Political Representation: Roles, representatives and the represented 9781138956278, 9781315665788

Table of contents :
Cover
Half Title
Title Page
Copyright Page
Table of Contents
List of illustrations
List of contributors
Acknowledgement
1. Introduction
A crisis of representative democracy?
Thinking about representation: sanction vs. selection
Theory meets empirics I – the relationship of representatives
and represented
Theory meets empirics II – issue congruence revisited
Exclusion – or who actually is “the represented”?
Note
References
2. A “selection model” for political representation
Introduction: sanctions v. selection
Accountability
Conclusion: the normative side of selection
Notes
References
3. Beyond trustees and delegates
Introduction
Trustees vs. delegates in historical context
Separating the three distinctions and keeping them separate
Whither representation?
Conclusion
Notes
References
4. Institutional constraints and territorial representation
Introduction
How constituencies are defined
Hypotheses
Representational role orientations in 13 European democracies
Institutional effects on territorial representation
Institutional effects on the focus of representation
Territorial representation by national and regional legislators
Conclusion
Notes
References
5. Promises and lies: an empirical comparison of Swiss MPs’ pre- and post-electoral positions
Introduction
Pledge fulfilment in the literature
What we look at: database and dependent variable
What factors might explain a change of mind?
Multivariate analysis and discussion of the results
Conclusion
Notes
References
6. Beyond congruence
Introduction
The normative underpinnings of congruence and its critical shortcomings
Situating congruence in the eighteenth-century prejudice against representation
From mandate to mobilization
Conclusion: interactivity – moving empirical research
beyond congruence?
Notes
References
7. Dimensionality of the European issue space
Introduction
Dimensionality of the issue space in European democracies
Method
Results
Discussion
Notes
References
8. Representation of political opinions: is the structuring pattern of policy preferences the same for citizens and elites?
Introduction
Dimensionality of political competition
Citizens’ and elites’ belief systems
The pattern of policy preferences of citizens and elites in the Swiss 2007 elections
Discussion
Notes
References
9. “Alignment of objectives” between parties and their electors: the role of personal issue salience in political representation
Introduction
Are we missing a dimension? On personal issue salience and political representation
Why does issue salience matter for the alignment of policy objectives?
Why does issue salience matter for citizens’ selection of parties?
Research hypotheses
Data and method
Empirical evidence from the Swiss case
Discussion
Notes
References
10. Studying the voter–party match: congruence and incongruence between voters and parties
Introduction
Policy representation and policy congruence
Determinants of voter–party policy congruence
Data and methods
Results
Conclusion
References
11. Measuring representation: rethinking the role of exclusion
Introduction
Measuring representation as inclusion
The benefits of exclusions
Lessons for measurers
Notes
References
12. Conclusion
Sanctions vs. selection
Selection vs. sanction – challenges to research on representation
The combination of selection and sanctions – a new view on the “crisis”
of representation?
Note
Appendix A
Appendix B
Appendix C
Appendix D
Appendix E
Index

Citation preview

Political Representation

In an ideal democracy, representatives would entirely reflect citizens’ views, preferences and wishes in their legislative work. However, real-life democracies do not meet this ideal and citizens’ policy preferences and priorities are mirrored only inadequately. This book provides new insights on political representation. It is guided by three questions:   

What roles should representatives play? Who is actually or should be represented? How are the representatives (or how should they be) connected with the represented?

Containing contributions from the perspectives of political theory and philosophy, as well as quantitative empirical studies, the volume demonstrates the need to adapt these established questions to new political realities. This text will be of key interest to scholars and students of political representation and parties, political theory, democratic theory, political philosophy and comparative politics. Marc Bühlmann is Executive Director of the Swiss Political Yearbook at the University of Bern. Jan Fivaz is a PhD candidate in Political Science at the University of Lausanne.

Routledge Research on Social and Political Elites Edited by Keith Dowding, Australian National University, and Patrick Dumont, University of Luxembourg.

Who are the elites that run the world? This series of books analyses who the elites are, how they rise and fall, the networks in which they operate and the effects they have on our lives. 1. Coalition Government and Party Mandate How coalition agreements constrain ministerial action Catherine Moury 2. The Selection of Ministers in Europe Hiring and firing Edited by Keith Dowding and Patrick Dumont 3. Parliamentary Elites in Central and Eastern Europe Recruitment and representation Edited by Elena Semenova, Michael Edinger and Heinrich Best 4. The Selection of Political Party Leaders in Contemporary Parliamentary Democracies A comparative study Edited by Jean-Benoit Pilet and William P. Cross 5. The Selection of Ministers around the World Edited by Keith Dowding and Patrick Dumont 6. Party Members and Activists Edited by Emilie van Haute and Anika Gauja 7. Political Representation Roles, representatives and the represented Edited by Marc Bühlmann and Jan Fivaz

Political Representation Roles, representatives and the represented

Add Add

Edited by Marc Bühlmann and Jan Fivaz

Add Add AddAddAdd AddAdd Add Add

First published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2016 selection and editorial material, Marc Bühlmann and Jan Fivaz; individual chapters, the contributors The right of Marc Bühlmann and Jan Fivaz to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Names: Bèuhlmann, Marc, editor. | Fivaz, Jan, editor. Title: Political representation : roles, representatives and the represented / edited by Marc Bèuhlmann and Jan Fivaz. Description: New York, NY : Routledge, 2016. | Series: Routledge research on social and political elites ; 7 | Includes bibliographical references. Identifiers: LCCN 2015025872| ISBN 9781138956278 (hdk.) | ISBN 9781315665788 (ebk.) Subjects: LCSH: Representative government and representation. Classification: LCC JF1051 .P5833 2016 | DDC 328.3/34--dc23 LC record available at http://lccn.loc.gov/2015025872 ISBN: 978-1-138-95627-8 (hbk) ISBN: 978-1-315-66578-8 (ebk) Typeset in Times New Roman by Taylor & Francis Books

Contents

List of illustrations List of contributors Acknowledgement 1 Introduction

vii ix xiv 1

MARC BÜHLMANN AND JAN FIVAZ

2 A “selection model” for political representation

12

JANE MANSBRIDGE

3 Beyond trustees and delegates

29

ANDREW REHFELD

4 Institutional constraints and territorial representation

48

AUDREY ANDRÉ, SAM DEPAUW AND KRIS DESCHOUWER

5 Promises and lies: an empirical comparison of Swiss MPs’ pre- and post-electoral positions

68

LISA SCHÄDEL, DANIEL SCHWARZ AND ANDREAS LADNER

6 Beyond congruence

85

LISA DISCH

7 Dimensionality of the European issue space

99

JAN KLEINNIJENHUIS AND ANDRÉ KROUWEL

8 Representation of political opinions: is the structuring pattern of policy preferences the same for citizens and elites?

117

JAN ROSSET, GEORG LUTZ AND KATHRIN KISSAU

9 “Alignment of objectives” between parties and their electors: the role of personal issue salience in political representation NATHALIE GIGER AND ZOE LEFKOFRIDI

135

vi

Contents

10 Studying the voter–party match: congruence and incongruence between voters and parties

152

JONAS LEFEVERE, STEFAAN WALGRAVE, MICHIEL NUYTEMANS AND KOEN PEPERMANS

11 Measuring representation: rethinking the role of exclusion

170

SUZANNE DOVI

12 Conclusion

187

MARC BÜHLMANN AND JAN FIVAZ

Appendix Appendix Appendix Appendix Appendix Index

A B C D E

194 198 201 203 204 208

List of illustrations

Figures 4.1 Plotted predicted probabilities of institutional effects on territorial representation 4.2 Plotted predicted probabilities of institutional effects on the focus of representation 5.1 Pre- and post-election spheres taken into account 8.1 Eigenvalues of factors for candidates 8.2 Eigenvalues of factors for citizens 9.1 Salience of the issue areas included in the study 9.2 Average individual distance on salient vs. non-salient issues 9.3 Congruence of Swiss parties with their electorates based on two different measures of congruence 10.1 Distribution of respondents over degrees of incongruence

58 60 72 126 126 144 144 145 161

Tables 3.1 Categories, names, and descriptions of three distinctions 3.2 Schematic conceptual space of three distinctions 4.1 Representational role orientations in 13 European democracies (in percentages) 4.2 Probit regression of institutional effects on territorial representation 4.3 The percentage changes of institutional effects on the focus of representation 5.1 Proximity matrix (congruence) between smartvote answer and legislative behaviour (dependent variable) 5.2 Number of MPs and roll call votes per party in the dataset 5.3 Value range and mean positional congruence per MP and party (grouped by party, percentages), ordered by decreasing mean value 5.4 Logit predictions for positional congruence between pre- and post-election sphere

36 38 55 57 59 73 73

74 78

viii 7.1 7.2 7.3 7.4 8.1 8.2 8.3 8.4 8.5 9.1 9.2 10.1 10.2 A.1 A.2 A.3 A.4 B.1 C.1 D.1 E.1

List of illustrations Number of EUP and EES respondents per country Factors that underlie the issue positions of European voters Average positions on issue dimensions within countries Countries with positive or negative correlations between orthogonal issue dimensions at the EU-level The 13 policy questions and the mean answers of citizens and candidates to them Loadings of the items on the three factors retained in the case of candidates Correlation between factors after rotation Loadings of the items on the four factors retained in the case of citizens Correlation between factors after rotation Ranking of parties based on congruence with their supporters National values compared Mean incongruence for party electorates, in 2003 and 2009 Results of independent sample t-tests comparing incongruence for various groups and election years Response rates of the partirep mp survey in 13 European democracies The effect of district magnitude on the focus of representation (interaction models) The effect of population size on the focus of representation (interaction models) The effect of district magnitude and population size on the focus of representation (restricted models) Description of the 34 smartvote statements / legislative votes (Chapter 5) Description of the 28 EU Profiler 2009 statements (Chapter 7) SELECTS questions / saliency categories (Chapter 9) List of issue statements

105 109 111 112 125 127 128 128 129 146 146 162 163 194 195 196 197 198 201 203 204

List of contributors

Audrey André is a Postdoctoral Researcher at the Vrije Universiteit Brussel. Her research focuses on the impact of electoral institutions on the behaviour of parties, legislators and voters. Key findings have been published in journals such as Electoral Studies, West European Politics, Party Politics, Acta Politica, and the Journal of Elections, Public Opinion & Parties. Marc Bühlmann is Director of the Swiss political yearbook Année Politique Suisse at the Institute of Political Science at the University of Berne. He received his Habilitation (post-doctoral qualification, venia docendi) in political science from the University of Bern for his work on the impact of political institutions on individual behaviour. His research and publications concentrate on political sociology, electoral and participation behaviour, methods, social capital, democratic theory and comparative politics. Sam Depauw is Assistant Professor and postdoctoral researcher at the Vrije Universiteit Brussel. His research concentrates on legislative and electoral studies. He has published extensively on political representation and party discipline in a.o. Comparative Political Studies, Journal of Legislative Studies, Parliamentary Affairs and Party Politics. Kris Deschouwer is Research Professor at the department of political science of the Vrije Universiteit Brussel. His research focuses on political parties, elections, political representation, federalism and consociational democracy. He is the author of The Politics of Belgium. Governing a Divided Society (Palgrave Macmillan 2012). Lisa Disch is Professor of Political Science and Women’s Studies at the University of Michigan. Her research interests include contemporary political thought, democratic theory, feminist theory and political representation. Her recent articles include “Toward a Mobilization Conception of Democratic Representation”, published in the American Political Science Review, and “Democratic Representation and the Constituency Paradox”, published in Perspectives. Lisa Disch earned her PhD from Rutgers University. She is the author of Hannah Arendt and the Limits of Philosophy, published by Cornell in 1994, The Tyranny of the Two-Party System, published by

x

List of contributors Columbia in 2002, and co-editor, with Mary Hawkesworth, of the Oxford Handbook of Feminist Theory, published by Oxford University Press in 2015.

Suzanne Dovi is an Associate Professor of the School of Government and Public Policy at the University of Arizona. Her research interests include democratic theory, representation (especially the representation of historically disadvantaged groups), feminist theory and human rights. Her work has appeared in the American Political Science Review, the Journal of Politics, Constellations, Polity, the LA Times and the Washington Post. Jan Fivaz studied contemporary history, political science and economics at the University of Bern. He has worked for the Swiss Federal Treasury Department and as a research associate for several research projects at the universities of Bern and Lausanne. Jan Fivaz is also a PhD candidate at the Autonomous University Institute IDHEAP in Lausanne. He has published on democratic theory, local democracy, political representation, electoral and legislative behaviour as well as on e-democracy and voting advice applications. Since 2003 he has been co-founder and co-director of the Swiss voting advice application (VAA) smartvote. Nathalie Giger is Assistant Professor of political behaviour at the University of Geneva. She holds a PhD from the University of Bern and before joining the department in Geneva she held various positions at the Universities of Konstanz, Mannheim and Zürich. Her research interests lie in the linkage between citizens and political elites, in particular in political representation and the electoral consequences of public policy. Her work has appeared in the American Journal of Political Science, the European Journal of Political Research, and West European Politics among others. Kathrin Kissau is Head of User Research at NET-Metrix AG in Zurich. She has a PhD in communication and political science from the University of Muenster, Germany. Her research and publications are centred around the topics: political participation and representation, new media developments and their political and societal effects, migrants’ media culture; and online research methods. Jan Kleinnijenhuis is Professor of Communication Science at the Vrije Universiteit Amsterdam. His research deals with the nature of political and economic news, with news selection, and with news effects. He publishes his work in, among others, the European Journal of Political Research, Communication Research, the Journal of Communication, Journalism, Social Networks and Political Analysis. André Krouwel teaches comparative political science and communication science at the Vrije Universiteit Amsterdam and is Academic Director of Kieskompas (Election Compass). Krouwel’s research focuses on political parties and elections. His most recent book is Party Transformations in European Democracies published by SUNY Press (www.sunypress.edu/

List of contributors

xi

p-5605-party-transformations-in-europe.aspx). Krouwel has published articles and book chapters on parliamentary and presidential elections, voting behaviour, political parties, party competition, populism and Euroskepticism. Furthermore, he investigates the impact of information on political attitudes and opinions, (negative) political emotions, beliefs in conspiracy theories, Eurosceptic attitudes and populism. Krouwel has also written on methodologies of estimating party positions in a multidimensional space and on weighting and matching data from non-probability and probability samples. Andreas Ladner is Professor for Political Institutions and Swiss Public Administration at the IDHEAP, University of Lausanne. His areas of research include political parties, municipalities, institutional change and e-democracy. He has conducted several major research projects at the Swiss National Science Foundation and authored books and articles on these topics. He is a co-author of the book Size and Local Democracy (Edward Elgar 2014). He has published in, among others, the International Political Science Review, the European Journal of Political Research, West European Politics, Electoral Studies and Party Politics, and regularly comments on Swiss politics in the media. Jonas Lefevere is Assistant Professor of Political Communication at the University of Amsterdam, and member of the Amsterdam School of Communication Research (ASCoR). His research interests deal with the impact of media on public opinion, media and politics, the behaviour of political elites, and congruence between elites and the mass public. He has published on these topics in Political Communication, Electoral Studies, West European Politics and Public Opinion Research. Zoe Lefkofridi is Joint Jean Monnet–Max Weber Postdoctoral Fellow at the Robert Schuman Centre for Advanced Studies (RSCAS) at the European University Institute (EUI) Florence, and Assistant Professor of Comparative Politics at the University of Salzburg. Her teaching and research focus on democracy and representation in Europe and beyond, with a focus on political inequalities and extremism. Her work appears in the European Political Science Review, European Union Politics and Electoral Studies, among others. Georg Lutz is currently Project Director of the Swiss Electoral Studies (Selects) at the Social Science research centre FORS, Lausanne and Professor of Political Science at the University of Lausanne. He holds a PhD in Political Science from the University of Bern. His work focuses on political institutions and political behaviour in a comparative perspective, as well as Swiss politics. Jane Mansbridge is the Adams Professor of Political Leadership and Democratic Values at the Harvard Kennedy School. She is the author of Beyond Adversary Democracy, an empirical and normative study of face-to-face

xii

List of contributors democracy and the award winning Why We Lost the ERA, a study of antideliberative dynamics in social movements based on organizing for an Equal Rights Amendment to the US Constitution. She is also editor or coeditor of the volumes Political Negotiation, Beyond Self-Interest, Feminism, and Oppositional Consciousness. Her current work includes studies of representation, democratic deliberation, everyday activism and the public understanding of collective action problems.

Michiel Nuytemans is Owner of Tree Company, a company that specializes in online applications on politics and participation, and scientific advisor to the Karel de Grote University College. He is also Research Associate of Political Science at the University of Antwerp. He has extensive experience in creating voting advice applications and has done research on VAAs and media and politics. Koen Pepermans is the Faculty Director of the Faculty of Social Sciences at the University of Antwerp (Belgium). He has a Masters in Political Sciences. Before his current position he was a teaching and research assistant in the domain of methodology at the same faculty. He has provided the necessary IT expertise since the first VAA in 2003 in his capacity as a part-time self-employed ICT professional (www.kpsoft.be). Andrew Rehfeld is Associate Professor of Political Science and Professor of Law (by courtesy) at Washington University in St. Louis. He has held the Fulbright Visiting Research Chair in the Study and Practice of Federalism at McGill University. His research focuses on contemporary democratic theory with related interests in the history of political thought and the philosophy of the social sciences. He has published one book, The Concept of Constituency (Cambridge University Press, 2005) and articles by him have appeared in, among others, the American Political Science Review, the Journal of Politics and the Annals of the American Academy of Political and Social Science. Since 2012 he has been the President and CEO of the Jewish Federation of St. Louis. Jan Rosset is a Postdoctoral Researcher at the Mannheim Centre for European Social Research (MZES), University of Mannheim, with a fellowship from the Swiss National Science Foundation. His research focuses on the relationship between economic and political inequality and on democratic political representation. Lisa Schädel received an MA in Political Science from the University of Zurich. She was a scientific collaborator at the University of Berne between 2010 and 2013. Her research interests included political representation as well as migration and Swiss politics. Daniel Schwarz received his PhD in political science from the University of Bern, Switzerland. Afterwards, he held a two-year position as visiting fellow at the Department of Methodology, London School of Economics

List of contributors

xiii

and Political Science. Currently he is a post-doctoral researcher and head of project of the Swiss voting advice application smartvote at the Centre of Competence for Public Management of the University of Bern. His main areas of study include legislative behaviour, political parties, and Swiss politics and institutions. Stefaan Walgrave is Professor of Political Science at the University of Antwerp in Belgium. His research deals with media and politics, social movements and political participation, and public opinion and elections. He currently is an ERC advanced grant holder with a project on how individual political actors process information.

Acknowledgement

This book is the result of the two-day workshop on Political Representation: New Forms of Measuring and Old Challenges at the University of Bern, Switzerland, organized in November 2009. The workshop offered a unique opportunity to bring together the two sides of political science research – scholars who work on political philosophy, and empirical political scientist – to discuss the different perspectives on their field of research. In our view the two worlds of theory and empirical research are too often separated. The book has its seeds in the quality of the presented papers and the lively discussions during the workshop. For the realization of our endeavour the support of Keith Dowding and Patrick Dumont, the editors of the Routledge series on Research on Social and Political Elites who decided to include our work in their series, was crucial. Without their support this book would not have been published. We would also like to thank Charlotte Endersby and Andrew Taylor from Routledge for their friendly and pleasant, but also very efficient and straightforward collaboration during the last year. Furthermore, the two anonymous reviewers provided essential feedback. Evelyne Wild helped us to bring the manuscript into a publishable form. Last but not least we thank all the authors, who stayed on board over several years. Marc Bühlmann and Jan Fivaz Bern, June 2015

1

Introduction Marc Bühlmann and Jan Fivaz

A crisis of representative democracy? In the eighteenth century, the idea of representation “transformed democracy from a doctrine suitable only for small and rapidly vanishing city-states to one applicable to the large nation-states of the modern age” (Dahl 1989: 29). Beginning with the thoughts of Mill (1991 [1861]), the topic of representation has become one of the most widely studied in political science, in theoretical research as well as in empirical research. In the 150 years of both philosophical thinking on and empirical analysing of democratic representation a huge range of ideas, models and findings have been presented and discussed. However, research about democratic representation is still guided by questions on the characteristics of the relationship between representatives and represented: is there or should there be some sort of relationship between representatives and represented, what should it look like and how does it actually work? In an ideal representative democracy representatives would be fully responsive to the citizens’ views, preferences and wishes and would entirely reflect them in their legislative work (Powell 2004). Yet real-life democracies do not meet this ideal and citizens’ policy preferences and priorities are mirrored only inadequately. We rather face a representational gap (Whitefield 2006), that is, more or less wide differences between the views and acts of representatives and the preferences of the represented. If this gap gets too large, the representational system will fail to work properly. In recent years – at least in the views of an increasing part of the public and according to the media discourse – almost all modern western democracies are struggling with decreasing political participation, loosening of party ties and affiliations as well as a loss of confidence of the citizens in parties, politics and the political institutions of the representative systems. The representational gap is seen as one of the most important reasons for these developments. Meanwhile, there is also an on-going debate whether western democracies are facing a general crisis or whether we are moving towards a “post democratic” society (e.g. Crouch 2008). Whereas the public and the media discourse tend to support such theories of crises, political scientists are much more cautious

2

Bühlmann and Fivaz

(Merkel 2015). However, there is a general consensus that modern democracies are facing important challenges. The first and probably most important challenge is the citizens’ loss of confidence in politics, parties and politicians. They have grown distrustful of politicians and parties, become sceptical about democratic institutions, and disillusioned about the functioning of the democratic process (Dalton 2007; Dalton and Wattenberg 2002). Given this loss of confidence of the citizens in their political elite, the question of the relationship between representatives and represented is of outmost importance. But there remain additional challenges, which should be taken into account: the changing roles of representatives due to growing mediatization and popularization of politics, the breaking of the chain of responsiveness in terms of the national elites’ loss of autonomy due to globalization, and the on-going global migration that challenges the question of congruence between representatives and the represented can be considered the most pressing challenges for democratic nations that all are based on representative political systems. The continued lack of confidence can have two potentially dangerous consequences for representative democracies (Cox 1996; Dailami 2000; Galbraith and Kum 2002; Longworth 1998; Puddington 2009; Rodrik 1997): first, disappointment about the behaviour of the representatives can lead to political abstention and apathy. Abstention negatively affects responsiveness because the preferences of non-voters are not considered by the representative system. This can lead to even more frustration, to political action bypassing the representative channels, or even violent and anti-democratic political behaviour (Bernstein 2001; Bühlmann et al. 2010; Kitschelt 1995; Mény and Surel 2000; Nye et al. 1997; Uslaner and Brown 2005). Second, decreasing public confidence strengthens the growth of populism, which profits from the critique of the political elite. Populists employ what Chambers (2004) calls “plebiscitary reason”. They try to manipulate the public by saying what the public wants to hear and by pandering to their audience’s desires only to pursue their own agendas. Is the decreasing confidence in the political elite a sign of interference or even of a rupture of the relationship between represented and representatives? In the United States as well as in Europe extremist and populist parties have been gaining ground in recent years and several of these parties represent ideas which are difficult to match with democratic principles and values. Without question, though, this development confronts democracies with a clear and severe challenge. But does it also necessarily stand for a dysfunctional representative system? This question is much less clear, because the populist parties are getting their votes in open and fair elections. Is the rise of populist parties a sign of a failure of the representative system or rather a sign for changing relationships between representatives and represented? A further challenge to modern democracies consists of the growing mediatization of politics, which leads to more superficial politics by increasing the importance of politicians’ images to the public and by decreasing the value of the substantive content of their politics. Politicians react to the economically

Introduction

3

driven news strategies of the media: the struggle for attention, as well as for the ability to define political issues, become a struggle for control of the news agendas and for the framing of the news (Wolfsfeld 1997). To get attention, politicians must come up with spectacular actions or arguments and/or present themselves as charismatic leaders. This leads to an on-going personalization and “spectacularization” of political communication (Mazzoleni and Schulz 1999). Thus, on the one hand, mediatization not only changes representatives’ roles but it also handicaps the citizens’ evaluation of the representatives. In other words, the relationship between representatives and represented is disturbed. However, modern and technological means of communication can also help to strengthen and widen this relationship. A third challenge arises due to globalization in terms of de-nationalization, that is, the supranationalization and regionalization of politics (Kriesi et al. 2013). One can observe a growing number of different stakeholders to whom a representative is accountable. This leads to the question of the legitimacy of these stakeholders. Representatives who pursue not only electoral goals but also hidden agendas that satisfy the stakeholders’ interests are accused of strategic shirking. Globalization further breaks the chain of responsiveness (Powell 2004). A growing number of political decisions are no longer made by national representatives who are accountable to the voters, that is, the citizens of a given country. Political decisions made in supranational organizations affect citizens within nation-states. However, these citizens cannot hold their representatives accountable for these decisions. Furthermore, elites can excuse unpopular national decisions with pressure from global markets or requirements of global institutions out of reach of national politics, and thereby blur the clarity of responsibility (Powell and Whitten 1993). This leads to the question of whether the de-nationalization does blur or at least change the relationship between represented and representatives. A last important political challenge for representative democracies is global migration. Depending on the migration laws, the openness and the inclusiveness of different established democracies varies greatly. This, however, raises the question of who the citizens to be represented actually are. Who gets represented? According to Dahl (1998), an important feature of democracy is equality, that is, the possibility for all persons who are affected by a political decision to take part in the respective decision-making process. However, this principle is undermined in countries where migrants do not have suffrage rights or must overcome high obstacles to become citizens. And there is no sign that global migration will decrease in the near future. Hence questions of inclusion and exclusion, that is, who gets represented and who actually are the actors in the relationship between representatives and represented, will become more and more important.

Thinking about representation: sanction vs. selection An important requirement for answering the questions raised above is the clarification of the nature of the relationship between representatives and

4

Bühlmann and Fivaz

represented. This book starts with a comparison of two powerful ideal-type models of representation brought forward by Jane Mansbridge (Chapter 2): the sanctions model and the selection model of representation. The sanctions model is the more classic model of representation and corresponds to the (neo-)liberal idea of self-interest as the main driver of individual behaviour. Both representatives and represented act according to their own benefits. Whether a representative is a shirking populist or a trustworthy agent is non-relevant because his or her role is only a means to reach the goal of getting (re)elected. According to the sanctions model, the attainment of this goal depends on how the represented are satisfied, that is, how their (self-)interests are fulfilled. Hence, the representatives only have extrinsic incentives to represent the wishes and needs of their voters. It is the fear of being sanctioned, that is, not getting elected at the next election, that drives them to act according to the preferences of their voters. In the sanctions model, elections are a means for both, responsibility and accountability (Kriesi et al. 2013). Voters hold their representatives accountable by threatening them with sanctions at the next election, while the representatives try to act responsibly, that is, in congruence with the voters’ wishes, to avoid these sanctions. The relationship between representatives and represented is seen as static. The main point of contact between the citizens and the elite is the election. In the election campaign, a representative tries to make clear the interests he or she will defend, whereas the voters choose among the different offers, re-elect the representatives that have fulfilled their pledges, or sanction those who did not keep their promises. In contrast to the sanctions model, Jane Mansbridge proposes an alternative model of how representation could work: the selection model (Chapter 2). In the selection model of representation the relationship is thought of as dynamic and – in its ideal conception – as an on-going deliberation between the representatives and the represented. In this model, intrinsically, self-motivated representatives try to align their objectives with those of the represented through deliberation and communication. The main aim of both representatives and represented is to align objectives of common interests. The societal and institutional context should provide mechanisms that make such alignment possible. Thus, the relationship between represented and representatives sought by the selection model is not hierarchical as in the sanction model but is based much more on mutual confidence. Of course, the hierarchy is present due to the function of representation. However, the closer the common interests are, the less citizens need power to control the representatives. Thus, the search for alignment of interests is much more important than monitoring and sanctioning. The represented prospectively select (rather than retrospectively deselect) intrinsically motivated, honest representatives who try to meet citizens’ preferences. Of course, both models are ideal-types. In fact, representative democracies will comprise elements from both models. However, the predominant model in representational research is definitely the sanctions model. Since Downs’ (1957) simplifying but empirically very fruitful comparison between political

Introduction

5

and market logic, the rational choice perspective on policy, polity and politics has become dominant in many fields of political science. The prevailing (empiric) view on representation is an economic one. Empirical scholars are normally very sceptical about intrinsically motivated representatives. The rational choice-based model of delegates, which empirical research (onesidedly) favours, defines representatives as rational actors aiming at (re)election and that do not shy away from using shirking strategies to reach this aim. The sanctions model also seems to dominate the public discussion and perception. Very often, media picture representatives as egoistic and corrupt, trying only to fulfil their own interests. It is lamented that the represented are more and more prevented from controlling and even sanctioning the representatives because there is no choice of honest politicians. Furthermore, the increasing lack of transparency makes it impossible for the people to control the political elite. Bearing in mind the alternative selection model, one could ask whether it is this perception of representation according to the sanctions model that is finally responsible for the statement of a crisis of representation. It is difficult to develop confidence in representatives who are solely described as corrupt and shirking egoists. An important role of political theorists is to challenge existing views and mirror specific circumstances. The recent theoretical discussions on representation indeed scrutinize the predominant economic view on the topic, that is, the sanctions model (see Chapters 2, 3, 6 and 11 in this book). Although this model seems quite useful for empirical investigations of representation, it is criticized by political theorists as being much too simple. Political philosophers plead for more parsimonious approaches to the very complex concept of representation. The question is whether the idea of a selection model of representation can help to rethink the relationship between representatives and represented and to open the quite one-sided empirical and societal discussion on the crisis of representation. While theorists should challenge existing views, political scientists should test theoretical ideas and challenge comforting research rituals. The question is whether the sanction model is more adequate to describe the representational relationship than the selection model or whether the selection model has the potential for new insights on representation. This book combines theoretical as well as empirical research. Beside Jane Mansbridge’s contribution, which provides a common ground for the other contributions in the book, it contains three further chapters with a theoretical perspective and six chapters based on empirical analyses. By jointly discussing recent developments in both the theoretical and the empirical research, by challenging normative assumptions with empirical insights, and by confronting real-life observations with theoretical concepts our book provides some new insights into the long-standing questions on how the relationship between representatives and represented should be structured and is structured in real life. The interaction between theory and empirics has turned out to be very fruitful.

6

Bühlmann and Fivaz

But this book is also a response to the renewed interest in representation within contemporary research on democratic theories and on political representation. On the one hand, it takes into account theoretical enlargements of the very notion of “representation”. The questioning of the so far predominant sanction model by the alternative selection model presents a first example of that, but the other theoretical chapters also provide new perspectives on the representational processes. On the other hand, the book includes new empirical findings to some extent based on new measurement tools or data sources that help us to understand how the actual relationship between representatives and represented is working.

Theory meets empirics I – the relationship of representatives and represented A number of chapters focus explicitly on the relationship between representatives and represented. Andrew Rehfeld (Chapter 3) provides a theoretical introduction to the different roles representatives can fulfil. He discusses the question of how closely a representative’s vote on legislation must normally correspond to the will of his or her constituents within the dichotomy between two classic ideal role types: the idea of trustees (i.e., representatives acting relatively detached from their voters, aiming at the best for the whole nation) and the idea of delegates (i.e., representatives acting closely in accordance with the preferences of their constituents). Andrew Rehfeld takes these two well-known ideal types only as a starting point and systematically enlarges this dichotomy, arguing that it obscures three fundamental distinctions: the aims of a representative (or according to Jane Mansbridge the intrinsic motivation, i.e., whether a representative aims at the good of all or the good of some), the source of judgement (i.e., whether a representative relies on his or her own judgement or the judgement of a third party), and responsiveness (i.e., the probability of sanctions). Based on these three distinctions, the dichotomy between trustee and delegate is expanded to describe eight different ideal types or roles of representatives. The presented conceptual enlargement provides an interesting roadmap for new empirical research. Indeed, Andrew Rehfeld argues that empirical scholars too often limit their analyses of the role conception of delegates to those role types that are easiest to operationalize and measure. However, his eight ideal types can serve as a good reference for empirical investigations into the various roles of representatives applying a much broader approach. The two following empirical chapters also enrich past research by presenting empirical evidence based on new datasets. Audrey André, Sam Depauw and Kris Deschouwer (Chapter 4) present the results of a comparative study based on data from candidate surveys in 13 countries. In the centre of their study is the question whether representatives apply a personal vote-seeking strategy (i.e. try to win elections with personalized campaigns and spending to support their own reputation) or a more party-centred strategy during election

Introduction

7

campaigns. In terms of Andrew Rehfeld’s conceptualization Audrey André et al. concentrate on the sources of role understanding, or what they call the “legislators’ campaign norms”. They argue that personal vote-seeking only makes sense in specific contexts. Electoral rules, district magnitude or pattern of intra-party competition build incentives or constraints in favour of or against personal vote-seeking. Audrey André et al. enrich the current research by including parties as actors and by pointing out the importance of the institutional context. Lisa Schädel, Daniel Schwarz and Andreas Ladner (Chapter 5) use data from smartvote, a Swiss voting advice application (VAA),1 and compare preelection statements with legislative voting behaviour of members of parliament (MP) in the Swiss lower house. Their starting point is the sanctions model. They monitor the extent to which MPs keep their pre-electoral promises after the election during their legislative work. By doing so, they address the already mentioned and widespread mistrust of citizens towards politicians: do MPs stick to their electoral pledges? Lisa Schädel et al. show that reality is much more complex. First, there is a surprisingly high congruence between preelectoral promises and legislative behaviour. MPs are much more honest than public perception would make us believe. Second, individual aspects as well as contextual factors explain whether an MP keeps electoral promises or not.

Theory meets empirics II – issue congruence revisited A series of five chapters focus on the concept of issue congruence – one of the most-discussed aspects of contemporary empirical research on political representation. However, the contributions develop very different perspectives. The starting point is made by Lisa Disch (Chapter 6) who posits a perspective of an interactive democracy – as she calls the idea of representation based on an on-going exchange between the representatives and represented. She directly challenges the idea of issue congruence and thus calls into question the basic assumptions of the subsequent empirical contributions on issue congruence in this book. Lisa Disch argues against the mandate model of representation that is based on assumptions that cannot be upheld. The mandate model assumes that there is something like a popular will that is incorporated into the governing process through elections, represented by mandate holders, and translated into political decisions. However, as Lisa Disch argues, neither the assumption of coherent citizen preferences nor the suggestion of independent preference formation holds. Representation cannot be understood as a dyadic relationship but must be seen as an interactive process in which interests are evoked, strengthened and renewed at the elite as well as at the citizen level. In this sense, representative democracy is not a means to represent a popular will, formulated before elections take place. Representation should rather be seen as a mode of production of citizen preferences by means of continuous processes of interaction between the represented and representatives – an alternative version of Jane Mansbridge’s alignment in her selection model.

8

Bühlmann and Fivaz

The assessment of the functioning of interactive representation would be more complex than looking for congruence between the opinions of citizens and MPs. The subsequent four empirical chapters further criticize the concept of issue congruence, even though not as radical as Lisa Disch. First, Jan Kleinnijenhuis and André Krouwel (Chapter 7) investigate the dimensionality of the European political space. They do so by analysing data generated by the EU Profiler, another VAA that was developed and used for the elections for the European Parliament in 2009. In recent years, VAAs have become very popular in European countries, with millions of voters using them, but political scientists have more or less ignored the possibilities that VAAs offer for research. Jan Kleinnijenhuis and André Krouwel provide an illustrative example of how this data can be used for research purposes. They use the voters’ answers given to the EU Profiler’s questionnaire during the 2009 election campaign to analyse whether there is a common European political space and how it is structured. Their study is of specific interest with regard to the question whether selection of candidates by voters based on issues is a feasible method in a heterogeneous context like the European elections. The contribution of Jan Rosset, Georg Lutz and Kathrin Kissau (Chapter 8) broadens Lisa Disch’s critique on the concept of issue congruence. They base their objections on empirical grounds, presenting some new evidence regarding the level of issue congruence in Switzerland by comparing the policy preferences of citizens and elites. The authors analyse whether the structures of policy preferences of representatives and voters do match. Their findings question the usefulness of the issue congruence concept: if the elites’ policy preferences are structured differently than those of the voters, how can the former act responsively to the preferences of the latter? The next empirically based critique of the idea of issue congruence is put forward by Nathalie Giger and Zoe Lefkofridi (Chapter 9). They argue that classic issue congruence studies are missing one important point: salience. The two authors develop the argument that models of representation should include the importance of different policies, conceptualized and measured at the citizen level. Their results show that personal salience makes a difference when analysing issue congruence. Thus, even if the authors insist on the importance of the concept of issue congruence, they enlarge the discussion by including the concept of saliency. In Chapter 10 Jonas Lefevere, Stefaan Walgrave, Michiel Nuytemans and Koen Pepermans argue that elections, that is, the first step in the representational process, are flawed due to a very low level of issue congruence. They use data from elections in Belgium and present evidence that a large proportion of voters do not vote “correctly” – that is, they do not vote for parties that are closest to their own political preferences. As the main reason for this mismatch, Jonas Lefevere et al. identify an information problem among voters, who often intend to vote for the parties with the best issue congruence but lack the necessary information to do so. Not only does this finding challenge the idea of issue congruence, but this mismatch also can lead to a spiral of

Introduction

9

distrust and alienation. Parties break their promises to their voters not because they are disloyal or because they betray their promises once they are in power but, simply, because they promised things other than what their voters thought they did. Incorrect voting, then, leads to incorrect expectations of what will be done once the representatives are elected. This mismatch between representatives and the represented further undermines the concept of issue congruence.

Exclusion – or who actually is “the represented”? Migration and globalization establish some sorts of “demoicracies” (Cheneval 2011). Belonging to different demoi is an increasing challenge to representative democracies. Often, there is only one demos within a nation-state that actually is represented (at least according to the sanctions model): the voters, that is, the citizens who have the right to vote. Thus, a further theoretical broadening of the discussion on political representation should cover the idea of inclusion and exclusion. In Chapter 11, Suzanne Dovi focuses on this question of inclusion and exclusion, providing some new insights into the question of who should be represented. Representational theories are often based on the seemingly simple assumption that representatives should represent the citizens. However, apart from the normative discussion on the role of representation, this assumption raises the question of what is meant by “citizens”, that is, who should be or actually is represented. Suzanne Dovi argues that political scientists equate democratic representation with inclusion, stating that the more citizens are included, the better representation will be. She demonstrates, however, that this is an incomplete framework for understanding representation. She pleads for a more adequate understanding of representation by taking exclusion into account. Exclusion, she argues, is inherent to representation because democratic institutions include and exclude simultaneously. By giving some examples of how exclusion takes place in democracies and by arguing that exclusion can even benefit democracies, Suzanne Dovi convincingly broadens the theoretical discussion on representation – even beyond the idea of sanctions or selection.

Note 1 VAAs are websites that provide voters with the opportunity to compare their own policy preferences with those of parties and candidates by answering a comprehensive questionnaire on a broad range of policy issues. For an overview on VAAs see Ladner and Fivaz 2012 or Garzia and Marschall 2014.

References Bernstein, Jeffrey L. (2001). “Linking Presidential and Congressional Approval During Unified and Divided Governments.” In: John R. Hibbing and Elizabeth Theiss-Morse (eds) What Is It about Government that Americans Dislike? Cambridge: Cambridge University Press, 98–117.

10

Bühlmann and Fivaz

Bühlmann, Marc, Laurent Bernhard and Lisa Müller (2010). “Swiss Democracy in Crisis? An Analysis of Exit and Voice.” In: Simon Hug and Hanspeter Kriesi (eds) Value Change in Switzerland. Lanham, MD: Lexington Books, 211–234. Chambers, Simone (2004). ‘Behind Closed Doors: Publicity, Secrecy, and the Quality of Deliberation.’ The Journal of Political Philosophy 12(4): 389–410. Cheneval, Francis (2011). The Government of the Peoples: On the Idea and Principles of Multilateral Democracy. New York: Palgrave Macmillan. Cox, Robert W. (1996). “Globalization, Multilateralism, and Democracy.” In: Robert W. Cox and Timothy J. Sinclair (eds) Approaches to World Order. New York: Cambridge University Press, 524–536. Crouch, Colin (2008). Postdemokratie. Frankfurt am Main: Suhrkamp. Dahl, Robert A. (1989). Democracy and its Critics. New Haven, CT: Yale University Press. Dahl, Robert A. (1998) On Democracy. New Haven, CT: Yale University Press. Dailami, Mansoor (2000). Financial Openness, Democracy, and Redistributive Policy. Policy Research Paper No. 2372. Washington, DC: World Bank. Dalton, Russell J. (2007). Democratic Challenges, Democratic Choices. The Erosion of Political Support in Advanced Industrial Democracies. Oxford: Oxford University Press. Dalton, Russell J. and Martin P. Wattenberg (eds) (2002). Parties Without Partisans. Political Change in Advanced Industrial Democracies. Oxford: Oxford University Press. Downs, Anthony (1957). An Economic Theory of Democracy. New York: Harper. Galbraith, James K. and Hyunsub Kum (2002). Inequality and Economic Growth. Working Paper No. 21. Austin, TX: The University of Texas Inequality Project. Garzia, Diego and Stefan Marschall (eds) (2014). Matching Voters with Parties and Candidates. Voting Advice Applications in a Comparative Perspective. Colchester: ECPR Press. Kitschelt, Herbert (1995). The Radical Right in Western Europe. Ann Arbor, MI: University of Michigan. Kriesi, Hanspeter, Sandra Lavenex, Frank Esser, Jörg Matthes, Marc Bühlmann and Daniel Bochsler (2013). Democracy in the Age of Globalization and Mediatization. Basingstoke: Palgrave Macmillan. Ladner, Andreas and Jan Fivaz (2012). “Voting Advice Applications.” In: Norbert Kersting (ed.) Electronic Democracy. Opladen: Barbara Budrichs Publisher, 177–198. Longworth, Richard C. (1998). Global Squeeze: The Coming Crisis for First World Nations. Chicago, IL: Contemporary Books. Mazzoleni, Gianpietro and Winfried Schulz (1999). ‘Mediatization of Politics: A Challenge for Democracy?’ Political Communication 16: 247–261. Mény, Yves and Yves Surel (2000). Par le peuple, pour le peuple. Le populisme et les démocraties. Paris: Fayard. Merkel, Wolfgang (ed.) (2015). Demokratie und Krise. Zum schwierigen Verhältnis von Theorie und Empirie. Wiesbaden: Springer VS. Mill, John Stuart (1991 [1861]). Considerations on Representative Government. Amherst, MA: Prometheus Books. Nye, Joseph S., Philip D. Zelikow and David C. King (1997). Why People Don’t Trust Government. Cambridge: Cambridge University Press. Powell, G. Bingham (2004) ‘The Chain of Responsiveness.’ Journal of Democracy 15(4): 91–105.

Introduction

11

Powell, G. Bingham and Guy D. Whitten (1993). ‘A Cross-National Analysis of Economic Voting: Taking Account of the Political Context.’ American Journal of Political Science 37(2): 391−414. Puddington, Arch (2009). ‘A Third Year of Decline.’ Journal of Democracy 20(2): 93–107. Rodrik, Dani (1997). Has Globalization Gone Too Far? Washington, DC: Institute for International Economics. Uslaner, Eric and Mitchell Brown (2005). ‘Inequality, Trust, and Civic Engagement.’ American Politics Research 33(6): 868–894. Whitefield, Stephen (2006). ‘Mind the Representation Gap: Explaining Differences in Public Views of Representation in Postcommunist Democracies.’ Comparative Political Studies 39(6): 733–758. Wolfsfeld, Gadi (1997). Media and Political Conflict. News from the Middle East. Cambridge: Cambridge University Press.

2

A “selection model” for political representation Jane Mansbridge

Introduction: sanctions v. selection1 Many political scientists, most economists, and almost all citizens who demand more “accountability” and “transparency” routinely rely on a sanctions model of principal–agent relations. In this model, the interests of the principals (the constituents) are assumed to conflict with the interests of their agent (the representative). The principals must therefore monitor the agent closely, rewarding good behaviour and punishing bad. Here I elaborate a contrasting model of principal–agent relations, which I will call a selection model. This model works only when a potential agent already has self-motivated, exogenous reasons for doing what the principal wants. The principal and agent thus have similar objectives even in the absence of the principal’s sanctions. As a general rule, the higher the ex ante probability that the objectives of principal and agent will be aligned, the more efficient it is for the principal to invest resources ex ante in selecting the required type rather than investing ex post in monitoring and sanctioning. A principal–agent model with more selection in the mix is also efficient when agents face unpredictable future decisions, must be flexible and adaptive, or are hard to monitor. If we view democratic political representation as a principal–agent problem, a constituent can reasonably want to put a representative in office and spend relatively little subsequent effort on monitoring and sanctioning whenever the representative’s established direction and policies are largely those the constituent desires and the representative also has a verifiable reputation for being both competent and honest.2 In the selection model, the representative’s accountability to the constituent will typically take the form of narrative and even deliberative accountability rather than accountability based on monitoring and sanctions, as the representative explains the reasons for his or her actions and even (ideally) engages in two-way communication with constituents, particularly when deviating from the constituents’ preferences. In practice both selection and sanctions are always at work, with the balance between the two depending on the character of the representative and the electoral setting. Selection and sanctions map directly onto the two commonly cited motivations of representatives, their desire to make good public

A “selection model” for representation

13

policy and their desire to be re-elected. These two motivations are almost always mixed. The alignment of principals’ and agents’ objectives will also never be perfect. Thus constituents will always have some reason to monitor their representatives in order either to try to induce new behaviour through the threat of sanction or simply to replace those representatives in the next election. Nor is the sanctions model ever pure. It is hard to imagine a representative with no intrinsic motivation to work for the policies he or she thinks good for the polity and hard to imagine any constituent voting for a representative whose preferences the constituent thought were always induced. Because both the sanction and the selection models are always mixed, it helps to think of the selection model as having selection at its core and sanctions at its periphery. In this core–periphery configuration, most of the congruence between the principals’ desires and the agent’s behaviour is accomplished by the voters selecting a representative who is honest, competent, and already has policy goals much like the constituents’. The strongest and most central mechanism for representing the constituents’ views is selection, while sanction works at the edges of the system, disciplining the selected representative’s tendencies to deviation only lightly and in the most important places. In politics, elections are the main instrument for both selection and sanction. The appropriate balance between selection and sanction depends on the costs and benefits of the two models in a given context. The benefits of selection are high when the agents will face unpredictable future situations, when they must act speedily, creatively, flexibly, and adaptively, when they must dedicate their powers to an evolving goal and adopt different means as the need demands, when the goals are long-run rather than short-run, and when the principals prefer relationships based on mutual trust and common goals rather than instrumental relationships. Selection is also attractive when its costs are low – that is, in contexts where sufficient numbers of potential agents are self-motivated, where the probabilities of aligned objectives between principals and agents are high, where principals can fairly easily gather sufficient information about agents to make a reasonably accurate choice at the time of selection, and where agents can effectively self-sort into appropriate roles. Finally, selection is attractive when the alternative – sanctions with monitoring – is impractical or very costly. In normative theory, participatory democrats who cast a critical eye on all representative relations have been particularly suspicious of selection models, as subject to elitism, corruption, and failures in citizen control of the representative. These objections have little purchase in some contexts. Normative theorists have not sufficiently considered either the possible efficiency gains of a selection model (which can perhaps be normatively discounted) or the quality of the constituent–representative relationship, which can be for many reasons more humanly satisfying in a selection model. In relatively uncorrupt democracies, the circumstances of political life frequently produce intrinsically motivated agents, the possibility of largely aligned objectives between voters and representatives, the capacity among voters to

14

Mansbridge

engage relatively easily in a reasonable selection process, and mechanisms by which relatively well aligned representatives sort themselves into their roles. When this is the case, the voter’s efforts are efficiently concentrated on selection at the front end of the process and the electoral system can normatively reinforce a commitment to the public good and a warranted relationship of trust and goodwill between representative and constituent. In the political science literature, the selection model, advanced in the early 1960s as one of the two paths to constituency control, was eclipsed by the sanctions model in the 1970s, despite data suggesting that in many circumstances the selection model had greater predictive power.3 This eclipse is coming to an end. My goal in this chapter is to restore the selection model fully to the status that it had in 1963, as descriptively and normatively an equal partner to the sanctions model. The selection model has three necessary components: self-motivated agents, aligned objectives, and a combination of selecting and sorting mechanisms that make such alignment possible. Self-motivated agents Management theorists have often distinguished between what McGregor (1960) called “Theory X”, appropriate to situations in which employees will avoid work if they can and do not care about organizational goals, and “Theory Y”, appropriate to situations in which employees intrinsically enjoy their work, just as they enjoy play or leisure, and can be internally committed to the aims of the organization. Self-motivated employees, McGregor theorized, would be the more likely to exercise imagination, ingenuity, and creativity in their work.4 In Theory Y, the intrinsic motivation of an employee can be of any kind – a desire to solve puzzles, to exercise outdoors, or to keep things neat. In politics also, the intrinsic motivation of a representative can be of any kind. A representative might, in principle, promote lower taxes simply because he or she personally had a lot of property subject to tax. Certain occupations, however – including politics in relatively uncorrupt democracies – attract agents whose intrinsic motivations include what might be called public spirit, that is, concern with the common good and at least some small willingness to make sacrifices in material interest for that good (Besley and Ghatak 2005). Although the mix of motivations in any given individual will almost always include some extrinsic motivation and will probably also include forms of intrinsic motivation not based on public spirit (such as the desires to feel useful, stand by one’s comrades, or not be overruled), in non-corrupt democracies where politics does not produce great pecuniary rewards, political office will tend to attract actors with some intrinsic public-spirited motivation. Yet public spirit is vulnerable to being undermined by three factors associated with the sanctions model: extrinsic incentives, monitoring, and a culture that assumes self-interest as the primary motivation. Psychological

A “selection model” for representation

15

research has established that in many circumstances extrinsic motivation drives out intrinsic motivation (although extrinsic rewards framed as either inherent to the task or honouring the actor’s intrinsic motivation do not have this negative effect). Extrinsic rewards can “crowd out” intrinsic rewards and motivation in real life as well as laboratory experiments, and such crowding out may be most likely when the agent has originally been selected for intrinsic motivation (Deci et al. 1999; Kruglanski et al. 1975; Deci and Ryan 1985; Frey 1997; Bohnet et al. 2001). Monitoring may also drive out intrinsic motivation, particularly when the subjects of the monitoring perceive it as expressing distrust. “Trust often invites reciprocal trust” in a virtuous spiral, while “institutionalized suspicion undermines trust” (O’Neill 2002: 19, 49, 25; Behn 2001: 83; Philip 2004: 22; Anechiaro and Jacobs 1996: 202; Goodin 1980, 1982). Even the act of describing most behaviour in self-interested terms tends to undermine public-spirited motivation, in part because norms tend to change in the direction of the perceived majority’s norms (Miller 1999: 1053). For these reasons, although some past institutional designers have argued that institutional arrangements based on self-interest are the most stable, such arrangements also have the potential for crowding out public spirit (Mansbridge 1990). In the United States, for example, the close races with little incumbency that are necessary to make the sanctions model work sometimes drive away the most public-spirited potential agents. Aligned objectives The selection model requires not only an internally motivated agent but also an alignment between the objectives of the principal and the self-motivated agent. From the perspective of democratic theory, the closer citizens come to having aligned objectives and common interests with their representatives, the less they need equal power as protection against the potentially greater power of the representatives. In the extreme, complete common interests could legitimate large inequalities of power, as long as – an important qualification – the less powerful could take back equal power as interests began to diverge and as long as the two other conditions of legitimately unequal power – the maintenance of equal respect and the opportunity for individual development – were met (Mansbridge 1980: 235–247). Such alignment of objectives can take place not only on the high ground of similar understandings of what is best for the nation as a whole but also on the ground of what is best for a particular individual or a community, such as farmers, miners, or inner city residents. Context affects the supply of representatives whose objectives are well aligned with those of their constituents. Such alignment appears more often in homogeneous districts (Fenno 1978: 124). Context-specific norms and institutions can also encourage public spirit and discourage the more obvious forms of corruption5 – as in the quite different interactions of culture and institutions in Sweden, the English, German, and French civil service, and the clergy in many forms of religion. Finally, widespread competence in potential

16

Mansbridge

representatives increases the supply of potential agents aligned with their constituents, both because competence is required to further those aligned objectives and because competence fosters internal motivation. Context also affects the demand for aligned objectives. Recognizing that no principal and agent ever have perfectly aligned objectives, a “contingency” application of the selection model asks when it makes most sense to tolerate some divergence in those objectives. Other things being equal, a selection model is preferable and a principal should rationally tolerate some divergence in objectives when the principal needs an agent capable of flexible, adaptive, and creative performance, discretion in negotiation, and dedication to longrun aims, as well as when tools for adequate monitoring and sanctioning are absent or expensive.6 In organizational theory, “high commitment” management practices, based implicitly on a selection model, are most effective in such contexts (Baron and Kreps 1999). In politics, descriptive representation (by geography, occupation, religion, tribe, race, gender, or any other factor), one form of a selection model, is most in demand when interests are crystalized and the representative must exercise discretion (Mansbridge 2003) or when the standard forms of monitoring and sanctioning are either counterproductive or weak. Thus constitutional conventions, which have no subsequent elections, rely heavily for their legitimacy on selection models of representation with a strong dose of descriptive representation. Selection and sorting The third necessary feature of an effective selection system involves accurate selection and sorting as well as the voters’ capacity to “de-select” a representative easily when circumstances change. For accurate selection, the principal must have sufficiently good information at the time of selection about the potential agent’s motivation and thus the likelihood of aligned objectives. In appropriate conditions, agents also sort themselves into jobs and organizations whose aims match as closely as possible their own intrinsic motivation. In political representation, ease of de-selection usually depends on the capacities of ancillary institutions such as political parties and the media. Voters often select not only on policy direction and competence but on intrinsic motivation, or “character”. The less a voter knows about policies, the more rational it is to select on character. In addition, the less easy it is to monitor a representative’s subsequent behaviour, the more rational it is to select on character.7 In these circumstances, both parties will benefit if the agent can send a signal that is hard to manipulate.8 Repeated interaction provides one form of expensive signal (Dixit 2002). In the absence of personal repeated interaction with a potential agent, principals can rely on the agent’s reputation within a network of individuals well placed to observe the agent over time (Shapiro 2005: 276; Bianco 1994: 152). Professional networks – including the professional civil service in such countries as France, England, Germany, and Sweden – help create reputations for

A “selection model” for representation

17

character. Political parties help create believable reputations for the political direction, the character, and the competence of legislative candidates. Voters in large-scale advanced democracies often look to the party identification of a candidate as the most important signal of at least political direction. De-selection (which economists often call “ex post selection”) generates an information problem even worse than the one the voters face at the initial selection. Although voters now have access to the public record of their representatives’ behaviour in office, the representatives also have an increased capacity to make their records look better than they are through easier access to the media and public events, the many communications to constituents paid for by the government, and the publicity attendant on constituency service. The selection model is not normatively tenable unless the citizenry retains sufficient capacity to remove representatives who have become out of touch with the prevailing objectives of the district’s voters. Thus party systems, other networks, and the media play a critical informational role, as does the internal competitiveness and vitality of the party system. Self-selection among potential agents can also serve as a critical sorting mechanism. Non-profit organizations often act differently from private firms because the managers, acting from more intrinsic motivations, “sort themselves, each gravitating to the types of organizations that he or she finds … most compatible” (Weisbrod 1988: 31–32).9 Potential political representatives engage heavily in this sorting process. Those who might otherwise be interested in running for office refrain from that investment if their personal policy orientations do not sufficiently reflect the orientations of their potential constituents (Fenno 1978; Lawless 2005). When the orientations of their constituents change (through changed boundaries in the district, changing demographics, or the entry of new generational cohorts with new political views), many elected representatives retire. Low compensation plays two contradictory roles in this sorting process. The pay for mission-orientated agents must be high enough to facilitate a sorting that selects for competence, but low enough to facilitate a sorting that selects for mission-orientation. The clergy and monks of the religions that demand celibacy or relative poverty have traditionally responded to these disincentives by sorting themselves on the dimension of mission-orientation. Low-paying “citizen legislatures” and public offices have generated the same self-sorting, with the same tension between sorting for competence and sorting for mission-orientation. Self-sorting by low compensation is highly inegalitarian, as members of different socio-economic groups bring greatly unequal resources to the sort. Unsalaried or uncompetitively compensated positions tend to attract individuals who have other resources to support themselves and their families. Sorting on “citizen duty” also disproportionately attracts members of the middle and professional classes, who have been socialized to value that form of duty (Prewitt 1970). Although religious, local, and ideological commitments sometimes cut across class to create a mission-orientated motivation to run for office, the frequent professional or

18

Mansbridge

middle-class bias of those who take underpaid offices can make aligned objectives and communication with working-class constituents more difficult. At the same time, low compensation increases the risk of corruption.

Accountability In standard principal–agent theory, accountability means that a principal “has powers to sanction or reward the agent” (Fearon 1999: 55). We may call this accountability as sanction. An earlier understanding of accountability, by contrast, stresses “giving an account” (rendre compte, Rechenschaft abgeben) (Behn 2001: 4; Philip 2004: 12; O’Neill 2002: 58). In a selection model of representation, where the principal and agent have largely aligned objectives, the focus of accountability shifts from monitoring and sanctioning to the agent’s giving reasons for his or her acts. When this communication is one-way, we may call this narrative accountability. When the communication is two-way, we may call it deliberative accountability.10 Narrative and deliberative accountability work best when the principal, even if unhappy with the result, can see that the intrinsic motivation underlying the aligned objectives remains unchanged. Representatives can relatively easily explain changes in policy by new facts and circumstances, while changes in principles throw doubt on the consistency of their character.11 The challenge for the selection model in practice is that representatives can rarely give full explanations in interaction with their constituents (Bianco 1994; Kingdon 1981: 48) and constituents rarely have the opportunity to interrogate their representatives in a forum that allows give and take (see Neblo et al. 2010 for an experimental exception). Accountability as simply giving an account must be married to some sanctions for the combination of selection core/sanction periphery to work. Yet the monitoring required for the sanctions need not be systematic and on-going. It can rely on individual citizens and interest groups to send in the alarm when they come across wrongdoing (“fire alarm” oversight) rather than on processes of continual monitoring (“police patrol” oversight) (McCubbins and Schwartz 1984; Aberbach 1990; Anechiaro and Jacobs 1996). Fire alarm oversight is particularly appropriate in selection-model systems, where intrinsic motivation on the part of the agent produces a high probability of honest, competent, and aligned behaviour. What organizational theorists call “network” – or “horizontal”, or “professional” – accountability can also substitute effectively in large part for the “vertical” accountability standard of the sanctions model. 12 If members of a network have a strong enough internal commitment to the norms of their profession, or even if the members have only a self-interested concern for the reputation of their network, they will have an incentive to monitor and sanction the behaviour of others in that network to keep potential defectors up to network standards. Parties often play this role, particularly in disciplined party systems. These networks of horizontal accountability, along with recruitment systems and larger social norms, help produce honesty and competence

A “selection model” for representation

19

outside any system of electoral sanctions. They can play an important role in the sanction periphery of a functional selection model. Non-hierarchical selection The selection model of representation should not be confused with either the “trustee” form derived from Edmund Burke13 or the elitism of Joseph Schumpeter. Trustee representation is only one instance of the selection model, and it is the least democratic kind. Both the word “trustee” itself and Burke’s own stance suggest that the trustee is wiser and more far-seeing than his constituents, and for this reason is more fit to rule (Herzog 1998). Indeed, in Burke’s era electoral representation implied for many “the principle of distinction”, that is, the idea that representatives “should rank higher than their constituents in wealth, talent, and virtue” (Manin 1997: 94).14 Schumpeter too did not hide his disdain for the “typical citizen”, who “drops down to a lower level of performance once he enters the political field” (Schumpeter [1942] 1962: 262). Yet the selection model of political representation need imply no more hierarchy than any representative process, whether based on sanctions or selection. The selection of a representative can rest purely on a division of labour, as the Abbé de Siéyès suggested years ago (Manin 1997: 3). When the US constitution was being debated in 1787–88 and the Federalists advanced their principle of distinction, the Anti-Federalists fiercely opposed that principle, arguing that representation required “likeness” and “resemblance” to constituents so that representatives could “possess their sentiments and feelings” (Storing 1981: 380; Manin 1997: 110). Both Federalists and Anti-Federalists assumed a model of representation based primarily on selection, but the criteria for selection were hierarchical in the one case and egalitarian in the other. The selection model in developed democracies today has lost many of the hierarchical trappings associated with its “trustee” form. In Timothy Besley’s formulation, for example, citizens select a “citizen-candidate” from among themselves. It is true that in practice in both sanction and selection models constituents generally choose representatives who are more highly educated and have more experience in politics than most of their supporters. Today, however, at least in the United States, when voters say they want to select a “good man” or “good woman” as a representative, they often seem to want someone like them, but with the interest, competence, and honesty to be a legislator. Candidates run for office with the slogan that they are “like you”. Edmund Burke did not present himself to his constituents as being “like them”. Homogeneity and extended incumbency – how evil, really? The selection model is associated in practice with two features usually considered troubling for the quality of democracy: homogeneous districts and extended incumbency. Yet positive cases can be made for both.

20

Mansbridge

The positive case for homogeneity is simple. The more homogeneous a district, the greater the proportion of voters who can be satisfied with their representative – a satisfaction that carries over, to a lesser degree, to the legislature as a whole (Brunell 2006; Buchler 2005; Issacharoff 2002a, 2002b; Persily 2002). If representation had only the goal of voter satisfaction, the ideal electoral district in a first-past-the-post single member system would always be one that included only constituents with exactly the same political preferences (Brunell 2006: 80; Persily 2002: 668). In a homogeneous district, communication between representatives and constituents is also easier. Representatives loosen up, feel relaxed, and communicate easily with their core constituencies. Constituents find it easier to contact a representative whom they view as “like them” (Fenno 1978; Gay 2002). On the negative side, homogeneity in practice is never total, and nearhomogeneity lets a representative ignore the minority (Guinier 1994: 135; Horowitz 2003). The more homogeneous a district is, therefore, the more critical it is for the minorities in the district to connect with critical media, likeminded citizens and associations that reach across district boundaries, and surrogate representatives in other districts. For the majority, homogeneity also risks increasing complacency and narrow-mindedness. In more homogeneous political spaces, neither voters nor representatives will have to pay the price of trying to communicate with others whom they do not immediately understand (Mutz 2006). They may also not understand the need for compromises at the legislative level. Lack of diversity may foster extremism (Sunstein 2003). In sum, short-run political satisfaction may be bought at the expense of long-run growth in understanding and capacity to compromise. Moving from homogeneity to incumbency, the positive case for incumbency is equally simple and somewhat similar to that for homogeneity. In non-corrupt systems, long stretches of repeated re-election and the absence of opposition often signal a satisfied constituency. Incumbency also promotes communication between representative and constituent. The longer the selected representative remains in office, the more chances constituents will have had to see the representative in person or to write and receive a reply, and the more likely they will be to have built a relationship that, although necessarily distant, facilitates, at least in the best case, warranted trust and further communication. Moreover, with a good, internally motivated representative, incumbency increases voters’ control over the legislature. Constituents are able to retain in office someone who has learned the ropes, made the contacts, become expert on relevant policies, and, as a result of experience, has more impact on legislative outcomes than a newcomer could have.15 In a selection model, voters do not lose power. They do not exercise power by exercising control over the representative, getting the representative to do something the representative would otherwise not do through the threat of being voted out. (In a pure selection model, voters would have zero control over the representative.) Rather, they exercise power over the legislature, getting it to do what it would otherwise not do by placing in it a representative who will pursue the policies they favour. The

A “selection model” for representation

21

voters’ ultimate goal is not control over a particular representative but their fair share of control over the entire legislature. In a selection model, incumbency usually increases the voters’ control over the legislature by making the instrument of their control – the representative – more effective. The negative side to incumbency derives primarily from the difficulty of removing the representative. Normatively, the viability of the selection model depends not only on good information at the time of initial selection but also on ease of removal when the selected representative is no longer aligned with the constituents’ needs and desires. Yet in contrast to incumbency per se, incumbency advantage “lowers electoral control over the institutions of government” in both a sanctions and a selection model (Ansolabehere and Snyder 2002: 329).16 When facing the requirement of re-election, incumbents can draw, unlike challengers, both on the renown that comes with their office and on the party’s and the voters’ gratitude for past policy work and constituency service. Incumbents also have government subsidies that allow them to communicate with their constituents, while challengers must find funds to pay for the equivalent publicity. Thus if the legislator changes (becoming lazy, corrupt, legislature-centred, or simply tired of going out on the hustings) or if the constituency changes (through district boundary changes, generational or mobility replacement, or an exogenous policy shock), the very capacities for communication and trust-building that are normatively positive features of incumbency when representative and constituents are well aligned make it harder for challengers to get their credentials before the public when that alignment slips. Because incumbency can indicate either genuine constituent satisfaction or constituent disempowerment, coming to a normative conclusion about any incumbent first requires asking some simple questions: Does the representative by and large promote policies and a larger political direction that the majority of constituents approve? How satisfied are the majority and minority of constituents with their representative? It also requires asking less simple and more contextual questions: Is satisfaction with the representative the result of ignorance or manipulation? Will the existing media system publicize departures from citizen preferences or interests? Are there lively interest groups that could bring attention to such departures and promote policies that would not otherwise be on the agenda? Are the citizens active in other forms of politics and therefore able to inform themselves easily and take action if their current representative no longer seems appropriate? Is the internal party system vital, self-policing, and continually infused with new activists and new ideas, or is it dormant, in the pocket of the incumbent, or relatively closed to newcomers? Although the answers to such questions are hard to come by, only they will allow us to judge the legitimacy of the citizens’ “tacit consent” to a particular incumbency. We need such answers before we can treat high rates of incumbency as evidence of democratic failure. In addition to possibly signalling reduced voter control over democratic institutions due to incumbency advantage or corruption, extended incumbency

22

Mansbridge

may also exact a price in citizen activism. One would expect even positive and genuinely chosen long incumbencies to undermine citizen political activity because once constituents have selected an honest and competent representative with approximately their own political views, they can turn to other matters, letting the representative get on with the business of politics. If representatives retain their constituents’ loyalty and thus face few challenges at election time, the voters also will not have the stimulus of a close election to bring their attention back to the political issues at stake. Yet the picture is not completely bleak. In some contexts, homogeneous districts with long-standing incumbents and satisfied constituents free up citizens for other forms of citizen activism. Rousseau wanted citizens to “fly to the assemblies” from a desire to do their part in bringing a good polity into being, not from fear that the other side might win.17 A politics that produces turnout through partisan enmity or conflicting material interests is not likely to cultivate attachment to the common interest. In sum, although a selection model of political representation may reduce citizen activism, that result is far from necessary and the close race that is the ideal of many a political scientist does not have indisputably positive normative qualities. When a selection model of political representation with long incumbencies is feasible, it may be better to look for incentives to citizen activism less in the process of electing representatives than in more direct forms of action.

Conclusion: the normative side of selection Ideally, politics should not be a market relationship. Voting for one candidate rather than another should not resemble buying one toaster rather than another. At its best political representation demands a communicative relationship. In political representation, “we”, plural, have many relationships with our representative and we do not always agree. “We” also have conflicting interests. In regard to one or another policy both the representative–constituent relationship and the constituent–constituent relationship may be fraught with tension. In this tension, if the representative cares only about votes and the minorities in the constituency do not have enough votes, cannot get them, and cannot work in coalition with the majority, that ends the story. In a selection model much depends on the character of a representative’s internal motivation. Normatively, such a representative should be open to persuasion. Even losing minorities should be able to respect the representative while making a deliberative case to both citizens and the representative that the representative’s policy choice was wrong. The characters of both the representative and the constituency are also formed in the representative–constituent relation. A sanctions model, based on inducing a representative’s preferences through the incentive of re-election, replaces an expectation of common interest between representative and constituent with an expectation that representatives will “shirk”, ideologically or in other ways, and that sanctions are therefore necessary to bring them to heel. Such expectations are sometimes amply warranted. But they are not costless.

A “selection model” for representation

23

These expectations affect who enters politics, what relationships they attempt to establish with their constituents, and how those constituents see the political system. A selection model, when feasible, will attract more internally motivated candidates into politics. The character of these representatives and their approaches to politics will in turn have subtle or even major effects on how the constituents think of their polity and what they want from it. Despite its disadvantages, in short, the selection model has several normatively satisfying features. It couples intrinsic motivation on the part of the representative with control (over the legislature, not the representative) on the part of the voter. Constituents will tend to get behaviour that is more dedicated, more adaptive to long-run goals, and more flexible in the means to those goals from an honest, competent representative who intrinsically wants to pursue the same kinds of policy they want than from a representative who, seeking only reelection, responds only to the promise of votes or the threat of withdrawing those votes. The harder it is to monitor and sanction, the more reason constituents will have to distrust a representative who acts only for extrinsic rewards.18 Constituents can also take some satisfaction from the quality of their relationship with an intrinsically motivated representative. A selection model privileges commitment to the common interest over the more self-interested motive of desiring re-election. Many constituents can also feel – correctly – that their representative is like them, at least in policy objectives and often in their overall approach to life. The constituent–representative relationship can be one of trust, even warmth. Although by necessity distant, it can partake of some of the characteristics of friendship, with an expectation of common interests. Without knowing each other personally, both partners in the relationship can wish the other well. Particularly on a national scale in a populous country, close communication between represented and representative is difficult, making this kind of relationship hard to foster. A selection model both builds on and creates such relationships. A contingency theory of representation adapts the size of the selection core and sanction periphery to the empirical context. While monitoring should never be absent and representatives should always be removable, a selection model works well in relatively uncorrupt polities where, by and large, representatives sort themselves into districts where they can present themselves as being what they are and internally want to be, and where constituents have sufficient information at the time of selection to make a reasonable choice. When feasible, a selection model will direct the attention of constituent and representative to the common interest and will foster that interest. It will also be more efficient in many instances than a model based on monitoring and sanctions. Thus, all other things being equal, societies that have available the ingredients necessary to base their representative systems primarily on uncorrupt selection should succeed in competition with societies in which the political system has to operate primarily through sanctions.

24

Mansbridge

Notes 1 This chapter is adapted from Mansbridge (2009). “A ‘Selection Model’ for Political Representation”, The Journal of Political Philosophy 17(4): 369–398. For funding my time to write, I thank the Radcliffe Institute for Advanced Study; for comments I thank the participants at the Austrian Political Science Association, the Oxford Political Thought Conference, the Centre for the Study of Democratic Institutions at the University of British Columbia, the Passmore Lecture series at the Australian National University, the E.U. Recon Workshop on Representation and Institutional Make-up in Brussels, and the Faculty Research Seminar at the Kennedy School of Government, and especially Rainer Brauböck, Iris Bohnet, Geoffrey Brennan, Archon Fung, Christian Grose, Mark Warren, and Richard Zeckhauser. 2 Some kinds of representation (e.g. surrogate representation, Mansbridge 2003) and representational claims (Saward 2006) do not take a principal–agent form. This chapter will discuss representation only as a principal–agent problem. It will also concentrate largely on the US case, which has generated most of the work promoting a sanctions model. In party list proportional representation systems and some other disciplined party systems, the party usually selects representatives primarily on party loyalty (including policy preference), competence and honesty, then adapts its own platform to a greater or lesser degree in response to potential voter sanctions. Parties themselves range from the most internally motivated or “gyroscopic” (responding largely to the members’ principles and ideology) to the most externally motivated (responding largely to electoral inducements or sanctions). 3 See Mansbridge 2009 unabridged, for a history of this eclipse and the beginning of a return to the selection model via, inter alia, the work of Fearon 1999 and Besley 2006. 4 See also the debate, reported in Mansbridge 2009 unabridged, between Carl Friedrich (1940) and Herbert Finer (1941) on the best way to secure responsible and ethical conduct among public officials. Friedrich argued for selecting self-motivated persons who sincerely wanted to work for the public interest and reinforcing those internal commitments, while Finer argued for external sanctions and controls. 5 By corruption here I mean only “behaviour which deviates from the formal duties of a public role because of private-regarding (personal, close family, private clique) pecuniary or status gains; or violates rules against the exercise of certain types of private-regarding influence” (Nye 1967: 419). 6 Although in a selection model voters usually give the representative more leeway, sanctions models are not necessarily inflexible. “Anticipatory representation”, in which the representative’s behaviour is induced by the prospect at the next election of the voters’ retrospective judgements on the representative’s past performance, is far more flexible than “promissory representation”, in which the representative’s behaviour is induced by fear of retaliation if he or she does not keep promises made at the first election (Mansbridge 2003). 7 Bohnet et al. (2001: 132) point out that in a “low-enforcement environment”, firms “heavily invest in screening of potential employees, stressing that character is more important than the possession of specific skills”. Similarly, when agents must act with discretion, a selection model is most valuable. 8 Bianco (1994: 56) and Mansbridge (1999: 305). Similarly, in voluntary interactions in a collective action problem, when both co-operators and defectors search for signals of a potential partner’s character type, the co-operators are most likely to succeed when such signals are most obvious (Frank 1988). 9 Besley and Ghatak (2005) thus advocate heterogeneity among missions, whether in the non-profit or other sectors (e.g. among different kinds of public school), to make this sorting process most efficient. Different kinds of incentive attract different kinds of individual (Bohnet and Oberholzer-Gee 2002), with the “currency” of reward helping in the sort (Brennan 1996).

A “selection model” for representation

25

10 The opinions of the US Supreme Court exemplify pure narrative accountability. In deliberative accountability principal and agent both ask questions and give answers, exploring whether or not they remain mutually aligned and whether the grounds of their alignment might have changed. 11 I thank Richard Zeckhauser for this suggestion. 12 See Goodin (2003: 12) for network accountability in non-profit institutions, arguing that accountability based on “praising or shaming and shunning” will work most effectively among actors motivated by the public good. See also Keohane and Nye (2001) on network accountability in international organizations. Yet networks can also function as “cosy cabals covering one another’s incompetence”; thus the networks themselves often require some external monitoring and sanctions (Goodin 2003: 41). For a theory of “dynamic accountability” that builds in some external monitoring and sanctions, see Sabel and Zeitlin 2008. For a contingency theory of accountability based on the sanction and selection models of representation, see Mansbridge 2014. 13 As Rehfeld (see Chapter 3) notes, Burke himself did not use the word “trustee”. 14 By contrast, in modern societies principals may decide that they want to select agents more talented and virtuous than they, but this not entailed by the model. 15 This pattern would hold even without institutional seniority advantages in the legislature. In the ideal, each citizen should have equal control over the legislature. It seems hard to produce this result, however, without forbidding the accumulation of experience among legislators, a course with normative and practical costs. 16 Incumbency advantage is no more than “the increment to the vote margin that a candidate gains by virtue of being the incumbent” (Erikson et al. 1993: 100); it does not include the underlying partisan advantage that accrues from a relatively homogeneous political district (Alford and Brady 1989; Oppenheimer 2005: 137). 17 Rousseau ([1762] 1947), opposing a representative conception of sovereignty, painted a compelling picture of the decline of a polity when its citizens fail to desire the common good. 18 See Warren and Mansbridge et al. 2013 for the importance of privacy in political negotiation, with a corresponding loss in the voters’ capacities to monitor the representatives’ behaviour.

References Aberbach, Joel (1990). Keeping a Watchful Eye: The Politics of Congressional Oversight. Washington, DC: Brookings. Alford, John R. and David W. Brady (1989). “Personal and Partisan Advantage in U.S. Congressional Elections.” In: Lawrence C. Dodd and Bruce I. Oppenheimer (eds) Congress Reconsidered, 4th edition. New York: Praeger Publishers, 156–169. Anechiaro, Frank and James B. Jacobs (1996). The Pursuit of Absolute Integrity: How Corruption Control Makes Government Ineffective. Chicago, IL: University of Chicago Press. Ansolabehere, Stephen and James M. Snyder (2002). ‘The Incumbency Advantage in U.S. Elections: An Analysis of State and Federal Offices, 1942–2000.’ Election Law Journal 1: 315–338. Baron, James N. and David Kreps (1999). Strategic Human Resources. New York: Wiley. Behn, Robert D. (2001). Rethinking Democratic Accountability. Washington, DC: Brookings Institution Press.

26

Mansbridge

Besley, Timothy (2006). Principled Agents? The Political Economy of Good Government. Oxford: Oxford University Press. Besley, Timothy and Maitreesh Ghatak (2005). ‘Competition and Incentives with Motivated Agents.’ The American Economic Review 95: 616–636. Bianco, William T. (1994). Trust: Representatives and Constituents. Minneapolis, MN: University of Michigan Press. Bohnet, Iris and Felix Oberholzer-Gee (2002). “Pay for Performance: Motivation and Selection Effects.” In: Bruno S. Frey and Margit Osterloh (eds) Successful Management by Motivation. Berlin: Springer, 119–139. Bohnet, Iris, Bruno S. Frey and Steffen Huck (2001). ‘More Order with Less Law: On Contract Enforcement, Trust, and Crowding.’ American Political Science Review 95: 131–144. Brennan, Geoffrey (1996). “Selection and the Currency of Reward.” In: Robert E. Goodin (ed.) The Theory of Institutional Design. Cambridge: Cambridge University Press, 262–272. Brunell, Thomas L. (2006). ‘Rethinking Redistricting: How Drawing Uncompetitive Districts Eliminates Gerrymanders, Enhances Representation, and Improves Attitudes Toward Congress.’ PS: Political Science & Politics 39: 77–85. Buchler, Justin (2005). ‘Competition, Representation and Redistricting: The Case Against Competitive Congressional Districts.’ Journal of Theoretical Politics 17: 431–463. Deci, Edward L. and Richard M. Ryan (1985). Intrinsic Motivation and Self-determination in Human Behaviour. New York: Plenum. Deci, Edward L., Richard Koestner and Richard M. Ryan (1999). ‘A Meta-Analytic Review of Experiments Examining the Effects of Extrinsic Rewards on Intrinsic Motivation.’ Psychological Bulletin 125(6): 627–668. Dixit, Avinash (2002). ‘Incentives and Organizations in the Public Sector: An Interpretative Review.’ Journal of Human Resources 37: 696–727. Erikson, Robert S., Gerald C. Wright Jr and John P. McIver (1993). Statehouse Democracy: Public Opinion and Policy in the American States. New York: Cambridge University Press. Fearon, James (1999). “Electoral Accountability and the Control of Politicians: Selecting Good Types Versus Sanctioning Poor Performance.” In: Adam Przeworski, Susan C. Stokes and Bernard Manin (eds) Democracy, Accountability, and Representation. New York: Cambridge University Press, 55–97. Fenno, Richard F. Jr (1978). Home Style: Members in Their Districts. Boston, MA: Little, Brown. Finer, Herman (1941). ‘Administrative Responsibility in Democratic Government.’ Public Administration Review 1: 335–350. Frank, Robert H. (1988). Passions Within Reason: The Strategic Role of the Emotions. New York: W.W. Norton & Co. Frey, Bruno S. (1997). ‘A Constitution for Knaves Crowds Out Civic Virtues.’ The Economic Journal 107: 1043–1053. Friedrich, Carl J. (1940). “Public Policy and the Nature of Administrative Responsibility.” In: Carl Friedrich and Edward S. Mason (eds) Public Policy: A Yearbook of the Graduate School of Public Administration. Cambridge, MA: Harvard University Press, 3–24. Gay, Claudine (2002). ‘Spirals of Trust: The Effect of Descriptive Representation on the Relationship Between Citizens and their Government.’ American Journal of Political Science 46: 717–732.

A “selection model” for representation

27

Goodin, Robert E. (1980). ‘Making Moral Incentives Pay.’ Policy Sciences 12: 131–145. Goodin, Robert E. (1982). Political Theory and Public Policy. Chicago, IL: University of Chicago Press. Goodin, Robert E. (2003). ‘Democratic Accountability: The Distinctiveness of the Third Sector.’ European Journal of Sociology 44: 359–396. Guinier, Lani (1994). The Tyranny of the Majority: Fundamental Fairness in Representative Democracy. New York: Free Press. Herzog, Don (1998). Poisoning the Minds of the Lower Orders. Princeton, NJ: Princeton University Press. Horowitz, Donald R. (2003). ‘Electoral Systems: A Primer for Decision-makers.’ Journal of Democracy 14: 115–127. Issacharoff, Samuel (2002a). ‘Gerrymandering and Political Cartels.’ Harvard Law Review 116: 593–648. Issacharoff, Samuel (2002b). ‘Why Elections?’ Harvard Law Review 116: 684–695. Keohane, Robert O. and Joseph S. Nye (2001). “The Club Model of Multilateral Cooperation and Problems of Democratic Legitimacy.” In: Roger B. Porter, Pierre Sauvé, Arvind Subramanian and Americo Beviglia Zampetti (eds) Efficiency, Equity, and Legitimacy: The Multilateral Trading System at the Millennium. Washington, DC: Brookings Institution Press, 264–294. Kingdon, John (1981). Congressmen’s Voting Decisions. New York: Harper and Row. Kruglanski, Arie W., Aviah Riter, Asher Amitai, Margolin Bath-Shevah, Leorah Shabtai and Daliah Zaksh (1975). ‘Can Money Enhance Intrinsic Motivation? A Test of the Content-consequence Hypothesis.’ Journal of Personality and Social Psychology 31: 744–750. Lawless, Jennifer (2005). It Takes a Candidate: Why Women Don’t Run for Office. Cambridge: Cambridge University Press. McCubbins, Matthew S. and Thomas Schwartz (1984). ‘Congressional Oversight Overlooked.’ American Journal of Political Science 28: 166–201. McGregor, Douglas (1960). The Human Side of Enterprise. New York: McGraw-Hill. Manin, Bernard (1997). The Principles of Representative Government. Cambridge: Cambridge University Press. Mansbridge, Jane (1980). Beyond Adversary Democracy. New York: Basic Books. Mansbridge, Jane (1990). “The Rise and Fall of Self Interest.” In: Jane Mansbridge (ed.) Beyond Self Interest. Chicago, IL: University of Chicago Press, 3–22. Mansbridge, Jane (1999). “Altruistic Trust.” In: Mark Warren (ed.) Democracy and Trust. Cambridge: Cambridge University Press, 290–309. Mansbridge, Jane (2003). ‘Rethinking Representation.’ American Political Science Review 97: 515–528. Mansbridge, Jane (2009). ‘A “Selection Model” of Political Representation.’ Journal of Political Philosophy 17(4): 369–398. Mansbridge, Jane (2014). “A Contingency Theory of Accountability.” In: Mark Bowens, Robert E. Goodin and Thomas Schillemans (eds) The Oxford Handbook of Public Accountability. Oxford: Oxford University Press, 55–68. Miller, Dale T. (1999). ‘The Norm of Self-interest.’ American Psychologist 54: 1053–1060. Mutz, Diana C. (2006). Hearing the Other Side: Deliberative Versus Participatory Democracy. Cambridge: Cambridge University Press. Neblo, Michael, Kevin Esterling, Ryan Kennedy, David Lazer and Anand Sokhey (2010). ‘Who Wants to Deliberate – and Why?’ American Political Science Review 104(3): 566–583.

28

Mansbridge

Nye, Joseph S. (1967). ‘Corruption and Political Development: A Cost-benefit Analysis.’ American Political Science Review 61: 417–427. O’Neill, Onora (2002). A Question of Trust. Cambridge: Cambridge University Press. Oppenheimer, Bruce I. (2005). “Deep Red and Blue Congressional Districts: The Causes and Consequences of Declining Party Competitiveness.” In: Lawrence C. Dodd and Bruce I. Oppenheimer (eds) Congress Reconsidered, 8th edition. Washington, DC: CQ Press, 135–157. Persily, Nathaniel (2002). ‘In Defense of Foxes Guarding Henhouses: The Case for Judicial Acquiescence to Incumbent-protecting Gerrymanders.’ Harvard Law Review 116: 684–695. Philp, Mark (2004). Accountability and Democracy: A Millean View. Unpublished ms. Prewitt, Kenneth (1970). ‘Political Ambitions, Volunteerism, and Electoral Accountability.’ American Political Science Review 64: 5–17. Rousseau, Jean-Jacques (1947 [1762]). The Social Contract. New York: Hafner. Sabel, Charles F. and Jonathan Zeitlin (2008). ‘Learning from Difference: The New Architecture of Experimentalist Governance in the EU.’ European Law Journal 14: 271–327. Saward, Michael (2006). ‘The Representative Claim.’ Contemporary Political Theory 5: 297–318. Schumpeter, Joseph (1962 [1942]). Capitalism, Socialism, and Democracy. New York: Harper and Row. Shapiro, Susan P. (2005). ‘Agency Theory.’ Annual Review of Sociology 31: 263–284. Storing, Herbert J. (1981). The Complete Anti-Federalist. Chicago, IL: University of Chicago Press. Sunstein, Cass R. (2003). Why Societies Need Dissent. Cambridge, MA: Harvard University Press. Warren, Mark, Jane Mansbridge, André Bächtiger, Maxwell A. Cameron, Simone Chambers, John Ferejohn, Alan Jacobs, Jack Knight, Daniel Naurin, Melissa Schwatzberg, Yael Tamir, Dennis Thompson and Melissa Williams (2015 [2013]). “Deliberative Negotiation.” In: Jane Mansbridge and Cathie Jo Martin (eds) Political Negotiation a Handbook. Washington, DC: Brookings Institution Press, 141–198. Weisbrod, Burton Allen (1988). The Nonprofit Economy. Cambridge, MA: Harvard University Press.

3

Beyond trustees and delegates Andrew Rehfeld

Introduction1 If there is an oldest question concerning political representation it might be whether representatives should act as trustees or delegates to their constituents. If there is a “new insight” into this question, it is that the distinction between “trustees” and “delegates” obscures three subordinate questions about which we are primarily concerned: 1

2 3

The representative’s source of judgement: Should representatives be selfreliant, or should they depend upon their constituents’ views about how to vote? The aims of legislation: Should representatives promote the good of all or, more narrowly, the good of their constituents? The representative’s responsiveness to sanction: Should a representative be more or less responsive to the prospect of re-election (or other sanction)?

Because “aims”, “judgement”, and “responsiveness” may vary independently of one another, there are eight possible combinations (2  2  2) of which “trustees” and “delegates” are but two. “Trustees” are generally described as (1) relying on their own judgement, concerning (2) the good of the whole, and (3) less responsive to sanctioning (acting instead according to civic virtue). By contrast “delegates” are generally described as (1) relying on their constituents’ judgement, concerning (2) the good of their constituents, all the time being (3) more responsive to sanctions (in particular, the hope of re-election). That leaves six other combinations left un-described and accounted for. Once the distinctions that are collapsed in the “trustee/delegate” debate are made apparent, I believe that we can see that they arise not because of anything particular to representation, but rather any time binding decisions are made and, in particular, when a decision maker wants to do something that in some sense ought to be done in cases where those to whom it is done do not approve. As long as lawmakers aim to make law that is substantively just and that corresponds to the preferences of those over whom it governs, the three subsidiary distinctions arise.2 This tension, then, applies with no more

30

Rehfeld

or less force in a direct democratic assembly in which all citizens participate 3 or in a nondemocratic monarchy in which a single individual wants to make law that is both just and that conforms to the preferences of those over whom it governs. Thus, despite its historical lineage, the main tensions described by the trustee/delegate distinction apply to political representatives on account of their taking on the role of decision makers rather than on their account of their being representatives per se.4 If this is correct, then the trustee/delegate problem is not a core problem of political representation per se, nor would it indicate a particular ethics of political representation in which representatives are said to take on unique obligations by virtue of their being representatives.5 Political representatives may (and almost certainly do) take on a particular obligation to care about a particular group (and not some other one) or to give a third party’s judgement (and not some other one’s) special consideration in their decision making. 6 But this is a specification of the core substantive problem that emerges from a broader normative account of decision making and location of authority rather than from some conceptual, normative, or empirical fact about political representation. The casting of the “trustee/delegate” problem as particular to political representation thus constitutes a substantive error that fails to distinguish the tension between citizen preferences and normative ideals from professional obligations that any decision maker (whether monarch, representative, or citizen in a direct democracy) takes on when he or she takes on the role of making law or other decisions. The analysis of this chapter also provides a way to understand how a core distinction that we attribute to democratic, representative governments applies no more or less to nondemocratic contexts, whether in the case of nongovernmental organizations (NGOs) or nondemocratic government. All decision makers whose decisions affect others will face trade-offs between following justice and following the preferences of those their decision affects, regardless of whether they are political representatives, democratic, or elected at all. The breadth of this treatment then is applicable both to familiar democratic contexts of representative government, as well as quasi- or nondemocratic decision makers who increasingly act on the world stage. In what follows, I do not address the question of where authority should be located. This may appear to ignore a key distinction because the traditional categories of “delegate” and “trustee” appear at first to denote different locations of authority about how a representative ought to vote. In “delegated” representation, we might imagine that authority is located directly with a home constituency, whereas with “trusteeship”, we might imagine that authority is located with the representative him- or herself, who now has the power to act. Thus, it is often said that the delegate is less autonomous, or independent, in acting than is the trustee (Pitkin 1967). In truth, however, the emphasis on authority and autonomy is of less analytical value than it first appears. On the surface, the authority/autonomy distinctions appear to emphasize who has the authority to cast a vote on

Beyond trustees and delegates

31

proposed legislation. Yet, this cannot be right because by presumption only the representative has the authority to cast the vote no more or less when bound by constituent instructions than when independent of them, no more, that is, when acting as a delegate or trustee. The question of authority must therefore concern who has the right to decide how to cast votes, or more generally, how to make decisions, and this question is at the core of the three underlying distinctions of aims, sources of judgement, and responsiveness. Once we specify these distinctions, we can then see that the location of authority is wholly independent of them. So-called trustees can be instructed, mandated, and bound by their home constituency’s authority to act as trustees, and so-called delegates might act that way because that is called for as the proper exercise of their independent authority. This creates certain apparent paradoxes like the “independent delegate” and the “mandated trustee” but only because the core concepts of trustee and delegate are obscuring so much to begin with. 7 One final note: In what follows, I bracket political parties and treat preferences as fixed in order to work through one set of conceptual and political puzzles. Although parties and mutable preferences will introduce greater complexity to the analysis, the core problems and puzzles articulated here would remain. Indeed, with reference to the questions of this chapter, political parties do not fundamentally alter the conceptual distinctions here. However, they do insert an additional level between voters and lawmakers. The trustee/delegate debate has focused on the relationship between voters and their individually elected representatives who are lawmakers because it emerged early as a political issue in the Anglo-American world where individual representatives are elected by individual constituencies. In a party system the trustee/delegate distinction (as well as its three underlying components) applies no more or less, but now describes the relationship either between voters and the parties, or the parties and the activity of individual members in the legislature. The insertion of the party as a mediating step between voters and policies is of course a separate issue; conceptually, however, the dynamic I describe here does not change.8 The argument continues over three more sections. First, I briefly trace the collapse of the three underlying distinctions into the “trustee/delegate” dichotomy, and provide context for and explain the relevance of the problem today. Second, I illustrate more fully the three conceptual categories that the “trustee/delegate” debate collapses – aims, sources of judgement, and responsiveness – and show the analytical value that such separation generates. Third, I conclude by explaining why the three distinctions are faced by political representatives only insofar as they take on the role of decision makers and not by virtue of the fact that they are representatives, let alone democratic, per se.

Trustees vs. delegates in historical context The treatment of the “trustee/delegate” problem as distinctive to democratic representative government (rather than a more general problem of decision

32

Rehfeld

making) may be due to the simultaneous emergence of the practice of political representation and notions of popular sovereignty in which political representatives were increasingly used to express the decisional authority of the people. The distinction between “trustees” and “delegates” traces back at least to Magna Carta in the early thirteenth century. At that time, representatives had little authority to act, let alone judge or enact law, but were merely to give their constituency’s (town or borough) assent to the king’s demand for money (Fasolt 1991). Over the next 500 years, that relationship would be complicated as representatives were given more decisional authority over law, and, simultaneously, popular sovereignty emerged as a governing ideal. By the time of the American Revolution, the question about whether representatives should be “trustees” or “delegates” of their constituents overlaid with the separate view that law should in some sense reflect the preferences of the people (Rehfeld 2005). The ideal that law should be both just and reflective of popular preferences – an ideal to which any kind of lawmaker might in practice aspire regardless of whether institutionally or legally bound to the people’s will – became hard to distinguish from the practice of using political representatives as lawmakers themselves. The underlying distinctions that were later subsumed by the “trustee/delegate” labels arose in debates about political representation in the eighteenth century. Scholars point to Edmund Burke as the exemplar of the “trustee” position and, in contrast, to a more defuse set of anti-Federalists to characterize the “delegate” position. In truth, their positions were more complicated and held by a wide variety of actors, each of whom had nuanced positions on the aims, sources of judgement, and responsiveness that should animate lawmakers more generally (Burke 1774; see also Kenyon 1966; Pitkin 1967; Rakove 1996; Rehfeld 2005). Consider, for example, the proposal to constitutionally protect the right of US citizens to instruct their representatives. The proposal, made by Representative Thomas Tudor Tucker in 1789, would have become part of the First Amendment to the US Constitution (concerning basic freedoms of speech, assembly, and religion). The proposal was soundly defeated in part by arguments, such as this one by Thomas Hartley (Kurland and Lerner 1987): If, in a small community, where the interests, habits, and manners are neither so numerous or diversified, instructions bind not, what shall we say of instructions to this body? Can it be supposed that the inhabitants of a single district in a State, are better informed with respect to the general interests of the Union, than a select body assembled from every part? Can it be supposed that a part will be more desirous of promoting the good of the whole than the whole will of the part? I apprehend, sir, that Congress will be the best judges of proper measures, and that instructions will never be resorted to but for party purposes, when they will generally contain the prejudices and acrimony of the party, rather than the dictates of honest reason and sound policy. (Annals 1: 733–745, 13: 39)

Beyond trustees and delegates

33

Hartley’s response illustrates two relevant points. First, Hartley identifies all three underlying distinctions that I claim comprise the “trustee/delegate” distinction. The aims of legislation should be the good of all (modified when appropriate by the good of a part); the judgement should be made by the lawmaker because of his or her proper deliberative position in the legislature. The lawmaker should also be less responsive to external sanctions that would reflect “the prejudices and acrimony of the party”, relying instead on his or her own “honest reason” and the desire for “sound policy”. Second, the language that Harley used conflates “representative” with the role of “lawmaker” (or, more generally, “decision maker”), a role that representatives take on once they are representatives. Although other speakers (and Hartley himself) would refer to “representatives” as well, his argument here is framed in terms of an ethics of law-making more generally. Stepping back from this particular example, we can characterize the trustee and delegate positions by systematically fixing their underlying three parts. The trustee position, often attributed to Edmund Burke, specifies that in making law, lawmakers should (1) aim at the national good; relying on (2) their own judgement of what constitutes that national good (after due deliberation); and in a manner that is (3) less responsive to electoral sanction, motivated instead by some form of civic virtue. Burke, for example, explicitly identifies the first two features. Defending the national good as the proper aim of legislation, he writes: If the local constituent should have an interest, or should form an hasty opinion, evidently opposite to the real good of the rest of the community, the member for that place [i.e., the representative] ought to be as far, as any other, from any endeavour to give it effect. (Burke 1774: 448) Separately, Burke defended the view that representatives rather than their constituents or other third party ought to be the source of judgement about what constitutes the national good. “Your representative owes you, not his industry only, but his judgment; and he betrays, instead of serving you, if he sacrifices it to your opinion” (Burke 1774: 446). Given the attribution of “trustee” to Burke’s conceptions, it may come as some surprise that he never uses the term in his “Speech to the Electors at Bristol”, keeping the parts more clearly separate. But two centuries later, those views have been conflated into the designation “trustee”. In her nuanced description, Pitkin attributes to Burke the “trustee” distinction, and rightly notes that in Burke’s view a representative ought to act from a sense of right and wrong and be less responsive to sanctions. “The first concept of representation encountered in Burke’s thought is thus an aristocracy of virtue and wisdom governing for the good of the entire nation” (Pitkin 1967: 218). Adrian Vermeule’s recent formulation of “trusteeship” is characteristic of this tendency to collapse all three features into a single term: “Burke’s trustee model of representation [is] a normative stance that

34

Rehfeld

sees a good representative as one who exercises independent judgment for the common weal, rather than simply acting so as to satisfy the preferences of a … constituency” (Vermeule 2007: 218). This is similarly reflected in more popular accounts of the trustee position.9 Delegate views of representation, like those Burke was countering, 10 press instead for a close correspondence between the views of an electoral constituency and the votes of representatives. The view was perhaps best captured by the opponents of the American Constitution in the 1780s, the so-called anti-Federalists. For the delegate (1) the aims of legislation are most often the good of a particular electoral constituency, (2) citizens are relied on to be the source of judgement about what constitutes that good, and (3) representatives are supposed to be highly responsive to the threat of sanction. Thus, so-called “delegates” seek to secure their constituents’ interests as their constituents so define them. Broadly speaking, then, the representative as “delegate” is one who (1) aims at the good of his or her constituents, (2) as judged by his or her constituents, and (3) is more responsive to external sanction (election or the avoidance of legal penalties). The representative as “trustee” is one who (1) aims at the good of the whole, (2) as judged by the representative, (3) is less responsive to sanction, but acts on some form of civic virtue instead.

Separating the three distinctions and keeping them separate Let us then formalize the three distinctions that underlie the “trustee/delegate” debate, denoting an independent and independently important problem to consider. Each of these problems has an empirical and a normative framing: (1) does or should a representative–decision maker pursue the good of the whole or a part? (2) does or should he or she rely on his or her own judgement, or that of a third party? and (3) does or should he or she be more responsive or less responsive to sanctions for his or her work? Indeed, if each of these three separate distinctions points at two directions (the whole vs. a part; the judgement of the representative–decision maker vs. the judgement of others; behaviour that is more or less responsive to sanction) and any combination of these three is possible, then there are eight possible ideal types (2  2  2). The use of the binary distinction “trustee” and “delegate” thus obscures six other ideal types, obscures how each variable may vary, and makes normative and empirical analysis less precise. We expand each of these in turn. First, we can broadly distinguish the ends at which representative decision makers aim based on their level of universality or particularity – whether, that is, they aim at the good of all or the good of some part. 11 At one extreme, we have representative–decision makers who aim at the good of all. Although the “good of all” has traditionally been described in terms of the good of “the nation”, there is no reason to conceptually limit this – the extreme would be the good of all humanity through conceptions of global justice. At the other extreme, we have representative–decision makers who aim toward the good of

Beyond trustees and delegates

35

a particular part. Although this particular part has traditionally been described in terms of the good of the representative’s electoral constituency, there is no reason to conceptually limit its application – the part might be an interest group or, at the extreme, the good of a single individual. This distinction is useful because it keeps our attention focused on how universal or particular a law or decision is or ought to be, quite apart from its substance. “Trustees”, as we have seen, are generally described as aiming at a more universal good – the good of all; “delegates” at the good of a part. Second, once we determine the ends toward which a law aims the representative–decision maker must make an epistemic determination about means and ends, judging whether a proposed bill is likely to achieve its particular goals: will it be good for the whole or the part at which he or she aims? Here we can distinguish among the sources of this judgement based on whether he or she relies on his or her own judgement or the judgement of someone else. This distinction is useful because it draws our attention to the source of judgement about the law (or decisions more generally), quite apart from the ends toward which the law or decision should aim. As we saw previously, “trustees” are generally described as using their own judgement; “delegates” as relying on the judgement of others. Finally, we can distinguish between the extent to which a representative– decision maker is responsive to sanctions.12 This distinction is useful because, with Mansbridge (see Chapter 2; also Mansbridge 2003), we might believe that the level of responsiveness is important for reasons of control or oversight of a representative–decision maker. In particular, to the extent we cannot effectively reward or sanction their every action, or merely do not want to take the time to do so, we might want representatives who do not need that sanctioning to do what they say they are going to do.13 As we saw, “trustees” are generally described as being less responsive to electoral sanctions (relying, historically, on “civic virtue”), whereas “delegates” are described as more responsive to electoral sanction. Each distinction is continuous, denoting a range of options. Lawmakers are likely to care (and perhaps ought to care) about the justice of a law for both the whole and the parts thereof, in some cases, more the one than the other. They are likely to rely (and ought to rely) sometimes and in part on their own judgement as to whether the law achieves its stated goals and sometimes on the judgement of others, particularly their constituents. Finally, they are likely to act (and ought to act), sometimes responding to electoral sanctions, sometimes not. This is why we speak of each of these three variables – aims, source of judgement, and responsiveness – as continuous rather than discontinuous, distinctions rather than dichotomies.14 We can now see that two of the three distinctions – aims and responsiveness – relate to some familiar labels. The pluralist/republican distinction, for example, highlights the debate about the proper aims of legislation. Pluralists such as Robert Dahl and David Truman took the view that lawmakers should promote the interests of their constituents directly (Dahl 1956; Truman 1953).

36

Rehfeld

Republicans, including Philip Pettit (1997) and Cass Sunstein (1993), argue that lawmakers should aim more directly at the good of all (Pettit 1997; Sunstein 1993).15 We can use Mansbridge’s idea of a “gyroscopic” representative (2003) to designate a lawmaker who is less responsive to sanctions, motivated instead by his or her own internal compass compared to the induced representative who is more responsive to sanction. 16 Because each distinction denotes a range, we can use these familiar labels to name the end points on the range, as summarized in Table 3.1. For the aims of legislation, we can call those who aim at advancing the good of all, “Republicans”, and those who aim at the particular good of some subgroup, “Pluralists”. For the sources of judgement, we can call those who rely on their own judgement “self-reliant”, and those who rely on some other person’s or group’s judgement “dependent”. For responsiveness, we can follow Mansbridge’s (2003) terminology, and call “gyroscopic” those lawmakers who are less responsive to sanction and call “induced” those lawmakers who are more responsive to sanctions. We can now see why the traditional “trustee/delegate” distinction obscures too much. The representative as trustee is an amalgam of someone who is (1) a Republican (i.e., aims at the good of all), (2) self-reliant (i.e., follows his or her own judgement about whether a proposed law achieves this aim) and (3) less responsive to external sanctioning. In contrast, the lawmaker as delegate is an amalgam of someone on the other side of each of these three distinctions, who is (1) a pluralist (i.e., seeks to promote his or her own constituents’ interests), (2) dependent (i.e., relies on his or her constituents’ or someone else’s

Table 3.1 Categories, names, and descriptions of three distinctions Category

Name

Description

Aim of Legislation

Republican Aims

Those who aim to promote the greater good (often, but not necessarily, the “nation”). Those who aim to promote the good of a part (often, but not necessarily, the “electoral constituency”). Those who rely on their own judgement. Those who depend on the judgement of others. Those who are less responsive to sanctions. Those who are more responsive to sanctions.

Pluralist Aims

Source of Judgement

Self-Reliant Dependent

Responsiveness

“Gyroscopic” Induced

Beyond trustees and delegates

37

judgement about whether a proposed law achieves this aim) and (3) more responsive to sanctions (especially by the prospect of future electoral success). But if the traditional “trustee/delegate” view is captured by these amalgams, why not use those familiar terms? The more significant problem with relying on the traditional language is that it emphasizes only two out of eight possible permutations of the answers to the three questions posed. In doing so, it makes it harder to recognize variation among the three elements that we might describe, track, and evaluate for empirical and normative reasons. It clouds our understanding of the social world rather than clarifying it. To demonstrate this, we can simply articulate how different combinations of each distinction have manifested themselves either in the academic literature or in practice. Table 3.2 presents a summary of these eight possible positions.17 Reading across the table, the top row, labelled “Republican aims”, includes all who aim at the public good, instead of acting to pursue their constituents’ or some other group’s particular welfare. Exemplars of this position, if imperfectly so, are Pettit and Sunstein. The bottom row, labelled “Pluralist aims”, includes all who promote the advancement of their constituents’ (or some other particular group’s) interests; exemplars here are Dahl and Truman. Table 3.2 is also divided into two large columns signifying the level of responsiveness to sanction – whether gyroscopic or induced. Exemplars here are Jane Mansbridge’s account of a “selection model”18 and, on the induced side, the treatment of what constitutes “good” political representation by empirical political scientists who often reduce “representation” to “policy correspondence” (Rehfeld 2009: 219). Individually, each of the eight cells corresponds to descriptions of a range of political actors, ideal types that have appeared in history. In cell A we find the Burkean trustee who, as we discussed previously, looks after the national good, as he or she judges it to be so, less responsive to electoral pressures, motivated instead to act by republican virtue. In cell B, we have someone who, like the Burkean trustee, is motivated internally to seek the good of all, but now depends on his or her constituents’ judgement to determine what that good actually is. We might consider as an exemplar of this position civil servants who may serve the common good, based in large part on a third-party’s judgement about that good (e.g., as dictated by the legislation they follow), and who cast themselves as unresponsive to sanctions.19 In cell C, we again have someone pursuing the general welfare who uses his or her own judgement to decide what that good is, but unlike the Burkean trustee or civil servant, he or she is extremely responsive to electoral sanctions. I have labelled this “Madisonian” because I believe this is precisely the kind of representation that James Madison was defending in his contributions to The Federalist (Rehfeld 2011: 633–634). In cell D, we are now far from the Burkean trustee: here is someone who, like Burke, aims at the good of all, but, unlike Burke, looks to his or her constituents (or some other particular group) to define what that good is, and does so only because he or she wants to get re-elected. Although anti-Federalists were more concerned overall with

Table 3.2 Schematic conceptual space of three distinctions

Republican Aims (e.g., Petit; Sunstein)

Less Responsive to Sanction (e.g.: Gyroscopic; Mansbridge’s “Selection Model”)

More Responsive to Sanction (e.g.: Induced; Presumptions of Empirical Political Science)

Self-Reliant Judgement

Dependent Judgement

Self-Reliant Judgement

Dependent Judgement

A. Burkean Trustees Those who seek the good of the whole by relying on their own judgement, and are less responsive to sanctions (often because they believe that is simply the right thing to do).

B. Civil Servants Those who seek the good of the whole by relying on the judgement of others, but are less responsive to sanctions (often because they believe that is simply the right thing to do). F. Ambassadors

C. Madisonian Lawmakers Those who seek the good of the whole by relying on their own judgement and who are more responsive to sanction.

D. AntiFederalists Those who seek the good of the whole by relying on the judgement of others and who are more responsive to sanction.

G. Professionals (Lawyer, doctor, financial advisor) Those who seek the good of a part (often that of their constituents) by relying on their own judgement, and who are more responsive to sanction.

H. Pared-down Delegates

E. Volunteers

Pluralist Aims (e.g.: Dahl; Truman)

Those who seek the good of a part (often that of their constituents) by relying on their own judgement and are less responsive to sanctions (often because they believe that is simply the right thing to do).

Those who seek the good of a part (often that of their constituents) by relying on the judgement of others and are less responsive to sanctions (often because they believe that is simply the right thing to do).

Those who seek the good of a part (often that of their constituents) by relying on the judgement of others and who are more responsive to sanction.

the particular interests of local communities, there was a strain that defended instructions with legal sanction in order to pursue the national good. Each of the four cells in the second row demonstrates a different kind of pluralist. On the left in cell E, we have the pluralist who aims at the good of his or her particular constituency, but uses his or her own judgement to determine what that good is. Committed first and foremost to the good of his or her constituents, they are also less responsive to sanctions. This might

Beyond trustees and delegates

39

describe a tin-eared, passionate volunteer working to better a community as he or she judges it would improve without regard to what the community believes is right and undeterred by the fact that he or she will soon be asked to leave. (Or, more positively, the same sort of person who acts out of dedication to the job, not for the rewards that he or she will receive.) In cell F, we find someone who, like the volunteer, both aims at a partial good and is less responsive to sanctions. But here he or she follows, rather than ignores, his or her constituents’ (or some other particular group’s) judgement about what that good consists of. This characterizes the ideals of foreign ambassadors who want to advance their home nations’ interests, as judged by the government they serve, but who are motivated by an internal sense of purpose (“service to country”) rather than induced by threat of sanction. It might also describe a true populist who is devoted to following the judgement of the people rather than his or her own re-election prospects (Rehfeld 2011). The exemplar of cell G is the “professional”: a hired lawyer, doctor, or perhaps financial advisor who is an induced pluralist seeking the good of his or her client for the sake of external reward (e.g., payment, reputation), but relies on his or her own judgement as to what the good of his or her client actually is. Finally, in cell H, we arrive at the pared-down version of the delegate model: someone who pursues the good of his or her constituents as his or her constituents see it, motivated simply by his or her desire for external reward (i.e., re-election).20 Now, in any one of the eight cells, we have a complex character that specifies how an actor behaves, or provides a model for how he or she ought to behave, but broken down into analytically distinct parts. Of course, it is difficult to find in any single cell a complete description of any single view, any single representative, or actor. This may explain why in analyses of the political world we tend toward complex characterizations – political behaviour is usually complex. Indeed, once framed this way, we imagine other subsidiary and familiar distinctions that we should also think hard about. First among these other issues is the temporal reference point for all three distinctions. For example, should we care that the law aims at a good today, at election time, or in the longer term (Thompson 2005)? Should we use their judgement today about the aims of law, or reference what we believe their judgement will be in the future (Przeworski, Stokes, and Manin 1999)? We can ask similar questions about constituent conditions: should we rely on constituent judgement as it is, or as it would be from an ideal, educated, and deliberative position? The point here is not to work through these distinctions in a hundred- or thousand-celled table. The three that we emphasize derive from the traditional “trustee/delegate” distinction and are enough to provide a useful conceptual space on which these other distinctions can unfold.

Whither representation? Unpacking the trustee/delegate distinction into its three component parts provides a level of conceptual clarity that the crude “trustee/delegate”

40

Rehfeld

distinction masks. But it raises a new problem for contemporary theories of representation, particularly newer ones that give an account of representation both inside and external to democratic contexts. These new theories – whether Mansbridge’s selection model (see Chapter 2), Michael Saward’s emphasis on “claims-making” (2010), or my own audience-dependent view (Rehfeld 2006) – distance themselves from Pitkin’s traditional democratic formulation in two ways. First, they no longer rely on a principle–agent model (to which Mansbridge’s selection model provides a particularly useful alternative within electoral democracies). Second, in the latter two cases, they provide an ontological account of representation itself that does not depend on electoral systems at all. By contrast, the “trustee/delegate” debate has been seen as a paradigmatic problem of political representation itself perhaps even a problem that arises out of the very nature of representation. The unpacking of the concept may help track what’s actually going on and allow us to use its component features in ways that do not require democratic electoral systems. There is an even more drastic point to be made here. Political representation operates in many different ways and for quite different purposes, and it is these purposes or functions that give rise to and dictate a set of different normative problems that representatives will face. The political representative called on to report on his or her constituents’ views is functioning differently than a political representative who is called on to deliberate and then decide on legislation for a nation. And a political representative of political prisoners, say, a member of the International Red Cross, may function in yet a different way. To be sure, these are not all democratic or elected representatives. But any theory of representation should account for these forms as well (or be merely a partial and increasingly less useful account of political representation – applicable only in electoral contexts). In these cases, it is proper to say that representatives take on different kinds of obligation and duty based not on account of their being a representative, but rather, on the function for which their particular case of representation is used. If this is right, then, the three distinctions collapsed in the “trustee/delegate” debate arise on account of an actor’s being a decision maker regardless of whether or not he or she is a political representative per se. Consider, for example, the usefulness of these three distinctions in the analysis of any kind of decision whether decision makers are making law, executing it, or making judgements about it as member of the judiciary. To take just one example, these distinctions have been used to analyse (empirically and normatively) how members of the US Supreme Court decide cases: do they rely on their clerk’s judgement, an outside interest group’s, or their own; do they think of pursuing the interests of the collective as it is now, or some partial group (say, the founders) within; are they responsive to public opinion and thus induced to act with an eye to a good reputation? Indeed, the examples used to illustrate each of the eight cells in Table 3.2 draw on all kinds of decision maker, not simply representative lawmakers.

Beyond trustees and delegates

41

Of course, members of the judiciary and other government officials may be political representatives as well.21 If so, then these examples simply beg the question: do the three distinctions arise on account of these people being political representatives, or on account of them being decision makers (regardless of whether they are political representatives)? Let’s turn to the clearer case, where two purportedly non-representative decision makers are deciding on law for a nation. In the first case, let us consider the citizen of a direct democracy; in the second case, let us consider an unelected monarch (who has complete legislative power). Neither of these fits any conventional view of being a “political representative”. In the case of a direct democracy, citizens are not political representatives because they are not standing in for anyone but acting as themselves; in the case of the monarch, we presume he or she has not met the sufficient requirements for being a political representative under the conventional view, having not been selected by his or her people, nor later held to account. 22 Yet, in both cases, they each must decide how to vote on law. And in both cases, the three distinctions arise as they did for the conventional “political representative”. In deciding whether to approve a law, the lawmaker – the citizen or the king/ queen – will be aiming at a good either for part of the whole or directly for all.23 Each will be either relying on his or her own judgement of whether the law promotes that good or someone else’s (perhaps their fellow citizens or the monarch’s advisors). And each will be more or less responsive to sanction. A king or queen might act to avoid revolution, a citizen in a direct democracy might act to avoid ostracism or simply ill reputation, or both might simply act gyroscopically on the basis of principle or policy without regard to sanction. Indeed, we might imagine a king or queen who fits the description of a “delegate” as described in cell H of Table 3.2: extremely responsive to his or her people (he or she wants to avoid revolution), he or she curries favour with them by aiming at their individual goods as they see it (say, by being loose with the purse) and relying on extensive polling to determine legislation. Similarly, we might imagine a citizen of a direct democracy who fits the description of the trustee: acting gyroscopically without any inducement from sanctions, he or she pursues what he or she believes is the good of all as he or she judges it.24 The fact that they are decision makers leads us to investigate and think about the three core distinctions of aims, sources of judgement, and responsiveness for empirical or normative purposes. Representation may still add something else to the equation. Representation may make salient the particulars of the three distinctions in a way that is important for democratic governance.25 So, for example, being a political representative might make one’s “constituency” the salient (and normatively important) referent for aims and sources of judgement, and might give some guidance to what level of responsiveness ought to guide the representative– decision maker. This is likely how the trustee/delegate distinction then plays out in the more familiar case of political representation: the tension arises between the decision maker’s judgement and that of his or her constituency because he or she is their political representative; similarly, the tension is between the aims

42

Rehfeld

of the whole and the aims of the constituency; and the sanctions that we perhaps ought to care about are re-election by the constituency rather than some other sanctions (e.g., revolution or pecuniary gain). These are of course important features of the political landscape, but conceptually they are but a particular specification of the underlying three distinctions.

Conclusion The reason that we care about the trustee/delegate debate is that we presumably care about how closely a representative’s votes on legislation correspond to the preferences and will of his or her constituents. If we consider that the location of authority itself is a question of who has the right to determine how to vote, and the question of how to vote itself is a matter of determining the content of the vote, we can see the structural parallels between these distinctions and the broader one of democracy and justice. If the three distinctions are important as a partial specification of how justice is pursued by lawmakers, the location of decisional authority determines who by right gets to decide how the content of justice in law is determined. In terms of democracy, a scheme of representation would be more democratic based inter alia on how closely the location of authority (to determine how to vote) is held by constituents. Roughly speaking, a decision itself – the vote that the representative or other decision maker will cast in the legislature – will be just or unjust in part (or perhaps completely), depending on how it is determined. If so, then a polity may be legitimate and democratic insofar as voters retain somewhat closely their right to determine how legislative votes are cast; a decision may be reasonably just insofar as it relies on the proper manner of determining its content (i.e., a proper setting of the three distinctions). Such a view would then shift our attention to the likelihood of voters, decision maker, or some other third party making the right kinds of judgement, about the right kind of good to be sought, responsive in the right degree to the right kinds of sanction. The problems here thus form a restatement of the long understood tension between democracy, legitimacy, and justice, pointing perhaps to an epistemic theory of both. 26 What the proper view of these issues should be is now well beyond the scope of this chapter. Here I have argued that the traditional “trustee/delegate” debate has been cast in overly broad terms that obscure three underlying conceptual distinctions of aims, source of judgement, and responsiveness. The three underlying distinctions form eight ideal types, of which the traditional terms “trustee” and “delegate” are but two. Although representation may provide a specification to the three distinctions towards whose good should we aim at, according to whose judgement, and responsive to whom, the three distinctions themselves do not formally arise as part of political representation. Instead, they arise on account of a person being a decision maker; the trustee/delegate debate thus does not arise as a feature of a theory of representation per se, but on account of the role that a representative takes on as a decision maker.

Beyond trustees and delegates

43

Notes 1 This chapter originally appeared in expanded form as (Rehfeld 2009). In addition to substantive clarifications it has been revised and redacted for inclusion in this volume in three important ways. First, the discussion here is limited to the conceptual development of the trustee/delegate debate that formed the core argument of the earlier article. Second, I have excluded the discussion of the relationship between the trustee/delegate distinction and normative accounts of authority. Third, I have excluded a section discussing how contemporary literature in political science has developed these themes, which included an extended critique of (Mansbridge 2003). For more on this last point, see also Mansbridge (2011) and Rehfeld (2011). For the original article on which this chapter is based, I am grateful to the anonymous reviewers for the American Political Science Review and its editors whose probing critiques reshaped much of the argument of the original manuscript, and particularly to Kirstie McClure for her extraordinary insight and substantive feedback at every step of the process. Earlier versions of this article were presented at the 2006 Annual Meeting of the American Political Science Association in Philadelphia, Pennsylvania, and at the Political Theory Workshops at Columbia University, Stanford University, and Washington University in St. Louis. I am also grateful to Marc Bühlmann, Jan Fivaz and Daniel Schwarz for their invitation to participate in the NCCR workshop on Political Representation for which the earlier published version of this paper was presented. I thank participants in that workshop for their feedback, as well as Jane Mansbridge, Jeffrey Lenowitz, Peter Stone, and Christopher Heath Wellman, respectively, for their formal responses in those contexts. I am also grateful to feedback from audience members at those occasions. For other feedback, I thank Charles Beitz, Randy Calvert, Dario Castiglione, Joshua Cohen, Lisa Disch, Bob Erikson, Clarissa Hayward, Zach Hoskins, Jack Knight, Frank Lovett, Kirstie McClure, Ian MacMullen, Mark Piper, Rob Reich, Melissa Schwartzberg, Anna Stilz, Nadia Urbinati, Mark Warren, and particularly Ron Watson. 2 I think any legitimate law will, as a condition of its legitimacy, need to both reflect a minimum of justice and correspond to the wills of those it governs. But this view takes us into long-standing controversies of the relationship between law and norms (Dworkin 1977; Hart 1961; Locke 1988 [1690]) and, more generally, between justice and democracy. Here, we can get at the central problems by framing them in terms of conditional intention, regardless of whether justice requires that law be substantively just or reflecting the will of those governed, and whether law, to be law, must encompass justice, anyone who wants to make reasonably decent law that is approved by the people it governs will face the trade-offs that are subsumed in the “trustee/delegate” distinction. 3 Indeed, Rousseau’s “general will” is simply a combination of laws that are just and that conform to the wills of those governed by them. For a particularly good discussion of this, see Lovett (2004). 4 The problems identified in this chapter contribute to a stronger argument that the concept of political representation derives its normative content only by virtue of the function to which it is put rather than on a feature of “representation” per se. For a defence of that position, see Rehfeld (2006). 5 For an insightful discussion of the role obligations of representatives as lawmakers and the tensions between justice and democracy at the heart thereof see Beerbohm (2007). For discussions of the ethics of representation, see Applbaum (1999), Dovi (2006), Sabl (2002), and Thompson (2005). 6 I thank Josh Cohen for framing the point in this way. 7 For more on the relationship between these issues and authority see Rehfeld (2009: 225–228).

44

Rehfeld

8 I thank Philippe Schmitter for emphasizing the importance of making clear the relationship between this analysis and political parties. 9 For a popular account, we refer to the following definition of “Trustee Representation”, found on Wikipedia: These “trustees” have sufficient autonomy to deliberate and act in favour of the greater common good and national interest, even if it means going against the short-term interests of their own constituencies. This model was formulated by Edmund Burke. In the trustee model, Burke argued that his behaviour in Parliament should be informed by his knowledge and experience, allowing him to serve the public interest. Essentially, a trustee considers an issue and, after hearing all sides of the debate, exercises their own judgment in making decisions about what should be done. (http://en.wikipedia.org/ wiki/Trustee_ model_of_representation, accessed on 12 December 2008). 10 Burke’s speech was made to counter a proposal that representatives be legally bound to the desires of their constituents as expressed through legally binding instructions. 11 As noted by Mansbridge (2011: 265), this distinction was first noted in Burke by Eulau et al. (1959) with terms that are less useful today. 12 I thank Melissa Schwartzbeg and Anna Stilz for helping clarify this point. 13 This is the core of the “selection model of representation” (see Chapter 2; Mansbridge 2009). 14 Mansbridge (2003) and Pitkin (1967) are both sensitive to the continuousness of these variables. On the usefulness of treating these sorts of phenomena as “distinctions” rather than “dichotomies”, see Putnam (2002). 15 Although we must keep in mind that the partial advocacy that is the hallmark of pluralism is a proximate one: pluralism itself is justifiable only by the argument that we all do better when each pursues his or her own partial good. The dichotomy thus focuses not on a comprehensive justification for either position, but rather from the substantive view of which one is more likely to succeed in attaining the good of all (Rehfeld 2005). 16 As discussed in Rehfeld (2009), Mansbridge’s “gyroscopic representative” was originally conceived as including “self-reliant judgement”. She later acknowledged the logical possibility of a dependent-gyroscopic representative, but saw little point in a separate category as such (Mansbridge 2011). For a defence of the value of this distinction and for a minimalist formulation of the gyroscopic representative, see Rehfeld (2011). 17 The examples referred to in these eight cells are meant only to illustrate the position. Mansbridge (2011) has taken issue with the seeming pointlessness of including dependent-gyroscopic representatives, and self-reliant representatives who are responsive to sanctions. I defend the utility of these particular classifications in Rehfeld (2011). 18 Mansbridge’s characterization of the selection model is also more narrowly focused on the good of the whole, rather than any particular constituency. But I take this to be an empirical observation – that most gyroscopic representatives, in fact, aim at the good of the whole – rather than a conceptual limitation. The key of her selection model is that representatives act gyroscopically, and are less responsive to sanction. See Mansbridge (Chapter 2, and also Mansbridge 2011; see Rehfeld 2011). 19 Mansbridge (2011) has criticized the category of representatives who are unresponsive to sanctions but dependent on constituent judgement as empty sets. This corresponds to the cases illustrated in cells B and F. Tellingly, the illustrations I give are of “civil servants” and “ambassadors” and thus out of the usual legislative realm. However a “true populist” who will only vote in a manner consistent with his or her constituents’ judgement, even when his or her constituents

Beyond trustees and delegates

20

21

22

23 24

25 26

45

tell him or her to use his or her own judgement instead would be a better illustration of the category. I explore that possibility elsewhere (Rehfeld 2011: 633–635). “Pared-down” refers to the fact that I leave open whether these delegates are authorized by their constituents. In contrast, the traditional formulation collapses the location of authority into the concept of delegation. For more on the relationship between the trustee/delegate debate and democratic authority see (Rehfeld 2009: 225–228). I thank Ron Watson for this point. The position I am suggesting depends on seeing the normative and empirical qualities of representation as separate from the function to which they are put. It treats representation as considerably broader than Pitkin’s and subsequent analysis in which “political representation” is confined to a more narrow range of cases where individuals have decisional authority. In so narrowing the function of representation in this way, Pitkin rejects symbolic and descriptive forms as well, which is the only point to which other theorists in the past have really objected (Phillips 1995). By limiting the scope of representation to “actors with decisional authority”, the overlap becomes perfect between features of those with “decisional authority” and “political representation”. In my view, representatives can take on very different functions, and need not be limited in this way. This would explain how descriptive representatives are still representatives, regardless of whether they have decisional authority and thus regardless of whether three distinctions arise. Because I cannot develop and defend this view here, the reader may treat it contingently. For an initial statement of my view, see Rehfeld (2006). The view that neither the citizen nor the monarch qualifies as a political representative is consistent with theorists from Rousseau to Pitkin. Citizens are not representing anyone, they are themselves; nonelected monarchs may be symbolic representatives of the people, but in their capacity to issue orders and make law unitarily, they are not usually seen as representatives. For more on the ethics of law making that touches on similar issues, see Beerbohm (2007). Arguably, this is what Rousseau’s “citizen” would be prior to casting a vote. After the vote, when the citizen found him- or herself in the minority, he or she would then rely on the judgement of others to realize that his or her view was wrong. In this, way Rousseau’s “citizen” can be seen as toggling between cells A and B of Table 3.2 (Rousseau 1978 [1690], Book IV, Chapter 2). I thank Josh Cohen for suggesting this formulation. I do not mean to assert that following the will of the people is not somehow subsumed in a complete principle of justice (although it need not be). Instead, I am separating one principle – “do what the people want” – from other “principles of substantive justice” in order to draw out this distinction. More formally, then, the conflict emerges when one particular plausible principle of justice (“respect the people’s views”) conflicts with other principles of justice that might counter the views of the people.

References Applbaum, Arthur Isak (1999). Ethics for Adversaries. Princeton, NJ: Princeton University Press. Beerbohm, Eric (2007). Democratic Complicity. Ph.D. dissertation. Department of Politics: Princeton University. Burke, Edmund (1774). Speech to the Electors of Bristol. New York: Harpers.

46

Rehfeld

Dahl, Robert A. (1956). A Preface to Democratic Theory. Chicago, IL: University of Chicago Press. Dovi, Suzanne (2006). The Good Representative. Oxford: Blackwell Publishing. Dworkin, Ronald (1977). Taking Rights Seriously. Cambridge, MA: Harvard University Press. Eulau, Heinz, John C. Wahlke, William Buchanan and Leroy C. Ferguson (1959). ‘The Role of the Representative: Some Empirical Observations on the Theory of Edmund Burke.’ American Political Science Review 53(3): 742–756. Fasolt, Constantin (1991). “Quod Omnes Tangit Ab Omnibus Approbari Debet: The Words and the Meaning.” In: Steven Bowman and Cody Blanche (eds) In Iure Veritas: Studies in Canon Law in Memory of Schafer Williams. Cincinnati, OH: University of Cincinnati College of Law, 21–55. Hart, Herbert L.A. (1961). The Concept of Law. Oxford: Oxford University Press. Kenyon, Cecilia (ed.) (1966). The Anti-Federalists. New York: Bobbs-Merrill. Kurland, Philip B. and Ralph Lerner (eds) (1987). The Founders’ Constitution. Chicago, IL: The University of Chicago Press. Locke, John (1988 [1690]). “Second Treatise of Government.” In: Peter Laslett (ed.) Two Treatises of Government. Cambridge: Cambridge University Press, 115–240. Lovett, Francis (2004). ‘Can Justice Be Based on Consent?’ Journal of Political Philosophy 12: 79–101. Mansbridge, Jane (2003). ‘Rethinking Representation.’ American Political Science Review 97: 515–528. Mansbridge, Jane (2009). ‘A “Selection Model” of Political Representation.’ Journal of Political Philosophy 17 (4): 369–398. Mansbridge, Jane (2011). ‘Clarifying the Concept of Representation.’ American Political Science Review 105(3): 621–630. Pettit, Philip (1997). Republicanism: A Theory of Freedom and Government. Oxford: Oxford University Press. Phillips, Anne (1995). The Politics of Presence. Oxford: Clarendon Press. Pitkin, Hanna (1967). The Concept of Representation. Berkeley, CA: University of California Press. Przeworski, Adam, Susan C. Stokes and Bernard Manin (eds) (1999). Democracy, Accountability and Representation. Cambridge: Cambridge University Press. Putnam, Hillary (2002). The Collapse of the Fact/Value Dichotomy. Cambridge, MA: Harvard University Press. Rakove, Jack (1996). Original Meanings: Politics and Ideas in the Making of the Constitution. New York: Alfred A. Knopf. Rehfeld, Andrew (2005). The Concept of Constituency: Political Representation, Democratic Legitimacy, and Institutional Design. Cambridge: Cambridge University Press. Rehfeld, Andrew (2006). ‘Towards a General Theory of Representation.’ Journal of Politics 68(1): 1–21. Rehfeld, Andrew (2009). ‘Representation Rethought: On Trustees, Delegates and Gyroscopes in the Study of Political Representation and Democracy.’ The American Political Science Review 103(2): 214–230. Rehfeld, Andrew (2011). ‘The Concepts of Representation.’ American Political Science Review 105(3): 631–641. Rousseau, Jean-Jacques (1978 [1762]). On the Social Contract. New York: St. Mary’s.

Beyond trustees and delegates

47

Sabl, Andrew (2002). Ruling Passions: Political Offices and Democratic Ethics. Princeton, NJ: Princeton University Press. Saward, Michael (2010). The Representative Claim. Oxford: Oxford University Press. Sunstein, Cass R. (1993). The Partial Constitution. Cambridge, MA: Harvard University Press. Thompson, Dennis (2005). Restoring Responsibility: Ethics in Government, Business, and Health Care. New York: Cambridge University Press. Truman, David (1953). The Governmental Process: Political Interests and Public Opinion. New York: Alfred A. Knopf. Vermeule, Adrian (2007). Mechanisms of Democracy: Institutional Design Writ Small. New York: Oxford University Press.

4

Institutional constraints and territorial representation Audrey André, Sam Depauw and Kris Deschouwer

Introduction Popular sovereignty is commonly understood in terms of people’s policy preferences somehow informing their representatives’ actions once elected (see Pitkin 1967). Elections in particular serve to translate people’s interests into public policies: on the one hand, constituents select the representatives who in office will pursue the policies they like; on the other hand, the risk of being sanctioned and voted out next time will prevent representatives from enacting policies they would otherwise favour if these promise to be unpopular with constituents (see Mansbridge 2009). Context, Mansbridge continued, further shapes relations between constituents and (would-be) representatives. How constituencies are defined in particular, we argue, constrains whom to represent. That is, in whose interest an elected representative will act is critically dependent on the institutions that govern his or her (re)election. Across the globe legislators are elected by districts: who’s included and who’s excluded is denoted by the districts’ boundaries on the map. That constituencies are defined in territorial terms creates the incentive for re-election seeking representatives to engage in territorial representation (Rehfeld 2005: 145).1 Whether would-be representatives who favour territorial representation are more likely to run if constituencies are defined in territorial terms, or whether representatives engage in territorial representation because they fear defeat at the polls if they do not, how constituencies are defined constrains legislators’ focus of representation. More specifically, we argue, district magnitude shapes a legislator’s views as to whom to represent and buttresses their efforts to represent not just the district but also the area centring on their hometown. Moreover, territorial notions of representation have not remained unaffected by the emergence of regional levels of elected government in many modern democracies. As a result decisionmaking is increasingly dislocated beyond the nation state in the multi-level environment (Hough and Jeffery 2006). The PartiRep MP survey offers a unique opportunity in this respect, because for the first time the cross-national comparison includes both national and regional legislators. A representative’s dilemma as to whether to look after the interests of the district or after the national interest has a long pedigree in understanding role

Constraints and territorial representation

49

orientations, dating back at least to Edmund Burke (Eulau et al. 1959). However, research into representational roles has thus far been unable to conclusively identify the sources of representatives’ role orientations – or to tie them to behaviour (see Jewell 1970; Searing 1994; Thomassen 1994). One reason is a dearth of comparative studies; the other evokes how waning party identifications and rising electoral volatility have brought to light the misspecification of role orientations (Wessels 1999b: 211–212). How constituencies are defined, Wessels argued, not only shapes a legislator’s uncertainty about district opinions, but also his or her electoral incentive to engage in territorial representation. The PartiRep survey uniquely allows us to separate the two mechanisms and offers a more fine-grained measurement of legislators’ focus of representation. The relevance of our findings is emphasized by electoral engineers’ propensity to set out to change representatives’ role orientations, working on the assumption that they are determined by the electoral institutions in particular (Norris 2004). The argument will proceed in six steps. Section one presents the theoretical foundations for the assumption that how constituencies are defined will shape patterns of political representation. Section two presents more precise hypotheses as to the effect of the twin processes of personalization of electoral competition and uncertainty as well as the effect of regionalization. Section three introduces the PartiRep data and the case selection of 13 advanced industrial democracies. Section four demonstrates how the conditions that representatives compete under for (re)election shape their commitment to territorial representation. Section five then aims to disentangle the effect on representatives’ focus of representation of the electoral incentive to cultivate a personal reputation from their uncertainty about whom to represent. Section six discusses the effect of regionalization and whether regional and national legislators differ in their commitment to territorial representation and their focus of representation, followed by a summary of our findings in section seven.

How constituencies are defined The definition of constituency constrains the kind of issues and interests that will gain representation (Rehfeld 2005). It is the assumption that how constituencies are defined will condition representatives’ commitment to territorial representation and their focus of representation. That is, electoral and governmental institutions shape the incentive that a representative has to look after the economic and social needs of the local area; they further constrain how they choose to define that constituency. The notion of constituency is commonly used to denote one of three groups (Rehfeld 2005: 35): (1) those who are eligible to vote for a representative, (2) those who voted for the representative, or (3) those in whose interest a representative acts.2 Electoral and governmental institutions matter because representatives will choose to act in the interest of those whose support contributes most to their re-election (Mayhew 1974).

50

André, Depauw and Deschouwer

Constituencies have to be “imagined”, Saward (2010: 50) pointed out. What is and what is not in their interest is “constructed” by the joint, and often conflicting, efforts of constituents and would-be representatives. The claims that the latter make to speak on constituents’ behalf present constituents with verbal and mental pictures of who they are and where their interest lies. Representatives have considerable discretion in deciding whom to represent: a constituency’s interest cannot be readily known (Manin 1997). Looking at his district, a US congressman for instance tends to see concentric circles of support (Fenno 1978): there are those who would always vote for him first in the primary and later in the general election; others would never vote for him. The erosion of traditional cleavages and the trend towards individualization among the electorate has only underlined representatives’ uncertainty about telling the difference (Andeweg 2003). While different choices are open to representatives as to whom to represent, there are “cultural constraints” to the claims that representatives can successfully make and these include electoral and governmental institutions (Saward 2010). The information cues that voters respond to are shaped by institutions (Shugart et al. 2005). One recent study has demonstrated that personal factors have but a limited impact on the role orientations of EU and national legislators: neither legislative seniority, education, nor party socialization can explain whether representatives favour representing (1) the district, (2) the party, or (3) the nation (Wessels 1999b: 221). By contrast, there were important differences between the levels of government, between EU and national legislators in terms of focus of representation. Moreover, both the level of competitiveness and uncertainty about the district’s interests were found to encourage EU and national legislators to shift their focus of representation from the district to either the party, or the population at large. On the one hand, electoral institutions render competition more candidate-centred or more party-centred. The conditions of competitiveness are shaped by the district’s magnitude. On the other hand, uncertainty is tied to the size of the district’s population. Large districts are not “communities of interest” sharing a common interest, Rehfeld (2005) agreed. In these circumstances representatives are less likely to define their “re-election constituency” in terms of the district, Wessels (1999a) hypothesized. But district magnitude and population size are highly correlated, however. The manner in which Wessels (1999b) has conceptualized electoral institutions thus resulted in easily ascribing the effect of one to the other. Another conceptualization will enable us to disentangle both. How constituencies are defined shapes the electoral incentive that representatives have to look after the economic and social needs of the local area, and thereby their focus of representation. This electoral incentive first stems from the personalization of electoral competition under various electoral rules. That is, electoral institutions determine the relative value to a representative’s re-election of his personal versus the party reputation and thereby shape the incentive they have to nurture a personal vote (Carey and Shugart 1995). The closer the electoral fortunes of co-partisans are tied together, the

Constraints and territorial representation

51

more representatives’ focus of representation will concentrate on mobilizing party voters. Alternatively, the more personalized electoral competition is, the more representatives will come to define their focus of representation in terms of the district (Wessels 1999b: 222). Representatives in five Westminster parliaments define their task in Parliament first and foremost in terms of representing the territorial district where there is intra-party competition (Heitshusen et al. 2005). In those circumstances candidates face a product differentiation problem (Cox and Thies 1998: 271), which magnifies with the number of co-partisan competitors they face. The party reputation by definition cannot help voters to discriminate between candidates of the same party. Candidates will seek to pre-empt defeat by nurturing a personal reputation that they think voters will respond to (see the notion of “anticipatory representation”: Mansbridge 2003: 516). Candidates who fail to stand out among competitors in this manner will be replaced over time (Norris 2004: 9).

Hypotheses There are two conceptualizations of how the personalization of electoral competition might be measured by the district’s magnitude. First, Wessels (1999b: 222) hypothesized that the smaller the district the more “personalized” the electoral competition will be. The smaller the number of candidates that run in the district, the larger is the number of voters that will be monitoring the positions and actions of individual candidates. Therefore as district magnitude grows, representatives will define their focus of representation less in terms of the district – irrespective of the form of the ballot. Their focus will be defined in partisan terms or by generalizing it to all people in the country (see Alpert 1979). Second, district magnitude has a differential effect on the incentive to cultivate a personal reputation according to Carey and Shugart (1995): (1) the incentive increases with district magnitude in open-list systems and decreases in closed-list systems as magnitude grows; and (2) at any given magnitude it is higher in open-list systems than in closed-list systems. 3 That is, the effect of district magnitude is conditional upon voters’ ability to alter the party’s preferred rank-order of candidates. When lists are open, the incentive to cater to narrow constituencies increases as the ratio of candidates running under the party label to the seats available grows. Therefore, Carey and Shugart (1995) propose to use district magnitude as a proxy for the scope of intra-party competition (but see Crisp et al. 2007; André and Depauw 2014). When lists are closed, shirking in terms of territorial representation is expected to increase with the number of co-partisans (Bowler and Farrell 1993). Or a representative’s ability to improve his chance of re-election by “putting a face on the party” and attracting additional votes to the party will decrease with district magnitude (Shugart 2008: 46).

52

André, Depauw and Deschouwer

Hypothesis 1: the higher the value to re-election of the personal reputation, the more committed legislators are to territorial representation. Furthermore, they are more likely to define the constituency they represent in terms of the district (and in particular some smaller part of it) and less likely to define it in partisan terms. 1a: the value of the personal reputation decreases with district magnitude. 1b: the value of the personal reputation increases with district magnitude in open-list systems and decreases in closed-list systems. Second, the process of representation, Fenno (1978: 10) observed, is tinged with uncertainty: representatives cannot be sure what constituents’ preferences are. Their uncertainty increases in the face of district heterogeneity, moreover. A legislator’s ability to know, and be instructed by, his constituents’ preferences, Converse and Pierce (1986) pointed out, critically depends on his district’s homogeneity. Previous research on the trustee-delegate styles of representation has emphasized how a legislator simply cannot be responsive to all constituents, if on most issues they are of a different opinion (Alpert 1979; Jewell 1970). District heterogeneity in its turn is first and foremost a product of the district’s population size: with each additional constituent grows the variety of interests to be represented and social complexity in the district (Wessels 1999a). In these circumstances representatives are less likely to make the district their focus of representation, Wessels (1999b) hypothesized. They opt instead to define their “re-election constituency” in partisan terms, or to generalize the representational focus to include all people in the country in order to reduce uncertainty. Hypothesis 2: the larger the district’s population size, the less likely it is that legislators define the constituency they represent in terms of the district. Third, patterns of territorial representation have been redesigned by the creation of regional levels of elected government in a number of modern democracies. On the one hand, regional legislators more readily define their focus of representation in territorial terms than do national legislators. Some interests are by their nature territorial: the location of a major road or power plant, for instance. The jurisdictions of regional parliaments focus more heavily on these kinds of issue (Lancaster 2002; Patzelt 2007). The early years of devolution in the United Kingdom, for instance, suggest that regional legislators have a greater commitment to constituency representation (Russell and Bradbury 2007; Bradbury and Mitchell 2007). Because of their jurisdictions regional legislators also have to deal more commonly with local interest groups and local government authorities (Patzelt 2007). That regional legislators more readily engage in territorial representation has also been explained by the fact that they are elected in smaller districts. If so, the effect is dependent on the previous arguments regarding the form of the ballot.

Constraints and territorial representation

53

Hypothesis 3: Regional legislators define representation more in territorial terms than do national legislators. Regionalization has been associated with the emergence of ethno-territorial parties, moreover. They tend to prioritize issues that are not related to the left–right dimension of party competition (Meguid 2008). Even if over time their platforms grow in size, they are associated with a single issue: that of representing the interest of a particular region. In this respect ethno-territorial parties are expected to prioritize territorial representation. At the same time their focus of representation need not concentrate on the local area: they will rather present their region as constituting a genuine “community of interest”, sharing a common interest that is not hampered by subregional differences. Hypothesis 4: Representatives of ethno-territorial parties are more committed to territorial representation than those of other parties of similar size, but their focus of representation is the region at large. The empirical analysis of these hypotheses will proceed in three steps. The first two parts address the way electoral competition and representational uncertainty impact legislators’ commitment to territorial representation on the one hand and their focus of representation on the other. The third section then focuses on the effects of regionalization and puts hypotheses 3 and 4 to the test. But before proceeding to the analysis, we first introduce the data and discuss the operationalization of the different dependent variables.

Representational role orientations in 13 European democracies To explore the consequences of institutional contexts for the representational role orientations of elected representatives, comparative elite data will be used that was collected as part of the PartiRep project.4 The project surveyed national legislators in 13 European democracies. The selection displays a wide diversity of electoral rules. Combined are systems where preferential voting significantly affects the allocation of seats to candidates (Belgium, Ireland, Poland, and Switzerland) and systems where preferential voting does not (Israel, Italy, the Netherlands, Norway, Portugal, the United Kingdom).5 Mixed-member systems by definition combine a nominal and proportional tier (Germany and Hungary); Austria adopted a system with complex districting, using varying electoral formulae and ballot types in the different tiers.6 In addition, the sample of countries varies in respect of the unitary or multi-level character of their state structure. Taking advantage of this variation, the PartiRep project also surveyed all regional legislators in Austria, Belgium, Portugal (in Madeira and the Azores), and Switzerland (excl. Appenzell-Innerrhoden) and a subset of regional legislators in Germany (in Brandenburg, Lower Saxony, Rhineland-Palatinate, and Thuringia), Italy (in Calabria, Campania, Lazio, Lombardia, Toscana, and Valle d’Aosta), and in the United Kingdom (in Scotland and Wales).

54

André, Depauw and Deschouwer

The selected parliaments further display a wide variety in district magnitude – ranging from single-member districts to at-large constituencies – and in the way constituencies are defined. General election constituencies map onto administrative boundaries in Belgium (provinces), Israel (the country), Italy (regions),7 the Netherlands (the country), Norway (counties), Portugal (districts), Switzerland (cantons and a handful of half-cantons), and in the upper levels of multi-tiered systems (the states in Austria and Germany; the counties and the country in Hungary). In Ireland and Poland district boundaries are drawn specifically for electoral purposes. Districts are assigned seats pro rata their population figure and the distribution is automatically adjusted following the census. Single-member districts are used in the lower tiers of Germany’s and Hungary’s mixedmember systems and in the United Kingdom. In order to retain equal ratios of population to representatives across the country frequent redistricting is required in those circumstances, which may easily disrupt local community identities (Steed 1985: 282). In Appendix A the number of responses in each parliament can be found. A correcting weight is used to bring the number of responses in the different countries and levels of government more in line with one another. In addition, responses are weighted by the size of the parliamentary party in each parliament. The PartiRep cross-national survey provides two distinct measures of the representational role orientations legislators ascribe themselves: their commitment to territorial representation and their representational focus. To measure their commitment to territorial representation, representatives were asked to rank order different duties and responsibilities that members of parliament engage in. More precisely, they weighed up the importance of “looking after the collective social and economic needs of the local area”, “influencing government policy”, “assisting individual voters in their dealings with public authorities”, and “liaising between PPG members and the party leadership and managing Parliament’s business”. Legislators who consider the advocacy of the local area’s interests as the single most important task they fulfil as elected representatives express a strong commitment to territorial representation and are assigned a value of “1”; those prioritizing other aspects of the job are coded as “0”. Defending local interests lies at the core of the role orientation of on average one in four legislators, but there are important cross-national differences in the selection (see Table 4.1). While their number exceeds 40% in Italy, local promoters are almost absent among Norwegian, Dutch, and Israeli legislators. To gain further insight into how they define their re-election constituency, representatives were also asked to indicate on a rising seven-point scale how important they consider it to represent the views and interests of (1) their municipality, (2) all the people in their district, (3) all the people who voted for their party, and (4) all people in the devolved region/country. As the distribution of these variables is strongly skewed to the left, they are dichotomized: only legislators attaching great importance to a particular focus of representation (i.e. a score of “6” or “7”) are coded as “1”. Table 4.1 displays

Constraints and territorial representation

55

Table 4.1 Representational role orientations in 13 European democracies (in percentages) Territorial representation

Austria Belgium Germany Hungary Ireland Israel Italy Netherlands Norway Poland Portugal Switzerland UK Mean

Focus of representation

commitment

district

municipality

party voters

all people

33.4 13.8 37.7 31.5 27.6 7.6 40.8 2.3 0.0 24.4 28.9 21.7 31.3 25.3

78.5 39.0 70.6 56.9 61.8 87.4 50.8 47.4 57.3 52.0 71.3 44.0 71.3 60.7

69.6 58.0 66.9 73.4 75.6 28.9 50.9 35.9 66.8 70.9 65.1 41.1 68.4 59.7

78.5 50.7 65.9 69.0 46.3 74.3 45.1 73.5 84.8 53.3 65.1 66.5 58.0 63.1

64.8 37.9 47.6 47.3 47.9 87.4 52.7 47.4 69.2 64.6 77.1 45.5 42.9 54.9

Note: Entries are percentages of representatives adopting a particular role orientation.

the percentages of representatives adopting a particular representational focus and reveals substantial differences across countries. The number of representatives who find it of the greatest importance to represent the views and interests of all people in the district, for instance, ranges from 39% to 87%. A similar cross-national gap of more than 40 percentage points can be observed for the other focuses as well. Whereas Norwegian and Dutch legislators have a strongly party-centred role orientation, legislators elected in single-member districts and open-list systems define representation more in territorial terms. Approximately 70% of the German and British elected representatives have a district-orientated interpretation of their representational function. In Hungary, Ireland, and Poland, a similar proportion of legislators emphasizes the importance of defending the interests of their hometown.

Institutional effects on territorial representation Institutions, it was argued, shape legislators’ commitment to territorial representation. In particular, the electoral incentives arising from the interaction between the district’s magnitude and the ballot structure are expected to affect their prioritization of promoting the local interest (H1). Alternatively, legislators may respond to the uncertainty about constituents’ interests attached to

56

André, Depauw and Deschouwer

the district’s heterogeneity (H2). Given the strong correlation between the number of seats a district is entitled to and the number of people living there, two separate models are estimated to test these hypotheses. The first model included the decimal logarithm of district magnitude, a binary indicator for open-list systems, and the interaction between both. In the second model, the decimal logarithm of the district’s population size measured in 1,000 inhabitants is substituted for district magnitude. To further explore the regionalization hypotheses, both models also include dichotomous variables indicating members of regional parliaments (H3) and representatives of ethno-territorial parties (H4). To isolate the effects of institutional incentives and constraints, the analysis further contains country dummies and control variables for the impact of seniority, the party’s ideological position on a left-right scale, and the size of the parliamentary party group. As none of these controls is consistently found to have an effect, discussion of the findings will focus by and large on the main hypotheses. Recall that the dependent variable can only take two values: representatives who identify “looking after the collective social and economic needs of the local area” the principal aspect of their representational task are coded “1” and those who identify other priorities are coded “0”. How they define the “local area” they claim to look after is left open to their own interpretation – be it a narrow bailiwick around their hometown or a wider territory. Given the dichotomous nature of the dependent variable, probabilistic regression is the proper estimation method. As in non-linear models the effect of any given independent variable depends on the values of all other explanatory factors, we have simulated the models using clarify in Stata (King et al. 2000). That is, legislators’ predicted probability of prioritizing territorial representation in preferential and nonpreferential systems was computed across a range of values of district magnitude and population size while keeping all other variables constant.8 To test whether the slope of the modifying variables is statistically significant in both list types, we performed a significance test of the first difference in predicted probabilities between the minimum and maximum simulated values. To ease interpretation, we present the results graphically. Bold lines indicate that the null hypothesis that district magnitude or population size has no effect can be rejected. Because we expect party differences in the role orientations of elected representatives, inflated significance levels are avoided by using clustered sandwich estimation of the standard errors (Steenbergen and Jones 2002). The first model in Table 4.2 disproves Wessels’ (1999b) hypothesis that the effect district magnitude has on legislators’ commitment to territorial representation is invariably negative in both non-preferential and preferential electoral systems. Rather, strong support is found for Carey and Shugart’s (1995) claim that the value of a favourable personal reputation among constituents is conditional upon the ballot structure that legislators compete under for re-election. When voters cannot alter the party’s predetermined ranking of candidates, elected representatives become increasingly unlikely to consider the advocacy of local interests their principal task as districts grow in magnitude. To estimate the

Constraints and territorial representation

57

Table 4.2 Probit regression of institutional effects on territorial representation Model 1 b. Open-list District magnitude (log) DM*Open-list

Model 2 s.e.

.087

s.e.

(.266)

Open-list

-.445

(.085)***

-.298

.693

(.149)***

Population (log) Population* Open-list

.171

(.116)

Regional Parliament Regionalist party Left–right PPG size (log) Seniority

.302

(.138)**

-.470

(.246)*

-.025 .163

(.023) (.121)

-.019

(.007)***

-.880 1742

(.404)**

Regional Parliament Regionalist party

.399

(.125)***

-.475

(.242)**

Left–right PPG size (log)

-.026 .180

(.024) (.120)**

Seniority

-.020

(.008)***

Constant N

-1.106 1742

(.374)***

Log pseudolikelihood LR(df) Nagelkerke r 2

b.

-837.990 256.240 20.400

(19)***

Constant N

.327

Log pseudo- -843.190 likelihood LR(df) 245.850 Nagelkerke r 2 19.600

(.391) (.063)***

(19)***

Note: The table displays the parameter estimates and standard errors (in parentheses) of a probit regression. The dependent variable is a legislator’s commitment to territorial representation.

slope of the effect of district magnitude in systems with effective preferential voting, the coefficient of the main effect and the interaction term need to be summed. The effect is clearly positive: an increase in the scope of intra-party competition encourages legislators to present themselves as assiduous local promoters. The simulated predicted probabilities and significance test of both slopes further increase confidence in our findings (see Table 4.3 and Figure 4.1). Changing district magnitude from a single-member district to an at-large district significantly increases a legislator’s predicted probability of identifying the advocacy of local interests as his single most important task by 15% in preferential systems, but decreases that probability by 12% in the absence of intraparty competition. As displayed in Figure 4.1 we can clearly distinguish between closed- and open-list systems at magnitudes greater than about eight – that is, the 90% confidence intervals for both list types diverge when the number of seats to be allocated in the district exceeds eight. Still, representatives of small and large parties differ in their reaction to electoral incentives. Even in open-list systems with high levels of intra-party competition, small parliamentary party

58

André, Depauw and Deschouwer 10

10

10

10

10

magnitude,

magnitude,

10

10

10

10

10

10

10 10

10

District magnitude, logscale

100

10

100

1000

20000

District's population, logscale

Figure 4.1 Plotted predicted probabilities of institutional effects on territorial representation

groups lack the sheer capacity to implement a “division of labour” in that some members are chiefly concerned with policy making and parliamentary responsibilities allowing others to prioritize territorial representation. A comparison of models 1 and 2 (Table 4.2) further underscores that the positive effect of district magnitude in preferential systems can be ascribed to the scope of intra-party competition rather than to larger district populations. In non-preferential systems, district magnitude and population size are strongly correlated at .9 and, as a result, both negative effects cannot be separated from one another. In effective preferential systems, on the other hand, the correlation between the district’s seat total and population figure is less strong at .4. Yet, substituting district magnitude by the district’s population size fails to produce similar results. The effect of population size in preferential systems is negative: as the district’s population increases elected representatives do not prioritize looking after the local area, they find territorial representation less important. The effect does not reach conventional levels of statistical significance however. While we cannot tell both list types apart in the smallest districts, for districts with more than 120,000 inhabitants a legislator’s predicted probability of considering the advocacy of local interests his principal task is systematically higher in open than in closed-lists systems. This provides further evidence of the fact that representatives respond strategically to the incentives fuelled by their re-election context: in a populous district that produces a high

Constraints and territorial representation

59

level uncertainty as to the interests to be represented, a legislator’s commitment to territorial representation is markedly stronger in systems requiring him to seek out preferential votes than in those that do not. In sum, intra-party competition rather than population size is the key to explaining representatives’ prioritization of looking after the interests of the local area. Their perception of the “local area” they claim to represent will be examined in the next section.

Institutional effects on the focus of representation How districts are defined affects in whose interest representatives claim to act. That is, the legal definition of districts has an impact on how important representatives think it is to represent the views and interests of the municipality, the district, the party, or all people in the country. As noted, representatives are said to adopt a given focus when they assign it a value of at least six on a rising seven-point scale. Analogous to the previous analysis, probabilistic regression models are estimated for each of these four focuses of representation. The models’ simulated predicted probabilities are reported in Table 4.3 and plotted in Figure 4.2. The exact coefficients, standard errors, and measures of model fit are reported in Appendix A. The districts’ boundaries on the map strongly affect how important representatives find it to represent the views and interests of the district. In this case, district magnitude and population size do produce the same results: legislators’ emphasis on representing the district diminishes with both the number of seats to be allocated and the number of people living there. The same negative pattern is noted in open- and closed-list systems (see Table 4.3 and Figure 4.2). A rise in magnitude from 1 to 100 decreases the predicted probability of assigning the greatest importance to representing the district by approximately 30%, irrespective of the ballot form. A rise in population size Table 4.3 The percentage changes of institutional effects on the focus of representation District Magnitude

Population Size

Interaction models

Restricted models

Interaction models

Restricted models

closed-list

open-list

open-list

closed-list

open-list

open-list

-30.4 +2.1 -3.0 -10.5

-4.9 +46.6 -34.4

-43.6 -13.9 +18.5 -2.0

-43.0 -60.0 +44.1 +30.2

district -27.5 municipality -9.4 party voters +10.7 all people -6.8

-42.9

-47.3 -14.4 +2.3 -9.0

Note: Entries are the percentage change in the predicted probability of adopting a particular focus of representation when district magnitude increases from 1 to 100. The district’s population size changes from 10,000 to 20,000,000 inhabitants for the interaction model (full sample) and from 10,000 to 4,000,000 for the restricted model (open-list systems only). Shaded cells indicate that the difference is statistically significant at the 90% level.

André, Depauw and Deschouwer

'"

'"

'"

'"

Figure

Figure

60

'" '" '"

'" '"

10 District magnitude, logscale

100

10

100 1000 District's population, logscale

20000

Figure

Interaction models - open-list systems

~

Q.

'"

'"

'"'"

'"

" '"'" ~

'"'"

Q.

'"

'" '"

10 District magnitude, logscale

10

100

100 1000 District's population, logscale

20000

Restricted models - open-list systems

'" '" '" '" '" '"

'" '"c::q i '". '" '"C"! '"

-