? 123 argue below that important work assuming inductive reasoning in economics can serve as a bridge with the ERP, although important differences remain, and considerable opportunities have yet to be realized.

THE ACCOMPLISHMENTS OF THE ECOLOGICALRATIONALITY PROGRAM The ERP is characterized by a heavy reliance on cognitive process models (which require some serious theorizing), empirical and experimental testing of these models, and an important methodological innovation: the preferred mode of testing relies on ‘out-ofsample’ prediction, or ‘cross-validation’ (Gigerenzer and Gaissmaier 2011). That is, performance is not measured by the best fit on an existing dataset but by the performance of a model on not yet known datasets, also done in Ericson et al. (2015) and Erev et al. (2017). There is no data-fitting after the fact. Cross-validation addresses the important bias–variance tradeoff (Gigerenzer and Brighton 2009). Simple models exhibit higher bias but typically less variance than complex models – it is the relative strength that determines which type of model outperforms the other in prediction. The key finding is that heuristics often exhibit little or no bias vis-à-vis more complex models, therefore the variance effect tends to dominate; we return to this below. Among ERP’s key successful demonstrations is that, when cross-validation is used, the performance of simple heuristics such as the recognition heuristic or the ‘take-the-best’ is better than that of complicated, computationally slow and greedy models such as multiple regression favored by economists (for example, Gigerenzer et al. 1999; Gigerenzer and Brighton 2009; Todd et al. 2012; Gigerenzer and Gaissmaier 2011). The simple, and rather intuitive, reason is that multiple regression essentially over-fits, looking backwards, without taking into account the noisiness that is inherent in datasets. An important implication is that the widely believed effort–accuracy trade-off (Payne et al. 1993) is often not something we need to worry about. Those using simple heuristics can have both. Less can be more. Much work has been done to understand these remarkable results; what drives the success of heuristics such as recognition and take-the-best is now much better understood (Baucells et al. 2008; Luan et al. 2011, 2014; Katsikopoulos 2013; Drechsler et al. 2014). There are three important environmental characteristics that are sufficient, but not necessarily necessary, that induce these striking results: non-compensatoriness of cues, dominance, and cumulative dominance.1 If at least one of these three is true, a lexicographic heuristic exhibits no bias vis-à-vis a linear rule, and is computationally less demanding. These theoretical results would not be important if these conditions were not found regularly in real environments. Şimşek (2013), however, found that these conditions are very common in 51 real-world datasets; consequently, a lexicographic heuristic performed as well as multiple linear regressions in the median dataset for approximately 90 percent of cases. Recent work has analyzed fast-and-frugal trees and successfully connected them to signal-detection theory (Luan et al. 2011, 2014); new heuristics such as the fluency and priority heuristics have been proposed (Hertwig et al. 2008; Brandstätter et al. 2006, 2008; Drechsler et al. 2014; but see also Johnson et al. 2008 on the priority heuristic); and a persuasive rationalization has been provided for the tendency of many

M4225-ALTMAN_9781782549574_t.indd 123

03/05/2017 08:20

124

Handbook of behavioural economics and smart decision-making

economists and psychologists to overlook the benefits of simplicity (Brighton and Gigerenzer 2015). Another important contribution of the ERP is the distinction between ‘decisionsfrom-description’ (DfD) and ‘decisions-from-experience’ (DfE) and the empirical validation of a robust gap between the two (for example, Barron and Erev 2003; Hertwig and Erev 2009). It is indeed intuitive that risk maps into, or maybe better invokes, DfD, and that uncertainty maps into, DfE. Furthermore, these map into Savage’s (1954) distinction between small (DfD) and large (DfE) world decision making (see Gigerenzer and Gaissmaier 2011 for a discussion). We also note parenthetically that in strategic environments DfD and DfE also map into eductive and evolutive (deductive and inductive) game theory (Binmore 1990; Friedman 1991). Hopefully, researchers both at ABC and ARC continue recent attempts at theory integration (for example, Schooler and Hertwig 2005, Luan et al. 2011, 2014) and related attempts to break down disciplinary boundaries (for example, Hutchinson and Gigerenzer 2005). This was, to some extent, also reflected in the make-up of the ABC Research Group but perhaps not as much as would have been desirable ex post in particular regarding the group’s engagement with economists. An increasing number of economists and researchers from management and organization (no wonder here, given where it all started: Simon 1955, 1956) have been attracted by the ER paradigm. For example, Åstebro and Elhedli (2006) have empirically demonstrated the usefulness of simple heuristics in forecasting commercial success for earlystage ventures. Eisenhardt and some of her colleagues (see, for a self-centered primer, Bingham and Eisenhardt 2014) have argued that successful repeated product innovation is best implemented through ‘simple rules’, or ‘semi-structures’, which define a path between too much and too little structure. Maitland and Sammartino (2014) have empirically demonstrated the use of simple decision rules for location choice by multinational companies when environments are politically hazardous. Indeed, Artinger et al. (2014) have provided a useful primer of heuristics as adaptive decision strategies in management but it seems clear that the use of heuristics in management and organization is understudied and remains a fruitful area of research. To see how understudied the topic is academically, Google strings such as ‘rules of thumbs to determine when projects pay off’ find more than 14 million hits and scores of lists of simple decision rules for everything from cash flow, real-estate investments, to other financial investments. While there can be no doubt that progress towards a science of heuristics has been tremendous and that the ABC group’s influence is increasing, there remain important blind spots though in our view.

THE ERP AND ECONOMICS – A MISSED OPPORTUNITY (SO FAR) The incompatibility of the ERP with economics has been emphasized by a number of ERP researchers. To a large extent, the ERP is positioned as an antithesis both to the HandBP and the neoclassical-economics program, including behavioral economics, which some view as a disguised extension of the neoclassical program (Berg and Gigerenzer 2010). We are sympathetic to the claims made, as far as they pertain to the overwhelming mass of research often dubbed behavioral economics. Exceptions to this exist, this book

M4225-ALTMAN_9781782549574_t.indd 124

03/05/2017 08:20

The beauty of simplicity? 125 serving as a case in point. For example, we ourselves have argued about the advantages of process models compared to as-if models (Spiliopoulos and Ortmann 2015). However, we fear that a purely antagonistic approach of emphasizing the divide has the unfortunate consequence of deepening the schism rather than fostering an exchange between these programs. The differences in opinions are well known; here we will attempt to highlight (perhaps surprising) similarities between these research programs; indeed in some cases we will find parallel, independent emergence of similar ideas. This suggests that there is significant scope for future exchange of ideas and productive collaboration between researchers from the two fields. Heuristics Extremely interesting work from economists like Manzini and Mariotti (for example, 2007, 2012a, 2012b, 2014; see also Mandler et al. 2012) seems to have developed in parallel to the work of the ABC Research Group. Parallel, yet mostly independent, work can be scientifically counterproductive in the sense that closer collaboration could have afforded increasing returns to research and the avoidance of duplication (for example, see Arkes and Ayton 1999 on the Concorde fallacy and related work in economics on sunk cost effects such as Friedman et al. 2007 and McAffee et al. 2010). Broadly inspired by the work of Gigerenzer and associates, the well-cited Manzini and Mariotti (2007) formalizes and axiomatizes a type of sequential eliminative heuristic demonstrating that boundedly rational choice procedures can be tested with observable choice (‘revealed preference’) data favored by more traditional economists. The more recent Manzini and Mariotti (2012a, 2014) builds on this earlier two-stage deterministic model of choice by providing models of stochastic choice when consideration sets are present (that is, agents fail to consider all feasible alternatives), a popular but typically less formalized approach in management science and marketing science that is related to random utility models that have been around for decades in economics. Mandler et al. (2012) provides procedural foundations for utility maximization, with the checklists in the title of their paper being the equivalent of the – preferably noncompensatory – cues central to the fast and frugal heuristics extensively analyzed by the ABC Research Group. The authors show that under specific conditions procedural utility maximization matches that of substantive utility. In Manzini and Mariotti (2012b), the authors extend and formalize a choice procedure introduced by Tversky (1969) that has recently also prominently featured in the work of Luan and colleagues (Luan et al. 2011, 2014). How to Choose Heuristics from the Adaptive Toolbox? Initial criticisms that the ERP had not adequately specified the heuristic selection method of the adaptive toolbox has prompted work directed at strategic selection. The most prominent response to this critique was to postulate a reinforcement learning mechanism over heuristics (Rieskamp and Otto 2006) – see also the RELACS model by Erev and Baron (2005). This is essentially the same solution proposed for strategic decision making by economists. For example, Aumann (1997, pp. 7–8) writes: ‘Ordinary people do not behave in a consciously rational way in their day-to-day activities. Rather, they

M4225-ALTMAN_9781782549574_t.indd 125

03/05/2017 08:20

126

Handbook of behavioural economics and smart decision-making

evolve “rules of thumb” that work in general, by an evolutionary process . . . or a learning process with similar properties.’ In the El Farol bar problem (Arthur 1994), agents hold a heterogeneous set of simple predictive models and learn to use the more effective rules (given their individual experience) over time; interestingly, such a learning process converges to the Nash equilibrium solution. Empirical work in repeated games by Stahl (1996, 1999, 2000) and Haruvy and Stahl (2012) find evidence that subjects learn to use relatively simple rules based on their prior performance – they refer to their model as rule-learning. These are concepts strikingly similar to those proposed by the ERP; however, the ERP studies were in the domain of individual decision making, whereas the economic studies are in strategic decision making. Clearly, there is potential here for both disciplines to interact and advance our knowledge of the strategy selection problem. What is the Appropriate Performance Metric for Model Comparisons? The ERP has promoted, rightly in our view, the use of cross-validation to compare the performance of heuristics to other more complex models, hence shifting the focus from explanation to prediction. This is a consequence of the effects of the bias–variance dilemma. More complex models will tend to fit better in-sample than simpler models (such as heuristics), but may perform worse on out-of-sample predictions. Friedman (1953) was an early proponent of the notion that theories should be evaluated on the basis of their predictive power; of course, ERP researchers would take issue with his contention that the processes (and underlying assumptions) are irrelevant – see, for example, the billiard player example in Friedman and Savage (1948). Studies published in prominent economics journals as far back as Camerer and Ho (1999), and including more recent work such as Wilcox (2011) and Spiliopoulos (2012, 2013), have also argued for, and used, cross-validation. See also Erev et al. (2017) and literature therein. What is the Appropriate Space for the Calculation of Deviations from Rationality? A further issue concerns how we measure deviations from rationality, if they exist at all. The ERP focuses on deviations in the consequence space, that is, comparing the actual loss in terms of the consequences of a behavior. Consequences can be actual payoffs, if they are well defined for a problem, or a metric based on the percentage of correct/wrong responses often used in binary tasks. Using deviations in the consequence space instead of the choice space is important, as seemingly large differences in choice may not translate into large deviations in the consequence space, particularly when computational costs are included. In the early history of Behavioral Economics, deviations from rationality were typically measured in the choice space, and this still occurs to a considerable extent. However, experimental economists have taken issue with experiments that have a flat payoff function around the normative solution, culminating in the payoff-dominance critique (Harrison 1989) that prompted a large debate in the field (see the comments and replies to this paper in the American Economic Review, 82 (5) in 1992). While originally intended as a critique of the design of many experiments in economics, implicit in the payoff-dominance critique is the notion that non-optimal behavior can only be identified when it is accompanied by large costs in the consequence space. A large deviation in the

M4225-ALTMAN_9781782549574_t.indd 126

03/05/2017 08:20

The beauty of simplicity? 127 choice but not consequence space can be thought of as suboptimal behavior with a low opportunity cost. The Interaction between Simple Decision Rules and the Environment The ERP is based on the premise that rationality should be assessed in the context of the environment, that is, Simon’s ‘scissors’ metaphor. In strategic settings, the definition of the environment must be extended to include institutions, market characteristics and the interactions between agents. Perhaps surprisingly, to ERP researchers, an early example of such interactions was given by Becker (1962) who analyzed a model of markets in which participants behaved irrationally or randomly. He found that seemingly rational behavior at the macro level (not only in the consequence space, but also in the choice space) could arise even from random behavior at the micro level. In this spirit, more recent developments in economics include the zero-intelligence program initiated by Gode and Sunder (1993) who examined the effects of the structure of continuous double-auctions on market outcomes. They found that simple agents, who made random bids with the only constraint that they do not make offers that would lead to a loss, converged and achieved near perfect allocative efficiency. The lesson to be learned from this research is that rationality cannot be ascribed to individual decision makers without explicit consideration of the environment. Cognitive Bounds and Behavior The premise that less is more with respect to the amount of information that decision makers use can be linked to bounds on cognition such as limitations in the amount of information that can be held in working memory (Cowan 2000) or the long-term memory retrieval system (Schooler and Anderson 1997). Economists have similarly been concerned with simple strategies that do not use all available historical information, dating back to the Axelrod (1984) tournament. Tit-for-tat and the win–stay/lose–shift strategies are examples of relatively simple heuristics that perform well in repeated games and are robust to the exact composition of types in the population and to noise. Explicit modeling of forgetting has been common in economic studies of learning in repeated games since Roth and Erev (1995) and Cheung and Friedman (1997). Finite-state automata are another methodological tool explicitly aimed at examining the effects of limiting the prior (in a temporal sense) information that a player conditions his/her strategies on (for example, Rubinstein 1986). Furthermore, it is well known in game theory that more information does not necessarily lead to better outcomes. Procedural Modeling An important characteristic of most ERP studies is the insistence that models should be procedural (or process based) in contrast to the majority of models in economics that are as-if models. The advantage of procedural models is that they make more specific predictions (choices and processes) than as-if models and are more falsifiable in the Popperian sense. For example, see Johnson et al. (2008) who argue that the process data is incompatible with that implied by the priority heuristic; this, of course, would not have been pos-

M4225-ALTMAN_9781782549574_t.indd 127

03/05/2017 08:20

128

Handbook of behavioural economics and smart decision-making

sible for an as-if model. It is perhaps here that cognitive psychologists have already exerted a unidirectional influence on economists. Early work in psychology employing processtracing techniques such as Mouselab (Johnson et al. 1989) and eye-tracking have spilled over to economics; see Crawford (2008) for an excellent overview. Providing process-level foundations to existing as-if models in economics, and highlighting the value-added of this, is another way of engaging economists with the ERP. For example, Fischbacher et al. (2013) modify economic theories of social preferences by imposing a decision tree structure to the order in which these variables are examined. Similarly, Spiliopoulos (2013) transforms a process-free model of pattern recognition in games (Spiliopoulos 2012) into a process-model encompassing both exemplar- and prototype-based categorization grounded in the ACT-R architecture. Reasoning by Similarity and Cases Reasoning by similarity can be a useful tool when confronted with the uncertainty of a new situation of which an agent has not had experience. Important theoretical contributions have been made by economists to case-based and analogy-based reasoning; see, for example, early work by Rubinstein (1988) and Leland (1994) on decision under risk and the extensive work of Gilboa and Schmeidler (1995, 2001). Other work by economists exploiting similarity in inductive inference involves the question of how agents play a new game (that they have not seen before); specifically, how prior experience from other games may spill over to new (unseen) games on the basis of similarity between games (for example, Mengel and Sciubba 2014). Also, Spiliopoulos (2013) shows that subjects learn from the similarity, not between games, but between patterns in the history of play during a single repeated game.

OPEN QUESTIONS AND CHALLENGES While the success of the ERP cannot be disputed, there remain many open questions in need of answers. We enumerate and discuss them next. First, what is the complete set of heuristics out there? This question may be unanswerable for the simple reason that, as illustrated by Ericson et al. (2015), there are probably as many definitions as there are researchers. Also, researchers very often have vested interests to differentiate their product (for example, Bingham and Eisenhardt 2014, or the already mentioned Ericson et al. 2015, who do not reference Gigerenzer et al. 1999). In other words, there will not be agreement on what is in the adaptive toolbox of heuristics any time soon. An answer to this question will become even harder as heuristics – which so far have been studied predominantly in non-strategic decision settings – will be addressed in strategic decision settings; see Vuori and Vuori (2014) for an excellent primer. An alternative approach is to first ask what is the set of building blocks that make up heuristics? A broad, but by no means complete, characterization is that these are comprised of search rules, stopping rules and decision rules. Second, how to choose the appropriate tool from that adaptive toolbox remains a prominent question in search of better answers – see Marewski and Link (2013) for a review. ERP researchers have made considerable progress on this issue, generating inter-

M4225-ALTMAN_9781782549574_t.indd 128

03/05/2017 08:20

The beauty of simplicity? 129 esting results about strategy selection (for example, Marewski and Schooler 2011). A predictable argument has it that strategy selection is the result of evolutionary pressure or strategy selection using a reinforcement learning mechanism over heuristics (Rieskamp and Otto 2006). We find that argument only partially persuasive. Our skepticism goes back to old debates about to what extent people take into account structural changes in the environment. There is some evidence that people, possibly moderated by market institutions, have in many circumstances surprisingly rational expectations but, of course, it is dependent on many things even without market institutions moderating. We do know that the use of heuristics changes when environmental conditions change (for example, the work of Hogarth and Karelaia 2005, see also Rieskamp and Otto 2006, Spiliopoulos et al. 2015, Spiliopoulos and Ortmann 2015) but we are far from understanding the issue of matching in their totality in a satisfactory manner. Ultimately, the complexity of the environment will determine the tools in the box. Third, while it is an interesting question to understand how changing environments can affect choice of heuristics, to what extent the use of heuristics can shape the environment is a question that brings about important issues of causality (for example, Hertwig et al. 2002 on parental investment) that strike us as under-researched. Fourth, ERP researchers have recently argued that the two programs of rationality not only have very different assessments of human rationality but also have very different policy implications identified as nudging and boosting (Katsikopoulos 2014; GrueneYanoff and Hertwig 2016). These issues strike us also as under-studied. We are also not certain that the real issue is that of nudging versus boosting. We do appreciate the fact that nudging might have some undesirable intertemporal consequences (for example, Carroll et al. 2009 and the literature that followed it) but submit that boosting is often an unavailable option. Despite the difficulties, this opens up important avenues for the ERP to have a significant impact at the policy level. Fifth, the ERP, it seems fair to say, has not managed to have much practical impact on management science and organization science. This is surprising given the intellectual origin of the key parts of the ABC agenda (Simon, anyone?). Despite the fact that many publications on the theoretical properties of heuristics have made their way into prominent management/organization science journals (for example, Hogarth and Karelaia 2005; Katsikopoulos 2013), we are unaware of any significant impact on this literature on organizations directly, or applied/empirical work on heuristics in organizations. This is particularly surprising given that bounded rationality has become an influential concept in management science and organization science and economics. An exception is the hiatus heuristic that predicts whether a customer is active or not, that is, will make future purchases. Wübben and Wangenheim (2013) not only find evidence of its use by executives, but also show using real-world data that simple heuristics can out-predict more complex models. Sixth, as (simple) heuristics are being discovered by management and organization sciences (for example, Loock and Hinnen 2015), the movement away from non-strategic decision making (the core of early ER research) to strategic settings brings in new complexities arising from strategic interactions. It is not that ABC has not started to struggle with these issues but the work in this area seems pedestrian compared with the rather more sophisticated work on non-strategic decision making. Promising examples include the collaboration between economists and psychologists in Fischbacher et al. (2013) mentioned

M4225-ALTMAN_9781782549574_t.indd 129

03/05/2017 08:20

130

Handbook of behavioural economics and smart decision-making

earlier, which we hope to see more of in the future. Another example is Stevens et al. (2011), who examine the effects of forgetting on the emergence of cooperative strategies in repeated interactions. Further bridging the different concepts of bounded rationality that psychologists and economists would be a fruitful endeavor. There are important differences across disciplines that we cannot fully discuss here – Katsikopoulos (2014) and Grüne-Yanoff et al. (2014) are excellent primers. Seventh, the topic of learning has not been broached successfully by the ERP; however, the potential exists for important work on simple heuristics of learning. A starting point is Selten’s learning direction theory (LDT), which is ultimately a simple story of ex post rather than ex ante rationality using minimal information – note again that this is an inductive model of reasoning. For example, LDT requires information only about the direction that would have led to an improvement in the outcome; reinforcement learning would also require the magnitude and regret-based learning would require information about counterfactual outcomes. As an aside, we draw the reader’s attention to the edited volume by Gigerenzer and Selten (2002). An excellent example of work along these lines is Bonawitz et al. (2014) who show that a simple heuristic (win–stay, lose–sample) can approximate computationally demanding Bayesian inference in non-strategic settings. Strategic interactions entail additional uncertainty – how often is the assumption of perfect information fulfilled in the real world? Do we know what the action space is, what the payoffs are, and the type/motives of our opponent? With so much uncertainty is strategic ignorance or bounded sophistication necessarily irrational? Ecological-rationality program researchers should note that economists have not ignored these important questions, such as uncertainty, as the literature is literally full of extensions and concepts specifically addressing them. On the other hand, ERP researchers can and should critique the characteristics of the solutions proposed by economists. For example, in many cases the extensions or refinements to equilibrium solution concepts that deal with these types of uncertainty may be orders of magnitude more complicated than those under perfect information. Again, however, we note that these solutions belong to the deductive strand of game theory, not the inductive strand; the latter should be far more palatable to psychologists. Eighth, and relatedly, some celebrated heuristics can easily be exploited (for example, default settings in a situation where the choice architect has vested interests: credit card companies, and so on). In general, it is necessary to assert to what extent the interests of the default-setter and the people that default are meant to nudge coincide. It would be a mistake to assume that it is always the case. Ninth, Goldberg (2005; see also Goldberg and Podell 1999) have argued that studying lotteries does not capture decision making in the real world in reasonable ways. The real issue is what to do with other problems that cannot be represented by lotteries with two or three outcomes? Tenth, the fast and frugal heuristics literature, in its insistence on avoiding the calibration of heuristics to empirical data, has glossed over the issue of behavioral heterogeneity.

CONCLUDING DISCUSSION We set out to sketch established facts and open questions about simple heuristics, while also pointing out some areas of similar thinking with the economics discipline that could

M4225-ALTMAN_9781782549574_t.indd 130

03/05/2017 08:20

The beauty of simplicity? 131 serve as a bridge for future work. As it turns out there is an increasing number of authors that lay claim to the term ‘simple heuristics’ which seems to originate with Gigerenzer et al. (1999). While sketching the history and different premises of the two big programs in the heuristics space, the ‘Heuristics-and-Biases’ program and the ‘Ecological-Rationality’ program, we have focused on the latter and discussed its undoubtable accomplishments and prospects. Among its considerable accomplishments are the successful demonstration that, when cross-validation is used, the performance of simple heuristics such as the recognition heuristic or the ‘take-the-best’ tends to be better than that of complicated, computationally slow and greedy models such as multiple regression favored by economists (for example, Gigerenzer et al. 1999, Todd et al. 2012; Brighton and Gigerenzer 2009; Gigerenzer and Gaissmaier 2011). The simple, and rather intuitive, reason is that multiple regression is prone to over-fitting to the noise in the data-generating process by only looking backwards. Another important implication is that the widely believed effortaccuracy trade-off is often not something to worry about. It has also been demonstrated persuasively that an important moderator of these findings is the way statistical information is presented. There remain many open questions and interesting research topics which we have tried to enumerate. We have tried hard to draw attention to work in economics that seems closely related to the ERP and to highlight where common ground exists for the two disciplines to initiate a dialogue and collaborate despite their differences. The reader will notice that the majority of research that we have cited in economics is firmly grounded in inductive (learning from experience) rather than deductive models. We believe that much of the criticism of economics by ERP researchers has been directed at normative solutions involving deductive reasoning. This, however, is a straw man of sorts, and does not acknowledge the richness of contemporary economics. We further draw attention to the fact that many of the studies in economics that we have cited are published in mainstream, highly ranked journals such as the American Economic Review, Quarterly Journal of Economics, Econometrica, and Games and Economic Behavior. Therefore, we believe that sufficient interest exists for work that can be related to the ERP, and for the ERP to make significant headway into the economics discipline. This attempt will be most successful by connecting new research to prior work in economics and simultaneously pointing out the similarities and differences. Economists would also be well advised to seek out common ground with psychologists beyond the (now) orthodox heuristics-and-biases program.

NOTES *

The authors are grateful for critical and helpful commentary on earlier versions from Morris Altman, Nathan Berg, Gerd Gigerenzer, Ralph Hertwig, Konstantinos Katsikopoulos, Elizabeth Maitland (whose suggestion inspired the title of our chapter), and Ben Newell. All errors in judgment and tone are ours. 1. Non-compensatoriness of cues is satisfied if the weight of a higher ranked cue is greater than the sum of all lower ranked cues. Consequently, lower-ranked cues can be ignored as regardless of the cue values, it is impossible for them to reverse a decision made using the higher ranked cue. Dominance is satisfied if the cue values of one object are all greater than those of the other object. Cumulative dominance is satisfied if the cue values of one object cumulatively dominate those of the other object. Further discussions and mathematical definitions of these concepts can be found in Şimşek (2013).

M4225-ALTMAN_9781782549574_t.indd 131

03/05/2017 08:20

132

Handbook of behavioural economics and smart decision-making

FURTHER READING Brandstätter, E., G. Gigerenzer and R. Hertwig (2006), ‘The priority heuristic: making choices without tradeoffs’, Psychological Review, 113 (2), 409–32. Brighton, H. and G. Gigerenzer (2015), ‘The bias bias’, Journal of Business Research, 68 (8), 1772–84. Dhami, M.K., R. Hertwig and U. Hoffrage (2004), ‘The role of representative design in an ecological approach to cognition’, Psychological Bulletin, 130 (6), 959–88. Gigerenzer, G. and H. Brighton (2009), ‘Homo heuristicus: why biased minds make better inferences’, Topics in Cognitive Science, 1 (1), 107–43. Gigerenzer, G. and W. Gaissmaier (2011), ‘Heuristic decision making’, Annual Review of Psychology, 62 (1), 451–82. Goldberg, E. and K. Podell (1999), ‘Adaptive versus veridical decision making and the frontal lobes’, Consciousness and Cognition, 8 (3), 364–77. Goldstein, D.G. and G. Gigerenzer (2011), ‘The beauty of simple models: themes in recognition heuristic research’, Judgment and Decision Making, 6 (5), 392–95. Grüne-Yanoff, T. and R. Hertwig (2016), ‘Nudge versus boost: how coherent are policy and theory?’, Minds and Machines, 26 (1), 149–83. Grüne-Yanoff, T., C. Marchionni and I. Moscati (2014), ‘Introduction: methodologies of bounded rationality’, Journal of Economic Methodology, 21 (4), 325–42. Hertwig, R. and I. Erev (2009), ‘The description-experience gap in risky choice’, Trends in Cognitive Sciences, 13 (12), 517–23. Hogarth, R.M. and N. Karelaia (2007), ‘Heuristic and linear models of judgment: matching rules and environments’, Psychological Review, 114 (3), 733–58. Katsikopoulos, K.V. and G. Gigerenzer (2008), ‘One-reason decision-making: modeling violations of expected utility theory’, Journal of Risk and Uncertainty, 37 (1), 35–56. Katsikopoulos, K.V., L.J. Schooler and R. Hertwig (2010), ‘The robust beauty of ordinary information’, Psychological Review, 117 (4), 1259–66. Luan, S., L.J. Schooler and G. Gigerenzer (2014), ‘From perception to preference and on to inference: an approach–avoidance analysis of thresholds’, Psychological Review, 121 (3), 501–25. Mandler, M., P. Manzini and M. Mariotti (2012), ‘A million answers to twenty questions: choosing by checklist’, Journal of Economic Theory, 147 (1), 71–92. Manzini, P. and M. Mariotti (2014), ‘Stochastic choice and consideration sets’, Econometrica, 82 (3), 1153–76. Marewski, J.N. and L.J. Schooler (2011), ‘Cognitive niches: an ecological model of strategy selection’, Psychological Review, 118 (3), 393–437. Pleskac, T.J. and R. Hertwig (2014), ‘Ecologically rational choice and the structure of the environment’, Journal of Experimental Psychology: General, 143 (5), 2000–2019. Schooler, L.J. and R. Hertwig (2005), ‘How forgetting aids heuristic inference’, Psychological Review, 112 (3), 610–28. Volz, K.G. and G. Gigerenzer (2012), ‘Cognitive processes in decisions under risk are not the same as in decisions under uncertainty’, Frontiers in Neuroscience, 6 (July), 1–6.

REFERENCES Arkes, H.R. and P. Ayton (1999), ‘The sunk cost and Concorde effects: are humans less rational than lower animals?’, Psychological Bulletin, 125 (5), 591–600. Arthur, W.B. (1994), ‘Inductive reasoning and bounded rationality’, American Economic Review, 84 (2), 406–11. Artinger, F., M. Petersen, G. Gigerenzer and J. Weibler (2014), ‘Heuristics as adaptive decision strategies in management’, Journal of Organizational Behavior, 36 (S1), S33–S52. Åstebro, T. and S. Elhedhli (2006), ‘The effectiveness of simple decision heuristics: forecasting commercial success for early-stage ventures’, Management Science, 52 (3), 395–409. Aumann, R.J. (1997), ‘Rationality and bounded rationality’, Games and Economic Behavior, 21 (1–2), 2–14. Axelrod, R. (1984), The Evolution of Cooperation, New York: Basic Books. Barron, G. and I. Erev (2003), ‘Small feedback-based decisions and their limited correspondence to descriptionbased decisions’, Journal of Behavioral Decision Making, 16 (3), 215–33. Baucells, M., J.A. Carrasco and R.M. Hogarth (2008), ‘Cumulative dominance and heuristic performance in binary multiattribute choice’, Operations Research, 56 (5), 1289–304. Becker, G.S. (1962), ‘Irrational behavior and economic theory’, Journal of Political Economy, 70 (1), 1–13.

M4225-ALTMAN_9781782549574_t.indd 132

03/05/2017 08:20

The beauty of simplicity? 133 Berg, N. and G. Gigerenzer (2010), ‘As-if behavioral economics: neoclassical economics in disguise?’, History of Economic Ideas, 18 (1), 133–65. Bingham, C.B. and K.M. Eisenhardt (2014), ‘Response to Vuori and Vuori’s commentary on “Heuristics in the strategy context”’, Strategic Management Journal, 35 (11), 1698–1702. Binmore, K. (1990), Essays on the Foundations of Game Theory, Oxford: Blackwell. Bonawitz, E., S. Denison, A. Gopnik and T.L. Griffiths (2014), ‘Win-stay, lose-sample: a simple sequential algorithm for approximating Bayesian inference’, Cognitive Psychology, 74 (C), 35–65. Brandstätter, E., G. Gigerenzer and R. Hertwig (2006), ‘The priority heuristic: making choices without tradeoffs’, Psychological Review, 113 (2), 409–32. Brandstätter, E., G. Gigerenzer and R. Hertwig (2008), ‘Risky choice with heuristics: reply to Birnbaum (2008), Johnson, Schulte-Mecklenbeck, and Willemsen (2008), and Rieger and Wang (2008)’, Psychological Review, 115 (1), 281–90. Brighton, H. and G. Gigerenzer (2015), ‘The bias bias’, Journal of Business Research, 68 (8), 1772–84. Carroll, G.D., J.J. Choi, D. Laibson, B.C. Madrian and A. Metrick (2009), ‘Optimal defaults and active decisions’, Quarterly Journal of Economics, 124 (4), 1639–74. Camerer, C.F. and T.-H. Ho (1999), ‘Experience-weighted attraction learning in normal form games’, Econometrica, 67 (4), 827–74. Camerer, C.F., G. Loewenstein and M. Rabin (eds) (2004), Advances in Behavioral Economics, New York: Princeton University Press. Cheung, Y.-W. and Daniel Friedman (1997), ‘Individual learning in normal form games: some laboratory results’, Games and Economic Behavior, 19 (1), 46–76. Cochrane, J. (2015), ‘Homo economicus or homo paleas’, The Grumpy Economist (John Cochrane’s blog), accessed 22 May 2015 at http://johnhcochrane.blogspot.com.au/2015/05/homo-economicus-or-homo-paleas. html. Cowan, N. (2000), ‘The magical number 4 in short-term memory: a reconsideration of mental storage capacity’, Behavioral and Brain Sciences, 24 (1), 87–114. Crawford, V. (2008), ‘Look-ups as the windows of the strategic soul’, in A. Caplin and A. Schotter (eds), The Foundations of Positive and Normative Economics, New York: Oxford University Press. Dhami, M.K., R. Hertwig and U. Hoffrage (2004), ‘The role of representative design in an ecological approach to cognition’, Psychological Bulletin, 130 (6), 959–88. Drechsler, M., K. Katsikopoulos and G. Gigerenzer (2014), ‘Axiomatizing bounded rationality: the priority heuristic’, Theory and Decision, 77 (2), 183–96. Edwards, W. (1956), ‘Reward probability, amount, and information as determiners of sequential two-alternative decisions’, Journal of Experimental Psychology, 52 (3), 177–88. Erev, I. and G. Barron (2005), ‘On adaptation, maximization, and reinforcement learning among cognitive strategies’, Psychological Review, 112 (4), 912–31. Erev, I., E. Ert, O. Plonsky, D. Cohen and O. Cohen (2017), ‘From anomalies to forecasts: toward a descriptive model of decisions under risk, under ambiguity, and from experience’, Psychological Review (in press), retrieved from http://departments.agri.huji.ac.il/economics/teachers/ert_eyal/CompDEPsychRev2016.12.22. pdf on February 6, 2017. Ericson, K., M. Marzilli, J.M. White, D. Laibson and J.D. Cohen (2015), ‘Money earlier or later? Simple heuristics explain intertemporal choices better than delay discounting does’, Psychological Science, 26 (6), 1–8. Fischbacher, U., R. Hertwig and A. Bruhin (2013), ‘How to model heterogeneity in costly punishment: insights from responders’ response times’, Journal of Behavioral Decision Making, 26 (5), 462–76. Friedman, D. (1991), ‘Evolutionary games in economics’, Econometrica, 59 (3), 637–66. Friedman, D., K. Pommerenke, R. Lukose, G. Milam and B.A. Huberman (2007), ‘Searching for the sunk cost fallacy’, Experimental Economics, 10 (1), 79–104. Friedman, M. (1953), Essays in Positive Economics, Chicago, IL: University of Chicago Press. Friedman, M. and L.J. Savage (1948), ‘The utility analysis of choices involving risk’, Journal of Political Economy, 56 (4), 279–304. Gigerenzer, G. (1991), ‘How to make cognitive illusions disappear: beyond “heuristics and biases”’, European Review of Social Psychology, 2 (1), 83–115. Gigerenzer, G. (1996)), ‘On narrow norms and vague heuristics: a reply to Kahneman and Tversky (1996)’, Psychological Review, 103 (3) 592–6. Gigerenzer, G. and H. Brighton (2009), ‘Homo heuristicus: why biased minds make better inferences’, Topics in Cognitive Science, 1 (1), 107–43. Gigerenzer, G. and W. Gaissmaier (2011), ‘Heuristic decision making’, Annual Review of Psychology, 62 (1), 451–82. Gigerenzer, G. and R. Selten (eds) (2002), Bounded Rationality: The Adaptive Toolbox, Cambridge, MA: MIT Press.

M4225-ALTMAN_9781782549574_t.indd 133

03/05/2017 08:20

134

Handbook of behavioural economics and smart decision-making

Gigerenzer, G., R. Hertwig and T. Pachur (2011), Heuristics: The Foundations of Adaptive Behavior, New York: Oxford University Press. Gigerenzer, G., P.M. Todd and the ABC Research Group (1999), Simple Heuristics That Make Us Smart, New York: Oxford University Press. Gilboa, I. and D. Schmeidler (1995), ‘Case-based decision theory’, Quarterly Journal of Economics, 110 (3), 605–39. Gilboa, I. and D. Schmeidler (2001), A Theory of Case-Based Decisions, Cambridge: Cambridge University Press. Gode, D.K. and S. Sunder (1993), ‘Allocative efficiency of markets with zero-intelligence traders – market as a partial substitute for individual rationality’, Journal of Political Economy, 101 (1), 119–37. Goldberg, G. (2005), The Wisdom Paradox: How Your Mind Can Grow Stronger as Your Brain Grows Older, New York: Penguin. Goldberg, E. and K. Podell (1999), ‘Adaptive versus veridical decision making and the frontal lobes’, Consciousness and Cognition, 8 (3), 364–77. Goldstein, D.G. and G. Gigerenzer (2011), ‘The beauty of simple models: themes in recognition heuristic research’, Judgment and Decision Making, 6 (5), 392–95. Grüne-Yanoff, T. and R. Hertwig (2016), ‘Nudge versus boost: how coherent are policy and theory?’, Minds and Machines, 26 (1), 149–83. Grüne-Yanoff, T., C. Marchionni and I. Moscati (2014), ‘Introduction: methodologies of bounded rationality’, Journal of Economic Methodology, 21 (4), 325–42. Harrison, G.W. (1989), ‘Theory and misbehavior of first-price auctions’, American Economic Review, 9 (4), 749–62. Haruvy, E. and D.O. Stahl (2012), ‘Between-game rule learning in dissimilar symmetric normal-form games’, Games and Economic Behavior, 74 (1) 208–21. Hertwig, R. and I. Erev (2009), ‘The description-experience gap in risky choice’, Trends in Cognitive Sciences, 13 (12), 517–23. Hertwig, R. and A. Ortmann (2004), ‘The cognitive illusions controversy: a methodological debate in disguise that matters to economists’, in R. Zwick and A. Rapoport (eds), Experimental Business Research III, Boston, MA: Kluwer, pp. 113–30. Hertwig, R., J.N. Davis and F.J. Sulloway (2002), ‘Parental investment: how an equity motive can produce inequality’, Psychological Bulletin, 128 (5), 728–45. Hertwig, R., S.M. Herzog, L.J. Schooler and T. Reimer (2008), ‘Fluency heuristic: a model of how the mind exploits a by-product of information retrieval’, Journal of Experimental Psychology: Learning, Memory, and Cognition, 34 (5), 1191–206. Hertwig, R., U. Hoffrage and the ABC Research Group (2013), Simple Heuristics in a Social World, New York: Oxford University Press. Heukelom, F. (2015), Behavioral Economics. A History, Cambridge: Cambridge University Press. Hogarth, R.M. and N. Karelaia (2005), ‘Simple models for multiattribute choice with many alternatives: when  it  does and does not pay to face trade-offs with binary attributes’, Management Science, 51 (12), 1860–72. Hogarth, R.M. and N. Karelaia (2006), ‘Regions of rationality: maps for bounded agents’, Decision Analysis, 3 (3), 124–44. Hogarth, R.M and N. Karelaia (2007), ‘Heuristic and linear models of judgment: matching rules and environments’, Psychological Review, 114 (3), 733–58. Hogarth, R.M. and M.W. Reder (eds) (1987), Rational Choice, Chicago, IL: University of Chicago Press. Hutchinson, J.M.C. and G. Gigerenzer (2005), ‘Simple heuristics and rules of thumb: where psychologists and behavioural biologists might meet’, Behavioural Processes 69 (2), 97–124. Johnson, E.J., J.W. Payne, D.A. Schkade and J.R. Bettman (1989), ‘Monitoring information processing and decisions: the Mouselab system’, unpublished manuscript, Fuqua School of Business, Duke University, Durham, NC. Johnson, E.J., M. Schulte-Mecklenbeck and M.C. Willemsen (2008), ‘Process models deserve process data: comment on Brandstätter, Gigerenzer, and Hertwig (2006)’, Psychological Review, 115 (1), 263–72. Kahneman, D. (2003a), ‘Maps of bounded rationality: psychology for behavioral economics’, American Economic Review, 93 (5), 1449–75. Kahneman, D. (2003b), ‘A perspective on judgment and choice: mapping bounded rationality’, American Psychologist, 58 (9), 697–720. Kahneman, D. (2011), Thinking, Fast and Slow, New York: Penguin. Kahneman, D. and A. Tversky (1979), ‘Prospect theory: an analysis of decision under risk’, Econometrica, 47 (2), 263–91. Kahneman, D. and A. Tversky (1996), ‘On the reality of cognitive illusions’, Psychological Review, 103 (3), 582–91.

M4225-ALTMAN_9781782549574_t.indd 134

03/05/2017 08:20

The beauty of simplicity? 135 Katsikopoulos, K.V. (2013), ‘Why do simple heuristics perform well in choices with binary attributes?’, Decision Analysis, 10 (4), 327–40. Katsikopoulos, K.V. (2014), ‘Bounded rationality: the two cultures’, Journal of Economic Methodology, 21 (4), 361–74. Katsikopoulos, K.V. and G. Gigerenzer (2008), ‘One-reason decision-making: modeling violations of expected utility theory’, Journal of Risk and Uncertainty, 37 (1), 35–56. Katsikopoulos, K.V., L.J. Schooler and R. Hertwig (2010), ‘The robust beauty of ordinary information’, Psychological Review, 117 (4), 1259–66. Leland, J.W. (1994), ‘Generalized similarity judgments – an alternative explanation for choice anomalies’, Journal of Risk and Uncertainty, 9 (2), 151–72. Loock, M. and G. Hinnen (2015), ‘Heuristics in organizations: a review and a research agenda’, Journal of Business Research, 68 (9), 2027–36. Lopes, L.L. (1992), ‘The rhetoric of irrationality’, Theory and Psychology, 1 (1), 65–82. Luan, S., L.J. Schooler and G. Gigerenzer (2011), ‘A signal-detection analysis of fast-and-frugal trees’, Psychological Review, 118 (2), 316–38. Luan, S., L.J. Schooler and G. Gigerenzer (2014), ‘From perception to preference and on to inference: an approach–avoidance analysis of thresholds’, Psychological Review, 121 (3), 501–25. Maitland, E. and A. Sammartino (2015), ‘Decision making and uncertainty: the role of heuristics and experience in assessing a politically hazardous environment’, Strategic Management Journal, 36 (10), 1554–78. Mandler, M., P. Manzini and M. Mariotti (2012), ‘A million answers to twenty questions: choosing by checklist’, Journal of Economic Theory, 147 (1), 71–92. Manzini, P. and M. Mariotti (2007), ‘Sequentially rationalizable choice’, American Economic Review, 97 (5), 1824–39. Manzini, P. and M. Mariotti (2012a), ‘Categorize then choose: boundedly rational choice and welfare’, Journal of the European Economic Association, 10 (5), 1141–65. Manzini, P. and M. Mariotti (2012b), ‘Choice by lexicographic semiorders’, Theoretical Economics, 7 (1), 1–23. Manzini, P. and M. Mariotti (2014), ‘Stochastic choice and consideration sets’, Econometrica, 82 (3), 1153–76. Marewski, J.N. and D. Link (2013), ‘Strategy selection: an introduction to the modeling challenge’, Wiley Interdisciplinary Reviews: Cognitive Science, 5 (1), 39–59. Marewski, J.N. and L.J. Schooler (2011), ‘Cognitive niches: an ecological model of strategy selection’, Psychological Review, 118 (3), 393–437. McAffee, R.P., H.M. Mialon and S.H. Mialon (2010), ‘Do sunk costs matter?’, Economic Inquiry, 48 (2), 323–36. Mengel, F. and E. Sciubba (2014), ‘Extrapolation and structural similarity in games’, Economics Letters, 125 (3), 381–5. Ortmann, A. (2015a), ‘Review of Floris Heukelom (2014), Behavioral Economics, A History’, Œconomia, 5-2, 259–67. Ortmann, A. (2015b), ‘Review of World Development Report 2015’, Journal of Economic Psychology, 48 (June), 111–20. Payne, J.W., J.R. Bettman and E.J. Johnson (1993), The Adaptive Decision Maker, Cambridge: Cambridge University Press. Peterson, C.R. and L.R. Beach (1967), ‘Man as an intuitive statistician’, Psychological Bulletin, 68 (1), 29–46. Pleskac, T.J. and R. Hertwig (2014), ‘Ecologically rational choice and the structure of the environment’, Journal of Experimental Psychology: General, 143 (5), 2000–2019. Rieskamp, J. and P.E. Otto (2006), ‘SSL: a theory of how people learn to select strategies’, Journal of Experimental Psychology: General, 135 (2), 207–36. Roth, A.E. and I. Erev (1995), ‘Learning in extensive-form games: experimental data and simple dynamic models in the intermediate term’, Games and Economic Behavior, 8 (1), 164–212. Rubinstein, A. (1986), ‘Finite automata play the repeated prisoner’s dilemma’, Journal of Economic Theory, 39 (1), 83–96. Rubinstein, A. (1988), ‘Similarity and decision-making under risk (is there a utility-theory resolution to the Allais paradox?)’, Journal of Economic Theory, 46 (1), 145–53. Savage, L.J. (1954), The Foundation of Statistics, New York: John Wiley and Sons. Schooler, L.J. and J.R. Anderson (1997), ‘The role of process in the rational analysis of memory’, Cognitive Psychology, 32 (3), 219–50. Schooler, L.J. and R. Hertwig (2005), ‘How forgetting aids heuristic inference’, Psychological Review, 112 (3), 610–28. Sedlmeier, P. and G. Gigerenzer (2001), ‘Teaching Bayesian reasoning in less than two hours’, Journal of Experimental Psychology: General, 130 (3), 380–400. Sent, E.-M. (2004), ‘Behavioral economics: how psychology made its (limited) way back into economics’, History of Political Economy, 36 (4), 735–60. Simon, H.A. (1947), Administrative Behavior, New York: Macmillan.

M4225-ALTMAN_9781782549574_t.indd 135

03/05/2017 08:20

136

Handbook of behavioural economics and smart decision-making

Simon, H.A. (1955), ‘A behavioral model of rational choice’, Quarterly Journal of Economics, 69 (1), 99–118. Simon, H.A. (1956), ‘Rational choice and the structure of the environment’, Psychological Review, 63 (2), 129–38. Şimşek, Ö. (2013), ‘Linear decision rule as aspiration for simple decision heuristics’, Advances in Neural Information Processing Systems, 26, 2904–12. Smith, V.L. (1991), ‘Rational choice: the contrast between economics and psychology’, Journal of Political Economy, 99 (4), 877–97. Smith, V.L. (2003), ‘Constructivist and ecological rationality in economics’, American Economic Review, 93 (3), 465–508. Spiliopoulos, L. (2012), ‘Pattern recognition and subjective belief learning in a repeated constant-sum game’, Games and Economic Behavior, 75 (2), 921–35. Spiliopoulos, L. (2013), ‘Beyond fictitious play beliefs: incorporating pattern recognition and similarity matching’, Games and Economic Behavior, 81 (September), 69–85. Spiliopoulos, L. and A. Ortmann (2014), ‘Model comparisons using tournaments: likes, “dislikes,” and challenges’, Psychological Methods, 19 (2), 230–50. Spiliopoulos, L. and A. Ortmann (2015), ‘The BCD of response time analysis in experimental economics’, doi:org/10.2139/ssrn.2401325. Spiliopoulos, L., A. Ortmann and L. Zhang (2015), ‘Attention and choice in games under time constraints: a process analysis’, 11 October, doi:org/10.2139/ssrn.2620163. Stahl, D.O. (1996), ‘Boundedly rational rule learning in a guessing game’, Games and Economic Behavior, 16 (2) 303–30. Stahl, D.O. (1999), ‘Evidence based rules and learning in symmetric normal-form games’, International Journal of Game Theory, 28 (1), 111–30. Stahl, D.O. (2000), ‘Rule learning in symmetric normal-form games: theory and evidence’, Games and Economic Behavior, 32 (1), 105–38. Stevens, J.R., J. Volstorf, L.J. Schooler and J. Rieskamp (2011), ‘Forgetting constrains the emergence of cooperative decision strategies’, Frontiers in Psychology, 1, art. 235, 1–12. Thaler, R. (1980), ‘Toward a positive theory of consumer choice’, Journal of Economic Behavior & Organization, 1 (1), 39–60. Thaler, R. (1985), ‘Mental accounting and consumer choice’, Marketing Science, 4 (3), 199–214. Thaler, R. (2015), Misbehaving: The Making of Behavioral Economics, New York and London: W.W. Norton. Todd, P.M., G. Gigerenzer and the ABC Research Group (2012), Ecological Rationality: Intelligence in the World, New York: Oxford University Press. Tversky, A. (1969), ‘Intransitivity of preferences’, Psychological Review, 76 (1), 31–48. Tversky, A. and D. Kahneman (1974), ‘Judgment under uncertainty: heuristics and biases’, Science, 185 (4157), 1124–31. Volz, K.G. and G. Gigerenzer (2012), ‘Cognitive processes in decisions under risk are not the same as in decisions under uncertainty’, Frontiers in Neuroscience, 6 (July), 1–6. Vuori, N. and T. Vuori (2014), ‘Comment on “Heuristics in the strategy context” by Bingham and Eisenhardt (2011)’, Strategic Management Journal, 35 (11), 1689–97. Wilcox, N.T. (2011), ‘Stochastically more risk averse: a contextual theory of stochastic discrete choice under risk’, Journal of Econometrics, 162 (1), 89–104. Wübben, M. and F. von Wangenheim (2013), ‘Instant customer base analysis: managerial heuristics often “get it right”’, Journal of Marketing, 72 (3), 82–93.

M4225-ALTMAN_9781782549574_t.indd 136

03/05/2017 08:20

8

Smart persons and human development: the missing ingredient in behavioral economics* John F. Tomer

INTRODUCTION There’s a growing sense among economists, especially behavioral economists, that the human actor in economics is not portrayed well by the economic man stereotype nor by the irrational, error-plagued person who is the stereotype deriving from psychological economics. The purpose of this chapter is, first, to explain about the inadequacy of these two stereotypical economic actors and, second, to develop an alternative, a more satisfactory stereotype known as the smart person. In the process, this chapter points the way to a better behavioral economics, a behavioral economics with smart people, a behavioral economics that is more realistic and more human. What is missing from the existing stereotypical actors, but present in the smart person actor, is the human who develops in stages along a number of developmental pathways over a lifetime. In contrast to the two existing stereotypes, the smart person’s character and capabilities are neither simply assumed nor inferred from the outcomes of narrow psychological laboratory experiments. The smart person’s character and behavior derive in good measure from the research of a variety of non-economist scientists and careful observers of human behavior. There is a tremendous need for a behavioral economics with smart persons in which the human actor, while far from perfect, develops, and all too often fails to develop, character and capabilities in a realistic way. The plan of the chapter is as follows. The next section explains what is missing from mainstream economics and psychological economics. The missing ingredient is the concept of human development. The third section carefully considers the characteristics of economic man, the human in mainstream economics. The following section carefully considers the character and capabilities of the human in psychological economics, particularly his or her lack of economic rationality. The fifth section develops a conception of an alternative human economic actor, an actor whose character and capabilities are much closer to the humans we know. This section explains how a behavioral economics with smart people has the potential to be a great improvement over the psychological economics version of behavioral economics with its error-prone stereotype.

THE MISSING INGREDIENT: THE CONCEPT OF HUMAN DEVELOPMENT The ingredient missing from economics is the conception of a human being as an individual who develops in many different ways along a sequence of stages, a maturational path. As wise thinkers through the ages have recognized, humans are capable of 137

M4225-ALTMAN_9781782549574_t.indd 137

03/05/2017 08:20

138

Handbook of behavioural economics and smart decision-making

attaining a very high level of development, involving a full flourishing of all their human capabilities in the broadest and highest sense over their entire life cycle. Clearly, the high human development (HD) envisioned by these thinkers involves much more than the acquisition of cognitive capability or workplace skill. High HD certainly involves social, psychological, emotional, and biological dimensions, among others, but the ideal or potential HD often fails to occur. Generally, only when the environment is favorable do humans have a chance of developing a high degree of their potential. Therefore, a key question is, what has to happen for individuals to develop to, or near to, their full potential? What kind of environment is necessary for favorable development? Among the necessary environmental conditions commonly recognized as necessary for reasonably high HD are a good education and the kind of early life nurturing usually provided by two loving parents. A supportive community and society are also important. For many, of course, the environment may not be favorable in some important respects, and as a consequence individuals may fail to negotiate significant stages of development. Thus, an individual may get stuck or partially stuck at a certain developmental stage and may fail to develop further without special developmental interventions. Without such help, it is likely that the individual will remain stuck at a level of HD that does not allow the full development of their talents. Conventional economic thinking provides little or no recognition of how individuals can advance along important developmental pathways and how they can overcome the types of difficulties that would otherwise prevent or inhibit their development. The concept of HD used here draws from a number of different traditions. First, it incorporates the perspective of developmental scientists whose field of study broadly encompasses HD in physical or biological, cognitive, and psychosocial domains or behaviors (see, for example, two HD texts: Kail and Cavanaugh 2007; Papalia et al. 2009). Second, the HD concept is inspired by the humanistic psychological perspective of Abraham Maslow (1943), notably his hierarchy of needs. Third, it is informed by research on neurodevelopment (see, for example, Perry 2002), particularly Perry’s work related to the developmental difficulties occurring in early childhood. Fourth, the HD conception here has been influenced by Ken Wilber’s (see, for example, Wilbur 2001, pp. 5–16) conception of how humans develop in an unfolding series of stages and levels from lower order to higher order along many dimensions or lines. The HD concept used here is related to, but distinctly different from, the HD concept pioneered by Amartya Sen, Martha Nussbaum and others. The latter concept which has been much used by international agencies (for example, the World Bank and the United Nations) concerned with economic development emphasizes a great number and variety of human functioning and capabilities. The Sen or Nussbaum HD concept is very useful for thinking about national and world economic development and how its progress can be measured. A good overview of this concept and its uses can be found in Alkire (2010). What this concept lacks is a conception of the stages of development in a human’s life and how human capacities and orientations change in predictable ways and sometimes fail to change. That is, there is no conception of the multidimensional developmental process that humans experience and the challenges a human typically encounters along the developmental pathways. To better understand the HD concept, it is important to illustrate graphically its main pathways and the sequence of development along each. For the purposes of this chapter,

M4225-ALTMAN_9781782549574_t.indd 138

03/05/2017 08:20

Smart persons and human development 139

Acquiring overall life direction, interests, outlooks, & motivation

4

3

Developing skills & talents: physical, academic, arts, technology Learning/appreciating many types of knowledge and acquiring academic discipline

2 Learning the basics: reading, writing, arithmetic 1 Figure 8.1

Educational and cognitive development

HD is represented as a three-sided pyramid. Each side represents a major developmental pathway. The three developmental pathways are (1) educational and cognitive development, (2) psychosocial, biological development, and (3) brain development (or neurodevelopment). In each case, the triangles representing the pathways start from very fundamental, early development and proceed stepwise to the highest level of development. The sequence of steps resembles in some respects Maslow’s (1943) hierarchy of needs in that, with some exceptions, earlier stages must precede later stages. Also, note that there is considerable interdependence among the three pathways. For economists, and presumably many academics, the easiest triangle or pathway to appreciate is the educational and cognitive development pathway. The side of the pyramid representing this pathway is shown in Figure 8.1. It starts at the bottom with ‘Learning the basics: reading, writing, arithmetic’. The second step is ‘Learning/appreciating many types of knowledge and acquiring academic discipline’. The third step is ‘Developing skills and talents: physical, academic, arts, technology’. The fourth and final step is ‘Acquiring overall life direction, interests, outlooks, and motivation’. The second pathway, psychosocial, biological development, is shown as the triangle in Figure 8.2. It starts with ‘Foundational neurodevelopment’ and proceeds to ‘Early learning, relating, and doing’ and then to ‘Becoming safe, secure, and satisfying physical needs’. The fourth step is ‘Finding oneself: competencies, motivations, values, and emotional intelligence’. The fifth step is ‘Finding oneself: friends, lovers, and loving family relations’. The sixth and final step is ‘Connecting to one’s highest values, spirituality, creativity, and aesthetics’.

M4225-ALTMAN_9781782549574_t.indd 139

03/05/2017 08:20

140

Handbook of behavioural economics and smart decision-making

Connecting to one’s highest values, spirituality, creativity, aesthetics

6

5

4

3 2 1 Figure 8.2

Finding oneself: friends, lovers, loving family relations

Finding oneself: competencies, motivations, values, emotional intelligence

Becoming safe, secure, and satisfying physical needs Early learning, relating, doing Foundational neurodevelopment

Psychosocial and biological development

The third pathway, brain development, is shown as the triangle in Figure 8.3. It starts with ‘Foundational neurodevelopment’ and proceeds to ‘Neurodevelopment associated with doing, achieving, relating, and learning’. The third step is ‘Overcoming brain development deficiencies and problems’. The fourth and final step is ‘Developing creativity and peak performance brain functioning’. Figure 8.4 shows how the three triangles described above combine to form the HD pyramid. No doubt a much more careful and micro elaboration of the pathways by a developmentally oriented behavioral scientist would include many more steps in each pathway than the number included here. The benefit of using the HD pyramid is that it focuses attention on three main ways that important human capabilities change, have the potential to change, or fail to realize their change potential. In Wilber’s (2001, pp. 5–6) view, human development involves an unfolding, emergent process marked by progressive subordination of older, lower-order behavior and capabilities to new higher-order behavior and capabilities along different pathways or lines. Using the HD pyramid helps us understand how change along one pathway may facilitate change along another pathway and how barriers to change in a pathway may result in lack of desired change along another pathway. It is important to note that society has a strong effect on an individual’s development. The society’s ethics, norms, rules, and basic institutions are integrated and have a cohesion that affects how far individuals develop (Wilber 1996, pp. 138–41). The society’s ‘cultural center of gravity acts like a magnet on individual development. If you are below the average level, it tends to pull you up. If you try to go above it, it tends to pull you down’

M4225-ALTMAN_9781782549574_t.indd 140

03/05/2017 08:20

Smart persons and human development 141

Developing creativity and peak performance brain functioning

4

3

Overcoming brain development deficiencies, problems Neurodevelopment associated with doing, achieving relating, learning

2 Foundational neurodevelopment 1 Figure 8.3

Brain development

Figure 8.3

Figure 8.2

Figure 8.1

Note that the three pathways are interdependent

Figure 8.4

Human development pyramid

(Wilber 1996, p. 139). The society’s developmental magnet helps you reach the expected level of development, but likely retards your earnest attempts to develop beyond the societal norm. As a consequence, relatively few people reach the highest developmental stages but many reach average levels.1

M4225-ALTMAN_9781782549574_t.indd 141

03/05/2017 08:20

142

Handbook of behavioural economics and smart decision-making

ECONOMIC MAN: THE HUMAN IN MAINSTREAM ECONOMICS Economic man, or Homo economicus, is the well-known human economic actor in mainstream economics. Because economic man’s behavior reflects the rational choice theory at the heart of mainstream economics, his or her behavior is machine-like in its perfect rationality (see Simon 1983, pp. 12–17). Economic man chooses in a narrowly self-interested way, using perfect logic and a complete knowledge of alternatives, and thus, selects the alternative that best enables attainment of his or her subjectively defined ends. If economic man is a consumer, the end is utility; if a producer, the end is profit. If economic man were human, we could say that he or she possesses infinite, or at least extremely high, cognitive capacity. In contrast, it seems to be implied that economic man has no or low non-cognitive capacity, that is, capacity relating to psychological, emotional, and social functioning. This further implies that economic man has zero capacity for pure empathetic (or other interest) motivation, the motivation opposite to self-interest. Economic man is also unreflective in the sense that he or she cannot stop to consider the appropriateness or rightness of his or her choices. It is fairly obvious to many, including a number of leading economic thinkers such as John Stuart Mill (1836), that economics does not consider the whole of man’s nature. Accordingly, economic man is a one-dimensional being who merely compares alternative ways to achieve his or her economic ends. The economic man concept continues to be widely used in economic modeling and analysis despite the fact that there are many economists who understand (1) that humans do not generally know the consequence of their actions for their long-term physical and mental health, and (2) that humans cannot be relied on to make decisions in their strict self and selfish interest. In their models, economists often use an economic actor who is a representative agent, a typical decision maker of a certain type. In mainstream economics, such agents are economic men who perform a particular role; they are, for example, consumers or decision makers in a firm. These agents presumably have made significant investments in human capital in order for them to carry out their economic role. Regardless of their human capital, the agents in these models behave in a perfectly rational manner, albeit in a particular context. Economic man’s behavior is in accord with the formal model of rational choice known as subjective expected utility (SEU) theory in which economic man chooses the alternative that maximizes his or her expected value of utility. As Herbert Simon (1983, p. 13) points out, SEU ‘is a beautiful object deserving a prominent place in Plato’s heaven of ideas’. Unfortunately, according to Simon, the ‘SEU theory has never been applied and can never be applied . . . in the real world’ (1983, p. 14). This is because ‘human beings have neither the facts nor the consistent structure of values nor the reasoning power at their disposal that would be required, even in . . . relatively simple [lab] situations to apply SEU principles’ (Simon 1983, p. 17). That is, the economic man concept is simplistic and unrealistic. Consider economic man from a developmental perspective. Economic man is unchanging; he or she has no history and no future. That is, the qualities possessed by economic man did not come about through a process of human development, and there is no prospect of future development that will cause these qualities to change. Economic man’s character is simply assumed; it is not an object of theoretical or empirical study. If

M4225-ALTMAN_9781782549574_t.indd 142

03/05/2017 08:20

Smart persons and human development 143 economic man’s character (perfect rationality) had come about through a developmental process, we might say that economic man had reached an egoistic stage of development in which he or she is aware of having many wants and is perfectly logical and persistent in the endeavor to satisfy those wants. Such a hypothetical development stage resembles Wilber’s (2001, p. 9) third stage of human development in which the individual is powerful, egocentric, and has a self that is distinct from his or her ‘tribe’. According to Wilber’s (2001, pp. 9–11) estimates, the great majority of people in the twenty-first century world develop beyond this egocentric stage during the course of their lives.

PSYCHOLOGICAL ECONOMIC MAN: THE HUMAN IN PSYCHOLOGICAL ECONOMICS Psychological economics is the prominent strand of behavioral economics that borrows from psychology, especially cognitive psychology, in order to achieve a more realistic understanding of human economic behavior than is possible with mainstream economics. Psychological economic man is the human in psychological economics (PE). Psychological economic man’s character, in sharp contrast to economic man, is very much an object of study, especially empirical study. Psychological economics is oriented to investigating human cognitive performance in relatively narrow and well-defined situations in order to isolate humans’ precise decision making and judgment behavior. Psychological economics researchers have focused to a large extent on exploring the degree to which human behavior systematically departs from economic rationality, that is, the extent to which psychological economic man is different from economic man. Overall, the findings of PE research are that humans are much less rational than mainstream economics assumes. That is, we humans are systematically and predictably irrational in all phases of our lives; we make many different kinds of errors in a great variety of particular situations (Ariely 2009, pp. 239–40). These errors derive from, among other things, the anchoring effect, judgment by representativeness, overconfidence, theoryinduced blindness, loss aversion, salience, use of mental accounts, framing, inconsistent preferences, defective affective forecasting, difficulties dealing with probabilities and time, the narrative fallacy, hindsight bias, confirmation bias, overestimating rare events, status quo bias, planning fallacy, and the availability and affect heuristics (Kahneman 2011). In light of these findings, it is not surprising that the psychological economic man stereotype is very much one of an irrational and error-prone being. In comparison to economic man, psychological economic man is decidedly not smart. This characterization of psychological economic man’s judgment and decision making is more realistic than that of the economic man stereotype precisely because it is based on a great amount of research. An important aspect of PE involves understanding two systems in the mind, system 1 and system 2. System 1, associated with intuition, is the aspect of our mind that ‘operates automatically and quickly, with little or no effort and no sense of voluntary control’ (Kahneman 2011, p. 20). Many of the predictable human errors which PE focuses on occur when our minds are in system 1 mode. If, in the face of a difficult question or issue, no easy system 1 solution comes to mind, that is when we typically switch to system 2. System 2 refers to effortful mental activities requiring concentration and self-control

M4225-ALTMAN_9781782549574_t.indd 143

03/05/2017 08:20

144

Handbook of behavioural economics and smart decision-making

(Kahneman 2011, p. 22). System 2 is slower; it may involve computation, deliberation, and constructing thoughts in an orderly series of steps. Psychological economics’ emphasis is less on explaining the reasons why humans commit cognitive errors and more on accurately characterizing humans’ cognitive performance. Nevertheless, a number of the leading researchers have offered explanations for humans’ error-proneness. According to Ariely (2009, p. 243), our senses and brain filter the information that comes to us so that the input to our decision making is not a fully accurate reflection of the reality of the situations we confront. In other words, the problem stems from ‘the basic wiring of our brains’ (Ariely 2009, p. 239). In Kahneman’s (2011, p. 51) view, errors of judgment and decision making often stem from ‘a self-reinforcing pattern of cognitive, emotional, and physical responses that [are] . . . associatively coherent’ (original emphasis). The errors often arise because our perception and cognition involve our body, not just our brain. Psychological economic man can learn and acquire the skill necessary to reduce the errors that typically occur when humans are operating in system 1 mode. When these error-causing difficulties are recognized, humans may switch to system 2 mode and may try harder in order to avoid significant mistakes, especially when the stakes are high (Kahneman 2011, pp. 25–8). To acquire these error reducing, decision-making skills ‘requires a regular environment, an adequate opportunity to practice, and rapid and unequivocal feedback’ on decision-making results (Kahneman 2011, p. 416). It should be noted that this learning does not amount to a move to a higher stage of HD. It is just psychological economic man’s regular mental mode of operation. It is also important to note that with respect to decision making and judgment, PE is largely concerned with humans’ cognitive functioning, not the non-cognitive functioning that would be part of humans’ move to a higher or lower stage of development. It is interesting to note that PE researchers, although they do not usually state it explicitly in their writings, strongly suggest that the systematic human departures from rationality that they find in their empirical research are ‘hardwired’ in the human brain and/or body. The term hardwired is understood to mean ‘pertaining to or being an intrinsic and relatively unmodifiable pattern’ (Etzioni 2014, p. 394). It is possible, though, that these cognitive errors are merely strong predispositions rather than the determinative attributes that PE researchers imply.2 If the errors and biases are not hardwired, it may be that these departures from rationality can be ‘corrected’ (Etzioni 2014, p. 397), possibly by virtue of education and training, by making a bigger effort, or via other interventions that take advantage of the human brain’s plasticity. The upshot of the above comparisons is that psychological economic man is more realistic than economic man, less rational than economic man, and is no better than economic man insofar as neither experiences human developmental stages.

NEEDED: A SMARTER PERSON IN A BETTER ECONOMICS Essence of the Smart Person Based on our analysis of the economic man of mainstream economics and the psychological economic man of psychological economics, there is clearly a need for a better

M4225-ALTMAN_9781782549574_t.indd 144

03/05/2017 08:20

Smart persons and human development 145 economics, a behavioral economics in which the actor is more human and less irrational. The desired economic actor should behave more realistically than economic man, be less error-prone than psychological economic man, and be more human in a developmental sense than either economic man or psychological economic man. Thus, the needed human actor should not only be a ‘smart person’, a person, who while far from being perfectly rational, is less fallible than psychological economic man, but should be a person whose capabilities and character develop in stages over his or her lifetime. The smart person (SP) is the boundedly rational decision maker whose decision-making behavior is generally in line with Herbert Simon’s understanding of how humans behave when making significant decisions. Therefore, in evaluating decision alternatives, SPs will generally consider a selected set of alternatives, evaluate each alternative sequentially, and then select the first satisfactory alternative, an alternative meeting the SP’s aspiration level (Simon 1955, pp. 110–12). This ‘satisficing’ decision-making procedure is boundedly rational in that it is intendedly rational. However, the SP’s rationality is limited by the human brain’s cognitive capacity and the complexity of the decision environment. It is only in the most simple and transparent situations that SPs can be perfectly rational in the utility maximizing sense (Simon 1959, p. 258). Thus, in the great majority of life decisions, SPs will be boundedly rational, reasonably competent decision makers.3 It is important to note that SP’s decision making can still be expected to manifest many of the errors and biases identified by psychological economics researchers, but these decision-making and judgment deficiencies will not be the defining characteristics of SPs’ decision making. With regard to HD, the SP actor is one who has the ability to develop to his or her potential, progressing along the three developmental pathways mentioned earlier (see Figures 8.1–8.4) as well as to develop along other unspecified paths, advancing stage by stage. The SP’s development may, however, fail to occur sometimes because the person’s developmental environment (parenting, community, society, and so on) has been unfavorable or for other reasons. As a consequence, in the absence of a helpful developmental intervention (for example, educational or therapeutic), the SP’s development along one or more pathways may become stuck. Further, owing to the interdependence of the pathways, progress or lack of progress along one pathway may affect progress or lack of progress along another pathway. Important Features of Human Development To appreciate the human development aspect of the SP, there are some aspects of HD that need further examination, particularly the non-educational aspects. In this regard, it is useful to give more attention to the neurodevelopment pathway. Neurodevelopment success and failure Bruce Perry’s (see, for example, 2002) work makes clear that we can only develop to our human potential if our brains develop to their potential. ‘Development [especially the neurodevelopment part] is a breathtaking orchestration of precision micro-construction that results in a human being’ (2002, p. 82). Eight key processes are involved in creating a mature, functional human brain: neurogenesis, differentiation, apoptosis, arborization, synaptogenesis, synaptic sculpting, and myelination (Perry 2002, pp. 82–5). It is not necessary here to consider each of these processes in detail. Suffice it to say that these processes

M4225-ALTMAN_9781782549574_t.indd 145

03/05/2017 08:20

146

Handbook of behavioural economics and smart decision-making

relate to neurons: their birth, movement, specialization, death, formation into dendritic trees, the formation of connections among neurons (synapses), the structuring of the synapsis, and the creation of efficient electrochemical functioning in the neural networks. These neurodevelopment processes occur in response to experience and are most responsive to experience in positive and negative ways during infancy and childhood (Perry 2002, p. 82). All of these processes must go well; otherwise, abnormal neurodevelopment occurs, causing profound brain dysfunction (Perry 2002, p. 85). ‘In order to develop properly, each [brain] area requires appropriately timed, patterned, repetitive experience’ (Perry and Szalavitz 2006, p. 248). For optimal neurodevelopment, it is crucially important that the lower brain systems develop first in a healthy fashion; otherwise, development of higher, more complex parts of the brain will not be able to occur satisfactorily. Full, healthy brain development may fail to occur for many reasons, most notably, because of adverse early childhood experiences that often involve toxic stress or trauma. Owing to such neurodevelopment deficits, both children and adults can get stuck or partially stuck at a relatively low stage of brain development with serious consequences for their later behavior and functioning.4 Other human development failures In addition to and often coexisting with adverse childhood experiences, three other important non-educational kinds of situations in which humans fail to develop satisfactorily deserve note (Tomer 2014): 1.

2.

3.

The molecules of emotion (different types of receptors and ligands in the brain and body) may fail to flow freely such as when emotions are repressed or denied. As a consequence, body and brain network pathways get blocked, and people get stuck in unhealthy patterns of behavior and experience negative emotional states (Pert 1997). People may fail to develop important emotional competencies (for example, inability to handle one’s distressing emotions) deriving from a lack of coordination between a person’s thinking brain (neocortex) and their lower brain areas (Goleman 2011). People may fail to develop the personality traits that are needed for their educational success, labor market success, health, and positive personal outcomes (Almlund et al. 2011).

The Time Pattern of Human Development There are several noteworthy features of the time pattern of human development. Non-cognitive versus cognitive development In early childhood just after birth, a child is not ready to develop cognitively. The development that is taking place is non-cognitive development, mainly occurring in the lower brain areas (Perry 2002, pp. 86–8; Perry and Szalavitz 2006, pp. 247–8). During very early child development, children are acquiring basic brain organization, a stable emotional basis, a secure attachment to their primary caregiver(s), and the basis for good social relationships. Inevitably, as the child grows older and non-cognitive development progresses, the relative amount of time devoted to non-cognitive development will decline. In other words, as the child matures and becomes more secure, independent, and confident, the

M4225-ALTMAN_9781782549574_t.indd 146

03/05/2017 08:20

Smart persons and human development 147 child’s need for the nurture and care of a parent will become less and less. Also, as the child’s higher brain develops, a greater proportion of the child’s development will be cognitive. More and more of the child’s development will involve learning and acquiring skills. It is useful to view expenditure of efforts and resources to aid both the non-cognitive and cognitive development of children as investments in human capital. After all, both kinds of developmental efforts involve investments of resources that enable humans to function at a higher level whether at home, in the workplace, in the community, or in relationships. Development in transitional periods It should be noted that in addition to early childhood, there are certain other important times during an individual’s lifespan when people typically make transitions from one stage of development to the next. One important example is the transition from middle childhood to adolescence (see, for example, Papalia et al. 2009, ch. 11). Although some young people may experience this transition favorably as an important growth opportunity, it is not unusual for others to experience this transition as difficult and stressful. In many cases, people, often with a great amount of effort and some distress, successfully make these transitions, moving on to the next stage of their life. However, in other cases, people may get stuck or partially stuck at their present developmental stage, and as a consequence of this developmental failure, certain later life opportunities may be precluded. It is useful to think of people who are dealing with transitions as making substantial investments in non-cognitive human capital, investments that sometimes require professional help, such as from social workers or psychologists.

ADULT DEVELOPMENTAL STAGES The developmental stages of children and adolescents have long been recognized, but adult developmental stages have only gained wide recognition in recent decades. Levinson’s (1978) study of adult development is arguably the single most important contribution to understanding the progression of adult lives over the years.5 To understand adult life stages, Levinson studied the life stories of a relatively small number of adults (40 men in four occupations in his 1978 study).6 His findings led him to conclude that an adult’s life has a universal pattern, an underlying systematic, non-genetic progression. According to Levinson (1978), the life course from age 17 to old age consists of a combination of stable periods and transitional periods. During stable periods, a person makes decisions and commits to building a life structure. During transitional periods, a person tends to review and evaluate the present structure of his or her life in order to decide what aspects of their life to keep and what aspects to reject. As Sheehy (2006, p. xvii) explains, humans have a resemblance to lobsters in that during parts of their lives they develop a series of hard protective shells, and during other life segments they shed the shell when it has become too small and confining. Similarly, humans at certain ages tend to find their life structure (a relatively fixed, stable life agenda) coming undone and deteriorating. This may evoke a sense of ‘crisis,’ or at least unsettling feelings, that provide them the impetus and opportunity to change their present life structure in order to incorporate life elements that were not previously part of their life’s agenda.

M4225-ALTMAN_9781782549574_t.indd 147

03/05/2017 08:20

148

Handbook of behavioural economics and smart decision-making

Based on his 1978 research on the pattern of adult development from age 17 to 60, Levinson has identified a number of developmental periods. First are two eras, early adulthood and middle adulthood. Early adulthood consists of two transition periods, early adult transition (age 17 to 22) and the age 30 transition (age 28 to 33), as well as two stable periods, entering the adult world (age 22 to 28) and settling down (age 33 to 40) (Levinson 1978, pp. 56–62). Middle adulthood consists of two transition periods, midlife transition (age 40 to 45) and the age 50 transition (age 50 to 55), as well as two stable periods, entering middle adulthood (age 45 to 50) and culmination of middle adulthood (age 55 to 60). Levinson mentions but did not study late adulthood (roughly 60 to 80) and late, late adulthood. During the stable periods, an adult develops a life structure which has important life components such as occupation, marriage-family, and friends. Adults seeks to create a structure that is simultaneously ‘viable in society and suitable for the self’ (Levinson 1978, pp. 53–4). Ideally, persons will decide on and build a life structure that will enable them to make their greatest contribution to society while enabling them to realize their dreams and values (Levinson 1978, pp. 51, 53–4, 324, 331). If the developmental tasks do not go well and a viable, motivating life structure is not created, the individual likely becomes stuck or partially stuck at an earlier stage of development (Levinson 1978, pp. 321–2). This generally is associated with decline, loss of vitality, imbalance, and stagnation. Erik Erikson’s (1982) writings on human development preceded Levinson’s, and they provide an interesting contrast with those of Levinson. Erikson (1982, pp. 32–3, 56–61, 69, 75) identified eight life (not just adulthood) stages: infancy, early childhood, play age, school age, adolescence, young adulthood, adulthood, and old age. Each stage is concerned with developing a basic strength and avoiding or fending off a core pathology or vulnerability. For example, in Erikson’s fourth stage, school age, children are developing competence and trying to avoid inertia and feelings of inferiority. In his eighth and final stage (old age), individuals are developing wisdom and integrity and avoiding despair and disdain. In the seventh stage (adulthood), individuals are developing generativity and care and avoiding stagnation and rejectivity. As Erikson (1982, p. 59) points out, ‘each [developmental] step is grounded in all the previous ones’. When any developmental step fails, the individual may not only realize the vulnerability or weakness associated with that stage but may regress to an earlier stage (Erikson 1982, p. 67). According to Levinson’s (1978, pp. 319–20) theory, the sequential developmental periods do not imply that adult development follows an ascending or hierarchical order. His view is that ‘the [developmental] tasks of one period are not better or more advanced than those of another, except in the general sense that each period builds upon the work of the earlier ones and represents a later phase in the cycle’ (Levinson 1978, p. 320). Thus, Levinson’s view is that the developmental periods are like seasons in that summer must follow spring, but that summer is not more developmentally advanced than spring. Consistent with this, when Levinson (1986, p. 12) refers to adolescence, he uses the term, adolescing, to mean ‘moving toward adulthood’ and, referring to adulthood, he uses the term senescing to mean ‘moving toward old age’ and death. In other words, when an adult grows older and thereby moves into a later developmental period, it does not imply that the individual’s capabilities have grown. I agree with Levinson to the extent that later developmental periods might simply allow a person to develop a greater range of abilities and interests.7 Levinson’s view, however, is contradicted by his findings indicating that

M4225-ALTMAN_9781782549574_t.indd 148

03/05/2017 08:20

Smart persons and human development 149 certain high-level abilities that middle and older aged people were able to develop could not have been developed unless they had developed certain prerequisite abilities in an earlier period of adulthood. It seems quite clear to me that adults in later life stages are in many cases acquiring capabilities that advance them to higher levels of the HD pyramid than would be possible for individuals in early adulthood. It is certainly true, though, that some older adults are only broadening their range of abilities and interests, not developing higher-level capabilities, and still others’ capabilities may unfortunately be declining as they age. Nevertheless, it is important to note that a significant number of Levinson’s findings seem to support the view that advancing to a later developmental period makes possible the development of certain types of higher capabilities. This viewpoint of ascending capabilities over the life course is more obvious in Erikson’s (for example, 1982) work. He makes clear that the full development of generativity and care must wait until middle to late adulthood even though it is based on seeds planted earlier. The full development of wisdom, integrity, and a number of other virtues must wait until relatively old age despite their basis in strengths developed earlier. From the research of Levinson, Sheehy and others, it is clear that adult human development is generally not a smooth process; stressful episodes and periodic crises are not uncommon. To a certain extent, this is inevitable, and adults need to figure out their developmental paths for themselves. However, as Levinson (1978, pp. 336–40) recognizes, it might make a lot of sense for society to try to smooth people’s developmental paths and to help developmentally failing adults. If a man’s early adulthood is dominated by poverty, recurrent unemployment, and the lack of a reasonably satisfactory niche in society, his adult development will be undermined. His energies will [then] go to simple survival rather than the pursuit of a Dream or the creation of a life structure that has value for himself and others. (Levinson 1978, p. 337)

If it were high on a nation’s priority list, much could be done to help improve adult developmental experiences especially in workplaces. In this regard, Levinson notes that ‘for large numbers of men, the conditions of work in early adulthood are oppressive, alienating and inimical to development’ (1978, p. 338). Levinson also notes that much could be done to provide ‘some degree of emotional support, guidance and sponsorship’ that would permit better development outcomes in early and middle adulthood. A society that does more along these lines is making the kind of investments in human capital that are likely to yield a high payoff for both individuals and society.

SMART PERSONS AND VIRTUE In addition to the various types of human growth that we customarily think of as elements of human development, humans may develop virtues. Smart persons can develop important virtues such as prudence, love of knowledge, courage, firmness, generosity, temperance, and justice. Virtues are acquired capacities or dispositions that enable persons to contribute in some generic way with a high degree of excellence to activities that are challenging and important (McCloskey 2006, p. 64; Roberts and Wood 2007, pp. 60–64). Virtues are not specific, technical skills and do not involve performing specific roles (for example, managing a business or playing basketball). Virtues are habits of the heart

M4225-ALTMAN_9781782549574_t.indd 149

03/05/2017 08:20

150

Handbook of behavioural economics and smart decision-making

(McCloskey 2006, p. 64), and they are deep, enduring, settled character qualities that are formed by education in the broadest sense (Roberts and Wood 2007, p. 69). Virtues may be perfections in the sense of perfecting our natural qualities. Or they may be correctives in the sense of correcting our natural human defects (Roberts and Wood 2007, pp. 68–69). Virtues generally enable us to achieve excellence in some sphere of activity such as the interpersonal, the political or civic, the intellectual, or the moral (Roberts and Wood 2007, pp. 60, 215). Virtues may also enable us to achieve the kind of excellence sought in a certain type of society. The predominant virtues people develop in a socialist or communist society are likely to be quite different from those developed in a capitalist society. In general, the virtues people develop will depend on political ideologies, religious ideals, and the prevailing vision of the good society, etc. As Deirdre McCloskey (2006) explains, capitalist societies, particularly Christian ones, tend to thrive when their citizens manifest the seven ‘bourgeois virtues’ (love, faith, hope, courage, temperance, prudence, and justice). Prudence is the central ethical virtue of the bourgeoisie. However, settling for prudence alone, as all too many economists recommend, is a recipe for societal disaster. A good, stable capitalism can only occur when prudence is conditioned by and integrated with the other six virtues. In other words, in a healthy capitalistic society, it is important that prudence, the profane (P) virtue, be sufficiently balanced by the sacred and social (S) virtues, the other six (see Klamer and Yalcintas 2004; Khachaturyan and Lynne 2010). Virtues, rather than being a product of activities or institutions in which the intended goal is to develop certain virtues, are, generally speaking, developed as a by-product of activities and institutions whose main purpose is something else. For example, in the home, parents’ values, teaching, and example contribute to their children’s later development of virtue. Similarly, school teachers’ values, teaching, and example are an important influence on children’s ultimate virtue development. Another important influence is children’s learning about admirable leaders in political, religious, business, military, entertainment, and athletic spheres. Young people’s virtue development is also influenced by their learning about important events in which the actions of persons in the news have demonstrated out-of-the-ordinary, inspiring qualities. These different experiences of young people may be instrumental in planting the seeds (values, ideals, and so on) that only later when opportunities present themselves develop (with much intentional practice) into full-fledged virtues. Note that with respect to intellectual virtues what is needed is ‘training that nurtures people in the right intellectual dispositions’ in order that they develop the ‘habits of mind of the epistemically rational person’ (Roberts and Wood 2007, p. 22, original emphases). This ‘regulatory’ activity would ‘provide procedural directions for acquiring knowledge, avoiding error, and conducting oneself rationally’ (Roberts and Wood 2007, p. 21). Also note that developing human virtues is an activity that is consistent with progressing to the highest level of development along all three developmental pathways. In other words, developing virtue(s) is consistent with: (1) acquiring overall life direction (pathway 1), (2) connecting to our highest values (pathway 2), and (3) developing creativity and peak performance (pathway 3). Moreover, it is consistent with the idea that virtues represent uncommon, extraordinary development of character (Roberts and Wood 2007, ch. 3). No doubt, the person who has developed a high degree of virtue is a wise person whose thinking and decision making reflect his or her wisdom. This wisdom is not the same as having a high intelligence quotient (IQ), knowing a lot, or having a good technique.

M4225-ALTMAN_9781782549574_t.indd 150

03/05/2017 08:20

Smart persons and human development 151 Wisdom is ‘the moral quality of knowing how to handle your own limitations,’ notably, ‘the ability to go against our lesser impulses [vanity, laziness, cowardice, and so on] for the sake of our higher ones’ (Brooks 2014). A wise person with many virtues is a person who has reached a very high level of HD. Arguably, a behavioral economics for smart people can help us to appreciate the possibility of a wise human actor, but such a high level of HD is not a conceptual possibility in mainstream economics or PE.

THE SMART PERSON RECONSIDERED It is more difficult to specify the qualities and character of the SP than it is for economic man or even for psychological economic man. This is because the qualities and character of the SP are determined by multi-stage developmental processes that do not have well-defined outcomes, even though much can be confidently said about the developmental processes themselves. For example, we now know a great deal about the process of neurodevelopment. However, for a specific person, the childhood neurodevelopment outcome will depend on factors such as the quality of the person’s early childhood environment and the person’s genetic endowment. A different set of factors will determine a person’s developmental progress in later life stages. In general, a person’s development will be determined by the kinds of life challenges the person encounters and how they respond to those challenges. Persons who both experience relatively favorable life situations and who rise to the challenges they face will no doubt develop much farther along the pathways than those for which this has not been the case. Also, a person’s development can go further if he or she has benefited from an intervention (an investment in intangible human capital) designed to help him or her overcome the difficulties that he or she has experienced in the transition from one life stage to the next (see Tomer 2008). In the absence of such an intervention, the person might have become stuck, unable to move on to the next stage.8 How does SP compare to economic man and psychological economic man from the standpoint of the stage of HD they resemble? As suggested earlier, economic man resembles Wilber’s third stage in the development of human consciousness, egocentrism. What stage do the other two stereotypical men resemble? First, psychological economic man’s characteristics cannot be said to resemble any of Wilber’s (2001, pp. 5–13) eight HD stages. This is because psychological economic man does not have a single characteristic way of relating to other humans. Second, SP’s character cannot be definitively specified, and, accordingly, cannot be said to have a close correspondence to the characteristic behavior associated with any of the particular stages of HD identified by Wilber. However, it is possible that the character of the SP economic actor could resemble one of Wilber’s five stages of HD above egocentrism. For example, SP’s character could be strongly conventional and conformist (level 4) or scientific, materialist, and achievement oriented (level 5) or any of the other stages up to level 8, integrative (uniting feeling with knowledge) (Wilber 2001, pp. 9–13). The actual position of a particular SP’s character on this HD hierarchy will depend on the developmental progress that the SP has made. Since very few people reach HD levels 7 and 8 and many reach levels 4, 5, and 6, SP’s development is likely to be in the latter range. More generally, SP’s character, because it is a developmental outcome, is determined by the quality and duration of his or her developmental

M4225-ALTMAN_9781782549574_t.indd 151

03/05/2017 08:20

152

Handbook of behavioural economics and smart decision-making

experience. In light of the above, perhaps we need to think more about the character of the people our societies are developing.

THE PROSPECT FOR A BEHAVIORAL ECONOMICS FOR SMART PEOPLE The development of a behavioral economics for smart people arguably could end up being very important. It could represent a significant step forward, not so much because it will replace earlier economic thought, but because it will strongly suggest both new thinking about what is possible with respect to developing human capabilities (a broadening of the human capital concept) and new thinking about the goals and prospects for economic policy. It could help economists and the public understand how humans, while not having super rational abilities, do have greater potential than previously understood. An important implication deriving from SP behavioral economics is that there is a great deal of human potential that has heretofore not been realized because of the blinders imposed on economic decision makers by prevailing economic thought. There is reason to believe that research in the SP behavioral economics vein will help to remove the blinders and point toward many of the ways in which individual human potential can be realized, and thereby, the potential of economies around the world can be realized.

CONCLUSIONS What economics needs is a behavioral economics with smart people. Unlike the economic man of mainstream economics and psychological economic man of psychological economics, the smart person develops capabilities and character in the course of advancing from stage to stage along a number of major developmental pathways during his or her lifetime. Because some persons will experience unfavorable environments without helpful interventions, they may get stuck and fail to develop very far. On the other hand, other people will advance to high levels of HD along a number of pathways. While smart people are far from being perfectly rational, they can improve their capabilities and character, learning to overcome many of their tendencies to error, thereby becoming competent, boundedly rational, virtuous, even wise, decision makers who make big and small decisions in their own best interests and in the best interests of their societies. A behavioral economics with smart people would presumably be a more optimistic economics. This is because it does not embrace the unrealistic rationality ideal of mainstream economics nor the entirely predictable irrationality of psychological economics. This is also because it embodies an understanding of how humans can in important ways improve themselves and their societies even though they may sometimes fail in the process.

NOTES *

I am indebted to both Leonard Marowitz and Betty Devine who read an early version of the manuscript and made comments and suggestions that have led to significant improvements.

M4225-ALTMAN_9781782549574_t.indd 152

03/05/2017 08:20

Smart persons and human development 153 1. Wilber (2001, pp. 9–13) has estimated the percentages of the population that reach eight major consciousness development levels. He estimates that less than 2 percent of the population reach the highest two levels of development. 2. The language used by PE researchers further implies that the cognitive errors they identify are universalistic, applying to all humans, not just particular groups of people (Etzioni 2014, pp. 394–5). There is, however, some evidence supporting the view that some groups of people do not behave in a predictably irrational manner. 3. Gerd Gigerenzer has explained in his research how human decision makers with limited computational ability and knowledge in the face of limited time and resources may make reasonably good decisions using heuristics, especially when they have had an opportunity for learning how to use them (see, for example, Gigerenzer and Goldstein 1996). 4. See, for example, the important empirical research of Anda et al (2006) and Felitti et al. (1998). 5. Levinson’s research builds to some degree on the research of Carl Yung and Erik Erikson (Levinson 1978, pp. 4–5); see, for example, Erikson (1982). 6. See also Levinson’s (1997) research on the adult developmental patterns of women. 7. It should be noted that there are certain types of human capabilities (athletic, mathematic, music, and so on) for which peak performance typically occurs prior to early middle age. For example, there is evidence that mathematicians generally make their greatest contributions prior to age 40. 8. Note that any society will have many characteristic developmental patterns, which include the typical developmental challenges faced and the typical levels of development reached by their citizens. In a particular society, the term SP presumably would refer to a person whose development outcome is somewhat typical for the society. Of course, in any society, there is considerable inequality in developmental outcomes. Thus, it can be useful to distinguish among groups with broadly different levels of development and competence. Recognizing this inequality may help us think more clearly about the kind of human capital strategy that would be best to achieve a country’s HD goals as well as its economic inequality goals.

REFERENCES Alkire, S. (2010), ‘Human development: definitions, critiques and related concepts’, OPHI Working Paper No. 36, 1 May, Oxford Poverty and Human Development Initiative, University of Oxford. Almlund, M., A.L. Duckworth, J.J. Heckman and T.D. Kautz (2011), ‘Personality psychology and economics’, NBER Working Paper No. 16822, March, National Bureau of Economic Research, Cambridge, MA. Anda, R.F., V.J. Felitti, J.D. Bremner, J.D. Walker, C. Whitfield, B.D. Perry et al. (2006), ‘The enduring effects of abuse and related adverse experiences in childhood: a convergence of evidence from neurobiology and epidemiology’, European Archives of Psychiatry and Clinical Neuroscience, 256 (3), 174–86. Ariely, D. (2009), Predictably Irrational: The Hidden Forces That Shape Our Decisions, revd edn, New York: HarperCollins. Brooks, D. (2014), ‘The mental virtues’, New York Times, 28 August, op-ed page. Erikson, E.H. (1982), The Life Cycle Completed: A Review, New York: W.W. Norton. Etzioni, A. (2014), ‘Treating rationality as a continuous variable’, Society, 51 (4), 393–400. Felitti, V.J., R.F. Anda, D. Nordenberg, D.F. Williamson, A.M. Spitz, V. Edwards et al. (1998), ‘Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults’, American Journal of Preventive Medicine, 14 (4), 245–58. Gigerenzer, G. and D.G. Goldstein (1996), ‘Reasoning the fast and frugal way: models of bounded rationality’, Psychological Review, 103 (4), 650–69. Goleman, D. (2011), The Brain and Emotional Intelligence: New Insights, Northampton, MA: More Than Sound. Kahneman, D. (2011), Thinking Fast and Slow, New York: Farrar, Straus, Giroux. Kail, R.V. and J.C. Cavanaugh (2007), Human Development: A Life-Span View, 4th edn, Belmont, CA: Thomson. Khachaturyan, M. and G.D. Lynne (2010), ‘Review of Deirdre N. McCloskey, The Bourgeois Virtues: Ethics for an Age of Commerce’, Journal of Socio-Economics, 39 (5), 610–12. Klamer, A. and A. Yalcintas (2004), ‘When being virtuous makes sense’, Aelementair, 3 (4), 1–4. Levinson, D.J. (1978), The Seasons of a Man’s Life, New York: Ballantine Books. Levinson, D.J. (1986), ‘A conception of adult development’, American Psychologist, 41 (1), 3–13. Levinson, D.J. (1997), The Seasons of a Woman’s Life, New York: Ballantine Books. Maslow, A. (1943), ‘A theory of human motivation’, Psychological Review, 50 (4), 370–96. McCloskey, D.N. (2006), The Bourgeois Virtues: Ethics for an Age of Commerce, Chicago, IL: University of Chicago Press.

M4225-ALTMAN_9781782549574_t.indd 153

03/05/2017 08:20

154

Handbook of behavioural economics and smart decision-making

Mill, J.S. (1836), ‘On the definition of political economy, and on the method of investigation proper to it’, London and Westminster Review, October. Papalia, D.E., S.W. Olds and R.D. Feldman (2009), Human Development, 11th edn, New York: McGraw Hill. Perry, B.D. (2002), ‘Childhood experience and the expression of genetic potential: what childhood neglect tells us about nature and nurture’, Brain and Mind, 3 (1), 79–100. Perry, B.D. and M. Szalavitz (2006), The Boy Who Was Raised as a Dog and Other Stories from a Child Psychiatrist’s Notebook, New York: Basic Books. Pert, C.B. (1997), Molecules of Emotion: Why You Feel the Way You Feel, New York: Scribner. Roberts, R.C. and W.J. Wood (2007), Intellectual Virtues: An Essay in Regulative Epistemology, Oxford: Oxford University Press. Sheehy, G. (2006), Passages: Predictable Crises of Adult Life, New York: Ballantine Books. Simon, H.A. (1955), ‘A behavioral model of rational choice’, Quarterly Journal of Economics, 69 (February), 99–118. Simon, H.A. (1959), ‘Theories of decision-making in economics and behavioral science’, American Economic Review, 49 (3), 253–83. Simon, H.A. (1983), Reason in Human Affairs, Stanford, CA: Stanford University Press. Tomer, J.F. (2008), Intangible Capital: Its Contribution to Economic Growth, Well-being and Rationality, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Tomer, J.F. (2014), ‘Integrating human capital with human development: toward a broader and more human conception of human development’, unpublished manuscript. Wilber, K. (1996), A Brief History of Everything, Boston, MA: Shambhala. Wilber, K. (2001), A Theory of Everything: An Integral Vision for Business, Politics, Science and Spirituality, Boston, MA: Shambhala.

M4225-ALTMAN_9781782549574_t.indd 154

03/05/2017 08:20

PART II ASPECTS OF SMART DECISION-MAKING

M4225-ALTMAN_9781782549574_t.indd 155

03/05/2017 08:20

M4225-ALTMAN_9781782549574_t.indd 156

03/05/2017 08:20

9

Behavioral strategy at the frontline: insights and inspirations from the US Marine Corps Mie Augier

1

INTRODUCTION

This chapter discusses and applies a few little ideas from the literature in behavioral organizational theory and strategy and discusses some implications for strategic managers of the behavioral strategy framework.1 In particular, it argues that behavioral strategy can help shed light on important decision making and organizational behaviors, illuminating but also helping to address key biases; and helping to contribute to the strategic design of the organizational and psychological architecture of organizations. This may help managers to understand the influences of key behaviors and improve the strategic management of them. The field of behavioral strategy has recently become successful as a scholarly framework (in particular within the strategic management literature); a framework which also can serve as an important lens to understand and address management issues such as behavioral biases. Behavioral strategy as an academic field is more recent than its practice, just as the field of organizations and management existed as practices well before the scholarly studies of them emerged. Indeed, it has been argued that ‘strategy’ as a practice is inherently ‘behavioral’ (Levinthal 2011; Fang 2013). In addition, strategy (and strategic management) is also inherently organizational; the strategic management of business firms and other organizations is the art and science of creating and sustaining competitive advantages in a world of competing organizations. The organizational and behavioral nature of strategy and strategic management combined make the nature of the manager’s task complex and challenging, dealing with complex environment, limited rationalities, and uncertainties and ambiguities; but it is also what makes it possible. There are traits, behaviors, norms, cultures, practices, and organizational mechanisms that alone and together make it possible to understand empirical behavior in organizations and, therefore, improve the management of them. The field of behavioral strategy, explicitly emerging within the organizational and behavioral realm, is therefore a promising field for scholars as well as practitioners; and, as it has evolved, it has contributed to the explication of several key dynamics of decision making and behaviors in organization (around themes such as learning, biases, and the interaction between individual psychologies and organizational characteristics), central to the domain of the strategic management of firms (Levinthal 2011; Powell et al. 2011). Behavioral organization studies and strategy holds valuable lessons for managers on several fronts. For instance, it views the organization as being shaped by its own history (as well as the interaction with others), but not entirely so, as there is room for proactively shaping the strategic environment and one’s performance in it. Behavioral strategy also provides important tools for implementing behavioral insights in practice (Lovallo and 157

M4225-ALTMAN_9781782549574_t.indd 157

03/05/2017 08:20

158

Handbook of behavioural economics and smart decision-making

Sibony 2010). For example, understanding organizational behaviors and decision biases are central to making strategic decisions in a proactive way and shaping outcomes without being trapped by biases and earlier decisions (including investment decisions). Also, (strategic) managers must be able to successfully identify strategic asymmetries in the competitive environment, and translate those into the building and maintaining of competitive advantages (preferably in a sustainable way). Moreover, managers must be able to embrace essential and unavoidable uncertainties in the competitive battlefield – while skillfully adapting their own organizations (with the inertias and competency traps that entails). Strategic management and leadership of organizations is not easy, but behavioral organizational strategy as a framework has valuable tools for understanding the strategic environment, for understanding individual and organizational traps and biases, for understanding strategic asymmetries which can be useful in building organizational capabilities and competitive advantages, and for adapting and implementing the steps as part of the process of organizational adaptation. The potential of the newly emerged perspective includes not only developing scholarship or deriving managerial implications, but also facilitating the interaction and mutual learning between the scholarship and practice, thus enabling further development of empirically realistic scholarship as well as real world management strategies. That is, behavioral ideas are useful not only for our theories about strategy and our management of them; but also for the further (strategic) development of scholarly concepts and ideas (Simon 1986, 1997). In particular, behavioral strategy can help (re)connect behavioral perspectives to the organizational context, relevant as most decisions take place in organizational contexts (March and Simon 1958; Simon 1991). To explicate the importance of behavioral strategy and behavioral ideas in strategy, the chapter looks at a few behavioral concepts and ideas (organizational identification, the exploration–exploitation balance and the importance of organizational ambidexterity, and the nurturing of organizational innovation). Each of these dimensions are central to organizational performance; yet difficult to manage. By using behavioral ideas about organizations, decision making, and strategy to understand examples of such behavioral issues in action, the chapter also uses the perspective as a lens to understand a few aspects of an organization which is seemingly very successful in recognizing and managing the behavioral and organizational aspects of strategy in practice. The rest of this chapter elaborates some of these tools as relevant to understanding the managerial implications of behavioral strategy (section 2), uses those concepts and tools to understand the dynamics of a particular organization, the United States Marine Corps (USMC) (section 3), before discussing insights from the Marine Corps for both management and behavioral strategy as a field (section 4). The closing summarizes the chapter and provides a few suggestions for further research. Throughout the chapter, behavioral strategy ideas are used to explicate three ideas/ concepts/mechanisms useful for understanding aspects of ‘behavioral strategy in action’ in the Marine Corps, and with important managerial implications (which, in turn, may also help develop further the field of behavioral strategy). First, organizational identification and loyalty are powerful altruistic forces which can help minimize costs as well as key biases in organizations (Simon 1991). How can organizations encourage and develop organizational identification and loyalty, and how can strategic managers help shape human motivation in organizations? Using behavioral

M4225-ALTMAN_9781782549574_t.indd 158

03/05/2017 08:20

Behavioral strategy at the frontline 159 strategy and examining how Marines cultivate – in part through design of organizational and psychological architecture of their organizations – a devotion to the organization, and the importance of this to their ability to adapt, have important implications for how organizations can cultivate organizational identification and counter interest biases. The second theme concerns balancing organizational routinization and innovation, and managing inertias. Important research has led to greater understanding of competency traps, stability biases, social biases; and the organizational tendency to, in particular as they grow in age and size, repress variation, creativity and other forces which could lead to innovation. Behavioral research has also shown that it is essential to try and generate and manage a balance between the two. Using behavioral ideas, this chapter looks at how the Marine Corps has tried to pursue both organizational rigidity/stability, and innovativeness (through their educational and organizational structures as well as their approach to leadership). Lessons for management include the importance of intellectual outliers; and a decentralized leadership and management structure allowing for new ideas to be heard, regardless of where in the organization it comes from. The third theme is the pursuit of organizational transformation through a strategy of evolution with design. Creating even small changes in organizations is very difficult and initiating organizational transformation even more so, but is also necessary in order to adapt to the changes in competitor behavior and structure, and other external (and internal) events. The Marines used behavioral ideas in action to guide their most comprehensive transformation in recent (if not all) history, which helped rebuild the organization with ambidextrous elements in the design, enabled learning from failures and from hypothetical futures, and focused on changing how the organization ‘think’, not just what it could do. Lessons for managers include the importance of understanding the future competitive environment (for example, minimize competitor neglect and other actionoriented biases), and to design the organization to embrace competencies and routines that improve efficiency but also with a key role for experiments and learning. All three aspects or themes are examples of behavioral organization theory and strategy ‘in action’, and its managerial implications, but can also be useful for the future development of the field of behavioral organization theory and strategy itself.

2

BEHAVIORAL STRATEGY, A WHIRLWIND OVERVIEW OF A FEW KEY IDEAS

‘Behavioral strategy’, although only recently emerging as a distinct perspective within strategic management, is in many ways embedded in a significant part of the foundation for organizational decision making and strategic management, as well as understanding (and improving) management and strategy making in practice. Behavioral ideas about organizations have been part of the foundation for strategic management theory as it has evolved, both in diagnosing and framing the field’s central questions and in shaping some of the main scholarly perspectives such as evolutionary perspectives and (dynamic) capabilities theory (see, for example, Rumelt et al. 1994; Winter 2000; Augier and Teece, 2007, 2008). Building on the earlier foundational work in organizational behavior and combining behavioral perspectives on organizations, research on behavioral and psychological decision making, behavioral strategy focuses on themes such as learning, attention, satisficing

M4225-ALTMAN_9781782549574_t.indd 159

03/05/2017 08:20

160

Handbook of behavioural economics and smart decision-making

and cognition. Recent behavioral strategy work also has called for a ‘new beginning’ to reconnect with behavioral economics and finance; to explicate the psychological grounding; and to integrate further psychology and strategy (Powell et al. 2011).2 Identified by merging ‘cognitive and social psychology . . . with strategic management theory and practice’ and aiming to ‘bring realistic assumptions about human cognition, emotions, and social behavior to the strategic management of organizations’ (Huy 2012, p. 240), the ‘new’ behavioral strategy is often largely rooted in individual behavior, but behavioral strategy as a whole also has significant organizational aspects, including insights into the development of organizational competency traps which may be beneficial for the organization in the short run, but inhibit ability to adapt to change in the longer run (March 1991). Implications for management and real-world strategy include illuminating certain key biases and bringing to light unconscious processes in organizational decision-making processes and asking key questions regarding the ongoing strategic processes in organizations, including if the strategic direction has to change, how to adapt and to what, and what key trends are important for shaping the strategic context for the firm. Research into the nature of organizational competency developments and the nature and need for innovation and change (as well as the psychological and organizational barriers to change) also include work on the mechanisms enabling and discouraging innovation and the possibilities and barriers to organization’s ability to adapt to disruptive changes (including technologies) (March 1991; Christensen 1997). Methodologically, behavioral strategy embraces multiple methods and methodologies in order to better understand real behavior and decision making in organizations.3 Using multiple methods, and disciplines, is important since understanding real-world behavior does not fit one or two lenses very neatly; yet it is also quite difficult for many reasons, including institutional and intellectual homophily which inhibits variation at the organizational as well as the ideas level. The mechanisms of homophily (operating in organizations as well as in communities of scholars) includes the inclination of organizations to recruit and retain people who are similar to each other in beliefs and competencies, as well as individuals within organizations often seeking to work with those individuals who are most like themselves. Over time, organizations forming groups of similar individuals are unlikely to create and encourage new and innovative thinking, so central to the healthy development of organizations and academic fields in the long run. Behavioral strategy, by embracing an empirically relevant perspective and using multiple methods, may be less likely (but not immune) to get stuck in such traps. Recent contributions relate behavioral strategy to related perspectives from behavioral neuroscience and cognitive neuroscience and reference point theory (Powell et al. 2011). Lovallo and Sibony (2010) emphasize also managerial implications, including biases which management can understand and discuss using the language of behavioral strategy. The biases are often manifestations of behavioral and organizational dynamics, and are shaped by such dynamics often in co-evolutionary and self-reinforcing ways. For example, ‘stability’ biases and organizational inertias are products of both individual psychologies and resistance to change (such as loss aversion and biases for status quo), as well as the way in which organizational routines and competencies develop, often leading to under-sampling of trying new things, and oversampling of doing more of the same, although for the organization to survive in the long run, a balance is central (March 1991).

M4225-ALTMAN_9781782549574_t.indd 160

03/05/2017 08:20

Behavioral strategy at the frontline 161 Another example around human motivation and conflicting interests in organizations opens the door for issues such as misaligned incentives (people pursuing individual goals at the expense of the organization’s goals) or misunderstanding (and misperceptions) of organizational goals (Cyert and March 1963). Understanding and awareness alone does not eliminate biases, but it is a central first step. For example, ideas from organizational behavior can help managers understand and design organizations that are more ambidextrous and that have mechanisms to cultivate creativity and ‘hot groups’ within organizations; groups that explore new ideas, despite stability, competency traps, and social biases for the status quo (March 1991; Leavitt 1995). Moreover, behavioral strategy can offer insight into the ‘strategic organizational design’ of organizational structures as well as the psychological and social architecture of the organization to help create a great sense of organizational identity and loyalty which can help curb issues of misaligned incentives and possibly help de-bias interest biases (Simon 1991). In both examples, integrating insights from both the older and more recent work in behavioral strategy into the behavioral and organizational nature of decision making can help us to better understand and address the biases. 2.1

The Importance of Organizational Identification

Herbert Simon’s contributions to behavioral organization studies and strategy include providing (with co-authors) many of the concepts which underlie behavioral economics and strategy (such as bounded rationality and satisficing) (March and Simon 1958). There is also in his work much untapped potential for strategy theory and management (Simon 1993; Augier and Sarasvarthy 2004). One source of insight and inspiration comes from his ideas on organizational identification, loyalty, and altruism; all ideas and mechanisms which counter self-interest seeking behavior and ‘interest biases’.4 A few decades ago, Simon pointed to the irony that despite the ubiquity of organizations, many of our scholarly theories of organizations were based on individual behavior and analysis of markets, rather than organizations (Simon 1991). He also emphasized the fact that real-world behavior in organizations often included motivational aspects that many theories could not explain, yet they were important to organizational dynamics and performance. Also, as Powell et al. pointed out, in behavioral strategy, ‘the whole question is how particular forms of behavior arise in and among organizations. If we do not show the mechanism, we do not explain the phenomenon’ (2011, p. 1375). A key in Simon’s ideas on identification is that while putting an organization’s goals ahead of our own may not be beneficial for individual organizational or societal members, it contributes to the overall adaptive and strategic fitness of the organization: ‘[S]ocial evolution often induces altruistic behavior in individuals that has net advantage for average fitness in society’ (Simon 1993, p. 160). He also finds that most economic and strategic theories ignore the powerful role that organizational identification can have in shaping both organizational member’s goals as well as their perceptions and cognition. Thus organizations with a higher degree of organizational identification and loyalty are more likely to be able to sustain elements of altruistic behaviors and trust, reducing further costs (for example, those associated with contracts and monitoring of possible agency problems), as well as strengthening the organization’s sense of identity and ‘culture’. Organizational identification can also help improve organizational coordination, as evolved shared goals

M4225-ALTMAN_9781782549574_t.indd 161

03/05/2017 08:20

162

Handbook of behavioural economics and smart decision-making

and norms serve as ‘focal points’, reducing the costs of coordination and also making it more adaptive (Simon 1991). Also, as organizational members develop a greater sense of loyalty, they become more inclined to identity with the organization’s goals rather than their own, thus reducing ‘interest biases’. How can managers cultivate, foster and sustain organizational identification? Most management literature gives little clues – building explicitly or implicitly on notions of self-interest seeking individuals (embedded in concepts of ‘opportunism’ or ‘guile’; Dosi 2004; Augier and March 2008). The difficulty of building organizational identification, given its importance, suggests that studying organizations which have a strong culture of organizational identification might be useful. 2.2

Nurturing Innovation, Creativity and ‘Hot Groups’ within Organizations to Help Develop Entrepreneurial and Dynamic Capabilities

Even though many organizations have evolved towards being smaller, flatter, and more decomposed over the past decades, still many organizations are quite large. Big organizations are great for many things, but they also have some weaknesses, at least from the point of view of contributing to innovation and innovative capabilities in the long run. Behavioral strategy is a useful lens for understanding and addressing key challenges of big organizations, including: ●

●

●

Stability biases and homophily dynamics. The tendency of organizations towards institutional and intellectual homophily, and towards reinforcing stability and social biases, and a tendency towards inertia rather than innovation. Inertia and competency traps. One of the most important concepts to capture the tendency for organizations to get ‘stuck’ doing more of the same is March’s idea of competency traps, and the need for organizations to balance exploring and exploiting activities (March 1991). March has argued how organizations, in order to be adaptive, must nurture and cultivate also the generation of new ideas, of creativity and other sources of innovation. The balancing between stability and change, exploration and exploitation is one of the most difficult things to do, and organizations such as Kodak are examples of failure to adapt owing to a fundamental imbalance between exploration and exploitation. Organizations as ‘unhealthy’ environments. Hal Leavitt adds a dimension to the competency trap argument by arguing that big organizations are ‘unhealthy’ environments for human beings, in part because their hierarchical structure undermines intellectual freedom and creativity which can lead to the generation of new ideas and innovations at the organizational and industry level (Leavitt 2007).

Other reasons for the stifling of entrepreneurial activities include social biases and norms leading to the elimination of outliers, and discouraging creative and ‘out-of-thebox’ thinking. Repressing creativity and the generation of new ideas is detrimental not only to the development of innovations in the longer run (and on the industry level), but also to the generation of organizational capabilities in the individual firms. Creative thinking and entrepreneurial behaviors are needed for managers too, and if not encouraged at all levels of the organization it is less likely that managers will exhibit entrepreneurial qualities.

M4225-ALTMAN_9781782549574_t.indd 162

03/05/2017 08:20

Behavioral strategy at the frontline 163 Managers must think outside of the box, set new goals, visions, aspirations for the organization and create new paths forward (Simon 1991). Thus, understanding better how some organizations cultivate individual outliers as well as hot groups is essential to the management of organizations (as well as giving inspiration to current managers as to how to generate such entrepreneurial dynamics within the firm) (Leavitt 2007; Augier et al. 2015). Carefully understanding the biases and dynamics of one’s organization (as well as others) can help management counter some of the biases and help determine and set clear and realistic strategies. The combined insights into behavioral and organizational aspects of decision making improve management’s ability to learn from and adjust its organizational and psychological architecture in order to develop robust capabilities for competing in the long run. The next sections take a look at some of the mechanisms and concepts of behavioral strategy ‘in action’.

3

COMPETING ON WARRIOR CAPABILITIES: THE ART AND SCIENCE OF BEHAVIORAL STRATEGY IN THE USMC

One organization with great potential for strategy/management/leadership insights is the USMC, a well-established organization, older and larger than most (business organizations at least) – yet it has received surprisingly little attention from strategy, organizations and management scholars, although it is an organization which can give a lot of insight and inspiration for practicing managers (as well as possibly for strategic management scholars), regarding the art and science of strategy. The Marine Corps provides useful examples of ‘behavioral strategy in action’, having both behavioral ideas about individual and organizational decision making (such as embracing uncertainty and being highly adaptive), as well as having mechanisms in place to counter some important biases.5 In particular, the Marine Corps’ ability to cultivate organizational identification and loyalty, its ability to maintain innovativeness, ambidexterity, and balancing of exploration and exploitation (even within a large organizational and seemingly hierarchical structure), and its adaptiveness through evolution with design might provide useful insights for managers and management scholars alike. The Marine Corps has a rich history of competing in several different environments (including peacetime) since its birth in 1775. The issue of ‘What makes marines a marine’ has been the source of puzzlement for decades, including the strong sense of organizational identification and unity – despite seemingly strong hierarchy – as manifested, for instance, in General Gray’s statement, ‘Every Marine is, first and foremost, a rifleman’ – putting functional specialties and individual interests aside. Indeed, a lesson from Gray is ‘always put your organization and your people ahead of yourself’ (personal conversation with General A. Gray), an interesting contrast to much of management theory’s emphasis on self-interest seeking. There are many components to what makes Marines seemingly more agile and able to adapt, including: issues of the organizational structure of the organization; how they cultivate their particular organizational culture; and if and how they attract those with more broader and curious minds. Here the focus is on only a subset of components that can help us see the relevance of behavioral strategy – for understanding organizational dynamics but also for learning from them to better manage our organizations.

M4225-ALTMAN_9781782549574_t.indd 163

03/05/2017 08:20

164

Handbook of behavioural economics and smart decision-making

At its most basic level, organizational capabilities include resources (physical, intellectual and intangible), organizational aspects facilitating and shaping motivation, structure and processes. As all organizational capabilities, the Marine Corps organizational capabilities are more than the sum of its parts; more than the resources, training, and organizational structures, it is the synergy between them that creates uniqueness. Elements of Marine Corps ‘warrior capabilities’ include the following. 3.1

Cultivating Learned Selflessness, Organizational Identification and Loyalty

Human motives change over time, responding to experience and the surprises of history. (Simon 1993, p. 160)

As mentioned above, behavioral strategy suggests that certain kinds of human motivation, such as loyalty, can increase organizational identification, which in turn can help minimize interest biases and improve alignment with organizational goals. How organizations train, mentor and educate people helps provide a better understanding of the organization’s goals, values and culture. The Marines have an exceptional sense of organizational loyalty and dedication to the organization’s goal. Part of that may be explained due to the purpose and mission of the Marine Corps (as it might attract people dedicated to higher causes), but there is also an important element of strategic organizational design, including design of the psycho-organizational mechanisms to build and cultivate team spirit, organizational identity and loyalty. The Marines cultivate and build organizational identification and loyalty that encourages a kind of ‘learned selflessness’ and concern beyond oneself, and an identification with the organization’s goals, rather than individual goals, thereby contributing to one of the most powerful altruistic forces in Simon’s discussion (Augier and Guo 2016). The Marines do that at several (interrelated and overlapping psychological, psycho-cultural, sociological, physical, intellectual) levels, including: ●

●

●

Sematic level: entering boot-camp, young Marines are no longer identified by their individual name but by reference to ‘this recruit’, ‘that recruit’, and so on; so the loss of ‘self’ relative to the identification with the organization is embedded even in the language used. Symbolic level: a first symbol of losing individual identity and giving up self is yellow footprints at the entrance to boot-camp, symbolizing a new path for the young Marines, stepping towards the disappearance of the individual and into a tradition and culture of selflessness and sense of duty. Stripping away individual characteristics include also the haircuts; sometimes making the young marines unrecognizable to themselves (which helps them shed their individualism and build team and organizational identity). At the physical level, Marine boot-camp is well known for its grueling exercises and tough physical standards. That too helps build team identity. For instance, by punishing recruits as a group for individual behavior (for example, being slow or otherwise not living up to the organization’s standards) helps instill the sense that any organization or team is only as strong as its weakest link (and discourages agency problem behavior). Several exercises are also designed to only be able to be

M4225-ALTMAN_9781782549574_t.indd 164

03/05/2017 08:20

Behavioral strategy at the frontline 165

●

successfully finished as teams. They also are taught to embrace uncertainty and fear, and to run towards the sound of the gun, not away from it. The fact that the Marine Corps builds a new identity instills in the young men an attitude that being a Marine is not a job; it is a calling. Such values help encourage an organizational identity and loyalty, and a pursuit of a logic of identity rather than one of consequences.

Thus the building of team identity, organizational identification and loyalty in the Marines has several levels and layers; including aspects of ‘strategic organizational design’ (for example, of exercises) as well as design of the psychological architecture and mental characteristics of marine training and education. This involves knowledge of what motivates humans and how to train and change psychology and behavior, as well as how individuals interact and are shaped by others and by the organizations they are in, in a coevolutionary way. The facilitation of organizational identification and loyalty can reduce possible conflict of interests as well as ambiguities and misperceptions about goals and interests. It also helps Marines to be highly adaptive on the battlefield. 3.2

Competing on Warrior Capabilities: Designing Ambidextrous Organizations to Counter Biases and Competency Traps

Marine training also signifies another interesting aspect of the organization; it exemplifies behavioral strategy in action and illustrates one way of managing the dynamics of inertia and innovativeness – not by balancing them (as in ‘either-or’), but by integrating them (both-and) into the heart of the organization’s capabilities. As the Marine Corps’ training guide states: ‘Training must be challenging. If training is a challenge, it builds competence and confidence by developing new skills. The pride and satisfaction gained by meeting training challenges instills loyalty and dedication. It inspires excellence by fostering initiative, enthusiasm, and eagerness to learn’ (USMC 1996, p. 4). It is not that marine training does not emphasize routines and rigidity; boot-camp is, after all, about the transmission of basic routines needed for successful military operations; marine training in history, martial arts, swimming, land combat – and training 37 000 recruits annually – requires much standardization, routinization and rigidity (Guo and Augier 2015). However, along with a reputation for rigorous training, the Marine Corps has also developed an instinct and inclination for innovativeness and a capacity for out-of-the-box thinking (for instance, the Marines have implemented one of the most visible and effective energy programs, the Expeditionary Energy Office). How can one organization simultaneously pursue rigidity and innovativeness? The routines underlying Marine training are focused enough to provide direction, yet flexible enough to accommodate the needs to adapt to changing conditions and commanders at all levels of the organization. Moreover, Marines are encouraged to think outside the box and ‘think strategically’ at all levels, despite a seemingly very hierarchical organization. The ability to listen to ideas from all levels is key to other intellectually innovative organizations in the past (Augier et al. 2015). The uniqueness of the Marine Corps from an organizational and capabilities perspective includes not only the important element of identification and altruism mentioned above, but also the fact that the Marine Corps is, effectively, ambidextrous by nature, mixing and integrating ‘core competencies’ of other services. While the Army is known for

M4225-ALTMAN_9781782549574_t.indd 165

03/05/2017 08:20

166

Handbook of behavioural economics and smart decision-making

and trains for land combat, the Air Force for its flying, and the Navy for sea capabilities, the Marine Corps has important land as well as sea and air components, so it has a built-in need for flexible capabilities as well as for a mindset that is both able to learn from the past (and from other organizations) as well as creating the future (and combining capabilities and ideas). The embracing and simultaneous pursuit of exploration and exploitation in the Marine Corps is made possible in part by the leadership style. Far from being about micro management, a key to how Marines operate is the shared understanding of the commander’s intent (again, made possible by organizational loyalty and identification). In particular when in combat, the ability to adapt and be innovative at the front line comes from the ability of junior leaders to make in-real-time decisions – based on their understanding of their leader’s intent.6 The shared understanding of the organization’s goals (minimizing interest biases through training and the building of loyalty) is also made possible through organizational communication; not formal communication channels but largely informal and implicit channels; almost shared mental and cognitive models or frames: We believe that implicit communication – to communicate through mutual understanding, using a minimum of key, well-understood phrases or even anticipating each other’s thoughts – is a faster, more effective way to communicate than through the use of detailed, explicit instructions. We develop this ability through familiarity and trust, which are based on a shared philosophy and shared experience. (USMC 1997, p. 72)

3.3

Organizational Transformation is Difficult but Not Impossible

A third lesson from the Marine Corps exemplifying behavioral ideas in action relates to the difficulties in creating strategic change and transformation in organizations. Organizational change and transformation is so central to organizational adaptation, yet also very difficult. Powerful mechanisms of individual and organizational inertia include stability biases (such as preference for status quo, anchoring and loss and risk aversion); social norms and biases; and individual and organizational and bureaucratic inertia which alone and together make organizational change exceptionally difficult. The Marine Corps has successfully changed and adapted to the changing strategic environment over the past centuries. One of the most recent comprehensive transformation of the Marine Corps into a more adaptive organization was in the late 1980s and 1990s, led by legendary Marine General Alfred Gray, which manifested behavioral strategy at organizational design level, living strategy and the strategic management of organizational change as a process of evolution with design. A core insight is the limited rationalities and psychologies (and dynamics of the changing strategic landscape) not only of the strategic competition (who might opponents be 30–40 years from now?), but also the limited rationalities and psychologies of our own organizations and people. Gray knew that in order to really improve the Marine Corps, he had to change ‘how the organization thought’. The transformation needed to be intellectual as well as organizational and, starting with a hard diagnostic look, needed to be about how to think, not what to think. Gray outlined a framework for ‘war-fighting’, which evolved as a symbol of adaptive and strategic thinking. A key text which is read by all Marines at all levels, ‘War-fighting’ was first described as a ‘doctrine’ but was much more than that; it is a philosophy and a strategic way of thinking instilled in all Marines regardless of rank and age, and to be

M4225-ALTMAN_9781782549574_t.indd 166

03/05/2017 08:20

Behavioral strategy at the frontline 167 applied in peacetime as well as in wartime. In Gray’s words, it is a ‘philosophy for action that, in war, in crisis and in peace, dictates the Marine Corps approach to duty’ (personal conversation with General A. Gray). The overall war-fighting framework also became a foundation and framework for the Marine Corps’ central initiatives, such as maneuver warfare and decentralized leadership.7 Central ingredients in the philosophy were to be able to understand competitor weaknesses (a strategic asymmetry) and to use agility, decentralized decision making and speed to their own advantage. In effect, maneuver warfare emerged almost as a behavioral and evolutionary alternative to previous static approaches. The broader context for war-fighting and maneuver warfare as paradigms for enabling strategic change within the Marine Corps had a lot to do with the strategic context in which the Marine Corps found itself after Vietnam. The concept and philosophy did not come out suddenly, but resulted from organizational adaptation and careful strategic organizational design in order for the organization to adapt to the changes in the external strategic environment. Gray had encountered maneuver warfare earlier (having spent a lot of his youth overseas and having done a lot of reading). He had read a lot of Clausewitz – but also developed an affinity for Sun Tzu, studied strategic deception, and was interested in learning from mistakes made in past conflicts (in particular, Vietnam and Korea). The end of Vietnam War was a time of crisis for the country but also for the Marine Corps as an organization. They had lost some of their traditional core competencies and had big moral problems and knew they needed to upgrade the quality and capability of the organization, starting with education; broad reading lists also became an integral part of the Marine Corp’s educational experience, helping young marines to be intellectually agile as well, and cultivating broad and curious minds and lifelong learning. Another aspect of the upgrading of Marine Corps capabilities had to do with learning from experience and from experiments including failures. To buffer innovation within the organization (and to protect it from mechanisms that would kill it), Gray set up a maneuver warfare board in the second marine division in order to show the rest of the organization that using maneuver warfare could be successful and was central for organizational adaptation. At the training level, Gray also set up a combined arms operations exercise to test ideas. This was a free-play type exercise in order to inspire and facilitate learning across all levels and to encourage creative thinking and after-action discussions in environments where subordinates could speak freely and contradict superiors without fear and commanders could learn to receive criticism. While the maneuver warfare board as an organizational experiment did not last long, it illustrates the kind of experimentation with ideas and organizations which behavioral strategy scholars have argued can lead to improved adaptation (March 1991). It also illustrates Gray’s profound belief in people and in ideas; when asked about how he managed to change organizations (despite all the reasons organizations resist changed), he said ‘It’s easy! Unleash the people with ideas; and protect them from bureaucrats, admin and paperwork’ (personal conversation with General A. Gray); a refreshing bottom-up approach to change, very much in keeping with the emphasis from March and others on the need for nurturing novel and new ideas. In keeping with this, the current Marine Corps commandant issued a call just last year for more disruptive thinkers in the organization.

M4225-ALTMAN_9781782549574_t.indd 167

03/05/2017 08:20

168

4

Handbook of behavioural economics and smart decision-making

MANAGING ORGANIZATIONS WITH THE TOOLS OF BEHAVIORAL STRATEGY AND WITH INSPIRATION FROM BEHAVIORAL STRATEGY IN ACTION

Building on the foundations and concepts of (old and new) behavioral strategy, and using some insights and inspirations from the USMC, we can find several ideas useful for strategic managers in organizations today. These include the following. 4.1

Building Organizational Loyalty

Leaders should think more about others than themselves . . . Being in the [organization] is not a job. We don’t work . . . We serve. (General Alfred Gray, 1991)

Recognizing the importance of mechanisms and behavioral ideas such as organizational loyalty, identification and altruism, how can managers help encourage and cultivate such behaviors? While business organizations probably will not develop ‘boot-camps’ for their people anytime soon, looking further into the psychological mechanisms of how the Marine Corps builds identification through boot-camps might provide insights that managers can be inspired to try to cultivate organizational identification.8 4.2

Embracing Uncertainty and the Simultaneous Pursuit of Exploration and Exploitation Can Help Counter Pressures towards Competency Traps

In the ‘fog of war’ there is chaos, and in that chaos opportunities present themselves. (Gray 1987, p. 18)

Behavioral strategy embraces uncertainty and ambiguity rather than trying to repress it. The competition organizations face involves uncertainty but if embraced, rather than assumed away, can also help shape the competition in the future. This involves understanding the psychologies of competitors, trying to create and utilize asymmetries in the competition to create and sustain competitive advantages. At the heart of this is a behavioral conception of decision making with individuals being limited in their rationalities and computational powers (March and Simon 1958; Cyert and March 1963).9 Thus the basis of the management of organizations in behavioral strategy is the ambiguity and uncertainty inherent in all decision making, giving rise to particular behaviors and necessitating the facilitation of shared perceptions and beliefs, starting with the leader’s vision and an understanding of the nature of the organization and its strategic environment. Technology has a place in uncertain situations, but not trying to reduce it. Marines use ‘technological advances to facilitate the human interface, not to chase after certainty in an inherently uncertain environment’, thus living the essential ‘ambiguities of experience’ (Mattis 2006, p. 16). Another insight from the Marine Corps is its ability to be adaptive in the battlefield owing to decentralized decision making and, essentially, satisficing (not maximizing) decisions. This is consistent with insights from behavioral decision making and strategy that have emerged with implications for both individual decision making and organizational behavior. The insight is that, in addition to perception, the human brain has other modes of thought – intuition and reasoning – each with its own characteristics. Intuition is

M4225-ALTMAN_9781782549574_t.indd 168

03/05/2017 08:20

Behavioral strategy at the frontline 169 particularly relevant to on-the-spot decision making in competition where there is no time for careful analysis of alternative options – much less any attempt to optimize. Instead, behavioral strategy and the Marines’ approach show that there is an evolutionary and adaptive value to being able to make on-the-spot ‘good enough’ solutions (Simon 1955). The research into and the importance of how Marines, chess players, fighter pilots, intensive care medical personnel and others use intuition and knowledge of past experiences without much analysis also invites further research on behavioral strategy into these processes and the synergies between the art and the science of strategic management. 4.3

Pursuing Organizational and Strategic Transformation through ‘Evolution with Design’

One of the paradoxes of organizations is that the more they build capabilities to do one thing, the less inclined they are to do others. Management scholars have pointed to the importance of ambidextrous organizations; those that can manage and balance both exploration and exploitation. Embracing a metaphor of organizational strategic management as one of evolution with design puts emphasis on the continuing strategic transformation and renewal of organizational capabilities as well as using and refining existing ones. Managers can help create better environments for this through strategic organizational design of the organizational and psychological architectures to facilitate learning, including from failures and counterfactual histories. Also central is an environment where ideas matter at least as much as rank; new ideas often do not come from the top of the organization, and if organizational members do not feel free to discuss them, these ideas will never reach the top. Google and others are famous for having setups for experimental thinking, and RAND, many years before, had carefully thought of this as well. It is essential to have an environment that encourages creative thinking. A former commandant of the Marine Corps University noted that he wanted a place where ‘freedom of thought was not only encouraged but rewarded. The idea is that the experimentation should be taken to the failure point . . . that only by reaching that point would we understand the unknowns’ (personal conversation with General A. Gray). Finally, at the level of the strategy making and the strategic leadership of such evolutionary and behavioral management processes, an important implication is captured in Henry Mintzberg’s image of a potter modeling clay into a piece of pottery being a better metaphor for how organizations actually develop strategies (rather than the planning metaphor in which senior decision makers formulate courses of action based on systematic, rational analysis of oneself and competitors, one organization’s strengths and weaknesses, and so one) (Mintzberg 1987). Such lines of thoughts suggest that we can generate implications for the generation of management strategies, including: learning about (and trying to cultivate) organizational identification and loyalty (to minimize interest biases; encouraging innovative and outof-the-box thinking (countering stability biases and competency traps); and embracing a vision of strategy as one of evolution with design, with decentralized leadership with emphasis on shared visions.

M4225-ALTMAN_9781782549574_t.indd 169

03/05/2017 08:20

170

5

Handbook of behavioural economics and smart decision-making

CLOSING

The early work by behavioral organization scholars on the limitations of rational decision making was central to the development of ‘behavioral economics’ and its applications to investment decisions (and the subfield of behavioral finance); and behavioral organization theory also was significant as a foundation stone for the very field of strategic management as it began to take off in the 1990s. Work in the field of strategic management has continually built on and integrated these ideas as it has also evolved concepts (such as core competencies and dynamic capabilities) that are behavioral in origin and with implications for management, and this has helped the field of behavioral strategy to take off. Discussing the new developments in the field of behavioral strategy, Powell et al. (2011, p. 1370) noted that practitioners are skeptical, ‘doubting whether the field can go beyond cognitive biases to produce useful framework that integrate psychology and strategy practice’. They then attribute a large part of the problem to ‘inadequate paradigm development’ within the behavioral strategy camp itself, partly due to terminological confusion and not being sufficiently embedded in the existing intellectual and institutional structure of the larger research community.10 However, such a ‘pre-paradigmatic’ state also offers opportunities. For example, as fields become more ‘mature’ and professionalized, centripetal forces often lead to an increase of conversations within itself, thus inhibiting interdisciplinary learning. Moreover, it is also interesting to note that other fields and subfields developed some of their most successful – and empirically relevant and operationalizable – contributions before they became too self-aware of their development as a ‘field’. This is not an argument against developing behavioral strategy further; that is key too. However, doing it in a way that is less concerned about its relative status within the academic fields, and more concerned about developing behavioral strategy in a ‘behavioral’ (and empirically realistic) way, means that the field is less likely to get too far trapped in the usual competency traps of specialization, and more likely to be able to retain core behavioral elements, such as using different research methodologies, as well as being able to help organizational and behavioral strategy remain empirically relevant in Simon’s sense (1997). Several decades of work in behavioral strategy ideas and perspectives have brought about important concepts and ideas on several fronts. For example, at the level of the development of ideas on management, the areas of management, organizations, economics, leadership and strategy have become enriched with behavioral ideas and concepts (such as bounded rationality, routines, slack, and learning), helping to motivate new subfields such as evolutionary and capability reasoning, which, in turn, are key inspirations for the ‘new’ behavioral strategy framework. Such a framework, especially when combining old and new behavioral ideas, invites research on organizational altruism, innovation, ‘hot groups’, intellectual outliers, and creativity – in addition to already semi-established sub-fields such as entrepreneurship and learning. In addition to being a useful lens for developing the field(s) of strategy organization, and understanding issues such as biases in organizations, the behavioral strategy framework is also possibly useful for examining practices not usually well understood in the strategy literature – including issues of organizational identification and loyalty. Although many of these go against much of the scholarly work on strategy, they are important parts of the real-world psychological and social processes in organizations; and also can deliver important value (contributing to issues such as retention, trust, and

M4225-ALTMAN_9781782549574_t.indd 170

03/05/2017 08:20

Behavioral strategy at the frontline 171 networks). By explicating such processes in real-world organizations, we may also, in time, help shape the future development of behavioral strategy as a field.

NOTES 1. 2. 3. 4.

5.

6.

7.

8. 9.

10.

The use of the term ‘little ideas’ is inspired by Jim March’s style and approach to research (Augier 2015; Maslach et al. 2015). We can further argue that since strategy is essentially organizational in nature, it might be relevant to reconnect behavioral strategy (again) with organizations (without downplaying the work on individual decision making), and to try and develop behavioral strategy in a ‘behavioral’ way (Simon 1986). This is an argument in both behavioral strategy and some of its intellectual foundational roots. See, for example, March (1965), Simon (1954) and Powell et al. (2011). Understanding the mechanisms of organizational identification and how it influences organizational performance and strategy can also help address a call from the ‘new’ behavioral strategy school: ‘Behavioral strategy has a long way to go in linking individual psychology with organizational strategies. One of the distinctive features of strategic management is its emphasis on collective behavior, and behavioral strategy must explain the psychological or social mechanisms by which mental processes affect organizations’ (Powell et al. 2011, p. 1374). Using empirical behavior in organizations as inspiration for understanding certain managerial behaviors (as well as for theory development) is consistent also with the ‘old’ behavioral organizational perspective underlying the field of strategic management as the ‘old’ field of behavioral organization studies started from (1) using different disciplinary perspective and ideas to get insights into real organizations and (2) using empirical studies of mechanisms in real-world organizations, to inform the further development of those theories as well as facilitating a better understanding of practice, thus embracing both an inter-disciplinarity in theories and methodologies as well as two-way street learning between scholars and practice (Simon 1986). ‘[S]ubordinate commanders must make decisions on their own initiative, based on their understanding of their senior’s intent, rather than passing information up the chain of command and waiting for the decision to be passed down. Further, a competent subordinate commander who is at the point of decision will naturally better appreciate the true situation than the senior commander some distance removed’ (USMC 1997, p. 71). We elaborate on the leadership aspect in Augier and Guo (2016). Inspired by Sun Tzu, and in keeping with the Marines’ embrace of uncertainty, maneuver warfare is about embracing and even creating ambiguity and uncertainty: ‘Maneuver warfare seeks to shatter the enemy’s cohesion through rapid, focus and unexpected actions which create a chaotic situation with which the enemy can not cope’ (USMC 1997, p. 73). This also appeals to a logic of identity rather than a logic of consequences, and doing things not because of their consequences, but because it is a calling and a duty; a sense of ‘doing what must be done’ (Gray 1991a). This has been at the heart of behavioral ideas since Simon’s landmark articles in the 1950s explicating the dynamics of the limits to rationality and satisficing. Also embraced by Marines: ‘A military decision is not merely a mathematical computation. Decision making requires both the situational awareness to recognize the essence of a given problem and the creative ability to devise a practical solution. These abilities are the products of experience, education, and intelligence’ (USMC 1997, p. 86). ‘The term “behavioral strategy” is not widely used and means different things to different people. Behavioral strategy does not have an agreed statement of purpose, conceptual framework, core research problems, methodological standards, communities of scholarship, or supporting institutions’ (Powell et al. 2011, p. 1370).

REFERENCES Augier, M. (2015), ‘The power of “little ideas”’, Journal of Management Inquiry, 24 (3), 322–3. Augier, M. and J. Guo (2016), ‘Overcoming negative leadership challenges through we-leadership: building organizational commitment with inspirations from the United States Marine Corps’, in D. Watola and D. Woycheshin (eds), Negative Leadership: International Perspectives, Kingston, Ontario: Canadian Defence University Press.

M4225-ALTMAN_9781782549574_t.indd 171

03/05/2017 08:20

172

Handbook of behavioural economics and smart decision-making

Augier, M. and J.G. March (2008), ‘Realism and comprehension in economics’, Journal of Economic Behavior and Organization, 66 (1), 95–105. Augier, M. and S. Sarasvathy (2004), ‘Integrating evolution, cognition and design: extending Simonian perspectives to strategic organization’, Strategic Organization, 2 (2), 169–204. Augier, M. and D. Teece (2007), ‘Competencies, capabilities and the neo-Schumpeterian tradition’, in H. Hanusch and A Pyka (eds), The Elgar Companion to Neo-Schumpeterian Economics, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Augier, M. and D. Teece (2008), ‘Strategy as evolution with design’, Organization Studies, 29 (8–9), 1187–208. Augier, M., J.G. March and A. Marshall (2015), ‘The flaring of intellectual outliers’, Organization Science, 26 (4), 1140–61 Christensen, C. (1997), The Innovator’s Dilemma, Boston, MA: Harvard Business School Press. Cyert, R. and J.G. March (1963), A Behavioral Theory of the Firm, Englewood Cliffs, NJ: Prentice Hall. Dosi, G. (2004), ‘A very reasonable objective still beyond our reach: economics as an empirically disciplined social science’, in M. Augier and J.G. March (eds), Models of a Man: Essays in Memory of Herbert A. Simon, Cambridge, MA: MIT Press. Fang, C. (2013), ‘Behavioral strategy’, in M. Augier and D. Teece (eds), The Palgrave Encyclopedia of Strategic Management, Basingstoke: Palgrave. Gray, A. (1987), ‘The art of command’, Marine Corps Gazette, 71 (10). Gray, A. (1991a), ‘Doing what must be done’, Marine Corps Gazette, 75 (2). Gray, A. (1991b), ‘A message from the Commandant of the Marine Corps’, Gazette, April. Guo, J. and M. Augier (2015), ‘The dynamics of rules, learning and adaptive leadership: inspirations and insights from the United States Marine Corps’, in D. Lindsay and D. Woychenshin (eds), Overcoming Leadership Challenges: International Perspectives, Kingston, Ontario: Canadian Defence Academy Press. Huy, Q.N. (2012), ‘Emotions in strategic organization: opportunities for impactful research’, Strategic Organization, 10 (3), 240–47. Leavitt, H. (1995), ‘The old days, hot groups, and manager’s lib’, Administrative Science Quarterly, 41 (2), 288–300. Leavitt, H. (2007), ‘Big organizations are unhealthy environments for human beings’, Academy of Management Learning and Education, 62 (2), 253–63. Levinthal, D. (2011), ‘A behavioral approach to strategy: what’s the alternative?’, Strategic Management Journal, 32 (13), 1517–23. Lovallo, D. and O. Sibony (2010), ‘The case for behavioral strategy’, McKinsey Quarterly, March, accessed 14 December 2016 at http://www.mckinsey.com/business-functions/strategy-and-corporate-finance/ourinsights/the-case-for-behavioral-strategy. March, J.G. (1965), ‘Introduction’, in J.G. March (ed.), Handbook in Organizations, Oxford: Blackwell. March, J.G. (1991), ‘Exploration and exploitation in organizational learning’, Organization Science, 2 (1), 71–87. March, J.G. and H.A. Simon (1958), Organizations, Oxford: Blackwell. Maslach, D., C. Liu, P. Madsen and V. Desai (2015), ‘The robust beauty of ‘little ideas’, the past and future of a behavioral theory of the firm’, Journal of Management Inquiry, 24 (3), 318–20. Mattis, J. (2006), ‘Commanding General’s command & control (C2) intend’, Marine Corps Gazette, 90 (8), p. 16. Mintzberg, H. (1987), ‘Crafting strategy’, Harvard Business Review, 65 (4), 66–75. Powell, T., D. Lovallo and C.R. Fox (2011), ‘Behavioral strategy’, Strategic Management Journal, 32 (13), 1369–86. Rumelt, R.P., D.E. Schendel and D.J. Teece (1994), ‘Introduction’, in R.P. Rumelt, D.E. Schendel and D.J. Teece (eds), Fundamental Issues in Strategy, Cambridge, MA: Harvard University Press, pp. 1–8. Simon, H.A. (1954), ‘Some strategic considerations in the construction of social science models’, in P. Lazarsfeld (ed.), Mathematical Thinking in the Social Sciences, Glencoe, IL: Free Press, pp. 388–415. Simon, H.A. (1955), ‘A behavioral model of rational choice’, Quarterly Journal of Economics, 69 (1), 99–118. Simon, H.A. (1986), ‘Some design and research methodologies in business administration’, in M. Audet and J.L. Maloutin (eds), La production des connaissances scientifique l’administration, Quebec: Les Presses de L’Universite, pp. 239–79. Simon, H.A. (1991), ‘Organizations and markets’, Journal of Economic Perspectives, 5 (2), 25–44. Simon, H.A. (1993), ‘Altruism and economics’, American Economic Review, 83 (2), 156–61. Simon, H.A. (1997), An Empirically Relevant Microeconomics, Cambridge, MA: MIT Press. US Marine Corps (USMC) (1996), ‘The Marine Corps’ philosophy and principles of training’, in Unit Training Management Guide MCRP 3-0A, November, Department of the Navy, Washington, DC, ch. 1. US Marine Corps (USMC) (1997), Warfighting, MCDP 1, June, Department of the Navy, Washington, DC. Winter, S. (2000), ‘The satisficing principle in capability learning’, Strategic Management Journal, 21 (10–11), 981–96.

M4225-ALTMAN_9781782549574_t.indd 172

03/05/2017 08:20

10 Feminist economics for smart behavioral economics Siobhan Austen

1

INTRODUCTION

Many of the key themes and concerns of feminist economics were summarized by Marianne Ferber and Julie Nelson in their introduction to the ten-year retrospective Feminist Economics Today: Beyond Economic Man. They noted that feminist economics is distinctive in the serious attention it gives to women, its challenging of the common confusion of gender and sex, and its challenging of the economics discipline in masculineonly terms (Ferber and Nelson 2003, pp. 1–2). They highlighted the social construction of both economic behavior and the contemporary discipline of economics. Several of the themes and concerns of feminist economics overlap those of smart behavioral economics. There is, for example, a shared critical perspective on mainstream economic models and a common concern with the particular issue of preference formation. In this chapter we elucidate key themes in feminist economics and highlight its relevance to smart behavioral economics. The discussion in this chapter is organized through the use of concepts drawn from the institutional analysis and design (IAD) framework developed by Elinor Ostrom and her colleagues. Although this framework is most commonly associated with new institutional, rather than feminist, economics, its concept of situated actors is relevant to one of feminist economics’ central themes: of the gendered nature of economic behavior. The IAD framework allows us to trace out the various ways that men’s and women’s preferences are shaped by socially learned expectations associated with being male or female. This is particularly useful for the analysis of observed sex-based differences in behavior, a field of research where the interests of many behavioral and feminist economists appear to intersect. Several additional aspects of the IAD framework, including the influence of mental models on individuals’ processing information, can be used to highlight other shared interests of feminist and smart behavioral economists. This chapter includes a discussion of how ideas about the structure and influence of mental models are in line with the feminist critique of the methods commonly used in studies of sex-based differences in behavior. In doing so, the chapter highlights a further important theme in feminist economics, that science is a socially constructed activity, with the social location, status and gender of scientists and scientific communities all playing a significant role in determining the methods and practices of science (Barker 1999, p. 325). As a meta-theoretical framework, the IAD also has the advantage of facilitating comparisons of different theories and models. This helps us identify some of the particular features of feminist and smart behavioral economics in comparison with other theoretical traditions in economics, including mainstream economics. Toward the end of this chapter 173

M4225-ALTMAN_9781782549574_t.indd 173

03/05/2017 08:20

174

Handbook of behavioural economics and smart decision-making

the discussion focuses on the feminist economics’ critique of the separate/soluble dichotomy in mainstream economics, whereby individuals in market situations are assumed to be atomized, self-interested, with exogenously determined preferences, while individuals in family situations are characterized as connected to each other, altruistic and engaged in a process of shaping preferences. Feminist economists have identified several problems associated with this dichotomy, including barriers to the economic analysis of the unique aspects of women’s lives. Reflecting one of feminist economics’ basic aims, of addressing the realities of women’s lives and their economic and other contributions (Harding 1999, p. 131), an alternative concept is thus advanced; that of ‘individuals-in-relation’. The chapter argues that this concept has the potential to guide future empirical and theoretical studies of men’s and women’s economic behavior. Some prominent feminist economists have already identified the strategic potential to link new institutional economics, which is Ostrom’s field, with feminist theory. Paula England and Nancy Folbre (2003, p. 62), for example, note the relevance of concepts such as endogenous tastes and reciprocity, which feature in new institutional (and smart behavioral) analysis, to notions about the gendered nature of economic behavior. However, many feminist economists contest other core concepts of new institutional and smart behavioral economics, such as the notion of boundedly rational economic agents. Julie Nelson (2003a, 2003b), for example, emphasizes the emotional and subjective aspects of decision-making, albeit the latter is incorporated in behavioral economics. Acknowledging these tensions, this chapter aims to further explore the potential connections between feminist, smart behavioral and new institutional economics. The organization of the chapter reflects these aims. The following section provides a brief introduction to the IAD framework. This is followed with a summary of the key features of feminist economics. Section 4 turns to a key research topic where the interests of feminist and smart behavioral economics appear to intersect, namely, the presence (or otherwise) of differences in the preferences and behavior of men and women. Section 5 explores the issue of (possible) differences in risk aversion in some detail, while section 6 considers the issue of altruistic preferences. Section 7 brings the discussion to a close with a summary of the key themes of feminist economics and some recommendations for smart behavioral economic research.

2

THE INSTITUTIONAL ANALYSIS AND DEVELOPMENT FRAMEWORK

The IAD framework is closely linked to the life work of Elinor Ostrom, the first (and thus far the only) woman to be awarded the Nobel Prize in economics. Ostrom described the IAD as a multi-level taxonomy of the universal components (organized in many layers) that are relevant to regularized social behavior (including interactions in markets, hierarchies and other situations). The broad features of the IAD framework are summarized in Figure 10.1. Of prime importance is the idea of an action arena. This is a ‘social space’ within which ‘participants with diverse preferences interact, exchange good and services, solve problems, dominate one another, or fight’ (Ostrom 2005, p. 14). The focus of IAD analysis tends to be on how the interaction between participants in different action situations is affected by

M4225-ALTMAN_9781782549574_t.indd 174

03/05/2017 08:20

Feminist economics for smart behavioral economics 175 Exogenous variables

Action arena

Biophysical/ material conditions Action situations Interactions Evaluative criteria

Attributes of community Participants Rules Outcomes Source: Ostrom (2005, p. 15).

Figure 10.1

A framework for institutional analysis

the characteristics of the situation itself, including the characteristics of the participants and their positions, preferences, levels of information, approaches to information processing, possible actions and potential payoffs. As Figure 10.1 indicates, interactions within situations lead to particular outcomes, which may be desirable or undesirable. The framework incorporates feedback loops to account for the way in which participants may respond to these outcomes by engaging in efforts to either change or reinforce the structure of the arena (as indicated by the line at the bottom of the figure). An important feature of IAD framework is its emphasis of the context of each action situation. Each action situation is understood to be ‘located’ within an action arena that is affected by a range of exogenous variables, including the attributes of the bio-physical world, the structure of the more general community (including the values generally accepted and the prevailing gender norms within the community), and the current set of rules in use, which will reflect the arena’s historical context. Some aspects of Ostrom’s work address the role of culture in shaping the mental models used by boundedly rational participants in different action situations. Ostrom (2005, pp. 106–7) highlights how the cultural environment, including its prevailing gender norms, shapes participants’ perceptions of what actions are possible, legitimate and desirable (or preferred), and it coordinates the actions of groups of participants. She also asserts that, because mental models are affected by culture, they are likely to be transmitted across generations, producing stability in patterns of behavior and outcomes over time. However, in Ostrom’s analysis, mental models can change/are not constant. Factors such as vividness and salience can be relevant to the type of model that is adopted and can be a source of change or difference in participants’ perceptions and actions (Ostrom 2005, p. 108).

M4225-ALTMAN_9781782549574_t.indd 175

03/05/2017 08:20

176

3

Handbook of behavioural economics and smart decision-making

FEMINIST ECONOMICS

The IAD can be used to explain key features of feminist economics. Feminist economics can be distinguished especially from mainstream economics by its concern for the influence of the contextual environment on the preferences, possible actions, payoffs and outcomes for men and women in market and family situations. The ‘situated’ nature of economic behavior is a fundamental concept in feminist economics. Informed by feminist philosophy of science, feminist economics considers how individuals’ (participants’) economic power, obligations, goals, interests and, ultimately, their economic outcomes are affected by their social roles and relationships, and how these, in turn, are affected by their ascribed social identities, including their gender, race, sexual orientation, and ethnicity. As its name suggests, feminist economics pays particular attention to the gendered nature of the contextual environment, and its implications for men’s and women’s economic roles, actions and outcomes. Gender is distinguished from sex, or the biological differences between males and females.1 It is understood that societies or communities assign different roles, norms, and meanings to men and women and their actions. For example, in most societies individuals are assigned to distinct social roles based on their gender (such as men to ‘breadwinner’ and women to ‘caregiver’ within the family). Men and women are also expected to comply with different norms of behavior (for example, men are expected to be brave, and women modest). Furthermore, psychological traits of masculinity and femininity are linked to gender norms (for example, women are considered virtuous if they comply with a norm of modesty but assertiveness can be considered a vice). Using the language of the IAD, a community’s gender norms affect various elements of the action arenas within which men and women participate. The norms influence the ability of men and women to participate in particular situations, the positions that they can take up within these situations, the range and nature of their possible actions, their access to information, and, potentially, the way they process this information, their payoffs from different actions, and, arguably most importantly, the quality of their outcomes. In turn, the gendered distribution of economic outcomes is likely to be reflected in patterns of action at various levels of the social hierarchy aimed at either entrenching existing norms or challenging them. The way in which these perspectives have influenced the feminist economic analysis of economic behavior and outcomes can be illustrated with examples relating to the labor market. Feminist economic analysis of occupational choice have focused on the impact of social structures and relationships on women and men’s work and career goals (Pujol 1997; Strassman 1997). Studies of the gender pay gap have explored the influence of social norms associated with providing care on the distribution of unpaid household work and, subsequently, on the gendered nature and configuration of work (Folbre 1994). Other studies have examined the failure of apparently gender-neutral market institutions to adequately value the commodities produced by women (Himmelweit 1995; Ironmonger 1996). Importantly, feminist economics’ emphasis on the social construction of behavior and outcomes has also influenced its relationship with the discipline of economics. Feminist economists have identified the influence of a range of gender norms and cultural biases on (using the language of the IAD) the action situations associated with the development and perpetuation of economic theory. These include the tendency for ‘culturally “mascu-

M4225-ALTMAN_9781782549574_t.indd 176

03/05/2017 08:20

Feminist economics for smart behavioral economics 177 line” topics, such as men and market behavior, and culturally “masculine” characteristics, such as autonomy, abstraction, and logic . . . [to] define the field’ (Ferber and Nelson 2003, p. 1). It is important to note that feminist economics challenges these definitions of economics, and devotes energy to exposing the biases in the discipline, in addition to focusing on ensuring that the lives and experiences of women feature in economic analysis, and attempting to remedy the common confusion of gender with sex.

4

FEMINIST ECONOMICS AND THE ANALYSIS OF OBSERVED DIFFERENCES IN THE PREFERENCES AND BEHAVIOR OF MEN AND WOMEN

As can be anticipated, given the description provided in the previous paragraphs, feminist economics’ analysis of observed differences in preferences and behaviors of men and women is distinguished by its focus on their social origins. Observed differences in the preferences and behavior between men and women are, thus, often the starting point of inquiry (into their origins), rather than the end point of an investigation of (apparent) differences in the ‘natures’ of men and women. Feminist economics’ focus on the social origins of observed differences in the preferences and behavior of men and women reflects an argument that preferences and behavior are gendered. For example, boys and girls in most communities are socialized into particular behavioral patterns; trained to different norms of bodily comportment from an early age. Gender norms in Western societies tend to emphasize physicality, aggression and indifference for boys and constraint for girls and, as a result, men and women are likely to find different types of behavior comfortable and achievable with a degree of fluidity. Performing the gendered actions might feel ‘natural’, be associated with positive ‘payoffs’, and result in positive ‘outcomes’. On the other hand, performing actions that are typically assigned to the opposite sex might illicit a sense of novelty, self-consciousness, and awkwardness (negative payoffs and outcomes that are evaluated as poor). There are also important feedback effects, with the experience of poor/good performance influencing the incentive to invest in gendered skills. Gendered socialization can also cause differences in the way men and women process information about a similar situation or arena. This is because representational schemes that are functional for different gender roles can make different kinds of information salient. For example, in traditional domestic settings, women may notice dirt that men do not, ‘not because women have a specially sensitive sensory apparatus . . . [but] because they have a role which designates the females of the household as the ones who have to clean up’ (Anderson 2009, n.p.). These processes may also result in cognitive styles that differ between men and women. For example, the tendency for men to be allocated positions associated with political and economic power that require detachment and control may encourage a cognitive style that is abstract, theoretical, disembodied, emotionally detached and analytical. The tendency for women to be assigned positions associated with the provision of care may encourage a cognitive style that is concrete, practical, embodied, relational and emotionally engaged. (England 2003, pp. 36–8). Patterns of value are also gendered. There is a cultural tendency in most communities

M4225-ALTMAN_9781782549574_t.indd 177

03/05/2017 08:20

178

Handbook of behavioural economics and smart decision-making

to link psychological traits considered ‘masculine’ with virtue when demonstrated by men, and ‘feminine’ traits with virtue when women demonstrate them. This influences the payoffs from different actions that can be performed by men and women and creates incentives for individuals to comply with prevailing gender norms. In academic work situations, for example, the quest for ‘masculine’ prestige may encourage the continued use of ‘masculine’ methods by men, and a rejection of methods associated with femininity or female-dominated fields of enquiry. For example, Nelson (1992) notes how the term ‘hard’ is often metaphorically attached to mathematical and quantitative analysis, and seen as positive and masculine. In contrast, the term ‘soft’ is attached to qualitative methods, is used as a pejorative, and is associated with femininity (see also Nelson 1996, 2003c). More generally, the material and other payoffs associated with different jobs or career paths can vary depending on whether the tasks entailed in the occupational role align with the individual’s gender. For example, men might perceive costs associated with their participation in types of work regarded as ‘feminine’, such as childcare; and women might attach costs to their involvement in types of work regarded as masculine, such as mining. The gendered distribution of power between men and women in many action arenas is an additional important influence on behavior and outcomes. It can cause ‘masculine’ actions to be valorized in particular privileged situations and women’s ability to participate in these situations to be limited. In academic situations, for example, the historical dominance of men has resulted in several formal and informal institutions that value (and thus produce positive payoffs for) ‘masculine’ forms of work and contributions to knowledge. The formal institutions include promotion criteria that emphasize a track record of journal publications and research grants. Often these criteria can only be satisfied by academics who are able to commit long working hours and have uninterrupted tenure, especially in their thirties. Gender differences (and inequity) in outcomes arise if men who conform to a traditional breadwinner role have some ability to achieve success in these situations, while other men and the many women who take on direct care roles in their families find it difficult to achieve positive outcomes. Finally, commonly held ideas about gender affect our perceptions of others (and their actions). A number of studies have demonstrated that the gender of a person affects the costs, benefits and probabilities that others assign to their actions (see, for example, Kahneman 2003). Barbara Reskin (2003), for example, highlighted how in employment situations managers might unconsciously attach certain behaviors, such as reliability or competitiveness, to particular individuals because of their gender. While the managers might consciously reject discrimination, their tendency to rely on familiar social categories might still cause them to think and ultimately act in ways that privilege individuals with a particular gender and disadvantage others. Paula England, Michelle Budig and Nancy Folbre’s (2002) analysis of the labor market outcomes of care workers has similar themes, highlighting how women are commonly perceived to be ‘naturally’ able to accomplish the work involved in caring for children, and for sick and elderly people. This is consequential because it tends to result in judgments of care work as something that does not require skill or effort, contributing to the low-wage outcomes of the many women engaged in care work (see also Austen and Jefferson 2014).

M4225-ALTMAN_9781782549574_t.indd 178

03/05/2017 08:20

Feminist economics for smart behavioral economics 179

5

FEMINIST ECONOMICS AND STUDIES OF SEX-BASED DIFFERENCES IN ATTITUDES TO RISK

We can consider now how feminist economists engage with the growing body of literature on sex-based differences in preferences and behavior. An important part of this literature deals with differences in the risk aversion of men and women. Much of it has been motivated by concern about evidence of an over-representation of women in relatively low-risk forms of assets and in particular occupations. This section provides an overview of these studies before introducing a critical perspective – informed by feminist economics – on the common conclusion that women are more risk averse than men. Studies of sex-based differences in risk preference have featured both studies of investment and insurance decision-making in the presence of risk and lottery or gamble experiments of risk-taking. Studies in the first group have used pension fund data to study the allocation of assets between investment options associated with different levels of risk. Studies in the latter group have included gambling experiments with student participants. They have typically focused on whether (and to what degree) the willingness to take a gamble or invest in a lottery is affected by the level of risk involved. Reflecting the acknowledged importance of both risk and loss aversion, many of these experiments have include scenarios where the possible outcomes are framed in terms of gains, while others are framed in terms of losses. Understandably, the analysis of sex-based differences has focused on the magnitude and statistical significance of observed differences in the choices of male and female participants. Commonly the studies have incorporated controls for other factors that might be relevant to a person’s risk preference, such as age. Several contextual environment experiments have involved students participating in computer-based simulated currency trading and stock market games (where the decision to enter particular currency markets or purchase particular securities involves risk). Apart from testing for sex-based differences in risk preference, these studies also examine the effects of factors such as ambiguity about the game’s outcomes, knowledge of financial markets, and confidence in financial decision-making. A recent study by Alison Booth, Lina Cardona-Sosa and Patrick Nolan (2014, p. 128) compared risk preferences exhibited by participants in mixed versus same-sex groups. Cathrine Eckel’s and Philip Grossman’s (2008, pp. 6–11) assessment is that neither the experimental nor the other studies provide conclusive evidence on the nature or extent of sex-based differences in risk preferences. Apparently this is ‘consistent with results from psychology, which tend to show differences in risk attitudes across environments for a given subject’ (Eckel and Grossman 2008, p. 6). Booth et al. (2014) also conclude that attitudes to risk are influenced heavily by contextual factors. In their study, the female participants’ willingness to invest varied substantially across the same-sex and mixed group settings of their experiments. Despite the mixed evidence from studies of the issue, Eckel and Grossman’s (2008, p. 6) general summary of the results of the gamble experiments is that they ‘suggest greater risk aversion by women’. Rachel Croson and Uri Gneezy (2009, p. 448) are more strident, claiming from their review of the literature on gender differences in preferences that ‘women are indeed more risk averse than men’. This is the starting point for an important review by prominent feminist economist Julie Nelson, who challenges the assumptions, methods and conclusions of behavioral

M4225-ALTMAN_9781782549574_t.indd 179

03/05/2017 08:20

180

Handbook of behavioural economics and smart decision-making

studies of sex-differences in risk preferences. In her 2012 paper ‘Are women really more risk-averse than men?’, Nelson reported the findings of a meta-analysis of published articles on the topic of sex and risk, including the studies canvassed by Croson and Gneezy (2009), Eckel and Grossman (2008), and Booth et al.’s more recent work. It examined the available evidence on the quantitative magnitudes of the differences between the average level of risk aversion observed for men and women, and the extent to which the observed distribution of risk aversion varies between male and female samples. In doing so, Nelson attempted to redress the tendency for behavioral studies to rely on measures of statistical significance in their judgments of the significance of observed differences in the risk aversion of men and women: ‘In the gender-and-risk literature, as in other literatures, however, judgments of “significant difference” are generally based on statistical significance alone. Discussions of the absolute size of the difference, much less its possible implications for society or policy, are rare’ (Nelson, 2012, p. 6). Nelson’s ‘alternative’ approach to assessing the evidence on gender differences in risk preference produced some revealing insights. Only 25 percent of the studies that she reviewed identified a difference favoring lower male risk aversion of more than half a standard deviation. Only two studies found a difference of more than one standard deviation of difference. Four studies identified differences that are statistically significant in the direction of greater female risk-taking. Thus, in Nelson’s assessment, an appropriate summary of the results of studies of sex differences in risk preference is that they point to ‘a statistically significant difference in mean risk aversion between men and women, with women on average being more risk averse’ (Nelson 2012, p. 2). This stands in important contrast to Croson’s and Gneezy’s (2009, p. 448) claim that ‘women are indeed more risk averse than men’. The latter statement implies that risk preference is a stable characteristic of people defined by their sex, and that a lower risk preference is universally true for every individual member of the class ‘women’ (as compared with ‘men’). As Nelson notes (2012, p. 3), ‘this exceedingly strong implication is not likely intended by those who write such statements’, noting that ‘just one example of a cautious man and a bold woman disproves it’. However, she goes on to draw our attention to how the more probable meaning of the statement (that women are, or are disposed to be more risk averse by virtue of being a woman) is also problematic. ‘In the current example, the statement would imply that greater risk aversion is an essential characteristic of womanliness – or, by parallel reasoning, that greater risk seeking is an essential characteristic of manliness’ (Nelson 2012, pp. 3–4). Reflecting themes in feminist economics introduced earlier, Nelson rejects the notion that risk aversion is a sex-linked ‘trait’ and, instead, locates the source of observed sexbased differences in risk preferences in patterns of gendered socialization and power. As such, observed sex-differences should not be the end-point of inquiries into risk preferences but, rather, the stimulus for further inquiry into the gendered norms and other institutions that influence men’s and women’s attitudes to risk. This potentially creates an important role for future studies of the issue in different cultural contexts. It is important to note that Nelson’s critique of the gender and risk literature also relates to another theme of feminist economics that was highlighted in earlier sections of this chapter, namely, its critical perspective on possible biases within the economics discipline. Nelson highlights the influence of mental models on the work of researchers; of how the inferences we derive from empirical data are likely to reflect ‘the structure of

M4225-ALTMAN_9781782549574_t.indd 180

03/05/2017 08:20

Feminist economics for smart behavioral economics 181 our inside worlds – that is, of evolved, developmental human cognition’ (Nelson 2012, p. 5). She notes that the models that we use are significantly influenced by our experiences of and beliefs about men and women and, thus, perhaps, it is not surprising that many studies ‘leap’ from evidence of a statistically significant difference in average levels of risk aversion to conclusions about men’s and women’s natures. That is, researchers are (as are others) prone to ‘confirmation bias’, whereby we tend to more readily absorb information that conforms to our pre-existing beliefs, including our beliefs about the ‘nature’ of men and women. This can be an important (and potentially dangerous) source of distortion in our work. For example, as Nelson points out, if we report a statistical significance in risk aversion that is not substantially significant we can reinforce common stereotypes about men’s and women’s ‘natures’. To the extent that this diverts attention from cultural and other institutional sources of differences in behavior, it can be an obstacle to the design of appropriate policy measures aimed at improved gender equity. That is, there is a risk that research into sex-based differences may contribute to the perpetuation of gender inequality, rather than help to reduce it.

6

FEMINIST ECONOMICS AND STUDIES OF ALTRUISM

Similar themes are apparent in feminist economic analyses of altruism. A number of studies of differences in altruism between men and women have been undertaken, motivated by a sense that they could lead to different patterns of charitable giving, bargaining, and household decision-making. As such, gender differences in altruism are potentially consequential for outcomes across a number of different market and family situations. In their 2008 paper ‘Altruism in individual and joint-giving decisions: what’s gender got to do with it?’, Linda Kamas, Anne Preston and Sandy Baum reviewed the experimental evidence on sex-based differences in altruism, and contributed the findings of their own study of the issue. The first part of this section draws heavily on their summary of the relevant literature. Broader feminist economic perspectives on the topic of altruism are considered in the latter part of this section. Here the focus of the discussion turns away from the question of whether women are more or less altruistic than men and toward the general importance attached to altruistic (and other other-regarding) preferences by feminist economists. As Kamas et al. (2008) describe, experimental studies of altruism typically assess participants’ willingness to sacrifice their own outcomes to improve the well-being of another either by using a dictator, ultimatum, public good or investment or trust game. The authors favor a dictator ‘game’, where the dictator decides how to allocate a sum of money between himself or herself and another player, on the grounds that it has the greatest ability to separate the effects of altruistic preferences on behavior from the effect of risk and competition. In their dictator games, the recipient of the money is a charity. Kamas et al. (2008) report findings from their own experiments that indicate significant gender differences in altruistic behavior, with women giving significantly more, on average, than men. Their finding was generally suggestive of a pattern of difference similar to that observed by James Andreoni and Lise Vesterland (2001, p. 293). However, several studies included in their review, such as those by Martin Dufwenberg and Astri Muren (2006), reported higher levels of generosity by men. Kamas et al. (2008, p. 25) conclude that in

M4225-ALTMAN_9781782549574_t.indd 181

03/05/2017 08:20

182

Handbook of behavioural economics and smart decision-making

the experimental literature there is no consensus on the substantive significance of gender differences in altruistic behavior. The explanation offered for the mixed results on altruistic preferences by Kamas et al. (2008, p. 25) centers on the differences in the experimental settings of the various studies. These differences relate to both the type of game used as well as ‘the experimental design or context . . . the framing of the experiment, the degree of anonymity, the subject population, and/or the manner in which the participants are chosen’ (Kamas et al. 2008, p. 25). Whether men or women are identified as the ‘more generous sex’ apparently varies with the price of giving, the degree of anonymity, and the possibility of reciprocity (see also Cox and Deck 2006). Several studies conclude that the gender of the recipient of an altruistic act also affects gift-giving. The study conducted by Kamas et al. (2008) found that gift-giving increased in mixed-sex team situations, and especially when the participants were able to negotiate a common gift. Kamas et al. (2008: 44) acknowledge (albeit in a footnote) that they do not provide an in-depth explanation of the causes of observed sex-based differences in altruism. However, they do allude to a number of influences stemming from the social environment, and these are potentially reflective of the processes of gendered socialization that were noted in earlier sections of this chapter. For example, their explanation for the observation of higher levels of gift-giving by mixed-sex teams includes a role for social information (about the social norm for gift-giving) and social image (a desire to be considered favorably by others) (Kamas et al. 2008, p. 27). As a reviewer of their paper apparently observed, it may also be possible that women are socialized to be more giving than men, and women’s identification as mothers or caregivers may lead to altruistic acts (Kamas et al. 2008, p. 45). The authors also acknowledge the possibility that as experiments of this type are conducted beyond the confines of the current set of developed western countries, the impacts of cultural and sociological forces on gender differences in altruism will become more apparent. The interpretation of experimental evidence offered by Kamas et al. contrasts that provided by Andreoni and Vesterland (2001). The latter appear to succumb to the various pitfalls involved in assessing sex-based differences in behavior that were noted by Julie Nelson. They infer from their experimental evidence (of a statistically significant gender difference in the observed levels of gift giving across 142 students in eight experimental settings) that ‘when altruism is expensive, women are kinder, but when it is cheap, men are more altruistic’ (Andreoni and Vesterland 2001, p. 293). The focus of their results is on the statistical significance of observed differences in means, rather than on the magnitude of these differences or the distribution of results. This is problematic when the observed gender gap is relatively small. Generally, the Andreoni and Vesterland study ‘essentializes’ the nature of men and women, reinforces common stereotypes, and fails to acknowledge the preferences of men and women who do not conform to group averages. It provides no insights into the possible sources of observed gender differences in altruism, and its discussion of policy implications is limited to a consideration of the consequences of gender differences in charitable gift giving and restaurant tipping. A more important feminist economic discussion of altruism shifts the focus of attention away from possible sex-based differences and toward the general importance of altruistic preferences (for men and women). This discussion forms part of a broader critique of the theoretical structure of mainstream economics by feminist economics. The critique

M4225-ALTMAN_9781782549574_t.indd 182

03/05/2017 08:20

Feminist economics for smart behavioral economics 183 argues against a narrow specification of the sources of individual motivation and argues instead for specifications that take account of various sources of motivation, including altruistic preferences, and the social influence on these motives. The critique of mainstream economic theory, developed by Paula England (see, for example, England 2003) focuses, first, on its assumption that individuals in market situations are atomized, self-interested, and have preferences no one can change. This is contrasted against the assumption that individuals in family situations are connected to each other, with interdependent preferences and engaged in a process of shaping the preferences and values of the young. As England explains, while the theory’s analysis of market situations features a ‘separative’ view of the self that presumes, amongst other things, that individuals lack sufficient emotional connection to others to feel any empathy – or to be altruistic, a ‘soluble’ self is assumed in its analysis of family situations, allowing both empathy and altruism to influence behavior and outcomes. England (2003, pp. 36–40) highlights the various problems with this theoretical structure. These include problems caused by incorrectly assuming pure self-interest in market situations, and by over-emphasizing the extent of empathy and altruism in family situations. Additional problems arise from the separative or soluble dichotomy and its relationship to gender dichotomies in western thought. England notes that in simple (sexist) formulations of western thought, men are seen as naturally separative, autonomous and individuated, while women are seen as naturally soluble, connected and yielding. Separation has been valorized in western thought, at least for men, while connectedness has been devalued. As a consequence, writing in economics and other fields has ‘failed to recognize that men are not entirely autonomous . . . whilst women’s nurturing work was taken for granted and excluded from . . . theory’ (England 2003, p. 38, original emphasis). These observations link to several of the themes of the feminist critique of mainstream economics noted in earlier sections of this chapter. For example, the valorizing of separation – and market situations – has contributed to a failure to adequately recognize the experiences and contributions of women. The gendered nature of the separative–soluble dichotomy helps to explain common confusion of gender with sex; of the tendency to identify ‘essential’ differences between men and women. Feminist economic analysis suggests that it is appropriate to assume that both male and female participants in market and family situations will have both ‘separative’ and ‘connective’ qualities; and that these qualities will have both positive and negative aspects. Core concepts, therefore, are of ‘individuals-in-relation’ or ‘relational autonomy’ (England 2003, p. 39). These concepts have obvious relevance for the feminist economic analysis of altruism (and for the analysis of the related concepts of cooperation and strong reciprocity). Altruism is potentially relevant to the preferences and behaviors of men and women. It should be considered as a source of motivation in market and other situations. There is a need for more theoretical and empirical studies of men’s and women’s altruistic (and other other-regarding) preferences. We need additional insights to how these preferences interact with other preferences, including self-regarding preferences; how they are shaped; and how they are influenced by different aspects of the contextual environment, such as general levels of altruism in the surrounding community. Taking this approach, studies of gender difference in altruism would ideally focus on women’s traditional association with the family sphere; how, therefore, they have been traditionally assumed or expected to be

M4225-ALTMAN_9781782549574_t.indd 183

03/05/2017 08:20

184

Handbook of behavioural economics and smart decision-making

altruistic; and the consequences of these assumptions or expectations for their observed behavior and their economic outcomes. This approach features in Nancy Folbre’s (1995) analysis of caring labor, a topic that calls into question the role and impact of altruistic preferences and which also has relevance for a range of important policy issues, including the future quality and cost of child and elder care, and pay equity. Folbre highlights that caring relies on a range of motivations, including reciprocity, altruism and responsibility. She also emphasizes that these motivations are constructed in a social environment. Folbre recognizes that caring labor is associated with tasks that women often specialize in, such as mothering. However, she also emphasizes that caring labor can (and is) undertaken by both men and women, and that it occurs in both family and market situations. Folbre is particularly concerned with the interactions between the different sources of motivation for caring labor. She acknowledges the role of altruism but notes that it interacts with long-run reciprocity and the fulfillment of obligation or responsibility. As such, she describes carers as being both ‘connected’ (through their altruistic preferences) and ‘separate’ (in their concern for their individual payoffs). In Folbre’s analysis, individuals may provide care out of a sense of affection or responsibility for others, but their motivation to care is likely to also be influenced by long-run expectations of reciprocity of either tangible or emotional services. Care motives are also described as being dependent, in part at least, on the level of altruism and reciprocity within the surrounding community. In turn, social norms are ascribed a potential role in helping prevent a coordination (or caring) failure. Folbre accounts for gender differences in caring labor in a variety ways. First, social norms, as well as notions of obligation, are gendered. As such, they result in a different structure of payoffs for men and women involved in caring and other roles. The historical context is also important, with women’s traditional roles in caring for others potentially affecting the nature and extent of their altruistic ‘preferences’ and, thus, their evaluation of caring and other roles. The outcomes from caring situations, described by Folbre, are often not positive for women. Caring labor is typically low paid and aspects of the work – including the responsibility, skill and effort involved are not generally reflected in wage and other outcomes. Given that at least part of the motivation for caring labor is self-interested, the low wages place at risk the ongoing supply of care. An appropriate policy response to this dilemma would be to improve the ‘rate of return’ from caring labor, regulating wage outcomes to ensure that low wages do not crowd out care motives. The contrast between Folbre’s analysis of caring labor and that offered by mainstream economists is stark. The latter tend to rely on the notion of non-pecuniary preferences, which are typically ‘lumped together’ and modeled as exogenously (and, presumably, biologically) determined. The independent determination of motivation in these models results in a prediction that if an individual gains positive utility from caring he or she will be willing to trade-off lower wages to ‘indulge’ this preference. Wage regulation is rejected on the assumption of the absence of social or other barriers to mobility. Indeed, in some analyses, higher wages for carers are viewed as a threat to caring labor – based on a belief that higher wages would encourage the participation of individuals with less altruistic preferences (Heyes 2005).

M4225-ALTMAN_9781782549574_t.indd 184

03/05/2017 08:20

Feminist economics for smart behavioral economics 185

7

CONCLUSIONS

This chapter has attempted to convey key themes in feminist economics of relevance to smart behavioral economists. It takes a novel approach to this task by structuring the discussion using concepts and terms drawn from the institutional analysis and design framework, developed by Elinor Ostrom. The framework was used to identify the distinctive features of feminist economics, including, perhaps most importantly, the emphasis it places on understanding the social influences on individual preferences, actions and outcomes. The ‘situated’ nature of economic behavior is a fundamental concept in feminist economics. Feminist economics pays particular attention to how individuals’ economic power, obligations, goals, interests and, ultimately, their economic outcomes are affected by their social roles and relationships, and how these, in turn, are affected by their ascribed social identities, including their gender. Gender is distinguished from sex, or an individual’s biological identity of being male or female. It refers to socially learned expectations and behaviors associated with being male or female. The chapter has highlighted the various ways in which the concept of a ‘situated actor’ influences feminist economists’ engagement with topics in smart behavioral economics. It has demonstrated that feminist economists tend to take a cautious approach to the analysis of observed differences in behavior between men and women. While feminist economists do not deny that these differences exist, they emphasize the need to explore their sources in the social environment, and they sound a strong warning about the dangers of drawing inferences about the essential ‘natures’ of men and women from these differences. The concept of a situated actor is also apparent in feminist economists’ perspectives on the theories and methods used in the analysis of economic behavior. The chapter highlighted the feminist perspective that academic inquiry itself is a fundamentally social process. As such, participants in academic work situations are subject to biases that arise from their own (essentially limited) set of experiences, including their experiences of and beliefs about men and women. This can be an important source of error in academic work, potentially contributing to a reinforcement of stereotypes about men and women, rather than promoting greater gender equity. An important concern of feminist economists is to minimize these sources of error by training economists and promoting the adoption and enforcement of methodological principles designed to check the influence of gender bias. The chapter also emphasized the feminist economics’ critique of the separate–soluble dichotomy in mainstream economics. The mainstream assumption is that individuals in market situations are ‘separate’, that is, essentially atomized, self-interested, with preferences no one can change. In contrast, this theory assumes that individuals in family situations are ‘soluble’, that is, connected to each other, with interdependent preferences, and engaged in a process of shaping the preferences and values of the young. The critique observes that the separate–soluble dichotomy that characterizes mainstream economics has a strong gender dimension, with market situations commonly associated with the activities of men, while family situations are commonly linked to the activities of women. This has various negative impacts, including the tendency for the lives and experiences of many women to be excluded from economic analysis, and the ‘essentializing’ of men’s and women’s natures (‘men are self-interested and autonomous, while women are caring and dependent’). The approach has limited the analysis of the range of motivations

M4225-ALTMAN_9781782549574_t.indd 185

03/05/2017 08:20

186

Handbook of behavioural economics and smart decision-making

(self-interested and other-regarding) affecting the behaviors of men and women in market and family situations. Feminist economics offers an alternative concept to guide future empirical and theoretical studies of behavior: that of ‘individuals-in-relation’. It conveys that men and women, in market and non-market situations, are likely to be influenced by self-interested and other-regarding preferences. The recommendation is for smart behavioral economics (smart decision-making) to continue to pursue studies of how different sources of motivation interact with each other; how preferences are shaped; and how they are influenced by different aspects of the contextual environment, such as general levels of altruism or reciprocity in the surrounding community. Further studies of gender differences in behavior are needed, but they should focus on how, for example, prevailing gender norms affect the positions men and women can participate in, the payoffs and value attached their alternative actions, and, ultimately, their economic outcomes.

NOTE 1. Increasingly, the analysis acknowledges multiple genders. This takes into account individuals whose gender identity differs from their biological sex.

REFERENCES Anderson, E. (2009), ‘Feminist epistemology and philosophy of science’, in E.N. Zalta (2012) (ed.), The Stanford Encyclopedia of Philosophy, Fall, accessed 17 December 2016 at https://plato.stanford.edu/archives/spr2009/ entries/feminism-epistemology/. Andreoni, J. and L. Vesterland (2001), ‘Which is the fair sex? Gender differences in altruism’, Quarterly Journal of Economics, 116 (1), 293–312. Austen, S. and T. Jefferson (2014), ‘Economic analysis, ideology and the public sphere: insights from Australia’s equal remuneration hearings’, Cambridge Journal of Economics, October, doi:10.1093/cje/beu042. Barker, D.K. (1999), ‘Feminist philosophy of science’, in P.A. O’Hara (ed.), Encyclopedia of Political Economy, London: Routledge, pp. 325–7. Booth, A., L. Cardona-Sosa and P. Nolan (2014), ‘Gender differences in risk aversion: do single-sex environments affect their development?’, Journal of Economic Behavior and Organization, 99 (March), 126–54. Cox, J. and C. Deck (2006), ‘When are women more generous than men?’, Economic Inquiry, 44 (4), 587–98. Croson, R. and U. Gneezy (2009), ‘Gender differences in preferences’, Journal of Economic Literature, 47 (2), 448–74. Dufwenberg, M. and A. Muren (2006), ‘Generosity, anonymity, gender’, Journal of Economic Behavior and Organization, 61 (1), 42–9. Eckel, C. and P. Grossman (2008), ‘Men, women and risk aversion: experimental evidence’, in C. Plott and V. Smith (eds), Handbook of Experimental Economics Results, vol. 1, New York: Elsevier, accessed 14 January 2015 at http://papers.ssrn.com/sol3/papers.cfm?abstract_id51883693. England, P. (2003), ‘Separative and soluble selves: dichotomous thinking in economics’, in M. Ferber and J, Nelson (eds), Feminist Economics Today: Beyond Economic Man, Chicago, IL: University of Chicago Press, pp. 33–60. England, P. and N. Folbre (2003), ‘Contracting for care’, in M.A. Ferber and J.A. Nelson (eds), Feminist Economics Today: Beyond Economic Man, Chicago, IL: University of Chicago Press, pp. 61–79. England, P., M. Budig and N. Folbre (2002), ‘Wages of virtue: the relative pay of care work’, Social Problems, 49 (4), 455–73. Ferber, M. and J. Nelson (2003), ‘Introduction: Beyond Economic Man, Ten Years Later’, in M. Ferber and J.  Nelson (eds), Feminist Economics Today: Beyond Economic Man, Chicago: The University of Chicago Press, pp. 1–33. Folbre, N. (1994), Who Pays for the Kids? Gender and the Structures of Constraint, Routledge: London.

M4225-ALTMAN_9781782549574_t.indd 186

03/05/2017 08:20

Feminist economics for smart behavioral economics 187 Folbre, N. (1995), ‘Holding hands at midnight: who pays for caring labour?’, Feminist Economics, 1 (1), 73–92. Harding, S. (1999), ‘The case for strategic realism: a response to Tony Lawson’, Feminist Economics, 5 (2), 127–33. Heyes, A. (2005), ‘The economics of vocation or “why is a badly paid nurse a good nurse?”’, Journal of Health Economics, 24 (3), 561–9. Himmelweit, S. (1995), ‘The discovery of unpaid work: the social consequences of the expansion of “work”’, Feminist Economics, 1 (2), 1–19. Ironmonger, D. (1996), ‘Counting outputs, capital inputs and caring labor: estimating gross household product’, Feminist Economics, 2 (3), 37–64. Kahneman, D. (2003), ‘A perspective on judgment and choice: mapping bounded rationality’, American Psychologist, 58 (9), 697–720. Kamas, L., A. Preston and S. Baum (2008), ‘Altruism in individual and joint-giving decisions: what’s gender got to do with it?’, Feminist Economics, 14 (3), 23–50. Nelson, J. (1992), ‘Gender, metaphor and the definition of economics’, Economics and Philosophy, 8 (1) 103–25. Nelson, J. (1996), Feminism, Objectivity, and Economics, New York: Routledge. Nelson, J. (2003a), ‘Confronting the science/value split: notes on feminist economics, institutionalism, pragmatism and process thought’, Cambridge Journal of Economics, 27 (1), 49–64 Nelson, J. (2003b), ‘Once more with feeling: feminist economics and the ontological question’, Feminist Economics, 9 (1), 109–18 Nelson, J. (2003c), ‘Separative and soluble firms: androcentric bias and business ethics’, in M. Ferber and J. Nelson (eds), Feminist Economics Today: Beyond Economic Man, Chicago, IL: University of Chicago Press, pp. 81–100. Nelson, J. (2012), ‘Are women really more risk-averse than men?’, Global and Development Institute Working Paper No. 12-05, Tufts University, Medford, MA. Ostrom, E. (2005), Understanding Institutional Diversity, Princeton, NJ: Princeton University Press. Pujol, M. (1997), ‘Broadening economic data and methods’, Feminist Economics, 3 (2), 119–20. Reskin, B. (2003), ‘Rethinking employment discrimination and its remedies’, in M. Guillen, R. Collins, P. England and M. Meyer (eds), The New Economic Sociology, New York: Russell Sage, pp. 218–44. Strassman, D. (1997), ‘Expanding the methodological boundaries of economics’, Feminist Economics, 3 (2), vii–ix.

M4225-ALTMAN_9781782549574_t.indd 187

03/05/2017 08:20

11 How regret moves individual and collective choices towards rationality Sacha Bourgeois-Gironde

1

IDENTIFYING REGRET

At first sight, regret, which is semantically very ambiguous and confined to psychological notions such as disappointment, remorse, or even repentance, is not the ‘smartest’ state of mind we can entertain in the course of our life. It is the feeling that we have been less than optimal in particular situations or longer stretches of time, where we could have acted on better lines, and we know it or realize it now, while in the throes of that negative emotion. It is ironical that the author most associated with a vindication of the virtuous role of emotions in decision-making in contemporary popular neurobiology, namely, Spinoza as interpreted by Damasio (Damasio 2004), emphasizes the irrationality of regret: ‘Repentance is not a virtue, that is, it does not arise from reason; instead, he who repents what he has done is twice wretched, that is, lacking in power’ (Spinoza 1677 [1996], p. 4). The individual not only has been practically suboptimal but a negative feeling accrues and prolongs her helplessness. Regret, thus viewed, does not represent the emotional side of a retrospective and corrective, cognitively driven process. To refute that view and make apparent how regret can help optimize decision-making processes rather than reinforce past failures, we need, first, to decompose between backward-looking and forwardlooking aspects of regret and, second, to better understand the articulation between its emotional and cognitive components. Taken altogether, these distinctions will permit us to define regret as a biologically anchored learning mechanism liable to direct decisions along an optimal path. In the absence of this psychological device, which apparently consists of dumbly lamenting over spilled milk, our decisions would really be dumber, as it is likely that we would not become aware, or at least sensitive, to our past mistakes. As for the emotional part, we are, by means of emotions such as regret, sensitized to suboptimality. However, the forward-looking and cognitive component is what makes regret a smart mechanism, and ourselves smarter by the same token, to the extent that the aversive aspect of felt pangs of regret drives up regret-avoidance behavior over the repetition of similar decision situations and, through the conscious anticipation of the emotional impact of comparatively bad or good consequences of a choice over several available options, it generalizes to a whole range of repeated or even one-off decisions. This pervasiveness of regret and its inherent cognitive nature triggered attempts at incorporating it in formal decision-theoretical frameworks. Regret, with its conceptual correlate disappointment, is one of the rare emotions which have been singularly identified and incorporated in formal decision-theory. Elster (1998) extends a long argument about how economists fail to account for specific emotional mechanisms. Most of the discussion about the relationship between emotions and rationality has been couched in general terms, falling short of phenomenological 188

M4225-ALTMAN_9781782549574_t.indd 188

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 189 characterization and a specific analysis of emotional mechanisms potentially associated with behavioral optimization in choice contexts. This amounts to deliberate ignoring or downplaying of game-theoretical models of guilt or envy and regret-based decisiontheory that were already developed in the 1980s and 1990s. However, this also triggers attention and effort towards the possibility to more finely integrate the psychological description of emotional mechanisms – including phenomenological, behavioral and neurobiological levels – and decision-theoretical normative issues. At the same time, this reveals the tension between descriptive and normative relative imports in accounting for human rationality. One obvious way to alleviate this tension is to uncover some normative aspects of emotions. What has distinguished regret among other emotions is that it definitely bears a cognitive component. It is an articulate, sensitive state of mind. Regret consists in feeling negatively the comparison between two states of affairs. This involves a series of subperformances, alternatively pointing to cognitive processes and hedonic states and, on the whole, their integration into a unified affective state. Imagine you have forgone an optimal option and now find yourself in a position to compare what you have got with what you could have got; this implies your ability to: ● ● ●

engage in counterfactual thinking; ascribe a value to a present state of affairs and to a counterfactual one; and compare actual and counterfactual values.

This becomes even more intricate when we envision anticipated regret. Anticipated regret is the heuristic we need to incorporate into smart regret-based behavioral decision-theoretical models, as, on its basis, individuals will tend to avoid negative future consequences. From a cognitive standpoint, the smartness then required amounts to the following: ● ● ● ●

representation of possible future states of affairs; hedonic simulation of the future (also known as mental time travel); ascription of value to future alternative states under different courses of actions taken; and comparison of utilities derived from possible alternative states.

The integration of these cognitive sub-processes functionally yields a future utilityweight in present decisions. Specific cerebral mechanisms underpinning each of these processes and their functional integration have been described (Gilbert and Wilson 2007; Boyer 2008; Schacter et al. 2008). It is worth noting from the outset that the way regret has entered decision-theory in the past three decades, as a rival to a standard model of expected utility theory, relies on a much more stylized version than that which cognitive and affective neuroscientists probe into our brains. However, we think there is enough overlap between the working definitions of regret respectively adopted in psychology and in economics (especially in experimental decision-theory) to allow a fruitful hybridization of the approaches and uncover regret as a smart mechanism both from a psychologically realistic and a formally tenable perspective. The main distinctions to be drawn in both contexts are in terms of available information, responsibility, and valence (in that order),

M4225-ALTMAN_9781782549574_t.indd 189

03/05/2017 08:20

190

Handbook of behavioural economics and smart decision-making

Table 11.1

Varieties of decision-theoretical emotions

Information Responsibility Valence

+

−

Regret Regret or remorse Rejoicing

Disappointment Deploring Regret

rather than in terms of mental abilities to project ourselves into alternatives and experience feelings about them (Zeelenberg 1999). Under that most basic view, to experience regret, individual X needs to know that her action A has led to consequence C and to further know that, had she taken another action, A’, the consequence would have been C’. Now, the difference, in terms of utility, u(C) – u(C’), if negative, is what regret amounts to. If we weaken or modify an element of this sequence, a typology of alternative ‘decisiontheoretical emotions’ derived from this definition of regret is generated (see Table 11.1). More forms of regret and related emotions can be encompassed if we consider the timing of the emotions with respect to the unfolding of the action; in particular regret can be post hoc, ex ante or online (we sometimes regret what we are doing while doing it). What has interested decision-theorists is the learning process associated with regret which ensures the transition from post hoc regret to anticipated regret and bias decisions to a predicted error minimization procedure (also known as minimax-regret). Young (2004) excellently accounts for this transition from experienced regret to anticipated regret. That transition goes from a backward-looking emotional state to a cognitive anticipation of future consequences and is at the core of what makes regret a potential smart mechanism in decision-making. In a past-payoff decision-model based on the elimination of regret, a single agent who faces a t-times repeated decision problem will try to eliminate or minimize regret over the period [1 − t]. The agent makes her choice over a set of actions A and obtains a payoff following each action taken. We do not have to consider the time that elapses between the action and its feedback, even though of course in a richer regret-based model this parameter is likely to play a role. The feedback of the action is not in itself the parameter to elicit regret: it has to be contrasted with the expected outcome or, differently, with the information about what she could have obtained had she taken another course of action, say, a. Let us suppose that second comparison is made possible at the end of period [1 − t]. As of period t, then, the agent’s regret for not having chosen action A is defined as the difference between two terms: the average payoff she would have obtained had she chosen a in periods 1 through to t, and the average payoff she actually obtained during those periods. Young defines an optimal regret-based strategy for this repeated decision problem: it satisfies a no-regret criterion if it ensures that for any sequence of action feedbacks, the agent’s regret for each of his actions becomes non-positive as t approaches infinity, all other things being equal. Some notable contributions made this stylized typology of regret and related emotions fit the investigation of underlying neurobiological mechanisms. We discuss them in the next section. By borrowing experimental paradigms from experimental decision-theory, neuroscientists indeed uncover specific mechanisms associated with the way repeated emotions are optimally biased by regret aversion. One of our concerns is to assess to

M4225-ALTMAN_9781782549574_t.indd 190

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 191 what extent these uncovered neurobiological mechanisms correspond to the more stylized notion of regret that has been integrated in some decision-theoretical frameworks in the view of formalizing how anticipated regret-based learning happens to be an optimal decisional heuristic.

2

THE NEUROBIOLOGICAL BASIS OF REGRET

Damasio (2008) has largely documented the effect of some bilateral lesion of the ventral medial prefrontal cortex on regret aversion driven decision-making. He and his colleagues have studied several cohorts of patients who, post-lesion, display deep difficulties in planning their life, establishing friendly associations, finding business partners, avoiding financial losses, keeping a stable occupation, refraining from certain impulsions and profanities. These behavioral patterns are in neat contrast with these subjects’ profiles before the lesion. However, the lesion has left unscathed the cognitive abilities: normal intelligence, linguistic understanding, and memory, normal visual, hearing and tactile acuity (Bechara et al. 2000). What essentially differs before and after the brain accident is the decision-making process which has become long and intricate, with all sorts of alternatives entertained and pondered, leading to choices which are found difficult to make and most often disadvantageous. When such choices are repeated, no learning apparently occurs, no lesson of bad past experiences is taken and consistency in suboptimal behavior is observed. This main difference affecting decision-making has been probed on a specific task, called the Iowa gambling task (IGT). Behavioral responses on this task were associated with a measure of variation of the physiological skin conductance response (SCR) and cardiac pace, considered as the main somatic markers of emotional states.1 In the IGT, participants face four decks of cards from which they can freely pick cards one after another. They know nothing about the relative value of the decks except that they are not equivalent. So, they must learn about this difference, which amounts clearly to a learning task. Unbeknown to them, the game will stop after 100 cards have been turned up. The decks differ in the variance between gains and long-term profit they present. Decks C and D present a small spread of value (small gains and small losses) to an extent that makes them profitable if the participant sticks to them. Decks A and B embed a few attractive large gains but also large losses such that sticking to them surely makes the participant lose his or her initial endowment. Control subjects, after having explored for a short while Decks A and B, turn and keep on playing on Decks C and D. Ventral medial patients stay on Decks A and B and ask for credit when their endowment happens to be exhausted in the middle of the game. Besides this behavioral particularity, patients exhibit a special physiological pattern compared with controls: they fail to exemplify the somatic markers that induce healthy subjects to avoid negative consequence choices in favor of optimal decision-making. This neurobiological grounding of corrective anticipatory behavior seems a key element of more explicit and cognitively driven regret-based learning processes. Ventral medial patients seem unable to relate together current actions, expectations and past failures. Only the internal building of this temporal binding between past, present and future is likely to sustain the process of regret-based learning guiding repeated decisions towards optimal payoffs. They also seem to be unable to evoke emotions and to make

M4225-ALTMAN_9781782549574_t.indd 191

03/05/2017 08:20

192

Handbook of behavioural economics and smart decision-making

them efficient mechanisms in action. Indeed, they are not deprived of all emotions and can abstractly judge that a situation (or an image or a fantasy) with a horrid component should trigger in them some felt negative emotion, but this emotion does not occur. They present the converse dissociation that the neuropsychologist Brenda Milner (1962) had observed with dorsal-lateral and hippocampal patients, who do the right thing but cannot verbalize it. The generation of this signal and its incorporation into executive function is an efficient mediation in decision-making. This signal amounts to an internal bias unconsciously diverting decisions from bad outcomes before it reaches the consciousness threshold, although this point has been contradicted by several authors (see, in particular, Persaud et al. 2007). Whether or not this signal is efficient before or after it reaches the individual’s consciousness, it is instrumental in defining anticipated regret as a mechanism whose cognitive and behavioral role is not independent of an embodied emotional alert system. The underlying cause of the decisional deficiency of ventral-medial patients is therefore, according to Damasio, a failed activation of covert emotional signals which are supposed to bias decision in a favorable direction in the long run. Damasio has, through his implementation of the IGT and the implicit role of SCR, made popular the view that the conscious explicit knowledge of preferences and choice-criteria is not enough to generate optimal decision behavior. Decision processes in the brain are distinct from other cognitive abilities governed by frontal lobes (for example, work-memory and responses inhibition) in the sense that they directly plunge in visceral pre-executive mechanisms associated with emotional autonomous arousal. More has to be said on how a post hoc experience of disappointment can change smoothly into an ex ante regret-aversive optimal behavior. We find some possible explanation when considering Damasio’s distinction between two types of internal signals (Damasio 2008). First, a genuine somatic loop generating ‘primary’ signals is triggered when individuals face choices under uncertainty and ambiguity. Second, a para-somatic loop can be activated by a mental representation of somatic states (consisting in cognitive states in which the subject simulates his being in a primary emotional state). This loop is selected by choices which, through previous repetitions of the similar situation, looms as quasi-certain or inevitable consequences. The transition from one loop to the other is driven by regret learning and, as we see, amounts to a deep change in the perception of the decision-theoretical context, from uncertainty to quasi-certainty. The study of this transition has given rise to several studies of what brain mechanisms help us adapt to different decisional structures. The fact that the choices are under ambiguity, uncertainty or risk is differentially treated by the brain (Hsu et al. 2005); the fact that we tend to systematically violate decision-theoretical axioms that are supposed to prevail in these distinct decisional-structures is also a feature that can be explained by the study of our brain fabric. In that context, regret has been made a paradigmatic case. For the purpose of studying specifically regret-aversion based decision-making, rather than conceptually and phenomenologically generic somatic markers, the IGT has been modified into a regret gambling task and used in several seminal studies (for example, Camille et al. 2004; Coricelli et al. 2005). In that task participants are invited to choose between two ‘wheels of fortune’, one on the left, one on the right. They display colored zones corresponding to possible gains or losses. For instance, in one typical choice, if subjects choose the gamble on the left, they might win €200 with 20 percent probability

M4225-ALTMAN_9781782549574_t.indd 192

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 193 or lose €50 with 80 percent probability; if they choose the gamble on the right, they might win or lose €50 with equal probabilities. The regret gambling task differs from the IGT in that it implements two contextual conditions in terms of feedback provided. Partial feedback shows only the outcome of the chosen gamble; to that extent it amounts to IGT and participants can be disappointed given their expectation that the needle stops on a positive zone of the wheel. Complete feedback allows for a comparison between the consequence of participants’ choice and that of the foregone option, eliciting possible regret or relief. Physiological responses (skin conductance responses), choice behavior and brain activity are influenced by these different levels of feedback. Coricelli et al. (2005) have shown that the same brain areas are activated when the brain faces a certain choice situation, before a decision is made, as when it processed the outcomes of similar choices over past repetitions. Precisely, the orbital frontal cortex and the amygdala mediate how past regret history biases subjects towards minimizing regret across similar choice situations in anticipation of its possible consequences. This result is consistent with Damasio’s result, according to which ventral medial structures support the integration between cognitive and emotional components of the entire process of decision-making. Besides the bottom-up mechanisms that make emotions inflect decisions in an optimal sense, the orbital frontal cortex specifically uses a top-down process in which cognitive components, such as counterfactual thinking, modulate emotion. This complex, double-looped, relation between cognition and emotion has been modeled by evolutionary and behavioral responses. However, such a lesson from neurobiology about the adaptive value of regret stresses the tension between the construal of regret in biology and in decision-theory. We proposed a study of the Allais paradox with ventral medial lesions to explore this tension. Allais (1953) showed that people tended to exhibit inconsistent choice patterns when presented with pairs of options that involved a contrast between quasi-certain and risky options (see Table 11.2). The presentation of the Allais paradox in Table 11.2 contrasts a pair of choices among pairs of lotteries. The individual is first invited to choose between Lottery A and Lottery B. He is then confronted with a list of probabilities of obtaining certain payoffs, as in every classical lottery. We can see, for instance, that in Lottery A and Lottery B, he has the same chance (89 percent) of obtaining €500 000. Likewise when confronted with the other choice between Lottery C and Lottery D, he has the same 89 percent chance of winning nothing across the two then concerned lotteries. If the subject rationally applied von Neumann and Morgenstern’s axiom of independence (von Neumann and Morgenstern 1944) which states that a choice between two options should be independent of what is Table 11.2

Matrix of the Allais paradox

Lotteries

Lottery A Lottery B Lottery C Lottery D

M4225-ALTMAN_9781782549574_t.indd 193

Probability of alternative states of nature S1, S2, S3 S1: 0.01

S2: 0.10

S3: 0.89

€500 000 €0 €500 000 €0

€500 000 €2 500 000 €500 000 €2 500 000

€500 000 €500 000 €0 €0

03/05/2017 08:20

194

Handbook of behavioural economics and smart decision-making

common (in terms of both payoffs and the probability of their realization, which is clearly the case for the fourth column of the table when we consider alternatively A and B and then C and D) between the options, he should realize that the pairs of lotteries A and C and B and D actually present the same contingencies. This lack of application of such an independence axiom is what makes the subject reverse his choice, without realizing most of the time that he does so, across the two pairs of choices. Most of the time, a same individual will tend to prefer A to B and at the same time, without realizing any reversal in her preferences, will tend to prefer D to C. In order to grasp the implicit violation of the independence axiom presented by this preference reversal, please delete the last column of the table, which presents common consequences to be bracketed, and now realize how A = C and B = D. Regret theory provides a simple explanation of Allais’s paradox. A person who has chosen option B has, if state of nature S1 materializes, strong reasons to regret his or her choice. A subject who has chosen option D would have much weaker reasons to regret his or her choice in the case of S1. When regret is taken into consideration, it seems quite reasonable to prefer A to B and D to C. Several psychological mechanisms have been hypothetically suggested to account for this type of preference reversal, among which the attractiveness of sure gains and the anticipation of regret if those sure gains would have been forgone and yet realized. The bottom line is that an emotional disposition towards possible future outcomes is involved in the Allais paradox. We tested this hypothesis on a population of patients suffering from behavioral variant frontotemporal dementia (bvFTD), a clinical population known to present ventromedial prefrontal cortex dysfunctions and deficits in experiencing emotional deficit in decision-tasks. We contrasted this group to matched controls and patients with Alzheimer’s disease (AD) who had no ventromedial prefrontal atrophy. Our results showed a drastic diminution of Allaisian behavior among bvFTD patients by contrast with controls and AD patients. We concluded that prefrontal regions are crucial in the production of a behavior that typically stands in contradiction with a basic axiom of rational decision. By contrast, impaired emotional mechanisms ironically produce hyperrational (non-Allaisian) behavior in bvFTD patients (see Bertoux et al. 2013). Decision-theory aims to provide an axiomatic – and thereby in principle intuitive – basis upon which we can assess whether actual choices and repeated decision patterns comply with norms of rationality these axioms are supposed to encapsulate. When these patterns deviate from what logically follows from axioms, we would be alternatively tempted to weaken the latter for the sake of psychological realism or discard evidence on behalf of a principled incommensurability between the descriptive and normative levels. Regret-based decision-theory is a unique attempt, in the recent history of decision-theory, to combine these opposed tendencies in a sort of reflexive equilibrium approach that would jointly increase the intuitiveness of the axiomatic basis and the cognitive adequacy of the proposed theory.

3

REGRET IN DECISION-THEORY

The Allais paradox gave rise to alternative decision-theoretic accounts. One explanation for the A-D pattern is that decision-makers anticipate regret if they choose B and find themselves in the state of nature S1. Note that at this juncture two comparisons with the

M4225-ALTMAN_9781782549574_t.indd 194

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 195 realized outcome are possible which can give rise to two distinct emotions, two forms of regret if one wishes. Depending on information about counterfactual outcomes or states of nature, the unfortunate decision-maker can express two forms of regret. Most naturally, he can regret his own choice B, since the state of affairs S1 has been the case and he won nothing. Had he chosen choice A he would have won a certain €500 000. However, not regretting his choice B, he could deplore his lack of luck and the fact that neither S2 nor S3 had been the case. To our best knowledge these two possible attitudes and types of regret when a risky decision has been made have not been studied. However, if that choice were repeated, it would be contrary to a rational regret-based decision to avoid B, which presents higher than expected utility, because there would be no grounded reason to think that the ‘universe’ is fatally stuck to the actualization of S1 (if it were, why would we speak of S2 and S3 in the first place and why would we have attributed the probability 0.01 to S1?). Per absurdum, we would systematically be deterred from choosing B, in a repeated sequence of choices, on the basis of anticipated regret if that individual had either developed extreme pessimism or, which amounts to the same, developed a distorted view of probabilities. This remark on the contrast of the relevance of taking into account regret in a single-choice versus repeated-choices situation makes the picture of how regret should enter into decision-theory more complex than at first sight. Regret-theory has been formalized to account for comparisons between actual and counterfactual outcomes within a single state of nature or ‘world’. Regret is relevant in that single world, as expressed through these introductory terms by Suhonen (2007, p. 11, emphasis added): The central idea behind regret theory is that, when making decisions, individuals take into account not only the consequences they might experience as a result of the action chosen, but also how each consequence compares with what they would have experienced under the same state of the world had they chosen differently . . . Then the overall level of satisfaction derived is a combination of the basic utility of the consequence actually experienced, and some decrement or increment of utility due to ‘regret’ or ‘rejoicing’.

In standard expected utility theory a prospect is evaluated according to the utility of each outcome irrespective of what the other possible outcomes can be. This is strictly forward-looking and consequentialist. As soon as an individual looks to past decisions to base his present choice on them or, more complexly, anticipate a future backwardlooking state of mind in which he anticipates he will be regretting the present decision he is liable to make, he immediately ceases to be strictly deciding along consequentialist guidelines. To incorporate such backward-looking attitudes in decision-making leads us to adopt a non-standard expected utility theory. For instance, an individual may wish to avoid uncertainty, or an individual may not be able to evaluate single payoffs per se but only by comparison with other possible outcomes, his mind being fit to reference points and relative status rather than to processing absolute values. Under a narrow Savagean interpretation of the Allais paradox (Savage 1954 [1972], consequences are identified with monetary payoffs. Under this restriction, expected utility theory is violated by most individuals (including Savage, according to the legend). Under a broader interpretation of what consequences are, or, equivalently, of what a decision-process amounts to, such as that provided by regret-theory, no violation exists. This way out may appear less than satisfactory, though once we relax axiomatic constraints every type of decisional pattern

M4225-ALTMAN_9781782549574_t.indd 195

03/05/2017 08:20

196

Handbook of behavioural economics and smart decision-making

can be interpreted as satisfying some revised state of the axiomatics, emptying the theory of all normative and descriptive content. As Tversky (1975) underlined, because we wish to maximize the predictive power of the theory, we are tempted to adopt a restricted interpretation of utility, such as the identification of outcomes with monetary payoffs. In this respect, we can consider regret-theory (Bell 1982; Loomes and Sugden 1982, 1987a) as an optimal trade-off between axiomatic revision and predictive power, a maximally conservative attempt at deviating from standard expected utility with a view to intuitively account for robust behavioral data. Loomes and Sugden (1987b) have built a model that generates testable prediction and compares with alternative theories, standard or not. Regret-theory consists of the intertwining of two factors in a single utility function that thereby incorporates two measures of satisfaction: utility of outcomes, as usual, and a quantitative measure of regret. The mixing of the two – which supposes their commensurability and, therefore, a conceptual assimilation of regret to a negative payoff – yields a moderately modified concept of utility. In Bell’s terms (1982, p. 963) regret is measured as: ‘the difference in value between the assets actually received and the highest level of assets produced by other alternatives’. Tracking this difference across choices is what regret amounts to. This is represented by a two-parameter function u(x, y) where x is the actually received payoff and y the difference just referred to. x and y cannot jointly increase and by construction x is maximal when y is null, which means that in this approach, the payoffs of options to be compared always sum to zero and there is always a dominant choice. Interestingly, this functional representation, thus interpreted, presents a potential contradiction with another heuristic supposed to make us smart, namely, Simon’s satisficing principle (Simon 1956, pp. 129, 136). If I reach a level xˉ of payoff above which I do not experience any utility increase, then y can freely increase, in the sense that I could have gotten more than xˉ , but without the difference Y − xˉ any longer generating any regret. The contradiction is swiftly spelled out, at the theoretical level first, if we consider that the use of a minimax-regret heuristic is compatible with the optimal determination of our satisficing threshold: given a certain difference between x and y, we can decide not to take it into account, either because it is too small or because it is too large. In that sense, it is the individual’s sensitivity to comparisons and regret they elicit that endogenously defines their satisfaction threshold. We can indeed assume that people are sensitive in a non-linear way to different payoff intervals between what they get and what they could have got. If the different is negligibly small or unrealistically too large, I can cease to be sensitive to the discrepancy between actual and forgone outcome. It is likely that the individual feels regret when the comparison falls between a certain perceivable and conceivable gap, beyond which it appears a vain feeling. In that sense minimization of anticipated regret and satisficing are not incompatible smart decision-making principles. This is consistent with some recent results about the behaviors of maximizers and satisficers with respect to their feeling regret about their decisions (Moyano-Diaz et al. 2014). As the authors of the study point out, a main difficulty in the study of decision-making is precisely the combination of the two coexisting partially incoherent aspects and dimensions that are maximization and satisfaction. Regret can be seen as playing a subtle mediational role between these two dimensions. Regret is more than an emotional reaction against bad consequences; it is also an internal transformation, as we have seen, of these bad consequences into an anticipatory process. Regret is then inherent to its future avoidance and is compatible with maximizing utility. As Schwartz et al. (2002) found and is reported by

M4225-ALTMAN_9781782549574_t.indd 196

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 197 Moyano-Diaz and his colleagues, maximizers tend to engage more often than satisficers in social comparisons and are more affected by them. Regret is for them a driving force towards optimization. Maximizing is not blind forward-looking or full-fledged consequentialism, then, as past errors and all alternatives are scrutinized. However, this type of regret-based maximization decision-making style bears a high psychological opportunity cost and potentially generates a lot of continuing frustration. In contrast, satisficers proceed more easily and are satisfied with their good enough option. In that sense, satisficers can be said to minimize the counterproductive use of first-order regret minimization. We then equate Simon’s satisficing to second-order minimax-regret in what we consider a fuller account of the psychological cost associated with the presence of bad decisions and the correlative first-order regret they tend to provoke. We could thus envisage a Simonbased regret model in which the individual is likely to experience regret when the psychological cost to do so does not exceed the benefits of correcting his or her present decision in view of future benefits. Admissible regret is thereby endogenously defined because future benefits are themselves bound by the individual’s level of satisfaction. When the latter is attained, there is no more valuable motive to admit regret as a reasonable emotion. This Simonian perspective on the link between satisfaction and regret suggests a deeper analysis of how the different ways of incorporating regret into decision-theory also convey different views of human rationality. In summary, Loomes and Sugden’s way is still essentially consequentialist, in the sense that regret is both included in a maximization process and that the functional representation they propose remains fundamentally compatible with a forward-looking attitude. Differently, the short analysis we have provided of the compatibility of satisficing and second-order regret-minimization amounts to a non-consequentialist view, decision-makers deciding now to ignore certain information and consequences of their choice beyond a certain threshold; satisfaction, in the Simonian sense, meaning not only to cease to adopt a utility maximizing attitude but also to cease to make any counterfactual comparison once a certain level of utility is reached. This way of blocking potential feelings of regret emphasizes, by contrast, the role of signals and omens in standard decision-making. We can imagine individuals ready to pay not to receive information about outcomes of forgone decisions. Karlsson and colleagues (2005), in an unpublished study, have documented a very similar phenomenon on financial markets, which they label an ‘ostrich effect’, people deliberately discarding information about their investments portfolios when markets go down. Regret aversion is seemingly an ordinary feature of decision-making related to a human propensity to respectively seek or avoid positive and negative omens and base our decision at this symbolic level rather than strictly focus on the evaluation of our choices’ consequences. We have studied regret in connection with the Newcomb problem in that perspective. This problem tackles deep philosophical issues around the nature of rationality, along the dividing line we have discussed above: can we rationally take into account information about our choices that in fact does not change the way consequences will be realized. Is there a way to vindicate the fact that, in some cases, we are sensitive to consequentially irrelevant information? This dividing line among decision-theorists – only among those of a philosophical bent though – has been labelled in terms of a difference between causal decision-theory (individuals make their choices according to basic stochastic dominance and independence principles) and evidential decision-theory (individuals may legitimately be influenced in their decision-making by symbols or information present in the choice

M4225-ALTMAN_9781782549574_t.indd 197

03/05/2017 08:20

198

Handbook of behavioural economics and smart decision-making

situation which, in principle, do not affect the consequences) (Joyce 1999). Nozick (1969) devised that paradox of rationality with the explicit purpose of contrasting consequentialist and symbolic forms of rationality. In this problem, two boxes, one opaque, one transparent, are presented to a decision-maker along with the following message: Imagine a being with great predictive powers. You are confronted with two boxes: B1 and B2. B1 is opaque and B2 is transparent, you can see that it contains €1. B2 contains €1; B1 contains either €10 or nothing. You may choose B1 alone or B1 and B2 together. If the being predicts that you choose both boxes, he does not put anything in B1; if he predicts that you choose B1 only, he puts C10 in B1. 5> What should you choose?

We had hypothesized that if this choice of alternatives, as Nozick thinks, coincides with different types of rationality, they could also elicit different levels of confidence in our choices. If I really believe in God, I might be inclined to accept a certain level of nonconsequentialism in my choices. However, disappointment can be greater in that case than if I had made a purely consequentialist choice (ignoring the omen) followed by a bad outcome. I therefore considered that regret, when I realized what I could have got had I made the other choice, is modulated by the type of rationality implied by our choices (Bourgeois-Gironde 2010). Let us label individuals one-boxers and two-boxers according to their decisions in the Newcomb problem (Nozick 1969). Two-boxers go against the prediction. The decisioncriteria they presumably follow have been characterized, as we did, as consequentialist, but also, in philosophical parlance, as we commented above, as ‘causalist’, by contrast with ‘evidentialist’. Two-boxers thus apparently exhibit a higher autonomy, that is, a higher level of decisiveness, in their choices than do one-boxers, although the latter’s faith in the omniscient predictor can also yield a high level of decisional confidence (think of Pascal’s wager; Pascal 1897). Integrating the decision-criteria predictions, signs and symbolic value may not be altogether irrational (Nozick 1994). It is pervasive enough, as, for example, in convincing ourselves of our good health or of the influence of our vote in national elections by going to vote, by accomplishing acts that amount to generate selfmanipulated positive signals (Quattrone and Tversky 1988). Shafir and Tversky (1992) ran the first empirical investigation of Newcomb problems. They submitted to their subjects a Newcomb problem. Their cover story was that ‘a program developed at MIT was applied during the entire experimental session to analyze the pattern of your preferences, and predict your choice (one or two boxes) with 85% accuracy’ (Shafir and Tversky 1992, p. 461). Although it was obvious that no deus ex machina intervened at the moment of choice, most experimental subjects opted for the single opaque box rather than for the dominant two-boxes strategy. It is as if they believed that by declining to take the money in box B2, they could change the amount of money already deposited in box B1. Adding on their test, we measured whether regret was different when negative outcomes are revealed to one-boxers and two-boxers. We observed – by means of a retrospective measure of satisfaction on a five-point Likert scale – that one-boxers, when facing negative outcomes, experience a significantly greater amount of regret than do two-boxers in the same situation. This is due, we speculate, to the lesser decisiveness or autonomy with which those choices are made, in spite of their greater faithfulness to the prediction. If a difference emerges between types of decision and amount of regret in the Newcomb problem, this can be considered as a step toward a better understanding of

M4225-ALTMAN_9781782549574_t.indd 198

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 199 how regret taps into rational antecedents of choices and can be modulated by competing criteria of rationality.

4

REGRET AS A COORDINATION DEVICE AND A SOCIAL MECHANISM

Bob wants to go to the opera with Ann tonight. However, in their last discussion, they split on a doubt about what they would do; Ann definitely seemed to wish to please Bob and to agree to go to the opera, although she had declared her preference for the boxing match in another part of town. When they left, Bob said clumsily but audibly that he would like to please her too. In the confusion they omitted to give each other their mobile numbers. Now Bob goes to the boxing match, and does not find Ann. In another scenario he goes to the opera but learns the next day that Ann had opted to go to the boxing match, having understood that Bob would join her there. One could say that an ex ante post hoc regret-minimizing based decision would have induced Bob to go to the fight, as he would have had the moral comfort to have tried to please Ann retrospectively. However, if Ann follows a similar strategy, they miss each other and fail to go out together to their joint detriment. Imagine this situation is repeated every day, with the same level of conversational confusion and omission of electronic devices coordination, by Bob and Ann. It is as if they both lived in ‘groundhog day’ with the characteristic fact that some external and internal elements (such as their inability to correct their spontaneous lack of coordination) fatally befell them. However, they can change their decision every day. There are many ways for them to get out of their predicament. They still have available external and internal resorts. They can use a binary coordinative device, like tossing a coin, or look at the sky, whether it is sunny or overcast, and silently fit their behavior, after a few learning trials, on this conventional signal. Or they can use regret, which is internal, on the hypothesis that they tend to feel the same, which was in the premise of the argument. The use of the coin, or the sky, or a traffic-light for that matter, is a coordinative device that leads to a correlated equilibrium (Aumann 1974). A correlated equilibrium arises in a situation where mis-coordination is likely, owing to several present Nash equilibria leading to suboptimal effects, and when players resort to a set of received signals or instructions by a neutral referee. A correlated equilibrium, technically, is a probability distribution over the players’ space of strategies realized by the referee (or by nature) and from which no player has a unilateral interest to deviate. Hart and Mas-Collel (2000) have shown in what sense regret is an adaptive heuristic in reaching correlated equilibria. Basing a decision on the one that simply minimizes our regret, in the ‘battle of the sexes’ between Bob and Ann described previously, could apparently lead to persistent mis-coordination. If one pleases the other in the same way, their paths will continue to diverge, but this could be solved by using the following heuristic: ‘Switch next period to a different action with a probability that is proportional to the regret for that action, where regret is defined as the increase in payoff had such a change always been made in the past.’ (Hart 2005, p. 1405). This works because it dynamically removes the players from a single-handed strategy (that would consist, for example, of blindly pleasing the other). Players engage in a learning dynamics that will make sources of regret endogenously

M4225-ALTMAN_9781782549574_t.indd 199

03/05/2017 08:20

200

Handbook of behavioural economics and smart decision-making

evolve and, thereby, by individually, but jointly, directing their effort to minimize this regret, happen to reach a correlated equilibrium in an efficient way. Regret owing to mere failure of coordination (not simply because of own fault or misplaced benevolence) will progressively guide the choice of strategies by pondering the probability of which of the strategies is played. It means that both players end up feeling the same type of regret based on their suboptimal payoffs, whatever the personal motives that previously led them to mis-coordinate. Interestingly, jointly minimized regret, although an internal endeavor and learning mechanism, plays the role of a public coordination device, avoiding any need for external communication or objective means of coordination, as is required in other ways of reaching correlated equilibrium. The use of regret represents a smart heuristic in the sense that it allows players to eschew the computation of highly complex objects such as repeated games strategies and beliefs. By means of Hart’s heuristics they can simply trade-off between past payoffs associated with a given course of action and match past frequency of success with the probability with which they will stick to this course of action. More precisely, regret is a ‘smart’ heuristic in the sense that it is cognitively parsimonious (to match past payoffs and future actions probabilities is relatively easy), but also in the sense that it is not dumb or fully blind either, as it requires a certain level of freedom of choice and self-modulation of the weight the individual wants to give to this signal. It then requires a certain degree of rationality. In evolutionary dynamics, by contrast with learning dynamics such as the use of a matching-regret heuristic, agents do not have to exhibit any level of rationality, as their phenotype (observable behavior) is deterministically dictated by their genotype entailing that they have no leeway to modulate their strategies and learning mechanism. They play relatively fixed actions that aggregate into group behavior. What can be called rational or irrational in that context is the collective dynamics of the population to the extent that it leads to optimal steady states. Our next question, then, is to ask to what extent assessments of collective rationality can be informed by regret-based decisional patterns at the individual level? Voting procedures are paramount social mechanisms that display this discrepancy between individual and collective rationality. It can go both ways. The paradox of voting is a typical instance of collective rationality (the possibility of an optimal social choice through aggregation of individual preferences) not being supported by individual rationality (in large groups, it is ‘irrational’, when we think in costs-benefits terms, to pay the cost of going to the voting booth). Arrow’s impossibility theorem (Arrow 1951 [1963]), on the other hand, is a deep illustration of how individual rationally structured preferences (with respect to their transitivity) do not necessarily aggregate into a transitive social preference. A solution of the paradox of voting has been proposed by Ferejohn and Fiorina (1974) in terms of the minimax regret criterion. Voters choose the action that yields a minimal regret in a worst case scenario. This implies a form of strategic voting which can seem contradictory with the fact that, if voters are aware that their vote is unlikely to be pivotal, they could still vote sincerely. The situation is far from being unrealistic. Also, collective regret is likely to arise in that situation. Regret being elicited by the possibility of comparing what one has got from what one could have got depends on the possibility of counterfactual learning at some point of the democratic process. The 2007 elimination of candidate Lionel Jospin in the first round of the 2007 French presidential

M4225-ALTMAN_9781782549574_t.indd 200

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 201 election may have aroused such an emotion among the voters for the minor left candidate, Taubira, leading to the far-right candidate Le Pen’s presence in the second round. The use of alternative voting procedures may not only have led to another result, but perhaps also to lesser pangs of regret felt by some of Taubira’s supporters (Baujard et al. 2014). Uninominal majority voting, approval voting, and evaluative voting differ in three comparable respects: (1) they are more or less expressive, in the sense that voters can convey more or less fine information about their preference in selecting their options; (2) it is more or less polarized, in the sense that negative feeling towards a candidate can be strengthened in evaluative voting if negative grades are allowed; (3) it can be more or less easily strategized. The parameters inherent to the voting procedure can be associated with its varying susceptibility to regret. If we have the opportunity to express a fuller and finer choice without jeopardizing the election of a consensual candidate (which is the case by use of approval voting in particular), regret, in case of non-success, will presumably be minimized. Tideman (1985) extended the minimax regret model of voting by borrowing from the Sugden and Loomes’ framework we discussed in section 3. He adds the concept of remorse and elation to the model, which are ‘emotions that arise as a consequence of being responsible for one’s circumstances by one’s own actions’ (Todeman 1985, p. 103). The key, there, is this feeling of responsibility which is a constitutive element of regret: losses and gains are accentuated if we are or feel responsible for the feared result. Guilt drives regret, even in a context where there is no objective influence of the voter on the outcome. It disproportionately distorts it and might explain high levels of voter turnout. Can it be smart in this social and political context? Emotions are an important element of the political game and influence behavior (individual or collective) (Groenendyk 2011). However, what has been less studied is how emotions endogenously depend on certain features of given democratic structures, in terms of choice procedures, aggregative mechanisms and informational filters and feedbacks. Understanding how the mechanisms of social choice generate by themselves emotional states and flux is crucial in view of the democratic regulation of the political game: democratic leaders do not want their institutions to be undermined by extreme political affects nourished by their constituencies. Several levels of analysis are relevant for this still widely open set of issues: (1) mechanisms of choice – essentially how the procedure used for voting and the aggregative mechanism associated with those rules trigger positive or negative emotions; (2) information and feedbacks about the political process and its results (pools and so on), on the basis on which payoffs comparison can be made and an adaptive regret heuristic launched; and (3) timing of choice – frequency of elections, possibility of voicing popular opinion in regular moments and channels – determining fluctuating emotional states in the population of voters. At a more general level, we could wish for a society that maximizes consensus, by minimizing the distance between the social choice and each individual vote (see Kemeny 1959) and that also minimizes regret in terms of minimizing individual loss functions. This is not strictly equivalent, and some dynamic emotional fluctuations in democracy might be due to the interplay between those two minimization functions. More generally, the issue is the analysis of how emotional dynamics in democracy might depend on fine structural choices on aggregative and informational political mechanisms. Again, we point here at a possible discrepancy between social consensus thus defined and individuals’ psychological

M4225-ALTMAN_9781782549574_t.indd 201

03/05/2017 08:20

202

Handbook of behavioural economics and smart decision-making

factors such as regret or satisfaction with respect to the effective social choice. Each individual feels more or less satisfied with the current social choice (for example, the result of the last election). Ignoring, for the sake of a simpler presentation, dynamic effects occurring in the course of office terms, these levels of regret or satisfaction can be envisioned, as a first approximation, as products of loss functions (minimax regret) for each individuals. From a computational point of view the aggregation of individual loss functions in the population and the measure of the distance between each individual preference of that population and the social choice are not identical. Incorporating individual levels of regret as a parameter of social satisfaction with respect to some social choice function is then an attempt at combining collective and individual rationality. It also potentially anchors back social choice mechanisms into adaptive heuristics upon which real individuals tend to make decisions, combining emotional and cognitive abilities, and fine-tuning our biological and social fabrics.

5

CONCLUSION

Anticipated regret is one of the most efficient heuristics that we can use in order to avoid suboptimal decision-making. It has been incorporated in decision-theory at an axiomatic level, both in individual decision-making and in interactive strategic situations. Moreover, it has been demonstrated an efficient learning mechanism, leading to optimal decision-making, and coordinative device in multiple equilibria game-theoretical contexts. Regret-theory is nevertheless compatible with bounded rationality paradigms such as Simon’s satisficing principle; the latter involving a form of second-order modulation of the amount of regret that an individual can reasonably experience in order to guide his or her decisions towards a satisfaction threshold. This view seems in agreement with what recent brain-imaging studies have taught us about the neurobiology of regret. The emotion of regret lies at the junction of the processing of aversive states and of the cognitive anticipation of future outcomes of the individual’s actions. For this set of reasons, we consider regret to be one of the best candidates with a view to unifying biological and decision-theoretical approaches to optimally bounded decision-making.

NOTE 1. Somatic markers such as perspiration, hence skin conductance of body parts such as fingertips, or heartbeats, pulsations, and so on form a particular class of measurable bodily states that Damasio and Bechara in a series of influential studies in the 1990s have shown to be correlated with so-called secondary emotions. The latter are feelings that have been associated, through past repeated experiences, to the learning and anticipation of future outcomes in certain typical choice situations. When a somatic marker is associated with a particular negative outcome it functions as an alarm bell and is a reliable signal. It is convenient to understand anticipated regret in the framework of this ‘somatic marker hypothesis’. Interestingly these predictive somatic markers can be effective without fully arising to consciousness, making them an automatic self-corrective mechanism. (For another type of experimental approach on feelings of errors and their predictive and corrective roles see Gangemi et al. 2015.)

M4225-ALTMAN_9781782549574_t.indd 202

03/05/2017 08:20

How regret moves individual and collective choices towards rationality 203

REFERENCES Allais, M. (1953), ‘Le comportement de l’homme rationnel devant le risque: critique des postulats et axiomes de l’école américaine’ (‘The attitude of the rational man to risk: critical assumptions and axioms of the American school’), Econometrica: Journal of the Econometric Society, 21 (4), 503–46. Arrow, K. (1951), Social Choice and Individual Values, 2nd edn 1963, New York: Wiley. Aumann, R.J. (1974), ‘Subjectivity and correlation in randomized strategies’, Journal of Mathematical Economics, 1 (1), 67–96. Baujard, A., H. Igersheim, I. Lebon, F. Gavrel and J.F. Laslier (2014), ‘Who’s favored by evaluative voting? An experiment conducted during the 2012 French Presidential Election’, Electoral Studies, 34 (June), 131–45. Bechara, A., D. Tranel and H. Damasio (2000), ‘Characterization of the decision-making deficit of patients with ventromedial prefrontal cortex lesions’, Brain, 123 (11), 2189–202. Bell, D.E. (1982), ‘Regret in decision making under uncertainty’, Operations Research, 30 (5), 961–81. Bertoux, M., F. Cova, M. Pessiglione, M. Hsu, B. Dubois and S. Bourgeois-Gironde (2014), ‘Behavioral variant frontotemporal dementia patients do not succumb to the Allais paradox’, Frontiers in Neuroscience, 8 (September), 287. Bourgeois-Gironde, S. (2010), ‘Regret and the rationality of choices’, Philosophical Transactions of the Royal Society B: Biological Sciences, 365 (1538), 249–57. Boyer, P. (2008), ‘Evolutionary economics of mental time travel?’, Trends in Cognitive Sciences, 12 (6), 219–24. Camille, N., G. Coricelli, J. Sallet, P. Pradat-Diehl, J.R. Duhamel and A. Sirigu (2004), ‘The involvement of the orbitofrontal cortex in the experience of regret’, Science, 304 (5674), 1167–70. Coricelli, G., H.D. Critchley, M. Joffily, J.P. O’Doherty, A. Sirigu and R.J. Dolan (2005), ‘Regret and its avoidance: a neuroimaging study of choice behavior’, Nature Neuroscience, 8 (9), 1255–62. Damasio, A. (2008), Descartes’ Error: Emotion, Reason and the Human Brain, New York: Random House. Damasio, A.R. (2004), Looking for Spinoza: Joy, Sorrow, and the Feeling Brain, New York: Random House. Elster, J. (1998), ‘Emotions and economic theory’, Journal of Economic Literature, 36 (1), 47–74. Ferejohn, J.A. and M.P. Fiorina (1974), ‘The paradox of not voting: a decision theoretic analysis’, American Political Science Review, 68 (2), 525–36. Gangemi, A., S. Bourgeois-Gironde and F. Mancini (2015), ‘Feelings of error in reasoning – in search of a phenomenon’, Thinking & Reasoning, 21 (4), 383–96. Gilbert, D.T. and T.D. Wilson (2007), ‘Prospection: experiencing the future’, Science, 317 (5843), 1351–4. Groenendyk, E. (2011), ‘Current emotion research in political science: how emotions help democracy overcome its collective action problem’, Emotion Review, 3 (4), 455–63. Hart, S. (2005), ‘Adaptive heuristics’, Econometrica, 73 (5), 1401–30. Hart, S. and A. Mas-Colell (2000), ‘A simple adaptive procedure leading to correlated equilibrium’, Econometrica, 68 (5), 1127–50. Hsu, M., M. Bhatt, R. Adolphs, D. Tranel and C.F. Camerer (2005), ‘Neural systems responding to degrees of uncertainty in human decision-making’, Science, 310 (5754), 1680–83. Joyce, J.M. (1999), The Foundations of Causal Decision Theory, Cambridge: Cambridge University Press. Karlsson, N., D.J. Seppi and G. Loewenstein (2005), ‘The “ostrich effect”: selective attention to information about investments’, unpublished manuscript, available at SSRN 772125. Kemeny, J.G. (1959), ‘Mathematics without numbers’, Daedalus, 88 (4), 577–91. Loomes, G. and R. Sugden (1982), ‘Regret theory: an alternative theory of rational choice under uncertainty’, Economic Journal, 92 (December), 805–24. Loomes, G. and R. Sugden (1987a), ‘Some implications of a more general form of regret theory’, Journal of Economic Theory, 41 (2), 270–87. Loomes, G. and R. Sugden (1987b), ‘Testing for regret and disappointment in choice under uncertainty’, Economic Journal, 97 (388a), 118–29. Milner, B. (1962), ‘Laterality effects in audition’, in V.B. Mountcastle (ed.), Interhemispheric Relations and Cerebral Dominance, Baltimore, MD: Johns Hopkins University Press, pp. 177–95. Moyano-Díaz, E., A. Martínez-Molina and F.P. Ponce (2014), ‘The price of gaining: maximization in decisionmaking, regret and life satisfaction’, Judgment and Decision Making, 9 (5), 500–509. Neumann, J.V. and O. Morgenstern (1944), Theory of Games and Economic Behavior, vol. 60, Princeton, NJ: Princeton University Press. Nozick, R. (1969), ‘Newcomb’s problem and two principles of choice’, in N. Rescher (ed.), Essays in Honor of Carl G. Hempel, Dordrecht: Springer, pp. 114–46. Nozick, R. (1994), The Nature of Rationality, Princeton, NJ: Princeton University Press. Pascal, B. (1897), Pensées et Opuscules (Thoughts and Minor Works), Edition Brunschvicg, Paris: Hachette. Persaud, N., P. McLeod and A. Cowey (2007), ‘Post-decision wagering objectively measures awareness’, Nature Neuroscience, 10 (2), 257–61.

M4225-ALTMAN_9781782549574_t.indd 203

03/05/2017 08:20

204

Handbook of behavioural economics and smart decision-making

Quattrone, G.A. and A. Tversky (1988), ‘Contrasting rational and psychological analyses of political choice’, American Political Science Review, 82 (3), 719–36. Savage, L.J. (1954), The Foundations of Statistics, New York: John Wiley & Sons, reprinted 1972, New York: Dover. Schacter, D.L., D.R. Addis and R.L. Buckner (2008), ‘Episodic simulation of future events’, Annals of the New York Academy of Sciences, 1124 (1), 39–60. Schwartz, B., A. Ward, J. Monterosso, S. Lyubomirsky, K. White and D.R. Lehman (2002), ‘Maximizing versus satisficing: happiness is a matter of choice’, Journal of Personality and Social Psychology, 83 (5), 1178–97. Shafir, E. and A. Tversky (1992), ‘Thinking through uncertainty: nonconsequential reasoning and choice’, Cognitive Psychology, 24 (4), 449–74. Simon, H.A. (1956), ‘Rational choice and the structure of the environment’, Psychological Review, 63 (2), 129–38. Spinoza, B. (1677), Ethica, ordine geometrico demonstrate, trans. E. Curley (1996), The Collected Works of Spinoza, vol. 1, Ethics, Princeton, NJ: Princeton University Press. Suhonen, N. (2007), Normative and Descriptive Theories of Decision Making Under Risk: A Short Review, Joensuu: University of Eastern Finland. Tideman, T.N. (1985), ‘Remorse, elation, and the paradox of voting’, Public Choice, 46 (1), 103–6. Tversky, A. (1975), ‘A critique of expected utility theory: descriptive and normative considerations’, Erkenntnis, 9 (2), 163–73. Young, H.P. (2004), Strategic Learning and Its Limits, Oxford: Oxford University Press. Zeelenberg, M. (1999), ‘Anticipated regret, expected feedback and behavioral decision making’, Journal of Behavioral Decision Making, 12 (2), 93–106.

M4225-ALTMAN_9781782549574_t.indd 204

03/05/2017 08:20

12 Is it rational to be in love? Paul Frijters and Gigi Foster

1

INTRODUCTION

Einstein reportedly once remarked that ‘only a life lived for others is a life worthwhile.’ Implicitly, Einstein thus viewed Rational Economic Man, who knows what he wants and only cares for himself, as not living a worthwhile life. To the degree that living a worthwhile life is rational, Einstein would have implicitly deemed ‘Rational’ Economic Man to be, well, not. The core process we aim to illuminate in this chapter consists of people – even very smart ones – being programmed, through economic and social means, to exhibit loyalty to both other people and abstractions. As members of families, nation states, religions, sports teams, professions, and friendships, most of us will have loyalties outside ourselves, deriving pleasure from seeing our loved ones thrive, but economists lack a tractable theory for how such loyalties come about. We therefore also lack an understanding of how groups employ institutions and strategies to make new generations of individuals adopt those loyalties that are useful to existing groups, as well as to successful functioning in their future lives, including as members of groups yet to emerge. In this chapter we try to fill that void, adding loyalty to the basic economic toolkit. Our theory of how people change their loyalties includes a large role for the unconscious mind: we will argue that changing loyalties is not a conscious choice. Based on introspection and simple observation of the human condition, we claim that people cannot consciously choose to increase their love to any level they want. They may consciously put themselves into circumstances that push them towards falling in love with something, but they cannot simply decide to love something and make this state of affairs come about instantaneously. In that sense, being in love is an unconscious process and thus rational in a limited way, on a par with other bodily processes that are unconsciously regulated, like maintaining blood sugar or testosterone levels. We start with a toy model of standard economic rationality wherein individuals never change their loyalties, which we then stretch and shape, via a series of intermediary models, into a model of fluid loyalties and rules of thumb as to how those changes come about. We illustrate how the fluid notion of loyalty throws light on group-mediated phenomena such as education, national symbols, group ideals, and adherence to the ideals of science. We use a simple public goods game to illustrate how particular loyalties to group abstractions held by a small minority help to coordinate a whole group of individuals on the optimal outcome for the group as a whole – a feature that has come in particularly handy during the course of human development. This then leads to a short discussion of the institutions via which group power is organized and maintained. The topics addressed in this chapter are thematically related to existing large economic literatures on household bargaining, parental investments, and reciprocity. However, as far as we know, these literatures have never discussed in mathematical terms how loyalties 205

M4225-ALTMAN_9781782549574_t.indd 205

03/05/2017 08:20

206

Handbook of behavioural economics and smart decision-making

arise and change, which is the main focus of our chapter. In that sense, we know of no prior work that we directly build upon. When we speak of ‘mainstream economics’ as the counterpoint to our models incorporating loyalty, we have in mind the notion of rational economic man that is taught to undergraduate students of economics, rather than the many expansions in various sub-literatures that the few who proceed to higher studies will encounter. The literatures adjoining this chapter and the content of first-year textbooks that we take as indicative of ‘mainstream economics’ thinking are both discussed in Frijters and Foster (2013). What this chapter adds to that book is an extended set of mathematical models that describe how loyalty arises. We also contemplate here a meta-question: is it advisable to resist changes in loyalties, seeking to remain immutable and fixed over time, like the rigid figure of mainstream economic models? Or is it smarter to be what we call in later sections a Rhytonian rationalist, whose notion of self changes over time via the ebbing and flowing of his bonds with other people and entities? We argue that from the view of the initial self, changing loyalties is like submitting to a form of premature death: a betrayal of what was originally cared for. From the view of the rational ever-experiencing self, however, developing loyalties to abstractions and people as they are encountered is likely to be the more adaptive and happiness-maximizing strategy. In this sense, we explicitly embrace self-delusion, belief in non-existing entities, and the jettisoning of loyalty towards prior notions of self as potentially quite ‘rational’ and evolutionary adaptive choices.

2

THE ARGUMENT FOR LOYALTY

We start with the observation that people are capable of forming lasting bonds both between themselves and other humans, and between themselves and abstractions that they conjure in their minds. These bonds are all held in the mind, but have large behavioral implications. The clearest evidence for this claim is pure introspection: is there truly nothing outside yourself that you care about, and with which you feel yourself to have long-lasting bonds? Is every favor you bestow on others the result of selfish maximization, providing no ‘warm glow’ (Andreoni 1990) or other positive internal reward? We need the concept of loyalty to explain behavior. Without a mental adherence to gods and spirits, we should not see lucky charms or private prayer, or any other activity that others cannot see us doing but that involves a time investment towards unseen and arguably non-existing entities. Additional evidence comes from our emotional responses to how we think others view us, showing a mental adherence to abstract ideals of behavior against which we believe we are judged (for example, fairness, chivalry, integrity): loyalty to these ideals is implicit in self-loathing internal experiences like guilt and shame that in turn have been known to drive behaviors from listlessness (Hentschel 2007) to self-harm (Inbar et al. 2013). A great deal of human behavior seen throughout history, from the child sacrifice of ancient cultures to the Australian Aboriginals’ ritual of pointing the bone, is extremely difficult to explain without the existence of some unseen link people sense they have to things outside themselves. How do these mental bonds arise? How does a person come to start ‘believing’ in social institutions, such as democracy, human rights, or equal opportunity? How does

M4225-ALTMAN_9781782549574_t.indd 206

03/05/2017 08:20

Is it rational to be in love? 207 he develop bonds with gods, or for that matter with other humans? Only once we have a working model for the development of such bonds will we be able to think in an organized way about when they arise and what their behavioral implications might be. Throughout this chapter, we refer to such bonds as ‘love’ or ‘loyalty’. We will offer a precise mathematical definition in later sections, but intuitively, true loyalty will be defined as the inclusion of a mental depiction of an outside entity within the mental depiction of the Self.1 Love will push the one who loves to take actions that support the loved entity, because the loving individual receives internal rewards (‘utility’) from doing so and pain from perceiving that his ideal is suffering, just as he would experience pleasure or pain when other parts of his Self (such as his physical person, or his self-esteem) are stimulated in positive or negative ways.

3

FROM GREEDY RATIONALITY TO RHYTONIAN RATIONALITY

In this section, we sequentially work through a series of simple models of the objective function that individuals are trying to maximize, starting with a standard mainstream economic model. Our final model nests a classical vision of rational individuals who are loyal only to a limited notion of Self, but also accommodates rational individuals who have fluid loyalties to a much broader notion of Self. To experience the difficulty of deciding which among the alternative personas these models accommodate is the most ‘rational’, a reader might approach them from the point of view of a concerned grandparent: which persona would you wish your own grandchild to have? 3.1

Greedy Rationality

To set the scene, suppose there are N entities. At least some subset of N must be thought of as actual individuals throughout all of our models, although later the set N will also include entities that do not exist. We focus on human decision maker i, a member of set N, where i will throughout the exposition denote the experiencing individual (that is, ‘the entity known as i that experiences utility’). Final consumption in period t of individual i is denoted by the vector of consumption goods Xit. Apart from consuming goods, individuals also have a resource that we call ‘power’, which can be intuitively understood as the ability to influence the environment, most notably other individuals. Power is intimately tied to the social environment. The power of individual i is denoted by sit and can be at least partly expressed as a function of the elements of Xit, as when some amount of power derives from the purchasing power denoted by a weighted average of consumption goods. However, power is not fully reducible to a function of consumption, as it can also come from an individual’s physical strength and other socially recognized promises and rights – elements that do not have a straightforward relation to traded consumption goods. The Xit vector can also include sit as an element, to accommodate the possibility that power may provide direct utility. We think of consumption goods – that is, Xit excluding the sit element, for all i and t – as transferable among entities, with the pre-transfer allocation (called the ‘endowment’ or ‘production’) of a good x to entity i in time t denoted by | xit. By design, if an entity is a

M4225-ALTMAN_9781782549574_t.indd 207

03/05/2017 08:20

208

Handbook of behavioural economics and smart decision-making

member of the subset of imagined but not real entities, its actual endowments of both consumption goods and power are zero. Each transfer of good x from i to j, which is part of a ‘trade’ as traditionally understood if j directly reciprocates, is denoted Txjit. Each use of the power of an individual i towards an entity j is denoted Tsjit. Transfers of goods can encompass trades between people, but also may include offerings to gods or ideals that do not exist but still appear in the set N of perceived market players. While we initially think of the transferable goods in this model as physical goods and readily observable services, we later allow for the possibility that some consumable goods do not exist at all: they can include imagined goods like salvation in the afterlife, the triumph of science over ignorance, and other such higher-order imaginary things that rely on the abstractive capacities of the real traders to be sustained as goods with consumption value. Individuals’ goods-transfer and power-application decisions at time t are based on expectations of the elements of Xi at time t and in future periods. This formulation allows for the possibility that people make investments (transfers) in order to have a higher Xi now or in the future that never in actuality materializes. The simplest form of ‘greedy rationality’ is then associated with an individual i who at time 0 (= today) makes choices based on his attempted maximization of EUi 5 E c a e2rtUit d 5 E c a e2rt (uc (Xit)) d t

t

xit 5 | xit 2 a Txijt 1 a Txjit j

(12.1)

j

E [ Txijt ] 5 Txijt; E [ Tsijt ] 5 Tsijt , where the final lines hold for all elements of X across all times periods, and for all i within the set N (meaning that expected and actual transfers of goods and power between any two entities are equivalent). r is a discount factor and uc(Xit) the utility enjoyed by individual i that is derived from his final consumption. In this model, the individual maximizes his utility subject to expectations (=E[.]) that conform to reality, so in that sense he is fully rational. The technology of exchange and power are inputs into a ‘reciprocation function’, Txjit, that can directly depend on both the first-move transfers from i to j (Txijt) and the power applied to j by i (Tsijt). Indeed, all the elements | xit, E[Txijt] and E[Tsijt] should be understood as functions of the other elements. Since we are not concerned here with finding analytical steady-state solutions for transfer levels, but rather with describing a theory of love and what it means to be a rational decision maker in a world with love, there is no ex ante restriction we must impose on these functions. The formulation above is a simplification of the types of utility set-ups suggested by the extensive literature on choice behavior (for example, Neumann and Morgenstern 1944; Fishburn and Rubinstein 1982; Prelec and Loewenstein 1991): it involves linear and intertemporal separability of particular utility items, exponential discounting, a lack of procedural irrationality (such as an inability to calculate or limited memory), no direct modelling of uncertainty, and no explicit utility role for procedural aspects of how the eventual allocation of goods comes about.2 In what follows, many of the mainstream extensions to this basic framework will emerge, but in a format different from that used by others, using this simple and flexible model as a starting point.

M4225-ALTMAN_9781782549574_t.indd 208

03/05/2017 08:20

Is it rational to be in love? 209 Within this formulation, the classic economic motivation of greed can be seen as a type of direct grab, whereby an individual can invest resources that beget an immediate payback in terms of Xit and/or sit. Such a ‘grab’ may involve a material quid pro quo, as is the case with classic voluntary trade, but may also involve threatened or actual physical domination, such as theft or rape, or other means of appropriation without (full market value) compensation. In its simplest form, a choice to be greedy is based on the expectation that an investment in an attempt to make a resource grab will have a direct payoff, that is, that a transfer of good a, Tajit, can be ‘paid for now’ by means of an investment in a good of recognized value b, Tbijt, or by the application of power: dE [ Tajit ] dTbijt

. 0,

dE [ Tajit ] dTsijt

.0

(12.2)

The expectation of payback from the ‘grab’ strategy can be rational because the individual expects a voluntary trade from entity j, or because he expects to get away with appropriating the desired good a from entity j. To make this model completely standard, the only adjustment required is to assume that power equals the market value of wealth (that is, that sit 5 g x | xit pxt , with pxt a ‘market’ price for good x that is identical and non-manipulable for everyone). However, in our use of the model we do not want to limit the concept of power to market income, since that would be equivalent to assuming that everything of value is for sale and that all individuals live in perfect markets as price-takers. Several subtle aspects of the representation above are of particular relevance to the topic of this chapter. First, the individual doing the maximizing is completely cognizant of his own feelings, as he is able to predict with perfection how he would feel about any possible state of the world. The decision-maker in the above model is hence a savant when it comes to himself, a paragon of self-knowledge that Socrates would have admired, allowing him to perform a quite incredible maximization routine. He has perfectly rational expectations about his future feelings when he is deciding to get married for the first time, have a child, buy his first car, or vote for the first time, and in that sense is not undergoing the excitement of ambiguity, nor is he discovering himself: he is calmly proceeding along the path that brings him greater expected utility than any other path he could have chosen. In reality, no one can be assumed to be that aware of himself or of the world that produces the final outcomes Xit and power quantity sit for each individual (in this sense, real individuals cannot avoid being only boundedly rational (Simon 1982)). This is unfortunate empirically, as there is no baseline population that truly behaves like the model, so the model’s use as a descriptive tool is exceedingly limited. To suggest the formulation above as possible even by approximation means viewing this incredible self-knowledge as an aspirational assumption – an obtainable goal for our decision-maker – rather than a reasonable descriptive assumption. There is an immediate and important corollary to the realization that the standard formulation is not a description of how any actual individual truly thinks. This corollary can be stated as the need to start with a different baseline model for any actual analysis of choices. Because this sounds (and is) a daunting task, it is tempting to see the baseline model as much more attainable than it really is and to present real-life behavior as, by

M4225-ALTMAN_9781782549574_t.indd 209

03/05/2017 08:20

210

Handbook of behavioural economics and smart decision-making

contrast, ‘irrational’ or ‘anomalous’. Other authors, for example, speak of deviations from rationality caused by perceptive or mental limitations (Simon 1955; Tversky and Kahneman 1974). This gives far too much credit to the aspirational model, however, which we argue should not be treated as a positive model: it merely defines a particular notion of rationality and admirable behavior, of which vastly slimmed-down versions may be useful in formulating analytically tractable models and empirical estimation. From that perspective, it is a second-order question what kind of mental limitation we should try to accommodate in extensions. Much more important is what whole class of behavior we should try to incorporate that the standard model has assumed to be irrelevant or beyond-scope. We argue that the loyalty of humans to others or, in other words, their social nature offers the simplest, most holistic, and hence most sensible direction to look towards when considering how to expand this aspirational model of mainstream economics. Socialization can influence individuals’ decisions of ‘what to be and what to aim for’, and can hence radically alter our understanding of the individual’s maximization problem, if only through the influence of parents and other groups that constrain and guide individuals as they develop. If it is accepted that socialization occurs and can be partly responsible for the real-world choices of individuals, this raises the question of whether it is truly ‘smart’ to aspire to operate like a greedy ‘rational’ human, and hence to neither truly care for others nor indulge in self-delusion of the sort that causes the set N to include unreal entities. From a prima facie empirical point of view, it seems quite unlikely that in the real world, this type of rational economic man is what a smart person would want to be: there is pervasive evidence that very religious and overly optimistic people, that is, those who would count as ‘self-deluded’ in the standard model, are quite a bit happier than others (Leung et al. 2005; Lewis et al. 2005). This should already give us pause for thought. The supposed ‘smartness’ reflected in the canonical parts of the standard model, namely, exponential discounting and rational expectations, is usually defended by pointing to the fact that an individual would change his mind if he did not discount exponentially, and that he could achieve better final outcomes (that is, a higher Xit at the end of period t) if he knew better how those outcomes came about. Hence, or so the argument goes, other ways of thinking and deciding would not be evolutionarily adaptive in a repeated setting with learning. While these claims are in themselves not always true in the presence of strategic considerations wherein pre-commitment matters,3 the main problem with such a defense of ‘rational man’ is that it does not point to a means of horse-racing the standard model against clearly articulated alternatives: in which type of operating environment should we consider the alternatives? Implicitly, an adherent of this view must presume that Uit is predicted perfectly by Xit (that is, that preferences are fixed), so that under perfect-market circumstances (that is, the absence of strategic considerations that open a role for irrationality) we can make the best plans to maximize our utility through our choice of consumption levels. Yet, having a modus operandi that maximizes consumption in a perfect market environment will not work out so well in other environments, nor when consumption is not the thing one aims for, nor when utility functions are flexible. Another important hidden element in the standard formulation is that an individual’s notion of ‘Self’ is taken to be self-evident and fixed: it is an entity towards which a person displays absolute altruism. The Self in the standard formulation is taken as an unchanging

M4225-ALTMAN_9781782549574_t.indd 210

03/05/2017 08:20

Is it rational to be in love? 211 entity about which an individual does, and even should, care about in the future to an unchanging degree. As Ng (1992) and others have said, that kind of formulation presumes the existence of something like a soul that is unchanging and that provides solidity to the notion of Self that an individual cares about. In reality of course, this is an idealized abstraction: people change in a myriad of ways over time and there is no a priori reason why it should be rational or evolutionarily adaptive for them to care about their future selves rather than, for example, just caring about the momentary pleasure obtained from their current self’s experiences. How strange it is, on reflection, to assume that the self is fixed. Individuals continuously change form and even change the make-up of their bodies, as they experience a constant exchange of fluids, solids, and even genetics with the outside world. Our genome is more like a moving cloud than a fixed point, with our genes exchanging genetic material with microbes and constantly changing as a result of cellular processes. Even in terms of mental traits, individuals routinely change their minds, their attitudes, and their preferences over time, often quite dramatically over the whole of the life course as new ideologies and religions come into being. What is fixed in reality is less the mind and body of an individual, but more his social endowments, such as his possessions, his passport and associated ‘rights’, his kin relations with others (as father, son, and tribe member, for example), and so on. The idea that it is somehow smart or optimal to care for some discounted flow of benefits towards a fixed ‘Self’ is not grounded in any economic logic or underlying foundation of rationality: it is itself a convention, an advocated position, a choice to buy into a particular abstraction being offered by a group (in this case, mainstream economists). Finally, the treatment of time in the standard model is also not as ‘rational’ as it might seem at first glance. Not only does the model assert that an individual is wholly uninterested in and unresponsive to the past, but the future is also purportedly seen as fundamentally different from the present: the present is taken as known, and the future is taken as expected. ‘Rational’ economic man does not care about his history or his ancestors, observes everything about the present with total objectivity, and is completely detached when forming expectations about the future. Purely from a psychological and neuroscientific point of view, this is an odd proposal. There is no clear way for a human mind to think in different ways about the past, present, and future. All these windows of time can be experienced and reacted to in the same way, using the same neuronal hardware and pathways. Consider fear, which can arouse great emotions in the present moment even if the fear does not materialize in the future, or can arise in reaction to a remembered childhood fairytale: the feared event is an imagined outcome that creates utility effects in the present moment when it is held in the mind, independent of the supposed timing of the feared event. The past, too, can therefore arouse emotions, and the image of the past is subject to constant re-writing and re-interpretation, much like the present is only experienced through the filter of our senses, rather than being objectively observed. The standard model’s artificial distinctions between a future that is coolly expected, a present that is certain, and a past that is coolly and totally ignored, are neither realistic nor obviously desirable from a welfare or evolutionary perspective. In sum, the greedy, rational individual lives in a delusion of fixity, and has accepted that his own current and future consumption are the only things that matter, but otherwise sees the world as it truly is. From this starting point, our next step is to include a more dynamic notion of Self.

M4225-ALTMAN_9781782549574_t.indd 211

03/05/2017 08:20

212

Handbook of behavioural economics and smart decision-making

3.2

Naïve Love

How can we conceptualize a changing Self, in a way that is both realistic and amenable to incorporation into the analytical framework of economics? Previous attempts to broaden the Self studied by economists (for example, Ainslie 1985) have been written in philosophical terms and/or focused on the formalization of only relatively small details of the arguably very complex problem. Our proposal by contrast is to consider a large addition – the possibility of being in love – and model this directly as an internal experience that drives change in the Self, leading to changed loyalties. We can incorporate both the process of change and the outcome into the economic toolkit by constructing individuals who have the capacity to start to care about a more expansive notion of Self that includes the feelings and experiences of others. An individual who has come to ‘love’ (or to ‘be loyal’) is someone who cares for another entity, j, to a degree bijt ≥ 0. Someone who loves in what we term a naïve way, but in all other ways conforms to the supposed rationality of the standard formulation, will then maximize EUi 5 E c a a bijt e2rtUjt d 5 E c a a bijt e2rt uj (Xjt) d 5 a a bij0 e2rt uj (Xjt) N

t

j51

E [ bijt ] 5 bij0

N

t

j51

N

t

j51

(12.3)

This formulation denotes a situation in which the individual anticipates at time 0 a fixed degree to which he cares for others in all periods of the remaining future (where this degree is denoted bij0), and this expectation of the fixity of love is a mark of his naïveté. He maximizes the streams of future utility towards those entities to which he is loyal, on the basis that his loyalties are unchanging from today onward, without wondering where those loyalties came from. The individual is now incredibly cognizant not only of his own psychology, but also of the psychology of those he presently loves, being able with perfect foresight to anticipate how they would feel under various circumstances. The naïvely loving individual is supremely loyal to a fixed notion of Self that now includes not merely his own ‘soul’ (via bii0, a term which is included in the summation above over all entities N), but also those of others. Despite the ignorance it assumes about how attachments change, this formulation allows for the same kind of rationality as before: the individual sees the world and himself as they truly are. 3.3

Being in Love

How then does bijt develop over time? This question essentially asks for a theory of love, not only for one’s own ‘experiencing self’ (for lack of a better word) but for any outside entity towards which one might plausibly develop love. We have proposed such a uniform theory of love elsewhere (Frijters and Foster 2013), which we call the love principle, and here extract its core implications for the mathematics and intuition of the present set of models. We simplify the argument by presuming throughout that individuals want the good a from the entity they will end up loving, just as we used that same good a as the desired acquisition target when describing greed. We argue that love increases when an individual believes unconsciously that increasing

M4225-ALTMAN_9781782549574_t.indd 212

03/05/2017 08:20

Is it rational to be in love? 213 his love for j will increase future transfers to him by j of the desired good a that cannot be obtained from j via direct grabbing. Because being in love involves the change of self, the individual is effectively investing part of his ‘soul’, in the expectation of reciprocity from the loved entity in the form of the transferral of the desired good. The simplest way to express mathematically what the loving mind perceives is as follows: dE [ Ta jit1s ] dbijt

. 0,

dE [ Ta jit1s ] dTbijt

0 bijt # 0,

dE [ Ta jit1s ] dTsijt

0 bijt # 0

(12.4)

Where the unconscious expects that the entity holding the desired good a will respond, in terms of goods transferral, to an investment of ‘self’ (dbijt > 0). The conscious belief that (additional units of) the desired good cannot be obtained by transferring goods or by using ] ] dE [ Ta dE [ Ta power without an investment of the self ( dTb 0 bijt and dTs 0 bijt) 4 is what makes  the investment of self optimal, as the usual ‘grabbing’ strategy of the conscious is thwarted. ] dE [ Ta Whether an individual consciously expects that d b .0 is ambiguous in our theory, but we argue the generic answer is likely to be ‘no’, meaning that the resulting shift in the Self comes as a surprise to most people. We view the period of change in loyalty thereby much like a unconsciously regulated bodily process such as our internal circadian rhythm: we can put ourselves in a situation where our internal clock gets reset, and can even take substances that help with that resetting, but we cannot consciously direct the dials of our internal clock. Purely on empirical grounds, it appears to be the same with love. The love principle presented in Frijters and Foster (2013) contends that the individual is not typically reflective about this expectation of the reciprocity initialized by love, and does not wonder exactly how the supposed reciprocal transfers will come about. Part of the evidence for this contention is the lack of interest that people show in questions about the mechanics of things like ‘Karma’ and ‘a good afterlife’. How are these things actually organized? How do ‘God’ and even ‘our partner’ actually come to care for us and give us transfers? This is a subject of surprisingly little critical thinking, which we argue is because it is not the conscious mind that does the expecting, but rather the unconscious. Frijters and Foster (2013) suggest that our expectation of reciprocity by the entities we love must be hidden from our critical thinking because our conscious mind would feel its self-esteem diminished by an open admission of weakness (embodied in the inability to ‘grab’ the outside entity’s resources). This then gives rise to rational self-delusion about the love mechanism itself: in order to submit but still feel good about it, we do not tell ourselves that we submit, but instead pretend that love arrives unexpectedly. Another aspect of reality that this formulation accommodates is that believed future transfers may never actually arise. Examples are transfers that supposedly take place after death, or that involve a cure for incurable illnesses. Actual transfers, such as when a nation rewards its war veterans, can of course also arise. Whether the expected transfer eventually occurs or not, our contention is that love is quintessentially about a potentially deluded unconscious mind that craves a transfer. We argue that it is important for the maintenance of love that the lover now and then sees believed confirmation of his expectation of reciprocity, such as signs from god or tokens of goodwill, but still, no actual transfers need ever occur in reality. ] dE [ Ta Looking more carefully at the process underlying the statement that d b . 0 we have jit 1 s

jit 1 s

ijt

ijt

jit 1 s

ijt

jit 1 s

ijt

M4225-ALTMAN_9781782549574_t.indd 213

03/05/2017 08:20

214

Handbook of behavioural economics and smart decision-making

in mind that an individual believes the entity possessing the desired good a is interested in our loyalty as well as other goods: it wants power over us (directly via bijt) but also actual goods transfers from us that prove that power over us. While in love, a person will therefore transfer goods k at time t to the loved entity, that is, Tkijt > 0, where k includes goods believed to be valuable to the loved entity. The bundle of {bijt, Tkijt} is then transferred in the expectation of a reciprocal but uncertain transfer of a in the future. Transfers to the loved entity in our theory thereby come about both from believing that the loved entity receives utility from the gift, and as a means of quelling internal doubt about the function E[Tajit+s]. In the background, an individual will implicitly have some notion of the elements in the utility function of the loved entity, even if that entity does not exist. Relevant to this, Frijters and Baron (2012) use a laboratory experiment to examine gift-giving to an abstract entity they called ‘Theoi’, which in reality was a computer algorithm randomizing its decisions about transfers in the form of ‘market prices’. The authors hypothesized that participants believed Theoi’s utility function was of the following form: UTheoi,t 5 a [ 2 TaTheoi,i,t1s 1 h (TaTheoi,i,t1s 2 pk *Tki,Theoi,t) ] i

hr . 0, hs , 0

(12.5)

where Tki,Theoi,t took the form of an allocation of real money, and TaTheoi,i,t+s the later setting of market prices. This formulation has the key characteristic, via the function h(.), that Theoi’s marginal utility of rewarding i with a gift of good a (=TaTheoi,i,t+s) increases with more valuable transfers from i to Theoi (= Tki,Theoi,t), with pk just picking up relative price effects (the locally perceived relative worth of good a compared to good k).5 The belief system of individual i about the loved entity’s utility function is crucial in determining the behavioral ramifications of loyalty. For example, if individual i believes that the entity only wants particular goods (like certain Greek gods who were allegedly only concerned with burned meat), then that is what is transferred. If the individual i believes the entity wants loyalty (like most Greek gods purportedly did), then that is what is invested.6 We do not discuss where such beliefs come from, because a cursory glance at history tells us that it is highly context-dependent. We merely note the bewildering array of things that people can believe non-existent entities care about, including adherence to rules of behavior, high art, dance, sex, poetry, life, and on and on. As a general rule we suggest that non-existing entities are often believed to care about exactly the same things that are deemed valuable in the society from which the believer comes, and in that sense Tki,Theoi,t will often overlap with the socially-determined notion of power dTs (that is, dTk . 0), such that we must transfer something of our own (purchasing) power in order to make an impression on the non-existing entity. If the unconscious were choosing optimally according to this belief, the optimizing investment would then solve ] 0E [ Ta 0EU dEU ] 50 0 E[Ta 1 dTk which would thus nail down the transfer Tki,Theoi,t if 0Tk 0Ta we make particular choices about the underlying functions. i, Theoi, t

i, Theoi, t

Theoi, i, t 1 s i, Theoi, t

i

Theoi, i, t 1 s

i

Theoi, i, t 1 s

i, Theoi, t

3.3.1 The love bargain If we think of all potentially loved entities j as being like the artificial entity ‘Theoi’ above in terms of real transfers, then we can map the reaction function of these entities into the

M4225-ALTMAN_9781782549574_t.indd 214

03/05/2017 08:20

Is it rational to be in love? 215 utility-maximization program of the rational individual. We have two choices as to how to model this: we could focus on individuals’ belief structures surrounding the entities j and incorporate those explicitly into the model, or we could focus on how the result of those belief structures adjusts the Self over time, as transfers are made to the entities j, and add the elements that this process implies to the maximization problem for individuals. We opt here for that second approach. We thus propose a mathematical description of how love changes in response to transfers, themselves optimally chosen based on the unconsciously imagined reactions of outside entities that are believed to demand our loyalty in return for entertaining a bargain. We propose that love increases in those time periods when an individual transfers more power to an entity than he receives back, where power now certainly includes purchasing power (that is, goods) but also anything else that would be understood by the individual to influence his environment and people around him: bijt11 5 f ( bijt,pi,kTkijt 2 E [ Tajit ]) 0 2f f (0,0) 5 0, f r . 0, ,0 0bijt 0 ( pi,kTkijt 2 E [ Tajit ]) dTsijt ~ pi,k . 0 dTkijt

(12.6)

This formulation states that individual i’s love towards entity j is path-dependent to an extent, but also increases when i transfers good k to entity j at a moment when j is not perceived to transfer the desired good a back to i, with pi,k the relative price as perceived by i (which allows for individuals to want different things). The imagined current expected transfer is denoted as E[Tajit], again allowing individuals to believe that they are getting transfers from non-existent entities. The investment of power is made under the expectation that the power transfer in turn engenders a transfer towards oneself in the future (in periods s > t), but does not involve other expected trades (a form of ceteris paribus condition: there are no other benefits expected from the investment of Self).7 The dependence on the existing loyalty is such that the higher the existing loyalty, the higher the transfer dTs must be that maintains that loyalty: 0b 0 (p Tk 0 f2 E [ Ta ] ) , 0. The assumption dTk ~ pi,k denotes the idea that the perceived value of the transfer to the loved entity in the mind of the person giving it is proportional to its relation with (socially defined) power: the higher is pi,k and thus the more an individual is giving up part of his social power for an outside entity in return for an uncertain return transfer in future periods, the faster the increase in loyalty towards that entity. This law of motion will vary in quality from person to person, depending on what the person believes the outside entity will react to, which could be time, fervency, physical goods, and so on – which we view, through the lens of the economist, as information about the outside entity’s utility function. The actual good transferred by the loving person reveals what that loving person believes the loved entity to care about. For instance, if the outside entity only cares about loyalty and nothing else, then the good k would merely be equal to bijt, but generically the outside entity will be presumed to care about loyalty as evidenced by visible transfers. Love of j by i, bijt, is decided upon this period on the basis of expected future transfers 2

ijt

M4225-ALTMAN_9781782549574_t.indd 215

i, k

ijt

ijt

jit

ijt

03/05/2017 08:20

216

Handbook of behavioural economics and smart decision-making

and an individual acts upon his love in this period by making transfers towards the loved entity. The transfers can range from writing love poems for a wooed woman, to taking risks in war for a loved cause, to burning meat for a wooed god. In a way, love is an answer to the problem of missing markets, either for goods that do not exist at all, or among entities that have no power to force transfers otherwise. It is intimately tied to the notion of power, in that the individual must care about whatever it is that he is giving up to the wooed entity dTs (and hence dTk is proportional to the relative price). If we think of power as purchasing power, which is a highly culturally specific interpretation of how one can influence others, then increasing love involves giving up some purchasing power (time, money, or goods) for the supposed benefit of the loved entity. Our love principle pre-supposes that individuals are not aware of the production technology giving rise to these expected future transfers Tajit+ s, as it is the unconscious mind, rather than the conscious mind, that makes the love bargain. ijt

ijt

3.4

Types of Rationality

If we take this line of reasoning a step further, different types of ‘rationality’ and notions of Self come into view based on the degree to which an individual is aware of how loves may ebb and flow over time. Using a simple linear parametrization, where qi will denote a level of rhytonia (explained below), gi a level of self-rationality, and hi a level of worldrationality, the maximization problem for individual i at t = 0 can be re-stated as: N N | | | EUi 5 E c a a b ijte2rtUjt d 5 a a b ijte2rt auj aE c Xjt 1 a TXmjt 2 a TXjmt d b b t

t

j51

m

j51

m

| b ijt 5 qi bijt 1 (12qi)bij0
F

0.5061 2.4023 0.9281 0.8156 2.0202 0.0124 0.0058

3 2 1 4 1 3 1

0.1687 1.2012 0.9281 0.2039 2.0202 0.0041 0.0058

0.56 4.39 3.24 0.69 7.35 0.01 0.02

0.6396 0.0150 0.0750 0.6014 0.0079 0.9978 0.8891

Notes: a 1 5 not important in selection of seed, 2 5 important, 3 5 very important. b 0 5 chronic stressor, 1 5 catastrophic stressor. c 0 5 none, 1 5 little, 2 5 much. d 0 5 no WTP, 1 5 WTP is greater than 0 *** Significant at the .01 level; ** significant at the .05 level; * significant at the .10 level.

As some farmers reduced their ratings of seeds’ drought tolerance trait while others increased their ratings, we take the absolute value of ratings changes for drought tolerance to test for whether different factors are associated with any change in seed trait ratings. For comparison, Appendix 14.2 includes tables with our evaluations of drivers of seed trait changes for excess rain tolerance, as several villages had significant changes in their use of seeds with this trait. To test whether the correlations are significant, we first conducted one-way analyses of variance (ANOVA) for the absolute change in ratings of seeds’ drought tolerance and baseline measures of farmer attitudes (Table 14.7). The only non-categorical independent variables – average yield loss and willingness to pay (WTP) – were divided into quintiles using the egen command in Stata. The results of the ANOVA tests (Table 14.7) indicate that there is not a statistically significant relationship between changes in farmers’ selection of seed toward droughtresistant seed and baseline willingness to take risks, average yield loss from drought, or willingness to pay to reduce the likelihood of catastrophic losses from drought. Perceived control over losses from drought, however, is significant at the .01 level, and farmers’ stated importance of drought-tolerance in seed selection and perceptions of drought as a catastrophic stressor are also significantly associated with changes toward droughttolerant seed, at the .05 and .10 level, respectively. To further test these associations, we conducted simple ordinary least squares (OLS) regressions of the absolute change in ratings of seed drought-tolerance on farmers’ primary plots against potential drivers of seed change, retaining baseline measures of willingness to take risks and the three variables that appear to have significant associations with changes in seed traits (Table 14.8). We also include some control variables that may be expected to influence seed selection, including age, literacy, and number of plots planted with maize. We hypothesize that farmers with a greater number of plots would be

M4225-ALTMAN_9781782549574_t.indd 266

03/05/2017 08:20

Do changes in farmers’ seed traits align with climate change? 267 Table 14.8

Results of OLS regression for absolute change in mean rating of seed drought-tolerance on primary plot

Variable Willingness to take risks Importance of drought-tolerant seed traita Perceived importance of drought stressorb Perceived control over losses from droughtc Age Knowledge of reading and writing Number of plots planted with maize Melchor Ocampo Roblada Grande Queretaro Observations Adjusted-R2

Coeff. (std. err.)

P > |t|

−0.0263 (0.0608) 0.2196 (0.1302) 0.1662 (0.1736) 0.7486 (0.3245) −0.0045 (0.0041) −0.2844 (0.1478) 0.0038 (0.0538) 0.1489 (0.1663) −0.1643 (0.1709) −0.0813 (0.1874) 90 0.1155

0.667 0.096* 0.341 0.024** 0.276 0.058* 0.944 0.373 0.339 0.666

Notes: a 1 5 not important in selection of seed, 2 5 important, 3 5 very important. b 0 5 chronic stressor, 1 5 catastrophic stressor. c 0 5 none, 1 5 little, 2 5 much. ** Significant at the .05 level; * significant at the .10 level.

more willing to change their selection of seed traits on a given plot, as they would still be able to use previously tested seeds on their other plots. In addition, we also include variables for farmer villages, as we have seen that farmers in certain villages have had more significant changes in their seed trait selection. As with the ANOVA analysis, we find no statistically significant relationship between changes in farmers’ rating of seed traits and willingness to take risks. We also find that the perception of drought as a catastrophic stressor is no longer statistically significant. Farmers’ stated importance of drought tolerance in seed selection and perceived control over losses from drought, however, remain statistically significant and associated with larger changes in the drought tolerance of seeds on the primary plot. Perceived control over losses from drought appears to have the largest effect on selection of seed traits, with differences in control between ‘none and a little’, or ‘a little and much’, associated with three-quarters of a unit change in the four-point seed rating scale. We find that farmers’ age, village, and number of plots planted with maize are not significantly associated with changes in drought-tolerant traits of planted seed.

M4225-ALTMAN_9781782549574_t.indd 267

03/05/2017 08:20

268

Handbook of behavioural economics and smart decision-making

Unexpectedly, both Roblada Grande and Queretaro have negative coefficients, indicating smaller absolute changes in seed drought tolerance from 2005–07, even though farmers in both villages increased their ratings of their seed’s drought tolerance, on average. This negative association may be because farmers in these villages had smaller absolute changes in drought tolerance ratings than farmers from Dolores Jaltenango, although these changes are not significant. Knowledge of reading and writing, however, is significantly but negatively associated with changes in drought tolerance ratings. To test the robustness of our findings, we also conducted regression analyses looking at relative rather than absolute changes in drought tolerance ratings and looking at the average change in drought tolerance ratings across all farmer plots as opposed to just on their primary plot (not presented). The significance of different variables varies somewhat across these models. When considering relative changes in drought tolerance ratings of seeds on the primary plot, only the dummy variable for the village of Melchor Ocampo is significantly associated with a change in ratings of seeds on the primary plot. For relative changes in ratings for seeds across all plots, the Melchor Ocampo dummy variable along with the stated importance of drought-tolerant traits and perceived control over losses from drought are all significant. For absolute differences in seed trait ratings across all plots, only perceived control over losses from drought is significant. We therefore observe that farmer’s stated importance of drought-tolerant seed traits is significantly associated with changes in selection of drought-resistant seed in two of the four models, while farmers’ perceived control over losses from drought is significant in three models. These findings suggest that farmers’ seed selection decisions are associated with their perceptions of climate change and of their ability to respond to climate change, though the association is not always clear. The survey does not ask farmers about their perceived importance of or control over excess rain as a stressor or about their willingness to pay to reduce losses from excess rain, but we conducted a similar analysis as for drought tolerance considering farmers’ selection of seeds with excess rain tolerance, without these variables (Appendix 14.2). While farmers’ stated importance of excess rain-tolerance in seeds is significant in the ANOVA analysis, none of the variables are significantly associated with absolute changes in excess rain tolerance ratings in the OLS regressions. In models using relative changes in excess rain tolerance ratings, the coefficient for Queretaro is significant when considering seed traits on all plots, while the coefficients for Roblada Grande and literacy are significant when considering seeds on the primary plot only. This finding suggests that village-level factors may play an important role in farmers’ seed selection decisions.

5

CONCLUSION

Our research shows that farmers in four villages of Chiapas, Mexico, changed their seed ratings of tolerance or resistance to four environmental stressors, most notably drought tolerance, although average changes differed by village. Changes in ratings of drought and excess rain tolerance are generally aligned with climate change predictions for temperature and precipitation in these villages, though the degree of alignment varies by village and depending on the climate model we use. Not unexpectedly, farmers’ changes in seed trait ratings do not perfectly correspond to climate change predictions, as climate variations

M4225-ALTMAN_9781782549574_t.indd 268

03/05/2017 08:20

Do changes in farmers’ seed traits align with climate change? 269 are uncertain, and as current seed trait choices may be based on more short-term climate change expectations than those in our models. While we cannot test whether changes in seed trait ratings are deliberate adaptions to climate change, we find that farmers’ baseline attitudes may partially motivate changes in seed trait ratings. Although willingness to take risks does not appear to affect farmer seed selection, farmers’ stated importance of drought tolerance in seed selection and their perception of control over losses from drought in 2005 are both associated with larger absolute changes in seed drought tolerance ratings between 2005 and 2007. On the other hand, literacy appears to decrease the likelihood of changes in ratings for this trait, though the possible reasons for this association are not clear. The concept of bounded rationality suggests that individual rationality in decisionmaking is constrained by information availability, individuals’ capacity to evaluate and process information, and time available to make decisions. Our results suggest that farmers’ selection of seed agronomic characteristics, whether knowingly or not, are aligned with long-term climatic fluctuations owing to climate change as predicted by climate models, and that baseline attitudes towards different stressors and farmers’ education may also play a role in selection of seed traits. Our findings are limited by the small sample size and by the relatively short timeframe of the study when compared with timelines for climate change, but are generally robust to several model specifications. This study lays a foundation for future investigation into what other variables may drive farmers’ climate adaptation behaviors given rational behavior under enormous uncertainty.

NOTES 1. Digital Climate Atlas of Mexico: http://uniatmos.atmosfera.unam.mx/ACDM/servmapas and National Meteorological Service of Mexico Weather Stations: http://smn.cna.gob.mx/index.php?option5com_cont ent&view5article&id542&Itemid575 (both accessed 11 January 2017). 2. Note that both sources had the same average for village of Queretaro in the HADGEM 1 temperature model (Figure 14.2). 3. The model at that time was the predecessor to the HADCEM 3 – the HADCM 2. Note that all HAD-rooted models stem from the Hadley Centre’s larger Unified Model, but vary depending on the necessary application (seasonal, decadal and centennial climate predictions). 4. The survey does not include questions on perceptions of or losses from excess rain. 5. Figures showing expected temperatures for the GFDL CM3 model are included in Appendix 14.1. 6. The baseline average temperatures for Queretaro are the same for both sources, hence only one line. 7. The grey parallel lines on the bar graphs represent the baseline averages from two different sources. 8. Note that we are suggesting a relationship between excess rain and rotting, as excess rain can lead to rotting of the maize crop, and as changes in seed ratings for these two traits appear to be associated with one another. 9. Farmers in Dolores Jaltenango did not significantly change their ratings of seed traits with the exception of wind tolerance, so we cannot evaluate whether changes align with climate change predictions.

REFERENCES Anderson, C.L., M. Bellon and L. Lipper (2012), ‘The effect of environmental sources of crop loss on farmer willingness to pay for maize seed varieties in Chiapas, Mexico’, Evans School Working Paper No. 2006-02, Evans School Policy Analysis & Research Group (EPAR), University of Washington, Seattle.

M4225-ALTMAN_9781782549574_t.indd 269

03/05/2017 08:20

270

Handbook of behavioural economics and smart decision-making

Bellon, M. and S. Brush (1994), ‘Keepers of the maize in Chiapas, Mexico’, Economic Botany, 48 (2), 196–209. Bellon, M and J. Hellin (2011), ‘Planting hybrids, keeping landraces: agricultural modernization and tradition among small-scale maize farmers in Chiapas, Mexico’, World Development, 39 (8), 1434–43. Bellon, M., M. Adato, J. Becceril and D. Mindek (2006), ‘Poor farmers’ perceived benefits from different types of maize germplasm: the case of creolization in lowland tropical Mexico’, World Development, 34 (1), 113–29. Bellon, M., D. Hodson and J. Hellin (2011), ‘Assessing the vulnerability of traditional maize seed systems in Mexico to climate change’, Proceedings of the National Academy of Sciences, 108 (33), 13432–7. Borja-Vega, C. and A. de la Fuente (2013), ‘Municipal vulnerability to climate change and climate-related events in Mexico’, World Bank Policy Research Working Paper 6417, World Bank, Washington, DC, doi:http:// dx.doi.org/10.1596/1813-9450-6417. Bryan, E., T.T. Deressa, G.A. Gbetibouo and C. Ringler (2009), ‘Adaptation to climate change in Ethiopia and South Africa: options and constraints’, Environmental Science & Policy, 12 (4), 413–26. Bubeck, P., W.J.W. Botzen and J.C.J.H. Aerts (2012), ‘A review of risk perceptions and other factors that influence flood mitigation behavior’, Risk Analysis, 32 (9), 1481–95. Burton, I. (1997), ‘Vulnerability and adaptive response in the context of climate and climate change’, Climatic Change, 36 (1–2), 185–96. Byerlee, D. and J.R. Anderson (1982), ‘Risk, utility and the value of information in farmer decision making’, Review of Marketing and Agricultural Economics, 50 (3), 231–46. Conde, C., F. Estrada, B. Martínez, O. Sánchez and C. Gay (2011), ‘Regional climate change scenarios for Mexico’, Atmósfera, 24 (1), 125–40. Conde, C. and D. Liverman, M. Flores, R. Ferrer, R. Araujo, E. Betancourt et al. (1997), ‘Vulnerability of rainfed maize crops in Mexico to climate change’, Climate Research, 9 (1–2), 17–23. Corral, J.A., N.D. Puga, J. de J.S. González, J.R. Parra, D.R.G. Eguiarte, J.B. Holland and G.M. García (2008), ‘Climatic adaptation and ecological descriptors of 42 Mexican maize races’, Crop Science, 48 (4), 1502–12. Fernández, A.T., T.A. Wise and E. Garvey (2012), ‘Achieving Mexico’s maize potential’, Global Development and Environment Institute Working Paper No. 12-03, Tufts University, Medford MA, accessed 12 December 2013 at http://www.ase.tufts.edu/gdae/Pubs/wp/12-03TurrentMexMaize.pdf. Food and Agriculture Organization (FAO) (2006), ‘Country pasture/forage resource profiles: Mexico’, FAO, Rome, accessed 8 May 2014 at http://www.fao.org/ag/AGP/AGPC/doc/Counprof/Mexico/Mexico.htm. Food and Agriculture Organization (FAO) (2014), FAOSTAT – crop production by country, FAO, Rome, accessed 8 May 2014 at http://faostat3.fao.org/faostat-gateway/go/to/download/Q/QC/E. Gebrehiwot, T. and A. van der Veen (2013), ‘Farm level adaptation to climate change: the case of farmers in the Ethiopian Highlands’, Environmental Management, 52 (1), 29–44. Hansen, J.W., S.M. Marx and E.U. Weber (2004), ‘The role of climate perceptions, expectations, and forecasts in farmer decision making: the Argentine Pampas and South Florida: final report of an IRI seed grant project’, IRI Technical Report 04-01, International Research Institute for Climate Prediction, Earth Institute, Columbia University, New York. Jones, P. and P. Thornton (2003), ‘The potential impact of climate change on maize production in Africa and Latin America in 2055’, Global Environmental Change, 13 (1), 51–9. Lobell, D. and M. Burke (2008), ‘Why are agricultural impacts of climate change so uncertain? The importance of temperature relative to precipitation’, Environmental Research Letters, 3 (3), doi:10.1088/1748-9326/3/3/034007. Lujala, P., H. Lein and J.K. Rød (2014), ‘Climate change, natural hazards, and risk perception: the role of proximity and personal experience’, Local Environment: The International Journal of Justice and Sustainability, 20 (4), 489–509. Maddison, D.J. (2007), ‘The perception of and adaptation to climate change in Africa’, World Bank Policy Research Working Paper 4308, World Bank, Washington, DC. Mase, A.S. and L.S. Prokopy (2014), ‘Unrealized potential: a review of perceptions and use of weather and climate information in agricultural decision making’, Weather, Climate, and Society, 6 (1), 47–61. Mercer, K. and H. Perales (2010), ‘Evolutionary response of landraces to climate change in centers of crop diversity’, Evolutionary Applications, 3 (5–6), 480–93. Mercer, K., H. Perales and J. Wainwright (2012), ‘Climate change and the transgenic adaptation strategy: Smallholder livelihoods, climate justice, and maize landraces in Mexico’, Global Environmental Change, 22 (2), 495–504. Moyo, M., B.M. Mvumi, M. Kunzekweguta, K. Mazvimavi, P. Craufurd and P. Dorward (2012), ‘Farmer perceptions on climate change and variability in semi-arid Zimbabwe in relation to climatology evidence’, African Crop Science Journal, 20 (2), 317–335. Pressoir, G. and J. Berthaud (2004a), ‘Patterns of population structure in maize landraces from the Central Valleys of Oaxaca in Mexico’, Heredity, 92 (2), 88–94. Pressoir, G. and J. Berthaud (2004b), ‘Population structure and strong divergent selection shape phenotypic diversification in maize landraces’, Heredity, 92 (2), 95–101.

M4225-ALTMAN_9781782549574_t.indd 270

03/05/2017 08:20

Do changes in farmers’ seed traits align with climate change? 271 Seo, S.N. (2012), ‘Decision making under climate risks: an analysis of sub-Saharan farmers’ adaptation behaviors’, Weather, Climate, and Society, 4 (4), 285–99. Servicio Meteorológico Nacional (Meteorological Service of Mexico) (SMN) (2014), Weather Stations, accessed 5 August 2014 at http://smn.cna.gob.mx/index.php?option5com_content&view5article&id542& Itemid575, last updated 2014. Simon, H.A. (1982), Models of Bounded Rationality: Empirically Grounded Economic Reason, vol. 3, Cambridge, MA: MIT Press. Smit, B. and M.W. Skinner (2002), ‘Adaptation options in agriculture to climate change: a typology’, Mitigation and Adaptation Strategies for Global Change, 7 (1), 85–114. Smithers, J. and B. Smit (1997), ‘Human adaptation to climatic variability and change’, Global Environmental Change, 7 (2), 129–46. Stokes-Prindle, C., B. Smoliak, A. Cullen and C.L. Anderson (2010), ‘Crops & climate change: maize’, White Paper, Evans School Policy Analysis & Research Group (EPAR), University of Washington, Seattle. Sutherland, L.A., R.J. Burton, J. Ingram, K. Blackstock, B. Slee and N. Gotts (2012), ‘Triggering change: towards a conceptualisation of major change processes in farm decision-making’, Journal of Environmental Management, 104 (August), 142–51. Tambo, J.A. and T. Abdoulaye (2012), ‘Climate change and agricultural technology adoption: the case of drought tolerant maize in rural Nigeria’, Mitigation and Adaptation Strategies for Global Change, 17 (3), 277–92. Ureta, C., E. Martínez-Meyer, H.R. Perales and E.R. Álvarez-Buylla (2012), ‘Projecting the effects of climate change on the distribution of maize races and their wild relatives in Mexico’, Global Change Biology, 18 (3), 1073–82. US National Oceanic and Atmospheric Administration (NOAA) (2004), North American Drought Monitor Maps, September–December 2004, accessed 9 January 2014 at http://www.ncdc.noaa.gov/temp-and-precip/ drought/nadm/nadm-maps.php. Vigouroux, Y., C. Mariac, S. De Mita, J. Pham, B. Gérard, B. et al. (2011), ‘Selection for earlier flowering crop associated with climatic variations in the Sahel’, PLoS One, 6 (5), e19563. Waddington, S.R. (2014), Personal email correspondence with C. Leigh Anderson, 9 January.

M4225-ALTMAN_9781782549574_t.indd 271

03/05/2017 08:20

272

Handbook of behavioural economics and smart decision-making

APPENDIX 14.1

GFDL CM3 MODELS Expected village temperature in April 2015–39

27.5 27.0