The Oxford Handbook of Legislative Studies 9780199653010, 0199653011

Table of contents :
Cover
THE OXFORD HANDBOOK OF LEGISLATIVE STUDIES
Copyright
Contents
About the Contributors
1 Introduction
Part I Theoretical Approaches in Legislative Studies
2 Formal Models of Legislatures
3 The Sociology of Legislators and Legislatures
4 Typologies and Classifications
Part II Methods of Legislative Research
5 Roll-Call Analysis and the Study of Legislatures
6 Words as Data: Content Analysis in Legislative Studies
7 Debate and Deliberation in Legislatures
8 Interviews and Surveys in Legislative Research
9 The Experimental Study of Legislative Behaviour
Part III Representation and Legislative Careers
10 Candidate Selection: Implications and Challenges for Legislative Behaviour
11 The Effect of Electoral Institutions on Legislative Behaviour
12 Gender and Legislatures
13 Roles in Legislatures
14 Legislative Careers
Part IV Organization and Rules
15 Procedure and Rules in Legislatures
16 The Politics of Bicameralism
17 Committees
Part V Parties in the Legislature
18 Political Parties and Legislators
19 Party Discipline
20 Legislative Party Switching
21 Legislative Institutions and Coalition Government
Part VI Policy-Making and Oversight
22 Institutional Foundations of Legislative Agenda-Setting
23 Lawmaking
24 Legislatures and Public Finance
25 Legislatures, Lobbying, and Interest Groups
26 Legislatures and Foreign Policy
27 Common Agency? Legislatures and Bureaucracies
Part VII Expanding the Scope of Legislative Analysis
28 Political Behaviour in the European Parliament
29 Sub-National Legislatures
30 The Study of Legislatures in Latin America
31 Legislatures in Central and Eastern Europe
32 Authoritarian Legislatures
33 Reluctant Democrats and Their Legislatures
Name Index
Subject Index

Citation preview

T H E OX F OR D HA N DB O OK OF

L E G I SL AT I V E ST U DI E S

THE OXFORD HANDBOOK OF

LEGISLATIVE STUDIES Edited by

SHANE MARTIN THOMAS SAALFELD and

KAARE W. STRØM

1

3 Great Clarendon Street, Oxford, United Kingdom

OX2  6DP,

Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark  of Oxford University Press in the UK and in certain other countries © The several contributors  2014 The moral rights of the authors‌have been asserted First Edition published in 2014 Impression:  1 All rights reserved. No part of this publication may be reproduced, stored  in a retrieval system, or transmitted, in any form or by any means, without  the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope  of  the above should be sent to the Rights Department, Oxford University Press,  at  the address  above You must not circulate this work in any other  form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University  Press 198 Madison Avenue, New  York, NY 10016, United States of America British Library Cataloguing in Publication  Data Data available Library of Congress Control Number: 2014935039 ISBN 978–0–19–965301–0 Printed and bound  by CPI Group (UK) Ltd, Croydon, CR0  4YY Links to third party websites are provided by Oxford in good faith  and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this  work.

Acknowledgments

Our sincere thanks to the authors for agreeing to contribute and for their cooperation throughout. This Handbook is their work, and without their timely and professional collaboration, it would have been an impossible task. Thanks also to Dominic Byatt, Commissioning Editor at Oxford University Press, for so wholeheartedly supporting the project. Our respective institutions, Dublin City University, the University of Bamberg, and the University of California, San Diego, hosted a number of editorial meetings, and we are grateful for the many ways in which they have supported our efforts. In addition to our home institutions, we thank the London School of Economics for hosting several editorial meetings at critical stages of the project. Martin gratefully acknowledges support for work on the Handbook from the Norwegian Research Council (FriSam Project No. 222442) and from the School of Law and Government and the Vice-President for Research, Dublin City University. We thank Matthias Bahr, Eric Bientzle, Wolfgang Goldbach, Margret Hornsteiner, Erik Jäger, David Ruß, and Akisato Suzuki for help in preparing the manuscript.

Contents

About the Contributors

1. Introduction Shane Martin, Thomas Saalfeld, and Kaare W. Strøm

xi

1

PA RT I   T H E OR E T IC A L A P P ROAC H E S I N L E G I SL AT I V E S T U DI E S 2. Formal Models of Legislatures Daniel Diermeier

29

3. The Sociology of Legislators and Legislatures Heinrich Best and Lars Vogel

57

4. Typologies and Classifications Amie Kreppel

82

PA RT I I   M E T HOD S OF L E G I SL AT I V E R E SE A RC H 5. Roll-Call Analysis and the Study of Legislatures Royce Carroll and Keith Poole

103

6. Words as Data: Content Analysis in Legislative Studies Jonathan B. Slapin and Sven-Oliver Proksch

126

7. Debate and Deliberation in Legislatures André Bächtiger

145

8. Interviews and Surveys in Legislative Research Stefanie Bailer

167

9. The Experimental Study of Legislative Behaviour James N. Druckman, Thomas J. Leeper, and Kevin J. Mullinix

194

viii  Contents

PA RT I I I   R E P R E SE N TAT ION A N D L E G I SL AT I V E C A R E E R S 10. Candidate Selection: Implications and Challenges for Legislative Behaviour Reuven Y. Hazan

213

11. The Effect of Electoral Institutions on Legislative Behaviour Audrey André, Sam Depauw, and Matthew S. Shugart

231

12. Gender and Legislatures Michelle M. Taylor-Robinson

250

13. Roles in Legislatures Rudy B. Andeweg

267

14. Legislative Careers Scott A. MacKenzie and Thad Kousser

286

PA RT I V   ORG A N I Z AT ION A N D RU L E S 15. Procedure and Rules in Legislatures Wolfgang C. Müller and Ulrich Sieberer

311

16. The Politics of Bicameralism William B. Heller and Diana M. Branduse

332

17. Committees Shane Martin

352

PA RT V   PA RT I E S I N T H E L E G I SL AT U R E 18. Political Parties and Legislators Thomas Saalfeld and Kaare W. Strøm

371

19. Party Discipline Christopher Kam

399

20. Legislative Party Switching Carol Mershon

418

21. Legislative Institutions and Coalition Government Lanny W. Martin and Georg Vanberg

436

Contents   ix

PA RT V I   P OL IC Y- M A K I N G A N D OV E R SIG H T 22. Institutional Foundations of Legislative Agenda-Setting Bjørn Erik Rasch

455

23. Lawmaking Sebastian M. Saiegh

481

24. Legislatures and Public Finance Joachim Wehner

514

25. Legislatures, Lobbying, and Interest Groups Anne Skorkjær Binderkrantz

526

26. Legislatures and Foreign Policy Tapio Raunio

543

27. Common Agency? Legislatures and Bureaucracies Mathew D. McCubbins

567

PA RT V I I   E X PA N DI N G T H E S C OP E OF L E G I SL AT I V E A NA LYSI S 28. Political Behaviour in the European Parliament Simon Hix and Bjørn Høyland

591

29. Sub-National Legislatures William Downs

609

30. The Study of Legislatures in Latin America Brian F. Crisp and Constanza F. Schibber

628

31. Legislatures in Central and Eastern Europe Josephine T. Andrews

647

32. Authoritarian Legislatures Paul Schuler and Edmund J. Malesky

676

33. Reluctant Democrats and Their Legislatures Gary W. Cox

696

Name Index Subject Index

717 737

About the Contributors

Rudy B. Andeweg  is Professor of Political Science at Leiden University. Audrey André  is Post-Doctoral researcher at the Vrije Universiteit, Brussels. Josephine T. Andrews is Associate Professor of Political Science, University of California, Davis. André Bächtiger  is Swiss National Science Research Professor at the University of Lucerne. Stefanie Bailer  is Assistant Professor for Global Governance at ETH Zurich. Heinrich Best  is Professor of Sociology and Chair of Social Science Research Methods/ Structural Analysis of Modern Societies at the Friedrich Schiller University of Jena. Anne Skorkjær Binderkrantz  is Professor of Political Science at Aarhus University. Diana M. Branduse  is Doctoral Candidate in the Department of Political Science, Binghamton University. Royce Carroll  is Associate Professor of Political Science at Rice University. Gary W.  Cox  is William Bennett Munro Professor of Political Science at Stanford University. Brian F. Crisp  is Professor of Political Science at Washington University in St Louis. Sam Depauw  is Assistant Professor and Post-Doctoral Researcher at the Department of Political Science, Vrije Universiteit Brussel. Daniel Diermeier  is IBM Professor of Regulation and Competitive Practice, Kellogg School of Management, Northwestern University. William  Downs  is Professor of Political Science and Associate Dean for Social and Behavioral Sciences at Georgia State University. James N. Druckman  is Payson S. Wild Professor of Political Science and Faculty Fellow at the Institute of Policy Research, Northwestern University. Reuven Y. Hazan  is Professor and Chair of the Department of Political Science, Hebrew University of Jerusalem.

xii   About the Contributors William B. Heller is Associate Professor in the Department of Political Science, Binghamton University. Simon Hix  is Professor of European and Comparative Politics at the London School of Economics and Political Science. Bjørn Høyland  is Professor of Political Science at the University of Oslo. Christopher Kam  is Associate Professor in the Department of Political Science at the University of British Columbia. Thad Kousser  is Professor of Political Science, University of California, San Diego. Amie Kreppel  is Jean Monnet Chair (ad personam) and Associate Professor in the Department of Political Science at the University of Florida. Thomas J. Leeper  is Post-Doctoral Researcher in the Department of Political Science and Government, Aarhus University. Scott A. MacKenzie is Assistant Professor of Political Science, University of California, Davis. Edmund J. Malesky  is Associate Professor of Political Science, Duke University. Lanny W. Martin  is Associate Professor of Political Science, Rice University. Shane Martin  is Reader in Comparative Politics, University of Leicester. Mathew D. McCubbins  is Professor of Political Science and Law, at Duke University. Carol Mershon  is Associate Professor, Department of Politics, University of Virginia. Wolfgang C. Müller  is Professor of Democratic Governance at the University of Vienna. Kevin J. Mullinix  is Doctoral Candidate in Political Science and Graduate Fellow at the Institute for Policy Research, Northwestern University. Keith Poole  is Philip H. Alston Distinguished Chair and Professor of Political Science at the University of Georgia. Sven-Oliver Proksch  is Assistant Professor of Political Science, McGill University. Bjørn Erik Rasch  is Professor of Political Science and Deputy Dean for Education at the Faculty of Social Sciences, University of Oslo. Tapio Raunio  is Professor of Political Science, University of Tampere. Thomas Saalfeld  is Professor of Political Science at the University of Bamberg and the founding Director of the Bamberg Graduate School of Social Sciences. Sebastian M. Saiegh  is Associate Professor of Political Science, University of California, San Diego.

About the Contributors  

xiii

Constanza F. Schibber  is Doctoral Candidate in the Department of Political Science, Washington University in St Louis. Paul Schuler  is Doctoral Candidate in the Department of Political Science, University of California, San Diego. Matthew S. Shugart  is Professor of Political Science, University of California, Davis. Ulrich Sieberer  is Research Group Leader in the Department of Politics and Fellow of the Zukunftskolleg at the University of Konstanz. Jonathan B. Slapin  is Associate Professor of Political Science, University of Houston. Kaare W.  Strøm  is Distinguished Professor of Political Science at the University of California, San Diego. Michelle M. Taylor-Robinson  is Professor of Political Science, Texas A&M University. Georg Vanberg  is Professor of Political Science at Duke University. Lars Vogel  is Research Fellow at the Department of Sociology, Friedrich Schiller Jena University of Jena, Germany. Joachim Wehner  is Senior Lecturer (Associate Professor) in Public Policy at the London School of Economics and Political Science.

C HA P T E R

1

INTRODUCTION SHA N E M A RT I N , T HOM AS S A A L F E L D, A N D KA A R E W. ST RØM

1.1 Introduction Legislatures are ubiquitous political bodies. In a recent survey, Loewenberg (2011, 18)  counted 191 such national assemblies. Many thousands could be found at the sub-national level. And at least in the case of the European Parliament, we have seen the slow but steady rise of a transnational assembly. In the broadest sense of the word, a legislature could be defined as “a body created to approve measures that will form the law of the land” (Norton 2013, 1). This definition focuses on law-making as the main purpose of a legislature and can be derived from the word’s etymological root, namely legis, the genitive of the Latin word for “law” (lex) and lator meaning “proposer” or “carrier” conjugated from the Latin verb ferre. The local names for legislatures include (in English translation) “Assembly,”“Diet,”“Congress,”“Parliament,” and many more. The term “parliament,” which is sometimes used interchangeably with “legislature,” has a different etymological root and points to a different set of activities with which legislatures tend to be associated—public deliberation, debate, transparency, and accountability. The French word parlement appeared in the twelfth century to denote an assembly or court where French monarchs (or their representatives) spoke to subjects, listened to their grievances, and explained policies. From the thirteenth century onward, the term parliamentum was used in medieval England for formal meetings between the monarch and representatives of the estates, which emanated from the Curia Regis of the eleventh century (Marschall 2005, 24–6). In his classic study of The English Constitution, Bagehot (1867) described these communicative activities under the rubrics “expressive,” “teaching,” and “informing” functions of the British Parliament. Yet, the most important function Bagehot ascribed to the nineteenth-century House of Commons as an “efficient part” of the English constitution is not captured in the terms “legislature” or “parliament” at all, namely the legislature’s “elective function.” The making, breaking, and maintaining of governments in

2  Introduction office is peculiar to legislatures in parliamentary systems of government, but does not exist in presidential systems based on a stricter separation of the executive and legislative branches. In parliamentary systems of government, the majority leadership controls the process of legislation via its majority and its powers to determine the agenda. In presidential systems, by contrast, the power to set the agenda does not rest with the executive. As a result, the executive needs congressional cooperation to get its legislation on the statute book. In parliamentary systems, the ultimate sanction of dismissing a government turns the communicative functions highlighted by Bagehot potentially into a very effective device to hold the executive accountable, both within the majority party (or parties) and in the permanent election campaign between majority and minority parties, between government and opposition (King 1976). In the past 50 years, legislative studies have become one of the major subfields of political science. Many national and international political science associations (e.g. the American Political Science Association, the [British] Political Studies Association, the International Political Science Association, or the European Consortium for Political Research) have standing sections on legislative studies. The number of articles on legislatures or legislators in leading international journals in the discipline is substantial. And there are a number of established academic journals specializing in legislative studies in political science (e.g. Parliamentary Affairs [founded in  1947], the Zeitschrift für Parlamentsfragen [since 1969], Legislative Studies Quarterly [since 1976], the Australasian Parliamentary Review [since 1986], or The Journal of Legislative Studies [since  1995]), and in other disciplines (e.g. Parliaments, Estates and Representation [since 1981], or Parliamentary History [since 1982]). Not only are legislatures ubiquitous bodies, they present political scientists with numerous intriguing puzzles, which have motivated research with implications for institutional analysis beyond the field of legislative studies itself. Why, and how, have these ancient assemblies, established in pre-democratic times and often characterized by aristocratic forms of behaviour, procedures, and symbols, survived the transition to mass democracies? How have they adapted? Why do they still exist, after all the diagnoses of anachronism, decline, irrelevance, and dysfunctionality? Why have they not been abolished in the light of widespread popular indifference even in those democracies that look back at a long-standing history of parliamentary government? What explains the similarities and differences in legislative rules, powers, and recruitment? What explains the inconsistency between the organizational and procedural structures they have developed and the original notion of an “assemblage of notables of equal status” (Loewenberg 2011, 14)? And, perhaps most intriguingly, what are the policy and other consequences of variation in how legislatures are organized and function? In short, legislatures pose a number of non-trivial puzzles. At the same time, they have always been relatively public and transparent bodies offering good opportunities to researchers interested in questions of institutional design, institutional adaptation, as well as rule-bound and strategic behaviour. A  great deal of information is available to researchers about the legislatures’ constitutional powers and internal rules as well as about their members’ backgrounds, public statements, revealed preferences,

Introduction   3

and behaviour in and around the chamber. Today, many legislatures provide excellent on-line access to records of plenary meetings, committee sessions, and the entire legislative process. This may explain why legislative studies have become such an important and increasingly dynamic sub-field of political science.

1.2 Legislatures Matter Some scholars have claimed the decline of parliaments in modern democratic politics. Such claims are not new. Lowell’s (1896), Ostrogorski’s (1902), and Bryce’s (1921) analyses of the effect of the rise of organized political parties in the nineteenth century led them to conclude that the power of legislatures had been undermined by partisanship (Ostrogorski, Bryce) and a decline in the quality of representatives (Bryce). This critique was echoed by foes of liberal democracy such as Schmitt (1988 [1923]) in the 1920s as well as by supporters of a more rational and democratic form of discourse in legislative decision-making, such as Habermas (1992 [1962]) in the 1960s. While political scientists have gradually come to understand how political parties can enhance the decisional efficiency and impact of legislatures (see, for example, Cox 1987; Aldrich 1995; or Cox and McCubbins 2005), the claim that legislatures are declining has remained. What has changed have been the external developments believed to cause this downturn: after the Second World War, a number of authors identified the growth of welfare states and executive agencies as causes for the alleged decline of European legislatures, as the growth of these programs and agencies has exacerbated the informational asymmetry between bureaucrats and elected politicians. In the 1970s, a variant of this causal argument was extended to the tendency for governments in some (again, mostly European) democracies to strike bargains with major producer groups in neo-corporatist arrangements. Richardson and Jordan (1979) thus spoke of a “post-parliamentary democracy.” In recent years, the narrative of declining legislatures has continued on two levels: some authors consider Europeanization (within Europe) and globalization (more generally) as crucial exogenous developments undermining the power of legislatures. Andersen and Burns (1996, 229) thus claim: “Western societies have become highly differentiated and far too complex for a parliament or its government to monitor, acquire sufficient knowledge and competence, and to deliberate on.” Parliamentary representation, they argue, is organized territorially, rendering legislative oversight ineffective in the contemporary world. In a recent article on the implications of the European rescue packages between 2010 and 2013, Zürn (2013, 10) restated the argument from the vantage point of transnational governance: “The crisis of parliamentary democracy in the West and the emergence of political authority beyond the nation state lead to a shift within the set of established justifications of legitimate authority.” His assessment has been echoed by students of public attitudes. Dalton (2004, 1), for example, observes that citizens “have grown distrustful of politicians, skeptical about democratic institutions, and disillusioned about how the democratic process functions.” And the available evidence

4  Introduction suggests that despite the relative transparency of legislative chambers, they are poorly understood and appreciated by democratic publics across the world (Loewenberg 2011, 92–103). Yet, the contributions to this handbook will provide overwhelming evidence that legislatures matter, and how they matter; how they have adapted to changes such as globalization, regional integration, judicialization; and how they have coped with the declining support base for the central agents of legislative politics, political parties. A few simple but striking examples corroborate our claim that legislatures are indeed consequential. In March 2013, the government of the Republic of Cyprus concluded an economic bailout agreement with the European Union, the European Central Bank, and the International Monetary Fund. This so-called troika imposed harsh conditions, including a loss of up to 10 percent of money held in Cypriot bank accounts. While all parties to the agreement expected negative reactions, neither the Cypriot government nor the troika considered the reaction of the parliament of Cyprus, through which the necessary enacting legislation would have had to be passed. Banks in Cyprus had to remain closed while the government tried to persuade Members of the House of Representatives to back the deal. On 19 March 2013 Cypriot MPs rejected the bailout terms, with more than half of the country’s 56 legislators voting against, and 19 MPs choosing to abstain. Amidst an atmosphere of crisis, the Cypriot finance minister then went to Moscow to seek help from Russia, and the European Central Bank threatened to withdraw the supply of cash to Cypriot banks. Finally, following new negotiations with the troika, a multiweek shutdown of banks, and much protest, Cyprus’s House of Representatives approved a different, and more favourable, bailout agreement at the end of April, by a majority of just two votes. Thus, even these powerful transnational actors had to take into account the preferences of legislators of a small state in an utterly weak bargaining position. Although the above account reminds us of the central role that legislature plays in times of crisis, legislatures are also central to the normal business of governing. In parliamentary systems, parliaments make and break governments (Bagehot 1867; Wheare 1963; Kluxen 1983; Laver and Shepsle 1996; Müller et al. 2003; Strøm 2003). The Federal Republic of Germany is a powerful illustration of this point. Between 1949 and 2013, only one complete transfer of power between government and opposition parties was triggered exclusively by a popular vote at elections—in the general election of 1998, the coalition of Christian Democrats and Liberals under Federal Chancellor Helmut Kohl was thrown out by the voters and replaced with a new coalition of Social Democrats and Greens led by Gerhard Schröder. All other governmental transitions (Adenauer to Erhard in 1963, Erhard to Kiesinger in 1966, Kiesinger to Brandt in 1969, Brandt to Schmidt in 1974, Schmidt to Kohl in 1982, and Schröder to Merkel in 2005) effectively resulted from negotiations between the parties in the legislature. This provides powerful evidence for the importance of Bagehot’s elective function in parliamentary systems of government. Although in presidential regimes legislatures tend not to make or break governments, their policy influence remains significant. As the passage of the Patient

Introduction   5

Protection and Affordable Care Act of 2010 (often referred to as the ACA, or Obamacare) in the United States exemplifies, the detail of legislative rules and procedures matter greatly. President Obama had identified healthcare reform as a policy priority of his first term in office. Yet, healthcare reform divided Americans (Brady and Kessler 2010). Normally, such divisive legislation would stand little chance of passing in the United States Senate, where individual Senators can filibuster bills—unless 60 percent of the members oppose the filibuster. Due to an embarrassing loss in a Senate by-election in Massachusetts, the president’s party did not have the filibuster-proof Senate majority required to pass the ACA. To circumvent a filibuster, the Democrats instead employed a little-known and arguably arcane parliamentary maoneuvre. Senate procedures allow proposals directly affecting the budget to be considered under a special budget reconciliation process. The healthcare bill had both legislative and budgetary components and it was therefore unclear which procedure would be used. The Parliamentarian (the Senate’s nonpartisan advisor on the interpretation of its rules, procedures and, critically, precedents) must decide whether bills are budgetary or legislative. Legislative bills can lead to long debates with the possibility of a filibuster, which can only be broken by a supermajority of 60 senators. In contrast, if a bill is budgetary, debate is limited to 20 hours with no opportunity to employ a filibuster. The interpretation of the Parliamentarian, Alan S. Frumin, was thus critical to the eventual passage of the ACA. As Oleszek (2013, 432), later noted: “Members [of the United States Congress] find new uses for old rules, employ innovative devices, or bypass traditional procedures and processes altogether to achieve their political and policy objectives.” And in this case, the fate of a very large and hotly contested bill depended on this arcane parliamentary ruling. As all of the above stories indicate, it is difficult if not impossible to study the politics of any established democracy without reference to its legislature or legislatures, and as some of the later chapters in this volume will suggest, legislatures sometimes play a critical role even where democracy has not been solidly entrenched.

1.3  The Evolution of Legislative Studies Legislative studies have always been a rich and pluralistic field of inquiry in political science. We want to offer three observations on the history of this field. Firstly, the study of legislatures pre-dates the evolution of political science as a discipline. Philosophers, historians, and lawyers were the first scholars to analyse legislatures. This has influenced the way legislatures have been—and are—studied. Political science does not have a monopoly on the study of legislatures, and in a number of countries, legal-constitutional and historical traditions retain a strong influence. Secondly, while pluralistic in its methodology and methods, legislative studies in political science have followed a general trajectory from (a) the “old” institutionalism prevalent between the late nineteenth century and the end of the Second World War, through (b) a neglect

6  Introduction of political institutions and a focus on individual behaviour in the 1950s and 1960s, to (c) more sophisticated macro-micro-macro perspectives from the mid-1980s onwards when the “new” institutionalism became a dominant force in contemporary political science. Thirdly, legislative studies have benefitted greatly from the growth of political science as a discipline after 1945, especially in the United States of America. The vast majority of scholars working in the field have always been based at US universities. This includes US-based scholars who have made significant contributions to research on non-US legislatures (e.g., Beer 1966; Huber 1996; Loewenberg 1967). As a result of the critical mass in US academia and other fortuitous circumstances, legislative studies have received major impulses from America. Our first point relates to the long history of studies of legislatures and the variety of academic disciplines that have covered legislatures and their activities. Scholarship about the process of law-making has a long and venerable tradition in a number of disciplines including philosophy, history, law, and, since the second half of the twentieth century, political science. Early typologies of government (including assemblies) can be found in ancient Greece in the Politics of Aristotle and Plato and, later in the Roman republic, in Cicero’s De Republica. We have fragmentary evidence of philosophical works on laws by Plato (Nomoi) and Cicero (De legibus), which were both primarily concerned with the content of laws rather than with the process of legislation. Throughout the Middle Ages authors such as Thomas Aquinas (1225–74) harked back to Greek (mainly Aristotelian) and Roman political thought in treatises on divine, natural, and man-made law. In early modernity, authors such as Bodin, Grotius, Hobbes, and Pufendorf took up questions of natural law and sovereignty in modern states, reflecting on the balance of power between sovereign monarchs and the various estates. The contractarian ideas of Hobbes and Locke, in particular, strongly influenced the understanding of the relationship between citizens and monarchs in subsequent liberal political thought. Locke in his Second Treatise on Government (1690) and Montesquieu in De l’esprit des Loix (1748) argued for a separation of power between the executive and legislative branches of government, an idea later taken up by the founding fathers of the US Constitution in the Federalist Papers (1787–8). In eighteenth-century Britain, Burke and Hume theorized about political parties in legislatures (which they saw largely as divisive expressions of particularistic interests), and Bolingbroke formulated a theory of legitimate, “patriotic” opposition (on Burke, Hume, and Bolingbroke see Mansfield, 1965). This list is far from exhaustive but suffices to demonstrate the importance of key areas of legislative studies in political theory for more than 2,000 years. In addition to its roots in philosophy, legislative studies owe a great deal to legal-constitutional scholarship. Until the eighteenth century, legal, theological, and philosophical works were hard to separate. Two developments favoured the “constitutionalization” of legislative politics: firstly, and most obviously, there was the rise of constitutionalism from the American and French Revolutions onwards. Combined with the rise of legal positivism, this development provided a major impetus for legal research. Secondly, with the growing volume and complexity of legislative business in some chambers, procedural rules had to be introduced to avoid chaos. Initially, these

Introduction   7

rules were written by politicians or clerks. Vice President Jefferson (1812) famously compiled his Manual of Parliamentary Practice for the US Senate when he held the office of vice president between 1797 and 1801. This influential statement of parliamentary law was informed by Hatsell’s (1818 [1718]) procedural manual for the British House of Commons and later “became a supplement to the rules of procedure that evolved in the U.S. House of Representatives and in territorial and state legislatures throughout the United States” (Loewenberg, 2011, 17). While such procedural rules have increasingly become a focus of research for political scientists (see the chapter by Müller and Sieberer in this volume), they have long been the territory of constitutional lawyers. In some countries the body of law pertaining to legislatures (such as in the German word “Parlamentsrecht”) have become a sub-field of public law (e.g., Schneider and Zeh 1989; Jack 2011). As democratic legislatures have evolved from medieval assemblies in Europe, many archaic procedures, symbols, and myths have continued to shape legislative practice despite the modernization and “rationalization” of legislatures since the nineteenth century (Manow 2010). Apart from a number of influential studies by eminent historians, there have been major attempts to coordinate and institutionalize historical research on legislatures through organizations such as the History of Parliament Trust in the United Kingdom (established in 1940), the International Commission for the History of Representative and Parliamentary Institutions set up by a number of European historians (1936), the German Kommission für Geschichte des Parlamentarismus und der politischen Parteien (1952), the Dutch Centrum voor Parlementaire Geschiedenis (1970), and the French Comité d’Histoire Parlementaire et Politique (2003). In short, political science is only one of several disciplines dealing with legislatures, and inter-disciplinary cooperation has often been lacking (Loewenberg 2011, 6). From the 1950s onwards and again from about 1980, legislative studies have nevertheless become a very dynamic sub-field of political science, especially in the United States. Our second main observation is that the history of legislative studies in political science is strongly influenced by the ebb and flow of institutionalism in political science and the social sciences more generally. Most early publications on legislatures are characterized by the attributes Peters (1999, 3–11) and others have identified as “old institutionalism” of the nineteenth and early twentieth centuries. This traditional form of institutional analysis is too diverse to have a coherent theory, methodology, or research program. Yet it is generally characterized by a strong focus on the formal institutions of government, constitutional law, and constitutional history. Loewenberg (2011, 108) thus observes that, some notable exceptions notwithstanding, US congressional research from the 1880s to the 1950s had been dominated by “a large volume of work on its constitutional powers, rules of procedure, and legislative enactments, as well as proposals for its reform.” “Old” institutionalists believed that structures had a clear causal influence on behaviour. In methodological terms their works tended to be holistic (rather than individualistic) with an emphasis on macro-level analyses. They also stressed the historical foundations of institutions both for an adequate understanding of their origins and the interpretations of the actors influenced by those institutions. Finally, and in line with the intellectual roots

8  Introduction of political science, there was a strong normative element in “old” institutional analysis, focusing on the institutional conditions for “good government.” In this sense, Tocqueville’s early description of the United States Congress in De la démocratie en Amérique (first published in French in 1835) may be a less-than-perfect example of the “old institutionalism” as the book reveals a great deal of interest in the fabric of American society, its relative egalitarianism, and the consequences of these social factors for the country’s political institutions. Nevertheless, Chapters V and VI present an early institutionalist account of Congress (Tocqueville 2000). Wilson’s (1885) study of Congressional Government half a century later can be considered to be a clear-cut early example of the “old” institutionalism in the United States. His focus on the formal institutions of Congress was coupled with the normative impetus of a practitioner who sought to influence behaviour through institutional reform. In Europe, the studies of Bagehot (1867), Bryce (1921), Jennings (1957), or Wheare (1963) are further examples. Wheare was unusual inasmuch as he dealt with more than one single legislature. However, systematic, theory-driven comparative analyses did not become widespread until the 1990s. Old institutionalism was largely a matter for area specialists, although a few scholars such as Polsby (1974) , Mezey (1979), or Lijphart (1984) contributed greatly to the development of typologies that could be used for comparative studies (see also Mershon’s contribution to this volume). Even the most cursory review of some older institutionalist literature, however, would be incomplete without reference to a large body of institutional analysis, especially during the 1920s and 1930s, that was deeply skeptical of legislatures and parliamentary democracy (e.g. Schmitt 1988 [1923]). In their most general form, these arguments—for example, the claim that legislatures had lost their character as public arenas for rational political discourse, or that they were ineffective and inefficient as decision-making bodies—continued to inform institutional debates after 1945. The literature on discourses and deliberative democracy are but two examples of the former theme (see Bächtiger’s contribution to this volume), whereas the institutional analysis of chaos results in parliamentary voting exemplify the latter (see Rasch’s chapter in this volume). In the 1960s, political scientists imported systems theory and structural functionalism from other social sciences to move beyond the old institutionalism of legislative studies both in the United States and elsewhere. The theoretical perspective underpinning systems theory proved to be a promising candidate for a general theory of political systems. Almond and Powell (1966), for example, saw legislatures as sub-system of a “political system” which helped to convert political inputs arising in the system’s environment into outputs. It is typical of this perspective that a great deal of scholarship was devoted to systemic “functions” legislatures might perform—Packenham (1970), for example, identified three main functions: (1) legitimation; (2) recruitment, socialization, and training of political leaders; and (3) decisional functions including law-making. This functionalism seemed to provide a framework for more analytic and comparative perspectives. Ultimately, however, systems theory and its derivatives in legislative studies remained abstract and highly aggregated and did little to address the growing criticism that “macrobehaviour” needed to be explained in terms of actors’ (e.g., legislators’) “micromotives” (Schelling 1978).

Introduction   9

In another departure from the “old” institutionalism, the 1950s and 1960s witnessed a growing interest in legislators and their individual perceptions and behaviour. This involved a radical shift to the micro level of analysis, which was inspired by the general rise of behaviourism in the social sciences. The interest in institutions, by contrast, declined: What really mattered in this view of politics was voting, interest group activity, and even less legal forms of articulations, which were then processed into “outputs.” In this conception of a political system the formal institutions of government were reduced to the “black box,” where the conversion of inputs into outputs occurred, almost magically it appeared to “critics of the approach” (Peters 1999, 14). Such criticism notwithstanding, modern legislative studies owe much, both substantively and methodologically, to a series of micro-level studies, such as a seminal study of representational role orientations among American state legislators, in which Wahlke and his collaborators (1962) interviewed 474 members of four such assemblies. With appropriate adjustments, their work inspired a large number of related studies across the world (Badura and Reese 1976; Chee 1976; Narin and Puri 1976; Converse and Pierce 1979; Hagger and Wing 1979; Maier et al. 1979; Clarke and Price 1981; Andeweg 1992; Searing 1994; Judge and Ilonszki 1995; Müller and Saalfeld 1997; Blomgren and Rozenberg 2012). Similarly, Fenno’s (1978) participant observation of the “home styles” of US Representatives generated interesting insights into the way US legislators carry out their work on Capitol Hill and also influenced international studies (e.g. Patzelt 1993; Brack et al. 2012), albeit not to the same extent as for research on role orientations. Despite the impressive amount of information that has become available on legislators, the micro-political approach to the study of legislatures had some deficiencies, which its leading proponents readily recognized. This behavioural research contributed to a host of legislature-specific in-depth studies; yet the question remained whether the fascinating results of these studies could be generalized beyond their specific samples. Furthermore, there was little leverage to explain the effect of institutional variation. Eulau (1996), one of the key proponents of the micro-political approach, saw this problem very clearly and emphasized the need to bridge the micro-macro divide. Finally, the data from such behavioural studies often made it difficult to draw clear causal inferences. Searing (1994, 6–7), a leading scholar in the field of legislative role analysis, thus conceded that the ideas underpinning much of his work were a descriptive framework rather than a theory from which causal explanations could suitably be derived. The “new” institutionalism that gradually came to dominate legislative studies from the 1980s onwards was not a renaissance of the structuralism and methodological holism of the “old” institutionalism. It was strongly informed by three decades of empirical research on the attitudes and behaviour of individual representatives in the US and elsewhere. In the language of Coleman’s (1986) macro-micro-macro model of (social) action, it benefitted from vastly increased knowledge of how legislators perceived macro-level constraints and incentives in their respective political environments (Coleman’s “macro-to-micro problem”). It therefore offered scholars some confidence in their respective micro-level assumptions.

10  Introduction Nor was the new institutionalism all of one kind; in fact it came in two distinct main varieties: one committed to rational choice analysis and the other much more sociological and norms-oriented. These rational-choice and sociological varieties largely differed in terms of their assumptions and explanatory strategies at the individual level. Rational-choice theorists were strongly influenced by the rise of formal modeling since the 1960s. The micro-level research of the past three decades had reassured Riker and his followers that their assumption of strategic, calculating actors was reasonable in the context of legislative institutions (Shepsle 2002). Loewenberg (2011, 116), though not a scholar of this tradition, conceded that Riker’s game-theoretic “approach was ideally suited to research on legislatures because they could readily be conceptualized as institutions composed of self-interested, purposeful members taking strategic actions within a set of rules to achieve their purposes.” Sociological institutionalists, by contrast, assumed that legislators and other political actors follow a “logic of appropriateness” (March and Olsen 1984), which they took to be acquired in the process of political socialization. Strategies and purposive behaviour exist, but political actors’ assessments of costs and benefits are influenced by the social norms that are prevalent within their communities and organizations. With its emphasis on voting procedures and other rules, the rational-choice variety of the “new institutionalism” came to dominate, as it offered interesting and often counter-intuitive explanations of legislative outcomes, addressing the “micro-to-macro problem” identified by Coleman (1986). Unlike the “old” institutionalists, behaviourists, or systems theorists, rationalchoice institutionalists, in particular, had a clear (and often fairly realistic) model of individual motivations. Unlike the first generation of rational-choice models of legislative politics (see, for example, Riker, 1962), however, institutions were not strictly endogenous to their models. Using fixed behavioural postulates about the motivations and preferences of relevant actors and the game-theoretic logic of institutions, they sought to derive propositions about observable behaviour and test hypotheses based on these propositions (Diermeier and Krehbiel 2003, 131). One crucial advantage of rational-choice institutionalism over its sociological “competitor” was that it lent itself much better to comparative analyses. Concepts such as “voting games,”“veto players,” or “agenda control” allowed researchers to strip a multitude of different institutions down to their institutional core logic. Veto-player models (e.g. Tsebelis 2002) or pivotal-player models (Krehbiel 1998) were general enough to be applied to and tested in a large number of different settings. The growing availability of cross-national data on legislative powers and procedures (e.g. Döring 1995; Fish and Kroenig 2009; Sieberer 2011) has allowed more sophisticated comparative analyses of the impact of legislative institutions. Although the rational-choice approach has provided the most influential neo-institutional approach in legislative studies, there have been other influential schools of scholarship, especially the sociological exploration of legislative roles, as well as historical approaches. The historical approach to the study of institutional development—including a much stronger cooperation between rational-choice theorists and historical institutionalists (e.g. Katznelson and Weingast 2005)—is a recent development

Introduction   11

and has largely remained confined to Congressional scholarship in the United States thus far (Katznelson 2011). Sociological institutionalism in legislative studies has its roots in the micro-political work on legislative roles cited above. The study of legislative roles continues to produce a number of fascinating findings on individual legislatures (see, for example, Blomgren and Rozenberg 2012), even though its proponents realized that “role theory is not a theory” (Searing 1994, 6–7). Nevertheless, sociological analyses, too, have experienced a certain revival in the context of the “new” institutionalism and have contributed to a much better understanding of the longer-term changes in the offices and profession of legislators cross-nationally (Best and Cotta 2000; Cotta and Best 2008; see also the contribution of Best and Vogel in this volume). Our third general observation on the field of legislative studies relates to the dominance of US scholars. This is not surprising given the size of the US political science community and the fact that both Congress and the US state legislatures are amongst the most powerful legislative bodies world-wide with a long and uninterrupted history of democratic government. As a result of this strength, Shepsle (2002, 390) was able to claim with a high degree of plausibility that “American politics has served the wider political science community by forging scientific tools and producing a laboratory in which they are tested, perfected and prepared for export.” The story told in this introduction supports Shepsle’s claim. While legislative studies were multi-disciplinary and multi-polar until the Second World War and while European research, in particular, may still be strong in areas such as the legislative history, the most important innovations since 1945 have emanated from research carried out in the United States. This has always included non-American researchers who had found an intellectual home there, and increasingly comparative research collaborations across the Atlantic and the Pacific oceans, as well as across the border between North and Latin America (e.g. Alemán 2013).

1.4  The Handbook Our handbook reflects the richness and promise of legislative studies and is therefore divided into a number of sections that survey and discuss various key themes and issues in the field. We begin with basics, with accounts of the different traditions in which insights about legislative politics have been aggregated and developed into theoretical bodies of knowledge concerning this aspect of political deliberation. The handbook contains three such chapters, to which we now turn.

1.4.1  Theoretical Approaches in Legislative Studies Like political science in general, scholars of legislatures approach the topic from different and, at least partially, competing theoretical perspectives. Diermeier (Chapter 2)

12  Introduction suggests that formal modeling and, in particular, game-theoretic approaches, are well suited to the comparative study of legislatures. Almost all formal models of legislatures fall into this tradition. Rational choice theory contends that political outcomes and decisions are to be explained as the aggregate results of the decisions of rational individuals, variously informed. Variants of rational choice theory differ in their view of how to model such collective choice processes. Social choice theoretic approaches model a collective choice process by its voting rule. A series of impossibility theorems, suggesting the collective choice should typically be chaotic, led many theorists to the conclusion that the sparse social choice theoretic approach was underspecified. This led to the development of institutionalist models, which add structure, such as committees or procedural prerogatives, to formal models of legislatures, to explain stabilities and biases that seem to exist in the real world of legislative politics. Early models of such structure-induced equilibrium have been increasingly replaced by ever more sophisticated game-theoretic approaches. As Diermeier argues, this framework not only allows for the representation of complex institutions but can also address questions of stability and the choice of institutions. Despite the dominance of rational choice, the sociological approach provides an alternative theoretical framework through which legislatures and legislators can be studied. The sociology of legislators and legislatures has two main units of analysis: legislators as social actors on the one hand, and legislatures as social institutions on the other. Best and Vogel (Chapter 3) focus on the social relations within parliaments (insider relations) and on the relations between parliaments and society (insider-outsider relations). They suggest that three types of social interaction manifest themselves: competitive struggle for votes; antagonistic cooperation; and hierarchical principal–agent relations. Kreppel (Chapter 4) provides a review of the historical evolution of legislative typologies and classification schemes, with particular attention to the implicit and explicit connections between the role and influence of legislatures and the broader political system. Her chapter includes an effort to clarify the terminology of legislatures and to identify various subtypes that exist based on their relationship with the executive and the relative emphasis they are likely to place on the primary activities of legislatures (including representation, linkage, control, and policy-making). Kreppel explores the implications for comparative legislative studies of the transition from the broad comparative “macro” typologies dominant in the 1970s and 1980s toward the micro analyses focused primarily on specific aspects of internal organization, including the shift from US-centred analysis to a more comparative lens. Kreppel concludes by introducing an alternative conceptual scheme based on the core concepts of institutional and individual autonomy.

1.4.2  Methods of Legislative Research Our second set of five chapters reviews the rapidly evolving methods and techniques that are currently deployed in the study of legislatures. Contemporary legislative studies thus make use of an increasingly diverse and sophisticated set of empirical tools to test

Introduction   13

an ever-increasing number of theory-driven hypotheses. Carroll and Poole (Chapter 5) explore what is arguably one of the most empirically studied aspects of legislative behaviour: voting. Scaling methods to evaluate the latent dimensions of political behaviour and choice have had a substantial impact on our understanding of the properties of legislative roll-call voting. As interest in spatial models of legislatures has grown, these methods have been employed to operationalize theories of the role of preferences and ideology in legislative politics. Scaling procedures and ideal point estimation have enabled scholars to assess the spatial properties of voting and to conduct numerous empirical investigations of preference-based theories of legislative politics.With the advent of new techniques and more computational power, such methods have become even more widespread in comparative politics. Carroll and Poole provide an overview of these methods and applications with a discussion of methodological challenges and recent developments in the field. Yet, legislator behaviour consists of much more than voting. As we noted in our introduction, the term “parliament” refers to the discursive aspects of politics in representative assemblies, and legislators leave a long record of recorded words in the form of legislative debates, parliamentary questions, and committee hearings and deliberations. Slapin and Proksch (Chapter 6) discuss recent technological advances that have provided scholars with the ability to analyse the vast amounts of textual data that legislatures produce each year. Researchers have taken advantage of these new tools to answer questions about constituency representation, agenda control, and party cohesion, among many other topics. Using texts as data has allowed researchers to develop a more nuanced understanding of ideology and position-taking in legislatures. However, a careful consideration of the data-generating process behind the text data is equally important as content analysis of that data. Just as with other forms of observational legislative data, text data is often generated by strategic interactions among legislators and legislative parties. Slapin and Proksch warn that when they consider how best to use available text data, researchers must be cognizant of the processes that generated the text. Bächtiger (Chapter 7) explores the role of debate and deliberation within legislatures. In the past two decades, scholars have begun to refine and challenge the conventional view that debate is a verbal contest between government and opposition without any direct policy impact. Bächtiger explains that parliamentary debate is more varied and consequential than commonly assumed. In trying to understand such debate, scholars have adopted three approaches:  (1)  a strategic and partisan-rhetoric approach, anchored in rational choice theory; (2) a deliberative approach which tries to identify institutional and issue-based conditions under which legislative speech can inform decisions; and (3) a discourse approach focused on the constitutive features of parliamentary debates which examines its norms and conventions. Bächtiger’s analysis of the contours of parliamentary debate compliments the previous chapter on text as data and will be of interest not just to scholars of deliberation but also to those who develop or employ the tools discussed by Slapin and Proksch in Chapter 6.

14  Introduction Surveys and interviews, which were so influential during the heyday of behaviourism, arguably continue to dominate empirical research designs in legislative studies. For one thing, they help give meaning to what legislators do within the legislative arena (voting and words). They also shed important light on legislators’ extra-legislative work, such as constituency activity and service. Arguably, surveys and interviews deliver the most direct measure of the thoughts and intentions of politicians, so that they are one of the most valuable sources of data for the study of legislative behaviour. Bailer (Chapter 8) reviews the use of surveys and interviews in legislative studies and offers specific guidance for drafting questionnaires and surveys questions, for conducting legislative surveys, and for tackling specific problems that may arise in elite interviews. Bailer discusses the major challenges of legislative interviews (quality control) and of legislative surveys (low response rates and selection bias), and offers potential remedies. The challenges of quality control and selection bias ultimately require that scholars consider alternative, or at least complementary, empirical strategies. The increasing popularity of experiments in political science suggests one possible avenue. While early work on legislative coalitions occasionally employed experiments, the bulk of progress in legislative research has been made with non-experimental work. This is likely to change quickly. Druckman, Leeper, and Mullinix (Chapter 9) argue that experiments provide unique strengths in legislative research. During the past decade, political science has experienced a mounting tide of experimental research directed at topics such as voting in legislatures, parliamentary coalitions, and legislative responsiveness. Druckman, Leeper, and Mullinix discuss variations in experimental design and their applications before concluding with a discussion of future challenges and opportunities for experimental work in legislative settings.

1.4.3  Representation and Legislative Careers As the five methodologically focused chapters will demonstrate, legislative scholars have a rapidly growing tool box of increasingly sharp instruments at their disposal. They can, and no doubt will, employ these tools to address a range of important questions. Regardless of theoretical orientation and empirical strategies, researchers in the field of legislative studies have produced an impressive body of substantive knowledge, which is explored in the next series of chapters in this handbook. Our analysis of this body of work begins, as do political campaigns, with the candidate selection process. Hazan (Chapter 10) argues that candidate selection has significant consequences for legislative politics. The rules and practices that govern this process open doors for some politicians and close them for others. Moreover, candidate selection rules may induce legislators to modify their behaviour, regardless of their fundamental preferences. Hazan argues that the behaviour of legislators is significantly influenced by their desire to be reselected, and that in order to fulfill this goal they will try to satisfy and respond to the selectorate. For example, if party cohesion has a positive

Introduction   15

effect on reselection, legislators will focus on party-centred activities and exhibit high levels of party unity. If more individualistic behaviour is valued in order to be reselected, legislators will engage in candidate-centred behaviour, leading to lower levels of party unity. Hazan is especially interested in the shift toward more inclusive candidate selection methods and argues that this will have a significant impact on legislative behaviour. André, Depauw, and Shugart (Chapter 11) review what has become a central topic in comparative politics—the relationship between electoral systems and legislative behaviour. Much research indicates that electoral institutions shape the behaviour of legislators and thereby strengthen or weaken their incentives to cultivate their personal reputation rather than that of their party. But research has largely missed the “theoretical link” tying legislators’ behavioural repertoires to the formal rules. In response, and by focusing on patterns of accountability as well as the competitiveness of electoral races, André, Depauw, and Shugart set out to examine how the mechanical effects of electoral institutions translate into incentives for legislators to cultivate a personal vote. They then turn to the psychological effects of electoral institutions, examining how personal vote incentives translate into behaviour such as initiating particularized legislation, undertaking constituency casework, and breaking with party discipline. The effects of electoral institutions on legislative behaviour should not be overstated, however. The electoral calculations of legislative candidates may thus be upset by the selective benefits (such as legislative particularism or mega-seats) handed out by party leaders, as well as the voters’ preferences regarding representation. Research on the representation of women in legislatures continues to uncover fundamental information of interest to legislative scholars and to scholars of gender and politics. Taylor-Robinson (Chapter 12) explores how institutional design may facilitate or hinder the election of women; how women and men legislators define their job and their constituents; how female and male legislators perform their constituency service, propose bills, and participate in deliberation. She also analyses the representation of women in legislative leadership positions. The author then proposes focus areas for future research, particularly whether and how descriptive representation is linked to substantive representation of women’s interests. She argues the need for greater attention to the diversity among women legislators and to how this diversity impacts substantive representation. More broadly she encourages research on the political opportunity structure that may promote or inhibit the promotion of women’s interests by either male or female representatives. In no other area of political science is the concept of “roles” as important as it is in legislative studies. It is sometimes argued that role analysis is currently regaining importance in this field in response to the dominance of institutionalist approaches. Andeweg (Chapter 13) is more skeptical, advocating the need for a move from mere description to theoretically grounded explanation. Roles make political institutions such as legislatures subjective: they are related to positions, but not identical to them. Perceived expectations, personal motivations, and strategic calculations are assumed to differentiate roles from positions. Eulau and Wahlke’s typology of representational roles, inspired by Burke and based on a study of US state legislatures, has long dominated the study of

16  Introduction legislative roles. Andeweg criticizes this conception on both theoretical and empirical grounds, in particular for its lack of predictive power. Searing’s typology, based on an inductive study of the UK House of Commons, has attracted the most attention as an alternative. Andeweg argues that both seminal studies are lacking due to the ambiguity of their treatment of political parties. Scholars have long recognized the theoretical link between political institutions and legislative careers. Nonetheless, questions remain about how basic institutional factors influence career development within legislative settings. Kousser and MacKenzie (Chapter 14) argue that addressing these questions can help scholars understand how legislative careerism affects legislative capacity. The institutional turn in this field, particularly work by Schlesinger (1966) and Polsby (1968), highlights the importance of roles and institutions to career choices. Subsequently, scholars have developed elaborate theoretical and empirical models to elucidate legislators’ career choices. Kousser and MacKenzie review research on term limits to illuminate how attending to the link between institutions and legislative careers can inform research on legislative capacity.

1.4.4  Organization and Rules The next section of the handbook is devoted to understanding the internal organizational structure and rules of legislative organization. Legislative rules are the universe of formal procedures governing the conduct of parliamentary business. In some cases (e.g., Jefferson’s Manual, see this chapter, section 1.3 entitled “The Evolution of Legislative Studies”), they have been surprisingly enduring. In most cases, they structure and constrain the behaviour of legislative actors. Rules emerge to avoid the plenary “bottleneck” problem, to economize proceedings, and to disentangle procedural and substantive issues. Historic cases show that legislative gridlock or regime collapse can happen when rules are absent or badly designed. Müller and Sieberer (Chapter 15) review the state of our knowledge of Western European parliaments with regard to rules governing elections in parliament, agenda-setting and voting, control of the executive, and reaching out to the public. The authors argue that research has been very unevenly distributed across these areas. While much is known about the rules that govern processes leading to authoritative decisions (cabinet inauguration, passage of legislation), much less is known about rules concerning other core functions of legislatures. The number of chambers in national legislatures has been the subject of much theoretically and empirically rich research. Arguably, bicameralism is one of the most common organizational features understood by the general political science profession. Heller and Branduse (Chapter 16) begin by noting that bicameralism is easy to identify but hard to measure. The fact that a constitution specifies two legislative chambers often obscures rather than illuminates the relative influence of the respective chambers, the effects of intercameral negotiations, or the extent to which consideration in a second chamber might alter legislative content. Moreover, studying bicameralism is problematic because, as with most political institutions, its effects emerge from processes that

Introduction   17

often are invisible to observers. Consequently, it can be difficult to identify whether or how bicameralism matters. Heller and Branduse build on previous studies of bicameralism and its effects to propose, first, areas of research that require more careful consideration of bicameralism and, second, an index of bicameralism based on second-chamber powers—that is to say, the extent to which bicameralism should matter. Virtually no democratic national legislature is without at least one committee. Legislative scholars have long sought to understand the origins, design, role, and significance of legislative committees. Today, the conventional notion is that a strong system of committees is a necessary if not sufficient condition for the legislature to operate effectively. Martin (Chapter 17) focuses on the exact reason for committees’ popularity as a form of legislative organization, reviewing the four most prominent theories of legislative organization in the United States: the distributional theory; the informational theory; the cartel-party theory; and the bicameral-conflict theory. Martin argues that examining committees in multiple systems represents an opportunity to reinvigorate the Congressional and comparative study of legislative organization and institutional research more generally.

1.4.5  Parties in the Legislature Along with committees, political parties are the most common and consequential substructure of most democratic legislatures. The next section of our handbook therefore explores the hows and whys of legislative parties, and coalitions of parties. Saalfeld and Strøm (Chapter 18)show that legislative parties are not only virtually ubiquitous, but also highly institutionalized, typically having well-developed structures of leadership as well as specialization. The authors review various explanations of political parties, including functional theories as well as the explanations based on the incentives of individual legislators. They also discuss challenges to these theories such as the influential work of Keith Krehbiel on US Congressional parties. Saalfeld and Strøm argue that it is meaningful to differentiate between various degrees of partyness among parties, and even among coalitions of parties, and review the institutional sources of variations in partyness, including the agenda control available to party leaders and the rewards and sanctions they control. Students of the United States Congress have long debated whether party cohesion is largely a function of leaders’ successful efforts to impose discipline on their followers or merely a product of party members’ shared preferences. Recent comparative contributions to this debate show that legislative institutions can and do supply party leaders with the capacity to enforce discipline even when their members’ disagree, and that discipline is a necessary aspect of party cohesion in any multiparty setting. Reviewing this literature, Kam (Chapter 19) suggests that political scientists should focus their efforts on understanding the connections between party cohesion and discipline, on one hand, and that the politics of candidate and leadership selection, on the other. After all, party leaders exercise their influence over a group of people who have chosen to join the party, whom the party has agreed to accept, and on whose support the leader may rely.

18  Introduction Legislative candidates and incumbent legislators can and sometimes do change party affiliation. Mershon (Chapter 20) highlights how the study of legislative party switching offers important lessons about parties, legislatures, and democratic politics more generally. For example, when we take seriously politicians’ willingness and ability to switch party, we shed fresh light on the wisdom that politicians treat parties as vehicles for winning votes, securing office, and wielding policy influence. An analysis of party switching provides new insights into legislative institutions as constraints on representatives’ behaviour and as resources in the hands of strategic politicians. Under parliamentary democracy, coalition governance requires multiple parties with conflicting policy goals to govern jointly. The need to compromise, coupled with their separate electoral accountability, introduces tensions into multiparty cabinets that are muted in single-party governments. Martin and Vanberg (Chapter 21) review recent scholarship that explores how parties that participate in coalitions make use of legislative institutions to confront these challenges. In particular, the legislative process can allow parties to engage in position-taking to distinguish themselves from their coalition partners, and to scrutinize and amend legislation introduced by ministers associated with other parties. The precise nature of legislative institutions—most importantly, the strength of the committee system—is crucial to these efforts. Martin and Vanberg provide cross-national data on the strength of legislative institutions and discuss the broader implications for the importance of legislatures in parliamentary systems. This chapter, combined with the authors’ earlier work on how legislative institutions play a critical role in facilitating multiparty governance, reminds us that legislatures are central players in the policy processes of law-making as well as oversight of the executive. It is to these themes that we turn next.

1.4.6  Policy-Making and Oversight Legislative agenda control consists in the scheduling of issues and timetable management, the ability to generate proposals, the ability to avoid or block proposals, and, finally, the ability to sequence or order options on the legislative floor. As Rasch (Chapter 22) demonstrates, work on agenda-setting is central to legislative research and the number of studies on the subject is large and rapidly growing, stimulated in part by important theoretical works such as McKelvey’s global cycling and agenda theorems and Romer and Rosenthal’s setter model. Pointing to the importance of procedural rights and various institutional mechanisms, Rasch shows that agenda-setting discretion in real-world assemblies is typically much more restricted than theoretical models would tend to suggest. Rasch argues that the agenda-setter in general seems to be less powerful, and the mechanisms by which the setter gains influence, are often subtle and intricate. One of the ultimate aims of legislative agenda manipulation is to control the enactment of laws. Indeed, law-making is perhaps the defining feature of national legislatures, which, to varying degrees, hold or share authority over what becomes or does not become law. Law-making is particularly complex under presidential government,

Introduction   19

where the separation of powers may add to the complexity of law-making. Such institutional variations might significantly affect the ability of political chief executives to get their legislative proposals enacted. Saiegh (Chapter 23), however, argues that variations in legislative passage rates flow from differences in uncertainty, and not from variation in partisan support. In particular, he identifies two major factors that shape law-making: the unpredictability of legislators’ voting behaviour, and whether buying votes is a feasible option. This explanation of the variation in governments’ legislative performance expands our understanding of law-making, agenda control under uncertainty, and party discipline in legislatures. The study of the effects of legislatures on public finance has become more explicitly comparative since the 1990s. Most research focuses on the budget approval function of legislatures, while there has been less research on the ex-post financial scrutiny in which they may engage. The comparative literature reviewed by Wehner (Chapter 24) highlights substantial differences in legislative budget authority and procedures, and shows that electoral systems and party majorities help explain when formal powers translate into legislative activism. Political economists have shown that powerful legislatures are associated with less disciplined fiscal policy outcomes. Yet, other possible impacts of legislative financial scrutiny, for instance on accountability and democracy, have been asserted but not yet rigorously examined. Moving forward, Wehner suggests that the comparative study of legislative financial scrutiny would benefit from expanding its geographical coverage and time horizon. The field should also investigate a broader array of potential impacts, and examine how these may result from more subtle differences in institutional design. Both law-making and budgetary politics are heavily influenced by interest groups and lobbyists. Congressional lobbying has always been central to American interest group studies. European scholars have traditionally focused more on corporatist structures that allow groups to interact primarily with civil servants. As Skorkjær Binderkrantz (Chapter 25) highlights, the last decades have seen a remarkable shift in attention with legislatures assuming a more central arena for interest groups across Europe—and, consequently, for scholars who study interest groups. There is still much to gain from studying the relations between interest groups and legislatures, including placing the legislature at the centre in studies of interest groups and their political influence, as well as investigating the complex set of (observable or less observable) linkages between interest groups and legislators. It is conventionally believed that foreign policy is very much dominated by the political executive, with legislatures wielding marginal or at best limited influence. Foreign policy is hence said to be determined by political and bureaucratic elites. Raunio (Chapter 26) provides a critical re-evaluation of such arguments, providing evidence that legislatures are definitely not—and have hardly ever been—marginalized in foreign policy. The United States Congress and European legislatures have several mechanisms to influence foreign policy and regional integration, from setting ex-ante limits on executive behaviour to holding the ultimate veto power in treaty ratification. By all accounts, the level of executive drift in foreign policy has decreased over the decades, perhaps in

20  Introduction part through the growing interdependence of domestic and foreign policy agendas. Raunio argues that, as foreign affairs have gained clearer and more varied domestic distributional consequences, legislators also have more incentives to engage themselves in this policy area. In all democracies—presidential or parliamentary—the chain of political delegation is complex where there is a multitude of principals, many of which are frequently replaced. Bureaucrats and politicians hence face common agency problems—i.e., they commonly have to answer to multiple principals. Because of the fractured and temporally unstable nature of democratic leadership, political oversight of the bureaucracy is a particularly problematic link in the process by which citizens control the government. As McCubbins (Chapter 27) notes, these problems have been recognized for a long time. Societies have long sought to create institutions that would strengthen oversight of the executive without simultaneously sacrificing the resoluteness of policy. McCubbins reviews the positive and normative literature on delegation and oversight and discusses the ways in which the contemporary legislative studies are attempting to understand these age-old problems.

1.4.7  Expanding the Scope of Legislative Analysis Studying national legislatures in advanced democracies has a long and venerable tradition, as the themes and research agendas discussed above demonstrate. Increasingly, however, scholars are expanding their horizons—both by studying legislatures in different, “non-traditional” settings or by looking back to the emergence of parliamentarism and parliamentary power in what are now well-established democracies. The European Parliament (EP)—the parliament of the European Union—has transformed from an unelected consultative assembly to a directly elected supranational legislature. The EP looks like a normal national parliament, albeit with stronger committees and somewhat weaker parliamentary political groups (parties) than most national legislatures in Europe. Moreover, its members are elected under different electoral formulas from country to country, providing intriguing opportunities to test theories concerning the effects of electoral institutions. Hix and Høyland (Chapter 28) consider the evolution of the European Parliament in three vital areas: (1) recruitment; (2) the formation of political groups and committees; and, finally, (3) the impact that parties and committees have on member behaviour, noting that the competition between and cohesion within parties has increased in tandem with the increase in the EP’s powers. Across the Atlantic, research on United States state legislatures both challenge and confirm many of the conventional theories of federal legislative behaviour. Downs’ (Chapter 29) analysis of sub-national legislatures charts the development of research on regional, provincial, cantonal, and other local-level assemblies by disaggregating the field into three core areas: recruitment and elections;legislative behaviour; and institutional design.Once treated as nominal institutions, sub-national legislatures are now the subject of studies characterized by increasing theoretical rigor and more frequent

Introduction   21

cross-national comparisons.The extant literature uses sub-national legislatures to test prevailing theoretical explanations of representation, deliberation, law-making, and supervision. It also specifically addresses the effects of the location of sub-national legislatures in systems of multi-level governance with respect to protections against central intrusion, countering executive dominance, and congruence of party alignments with those at national level.Downs points to promising research frontiers, including post-election government formation, the international activities of sub-national institutions, and the use of media technology and e-democracy in sub-state assemblies. Perhaps one of the most promising research frontiers of the last decade has involved the increasing attention shown to legislatures throughout Latin America. Much of the work on legislative politics in Latin America has taken advantage of theories about legislative politics originally developed to study other parts of the world. Crisp and Schibber (Chapter 30) review this scholarship, identifying 31 books and 151 journal articles focused on the study of legislative politics in Latin America. Their review groups articles based on their substantive focus on the electoral connection, on cameral politics, or on legislative-executive relations. Crisp and Schibber highlight how scholars working in the region have stretched theories or made entirely original contributions of their own—as well as amassing impressive amounts of systematic, empirical research. The fall of communism in Central and Eastern Europe and the creation of new institutions of democracy have provided scholars with unique opportunities to observe in real time the origin and early evolution of legislative bodies. Andrews (Chapter 31) reviews scholarship on the strength of legislatures in the new democracies of Central and Eastern Europe, focusing on three important lines of research: (1) the constitutional powers attributed to the legislature versus the executive; (2) the partisan structure of the legislative membership and its implications for policy-making and government stability; and (3) the institutional capacity implied by these legislatures’ internal organizational features. Andrews creates a unique scale of legislative strength based on constitutional powers and finds that, contrary to many accounts, legislatures in Central and Eastern Europe are as powerful as those in Western Europe, even considering the popularity of the dual executive in this newly democratic region. For the legislatures’ powers to be fully realized, however, more time is needed for party systems to stabilize and for the development of more structured legislative organization. Clearly, legislatures are central to modern representative democracy and, as we have seen above, to the process of democratization. But what role, if any, do legislatures play in non-democracies? As Schuler and Malesky (Chapter 32) note, both the number of authoritarian states with active legislatures and the role those bodies play in policy-making have increased dramatically over time. Yet, the costs and risks to autocrats of maintaining a permanent assembly are high, so the growth of such assemblies is puzzling. Leading scholars have offered answers that employ innovative theories and creative methods for analysing some of the world’s most opaque polities.At the same time, others have been critical of this endeavor, characterizing the work as functionalist, and criticizing empirical approaches that have difficulty separating the role of assemblies from other institutions, such as regime parties. Schuler and Malesky review the

22  Introduction history of how institutions have been conceived by authoritarian scholars, before summarizing the theoretical arguments specific to authoritarian legislatures. They then survey the empirical evidence on the preconditions and effects of legislative institutions under autocracy. Many rulers embrace the institutional trappings of democracy—elections and legislatures—only under domestic and international pressure. In the final essay in this volume, Cox (Chapter 33) considers how such reluctant democrats manage their legislatures, arguing that most have re-engineered the budgetary reversion to defang their legislature’s power over the purse. Cox provide examples of budgetary reversions, develops concepts to analyse their differences, and explores some of the institutional and behavioural correlates of executive-favouring reversions—viz., weaker confidence procedures, weaker control of executive decrees, less credible sovereign debt, and more frequent use of extra-constitutional means to attain power. He argues that unless the budgetary reversion point favours the legislature, its powers, and ultimately the prospects for democracy, are fragile. Thus, as much as legislatures are ubiquitous and their functions in advanced societies seem secure, this fascinating chapter reminds us that popular sovereignty exercised through such representative bodies can never be taken for granted.

References Aldrich, J. H., 1995. Why Parties? The Origin and Transformation of Political Parties in America. Chicago: The University of Chicago Press. Alemán, E., 2013. Latin American Legislative Politics: A Survey of Peer-Reviewed Publications in English. Journal of Politics in Latin America, 1: 15–36. Almond G. and Powell, B., 1966. Comparative Politics:  A  Developmental Approach. Boston: Little Brown. Andeweg, R. B., 1992. Executive-Legislative Relations in the Netherlands:  Consecutive and Coexisting Patterns. Legislative Studies Quarterly, 17(2): 161–82. Badura, B. and Reese, J., 1976. Jungparlamentarier in Bonn—ihre Sozialisation im Deutschen Bundestag. Stuttgart: Frommann Holzboog. Bagehot,W., 1867 [1936]. The English Constitution. London: Humphrey Milford. Beer, S. H., 1966. The British Legislature and the Problem of Mobilizing Consent. In E. Frank and R. K. Huitt (eds.). Lawmakers in a Changing World, pp. 30–48. Englewood Cliffs: Prentice-Hall. Best, H. and Cotta, M., 2000. Parliamentary Representatives in Europe 1848-2000: Legislative Recruitment and Careers in Eleven European Countries. Oxford: Oxford University Press. Blomgren, M. and Rozenberg, O. (eds.), 2012. Parliamentary Roles in Modern Legislatures. London: Routledge. Brack, N., Costa, O., and Teixeira, C. P., 2012. Attitudes Towards the Focus and Style of Political Representation Among Belgian, French and Portuguese Parliamentarians. Representation, 48(4): 387–402. Brady, D. W. and Kessler, D. P., 2010. Who Supports Health Reform? PS: Political Science and Politics, 43(1):161–75. Bryce, J., 1921. Modern Democracies. 2 Volumes. London: Macmillan.

Introduction   23

Chee, C. H., 1976. The Role of Parliamentary Politicians in Singapore. Legislative Studies Quarterly,1: 423–41. Clarke, H. D. and Price, R. G., 1981. Parliamentary Experience and Representational Role Orientations in Canada. Legislative Studies Quarterly, 6: 373–90. Coleman, J. S., 1986. Social Theory, Social Research, and a Theory of Action. American Journal of Sociology, 91(6): 1309–35. Converse, P. E. and Pierce, R., 1979. Representative Roles and Legislative Behavior in France. Legislative Studies Quarterly, 4: 525–62. Cotta, M. and Best, H., 2008. Democratic Representation in Europe:  Diversity, Change, and Convergence. Oxford: Oxford University Press. Cox, G. W. 1987. The Efficient Secret. Cambridge: Cambridge University Press. Cox, G. W. and McCubbins, M. D., 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press. Dalton, R. J., 2004. Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies. Oxford: Oxford University Press. Diermeier, D. and Krehbiel, K., 2003. Institutionalism as a Methodology. Journal of Theoretical Politics, 15(2): 123–44. Döring, H. (ed.), 1995. Parliaments and Majority Rule in Western Europe. Frankfurt: Campus Verlag. Eulau, H., 1996. Micro-Macro Dilemmas in Political Science:  Personal Pathways Through Complexity. Norman: University of Oklahoma Press. Fenno, R. F., 1978. Home Style: House Members in Their Districts. Boston: Little, Brown. Fish, M. S. and Kroenig, M., 2009. The Handbook of National Legislatures: A Global Survey. New York: Cambridge University Press. Habermas, J., 1992 [1962]. The Structural Transformation of the Public Sphere: Inquiry into a Category of Bourgeois Society. Oxford: Polity Press. Hagger, M. and Wing, M., 1979. Legislative Roles and Clientele Orientations in the European Parliament. Legislative Studies Quarterly, 4: 165–96. Hatsell, J., 1818 [1712]. Precedents of Proceedings in the House of Commons with Observations. London: Hansard. Huber, J. D., 1996. The Vote of Confidence in Parliamentary Democracies. American Political Science Review, 90(2): 269–82. Jack, M., 2011. Erskine May: Parliamentary Practice. 24th ed. London: Butterworth. Jefferson, T., 1812. A Manual of Parliamentary Practice, Composed Originally for the United Sates Senate, 2nd ed. Washington, DC: Milligan and Cooper (reprinted with an extensive commentary in Constitution, Jefferson’s Manual and Rules of the House of Representatives of the United States. 111th Congress, Washington, DC: US Government Printing Office 2009, 123–328). Jennings, I., 1957. Parliament. 2nd ed., Cambridge: Cambridge University Press. Judge, D. and Ilonszki, G., 1995. Member-Constituency Linkages in the Hungarian Parliament. Legislative Studies Quarterly, 20: 161–76. Katznelson, I., 2011. Historical Approaches to the Study of Congress: Toward a Congressional Vantage on American Political Development. In E. Schickler and F. E. Lee. (eds). The Oxford Handbook of the American Congress, pp. 115–40. Oxford: Oxford University Press. Katznelson, I., and Weingast, B. R. (eds.) 2005. Preferences and Situations: Points of Intersections between Historical and Rational Choice Institutionalism. New York: Russell Sage Foundation.

24  Introduction King, A., 1976. Modes of Executive-Legislative Relations:  Great Britain, France, and West Germany. Legislative Studies Quarterly, 1(1):11–36. Kluxen, K., 1983. Geschichte und Problematik des Parlamentarismus. Frankfurt am Main: Suhrkamp. Krehbiel. K., 1998. Pivotal Politics: A Theory of U.S. Lawmaking. Chicago: The University of Chicago Press. Laver, M. and Shepsle, K. A., 1996. Making and Breaking Governments: Cabinets and Legislatures In Parliamentary Democracies. Cambridge: Cambridge University Press. Lijphart, A., 1984. Democracies:  Patterns of Majoritarian and Consensus Government in Twenty-one Countries. New Haven: Yale University Press. Loewenberg, G., 2011 On Legislatures:  the Puzzle of Representation. Boulder:  Paradigm Publishers. Loewenberg, G., 1967. Parliament in the German Political System. Ithaca: Cornell University Press. Maier, H., Rausch, H., Hübner, E., and Oberreuter, H., 1969. Parlament und Parlamentsreform: Zum Selbstverständnis des fünften Deutschen Bundestages, 2nd ed., München: Vögel 1979. Manow, P., 2010. In the King’s Shadow: The Political Anatomy of Democratic Representation. Oxford: Polity Press. Mansfield, H. C., 1965. Statesmanship and Party Government: A Study of Burke and Bolingbroke. Chicago: University of Chicago Press. March, J. G. and Olsen, J. P., 1984. The New Institutionalism: Organizational Factors in Political Life. American Political Science Review, 78: 734–49. Marschall, S., 2005. Parlamentarismus: Eine Einführung. Baden-Baden: Nomos. Mezey, M. L., 1979. Comparative Legislatures. Durham: Duke University Press. Müller, W. C., Bergman, T., and Strøm, K., 2003. Parliamentary Democracy:  Promise and Problems. In K. Strøm, W. C. Müller, and T. Bergman (eds.) Delegation and Accountability in Parliamentary Democracies, pp. 3–32. Oxford: Oxford University Press. Müller, W. C. and Saalfeld, T. (eds.), 1997. Members of Parliament in Western Europe: Roles and Behaviour. London: Frank Cass. Narin I. and Puri S. L., 1976. Legislators in an Indian State: A Study of Role Images and the Pattern of Constituency Linkages. Legislative Studies Quarterly, 1: 315–30. Norton, P., 2013. Parliament in British Politics. 2nd ed. Basingstoke: Palgrave Macmillan. Oleszek, W. J., 2013. Congressional procedures and the policy process. Washington DC: CQ Press. Ostrogorski, M., 1902. Democracy and the Organization of Political Parties. London: Macmillan. Packenham, R. A., 1970. Legislatures and Political Development. In A. Kornberg and L. D. Musolf (eds.). Legislatures in Developmental Perspective, pp. 521–37. Durham: Duke University Press. Patzelt, W. J., 1993. Abgeordnete und Repräsentation: Amtsverständnis und Wahlkreisarbeit. Pas sau: Wissenschaftsverlag Rothe. Peters, B. G., 1999. Institutional Theory in Political Science:  The ‘New Institutionalism’. London: Pinter. Polsby, N. W., 1968. The Institutionalization of the U.S. House of Representatives. American Political Science Review, 62: 144–68. Polsby, N. legislatures in parliamentary d W., 1974. Political Promises: Essays and Commentary on American Politics. New York: Oxford University Press. Richardson, J. J. and Jordan, A. G., 1979. Governing Under Pressure: The Policy Process in a Post-Parliamentary Democracy. Oxford: Martin Robertson. Riker, W. H., 1962. The Theory of Political Coalitions. New Haven: Yale University Press.

Introduction   25

Schelling, T. C., 1978. Micromotives and Macrobehavior. New York: W. W. Norton. Schlesinger, J. A., 1966. Ambition and Politics. Chicago: Rand MacNally. Schmitt, C., 1988 [1923]. The Crisis of Parliamentary Democracy. Translated by Ellen Kennedy. Cambridge, MA: MIT Press. Schneider, H. P. and Zeh, W. (eds.), 1989. Parlamentsrecht und Parlamentspraxis in der Bundesrepublik Deutschland: Ein Handbuch. Berlin: de Gruyter. Searing, D. D., 1994. Westminster’s World: Understanding Political Roles. Cambridge: Harvard University Press. Shepsle, K. A., 2002. Assessing Comparative Legislative Research. In G. Loewenberg, P. Squire, and D. R. Kiewiet (eds.). Legislatures: Comparative Perspectives on Representative Assemblies, pp. 387–97. Ann Arbor: University of Michigan Press. Sieberer, U., 2011. The Institutional Power of Western European Parliaments: A Multidimensional Analysis. West European Politics, 34(4): 731–54. Strøm, K., 2003. Parliamentary Democracy and Delegation. In K. Strøm, W. C. Müller, and T. Bergman (eds.). Delegation and Accountability in Parliamentary Democracies, pp. 55–106. Oxford: Oxford University Press. Tocqueville, A.  de, 2000 [1835]. Democracy in America. The Complete and Unabridged Volumes I  and II. Translated by Henry Reeve. With Introduction by Joseph Epstein. New York: Bantam. Tsebelis, G. 2002. Veto Players:  How Political Institutions Work. Princeton:  Princeton University Press. Wahlke, J. C., Eulau, H., Buchanan, W., and Ferguson, L. C., 1962. The Legislative System: Explorations in Legislative Behavior. New York: Wiley. Wheare, K. C., 1963. Legislatures. Oxford: Oxford University Press. Wilson, W., 1885. Congressional Government: A Study in American Politics. Boston: Houghton Mifflin. Zürn, M., 2013. Die schwindende Macht der Mehrheiten. Weshalb Legitimationskonflikte in der Demokratie zunehmen werden. WZB-Mitteilungen, No. 139: 10–13.

PA R T  I

T H E OR E T IC A L A P P ROAC H E S I N L E G I SL AT I V E ST U DI E S

C HA P T E R

2

FORMAL MODELS OF L E G I S L AT U R E S DA N I E L DI E R M E I E R

2.1 Introduction Since its very beginning, the study of legislatures has been one of the cornerstones of formal modeling in politics (e.g. Black 1958).1 Formal theory begins with the mathematical representation of a political domain, here legislative decision-making. The first step consists in specifying the possible states of the political domain in question. These states may just consist in the outcomes of a collective choice process, e.g. the policies adopted, or may also contain representation of behaviour by politicians, e.g. who voted for or against a given bill. The formal representation will set aside all unnecessary detail of the political domain. For example, the set of policies may simply be represented as a finite set or an interval. In such representations, e.g. the wording of a bill or its length, are considered unimportant. What is and what is not relevant is, of course, determined by the judgment and intent of the modeler. The formal representation may serve various purposes. Many versions of formal political theory attempt to predict and explain the outcomes of collective decision processes. Others have a systematizing purpose and aim to derive known models from a more general framework. Others are concerned with normative questions, e.g. policy and welfare consequences of certain political institutions. And yet others are methodological in nature, trying to identify problems with a particular formal approach or methodology. Almost all existing formal models of legislative decision-making fall into the rational choice tradition. Rational choice theory contends that any explanation of political phenomena must be grounded in the choices of rational individuals. That is, a complete theory must, at the very least, include a list of actors and a preference profile: a list of transitive and complete preference relations for every actor.2 Variants of rational choice theory then differ in their view on what else is needed to explain a political

30   Daniel Diermeier phenomenon. The importance of these modeling decisions is frequently overlooked by researchers who are not closely engaged with formal theory building, but are crucially important, especially in domains that are intrinsically comparative in nature such as the study of legislatures. The reason lies in the very nature of formal modeling. The discipline imposed by mathematical representations forces the theorists to be very clear about what is assumed and what is not. Moreover, since only very simple models can be solved analytically, the trade-offs of what can and cannot be included in a model are typically quite restrictive. This means that the modeler must repeatedly make decisions about which aspects of the phenomenon should be included and which should be set aside. These constraints, trade-offs, and considerations are importantly shaped by the underlying formal methodology and may reflect different visions of the purpose of formal modeling. In the context of a theory of legislatures, a formal methodology that is suitable for applications in comparative research needs to satisfy at least three criteria3: 1. Universality. The methodology needs to be able to represent a broad variety of institutional differences across legislatures. 2. Empirical content. The formal methodology must have empirical content in many applications. There are two aspects to this criterion: a. The solution concepts associated with a formal methodology must generally admit a solution; they cannot be empty or lack existence. b. The solution concepts must be predictive; they must clearly divide the possible outcomes of the process into those that are consistent with the theory and those that are not. A theory that is consistent with every possible outcome does not have much explanatory power. 3. Equilibrium institutions. In a comparative context, formal models will contain representations of the relevant rules and institutions that shape decision-making. But many of these institutions are themselves objects of choice. Any satisfactory formal methodology must work at all levels of the collective choice process—at the level where policies are determined and at the level where institutions are chosen. It must be able to explain why some institutional features come into existence and persist, while others are either nonexistent or transient. Next, we review the three main formal methodologies used in the theory of legislatures are reviewed, and their respective strengths and weaknesses discussed. We then survey the two main subject domains where formal models have been used: the US Congress (and, to a much lesser extent, non-US legislatures) and coalition government. Next, we discuss some of the attempts to integrate these two domains and create comparative formal models and outline some desirable features of a general theory of legislatures. Many of these topics are vast and active areas of research. Our goal is not to discuss every, or even most of the relevant contributions—an impossible task—but to trace the main developments and debates of this vibrant literature.4 The purpose is to provide a road-map that provides enough structure to locate the various approaches and also to point out some major gaps. Our hope is that it will provide some orientation for any

Formal Models of Legislatures  

31

researcher who wants to explore this sometimes bewildering, and, occasionally, even intimidating research tradition in more detail.

2.2  Abstract Properties of Collective Choice Variants of rational choice models differ by what they add to the common components of any rational choice model (the set of states, actors, and preferences). The oldest general rational choice methodology, social choice theory, pioneered by Arrow (1951), adds one more component to its methodological vocabulary: an aggregation function. Formally, an aggregation rule is simply a function from the set of preference profiles to a single preference relation, which captures collective preferences. In most cases of interest to legislative scholars, all the information contained in the definition of an aggregation function is given by its set of decisive coalitions. For example, in the case of majority rule (with an odd number of legislators), any group at least as large as a majority would constitute a decisive coalition.5 In these cases, all we need to know to derive the outcomes of a collective decision process are three things: the set of possible outcomes, a preference profile, and a list of decisive coalitions. Needless to say, this is a very sparse description of legislative decision-making. For example, any differences among legislators (constituency, electoral rule, party, age, gender, education, professional background, ethnicity, wealth, etc.) are set aside unless they are captured by preference orderings over policies. Similarly, institutional differences such as complex recognition procedures, voting agendas, and committee systems are only captured in as much as they influence the list of decisive coalitions. Note also, that, strictly speaking, social choice theory does not predict behaviour (e.g. the voting behaviour of individual legislators or parties) but only policy outcomes. That is, a given outcome may be consistent with very different voting patterns. Such sparse representation is not necessarily a flaw. On the contrary, it can be a supreme virtue in as much as it leads to general results. What matters is whether the formal methodology allows researchers to capture the relevant details of legislative decision-making. This early age of formal modeling was based on the conviction that only very few components were relevant to understand political decision-making; its goal was to develop a general theory of politics, similar to general equilibrium theory in economics. Indeed, in as much as social choice theory is able to explain legislative decision-making in terms of only a few parameters, it makes a strong case that any additional details of the choice process, e.g. legislative procedures, simply do not matter as much as they may appear to; irrelevant institutional complexity would be reduced to a few parameters that were truly essential (McKelvey 1986). But this approach quickly ran into severe difficulties. Social choice theory is known for its negative results: from Arrow’s Impossibility Theorem (Arrow 1951) to the generic

32   Daniel Diermeier emptiness of the core (Plott 1967; McKelvey 1976, 1979; McKelvey and Schofield 1987). Of course, these results have provided tremendous insights into the nature of political decision-making, but, for an explanatory or predictive purpose, non-existence results are problematic: they violate the second criterion for formal methodologies (2.a listed earlier). The problem is that, methodologically, impossibility results fail to generate predictions in many settings. These results, especially McKelvey (1976), have often been interpreted as “predicting chaos.” But this interpretation is problematic and certainly does not correspond with the mathematical notion of “chaos” used in chaos theory (e.g. Lorenz 1960, 1993; Poincaré 2007 [1908]). The problem with social choice theory is not that the theory predicts dramatic instability, but that in many domains it predicts nothing at all.6 This does not mean that social choice theoretic results are somehow not important; far from it. They investigate the question of how far one can go by just looking at preference profiles and decisive coalitions. The fact that this research program is most famous for its impossibility results establishes that these features are not sufficient to explain political outcomes. Indeed, it is exactly the “negative” results of social choice theory that makes this point most forcefully. Subsequent work in this tradition has tried to address these problems by changing solution concepts, moving from the core, the most important solution concept in social choice theory, to concepts such as the top cycle set (Miller 1977), the uncovered set (Miller 1980), and various others. These attempts face various problems (Diermeier 1997, 2014). Either the set of predicted outcomes is too large to be useful—this is the case for the top cycle (McKelvey 1976, 1979)—or the solution concept implicitly imposes a particular agenda structure violating the ambition of “institution-free properties” (McKelvey 1986). For example, Banks (1984) showed that the uncovered set contains sophisticated voting outcomes (Farquharson 1969) if voting is conducted according to an amendment agenda. But this does not hold for other agenda types, such as elimination agendas (Ordeshook and Schwartz 1987) or two-step agendas (Banks 1989). If voting is conducted according to these agenda types, sophisticated voting outcomes may lie outside the uncovered set. This means that in the context of legislative decision-making alternative solution concepts, such as the uncovered set, implicitly assume a particular, quite specific agenda structure. It does not cover all binary voting agendas, let alone more general agenda structures. But if social choice theoretic approaches implicitly depend on institutional details, one may want to analyse the institutional detail directly. This is the path chosen by the (Neo-)Institutionalists.

2.3  The Institutionalist Alternative The formal analysis of political institutions dates back to very beginning of mathematical modeling in political science, beginning with Black (1958), Downs (1957), Riker (1962), and Farquharson (1969). These approaches concentrated on the analysis of specific institutions such as committees (Black), elections (Downs), coalitions (Riker), and agendas (Farquharson).

Formal Models of Legislatures  

33

Institutionalism, however, is not simply the analysis of institutions. Rather, it constitutes an alternative methodological approach to the social choice theoretic tradition. In its modern form as “new institutionalism” (Shepsle 1979), it has become the dominant research area in formal modeling today. What is common across all versions of institutionalism, is a rejection of the social choice research program of finding institution-freegeneral theorems of collective choice. Instead the goal is to develop formal methods that enable the comparison of political institutions in a structured and systematic manner. The theoretical unity of institutionalism thus does not consist in a hierarchy of ever more general results, but in sets of methods and concepts that make such comparisons possible. To capture such institutional variation, the formal apparatus of institutionalism needs to be enriched beyond the list of decisive coalitions. These additional structural parameters, however, are not intermediate steps on the path to a more general theory, but are necessary to explain political phenomena.7 This view of “institutionalism as a methodology” has been formulated by Diermeier and Krehbiel (2003) as follows:

1. Define and hold fixed behavioural postulates for political actors within the collective choice setting to be studied. 2. Characterize formally the institutions in effect (as defined in step 1). 3. Deduce the behaviour that arises within the institutional setting given the behavioural postulate, and characterize the outcome that results from the behaviour. 4. Compare the derived implications with empirical regularities and data.

Steps 1 and 2 are exogenous within the context of a well-specified institutional theory: they are the assumptions of the theory. They will usually include the set of decisive coalitions, but may contain additional institutional detail. Steps 3 and 4 are endogenous; they are what is derived, predicted, and explained. It is crucial that behavioural postulates remain fixed and consistent within and across studies. Since a key element of institutional analysis is to vary institutional features while keeping the behavioural postulate constant, it is of great importance that the equilibrium concept is applicable in many collective choice problems. Step 3 captures the implications of the model. The model’s implications may only pertain to outcomes (e.g. whether a certain bill will pass) or to behavioural regularities (e.g. which legislator would vote for or against the bill). Step 4 consists of empirical assessments of the predictions of the model. Although issues of testing are beyond the scope of this chapter, the testing of institutional theories is essential in the evaluation of institutionalist models as good explanations for empirical phenomena. Any institutionalist model, however, immediately faces a serious problem. If it is correct that policy outcomes depend on the collective decision-making institutions, then different institutions will yield different outcomes. But this means that rational legislators would anticipate the relationship between institutions and outcomes, and form preferences over institutions. That is, they will prefer one voting agenda over another, compete over committee assignments, or fight over who has the right to propose or

34   Daniel Diermeier amend a bill. Of course, this insight is not new to observers of legislative politics and is frequently expressed by the legislators themselves. Former US House Minority Leader Robert H. Michel (Republican-Illinois) stated this sentiment as follows:8 Procedure hasn’t simply become more important than substance—it has, through a strange alchemy, become the substance of our deliberations. Who rules House procedures rules the House.

This issue is particularly pressing if legislatures can fully determine their own organizational structure by the same voting rule (e.g. simple majority rule) that is used to pass bills, as is the case in the US Congress and many other legislatures.9 The question then becomes how these institutions can function as constraints on legislators at all. Riker expressed this point as follows (1980, 445): In that sense rules or institutions are just more alternatives in the policy space and the status quo of one set of rules can be supplanted with another set of rules. . . . If institutions are congealed tastes and if tastes lack equilibrium, then also do institutions, except for short-run events.

This argument points to an important methodological requirement for institutionalist models: a satisfactory formal methodology must work at all levels of the collective choice process—at the level where policies are determined and at the level where institutions are chosen. The goal of a theory of institutions is to explain why some institutional features come into existence, and why they persist. In other words, institutional features that were assumed to be exogenous, e.g. a committee system, need to be justified as equilibria in some underlying model of institutional choice; the institutions that constitute institutional equilibria must be equilibrium institutions as well (Shepsle 1986). It was this concern that led to our methodological criterion No. 3 (equilibrium institutions). To satisfy this requirement, a formal methodology must be applicable at every stage of institutionalist choice. As we will see in the next section, this turns out to create difficulties for certain variants of institutionalism.

2.4  Variants of Institutionalism 2.4.1  Structure-Induced Equilibria Historically, the first version of formal institutionalism was the Structure-Induced Equilibrium (SIE) approach (Shepsle 1979; Shepsle and Weingast 1981). The success of the SIE approach was largely due to the fact that it combined the powerful tools of formal analysis with a focus on institutional detail, especially in the study of US Congress. SIE models were able to formalize and analyse phenomena, such as the committee system, that had already been identified as important features of congressional decision-making.

Formal Models of Legislatures  

35

In his seminal paper, Shepsle (1979) formally introduced institutional structure as a committee system which assigns each individual to exactly one issue. Intuitively, a committee has exclusive jurisdiction on a given issue and determines the outcome by majority rule. Given convex and continuous preferences, the majority core is non-empty for each issue. The explanatory concept of the political system then consists in those outcomes that are in the issue-by-issue core. Using a fixed-point argument by Kramer (1972), one can then show that for any committee system, the issue-by-issue core is non-empty. Notice that in contrast withthe abstract social choice theoretic results discussed above, the theoretical apparatus of SIE contains an additional argument—an assignment of individuals to issues or dimensions, which ensures that the model always creates a prediction. While this may at first seem like a minor departure from social choice theory, it marks a rejection of the reductive research program that underlies it. Rather than searching for general explanations that apply in many institutional settings, the model is intended to capture a particular institutional arrangement, here the congressional committee system. A different application, say of cabinet decision-making, for example, would require a different institutional structure. Indeed, such a model was later proposed by Laver and Shepsle (1990). While SIE models avoid the non-existence problem of social choice theoretic approaches, they fall prey to the institutional choice problem identified by Riker (1980). This was shown in a critique of the Laver and Shepsle (1990) model of cabinet decision-making by Austen-Smith and Banks (1990). In their model, Austen-Smith and Banks (1990) ask the following question: Do the assignments of cabinet portfolios to parties constitute institutional equilibria? Formally, are the jurisdictional assignments themselves in the (majority) core of a model of institutional choice? Austen-Smith and Banks show that if there are three parties and all parties have Euclidean preferences, then a portfolio core exists. Intuitively, this means that under these conditions, we can have “stable” portfolio assignment. More precisely, there exist assignments of portfolios to parties that cannot be beaten by any other assignment via majority rule. Unfortunately, none of the conditions can be relaxed. That is, one can easily construct counter-examples for four parties or for non-Euclidean preferences, where a core does not exist. In other words, other than in the knife-edge case of three parties and Euclidean preferences, the portfolio core may be empty. Thus, the structure-induced equilibrium approach cannot escape the Riker critique:  the institutions that secure institutional equilibria do not generally constitute equilibrium institutions.

2.4.2  Non-Cooperative Game Theory Inthe mid-1980s, a new approach emerged in the study of politics that used a very different methodology: non-cooperative game theory. Like Structure-Induced Equilibrium theory, this approach rejects the reductive research program of social choice theory and, instead, focuses on the institutional details of collective choice. But in contrast to Structure-Induced Equilibrium theory, it completely abandoned the core

36   Daniel Diermeier or other cooperative solution concepts. Political interaction was now modeled as a non-cooperative game. Given the known sensitivity of game-theoretic analysis to the details of the game form, this approach, from the very beginning, precludes any notion of an “institution-free” theory. Rather, the sensitivity of institutional detail was viewed as a strength that allows the varying of institutional details in a systematic way while keeping the explanatory concept fixed (Myerson 1996). The radical departure from existing models can be seen in one of the first and most influential applications of game-theory to legislative decision-making:  the agenda-setting model of Romer and Rosenthal (1978). Originally, developed to model school-board referenda, it quickly became one of the central building blocks of formal models of legislative decision-making (Denzau and Mackay 1983). In the basic model, an agenda-setter makes a policy proposal, which is then pitted against a default alternative in an up-or-down majority vote. This approach captures a prominent feature in legislative decision-making: there is typically some agent (a chair, committee, party leader, minister, and so forth) that effectively holds agenda control, yet that agent’s power is nonetheless checked by the requirement of majority approval. The model yielded two fundamental insights. First, the existence of the “power to propose” provides an agenda-setter with the ability to bias policy outcomes in his favour even in the case where a median voter exists, as the agenda-setter can make a policy proposal to satisfy a bare majority of the “cheapest” voters necessary to ensure approval and move the outcome away from the median.10 Second, the policy outcome not only depends on voters’ preferences but also on the location of the default policy, which defines the reservation utility of the voters. The Romer-Rosenthal model was originally developed in a uni-dimensional policy space, where a median voter exists. However, that does not mean that it is restricted to a uni-dimensional policy space. Indeed, it can be applied in multidimensional policy environments and general preference configurations without difficulty. In such policy environments, the preferences of the decisive coalitions define so-called “win-sets” and proposers simply pick their most preferred point within the win-set. But the application to uni-dimensional case serves an important theoretical purpose. The model implies that even if a median exists, the predicted policy outcome will (typically) not be the median. This insight constitutes the fundamental difference between non-cooperative models and the early formal traditions that emerged from social choice theory. As we will see later, this difference has shaped the theoretical discussions in formal modeling of legislatures until today. It is at the heart of the most important debates on the internal organization of the US Congress, the proper models of coalition governments, decision-making in the European Union, and many others. The Romer-Rosenthal model is a very stylized model of legislative decision-making. In particular, it makes three main assumptions:

1. An exogenous status quo. 2. Decision-making on a given issue ends when a majority approves a proposal. 3. Complete information.

Formal Models of Legislatures  

37

Each one of these three assumptions was subsequently relaxed, thereby generating extensive sub-literatures. The assumption of an exogenous status is potentially problematic, as it determines the extent of possible policy bias. By far the most influential model without an exogenous status quo is the Baron-Ferejohn model (1989a).11 In all variants of the Baron-Ferejohn model, a proposer is selected according to a known rule (usually modeled as a probability distribution). The proposer then proposes a policy or an allocation of benefits to a group of voters. According to a given voting rule (usually majority rule), the proposal is either accepted or rejected. If the proposal is accepted, the game ends and all actors receive pay-offs as specified by the accepted proposal. Otherwise, another proposer is selected, etc.12 The process continues until a proposal is accepted or the game ends. In many applications, the game continues indefinitely until a proposal is accepted. In equilibrium, proposers select the “cheapest coalition,” i.e. the players with the lowest reservation values that, if added to the proposer, constitute a winning coalition (proposers randomize if there are multiple coalitions), and then keep the remainder of the pay-off for themselves. The ability to extract additional surplus rests in the power to propose (Baron and Ferejohn 1989b). Instead of an exogenously given status quo, as in the Romer-Rosenthal model, proposers consider each player’s continuation value, i.e. the expected equilibrium pay-off if bargaining continues. The winning coalition consists of the non-proposing legislators most eager to be included. Intuitively, proposers are able to exploit the impatience of non-proposing legislators. This impatience not only follows from the discounting of future pay-offs, but also includes the concerns of being out of any future coalition (Eraslan and Merlo 2002). Note that the Baron-Ferejohn model (as well as the Romer-Rosenthal model) applies to multidimensional environments, including the paradigmatic “divide-the-dollar” framework, where agents need to collectively decide on how to divide a fixed, transferable benefit (a “dollar”). The divide-the-dollar problem is the classical formulation of a collective decision problem with an empty core. Yet, this fact is not even mentioned in the seminal Baron-Ferejohn paper (Baron and Ferejohn 1989a). Rather, the approach is presented as a free-standing, self-sufficient, formal methodology capable of modeling complex institutions. Both Romer-Rosenthal and Baron-Ferejohn models assume that once a majority approves of a proposal, the game ends. This assumption is appropriate in many domains, e.g. budgetary decision-making, but it needs to be modified in some important policy domains where default policies persist until they are changed. Examples include entitlement programs such as social security, allocation of property rights, and regulatory regimes. In a dynamic policy environment, the passage of a bill does not prevent the legislature from coming back to the same policy issue at a later date. Rather, the passage of a bill merely sets the default for subsequent rounds of policy-making. Baron (1996) was the first to explore such a model in the context of a uni-dimensional policy space. Other approaches include Kalandrakis (2004), Bernheim, Rangel, and Rayo (2006), and Diermeier and Fong (2011). These papers have yielded important and sometimes surprising insights, but they also highlighted some of mathematical

38   Daniel Diermeier difficulties of solving sequential bargainingmodels. It is an area with many open questions including a complete analysis of multidimensional choice environments for general proposal rules. The third variation introduced incomplete information into the Romer-Rosenthal framework. By far the most influential approach is due to Gilligan and Krehbiel. In a series of papers, Gilligan and Krehbiel (1987, 1989a, 1989b, 1990) present an informational rationale for granting committees special proposal and amendment prerogatives. The idea is that legislators are usually unsure about the policy consequences of a bill. Risk-averse legislators then benefit from reducing such policy uncertainty. To reduce such uncertainty, legislators can “specialize,” i.e. acquire private information, about the policy consequences of a bill. Specialization is costly, but provides the legislators with private information that is useful to the legislature as a whole. Gilligan and Krehbiel model the interaction between the committee (“the proposer”) and the floor as a signaling game. In the case where the committee has no procedural prerogatives (i.e. where the floor can amend the committee’s proposal without restrictions), the committee proposal is merely cheap talk and does not convey any credible information to the floor. Since that is the case, a committee will have no incentive to acquire costly information about policy consequences. If, however, the floor grants the committee procedural prerogative (e.g. a “closed rule” where no amendments are permitted), the committee can credibly transmit its private information in equilibrium. This allows the committee to bias the policy outcome in its direction, but, for a broad set of parameter values, the cost of policy bias is smaller than the benefits from specialization. Hence, the chamber grants the committee proposal prerogatives, the committee specializes and credibly transmits its information to the floor, leading to a biased policy outcome. When discussing formal models of legislatures, three desiderata for a formal methodology were identified: (1) universality (i.e. the ability to capture relevant institutional detail); (2) empirical content (i.e. existence and sharpness of the solution concept); and (3) equilibrium institutions (i.e. institutions must constitute equilibria at the level of institutional choice). Universality is not a problem for non-cooperative models. Game forms are very general and flexible frameworks to capture institutional detail. Indeed, one weakness of non-cooperative game theory is that its models are too detailed. And the results may depend too much on these details; who moves when, what are the countermoves, and so forth. Existence similarly is not a significant problem. Nash Equilibria exist under very broad conditions. Indeed, the typical reaction to a non-cooperative model where a Nash Equilbrium does not exist is not that this lack of existence points to a deep theoretical problem, but that the model was poorly designed. Rather, the Achilles’ heel of non-cooperative models is their lack of explanatory power: game theoretic analysis does not necessarily lead to sharp predictions. Multiple equilibria and outcomes are a feature of many games. And in some cases, e.g. the famous “Folk Theorems” for repeated games, all individually rational outcomes can be supported in equilibrium.

Formal Models of Legislatures  

39

Multiple equilibria already occur in simple voting games. Consider the case of three players with two alternatives, x and y, where each player strictly prefers x over y. If the winning alternative is chosen by majority rule, a unanimous vote for y is still a Nash Equilibrium, as no player has an incentive to deviate from the prescribed strategy “play y.” To avoid these problems, formal modelers have added additional requirements to Nash Equilibria. Such so-called “refinements” come in various forms. In legislative models, the two most popular refinements are to iteratively exclude weakly dominated strategies (this takes care of the problem of voting equilibria that are independent of legislators preferences) and stationarity (this precludes the Folk Theorem). These assumptions have effectively become standard modeling practice. That said, they are, strictly speaking, additional assumptions that require additional arguments. In the area of incomplete information games, these issues are considerably less settled. This is particularly true in cheap-talk games, as in Gilligan and Krehbiel’s (1987) open rule case. Whatever refinements are used, however, they must be commonly applied across many situations. Otherwise, it would be impossible to disentangle the effect of different institutions from assumptions about refinements. What about equilibrium institutions and the Riker critique? What happens if the choices of institutions, such as gatekeeping power or committee assignments, are endogenized? Thanks to the power of Nash Equilibrium, an answer can easily be given. It was first formulated by Gilligan and Krehbiel (1987) as the question of optimal institutional choice, then broadened by Calvert (1995). The general strategy is straightforward. First, the game form is extended such that actors can choose institutions. Each institution then corresponds to a subgame. Second, the equilibria to these subgames are characterized. This allows one to derive (induced) preferences over the corresponding institutions and solve for the equilibria in the extended game of institutional choice. In the legislative context, this means that institutions, such as a seniority system or the existence of powerful committees, that were originally modeled as features of the game form, can now be interpreted as particular equilibria in an underlying game of institutional choice (e.g. McKelvey and Riezman 1992; Diermeier 1995; Diermeier and Myerson 1999). Of course, this game of institutional choice itself is governed by (higher-order)institutions, such as the voting or proposal rules used for deciding on procedure. But again, one can push the analysis one step further and interpret voting rules or even constitutions as equilibria in some more fundamental game. This approach has been used with great promise in the study of coalition government (Lupia and Strøm, 1995; Diermeier and Merlo, 2000), the internal organization of the US Congress (McKelvey and Riezman 1992; Diermeier and Vlaicu 2011a; Eguia and Shepsle 2012), and the comparison between parliamentary and presidential democracies (Diermeier and Feddersen 1998). To understand these issues in more detail, we now turn to the two main domains of formal theory in legislatures: the US Congress and coalition government in multiparty democracies.

40   Daniel Diermeier

2.5  Formal Modeling in Action: Committees and the Internal Organization of the US Congress Legislative voting rules are, with few exceptions, anonymous and egalitarian. It does not matter who votes for a proposal, and each vote counts equally. Yet, actual legislatures are characterized by a variety of organizational structures that lead to significant differences in the influence of individual members on policy choices. Examples include committee chairmen, party leaders, and speakers. These institutions, and the norms that support them, are rarely mandated by a country’s constitution. Rather, legislative majorities choose to setup and maintain their internal organizational structure. This leads to two questions. First, what are the consequences of these institutions? Second, why are they stable and persistent? The prime example of this phenomenon is, of course, the United States Congress, with its elaborate system of powerful committees and norms. Formal theories of the internal organization of legislatures were largely developed in the context of the United States Congress, especially the congressional committee system. Structure-induced equilibrium theory, in particular, was used to provide a distributive explanation of powerful committees (Shepsle 1979; Shepsle and Weingast 1981; Shepsle and Weingast 1987; Weingast and Marshall 1988).13 Procedural prerogatives of committees, such as gatekeeping rules, are interpreted as commitment devices that enforce and maintain distributive agreements between committee members that deeply care about a particular policy dimension. The models showed that these agreements, while beneficial for the committees, may lead to inefficient outcomes for the legislature as a whole. But this raises the question why such arrangements are stable, since, as we saw in our discussion of structure-induced equilbria, there typically exist alternative committee structures that would be majority preferred to any current one. The main alternative to the distributive approach is the informational theory of committees (Gilligan and Krehbiel 1987, 1990, 1997) discussed earlier.14 In this theory, the legislature is willing to commit to granting the committee restrictive amendment procedures because this induces the committee to acquire and share information. But, as in the case of distributive approaches, the argument crucially depends on an assumption of “procedural commitment” (Krehbiel 1991). Without procedural commitment, a majority grants the committee a restrictive procedure only if it yields a policy outcome similar to that occurring under an open rule (Gilligan and Krehbiel 1989a, 1989b; Banks 1999). But then the committee has no incentive to specialize in the first place. This leads to the question of how the legislature is able to procedurally commit to not renege on restrictive procedures.15 A comparatively small, usually empirically oriented literature has tried to apply these approaches in other legislatures (e.g. Huber 1996; Döring and Hallerberg 2004).16 Many other legislatures also exhibit complex committee systems, among them Germany,

Formal Models of Legislatures  

41

Italy, Japan, Norway, and Sweden (e.g. Rogowski 1990). However, the details of internal organization differ widely. While individual committees are comparatively strong in Congress, they are weak in most parliamentary bodies (Lees and Shaw 1979; Mattson and Strøm 1995). More importantly, legislative functions that are exercised by committees in Congress may be fulfilled by other actors in a different setting (see Martin, this volume). In parliamentary democracies, for instance, ministries rather than committees acquire policy expertise, and they are the institutions where bills are drafted and markedup. In other words, while US Congressional committees can be interpreted as the proposers and gatekeepers in the legislative process, the same cannot be said about parliamentary committees. That role is primarily occupied by the cabinet, an insight already pointed out by Bagehot (1963 [1867], 66–7): “By [the cabinet] we mean a committee of the legislative body selected to be the executive body.” Bagehot’s dictum creates some new problems for a comparative theory of legislatures as it needs to be expanded to include the study of cabinets, which in the typical case of multiparty democracies is intimately connected with the study of coalition government.

2.6  Coalition Government It is the distinctive characteristic of parliamentary democracies that the executive derives it smandate from and is politically responsible to the legislature. This has two consequences. First, unless one party wins a majority of seats, a rare case in electoral systems under proportional representation, the government is not determined by an election alone, but is the result of an elaborate bargaining process among the parties represented in the parliament. Second, parliamentary governments may lose the confidence of the parliament at any time, which leads to their immediate termination. Thus, historically, two questions have dominated the study of coalition government: Which governments will form? And how long will they last? Much of the early formal research on coalition government was cooperative in nature.17 The focus of this literature was on which parties ended up in the governments and how they allocated ministries among the members of the coalition. All this changed in the late 1980s and early 1990s, which witnessed a dramatic shift towards institutionalist models, paralleling earlier developments in the study of the US Congress. In 1989, Baron published the first applications of non-cooperative game theory (the Baron-Ferejohn model) to the study of governmental coalitions. A year later, Laver and Shepsle (1990) used structure-induced equilibrium models to study the formation and stability of cabinets. This institutionalist shift was significantly advanced by two empirical contributions: Kaare Strøm’s work on minority cabinets (1985, 1990) and Browne’s et al. study of cabinet stability (1984, 1986, 1988). Strøm’s work was important in various respects. First, it focused on a puzzling, but prevalent case of coalition governments—minority governments, i.e. cabinets where the parties that occupy portfolios together do not control a

42   Daniel Diermeier majority of seats in the legislature. As Strøm showed, their existence was neither rare nor occurred only during a crisis phenomenon. Moreover, many of the minority governments were quite stable, though less stable than majority coalitions. In addition, many countries (e.g. Italy) were usually governed by surplus coalition, i.e. governments where certain parties were expendable without losing a chamber majority. The existence of minority and surplus coalitions was inconsistent with most existing theories of coalition formation and reoriented the theoretical literature. Importantly, it refocused the government formation question from “which government will form?” to “which type of government will form?” The second empirical contribution was the study of cabinet stability pioneered by Browne et al. (1984, 1986, 1988) and then further advanced by Paul Warwick and others (King et al. 1990; Warwick 1992a, 1992b, 1992c, 1994; Warwick and Easton 1992; Diermeier and Stevenson 1999, 2000). This literature not only focused attention on the issue of cabinet stability but identified a list of institutional features that correlated with cabinet stability. This line of research provided rich new empirical regularities. But a compelling theoretical framework was lacking. Competing candidates for such a framework, however, were being developed independently using different methodological approaches in their analysis of coalition formation: the Laver-Shepsle model and Baron’s sequential bargaining models. Both contributions put the role of institutions front and centre.18 The point of this new approach was not that political institutions had been overlooked as a potentially important object of study. Rather, both approaches fully embraced the institutionalist credo: a formal theory of politics is a theory of institutions. This shift in perspective has had a profound effect on the subsequent development of formal models of coalition government. In a highly influential paper, Laver and Shepsle (1990) apply this methodology to the study of cabinets where portfolios are interpreted as issue assignments as in Shepsle (1979). In contrast to the application to the US Congress, where the issue median was decisive, here ministers can unilaterally determine the policy-choice in their portfolio. Laver and Shepsle then ask under what circumstances such portfolio assignments are stable; i.e. whether they are in the majority core of a voting game on cabinet assignments. Importantly, this may include minority governments. The model could also be applied to the issue of cabinet stability, e.g. by considering shifts in the policy position of the parties represented in parliament.19 Laver and Shepsle’s work proved highly influential, in part due to its institutionalist methodology that linked the study of the US Congress with the study of coalitions, in part due to its ability to provide institutionalist explanations for the previously identified empirical regularities. As we have discussed earlier, however, Austen-Smith and Banks (1990) identified some methodological weaknesses of the Laver-Shepsle approach. These difficulties led to the development of an alternative modeling approach—the use of sequential bargaining models in the tradition of Baron and Ferejohn. Due to their reliance on non-cooperative game theory, sequential bargaining models avoid the non-existence problems associated with structure-induced equilibrium, even in environments (e.g. dividing a fixed benefit under majority rule) where the core is empty.

Formal Models of Legislatures  

43

These features make the model very suitable for institutional analysis. Compared to the Laver-Shepsle model the Baron-Ferejohn model, however, is much more difficult to work with, especially if one leaves the purely distributive environment and considers (multidimensional) policy preferences. Finally, the first applications of the Baron-Ferejohn model (e.g. Baron 1989a, 1991b) could not account for minority governments and were not designed to address issues of government stability. After all, in the classic Baron-Ferejohn model, the game ends when an agreement is reached. These difficulties lead to the development of alternative approaches. These approaches were also formulated in the language of non-cooperative game theory, but relied on a different bargaining protocol. For our purposes, the most relevant approach was the “efficient bargaining approach” (Merlo and Wilson 1995; Merlo 1997; Diermeier and Merlo 2000; Baron and Diermeier 2001; Diermeier, Eraslan, and Merlo 2003). In these models, actors bargain under unanimity rule until agreement is reached or the game ends. Originally developed in pure game theory, the approach became influential once it was recognized that such bargaining protocols could capture an old idea from coalition theory: the concept of a proto-coalition (Axelrod 1970). That is, coalition bargaining processes are often characterized by multiple government formation attempts, where a party leader (usually called a “formateur”) selects a proto-coalition that will negotiate to form a new government. This bargaining process continues until agreement is reached or the parties abandon the coalition formation process. Bargaining within a proto-coalition is then modeled using efficient bargaining theory. Efficient bargaining models have many appealing properties. Bargaining outcomes are always efficient even in cases with randomly changing payoffs (Merlo and Wilson 1995)  or a multi-dimensional policy space (Baron and Diermeier 2001). But more importantly, efficient bargaining models satisfy a separation property: the outcome of the bargaining game is (at least partially) independent of the identity of the proposer and the exact bargaining protocol. For example, in the Merlo-Wilson model, the set of states where parties reach agreement must be independent of the proposer’s identity. In the Baron-Diermeier model, the policy location agreed upon by the parties does not depend on the details of the bargaining process such as recognition probabilities or discount factors but only on the ideal points of the members of the proto-coalition. This is a crucial advantage as it avoids some of the technical difficulties of the Baron-Ferejohn bargaining protocol. As a consequence, efficient bargaining models could then be applied to more complex environments such as full equilibrium models that combine cabinet formation with voting under proportional representation with a static (Baron and Diermeier 2001) or dynamic policy environment (Baron, Diermeier, and Fong 2012). Most importantly, efficient bargaining models can address the original questions posed by the empirical research: government-type (especially the occurrence of minority and surplus majority) and cabinet stability. The approach was first formulated in a paper by Lupia and Strøm (1995); cabinets were interpreted as equilibria of an underlying bargaining process that must be sustained over time as the political environment and thus the distribution of bargaining power among governing parties changes. The long-term impact of the Lupia-Strøm model was mainly conceptual. The actual

44   Daniel Diermeier bargaining model was too simple to capture dynamics of coalition bargaining.20 What was needed was a fully dynamic model to capture the changing cabinet bargaining environment. In a first step to provide such a model, Diermeier and Merlo (2000) embeded the Baron-Diermeier efficient bargaining model in a multi-stage game with random shocks to both electoral prospects and to the parties’ reservation values. Such shocks may lead to the termination of government before the end of the game. In that case, a new cabinet would take office. On the other hand, cabinets may avoid termination by redistributing (“reshuffling”) distributive benefits among the incumbent parties in response to external events. Diermeier and Merlo show that both minority and super-majority cabinets can occur as equilibrium phenomena. Moreover, minority and surplus coalitions are not rare exceptions, but may form for all parameter values. However, they are less stable than minimal winning coalition which balance smaller short-term pay-offs with higher stability. Depending on the parameters of the model, formateurs may thus choose any cabinet type. For some parameter values, minimal winning cabinets are the best deal. For other parameter values, surplus or minority coalitions are optimal. So all observed cabinet types can be recovered as equilibrium phenomena. One important consequence of the Diermeier and Merlo framework is that formateurs anticipate the benefits and costs of maintaining a given cabinet over time. Both costs and benefits depend on the cabinet type. So, depending on the parameters, formateurs will choose different cabinet types, including, sometimes, minority cabinets that are likely to terminate early. The stability and the relative occurrence of different types of governments are thus jointly determined in equilibrium. This insight has important consequences for the empirical study of cabinet stability, which had been dominated by regression models, i.e. a suitably defined stochastic process with sets of covariates that influence termination probabilities. Since Strøm (1985), all these models have contained institutional characteristics, both of the polity (e.g. constitutional features) and of the cabinet (e.g. its majority status). But once we model cabinets as equilibria, expectations about cabinet duration influence the choice of initial cabinets; cabinets and their durations are jointly selected in equilibrium. Empirically, this implies that we cannot treat cabinet-specific features as proper independent variables in a regression model. Recent papers (Merlo 1997; Diermeier, Eraslan, and Merlo 2003) have tried to resolve these problems by using a structural estimation approach. The approach consists in specifying a bargaining model of government formation, estimating the model’s parameters, assessing the ability of the model to account for key features of the data, and then using the estimated structural model to conduct (counterfactual) experiments of comparative institutional analysis. The theoretical developments in the study of coalition government had important consequences for a general theory of legislatures. Firstly, the models demonstrated how the methodologies developed in the context of the US Congress can be applied to non-US legislatures and capture coalition government, the crucial phenomenon in multiparty parliamentary democracies. Secondly, the existing empirical research on cabinet stability and cabinet types put the question of institutional choice and

Formal Models of Legislatures  

45

stability front and centre. While in models of the US Congress it was not unreasonable to assume that, in the short-run, the institutional structure (e.g. committee system, seniority norms, etc.) could be considered fixed, such an assumption was simply inappropriate in the study of governing coalitions. But once cabinets are viewed as legislative institutions (as Bagehot argued) that must be sustained by majority support, one can ask the same question about powerful committees and legislative leadership structures. This question is not only important for comparativists, but lies at the centre of some of the crucial debate among the leading scholars of the US Congress. Being a member of a powerful committee or of the cabinet bestows significant procedural advantages to certain legislators. The question then becomes why a chamber majority bestows such prerogatives to some of its members and why such agreements are stable. In the context of the debate over parties in the US Congress, Krehbiel (2004, 117)has put this argument as follows21: Majoritarianism and procedural endogeneity imply that the median voter is pivotal in a procedurally neutral setting. Why would such a legislative voter consent to procedures whose foreseeable consequences are tantamount to an abdication of power?

This argument not only applies to the US Congress, but to any legislature. How do we explain the variety and stability of legislative institutions if such institutions have important distributional and policy consequences? Thirdly, the review of the literature also pointed out the significant technical difficulties in creating formal models that can address these problems. Social choice-based models are elegant, but cannot capture sufficient institutional detail or model dynamics, and face severe existence problems of its solution concepts. Structure-Induced Equilibrium models can capture more institutional detail, but inherit some of the existence problems of social choice theory, and dynamic considerations can only be captured as shocks to the system. Non-cooperative models can capture rich institutional structures as well as dynamics, but are difficult to use.22 It is therefore not surprising that a general theoretical framework to study legislatures is still lacking. In the final section we discuss some of the existing approaches and conclude with some goals for future research.

2.7  A Comparative Perspective Existing formal research of legislatures has largely progressed along geographical divisions. The extensive research on the US Congress has developed in isolation from the equally lively research on coalition government in parliamentary democracies. While some brave minds have tried to cross the divide between the two research traditions

46   Daniel Diermeier (e.g. Baron 1989; Laver and Shepsle 1990) a systematic formal approach that can accommodate not only the two main traditions, but also can accommodate any legislative systems is still lacking. Currently, there are two approaches that can serve as candidates for such a paradigm: the veto players approach (Tsebelis 1995, 2002; Tsebelis and Money 1997) and institutional equilibria (Calvert 1995; Diermeier and Feddersen 1998; Diermeier and Myerson 1999; Diermeier and Krehbiel 2003). Tsebelis (2002) proposed a formal analysis of legislative institutions based on the concept of veto players. Its purpose is to provide a framework that can represent various political systems in a unified fashion and then use that representation to study policy consequences, especially with respect to policy stability. A veto player is an individual or collective actor, such as the House of Representatives, whose agreement is necessary for a policy change. That is, a veto player has the power to block a change from the status quo policy. Policy change can occur when there are policies preferred to the status quo by all veto players. These points constitute the win set of the status quo; the larger the win set, the higher the potential to enact another policy, the lower the stability of the status quo. Thus, by identifying the veto players and the associated win set and comparing them across polities, we can make inferences about the relative stability or, conversely, the opportunity for policy change. While in principle, adding veto players will increase policy stability, this will not hold if veto players are redundant, i.e. if the win set does not change as a veto player is added. One of the attractive features of this approach is that it imposes a common structure on political institutions that facilitates comparative analysis and can create new insights. For example, both a supreme court that can nullify legislation and the IMF that can block a certain fiscal policy can be modeled as veto players. Similarly, when comparing constitutions, e.g. presidential and parliamentary polities, all that matters are the resulting veto players. This means that a parliamentary system with an independent supreme court may behave similarly to a presidential one where the president can veto legislation. Methodologically, this approach falls into the social-choice theoretic tradition. That is, the formal representation of legislative decision-making consists of three elements: the set of possible outcomes (here, a Euclidean, multidimensional policy space); a preference profile (here, quadratic preferences); and a list of decisive coalitions (here, the list of individual and collective veto players). It thus inherits all the difficulties of this approach and the associated solutions concepts such as the uncovered set. This is particularly true of the concept of “collective veto players.” In addition, the restriction to social-choice theoretic concepts limits the scope of institutional detail that can be represented formally. For example, while Tsebelis repeatedly discusses the importance of agenda-setting, his framework does not allow a formal representation of agenda-setting power. Hence, any discussion of proposal power and agenda-setting must remain informal. Tsebelis utilizes various other solution concepts (the uncovered set, the “yolk,” the win-circle) but, as he acknowledges, most of the problems remain and theoretical propositions cannot be obtained. What we are left with are “regularities,”23 that seem to hold in most preference configurations, but not in general. Finally, the approach cannot account for the stability and equilibrium properties of veto players. For example, both

Formal Models of Legislatures  

47

an independent regulatory commission and a legislative committee can serve as equivalent veto players, but, in a bicameral system, the committee can be stripped of its powers by a simple chamber majority, while a regulatory commission can only be abandoned if both chambers (and, if applicable, the president) concur. More generally, within its formal framework the veto player approach cannot distinguish between endogenous and exogenous veto players: which veto players are to be considered as constraints and which should be viewed as institutional equilibria in an underlying game of institutional choice. The concept of institutional equilibria is central to the main alternative to the veto player approach:  the equilibrium institution approach. We have seen above how the approach has been applied in isolation to both the US Congress and study of coalition government. What has been missing, however, has been a genuine comparative point of view that can account for various institutional structures across countries. To accomplish this goal, the concept of institutional equilibria is critical. To see this, recall that (a), cabinets are the critical policy making institutions in parliamentary democracies, (b)  cabinets need the continual support of the legislature. In the language of non-cooperative game theory, they must constitute (institutional) equilibria in an underlying game of coalition bargaining. As discussed, both the solution concepts associated with social-choice theory and the structure-induced equilibrium theory approach are often empty. Hence, there may not be any stable institutions in a model of institutional choice. It is here where non-cooperative game theory provides a crucial methodological advantage. To see why, note the problem of institutional choice defines a hierarchy of institutional models where the institutions of each lower level become objects of choice at a higher level. Diermeier and Krehbiel (2003) state this as follows:

1. Define and hold fixed behavioural postulates for political actors in collective choice process in which well-defined institutions are explicit objects of choice. 2. Conduct and/or embed institutional analysis (i.e. steps 1–4 discussed earlier) for each institution identified in A. 3. Characterize formally the (second order) institutions that constrain the choice of institutions defined in A. 4. Deduce the set of institutions that will be chosen given steps A–C 5. Compare the derived implications with empirical regularities and data.

The critical advantage of game-theoretic analysis lies in the concept of “subgames,” that can be used to formally model each feasible institution. The theorist then solves for the equilibria of each subgame and embeds each subgame in a larger game of institutional choice. By solving that larger game, we then have a precise concept of institutional equilibria. Non-cooperative game theory can accomplish this task for two reasons. First, extensive form games offer a very rich formal framework to capture institutional variety. Second, Nash equilibria exist for very broad classes of games. So, they usually will exist at all the levels of each subgame and at the level of the institutional choice game.The following examples serve to illustrate this approach.

48   Daniel Diermeier McKelvey and Riezman (1992) provide a formal model that yields a seniority system (e.g. Polsby, Gallagher, and Rundquist, 1969) as an equilibrium institution.24 Seniority systems give legislatures with longer tenure disproportionate proposal rights which result in a disproportionate share of benefits allocated to districts with more senior legislators. But then voters have an incentive to re-elect senior legislators. The senior system thus creates an incumbency advantage, which will be supported by legislators interested in re-election. The seniority system thus is not assumed as a constraint on collective decision-making, but is sustained in equilibrium in a game combining legislative bargaining and electoral competition. Diermeier and Myerson (1999) show how bargaining between multiple chambers or an independently elected president with veto powers lead to the existence of internal veto players (e.g. strong committees) or super-majority rules (e.g. a filibuster rule as in the US Senate). In contrast, unicameral legislators provide incentives to delegate power to a single actor such as a prime minister or party leader. Thus the model, not only explains the existence of certain institutions, but also their variance across political systems. The model points out that institutions that appear to be quite different (e.g. cabinets and US committees, or veto players and super-majority rules) may be functional equivalents. This is similar to Tsebelis’ veto-player approach, but the modeling approach is quite different. Diermeier and Feddersen (1998) present a multi-period model of distributive legislative bargaining in the Baron-Ferejohn tradition, where legislators need to distribute a fixed benefit in each policy period. In each period, legislators are randomly assigned a reservation value (i.e. a payoff that they will receive if the proposal fails), and a proposer is selected according to a recognition rule. However, these recognition rules are not fixed, but voted on in an organizational period where legislators bargain over recognition rights for the policy periods. Diermeier and Feddersen then apply their model to two different constitutional settings: a presidential democracy reminiscent of the US Congress and a parliamentary democracy with multiple parties. The key institutional difference between the two settings is the existence of the vote of confidence procedure,25 which allows a governing coalition to link voting on a particular bill with the survival of the governing coalition. Diermeier and Feddersen show that this constitutional feature has dramatic consequences; it permits legislative majorities to vote cohesively on multiple issues and to capture more of the legislative benefits compared witha presidential system. The model thus can explain the difference in voting cohesion between parliamentary and presidential systems. Note that the explanation does not depend on party discipline or electoral competition. Rather, it depends on the expectations of future benefits if the current governing coalition can stay in office. Therefore, the model can also explain why in parliamentary democracies’ voting cohesion is greater across parties than within parties in the US Congress. As illustrated by this example, we can see three main characteristics of the institutional equilibria approach. First, unless specified by the constitutions, any legislative institutions must be explained as an equilibrium in an underlying game of institutional

Formal Models of Legislatures  

49

choice. Second, institutions from different systems and countries are represented as formal equivalents. That allows models to compare different constitutions. Third, the models not only try to explain the existence of legislative institutions, but their variance across political systems.26 But nothing in life is free. The methodological advantages of non-cooperative game theory also come with substantial costs. First, as discussed above, non-cooperative games often have more than one equilibrium. This frequently requires stronger equilibrium concepts than Nash-Equilibrium (Subgame-perfection, Markov-Perfection, iterated elimination of weakly dominated strategies). The McKelvey-Riezman model, for example, uses all of these equilibrium refinements. In practice, these restrictions are widely accepted, but they do constitute additional assumptions. The second issue is the institutional richness of non-cooperative models, which can be both a blessing and a curse.27 In some applications, it is a blessing because it allows us to model the exact sequence of moves as specified in a political constitution. But in other contexts (e.g. intra-party bargaining among rival factions), the precise rules may either not be known, or decisions may result from informal negotiations. Non-cooperative game theory is not very suitable to model “informal institutions.” It forces the modeler to make decisions (“who can make offers?,” “who moves first?,” and so forth) even if such modeling decisions may seem to be arbitrary. This becomes particularly important if the respective modeling variants are associated with qualitatively different equilibria. There is no simple solution to this problem other than the laborious process of creating models of increasing generality and robustness that can then be compared with data. This is one reason why a close connection with empirical regularities plays such an important role in institutionalist models. Finally, the analysis of non-cooperative models often presents formidable technical difficulties, especially in multidimensional policy spaces. This does not reflect a methodological problem, just the sheer mathematical challenges associated with solving non-cooperative models.

2.8  Conclusion: Toward a General Theory of Legislatures This line of argument suggests some directions for the development of a general theory of legislatures. First, the goal should be to develop a genuine comparative orientation. The split between the study of the US Congress, coalition government, and other legislatures must be bridged. The comparative study of elections can serve as a potential reference case (e.g. Cox 1997). While electoral rules vary widely, much of their complexity can be reduced to some key feature such as district size (which influences the number of parties in an electoral system) and the degree of inter-party competition within the same district (which influences party discipline).

50   Daniel Diermeier Second, bridging such divides requires the ability to represent different institutions from different countries in a common formal framework. This may redefine traditional distinctions, e.g. between presidential and parliamentary democracies or between committees and cabinets. Both the veto player and the institutional equilibrium approach offer possible frameworks. Third, the only exogenous institutions are the general rules specified in a country’s constitution. All other legislative institutions, even if they are stable and impactful, need to be explained as institutional equilibria. Fourth, general theories must be able to explain not only the existence of legislative institutions but their variance. In this context, it is important that the models get the major differences right. That is, a model that can account for the (relatively minor) difference in power between the Norwegian and the German committee system, but cannot account for the difference between committees in the UK parliament and the US Congress, is of limited use. While the formal study of legislative politics has come a long way, much remains to be done. On the positive side, in recent years we have seen the emergence of research programs that are genuinely comparative in nature. On the negative side, the technical challenges of developing models in common frameworks remain formidable.

Notes 1. I will use the term “legislatures” throughout this chapters rather than “parliaments” or “committees.” In my view, it best expresses the ambition of legislative theory to apply to any decision-making body of moderate size (at least threeand no more than a few hundredactors) whose members repeatedly come together to deliberate and make binding decisions. The term “committees” is too technical, and the term “parliaments” often connotates the legislatures of parliamentary democracies thus excluding the US Congress, Latin America, and many other legislatures in Europe, Asia, and Africa. 2. This is enough for models with finitely many alternatives. In the continuous case, it is usually also assumed that preferences are continuous and strictly convex. 3. For a detailed discussion of these issues, see Diermeier (1997). 4. For recent overviews on coalitions, see Diermeier (2006, 2011), for overviews on the US Congress, see Krehbiel(2006), Cox (2006), and Volden and Wiseman (2011). 5. Additional concepts such as veto players can then be defined using the concept of decisive coalitions. 6. For more details, see Diermeier (1997, 2014). 7. See Riker (1980) for a forceful argument about the irreducibility of dependence on institutions 8. Testimony before the GOP Task Force on Congressional Reform, 16 December 1987. Cited from Diermeier, Prato, and Vlaicu (2014). 9. In the case of the US Congress, this is clearly stated in the constitution: “Each House may determine the Rules of its Proceedings” (US Constitution Article 1, Section 5, Clause 2). 10. To be sure, there are cases where the median policy prevails, e.g. when the status quo is located exactly on the median’s ideal point. Here we focus on the typical case where the Romer-Rosenthal model implies an outcome that is distinct from the median.

Formal Models of Legislatures  

51

11. Variants of the model have been used in the study of legislative voting rules (Baron and Ferejohn 1989a), committee power (Baron Ferejohn 1989b), pork-barrel programs (Baron 1991a), government formation (Baron 1989, 1991b), multi-party elections (Austen-Smith and Banks 1988; Baron 1991b; Chari, Jones, and Marimon 1997), and inter-chamber bargaining (Diermeier and Myerson 1994). 12. A variant of this set-up allows (nested) amendments to a proposal before it is voted upon. This is the case of an open amendment rule (Baron and Ferejohn 1989a). 13. For an overview see Shepsle and Weingast (1994). 14. Other contributions include Epstein (1997), Baron (2000), Collander (2008), and Kim and Rothenberg (2008). 15. For a model that sustains commitment in equilibrium using an overlapping generation’s model see Diermeier (1995). See also Muthoo and Shepsle (2012). McKelvey and Riezman (1992) use a model that involves both legislative and electoral stages to show that maintaining a seniority system with proposal power encourages the re-election of incumbents. See also Callander (2008) for a model where policy complexity provides a solution to the commitment problem. 16. In recent years, there has been a third important approach that has explained the institutional structure of theUS Congress as a consequence of majority party control. The partisan approach argues that party majorities allow their party leadership to control the agenda because sacrificing their individual policy goals in favour of the party median helps maintain the party “brand” and with it the party members’ re-election prospects (Cox and McCubbins 1993, 2005). Deferring to the party leadership may be supported either through disciplined voting on bills (Aldrich and Rohde, 2001) or on procedures (Cox and McCubbins 2005). Diermeier and Vlaicu (2011a) show that even in the absence of exogenously imposed party voting discipline preference, affinities among a majority are sufficient to induce partisan voting cohesion over asymmetric procedures. This approach has been highly influential in the study of the US Congress, but it is less important in other legislatures, which are often characterized by multiple, disciplined parties. 17. See Laver and Schofield (1990) for a detailed review of the field up to the late 1980s. 18. In parallel, theoretical work in the social choice tradition has continued. See, for example, Schofield and Sened (2006) and Schofield (2008). 19. See Laver and Shepsle (1994, 1996) for details. 20. See Diermeier and Stevenson (2000) and Diermeier and Merlo (2000) 21. A similar debate occurred a decade earlier on the role of committees (Krehbiel 1991, 2004). 22. The model in Diermeier and Merlo (2000), for example, has only been solved for three parties with quadratic preferences. 23. Tsebelis refers to them as “conjectures.” 24. See also Eguia and Shepsle (2012). 25. See also Huber (1996) and Baron (1998). 26. Adding uncertainty to the Diermeier-Feddersen model can explain cross-country difference in legislative success rates (Diermeier and Vlaicu, 2011b). 27. Note also that the approaches use different models of policy decision-making. McKelvey and Riezman as well as Diermeier and Feddersen use sequential bargaining models in the Baron-Ferejohn tradition. Diermeier and Myerson use a vote-buying model (Groseclose and Snyder 1996), though the general results hold for more general approaches. These modeling choices are driven by technical convenience and may not matter in the end, but the models have yet to be integrated into a common framework.

52   Daniel Diermeier

References Aldrich, J. H. and Rohde, D. W., 2001. The Logic Of Conditional Party Government: Revisiting the Electoral Connection. In: L. C. Dodd and B. I. Oppenheimer (eds.). Congress Reconsidered (7th ed.), pp. 269–92. Washington, DC: Congressional Quarterly Press.  Arrow, K. J., 1951. Social Choice and Individual Values. New York: Wiley. Austen-Smith, D., and Banks, J. S. 1988. “Elections, Coalitions and Legislative Outcomes.” American Political Science Review 82(2):405–22. Austen-Smith, D. and Banks, J. S., 1990. Stable Governments and the Allocation of Policy Portfolios. American Political Science Review, 84: 891–906. Axelrod, R. M., 1970.  Conflict of Interest:  A  Theory of Divergent Goals with Applications to Politics. Chicago: Markham Publishing Company. Bagehot, W., 1963 [1867]. The English Constitution. London: Collins. Banks, J. S., 1984. Sophisticated Voting and Agenda Control. Social Choice and Welfare, 1: 295–306. Banks, J. S., 1989. Equilibrium Outcomes in Two Stage Amendment Procedures. American Journal of Political Science, 33: 25–43. Banks, J. S., 1999. Committee Proposals and Restrictive Rules.  Proceedings of the National Academy of Sciences, 96: 8295–300. Baron, D. P., 1989. A non-cooperative Theory of Legislative Coalitions. American Journal of Political Science, 33: 1048–84. Baron, D. P., 1991a. Majoritarian Theory of Legislative Control. American Journal of Political Science, 35: 57–90. Baron, D. P., 1991b. A Spatial Bargaining Theory of Government Formation in Parliamentary Systems. American Political Science Review, 85: 137–64. Baron, D. P., 1996. A Dynamic Theory of Collective Goods Programs. American Political Science Review, 90: 316–30. Baron, D. P., 1998. Comparative Dynamics of Parliamentary Governments. American Political Science Review, 92: 593–609. Baron, D. P., 2000. Legislative Organization with Informational Committees. American Journal of Political Science, 44: 485–505. Baron, D. P. and Diermeier, D., 2001. Elections, Governments, and Parliaments in Proportional Representation Systems. Quarterly Journal of Economics, 116: 933–67. Baron, D. P. and Ferejohn, J. A., 1989a. Bargaining in Legislatures. American Political Science Review, 83: 1181–206. Baron, D. P. and Ferejohn, J. A., 1989b. The Power to Propose. In P. C. Ordeshook (ed.). Models of Strategic Choice in Politic, pp. 343–66. Ann Arbor: University of Michigan Press. Baron, D. P., Diermeier, D., and Fong, P., 2012. A Dynamic Theory of Parliamentary Democracy. Economic Theory, 49: 703–38. Bernheim, B. D., Rangel, A., and Rayo, L., 2006. The Power of the Last Word in Legislative Policy making. Econometrica, 74: 1161–90. Black, D., 1958.  The Theory of Committees and Elections. Cambridge:  Cambridge University Press. Browne, E. C., Frendreis, J. P., and Gleiber, D. W., 1984. An “Events” Approach to the Problem of Cabinet Stability. Comparative Political Studies, 17: 167–97. Browne, E. C., Frendreis, J. P., and Gleiber, D. W., 1986. The Process of Cabinet Dissolution: An Exponential Model of Duration and Stability in Western Democracies. American Journal of Political Science, 30: 628–50.

Formal Models of Legislatures  

53

Browne, E. C., Frendreis, J. P., and Gleiber, D. W., 1988. Contending Models of Cabinet Stability: A Rejoinder. American Political Science Review, 82: 923–41. Callander, S., 2008. A Theory of Policy Expertise. Quarterly Journal of Political Science, 3: 123–40. Calvert, R., 1995. The Rational Choice Theory of Social Institutions: Cooperation, Coordination, and Communication. In J. S. Banks and E. A. Hanushek (eds.). Modern Political Economy: Old Topics, New Directions, pp. 216–55. Cambridge: Cambridge University Press. Chari, V. V., Jones, L. E., and Marimon, R., 1997. The Economics of Split-ticket Voting in Representative Democracies. American Economic Review, 87: 957–76. Cox, G. W., 1997. Making Votes Count: Strategic Coordination in the World’s Electoral Systems. Cambridge: Cambridge University Press. Cox, D. R., 2006. The Organization of Democratic Legislatures. In B. R. Weingast and D. A. Wittman (eds.). The Oxford Handbook of Political Economy,  pp. 141–61. Oxford:  Oxford University Press. Cox, G. W. and McCubbins, M. D., 1993. Legislative Leviathan: Party Government in the House. Berkeley: University of California Press. Cox, G. W. and McCubbins, M. D., 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press. Denzau, A. T. and Mackay, R. J., 1983. Gatekeeping and Monopoly Power of Committees: An Analysis of Sincere and Sophisticated Behavior.  American Journal of Political Science, 27: 740–61. Diermeier, D., 1995. Commitment, Deference, and Legislative Institutions. American Political Science Review, 89: 344–55. Diermeier, D., 1997. Explanatory Concepts in Positive Political Theory. Working Paper. Graduate School of Business. Stanford University. Diermeier, D. 2006. Coalition Government. In B. Weingast and D. Wittman (eds.), Oxford Handbook of Political Economy (162–79). Oxford University Press. Diermeier, D. 2011. Coalition Experiments. In J. N. Druckman, D. P. Green, J. H. Kuklinski, and A. Lupia (eds.), Handbook of Experimental Political Science (399–412). New York: Cambridge University Press. Diermeier, D. 2014. Positive Political Theory. Encyclopedia of Political Thought. Hoboken: Wiley-Blackwell. Diermeier, D. and Feddersen, T. J., 1998. Comparing Constitutions: Cohesion and Distribution in Legislatures. European Economic Review, 42: 665–72. Diermeier, D. and Fong, P., 2011. Legislative Bargaining with Reconsideration. The Quarterly Journal of Economics, 126: 947–85. Diermeier, D. and Krehbiel, K., 2003. Institutionalism as a Methodology. Journal of Theoretical Politics, 15: 123–44. Diermeier, D. and Merlo, A., 2000. Government Turnover in Parliamentary Democracies. Journal of Economic Theory, 94: 46–79. Diermeier, D. and Myerson, R.B., 1994. Bargaining, Veto Power, and Legislative Committees. Center for the Mathematical Studies in Economics and Management Science. Working Paper No.1089. Northwestern University. Diermeier, D. and Myerson, R. B., 1999. Bicameralism and its Consequences for the Internal Organization of Legislatures. American Economic Review, 89: 1182–96. Diermeier, D. and Stevenson, R., 1999. Cabinet Survival and Competing Risks. American Journal of Political Science, 43: 1051–68.

54   Daniel Diermeier Diermeier, D. and Stevenson, R. T., 2000. Cabinet Terminations and Critical Events. American Political Science Review, 94: 627–40. Diermeier, D. and Vlaicu, R., 2011a. Parties, Coalitions and the Internal Organization of Legislatures. American Political Science Review, 105: 359–80. Diermeier, D. and Vlaicu, R., 2011b. Executive Control and Legislative Success. Review of Economic Studies, 78: 846–71. Diermeier, D., Eraslan, H., and Merlo, A., 2003. A Structural Model of Government Formation. Econometrica, 71: 27–70. Diermeier, D., Prato, C., and Vlaicu, R., N.D. (in press). Procedural Choice in Majoritarian Organizations. American Journal of Political Science. Döring, H. and Hallerberg, M., 2004. Patterns of Parliamentary Behavior: Passage of Legislation across Western Europe. Aldershot: Ashgate. Downs, A., 1957. An Economic Theory of Democracy. New York: Harper & Brothers. Eguia, J. and Shepsle, K., 2012. Endogenous Agendas and Seniority Advantage. Working Paper. Harvard University. Epstein, D., 1997. An Informational Rationale for Committee Gatekeeping Power.  Public Choice, 91: 271–99. Eraslan, H. and Merlo, A., 2002. Majority Rule in a Stochastic Model of Bargaining. Journal of Economic Theory, 103: 31–48. Farquharson, R., 1969. Theory of Voting. New Haven: Yale University Press. Gilligan, G. and Krehbiel, K., 1987. Collective Decision-making and Standing Committees: An Informational Rationale for Restrictive Amendment Procedures. Journal of Law, Economics and Organization, 3:287–335. Gilligan, G. and Krehbiel, K., 1989a. Collective Choice without Procedural Commitment. In P. C. Ordeshook(ed.). Models of Strategic Choice in Politics, pp. 295–314. Ann Arbor: University of Michigan Press. Gilligan, G. and Krehbiel, K., 1989b. Asymmetric Information and Legislative Rules with a Heterogeneous Committee. American Journal of Political Science, 33: 459–90. Gilligan, G. and Krehbiel, K., 1990. Organization of Informative Committees by a Rational Legislature. American Journal of Political Science, 34: 531–64. Gilligan, G. and Krehbiel, K., 1997. Specialization Decisions within Committee. Journal of Law, Economics and Organization, 13: 366–86. Groseclose, T. and Snyder, J. M., 1996. Buying Supermajorities. American Political Science Review, 90: 303–15. Huber, J. D., 1996. Rationalizing Parliament: Legislative Institutions and Party Politics in France. Cambridge: Cambridge University Press. Kalandrakis, A., 2004. A Three-player Dynamic Majoritarian Bargaining Game. Journal of Economic Theory, 116: 294–322. Kim, J. and Rothenberg, L. S., 2008. Foundations of Legislative Organization and Committee Influence. Journal of Theoretical Politics, 20: 339–74. King, G., Alt, J. E., Burns, N. E., and Laver, M., 1990. A Unified Model of Cabinet Dissolution in Parliamentary Democracies. American Journal of Political Science, 34: 846–71. Kramer, G. H., 1972. Sophisticated Voting Over Multidimensional Choice Spaces. Journal of Mathematical Sociology, 2: 165–80. Krehbiel, K., 1991.  Information and Legislative organization. Ann Arbor:  University of Michigan Press. Krehbiel, K., 2004. Legislative Organization. Journal of Economic Perspectives, 18: 113–28.

Formal Models of Legislatures  

55

Krehbiel, K., 2006. Pivots. In B. R. Weingast and D. A. Wittman (eds.). The Oxford Handbook of Political Economy, pp. 223–40. Oxford: Oxford University Press. Laver, M., and Shepsle, K. A., 1990. Coalitions and Cabinet Government. American Political Science Review, 84: 873–90. Laver, M. and Shepsle, K., 1994. Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge University Press. Laver, M. and Shepsle, K., 1996. Making and Breaking Governments: Cabinets and Legislatures in Parliamentary Democracies. New York: Cambridge University Press. Laver, M. and Schofield, N.,1990. Multiparty Government: The Politics of Coalition in Europe. Oxford: Oxford University Press. Lees, J. D. and Shaw, M., 1979. Committees in Legislatures:A Comparative Analysis. Durham: Duke University Press. Lorenz, E. N., 1960. The Statistical Prediction of Solutions of Dynamic Equations. Symposium on Numerical Weather Prediction in Tokyo, 629–35. Lorenz, E. N., 1993. The Essence of Chaos. Seattle: University of Washington Press. Lupia, A. and Strøm, K., 1995. Coalition Termination and the Strategic Timing of Parliamentary Elections. American Political Science Review, 89: 648–65. Mattson, I. and Strøm, K., 1995. Parliamentary Committees. In H. Döring (ed.). Parliaments and Majority Rulein Western Europe, pp. 249–306. New York: St Martin’s Press. McKelvey, R. D., 1976. Intransitivities in Multidimensional Voting Models and Some Implications for Agenda Control. Journal of Economic Theory, 12: 472–82. McKelvey, R. D., 1979. General Conditions for Global Intransitivities in Formal Voting Models. Econometrica, 47: 1085–112. McKelvey, R. D., 1986. Covering, Dominance, and Institution-free Properties of Social Choice. American Journal of Political Science, 30: 283–314. McKelvey, R. D. and Riezman, R., 1992. Seniority in Legislatures. American Political Science Review, 86: 951–65. McKelvey, R. D. and Schofield, N., 1987. Generalized Symmetry Conditions at a Core Point. Econometrica, 55: 923–33. Merlo, A., 1997. Bargaining Over Governments in a Stochastic Environment. Journal of Political Economy, 105: 101–31. Merlo, A. and Wilson, C., 1995. A Stochastic Model of Sequential Bargaining with Complete Information. Econometrica, 63: 371–99. Miller, N. R., 1977. Graph-theoretical Approaches to the Theory of Voting. American Journal of Political Science, 21: 769–803. Miller, N. R., 1980. A New Solution Set for Tournaments and Majority Voting: Further Graph-theoretical Approaches to the Theory of Voting. American Journal of Political Science, 24: 68–96. Muthoo, A. and Shepsle, K., 2012. Seniority and Incumbency in Legislatures. Working Paper. Harvard University. Myerson, R. M., 1996. Foundations of Social Choice Theory. CMSEMS Discussion Paper No. 1162 Northwestern University. Ordeshook, P.C. and Schwartz, T., 1987. Agenda and the Control of Political Outcomes. American Political Science Review, 81: 179–99. Plott, C. R., 1967. A Notion of Equilibrium and its Possibility Under Majority Rule. American Economic Review, 57:787–806. Poincaré, H., 2007 [1908]. Science and Method. New York: Cosimo Classics.

56   Daniel Diermeier Polsby, N. W., Gallaher, M., and Rundquist, B. S., 1969. The Growth of the Seniority System in the U.S. House of Representatives. American Political Science Review, 63: 787–807. Riker, W. H., 1962. The Theory of Political Coalitions. New Haven: Yale University Press. Riker, W. H., 1980. Implications from the Disequilibrium of Majority Rule for the Study of Institutions. American Political Science Review, 74: 432–46. Rogowski, Ronald., 1990. The Information-economizing Organization of Legislatures. Working Paper, UCLA. Romer, T. and Rosenthal, H., 1978. Political Resource Allocation, Controlled Agendas, and the Status Quo. Public Choice, 33: 27–43. Schofield, N., 2008. The Spatial Model of Politics. New York: Routledge. Schofield, N. and Sened, I., 2006. Multiparty Democracy:  Elections and Legislative Politics. Cambridge: Cambridge University Press. Shepsle, K. A., 1979. Institutional Arrangements and Equilibrium in Multidimensional Voting Models. American Journal of Political Science, 23: 27–59. Shepsle, K. A., 1986. Institutional Equilibrium and Equilibrium Institutions. Political Science: The Science of Politics, pp. 51–81. New York: Agathon Press. Shepsle, K. A., and Weingast, B. R., 1981. Structure-induced Equilibrium, and Legislative Choice. Public Choice, 37: 503–19. Shepsle, K. A. and Weingast, B. R., 1987. The Institutional Foundations of Committee Power. American Political Science Review, 81: 85–104. Shepsle, K. A. and Weingast, B. R., 1994. Positive Theories of Congressional Institutions. Legislative Studies Quarterly, 19: 149–79. Strøm, K., 1985. Party Goals and Government Performance in Parliamentary Democracies. American Political Science Review, 79: 738–54. Strøm, K., 1990.  Minority Government and Majority Rule. Cambridge:  Cambridge University Press. Tsebelis, G., 1995. Decision Making in Political Systems: Veto Players in Presidentialism, Par­ liamentarism, Multicameralism and Multipartyism .  British Journal of Political Science, 25: 289–325. Tsebelis, G., 2002.  Veto Players:  How Political Institutions Work. Princeton:  Princeton University Press. Tsebelis, G. and Money, J., 1997. Bicameralism. Cambridge: Cambridge University Press. Volden, C. and Wiseman, A. E., 2011. Formal Approaches to the Study of Congress. In F. Lee and E. Schickler (eds.), Oxford Handbook on the American Congress, pp. 36–65. New York: Oxford University Press. Warwick, P., 1992a. Ideological Diversity and Government Survival in Western European Parlia­mentary Democracies. Comparative Political Studies, 25: 332–61. Warwick, P., 1992b. Rising Hazards: An Underlying Dynamic of Parliamentary Government. American Journal of Political Science, 36: 857–76. Warwick, P., 1992c. Economic Trends and Government Survival in West European Parliamen­ tary Democracies. American Political Science Review, 86: 875–87. Warwick, P., 1994. Government Survival in Parliamentary Democracies. Cambridge: Cambridge University Press. Warwick, P. and Easton, S. T., 1992. The Cabinet Stability Controversy: New Perspectives on a Classic Problem. American Journal of Political Science, 26: 122–46. Weingast, B. R. and Marshall, W. J., 1988. The Industrial Organization of Congress; or, Why Legislatures, Like Firms, are not Organized as Markets.  Journal of Political Economy, 96: 132–63.

C HA P T E R

3

THE SOCIOLOGY OF L E G I S L AT O R S A N D L E G I S L AT U R E S H E I N R IC H BE ST A N D L A R S VO G E L

3.1 Introduction The sociology of legislators and legislatures has two main objects: legislators as social actors and legislatures as social institutions. Not only does the sociological approach investigate the impact of institutions on the behaviour and preferences of actors, it also analyses the social prerequisites and conditions underlying the establishment, maintenance and transformation of institutions, as well as the impact of social actors on these processes. This chapter focuses on the social relations within parliaments (insider–insider relations) and on the relations between parliaments and society (insider–outsider relations), which manifest themselves in three types of social interaction: competitive struggle for votes (i.e. the competition between legislators organized in political parties for the support of the population); antagonistic cooperation (i.e. the consensus of legislators regarding the institutions and informal rules of conflict resolution); and principal–agent relations (i.e. the asymmetric interdependence between legislators and their constituents based on delegation and accountability). In a general sense, social institutions can be defined as enduring systems of formal and informal rules, norms, and values structuring social interactions and relations; institutions make them durable, binding, and meaningful for the actors engaged in them. In a narrower sense, political institutions such as legislatures “are formal arrangements for aggregating individuals and regulating their behaviour through the use of explicit rules and decision processes enforced by an actor or set of actors formally recognized as possessing such power” (Levi 1990, 405). Political institutions are established (or radically transformed) at “critical junctures” (Rokkan 1999, 34). In such situations actors enjoy a wide degree of freedom because institutional constrains do not exist or

58   Heinrich Best and Lars Vogel have become ineffective for solving problems of social regulation and integration. This wide leeway results in insecurity, impeding cooperation between legislators due to the reduced predictability of behaviour and the lack of institutional resources for exercising and retaining power. In such situations influential actors seek to establish new institutions to provide a stable framework for calculating the benefits and costs of available options for behaviour (i.e. to reduce transaction costs). After this new institutional arrangement is successfully established, it reduces the leeway of the actors by restricting certain behaviours and promoting others. The initial institutional choices made by influential actors at critical junctures have long-term and pervasive effects on the behaviour of their successors, as institutions generate positive feedback (Pierson 2004). However, path dependency is limited by potential path inefficiency; if the decrease in transaction costs generated by maintaining an established institutional order is outweighed by the opportunity costs of maintaining it (i.e. the institutional arrangement is inefficient to solve problems of social regulation and integration), another critical juncture is reached, and the existing institutional order is challenged (Best 2010b). Legislatures in democracies are core elements of an institutional framework in which the struggle for power is mediated by peaceful competition for votes between antagonistic collective actors, who are usually organized in parties (Schumpeter 1959, 259, 269). The pacification of the struggle for power is a demanding task in terms of normative and institutional prerequisites. For centuries, the struggle for power was not peaceful. There are many examples of instances in which peaceful democratic competition was substituted by autocratic regimes that tried—often successfully—to win the political struggle by eliminating their competitors, sometimes even physically. The question therefore arises: why do legislators respect the institutional rules of democracy after acquiring the reins of state power? The answer is that the competitive struggle for votes and power is peaceful only if there is a basic consensus among otherwise competing actors on the rules of the political competition and a mutual acceptance of the political rivals as legitimate. The American sociologist William G. Sumner coined the term “antagonistic cooperation” to denote how adversaries may enter into limited but durable partnerships in order to pursue common interests and maintain a mutually beneficial social order (Best 2010a, 102). The competitive struggle for votes also shapes the relationship between legislators and their electorate. In Schumpeter’s (1959) pervasive analysis this relation is depicted as a leader democracy to underline that political leaders impose their preferences on voters. Today, however, the direction of influence is debated, both theoretically (Körösényi 2010) and empirically (see 3.4). On the most basic level, the social logic underlying the interaction between legislators and voters is a principal–agent relation. In the simplest terms, the population can be seen as the principal, who selects agents (legislators) and delegates to them the task of acting on their behalf in politics. The population holds their agents accountable for the outcomes they achieve (e.g. Strøm, Müller, and Bergman 2006). Legislatures can be described as political institutions that regulate both the competitive struggle of legislators for votes, based on antagonistic cooperation and the

The Sociology of Legislators and Legislatures  

59

interaction between legislators and their constituents, based on a principal–agent relation. Both the insider–insider relation and the insider–outsider relation are interdependent. While political institutions such as legislatures provide a stabilizing frame for such interactions, the stability of these institutions is simultaneously reinforced or eroded by the behaviour of the actors involved. This chapter investigates three facets of the relationships outlined above:  how is antagonistic cooperation learned and reinforced in the process of legislative socialization? In what way are the social interests and preferences of the population articulated and mediated in the legislature by virtue of the qualities and qualifications of the selected agents (legislators), and through the interactions between legislators and their constituents?

3.2  Institutional Socialization: The Learning of Antagonistic Cooperation Scholars of legislative studies have identified a broad range of informal rules and norms influencing the behaviour of legislators; these rules and norms are independent of both the partisan and the personnel composition of parliament (Matthews 1959; Cox 2000). Newcomers elected to parliament for the first time are faced with the challenge of acquainting themselves with these informal rules and norms. Although there is no agreement about the scope of attitudes and preferences that parliamentary socialization comprises, there is agreement that this process can be modeled as adaptation of newcomers to the senior members (Asher 1973; Badura and Reese 1976; Bell and Price 1975; Fenno 1962; Hedlund 1968; Mughan, Box-Steffensmeier, and Scully 1997). Research on political socialization thereby contributes to the understanding of the maintenance of institutions and how they impart their regulative and constitutive rules and norms to their members. Empirical research has repeatedly demonstrated that institutional socialization is, in large part, learning how to cooperate with competitors both within and outside one’s party. Fenno (1962, 320) described socialization in the Appropriations Committee of the US House of Representatives as an apprenticeship in which the perceptions, attitudes, and behaviours of newcomers must adapt to the consensus among senior members and “serve[s]‌as a basis for Committee integration.” Newcomers are offered little freedom or flexibility and are expected to follow the rules unquestioningly. In addition to this component of apprenticeship, the content of the consensus can be read as a guideline for antagonistic cooperation, including specialization (the deliberate distinction of responsibilities, which reduces conflicts of interest), reciprocity (do ut des as the predominant mode of interaction) and (sub-)committee unity (compromise or consensus as overarching goal) (Fenno 1962, 316). Other studies have identified precisely these norms as being gradually internalized by newcomers to both the US House

60   Heinrich Best and Lars Vogel of Representatives (Asher 1973) and German state legislatures (Reiser et al. 2011, 834). Davidson (1969, 72) found that members of the US Congress shift from “insiders to outsiders” with tenure. The longer they hold their positions, congressional representatives increasingly focus on internal affairs such as legislation and committee assignment and come to perceive their mandate to pursue the particular interests of their districts as decreasingly important. This shift in priorities indicates that the requirements of interest representation are attenuated in favour of stronger insider cooperation. Similar results were presented by Mughan et al. (1997, 50), who demonstrated an association between “deradicalization” (Searing 1986, 341) in several policy areas and tenure among members of the British House of Commons. Most of the recent research on legislative socialization has been conducted on the European Parliament, with the key question being whether its legislators are “going native,” i.e. does lengthy tenure foster a more pan-European perspective, with an increasing endorsement of pro-European policies (Scully 2005, 45)? The empirical results gathered, however, give only marginal indication for such a process (Franklin and Scarrow 1999, 57; Scully 2005, 54, 133). Studies of legislative socialization emphasize that the ambition of newcomers to gain political influence and their lack of the resources necessary to achieve this goal are the central incentives for these junior legislators to adapt to the internal rules and norms in parliament. Those who are successful have learned “that their conformity to [... ] norms [in legislatures, HB/LV] is the ultimate source of their influence” (Fenno 1962, 322). This learning process occurs in the daily interactions between the legislators, in which deviations are answered by sanctions ranging from subtle non-verbal signals to the refusal of further promotion in the parliamentary hierarchy. The significant others in these interactions are located within the specialized sub-organizations of legislatures bodies, such as committees and party caucuses. While the preceding section gives a good organizational overview, the full picture is more nuanced and ambiguous. In their work on first-term legislators in the California state assembly, Bell and Price (1975, 166) demonstrated that socialization is not a linear process. They found that some of the newcomers’ attitudes did not appear to converge after two or more years, and, more strikingly, senior members may also adapt to newcomers. There is also evidence that political socialization is not restricted to the legislature itself, but includes the pre-parliamentary career; as a result, many newcomers are acquainted with legislative rules and norms prior to their entry into parliament (Asher 1973, 512; Bell and Price 1975, 138). These findings have important implications: there are indications that pre-parliamentary socialization is actually the dominant factor structuring legislative behaviour, more important than the actual institutional environment. This occurs especially in newly established institutions, such as the Scottish regional parliament (Stolz 2010, 284), which has no senior members to provide models for newcomer behaviour. Finally, Best and Vogel (2012a, 56f) have shown that change in the representational focus of German legislators, which was initially attributed to newcomer status, vanishes if membership in formal positions in parliament is statistically controlled. In recent years, institutional socialization has been a neglected area of legislative studies. The majority of the studies mentioned above were conducted before 1990, and these

The Sociology of Legislators and Legislatures  

61

mainly pertain to the US. None of these follow a comparative approach, and only a few rely on quantitative individual-level panel data with control groups (Asher 1973; Bell and Price 1975; Reiser et al. 2011). This is the only adequate method to distinguish institutional socialization from both confounding cohort effects (if socialization is measured by tenure) and general change among the totality of legislators in parliament (which is not observable without a control group of elder members). In view of the desiderata sketched above, comparative empirical analyses of institutional socialization are needed in order to understand this important precondition for antagonistic cooperation. Such studies should include legislatures as the central institutions in which the norms of reciprocity and compromise are communicated. However, given that preconditions for successful socialization are also provided by the pre-parliamentary career, research needs to investigate legislatures as just one part of legislators’ political career, in which they become recruited into their first mandate.

3.3  Legislative Recruitment: Equality versus Social Closure Legislators play a central role in legitimizing representative democracies (Best and Cotta 2000, 7). The qualities and qualifications of legislators therefore have a symbolic and functional significance for the performance and stability of representative democracies (Sartori 1987). Research in legislative recruitment has strongly confirmed this view and has identified parliaments as intersections of different segments of the elite system, a mixtum compositum of representatives of different societal interests and from various political followings. However, the assumption that representative assemblies socially “mirror” the societies from which they are recruited is a normative construct and is never found in a polity where there is “free competition for a free vote” (Schumpeter 1959, 271; Norris 1997). The parliaments that came closest to this “mirroring” ideal were characteristically those of Eastern European people’s democracies where the cadres’ offices of Communist parties controlled admission to the assemblies according to elaborate quota systems based on criteria such as gender, social origin, and ethnic background. However, even under these special conditions, the mirrors were distorted in favour of meritorious party activists and veterans (Patzelt and Schirmer 2002, 386– 441). The paradox that defines sociological research into legislative recruitment is that through processes of selection and election which are—in principle—egalitarian, inclusive, and free, representative democracy emerges as an order of inequality. Legislative recruitment is the crucial point of the democratic game when it comes to the decision about who has the right of representation. These rules are inherent to the “competition for political leadership,” which was called “free” by Schumpeter (1959, 295) “in the same sense in which everyone is free to start another textile mill,” meaning anyone can be considered free to enter into the competition. The basic Schumpeterian

62   Heinrich Best and Lars Vogel design has to be amended, however, in two respects that are insufficiently covered and dealt with in his work. The first is the fact that, like economic elites, political elites try to curb the effects of competition by introducing cartels and restricting access for new competitors. The second is the observation that the competition for power is a two-layered process: Parties as selectorates compete for a bigger share of the positions in the political system, which are directly or indirectly distributed through elective competition. Individual contenders compete for the backing of selectorates and for access to the valuable elective offices. In short, parties (i.e. selectorates) want to improve their competitiveness through good candidates who can serve their parties’ external and internal needs in their struggle for power. Meanwhile, contenders display and employ the assets they have that give them an edge over their competitors, establishing good starting positions in their race for offices (Norris and Lovenduski 1995). The changing nature of this competition is the driving force behind the long-term transformation of recruitment patterns of representative elites. A  supply-and-demand model has been very useful in conceptualizing the dynamics and constraints of the recruitment process. In particular, such a model helps to promote a better understanding of why long-term changes in recruitment patterns of European representative elites seem to follow regular trends, notwithstanding some erratic fluctuations at historical turning-points and caesurae of recent European history (Norris and Lovenduski 1995; Best and Cotta 2000, 9–16). In the simplest version of the recruitment function there is a demand and a supply side and sets of formal rules and informal practices which determine how both sides are matched. The rules and practices include the criteria that determine who participates in which role in the competition for parliamentary seats and what rewards or risks contenders may expect in the competition. The main actors representing the supply or demand sides of the recruitment process are contenders, selectorates, and electorates. Contenders are those actors “who are stimulated to enter the competition for offices by individual incentives like prestige, power, material rewards, spiritual or ideological commitments” (Best and Cotta 2000, 11). They dispose of certain resources that qualify them for entry into the electoral competition and determine their starting position in the race for mandates and offices. Attributes and affiliations of contenders give a favourable or unfavourable momentum to their passage through the recruitment process. Selectorates are collective actors who select candidates according “to complex choices considering the probable value of the contenders’ resources for electoral success, to their ideological fit with, their instrumental function for, and their loyalty to the selectorates” (Best and Cotta 2000, 11). In a distant past selectorates were informal caucuses made up of dignitaries or state officials involved in the selection of candidates. Today selectorates in most liberal democracies tend to be institutionalized in the form of party organizations. These selectorates have an intermediate position in the recruitment market, matching the offer of contenders with the perceived preferences of electorates. Electorates are the “end consumers of offers on the electoral support markets” and the final judges of the outcomes of legislative recruitment (Best and Cotta 2000, 12). Their perceived preferences for a certain type of parliamentary representation are one factor shaping the lists of candidates drawn up by selectorates. The given makeup of a parliament can therefore

The Sociology of Legislators and Legislatures  

63

be regarded as the final balance of advantageous and disadvantageous factors working in the (self-)selective process preceding the act of recruitment (Best and Cotta 2000, 12). The interactions of the actors involved in the recruitment process are partly concealed by the secrecy of the ballot box and the seclusion of the backrooms where caucuses and party dignitaries meet. However, much of the process of parliamentary recruitment is open to public scrutiny and media attention. This directs, adjusts, and intensifies the public’s perceptions and expectations concerning the qualities and qualifications of contenders (Hetherington 2001). Legislative recruitment should therefore be considered part of a construction of reality by which groups of selectorates—today, mostly within political parties—try to influence the competition for power in their own favour. Their lists of candidates are an important element of the face parties present to voters and may be indicative to their “closeness” to certain quarters of the electorate. The makeup of a party’s parliamentary representation is therefore both a potential attractor of votes and a “tracer” for the groups it targets in the electorate (Best 2007). To the extent that the properties and qualities of contenders attract voter support in electoral campaigns, they are assets in the competition for legislative seats (see Fig. 3.1). Selectorates will prefer contenders with such valuable characteristics, as they provide an advantage in the struggle for power. Symbolic representation, i.e. the choice of representatives according to the image they transmit to both the public in general and to their constituents in particular, is based on innate or primordial qualities of contenders that relate to fundamental political issues. Examples are skin colour, gender, religion (as far as it is inherited), or social origin (such as working-class background). This symbolic representation can be distinguished from a second type of representation based on acquired attributes that are also attractive to voters, but stem from representatives’ personal reputations. In this category we find “heroes,” “martyrs,” or well-tried leaders. This type of representation could be referred to as “deferential,” as it is based on voters’ deference to achievements of those who are supposed to represent them. Seen from the standpoint of selectorates (in particular, political parties), both types of representation have an external focus: here, the main concern of selectorates is how their supply of representatives affects their parties’ images among voters. In addition to the external focus, selectorates also choose candidates for parliamentary office according to an “internal focus,” i.e. the functional requirements of party Focus External

Internal

Acquired

deferential

instrumental

Attributed

symbolic

relational

Origin

(Best 2007, 90) FIG. 3.1  Assets

for legislative recruitment and careers

64   Heinrich Best and Lars Vogel organizations and parliamentary parties for experts on certain policy issues, intermediaries connecting them to pressure groups and integrators capable of uniting divergent segments of party organizations or tiers of the political system (see Fig. 3.1). Again it is useful to distinguish between two types of resources available to contenders. The first type is relational and based on representatives’ relations to the extra-parliamentary sphere, built up through linkage positions held in organizations and networks. The second type of resource with an internal significance to party apparatuses refers to acquired competencies and qualifications of contenders that are instrumental for the exertion of the representational role. Inter alias level of education, area of study, pre-parliamentary professional skills, and previous political experience fall into this category of resource. The “internal–external” dichotomy of the taxonomy depicted in Fig. 3.1 refers to the demand side of the recruitment function by specifying the double task of selectorates: to supply parliamentary parties with personnel able (a) to attract voters by representing their diverse interests credibly, and (b) to fulfil the function of legislators competently by providing effective leadership and good governance. These competing and sometimes conflicting demands have been captured in the trustee and delegate conception of representations, whereby the delegate is a derivative of constituents’ preferences and the trustee is tied up in the institutional constraints of the policy-making process (Eulau et al. 1959; Mansbridge 2003). These two different conceptions of representation are both present in the institutional fabric and political practice of representative democracies and have to be accommodated within the collective and the individual actors involved; however, the emphasis on one or the other conception shifts over time and varies between polities. For example, it is plausible to assume that the extension of suffrage and voting eligibility in the process of mass democratization was accompanied by a shift from an internal to an external focus of parliamentary recruitment and toward “descriptive representation” which allows constituents to recognize themselves in those who represent them (Pitkin 1967). The “attributed–acquired” dichotomy in Fig. 3.1 refers to the supply side of the recruitment function. It considers the fact that representatives owe their parliamentary office not only to their personal virtues, qualifications, and skills, but also to the support of powerful organizations or factions of the selectorate who support them with patronage and sponsorship, expecting loyalty and services in return. In theory, the number of contenders in mass democracies is equal to the total size of the constituency; in fact, however, the supply of contenders is dramatically reduced by informal requirements for those who enter the competition. The “free competition for a free vote” is limited by a process that is commonly and somewhat euphemistically called “political professionalization,” i.e. a configuration of social processes and informal structures restricting access to parliamentary seats and political offices (Best 2003, 370). Political professionalization “defines the rules and rites of access of the group, what holds the members of the group together, and what sets them apart from other individuals in larger society” (Beaver and Rosen 1978, 66–7). In short, political professionalization establishes an insider–outsider differential and provides the social mechanisms that integrate professional politicians collectively into the “political class” (Borchert and Zeiss 2003). In Europe, the

The Sociology of Legislators and Legislatures  

65

mechanisms of candidate selection, which have developed over the past 150 years, are now predominantly controlled by party organizations, converting parliaments into quasi internal labour markets for parties. If the access to and support of organizational power has become a crucial advantage in the competition for seats, one could expect an increasing emphasis on relational assets in the recruitment process. It seems that there is an inherent contradiction in the recruitment function: on the one hand it entails the promise of democratization by an “opening of political societies” and an “expansion of choice opportunities” (Blondel 1997, 96). In the long run, it fosters a rise in self-expressive values, shifting cultural norms toward a greater emphasis on responsive and inclusive elites (Welzel 2002). On the other hand restrict the self-interest of established elites and the functional requirements of highly complex policy-making processes the chances of potential contenders and the responsiveness of those in office. In fact, Max Weber (1958 [1919]) already pointed to the fact that political professionalization is positively correlated with incumbency, and Robert Michels (1915) based his work on the trade-off between political democratization and the accompanying emergence of oligarchies in mass organizations. The gap between the practices of parliamentary representation and the promises of mass democracy seems to be widening with a negative impact on the legitimacy of representative democracy and the reputation of those who represent it (Dalton 2004). Previous research into legislative recruitment has analysed the contradictory coevolution of mass democracy and political professionalization in a series of country-specific studies. This approach was justified because the decisive context of legislative recruitment is the nation state, within whose perimeters the creation and redistribution of wealth, the acquisition and attribution of power, the definition and assertion of collective identities, the institutionalization of norms in legal systems, the formation of large bureaucratic structures, the aggregation of interests, and the emergence of platforms for collective political action (e.g. parties) are performed (Rokkan 1999). If legislative recruitment and its outcomes are path-dependent, it has been the nation state that paved and maintained the way. The formal structure of opportunities for access to office, such as electoral laws and eligibility rules, the supply of and demand for contenders, and the composition and mode of operation of selectorates (such as caucuses or parties) were all defined by the national boundaries of polities and societies, although the actual act of recruitment might have taken place on a local or regional level. No wonder, therefore, that political elites have been expected to represent the national distinctiveness of polities and societies (Mosca 1939). Thanks to regional diversity in the processes of state- and nation-building and the emergence of different paths toward democracy, the polities of Europe offer themselves as the richest field of observation for such a study. Research into the long-term trends of representation recruitment has departed from the heuristic assumption that the long-term change of legislative recruitment was driven by a single harmonious momentum for “political development” or “political modernization” (Best and Cotta 2000). However, even if one applies the most general notion of political modernization, we see at best a contradictory picture. While European parliaments have long ceased to be exclusive clubs for the wealthy and those

66   Heinrich Best and Lars Vogel who inherited positions of high social status, and while women have increasingly found their way into national assemblies, new barriers have emerged to replace those of class and gender. Legislative recruitment is now less shaped by the status hierarchies and value systems prevalent in societies at large; the decisive barriers and filters in the process of selecting candidates are now located within the narrower realm of political systems (Best and Cotta 2000).This shift is indicated by the gradual exclusion from the ranks of legislators of those who have a background in productive or distributive economic activities (such as workers and farmers), the corresponding increase in public servants and (for some time) in officials of pressure group organizations and parties, the growing accumulation (sequential and simultaneous) of local and regional offices and the increasing embedding of contenders into the higher ranks of party hierarchies. The abolition of formal barriers of access to parliaments was thus complemented by the establishment of an informal insider–outsider differential, firmly guarded and perpetuated by selectorates and party organizations. Today, those who are available for elective public offices, who have qualifications and skills deemed useful for a political career (preferably certificated by an academic degree of some kind), and who are willing and able to implant themselves in local or party offices, stand a greater chance of penetrating the filters and overcome the barriers on their way to a parliamentary seat. The rerise of the public sector (after its early heyday in the nineteenth century and its decline between the two World Wars) as the preferred supplier of parliamentary representatives in Europe, can be associated with the emergence of parties that rely “increasingly for [their] resources on the subventions and other benefits and privileges afforded by the state” (Katz and Mair 1995, 20–1). With the goals of politics becoming more self-referential and politics becoming a profession in its own right, representatives with a background in public service embody the fusion between party and state; their state employer sponsored them when they were amateur or semi-amateur politicians through generous exemptions and offered them a safe haven if their political career was threatened (Cotta and Almeida 2007). On the other hand, disposes their background and actual interest them to act as “agents of the state” in their representational role (Katz and Mair 1995, 18). Contenders from other professional backgrounds do not enjoy the same privileges and have to face a disproportionately unfavourable risk–benefit relationship when they pursue their political careers. However, the rise of public service to become the main societal sector for parliamentary recruitment not only reflects the cost–benefit calculations of selectorates and contenders, but can also be linked to the main challenge faced by Western polities in the bipolar world after the Second World War: the establishment of consensually unified polities and societies as a primary condition for the containment of communism. The mediation of conflicts and the integration of societies was the order of the day, and corporate interest mediation and the extension of welfare state benefits, particularly, were the most important consensus-creating policies. This “consensus challenge” found a response in parliamentary recruitment whereby redistribution specialists, who are predominantly found in the public sector, prevailed during this period (Best 2007).

The Sociology of Legislators and Legislatures  

67

It has been suggested to extend the challenge–response model into a general explanatory scheme for the long-term transformation of European legislative recruitment (Best 2003; Best 2007). It assumes that challenges which originate in or are mediated by the social and economic structure and the institutional order have an impact on the elite structure and agenda, which in turn influence elite responses to the same challenges. Elite responses may have repercussions on elite structures and agendas, the social structure and the institutional setting of the political order, separately or in combination. Rather than a linear development following the general transformation of social structures, there has been a pattern of change in parliamentary leadership groups reflecting the timeline of central challenges for polities and societies since the Western world entered the era of democratization and industrialization. Thus, the first period of public service dominance in many national parliaments during the nineteenth century coincides with the era of state- and nation-building. During this period, “symbol specialists” (Lasswell 1954) and specialists in the application of executive power, both of which were to be found in the higher ranks of the public sector, had a dominant role. The second challenge had to be taken up in the period of accelerated economic change during which most European societies faced the full impact of industrialization. In this period, specialists in the creation and appropriation of wealth, such as entrepreneurs and landowners, prevailed in parliament (Best 2007, 98–9). The third challenge was the development of mass democracy and the accumulation of organizational power outside state apparatuses (such as political parties and pressure groups). This period saw the rise of specialists in mass mobilization and the running of intermediary organizations (Best 2007, 98–9). A  link can be established here to earlier typologies of parties, with the “elite party” providing a political arena for high-ranking state officials and economic elites such as entrepreneurs and large landowners, and the “mass party” providing career opportunities for party and pressure group officials, and with the “catch-all party” forming the seedbed for the “redistribution specialists” from public service stock (Katz and Mair 1995). The indications for a convergence of legislative recruitment and career patterns in Western democracies after the Second World War, however limited, can therefore be attributed to a growing synchronization between the developments in party systems and the main policy alternatives faced by Western democracies; at the same time, the impact of changes in the formal structures of opportunities, such as electoral laws and eligibility rules, as well as general societal change at large, has lost momentum. The single most important factor that had synchronized the developments of post-Second World War Western democracies—the communist challenge—disappeared rapidly between 1989 and 1990. If there is an empirical pattern of “challenge and response,” we would expect that the disappearance of the communist threat would have resulted in a change of legislative recruitment and career patterns. In particular, this change should have affected legislative recruitment from the public sector. Indeed, our data confirm this expectation. The time series for public-sector representation in Western European parliaments reached its turning point at the end of the Cold War

68   Heinrich Best and Lars Vogel and has decreased considerably since this time (Best 2007, 100). This change is most pronounced in the case of the teaching profession, which is the single most important sub-category among European legislators from the public sector. Since the beginning of the 1990s the average share of members of the teaching profession in the parliaments of Western Europe has dropped by more than 20 percent from its peak at the beginning of the 1990s, meaning that they have lost about half of their previous gains since the beginning of the 1970s (Best 2007, 100). As the dominance of legislators with a professional background in public service has not yet been contested by any other professional category, a diversification of recruitment channels can already be seen. Assets such as loyalty toward the established political order, a “neutral” position in class conflicts, and command of the policies of redistribution, which can all be ascribed to contenders from the public service, are less valued after the consensus challenge has passed. The change in trends of legislative recruitment patterns was accompanied by a sharp increase in the turnover of individual legislators, whereby the average turnover rates of Western European parliaments nearly doubled between the end of the 1980s and the mid-1990s. The time series for newcomers forms a distinct peak during these years that was only exceeded during the periods of post-crisis recruitment following the First and Second World Wars. Although turnover rates have leveled off since the mid-1990s they are still above the average levels of the post-Second World War era. It is no mere coincidence that these changes occurred during and after the period of regime transition in Eastern Europe: the fall of communism in the East marked the end of the consensus challenge in the West. Recent developments such as the increasing diversification of recruitment patterns or the increase in turnover have been interpreted as responses to a legitimacy challenge that has emerged within the political systems of Western democracies (Dalton 2004). This new challenge targets elite quality, i.e. the ability of a representative democracy to produce efficient and accountable political elites. Institutional settings for elite recruitment, such as the cartel party, which are based on arrangements between politicians to appropriate and share the resources of the state (Katz and Mair 1995), might be suitable to meet such challenges and create a political elite that is united by common material interest. In the long run, however, these tendencies might serve to further undermine the legitimacy of representative democracy, as the ingroup–outgroup differential may become too large and may not be justifiable by the incumbents’ performance; the “true nature of democracy” (Schumpeter 1959, 295) is blurred if the competitive struggle for power is impeded. The closure of the political market through political professionalization and the pooling of interests between formally competing parties is an autocatalytic process that may jeopardize the working of representative democracy. In this respect, the emergence of the legitimacy challenge is an indication that democracies are able to generate countervailing mechanisms, infusing new competition into the system, and making the political profession riskier than most of its protagonists would like it to be.

The Sociology of Legislators and Legislatures  

69

3.4  Representation: Legislators and Their Constituents Legislative recruitment is a powerful, albeit indirect, mechanism for the transfer of societal preferences and interests to parliament via the qualities and qualifications of the legislators. Legislators are, however, also involved in a multitude of direct interactions with their constituents, both during and between elections. The manner in which societal preferences influence the behaviour of legislators in parliaments and the repercussions of this behaviour on societal preferences can be described in more general terms as a principal–agent relation. This basic constellation results in an asymmetric interdependency:  the voters expect their legislators to advocate their interests and preferences; the legislators are in turn dependent on voter support to get (re)elected. The relation is asymmetric as the legislators are not only better informed and better equipped to deal with political matters than their constituents, but (unlike their voters) they additionally conduct politics professionally. Therefore, legislators provide leadership in the principal–agent relation and significantly frame the definition of political success upon which their constituents evaluate them. The asymmetry in the relationship between voters and legislators adds to the basic problem of principals in relation to their agents: how to secure effective and efficient supervision of agents, in order to make sure that they act not only on behalf of their principals, but also in their principals’ interest? In normative theory the most important mechanism of control by the electorate is the competitive struggle for votes between legislators: with multiple potential agents legislators are supervised and evaluated by contenders both inside and outside parliament. The risk of being voted out in the next election constitutes a strong incentive for incumbent legislators to be responsive to their constituents. In this respect, parliaments are institutions in which legislators interact in order to advocate the interests of their principals. For a peaceful and efficient competitive struggle, there must be consensus among the legislators about the legitimacy of parliament and its rules and norms. Such antagonistic cooperation (i.e. cooperation between otherwise conflicting actors) comprises both the consensus about politics (i.e. the formal and informal rules institutionalized in parliament and its surrounding institutions) and the willingness and ability to make political compromises (i.e. achieving policy decisions in the face of competing interests). As the interaction between legislators is a balancing act between cooperation and competition, the legislators need leeway for their actions. On the one hand, this serves the interest of the principals, as it allows their agents to advocate their interests; on the other hand, it necessarily implies at least a partial deviation of the agents from the preferences of their constituents. The internal relations between legislators, as structured by legislatures and the relationship between legislators and their constituents, are thus inextricably intertwined.

70   Heinrich Best and Lars Vogel The combination of asymmetric interdependence, the competitive struggle for votes and antagonistic cooperation establishes diverging preferences for legislators and citizens. The population demands both close ex-ante and ex-post screening and as many restrictions as possible on the behaviour of their legislators (Lupia 2006). In contrast, legislators are keen to increase their autonomy. However, the preferences of legislators and the electorate may converge, as the majority of voters is unwilling and/or unable to get actively involved in politics. The population thereby accepts the delegation of authority and the leadership role of legislators, who are kept responsive and responsible to their constituents by the competitive struggle for votes. In this way, the social interaction between legislators and their constituents includes the potential for both convergence and divergence, responsiveness and leadership. Any valid investigation of this social interaction and its determining institutional and social factors requires the inclusion of comparisons of legislators and constituents. Such comparisons are rarely conducted in legislative studies per se (Leston-Bandeira 2012; Norton 2002) but are common in studies on public opinion, representation, and political elites. With their pioneering work, Miller and Stokes (1963) were the first to analyse relative congruence between legislators and their constituents as the extent to which “opinions in districts [... ] that are more supportive of a policy in a particular direction are reflected through more supportive roll call voting in that direction” (Shapiro 2011, 990). Despite many differences in research design, data, and theoretical concepts, Miller and Stokes’ study and its successors have demonstrated that in the US (Stone 1982; Powell 1982; Page et al. 1984), and in Canada (Soroka, Penner, and Blidook 2009) district opinions are reflected in legislators’ preferences and behaviour. The degree of this relative congruence is increased by homogeneity of district opinion (Harden and Carsey 2012), saliency (Kuklinski and Elling 1977), low complexity, rootedness of issues in inter-party conflicts (Hurley and Hill 2003, 308), and the commitment of legislators to responsiveness (McCrone and Kuklinski 1979; but see Friesema and Hedlund 1974). While dyadic representation focuses on individual legislators, collective representation is considered as the systemic property of a polity and is thus measured by comparing whole legislatures with the totality of the population (Pitkin 1967, 126ff; Weissberg 1978). However, studies on collective representation are rarely conducted. Rather, research has concentrated on the way changes in public opinion are reflected in changing preferences of legislators and ultimately in policy output. These studies have shown responsiveness at work in the US (Page et al. 1984; Stimson, Mackuen, and Erikson 1995, 985; Shapiro 2011), and in Germany (Brettschneider 1995). It was especially in light of this strand of research that the underlying model of representation was revised. Miller and Stokes based their research on the assumption of a demand–input model (Wahlke 1971, 272–3), presuming constituents to be active in formulating policies to which legislators are supposed to remain faithful if they intend to pursue (re-)election. The mechanisms that are meant to ensure legislators’ responsiveness are electoral turnover and the rational anticipation of electoral outcomes by the legislators between elections (Stimson, Mackuen, and Erikson 1995, 544–5). The asymmetric interdependence of the principal– agent relation was introduced as an amendment to this theorem of rational anticipation.

The Sociology of Legislators and Legislatures  

71

In consequence, legislators are presumed to adjust their behaviour in anticipation of the preferences of their constituents in the upcoming election and to actively form these preferences in an advantageous manner (Mansbridge 2003; Körösényi 2010). In the US and Europe it is now widely accepted—however, not always considered in empirical studies—that the impact between legislators and constituents is reciprocal (Jacobs and Shapiro 1994; Wlezien 1995; Steenbergen, Edwards, and de Vries 2007; Dalton, Farrell, and McAllister 2011). Although Miller and Stokes demonstrated that political parties in the US impact legislative behaviour significantly in those policy domains that cut across regional boundaries (Miller and Stokes 1963, 56), it has been primarily the research on Europe that focuses on collective congruence between legislators and partisan voters of the same party. This approach is justified, as geographic electoral districts appear to be less relevant to the behaviour of legislators in parliamentary systems with distinct party cohesion and discipline (Thomassen 1994). Most studies (Barnes 1977; Farah 1980; Converse and Pierce 1986; Esaiasson and Holmberg 1996; Miller et al. 1999) regard voters as inactive, but responsive to legislators, who offer policy packages and try to find support for them in elections (responsible-party model). The influence of electoral systems on the congruence between legislators and constituents has been analysed, showing different patterns for majoritarian and proportional systems: legislators tend to be closer to their party voters in the latter system and closer to the overall median voter in the former system (Dalton 1985, 286f Huber and Powell 1994; Wessels 1999). The consequences of this pattern for the level of overall collective congruence at the national level differ with the proportion of partisan voters in the electorate (i.e. social composition) (Rohrschneider and Whitefield 2012). In addition to the electoral system, other factors impact collective congruence; generally speaking, the more hierarchically centralized parties’ organization (Dalton 1985, 291) and the higher the polarization between parties, (Wessels 1999, 158) the higher the collective congruence becomes. Catch-all parties consider the median voter as their point of reference, while smaller parties representing more homogenous constituents cater more closely to their partisans (Ezrow et al. 2011). Testing the responsible-party model for several countries in Latin America has provided evidence that the degree of institutionalization of the party system (i.e. a mature record of party competition for votes) contributes significantly to the congruence between legislators and their partisan voters (Luna and Zechmeister 2005). In general, collective congruence is less pronounced with regard to particular policy issues, but manifests itself more in the form of ideological agreement, especially concerning the left–right continuum. Ideological congruences or differences between voters and legislators reflect the cleavages and polarities of the party system, thereby increasing party–voter congruence for issues related to party conflicts (Hurley and Hill 2003; Valen and Narud 2007). The party system largely simplifies policy issues to a dominant ideological dimension; this helps to overcome problems of social choice and is therefore a precondition for collective congruence (Thomassen 1994, 254–5). Observed patterns regarding ideology primarily display relative congruence: the positions of the legislators on the left–right continuum tend to correspond to the differences between

72   Heinrich Best and Lars Vogel their respective parties’ voters, but they tend to be either more or less pronounced than among the population (“polarized” and “moderating” trusteeship: Kitschelt et al. 1999, 82; for an overview: Valen and Narud 2007, 296), or all legislators are shifted to the left of their voters (“Elite-mass displacement” Putnam 1976, 118) or (less often), shifted to the right of their voters (Dalton 1985, 275; Esaiasson and Holmberg 1996, 94; Andeweg 2011, 47). While polarization between legislators indicates the predominance of the competitive struggle for votes, the tendency toward the median signifies pronounced antagonistic cooperation. The left–right shift can be partially explained by the distinct social composition of the legislators—especially their higher educational status—but there seems to be additional influence of individuals’ status as legislators (McAllister 1991, 259)  (i.e. distance arises out of the different positions in the prinicipal–agent constellation). The aforementioned studies focus on policy congruence with regard to both issues and ideology. In contrast, congruence in preferences regarding political processes and institutions has been rarely analysed for legislators. However, the interest in what might be termed “congruence concerning politics” has increased lately, as it appears that the evaluation of politicians and the political system is not only driven by (in)congruence regarding policies, but, increasingly, by the extent to which legislators and citizens agree or disagree on “process preferences” (i.e. preferences about how politics ought to be conducted; Hibbing and Theiss-Morse 2001). In recent decades the distribution of such process preferences among citizens has moved in the direction of increased demand for direct political participation and legislators’ responsiveness (Mair 2011). Against this background, the decreasing political support and trust in democratic institutions, actors and processes (Dalton 2004) may be seen as being driven by a lack of congruence regarding political processes. Despite these evident changes, only a few recent studies have analysed the process preferences of citizens (Carman 2006; Carman 2007; Bengtsson and Wass 2011) while even fewer studies compare these preferences to those of legislators (Méndez-Lago and Martínez 2002; Bengtsson and Wass 2012; Esaiasson and Holmberg 1996). With regard to basic principles and liberties, political elites and legislators are generally more supportive of core elements of liberal democracy than the general population (McClosky and Brill 1983; Pettersson 2010, 132). However, it has been shown repeatedly that legislators are less favourably disposed than the general population to direct political participation, which would most severely limit their discretion in the principal–agent relationship outlined above. In addition to such fundamental normative principles, the process preferences of legislators regarding their own behaviour are of decisive importance, because the population perceives institutions such as legislatures primarily through the observable behaviour of their members, which thereby constitutes the pivotal benchmark for constituents’ evaluations of both their elected legislators and the legislature. Some studies in this area, in the tradition of Eulau et al. (1959), centre on the focus and style of representation. These studies have demonstrated high levels of congruence regarding process preferences between the general population and legislators for Finland (Bengtsson and Wass

The Sociology of Legislators and Legislatures  

73

2012), contrasted with low levels of congruence for Spain (Méndez-Lago and Martínez 2002) and Sweden (Esaiasson and Holmberg 1996). Andeweg and Thomassen (2005) analysed new dimensions of role orientations. They found that, in the Netherlands, clear majorities in both the general population and the body of legislators demand that political control is conducted based on mandates given ex-ante in elections, as opposed to control via ex-post evaluation. However, the general population and legislators disagree on the impact that legislators should have on the formation process of political preferences: while the population prefers responsiveness, legislators tend to favour political leadership. One recent study on Germany (Best and Vogel 2011) analysed attitudes toward a wide range of politics and found that legislators generally prefer a narrow scope of state responsibilities, with fully professional politicians organized in parties and exerting strong political leadership. In contrast, the general population tends to support a large scope for state responsibilities, with politicians being primarily responsive to their constituents, less connected to political parties, and conducting politics merely as a part-time occupation. A recent comparison of legislators and the general population in almost all member states of the European Union revealed that legislators are much more Europeanized than their respective populations, especially with regard to the preferred level of integration (i.e. the politics of competence allocation). Legislators evaluate the allocation of policy competences at both the EU and national level differently, considering whether the assumed benefits of the status quo or the reallocation of competences in a certain policy area would lead to their empowerment (Best 2012b, 240). In contrast, the population prefers competences to be allocated at the level—usually national—which preserves constituent influence on their legislators (Best 2012a, 214). To summarize, the relation between legislators and their constituents is shaped by the underlying social constellation. Diverging preferences regarding politics are due to both the asymmetric interdependence of the principal–agent relation and the necessity for antagonistic cooperation. This constellation generally results in legislators who emphasize their personal autonomy and prerogative for flexible behaviour and who are more supportive of political systems that grant them this autonomy. Disagreement between legislators and citizens regarding politics (politics gap) is thus inevitable as it is rooted in the social foundation of their relationship. As long as politics remain a low priority on the public agenda, legislators are able to detract from this politics gap by referring to policy (especially economic outcomes) that they present as being direct outcomes of their professional political competence. However, as popular demand for control and responsiveness of legislators increases, and the ability of legislators to deliver their constituents’ desired outcomes decreases, the politics gap may develop into a legitimacy challenge (Best 2007, 108–11). With regard to policies, it seems necessary to distinguish between two situations. In the first case, preferences are organized along ideological lines and rooted in identifiable social groups, such as homogenous districts or parties. In this case, relative congruence between legislators and their constituents in districts and parties is achievable. Relative congruence is less likely if issues are loosely connected to ideology; in such cases, constituents are heterogeneous and therefore do not provide sufficient cues for their legislators. In any case, absolute congruence (i.e. identity of

74   Heinrich Best and Lars Vogel legislators and constituents) is exceptional. This pattern of simultaneous relative correspondence and absolute deviation from the population indicates that the behaviour of legislators is not determined by the preferences of their constituents, but rather that their behaviour is jointly constrained by their constituents and by the positioning of competing parties. Within this space, legislators can use the asymmetry in the principal–agent relation to impact the preferences to which they are, in turn, responsive. Political leadership conducted by legislators is influenced by the internal relations among parliamentary contenders. If the demand for antagonistic cooperation is of primary interest, legislators may use their autonomy to attenuate policy differences. In contrast, legislators may emphasize policy differences if distinctiveness in the competitive struggle for votes is their priority. In turn, both antagonistic cooperation and the competitive struggle between legislators are influenced by the principal–agent relation: close control by constituents may result in a reluctance to engage in antagonistic cooperation, while inactive and heterogeneous constituents facilitate legislators’ convergence. There have been only a few empirical studies on the reciprocal influence of insider–insider and insider–outsider relations. After German unification, Eastern German legislators—with some exceptions among the Post-Communists—at the national and sub-national levels tended to adapt to their Western counterparts in terms of politics and, less consistently, in terms of policies. This rapid integration, which took place without much overt reluctance, seems to have enlarged the distance today between legislators from Eastern Germany and their constituents to an extent well beyond that which has been common in Western Germany. This distance is posited to be one of the driving forces for the lower levels of political support in the present day among the general population in Eastern Germany, as compared to Western Germany (Best and Vogel 2012b). One of the strategies that could be pursued by legislators to cope with the increasing demand for responsiveness, coupled with the difficulty in providing the desired political and economic outcomes (Wessels 2011, 96), would be to intensify antagonistic cooperation, with the goal of attenuating the competitive struggle for votes and thus lowering the risk of electoral defeat. The employment of such a strategy could ease insider–insider relations, but would presumably lead to deterioration in insider–outsider relations. Legislatures are the central institution in which legislators interact, with their internal rules and norms structuring the conditions for antagonistic cooperation. If both insider–insider and insider–outsider relations are considered to be interdependent, studying insider–insider relations in parliament would contribute to the understanding of the linkages between legislators and the general population in representative democracies.

3.5 Conclusion The sociology of legislators and legislatures is, by its very nature, an interdisciplinary approach which investigates the social interactions, groups, and institutions involved

The Sociology of Legislators and Legislatures  

75

in the process of parliamentary representation. From this point of view the institutions of representative democracies are not just a given, but are the creation of the people running them: representatives are therefore both the creations and creators of the institutional settings in which they act. This double perspective gives credit to both rational-choice theory and new institutionalism. According to the rational-choice framework, institutions are presumed to provide systems of rules and regulations that constrain and enable behaviour by determining the strategies that actors chose to pursue their otherwise fixed and exogenous ambitions and preferences. The sociological approach does not deny that legislators make their choices in a rational manner; however, it focuses on the volatile and endogenous character of preferences and choices that emerge out of social constellations and may be transformed by the interactions of legislators. New institutionalism, which identifies institutions as the main context framing the preferences of the actors involved, is extended by sociological approaches maintaining that the stability and proper function of institutions depends on the social relations and interactions taking place within them. Three types of social relations underlie and constitute the interactions of legislators both inside and outside of parliament: the competitive struggle for votes, antagonistic cooperation, and principal–agent relations. Those relations are intertwined and determine the behaviour of legislators vis-á-vis their fellows, their competitors, and their constituents. Through the process of institutional socialization, antagonistic cooperation is learned and deepened. Patterns of legislative recruitment indicate that the logic of antagonistic cooperation leads to social closure of access to legislatures, transforming parliaments into internal labour markets for parties, and impeding the competitive struggle for votes. However, recent changes in recruitment patterns can be interpreted as an indication of the emergence of a legitimacy challenge and a countermovement against political professionalization. Such developments may be founded in the dynamics of the principal–agent relation between legislators and constituents, by which legislators are keen to enlarge their own autonomy while constituents tend to restrict it. When there is an increased awareness of moral hazards, legislators tend to be put on a shorter leash. Political parties have responded to this legitimacy challenge by opening access to parliament for new brands of politicians and by limiting tenure. There is an intrinsic gap between principals and agents regarding the preferences of how politics should be pursued, but legislators are only constrained and not entirely determined by the policy preferences of their constituents. Leadership by legislators and parties is thus facilitated. While much is known about these processes at the aggregate level, little investigation has been conducted on how principal–agent relations and antagonistic cooperation translate into actual social interactions and vice versa. Because of the contingencies between insider–insider relations and insider–outsider relations these interactions are embedded in highly complex contexts. Legislators and parties are faced with the dilemma of either optimizing their responsiveness to their constituents while eroding internal cooperation with their political opponents, or boosting internal cooperation while distancing themselves from their constituents. Stable representative democracies are therefore institutional frameworks and informal arrangements which achieve an

76   Heinrich Best and Lars Vogel equilibrium between the competing demands resulting from two divergent social relations. How this situation affects the daily interactions of legislators is largely unknown. Little systematic knowledge is available about informal contacts besides and beyond formal institutional or organizational settings, although the importance of bonding between rivals has been shown in earlier research (Best 2010a, 112). The sociology of legislators and legislatures also suffers from a lack of longitudinal data, which would allow a better understanding of the social mechanisms underlying the emergence, adaptation, and transformation of preferences, norms, and rules (i.e. for cooperation between competitors) prevalent among legislators. Further research is also needed to understand how the internet influences relations between legislators and their constituents. In contrast to those who see the internet as a one-way road to a more participatory e-democracy (Browning 1995; White 1997; for an overview: Coleman 2005; Ward and Gibson 2009), a sociological approach to this area of inquiry would shed light on how information technology is superimposed by the logic of existing social relations between its users. Considering the logic and the dynamics of principal–agent relations, the internet has the potential to establish and strengthen the direct links between legislators and constituents, in addition to the processes that take place via established channels such as elections and political parties. In this way, the internet might increase the ability of principals to control their agents. However, keeping in mind the underlying social logic of the principal–agent relation, legislators will presumably act, at the same time, to prevent any restriction on their autonomy. One can therefore expect the internet activities of legislators to be shaped by the social logic of the insider–outsider relation, rather than solely by the technological potentials of facilitated internet communication.

References Andeweg, R. B., 2011. Approaching Perfect Policy Congruence: Measurement, Development, and Relevance for Political Representation. In M. Rosema, B. Denters, and K. Aarts (eds.). How Democracy Works. Political Representation and Policy Congruence in Modern Societies, pp. 39–52. Amsterdam: Pallas Publications. Andeweg, R. B. and Thomassen, J. J. A., 2005. Modes of Political Representation: Toward a New Typology. Legislative Studies Quarterly, 30, 507–28. Asher, H. B., 1973. The Learning of legislative Norms. American Political Science Review, 67, 499–513. Badura, B. and Reese, J., 1976. Jungparlamentarier in Bonn-Ihre Sozialisation im Deutschen Bundestag. Stuttgart: Fromann-Holzboog. Barnes, S. H., 1977. Representation in Italy:  Institutionalized Tradition and Electoral Choice.Chicago: Chicago University Press. Beaver, D. and Rosen, R., 1978. Studies in the Scientific Collaboration Part I: The Professional Origins of Scientific Co-Authorship. Scientometrics, 1,65–84. Bell, C. G. and Price, C. M., 1975. The First Term: A Study in Legislative Socialization. Beverly Hills: Sage.

The Sociology of Legislators and Legislatures  

77

Bengtsson, Å. and Wass, H., 2011. The Representative Roles of MPs: A Citizen Perspective. Scandinavian Political Studies, 34,143–67. Bengtsson, Å. and Wass, H., 2012. Congruence between MPs’, Non-elected Candidates’ and Citi­zens’ Preferences for Representational Roles. Annual meeting of the American Political Science Association. 30 August–2 September, New Orleans. Best, H., 2003. Der langfristige Wandel politischer Eliten in Europa 1867–2000. Auf dem Weg der Konvergenz? In S. Hradil and P. Imbusch (eds.). Oberschichten—Eliten—herrschende Klassen, pp. 369–400.Opladen: Leske + Budrich. Best, H., 2007. New Challenges, New Elites? Changes in the Recruitment and Career Patterns of European Representative Elites. Comparative Sociology, 6, 85–113. Best, H., 2010a. Associated Rivals:  Antagonism and Cooperation in the German Political Elite. In J. Higley and H. Best (eds.) Democratic Elitism: New Theoretical and Comparative Perspective, pp. 97–116. Leiden, Boston: Brill. Best, H., 2010b. Transitions, Transformations and the Role of Elites. In H. Best, K. Blum, M. Fritsch, and R. K. Silbereisen (eds.). Transitions—Transformations:  Trajectories of Social, Economic and Political Change after Communism (Special Issue 35, Historical Social Research), pp. 9–12. Köln: Center for Social Research. Best, H., 2012a. Elite foundations of European Integration:  A Causal Analysis. In H. Best, G. Lengyel and L. Verzichelli (eds.). The Europe of Elites. A Study into the Europeanness of Europe’s Political and Economic Elites, pp. 208–33. Oxford: Oxford University Press. Best, H., 2012b. Elites of Europe and the Europe of Elites: A Conclusion. In H. Best, G. Lengyel, and L. Verzichelli (eds.). The Europe of Elites. A  Study into the Europeanness of Europe’s Political and Economic Elites, pp. 234–41. Oxford: Oxford University Press. Best, H. and Cotta, M., 2000. Elites Transformation and Modes of Representation since the Mid-Nineteenth Century:  Some Theoretical Considerations. In H. Best and M. Cotta (eds.).  Parliamentary Representatives in Europe 1848–2000. Legislative Recruitment and Careers in Eleven European Countries, pp. 1–28. Oxford: Oxford University Press. Best, H. and Vogel, L., 2011. Politische Eliten im vereinten Deutschland. Strukturen— Einstellungen—Handlungsbedingungen. In A. Lorenz (ed.). Ostdeutschland und die Sozialwissenschaften. Bilanz und Perspektiven 20 Jahre nach der Wiedervereinigung, pp. 120–52. Berlin: Budrich. Best, H. and Vogel, L., 2012a. The Emergence and Transformation of Representative Roles. In O. Rozenberg and M. Blomgren (eds.). Parliamentary Roles in Modern Legislatures, pp. 37– 65. Routledge. Best, H. and Vogel, L., 2012b. Zweimal Deutsche Vereinigung: System- und Sozialintegration der politischen Eliten nach 1871 und 1990 im Vergleich. In H. Best and E. Holtmann (eds.). Aufbruch der entsicherten Gesellschaft. Deutschland nach der Wiedervereinigung, pp. 85–103. Frankfurt, New York: Campus Verlag. Blondel, J., 1997. Political Progress. Reality or Illusion? In A. Burgen, P. McLaughin,and J. Mittelstraß (eds.). The Idea of Progress, pp. 77–101. Berlin, New York: de Gruyter. Borchert, J. and Zeiss, J. (eds.), 2003. The Political Class in Advanced Democracies.Oxford: Oxford University Press. Brettschneider, F., 1995. Öffentliche Meinung und Politik.Opladen: Westdeutscher Verlag. Browning, G., 1995. Electronic Democracy: Using the Internet to Influence American Politics. Wilton: Pemberton Press. Carman, C. J., 2006. Public Preferences for Parliamentary Representation in the UK:  An Overlooked Link? Political Studies, 54, 103–22.

78   Heinrich Best and Lars Vogel Carman, C. J., 2007. Assessing Preferences for Political Representation in the US. Journal of Elections, Public Opinions and Parties, 17,1–19. Coleman, S., 2005. New Mediation and Direct Representation:  Reconceptualizing Representation in the Digital Age. New Media & Society, 7(2),177–98. Converse, P. E. and Pierce, R., 1986. Political Representation in France.Cambridge, MA: Harvard University Press. Cotta, M. and Almeida, P. T., 2007. From Servants of the State to Elected Representatives: Public SectorsBackground among Members of Parliament. In M. Cotta and H. Best (eds.). Democratic Representation in Europe. Diversity, Change, and Convergence, pp. 51–76. Oxford: Oxford University Press. Cox, G. W., 2000. On the Effects of Legislative Rules. Legislative Studies Quarterly, 25, 169–92. Dalton, R., 2004. Democratic Challenges. Democratic Choices. The Erosion of Political Support in Advanced Industrial Democracies. Oxford: Oxford University Press. Dalton, R. J., 1985. Political Parties and Political Representation. Party Supporters and Party Elites in Nine Nations. Comparative Political Studies, 18 (3), 267–99. Dalton, R. J., Farrell, D. M., and McAllister, I., 2011. The Dynamics of Political Representation. In M. Rosema, B. Denters, and K. Aarts (eds.). How Democracy Works. Political Representation and Policy Congruence in Modern Societies, pp. 21–38. Amsterdam: Pallas Publications. Davidson, R. H., 1969. The Role of the Congressman.New York: Pegasus. Esaiasson, P. and Holmberg, S., 1996. Representation from Above: Members of Parliament and Representative Democracy in Sweden. Aldershot: Dartmouth. Eulau, H., Wahlke, J. C., Buchanan, W., and Ferguson, L. C., 1959. The Role of the Representative: Some Empirical Observations of the Theory of Edmund Burke. American Political Science Review, 53, 742–56. Ezrow, L., De Vries, C., Steenbergen, M., and Edwards, E., 2011. Mean Voter Representation and Partisan Constituency Representation: Do Parties Respond to the Mean Voter Position or to their Supporters? Party Politics, 17,275–301. Farah, B., 1980. Political Representation in West Germany. Ann Arbor: University of Michigan, unpublished Dissertation. Fenno, R., 1962. The House Appropriations Committee as a Political System: The Problem of Integration. American Political Science Review, 56(2), 310–24. Franklin, M. N. and Scarrow, S. S., 1999. Making Europeans? The Socializing Power of the European Parliament. In R. Katz and B. Wessels (eds.). The European Parliament, the National Parliaments, and European Integration, pp. 45–60. Oxford: Oxford University Press. Friesema, H. P. and Hedlund, R. D., 1974. The Reality of Representational Roles. In N. R. Luttbeg (ed.). Public Opinion and Public Policy, pp. 413–17. 2nd ed. Homewoos: Dorsey Press. Harden, J. J. and Carsey, T. M., 2012. Balancing Constituency Representation and Party Respon­ siveness in the US Senate: The Conditioning Effect of State Ideological Heterogeneity. Public Choice, 150,137–54. Hedlund, R. D., 1968. Legislative Socialization and Role Orientations: A Study of Iowa Legislature. Ann Arbor. Hetherington, M., 2001. Resurgent Mass Partisanship: The Role of Elite Polarization. American Political Science Review, 95,619–31. Hibbing, J. R. and Theiss-Morse, E., 2001. Process Preferences and American Politics: What the People want Government to Be. American Political Science Review, 95, 145–53. Huber, J. D. and Powell, B. G., 1994. Congruence between Citizens and Policymakers in Two Visions of Liberal Democracy. World Politics, 46, 291–326.

The Sociology of Legislators and Legislatures  

79

Hurley, P. A. and Hill, K. Q., 2003. Beyond the Demand-input Model:  A  Theory of Representational Linkages. Journal of Politics, 65, 304–26. Jacobs, L. R. and Shapiro, R. Y., 1994. Studying Substantive Democracy. PS: Political Science and Politics, 27, 9–17. Katz, R. and Mair, P., 1995. Changing Models of Party Organization and Party Democracy. The Emergence of the Cartel Party. Party Politics, 1(1):5–28. Kitschelt, H., Mansfeldova, Z., Markowski, R., and Tóka, G., 1999. Post-Communist Party Systems. Competition, Representation, and Inter-party Cooperation. Cambridge: Cambridge University Press. Körösényi, A., 2010. Beyond the Happy Consensus about Democratic Elitism. In H. Best and J. Higley (eds.). Democratic Elitism. New Theoretical and Comparative Perspectives, pp. 43– 60. Leiden, Boston: Brill. Kuklinski, J. H. and Elling, R. C., 1977. Representational Role, Constituency Opinion, and Legislative Roll-Call Behaviour. American Political Science Review, 21, 135–47. Lasswell, H. D., 1954. The Elite Concept. In H. D. Lasswell, D. Lerner, and C. E. Rothwell (eds.). The  Comparative Study of Elites. An Introduction and Bibliography, pp. 6–21. Stanford: Stanford University Press. Leston-Bandeira, C., 2012. Studying the Relationsship between Parliament and Citizens. Journal of Legislative Studies, Special Issue: Parliaments and Citizens, 18, 265–74. Levi, M., 1990. A Logic of Institutional Change. In K. S. Cook and M. Levi (eds.). Limits of Rationality, pp. 402–18. Chicago: University of Chicago Press. Luna, J. and Zechmeister, E., 2005. Political Representation in Latin America—A Study of Elite-mass Congruence in Nine Countries. Comparative Political Studies, 38, 388–416. Lupia, A., 2006. Delegation and its Perils. In K. Strøm, W. C. Müller, and T. Bergman (eds.). Delegation and Accountability in Parliamentary Democracies, pp. 33–54. Oxford:  Oxford University Press. Mair, P., 2011. Is Governing Becoming more Contentious? In M. Rosema, B. Denters, and K. Aarts (eds.). How Democracy Works. Political Representation and Policy Congruence in Modern Societies, pp. 77–86. Amsterdam: Pallas Publications. Mansbridge, J., 2003. Rethinking Representation. American Political Science Review, 97, 515–28. Matthews, D. R., 1959. The Folkways of the United States Senate. American Political Science Review, 53, 1064–89. McAllister, I., 1991. Party Elites, Voters and Political Attitudes: Testing Three Explanations for Mass-Elite Differences. Canadian Journal of Political Science/Revue canadienne de science politique, 24, 237–68. McClosky, H. and Brill, A., 1983. Dimensions of tolerance. What Americans believe about civil liberties. New York: Russell Sage Foundation. McCrone, D. J. and Kuklinski, J. H., 1979. The Delegate Theory of Representation. American Journal of Political Science, 23, 278–300. Méndez-Lago, M., and Martínez, A., 2002. Political Representation in Spain: An Empirical Analysis of the Perception of Citizens and MPs. The Journal of Legislative Studies, 8, 63–90. Michels, R., 1915 [1911]. Political Parties. A Sociological Study of the Oligarchical Tendencies of Modern Democracies. New York: Free Press. Miller, W. E., Pierce, R., Thomassen, J., Herrera, R., Holmberg, S., Esaiasson, P., and Wessels, B. (eds.). 1999. Policy Representation in Western Democracies. Oxford: Oxford University Press. Miller, W. E. and Stokes, D. E., 1963. Constituency Influence in Congress. American Political Science Review, 57,45–56.

80   Heinrich Best and Lars Vogel Mosca, G., 1939 [1911]. The Ruling Class. New York: McGraw Hill. Mughan, A., Box-Steffensmeier, J., and Scully, R., 1997. Mapping Legislative Socialization. European Journal of Political Research, 33, 93–106. Norris, P., 1997. Introduction. Theories of Recruitment. In P. Norris (ed.). Passages to Power: Legislative Recruitment in Advanced Democracies, pp. 1–14. Cambridge: Cambridge University Press. Norris, P. and Lovenduski, J., 1995. Political Recruitment: Gender, Race, and Class in the British Parliament. Cambridge, New York, Melbourne: Cambridge UniversityPress. Norton, P., 2002. Introduction:  Linking Parliament and Citizens. In P. Norton (ed.). Parliaments and citizens in Western Europe (Parliaments in contemporary Western Europe, Vol. 3), pp. 1–18. London: Cass. Page, B. I., Shapiro, R. Y., Gronke, P. W., and Rosenberg, R. M., 1984. Constituency, Party, and Representation in Congress. The Public Opinion Quarterly, 48, 741–56. Patzelt, W. J. and Schirmer, R., 2002. Die Volkskammer der DDR. Sozialistischer Parlamentarismus in Theorie und Praxis.Wiesbaden: Westdeutscher Verlag. Pettersson, T., 2010. Pro-democratic Orientations, Political Shortcuts and Policy Issues: Comparative Analyses of Elite-mass Congruence in Old and New Democracies. In. U. J. van Beek (ed.) Democracy under Scrutiny: Elites, Citizens, Cultures, pp. 117–46. Opladen: Verlag Barbara Budrich. Pierson, P., 2004. Politics in Time: History, Institutions, and Social Analysis. Princeton: Princeton UniversityPress. Pitkin, H. F., 1967. The Concept of Representation. Berkeley: Berkeley University Press. Powell, L. W., 1982. Issue Representation in Congress. Journal of Politics, 44,658–78. Putnam, R. D., 1976. The Comparative Study of Political Elites. Englewood Cliffs: Prentice-Hall. Reiser, M., Hülsken, C., Schwarz, B., and Borchert, J., 2011. Das Reden der Neulinge und andere Sünden. Parlamentarische Sozialisation und Parlamentskultur in zwei deutschen Landtagen. Zeitschrift für Parlamentsfragen, 4, 820–34. Rohrschneider, R. and Whitefield, S., 2012. The Strain of Representation. How Parties Represent Diverse Voters in Western and Eastern Europe. Oxford: Oxford University Press. Rokkan, S., 1999. State Formation, Nation-Building, and Mass Politics in Europe. The Theory by Stein Rokkan; Based on his Collected Works (edited by P. Flora with S. Kuhnle and D. Urwin). Oxford: Oxford University Press. Sartori, G., 1987. Democratic Theory Revisited.Chatham: Chatham House. Schumpeter, J. A., 1959 [1940]. Capitalism, Socialism and Democracy.London: George Allen and Unwin. Scully, R., 2005. Becoming Europeans? Attitudes, Behaviour, and Socialization in the European Parliament.Oxford: Oxford University Press. Searing, D. D., 1986. A Theory of Political-socialization—Institutional Support and Deradicalization in Britain. British Journal of Political Science, 16, 341–76. Shapiro, R. Y., 2011. Public Opinion and American Democracy. Public Opinion Quarterly, 75, 982–1017. Soroka, S., Penner, E., and Blidook, K., 2009. Constituency Influence in Parliament. Canadian Journal of Political Science/Revue Canadienne De Science Politique, 42, 563–91. Steenbergen, M. R., Edwards, E. E., and de Vries, C. E., 2007. Who’s Cueing Whom? Mass-elite Linkages and the Future of European Integration. European Union Politics, 8, 13–35. Stimson, J. A., Mackuen, M. B., and Erikson, R. S., 1995. Dynamic Representation. American Political Science Review, 89, 543–65.

The Sociology of Legislators and Legislatures  

81

Stolz, K., 2010. Towards a Regional Political Class? Professional politicians and regional institutions in Catalonia and Scotland. Manchester: Manchester University Press. Stone, W. J., 1982. Electoral Change and Policy Representation in Congress—Domestic Welfare Issues from 1956–1972. British Journal of Political Science, 12, 95–115. Strøm, K., Müller, W. C., and Bergman, T., 2006. Delegation and Accountability in Parliamentary Democracies.Oxford: Oxford UniversityPress. Thomassen, J., 1994. Empirical Research into Political Representation: Failing Democracy or Failing Models? In M. K. M. Jennings, and E.Thomas (eds.). Elections at Home and Abroad. Essays in Honour of Warren E.  Miller, pp. 237–64. Ann Arbor:  University of Michigan Press. Valen, H. and Narud, H. M., 2007. The Conditional Party Mandate: A Model for the Study of Mass and Elite Opinion Patterns. European Journal of Political Research, 46, 293–318. Wahlke, J. C., 1971. Policy Demands and System Support—Role of Represented. British Journal of Political Science, 1, 271–90. Ward, S. and Gibson, R., 2009. European Political Organization and the Internet: Mobilization, Participation, and Change. In A. Chadwick and P. N. Howard (eds.). Routledge handbook of Internet politics, pp. 25–39. London, New York: Routledge. Weber, M. 1958 [1919]. Politics as a Vocation. In H. H. Gerth and C. W. Mills (eds.). Max Weber: Essays in Sociology, pp. 77–128. New York: Oxford University Press. Weissberg, R., 1978. Collective vs. Dyadic Representation in Congress. American Political Science Review, 72, 535–47. Welzel, C., 2002. Effective Democracy, Mass Culture and the Quality of Elites: The Human Devel­opment Process. International Journcal of Comparative Sociology, 43, 317–40. Wessels, B., 1999. System Characteristics Matter. Empirical Evidence from Ten Representation Studies. In W. E. Miller, R. Pierce, J. Thomassen, R. Herrera, S. Holmberg, P. Esaiasson, and B. Wessels (eds.).Policy Representation in Western Democracies, pp. 137–61. Oxford: Oxford University Press. Wessels, B., 2011. Performance and Deficits of Present-Day-Representation. In S. Alonso, J. Keane, and W. Merkel (eds.). The Future of Representative Democracy, pp. 96–123. Cambridge: Cambridge University Press. White, C. S., 1997. Citizen Participation and the Internet. Prospects for Civic Deliberation in the Information Age. Social Studies, 88, 23–28. Wlezien, C., 1995. The Public as Thermostat: Dynamics of Preferences for Spending. American Journal of Political Science, 39, 981–1000.

C HA P T E R

4

T YPOLOGIES AND C L A S S I F I C AT I O N S A M I E K R E PPE L

4.1 Introduction Legislatures exist in almost every country of the world so it is not surprising that scholars have long been interested in studying them and comparing them. However, the diversity among legislatures has led to the proliferation of different methods of comparison, each concentrating on different aspects of legislatures, although all, to one degree or another, include an assessment of the relative power of the legislature within the broader political context. What varies most are the attributes that are considered vital elements of a legislature’s power. Many analyses focus on the institutional characteristics of political systems as a whole and the legislatures themselves. Other studies are more concerned with the various functions performed by legislatures within the broader political system. In other words, some analyses are focused on what legislatures do, while others concentrate more on how they do it. In general, the most significant contributions to the legislative studies literature implicitly or explicitly recognize that these two aspects of legislatures are intrinsically linked. In particular, the connection between structures and their impact on the capacity of legislatures to successfully engage in the full gambit of functions generally ascribed to them is at the core of comparative legislative studies. The result of many of these analyses, particularly in the pantheon of “classical” comparative legislative research, is the creation of a typology or classification scheme that groups legislatures into various categories.1 While this is a useful exercise in that it facilitates easy comparisons and a common language for the labeling of legislatures, in most cases this approach provides little in the way of explanation. Classification schemes allow us to label types of legislatures and compare their relative powers, tasks, and structures, but they do not provide a comprehensive theoretical explanation of why legislatures are of one type or another. Moreover, classification schemes tend to be static. Legislatures are categorized as one type or another—with little possibility for shifts in

Typologies and Classifications  

83

classification to occur across time, particularly within shorter periods such as from one election to the next. This rigid character fundamentally limits the utility of most typologies, particularly when studying dynamic legislatures or less stable political systems. The goal of this chapter is to revisit the primary legislative typologies and classification schemes still prevalent in the field and examine the extent to which these can be collectively expanded and embedded within a coherent theoretical framework that moves beyond the classification of different legislatures towards an explanation of why they differ. In particular, this chapter will present a theoretical framework that provides both a classification scheme to enable easy comparisons and a theoretical explanation of legislative power that is flexible enough to explain changes in power and influence over time, even in the absence of systemic institutional change. Before we can move forward in accomplishing this task, however, it will be necessary to establish some common definitions of key terms to ensure semantic clarity.

4.2  Establishing Common Terminology: The First Classification Scheme One of the most problematic aspects of trying to study legislatures comparatively is the profusion of terms such as “assembly,” “congress,” or “parliament” that are often used interchangeably with the term “legislature.” Since this lack of semantic rigor can lead to confusion and misperceptions about the underlying character of the legislature being analysed (as well as the broader political system in which it functions), it is worthwhile to develop some clear definitions and basic terminological classifications. Of the words frequently used to describe a legislature,“assembly” is the most general. Indeed, assemblies need not be explicitly political and the word (uncapitalized) refers simply to the coming together of a group of people for some purpose; for example, a school assembly. It is only when we implicitly or explicitly add the qualifier “political” or “legislative” that we begin to think of assemblies in the same context as legislatures, parliaments, and congresses. By the same token, parliaments and congresses, generically, can be best understood as specific types of legislatures. This understanding of these four terms creates a hierarchy of institutions from the most general (an assembly) to the most specific (congress and parliament) which are understood as types of the mid-level category of “legislature” (Fig. 4.1).2 From this perspective, legislatures are those assemblies that are organized for specifically political, and to some degree legislative, purposes. While this definition serves to differentiate legislatures from other assemblies, it is still a very general category. To be able to effectively compare legislatures cross-nationally we need additional definitions and clarification to help us effectively distinguish between the two primary types of legislature: parliament and congress.3 It is fundamentally the structural characteristics of

84   Amie Kreppel

ASSEMBLIES

LEGISLATURES (Legislative Assemblies)

parliaments

FIG.

congresses

4.1  A  hierarchy of institutions

the political systems in which these two types of legislature are located that distinguishes between them.4 Regardless of whether or not a political system can be categorized as democratic, if there is a legislature as well as an executive branch, it is the relationship between these two that will determine the core characteristics of the legislature and determine whether it is a parliament or a congress. The central determinative characteristic is the relative level of interdependence between the branches. In what are commonly referred to as parliamentary systems, the executive branch (prime minister and cabinet) is elected by the legislature, usually from within its own ranks.5 Furthermore, the executive branch or “government” is formally responsible to the legislature throughout its tenure. This means that it can be removed from office at any time should a majority of the legislature oppose it, regardless of the electoral cycle.6 In turn, removal of the executive by the legislature may be accompanied by early elections for the legislature, thus creating the possibility for “mutually assured destruction” for both branches. Thus, there is a high degree of mutual dependence between the government and the legislature. For this reason these types of systems are known generically as “fused-powers” systems and their legislatures are usually referred to as “parliaments,” regardless of their formal national title. In contrast, within what are popularly referred to as presidential systems the executive is elected by the people (generally through direct election), rather than by the legislature. Thus, unlike parliamentary systems, there is no direct tie of responsibility between the executive and legislative branches, only a very limited possibility of removal of the executive by the legislature impeachment), and no provision for the dissolution of the legislature by the executive branch. For this reason presidential systems are considered

Typologies and Classifications  

85

“separation-of-powers” systems in the sense that the core institutional branches themselves are separate from one another. Although separation-of-powers systems are most often presidential in character, this is not a requirement and non-presidential separation-of-powers systems exist (e.g. the European Union). Unlike legislatures in fused-powers system, there is no commonly accepted general category for legislatures in separation-of-powers systems, though many have the formal title of “congress.” However, “congress,” like “parliament,” can also be understood as a general type of legislature, with the defining characteristic being the relative independence of the executive and legislative branches. Although this differentiation is intended only to provide semantic clarity, it does link to one of the more common metrics used to evaluate legislatures in many typologies. Both selection and oversight of the executive are key elements of the “control function” often exercised by legislatures, but only the latter is employed by congress-type legislatures. A brief review of the core functions of legislatures and the institutional characteristics that are often linked to them will facilitate a better understanding of the logic supporting most existing typologies and classification schemes and provide the foundation for the analysis that follows.7

4.3  Historical Evolution of Legislative Typologies and Classifications Traditionally, legislatures are tasked with four primary functions: linkage; representation; control/oversight; and policy-making (Bagehot 1872; Blondel 1973; Loewenberg and Patterson 1979; Olson 1994; Loewenberg 2011). Most legislatures engage in all four of these tasks to some degree, but what varies between them is their relative level of power or influence within each arena and the extent to which each aspect is prioritized, though these are clearly linked. A quick review of what each of these tasks entails will serve to underscore their significance in understanding and evaluating the role of legislatures within the broader political system and the extent to which existing typologies incorporate one or more of these core tasks in their assessments of legislative power and influence. Perhaps the most critical task of legislatures in terms of democratic theory is to link citizens to the government. Even when the legislature is weak in terms of its other roles, it is almost always able to serve “as an intermediary between the constituency and the central government” (Olson 1980, 135). In this context, legislatures act as a conduit of information (through their individual members) allowing local level demands to be heard by the central government and the policies and actions of the central government to be explained to citizens. Although all legislatures perform this task to a certain extent, their ability to serve as effective tools of communication, as well as the relative importance of this role can vary significantly. The linkage role of the legislature will

86   Amie Kreppel be more important in political systems in which citizens do not have the opportunity to communicate their preferences and satisfaction directly to the executive branch through the electoral process. Thus, in parliamentary systems, as well as in most non-democratic regimes, the linkage function of the legislature is likely to be more important because it may be the only stable mechanism of communication between citizens and the executive. In addition to providing a link between citizens and government, the individual members of a legislature are also expected to represent their constituents. In other words, legislators are responsible for advocating for their constituents in their stead, insuring that the opinions, perspectives, and values of citizens are present in the policy-making process (Pitkin 1967). Legislators are the primary conduits of political representation for citizens. The comparatively large number of members within most legislatures (as compared to the executive branch) and the relatively small scale of their electoral districts make the effective representation of a variety of different minority communities possible. In contrast, smaller minority groups may have a difficult time accessing or influencing the executive branch as a result of the latter’s need to represent the interests of the country as a whole.8 Unlike the executive branch, which generally includes a small number of actors who share unique (or at least similar) political beliefs, legislatures are larger and, in all but the most repressive systems, contain representatives of the governing parties and opposition groups. This plural characteristic of legislatures not only increases their representative capacity, it also enables them to serve as a public forum of debate, in which diverse opinions and opposing views can directly engage with one another with the goal of influencing public opinion and policy outcomes. The ability of the governed to control the government is one of the foundational tenets of representative democracy.9 The primary tool used to achieve this goal is regularly scheduled free and fair elections. The type of executive control practiced by the legislature is directly linked to the nature of the relationship between voters and the executive branch and between the legislature and the executive branch.10 The character of the control function of the legislature is much more prominent in fused-powers systems than it is in separation-of-powers systems. In the former the legislature has the power of investiture and/or censure, while in the latter executive control is limited to confirmation (cabinet) and impeachment. Although control generally refers to selection and removal of the executive, legislative oversight of the executive branch is a much broader concept, entailing the monitoring of executive agencies tasked with the implementation of policy decisions, and regular engagement with the political executive to insure it is meeting its commitments to the public and adequately addressing the various policy needs of the country. Legislatures also play a critical role in insuring proper oversight of the executive branch through budgetary authority and the right to review the budgetary implications of policies. Legislatures may even be able to exercise these types of “oversight” functions in non-democratic systems, even if they are unable to effectively “control” the executive branch as a whole.

Typologies and Classifications  

87

In all of these activities legislatures are largely reactive and have only indirect policy influence. In many political systems, however, legislatures also have the ability to directly and even proactively participate in the policy-making process. Such direct engagement can vary significantly, ranging from providing consultative opinions to making significant amendments, and from initiating independent proposals to vetoing those of the executive branch. Because of the central role of policy-making within the political process more generally, most typologies depend heavily on this critical aspect of legislative power as a primary (if not sole) element of their classification scheme. As noted above, however, there isa broad variety of tasks regularly accomplished by legislatures, and in many cases legislating is not one of the most important.11 One way to understand the critical differences between the four core tasks outlined above (linkage, representation, control/oversight, and policy-making) is to cluster them according to how they shape the role of the legislature. When acting as a linkage mechanism or representative institution, legislatures are functionally the agents of the voters who elected them. However, when they engage in executive oversight and control functions, legislatures become the principal while the executive takes on the mantle of agent. Finally, when participating in policy-making the legislature incorporates both roles, and acts as the agent of citizens, as well as a principal overseeing the activities of the executive (Kreppel 2010). It is the relative balance between these diverse functions that serves as the foundation for most typologies of legislatures. Although many, if not most, classification schemes are primarily focused on the relative policy-making power of the legislature (Blondel 1970; Polsby 1975; Olson 1980), there are those that weigh other aspects of legislative behaviour as much or more, particularly those classification schemes that strive to include legislatures from non-democratic and democratizing political systems (Mezey 1979; Kim, et al. 1984). The task of most classification schemes is to provide a framework within which we can compare legislatures, across time and especially across political systems. These typologies offer a kind of shorthand by providing a substantial amount of information in a relatively concise and cohesive package. Descriptive typologies such as that presented in section 4.2 describing parliaments and congresses, provide information about the constitutional structure of the legislature and its institutional ties to other aspects of the political system, but little beyond that. Thus, it is possible to differentiate between the Mexican and Italian legislatures by noting that the former is a congress while the latter is a parliament. This tells us that Mexico has a separation-of-powers system, while Italy has a fused-powers system. In turn, from this we know something about the mechanisms of executive control that can be wielded by the legislature (confirmation and impeachment in the former versus investiture and/or censure in the latter). These are only the most general characteristics of the legislature, however. They are useful to some degree and the classification can lead to different expectations for each legislature, but they do not provide much support for substantive comparisons. For that it is necessary to engage in a deeper investigation into the extent to which a legislature is able to pursue the four functions listed above to better evaluate the impact the legislature has on the political system as a whole (or on the policy process more specifically).

88   Amie Kreppel Most contemporary research on comparative legislatures begins with the seminal work of Jean Blondel (1970; 1973).12 Blondel began from the institutional characteristics that define legislatures and focusing on parliaments (implicitly in most cases) he examined their legislative oversight and control functions through the lens of what he termed “viscosity.” This was essentially a conceptual mechanism to measure the negative powers of the legislature in the policy process. The greater a legislature’s viscosity, the more it is able to delay or even reject the policy proposals of the executive branch or the “rule-making” process initiated by the executive (Blondel 1970). Although Blondel recognized that legislatures have other functions within the political system, including linkage and representation, he also noted that law-making “is the one which is most commonly referred to in textbooks” (1970, 71). Legislative viscosity was measured based on the tools at the disposal of the legislature to slow down the legislative process, including the duration of the debate, the tabling, length and success of amendments, as well as their origin (majority or opposition parties). A legislature is deemed to be more or less “subservient” or “free” based on the viscosity of the policy-making process, or the extent to which it can slow down the seemingly inexorable legislative machine of the executive. This interpretation of legislative power as the negative ability to slow down or stop policy is reflected in the more recent literature on institutional veto players and their impact on policy change (Tsebelis 2002). What is notable about Blondel’s analysis is that there is no variation on the parliament/congress axis because of the cases selected (France, India, Ireland, Sweden,and the UK). In the end, perhaps unsurprisingly, Blondel finds that strictly majoritarian parliamentary systems (such as the UK and Ireland), are less viscous than those of Sweden and especially France. Blondel goes a step further to note that the breadth of the policy arenas that are available for the legislature to take an active role in is also significant. For example, in his analysis the French legislature appeared superficially to be more “free,” but only within a “restricted domain” (1970, 85). Despite this additional conclusion, the primary focus of Blondel’s typology or classification of legislatures depends critically upon their ability to impact the policy-making process. More specifically, Blondel focused on the negative powers of the legislature, which provide it with some opportunity to affect policy outputs and was narrowly constrained to include only the executive-legislative relationship within this process.13 Other scholars following Blondel increased the comparative scope of their analyses by including a greater diversity of legislatures and expanding the set of roles discussed. One such effort was introduced by Nelson Polsby (1975) with his conceptualization of “arena” and “transformative” legislatures. These two categories were not presented as a dichotomy, but rather an axis along which legislatures could be placed with the US Congress serving as the ideal for the transformative type and the UK House of Commons playing that role for the arena type (Polsby 1975, 130–1). While much of the more detailed analysis focuses on the US case and the methodology tends toward description more than rigorous empirical analysis, in many ways Polsby’s analysis contains the seeds for a much more comprehensive theoretical framework for comparative legislative studies.

Typologies and Classifications  

89

To begin with, the analysis immediately differentiates between the two ideal types of legislatures in terms of not only their legislative role, but also their respective loci of power. To understand transformative legislatures and their role in the policy process, according to Polsby, the critical subjects of analysis are the internal structures of the legislature. In contrast, one must look outside the legislature for the source of decision-making power in arena-type legislatures.14 This is summarized by his conclusion that “for arenas the impact of external forces is decisive in accounting for legislative outcomes. For transformative legislatures, what is decisive are variables depicting internal structure and sub-cultural norms” (Polsby 1975, 141).15 While the specific external actors with the power to shape the policy process are not discussed in any detail, the effect of party organization on what Polsby labels legislative “independence” is incorporated into his comparative analysis. Indeed, the “independence and hence transformative capacity of the legislature in modern democratic political systems lies in the character of parliamentary parties” (Polsby 1975, 142). In particular, there are three aspects of internal party organization that influence legislative independence, all linked to the majoritarian or consensual character of the parliamentary parties.16 These include the breadth of the coalition of the dominant parliamentary group, the extent to which the legislative parties are hierarchically organized, and whether or not legislative majorities on specific policy issues are “fixed and assured” (Polsby 1975, 142). More simply, arena-type legislatures are those in which the parliamentary party limits the independence of members through hierarchical structures and centralized decision-making among a narrow group of actors. In contrast, internal party structures play a diminished role within transformative legislatures. Thus, organizing majorities can be highly coalitional (transformative) or narrowly based (arena) and party management can be very decentralized (transformative) or centralized (arena). Finally, successive policy majorities can be very flexible (transformative) or fixed (arena). The combination of these classifications leads to a typology of legislative independence ranging from highly transformative (USA) to arena (UK). Although Polsby is not explicit about the broader role of political institutions in the development of this typology, the implication is that transformative legislatures exist in separation-of-powers systems (congress-type legislatures), while arena legislatures are “parliaments” within fused-powers systems. Partially this conclusion is a function of the cases he studies (USA, Netherlands, Sweden, Germany, Italy, UK, Belgium, and the French IV and V Republics). However, it is also the result of a failure to include explicitly the character of executive-legislative relations and the exogenous party structure within the analysis.17 The absence of the role of the exogenous party is not the only limitation of the arena-transformative typology introduced by Polsby. Like its predecessor it is fundamentally limited to democracies and to the internal characteristics of the legislature. The next typology introduced into the field of comparative legislative studies moved beyond the confines of policy-making influence to conceive of legislative power in terms of support and durability as well (Mezey 1979). Rather than relying on a single axis of decision-making viscosity or transformative power, this typology also includes a second axis incorporating the extent of popular support for the legislature. The

90   Amie Kreppel addition of support into the analysis not only increases the complexity of the typology, but also allows it to be applicable to a broader array of legislatures, including those in non-democratic systems. Moreover, support is exogenous to the legislature and has the potential to reflect the impact and significance of both the linkage and representative roles more directly than policy-making power. Despite the addition of the support variable, Mezey, like his predecessors, centres his typology in large part on measures of legislative power, understood as “the importance of the legislature in the policy-making process” (Mezey 1979, 151). For Mezey, legislatures fall into three general categories of policy-making power; strong, modest, and little to none. What differentiates Mezey’s approach most significantly is the idea that external support can influence internal policy-making and the power of the legislature more generally. In particular, Mezey focuses on the ability of the legislature to constrain the executive, arguing that “legislatures will be salient in the policy-making process to the extent that their presence and prerogatives act as a constraint on the executive,” without making them vulnerable to dissolution (Mezey 1979, 153). In his analysis, exogenous popular support is the bulwark against executive dissolution of the legislature, thus those that can rely on popular support are better able to use whatever institutional policy-making resources they have at their disposal. In contrast, legislatures with low popular support must always fear that the executive will formally or informally remove the legislature as an obstacle to achieving its policy goals.18 This expansion to include popular support as a factor in understanding the relative power of a legislature is critical in that it broadens the analysis to the political landscape outside of the legislature itself and recognizes the importance of exogenous actors.19 It also underscores the potential impact that linkage and representative functions can indirectly have on the policy-making influence of the legislature. The combination of these two axes, formal policy making powers and external support lead to a five-part typology of legislatures (see Table 4.1). Unsurprisingly the “active” and “reactive” categories overlap to a good degree with Polsby’s transformative and arena-type legislatures. What the concept of external support adds to our understanding of legislative power is the ability to distinguish between those legislatures that are able to make the most of their formal decision-making powers (USA) and those that are not (Italy, French IV Republic) because they are “vulnerable” to being bypassed or dismantled by the executive. Moreover, the inclusion of the minimal and marginal categories extends the applicability of the typology to non-democracies Table 4.1  Mezey’s typology of legislatures Policy-Making Power

Less Supported Legislatures

More Supported Legislatures

Strong

Vulnerable Legislatures

Active Legislatures

Modest

Marginal Legislatures

Reactive Legislatures

Little or none

NA

Minimal Legislatures

Typologies and Classifications  

91

and allows us to distinguish between them on the basis of popular support (and likely durability). For these reasons Mezey’s typology stands as the most comprehensive example of this type of comparative legislative work to date. Without detracting from the significance of his contribution, this is in part due to the rapid decline of the field of comparative legislative studies. Indeed, other than a few edited volumes in the 1980s and 1990s very little similar macro comparative work has been published since.20 This decline is at least partially due to the simultaneous rise in micro-level legislative studies that are often ill-suited to broad cross-national analysis.21 The shift from broad macro-comparative analyses of legislative power and durability to studies of specific aspects of the internal organizational structure of legislatures (micro analysis) is in part a result of Polsby’s original work on transformative and arena legislatures. By highlighting the importance of internal institutional structures in the decision-making processes of transformative legislatures such as the US Congress, he set the stage for research on internal legislative structures rather than cross-national comparisons. Indeed, Polsby’s own work on the institutionalization of Congress was at the vanguard of this micro approach (Polsby 1968). As others responded to the challenge there was a substantive shift, most notably in the American Congress literature, toward a focus on all aspects of internal decision-making and organizational structures. This new work concentrated on investigating specific aspects of legislatures such as agenda-setting, budgetary authority, committee power, caucuses, and leadership structures, rather than broad comparisons of the legislatures themselves. The result of this new wave of research was the transformation of the field of legislative studies as a whole, as scholars emulated the increasingly rich and detailed micro-level research predominant in the American Congressional literature (Döring 1995; Loewenberg, Squire, and Kiewiet 2002; Döring and Hallerberg 2004). The influence of the micro-level approach spread well beyond the confines of the American Congress literature to inspire those studying legislatures in almost every geographic region to attempt to follow suit (with varying degrees of success).22 While the benefits of this trend have been substantial in terms of the accumulation of a wealth of detailed studies on particular aspects of the internal organizational structures of a broad variety of legislatures, the cost has been substantial as well. While there is a wealth of detailed analyses of the impact of particular aspects of internal legislative structures within one or a few legislative contexts, there is comparatively little in the way of a shared theoretical framework with which we can effectively utilize those studies to make macro-level cross-national comparisons, or even cross-time comparisons within a single national context. This is not to suggest that there has not been any recent high quality work comparing legislatures cross-nationally (see for example Sieberer 2011 or in a different way Fish and Kroenig 2009), but this work is scarce and has avoided the development of new typologies that facilitate systemic comparisons. The question raised by this brief analysis is whether it is possible to develop a theoretical framework that allows for the integration of the now-prevalent micro analysis of legislatures and the type of comparative typologies that flourished in the past. This approach would recognize the critical impact of internal organizational structures,

92   Amie Kreppel a variety of exogenous actors and variables, as well as the broader systemic political variables that influence the character of the legislature as a whole. The resulting typology would be able to include parliaments as well as congresses, would be flexible enough to provide explanations for change over time within a single legislature (even in the absence of substantive institutional change), and variations across legislatures. This type of approach is possible, but it requires a shift away from focusing primarily on measuring legislative power, which is the dependent variable most often of central interest to scholars, and focusing instead on the institutional and environmental independent variables that serve as the origins of legislative power. Understanding how these variables are linked and interact can serve as the basis for the development of a new typology that bridges the micro and macro approaches to comparative legislative studies. In other words, it is necessary to move away from simply measuring legislative power toward an understanding of why legislatures are powerful (or not).

4.4  Comparative Legislative Autonomy Most analyses of legislatures focus on the relative decision-making power of the legislature, leaving aside explicit analyses of the representative and linkage tasks, and only implicitly examining control functions (Döring 1995; Fish and Kroenig 2009). This is an understandable approach given the central importance of policy-making within the broader political process. Indeed, control, or at least influence over policy outcomes is the raison d’etre of most political activities and actors. Although there are certainly some who seek power for power’s sake, most would agree that it is the power to determine policy outcomes that drives political actors from political parties, those from lobbying groups, and individual voters deciding how to vote.23 The ability of a legislature to influence decisions over policy outputs is the dependent variable of this typology as well. What is new, however, is the incorporation and central importance of a series of independent variables that together provide critical information regarding the level of “legislative autonomy.” The relative autonomy of a legislature can then be used to predict its potential to influence the policy-making process. The goal is to provide a theoretical construct within which we can embed the knowledge we already have about how legislatures work and how they are impacted by other aspects of the political system. To be comprehensive this approach must include the impact of the broader political environment, the internal structures and resources of the legislature, as well as the partisan framework that affects individual members and groups within the legislature. The concept of “legislative autonomy” allows us to achieve this by uniting a central aspect of all of these elements of the political system—decision-making independence. This approach focuses on the relative level of exogenous control over the actions of the legislature and its members, the ability of exogenous actors to influence the outcomes of the legislative process directly or indirectly.

Typologies and Classifications  

93

Institutional Autonomy

External influence can originate from direct control such as an executive decree, or indirect influence such as partisan control over the careers of party members. In either case, actors beyond the confines of the legislature are able to impact the role of the legislature and/or its members in the policy-making process. In doing so, these exogenous actors effectively weaken the influence of the legislature as an institution and/or its members as individual actors. The majority of variables highlighted by the recent spate of micro-focused studies as significant factors in the policy-making power of legislatures are directly linked to the relative decision-making autonomy of legislatures and their members. These variables can be roughly divided into two broad categories—institutional and partisan. The former affects the legislature as a whole, while the latter affects the individual members and party groups within the legislature. The variables included within the institutional autonomy axis are the most familiar within the comparative-legislatures literature and include both the broad structure of the political system as a whole and the internal organizational characteristics that have been the focus of most micro-level legislative research.24 These institutional variables can be divided into two categories:  those that are constitutional (and comparatively rigid), and those that are institutional and more easily modified. As noted earlier, the core distinction between congresses and parliaments is their relative level of institutional autonomy from the executive branch. Congress-type legislatures benefit from a much higher level of institutional autonomy because of their constitutionally protected independence. In contrast, “parliaments” are weakened by the “fused” character of the executive-legislative relationship, which renders them less institutionally autonomous. This basic division between congresses and parliaments is visualized in Fig. 4.2 by the location of both types of congresses in the upper two quadrants, while parliaments are

Weak Congress

Strong Congress

Weak Parliament

Strong Parliament

Individual/Partisan autonomy

FIG.

4.2  Institutional and partisan autonomy

94   Amie Kreppel in the lower two.25 The precise location of a specific congress or parliament along the vertical axis will depend on a variety of secondary institutional variables within the political system. Secondary institutional characteristics can be formal or informal and are much more likely to change over time. These include a broad variety of structures that make up the subject of much of micro-legislative studies literature including agenda-control, committee structure and powers, as well as other institutional resources such as time, staff, and the quality/full-time status of members. The fact that these institutional characteristics are more susceptible to change than the constitutional differentiation between fused and separation-of-powers systems in no way suggests that they are less important in terms of their impact on legislative power. Instead, these two aspects of institutional autonomy should be understood as interactive—both combining to determine the institutional autonomy of the legislature as a whole. Thus, the greater the institutional autonomy (constitutionally or internally created) the higher the legislature will be placed along the vertical axis (see Fig. 4.2). While the classification of a legislature as a congress or parliament depends on the constitutionally created relationship between the legislative and executive branches of government, the differentiation between strong and weak versions of each type of legislature (horizontal axis) depends on a different kind of autonomy, one that is generated by the characteristics of the party system and the individual parties themselves. This type of partisan autonomy depends on a broad variety of variables including the level of centralized control within individual parties, party penetration within society, campaign finance rules, the laws governing ballot access, and the character of the electoral system as a whole among others.26 As with the vertical axis, the critical attribute is autonomy, particularly decision-making autonomy within the policy process (legislative influence). However, because of the possibility for variations between individual members, party caucuses, and external party organizations, the relationships can be more complex than those found in the institutional arena. At the individual level, the greater the ability of a party (external or internal party caucus) to control the political future of its members in the legislature (through ballot access, funding resources, internal assignments, or other critical benefits) the lower the individual autonomy of members of the legislature will be, meaning that members will have little ability to autonomously impact policy outcomes. Instead members will closely follow party policy directives (party voting) regardless of their personal or even constituency preferences. A similar dynamic exists between party groups/caucuses within the legislature and the external party organization (party in the electorate). If the hierarchy of the external party overlaps with that of the internal party organization, and/or if the external party controls party resources, candidate selection, etc., then the internal party groups will be less autonomous and have less independent ability to impact policy outcomes. However, when there is a parallel leadership structure between internal and external party organizations, and/or when the internal party caucus controls resources, candidate selection, etc., the internal leaders of the party will be better able to directly impact

Typologies and Classifications  

95

policy outcomes independently from the external party organization.27 It is important to note that individual member autonomy and internal party caucus autonomy need not be linked or systematic. Thus, individual autonomy may exist in cases where there is either high or low internal party caucus autonomy and there can be substantial variations between political parties within a single national party system. The greater the autonomy of individual members and/or party organizations within the legislature from exogenous partisan control (i.e., from the “party in the electorate” and/or the “party in the executive”) the higher the level of partisan autonomy will be for the institution as a whole. Those legislatures with higher levels of (individual and group) partisan autonomy will be located further to the right along the horizontal axis.28 Changes along the horizontal axis are more likely than variations in institutional characteristics since these can be achieved by alterations to electoral laws, campaign finance regulations, internal party statutes, as well as simple shifts in the electoral fortunes of the parties themselves, leading to a change in the partisan control of the executive. This approach to understanding and comparing legislatures may at first glance seem to lead to another typology that simply divides legislatures into four categories—not entirely dissimilar to those created by Polsby (1975). However, employing the concept of autonomy as a mechanism for evaluating legislative power provides a theoretical explanation of the policy-making influence of legislatures capable of incorporating the numerous micro analyses that have been prevalent in legislative studies over the last three decades. It also highlights the erroneous assumption that parliaments are necessarily weak, while congresses are strong(er) in terms of policy influence that is inherent in much of the early comparative-legislatures literature. This misconception originates in the implicit merging of the two axes presented in Fig. 4.2 due to the assumption that parliaments have strong parties and congresses have weak(er) ones (Sartori 1997; Linz 1990). Once these two axes are separated and considered individually through an analysis of the variables outlined earlier it is possible to distinguish clear examples of all four types of legislatures. While the US Congress and the British Parliament serve as the emblematic examples of the strong congress and weak parliament respectively, it is also possible to point to other cases that represent the weak congress and strong parliament categories such as the Mexican Congress and Italian Parliament respectively. By bringing together the institutional (internal) and partisan (external) variables that affect legislative autonomy, and thus policy-making influence, this approach serves as a first step at integrating the micro and macro approaches to legislative studies. In doing so it allows for the development of a theoretically driven typology applicable across a broad variety of political contexts and capable of explaining change over time. Movement from one type of legislature to another (i.e. from parliament to congress) is unlikely given that this classification is determined by constitutional provisions. However, there can be substantial movement within quadrants as a result of changes to less fundamental aspects of the structural and institutional characteristics of the legislative process or the institutions themselves. Thus, movement along the vertical axis is possible, even if it is generally constrained to movement within quadrants. In contrast, movement along the

96   Amie Kreppel horizontal axis, even movement from the “strong” category to the “weak” one (between quadrants) can occur in a number of ways and may result simply from a change in the electoral fortunes of one political party or another.29

4.5  Unanswered Questions and Future Research The proliferation of micro-oriented studies of the US Congress that followed the ground-breaking work of Polsby and others has fundamentally changed the character of comparative legislative studies. There is now widespread agreement on the significance of a variety of internal organizational and structural elements of legislatures. The importance of amendment and voting rules, the committee system, and a myriad of other characteristics of legislatures are now well recognized and amply studied, not only within an American context, but also internationally through case studies and cross-national collaborative projects (Döring 1995; Loewenberg, Squire, and Kiewiet 2002; Hallerberg and Döring 2004; Fish and Koenig 2009). While unquestionably important, these micro level analyses have two very important weaknesses that hinder the further development of comparative legislative studies. First, they often fail to fully incorporate the interactive effects of these diverse internal structures and rules. Studies of agenda-setting, committee powers, or organizational structures rarely extend their analyses to include other aspects of the organizational structure of the legislature, which may have a profound impact on the aspect studied. Legislatures are complex organizations with profoundly integrated structures—focusing on why or how just one element of this structure in isolation impacts the policy-making role of the legislature as a whole is bound to result in incomplete understanding, if not actual errors of interpretation. A theoretically grounded typology that incorporates a broad spectrum of structural/ organizational variables would help mitigate this concern. A second weakness of the micro approach as broadly applied in the comparative legislatures literature has been the focus on the internal structures of the legislature to the exclusion of the broader political environment, and especially the character of the party system and individual parties. Regardless of the internal organization of a legislature, if political parties are able to dominate the actions of legislators and control legislative activity from outside the institution, the relative power of the legislature itself will be minimal. Exogenous manipulation of policy outcomes through control over individuals and/or legislative party groups is not uncommon, and is readily recognized in the literature, but it is rarely integrated into micro analyses of legislative structures. This absence neglects an important component of actual legislative control over policy outcomes. A typology that includes both internal and external variables (and their interaction) allows for a more comprehensive understanding of the ability of the legislature to autonomously influence policy decisions.

Typologies and Classifications  

97

Finally, while the development of a theoretically driven typology that can bridge the micro and macro approaches to comparative legislative analysis provides a possible road map for future research, it does not solve the second major problem that faced scholars back in the 1970s: the lack of large sets of comparable data. While it is certainly the case that the number of empirical individual cases studies, limited comparative cases studies (small N), and even broad cross-national analyses (large N studies) of a variety of internal legislative characteristics has increased over the past 30–40 years, it is also the case that comparability between cases, data sets, and analyses remains a challenge. In part this is because there has not been a macro approach to provide a framework for linking these diverse studies. Fundamentally, it is also because all political systems are different and what is important in one may not be important in another. Furthermore, data collection reflects both the interests and the resources of the scholar. Thus, even with a unifying framework the underlying challenge of detailed large N comparisons will remain. This does not mean that we should not aspire to overcome these challenges. Indeed, the proliferation of legislative scholars, inspired in part by the flourishing of the American Congress literature over the past few decades suggests that there is much to be optimistic about. To the extent that scholars can work together, share resources, and collaborate on refining and improving shared understandings and interpretations of legislative influence and importance, the field as a whole will be significantly improved. The current volume is a tribute to this type of collaboration, as are other large-scale joint efforts that have begun to flourish in recent years (Fish and Koenig 2009).

Notes 1. For the purposes of this chapter typology and classification schemes are used interchangeably. 2. It should be noted that this interpretation of parliaments and congresses as types of legislatures is not necessarily exhaustive and does not preclude the potential existence of other types of legislatures. Within democratic political systems, however, these types are able to accommodate all legislatures with which the author is familiar. 3. While most are familiar with the roots of the word “parliament—derived from the French verb parler, “to speak”—few recognize that the word “congress” is derived from the Latin congressus,“a meeting or [hostile] encounter; to contend or engage” (Harper 2001). The use of congress as a general term to denote legislatures within separation-of-powers systems is justified by the more frequent policy-making focus of their activities and the greater likelihood of conflict between the legislative and executive branches when compared to fused-power systems. 4. Recognition of the impact of broader institutional structures on legislative type is not new, however the terminological hierarchy here is intended to clarify continuing ambiguities. For example, nearly a half century ago Wheare (1963) differentiated between parliaments and congresses—but without directly tying it to the underlying relationship between the legislative and executive branches. In contrast, Laver (2008) highlights the critical importance of this relationship, but creates a legislature–parliament dichotomy instead of allowing legislatures to remain the broader classification.

98   Amie Kreppel 5. Note that although this is the norm, there are parliamentary systems in which it is possible, or even the norm to select members of the cabinet from outside the legislature (Andeweg and Nijzink 1995). 6. This is, of course, a simplification and ignores possible obstacles such as requirements for a constructive vote of no-confidence, for example. 7. The preceding section 4.2 draws from the typology first published in Kreppel (2010). 8. This is especially true in presidential systems and those parliamentary systems in which a single party wins an absolute majority of seats in the parliament. In contrast, coalition governments that include parties representing a variety of minority interests can serve to increase the representative function of the executive branch as a whole. 9. Readers should note that the ability to select the executive within parliamentary-type legislatures is incorporated within the “control” function. However, because congress-type legislatures do not have this function it is not included as a distinct category. 10. This, of course, assumes at least a moderately democratic political system. The control function of legislatures in non-democratic systems is marginal at best. 11. This is particularly true in non-democratic systems, effectively limiting the applicability of many typologies to the democratic context. 12. Prior to Blondel there was some comparative work, beginning with Walter Bagehot, (1872) and Lord Bryce (1921), that looked at the changing role and character of legislatures in the broader political system. However, the former focused largely on the specifics of the UK and US systems, while the latter instead lamented the decline of legislatures (and the rise of political parties) rather than present a fully developed comparative typology. 13. The potential positive impact of negative power is discussed to some degree in that very “viscous” legislatures are assumed to be able to cajole concessions from the executive during the policy-making process through the use, or threatened use, of their negative powers (Blondel 1970). 14. Polsby notes (p. 130) that the external source is likely to be political parties, but leaves open the possibility for other actors as well (military, bureaucracy, etc.). He does not, however, examine the role of external political party organizations in particular, despite noting their clear significance. 15. The importance of the internal organizational structures is linked to the role these play in the transformation of policy, through committee review, amendments, etc. In contrast, the primary function of arena-type legislatures is the discussion of policy alternatives constructed elsewhere. 16. Note that this terminology associated with Arend Lijpart’s ground-breaking work on democracies in plural societies (1977) was not yet available to Polsby. 17. It is this lacuna that the concept of “legislative autonomy” based on both structural and partisan autonomy (discussed in section 4.4) attempts to correct. 18. While the abolition of the legislature is not a particularly relevant threat in established democracies, there are other tactics that executives can employ to effectively go around legislatures (executive orders and decrees). 19. Polsby does this as well when he highlights the importance of external actors for arena-type legislatures while Mezey credits external support with providing some of the institutional power that allows transformative legislatures to achieve their full potential. 20. See, for example, Olson 1980; Ornstein 1981; Kim, Barkan, Turan, and Jewell 1984; and Copeland and Patterson 1994.

Typologies and Classifications  

99

21. It should be highlighted that a great deal of excellent “institutionalist” research on legislatures has been conducted since this period, but much has been focused either on specific internal characteristics/micro analyses of legislatures (Döring 1995 is an excellent compendium of such work), or on broader systemic analyses (Lijphart 1999; Tsebelis 2002) rather than focusing on comparative legislatures per se. 22. Many who pursue the micro approach and focus on the internal attributes of a legislature are unaware of Polsby’s original exhortation to look at internal organizational structures to understand the functioning of transformative-type legislatures, but to focus on external influences on legislative behaviour to understand arena-type legislatures. As a result there is a tendency to decry the “decline of parliaments” by those seeking a transformative role from an arena-type legislature. 23. Indeed, in many respects linkage and representation matter because citizens want to be connected to and have their preferences represented in the policy process. 24. The interpretation of institutional autonomy used here is similar to, but broader than that used by Fish and Kroenig (2009). 25. The implication of the quadrants is that even the weakest congresses will be more autonomous than the strongest parliaments. 26. These exogenous variables encompass those anticipated to be of significant for “arena”type legislatures by Polsby (1975), however, they are important for both congresses and parliaments despite the very different role that parties are generally assumed to play within these two institutional contexts. 27. This differentiation between party in the legislature and party in the electorate can be particularly important, for example, when the party in the electorate is controlled by party activists (less moderate) and the party in the legislature is in a coalition—which generally requires compromises that can be unpopular with the party base. 28. It should be noted that the two axes are not independent. Fused-powers systems tend to have stronger party systems with more centralized control of members, but the relationship between the two axes are tendencies rather than iron laws. 29. This is especially possible in systems that have wide variation in the internal organization of political parties. If some parties have strong control over members while others are less centralized a switch from one party in government to the other can have a significant impact on the ability of the legislature to act independently. Poland in the 1990s experienced this type of change as the majority shifted between the highly centralized post-Communist SLD and the deeply fragmented post-Solidarity AWS.

References Andeweg, R. and Nijzink, L., 1995. Beyond the Two-Body Image:Relations between Ministers and MPs. In H. Döring (ed.). Parliaments and Majority Rule in Western Europe, pp. 152–78. New York: Palgrave Macmillan. Bagehot, W., 2001 [1872]. The English Constitution. Cambridge: Cambridge University Press. Blondel, J., 1970. Legislative Behaviour:  Some Steps towards Cross-national Measurement. Government and Opposition, 5: 67–85. Blondel,J., 1973. Comparative Legislatures. Englewood Cliffs: Prentice Hall. Bryce, L., 1921. Modern Democracies. London: Macmillan.

100   Amie Kreppel Copeland, G. and Patterson, S., 1994. Parliaments in the Modern World. Ann Arbor: University of Michigan Press. Döring, H. (ed.), 1995. Parliaments and Majority Rule in Western Europe.New York: Palgrave Macmillan. Döring, H. and Hallerberg, M. (eds.), 2004. Patterns of Parliamentary Behaviour: Passage of Legislation Across Western Europe. Aldershot: Ashgate Press. Fish, S. and Kroenig, M., 2009. The Handbook of National Legislatures. New York: Cambridge University Press. Harper, D., 2001. The Online Etymology Dictionary (). Kim,C. L., Barkan, J., Turan, I., and Jewell, M., 1984. The Legislative Connection. Durham: Duke University Press. Kreppel, A., 2010. Legislatures. In D. Caramani (ed.). Comparative Politics (2nd ed.), ch. 7. Oxford: Oxford University Press. Laver, M., 2008. Governmental Politics and the Dynamics of Multiparty Competition. Political Research Quarterly, 61: 532–36. Lijphart, A., 1977. Democracy in Plural Societies. Princeton: Yale University Press. Lijphart, A., 1999. Patterns of Democracy. Princeton: University of Princeton Press. Linz, J., 1990. The Perils of Presidentialism. Journal of Democracy, Vol. 1, No. 1, Winter: 51–69. Loewenberg, G. and Patterson, S., 1979. Comparing Legislatures. Boston: Little Brown. Loewenberg, G., 2011. On Legislatures. Boulder: Paradigm Publishers. Loewenberg, G., Squire, P., and Kiewiet, R. (eds.), 2002. Legislatures. Ann Arbor: University of Michigan Press. Mezey, M., 1979. Comparative Legislatures. Durham: Duke University Press. Olson, D., 1980. The Legislative Process. New York: Harper and Row. Olson, D., 1994. Democratic Legislative Institutions. London: M.E. Sharpe. Ornstein, N. (ed.), 1981. The Role of the Legislature in Western Democracies. Washington: American Enterprise Institute. Pitkin, H., 1967. The Concept of Representation. Berkeley: University of California Press. Polsby, N., 1968. The Institutionalization of the U.S. House of Representatives. American Political Science Review, 62: 144–68. Polsby, N., 1975. Legislatures. In F.I. Greenstein and N. Polsby (eds.). Handbook of Political Science (Vol. V). Reading: Addison-Wesley Press. Sartori, G., 1997. Comparative Constitutional Engineering. New York: New York UniversityPress. Sieberer, U., 2011. The Institutional Power of Western European Parliaments: A Multidimensional Analysis. West European Politics, 34: 731–54. Tsebelis, G., 2002. Veto Players. Princeton: Princeton University Press. Wheare, K.C., 1963. Legislatures. Oxford: Oxford University Press.

PA R T  I I

M E T HOD S OF L E G I SL AT I V E R E SE A RC H

C HA P T E R

5

R O L L - C A L L A NA LYS I S A N D T H E S T U DY O F L E G I S L AT U R E S ROYC E C A R ROL L A N D K E I T H P O OL E

5.1 Introduction The application of spatial models of choice and judgment to measure the behaviour of legislators is built upon applying statistical procedures that analyse observed data and extract latent (i.e. abstract) dimensions upon which the objects or subjects can be placed. In political science, scholars are generally interested in policy or ideological scales using individuals’ judgments or observed voting behaviour. The results of these scales, usually called “ideal point estimates,” uncover the basic dimensionality of choice behaviour and have the potential to reveal the underlying latent preferences behind that behaviour. In political contexts, these might be basic differences such as left-right, liberal-conservative, secular-religious, or regional cleavages that provide a common thread across numerous policy choices. For legislative scholars, these tools can provide an invaluable means for testing hypotheses relating to legislators’ preferences or for exploring the basic patterns behind otherwise complex data. With numerous arguments in the study of legislatures relying on spatial analogies of the policy distances between political actors, the empirical methods of quantifying distance in legislative behaviour have provided enormous utility to the field. These methods have become central to the study of the US Congress and as scholars adapt these methods for comparative purposes they are increasingly important for studies of other legislatures as well.

104   Royce Carroll and Keith Poole

5.2  The Evolution of Scaling Techniques and Spatial Analysis The scaling methods used to analyse roll-call voting mostly originate from research in psychology that was generally interested in deriving latent properties such as intelligence from patterns of data. The foundational work in this area was conducted by Charles Spearman (1904) in the form of factor analysis and by Karl Pearson (1901) in the form of Principle Component Analysis (or eigenvector-eigenvalue decomposition). These techniques were advanced by L.  L. Thurstone who succeeded in developing a method for extracting multiple factors from a correlation matrix (1931; 1947).1 Parallel to these innovations in statistics were developments in economic models. Hotelling, who also gave principal components a solid statistical foundation (Hotelling 1933), produced the seminal theoretical work on the stability of competition, which is now generally recognized as the beginnings of the spatial (geometric) model of voting (1929). This work later led to the development of the median voter theorem by Black (1948; 1958) and the influential exploration of its implications for political competition by Downs (1957). In 1936 the work of Eckart and Young (1936) along with Young and Householder (1938) provided the foundations for classical multidimensional scaling (MDS), which was developed by Torgerson (1952; 1958). MDS methods are applied to relational data, such as similarities and preferential choice data that can be regarded as distances. Multidimensional scaling methods represent measurements of similarity between pairs of stimuli as distances between points in a low-dimensional (usually Euclidean) space. The methods locate the points in such a way that points corresponding to very similar stimuli are located close together while those corresponding to very dissimilar stimuli are located further apart.2 Shepard (1962a; 1962b) developed nonmetric multidimensional scaling (NMDS) in which distances are estimated that reproduce a weak monotone transformation (or rank ordering) of the observed dissimilarities.3 Kruskal’s (1964a; 1964b; 1965) monotone regression procedure led to the development of a powerful and practical nonmetric MDS computer program (Kruskal, Young, and Seery 1973).4 At the same time Guttman (1944; 1950)  developed scalogram analysis (Guttman Scaling),whichformsthe basis of modern Item Response Theory,5 while Clyde Coombs developed unfolding analysis (Coombs 1950; 1952; 1958; 1964)  for ranked preference data. In this work, Coombs introduced the idea of an ideal point and a single-peaked preference function; the purpose of an unfolding analysis was to arrange the individuals’ ideal points and points representing the stimuli along a scale so that the distances between the ideal points and the stimuli points reproduced the observed rank orderings.6 By the mid-1950s these techniques began to appear in the work of political scientists interested in legislative voting. Duncan MacRae’s pathbreaking work on voting in

Roll-Call Analysis and the Study of Legislatures  

105

the US Congress (MacRae 1958; 1970) utilized both factor analysis and scaling methods to analyse correlation matrices computed between roll calls and between legislators to uncover the dimensional structure of roll-call voting.7 In the 1980s Poole and Rosenthal combined the random utility model developed by economists (McFadden 1976), the spatial model of voting, and alternating estimation methods developed in psychometrics (Chang and Carroll 1969; Carroll and Chang 1970; Young, de Leeuw, and Takane 1976; Takane, Young, and de Leeuw 1977)  to develop NOMINATE, an unfolding method for parliamentary roll-call data (Poole and Rosenthal 1985, 1991; 1997; Poole 2005). The NOMINATE model is based on the spatial theory of voting.8 Legislators have ideal points in an abstract policy space and vote for the policy alternative closest to their ideal point. Each roll-call vote has two policy points—one corresponding to “yea” and one to “nay.” Consistent with the random utility model, each legislator’s utility function consists of (1) a deterministic component that is a function of the distance between the legislator and a roll-call outcome; and (2) a stochastic component that represents the idiosyncratic component of utility, which captures the aspects of voting behaviour that are not explained by the spatial dimension. The deterministic portion of the utility function is assumed to have a normal distribution and voting is probabilistic. An alternating method is used to estimate the parameters, meaning that an iterative process alternates between estimation of the ideal points and the roll-call parameters. Given starting estimates of the legislator ideal points, the roll-call parameters are estimated. Given these roll-call parameters, new legislator ideal points are estimated, and so on. Classical methods of optimization are used to estimate the parameters. As advances in computing power have popularized simulation methods for the estimation of complex multivariate models, these methods were fused with long-standing psychometric methods. Specifically, Markov Chain Monte Carlo (MCMC) simulation (Metropolis and Ulam 1949; Hastings 1970; Geman and Geman 1984; Gelfand and Smith 1990; Gelman 1992) within a Bayesian framework (Gelman, Carlin, Stern, and Rubin 2000; Gill 2002) have been increasingly used to perform an unfolding analysis of parliamentary roll-call data in legislatures and courts, especially due to the influential work of Martin and Quinn (2002) and Clinton, Jackman and Rivers (2004). The difference is that estimations are based on sampling from conditional distributions for the legislator and roll-call parameters using the Gibbs sampler (Geman and Geman 1984; Gelfand and Smith 1990), especially the implementations by Jackman (2008) and Martin and Quinn (2009). Although not intrinsic to the estimation method, Bayesian MCMC applications usually use a quadratic deterministic utility function.9

5.3  Spatial Models of Voting The common thread of all spatial (geometric) models of voting is the notion that a vote is based on the distance between a respondent’s or legislator’s ideal point and a policy

106   Royce Carroll and Keith Poole proposal. We focus on the legislative voting model of binary choice (roll-call voting) because it is the most widely used, but our development below can be easily extended to mass survey data (ratio scale, rank orders, and nominal data). An important concept here is the notion of spatial “error.” By error, we simply mean instances of voting that are not in line with spatial preferences. For instance, if a legislator encounters a proposal that provides disutility relative to the alternative (the status quo), that voter should vote “nay,” while a legislator should vote “yea” if that proposal is closer to his or her ideal point. In short, if there were no error of this sort, the legislator would always vote for the closest alternative in the policy space on every roll call. This results from the fact that the utility function is symmetric. Let the two policy outcomes corresponding to yea and nay on the jth (j=1,. . . ,q) roll call be represented by Ojy and Ojn respectively. In most cases it is more convenient to work with the midpoint of the two outcomes:

Zj =



O jy + O jn 2



In one dimension Zj is known as a cutting point that divides the yeas from the nays. With perfect spatial voting, all the legislators to the left of Zj vote for one outcome and all the legislators to the right of Zj vote for the opposite outcome. In two dimensions, a cutting line will divide the yeas and nays. In one-dimensional perfect roll-call voting both the legislators and the roll-call midpoints are represented by points—Xi and Zj respectively—and a joint rank ordering of the legislators and roll-call midpoints can be found that exactly reproduces the roll-call votes (Poole 2005). The process of eigenvector decomposition (described in fn.1) of the voting pattern in Fig. 5.1, for example, will produce a rank order (from X1 to X6) that corresponds to the ideal points depicted for the legislators. Legislators X1

X2

X3

X4

X5

X6

Z1

Y

N

N

N

N

N

Z2

Y

Y

N

N

N

N

Z3

Y

Y

Y

N

N

N

Z4

Y

Y

Y

Y

N

N

Z5

Y

Y

Y

Y

Y

N

Votes

FIG. 5.1  Perfect

spatial voting in one dimension

Roll-Call Analysis and the Study of Legislatures  

107

1.0 N

Y

0.5

Normal Vector

Oy

N

Y

On N

0.0 N −0.5

Y

Y

Y

N N N

Cutting Plane

−1.0 −1.0 FIG. 5.2  Example

−0.5

0.0

0.5

1.0

of a cutting line in two dimensions

In two or more dimensional perfect voting a legislator is still represented by a point—the s by 1 vector Xi where s is the number of dimensions—but a roll call is now represented by a plane that is perpendicular to a line joining the yea and nay policy points—the s by 1 vectors Ojy and Ojn—and passes through the midpoint, the s by 1 vector Zj. The normal vector to this cutting plane is parallel to the line joining the yea and nay policy points. Fig. 5.2 shows a simple example of 12 legislators in two dimensions, where the cutting line is shown that separates the yea (Y) and nay (N) voters on a particular proposal with outcomes of Oy if passed and On if defeated. Because each voter can be separated by this cutting line, this two-dimensional example produces no errors. In two dimensions, if a variety of voting coalitions form amongst the legislators, then the q cutting lines will criss-cross one another in a myriad of directions creating a very large number of what are called “polytopes.” With perfect voting a legislator is only defined up to a polytope. That is, each legislator could be anywhere in the polytope that corresponds to his or her roll-call choices (Poole 2005). A non-parametric method of unfolding binary choice data is Optimal Classification (Poole 2000; Poole et al. 2012b), which optimizes cutting-line locations in order to minimize the number of incorrectly predicted votes (errors). In this method, the coordinates are obtained by finding their optimal location among the cutting lines that minimizes the number of errors. Fig. 5.3 presents an example of the process of locating members in two dimensions using this method, and employing data from the sixth EU Parliament (Hix, Noury, and Roland 2009), with the MEP coordinates and a sample of 100 cutting lines plotted.10 If error is present, then the problem of estimating cutting planes is equivalent to a probit or logit analysis depending upon the assumptions made about the error. The parametric methods of roll-call analysis make use of some form of assumption about error in order to obtain interval information. As an illustration, Fig. 5.4 presents a spatial map of the final passage vote of the landmark 1964 Civil Rights Act in the US Senate using

108   Royce Carroll and Keith Poole 0.8

0.6

EUL/NGL IND/DEM EPP-ED ALDE SOC UEN G/EFA

2nd Dimension

0.4

0.2

0.0

−0.2

−0.4 −0.4

−0.2

0.0

0.2

0.4

1st Dimension FIG. 5.3  Optimal

classification coordinates and cutting lines for the sixth EU parliament

DW-NOMINATE. The left panel shows all the senators and the right panel shows just the five senators who were errors in the DW-NOMINATE analysis. Each senator’s location in the map is a function of all the roll-calls the senator participated in during his/ her career. The cutting line is specific to the roll-call and divides those senators who are predicted to vote yea from those who are predicted to vote nay. Those senators who are incorrectly predicted—the “errors”—are indicated in the right side plot. The descriptive labels and the relative positions of the party tokens in the map show that a coalition of Republicans and Northern Democrats voted for the act and a coalition of Southern Democrats and a few Republicans voted against the act. The major probabilistic models of parliamentary voting, such as DW-NOMINATE in this example, are based on the random utility model described above. To recap, in the random utility model a legislator’s overall utility for voting year is the sum of a deterministic utility and a random error. Drawing on the description in Poole (2005), suppose there are p legislators, q roll calls, and s dimensions indexed by i=1,...,p, j=1,. . . ,q, and k=1,. . . ,s, respectively. Legislator i’s utility for the yea outcome on roll call j is:

U ijy = u ijy + ε ijy



Roll-Call Analysis and the Study of Legislatures  

1964 Civil Rights Act: All Senators 19 June 1964

0.5

0.0

−0.5

−1.0

1964 Civil Rights Act: Errors Only 19 June 1964

S SSS SSSS S S D S S D Predicted S DD S Against D D D D DS R DDDDSD DD DDD D DD RR D DDD D D DDD R R RR D DD D D R DD DD D D RD RRRR R D RR R R R R R R Predicted For RRR R R R R RR Cutting Line S

−1.0

−0.5

0.0

0.5

Cutting Line

1.0

Region (Civil Rights)

Region (Civil Rights)

1.0

1.0

Predicted Against

S

0.5

0.0

R R R

−0.5

R

Predicted For −1.0 −1.0

Liberal-Conservative Dimension FIG. 5.4  Cutting

109

−0.5

0.0

0.5

1.0

Liberal-Conservative Dimension

line and errors on 1964 Civil Rights Act in the US Senate

where uijy is the deterministic portion of the utility function and εijy is the stochastic portion of the utility function. If there is no error, then the legislator votes yea if Uijy > Uijn. Equivalently, if the difference, Uijy–Uijn, is positive, the legislator votes yea. With random error the utility difference is: U ijy − U ijn = u ijy − u ijn + ε ijy − ε ijn





So that the legislator votes yea if: u ijy − u ijn > ε ijn − ε ijy





That is, the legislator votes yea if the difference in the deterministic utilities is greater than the difference between the two random errors. Since the errors are unobserved, we must make an assumption about the error distribution from which they are drawn. We can calculate the probability that the legislator will vote yea. That is:

(

) (

(

) (

) ( < 0) = P (ε

P Legislator i votes yea = P U ijy − U ijn > 0 = P ε ijn − ε ijy < u ijy − u ijn P Legislator i votes nay = P U ijy − U ijn

So that P(yea)+P(nay) = 1.

ijn

− ε ijy > u ijy − u ijn

) )

110   Royce Carroll and Keith Poole Poole and Rosenthal’s NOMINATE (Nominal Three-Step Estimation) model is based on the normal distribution utility function. The normal distribution concentrates the utility near the individual’s ideal point with tails that quickly approach zero as the choices become more and more distant. With the normal distribution utility model, legislator i’s utility for the yea outcome on roll call j is:

u ijy = βe



 1 s  2 w k d ijky −   2 k=1 

∑



2 where dijky is the squared distance of the ith legislator to the yea outcome on the kth dimension:

d 2ijky = (X ik − O jky )2





the wk are salience weights (wk> 0); and because there is no natural metric β “adjusts” for the overall noise level and is proportional to the variance of the error distribution. The wk allow the indifference curves of the utility function to be ellipses rather than circles. The difference between the deterministic utilities is:    1 s    1 s 2  2  − ∑ w k d ijkn   u ijy − u ijn = β   − 2 ∑ w k dijky − 2   k =1 e k=1 e 





This equation cannot be further simplified. Despite this apparent complexity, it is not difficult to work with computationally. With the quadratic distribution deterministic utility model, legislator i’s utility for the yea outcome on roll call j is just: s

(

u ijy u ijy = −d 2ijy = − ∑ X ik − O jky



k=1

2

)



The difference between the deterministic quadratic utilities is:





s

s

k =1

k =1

u ijy − u ijn = − ∑ (X ik − O jky )2 + ∑ (X ik − O jkn )2

s

s

k=1

k=1



= −2∑ X ik (O jkn − O jky ) + ∑ (O jkn − O jky )(O jkn + O jky )



Roll-Call Analysis and the Study of Legislatures  

111

This is isomorphic with the two parameter Item Response Model by setting:

s

α = ∑ (O jkn k =1



 −2(O j1n − O j1y )  −2(O − O ) j2y  j2n − O jky )(O jkn + O jky ) and β j =        −2(O jsn − O jsy ) 



where βj is an s by 1 vector. This allows the difference between the latent utilities for yea and nay to be written in the same form as the item response model; namely:

y ij* = U ijy = U ijn = α j + X i′ β j + ε ij



where yij* is the difference between the latent utilities and ε ij = ε ijn − ε ijy ~ N (0, 1)





This is known as the Quadratic-Normal Model (Poole 2001; 2005). Both the NOMINATE model and the Quadratic-Normal can be estimated with maximum likelihood (Poole 2005; Poole et al. 2012a). An alternative to maximum likelihood estimators for spatial voting models is Monte Carlo simulation (Metropolis and Ulam 1949; Hastings 1970; Geman and Geman 1984; Gelfand and Smith 1990; Gelman 1992), which makes use of a “random tour” of the parameter space in a Bayesian framework (Jackman 2000a). The random tour involves recording the probability function at hundreds of thousands or millions of points in the parameter space. From this process, the shape of the distribution over the parameter space is known with some degree of certainty at the end of the tour. The means and standard errors of the parameters are calculated based on this information. The Markov-chain Monte Carlo (MCMC) method used to generate the random tour is computationally intensive, but produces both estimates of the legislator ideal points and roll-call parameters and measures of uncertainty for those estimates. The intuition behind the Bayesian simulation approach can be seen by looking at the simple formulas for conditional probability. Let θ and Y be two events, then in classical probability theory:



P (θ| Y ) =

P(θ ∩ Y) P(θ ∩ Y) and P ( Y| θ) = P(Y) P(θ)

hence

P(θ ∩ Y) = P (θ | Y ) P ( Y ) = P ( Y | θ) P (θ )



112   Royce Carroll and Keith Poole And P (θ | Y ) =



P(Y | θ)P(θ) P(Y)

In the Bayesian framework, Y is the observed data, the θ are the parameters, P(Y |θ) is the likelihood function of the sample, P(θ) is the prior distribution of the parameters, P(Y) is the marginal distribution of the sample, and P(θ | Y) is theposterior distribution. Because P(Y) is a constant, the posterior distribution is proportional to the product of the likelihood function and the prior distribution; that is: P (θ | Y ) ∞ ( Y | θ ) P (θ )





The researcher specifies the prior distribution of the parameters, P(θ). When this framework is applied to roll call data, Y is the p by q matrix of choices and θ is the vector of legislator ideal points and roll call parameters. In the Bayesian NOMINATE model (Carroll et al. 2013) the Likelihood function is:

(

L O jy , O jn , X | Y



)

where X is the p by s matrix of legislator coordinates, and Ojy and Ojn are q by s matrices. The prior distributions are all normal distributions so the posterior distribution is:

(

) (

)

( )

ξ O jy ,O jn ,X | Y = L O jy ,O jn ,X | Y ξ(O1y )ξ (O1n ) … ξ X p

Similarly, the Posterior distribution for the Bayesian Quadratic Normal (IRT) model is:

(

ξ α j ,βj , X | Y

)

Each of these models can be estimated using standard MCMC methods (Clinton, Jackman, and Rivers 2004; Carroll et al. 2009a; Carroll et al. 2013). For the same models, Bayesian estimation generally provides similar estimates to those of MLE approaches (Carroll et al. 2009a) but the former allows greater flexibility in the use of exogenous information and a measure of uncertainty obtained via simulations, at the expense of requiring greater computational resources. In addition, many recent efforts to tailor models to specific problems have made use of the BUGS language11 for statistical

Roll-Call Analysis and the Study of Legislatures  

113

programming which enables the flexible use of Bayesian MCMC methods (e.g. Clinton and Jackman 2009; Zucco and Lauderdale 2011).

5.4  Seminal Applications of Scaling Methodology in Legislative Studies As noted earlier, the initial substantive applications of early scaling techniques were applied to the US Congress (e.g. Clausen and Cheney 1970; Clausen 1973; see Collie 1984 for review). As spatial techniques were developed, several waves of research emerged employing NOMINATE, notably Poole and Rosenthal’s work identifying the growing ideological polarization in Congress (1984; recently extended by McCarty, Poole, and Rosenthal 2006) and the debate surrounding the dimensionality of Congress (Poole 1984; Poole and Rosenthal 1985; Koford 1989; Wilcox and Clausen 1991). Using the measures made possibly by the DW-NOMINATE dynamic scaling technique, Poole and Rosenthal (1997) completed a landmark study of the evolution of the American Congressional party system. This work provided a new understanding of the shifts between the major eras of political, economic, and social conflict—through the lens of the latent patterns of legislative voting behaviour. Much of the subsequent development of the method itself has been driven by the desire to derive preference measures for Congress to operationalize important theories in the field, especially those employing formal spatial models. Among these questions one of the most prominent has been the relative importance of party, preferences, and constituency on voting (Levitt 1996; Snyder and Groseclose 2000; Ansolabehere, Snyder, and Stewart 2001; McCarty, Poole, and Rosenthal 2001; Cox and Poole 2002). Many theorists examining influential Congressional organization theories have also made use of these preferences measures, including informational and distributive theories of committees (Londregan and Snyder 1994), “conditional party government” theory (Aldrich and Rhode 1998; Forgette and Sala 1999), “party cartel” theory (Cox and McCubbins 2005), and “pivotal politics” theory (Krehbeil 1998; Chiou and Rothenberg 2003). Numerous studies have also attempted to distinguish between partisan and “floor median” control of Congress using scaling techniques (Krehbiel et  al. 2005; Clinton 2007; 2011; Stiglitz and Weingast 2010). Much of this work on parties has extended to applications in the American state legislatures (Aldrich and Battista 2002; Wright and Shaffner 2003; S. Jenkins 2006; Shor et al. 2010; Carroll and Eichorst 2013) and historical contexts (J. Jenkins 1999). With the solidification of scaling methods as a standard tool in the study of Congress, there has been a dramatic increase in the use of such methods in comparative contexts. Again, many applications have been focused on identifying the

114   Royce Carroll and Keith Poole dimensional basis of voting divisions. Hix (2001) and Hix, Noury, and Roland (2006), provided early and important applications of ideal point estimation to identify the origins for voting cleavages in the EU Parliament. Similar analyses have examined the key dimensions of voting behaviour in the UN (Voeten 2000; 2004), Korea (Hix and Jun 2009), Ireland (Hansen 2009), the Weimar Republic (Hansen and Debus 2011), Canada (Godbout and Hoyland 2011), Switzerland (Hug and Schulz 2007), and the Czech Republic (Lyons and Lacina 2009). Rosenthal and Voeten (2004), focusing on the French Fourth Republic, employ Optimal Classification to exploit its appropriateness for situations of “perfect” spatial voting (that is, low rates of spatial voting “error”) which is especially common in chambers with high levels of party unity. Others have employed some form of spatial analysis method to assess the basis of voting blocs in legislative data in cases such as Argentina (Jones, Hwang, and Micozzi 2009), Chile (Morgenstern 2004; Alemán and Saiegh 2007), Brazil and Uruguay (Morgenstern 2004), France (Sauger 2009), Hungary (Ágh 1999), Denmark (Hansen 2008), Russia (Bagashka 2008), and the EU Council (Hagemann 2007; Hagemann and Hoyland 2008). Parallel to the studies of Congress built around testing formal and informal spatial and ideological arguments, numerous recent comparative studies have used ideal point estimates to measure policy preferences of individual legislators to allow testing of hypotheses requiring a proxy measure for ideological preference information. These applications have been employed to examine questions such as the reasons for switching parties in Brazil (Desposato 2006), legislative organization theory in Russia (Myagkov and Kiewiet 1996) and Italy (Curini and Zucchini 2010), and the role of ideology in legislative voting in Brazil (Zucco 2009) and the UK (Kam 2001; Schonhardt-Bailey 2003). Jones and Hwang (2005) use ideal point estimation as part of their extension of Cox and McCubbins’ (1993) cartel theory to the Argentine Congress.

5.5  Trends, Extensions, and Improvements Among the most important extensions to empirical spatial models are those addressing the issue of comparability across time. One such challenge is to make ideal points comparable across time by constraining them to a common scale. The most well-known approach of this sort is the DW-NOMINATE model (Poole and Rosenthal 1997) which constrains ideal points to be a polynomial function of time. An alternative approach has made use of “bridge” observations across time (Bailey 2007), such that voters in different time periods are treated as having voted on the same issues. A second major concern of researchers, so far mainly in the US context but relevant elsewhere, has been establishing comparability between legislators in different chambers. One approach to this is to make use of the common members between two or more

Roll-Call Analysis and the Study of Legislatures  

115

chambers to unify the roll-call data, an approach taken by Poole (1998), Bailey (2007), and Battista et al. (2013). Another approach in the same vein is to rescale coordinates based on a common metric, an approach taken by Shor, Berry, and McCarty (2010) in their work designed to place US State legislators on a common scale. Bridging techniques also allow other institutions and actors to be considered on the same scale as legislators, as has been done in the case of US presidents (e.g. McCarty and Poole 1995; Bailey and Chang 2001; Bailey 2007; Treier 2011). Another issue of particular concern for comparative applications is the matterof abstentions. Since abstentions are numerous in many chambers outside Congress, questions of how abstentions should be treated have led to the application of multichotomous choice models (e.g. Rosas and Shomer 2008). Finally, a major source of innovation recently in the broader genre of preference measures has been the application of these methods to data sources other than roll-call votes. These range from analogous alternatives such as early day motions in the UK (Kellermann 2012) to data with very different properties such as legislative co-sponsorship (Talbert and Potoski 2003; Alemán et al. 2009; Desposato, Kearney, and Crisp 2011; Calvo and Sagarzazu 2011; Barnes 2012), speech (Proksch and Slapin 2008; Bernauer and Bräuninger 2009), and other sources of data with valuable choice information (e.g. Hix and Crombez 2005). Co-sponsorship, in particular, has been a source of interest for the closely related method of network analysis (Fowler 2006; Alemán 2009).

5.6  Problems and Issues with Scaling Techniques Although it is a widespread practice to use ideal point estimates as a form of raw data processed through subsequent statistical analysis, both as independent and dependent variables, it is important to note that underlying each estimate is a stochastic process that ideally should be taken into account for many applications. One initial limitation of estimation via the NOMINATE model on this front, and an attraction of Bayesian estimators, was the lack of reliable measures of uncertainty for the ideal point estimates (see Poole and Rosenthal 1997). Such uncertainty estimates are often important in establishing the statistical meaning in the difference between the ideal points of different voters or between those of the same voter across time. This was addressed in the NOMINATE framework through use of the parametric bootstrap (Lewis and Poole 2004; Carroll et al. 2009b; Poole et al. 2011), providing standard errors to enable statistical tests of the difference between legislator points. For all parametric methods, a significant issue persists in obtaining reliable interval information for ideal points—that is, a meaningful basis for cardinal distances between legislators’ ideal points. Although the recovery of rank orders has proven extremely

116   Royce Carroll and Keith Poole robust via numerous methods, the meaning of the distance between points remains an area of substantial concern. In particular, with low levels of error, there is no basis for the recovery of interval-level parameter estimates in spatial (geometric) models of parliamentary voting. This problem is especially severe for the legislators that would be interpreted as “extreme.” An inherent problem of probabilistic estimators is the issue of “errorless” voting by a legislator at the exterior of the preference dimension—those at the exterior of the policy space—which is particularly problematic in smaller chambers (such as committees or courts) where these constitute a larger proportion of the membership. Errorless voting is often better assessed through the non-parametric method of Optimal Classification (Rosenthal and Voeten 2004), although this method lacks the uncertainty measures available in other methods (Poole 2005). Perhaps the greatest limitation for the spatial analysis of roll-call data is simply the degree to which the data fit the assumptions of the model. For researchers aiming to obtain a measure of preferences, roll-call votes are only as useful as the underlying process by which they are generated. Despite the wide acceptance of the face validity of data generated from US Congressional voting, in some cases roll calls simply do not contain information on the sincere preferences of individual legislators and are not useful for such applications. Thus, scholars can fruitfully apply these methods only with deep knowledge of the meaning behind the input data. While scaling methods can always identify latent dimensions of difference among legislators, what these differences represent may vary dramatically from case to case. Hence one cannot assume comparability between the type of information uncovered in any two legislative environments. In situations where strategic voting is widespread—where voting is motivated by factors unrelated to one’s preferences regarding the content of the policies under consideration—one cannot assume that the variation in the data can be reasonably interpreted as reflecting the preferences of individual legislators. For instance, in systems with very high party discipline, defections may not be directly a function of preferences and therefore will not provide reliable preference information on the individuals within the party (Spirling and McLean 2007; Spirling 2010). Of course, behaviour in such cases may nevertheless provide meaningful information on group-level preferences or on the position-taking strategies involved with these voting decisions. Another widespread challenge for interpreting the meaning of ideal point estimates occurs when there is a biased selection mechanism for determining which votes occur (e.g. Carrubba et al. 2006; Carrubba, Gabel, and Hug 2008; Hug 2010; see also Clinton and Lapinski 2008). Without choice data that provides information distinguishing legislators on issues where they disagree, no technique can uncover differences in underlying preferences—even if the voting itself reflects those preferences. Perhaps the broadest and most obvious limitation on further expanding the traditional application of spatial models of choice is the lack of recorded voting data. Many countries do not record roll-call votes in significant numbers at the individual level (Carey 2009). As a result, there has been a growing interest in alternative inputs that meet the needs of researchers looking for measures of policy preferences (e.g. Saiegh

Roll-Call Analysis and the Study of Legislatures  

117

2009). Hence, the family of scaling methods discussed will continue to be applied to legislative environments far beyond the existence and reliability of roll-call data. Of course, for studies intending to explain voting behaviour itself, issues regarding limitations in the selection of recorded votes must be addressed directly.

5.7 Conclusion Spatial methods of analysing choice data have made an enormous contribution to our understanding of legislatures. Beyond the countless studies dealing with the US Congress, numerous quantitative analyses of state legislatures and assemblies outside the US have been enabled by applying spatial methods to roll-call data. However, substantial limitations exist in our ability to export standard interpretations of ideal point estimates derived from the work on the US Congress. This does not mean spatial roll-call analysis cannot provide useful information in most contexts, but each application must fundamentally reassess the meaning of these measures. Doing so requires a thorough evaluation of the numerous variables leading up to the choice in question before one chooses to interpret the data through a spatial model. That, in turn, means acquiring deep knowledge of the processes and norms of the legislatures we study.

Notes 1. See Poole (2008), on which this section draws, for a thorough review of this work. 2. The method is as follows. First, transform the observed similarities/dissimilarities into squared distances. (For example, if the matrix is a Pearson correlation matrix subtract all the entries from 1 and square the result.) Next, double-centre the matrix of squared distances by subtracting from each entry in the matrix the mean of the row, the mean of the column, adding the mean of the matrix, and then dividing by -2. This has the effect of removing the squared terms from the matrix leaving just the cross-product matrix (see Gower 1966). Finally, perform an eigenvalue-eigenvector decomposition to solve for the coordinates. 3. Graphing the “true” (that is, the estimated or reproduced) distances—the d’s—versus the observed dissimilarities—the δ’s—revealed the relationship between them. This became known as the “Shepard diagram.” 4. This became known as KYST. Later de Leeuw (1977; 1988) and de Leeuw and Heiser (1977) modernized KYST with the SMACOF algorithm. 5. It is a set of items (questions, problems, etc.) that is ranked in order of difficulty so that those who answer correctly (agree) on a more difficult (or extreme) item will also answer correctly (agree) will all less difficult (extreme) items that preceded it. 6. Both unfolding analysis and scalogram analysis deal with individuals’ responses to a setof stimuli. But whereas unfolding analysis assumes a single-peaked (usually symmetric) utility function, Guttman scaling and IRT models are based on a utility function that is always monotonically increasing or decreasing over the relevant dimension or space. Yet, these

118   Royce Carroll and Keith Poole models are observationally equivalent in the binary choice context typical of parliamentary voting (Weisberg 1968; Poole 2005). 7. See Cromwell (1982; 1985) for similar applications on divisions in the UK parliament, Best (1982; 1995) for applications to parliaments in Europe. See Broach (1972) and Harmel and Hamm (1986) for early applications to US state legislatures and Collie (1984) for a broad review of later work. 8. See Poole (2005), on which this section draws for a more extensive discussion of these models. 9. With a quadratic deterministic utility function the simple item response model (Rasch 1961) is mathematically equivalent to the basic spatial model if legislators have quadratic utility functions with additive random error (Ladha 1991; Londregan 2000; Clinton, Jackman, and Rivers 2004). This has the effect of making the estimation quite straightforward as it boils down to a series of linear regressions. 10. Shapes vary by party group as follows, European People’s Party–European Democrats (EPP–ED), Party of European Socialists (SOC). Alliance of Liberals and Democrats for Europe (ALDE), Union for Europe of the Nations (UEN), the Greens–European Free Alliance (G/EFA), European United Left–Nordic Green Left (EUL/NGL), Independence/ Democracy (IND/DEM). 11. See Lunn et al. (2000). See Armstrong et al. (2014) for an overview of these and other software implementations.

References Ágh, A., 1999. The Parliamentarization of the East Central European Parties: Party Discipline in the Hungarian Parliament, 1990-1996. In S. Bowler, D.M. Farrell, and R.S. Katz (eds.). Party Discipline and Parliamentary Government, pp. 167–88. Columbus:  Ohio State University Press. Aldrich, J. H. and Battista, J. S. C., 2002. Conditional Party Government in the States. American Journal of Political Science, 46: 164–72. Aldrich, J. H. and Rohde, D. W., 1998. Measuring Conditional Party Government. Presented at the Annual Meeting of the Midwest Political Science Association, April 23–25, Chicago. Alemán, E., 2009. Institutions, Political Conflict, and the Cohesion of Policy Networks in the Chilean Congress, 1961-2006. Journal of Latin American Studies, 41: 467–91. Alemán, E. and Saiegh, S. M., 2007. Legislative Preferences, Political Parties, and Coalition Unity in Chile. Comparative Politics, 39: 253–72. Alemán, E., Calvo, E., Jones, M. P., and Kaplan, N., 2009. Comparing Cosponsorship and Rollcall Ideal Points. Legislative Studies Quarterly, 34: 87–116. Ansolabehere, S., Snyder, J. M., and Stewart III, C., 2001. Candidate Positioning in U.S. House Elections. American Journal of Political Science, 45: 136–59. Armstrong, D., Bakker, R., Carroll, R., Hare, C., Poole, K., and Rosenthal, H., 2014. Analyzing Spatial Models of Choice and Judgment with R. Boca Raton: Chapman & Hall/CRC. Bagashka, T., 2008. Invisible Politics: Institutional Incentives and Legislative Alignments in the Russian Duma, 1996–99. Legislative Studies Quarterly, 33: 415–44. Bailey, M. A., 2007. Comparable Preference Estimates across Time and Institutions for the Court, Congress, and Presidency. American Journal of Political Science, 51: 433–48. Bailey, M. A. and Chang, K. H., 2001. Comparing Presidents, Senators, and Jus­tices: Interinsti­ tutional Preference Estimation. Journal of Law, Economics, and Organization, 17: 477–506.

Roll-Call Analysis and the Study of Legislatures  

119

Barnes, T., 2012. Gender and Legislative Preferences: Evidence from the Argentine Provinces. Politics and Gender, 8: 483–507. Battista, J. C., Peress, M., and Richman, J., 2013. Common-space Ideal Points, Committee Assignments, and Financial Interests in the State Legislatures. State Politics & Policy Quarterly, 13: 70–87. Bernauer, J. and Bräuninger, T., 2009. Intra-party Preference Heterogeneity and Faction Mem­ bership in the 15th German Bundestag: A Computational Text Analysis of Parliamentary Speeches. German Politics, 18: 385–402. Best, H., 1982. Recruitment, Careers and Legislative Behavior of German Parliamentarians, 1848-1953. Historical Social Research/Historische Sozialforschung, 23: 20–54. Best, H., 1995. “Disorder Yields to Order Fair the Place”: The Emergence of Political Parties in Western and Central Europe. Parliaments, Estates & Representation, 15: 133–42. Black, D., 1948. On the Rationale of Group Decision-making. Journal of Political Economy, 56:23–34. Black, D., 1958. The Theory of Committees and Elections. Cambridge: Cambridge University Press. Broach, G. T., 1972. A Comparative Dimensional Analysis of Partisan and Urban-rural Voting in State Legislatures. The Journal of Politics, 34: 905–21. Calvo, E. and Sagarzazu, I., 2011. Legislator Success in Committee: Gatekeeping Authority and the Loss of Majority Control. American Journal of Political Science, 55: 1–15. Carey, J. M., 2009. Legislative Voting and Accountability. Cambridge: Cambridge University Press. Carroll, J. D. and Chang, J. J., 1970. Analysis of Individual Differences in Multidimensional Scaling via an N-way Generalization of “Eckart-Young” Decomposition. Psychometrika, 35: 283–320. Carroll, R. and Eichorst, J., 2013. The role of party: The legislative consequences of partisan electoral competition. Legislative Studies Quarterly, 38: 83–109. Carroll, R., Lewis, J. B., Lo, J., Poole, K. T., and Rosenthal, H., 2009a. Measuring Bias and Uncer­ tainty in DW-NOMINATE Ideal Point Estimates via the Parametric Bootstrap. Political Analysis, 17: 261–75. Carroll, R., Lewis, J. B., Lo, J., Poole, K. T., and Rosenthal, H., 2009b. Comparing NOMINATE and IDEAL:  Points of Difference and Monte Carlo Tests. Legislative Studies Quarterly, 34: 555–91. Carroll, R.,Lewis, J. B., Lo, J., Poole, K. T., and Rosenthal, H., 2013. The Structure of Utility in Spatial Models of Voting. American Journal of Political Science, 57: 1008–28. Chang, J. J. and Carroll, J. D., 1969. How to use MDPREF: A Computer Program for Multidi­ mensional Analysis of Preference Data. Multidimensional Scaling Program Package of Bell Laboratories. Murray Hill:Bell Laboratories. Carrubba, C., Gabel, M., and Hug, S., 2008. Legislative Voting Behavior, Seen and Unseen: A The­ory of Rollcall Vote Selection. Legislative Studies Quarterly, 33: 543–72. Carrubba, C. J., Gabel, M., Murrah, L., Clough, R., Montegomery, E., and Schambach, R., 2006. Off the Record: Unrecorded Legislative Votes, Selection Bias, and Roll-call Vote Aote analysis. British Journal of Political Science, 36: 691–704. Chiou, F.-Y. Y. and Rothenberg, L. S., 2003. When Pivotal Politics Meets Partisan Politics. American Journal of Political Science, 47: 503–22. Clausen, A. R., 1973. How Congressmen Decide: A Policy Focus. New York:Palgrave Macmillan. Clausen, A. R. and Cheney, R. B., 1970. A Comparative Analysis of Senate-house Voting on Economic and Welfare Policy, 1953-1964. American Political Science Review, 64: 138–52. Clinton, J. D., 2007. Lawmaking and Roll Calls. Journal of Politics, 69: 455–67.

120   Royce Carroll and Keith Poole Clinton, J. D., 2012. Congress, Lawmaking and the Fair Labor Standards Act, 1971-2000. American Journal of Political Science, 56: 355–72. Clinton, J. D., 2012. Using Roll Calls to Test Models of Politics. Annual Review of Political Science, 15: 79–99. Clinton, J. D. and Lapinski, J., 2008. Laws and Roll Calls in the U.S. Congress, 1891–1994. Legislative Studies Quarterly, 33: 511–41. Clinton, J. D., Jackman, S. D., and Rivers, D., 2004. The Statistical Analysis of Roll Call Data: A Unified Approach. American Political Science Review, 98: 355–70. Clinton, J. D., Bertelli, A., Grose, C., Lewis, D., and Nixon, D., 2012. Separated Powers in the United States:  The Ideology of Agencies, Presidents and Congress. American Journal of Political Science,56: 341–54. Collie, M. P., 1984. Voting Behavior in Legislatures. Legislative Studies Quarterly, 9: 3–50. Coombs, C., 1950. Psychological Scaling without a Unit of Measurement. Psychological Review, 57: 148–58. Coombs, C., 1952. A Theory of Psychological Scaling. Engineering Research Bulletin Number 34. Ann Arbor: University of Michigan Press. Coombs, C., 1958. On the Use of Inconsistency of Preferences in Psychological Measurement. Journal of Experimental Psychology, 55: 1–7. Coombs, C., 1964. A Theory of Data. New York: Wiley. Cox, G. W. and McCubbins, M. D., 1993. Legislative Leviathan: Party Government in the House. Berkeley: University of California Press. Cox, G. W. and Poole, K. T., 2002. On Measuring Partisanship in Roll Call Voting: The U.S. House of Representatives, 1877–1999. American Journal of Political Science, 46: 477–89. Cox, G. W. and McCubbins, M. D., 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press. Cromwell, V., 1982. Mapping the Political World of 1861: A Multidimensional Analysis of House of Commons’ Division Lists. Legislative Studies Quarterly, 7: 281–97. Cromwell, V., 1985. Computer Analysis of House of Commons’ Division Lists, 1861– 1936: A Report on Current Research. Parliaments, Estates and Representation, 5: 33–6. Curini, L. and Zucchini, F., 2010. Testing the Law-making Theories in a Parliamentary Democracy: A Roll Call Analysis of the Italian Chamber of Deputies (1988–2008). In T. König, M. Debus, and G. Tsebelis (eds.).Reform Processes and Policy Change, pp. 189–212. New York: Springer New York. Desposato, S. W., 2006a. Parties for Rent? Ambition, Ideology, and Party Switching in Brazil’s Chamber of Deputies. American Journal of Political Science, 50: 62–80. Desposato, S. W., 2006b. The Impact of Electoral Rules on Legislative Parties: Lessons from the Brazilian Senate and Chamber of Deputies. Journal of Politics, 68: 1018–30. Desposato, S. W, Kearney, M. C., and Crisp, B. F., 2011. Using Cosponsorship to Estimate Ideal Points. Legislative Studies Quarterly, 36: 531–65. De Leeuw, J., 1977. Applications of Convex Analysis to Multidimensional Scaling. In J. R. Barra, F. Brodeau, G. Romier, and B. van Cutsem (eds.).Recent Developments in Statistics, pp. 133– 45. Amsterdam: North-Holland. De Leeuw, J., 1988. Convergence of the Majorization Method for Multidimensional Scaling. Journal of Classification, 5:163–80. De Leeuw, J. and Heiser, W. J., 1977. Convergence of Correction-matrix Algorithms for Multidimensional Scaling. In J. C. Lingoes, E. E. Roskam, and I. Borg (eds.). Geometric Representations of Relational Data,pp. 735–52. Ann Arbor: Mathesis Press.

Roll-Call Analysis and the Study of Legislatures  

121

Dewan, T. and Spirling, A., 2011. Strategic Opposition and Government Cohesion in Westminster Democracies. American Political Science Review, 105: 337–58. Eckart, C. H. and Young, G., 1936. The Approximation of One Matrix by Another of Lower Rank. Psychometrika, 1: 211–18. Forgette, Richard and Sala, Brian R., 1999. Conditional Party Government and Member Turnout on Senate Recorded Votes, 1873-1935. Journal of Politics, 61: 467–84. Fowler, J. H., 2006. Connecting the Congress: A Study of Cosponsorship Networks. Political Analysis, 14: 456–87. Geman, D. and Geman, S.,1984. Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 6:721–41. Gelfand, A. E. and Smith, A. F. M., 1990. Sampling-based Approaches to Calculating Marginal Densities. Journal of the American Statistical Association, 85: 398–409. Gelman, A., 1992. Iterative and Non-iterative Simulation Algorithms. Computing Science and Statistics, 24: 433–38. Gelman, A., Carlin, J. B., Stern, H. S., and Rubin, D. B., 2000. Bayesian Data Analysis. New York: Chapman and Hall/CRC. Gill, J., 2002. Bayesian Methods: A Social and Behavioral Sciences Approach. Boca Raton: Chapman andHall/CRC. Godbout, J.-F. and Høyland, B., 2011. Legislative Voting in the Canadian Parliament. Canadian Journal of Political Science/Revue canadienne de science politique, 44: 367–88. Gower, J. C., 1966. Some Distance Properties of Latent Root and Vector Methods Used in Multivariate Analysis. Biometrika, 53: 325–38. Guttman, L. L., 1944. A Basis for Scaling Qualitative Data. American Sociological Review, 9: 139–50. Guttman, L. L., 1950. The Basis for Scalogram Analysis. In Stouffer, S. A., Guttman, L., Suchman, E. A., Lazarsfeld, P. F., Star, S. A., and Clausen, J. A. (eds.). Measurement and Prediction: The American Soldier,Vol. IV, pp. 60–90. New York: Wiley. Hagemann, S., 2007. Applying Ideal Point Estimation Methods to the Council of Ministers. European Union Politics, 8: 279–96. Hagemann, S. and Hoyland, B., 2008. Parties in the Council? Journal of European Public Policy, 15: 1205–21. Hansen, M. E., 2008. Reconsidering the Party Distances and Dimensionality of the Danish Folketing. The Journal of Legislative Studies, 14: 264–78. Hansen, M. E., 2009. The Positions of Irish Parliamentary Parties 1937–2006. Irish Political Studies, 24: 29–44. Hansen, M. E. and Debus, M., 2011. The Behaviour of Political Parties and MPs in the Parliaments of the Weimar Republic. Party Politics, 18: 709–26. Harmel, R. and Hamm, K. E., 1986. Development of a party role in a no-party legislature. Political Research Quarterly, 39: 79–92. Hastings, W. K., 1970. Monte Carlo Sampling Methods Using Markov Chains and their Applications. Biometrika, 54: 97–109. Hix, S., 2001. Legislative Behaviour and Party Competition in the European Parliament: An Application of Nominate to the EU. JCMS: Journal of Common Market Studies, 39: 663–88. Hix, S. and Crombez, C., 2005. Extracting Ideal Point Estimates from Actors’ Preferences in the EU Constitutional Negotiations. European Union Politics, 6: 353–76. Hix, S. and Jun, H.W., 2009. Party Behaviour in the Parliamentary Arena. Party Politics, 15: 667–94.

122   Royce Carroll and Keith Poole Hix, S., Noury, A., and Roland, G., 2006. Dimensions of Politics in the European Parliament. American Journal of Political Science, 50: 494–520. Hix, S., Noury, A., and Roland, G., 2009. Voting Patterns and Alliance Formation in the European Parliament. Philosophical Transactions of the Royal Society B, 364: 821–31. Hotelling, H., 1929. Stability in Competition. Economic Journal, 39: 41–57. Hotelling, H., 1933. Analysis of a Complex Statistical Variables with Principal Components. Journal of Educational Psychology, 24: 498–520. Hug, S., 2010. Selection Effects in Roll Call Votes. British Journal of Political Science, 40: 225–35. Hug, S. and Schulz, T., 2007. Left-right Positions of Political Parties in Switzerland. Party Politics, 13: 305–30. Jackman, S. D., 2000a. Estimation and Inference via Bayesian Simulation: An Introduction to Markov Chain Monte Carlo. American Journal of Political Science, 44: 375–404. Jackman, S. D., 2000b. Estimation and Inference are Missing Data Problems: Unifying Social Science Statistics via Bayesian Simulation. Political Analysis, 8: 307–32. Jackman, S. D., 2001. Multidimensional Analysis of roll Call Data via Bayesian Simulation: Identification, Estimation, Inference and Model Checking. Political Analysis, 9: 227–41. Jackman, S. D., 2008. pscl:  Classes and Methods for R Developed in the Political Science Computational Laboratory. . Jenkins, J. A., 1999. Examining the Bonding Effects of Party:  A  Comparative Analysis of Roll-call Voting in the U.S. and Confederate Houses. American Journal of Political Science, 43: 1144–65. Jones, M. P. and Hwang, W., 2005. Party Government in Presidential Democracies: Extending Cartel Theory Beyond the U.S. Congress. American Journal of Political Science, 49: 267–82. Jones, M.P., Hwang, W., and Micozzi, J.P., 2009. Government and Opposition in the Argentine Congress, 1989-2007: Understanding Inter-party Dynamics Through Roll Call Vote. Journal of Politics in Latin America 1: 67–96. Kam, C., 2001. Do Ideological Preferences Explain Parliamentary Behaviour? Evidence from Great Britain and Canada. The Journal of Legislative Studies, 7: 89–126. Kellermann, M.,2012. Estimating Ideal Points in the British House of Commons Using Early Day. American Journal of Political Science, 56: 757–71. Koford, K., 1989. Dimensions in Congressional Voting. American Political Science Review, 83: 949–62. Krehbiel, K., 1998. Pivotal Politics. Chicago: University of Chicago Press. Krehbiel, K., Meirowitz, A., and Woon, J., 2005. Testing Theories of Lawmaking. Social Choice and Strategic Decisions: 249–68. Kruskal, J. B., 1964a. Multidimensional Scaling by Optimizing a Goodness of Fit to a Nonmetric Hypothesis. Psychometrika, 29: 1–27. Kruskal, J. B., 1964b. Nonmetric Multidimensional Scaling:  A  Numerical Method. Psychometrika, 29: 115–29. Kruskal, J. B., 1965. Analysis of Factorial Experiments by Estimating Monotone Transformations of the Data. Journal of the Royal Statistical Society B, 27: 251–63. Kruskal, J. B. and Wish, M., 1978. Multidimensional Scaling. Beverly Hills: Sage. Kruskal, J. B., Young, F. W., and Seery, J. B., 1973. How to Use KYST: A Very Flexible Program to do Multidimensional Scaling and Unfolding. Multidimensional Scaling Program Package of Bell Laboratories. Murray Hill:Bell Laboratories. Ladha, K. K., 1991. A Spatial Model of Legislative Voting with Perceptual Error. Public Choice, 68: 151–74.

Roll-Call Analysis and the Study of Legislatures  

123

Lauderdale, B. E., 2010. Unpredictable Voters in Ideal Point Estimation. Political Analysis, 18: 151–71. Levitt, S., 1996. How do senators vote? Disentangling the Role of Voter Preferences, Party Affiliation, and Senator Ideology. American Economic Review, 86: 425–41. Londregan, J., 2000. Estimating Legislators’ Preferred Points. Political Analysis, 8: 35–56. Londregan, J. and Snyder, J. M., 1994. Comparing Committee and Floor Preferences. Legislative Studies Quarterly,19: 233–66. Lunn, D. J., Thomas, A., Best, N., and Spiegelhalter, D., 2000. WinBUGS-a Bayesian Modelling Framework: Concepts, Structure, and Extensibility. Statistics and computing, 10: 325–37. Lyons, P. and Lacina, T., 2009. An Examination of Legislative Roll Call Voting in the Czech Republic Using Spatial Models. Czech Sociological Review, 45: 1155–90. MacRae, D., Jr., 1958. Dimensions of Congressional Voting. Berkeley: University of California Press. MacRae, D., Jr., 1970. Issues and Parties in Legislative Voting. New York: Harper and Row. Martin, A. D. and Quinn, K. M., 2002. Dynamic Ideal Point Estimation via Markov Chain Monte Carlo for the U.S. Supreme Court, 1953-1999. Political Analysis, 10: 134–15. Martin, A. D. and Quinn, K. M., 2009. MCMCpack: Markov Chain Monte Carlo (MCMC). R Package. . McCarty, N., Poole, K. T., and Rosenthal, H., 2001. The Hunt for Party Discipline in Congress. American Political Science Review, 95: 673–87. McCarty, N. M. and Poole, K. T., 1995. Veto Power and Legislation: An Empirical Analysis of Executive and Legislative Bargaining from 1961 to 1986. Journal of Law, Economics, and Organization, 11: 282–312. McCarty, N. M., Poole, K. T., and Rosenthal, H., 2006. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge: MIT Press. McFadden, D., 1976. Quantal Choice Analysis:  A  Survey. Annals of Economic and Social Measurement, 5: 363–90. Metropolis, N. C. and Ulam, S., 1949. The Monte Carlo Method. Journal of the American Statistical Association, 44: 335–41. Montinola, G. R., 1999. Politicians, Parties, and the Persistence of Weak States: Lessons from the Philippines. Development and Change, 30: 739–74. Morgenstern, S., 2004. Patterns of Legislative Politics: Roll-call Voting in Latin America and the United States. Cambridge: Cambridge University Press. Myagkov, M. G. and Kiewiet, D. R., 1996. Czar Rule in the Russian Congress of People’s Deputies? Legislative Studies Quarterly, 21: 5–40. Pearson, K. P., 1901. On lines and Planes of Closest fit to Systems of Points in Space. The London, Edinburgh and Dublin Philosophical Magazine and Journal, 6: 559–72. Poole, K.,1984. Least Squares Metric, Unidimensional Unfolding. Psychometrika, 49: 311–23. Poole, K. T., 1988. Recent Developments in Analytical Models of Voting in the U. S. Congress. Legislative Studies Quarterly, 13: 117–33. Poole, K. T., 2000. Non-parametric Unfolding of Binary Choice Data. Political Analysis, 8:211–37. Poole, K. T., 2001. The Geometry of Multidimensional Quadratic Utility in Models of Parliamentary Roll Call Voting. Political Analysis, 9: 211–26. Poole, K. T., 2005. Spatial Models of Parliamentary Voting. New York: Cambridge University Press. Poole, K., 2008. The Evolving Influence of Psychometrics in Political Science. In J. Box-Steffensmeier, H.E. Brady, and D. Collier (eds.).The Oxford Handbook of Political Methodology, pp. 199–221. Oxford: Oxford University Press. Poole, K. T. and Lewis, J. B., 2004. Measuring Bias and Uncertainty in Ideal Point Estimates via the Parametric Bootstrap. Political Analysis, 12: 105–27.

124   Royce Carroll and Keith Poole Poole, K. T. and Rosenthal, H., 1985. A Spatial Model for Legislative Roll Call Analysis. American Journal of Political Science, 29: 357–84. Poole, K. T. and Rosenthal, H., 1991. Patterns of Congressional Voting. American Journal of Political Science, 35: 228–78. Poole, K. T. and Rosenthal, H., 1997. Congress: A Political-Economic History of Roll Call Voting. New York: Oxford University Press. Poole, K., Lewis, J., Lo, J., and Carroll, R.,2011. Scaling Roll Call Votes with Wnominatein R. Journal of Statistical Software, 42: 1–21. Poole, K. T., Lewis, J. B., Lo, J., and Carroll, R., 2012a. wnominate: WNOMINATE Roll Call Analysis Software. R package. . Poole, K. T., Lewis, J. B., Lo, J., and Carroll, R.,2012b oc: Optimal Classification Roll Call Analysis Software. R package. . Proksch, S. O. and Slapin, J. B., 2010. Position Taking in European Parliament Speeches. British Journal of Political Science, 40: 587–611. Quinn, K. M., 2004. Bayesian Factor Analysis for Mixed Ordinal and Continuous Responses. Political Analysis, 12: 338–53. Quinn, K. M. and Martin, A. D., 2002. An Integrated Computational Model of Multiparty Electoral Competition. Statistical Science, 17: 405–19. Quinn, K. M., Martin, A. D., and Whitford, A. B., 1999. Voter Choice in Multi-party Democracies:  A  Test of Competing Theories and Models. American Journal of Political Science, 43: 1231–47. Rasch, G., 1961. On General Laws and the Meaning of Measurement in Psychology. Proceedings of the IV Berkeley Symposium on Mathematical Statistics and Probability, 4: 321–33. Rosas, G. and Shomer, Y., 2008. Models of Nonresponse in Legislative Politics. Legislative Studies Quarterly, 33: 573–601. Rosenthal, H. and Voeten, E., 2004. Analyzing Roll Calls with Perfect Spatial Voting: France 1946–1958. American Journal of Political Science, 48: 620–32. Saiegh, S. M., 2009. Recovering a Basic Space from Elite Surveys: Evidence from Latin America. Legislative Studies Quarterly, 34: 117–45. Sauger, N., 2009. Party Discipline and Coalition Management in the French Parliament. West European Politics, 32: 310–26. Schonhardt-Bailey, C., 2003. Ideology, Party and Interests in the British Parliament of 1841–47. British Journal of Political Science, 33: 581–605. Shepard, R. N., 1962a. The Analysis of Proximities: Multidimensional Scaling with An Unknown Distance Function. I. Psychometrika, 27: 125–39. Shepard, R. N., 1962b. The Analysis of Proximities: Multidimensional Scaling with An Unknown Distance Function. II. Psychometrika, 27: 219–46. Shor, B., Berry, C., and McCarty, N., 2010. A Bridge to Somewhere:  Mapping State and Congressional Ideology on a Cross-institutional Common Space. Legislative Studies Quarterly, 35: 417–48. Snyder, J. and Groseclose, T., 2000. Estimating Party Influence in Congressional Roll-call Voting. American Journal of Political Science, 44: 193–211. Spearman, C. E., 1904. General Intelligence’ Objectively Determined and Measured. American Journal of Psychology, 15: 201–93. Spirling, A., 2010. Identifying Intraparty Voting Blocs in the U.K. House of Commons. Journal of the American Statistical Association, 105: 447–57.

Roll-Call Analysis and the Study of Legislatures  

125

Spirling, A. and McLean, I., 2007. UK OC OK? Interpreting Optimal Classification Scores for the U.K. House of Commons. Political Analysis, 15: 85–96. Stiglitz, E. H. and Weingast, B. R., 2010. Agenda Control in Congress: Evidence from Cut Point Estimates and Ideal Point Uncertainty. Legislative Studies Quarterly, 35: 157–85. Takane, Y., Young, F. W., and de Leeuw, J., 1977. Nonmetric Individual Differences Multidimensional Scaling:  An Alternating Least-squares Method with Optimal Scaling Features. Psychometrika, 42: 7–67. Thurstone, L. L., 1931. Multiple Factor Analysis. Psychological Review, 38: 406–27. Thurstone, L. L., 1947. Multiple Factor Analysis. Chicago: University of Chicago Press. Torgerson,W.S.,1952.MultidimensionalScaling: I. TheoryAndMethod.Psychometrika,17: 401–19. Torgerson, W. S., 1958. Theory and Methods of Scaling. New York: Wiley. Treier, S., 2011. Comparing Ideal Points Across Institutions and Time. American Politics Research, 39:804–831. Voeten, E., 2000. Clashes in the Assembly. International Organization, 54: 185–215. Voeten, E., 2004. Resisting the Lonely Superpower: Responses of States in the United Nations to U.S. Dominance. Journal of Politics, 66: 729–54. Weisberg, H. F., 1968. Dimensional Analysis of Legislative Roll Calls. Doctoral dissertation, University of Michigan. Wilcox, C. and Clausen, A., 1991. The Dimensionality of Roll-call Voting Reconsidered. Legislative Studies Quarterly, 16: 393–406. Wright, G. C. and Schaffner, B. F., 2002. The Influence of Party:  Evidence from the State Legislatures. American Political Science Review, 96: 367–79. Young, G and Householder, A. S., 1938. Discussion of a Set of Points in Terms of their Mutual Distances. Psychometrika, 3: 19–22. Young, F. W., de Leeuw, J., and Takane, Y., 1976. Regression with Quantitative and Qualitative Variables:  An Alternating Least Squares Method with Optimal Scaling Features. Psychometrika, 41: 505–29. Zucco, C., 2009. Ideology or what? Legislative Behavior in Multiparty Presidential Settings. The Journal of Politics, 71: 1076–92. Zucco, C. and Lauderdale, B. E., 2011. Distinguishing between Influences on Brazilian Legislative Behavior. Legislative Studies Quarterly, 36: 363–96.

C HA P T E R

6

WO R D S A S DATA :  C O N T E N T A NA LYS I S I N L E G I S L AT I V E STUDIES JONAT HA N B. SL A PI N A N D S V E N - OL I V E R PROK S C H

6.1 Introduction Every year, democratic legislatures produce reams upon reams of documents. The stack of paper includes draft bills, amendments to bills, committee reports, transcripts of floor debates, parliamentary questions, reports of special investigations, adopted legislation, and press releases. Until recently most of this material went unexplored, largely due to the daunting task of compiling it into a useable format and the lack of appropriate content analytic techniques. Sifting through the entire legislative record without the appropriate tools would be a fool’s errand. Instead, legislative studies have been largely confined to the analysis of roll-call votes—a small, but relatively easily defined and collected portion of the legislative record. As parliamentary records have gone electronic, with parliaments storing documents in easily searchable on-line databases, their content has become more accessible to researchers and available for large-scale data analysis. Advances in computing power, machine learning techniques, and statistical methods have allowed researchers to develop tools to extract systematic meaning from the large corpus of text that legislatures produce. Taken together, these advances have enabled political scientists to take advantage of this treasure trove of data to study representation, party politics, policy-making, and legislative behaviour. Over the last decade, advances in content analysis techniques have led to an improved understanding of party politics and legislative behaviour, as researchers have been able to explore previously unanswerable questions. We begin by describing content analysis, in particular recent approaches that treat “words as data,” describe some important applications to comparative legislative studies, and then move

Words as Data: Content Analysis in Legislative Studies  

127

on to discuss the variety of challenges scholars face moving forward. Lastly, we summarize the continuing trends in the field.

6.2  Words as Data: Classical Content Analysis Kimberly Neuendorf describes content analysis as “the systematic, objective, quantitative analysis of message characteristics” (Neuendorf 2002, 1). The goal of content analysis is to extract meaningful content from an entire corpus of text in a systematic way. To do so effectively, scholars must analyse texts relevant to the research questions they wish to answer, consider the actors and processes that produced these texts, consider the techniques to analyse the content of texts, and consider what can be learned from the analysis. Content analysis has a long history in the social science and communications literature. It was originally used to explore trends of foreign policy coverage in newspapers, and during the Second World War, it was used as an intelligence tool to gain information about changes in German policy, especially toward the Soviet Union. Following the Second World War, political scientists, sociologists, and communications scholars took up content analysis to study electoral campaigns and media, among other topics (for a detailed history, see Krippendorff 2004). Today, legislative scholars in political science, in particular, use content analytic techniques to learn about the nature of party ideology (e.g. Budge et al. 2001; Laver, Benoit, and Garry 2003; Slapin and Proksch 2008; Schonhardt-Bailey 2008), representation (e.g. Martin 2011; Grimmer 2011), issue salience (e.g. Quinn et al. 2010), or the ideological leaning of news media (e.g. Groseclose and Milyo 2005), to name just a few of the many possible applications. Most of the work in these areas can be classified as quantitative content analysis, in which researchers turn words into numbers—they may look at the overall length of texts, or classify or scale documents based on word counts or sentences. This is in contrast with qualitative approaches, which typically involve reading all documents and offering a synthesis or interpretation of their content. Before the advent of computer-aided techniques, quantitative content analysis often involved physically measuring the length of text (i.e. taking a ruler to the front page of a newspaper), counting headlines, or manually classifying large quantities of text using human coders, a very labour- and time-intensive process (see Krippendorff 2004, 5–9). In recent years, quantitative content analysis has been greatly aided by advances in computer science, statistics, and machine learning techniques. It has become much easier to parse documents into words and sentences, count these units, and extract meaning from their relative frequencies across documents. Of course, computer-based content analysis is not without its problems (and, naturally, its critics). We explore the various

128   Jonathan B. Slapin and Sven-Oliver Proksch advantages and disadvantages of using computers to machine read and extract meaning from texts below, specifically with regard to legislative and party documents.

6.3  Content Analysis and the Comparative Study of Parties and Legislatures Legislative scholars use content analysis to examine at least seven features of legislatures:  party ideology and polarization, government positioning, parliamentary scrutiny, constituency-based representation, policy agenda, quality of debate, and the role of media. Table 6.1 lists these themes together with their main approaches, core assumptions, requirements, typical data sources, and examples.

6. 3.1  Ideology In the vast majority of multiparty democracies, political parties publish manifestos during parliamentary election campaigns. Typically written by the party leadership, manifestos outline the party’s program for government should it get elected, and structure the party’s election campaign. Perhaps the most ambitious content analysis research program in political science is the hand-coding of these party manifestos by the Manifesto Project (Budge et al. 1987; 2001; Klingemann et al. 2006). Formerly known as the Manifesto Research Group and later as the Comparative Manifestos Project (CMP), the project examines party manifestos and government declarations, and the resulting estimates of party ideology and government positions produced by the project have been used in countless studies of legislative behaviour, coalition formation, policy-making, and representation (e.g. Martin and Stevenson 2001; Adams et al. 2006, 2011; Bawn and Rosenbluth 2006; Walgrave and Nuytemans 2009; Lipsmeyer and Pierce 2011; McDonald and Budge 2005; Warwick 2011). The project attempts to uncover the salient dimensions of party competition in a comparative manner across advanced, industrialized democracies. The fundamental notion of party competition guiding the project “was that parties argued with each other by emphasizing different policy priorities rather than by directly confronting each other on the same issues” (Budge et al. 2001, 6–7). The assumption that policy priorities reflect ideological differences has been referred to as a salience theory of party competition. To capture voters, parties tend to arrive at similar positions on the various issues (e.g. everyone has to agree that environmental protection is good—no one wants to compete on a platform of environmental degradation). However, some parties are perceived as being “better” on some issues than others—Green parties are more credible on issues of the environment than liberal parties, while liberal parties are more credible on business issues.

Table 6.1  Examples of Content Analysis Approaches to the Study of Legislatures Approach

Assumption

Requirements

Data

Example

Wordscores

Word counts reflect ideological differences given that words in reference texts constitute relevant dictionary, word meanings are constant

References texts and reference scores

Party manifestos, parliamentary speeches

Laver, Benoit, and Garry (2003)

Wordfish

Ideology is latent and documents can be scaled, word counts reflect ideological differences, word meanings are constant

Identical data-generating Party manifestos, mechanism across parliamentary speeches documents

Slapin and Proksch (2008)

Coding scheme, human coders

Party manifestos or equivalent

Budge et al. (2001)

Coding scheme, human coders

Governmental declarations, coalition agreements

Budge et al. (2001)

Party Ideology

Manifesto Project (CMP) Policy priorities reflect ideological differences, identical coding scheme applicable across countries and time Government Positioning Hand coding

Policy priorities in governmental declarations reflect ideological differences, identical coding scheme applicable across countries and time

Policy Agenda Document classification Parliamentary speeches or press releases reflect and topic modeling latent political priorities

Number of topics needs to Parliamentary speeches, be chosen party press releases

Quinn et al. (2010), Grimmer (2010) (Continued)

Table 6.1 (Continued) Approach

Assumption

Requirements

Data

Example

Oversight and Parliamentary Scrutiny Speech length

Longer parliamentary speeches of coalition parties reflect partisan positioning

Parliamentary speeches

Martin and Vanberg (2008b)

Bill changes

Number of successful article changes in bills reflect more intense coalition partner monitoring

Government bills

Martin and Vanberg (2005, 2011)

Bill length

Proxy for legislative constraints on the bureaucracy Proxy for complexity

Legislative database

Huber and Shipan (2002) Maltzman and Shipan (2008)

Constituency Representation Geography-based dictionary

Geography-based references in questions reflect constituency representation

Dictionary, human coders Parliamentary questions

Martin (2011)

List of think tanks

Newspaper articles, parliamentary speeches

Grosclose and Milyo (2005)

Coding scheme, human coders

Parliamentary speeches

Steiner et al. (2004)

Positions of Media and Legislators Citation patterns of think Legislators and media outlets cite think tanks tanks according to their ideological positions Quality of Debate Discourse analysis

Speeches can be coded according to level of persuasion, justification, or respect

Words as Data: Content Analysis in Legislative Studies  

131

During election campaigns, parties emphasize the issues they are perceived to “own” and ignore others. The CMP coding scheme seeks to capture this facet of party competition. The project sorts each sentence of each manifesto into one of 56 categories. Most of the categories are positive in nature (e.g. positive mentions of the environment on the theory that parties do not make negative mentions of the environment). Thus, the coding scheme captures the relative emphasis a party gives to any one category over the other categories. The goal is to create a single coding scheme valid across countries and time. The most recent iteration of the project coded 1,314 manifestos written by 651 parties over 185 elections in 51 countries (Budge et al. 2006). Each manifesto is coded by hand—each sentence (or quasi-sentence) is read by a researcher and placed into one of the 56 categories (or remains uncoded if it is deemed to fit none of the categories). The end result of the coding process is the percentage of each manifesto falling into each category. These percentages can then be aggregated to produce general left-right party position scales or scales on other policy dimensions (Laver and Budge 1992; Lowe et al. 2011). In recent years, the CMP has been subject to trenchant critiques. The reading and coding of these manifestos is very labour-intensive and prone to inconsistencies, which the CMP does not sufficiently account for (Mikhaylov et al. 2012), and the manual coding protocol does not account for uncertainty in the text generation process (Benoit et al. 2009). Moreover, it is not clear the rigid coding scheme that was devised in the 1980s can be applied in the same way across time and across countries as the manifesto project assumes (Benoit and Laver 2007). To alleviate some of these problems, researchers have attempted to harness the power of computers to extract positions from text. The first attempts to use computers to categorize political texts (and party manifestos in particular) generated dictionaries of words representing a particular ideology, generally culled from manifestos themselves, and then used computers to count the number of times the words in the dictionary appeared in any other given document of interest (Laver and Garry 2000). Laver, Benoit, and Garry (2003) built on this framework to create an automated technique for estimating ideology from manifestos, or any other political text. In creating Wordscores they developed a technique to compare word counts in reference documents that are carefully chosen by the researcher to word counts in documents of interest (virgin documents in the parlance of Laver, Benoit, and Garry). Based on the relation between the relative frequency of words in the reference documents (and the reference scores assigned to them to anchor the political space) and relative frequency of words in the documents of interest, the documents of interest are placed in a one-dimensional policy space. The assumption is that the relative frequency with which parties use words tells researchers something about their latent positions in a policy space. This technique has been applied successfully to examine legislative behaviour and party positions using as data both party manifestos (e.g. Proksch and Slapin 2006; Debus 2009) and parliamentary speeches (e.g. Giannetti and Laver 2005; Hakhverdian2009; Klemmensen et al. 2007; Bernauer and Bräuninger 2009).

132   Jonathan B. Slapin and Sven-Oliver Proksch However, its use has also raised some issues and concerns. First, scholars are tempted to use it as a method for estimating long time-series, but the question of how one chooses reference texts to create comparisons over time is not clear, and the choices one makes with regard to reference texts and positions can greatly affect the analysis (Budge and Pennings 2007). Second, the original transformation of word scores into document scores yields estimates of virgin text positions and reference texts on a different and incomparable scale. As a result, there has been some controversy over the best way to transform the virgin text scores to make them comparable with the reference texts (Lowe 2008; Martin and Vanberg 2008a; Laver and Benoit 2008). Lastly, the choice of reference texts is particularly important when attempting to estimate positions on different policy dimensions. Laver, Benoit, and Garry proposed to use all text in the manifestos in both the virigin and reference documents and to estimate new word scores simply by changing the reference values assigned to the reference documents without changing the actual text input. This means, however, that one poorly chosen reference text or reference value can drastically change the results. In the extreme, the estimated dimension might not be related at all to the dimension of interest since the reference texts do not contain language that discriminates on this dimension. More recent work has attempted to estimate the ideology of parties and individual legislators without relying on dictionaries or reference documents (e.g. Monroe andMaeda 2004; Slapin and Proksch 2008; Proksch and Slapin 2010). Slapin and Proksch (2008) present and estimate a parametric scaling model called Wordfish to place documents in a one-dimensional space based solely on relative word frequencies. Rather than taking advantage of information contained in reference documents, as Wordscores does, Wordfish assumes word frequencies within documents are driven by a statistical model (namely that they are Poisson distributed), and are the result of underlying latent policy positions that can be estimated. The technique is thus applicable in instances where finding reference texts is difficult. To estimate positions on policy dimensions, the authors have proposed to apply the scaling model to relevant sections of the manifesto. Identifying those sections can be done manually as the manifestos are often structured into policy-specific paragraphs. However, this only works for broad policy dimensions (Proksch and Slapin 2008). Alternatively, scholars can rely on keywords from legislative databases that clearly identify policy dimensions to tag relevant sentences in the manifesto (König et al. 2010). Independent of the approach chosen, the important point is that policy dimensions are defined a priori and on substantive grounds. As the approach is unsupervised and relies on the available data without requiring reference documents or a dictionary, the demands for data quality are quite high. If one wishes to estimate ideology from a set of documents, the principal purpose of the authors in writing the documents must have been to express a particular ideology. In particular, the Wordfish technique (and Wordscores, too, for that matter) best picks up ideological differences between documents when the data-generating process allows authors to emphasize or de-emphasize any issue they want. Researchers need to keep this in mind when applying the methods off-the-shelf. In particular, agenda effects may impact the estimation. Agenda effects may exist over time if politicians or parties use a

Words as Data: Content Analysis in Legislative Studies  

133

different vocabulary to express ideology in different time periods, something that seems natural over long time periods but may not be as problematic over shorter ones. For example, parties may change how they express their positions on environmental regulation as new environmental problems arise. Whereas clean air or water standards may be the dominant theme in one decade, issues concerning global climate change may prevail in another. Thus, changes in the policy agenda over a long time period may alter the use of language for this policy domain in a way that distinguishes all parties in one decade quite well from those in the previous decade. Such differences, however, could be simply a function of the changing topics within the policy domain, and not the result of an underlying change in party positions on environmental regulation. Long time-series analysis of political text therefore pose particular challenges that must be taken into consideration. Agenda effects can also occur when politicians must use the same vocabulary to talk about very specific issues. In such an instance, the agenda is pre-determined and parties or their members are not free to emphasize or de-emphasize issues as they can in a manifesto. For example, this may be problematic in parliamentary debates, where the topic of the debate is predetermined (and procedural language is more common than in manifestos). Finally, agenda effects may occur when researchers mix different types of documents that do not follow the same data-generating process. In this instance, the principal dimension extracted from the text data may simply reflect the different data generating processes rather than different ideologies. In short, techniques such as Wordfish or Wordscores have the potential to estimate positions for large collections of political text documents, but researchers should be aware of the assumptions behind these techniques in applied work.

6.3.2 Policy Agenda Certainly not all content analysis in legislatures is aimed at estimating party or legislator positions, or relies upon party manifestos as data. An emerging field of research examines the topics of the policy agenda as expressed in legislative speeches or other text data such as blogs and press releases. Often, these techniques make use of so-called topic models to identify latent themes in the text corpus (e.g. Quinn et al. 2010; Grimmer 2010). Quinn et  al., for example, explore the policy agenda of the US Senate. They uncover 42 latent topics and examine how attention to these topics vary over time, demonstrating that their measure tracks with the Senate’s voting agenda, but unveils more nuanced information about the Senate’s attention to various issues. Moreover, they present models to explain why Senators participate in debate on the latent topics they estimate. They consistently find Senators speak more often on topics when they serve on a relevant committee, however committee chairs do not tend to speak more than other committee members. In addition, ideological extremists tend to speak more, specifically on controversial topics such as judicial affairs, labour policy, taxes, and the budget. Topic models can therefore provide interesting insights into the operations of

134   Jonathan B. Slapin and Sven-Oliver Proksch legislatures and representation. The models are unsupervised, and like text scaling models, require the researcher to interpret the estimates ex post. Another requirement for topic models is that researchers need to set the number of latent topics. This may not always be a straightforward task, given that important topics may stay undetected if the number is too small, and be artificial if the number is set too high.

6.3.3  Beyond Policy and Ideology: Oversight, Representation, and the Media The use of content analysis in legislative scholarship has expanded beyond ideology and political attention to fields as varied as representation, oversight, and the media. Other work, for example, has examined bills to study oversight or the stability of laws. Huber and Shipan (2002) examine legislative oversight of the bureaucracy by looking at the length of bills, assuming that longer bills place greater limits on what bureaucrats can do. To achieve cross-national comparison, they standardize the page formatting of bills and account for linguistic differences by applying a verbosity multiplier capturing the efficiency of different languages. Their underlying theoretical notion is that greater policy conflict in the legislature may lead to more discretion for bureaucrats. Therefore, legislators write more detailed (i.e. longer) bills to prevent agency loss when delegating powers to the bureaucracy. Maltzman and Shipan (2008), in contrast, use page length of legislation as a measure of law-specific complexity. They use this measure to test the hypothesis that complex laws are more likely to be amended in the future. Thus, in these instances bill length serves as a proxy for both detail and complexity, which may or may not be related. One could have a very complex, but vague law covering many different areas, or a bill of similar length that is narrowly focused, but highly detailed. More research on the full text of legislation seems desirable, but it is ultimately very difficult. Bills are of a highly technical nature, their legal language does not compare well to position statements by politicians, and the significance of bills is likely independent of their length. Other recent work on parliamentary oversight has focused on how members of parliament scrutinize government behaviour by examining parliamentary questions (Proksch and Slapin 2011) and speech (Martin and Vanberg 2008b) rather than bill length. Proksch and Slapin find that members of the European Parliament from national opposition parties use their institutional position at the European level to gather information and scrutinize their home government’s behaviour. Martin and Vanberg focus on legislative speech to examine how parties in coalition governments make use of backbenchers to keep tabs on their coalition partners and highlight differences between the partners. Scholars have also turned to content analysis to examine the connection between MPs and their constituents. Martin (2011) uses parliamentary questions to measure constituency service, while Grimmer (2010) presents a topic model to understand how US senators use press releases to connect with voters. Martin, together with a team of research assistants, hand-coded approximately 124,000 questions asked in the Irish Dáil between

Words as Data: Content Analysis in Legislative Studies  

135

1997 and 2002 to determine whether they addressed local issues. He finds that about 44 percent of all questions asked in the Dáil have a local, constituency service focus. Grimmer (2010), in contrast, uses a Bayesian hierarchical topic model to examine how senators explain their work in Washington to their constituents. Grimmer’s modified topic model takes into account the hierarchical structure of press releases issued by representatives (thus, political statements constitute the lower level and authors the higher level). He applies this model to estimate the issues senators emphasize in communication with their constituents. On the basis of the estimates of the expressed agenda, he shows that committee leaders in the US Senate pay more attention to issues in their policy jurisdiction than the average senator and that the senators from the same state tend to have a more similar expressed agenda than senators from other states. Other work by Groseclose and Milyo (2005) examine US Congress speeches and media reports for references to think tanks. They use these think-tank citations as bridging observations to locate the media in the same ideological space as members of Congress to assess the presence of media bias. The underlying notion is that members of Congress receive higher utility when citing think tanks that hold a similar ideology to them, as do journalists. Their evidence suggests that there is a left-wing bias in the mainstream American press compared with the median member of Congress. Lastly, other content analytic work approaches legislative speeches from a deliberative politics perspective. Rather than estimate ideology or representation, the goal of this work is to examine the quality of debate as captured by levels of persuasion, justification, or respect, and relies on hand-coding of a number of selected speeches (Steiner et al. 2004; Bächtiger, this volume). The argument for this procedure is that higher quality debates lead to better democracy.

6.4  Issues in Content Analysis While advances in access to legislative documents and content analytic techniques have certainly pushed legislative studies forward, there are issues of research design to which anyone using content analysis must attend. First, and foremost, researchers typically employ content analysis to measure latent variables—variables that are not directly observed and can only be indirectly measured through indicators. Party ideology, for example, is a latent variable. The researcher can never directly observe a party’s ideology. Instead, the researcher may observe indicators, such as the content of speeches made by party leaders in parliament, the direction of votes cast by party MPs, answers by party supporters to issue questions in surveys, and so on. From these indicators, the researcher infers the ideological stance of the party. Even this is not so simple, as it involves decisions regarding the dimensionality of the policy space, which itself is latent (Benoit and Laver 2006). When examining the indicators of party ideology (typically extracted from documents via some form of content analysis) how much information should be maintained? Is it desirable to boil down all the information to a single, general dimension?

136   Jonathan B. Slapin and Sven-Oliver Proksch Or is it more desirable to estimate positions separately for economic, social,and foreign policy dimensions? Researchers must address these important questions before even starting to measure the positions of parties or MPs. Any attempt to measure such latent concepts must take into account the validity and reliability of the proposed measure. Validity refers to the degree to which the given measure captures the nature of the underlying latent concept, and is often thought of as the (lack of) bias. Because the concept we wish to measure is latent, there is no direct way to assess validity. Instead, researchers typically assess validity by comparing two or more different measures of the same latent concept, in the hope that they correlate, or even better they agree (see Mikhaylov, Benoit, and Laver 2012 for a discussion of the difference between correlation and agreement). Alternatively, researchers examine construct validity by asking whether the indicators they are using are theoretically related to the underlying concept they wish to measure. When choosing documents to content analyse, researchers must be concerned with construct validity. Were the documents written in such a way as to provide information about the underlying concept of interest? If one is interested in measuring ideology, for example, it probably makes more sense to content analyse MPs’ floor speeches, in which researchers can reasonably assume MPs express their positions with regard to legislation, than to analyse MPs’ Twitter statements, in which they may issue brief comments on current events, but not necessarily on current legislation. These questions of validity suggest that researchers must pay careful attention to the strategic environment that produced the texts they wish to analyse. They must carefully examine the data generating process behind the texts. If a researcher wishes to use legislative speeches to examine party ideology, he or she must carefully consider why it is that legislators give speeches, and what pressures they face when they do so. Are they attempting to communicate with their constituents, their fellow party members, or members of other parties? To what extent does their party leadership control what they say on the floor, or even control their access to the floor? The audience and degree of party pressure will determine the content of the speech, and the extent to which the speech captures the legislator’s own ideological position versus the position of some other group (e.g. the party or the constituency). Alternatively, speeches may not be about ideology at all, but may instead provide an opportunity to talk about constituency service and the government funds the legislator has brought back to the district or simply to criticize the government’s policy decisions. In this instance, analysing speeches may provide information about the importance of pork-barrel politics and about the difference between government and opposition MPs, but not about a particular ideological stance. In other words, it is impossible to extract information from a document that the document was not meant to convey. Proksch and Slapin (2012) argue, for example, that institutional constraints and electoral considerations greatly affect who may give a speech on the floor of parliament and the content of their message. Unless researchers account for the strategic nature of speech when performing content analysis, they will not have a valid measure of the latent concept they wish to capture. This is why party manifestos have proved so useful when trying to capture party ideology. They

Words as Data: Content Analysis in Legislative Studies  

137

are written with the aim to present the party’s primary ideological program in the midst of electoral competition. The data-generating process is clearly related to the latent variable researchers wish to measure. Relatedly, researchers must be concerned with issues of reliability. A reliable method is one that produces the same result each time the analysis is conducted, and is related to the statistical concept of precision (the inverse of the variance). A reliable measure may or may not be valid, but if it is not valid, it is not valid in the same way each time, regardless of who is doing the measuring. These issues, validity and reliability, get to the heart of the debate over whether human coding or machine coding is better when it comes to content analysis. Proponents of machine coding argue that their techniques are 100 percent reliable (even if they may not be perfectly valid). When using Wordscores or Wordfish (or indeed any computer algorithm) to estimate party positions from policy documents or manifestos, researchers making the same research design decisions (e.g. which documents to include in the analysis, whether to stem words to their roots or not, and whether to exclude stopwords, i.e. words that have no ideological content such as prepositions and conjunctions, from the analysis) will obtain the same results every time they conduct the analysis. This is not necessarily true for projects using human coding. Two human coders may read a document in a different way and categorize it differently. Perhaps, on average, they will get the same result as machine coding would, but there is likely to be more noise in the process. Mikhaylov, Benoit, and Laver (2012) find that there is a great deal of noise in the process of human-coding manifestos in the manner of the CMP, for example. This is not to say that machine coding is superior to human coding. Those who argue in favour of human coding often suggest that validity is higher with human coding. Computers may always produce the same answers, but those answers are not necessarily right. Humans are better able to draw on a wide variety of knowledge to “correctly” classify texts, whereas computers can only draw on the information human researchers give them. Of course, one means of assessing reliability among human coders is to examine intercoder reliability. If two or more people are confronted with the same coding task, to what extent do they arrive at the same answer? If multiple coders are unable to arrive at the same answer, perhaps the coding assignment is too difficult, or the instructions given to coders is not sufficiently clear. This has been a criticism of the CMP, for example. Typically, the CMP has only employed one coder per document, making it impossible to assess inter-coder reliability. Recent work has examined the extent to which multiple human coders are actually able to classify sentences on the basis of the CMP scheme (Mikhaylov, Benoit, and Laver 2012). The results were not encouraging. Common measures of inter-coder reliability (namely Fleiss’s Kappa) show that coders were not able to classify sentences in a reliable manner. In its latest installment, the manifesto project now uses two coders to analyse each manifesto, but it does retain the coding scheme. Lastly, when analysing a corpus of text, researchers must ensure that all texts within that corpus are comparable. For example, it would not make sense to analyse speeches, bills, and manifestos all at the same time as part of a single corpus. As discussed above, these texts were written for different purposes, have different intended audiences, use different language, and are subject to different constraints. In short, they were produced

138   Jonathan B. Slapin and Sven-Oliver Proksch using different data-generating processes and are, therefore, not directly comparable. While it may seem obvious that it does not make sense to compare speeches and manifestos at the same time, oftentimes comparisons are made that are less obviously problematic, but nonetheless raise important issues. For example, one may question whether it is appropriate to compare the content of manifestos across countries and time. The underlying dimensions of political competition are not the same everywhere, and also vary across elections—the most salient dimensions of partisan conflict in the legislature may differ in Germany and Japan; and the most salient dimensions of conflict in the 1970s may or may not be the most salient dimensions of conflict today. Likewise, while parliamentary questions may primarily serve as a means of oversight in one parliament, elsewhere they may serve primarily as a means to address constituency service issues. The differences may be due to institutions governing electoral incentives, party competition, or other considerations. Regardless, even if the type of document is nominally the same, if the underlying strategic environments that produce the documents are different, the content of the documents may not be directly comparable.

6.5  Ways Forward Recent trends in content analysis and the study of legislatures involve both applying new technological techniques and examining new sources of data to answer new substantive questions. Methodologically, there has been a rapid move toward importing new methods of machine learning from the fields of computational linguistics and computer science (e.g. Hopkins and King 2010; Quinn et al. 2010; Diermeier et al. 2012; Spirling 2012). Other recent advances involve taking advantage of the new legislative databases to examine forms of parliamentary behaviour that were largely ignored previously—e.g. parliamentary questions. The recent studies discussed earlier all use the newly available data to study questions regarding constituency service, representation, and much more in ways not previously possible. These studies have led to new insights regarding party competition, and have also allowed researchers to peer into the black box of intraparty dynamics to explore how backbenchers interact with party leaders (e.g. Martin and Vanberg 2008b; Proksch and Slapin 2012). To use these data in the most effective manner, researchers must pay careful attention to the (often unobserved) processes generating the text they use as data, as well as the assumptions of the methods they employ. Next, we outline a series of points that we feel content analytic studies of legislative politics must address.

6.5.1  Strategic Data-Generating Process Politicians are strategic communicators. This means that all content analysis needs to consider the data-generating process of the specific textual data being used. Not all

Words as Data: Content Analysis in Legislative Studies  

139

political text in legislatures is created equally. When delivering speeches or asking parliamentary questions, politicians may have a specific audience in mind. Draft legislation has its own technical features that makes it look very different from positional statements by parties and their MPs. Text as data approaches ask a lot from the data. For instance, when applying Wordscores or Wordfish the assumption that word counts are generated according to one-dimensional ideological preferences needs to be satisfied if this extracted dimension is interpreted as ideology. If the agenda is prestructured and actors are not free to emphasize or de-emphasize issues (e.g. on the aggregate in parliamentary speeches), this may be more problematic than in a context where actors are free to choose issues (e.g. during election campaigns or in parliamentary debates on specific topics such as the budget). Another implication is that it might not be advisable to apply the same estimation technique across different types of documents, in particular if they are generated differently. For instance, comparing positions based on speeches and manifestos may be particularly difficult because the audience and the context for the message is quite different. Finally, it is important for scholars to consider the institutional context. In some political systems, for instance, it is easier for backbenchers to participate in parliamentary debates than in others (Proksch and Slapin 2012). Thus, there is the possibility that even though the distribution of preferences is identical across political systems, the institutional hurdles in one system prevent the expression of preferences, but allow them in another. Failure to account for such differences is likely to lead to biased inferences.

6.5.2  Level of Analysis Some research questions involve measurement at a very detailed level. For instance, when the goal is to find out if MPs ask parliamentary questions regarding the oversight of legislation in force, then the measurement process must identify questions asked by individual members and perhaps hand code features of those questions. In other instances, the goal may be simply to find out the ideological dispersion of the party system in parliamentary debates or the topics on the legislative agenda. Here aggregating all speeches by members of the same party and applying unsupervised scaling or classification techniques may be sufficient. The measurement approach used needs to fit the question at hand. In some instances, a more labour-intensive and costly hand-coding approach, ideally involving multiple coders, may be more advisable, whereas in others computer-assisted techniques seem more promising when dealing with large collections of text. Thus, content analysis techniques should be chosen according to the level of analysis, which in turn should be motivated by the substantive research question of the project.

6.5.3  Latent Variables Ideology and the salient dimensions of partisan conflict are extremely difficult concepts to measure. These variables, and other related concepts (e.g. intra-party and party

140   Jonathan B. Slapin and Sven-Oliver Proksch system cohesion) are fundamentally unobservable, but they are indirectly expressed in various forms by members in parliament. MPs express their latent positions through political speech, voting in parliament, sponsoring legislation, or through statements in media and during election campaigns. This means that the appropriate way to approach the study of ideology is to treat it as a latent variable and apply a relevant measurement model. But developing appropriate measures of these latent concepts also means accounting for the strategic nature in the legislative environment that produces the data—e.g. what pressures are exerted by parties on members when speaking, voting, participating in committees, etc. The latent variable approach stands in contrast to other approaches, for example, the CMP, that assume there is a correct “gold standard” that coders can attain given sufficient training.

6.5.4 Uncertainty Both human-based and computer-assisted content analysis involves measurement uncertainty. Multiple human coders will never perfectly agree on how to code text. Existing hand-coded approaches typically provide intercoder reliability measures based on selected codings, but rarely incorporate uncertainty into the coding framework. Computer-assisted analysis needs to work with available text. Classical dictionary-based analysis makes it difficult to gauge measurement error, whereas more recent latent variable models offer ways to take uncertainty into account, either in a frequentist or Bayesian framework. Ultimately, uncertainty in political text analysis has multiple components. First, the text generation process itself is stochastic, meaning that MPs or parties can use different text to express the exact same position (Benoit et al. 2009). Second, there is measurement uncertainty associated with the chosen statistical model. Capturing measurement error is, of course, not an end in itself. Ideally, it should be carried over into secondary analysis that use quantities of interest extracted from text as independent or dependent variables (e.g. Benoit et al. 2009; Proksch and Slapin 2010).

6.5.5  Consider What Legislators Do Not Say All content analysis techniques rely on available political text to make inferential statements, but techniques differ in the way they include information about what is not said. A strategic model of political language may predict that certain parliamentary actors receive less floor time for electoral and institutional reasons (e.g. Proksch and Slapin 2012). Party leaders, for example, may wish to prevent dissidents from taking the floor of parliament to present views that run contrary to the party’s primary platform. When leaders systematically prevent certain party factions from giving speeches, parliamentary speeches may not offer an accurate picture of the true views (and diversity of views) within the party. If this is true, then models that fail to account for selection effects will produce results that are invalid or biased (with the direction and extent of bias not necessarily being clear a priori). Therefore, the analysis of missing values may allow

Words as Data: Content Analysis in Legislative Studies  

141

researchers to examine under what conditions actors choose to remain silent and what this means for inferences regarding political conflict in parliaments and governments.

6.6 Conclusion Content analysis is a powerful tool in studies of parties and legislatures. New advances in data and methods make it a particularly interesting and dynamic technique for examining latent variables such as ideology. More and more data are becoming available, and new statistical techniques and text as data approaches are being developed to analyse them. To effectively apply content analysis to the study of legislatures, though, researchers must carefully consider the assumptions of the estimation techniques they are applying, and the nature of the data they are using. These considerations include, for example, whether hand-coding or computer-coding is best, whether a supervised or unsupervised learning technique is most appropriate, and whether, given the data-generating process, extracting the desired information from the text is even possible. Further advances are most likely to occur at the intersection of a variety fields. Scholars of legislative politics who apply content analysis will likely greatly benefit from advances in the fields of computational linguistics, statistics, and computer science. In the absence of a general theory of political speech (or text), scholars must continue to adapt content analysis strategies to the problem at hand. New data mean that existing measurement models for text may be of limited use when applied in a context other than the model’s original intent. Furthermore, new types of text data are always becoming available. Comprehensive on-line archives of parliaments have led to better access to text data for content analysis. However, parliaments can change the structure of their databases, meaning the access to data and the way in which it is accessible may change without notice. There is a need for political scientists to capture the information provided by parliaments and store it in a more permanent and stable format through the construction of databases to ensure the replicability of studies using data from parliamentary archives. While text as data techniques hold great promise in the study of legislatures, careful application is required. The five points outlined in the previous section may serve as a guideline for researchers when applying quantitative content analysis techniques. Finally, valid comparisons across different types of documents may be a desirable research goal, but one that is difficult. We see the conceptual and empirical integration of these different data sources, for example, draft legislation and policy statements of parties and legislators, as one of the future challenges in the text-as-data literature.

References Adams, J., Clark, M., Ezrow, L., and Glasgow, G., 2006. Are Niche Parties Fundamentally Different From Mainstream Parties? The Causes and the Electoral Consequences of Western European Parties’ Policy Shifts, 1976-1998. American Journal of Political Science, 50: 513–29.

142   Jonathan B. Slapin and Sven-Oliver Proksch Adams, J., Ezrow L., and Somer-Topcu, Z., 2011. Is Anybody Listening? Evidence that Voters Do Not Respond to European Parties’ Policy Statements During Elections. American Journal of Political Science, 55: 370–82. Bawn, K. and Rosenbluth, F., 2006. Short versus Long Coalitions: Electoral Accountability and the Size of the Public Sector. American Journal of Political Science, 50: 251–65. Benoit, K. and Laver, M., 2006. Party Policy in Modern Democracies. New York: Routledge. Benoit, K. and Laver, M., 2007. Estimating Party Policy Positions: Comparing Expert Surveys and Hand-Coded Content Analysis. Electoral Studies, 26: 90–107. Benoit, K. and Laver, M., 2008. Compared to What? A Comment on “A Robust Transformation Procedure for Interpreting Political Text” by Martin and Vanberg. Political Analysis, 16: 101–11. Benoit, K., Laver, M., and Mikhaylov, S., 2009. Treating Words as Data with Error: Uncertainty in Text Statements of Policy Positions. American Journal of Political Science, 53: 495–513. Bernauer, J. and Bräuninger, T., 2009. Intra-Party Preference Heterogeneity and Faction Membership in the 15th German Bundestag: A Computational Text Analysis of Parliamentary Speeches. German Politics, 18: 385–402. Budge, I., Robertson, D., and Hearl, D. (eds.), 1987. Ideology, Strategy, and Party Change: Spatial Analyses of Post-War Election Programmes in 19 Democracies. Cambridge:  Cambridge University Press. Budge, I., Klingemann, H.-D., Volkens, A., Bara, J., and Tanenbaum, E. (eds.), 2001. Mapping Policy Preferences: Estimates for Parties, Electors, and Governments, 1945-1998. Oxford: Oxford University Press. Budge, I. and Pennings, P., 2007. Do They Work? Validating Computerised Word Frequency Estimates Against Policy Series. Electoral Studies, 26: 121–29. Debus, M., 2009. Pre-Electoral Commitments and Government Formation. Public Choice, 138: 45–64. Diermeier, D., Godbout, J.-F., Yu, B., and Kaufmann, S., 2012. Language and Ideology in Congress. British Journal of Political Science, 42: 31–55. Giannetti, D. and Laver, M., 2005. Policy Positions and Jobs in the Government. European Journal of Political Research, 44: 91–120. Grimmer, J., 2010. A Bayesian Hierarchical Topic Model for Political Texts:  Measuring Expressed Agendas in Senate Press Releases. Political Analysis 18: 1–35. Groseclose, T. and Milyo, J., 2005. A Measure of Media Bias. The Quarterly Journal of Economics, 120: 1191–237. Hakhverdian, A., 2009. Capturing Government Policy on the Left–Right Scale: Evidence from the United Kingdom, 1956–2006. Political Studies, 57: 720–45. Hopkins, D. J. and King, G., 2010. A Method of Automated Nonparametric Content Analysis for Social Science. American Journal of Political Science, 54: 229–47. Huber, J. D. and Shipan, C. R., 2002. Deliberate Discretion: The Institutional Foundations of Bureaucratic Autonomy. Cambridge: Cambridge University Press. Klemmensen, R., Hobolt, S. B., and Hansen, M. E., 2007. Estimating Policy Positions using Political Texts: An Evaluation of the Wordscores Approach. Electoral Studies, 26: 746–55. Klingemann, H.-D., Volkens, A., Bara, J., Budge, I., and McDonald, M. (ed.). 2006. Mapping Policy Preferences II: Estimates for Parties, Electors and Governments in Central and Eastern Europe, European Union and OECD, 1990-2003. Oxford: Oxford University Press. König, T., Luig, B, Proksch, S.-O., and Slapin, J., 2010. Measuring Policy Positions of Veto Players in Parliamentary Democracies. In T. König, G. Tsebelis, and M. Debus (eds.).Reform

Words as Data: Content Analysis in Legislative Studies  

143

Processes and Policy Change:  Veto Players and Decision-Making in Modern Democracies, pp. 69–95. New York: Springer. Krippendorff, K., 2004. Content Analysis:  An Introduction to Its Methodology. Thousand Oaks: Sage Publications. Laver, M, and Budge, I. (eds.), 1992. Party Policy and Government Coalitions. London: Macmillan. Laver, M. and Garry, J., 2000. Estimating Policy Positions from Political Texts. American Journal of Political Science, 44: 619–34. Laver, M., Benoit, K., and Garry, J., 2003. Extracting Policy Positions from Political Texts using Words as Data. American Political Science Review, 97: 311–32. Lipsmeyer, C. S. and Pierce, H. N., 2011. The Eyes that Bind: Junior Ministers as Oversight Mechanisms in Coalition Governments. Journal of Politics, 73: 1152–64. Lowe, W., 2008. Understanding Wordscores. Political Analysis, 16: 356–71. Lowe, W., Benoit, K., Mikhaylov, S., and Laver, M., 2011. Scaling Policy Positions from Coded Units of Political Texts. Legislative Studies Quarterly 36: 123–55. Maltzman, F. and Shipan, C. R., 2008. Change, Continuity, and the Evolution of the Law. American Journal of Political Science, 52: 252–67. Martin, L. W. and Stevenson, R. T., 2001. Government Formation in Parliamentary Democracies. American Journal of Political Science, 45: 33–50. Martin, L. W. and Vanberg, G., 2005. Coalition Policymaking and Legislative Review. American Political Science Review, 99: 93–106. Martin, L. W. and Vanberg, G., 2008a. A Robust Transformation Procedure for Interpreting Political Text. Political Analysis, 16: 93–100. Martin, L. W. and Vanberg, G., 2008b. Coalition Government and Political Communication. Political Research Quarterly, 61: 502–16. Martin, L. W. and Vanberg, G., 2011. Parliaments and Coalitions:  The Role of Legislative Institutions in Multiparty Governance. Oxford: Oxford University Press. Martin, S., 2011. Using Parliamentary Questions to Measure Constituency Focus: An Application to the Irish Case. Political Studies, 59: 472–88. McDonald, M. D. and Budge, I., 2005. Elections, Parties, Democracy: Conferring the Median Mandate. Oxford: Oxford University Press. Mikhaylov, S., Laver, M., and Benoit, K., 2012. Coder Reliability and Misclassification in the Human Coding of Party Manifestos. Political Analysis, 20: 78–91. Monroe, B. L. and Maeda, K., 2004. Talk’s Cheap: Text-based Estimation of Rhetorical Idealpoints. Paper prepared for the 21st Meeting of the Society for Political Methodology, 29–31 July. Neuendorf, K. A., 2002. The Content Analysis Guidebook. Thousand Oaks: Sage Publications. Proksch, S-O. and Slapin, J. B., 2006. Institutions and Coalition Formation:  The German Election of 2005. West European Politics, 29: 540–59. Proksch, S-O. and Slapin, J. B., 2010. Position taking in European Parliament Speeches. British Journal of Political Science, 40: 587–611. Proksch, S-O. and Slapin, J. B., 2011. Parliamentary Questions and Oversight in the European Union. European Journal of Political Research, 50: 53–79. Proksch, S-O. and Slapin, J. B., 2012. Institutional Foundations of Legislative Speech. American Journal of Political Science, 56: 520–37. Quinn, K., Monroe, B. L., Colaresi, M., Crespin, M. H., and Radev, D. R., 2010. How to Analyze Political Attention with Minimal Assumptions and Costs. American Journal of Political Science, 54: 209–28.

144   Jonathan B. Slapin and Sven-Oliver Proksch Schonhardt-Bailey, C., 2008. The Constitutional Debate on Partial-Birth Abortion: Constitutional Gravitas and Moral Passion. British Journal of Political Science, 38: 383–410. Slapin, J. B. and Proksch, S-O., 2008. A Scaling Model for Estimating Time-Series Party Positions from Texts. American Journal of Political Science, 52: 705–22. Spirling, A., 2012. U.S. Treaty Making with American Indians: Institutional Change and Relative Power, 1784–1911. American Journal of Political Science, 56: 84–97. Steiner, J., Bächtiger, A., Spörndli, M., and Steenbergen, M. R., 2004. Deliberative Politics in Action: Analysing Parliamentary Discourse. Cambridge: Cambridge University Press. Walgrave, S. and Nuytemans, M. 2009. Friction and Party Manifesto Change in 25 Countries, 1945–98. American Journal of Political Science, 53: 190–206. Warwick, P. V., 2011. Voters, Parties, and Declared Government Policy. Comparative Political Studies, 44: 1675–99.

C HA P T E R

7

D E B AT E A N D D E L I B E R AT I O N I N L E G I S L AT U R E S A N DR É BÄC H T IG E R

7.1 Introduction Despite the central role of speech and debate in parliament, parliamentary speech has not figured prominently on the scholarly agenda in contemporary political science. As Proksch and Slapin (2012, 520) note: “Participation in legislative debates is among the most visible activities of members of parliament (MPs), yet debates remain an understudied form of legislative behavior.” The conventional view holds that parliamentary debate does not have any impact on policy-making and is little more than the “public displays of the policy platforms of both government and opposition” (Brennan and Hamlin 1993, 447). Consequently, the purpose of parliamentary debate seems largely symbolic. In the last two decades, however, scholars have begun to refine as well as to challenge the conventional view. Not only is legislative speech more varied than commonly assumed, some scholars also argue that under specific conditions, legislative speech may be much more than the sterile argumentative confrontation of government and opposition and may even display features that have normative appeal. Three approaches can be distinguished: a strategic and partisan-rhetoric approach (anchored in rational choice theory), a deliberative approach, and a discourse approach. The strategic and partisan-rhetoric approach provides a systematization and extension of the conventional view. It starts from the assumption that legislative speech is cheap talk and its main purpose partisan and electoral. Yet it offers some intriguing insights in its variability, especially by adopting a comparative approach and by exploring the effects of different institutional arrangements and intraparty politics on the structuring of parliamentary debate. The deliberative approach, in turn, challenges the conventional understanding of legislative speech. It tries to identify institutional and issue-based conditions under which legislative speech can be deliberative, i.e., reasoned, respectful, informed, and oriented toward finding agreement. The discourse approach, finally, focuses on the

146   André Bächtiger constitutive features of parliamentary debates and offers an in-depth look at its norms and conventions. Recent variants of the discourse approach have also started to analyse specific framings of legislative speech, sometimes also in connection with normative expectations. This chapter gives a broad overview of the strategic/partisan-rhetoric, deliberative, and discourse approaches to parliamentary debate. It starts with the basic theoretical assumptions of the three approaches, followed by a presentation of some major empirical findings. Given deliberation’s prominence in contemporary political theory, its different take on legislative speech, and the big controversy about deliberation’s viability in the realm of legislatures, this chapter takes a special focus on the deliberative approach to legislative speech.1 At the end, the chapter will discuss a number of commonalities, remaining tensions, as well as some avenues for future research.

7.2  Strategic and Partisan-Rhetoric Approach The rational choice (or formal) approach to legislative speech provides a systematization and extension of the conventional view on legislative debate. Two versions can be distinguished. Austen-Smith (1990) presented a model where legislative debate is viewed as “cheap talk” and may influence decision-making through information revelation, i.e., by exchanging information that other legislators do not possess. Austen-Smith’s formal model posits that the role of debate is limited and one of timing. A debate stage in the policy-making process reveals all the information that otherwise would be revealed through agenda-setting. Or, as Austen-Smith (1990, 144)  succinctly puts it:  “debate does not elicit information that otherwise would not be made available during the decision-making process.” Generally, the revelation of private information may affect legislation, but only under some very restrictive conditions, namely when the distribution of legislators’ preferences over consequences is not too dissimilar. In contrast, when legislators’ preferences are dissimilar, they have an incentive to misrepresent information and manipulate the decision-making process in their favour. Since similar preferences of legislators are a very rare scenario in legislative politics, information revelation through debate will only very rarely affect the final outcome. As Proksch and Slapin (2012, 521) concur, if legislators mainly cared about the persuasive impact of legislative speech onother MPs, then “we would expect not to find much legislative debate at all.” Consequently, there must be other rationales for why MPs engage in legislative debate. A second view in the rational choice tradition starts from the assumption that legislators are concerned with electoral and partisan considerations when giving speeches in parliament. Legislative speech enables MPs to give a party or personal message to voters and partisan rank-and-file, as legislative debate receives coverage in all types of media. Legislative speech thus provides an opportunity for position-taking, advertising, and

Debate and Deliberation in Legislatures  

147

credit-claiming, which are key strategies to enhance the re-election chances of an MP (Mayhew 1974). Moreover, rational choice theorists have also started to take an in-depth look at intraparty politics (e.g., the relationship between parties and their voters) and institutional arrangements (e.g., different electoral systems), which are considered key factors for explaining variation in legislative speech. However, following Austen-Smith’s pessimistic expectations about the importance and influence of debate, a core assumption of the partisan-rhetoric view is that legislative speech can generally have no persuasive effect on policy-making. As we shall see later, this is in stark contrast with the deliberative approach, arguing that legislative speech may not be fully captured by a strategic model of position-taking where communication is exclusively tailored to secure or enhance electoral success. In recent years, rational choice scholars have made a dedicated effort to examine parliamentary debate empirically. One important area of research is to use legislative speech in order to analyse positions of members of parliament or political parties (e.g., Laver and Benoit 2002; Monroe and Maeda 2004; Proksch and Slapin 2010; Diermeier et al. 2012). This is mostly done through computer-based content analysis (such asWordfish), a topic that is dealt with in Proksch and Slapin’s chapter in this volume. This research, however, is not directly concerned with the functioning of legislative debate and therefore not addressed in this chapter. A further area of empirical research examines why legislators actually deliver speeches. In a pioneering study, Maltzman and Sigelman (1996) examined which factors influence the use of unconstrained floor time in the US Congress. Drawing on Fenno’s (1973) claim that MPs are motivated by both electoral and policy considerations, they focus on electoral and policy-based explanations for the propensity of delivering speeches. Empirically, Maltzman and Sigelman (1996) find that electorally based explanations do not perform well. For instance, more electorally vulnerable members used less unconstrained floor time than their more electorally secure colleagues. In contrast, policy-based explanations fare much better: party leadership, minority status, and ideological extremism were all conducive to longer floor speeches. While party leaders may have an interest in protecting and promoting the party’s brand, minority party members and political extremists may find it difficult to shape policy outcomes, thus have an incentive to speak up on the floor. Maltzman and Sigelman conclude that the failure of electoral explanations does not necessarily mean that MPs would not be concerned with them: “It is conceivable that electorally vulnerable members find that their time is better spent in other activities, such as fundraising, than in delivering speeches to the C-SPAN audience” (Maltzman and Sigelman 1996, 827–8). Building on the work of Maltzman and Sigelman (1996), Morris (2001) focused on the frequency with which MPs made partisan-oriented one minute speeches in the 104th Congress. While confirming some of Maltzman and Sigelman’s findings, he also found that partisanship had a strong impact on debating behaviour. For instance, MPs with less electoral security used more partisan rhetoric. In this regard, Quinn et al. (2010) also demonstrate that in election times, politicians put more emphasis on symbolic and social issues in speeches.

148   André Bächtiger Recent studies in the rational choice tradition take an in-depth look at intraparty politics and institutional arrangements that shape the structure of parliamentary debate. Martin and Vanberg (2008) argue that coalition parties in parliamentary democracies must communicate to their voters that they have not violated their electoral promises when they agree to policy compromises with their coalition partners. One pathway to accomplish this goal is through legislative debate. Martin and Vanberg hypothesize that coalition parties will communicate more extensively on issues that divide them from their coalition partners. This combines with an electoral cycle argument:  the closer elections are, the more parties will engage in this activity. Using the length of legislative speeches as a proxy for the extent to which parties communicate with their constituents, they demonstrate that coalition parties indeed communicate more extensively on issues that divide them from their coalition partners and do so increasingly when elections approach. Even though the mere length of speaking does not give direct insight into the content of speeches, the study still provides circumstantial evidence that coalition parties use legislative debate to convince constituents “that they have not given away the store in exchange for the spoils of office” (Martin and Vanberg 2008, 513–14). Proksch and Slapin (2012) present a comparative institutional approach to legislative speech that is based on a strategic model of intraparty politics. They start from the assumption that in order to uphold the party’s brand, party leaders must prevent their MPs from delivering speeches that contradict the party’s central message. How much control is exerted by party leaders, however, depends on institutional factors, namely the configuration of the electoral system. In this regard, Proksch and Slapin focus on four scenarios. In scenario one, a presidential system with a majoritarian electoral system, the electoral independence between legislature and executive means that party unity is not very important and party leaders will not exert much control when MPs take the floor. In scenario two, a parliamentary system with a majoritarian electoral system, party leaders will generally place more weight on party unity than party leaders in a presidential system. Nevertheless, in a first-past-the-post electoral system thatemphasizes the importance of individual candidates, party leaders place less weight on party unity and exert less control when MPs take the floor. In scenario 3, a parliamentary system with closed-list proportional representation, party leaders will place high value on party unity, as voters rely heavily on party “labels” to make their electoral choices. Consequently, party leaders will control the public exposure of MPs. In scenario 4, a parliamentary system with mixed-member proportional representation, elements of closed-list PR are combined with single-member districts. This means that while party leaders put a high value on party unity, MPs have different incentives depending on whether they are elected off the party list or from single-member districts. While MPs elected from the party list will value the party’s position highly, this is less true for MPs elected in a single-member district. Consequently, party leaders will try to keep such MPs off the floor, since the latter will have an incentive to present their own positions to voters which may deviate from the party’s central message. Proksch and Slapin test the implications of their theoretical model with data on the amount of legislative speeches made by party leaders and backbenchers in the UK (an example

Debate and Deliberation in Legislatures  

149

of parliamentary democracy with a majoritarian electoral system) and Germany (an example of a mixed-member proportional representation system). The findings corroborate the theoretical expectations: party leaders in Germany are more active on the floor than backbenchers. By contrast, in the UK, the majoritarian electoral system provides leaders with fewer reasons to keep backbenchers off the floor. Moreover, German MPs who are ideologically distant from their party leaders are less likely to take the floor; in the UK, ideologically distant MPs are more likely to take the floor than MPs who toe the party line. Finally, the German data show that MPs elected in single-member districts tend to make fewer speeches than MPs elected from the party list, but this effect weakens when the ideological distance between backbenchers and leaders is taken into account. The study not only suggests that legislative debates are strongly affected by partisan strategizing within specific electoral contexts, it also shows that party cohesion is inversely related to party positions communicated in legislative speeches:  “Speeches may [...] underestimate the ideological differences within parties in many parliamentary systems” (Proksch and Slapin 2012, 535). This has important implications for the use of legislative speech as a resource to estimate party preferences and party cohesion. These two examples nicely illustrate how rational choice analysis can deepen our understanding of legislative speech, even though the basic assumptions are fully in line with the conventional and partisan-rhetoric understanding of parliamentary debate.

7.3  Deliberative Approach The deliberative approach to parliamentary debate brings a normative dimension back to legislative politics. In its classic version, deliberation means that actors justify their positions with a focus on the common good, weigh alternative arguments and positions with respect, are willing to yield to the force of the better argument, and try to find a reasoned consensus on validity claims (Habermas 1983, 1996; Chambers 1996; Gutmann and Thompson 1996). The goals of deliberation are geared to both epistemic fruitfulness and consensus: deliberation should generate decisions that are better reasoned and informed, more public-orientated and consensual, and consequently more legitimate and effective. In the past decade, deliberation has become one of the “most active areas of political theory” (Dryzek 2007, 237). The deliberative approach has not only given a new language to analyse political talk, it has also prompted a vigorous debate onhow politics should ideally function. In this regard, deliberative theory has taken issue with the standard aggregative models in legislative studies (and political science in general) where outcomes are determined by numbers, i.e., who has the most votes, and not by reasons, i.e., who has the best arguments. Prominent deliberative theorists have repeatedly argued that parliaments are an important sphere of deliberation because they serve essential legitimizing and social integrative functions (e.g. Habermas 1996). Habermas (2005, 389) even states that legislative deliberation “reaches just to the centre of the whole approach to deliberative

150   André Bächtiger politics.” In his book Mild voice of reason, Joseph Bessette (1994) has also provided the first deliberative reading of policy-making in the US Congress. Here, Bessette contrasts political (or electoral) and deliberative explanations for key aspects of legislative politics. For instance, while a political explanation views the function of floor debate only pro forma to enhance the standing with constituents, a deliberative explanation views floor debates as the “final opportunity to hear the strongest arguments pro and con” (Bessette 1994, 152), while simultaneously serving as an information source for the contents of complex bills. Or, while a political explanation views committee hearings as publicizing devices “to mobilize support outside of Congress,” a deliberative explanation views hearings as fora “to elicit the information and arguments necessary to make informed judgments” (Bessette 1994, 152). Yet, many political scientists—and even some deliberative scholars (e.g., Fishkin and Luskin 2005)—strongly question whether parliaments can entail genuine deliberation. The key criticism is that parliamentary debate is oriented toward voting, not toward collecting and aggregating information (see Ferrié 2008; Rasch 2011). Drawing from the partisan-rhetoric approach, critics hold that arguments in legislative speech are not directed toward persuading other MPs but toward mobilizing an outside audience— voters, citizens, as well as partisan rank-and-file. Deliberation in parliament is also constrained by the fact that modern legislatures operate under severe time constraints and have strictly regulated access to delivering speeches (Rasch 2011). Time constraints and access regulations give parliamentary debates a ritualized and rigid character. Speeches are not spontaneous, but frequently prepared in advance. Rasch (2011, 20) summarizes the critical view on parliamentary deliberation as follows: Arguing seldom affects information and preferences in a way that become important at the final voting stage. Outcomes almost always are known in advance. Plenary debates lack the dynamic elements that are central to any deliberative process marked by conflicting preferences.

In other words, a deliberative lens on parliamentary debate, assuming that parliamentary debate can change the opinions of MPs seems aspirational and critical rather than descriptive (see Esterling 2011, 191). Current deliberative scholarship acknowledges that electoral, partisan, and representative pressures loom large in parliaments, constraining full-fledged deliberation. They nonetheless argue that the conventional and partisan-rhetoric approach may be too limiting. First, much of the criticism against the possibility of deliberation in parliaments stems from an analysis of parliamentary systems as well as of contemporary American politics. As we shall see later, critics are certainly not misled with regard to deliberative failures in parliamentary systems as well as in the contemporary US Congress; but critics tend to ignore that a different institutional organization of legislatures—in combination with issue type and partisan strategies—may be conducive to a much higher potential for deliberative action. In this regard, deliberative scholars also claim that the role of institutions and contexts in shaping political behaviour is more

Debate and Deliberation in Legislatures  

151

profound than rational choice scholars tend to assume. They think that institutions and other contexts do not only affect the strategic incentives of legislators, but may also create spaces for different action logics (including deliberation). In this view, strategic action is not the only action logic in politics and the importance of strategic action may vary according to different contexts. The claim is that when stakes are high, actors will follow a course of action that is highly compatible with the predictions of rational choice theory. But if stakes are lower, actors have more space to behave differently: they may remain strategic but they may also become genuine deliberative actors (see Bächtiger and Hangartner 2010). Second, while deliberative scholars concede that fundamental opinion changes in core values are unlikely in parliaments, even under the best circumstances, deliberation can still contribute to the theory-component of legislator’s preferences (Vanberg and Buchanan 1989) by allowing legislators to update their preferences or learn about unforeseen policy consequences. Concrete policies are complex entities, comprising smaller elements on which participating politicians can learn and change their minds. Third, critics also (partially) misrepresent the link between deliberation and voting. Whether persuasion actually occurs is not always a major consideration in deliberative theory, since this literature accords normative importance to the quality of argument and discussion itself (see Esterling 2011, 192). Thus, prepared speeches may be deliberative, i.e., they may display high justification rationality and respect, or even document opinion change, even though they are not fully interactive. Fourth, and perhaps most importantly, we should not pre-judge the possibility of parliamentary deliberation: claims for and against the possibility of deliberation in legislatures need to be proven or disproven empirically. But how much deliberative action is there really in legislatures, and how does it work? For this purpose, let us focus on three different legislatures, namely those of Switzerland, the United States, and Germany (see Steiner et al. 2004; Bächtiger 2005). Specifically, the Swiss parliament serves as an example of a non-parliamentary consensus system, the United States’ Congress as an example of a competitive presidential system, and the German parliament as an example of a competitive parliamentary system. The analysis presented here is a reanalysis of the study by Bächtiger (2005). It considers a total of 52 recorded debates in the three countries, stemming mostly from the late 1980s and the 1990s. These debates cover topics with high partisan polarization such as economic policies and abortion, as well as less polarized topics such as rights of the disabled, animal welfare, and crime prevention. The debates comprised nearly 4,500 speeches, which were coded for their deliberative quality. Notice that in the Swiss and German legislatures, recorded debates in non-public committees could be used for a quantitative analysis of deliberate quality. The quality of deliberation is measured by the Discourse Quality Index (DQI; Steenbergen et al. 2003). Rooted in a Habermasian understanding of deliberation, the DQI employs a number of philosophically derived indicators of deliberative quality. The focus in this chapter is on four key indicators: (1) justification rationality (do speakers give elaborated reasons for their positions or do they forward demands and positions with no or only simple reasons?); (2) common good orientation (do speakers cast

152   André Bächtiger their justifications in terms of conceptions of the common good or in terms of narrow group or constituency interests?); (3) respect toward demands and counterarguments (do speakers degrade, treat neutrally, value, or agree with demands and arguments from other speakers?); and (4) constructive politics (do speakers sit on their positions or submit alternative or mediating proposals?). These four indicators capture two essential concepts underlying deliberative theory: the concept of rational argument—measured by justification rationality and common good orientation—and the concept of weighing positions and reasons with a “favourable attitude toward, and constructive interaction with, the persons with whom one disagrees” (Gutmann and Thompson 1990, 85)—measured by respect and constructive politics. The DQI has proven to be a reliable measurement instrument, i.e., there is generally broad agreement where a particular speech act falls on the fourindicators. For the debates considered in this article, two independent coders scored a subset of the speech acts. The rate of inter-coder reliability ranges from a low of .919 for respect toward counterarguments to a high of 1 for content of justification. Cohen’s kappa, which controls for inter-coder agreements by chance, ranges from .881 for respect toward counterarguments to .954 for constructive politics. These figures indicate excellent inter-coder reliability.2 The DQI has also met with considerable support from deliberative theorists (Habermas 2005; Thompson 2008). For example, Habermas (2005, 389) writes that the DQI captures “essential features of proper deliberation.” Thus, the DQI appears to have good construct validity. Let us first focus on the extentof deliberative action in the three legislatures. Overall, the amount of high-quality deliberation is fairly limited, which seems to corroborate pessimistic expectations:  sophisticated justifications amount to 39 percent, common good appeals to 15 percent, explicit respect toward (and agreement with) demands or counterarguments to 12 percent, and mediating proposals to 9 percent. However, there is considerable variability of deliberative quality across the three legislatures as well as across different arenas and issues within these legislatures. Let us now explore what drives this variability. The data constitute a multilevel structure with nearly 4,488 speeches nested in 52 debates, which in turn are nested in seven legislative institutions (the German Bundestag, Bundesrat, and Vermittlungsausschuss; the Swiss Nationalrat and Ständerat; and the US House of Representatives and Senate). Because of this data structure, multilevel analysis is used. The analysis includes fixed effects for the different predictors and variance components at the levels of legislature, debate and, where appropriate, speeches (i.e., a random intercept model was estimated). System type and chamber vary across legislatures, whereas the remaining predictors vary across debates. There are no predictors at the level of speeches. Table 7.1 displays the results of the comparison of deliberative quality among the three legislatures. We see that respect is the most sensitive deliberative indicator, while other indicators, especially constructive politics, showed remarkable resilience to institutional and issue variation. A robust finding is that the Swiss grand coalition setting enhances respectful behaviour of MPs. On a 7-point respect scale, the difference of the Swiss parliament to the US Congress and the German parliament is 0.4. The institutional argument is that coalition arrangements open up spaces for less politicized interactions,

Debate and Deliberation in Legislatures  

153

Table 7.1  Predicting the quality of deliberation in three legislatures (Switzerland, United States, and Germany)

Predictor Fixed Effects: Consensus Democracy

Sophisticated Justification

Common Good Appeals

Respect Demands and CounterArguments

Constructive Politics

.45 (.86) .15 (.78) -2.33** (.28) .52** (.19) 1.37** (.37) .13 (.23) -.77 (.71)

-.23 (.20) .24 (.15) .86** (.17) -.03 (.15) -.31 (.22) -.20 (.18) .63** (.21)

.40** (.07) .36** (.07) .31** (.06) .43** (.07) .09 (.09) .03 (.09) 1.74** (.10)

-.27 (.18) -.22 (.17) -.11 (.08) -.09 (.08) .16 (.13) -.01 (.10) .79** (.18)

.97+ (.55) .23** (.07)

.00 (.00) .11** (.04)

.00 (.00) .03** (.01) 1.36** (.03)

.04 (.03) .03* (.01)

Number of Speeches

4464

4464

4464

4464

Method

Logit

Gompit

Linear

Gompit

Second Chamber Non-public Arena Low Issue Polarization Low Party Discipline Strong Veto Power Constant Variance Components: Legislature Level Debate Level Speaker Level

Notes: Table entries are maximum likelihood multilevel model estimates with estimated standard errors in parentheses. The linear models were estimated using an iterative generalized least squares (IGLS) algorithm. The Logit and Gompit models were estimated using secondorder penalized quasi likelihood (PQL) estimation. Gompit models take into account extremely skewed distributions; note that in Gompit models, a score of 0 is predicted. All estimates were obtained using MLwiN version 2.01. ** p