Segmented Representation: Political Party Strategies in Unequal Democracies (Oxford Studies in Democratization) [Illustrated] 9780199642649, 0199642648

Citation preview

OX F O R D S T U D I E S I N D E M O C R AT I Z AT IO N Series editor: Laurence Whitehead

Oxford Studies in Democratization is a series for scholars and students of comparative politics and related disciplines. Volumes will concentrate on the comparative study of the democratization processes that accompanied the decline and termination of the Cold War. The geographical focus of the series will primarily be Latin America, the Caribbean, Southern and Eastern Europe, and relevant experiences in Africa and Asia.

OTHER BOOKS IN THE SERIES Democracy, Agency, and the State: Theory with Comparative Intent Guillermo O’Donnell Regime-Building: Democratization and International Administration Oisín Tansey Rethinking Arab Democratization: Elections without Democracy Larbi Sadiki Accountability Politics: Power and Voice in Rural Mexico Jonathan A. Fox Regimes and Democracy in Latin America: Theories and Methods Edited by Gerardo L. Munck Democracy and Diversity: Political Engineering in the Asia-Pacific Benjamin Reilly Democratic Accountability in Latin America: Edited by Scott Mainwaring and Christopher Weina Democratization: Theory and Experience Laurence Whitehead The Politics of Uncertainty: Sustaining and Subverting Electoral Authoritarianism Andreas Schedler The Politics of Accountability in Southeast Asia: The Dominance of Moral Ideologies Garry Rodan, Caroline Hughes

‘Juan Pablo Luna’s study offers three important innovations in the study of citizen-politician linkage relations that should be heeded by future scholarship. First, parties’ linkage strategies often involve a complex combination of programmatic appeals and clientelistic inducements, cultivation of party identification and party leadership personalities, no straight-forward zero-sum trade-offs. Parties therefore have to cope with the challenge to coordinate politicians and “harmonize” constituencies that are attracted by diverging linkage strategies. Second, the differentiation of linkage strategies is often made possible and sustained by sub-national differentiation of party strategies. Third, the political economy of inequality is critical for citizen-party linkage strategies. Under conditions of high economic inequality, where conventional theory would predict the demise of democracy, politicians’ strategic deployment of clientelism may in fact sustain social pacification and preserve democracy.’ Herbert Kitschelt, Duke University ‘Segmented Representation is one of the most important books on Latin American parties to be published in decades. Drawing on painstaking research in Chile and Uruguay, Luna shows that the kinds of programmatic appeal made by parties in advanced industrialized (and often taken for granted in theories of party behavior) are less viable in highly u ­ nequal societies, and that successful parties must make segmented appeals to diverse—even seemingly opposed—constituencies. The concept of segmented linkages will diffuse widely among scholars. Luna’s work helps us understand why parties operate differently in a context of extreme inequality, while at the same time showing how segmented linkages work to reinforce existing inequalities. The book will make a major contribution to the study of political parties, not only in Latin America but throughout the developing world.’ Steven Levitsky, Professor of Government, Harvard University ‘High levels of income inequality pose a challenge to politicians. They need to keep the rich happy to keep investment and donations flowing and the poor happy to secure their votes. How do politicians and parties reconcile the demands of these different constituencies in the most unequal part of the world? In this pathbreaking study, Luna suggests that they simultaneously deliver policies to the rich and pork to the poor. He supports his argument with rich empirical work drawn from the experiences of two new Latin American democracies, Chile and Uruguay, as well as other cases from the region and India. This book makes a formidable contribution to our understanding of the mechanisms underpinning electoral representation in unequal democracies. It will be a must-read for students of representation, electoral behavior and political parties in established as well as new democracies, and will also prove invaluable to practitioners of approaches to empirical research that draw on mixed-methods.’ M. Victoria Murillo, Professor of Political Science and International Affairs, Columbia University ‘This is a path-breaking study of democratic representation in highly unequal societies. Juan Pablo Luna convincingly demonstrates that segmented societies create both incentives and opportunities for political parties’—on both the left and the right—to use different types of electoral appeals to attract distinct groups of voters. Drawing from survey data and field research in Chile and Uruguay, Luna shows how parties craft electoral strategies that combine programmatic, clientelistic, and personalistic linkages to voters, allowing them to generate support among heterogeneous but socially segmented constituencies. The result is a masterful analysis of how inequality structures democratic representation—and, ­ultimately, how democracy can reproduce social inequality.’ Kenneth M. Roberts, Cornell University

Segmented Representation Political Party Strategies in Unequal Democracies

J UA N PA B L O   L U N A

3

3 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom 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 © Juan Pablo Luna 2014 The moral rights of the author‌have 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: 2014934287 ISBN 978–0–19–964264–9 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.

Contents List of Illustrations List of Tables List of Abbreviations Prologue

xi xv xvii xix

Introduction Segmented Democratic Representation in Unequal Societies Why Segmented Representation Matters for Redistribution and for Party System Change Research Design Case Selection Multi-level Research Design: Paired District Comparison Tools for Observing and Analyzing Different Types of Linkages Plan of the Book

1 3 8 12 12 14 15 15

Part I: Descriptive Inference 1. Segmented Electoral Appeals: A Descriptive Framework Towards an Adapted Framework Types of Party–Voter Linkages: A Minimal Definition Party–Voter Linkages and the Assumption of Trade-Offs The Structure of Segmentation Party–Voter Linkage Types and Strategic Harmonization Segmentation and Harmonization of Electoral Appeals Operationalization and Identification Strategy: Mapping Segmented Linkages Level of Analysis: System or Individual Parties? Unit of Observation and Units of Analysis Recovering Historical Patterns of Segmentation and Harmonization (Independence to 1973) Observing Socioeconomic Segmentation at the Aggregate Level in the Contemporary Period Observing Territorial and Socioeconomic Segmentation at the District Level Observing Party Strategies

21 22 23 29 32 34 41 45 45 47 48 48 50 52

vi

Contents Observing Structures of Segmentation in Other Cases Summary of the Operationalization Strategy for Measuring the Dependent Variable Causes of Party–Voter Linkage Types: Available Arguments Causes of Different Types of Linkages and Structures of Segmentation Which Types of Linkages and What Degree of Segmentation Are Observed in the System? Is Segmentation Mixed or Dual? Can Parties Strategically Harmonize Segmented Appeals? Expected types of Causation and Hypothesized Causal Configurations Across Cases

2. Patterns of Party–Voter Linkages in Chile and Uruguay: A Stylized Description of the Pre-1973 and Post-Transitional Periods Introduction The Chilean and Uruguayan Pre-Authoritarian Party Systems A General Characterization of the Chilean Party System A General Characterization of the Uruguayan Party System Applying the Analytical Framework to the Pre-Authoritarian Party Systems The Historical Configuration of Party–Voter Linkages in Chile and Uruguay Party–Voter Linkages in Post-Authoritarian Chile and Uruguay, 1990–2010 Post-Transitional Party–Voter Linkages in Chile Post-Transitional Party–Voter Linkages in Uruguay Applying the Analytical Framework to Post-Transitional Party Systems Recent Evolution of the Two-Party Systems Chile’s Bureaucratic-Authoritarian Regime (1973–1989) Uruguay: The Bureaucratic-Authoritarian (BA) Regime and Transition to Democracy (1973–1984) Post-Transitional Party-Systems and Programmatic Party–Voter Linkages Chile: 1989–2010 Uruguay: 1984–2010 Summing Up: The Contemporary Configuration of Party-Voter Linkages in Chile and Uruguay

52 53 53 59 60 62 63 66 70 70 71 71 72 75 83 83 85 86 88 94 94 103 106 106 109 115

Contents

vii

3. Socioeconomic Segmentation of Party–Voter Linkages in Post-Transitional Chile and Uruguay 117 Introduction 117 Methodological Strategy: Programmatic Linkage Proxies 117 The Evolution and Socioeconomic Segmentation of Programmatic Linkages in Chile (1989–2010) and Uruguay (1984–2010) 120 Significant Partisan Differences Across Programmatic Divides 121 Relative Programmatic Placements of Partisan Electorates and Congressional Delegations 122 Procrustes Analysis: Multidimensional Issue Congruence 129 Cross-Validation: Interview and Expert Survey Evidence 133 Alternative Types of Linkages and their Patterns of Socioeconomic Segmentation 139 Conclusion 141 4. Territorial Segmentation of Party–Voter Linkages in Post-Transitional Chile and Uruguay 144 Introduction and Methodological Strategy 144 Comparative Notes on Party–Voter Linkages at the Local Level in Chile and Uruguay 148 Chile: Overall Dynamics of Party–Voter Linkages at the District Level 150 Linkage Segmentation in Chilean Districts: A Classification of District Types 153 Upper-Income Sectors: “Rightist Ghettos” 154 Lower-Income Sectors Without Available Territories: “Leftist Ghettos” and Pinochet’s Successful Mayors in Congress 156 Middle and Lower-Income Districts with Available Territories: Heterogeneous Societies and Intra-District Segmentation 162 Uruguay: Overall Dynamics of Party–Voter Linkages at the District Level 171 Ideological Families and the Recrafting of Local Party Systems: A Tentative Characterization of District Types in Uruguay 171 Towards National–Local Dealignment? The Characteristics of Recent Political Leaderships at the District Level 173 The Evolution of Linkage Types at the District Level 176 The Receding of Clientelism 177

viii

Contents

The Mutation and (Municipal) Refuge of Clientelism and Constituency Service The Current Nature of Non-Programmatic Linkages in Uruguay: Overarching Implications 5. Strategic Harmonization of Segmented Linkages: The UDI and FA in Comparative Perspective Introduction Segmentation and Harmonization Strategies of the UDI and FA Setting Comparative Parameters: Less Successful Parties and Strategic Harmonization The UDI in Chile Origins and Social Base The Core Constituency The UDI’s Outreach to the Popular Sectors: Authoritarian Clientelism  The Political Opportunity Structure for Seducing the “Soft Vote” Political Organization and the UDI’s Partisan Apparatus The “Popular Party” and Lavin’s National Leadership Corollary: The UDI’s Harmonized Dual Strategy The FA in Uruguay Origins and Social Base The FA’s Core Constituency The Political Opportunity Structure of the 1990s and 2000s Reaching the FA’s Peripheral Constituency Moderation, Leadership Renewal, and Electoral Alliances  The Municipal Government of Montevideo Reaching the FA’s Emerging Constituency: The AU and the MPP The FA’s Harmonized Mixed Strategy The UDI and FA in Comparative Perspective

182 186 191 191 192 195 202 204 210 213 214 215 220 223 224 225 231 239 242 242 245 247 252 254

Part II: Causal Inference 6. Causal Induction: Explaining Linkage Structures in Chile and Uruguay The Variance to Be Explained

259 259

Contents The Nature of the Causal Argument: Types of Causation and Coding of Each Case Explaining the Scope of Socioeconomic Segmentation and Predominant Types of Party–Voter Linkages Social Inequality Privatization, Social Policy Provision, and Decentralization Timing of Market and State Reforms Authoritarian Legacies and the Politics of the Transition to Democracy Institutional Factors Explaining Patterns of Territorial Segmentation Explaining Types of Strategic Harmonization District Magnitude Separation of Municipal and National-Level Elections Nomination Procedures Campaign Finance Regulations Privatization and State Reform Individual Parties’ Access to Material and Symbolic Resources Conclusion 7. Plausibility and Scope: Out-of-Sample Tests Operationalization: Identifying Non-Segmented, Segmented and Harmonized, and Segmented and Non-Harmonized Linkage Strategies Strategy for Operationalizing Other Variables Assessing the External Validity of Descriptive Propositions and an Overarching Causal Claim External Validity of Explanatory Propositions Scope of Segmentation, Predominant Linkage Types Strategic Harmonization Individually Successful Parties in Highly Unequal Societies The Bharatiya Janata Party (BJP) The Alianza Republicana Nacional (ARENA) The Partido dos Trabalhadores (PT) The Movimiento al Socialismo (MAS) Comparative Analysis of the Shadow Cases’ Causal Configurations Conclusion

ix 263 264 264 267 270 271 271 272 274 276 276 277 277 278 278 279 281 282 288 289 291 291 300 303 303 306 310 315 319 323

x

Contents

Conclusion Methodological Implications Substantive Implications Linkage Strategies’ Redistributive Impacts Inter-Temporal Party and Party-System Dynamics, and States’ Bureaucratic Capacity Theoretical Implications

325 327 328 328

Bibliography Name Index Subject Index

341 365 371

336 339

List of Illustrations

L I S T O F F IG U R E S 1.1 1.2 1.3 1.4 1.5 2.1 2.2 2.3 2.4 5.1 5.2 7.1

Pure strategy 42 Harmonized dual strategy 43 Non-harmonized dual strategy 43 Harmonized mixed strategy 44 Non-harmonized mixed strategy 44 Territorial linkage segmentation in Chile (pre-1964 and 1964–1973) 78 Territorial linkage segmentation in Uruguay (until 1973) 82 Territorial linkage segmentation in Chile (post-1989) 90 Territorial linkage segmentation in Uruguay (post-1984) 93 The UDI’s linkage strategy 194 The FA’s linkage strategy 195 Set-representation of causal configurations explaining the presence of NS-OP linkage structures 294

LIST OF EX AMPLES 1.1 Party-based appeals 38 1.2 Candidate-based appeals 38 1.3 Candidate-based appeals organized by the party (with potential harmonization) 39

xii

List of Illustration LIST OF GRAPHS

2.1 Electoral participation rates in Chile and Uruguay (post-transitional period) 108 2.2 Electoral evolution of traditional and challenger parties in Uruguay (1942–2004) 110 3.1 Summary of procrustes results obtained for different electoral cross-sections and model specifications (two/three dimensions, oblique/orthogonal rotations) 131 3.2 Expert survey estimates of the use of different mobilization strategies in Chile and Uruguay 138 3.3 Socioeconomic segmentation of party identification in Chile and Uruguay 140 5.1 The Concertación’s social bases (lower-chamber elections) 207 5.2 Alianza’s (and UDI’s) social bases (lower-chamber elections) 208 5.3 Electoral evolution of the UDI across social strata (lower-chamber elections, 1997 = 100) 209 5.4 Votes for the UDI in presidential elections (first round: 1999 and 2005) 209 5.5 Dissident roll-call votes from within the Alianza (2006–2008) 211 5.6 Average self-reported campaign expenditures by parties’ Lower-chamber candidates (2005) 212 5.7 Average self-reported campaign expenditures by parties’ municipal candidates (2008) 212 5.8 Yearly average popularity ratings for selected presidential leaders (2002–2005) 223 5.9 Electoral growth of the FA across socioeconomic and geographic segments 228 5.10 Electoral performance of four FA fractions across social segments (1999) 230 5.11 Partisan affiliation of PIT-CNT’s directorate (2003 congress) 232 5.12 Proportion of union leaders in top positions on respective fractions’ senate and deputy electoral lists 233 5.13 Total union membership and public vs. private origin of union members (1987–2003) 234



List of Illustration

xiii

5.14 Parties’ campaign expenditures: 1994 and 1999–2000 electoral cycles (estimated amounts expressed in 1995 US millions of dollars) 236 5.15 Parties’ TV advertising expenditures: 2004 electoral Cycle (amounts reported in US dollars) 236 5.16 Percentage of militant activists among political party sympathizers 237 5.17 Number of party activists officially registered by the Servicio Electoral circa 2011 (Chile) 239 5.18 Occupational background of shantytown dwellers 250 6.1 Income ventile distributions 265 6.2 Pension benefit structure by income decile in Chile (2003) and Uruguay (2005) 267 7.1 Linkage “mixes” in six countries (system and recently most electorally successful party) 290 7.2 Linkage segmentation, socioeconomic development, and social inequality 292 7.3 GDP per capita, inequality, and programmatic linkages 296 7.4 Health expenditures as % of GDP and programmatic linkages 298 7.5 Levels of fiscal decentralization and leadership-based appeals 300 8.1 Poverty estimates, municipal enrollment, and PSUn scores in Greater Santiago municipalities included in the electoral district sample (Chapter 4). 334 8.2 Ideal vs. perceived role of politicians in terms of favoring the “poor” or the “rich” circa 2012 (Chile and Uruguay) 336

LIST OF MAPS 3.1 Chilean and Uruguayan voters’ positions along the state–market and authoritarian–democratic divides (1988) 123 3.2 Chilean and Uruguayan congressional representatives’ positions along the state–market and authoritarian–democratic divides (1993/1997–1995–2000) 125

xiv

List of Illustration

3.3 Chilean and Uruguayan congressional representatives’ positions along the state–market and liberal–conservative divides (1993/1997–1995–2000) 125 3.4 Chilean and Uruguayan partisan representatives on the state– market and liberal–conservative divides (2006/2010–2005/2010) 127 3.5 Chilean and Uruguayan partisan electorates on the state–market and liberal–conservative divides (2006) 128 3.6 Chilean partisan electorates on the state–market and liberal– conservative divides (2006) 128 3.7 Uruguayan partisan electorates on the state–market and liberal–conservative divides (2006) 129 3.8 Multidimensional issue congruence, Chile 2006 132 3.9 Multidimensional issue congruence, Uruguay 2006 133 3.10 Multidimensional issue congruence, Chile 2006: state–market/liberal–conservative (low-educated voters only) 134 3.11 Multidimensional issue congruence, Uruguay 2006: state–market/liberal–conservative (low-educated voters only) 135 6.1 Social fragmentation and district boundaries (each elects two congressional representatives) in the metropolitan region of Santiago 273 6.2 Social fragmentation in the district of Montevideo (elects forty-four congressional representatives) 274

List of Tables 1.1 1.2

Linkage ideal types 27 Structure of segmentation along territorial and socioeconomic divides 33 1.3 Structure of required linkage harmonization of electoral appeals 37 1.4 Data sources for measuring different dimensions of the dependent variable 54 1.5 Summary of theoretical arguments on determinants of party–voter linkages 56 2.1 Socioeconomic linkage segmentation in Chile (pre-1964) 76 2.2 Socioeconomic linkage segmentation in Chile (1964–1973) 78 2.3 Socioeconomic linkage segmentation in Uruguay (until 1973) 79 2.4a Left–right self-placement and polarization of university students in pre-authoritarian Chile 80 2.4b Left–right self-placement and polarization of university students in pre-authoritarian Uruguay 80 2.5 Comparative analysis of mobilization strategies in Chile and Uruguay (pre-1973) 84 2.6 Socioeconomic linkage segmentation in Chile (post-1989) 89 2.7 Socioeconomic linkage segmentation in Uruguay (post-1984) 91 2.8 Comparative analysis of mobilization strategies in Chile and Uruguay (post-transitional period) 95 2.9. Citizen’s justifications for their vote in the 1988 plebiscite 97 3.1 Preferred presidential candidate by social strata and ideological leaning (2001 and 2005) 141 4.1 Districts’ socioeconomic characteristics 145 4.2 Districts’ political trajectories 146 4.3 Summary characterization of district types 149 5.1 Socioeconomic linkage segmentation in the electoral strategy of the UDI 193 5.2 Socioeconomic linkage segmentation in the electoral strategy of the FA 193

xvi

List of Tables

5.3 Mechanisms required for achieving harmonization on the basis of each party’s predominant linkage strategies 203 5.4 Lower-chamber (1989–2001) and municipal election (2004) results per electoral pact and mainstream parties (%) 206 5.5 Electoral results 1942–1999 (%) 227 6.1 Party–voter linkage change in Chile and Uruguay: a summary of the descriptive argument 260 7.1 Operationalization of fuzzy sets representing different segmentation and harmonization scenarios 285 7.2 Party-system’s membership into four fuzzy sets 286 7.3 Linkage and strategy classification of six case studies 288 7.4 Observed linear correlation between linkage structure types and socio-structural variables 293 7.5 Observed linear correlations between institutional incentives, candidate-based linkage types, and linkage segmentation outcomes 299 7.6 Observed linear correlations between linkage types, institutional incentives, and linkage harmonization outcomes. 301 7.7 Fiscal decentralization as a necessary/sufficient cause of segmented and harmonized linkage strategies? 302 7.8 Fiscal decentralization as a necessary/sufficient cause of segmented and non-harmonized linkage strategies? 302 8.1 Cerro Navia and Vitacura: educational policy over time (2001–2010) 333

List of Abbreviations ARENA Alianza Republicana Nacional (El Salvador) AU Asamblea Uruguay (Uruguay) BA bureaucratic-authoritarian regime BJP Bharatiya Janata Party (India) BSP Bahujan Samaj Party (India) CCZ Centros Comunales Zonales (Uruguay) CUT Central Única de Trabajadores (Chile) DC Partido Demócrata Cristiano (Chile) DSV double simultaneous vote EP Encuentro Progresista (Uruguay) FA Frente Amplio (Uruguay) FA Nacionalista Frente Amplio Nacionalista (El Salvador) FMLN Farabundo Martí de Liberación Nacional FPMR Frente Patriotico Manuel Rodríguez (El Salvador) ISI import-substitution industrialization JD Janata Dal (India) MAPU Movimiento de Acción Popular Unitaria (Chile) MAS Movimiento al Socialismo (Bolivia) MIR Movimiento de Izquierda Revolucionaria (Bolivia) MLN-T Movimiento de Liberación Nacional Tupamaros (Uruguay) MNR Movimiento Nacional de Rocha MPP Movimiento de Participación Popular (Uruguay) MSM Movimiento Sin Miedo (Bolivia) MST Movimento dos Trabalhadores sem Terra (Brazil) NM Nueva Mayoría (Uruguay) NS not segmented party-voter linkage strategy OnP only non-programmatic party voter linkage strategy OP only programmatic party-voter linkage strategy OPP Oficina de Planeamiento y Presupuesto (Uruguay) ORDEN Organización Democrática Nacionalista (El Salvador) PAIS Partido Amplio de Izquierda Socialista (Chile) PC-Chi Partido Comunista (Chile) PC Partido Colorado (Uruguay) PCN Partido de Conciliación Nacional (El Salvador) PID Party-identification

xviii PIT-CNT PCU PN PPD PR PS PSU PSUn PSDB PT RN UDI UP WVS

List of Abbreviations Plenario Intersindical de Trabajadores (Uruguay) Partido Comunista (Uruguay) Partido Blanco/Partido Nacional (Uruguay) Partido por la Democracia (Chile) proportional representation Partido Socialista (Chile) Partido Socialista (Uruguay) Prueba de Selección Universitaria (Chile) Partido de la Social Democracia Brasileira (Brazil) Partido dos Trabalhadores (Brazil) Renovación Nacional (Chile) Unión Demócrata Independiente (Chile) Unidad Popular (Chile) World Values Survey

Prologue This book has come a long way. The project started in 2001 as my dissertation proposal at the Department of Political Science of the University of North Carolina at Chapel Hill. It came to an end in the Spring of 2013, while I was a visiting scholar at the David Rockefeller Center for Latin American Studies of Harvard University. In between, I spent most of my time in Santiago de Chile, as a scholar at the Instituto de Ciencia Política of the Pontificia Universidad Católica de Chile (ICP at PUC). I worked on a first version of the manuscript while on sabbatical at the Program for Latin American Studies at Princeton University in 2008; and then, once again, came back to the project (after having dropped it for the second time), while enjoying a short but incredibly fruitful stay at the Watson Institute for International Studies at Brown University. During the twelve years in which the manuscript was protracted, I was distracted by life-changing events. I started this project as the single-child of two elder parents, and as a job-less individual. I now end it as the happy father of Joaquina and Santiago, as the older surviving male in the family, and as a happily placed professional. Yet, my difficulty in driving this project to conclusion relates more to intellectual struggles than to family and professional developments. I “definitively” dropped the project between 2005 and 2008, and then again, from 2009 to early 2011. In retrospect, during those times, I went back and forth between the feeling that the manuscript was “just a footnote to conventional wisdom” (as put by a member of my dissertation committee) and the belief that it was still worthy of seeing the light. The fact that turning my dissertation into a viable book would require much extra work (as correctly anticipated by another member of my committee) only increased the opportunity costs of the enterprise. Along with Evelyne Huber’s tender insistence and unfaltering intellectual and emotional support during all these years, the 2008 Juan Linz Best Dissertation Award of the Comparative Democratization Section of the American Political Science Association and a Post-Doctoral Fellowship awarded by the Latin American Program of the Woodrow Wilson Center for Scholars gradually convinced me that it was still worth giving it a shot. Subsequent political events observed in both Chile and Uruguay provided indications of external validity to the original argument, also contributing to encourage me to pursue the project further. However, while I was finally convinced to finish the manuscript, I started to struggle with a second problem. I knew I had a sound critique of established approaches to party’s electoral strategizing and a host of empirical evidence to support it. At the same time, I increasingly felt challenged to turn such critique

xx

Prologue

into a potentially useful alternative. As Deborah Yashar once put it in one of our conversations during Fall of 2008, she could not yet hear “my voice.” Although my own limitations militated against finding such “voice” and turning it into harmonic sound, I also began to feel increasingly uncomfortable with the paradigmatic shift taking place in mainstream political science. To me, the emerging orthodoxy felt like a Chinese shoe in which my argument would not fit. My tropism towards splitting (instead of lumping) and my sense that we still needed to work much more in descriptive inference before jumping into sophisticated causal inference (more often than not, on the basis of rough proxies) did not fit well with academic orthodoxy. Again, why bother to write a book that no reviewer attuned to the current orthodoxy would view as publishable? Only when assured that I  could live a decent and high-quality life, surrounded by supportive and fun-to-work-with colleagues at the ICP without the perils of worrying too much about tenure, I decided to insist on the enterprise. Laurence Withehead and Dominic Byatt’s interest in the manuscript also helped greatly to spearhead my decision to write the final version. In doing so, I finally felt that I could freely write as rigorously as I was able, but without having to comply fully with a set of epistemological and methodological orthodoxies that had tied my hands before. So much for the intellectual and “micro-foundational” trajectory of the project. Prologues are also the place to thank those that gave you headaches; and those that provided analgesics and healing along the way. Before arriving at Chapel Hill, Fernando Filgueira provided excellent advice and encouragement. During these years, I continued to have the privilege of his friendship and brilliant scholarly advice. I also enjoyed the friendship of his father Carlos, whose death left, beyond a personal loss, a vast vacuum in the Uruguayan social sciences. Diego Hernández has also been a continuous source of friendship and intellectual exchange since our undergraduate years at the Universidad Católica in Montevideo through the completion of this book (he helped me put together the district maps that appear in Chapter 6). In spite of his busy schedule as Chair of the Department of Political Science at the time, Jonathan Hartlyn always found the time to meet and discuss my research with him, also providing extremely helpful advice and ideas over the years. Especially when enthusiasm was lacking, Lars Schoultz was ready to “meet for coffee at the Daily Grind,” and was a continuous source of charm, friendship, and savvy professional and personal counseling. None of “our Karinas” know how much was said about them by the small tables under the sun (or around the bookshelves, when winter hit). At UNC I also met Manuel Alcántara, with whom I  later engaged in collaborative research that benefited this project. During my stay at UNC and through a series of meetings across the United States,

Prologue

xxi

my participation in the research group headed by Herbert Kitschelt on Latin American party-systems was fascinating. Herbert has also embraced my criticism to his framework with intellectual openness and generosity. Additionally, research collaboration with Elizabeth Zechmeister, Guillermo Rosas, and Kirk Hawkins was both enriching and enjoyable. Since my arrival to Chapel Hill in 1999, Ngetta Kabiri contributed significantly in creating a home away from home. I will always remember our talks, very late at night, in-between reading assignments and reaction papers to be written by the next morning. Oscar Chamosa, Maria Marta Sobico, Leticia Ruíz-Rodríguez, José Miguel Sandoval, Indira Palacios, and Merike Blofield also became close friends. Such friendship has stood the tests of time and distance and constitutes an unanticipated gift that I  took from Chapel Hill. Christina Ewig also supported me through the intricacies of my first year of graduate student life in the department and provided much needed advice. After coming back from fieldwork in 2003, most of my old friends were no longer in town. However, I was fortunate to meet Jennifer Pribble, Agustina Giraudy, and Mireya Dávila, whose friendship and company I now cherish. Phyllis Howren, Laura Doerfer, and her children frequently helped to put things in perspective and extended their affection to Karina and me. All of them made our life in North Carolina substantially more enjoyable. At UNC, the research was made possible by several rounds of funding provided by the pre-dissertation grants (2000, 2001, and 2002) awarded by the Institute of Latin American Studies and funded by the Ford Foundation, the Tinker Foundation, and the Mellon Foundation. A  CLACSO fellowship for young Latin American scholars awarded in 2001 was also helpful in surveying the literature, analyzing preliminary evidence, and refocusing the project. The full research project could have not been pursued without the financial support provided by the 2002 University Center for International Studies Inaugural REACH Fellowship. While pursuing dissertation fieldwork in Chile I  was associated with the Universidad Andrés Bello and the Universidad de Chile. I am especially thankful to Juan Esteban Montes and Alfredo Joignant for supporting my research and for providing advice during my stay at those institutions. David Altman, Rossana Castiglioni, Tomás Moulián, Norbert Lechner, Rodrigo Márquez, Jorge Rodríguez, Cristián Beltrán; Carolina Stefoni, Miguel Angel López, and Rodrigo Araya provided helpful advice and much needed networking during my early days in Santiago. During my fieldwork in Uruguay, I  also benefitted extensively from the help of Ignacio Zuasnábar, Cecilia Rossell, Rosario Queirolo, Gerardo Caetano, Constanza Moreira, José Fernández, María José Alvarez, and my uncle, Enrique Fariña, who provided invaluable contacts in every one of the districts in the interior of the country. My father

xxii

Prologue

also contributed significantly through his own contacts. Nicolás Puljak, Marianela Fernández, María José Ramos, Claudia Italiano, Denisse Gelber, Cecilia Blummeto, Martín Fariña, and Maria Clara Caponi helped me to transcribe more than a hundred hours of tape. Many others still remain to be processed. My students of Sistema Político I at the Universidad Católica del Uruguay took part on a fieldwork experience in several Centros Comunales Zonales and debated my hypotheses during our class. Many of them enthusiastically embraced the project and I  am extremely thankful to them for that. In particular, Denisse Gelber and the late Pablo Alegre, a bright young scholar and human being whose death I still cannot come to terms with, provided high-quality research assistance and feedback at that time. Their own complementary fieldwork was partially funded by the Facultad de Ciencias Humanas and IPES at the Universidad Católica del Uruguay. IPES Director at the time, Rubén Kaztman, supported that project from the first day and I am very thankful for that. Before completing the dissertation, the Instituto de Ciencia Política of the Universidad Católica provided an ideal working and academic environment. I am particularly thankful to Tomás Chuaqui and the Faculty Board, who decided to hire me trusting that I would finish the dissertation in due time. David Altman, Rossana Castiglioni, Claudia Heiss, Claudio Fuentes, and Robert Funk provided valuable feedback as well as encouragement and support during the final stages of the project, while my other friends and family supported the enterprise from a safe distance. When work seemed insurmountable, a horse-ride at home or un partidito de Frontón with Martín, Andrés, Zapata, or Fernando in Los Titanes or at Eutko Etxea in Santiago did the trick. Dissertation writing was also enriched by further comments from George Rabinowitz, Marco Steenbergen, James Stimson, Gary Marks, Jennifer Pribble, Mireya Davila, Agustina Giraudy, and Ryan Carlin. During the book writing stage, a conversation with Gary Goertz at APSA 2011 became pivotal for substantially improving the packaging and illustration of the argument. Ken Roberts, Richard Snyder, Vicky Murillo, Juan Linz, Gerry Munck, and Deborah Yashar also provided detailed feedback at this stage. After a preliminary version of Chapter 1 of the book was bashed at a seminar at Columbia in the Spring of 2011, Bob Kaufman provided charm, professional wisdom, and incredibly useful feedback over lunch at his place on how to repackage the argument more convincingly. Also in 2011, David Collier, Jennifer Cyr, Maria Paula Saffon, Santiago Anria, Andreas Feldmann, Omar Sánchez, Yanilda González, Carlos Meléndez, and Sara Niedzwiecki provided valuable advice during an ICQRM research section. In March 2012, Anthony Pezzola organized a book-worshop at the ICP, in which I had the opportunity to discuss a complete first draft with an incredible group

Prologue

xxiii

of colleagues composed by:  Julieta Suarez Cao, Cassandra Sweet, Marissa Von Bulow, Rodrigo Mardones, Valeria Palanza, Pierre Ostiguy, Andreas Feldmann, Richard Snyder, Ken Eaton, Dan Brinks, Evelyne Huber, Hillel Soifer, and Agustina Giraudy. Several doctoral students at the ICP, Sergio Toro, Germán Bidegain, Ana María Farías, Roody Reserve, and Felipe Monestier, took part in the workshop and also raised difficult questions and critiques. After the workshop, Pierre Ostiguy provided detailed (page-by-page) and very rich feedback in two meetings that lasted well after midnight. After the rewriting was done, Valeria Palanza volunteered to organize two mini-workshops to discuss new versions of the most problematic chapters. Those workshops were also crucial in fine-tuning the last version of the manuscript. Valeria Palanza, Rodrigo Mardones, Andreas Feldmann, Francisca Reyes, Julieta Suarez Cao, Pierre Ostiguy, Marissa Von Bulow, Roberto Durán, and Diego Rossello provided detailed feedback and ideas at those instances. During the writing-up process I  also received feedback and help with different sections of the manuscript from Rossana Castiglioni, Jennifer Pribble, Manuel Antonio Garretón, David Altman, Juan Bogliaccini, Jorge Valladares, Santiago Anria, Alfonso Ferrufino, Cesar Zucco, Daniela Campello, Mitchell Seligson, Robert Funk, Marcela Ríos, Bill Smith, Noam Lupu, Mariano Tomassi, Dietrich Rueschmeyer, Patrick Heller, Steve Levitsky, Jorge Domínguez, Frances Hagopian, the late Guillermo O’Donnell, James Mahoney, John Stephens, Arturo Valenzuela, Marcelo Leiras, and Gabriel Vommaro. Research assistance during this period was provided by a set of incredibly talented undergraduate students at the ICP: Pilar Giannini, Diego Díaz-Rioseco, Rodrigo Nuñez, Héctor Bahamonde, Rodolfo López, Martín Ordoñez, Giancarlo Visconti, and Valentina Salas. Several arguments were also exposed and enriched by exchanges with the first four generations of doctoral students at the ICP. In particular, Sergio Toro, Rafael Piñeiro, Ana María Farías, Antoine Maillet, Germán Bidegain, Roody Reserve, and Felipe Monestier have all influenced my thinking and rewriting of significant sections of the book. Among that group of graduate students, Fernando Rosenblatt assisted me in writing up the final version. Fernando has undoubtedly read the final version of the manuscript even more than I. We iterated versions of every chapter, reaching a record of twenty-seven versions for one of them. Even after version twenty-six, Fernando still had sound and useful comments to make. I cannot thank him enough for his help and intellectual contribution to the project during the final stretch of writing during most of 2012 and early 2013. During that time, Elaine Thomas also managed to turn my unruly writing into sound prose, while also dropping very useful substantive comments along her way throughout the manuscript. Dania Straughan and Nikolai Stieglitz also helped significantly in preparing the last version for review and in producing

xxiv

Prologue

the last wave of images for the book. Of course, I am the only one to be held responsible for the errors and omissions that remain. Previous versions of the argument presented in Chapter 5 were published in the following two articles:  “Frente Amplio and the Crafting of a Social Democratic Alternative in Uruguay,” Latin American Politics and Society, vol. 49 (4):  1–30; and “Segmented Party-Voter Linkages in Latin America:  The Case of the UDI,” Journal of Latin American Studies, vol. 42: 325–56. I thank Latin American Politics and Society and the Journal of Latin American Studies for their permission to reproduce extensive passages of both articles here. The data used in this project originates in the Projecto de Elites Parlamentarias of the Universidad de Salamanca, the World Values Survey, the Projeto Cone Sul, the Americas Barometer and LAPOP Surveys, the Democratic Accountability and Linkages Project, and additional data generously provided by Nestor da Costa and Equipos-Mori for Uruguay. Without having access to that data, the project, in its current form, would have been impossible to develop. Also, this project could not have been possible without the collaboration of close to two hundred individuals who opened their houses, offices, and lives to a complete stranger who came to ask bothersome questions. Those individuals ranged from highly reputed congress-members and ministers to shanty-town dwellers, from the extreme right of the ideological spectrum to the most radical left. From them, I learned how “politics really works” in Chile and Uruguay and I am enormously thankful for their generosity in sharing their time and knowledge with me. Although I cannot thank them all here, I cannot avoid mentioning Margarita Cofre in Lo Hermida (Santiago) and “El Negro” Miranda in San Luis (so close to home, in the “golden coast” of Canelones). Memories of their sense of dignity in facing the rigors of a tough life will always overwhelm me. In a way, I hope that some of this work helps to improve the lives of others like them in the future. During the writing-up phase I received generous funding from the following research projects: FONDECYT numbers 1060749, 1060760, 1090605, and 1110565; and MILENIO # NS 100014. Along the fascinating (and many times, painfully boring) journey that ends with this book, I have also accumulated massive debts to my family. I regret my father could not see the project finished. I am sure he would have been thrilled to have the book in his hands, even if he could not understand a word of English. I am glad my mother will have the opportunity to see it finished. Although her own academic career has undoubtedly inspired mine, her most decisive contribution to this project was made unconsciously (to both of us) many years ago, during Uruguay’s transition to democracy. Accompanying my mother to instances of grassroots political mobilization against the authoritarian regime and then, to campaign events for the 1984 elections exposed me to a fascinating dimension of political life. At that same time, my father’s career

Prologue

xxv

in journalism provided the opportunity to meet several—and, at dinner, hear about many more—frontline Latin American politicians of the time. I believe my “dual” and premature exposure to politics at that time, still transpires in my understanding of politics. Such exposure would have not been possible if my parents had treated me “only” as the small child I was at that time. My deepest gratitude goes to Karina and to our children, Joaquina and Santiago. They have always stood by providing understanding, love, and much needed distraction and perspective. During these years, the three of them have taught me (not always nicely!) what I was missing by focusing too much on academic life. I cannot be more thankful for that lesson; my life is richer and substantially more meaningful because of that. As a result of such learning, but also due to my usual dissatisfaction with political science’s “mainstream,” I  now often find myself enjoying prologues more than the substantive stuff that follows. Prologues reveal the human and collective dimension of research that usually gets unseen behind the individualism and professional postures that are so prevalent in academia. I hope this prologue has done so, and that you regardless enjoy what follows. Juan Pablo Luna Pirque

Introduction The city of Santiago in Chile has an endemic problem with street dogs. In the early 2000s, a political party that controlled several mayoralties in Santiago’s metropolitan area devised a solution to that problem, without having to sterilize or euthanize street dogs. The latter would have run against the party’s ideology, and would have given rise to a strong reaction by dog-lovers and animal protection groups. The former would have also risked alienating some traditional supporters of the party who oppose contraception. Instead, the party decided to gradually (and secretly) hunt and trap the dogs populating upper-class and downtown neighborhoods, to then “ship” them a couple of miles away to lower-class municipalities also controlled by mayors of the party. Pursued over time this strategy would not attract much attention in the short run, but would improve quality of life in upper-class municipalities. Concurrently, the mayors of the party “willing” to accept street dog shipments would receive a “compensatory fee,” which included key resources for engaging in privately funded social assistance and particularism in their municipalities. Those mayors also gained (concealed) political recognition from the leaders of the party. Meanwhile, the party’s mayors governing upper-class municipalities could later proclaim at election time, without having to explain too much, their contribution to improving quality of life in their municipalities. Such mayoral success in providing valuable, visible public goods could then be useful for launching more ambitious political careers for national office. This vignette conveys the gist of the argument that unfolds throughout this book. The argument entails seven related but distinct claims, all of them at play in this example. First, parties can pursue highly segmented strategies to appeal to different electoral bases, especially when competing in unequal social settings (e.g., providing a public good by reducing the street dog population in upper-class districts while instead disbursing social assistance in lower-income districts). Second, if the society is at all socially fragmented and the party succeeds at simultaneously segmenting and harmonizing its appeals, the overall strategy is unlikely to be noticed and pinpointed by either constituents or the press,

2

Segmented Representation

even if the physical distance separating districts is not great. (The districts in the vignette above, for example, were only a couple of miles apart.) Third, though segmenting electoral appeals is promising for political parties, it could lead to inconsistencies or trade-offs (e.g., if the strategy becomes publicly known and/or if some mayors defect or do not fully comply with their side of the agreement). In such a case, the same party claiming credit for quality of life improvement through dog reduction in one district would be found to be actively fueling that very problem in another district. Fourth, to avoid such trade-offs, it is crucial for party leaders to harmonize segmented appeals. Such harmonization requires the capacity to command and strategically allocate symbolic and material resources for mobilizing and disciplining party candidates and activists. For instance, in this example, party leaders needed to be able to allocate political support and resources for social assistance to local party leaders (e.g., mayors in poor districts receiving dogs). Those resources were crucial in facilitating clientelistic mobilization. Meanwhile, party leaders helped nationally visible leaders to craft powerful reputations for effectively providing public goods, without taking measures counter to the party’s programmatic commitments to its original support base. Fifth, not every party can segment and harmonize appeals this way. Parties’ ability to engage in different segmentation schemes is contingent on their strategic situation in the party system, and on the range of material and symbolic resources at their disposal at a given point in time. For the dogs deal to work, the party needed to have mayors in both privileged and popular districts. It also needed access to private resources for funding the social assistance benefits given to dog-taking mayors in lower-income municipalities. And, lastly, the party needed a group of influential and respected leaders so that the mayors would abide by the decisions taken at the top. Sixth, segmented political strategies are facilitated by the kinds of territorial and political separation of social sectors common in unequal societies. The strategy for dealing with street dogs in Santiago would have not been feasible in a city where socioeconomic and territorial segmentation did not overlap or where district boundaries cut across the geographic boundaries of upper- and lower-class neighborhoods. Seventh, segmented strategies produce tangible distributive results. Whereas those living in wealthy municipalities benefit from improving public goods (e.g., disappearance of stray dogs from their streets and parks), those living in poorer municipalities receive deteriorating public goods (e.g., growing numbers of stray dogs outside). Although unknown to this last group of citizens, the deterioration of public goods in their municipalities increases the availability of short-term social assistance handouts.

Introduction

3

In light of my evidence and findings, current ideas of political representation in contemporary democracies need to be revisited. At the same time, our methods for assessing and comparing parties’ electoral strategies also need to be modified. This book aims to contribute to each of these larger undertakings.

SE G M E N T E D D E M O C R AT IC R E P R E SE N TAT IO N I N U N E Q UA L S O C I E T I E S We tell our friends: “to represent you, we should get more votes in the popular sectors, not at the elite level where we cannot get more.” So we ask them for their financial support, but we also ask them to abstain from showing up with us; in the snapshot, we will always be with the poor, not with them. (. ..) It was hard for them to understand that we needed to appeal to the poor, but that’s where there are more votes to grasp. (Anonymous UDI national leader, personal interview, 2008)

Imagine a rightist party that has a small but financially powerful core constituency that provides massive financial resources to its leadership in exchange for interest representation in Congress. As such, and in the context of an unequal society in which all social groups are politically enfranchised, our rightist party is condemned to a minority position in the party system. Yet, that party might seek to make electoral inroads into a peripheral constituency by deploying the financial resources it obtains from its relationship with its core constituency to pursue a non-programmatic strategy to attract poor voters. The peripheral constituency brings the “numbers” that the party needs to be electorally successful. Electoral success, in turn, allows the party to pay off its core constituency through favorable legislation and interest representation in Congress. This first example presents a stylized account of the electoral strategy implemented by the Unión Demócrata Independiente (UDI) of Chile, a party with strong roots in the Pinochet dictatorship that gradually became Chile’s most successful party, with more members than any other party elected to the lower chamber in the last three elections (2001, 2005, and 2009). Consider now the case of a leftist party with a long-standing programmatic platform, upon which it has developed a solid partnership with the union movement. The union movement is interested in protecting the vested interests of middle-class sectors dependent on public jobs and state transfers (i.e., pensions and social benefits targeted to those in the formal sector of the economy). While the leftist party has courted this core constituency, market reforms have contributed to weakening the clientelistic networks of its rival party, making a new constituency available for the leftist party’s electoral

4

Segmented Representation

expansion. Such expansion is needed to craft a winning electoral coalition, because the party’s core constituency is too small to provide a sizable plurality. However, the conventional programmatic appeals upon which the party relies for mobilizing its core constituency are unlikely to succeed in mobilizing the emergent (and electorally pivotal) peripheral constituency because the unemployed, as well as informal workers, have distributive preferences that differ from (and run counter to) those of formal sector workers. This second illustration presents a stylized account of the electoral strategy implemented by the Frente Amplio (FA) of Uruguay. Confronting this strategic scenario, the party had to mobilize its emergent constituency programmatically by prioritizing issues that do not conflict with the interests of its core constituency, and also develop non-programmatic appeals to pursue electoral expansion. After challenging one of the most enduring traditional party systems in Latin America, the FA, like the UDI in Chile, became Uruguay’s most successful party, receiving the most votes every election cycle since 1999. In 2004 and 2009, the party won the presidency, as well as an absolute congressional majority. These two examples illustrate the general argument of this book. Particularly in the context of unequal societies, political parties, as well as individual political entrepreneurs, will seek to build an electoral base by ideologically representing the views (and material interests) of a given portion of their electoral constituency. However, due to social inequality, that “programmatic constituency” will not provide enough votes to secure an electoral majority. Under such conditions, only those parties or candidates that are able to segment their electoral appeals can grow. I define segmentation as a process through which political parties and candidates deploy different types of electoral appeals to mobilize distinct constituencies. Both examples, as well as other cases analyzed in the book such as those of the Bharatiya Janata Party (BJP) in India, the Alianza Republicana Nacionalista (ARENA) in El Salvador, the Movimiento al Socialismo (MAS) in Bolivia, or the Partido dos Trabalhadores (PT) in Brazil also suggest that incentives to segment electoral strategies apply to parties of both the right and the left. Inequality creates straightforward constraints for a rightist party seeking to expand its electoral constituency by programmatically mobilizing its voters on the basis of their distributive preferences. Those constraints may initially seem less salient in the case of a leftist party. Why would it be difficult for a leftist party competing in an unequal society to rally an electoral plurality by programmatically favoring a radical redistributive agenda? Two constraints are crucial for the left. First, in a globalized economy governments are “structurally dependent on capital”

Introduction

5

(see Przeworski and Wallerstein 1988; Boix 2003; Acemoglu and Robinson 2006; Campello 2011). That dependence reduces room for promoting and implementing policies that could produce massive capital flight (Boix 2003; Acemoglu and Robinson 2006; Campello 2011), or long-term decline in economic growth.1 Second, but equally important, the target constituency of the left is not internally homogenous.2 As a result of this second constraint, even when targeting the middle and lower-middle classes, leftist parties also confront the challenge of courting socially divided constituencies with opposing distributive interests.3 My findings suggest that leftist parties also respond to this challenge by socially segmenting their electoral appeals. In spite of their many differences, the parties analyzed in the book share a common trait: their ability to deploy simultaneously segmented and harmonized linkages to distinct socioeconomic constituencies. Both the UDI and the FA (as well as the BJP, the MAS, the PT, and ARENA) have pursued a segmented strategy to appeal to different socioeconomic groups. To build a large and stable constituency while competing in unequal societies, those parties sought to combine the programmatic mobilization of distributive appeals to mobilize their core constituency with varied electoral appeals to attract other groups of voters. In some cases, the appeals directed at peripheral constituencies have involved programmatic mobilization of alternative (non-distributive) issues. In other cases, clientelistic or other kinds of non-programmatic appeals have been pursued. In combining different sets of electoral appeals to target different constituencies, the parties analyzed here have pursued segmented electoral strategies.4 However, as introduced above, although appeal segmentation resolves the daunting dilemma of confronting significantly different

  1  Leaders or parties with access to massive state rents (e.g., oil, gas, or mineral exporters) enjoy greater autonomy from capital, and can therefore engage in more radical redistributive programs. This possibility might restrict the external validity of my claim to non-rentier states.   2  The effects of inequality in shaping preference formation are also consequential in this regard. I explore those effects in Chapter 1.   3 This second condition merits some illustration. The divide between the working class included in the formal sector of the economy and informal workers creates conflicts of interest within the target constituency of the left (Przeworski and Sprague 1986). Those conflicts of interest are not abstract, but rather visible and politically salient. For instance, in Latin America pension benefits are regressive because they were traditionally tied to enrollment in formal sector jobs, which were traditionally occupied by middle-class sectors (see Mesa-Lago 1997; Filgueira 2005). Whereas pension beneficiaries have a stake in protecting the policies in which those benefits are embedded, informal sector workers would be better off if the current system was dismantled and a (universalistic) pension system reform was implemented (Filgueira 2005; Huber et al. 2010).   4  See Gibson (1996 and 1997) for a similar argument.

6

Segmented Representation

electorates, it can introduce contradictions in parties’ overall electoral strategies. In order to be successful, the parties analyzed here have harmonized their electoral appeals into a coherent strategy. The pursuit of segmented but collectively harmonized electoral appeals has given each party an electoral edge over rival parties that were unable to segment their appeals to different constituencies and/or unable to harmonize segmented appeals into a coherent overall party strategy. Pervasive as the segmentation of electoral appeals is and has been, its substantive and theoretical implications remain little understood. In order to fully grasp these implications, our conventional theoretical and methodological approach to the study of electoral mobilization needs to be amended. The framework for understanding linkages between parties and voters originally proposed by Herbert Kitschelt has great potential for illuminating segmented patterns of representation. However, conventional applications of that framework in the analysis of programmatic and non-programmatic mobilization strategies have neglected the presence of socially segmented party–voter linkages. Instead of integrating segmentation into the analysis, analysts have tended to favor aggregate assessments, describing parties and party systems as either “programmatic,” “clientelistic,” or “personalistic.” While sometimes useful for comparative analyses, those aggregate assessments obscure the fact that given individual parties and even given leaders often rely heavily on more than one such type of appeal, shifting among them as they seek to appeal to distinct constituencies. Mainstream approaches to the analysis of electoral strategizing and party politics theoretically neglect the presence of segmented strategies. Such neglect is further reinforced by methodological approaches that make segmentation invisible to the analyst. There is a pressing need to correct the biased descriptive inferences we obtain from applying the conventional party–voter linkage framework. These biased descriptive inferences lead to deceptive causal analyses, subsequently obscuring assessments of the determinants of different linkage types, their distribution among different social groups and across societies, and their political and distributive consequences. To be sure, the segmentation of electoral appeals is not a new discovery (see, e.g., Valenzuela 1977; Gibson 1992, 2005; Coppedge 1998; Levitsky 2003; Hagopian 2009; Morgan 2011; Calvo and Murillo n.d.).5 Taylor-Robinson   5  For instance, as argued by Coppedge (1998, p. 552): “Before we can assess more accurately how ideological Latin American parties and party systems are, it is necessary to clarify three issues. First, personalism and ideology are not necessarily mutually exclusive qualities. Some of the most rigidly ideological parties in the world have been closely identified with, and tightly controlled by, strong personalities, and parties that are known primarily as vehicles for strong

Introduction

7

(2010) has even recently presented an explicit argument on segmentation.6 Although its implications are in some respects similar, my argument differs from Taylor-Robinson’s in several important ways. First, in my framework parties eventually have the capacity to avoid “structural determinism” and mobilize “poor voters” programmatically, provided favorable partisan agency and favorable structural conditions exist. Second, in my framework the agent could either be a party or an individual candidate. I therefore place equal emphasis on the need to segment linkages and to harmonize mobilization strategies at the partisan (collective) level. Last, but not least, my argument does not only consider programmatic and clientelistic linkage strategies. Other linkage strategies, some of which are not contingent on short-term interactions between a political party and its voters (e.g., party identification, which is usually built on a long-term relationship between the party and its identifiers), are also crucial in understanding the complex strategies that parties can pursue. The evidence I  present in this book suggests that segmented electoral appeals are an enduring characteristic of unequal democracies. The structure of segmentation and the types of electoral appeals that get segmented change over time, in accordance to parties’ adaptation to different opportunity structures. Historical and contextual conditions and the patterns of social inequality shape the relative sizes and profiles of different social constituencies, which are differentially prone to becoming mobilized electorally through diverse strategies.7 Electoral segmentation itself is ubiquitous and can be structured personalities may nevertheless stake out clear ideological positions. Second, clientelism and ideology are not necessarily mutually exclusive. Many successful parties all over the world trade personal favors for political support. Even in supposedly highly ideological Chile, party officials of all tendencies engaged in the same sorts of clientelistic activities (Valenzuela, 1977: 166). Clientelism is merely a means to build and maintain a power base; ideology, where it exists, is what guides what that power is used for. Many parties are to some degree clientelistic, to some degree personalistic and to some degree ideological; these three qualities vary independently.”   6  “Clientelism may help a legislator to represent both rich and poor constituents. The rich may want national policies that the poor cannot monitor, or that poor people view as unlikely to be implemented in a timely fashion or to directly affect them (e.g., increasing the safety of bank deposits or regulating private school tuition). In that case a legislator can address the rich person's national policy preferences without the poor person wanting to sanction. Meanwhile, the legislator can deliver clientelistic benefits to poor people. Clientelistic benefits may be the policy/service package the poor person desires because they are cheap to monitor and their value is clear. The legislator's clientelistic work could also win favor with the rich person who receives the contract to build a local public works project. In this scenario, poor and rich people do not have incompatible policy preferences; they are interested in different things and representing both should be feasible, unless the government's budget is so limited that it cannot afford to fund both policies and services” (p. 48).   7  For a recent argument and evidence on the structure of inequality and its effects on voters’ distributive preferences, see Lupu and Pontusson (2011).

8

Segmented Representation

in terms of socioeconomic categories, territorial dimensions, or both. In ethnically divided societies, this type of segmentation can also be implemented along ethnic or religious lines.8 This book contributes to the analysis of electoral mobilization by providing a systematic framework for analyzing segmented linkage strategies. That framework also considers different mechanisms that may help parties harmonize segmented linkages, avoiding trade-offs and exploiting the eventual synergies that linkages may create. Different partisan endowments give individual parties greater or lesser capacity to develop such complex linkage strategies and adapt them over time to new opportunity structures.

W H Y SE G M E N T E D R E P R E SE N TAT IO N M AT T E R S F O R R E D I S T R I BU T IO N A N D F O R PA RT Y SYS T E M   C HA N G E Where the political force of the idea of democracy came from in this new epoch was its combination of formal social equality with a practical order founded on the protection and reproduction of an increasingly dynamic system of economic inequality . . . No one at all in 1750 either did or could have seen democracy as a natural name or an apt institutional form for the effective protection of productive wealth. But today we know better. In the teeth of ex ante perceived probability that is exactly what representative democracy has in the long run proved. (Dunn 2003, p. 22)9 In politics we will be recognizing the principle of one man, one vote and one vote, one value. In our social and economic life, we shall, by reason of our social and economic structure, continue to deny the principle of one man, one value. How long shall we continue to live this life of contradictions? How long shall we continue to deny equality in our social and economic life? If we continue to deny it for long, we will do so only by putting our political democracy in peril. (B. R. Ambedkar, Chairman of the Drafting Committee of the Indian 1950 Constitution)10

Contemporary politics is shaped by two underlying trends:  the massive incorporation of previously excluded groups into electoral democracy, and an unprecedented increase in socioeconomic inequality.11 It follows that   8  See e.g., Chandra (2007).   9  Cited in Przeworski (2010, p. 84). Cf. Marx (1978 [1844]).   10  Cited in Guha (2008, p. 133) and reproduced in Przeworski (2010, p. 93).   11  See Przeworski (2010, pp. 84–98) for a provocative argument and evidence on the “dual” character of democracy as the realization of formal political equality, and as an institutional form that has been compatible with the protection and perpetuation of economic inequality.

Introduction

9

the political inclusion of the masses has not produced consistent results in reducing social inequality. Yet, the interaction between democracy and social inequality has normally been expected to produce significant sociopolitical change. Some expect democracy to languish under conditions of social inequality. Others instead expect democracy, when functioning in the context of high inequality, to produce social redistribution.12 These expectations are not borne out by recent experience in Latin America, the most unequal region in the world. High inequality and democratic politics have coexisted in most of Latin America for over three decades without significant shifts towards either redistribution or regime backlashes.13 Theories examining the link between democracy and inequality miss a crucial causal mechanism shaping how and to what degree social inequalities are translated into the political process, and thus how and whether significant sociopolitical change does result (Huber and Stephens 2012; Kaufman 2009). Formal models inspired in Meltzer and Richard’s (1981) influential analysis assume that parties will be universally and automatically responsive to the median voter’s distributive preferences, on the basis of a single-issue spatial model. The existence of programmatic linkages between parties and voters and the centrality of distributive preferences in forging such linkages is thus taken for granted. Furthermore, these theories assume that different social groups all have equal access to programmatic representation. Yet, programmatic representation is the exception, not the rule in Latin America.14   12  Accordingly, in highly unequal societies the median voter is expected to have strong redistributive preferences (e.g., Meltzer and Richard 1981), which will induce redistributive policymaking. Elite fear of redistribution might, at least under certain conditions (e.g., territorially bounded capital investments), lead to resistance to democratizing the political regime or to authoritarian backlashes (e.g., Boix 2003; Acemoglu and Robinson 2006). Even where elite assets are not fixed and capital mobility is possible due to globalization, these theories expect to see redistributive policies in response to median voters’ preferences. Although other works expect the poor to show less interest and engagement in collective political action due to their relative lack of resources for paying the costs of political participation (see, e.g., Ansolabehere et al. 2003; Goodin and Dryzek 1980), the “distributive struggle” approach, anchored in median-voter models (e.g., Boix 2003; Acemoglu and Robinson 2006)  has gained increasing popularity in recent years.   13  Although recent studies indicate the consolidation of a modest, yet significant, downward shift in economic inequality in the region (López-Calva and Lustig 2010; Huber and Stephens 2012; Economist 2012), most analysts coincide in signaling economic growth and the expansion of conditional cash-transfer programs as the most likely causes of this decline in inequality. In short, inequality has declined in Latin America, but seemingly not through the classical redistributive mechanisms envisioned in the political economy literature and problematized in this work (Huber and Stephens 2012).   14  Kitschelt et al. (2010). In a democratic society, political parties play a crucial mediating role in translating social inequality into policy outcomes, via the political representation of different

10

Segmented Representation

How does social inequality translate into the political process, in ways that hinder democratic politics in achieving its “equalizing” role? Answers to that question should deal with two related but analytically different issues:  the structure of political representation, and the translation of electoral preferences into policy outcomes. In this book, I focus on the first of those processes, seeking to analyze how inequality gets politicized in society, and how inequality shapes partisan mobilization strategies in ways that turn distributive issues into either politically salient or inconsequential ones. To address the last two issues, this book focuses on inequality’s role in shaping political parties’ electoral strategies. I thus seek to analyze how parties of the left and the right are able to craft viable electoral strategies while competing in a context of social (and territorialized) inequality and mass political incorporation. I argue that social inequality provides parties with both incentives and opportunities to segment their strategies for appealing and linking to distinct electoral constituencies. On the one hand, social inequality makes the pursuit of segmented linkages an attractive vote-gathering strategy when courting a socially heterogeneous constituency. On the other hand, when competing in unequal societies parties are better able to segment their linkage strategies while reducing the electoral trade-offs that might emerge than they are in the context of less socially fragmented civil societies (Weitz-Shapiro 2012; Kitschelt 2000). While segmented linkage strategies stabilize democracy by reducing threats to elites and thus polarization, they can also reinforce inequalities in citizens’ access to political representation, which is often already socially stratified. Socially stratified representation, in turn, would bias or inhibit redistribution. Therefore, social inequality would not be corrected through the political process (as assumed by existing theories), but rather replicated by asymmetries in political representation. Quite ironically, and contrary to common-sense predictions, inequality would then become harder to tackle democratically when income distribution is most socially skewed. To be sure, inequality can alternatively lead to polarization (i.e., via intense populist mobilization) and eventually to the demise of democracy. Such an outcome is consistent with the predictions of contemporary political economy models (see Boix 2003; Acemoglu and Robinson 2006). Yet, this book brings to light an alternative scenario in which democratic politics and high inequality societal interests. Latin American party systems have undergone massive turmoil produced by the decay (or disappearance) of traditional parties, their subsequent dismissal or entrenchment in new movements (Roberts n.d.) and the emergence of new parties and leaders (many of them outsiders, rallying an anti-party constituency).

Introduction

11

combine into a stable equilibrium. Such equilibrium runs contrary to conventional expectations (Meltzer and Richard 1981; Boix 2003; Acemoglu and Robinson 2006). One key element explaining the durability of both democracy and inequality in such cases is parties’ mobilization strategies. The remainder of this book examines these strategies and how social inequality contributes to shaping them. The ways in which political representation translates into policy outcomes via the policy process constitutes another element key to explaining such equilibrium scenarios, one that I do not discuss in this work.15 The types of electoral appeals that candidates make when mobilizing different electoral segments also have implications for understanding the dynamic evolution of party systems. This understanding not only applies to systems in which massive and rapid changes occur, in the aftermath of severe representation crises (Mainwaring et al. 2006). It also applies to cases that have seemingly more stable party systems like those analyzed in-depth in this book: Chile and Uruguay. The nature of successful electoral appeals feeds back into future party-system characteristics. For instance, at time t, parties and candidates might face an incentive to appeal to a significant share of voters on the basis of personalized, non-partisan (and frequently anti-party) mobilization tactics.16 When these electoral strategies are successful at the polls, citizens become increasingly disaffected from partisan politics at time t + 1. At this point, outsider candidates and non-partisan politicians will have a better electoral shot than they did at time t. This is an outcome that I describe for contemporary Chile, which was almost always considered a highly institutionalized and stable party system until recently. In turn, if at time t parties and their candidates face an opportunity structure that fosters a collective interest in appealing to voters on the basis of partisan identities and programmatic labels, systemic actors will be strengthened at time t + 1. The strengthening of systemic players does not necessarily configure a “closed” system, but may be compatible with more incremental processes of party-system reconfiguration, which accommodate significant change within an institutionalized party system. This type of institutionalized change fits the trajectory of the Uruguayan party system.

  15  See Pribble (2013) and Fairfield (2010) for arguments and evidence on how the nature of political representation translates into relevant policy outputs such as social policy or tax regimes.   16  Indeed, parties can, under given contextual conditions, engage in “brand dilution” (Lupu 2010).

12

Segmented Representation R E SE A R C H   D E SIG N

Although usually less valued than causal inference, descriptive inference is fundamental for causal analysis. We cannot hope to explain a phenomenon that we cannot descriptively apprehend. My research design is based on that premise. I thus focus, primarily and fundamentally, on trying to develop better descriptive inferences regarding the nature, scope, and social distribution of different types of party–voter linkages at the local, party, and systemic levels. Case Selection Taken together and compared to others in the region, the cases of Chile and Uruguay represent “best case scenarios” for stable, programmatic party–voter linkages in Latin America. Indeed, Chile and Uruguay’s party systems are conventionally seen as the most institutionalized in the region (Mainwaring and Scully 1995; Mainwaring 1999; Jones 2005). They have also been characterized as the most programmatic party systems (Kitschelt et al. 2010). In terms of this portrait, I have selected on the dependent variable. While selection on the dependent variable is not a potential problem for the conceptual innovation and descriptive inference presented for the analysis of segmented party–voter linkage strategies, it could be thought problematic for the kind of causal inference required to identify the causal mechanisms that explain the adoption of different linkage strategies. In reality, however, Chile and Uruguay’s party systems are less similar than many existing studies would lead us to believe. While comparatively valid in relation to other cases in the region, the conventional characterization of these two party systems is only partial, and is incapable of explaining continuity and change in either system. Indeed, this conventional wisdom risks obscuring our understanding of these cases. Perhaps most strikingly, the general level of connection between parties and voters in the two cases is radically different. In 2006, approximately 25 percent of Chilean citizens supported a given political party, while in Uruguay more than 60 percent identified with a party. In 2010, only 10 percent of Chileans sympathized with a party, while more than 60 percent of Uruguayans still did. In fact, in a comparison of twenty-three cases based on data from the LAPOP Survey 2010, while Chile had the lowest percentage of party sympathizers (12 percent), Uruguay had the highest (66 percent). The different levels of party identification in the two cases are traceable to significant mutations in the ways political parties have structured linkages to

Introduction

13

civil society, in the midst of socio-structural and institutional change. And the differences in party identification are just the tip of the iceberg. Re-examining Chile and Uruguay using my new analytic framework reveals that parties’ ways of combining programmatic, clientelistic, and leadership-based appeals in their strategies for mobilizing supporters are quite distinct across cases. Moreover, whereas Chile and Uruguay might be similar at the aggregate level, my unpacking of the party and district levels uncovers much more variance than these aggregate-level characterizations might lead one to suspect. Thus, the evidence I present throughout the book suggests that the alleged selection bias is far less significant than it might superficially appear. The diachronic evolution of programmatic and non-programmatic linkages in these cases has also run in opposite directions. Historically, both countries presented relatively pervasive non-programmatic linkages. However, the nature and scope of such linkages have been drastically redefined over time. In this respect too, the two cases have proven strikingly different. Causal theories designed to clarify why these two cases have similar aggregate levels of programmatic structuring cannot explain the different linkage-type configurations and trajectories I identify. Kitschelt et al. (2010), for instance, develop such an argument on the basis of a path-dependent causal chain, tracing back to the 1920s, in which the scope of import-substitution industrialization achieved up to the 1970s figures decisively (Kitschelt et al. 2010). My analysis emphasizes that, in spite of their similar long-term trajectories, both the political economies of state and decentralizing reforms and the character of the two countries’ transitions to democracy have diverged considerably. Their respective electoral institutions as well as other institutional features shaping incentives for party organization and campaigning also differ. Comparing Chile and Uruguay therefore helps shed new light on how short-term factors have contributed to shaping differing combinations of party–voter linkage strategies as well as successful and unsuccessful individual party adaptations. My case selection ultimately pays off analytically in two important ways. First, while allowing for a detailed unpacking illustrating the limitations of current single linkage descriptions, it underscores the complementary role of short-term, socio-structural changes in triggering segmented and distinct linkage strategies. The pairing of the Chilean and Uruguayan cases is able to highlight the role of such short-term factors because of the countries’ unusual similarity when it comes to long-term drivers of programmatic structure. If the scope and evolution of their programmatic linkages differs, including over time, it is therefore plausible to conclude that short-term variables have played at least a relevant mediating role in shaping these outcomes. Second, precisely because Chile and Uruguay are such “exceptionally favorable” cases,

14

Segmented Representation

if the extent of programmatic linking is limited and extensive segmentation is required for parties to succeed even there, then linkage segmentation should be all the more salient in countries lacking such favorable historical and structural conditions. The nature of the initial case selection therefore allows for some general but important logical deductions regarding other cases with less favorable long-term drivers. The Western European literature has acted as an implicit benchmark in analyses of party systems in developing countries, which usually fail to meet it. While the findings I present in this book might be seen as another “failure” of this sort, they also illustrate a possible path for the evolution of party systems in advanced industrial societies, especially those that have recently witnessed rising inequality, state retrenchment, and political processes of regional and local devolution. The large-N analysis I  present in Chapter  7 suggests that segmented party–voter linkage strategies also characterize the contemporary party systems of some Western European countries such as France, Greece, and Italy.

Multi-level Research Design: Paired District Comparison The empirical analysis that unfolds in the book is based on a multi-level research design structured around comparison of twelve congressional districts (plus nineteen municipalities) and two national cases: Chile and Uruguay. For each case, I  selected a sample of congressional districts maximizing variance in socio-structural configurations (socioeconomic background, social heterogeneity) and political trajectories (political competitiveness, electoral evolution). The characteristics of the district sample for each case are discussed in Chapter 4. This multi-level research design permits multiple, structured comparisons, conferring greater levels of internal validity to my causal inferences, even where there are limited degrees of freedom. Those structured comparisons permit me to account for the effects of district characteristics, individual partisan factors, and systemic traits on party–voter linkage strategies. This multi-level comparison between Chile and Uruguay allows for a thorough measurement of the dependent variable: the extent to which different linkage types are segmented, their deployment in mobilizing different electoral constituencies, and their evolution over time. Additionally, the combination of comparative and cross-sectional analyses across districts both within and between cases allows for investigating the determinants of representational gaps between different social groups. Finally, analyzing district-level variation (controlling for variation at the national level) enables one to analyze an

Introduction

15

additional set of district-level independent variables (electoral competitiveness, socioeconomic development and social heterogeneity, degrees of state intervention in the local economy, and incumbency). The multi-level comparison between Chile and Uruguay is thus advantageous for a number of reasons. Tools for Observing and Analyzing Different Types of Linkages To analyze programmatic linkages and their evolution over time, as well as their cross-sectional distribution, I also statistically analyze data from available survey research (the 1983 Proyeto Cone Sul; the second, third, fourth, and fifth waves of the World Values Survey of the University of Michigan and a parallel survey conducted in Uruguay in 2002; the LAPOP Survey for 2006– 2010; and the first and fourth waves of the Encuesta de Elites Parlamentarias de América Latina of the Universidad de Salamanca). For this part of my analysis, I  triangulate citizen and party representative issue positions on a set of relevant programmatic dimensions, including preferences on:  social policy, economic development strategy, political regime, and value and moral predispositions. To analyze the nature and scope of non-programmatic linkages, I pursued extensive fieldwork at the district level. In the field, I observed political activities during campaign- and off-season ethnographically from 2002 to 2009, and conducted in-depth interviews with more than 180 party activists, qualified observers, congressional and local council members, and national party leaders. I conducted a first round of interviews in both countries in 2002 and 2003, and a smaller second round in 2008–2011. Drawing on extensive interview material, I illustrate the nature of successful partisan adaptation to the emerging configuration of programmatic and non-programmatic linkages observed in each country.

PLAN OF THE BOOK Part I  of the book focuses on conceptualizing party–voter linkage types, disaggregating linkage analyses to the local level, and explicitly theorizing ways in which parties can coordinate different linkage types into a coherent overall party strategy. Part I also applies my proposed framework to descriptively assessing party–voter linkage configurations in pre-authoritarian and post-transitional Chile and Uruguay (Chapters  2, 3, and 4). My framework

16

Segmented Representation

does not yield ready-made estimates of linkage segmentation at the national level. In other words, my framework is not designed to generate aggregate characterizations of a party’s mobilization strategy as either “programmatic,” “clientelistic,” or “charismatic.” Instead, the framework I  present and apply in Part I  seeks to unpack the electoral appeals that different parties pursue when mobilizing distinct constituencies. In closing Part I, Chapter 6 provides a comparative analysis of the strategies of individual parties, comparing the cases of the UDI (Chile) and the FA (Uruguay) to those of less electorally successful parties in each system. Drawing on the conceptualization and description offered in Part I, Part II of the book identifies the possible causal mechanisms explaining the evolution and contemporary configuration of linkage types in Chile and Uruguay (Chapter 6). My argument is inductively inferred from a comparative historical analysis of these two cases. Such comparative historical analysis takes into account the political economies of the pre-1973, transitional, and post-1990s party systems as significant drivers of change in the types and segmentation of electoral appeals. Also in Part II, Chapter 7 analyzes the external validity of my argument by using a cross-national analysis for testing the argument derived inductively from the two core case studies. In pursuing this large-N comparison, I combine fuzzy-set analyses with simple econometric models. I also draw on four comparative vignettes developed to briefly test the potential for generalizing my descriptive and explanatory framework to cases lacking the structural and historical conditions found in Chile and Uruguay. This exercise complements the most-similar systems design applied in the book with a most-different systems exploration, analyzing the segmentation of electoral appeals in the following cases: the PT (Brazil), ARENA (El Salvador), the MAS (Bolivia), and the BJP (India) The comparative analyses I pursue in Part II suggest that the scope of the argument is not limited to the cases of Chile and Uruguay:  two relatively small and stable Latin American democracies, usually characterized as having the two most institutionalized and programmatic party systems in the region (Mainwaring and Scully 1995; Luna and Zechmeister 2005; Kitschelt et  al. 2010). Indeed, the argument applies to a larger set of cases in which mass democratic politics and high levels of social inequality are jointly present (Brazil, India, El Salvador, and Bolivia), irrespective of significant variance regarding other theoretically relevant independent variables (e.g., different levels of socioeconomic development, unitary vs. federal systems, ethnically/religiously divided vs. homogeneous societies). Moreover, the argument applies to successful parties of both the left (the FA, the PT, the MAS) and the right (the UDI, the BJP, and ARENA).

Introduction

17

The book concludes by exploring the implications of my analysis for future research on the link between democracy, inequality, and structures of political representation. I also briefly speculate on how the types of political mobilization strategies described in the book might relate to policy outcomes observed in Chile and Uruguay over the last two decades. The conclusion also discusses the methodological and theoretical implications of my findings for the study of party–voter linkages.

PA R T   I : D E S C R I P T I V E INFERENCE

1 Segmented Electoral Appeals: A Descriptive Framework In contexts of significant inequality, it is very difficult to successfully mobilize voters using just one type of appeal. While competing in unequal societies, parties’ main challenges consist of combining socioeconomic and territorially segmented linkage strategies, avoiding strategic contradictions between different types of appeals aimed at different social constituencies, and at the same time persisting as meaningful organizations. While parties that do not segment their appeals risk electoral decline, those that segment but fail to harmonize their strategies into a coherent party strategy risk fragmentation and atomization. More segmentation does not necessarily reduce a party’s capacity to harmonize its strategy. The two most successful cases analyzed in this book are the UDI in Chile and the FA in Uruguay. These are the parties that segment their linkage strategies the most within their respective party systems, and also those that best harmonize their segmented mobilization strategies. Therefore, political parties seeking to persist and grow over time need to both segment and harmonize diverse linkage strategies.1 To empirically grasp this reality, parties’ ability to harmonize and combine socially segmented appeals

  1  Many parties in Latin America have failed to solve the dilemma of segmenting their linkage strategies while holding their organization together. In cases of “democratic representation crises” and party system collapse, there have been instances of both linkage segmentation without harmonization resulting in partisan atomization across districts and of electoral collapse. Such results have been typical in Andean countries (Mainwaring et al. 2006; Morgan 2011). In other cases, national coalitions have been crafted through pragmatic alliances between heterogeneous provincial and local bosses commanding powerful clientelistic machines in poor areas. This arrangement has been characteristic of federal systems, and specifically of Argentina and Mexico (Gibson and Calvo 2001; Calvo and Escolar 2005; Gibson and Suarez-Cao 2010). According to conventional wisdom, the cases of Chile and Uruguay represent a third, more positive outcome. In these cases, stable and programmatically oriented party systems seem to have consolidated. Brazil is also increasingly seen as a party system approaching this configuration (Hagopian 2009; Hagopian et al. 2009; Zucco 2008).

22

Segmented Representation

needs to be analyzed, taking into account the relationship between each party and its different electoral bases. Edward Gibson’s (1992 and 2005) approach to electoral coalition-making in Latin America provides a very useful guide to doing so. According to Gibson, analyses of (conservative) party strategies should begin by distinguishing between the party’s core and non-core constituencies. The core constituency provides ideological and financial resources, and is the most important group in defining the party’s identity. However, the core constituency does not provide enough votes to make the party electorally viable. Conservative parties therefore need to make significant electoral inroads into non-core constituencies. The electoral strategy for attracting non-core constituents is necessarily different from that directed at core supporters, and usually entails de-emphasizing ideological appeals. Such parties thus face the challenge of harmonizing segmented electoral strategies to craft multi-class social bases. Following Gibson’s lead, one can then ask how parties bring and hold together activists and candidates crafting electoral appeals that require different types of symbolic and material resources, while fostering the party’s collective interest. In this chapter, I develop a new conceptual and analytical framework that captures the nature of parties’ strategic efforts to deploy and harmonize segmented electoral appeals by adapting the party–voter linkage framework. This adapted framework is needed for better understanding the gaps in political representation characteristic of highly unequal societies and the reasons why, contrary to received wisdom, unequal societies over time can remain both stable democracies and stubbornly inegalitarian in their patterns of distribution.

T OWA R D S A N A DA P T E D F R A M EWO R K 2 Observing the nature of linkages between parties and voters, and the partisan technologies that enable them, provides an ideal analytical window through which to observe how parties and party systems compete for political office by pursuing distinct electoral appeals. Kitschelt (2000) and Kitschelt and Wilkinson (2007) propose both such a project and an innovative framework for pursuing it—one explicitly devised to account for parties’ distinctive ways of forging relations of accountability with voters. Kitschelt’s invaluable

  2  In the remainder of the book, I use the following three terms interchangeably: “electoral appeals,” “mobilization strategies,” and “party–voter linkages.” The three terms refer to parties’ different types of attempts to mobilize voters.



Segmented Electoral Appeals

23

framework has inspired a substantial body of work analyzing both programmatic and non-programmatic electoral appeals. A limitation of the party–voter linkage framework, as it has been formulated and applied in such studies, is that it has impeded detection of segmented linkage strategies. Linkage segmentation can be structured along territorial or socioeconomic divides, and involves the pursuit of not only programmatic but also non-programmatic (e.g., charismatic, party identity, or particularistic) mobilization strategies. There are thus important challenges that need to be tackled in order to more realistically describe party–voter linkage configurations, and more accurately analyze their causal relations to other variables. Types of Party–Voter Linkages: A Minimal Definition In the course of my fieldwork I  observed many types of linkages. The vast majority of linkages I  observed are difficult to characterize using the usual categories derived from the party–voter linkage framework (i.e., clientelistic, personalistic, or programmatic).3 The great majority of the empirical instances I describe fall within a grey zone with respect to currently available conceptualizations of party–voter linkages. Yet, some set of clear conceptual demarcations is needed for guiding theoretical and empirical work. Therefore, in this section I present a set of minimal conceptual demarcations, adopting an initially “agnostic” view as to the positive nature of the linkage types I identify. By doing so, I aim to avoid two of the main pitfalls of much of the existing literature in this area: introducing overly rigid conceptual demarcations that risk hindering the analysis itself or, alternatively, drawing on excessively stretched conceptualizations.4 Whereas the pitfalls of   3 See Piattoni (2001), Stokes (2009), Kitschelt and Wilkinson (2007), Lyne (2008), Díaz-Cayeros et al. (n.d.), and Luna et al. (n.d.) for different attempts at developing linkage-type taxonomies.   4  The party–voter linkage framework suffers from definitional problems caused by conceptual stretching. First, criteria for applying the labels “programmatic” and “non-programmatic” are highly variable, and these categories are not clearly demarcated. For instance, while some would see the implementation of pro-union legislation by a leftist party with a significant labor base as the fulfillment of its programmatic platform, others would classify this as a clientelistic pact, through delivery of a club good. Still others would characterize this type of linkage as one typical of interest-group politics. Meanwhile, a candidate who gives away TV sets door-to-door but lacks a device for monitoring and punishing voters who then do not turn out for her might be classified by some as pursuing clientelistic linkages. But given the lack of coercion and monitoring, others would consider this not clientelism, but just regular advertising (Stokes 2005). This conceptual stretching is matched by different normative views on “clientelism,” “pork-barreling,” “vote-buying,” “constituency service,” and “programmatic politics.” In the American politics

24

Segmented Representation

conceptual stretching are familiar (Sartori 1970; Collier and Levitsky 1997), working with an excessively rigid conceptualization is also problematic.5 While useful in other contexts, applying a stringent conceptualization might distract from the analytical work I seek to advance in this book. When seeking to classify cases using categories with boundaries that are overly rigid, analysts are either forced to provide very stylized (and usually unrealistic) case descriptions, or to invest a great deal of effort in proving that a given case empirically fulfills all the necessary conditions for full membership in the conceptual category (e.g., for every instance classified as clientelism, one should provide evidence that there does exist an enforced contingent exchange between candidates, their brokers, and their voters). In the context of this book, I  am more interested in observing how parties segment and combine different types of linkages than in achieving precise conceptual and empirical demarcations of specific electoral appeals. Of course, the first objective is not feasible without having at least some criteria for distinguishing among different types of appeals. In sum, I prefer to work with somewhat fuzzier definitions of linkages that can still hopefully guide the analysis and provide some theoretical demarcation. This is also convenient because I am not interested in exploring only one type of linkage and demarcating it from others, but in mapping how parties combine largely different types of appeals to rally different social segments. I will work on the basis of a two-dimensional typology. First, I  consider the type of linkage that is implemented. Second, I distinguish between linkages implemented by individual candidates from those deployed by parties as collective institutions. Regarding the first dimension, I draw on previous work by Kitschelt (2000), Kitschelt and Wilkinson (2007), and Stokes (2009). On this basis, I classify linkages or appeals in terms of the type of resources on which they are based, and identify two basic types of resources used for literature, activities of incumbent candidates such as pork-barreling and constituency service are thought to provide stability and to facilitate policymaking in the congressional arena (e.g., Fenno 1978; Evans 2004). Yet, when observed in the developing world, the same types of activities are often considered inimical to modern politics (e.g., Stokes 2005; Schaffer 2007).   5  Consider for instance Susan Stokes’s (2005) conceptualization of clientelism as a mobilization strategy that necessarily requires monitoring and enforcement of the clientelistic deal. While allowing more elegant theorizing, assuming that clientelism necessarily requires monitoring excludes “softer” but more extensive and empirically relevant instances of the phenomenon. Much of Javier Auyero’s (2001) praised description of poor people’s politics in Argentina, which his analysis frames as clientelism, would be excluded from Stokes’s conceptualization. Similarly, much of Ostiguy’s (2009) analysis of the “programmatic orientation” of the Argentinean electorate and its leaders would not fit the conceptualization of programmatic linkages put forth by Kitschelt et al. (2010) and Luna et al. (n.d.).



Segmented Electoral Appeals

25

political mobilization: material and symbolic. Also in line with the literature, I then distinguish between material exchanges based on the provision of public goods and those based on the provision of private goods. Candidates can thus link to their voters by providing or promising them access to excludable benefits (private goods) or through general policies distributing resources among large groups of citizens. Parties use not only contingent and strategic exchanges, but also retrospective and non-strategic types of linkages. In the party–voter linkage framework, linkages are theoretically conceived of as the outcome of contingent (that is, on the spot) strategic interactions between candidates and voters. This view is particularly clear in contemporary analyses of clientelistic linkages (e.g., Stokes 2005; Dixit and Londregan 1996; Kitschelt and Wilkinson 2007). Analyzing party–voter linkages this way has important advantages, but also limitations. On the one hand, centering analyses on strategic interactions between two rational individuals (or between a candidate and a small group) provides more opportunities for formal modeling, yielding more elegant theories. Such theories are better able to identify precise causal mechanisms, which are in turn more easily translated into empirically testable implications. On the other hand, such theories can be quite simplistic. Voter agency has frequently been neglected in mainstream analyses of clientelistic interactions. Such analyses have focused almost exclusively on parties’ or candidates’ strategies, overlooking voters’ own strategizing (see Lyne 2008 and Nichter 2008 as exceptions). Furthermore, party–voter linkage theories centered on strategic exchanges depend on unrealistic assumptions, and neglect factors such as the “irrational” attachment of a “client” to her “patron.” Yet such attachment is frequently described in ethnographic accounts of clientelistic politics (e.g., Auyero 2001). Nor can such theories account for candidates’ willingness to disburse clientelistic side-payments in the absence of monitoring and coercion. This also reflects the centrality of “machine politics” systems (such as Argentina and Mexico) as reference cases underpinning formal theorizing on clientelism. The assumed contingency of the strategic interaction is another unrealistic assumption embedded in this framework. Most of the theoretical literature on party–voter linkages (both for clientelism and programmatic voting) sees the exchange of “favors” and “programmatic stances” for votes as contingent in the short run. That is, a party or candidate should continuously (in every election) “buy” the vote of its clients or commit to a programmatic platform that represents the views and interests of its constituency in order to secure its electoral base. However, especially if parties have a long-standing relationship with a segment of the electorate, at least some voters or constituencies will respond to either programmatic or clientelistic exchanges crafted in the past, thereby remaining loyal. For instance, the core constituencies of the

26

Segmented Representation

Frente Amplio (FA) in Uruguay and of the Unión Demócrata Independiente (UDI) in Chile have strong partisan loyalties that were crafted back in the 1970s (FA) and 1980s (UDI). Uruguay’s traditional parties were also able to profit from past clientelistic transactions long after their capacity to efficiently deliver side-payments started to decline. Similarly, the core constituency of the center-left in Chile still adheres to the Concertación thanks to its programmatic stance (and actions) during the transition to democracy. In sum, parties have reputations, which may build upon their past history of programmatic, non-programmatic, and leadership-based appeals. Sometimes these partisan reputations can cement strong identifications anchored in the past. Party identification is also fostered by citizens’ socialization during violent or divisive conflicts. When such identifications exist, the electoral allegiance of a potentially sizeable segment of the party’s electoral coalition becomes inelastic in the face of short-term campaign dynamics. Moreover, although usually obscured in contemporary analyses of clientelism focused on contingent exchanges and election-day mobilization, inter-temporal dynamics are also crucial in “real world” clientelistic relations (Auyero 2001). Parties with long traditions in the party system can therefore rely on their “usable pasts” (Grzymala-Busse 2002)  for mobilizing at least a segment of their electoral constituency. Those symbolic resources and collective identities are also pivotal for mobilizing party activists. At the same time, they provide coherence and unity to partisan organizations (Rosenblatt 2013). Granted, the retrospective nature of some linkages might not apply to the analysis of new political parties or, at the systemic level, to the analysis of more fluid party systems where partisan loyalties are weak or non-existent. Nonetheless, the contingency of party–voter linkages assumed in the most frequent applications of the party–voter linkage framework is theoretically problematic. Parties per se can mobilize electoral constituencies on the basis of not only contingent but also non-contingent mobilization strategies. The relative weight of these alternative mobilization strategies is theoretically consequential, and cannot be presumed a priori in analyzing linkages between parties and voters. This yields three types of resources deployed in the crafting of electoral appeals: provision of public goods (i.e., public policies), linkages based on the distribution of excludable or private goods (i.e., handouts), and symbolic appeals. Each such type of appeal may be structured either at the party level or around an individual candidacy. The capacity to structure programmatic or clientelistic linkages has an elective affinity with strong parties, whereas charismatic linkages are usually structured by individual candidates. These tendencies notwithstanding, linkages based on the provision of public goods, private goods, or the mobilization of symbolic resources can theoretically be structured by either individual



Segmented Electoral Appeals

27

Table 1.1.  Linkage ideal types Resources relevant for the appeal Relevant agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate Party

Candidate traits Partisan identity

Electoral platform Programmatic

Particularistic Clientelistic

Source: Author’s own construction.

candidates or by political parties. Elective affinities might also be assumed to exist between district-level politics and personalized linkages, as well as between national-level politics and party-based appeals. Yet, there are plenty of examples in which local candidates compete on the basis of linkages crafted around their parties’ traits, and not on their individual characteristics. These are linkages in which the party continues to be the relevant unit for crafting the appeal. Other times, national-level candidates (e.g., Joaquín Lavín or Michelle Bachelet in Chile) have competed in presidential elections by seeking to detach themselves from their respective parties. In short, the partisan or candidate-based nature of linkages should not be presumed ex ante; rather, it must be determined empirically. Jointly considering the type of resource used in the appeal and the agent who seeks to structure it, as shown in Table 1.1, I identify six ideal types of linkages: programmatic, clientelistic, party-identification, candidate platform, particularistic, and candidate traits. Whereas political parties as collective organizations are central to the pursuit of the first three of these linkages, the remaining three types are based on the qualities of individual candidacies. In this regard, the three party-based linkages in the bottom row of Table 1.1 share a common and important trait. Even if on the ground particular individuals craft these linkages when appealing to a given constituency, these three types of linkages involve the deployment of collective resources that are not available to individual candidates running on the basis of candidate-based linkages identified in the row above. When implementing party-based linkages (programmatic, clientelistic, and party-identification appeals), party membership becomes an important and valuable resource. To illustrate the point, consider first programmatic linkages. I define “programmatic” linkages and appeals minimalistically, as those crafted on the basis of general ideological stances or platforms that the party or candidate primes during electoral campaigns. These appeals are crafted around positions on general issues, or regarding the provision of a given set of public goods. When programmatic linkages predominate, parties compete by publicizing

28

Segmented Representation

their programmatic platforms (and implementing them while governing), and voters select parties in terms of their programmatic proximity to their own policy preferences. To make this type of linkage feasible, programmatic party organizations should select, socialize, and bring together office-seeking politicians with similar programmatic agendas, which are then enforced through organizational mechanisms structuring how they can pursue their careers effectively within the party. In turn, programmatic cohesion facilitates ideologically consistent policy-making, which is needed to reinforce the party’s programmatic credibility over time. Programmatic cohesion thus reinforces the ideological content of the “party label,” which becomes valuable for structuring linkages to voters (Kitschelt et al. 2010). Party labels then simplify complex issue spaces, facilitating voters’ decisions (Aldrich 1995; Hinnich and Munger 1994). Let us now consider clientelistic linkages. An efficient clientelistic party machine both extracts resources (from either the state or private sector) and deploys them on individual voters or constituency groups. This type of linkage works around a hierarchical (and usually center-periphery) network of activists, through which material resources are first obtained and then distributed through a top-down process, while votes are delivered from the bottom up. Partisan networks and grassroots activists provide those on top with crucial information for deciding which individuals or constituency groups should be targeted and which partisan activists should be privileged when distributing clientelistic handouts (Calvo and Murillo 2007; Stokes 2005). In turn, those on top control access to material resources available to the party as a collective organization, not to individual candidates. Finally, consider the case of a party that draws on voters’ partisan identities (henceforth, PID) to craft its electoral appeal. As in the case of programmatic appeals, candidates sponsored by this type of party have an interest in being closely associated with the party label (to which voters are attached). In turn, the party also has an interest in fostering its particular identity and reproducing it over time (Hartlyn 1988). Parties should therefore invest resources in socializing their members and voters into their particular partisan subcultures. This socialization then provides a crucial electoral advantage to candidates perceived as closely representative of the party’s identity. Citizens who strongly identify with a political party cast their votes in response to their affective relation to the party (Campbell et al. 1980), placing less emphasis on candidates’ contingent mobilization efforts (whether clientelistic or programmatic). By contrast, when pursuing linkages based on candidate traits, electoral platforms, and particularistic goods, candidates do not need to rely on partisan resources. Moreover, in contexts of mass disaffection with established



Segmented Electoral Appeals

29

parties, those candidates might profit from the anti-partisan mood of voters. Like parties, individual candidates can appeal to voters on the basis of different types of material and symbolic resources. When they distribute private goods and handouts (either privately or state funded), to which they have individualized access, candidates appeal to voters on the basis of particularistic linkages. Where they promise, through success of their individual candidacies, to deliver public goods, candidate platforms are at the crux of the linkage with voters. In general, with the exception of small districts or candidates with access to massive resources for funding a top-down organizational infrastructure, candidates lack the capacity to target the allocation of handouts with an eye to boosting the electoral return of the material resources invested. These limitations set particularistic linkages apart from clientelistic mobilization by political parties. If campaigning on the basis of their position on issues or on their ideological profiles, candidates pursue candidate-platform appeals. Yet candidates running as individuals cannot rely on partisan reputations to enhance their credibility with voters. In this regard, the ability of these independent and/or outwardly non-partisan candidates to compete on programmatic platforms involving promises regarding the provision of specific public goods is more restricted than that of political parties. Individual candidates can also compete on the basis of symbolic resources. In this case, appeals are crafted around candidate traits, which can range from a candidate’s charisma or empathy in relating to voters to his perceived capacity for governing or implementing a set of universally preferred policies. Name recognition, as well as socially valued professions (such as medical doctors or veterinary ones in rural communities) might also become powerful candidate traits. Party–Voter Linkages and the Assumption of Trade-Offs In his seminal 2000 article, Kitschelt discusses three overarching party–voter linkage strategies:  programmatic, charismatic, and personality-based linkages (Kitschelt 2000). While Kitschelt theoretically entertains and explores the possibility of linkage segmentation, he considers it unlikely. According to Kitschelt, parties can only successfully pursue one type of linkage at a time: [T]‌he incompatibilities between charismatic, clientelist, and programmatic linkages are not absolute. At low doses, all linkage mechanisms may be compatible. As politicians intensify their cultivation of a particular type of linkage, however, they reach a production possibility frontier at which further intensifications of one linkage

30

Segmented Representation

mechanism can occur only at the expense of toning down other linkage mechanisms. (Kitschelt 2000, p. 855)

In theory, this conclusion appears very compelling. And indeed, it often does seem to hold when non-programmatic linkages are financed through state resources that are transformed into subsidies, clientelistic side-payments, or “pork-barrel spending.” A party cannot sustain high levels of programmatic linkages with constituents (based on the provision of collective goods such as economic welfare or efficient regulation) while simultaneously structuring clientelistic linkages financed through state resources. For incumbent parties, pursing clientelistic and programmatic linkages will ultimately lead to fiscal imbalances or corruption, eroding governing parties’ capacity to provide public goods (Weitz-Shapiro 2012). There is also good reason to expect trade-offs between linkages based on charisma and more programmatic ones. Features of policy-making under “charismatic” leaderships can also erode the provision of public goods by implementing “populist” public policies that lack long-term viability and end up alienating middle-class voters. Recent analyses on the policy-making dynamic found in “populist left” cases in contemporary Latin America resemble this type of scenario (Weyland et al. 2010). Other trade-offs among different linkage ideal types contained in Table 1.1 also seem quite plausible in theory. For instance, parties that link to constituents on the basis of strong partisan reputations might be seen as unable to simultaneously link to voters on the basis of their candidate traits. In practice, however, the feasibility and empirical frequency of linkage segmentation nonetheless defies the assumption of a generalized trade-off when it comes to politicians’ ability to pursue linkages of different types. Indeed, successful candidates often do combine different linkage strategies in their campaigns, while minimizing the theoretical trade-offs attributed to linkage segmentation (see Valenzuela 1977; Przeworski and Sprague 1986; Gibson 1996; Gibson 1997; Coppedge 1998; Levitsky 2003; Magaloni et al. 2007; Ostiguy 2009; Taylor-Robinson 2010; Morgan 2011; Kitschelt and Singer 2011). One reason that linkage segmentation is often more viable in reality than theory would have it, is that the nature and role of private financial contributions to campaigns has typically been oversimplified and underestimated. To be sure, Kitschelt (2000) does recognize the role of private sector donations in campaign finance, and explicitly theorizes a dual-representation strategy in which resource-rich, vote-poor donors contribute to a party that seeks to get the vote of resource-poor, vote-rich voters. However, his framework systematically links donations from asset-rich private contributors to rent-seeking



Segmented Electoral Appeals

31

exchanges between business interests and the state (see also Lyne 2008). One possible explanation for this apparent oversimplification is that the bulk of theory and research dealing with clientelistic linkages is based on strategic analyses of incumbent parties or candidates. Ideally, however, the analysis of parties and party systems should not deal only with incumbent parties and leaders, but rather be able to compare parties across the political spectrum at the same point in time. This prevailing way of thinking about private sector donations also downplays the fact that they could be obtained in exchange for programmatic or interest representation, such as tax breaks or favorable regulatory legislation. Access to privately donated funds is thereby conceptually underestimated, making the trade-off between linkage types look stronger than it really is. Private funding could also be provided by illegal organizations seeking judicial immunity. Electoral funding provided by mafia organizations certainly might undermine the provision of some public goods by eroding transparency, justice, and public security. However, like private contributions from other wealthy social interest groups, it might not immediately translate into the same type of trade-offs envisaged when state-financed linkages are presumed. The role of private campaign financing tends to be especially important when analyzing societies where wealthy and powerful non-state actors are able to finance party–voter linkages, thereby providing parties and candidates with resources enabling them to provide material patronage to supporters from other social groups or geographical areas without thereby directly distorting state budgets. In fact, it is my contention that the levels of inequality, social fragmentation, and territorial segmentation observed in many contemporary societies make linkage segmentation not only feasible, but often imperative. In fact, in unequal societies, exclusive reliance on just one type of linkage (programmatic, clientelistic, or charismatic) can itself lead to trade-offs, especially when socioeconomically different constituencies are targeted. Imagine an exclusively programmatic party that has two target constituencies with different policy preferences and thus potential conflicts of interest. If the party uses different programmatic appeals to reach each constituency (therefore avoiding non-programmatic linkages), it runs the risk of a strong electoral trade-off. Such a trade-off is not the result of linkage segmentation. On the contrary, it is reliance on only one type of linkage (albeit targeted to socially diverse constituencies) that becomes electorally disadvantageous due to the salience of distributive conflicts within the party’s heterogeneous electoral constituencies. In sum, the potential trade-offs resulting from segmented strategies combining different types of linkages (e.g., programmatic and clientelistic) are variable. Depending on the context, some combinations may be more feasible

32

Segmented Representation

than others (Weitz-Shapiro 2012). The trade-off assumption therefore needs to be empirically assessed, and the presence of trade-offs should be treated as a variable, not assumed as a universal constant. One must thereby begin by analyzing how specific partisan trajectories and material or symbolic resources have enabled or disabled parties’ pursuit of segmented linkage strategies in different contexts. My framework focuses attention to two complementary devices through which parties can avoid trade-offs. First, parties can segment their linkage appeals along socioeconomic and/or territorial lines. Especially in highly unequal societies, the overlap between socioeconomic and territorial segregation helps make segmentation less visible to voters, reducing the electoral risks of combining seemingly “incompatible” forms of appeals. Second, at least some parties have access to symbolic and material resources that help in centrally harmonizing segmented linkage strategies. Such harmonization, which I will refer to as “strategic harmonization,” also enables those parties to implement highly segmented strategies while further minimizing trade-offs.6 The Structure of Segmentation Linkage segmentation can be structured around various social divisions, which are more salient in some societies than others. In this book I  am interested in analyzing two particular social divisions: territorial and socioeconomic. I  propose to observe whether politicians segment their electoral appeals when targeting different geographic groups (territorial segmentation) and/or socioeconomic groups (socioeconomic segmentation). Socioeconomic segmentation could be considered a subset of functional segmentation, which in different societies could be structured around ethnicity, religion, or salient cultural identities. Table 1.2 identifies the four possible combinations of socioeconomic and territorial segmentation. Since I assume that in highly unequal societies socioeconomic segmentation is ubiquitous, in the context of this work I  focus only on the two possible combinations that occur when high socioeconomic segmentation is observed.7 Those two possible combinations relate to the interaction between high levels of socioeconomic segmentation and different   6  I use the term “harmonization” instead of “coordination” to avoid conceptual confusion regarding the notion of “electoral coordination” in the neo-institutionalist literature on parties and party systems.   7  Yet, in other contexts, the presence of territorial segmentation in the absence of socioeconomic segmentation might be of special interest.



Segmented Electoral Appeals

33

Table 1.2.  Structure of segmentation along territorial and socioeconomic divides Socioeconomic segmentation Territorial segmentation

Yes

No

Yes

Dual linkages

No

Mixed linkages

Irrelevant for highly unequal societies Trade-off: no segmentation

Source: Own construction.

patterns of territorial segmentation. Territorial segmentation will likely vary depending on the overlap between electoral districts and the levels of territorialized inequality observed in each case (Rodden 2010). In the first scenario, there is socioeconomic segmentation, but not territorial segmentation. Thus, in this scenario, parties segment their appeals to different socioeconomic groups, but implement homogeneous linkage strategies across geographic units (electoral districts). In the second scenario, socioeconomic and territorial segmentation are mutually reinforcing, generating differences across districts and across social groups. In the remainder of this book I will refer to the first scenario as one in which, at the district level, mixed linkages predominate. In turn, the second scenario is one in which dual linkages predominate, both within and across districts. How do dual and mixed strategies relate to inequality? Ceteris paribus, high inequality correlates with electoral districts that are more socially and spatially segregated, causing them to be more internally homogeneous and externally heterogeneous than those observed in less unequal societies. In contexts of high inequality, individual districts therefore have greater chances of becoming detached electoral “micro-cosmoses,” in which parties can implement segmented party–voter linkages along territorial and socioeconomic lines (dual linkages). Dual strategies are theorized as being those that generate the most regressive distributive impacts across districts and, assuming that spatial segregation correlates to greater inequality, socioeconomic groups. Imagine a party that has the ability to segment its linkage strategies, providing interest or programmatic representation to a given set of districts, while engaging in charismatic or other non-programmatic forms of mobilization in a different set of districts. Especially in cases where the party has the ability to strategically harmonize such a strategy and pursue it systematically, public policy will be significantly biased in favor of the set of districts in which programmatic or interest-based mobilization predominates.

34

Segmented Representation

Mixed strategies also have the potential to produce regressive distributive outcomes among groups that relate to parties on the basis of distinct types of linkages. Strategic harmonization across districts might allow a party to target different potential socioeconomic groups (i.e., middle classes versus the informal sector) in a segmented fashion. More regressive distribution could therefore also be observed under a mixed linkage scenario. Where that is the case, regressive distributive impacts will not run along territorial lines, but instead along strictly socioeconomic or interest group lines. Although potentially large, I expect that the regressive distributive effects of mixed strategies will be milder than those observed when territorial and socioeconomic segmentation reinforce each other. Parties pursuing mixed or dual strategies confront different challenges when it comes to avoiding trade-offs. On the one hand, parties pursuing dual-linkage strategies might have a more difficult time harmonizing different district-level appeals into a coherent overall party strategy. On the other hand, parties engaging in mixed linkage strategies have a more difficult time pursuing distinct appeals in the same district without being noticed. Different contextual conditions (e.g., district sizes), as well as partisan endowments, might help parties to overcome those challenges.

Party–Voter Linkage Types and Strategic Harmonization Parties can be understood as either unified actors with clearly defined goals and precisely identified “choices” and actions (see Downs 1957), or as complex organizations, formed by different individuals with distinct and sometimes competing preferences and goals (see Panebianco 1988; Kitschelt 1989; Kitschelt 1994; Levitsky 2003). Alternatively, parties could also be conceived of as epiphenomenal to the choices and preferences of individual actors in pursuit of specific interests (e.g., election and re-election). This latter conception also understands parties as complex and collectively created organizations (see Aldrich 1995). Yet that approach, which has become very influential in the analysis of American politics, focuses more on modeling the action of different clusters of individual actors (i.e., candidates, MPs) on the basis of their often narrowly conceived “interests” than on the analysis of parties as complex organizations. Since I  am only interested in observing parties in their pursuit of electoral goals, I deploy a simplified notion of what a party is. Notwithstanding, I  conceive of parties as collective and complex organizations that provide their members access to different sets of resources (from material benefits, to



Segmented Electoral Appeals

35

funding for electoral campaigns, to symbolic resources such as a party’s programmatic reputation or “epic”). In a given party system at a given time, individual parties can differ in terms of their access to such resources. This is contingent on their own characteristics and historical trajectories, and on the interaction between a party’s own features and the electoral context in which it competes. Given the importance of parties’ historical trajectories, new parties might lack access to critical resources. Yet, especially in the context of rising discontent with established parties, being a newcomer might in itself be a powerful competitive resource. There might also be a mismatch between parties’ resources and strategic needs. For instance, due to its particular trajectory in a party system, a party could command disproportionate access to a decentralized network of activists. This is typical of mass-based electoral parties that consolidated their organizations in the early and mid-twentieth century. Whereas such a network might remain an important resource for pursuing some types of linkages (e.g., clientelism, PID), it might have become less useful for the pursuit of other types of linkages that nowadays can be pursued at a distance and “by air” (e.g., programmatic, leader-based appeals). If a mass party pursues one of those types of appeals, its network at the grassroots is certainly not a key resource, and might even become a hindrance (e.g., fueling grassroots discontent with leaders who have “abandoned” their bases). Moreover, in contexts where individual candidates are allowed to raise funds for their electoral campaigns in a decentralized manner, the value of party-provided funds decreases. Their value further decreases in contexts where parties, as collective institutions, have less access to campaign financing. In sum, though valuable in principle, the electoral payoffs to be derived from party-provided resources are subject to significant contextual variation. Moreover, although not merely an externality of agents’ pursuit of their individual goals as conceived in Aldrich (1995), over time political parties benefit from (or become disadvantaged by) the predominant strategies that their candidates seek to pursue while trying to win elections. In short, depending on context, party resources may further some forms of electoral appeals more than others, and, ceteris paribus, some kinds of candidate appeals contribute more than others to strengthening parties as collective organizations over time. In fact, reliance on candidate-based appeals has precisely the opposite effect. Although different in nature, for the three types of linkage included in the lower row of Table  1.1 (clientelistic, programmatic, and PID), political parties provide crucial symbolic and material resources that are deployed in their candidates’ campaigns. Those resources contribute to maintaining a partisan reputation, and providing a party label that voters can use as a cognitive

36

Segmented Representation

shortcut for more efficiently linking themselves to individual candidates pursuing clientelistic, programmatic, or PID linkages. Independent of the type of party-centered linkage in question, when parties are the relevant actor for structuring the appeal, a candidate’s own position and personal characteristics have less value vis-á-vis those of her party, which serve as a “guarantee” of the candidate’s qualities. Those “qualities” could be the candidate’s programmatic credibility, her capacity to deliver excludable goods, or her own socialization and membership in the party’s subculture. While constraining the behavior of individual politicians, the three types of party-centered electoral appeal all compensate office-seekers for their commitment to the collective organization by providing them with access to valuable campaign resources. In turn, candidates’ dependence on those collective resources contributes to strengthening partisan organizations over time and furthering party leaders’ capacity to harmonize segmented electoral strategies. Candidate-centered appeals, on the other hand, (whether particularistic, candidate-trait, or candidate-platform), can be detrimental to harmonized electoral strategies. For instance, candidates running under the same party list can compete on the basis of different types of linkages. Moreover, the party’s candidates may deploy different appeals across districts without strategically harmonizing them. When this type of scenario consolidates, as in cases where political parties become unpopular, the party, as a source of collective goods for electoral mobilization, is devalued, especially where its symbolic resources are concerned. Moreover, when candidate-based appeals persist over time, the party as such becomes decreasingly important and powerful, eventually endangering its electoral efficiency and political survival. Additionally, electoral trade-offs are more likely to emerge due to inconsistencies or contradictions between appeals crafted by different candidates without strategic party-level coordination. Finally, a more complex scenario arises where parties pursue electoral strategies combining both party-centered and candidate-centered appeals. Here parties need to leverage the resources generated through continued, partial reliance on party-centered appeals, and use them efficiently to harmonize both party and candidate-centered appeals at the party level. Sometimes parties succeed in meeting this challenge; other times they do not. Thus, depending on the mix of appeals in play, three potential types of scenarios may arise, each involving different types and possibilities of strategic harmonization. These three scenarios are identified in Table 1.3. When only party-centered (programmatic, party-identification or clientelistic) appeals predominate, strategic harmonization is achieved without party leaders needing to resort to purposive short-term action. This is



Segmented Electoral Appeals

37

Table 1.3.  Structure of required linkage harmonization of electoral appeals Candidate-centered appeals are used Party-centered appeals are used

Yes

No

Yes

Harmonization via party leadership needed and feasible

No

Decentralized harmonization devices needed and difficult to develop

Self-harmonized (no purposive short-term action by party leadership is required)

Source: Author’s own construction.

a scenario of “self-harmonization”:  the three types of party-centered linkages require harmonization, and they also provide parties with valuable institutional and organizational resources to pursue it. To the contrary, when only candidate-centered (electoral platform, candidate traits, or particularistic) appeals are prevalent, the party’s electoral strategy is very difficult to harmonize. Each individual candidate competes on the basis of decentralized appeals, using their individual resources and incentives. In this case, a complex (and unusual) system of decentralized harmonization would be required to achieve party-level coherence. Finally, if a party pursues a mix of candidate and party-centered appeals, then strategic harmonization is more feasible. The use of party-centered appeals provides party leaders with valuable resources that can then be deployed in seeking to harmonize the electoral appeals of candidates running on candidate-centered linkages. Yet, harmonization requires party leaders to explicitly draw up a consistent electoral strategy and efficiently enforce its pursuit. This job is challenging, and parties in this situation may also fail to harmonize their strategies, thus eroding their value and efficiency over time. Examples 1.1, 1.2, and 1.3 illustrate the three alternative scenarios just identified. Example 1.1 presents two photographs taken at campaign activities of the Movimiento de Participación Popular (MPP) fraction of the FA during the 2009 elections. Both images clearly convey the “collective” (party-centered) character of the campaign. Even though personalized publicity for José Mujica’s presidential candidacy is present in the background, this display combines party fraction symbols: the three-colored banners in red, blue, and white represent the FA, whereas the single- (red) or double-colored banners (red and black) identify the MPP.8   8  Although black and white images are reproduced here, color photos are included in the online appendix. The online appendix is available at http://thedata.harvard.edu/dvn/dv/jpluna.

38

Segmented Representation

Example 1.1.  Party-based appeals Source: campaign observation in Uruguayan districts, 2009.

Example 1.2.  Candidate-based appeals Source: author’s campaign observation in Chilean districts.

Example  1.2, in turn, exhibits two illustrations of personalized, candidate-based appeals. Here, candidates draw on their individual characteristics and personal traits to attract voters, usually obscuring their identification with a political party. The first of those images (left panel) is a campaign flyer for a Partido Demócrata Cristiano (Chile) candidate competing for a congressional seat in 2005. Nowhere does the flyer indicate the candidate’s partisan affiliation. Indeed, the colored diamonds appearing in the upper-left corner are the symbol of Santiago de Chile’s metro system, which was extended to the district before the election. Implicitly, the candidate is cannily taking the



Segmented Electoral Appeals

39

Example 1.3.  Candidate-based appeals organized by the party (with potential harmonization) Source: author’s campaign observation in Chilean districts.

credit for that extension in his campaign flyer. The candidate also presents himself as a “man dedicated to service,” and highlights his profession (medical doctor) to encourage voters to support his candidacy. The second image included in Example 1.2 (right panel) is a campaign advertisement for the mayoral race of 2012. The candidate is seeking to establish name recognition by associating his last name (Carter) with a well-established industrial brand (Caterpillar). To do so, the banner emulates Caterpillar’s logo, font, and colors, while substituting the brand’s name for Carter. Party-centered appeals figure nowhere in his campaign publicity, which is also notably devoid of partisan symbols. Example  1.3 contains four pictures from a campaign also employing candidate-centered appeals like those figuring in Example 1.2. The campaign in Example 1.3, however, is implemented by the party (the UDI) as an organization. Whereas the party does not show up in the campaign material, the organization is pivotal in funding and implementing the presented strategy across districts. The first image included in Example 1.3 (upper-left corner) represents an UDI lower-chamber congressional candidate handing out

Segmented Representation

40

reading glasses to a woman. Along with the glasses, he includes a flyer with his campaign motto: “Pa que Leay bien” (“For you to read well”). His campaign motto draws on a happy coincidence: his last name (Leay) sounds, exactly like the Chilean slang for the verb “to read.” Like the first candidate in Example 1.2, a central objective of his advertising is to establish name recognition with the help of a mnemonic. The remaining images also show candidate-centered appeals organized by the UDI. The second image displays a mission organized by party activists and collaborators in which a group of optometrists prescribe glasses for shantytown dwellers in Santiago de Chile. The third and fourth images (bottom left and right corners) display two types of eyeglass cases handed out by other UDI congressional candidates (one for the lower chamber and one for the Senate) in different districts of the city. These two last photographs suggest that, despite being deployed in a personalized fashion at the district level, the strategy was deployed by the party as an organization. The UDI not only devised this strategy and centrally bought eyeglasses in China, but also provided the funds and the logistics to implement it in several working-class districts.9 Examples  1.1 through 1.3 illustrate differences in parties’ capacities and methods for harmonizing segmented strategies. Example 1.1 represents a situation in which the party and its collective organization (structured around fractions) figure very visibly and are pivotal for implementing the strategy. In this type of scenario, candidates are also significantly constrained by the organization and its symbolic resources. Examples 1.2 and 1.3 represent situations in which the party is not pivotal for crafting party–voter linkages. Yet, in Example 1.3 the strategy nonetheless relies on the party’s capacity to provide the resources and logistics needed for successfully deploying the candidate-based appeals. The party’s ability to do so might in turn provide its leaders with efficient levers for constraining the behavior of its individual candidates, even if the candidates compete in a decentralized way without using the party’s symbols or overtly drawing on its material resources. In Example 1.2, the party as a collective institution is much less able to constrain the behavior of the candidates it formally endorses. In sum, in situations like the ones represented by Example 1.1, self-harmonization is more likely. In situations like those in Example 1.2, decentralized harmonization devices are both needed and very difficult to develop. Finally, in situations like those in Example 1.3, harmonization is achieved through the party leadership.  

9 

Based on interviews with party activists in 2011.



Segmented Electoral Appeals

41

To be sure, as treated here, strategic harmonization should not be confounded with the notion of strategic coordination, which deals with different types of party-system outcomes regarding the coordination of supply (strategic entry) and demand (strategic voting) under different electoral formulas (Duverger 1951; Cox 1997). Strategic harmonization is also conceptually different from party-system nationalization (see, e.g., Chhibber and Kollman 2004; Caramani 2006; Mainwaring and Jones 2003; Morgenstern and Potthoff 2005)  and vertical integration (Hicken 2009). Party-system nationalization and vertical integration refer to the evenness in the geographical distribution of electoral support for the party. Instead, strategic harmonization, as conceived here, relates to the party’s capacity to implement segmented linkages across constituencies in ways that reduce possible trade-offs and enhance synergies among distinct mobilization attempts. Those strategies should exploit the party’s competitive advantages and target specific constituencies, which may be more or less evenly distributed across districts. If the key electoral constituencies of a party implementing a segmented and coordinated strategy were territorially concentrated in a relatively small set of districts, such a party could both display low levels of nationalization (or vertical integration) and high levels of strategic harmonization. Segmentation and Harmonization of Electoral Appeals My framework for analyzing segmented party–voter linkages identifies five ideal-type outcomes. These outcomes are defined by both the structure of segmentation and the levels of strategic harmonization observed in each case. Four of those ideal types correspond to situations in which political parties pursue segmented linkage strategies. The fifth type, which I call “pure linkages,” represents a situation in which the party pursues only one type of linkage throughout all districts. This pure linkage scenario is consistent with the trade-off assumption built into the original party–voter linkage literature. Figure 1.1, which considers three districts, graphically illustrates a pure linkage scenario. As observed in the figure, in each of the three districts the same type of linkage (as represented by a given pattern) is pursued, and logically coincides with that of the national party. This pure linkage scenario, like the others described in the rest of this section, is compatible with different types of party–voter linkages. For instance, it can represent a party relying on programmatic linkages and thus able to harmonize electoral mobilization efforts and achieve partisan discipline (see Aldrich 1995). However, the figure could also illustrate the case of a clientelistic party achieving strategic harmonization by putting

Segmented Representation

42

PARTY

District 1

District 2

District 3

Figure 1.1.  Pure strategy

together a hierarchical organization. Or, Figure 1.1 could illustrate a party drawing on charismatic mobilization consistently across districts, or a party with strong symbolic ties to electoral segments (distributed in different districts) that reproduces electoral support by consistently drawing on party identifications. The remaining four strategic scenarios apply to parties simultaneously using more than one type of linkage. They are distinguished by whether: (a) linkage strategies are harmonized into a consistent national party strategy, and (b) linkage segmentation occurs primarily between or within districts. When the national party is able strategically to harmonize segmented linkage strategies into a consistent overall strategy, I classify the scenario as “harmonized.” Where parties are weaker than local candidates and therefore lack resources to pursue a strategic harmonization of segmented strategies, linkage strategies are not harmonized by default. My framework thereby explicitly considers the extent to which linkage strategies are harmonized across districts, distinguishing between scenarios characterized by greater or lesser strategic harmonization. Following the language introduced in Table 1.2, when segmentation occurs between districts I classify the strategy as “dual.” When segmentation instead occurs within districts, I classify the strategy as “mixed.” Although in highly unequal and geographically segregated societies territorial and socioeconomic divisions do tend to overlap, distinguishing between dual and mixed strategies allows my framework to distinguish socioeconomic segmentation patterns that are reinforced through geographic segmentation (dual segmentation) from patterns of socioeconomic segmentation that are relatively tamed by the lack of territorial segmentation Figure 1.2 represents a “harmonized dual strategy” scenario. Here, a party pursues a dual strategy (thus implementing different types of linkages in



Segmented Electoral Appeals

43

PARTY

District 1

District 2

District 3

Figure 1.2.  Harmonized dual strategy

PARTY

District 1

District 2

District 3

Figure 1.3.  Non-harmonized dual strategy

different districts), which is harmonized into an overarching partisan strategy. Different box patterns represent distinct linkage types. Figure 1.3, in turn, represents a “non-harmonized dual strategy” in which district-level segmentation is observed, but in which the party lacks the capacity to strategically harmonize those segmented strategies into a party-level strategy. In other words, the dual strategy is pursued in a decentralized fashion, with the party leadership unable to harmonize it strategically. The remaining two possible outcomes are defined by the presence of mixed strategies, in which parties or candidates pursue different appeals for targeting different segments within a district. Figure  1.4 represents the “harmonized mixed” scenario, in which the party is able to deploy multiple strategies within a district. Although this outcome is equivalent at the aggregate level to the one observed in Figure  1.1, its district-level composition and the challenge it presents to the party are different. The main challenge the party faces in achieving a “harmonized dual” outcome is strategically harmonizing different

Segmented Representation

44

PARTY

District 1

District 2

District 3

Figure 1.4.  Harmonized mixed strategy

PARTY

District 1

District 2

District 3

Figure 1.5.  Non-harmonized mixed strategy

strategies across districts. By contrast, for achieving a “harmonized mixed” outcome, the party’s main challenge is avoiding trade-offs when implementing segmented strategies within districts. Available evidence suggests that this challenge might be more daunting than that of harmonizing linkages across districts (e.g., Weitz-Shapiro 2012). Finally, Figure 1.5 illustrates a “non-harmonized mixed” scenario, in which segmented linkages within districts are not strategically harmonized into a consistent national-level strategy. The outcome is a decentralized strategy in which the party, at the aggregate level, lacks the capacity to enforce strategic harmonization. In such a scenario, as well as in non-harmonized dual scenarios, the party label gradually becomes less valuable for pursuing different types of appeals. In Chapters 2 through 5, I apply this framework to party–voter linkages in Chile and Uruguay. The evidence presented in the next four chapters illustrates the advantages of my framework for describing party–voter linkages at



Segmented Electoral Appeals

45

the district level, as well as their translation into distinct patterns of electoral competition at the national level. In the remainder of this chapter I discuss the operationalization strategy I apply to map segmented electoral appeals in Chile and Uruguay. I also explore the possible causal triggers of the empirical configurations I seek to describe in the remainder of Part I.

O P E R AT IO NA L I Z AT IO N A N D I D E N T I F IC AT IO N ST R AT E G Y:   M A P P I N G SE G M E N T E D L I N KAG E S Level of Analysis: System or Individual Parties? The party–voter linkage framework lacks precision when it comes to the unit of analysis. Interestingly, studies focusing on non-programmatic linkages usually analyze a single party (cf. Auyero 2001; Stokes 2005; Díaz-Cayeros et al. n.d. on clientelism; Hawkins 2010 on charismatic mobilization). And when comparative analyses of non-programmatic linkages are pursued, comparisons are usually drawn across party systems, not within them. Therefore, the very possibility that not all parties within a given system implement the same type of linkage strategy is overlooked by the research design of such studies. In turn, analyses looking at programmatic linkages tend to be structured at the party system level. This is the case, in part, because programmatic competition is assumed to require differentiation of platforms between at least two parties (cf. Kitschelt et al. 1999; Kitschelt et al. 2010; Luna et al. n.d.). In this literature, systemic comparisons are the norm. The specific focus of each literature coincides with expectations regarding the possibility of linkage-type contagion. Since Duverger famously introduced the idea of “contagion from the left,” many have subscribed to the hypothesis of systemic contagion (Duverger 1951; Roberts 2002; Boas 2010). According to this idea, successful strategies are benchmarked and diffused across the system, leading to a systemic outcome in which partisan strategies converge, producing a unified systemic outcome. However, recent research entertains the hypothesis of divergence. In clientelistic systems, one party often seems better able to establish and hegemonize the dense social network needed for enforcing clientelistic transactions (Calvo and Murillo n.d.). In contrast, given their lack of access to material resources, externally mobilized parties have been thought more inclined to pursue programmatic strategies (Shefter 1994; Hunter 2010). In short, the emerging literature on party–voter linkages lacks explicit agreement on whether parties or party systems are the appropriate unit of analysis. This lack of agreement regarding the unit of

46

Segmented Representation

analysis, coupled with the trade-off assumption, is responsible for producing incomplete assessments of the type of linkage strategies often implemented by political parties. Those assessments neglect the presence of linkage segmentation and its causal analysis. The framework I propose in this work simultaneously describes linkage patterns at both the systemic and individual party levels. This two-level analysis is necessary because I expect institutional, historical, and given socio-structural configurations (i.e., patterns of social inequality, territorial or interest group fragmentation, etc.) at the national and district levels to trigger converging partisan strategies for pursuing specific kinds and combinations of linkages. I  also expect electorally successful linkage strategies to disseminate, yielding systemic convergence (Boas 2010). Yet, individual parties carry different symbolic, material, and organizational endowments. Those differential endowments inhibit rapid and complete convergence, inducing continued divergence at the individual party level. Moreover, parties target different constituencies according to which distinct types of linkages might best pay off. In addition, political parties also have different ideological preferences regarding the types of linkages they should implement. As a result, at a given time, parties may differ in their capacity to structure particular kinds of party–voter linkage strategies. In sum, where parties’ strategies are not all identical, neither systemic nor individual-level analysis alone is sufficient. The two need to be jointly considered and empirically compared across time to sort out whether contagion or divergence is the dominant tendency. I thus propose to treat the “systemic” level as one that, along with structural, historical, and institutional variables, configures a given opportunity structure within which individual parties compete. Parties seek to adapt to such an opportunity structure by investing a set of available symbolic and material resources in the deployment and harmonization of segmented linkage strategies. Yet, at a given moment (i.e., when confronting a specific opportunity structure) individual parties are differentially endowed for adapting successfully. Over time, this explains why certain parties consolidate and grow in the system, pursuing successful partisan adaptation processes, while others decline. On this basis, my framework is compatible with a scenario of “convergent divergence,” in which systemic traits (i.e., contagion) are present, while divergent individual party adaptation is also feasible. Simultaneously observing the systemic and individual party levels is the only way to calibrate the relative importance of systemic convergence and individual parties’ divergence.



Segmented Electoral Appeals

47

Unit of Observation and Units of Analysis In Kitschelt’s (2000) original formulation, the party–voter linkage framework does not clearly indicate the level at which linkages should be observed. Should they be observed nationally, or at the district level? This imprecision has contributed to the proliferation of aggregate-level analyses, whose very method of inquiry implicitly neglects segmentation. Usually, observation of one type of linkage is taken as evidence of a lack of other types. For instance, non-programmatic linkages are presumed to be less prevalent in cases were aggregate cross-national analyses show greater levels of programmatic structure. Meanwhile, ethnographic and small-N evidence of clientelism is frequently used to infer that a party system is pervaded by non-programmatic linkages, and consequently lacks room for programmatic mobilization. As a result, available research on party–voter linkages tends to present two limitations: focusing almost exclusively on either the national or the local level, and relying almost exclusively on either cross-national survey analyses privileging institutional indicators or on qualitative case studies. Assuming that whatever form of linkage one has observed is and must then be the only kind of linkage that matters in a given case is consistent with the trade-off assumption. This leads to biased descriptive assessments, which ultimately obscure three crucial facts. First, different electoral cross-sections can link to parties on the basis of different types of linkages. Second, different parties within a given party system can pursue different electoral and organizational strategies depending on the competitive opportunities they face and their relative competitive endowments. Third, the same political party can use different types of linkages for mobilizing constituencies depending on its strategic situation. Biased descriptive assessments, in turn, provide an unreassuringly shaky basis on which to construct comprehensive theoretical analyses of the determinants and implications of different types of linkages between parties and their constituents. I argue that only parties good at both segmenting and harmonizing their linkage strategies can persist as successful electoral organizations in unequal societies. To demonstrate this claim, I first observe linkage strategies at both the national and district levels, analyzing whether the same party, competing in different local contexts and linking to different social segments, diversifies its linkage strategies territorially, and/or across socioeconomic groups.10 The units of observation at the district and municipal levels are the congressional campaigns of each party contesting elections in each district.11 My   10  It follows that I focus primarily on national political parties (i.e., those that seek to compete at the national level).   11  In my two primary cases (the UDI and the FA), district boundaries could be considered exogenous to partisan interests. In Uruguay, the district and municipal boundaries have not

48

Segmented Representation

use of a comprehensive district sample subject to a wide variety of contextual factors potentially shaping local-level party competition increases the leverage of my analysis. Having observed and codified each district-level campaign, I then characterize the party’s national-level strategy, identifying potential conflicts among different segmented strategies in the process. I examine how and whether each party meets the challenge of harmonizing the decentralized deployment of segmented electoral appeals. Recovering Historical Patterns of Segmentation and Harmonization (Independence to 1973) In order to set the baseline for assessing change in the post-transitional era, the first part of Chapter 2 describes the main features of the traditional party-systems of Chile and Uruguay. The chapter presents a stylized narrative of both party systems, classifying the types of linkages and structure of segmentation observed from independence until 1973.12 The narrative reveals that segmentation was ubiquitous during many historical periods, and that the political economy of those different periods shaped incentives for pursuing distinct linkage and segmentation strategies. My narrative is largely informed by secondary and historical sources. Observing Socioeconomic Segmentation at the Aggregate Level in the Contemporary Period The second part of Chapter 2, along with Chapters 3 and 4, describe contemporary linkage strategies in Chile and Uruguay. This analysis provides insights on the current distribution and equity of different types of party–voter linkages to citizen groups with different educational backgrounds in each society. My analysis in this section draws on two types of primary evidence: mass- and elite-level surveys, using semi-structured interviews and participant observation in both cases. Survey-based evidence is primarily used to describe linkage types and the structure of socioeconomic segmentation at the aggregate level.

changed since the early 1900s. In Chile, the last instance of gerrymandering was observed during the military regime, in preparation for the 1988 plebiscite (see Navia 2003).   12  In Chapter 2, both for the sake of space and due to data restrictions that preclude pursuing a more fine-grained analysis, I do not pursue analyses at the level of individual parties.



Segmented Electoral Appeals

49

In particular, my analysis focuses on seeking to assess the degree to which parties pursue segmented programmatic linkages across socioeconomic groups. Although I present a series of evidence on other types of linkages at the aggregate level, I devote significantly more space to programmatic linkages based on the promised distribution of collective goods than to linkages based on symbolic resources or the exchange of private goods. This decision derives from the need to empirically problematize the alleged centrality and cross-national equivalence of programmatic linkages in and across Chile and Uruguay (Kitschelt et al. 2010). Measuring and comparing the levels of programmatic structuration (i.e., the prevalence of programmatic linkages in a party system) is difficult and controversial (Luna et al. n.d. Kitschelt et al. n.d.). No established technique exists for doing so, and there are multiple data limitations (Powell Jr. 2004). My main analytical strategy is using a series of case-specific confirmatory factor analyses extracted to test the empirical configuration of three potential competitive divides:  regime, moral/religious values, and state/market. Such factors are then used to map the distribution of partisan groups and the internal coherence of partisan placements at both the elite and voter levels. In this respect, I am applying a methodology similar to that applied by, among others, Barnes and Kaase (1979), Inglehart (1984), Dalton (1988), Kitschelt (1994), Kitschelt et al. (1999), and Moreno (1999). Given my interest in the societal distribution of linkage strategies, I also present evidence of the internal ideological cohesiveness of partisan groups segmented on the basis of their educational levels, which I use as a proxy for socioeconomic status.13 Finally, I apply procrustes analysis, a variety of multidimensional scaling, to compute an issue congruence estimate for each case, cross-section, and time period.14 To be sure, this operationalization strategy is based on detecting traces of programmatic linkage types and segmentation strategies at the systemic level. This is an admittedly suboptimal strategy for at least three reasons. First, my measures only constitute a distant (yet, theoretically reasonable) proxy for the phenomena I seek to observe. The assumption behind this operationalization strategy is that when parties seek to pursue programmatic mobilization strategies, party voters and leaders would have more consistent and well-defined programmatic views on salient competitive dimensions. Second, my operationalization strategy is suboptimal because analyzing “traces” of programmatic linkages might also conflate cases in which programmatic mobilization

  13  See Chapter 4 for a description and justification of such an operationalization strategy.   14  The online appendix includes a non-technical introduction to procrustes analysis and its proposed application to estimate issue-congruence.

50

Segmented Representation

was strong in the past and continues to shape elite and citizen attitudes with cases that use programmatic mobilization strategies in the contemporary period. Third, it is important to stress and clarify at the outset that, at the citizen level, the evidence draws on responses by those who declared themselves to have voted for a given party. Especially where alienation of citizens from parties is high, the cross-section interviewed might not be representative of the voting population. Estimating the presence of other types of linkages is also complex. Whereas the strength of partisan identifications and the importance of candidate traits is conventionally estimated through mass-survey instruments, estimating the prevalence of clientelistic/particularistic exchanges through survey analysis is also problematic, at least when social-desirability biases have not been controlled through the questionnaire design. (For exceptions in this regard see González-Ocantos et al. 2011; see also Calvo and Murillo n.d.) Given these limitations, I seek to cross-validate survey-based evidence with evidence produced through semi-structured interviews and secondary sources available for each case. Observing Territorial and Socioeconomic Segmentation at the District Level Switching to the district level, Chapter 4 traces the recent evolution of local party–voter linkages. To do so, I rely on qualitative evidence gathered from fieldwork research on seven electoral districts and nineteen municipalities in Chile, and from seven Uruguayan municipalities, which are also senatorial and lower-chamber districts. Interviews were conducted in three rounds (2001, 2002, and 2003). An additional round of campaign observation and ten interviews was conducted between 2008 and 2011 in both countries to validate the narrative inferred from the previous three rounds of interviews. The sample of districts and their socioeconomic and political characteristics are described in Chapter 4. Due to the small size of Chilean districts, the sample for that country is largely urban and drawn from the metropolitan region (six of the seven districts and sixteen out of nineteen municipalities belong to Greater Santiago). Complementary research in three towns (Puerto Montt, Cochamó, and Río Puelo) of district 57 of the southern 10th region of the country was used to control for urban–rural and regional differences.15 Although I summarize the   15  Whereas Puerto Montt is a city and the regional capital, both Cochamó and Río Puelo are small rural towns near the Argentinean border.



Segmented Electoral Appeals

51

evidence obtained in those locations, the sample is not sufficiently large to make valid inferences about rural areas. That said, a second round of interviews pursued in 2008 and then again in 2010 and 2011 to survey the views of party leaders with political careers in different congressional districts largely confirms the party–voter linkage patterns I describe. The configuration of Chilean congressional districts, which are usually comprised by more than one municipality, adds an important complication to the analysis. In my characterization of district types, I identify two configurations that generally apply to all municipalities belonging to the same district. Within several of the districts I studied (23, 18, 27, 26) segmentation was low. Meanwhile, in the socioeconomically diverse municipalities within Districts 29, 57, and 24, I found different types of linkages. These cases thus increased the overall differentiation and heterogeneity of linkages observed within districts. My description of the internal dynamics of Districts 24, 29 and 57 seeks to account for these nuances. In turn, the Uruguayan sample included both urban and rural sectors, which were present within each of the districts I analyzed. Due to their large size and social heterogeneity, in the most populous districts I  focused my research on particular neighborhoods or towns in order to cover the widest possible set of contextual conditions, to which I refer in the characterization of each district.16 The overlap between districts and municipalities simplifies the description of local politics in Uruguay. Moreover, the scope of the sample allows for broader generalizations than in the Chilean case. My fieldwork was conducted during the harshest economic and financial crisis endured by Uruguay since the 1930s. This probably made social assistance activities devised to cope with unemployment and the social emergency more salient than in normal times. Thus, while the political dynamics I describe are structural, the centrality of linkages crafted around the distribution of social assistance might have been amplified in the evidence I report for this case. The evidence on which the district-level analysis is based was collected through semi-structured interviews with relevant political and societal figures in each district. My sample included the following target categories: (a) successful and unsuccessful candidates who had contested recent elections; (b) current and past legislators, mayors, and local council members, as well as salient political figures in the district; (c) leaders of community organizations; and (d)  key informants (party strategists and representatives from partisan think-tanks or congressional delegations, school teachers, priests, etc.). The use of interview material is oftentimes viewed as inadequate for causal inference, as it represents the revealed preferences of the interviewee and not  

16 

These neighborhoods and towns are also identified and briefly described in Chapter 4.

52

Segmented Representation

necessarily her honest opinion. While acknowledging this critique, I nonetheless make intensive use of interview material. I justify this decision on two grounds. First, the great majority of the interviews I cite in the text reflect interests and preferences that cannot be judged as “politically correct” (i.e., many political actors responded against their presumed true interest). In this regard, I think the evidence I obtained in my fieldwork reflects more than just revealed preferences. Second, but more crucially, I  draw on interview material to illustrate a descriptive argument, not a causal inference, constructed from analyses of primary and secondary sources. This descriptive argument is founded not only or primarily on the specific interview extracts included in the book, but rather on a narrative that progressively coalesced out of the many interviews I conducted for each argument and case. The emerging narrative was cross-checked both through subsequent rounds of interviews, analysis of available secondary sources, and my participant observation in each district. The extracts of these interviews selected for inclusion in the text should thus be regarded as an illustration of my narrative concerning each case and process. Passages of the interviews were obviously not selected randomly or comprehensively for inclusion. Their selection was made with two ends in mind: (a) to represent conflicting views or different political camps when distinct narratives emerged on a given issue, and (b) to select extracts that were both synthetic and eloquent regarding the narrative presented in the text. Observing Party Strategies In Chapter  5, I  offer a comparative analysis of the electoral strategies pursued by the most successful individual party in each system (the UDI in Chile and FA in Uruguay). I first analyze each party’s competitive trajectory and the composition of its electoral base, drawing on secondary sources and available historical narratives for each case. I then triangulate between the analysis based on these sources, my interview results, and survey data to illustrate the nature of each party’s electoral strategy and the different ways partisan elites pursue harmonization in each case. Alongside the “success” cases of the UDI and FA, I also analyze parties that have displayed declining capacity to segment and harmonize segmented electoral appeals in each party system. Observing Structures of Segmentation in Other Cases In Chapter 7, I seek to test the external validity of my descriptive inference in a set of cases outside my original sample. To do so, I pursue two types of



Segmented Electoral Appeals

53

analysis. First, I qualitatively analyze a set of four parties that, irrespective of other contextual conditions, operate in areas with high levels of social inequality. For each party, I seek to qualitatively assess the levels of linkage segmentation and harmonization observed. The analysis of these cases is based on available secondary sources. Second, I also seek to test my descriptive inference in a large-N sample. To do so, I draw on the results obtained by the Democratic Accountability and Linkages Project implemented in 2009 (hereafter, the expert survey), under the leadership of Herbert Kitschelt in eighty-five cases worldwide.17 Using those results, I construct indicators measuring the levels of segmentation and harmonization observed in each national case and across a sample of individual parties. I then use that indicator in quantitative and fuzzy-set analyses that seek to test the main observable implications of the theory derived from my small-N analysis of Chile and Uruguay to explain the evolution and current configuration of party–voter linkage strategies. Summary of the Operationalization Strategy for Measuring the Dependent Variable The operationalization strategy just outlined draws on the analysis of six types of data sources, most of which are available for some periods and dimensions, and not for others. Table 1.4 presents the overall operationalization strategy, identifying with a “1” the technique that provided the foundation for my coding of each case. While sources identified with a “2” acted as complementary ones, those identified with a “3” were used for cross-validation. As observed in the table, the use of secondary sources was vital for historically characterizing each system. Meanwhile, primary and secondary sources were critical for measuring the different dimensions of the dependent variable in the contemporary period.

C AU SE S O F PA RT Y– VO T E R L I N KAG E T Y P E S :   AVA I L A B L E A R G UM E N T S Part II of the book presents a series of empirically inferred propositions regarding the observed causal patterns in Chile and Uruguay. These propositions

  17  The Project is entitled “Democratic Accountability and Linkages Project” and was directed by Herbert Kitschelt at Duke University.

Segmented Representation

54

Table 1.4.  Data sources for measuring different dimensions of the dependent variable Secondary sources Historical Official Mass sources statistics surveys Historical narrative Type of linkage Socioeconomic segmentation Territorial segmentation Individual partisan strategy Out-of-sample cases

1

2

3

2

1 1

Primary sources

Elite Expert surveys surveys

1 1

2 2

3

2

1

3

1

3

Interviews and participant observation

3 3

1 2

3

1

3

1

1

Source: Author’s own construction Key   1  = Primary source for coding the case   2  = Secondary or complementary source for coding the case   3  = Source for cross-validating the coding of each case

are contextualized with an overview of existing theoretical arguments about party–voter linkage types. To be sure, those arguments, founded on the trade-off assumption, were crafted for explaining one type of linkage’s predominance over others, not to analyze the structure of segmentation of party– voter linkages. Notwithstanding, I  draw on these theoretical arguments for proposing a novel explanation of linkage segmentation and its structure in post-transitional Chile and Uruguay. Available explanations essentially deal with four types of linkages:  programmatic, clientelistic, party-identification, and candidate-based. As my own new scheme of classification clarifies, the last can actually take several different forms (candidate-platform, candidate-traits, or particularistic). Four bundles of theoretical factors feature prominently in explanatory accounts of party–voter linkage types. Table  1.5 presents a summary of those explanatory accounts. Whereas some of the featured explanatory arguments do not directly predict specific types of party–voter linkages, authors using them look at dependent variables that arguably have direct implications for the pursuit of some types of electoral appeals and for precluding the pursuit of others.18   18  For instance, while Carey and Shugart (1995) seek to predict the personalization of electoral races, and Harbers (2010) seeks to explain party-system nationalization, both arguments



Segmented Electoral Appeals

55

One subset of the electoral politics literature presents arguments on the effects of different electoral formulas on candidates’ incentives to cultivate “personal” versus “party-based” allegiances. For instance, Carey and Shugart (1995), Cox (1987), and Ames (2001) have variously identified electoral incentives such as district size, the electoral system, or the electoral list structure as leading to more or less personalized systems. More personalized systems, in turn, are seen as providing greater opportunities for the pursuit of candidate-based or clientelistic appeals. A different strand of institutional analyses looks at the effects of decentralization (fiscal or political) on partisan institutions. By territorially fragmenting and diffusing authority, decentralization provides greater opportunities for local actors such as mayors or governors to break free from national-level leadership and partisan organizations (Eaton 2004). Decentralization can thereby foster the denationalization of party systems (Harbers 2010), increasing the likelihood of candidate-based appeals while decreasing the likelihood of programmatic and PID appeals. Moreover, decentralization policies could also reduce the stock of material incentives available to national party leaders for maintaining partisan networks and keeping them alive and aligned. I recover these arguments in the explanation I propose. Second, several authors analyze the effects of particular historical factors on different types of linkages. Analyzing the nature of electoral volatility across a large-N sample of party systems in contemporary democracies and semi-democracies, Mainwaring and Zocco (2007) conclude that party systems that originated in earlier periods (e.g., before the expansion of mass media) display lower levels of electoral volatility. The authors argue that “older” party systems were able to craft more solid partisan identities, which were then transferred inter-generationally. Mainwaring and Zocco thus highlight the historical causation of at least some linkage types. In turn, Dalton and Weldon (2007), among others, have long argued that the high penetration of mass media and their influence in politics and campaigns have contributed to reducing the importance of partisan identification in contemporary party systems, thereby privileging candidate-centered campaigns. According to these authors, the importance of “air” campaigns (versus land-based ones) also reduces the importance of clientelistic exchanges in contemporary societies. A long tradition in the American politics literature seeks to explain partisan identifications. Whereas some authors argue that resilient

concern factors that could in turn very well affect the relative salience of candidate or party-based linkages in a given party system.

Table 1.5.  Summary of theoretical arguments on determinants of party–voter linkages Theoretical bundle

Authors

Electoral Cox (1987); institutions Carey and Shugart (1995) Eaton (2004); Harbers (2010) Historical Mainwaring and Zocco (2007) causes Shefter (1994) Keefer (2008) Kitschelt et al. (2010)

Structural

Mainwaring et al. (2006); Morgan (2011); Hawkins (2010); Fiorina (1981) Kitschelt and Wilkinson (2007) Stokes (2005); Lyne (2008); Taylor-Robinson (2010) Piattoni (2001) Kitschelt and Wilkinson (2007)

Individual level

Independent variable

Programmatic Clientelistic

PID

Electoral formulas that “personalize races” Decentralized political systems

Decrease

Increase

Decrease Increase

Decrease

Increase

Decrease Increase

Late timing of democratization and party origin Mass mobilization after bureaucratic development Long-term democratic experience Long-term democratic experience + early socioeconomic development + grievances over economic model Failed/“bankrupt” representation Socioeconomic development Poverty Clientelistic tradition in spite of greater development State’s share of economic activity (public goods) Ethnic fragmentation

Kitschelt and Wilkinson (2007); Chandra (2007) Dalton and Weldon (2007) Mass media influence Campbell et al. (1960); Lazarsfeld Early political socialization, socialization et al. (1944); Fiorina (1981) into secondary groups, “running tally”

Source: Author’s own construction on the basis of literature review.

Candidate based

Decrease Increase

Decrease

Increase Increase

Increase

Increase

Decrease

Had decreased Had decreased Decrease Increase before before (Morgan) (Morgan) Increase Decrease Increase Increase Increase Increase Decrease

Decrease Increase Increase



Segmented Electoral Appeals

57

partisan identities arise from vehicles of primary socialization (e.g., family, inter-generational transmission) (Campbell et al. 1960), others explain them as resulting from secondary socialization (e.g., from the influence of friends, co-workers, etc.) (Lazarsfeld et al. 1944). Fiorina (1981), as well as other authors looking at the dynamic evolution of partisanship (see e.g., Erikson et  al. 2002), provide a less static view of partisan identities. In these latter accounts, performance in office is a crucial determinant of parties’ capacity to retain loyal followers. Other analysts stress countries’ respective histories of democratization. Shefter (1994) argues that where democratization preceded bureaucratic expansion, clientelism is more likely than in cases where bureaucratic expansion preceded democratization. Similarly, both Keefer (2008) and Kitschelt et al. (2010) argue that long-term democratic experience influences countries’ types of party–voter linkages. According to Keefer (2008) democratic experience reduces the role of candidate-centered appeals relative to other alternatives. According to Kitschelt et  al. (2010), the longer a country has been democratic the more likely it is to have programmatic linkages. Kitschelt et al. (1999) and Kitschelt et al. (2010) explain current levels of programmatic structuring observed in East European and Latin American party systems as a result of countries’ complex, long-term trajectories in the twentieth century. In Eastern Europe, they argue, the legacies of communist rule and the capacity for mobilizing voters programmatically on that basis shaped these systems’ subsequent capacity for developing programmatic linkages. Meanwhile, in Latin America, the paucity of democratic contestation during the twentieth century, lack of societal capacities, and low salience of economic grievances reduced party systems’ capacity for structuring programmatic linkages. According to Kitschelt et al. (2010), Chile and Uruguay are exceptional in this regard, and are also the cases that showed greater levels of programmatic structuring in the late 1990s. Another set of authors also uses historical sequence to explain a functionally opposite result: the relatively rapid collapse of previously stable party systems (Mainwaring et al. 2006, Morgan 2011, Hawkins 2010). In this case, failures of representation as well as successive governments’ bad performance lead the way towards partisan dealignment and eventually party-system collapse. Fiorina’s (1981) model of partisanship as a “running tally” that voters update each election on the basis of government performance, fits this argument well. When the running tally turns increasingly negative, more room is open for candidate-based appeals, usually crafted around charismatic linkages seeking to fill the representation void (Hawkins 2010). This type of outcome has not been observed in Chile and Uruguay at the systemic level. Yet, traditional parties in both cases have alienated a segment of their original constituencies over

58

Segmented Representation

time, arguably because they have been increasingly unable to provide for them over a series of recent elections. Third, a series of authors point to structural factors as drivers of different types of linkages. According to Kitschelt and Wilkinson (2007), who draw on a long line of research, ceteris paribus socioeconomic development, as well as the relative greater size of the state sector in the economy, makes programmatic linkages more prevalent and clientelistic linkages less likely. Complementing this argument, and in line with traditional views in the literature, other authors point to high levels of poverty as fundamental drivers of clientelistic appeals (Stokes 2005; Lyne 2008; Taylor-Robinson 2010). Such an argument has received significant empirical backing in the literature (Kitschelt and Wilkinson 2007), and remains theoretically plausible, as the cost effectiveness of clientelism is enhanced in the context of mass poverty. Yet, Piattoni (2001) has argued that clientelism remains ubiquitous in highly developed societies (e.g., Italy and Greece), which display a long-lasting tradition of clientelistic intermediation. Piattoni’s argument implies the need also to look at possible “mutations” of clientelistic exchanges where socioeconomic conditions improve in previously clientelistic contexts. Also looking at structural factors, both Kitschelt and Wilkinson (2007) and Chandra (2007) point to ethnic divisions as a factor that can reduce programmatic politics and bring about greater levels of clientelism. Their argument implies that representatives of different ethnic groups would extract and distribute state resources as club goods to members of their own ethnic groups, instead of competing around the provision of larger public goods. This argument is less relevant to my analysis, given the low political salience of ethnic divides in the two societies I examine. Finally, a series of works seeks to explain individual party adaptation and are thus not reflected in Table  1.5, which focuses on system-level variance. Partisan adaptation is crucial in accounting for the electoral success of a specific party in a system (cf. Kitschelt 1989; Kitschelt 1994; Katz and Mair 1995; Levitsky 2003; Burgess and Levitsky 2003; Hunter 2010). Individual-level party explanations stress the role of relative partisan endowments (e.g., intangible programmatic affinities with the median voter at a given moment, material resources for funding a given electoral campaign) in causing adaptation and success. Parties’ organizational features might also assist them in pursuing and harmonizing certain linkage strategies (Panebianco 1988; Przeworski and Sprague 1986; Kitschelt 1994; Levitsky 2003). Although analyzing a single (usually successful) case of individual party adaptation inhibits the type of system-level analysis in which I propose to embed it, party-level explanations (e.g., Levitsky 2003; Hunter 2010)  provide extremely valuable narratives of successful instances of party adaptation to changing conditions.



Segmented Electoral Appeals

59

C AU SE S O F D I F F E R E N T T Y P E S O F L I N KAG E S A N D ST RU C T U R E S O F SE G M E N TAT IO N The main objective in this book is to improve descriptive inference regarding the structure of segmentation and strategic harmonization of party–voter linkages in contemporary societies. But I additionally propose a causal explanation of the evolution of party–voter linkages in Chile and Uruguay, which is then tested in out-of-sample cases. The causal argument partly supports several existing claims in the literature but contradicts others. While theoretically plausible, I do not find definitive empirical support for the first set of institutional arguments linking electoral rules to specific types of party–voter linkage outcomes.19 These institutional factors cannot explain shifts in linkage patterns over time in the context of constant institutional rules, or patterns of sub-national variance in party–voter linkages within the same set of electoral rules. At the individual party level, electoral rules fail to explain why certain parties adapt better than others to competitive scenarios in which particular multiple linkage strategies are electorally more efficient. I therefore conclude that electoral institutions are relatively insignificant in determining the specific types of party–voter linkages observed in different cases. Yet, they can be very consequential for the types of segmentation observed, and for how able parties in a given system are to harmonize different electoral strategies. The evidence in Chapters 2–4 also challenges Kitschelt et al.’s (2010) descriptive and causal arguments in two respects. Descriptively, these authors identify both Chile and Uruguay as having the most “programmatic structure” in the region, a similarity for which their causal variables seek to account. I show that this apparent similarity between the two cases holds descriptively only at the aggregate level. Applying my analytic framework reveals that the distribution, nature, and levels of segmentation of programmatic party–voter linkages are all actually very different in the two cases. Moreover, over time the two party systems have moved in opposite directions. While Chile lost programmatic strength during the 1990s and 2000s, Uruguay gained it. The causal factors stressed by these authors in explaining Chile and Uruguay’s cross-national convergence on high levels of programmatic structure cannot also explain their remarkable differences. However, I draw fundamental insights from Kitschelt et al.’s (2010) argument, assigning an important role to socio-structural and path-dependent factors that shape different

  19  In this regard, my findings are consistent with those of a recent wave of cross-national studies (Kitschelt and Kselman 2011; Kitschelt et al. 2010; Luna et al. n.d.).

60

Segmented Representation

opportunities for the pursuit of programmatic linkages in each system. In addition, in line with Kitschelt and Wilkinson’s (2007) claims regarding the size of the public sector, I also find the salience and size of potential electoral constituencies interested in the provision of large and encompassing public goods to be a fundamental driver of programmatic appeals. Which Types of Linkages and What Degree of Segmentation Are Observed in the System? I argue throughout that social inequality contributes to linkage segmentation. It does so by increasing the electoral benefits of segmented strategies, while reducing trade-offs and therefore the costs of such strategies. The impact of inequality in Latin America cannot be understated. The dramatic levels of social inequality observed in Latin America (as well as in cases such as the US and India) can affect democratic representation through multiple and complementary causal mechanisms. High levels of inequality increase the social salience of socioeconomic issues for a large proportion of the population. Yet, the translation of inequality into salient redistributive preferences is complex, and the assumption that one simply reflects the other is unwarranted. Such translation is contingent on both socio-structural manifestations of inequality and on political agency (Piñeiro and Rosenblatt 2011). Moreover, high inequality might lead to polarized distributive preferences, making it difficult for parties to craft relatively encompassing cross-class alliances (and therefore a winning electoral coalition) around a set of shared policy preferences. Paradoxically, high inequality might therefore reduce parties’ chances of mobilizing large constituencies around a socially salient divide on the basis of redistributive platforms (Blofield and Luna 2011). Alternatively, in cases where leaders pursuing “populist” mobilization attempt to push for radical distributive agendas, democratic rule may be endangered.20

  20  The political effects of inequality are also indirect. For instance, inequality might shape the demand side of the representation equation by inducing a series of psychological effects at the individual level. Accordingly, high inequality might contribute to consolidating cognitive biases that could be consequential for political preference formation (see Shapiro 2002). “Empathy” and “physical distance” gulfs might emerge among individuals living immersed in different and segregated socioeconomic settings (Shapiro 2002). Those cognitive gulfs distort the translation of effects of inequality into political preferences across segregated settings; for instance, by making differences among localities or social groups invisible (and therefore unconsequential for



Segmented Electoral Appeals

61

The potential effects of social inequality also interact with those of state and market reforms. Independent of their short- and long-term socioeconomic impact, market reforms have important implications for the configuration of parties’ electoral constituencies. First of all, liberalizing market reforms restrict the scope of state-supplied public goods that are provided either universally or to large social groups. This reduces the likelihood of mobilization strategies aimed at large constituencies on the basis of preferences regarding the provision and quality of public goods. Additionally, labor and pension reforms, as well as the privatization of state-owned enterprises, tend to reduce the scope of organized interest groups that can be targeted by political parties, such as unions and large organizations of social policy beneficiaries. Large constituent groups historically organize around the state-centered model become fragmented and lose their potential for collective action and collective preference formation (cf. Roberts 2002; Kurtz 2004). When applied in the context of already unequal societies, market reforms reduce the salience of large public goods around which electoral coalitions with alternative preferences regarding their provision can emerge and mobilize. The timing of reforms is also crucial. If market reforms are attempted under democracy, potential constituent groups can emerge by bringing together the “winners” or “losers” of the reform (cf. Hagopian 2009; Kitschelt et al. 2010). By contrast, when reforms are pursued under an authoritarian regime, they more directly and unequivocally contribute to the unraveling of large constituent groups. Along with socio-structural conditions (e.g., high inequality) and short-term, political-economic factors (e.g., state and market reforms), the long-term traits of a party system can also affect segmentation. If cleavage mobilization was important in the past, it might provide a basis for recrafting encompassing programmatic appeals in the present, thus reducing segmentation. If clientelistic mobilization was significant in the past, it might prompt parties to revert to their historical mobilization strategies, thus avoiding programmatic mobilization altogether. Or, as argued by Pasotti (2010), it might induce parties to compete by “branding” new styles (turning to new mobilization strategies and vilifying the “old ways”). In turn, if party identifications remain strong as a result of the endurance of historical parties in the system, that may leave less room for innovative party–voter linkage strategies political action). Different “framing” effects could also contribute to the emergence of distinct focuses of attention across social settings and groups, whose preference formation is primed, once again, in a socially segmented fashion. Overall, those processes make it more difficult for parties to engage in distributive preference formation (across social groups) and mobilization.

62

Segmented Representation

to replace them. Each of these scenarios makes certain types of linkages politically unviable, thus restricting the menu of linkage options available. Is Segmentation Mixed or Dual? The geographic nature of electoral districts (their size and territorial scope) is a function of electoral rules such as the average district size, district magnitude, and the nature of districts’ boundaries. Smaller district sizes can contribute to the configuration of socially homogeneous districts, making social stratification between districts greater than within them. That is especially the case in unequal societies, where territorial units tend to become more socially segregated and internally homogeneous. Smaller districts and smaller district magnitudes should thus correlate with lower levels of segmentation within districts (thereby making mixed strategies less likely). When the number of candidates is small (a result of lower district magnitudes), it might become more difficult for parties to structure segmented linkages within a district. In a nutshell, one candidate has more difficulty in simultaneously structuring programmatic, clientelistic, and/or charismatic linkages, than a team of candidates from the same party in which different candidates can specialize in a linkage to target selected constituencies within a district. If district boundaries overlap with socially segregated neighborhoods, opportunities for segmenting between districts (dual-linkage strategies) increase, while the incentives for segmenting within districts (mixed-linkage strategies) decrease. Parties’ capacity to mobilize one constituency on the basis of programmatic linkages across districts (e.g., mobilizing poor voters or union members, independently of their specific geographical distribution) is at least partly determined by the strength of transversal interest groups. Processes such as privatization or the decentralization of services usually fragment and weaken transversal interest groups. On the one hand, those reform processes generate “exit” options for upper segments of society. Those “exit” options, while reinforcing a dual distribution of access to previously universal public goods such as education, health, public security, and infrastructure, also contribute to diminishing the political “voice” of those who cannot opt out of the public good (Filgueira and Filgueira 1997a). Moreover, even in contexts of rising inequality, increased access to mass consumption and increasing (but segmented) social mobility also make the emergence of broad constituencies formed around convergent preferences on the provision of public goods less likely. Unequal and fragmented civil societies, and less universal public goods provision, make it more difficult for a functional equivalent to the factors favorable to class-mobilization and cross-class coalition-making that



Segmented Electoral Appeals

63

allowed programmatic mobilization to occur in post-WWII Europe to emerge (Mair 1990). In sum, the interaction of long-term socio-structural traits (inequality) with market and decentralizing reforms segments and fragments civil societies along territorial and socioeconomic lines, making districts more internally homogeneous and externally socially segregated, while contributing to weakening and fragmenting large constituency and interest groups concerned with provision of public goods. Ceteris paribus, this makes it more difficult to craft cross-class political coalitions and develop more encompassing programmatic platforms. Eventually, at the individual level, these processes also shape the ways in which political preferences are constructed and might further contribute to the decline of organized collective action. Can Parties Strategically Harmonize Segmented Appeals? If parties segment their strategies (either between or within districts), strategic harmonization is needed to hold the partisan organization together. Pure linkage strategies automatically solve strategic harmonization dilemmas, integrating partisan activities across multiple arenas (electoral, territorial, congressional, etc.). For instance, a party that relies on programmatic linkages to its constituents also has the ability to harmonize electoral mobilization efforts and achieve party discipline on that basis (see Aldrich 1995). A clientelistic machine also solves harmonization dilemmas by putting together a hierarchical organization that distributes benefits and penalties to party activists who contribute to the collective well-being of the party, efficiently exchanging favors (private or club goods) for votes. Yet, if the trade-off assumption is relaxed, as proposed here, one can no longer assume that programmatic structure at the national level disables or squeezes out non-programmatic linkages at the district level, especially when programmatic linkages are more prevalent in specific electoral cross-sections. On the one hand, budget constraints required by both anti-inflationary policies and economic openness have reduced the state’s economic leverage, shrinking the scope for economic policy divergence. In societies where economic concerns are highly salient, the reduced scope of policy divergence on the state’s economic relation to the market can limit parties’ capacities to compete on significantly different programmatic platforms. Eventually, polarized distributive preferences (caused by high inequality), matched by decreasing room for competing on distinct economic policy packages place exclusively programmatic political parties on shaky ground.

64

Segmented Representation

On the other hand, the combination of privatization and resulting budget constraints may also challenge the survival of state-dependent clientelistic machines and national corporatist arrangements by making fewer discretionary resources available to governing politicians. In turn, decentralization of public services may also limit massive clientelistic strategies by segmenting and localizing client–patron networks. Economic openness also constrains clientelism: pervasive clientelism induces economic inefficiency and contributes to an investment-hostile environment. Decentralization also poses daunting organizational challenges to national parties. Centralized polities enhance parties’ capacity for enforcing discipline and local-national alignments, particularly by enabling them to control deployment of valuable resources from the center to the periphery (see Valenzuela 1977). National–local brokerage networks were instrumental in keeping parties institutionally strong under the state-centric model (Valenzuela 1977; Borzutzky 2002; González 1991). Partisan territorial committees and cells were usually financed on the basis of this center–periphery resource flow. With decentralization, local politicians gained financial and political autonomy from national party leaders. While this might yield a welcome democratization of local politics, it also hinders national partisan organizations (Harbers 2010). Increasingly personalistic leaderships can also begin to flourish at the local level, consolidating multiple “electoral microcosms” within each party’s electoral base. While this might strengthen local party apparatuses, it can also generate greater needs for harmonization at the national level. Paradoxically, national party leaders can then find themselves faced with a greater need to achieve harmonization, while having fewer instruments at hand to do so; programmatic alignments become less salient even while there are fewer resources available for deploying harmonized non-programmatic strategies. State and market reforms then complicate political parties’ unified action, constraining policy divergence and therefore programmatic mobilization, and together with decentralization, reduce parties’ room for implementing non-programmatic mobilization strategies on the basis of national-level fiscal resources. After state and market reforms are introduced, two sets of factors can affect parties’ capacity to pursue strategic harmonization. On the one hand, electoral institutions might provide political parties with the capacity to organize across districts. For instance, if parties have the capacity to control the nomination process, have access to material resources that can be centrally allocated to reward party loyalists, or operate in the context of an electoral system that favors political parties vis-à-vis individual candidacies, party leaders will then have more opportunities for strategically harmonizing linkage strategies. Such institutional incentives operate at the systemic level, thus applying equally to every political party competing in the system.



Segmented Electoral Appeals

65

On the other hand, parties diverge in their access to material and symbolic resources helpful for harmonizing segmented linkage strategies. For instance, in cases where traditional parties are strong, political parties might benefit from strong party identifications or well-crystallized programmatic reputations. Parties whose core constituencies have more access to financial resources may benefit from access to private donations. In turn, parties that have developed and maintained a mass-level organization can rely on their activists for developing labor-intensive campaign activities. In sum, different parties have access to different resources useful for harmonizing strategy across districts. Yet, parties could also have access to resources that do not provide means for efficiently segmenting and harmonizing linkage strategies at a given point in time. For instance, in an anti-establishment climate, having a strong and long-lived party label might prove disadvantageous. In sum, differential access to material, symbolic, and organizational resources for crafting segmented and harmonized linkage strategies at a given point in time can explain divergence among parties competing in the same system, preventing contagion. As a result, campaign strategies might not necessarily spread from a successful party to competitors in the same system. In this book, I  am not interested in analyzing party organizations more broadly. Yet, the evidence I present in Chapter 5 does suggest important differences in the internal organization of parties falling into each of the three cells identified in Table 1.1. Such evidence also suggests the presence of feedback effects between linkage types and organizational features, as already suggested by previous analyses of Latin American parties. For instance, as claimed by Levitsky (2001, p.  103), clientelism in Brazil is consistent with highly individual politics and frequent party switching (Mainwaring 1999). Meanwhile, in parties characterized by strong charismatic leaderships (e.g., the Argentine Partido Justicialista or the Mexican Partido Revolucionario Institucional), the party leader exercises a good deal of internal power by disciplining the rank and file (Bruhn 1997; McGuire 1997). In the analysis that follows, I  find that parties with predominantly candidate-centered appeals—like Renovación Nacional and those of the Concertación in Chile—have less harmonized strategies, as well as increasingly less powerful partisan organizations than the UDI. Meanwhile, parties using predominantly party-based appeals—like those in Uruguay, and especially the FA—achieve greater harmony yet flexibly include many fractions. Finally, in parties combining party and candidate-based appeals—like the UDI in Chile—the party may remain central in providing resources to candidate-based campaigns and thus still be able to achieve harmonization. Organizationally, such harmonization coincides with a top-down and strictly hierarchical leadership structure.

66

Segmented Representation

However, comparison of the UDI and the FA as cases successfully combining segmentation and harmonization does not point to a single, optimal organizational model. Indeed, whereas the FA is a highly factionalized mass-based party, the UDI is an internally homogeneous and very hierarchical party. In terms of its organization, it approximates a professional electoral party. The hierarchical and top-down structure of the UDI is at odds with conventional expectations regarding more open party organizations’ supposedly superior ability to promote successful party adaptation (Panebianco 1988; Kitschelt 1994; Roberts 1998; Levitsky 2003). Expected Types of Causation and Hypothesized Causal Configurations Across Cases The approach to causation I advocate here is not linear, but conjunctural (see Ragin 1987). I  assume that different combinations of causes can eventually lead to the same outcome across cases, a pattern sometimes called “equifinality,” and that, at the same time, largely similar causal configurations occurring across different contexts can lead to divergent outcomes (“multi-finality”). The two case studies developed in the major part of this book clearly suggest that such an approach to causation is warranted. Yet, my two case studies are explicitly aimed at empirically applying the proposed descriptive framework and at generating hypotheses that could be tested in future studies in which, unlike the two presented here, relevant causal interactions can be more clearly and unambiguously isolated. Regarding segmentation and the types of party–voter linkages pursued in different cases, I  have four complementary expectations. First, I  expect greater levels of social inequality to produce more linkage segmentation, and to reduce the scope of programmatic mobilization. Inequality is today greater in Chile than in Uruguay. Over time inequality has likely increased in Chile, while remaining fairly constant in Uruguay. On this basis I expect to observe increasing levels of socioeconomic segmentation of electoral appeals in Chile, as well as reduced scope for programmatic mobilization. For Uruguay, I expect both these levels to be more constant over time. Second, privatizing and decentralizing state reforms as well as market liberalization were massive in Chile under the authoritarian regime headed by Pinochet. I expect these reforms and liberalization to have further reduced the scope for programmatic mobilization in Chile, where I  also expect to find more fragmentation of transversal interest groups than in Uruguay. While neoliberal reforms were attempted in Uruguay after re-democratization, their reach was restricted, largely because they



Segmented Electoral Appeals

67

activated a veto coalition in which the FA and interest groups coalesced to defend the remnants of the import-substitution-industrialization model. Interest groups, faced with such reform attempts under democratic conditions with lower risk of repression, mobilized at the national level and avoided the high levels of fragmentation and demobilization observed in Chile. Third, fiscal and budgetary constraints increased in Uruguay during the post-transitional period, due to the crisis of the statist model. As a result, I  expect parties’ ability to feed encompassing clientelistic machines to have declined. I  expect to observe the same outcome in Chile, although as a result of a different political economy. In the Chilean case, I  expect market and decentralizing reforms to have also eroded the state apparatus’ capacity to fund traditional clientelistic networks. The scope of market activity and strength of the private sector in Chile may also have substituted state funding for private financing of clientelistic or particularistic mobilization strategies. Private funding of political campaigns should be less pivotal in the still heavily statist Uruguay. Fourth, I expect the politics of the transition to democracy also to shape the centrality of partisan identities and parties’ incentives for nurturing certain mobilization strategies rather than others. In the case of Chile, the regime divide running between pro-Pinochet forces and the opposition to the authoritarian regime gained prevalence over alternative political divisions. The binominal electoral system meanwhile provided overwhelming incentives for the convergence of different forces within each of the two principal camps in the form of electoral pacts. As time passed, each coalition (the originally pro-Pinochet Alianza and the pro-democratic Concertación) overshadowed the different political parties comprising them. Other “authoritarian enclaves” included in the 1980 constitution also promoted a highly consensual political style, which gradually blurred the programmatic divide between both political camps. This meant that each coalition’s political identity, as well as electoral competition between them, remained anchored in the political divide of the 1970s and 1980s. In Uruguay, by contrast, political parties and their individual partisan identities remained central for electoral mobilization. Moreover, during the authoritarian period and the transition to democracy, the FA emerged as a powerful political subculture in society. Over time, both traditional parties (Blancos and Colorados) and the FA engaged in intense electoral competition, reinforcing a bi-polar competitive logic. This logic became pivotal in reinforcing programmatic mobilization along the state–market divide, with the FA defending the statist model, and both traditional parties promoting economic reform while in office during the 1990s.

68

Segmented Representation

In Uruguay, I  therefore expect to find programmatic linkages mobilizing voters around distributive issues. I also expect partisan identifications to be stronger in Uruguay than in Chile, as in the latter the authoritarian–democratic divide is mobilized through internally heterogeneous electoral pacts. State-funded clientelism and particularism should in turn have decreased in both cases. In reality, however, this expectation seems to have been fully borne out only in Uruguay, whereas privately funded clientelism or particularism appears to have remained more prominent in Chile. Regarding the structure of linkage segmentation, I expect to observe more instances of dual segmentation in Chile than in Uruguay, where mixed linkage configurations should be more common. This expectation is justified on two grounds. First, the interaction between Chile’s electoral system and the patterns of socioeconomic segregation characteristic of highly unequal social structures generates internally homogeneous and externally heterogeneous districts. In Uruguay, districts are more internally heterogeneous socioeconomically, making dual patterns of linkage segmentation less likely. Those patterns are also reinforced through the electoral system. In Uruguay, with larger district magnitudes, different MPs and congressional candidates can specialize in mobilizing different socioeconomic segments within a district, thus reinforcing a mixed linkage structure. This is virtually impossible in Chile, where district magnitude is fixed at 2, and each electoral pact usually elects one MP per district. Second, the organization of interest groups across district boundaries provides greater incentives for parties in Uruguay to mobilize voters on the basis of transversal issues and programmatic stances. This should therefore decrease the likelihood of observing dual-linkage structures, in which every district is treated as a “parallel reality.” Given that interest groups in Uruguay are more organized while those in Chile are more fragmented, I expect to find more mixed segmentation in the former country and more dual segmentation in the latter. When it comes to strategic harmonization, institutional features can contribute to providing incentives that either facilitate or hinder coordination by the party leadership. Ceteris paribus, small district magnitudes, personalized list structures, and a split between local- (municipal) and national-level elections (congressional, presidential) should be expected to strengthen individual candidates relative to their parties. In both Chile and Uruguay, localand national-level elections are split; thus, my cases do not provide leverage for testing my expectation regarding this last institutional independent variable. District magnitudes are variable in Uruguay, reaching a maximum of forty-four in Montevideo, and a minimum of two in the least-populated districts. In Chile, as noted above, magnitude is fixed at 2. On this basis, I expect



Segmented Electoral Appeals

69

to observe more personalization in Chile and in low-magnitude districts in Uruguay than in the most populated districts in the latter case. The structure of electoral lists in each case is expected to reinforce such trends, since Chile’s system is more personalized. Formally, candidates’ nomination is contingent on party leaders in both countries. Yet, due to the strong incumbency advantages observed in Chile, control over nominations has informally shifted from party elites to sitting MPs (Navia 2008). This should also contribute to greater personalization, generating a need for more strategic coordination by party leaders to achieve harmonization. In the Uruguayan case, nomination, as well as the relative ranking of each candidate within the list, is controlled by fraction and party leaders, thus facilitating strategic coordination. Finally, party leaders can also seek to promote strategic harmonization by allocating more generous financial resources to more loyal candidates. However, it is harder for party leaders to capitalize on this strategy where fundraising is decentralized and takes place at the individual candidate level. Decentralized fundraising is generally observed in Chile; thus further decreasing leaders’ capacity to strategically coordinate electoral appeals. In Uruguay, campaign contributions are raised at the fraction level, thus facilitating the pursuit of strategic harmonization by party or fractional leaders. Turning to the individual party level, I  argue that parties that have disproportionate access to material and symbolic resources for segmenting and harmonizing appeals will be more successful electorally. Yet, the nature of those required resources varies according to both time and place. How such resources are administered and deployed through differently structured party organizations translating them into electoral outcomes also varies widely across cases. In sum, in accordance with equifinality, even if the same value of a given independent variable can produce the same result across cases, it may do so via different and contextually variable causal mechanisms and via distinct interactions with third variables.

2 Patterns of Party–Voter Linkages in Chile and Uruguay: A Stylized Description of the Pre-1973 and Post-Transitional Periods

I N T R O D U C T IO N This chapter has three purposes. First, it substantiates the claim that in unequal societies segmented electoral appeals are an enduring characteristic of durable democracies and stable party systems. While traditional analyses of Chilean and Uruguayan party systems tend to neglect or underestimate segmentation by focusing on one particular type of linkage, linkage segmentation is actually a long-standing feature of both party systems. Segmentation is shaped by the political economy of each society as well as the political opportunity structures that parties confront in a given time-period, but its presence is almost constant. Parties’ relative capacity to harmonize segmented linkage strategies also varies contextually, depending on the resources to which they have access. Second, the chapter seeks to describe the main features of the traditional Chilean and Uruguayan party systems in order to establish a basis for assessing change in the post-transitional era. To this end, I present a stylized account of both systems. A  full-fledged historical narrative describing the evolution of both party systems since independence and until 1973 is available in the online appendix of the book.1 Third, the chapter characterizes party–voter linkages since Chile and Uruguay re-democratized, in 1989 and 1984 respectively. The framework proposed in Chapter 1 is then used to compare the two cases’ contemporary linkage patterns, providing an analysis for the contemporary period equivalent to that offered for the years prior to 1973. This comparison of the two periods provides the basis for the causal analysis of changes in party–voter linkages that Chapter 6 presents.  

1 

The online appendix is available at .



Party–Voter Linkages in Chile and Uruguay

71

T H E C H I L E A N A N D U RU G UAYA N P R E - AU T HO R I TA R IA N PA RT Y SYS T E M S A General Characterization of the Chilean Party System The pre-authoritarian party system of Chile has received widespread academic attention, and has been consensually characterized as one of the most stable, institutionalized, and ideologically structured party systems in Latin America (Gil 1966; Valenzuela 1977 and 1995; Garretón 1988; Dix 1989 Scully 1992; Mainwaring and Scully 1995; Siavelis 1999; Roberts n.d.).2 Dix (1989) therefore presents the Chilean party system as the one in Latin America most closely resembling the Western European model. Yet, Chilean parties failed to clearly capture specific social bases. Although the left was stronger among the working class, and the center and right had their electoral strongholds in the middle and upper classes, all parties had heterogeneous social bases and recruited mainly poorer voters (Valenzuela and Valenzuela 1986, p. 197). Miners’ strong support for the Partido Comunista (PC-Chi) was the sole partial exception to this general rule (Valenzuela and Valenzuela 1986, p. 197). At the elite level, Chilean political parties not only differed ideologically, but also constituted cultural and organizational “pillars” through which society was organized. Collective, and especially local interests were channeled and represented through ubiquitous partisan organizations (Valenzuela 1977; Scully 1992; Mainwaring and Scully 1995; Roberts forthcoming). Chilean parties became the “brokers” that linked society with the state and constituted the “backbone” of the pre-authoritarian sociopolitical matrix, penetrating every aspect of society and overshadowing interest groups, trade unions, social movements, and community organizations (Gil 1966; Garretón 1988; Valenzuela 1999). The twentieth-century party-system has also generally been characterized as one in which a system of political compromise predominated.3 Compromise resulted from a tradition of coalitional rule and elite cooperation, in turn facilitated by the presence of a moderate political center (Valenzuela 1978; Moulián 1985; Scully 1992). The division of the electorate into “ideological thirds,” and

  2  See Montes et al. (2000) for an exception to this view.   3  This of course applies to the pre-1973 period. After the transition to democracy the tradition of political compromise was revived but came to be known as the democracia de los consensos (consensual democracy). See Chapter 3 for further detail on this latter period.

Segmented Representation

72

the relatively high levels of state autonomy vis-à-vis individual parties, reinforced the centrality of electoral contestation and promoted competition (particularly in local and congressional elections). However, at the same time it also encouraged coalition building and compromise in presidential elections and congressional politics (Scully 1992; Valenzuela 1977, 1978, and 1999). Operating in that context, parties were also able to extract resources from the state in order to develop encompassing national organizations centered on vertical patronage and clientelisitc networks redistributing mineral rents to the periphery (Valenzuela 1977; Borzutzky 2002). In this regard, the horse-trading and pork-barrel used to build governing coalitions allowed parties to extract state resources to finance clientelistic electoral mobilization. The combination of ideology and clientelism was a key feature of the pre-1973 system (Borzutzky 2002, p.  26). Whereas ideology was diffused through each party’s organization and provided “worldviews” helping them propose solutions to the problems of the day, clientelism, including on the left, secured political parties’ electoral bases. It also provided incentives for pursuing accommodation and moderation in the congressional and policy-making arenas (Borzutzky 2002, p.  26), in order to obtain access to resources that could be deployed in clientelistic appeals. In turn, frequent political turnover among political parties with distinct ideologies was also observed, signaling the weakness of deep-rooted partisan identities. With a few partial exceptions (1891 and the late 1920s), the combination of programmatic and clientelistic mobilization observed in Chile contributed to keeping constitutional rule in place from 1830 until the mid-1960s (Moulián 1985). From that point however, the increasing emphasis on programmatic mobilization reduced segmentation and led to significant sociopolitical polarization before the military coup of 1973. A General Characterization of the Uruguayan Party System In Uruguay, political parties represented cross-sectional cuts of society. They established links with constituents by combining resilient partisan identities with an expanding system of clientelistic side-payments and state patronage. From the nineteenth century to 1971, the party system was structured around two major parties, the Partido Blanco and the Partido Colorado.4 If we trace back these parties’ origins to the confrontations between rural caudillos in the

 

4 

In this work I use Partido Blanco and Partido Nacional as interchangeable terms.



Party–Voter Linkages in Chile and Uruguay

73

post-independence years, both managed to adapt and survive for more than 170 years, maintaining control of the presidency until the election of 2004. Until 1971, the Blancos and Colorados managed together to obtain more than 90 percent of the votes, successfully frustrating the entry of third parties into the political arena.5 The centrality of the two parties (lemas) and the traditional party system were both powerfully reinforced by Uruguayan electoral institutions. Although Uruguay combined a single plurality rule for electing the president with a proportional representation (PR) system for Congress, the centrifugal effects of PR were contained by a system of concurrent and “fused” elections (at the presidential, congressional, and local levels) and the use of the double simultaneous vote (DSV). The DSV was instituted in 1910, and became the electoral masterpiece ensuring continuity in the system.6 The DSV means that voters vote simultaneously for a party and for a specific set of candidates within the party, although they have to select lists—which cannot be modified—from among those presented by rival fractions within each party. Once they choose the party, they have as many options as there are lists presented by the party. An obvious consequence of voting for a party is that split tickets are not allowed. The party that obtains a plurality of the votes is the winner; and the winner of a plurality within that party becomes president. From 1934 until the electoral cycle of 1999–2000 all elections occurred at the same time. This simultaneity obviously reinforced the effects of the ban on split tickets. The DVS vote, reinforced by the PR electoral system, provided incentives for emerging leaderships in the party to compete with established ones by opening a parallel electoral list (sub-lema) within the overall lema. While allowing emerging leaders to test their electoral potential and challenge established party leaders, such parallel lists also accumulated votes for the lema while diversifying the party’s overall electoral menu (González 1995, pp. 146–7). These electoral institutions and the flexibility inherent in a highly fractionalized system allowed the traditional party system to endure and resist subsequent challenges brought about first by socioeconomic modernization, and then by stagnation in the post-World War II period (Rial 1984). By the mid-1950s, social discontent borne of economic stagnation increasingly   5  Indeed, prior to the 1960s only a handful of non-traditional parties had ever achieved congressional representation: the Union Liberal founded in 1855, the Partido Radical founded in 1873, the Partido Constitucional founded in 1880, the Partido Socialista that was founded at the turn of the century and gave rise to the Partido Comunista in 1921 following an internal split, and the Unión Cívica founded in 1910 and that gave rise to the Partido Demócrata Cristiano in 1960 after a split from its left wing.   6  For a detailed characterization of the DSV, see Piñeiro (2007).

74

Segmented Representation

strained the system (Finch et al. 1980). This strain resulted in the emergence of the FA as a significant third party in the election of 1971. Prior to the FA’s emergence, the traditional two-party system witnessed an unprecedented cycle of alternation between the traditional parties, with the Partido Blanco winning the presidential election for the first time in 1958, and a steady increase in their internal fractionalization.7 In terms of their internal organization, both traditional parties were “rather loose coalitions of fractions.”8 Uruguayan parties were therefore able to accommodate people with all sorts of widely differing ideas, while fractions of rival parties “professing . . . exactly the same ideals” could also at times be observed.9 The high fractionalization and internal heterogeneity of Uruguayan parties, as well as frequent collaboration between ideologically proximate fractions of different parties, led many observers to question the existence of a truly bipartisan system in the country, and to claim that Uruguay instead had a multi-party system, albeit one hidden behind a two-party façade (see e.g., Lindahl 1962; Real de Azúa 1971; Sartori 1976; Errandonea 1989).10 However, although they were “catch-all” organizations, the traditional parties were nonetheless “meaningful” (Gonzalez 1995, p. 143). Strong partisan identifications that emerged in the civil wars of the nineteenth century were crucial in demarcating the boundaries between the Blanco and Colorado camps. These respective partisan identities were held together by deep-rooted political traditions (Pérez Antón 1988). Party fractions competed by combining candidate-trait and clientelistic appeals, with some emphasis on their programmatic profiles. Whereas the latter were more prominent when mobilizing middle- and upper-class voters, the first two were important with middle and lower socioeconomic groups. From the nineteenth century onwards, the division of labor between caudillos and doctores within each party and partisan fraction facilitated this segmentation of appeals. Caudillos, of popular extraction, leveraged clientelistic and leader-based appeals, both in the rural sector and in the cities. Some of these caudillos also developed distinctive programmatic platforms. Doctores, or urban professionals of upper- and middle-class extraction, were more

  7  Whereas from 1925 to 1931 the number of lower chamber electoral lists averaged 143, for 1946–1971 that figure rose to 314, peaking in 1971 with 590 lower-chamber lists (see González 1991 and 1995). However, the real impact of fractionalization might have been exaggerated by the empirical focus on lists contesting the elections instead of on those effectively winning congressional seats (see Buquet et al. 1998).   8  González (1995, p. 142)   9  Julio Martínez Lamas (1946) cited in González (1995, p. 142).   10  See González 1995 for a critical appraisal and review of this literature.



Party–Voter Linkages in Chile and Uruguay

75

clearly engaged in programmatic mobilization, also usually combined with candidate-trait appeals. Meanwhile, political parties competed by mobilizing strong partisan identities that were pivotal in appealing to voters from all social segments. Parties and their fractions thus deployed a wide range of electoral appeals. In contrast to Chile, however, the socioeconomic segmentation of parties’ appeals seems to have been lower in Uruguay. Clientelism was not restricted solely to mobilizing lower-class and/or rural voters, and programmatic platforms and leader-based appeals were used to mobilize not only upper but also middle and lower social segments (Real de Azúa 1971; Rama 1971; Fa Robaina 1972; Aguiar 1984; Vernazza 1989). Meanwhile, partisan identifications were equally instrumental in getting the vote of lower and upper classes. The relatively lower levels of inequality and social differentiation between social classes historically observed in Uruguay paralleled relatively low levels of socioeconomic inequality (Real de Azúa 1971). Notwithstanding, the range of different electoral appeals used to mobilize voters in pre-1973 Uruguay was arguable wider than in Chile. Voters could thus display a variety of “motives” for adhering to a particular subfraction (agrupación), fraction, and party (lema). Appeal segmentation contributed to providing and reinforcing such motivations. As in Chile, political parties in Uruguay became central in brokering state– society relations in the pre-1973 period. Their central role in this regard was based on both the pursuit of segmented linkage appeals and of consensual politics at the elite level. The latter helped to stabilize democracy, while providing rival parties with access to state resources for the pursuit of clientelism. Starting in 1955, a long-term economic crisis progressively challenged the survival of the system by limiting key resources for the electoral reproduction of the traditional party system. While the drainage of state resources reduced the capacity to supply clientelistic side-payments, voters’ mounting discontent also challenged traditional party identifications and parties’ public policy records. In this context, before the 1973 military coup the system witnessed increasing polarization and unrest, as well as the emergence of a significant third-party challenger in 1971.

A P P LY I N G T H E A NA LY T IC A L F R A M EWO R K T O T H E P R E - AU T HO R I TA R IA N PA RT Y SYS T E M S This section uses the framework proposed in Chapter 1 to analyze the characteristics of party–voter linkages in the party systems of Chile and Uruguay

76

Segmented Representation

Table 2.1.  Socioeconomic linkage segmentation in Chile (pre-1964) Resources used in the appeal Agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate Party

Candidate traits Partisan identity

Electoral platform Programmatic (upper and middle classes, some working class in mining areas)*

Particularistic Clientelistic (popular and rural sectors)+

Key Bold: Predominant strategies in system. + Predominant strategy for peripheral constituency. * Predominant strategy for core constituency.

up until 1973. Given the lack of precise case narratives at the district level, my historical characterization is limited to the systemic level, and to “gross” territorial demarcations at the regional level. Whereas Chile in this era had two distinct periods (up to 1964 and from 1964 to 1973), Uruguay exhibited more of a constant general pattern, especially from 1919 onwards. Tables 2.1. and 2.2 show the relative incidence of different types of linkages in mobilizing distinct socioeconomic constituencies in the pre-1973 era in Chile. Before 1964, Chilean political parties drew on a combination of clientelistic and particularistic linkages to mobilize poor and rural constituents at the local level, and used programmatic appeals to mobilize elite and urban middle-class groups. Working-class voters in mining areas were also mobilized programmatically. Whereas classic works on the Chilean party system have focused more strongly on one type of linkage (e.g., Scully 1992 on cleavage mobilization, Borzutzky 2002 on clientelism and local–national patronage networks), the party system actually combined both types of mobilization. In Valenzuela (1977), the combination of programmatic politics (at least at the elite level) and particularistic mobilization (especially at the local level) is explicitly acknowledged and thoroughly described. Political parties were the main agents implementing and thus harmonizing both of these kinds of electoral appeals. At the elite level, partisan subcultures and strong partisan identifications were present. Yet, high electoral competitiveness, continuous partisan turnover, and instances of successful anti-party leaderships reduced the overall role of strong partisan identifications, especially in regard to non-elite groups (Montes et  al. 2000). In this sense, ideological currents on the programmatic side, and brokers positioned in the national–local patronage and



Party–Voter Linkages in Chile and Uruguay

77

clientelistic networks, were probably more salient for structuring linkages than strong party identifications. Considerable territorial segmentation existed between urban, urban poor, and rural districts. I  therefore classify Chilean parties’ linkage strategies up to 1964 as having approximated the “harmonized dual type,” achieved through parties’ self-harmonizing mechanisms. Chilean parties simultaneously specialized in diffusing their ideology to the mass level and developing ever-expanding brokerage networks distributing state resources to the territorial periphery. Both types of activities (i.e., the building of mass-based organizations and the brokerage of state resources) strengthened parties as collective institutions. Linkage segmentation was also sometimes present within districts where parties targeted different socioeconomic groups of constituents (e.g., urban poor versus urban middle and upper segments, or mining workers versus rural workers). But for the most part parties segmented linkage types across districts with different socioeconomic profiles. Whereas urban dwellers were more frequently mobilized through cleavage mobilization, the urban poor and the rural poor (especially those to the south of the metropolitan region in peonage) were frequently mobilized through clientelism and vote-buying transactions (Gil 1966; Moulián 1985; Loveman 1976; Valenzuela 1999). Panel 1 in Figure 2.1 illustrates the configuration of linkage types observed in Chile pre-1964. Table 2.2 characterizes the 1964–1973 Chilean party system. By the early 1960s, parties of the center and left had begun systematically pursuing programmatic mobilization efforts in parts of the countryside (especially in the south and central regions). This was especially true of the Demócrata Cristiano, who gradually displaced the Partido Radical. Land reforms implemented under Presidents Frei Montalva and Allende were the corollary of those mobilization efforts (Loveman 1976). Similar mobilization initiatives were pursued in the urban centers regarding pobladores (shantytown dwellers) with the Movimiento de Izquierda Revolucionaria and the PC-Chi organizing land takeovers in the urban periphery (Espinoza 1998). Those efforts weakened conservative parties’ clientelistic base in the rural sectors, contributing to ideological polarization in the pre-coup years as conservatives’ stake in congressional bargains thereby weakened (Valenzuela 1999; Scully 1992; Moulián 1984). At the same time, however, these more programmatic mobilization initiatives more closely aligned the popular sectors with parties of the left and center, enhancing programmatic representation in the system and arguably reducing the segmentation of electoral appeals. In the post-1964 period, something approaching the “pure linkage strategy” emerged. Linkage segmentation across socioeconomically distinct types of districts was reduced, while parties competed more strongly on

Segmented Representation

78 Chile Pre-1964

PARTY

RURAL (SOUTH)

URBAN

RURAL (MINING)

Chile Post-1964 PARTY

District 1

District 2

District 3

Programmatic Clientelism

Figure 2.1.  Territorial linkage segmentation in Chile (pre-1964 and 1964–1973) Table 2.2.  Socioeconomic linkage segmentation in Chile (1964–1973) Resources used in the appeal Agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate Party

Candidate traits Partisan identity

Electoral platform Programmatic (all social segments)

Particularistic Clientelistic (local level)

Key Bold: Predominant strategies in system.



Party–Voter Linkages in Chile and Uruguay

79

Table 2.3.  Socioeconomic linkage segmentation in Uruguay (until 1973) Resources used in the appeal Agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate

Candidate traits (all social segments)*+ Partisan identity (all social segments)*+

Electoral Platform

Particularistic

Programmatic (more upper and middle classes)*

Clientelistic (marginally more in lower and rural segments)+

Party

Key Bold: Predominant strategies in system. + Predominant strategy for peripheral constituency. * Predominant strategy for core constituency.

the basis of programmatic mobilization. Of course, clientelism, patronage, and “pork-barreling” continued at the local level, but the overall levels of between-district segmentation were reduced. In this period, the reduction of segmentation logically should have also reduced the need to harmonize parties’ electoral appeals. Panel 2 in Figure 2.1. illustrates the transformation of linkage patterns observed for the 1964–1973 period, in contrast to those described for the earlier period. Except for the similar incidence of clientelistic linkages, pre-1973 Uruguay contrasted sharply with Chile. On the one hand, as illustrated in Table  2.3, party-identification, clientelism, and leader-based appeals were all predominant in the Uruguayan system, and linkage strategies were less segmented along socioeconomic lines. Whereas party identifications were key in establishing linkages between political parties (lemas) and voters, clientelism and leader-based appeals (including not only candidate traits but also programmatic platforms) were predominant in determining voters’ allegiance to different fractions within parties. Drawing on two roughly comparable surveys during the pre-authoritarian period (conducted in 1964 Chile and 1966 Uruguay), Tables 2.4a and 2.4b compare the levels of left–right polarization observed in each system prior to the authoritarian period using partisan self-identification figures for Chilean (Table 2.4a) and Uruguayan (Table 2.4b) university students in the mid-1960s. Unfortunately, these surveys were based on samples of only university students and therefore do not allow one to draw inferences beyond this social group. Nonetheless, they provide a basis for some rough comparative analysis of features of the two party systems at this time. The comparison reveals a significantly greater level of ideological polarization in the Chilean

80

Segmented Representation Table 2.4a.  Left–right self-placement and polarization of university students in pre-authoritarian Chile Chilean students Party with which student most identified

Left–right self-placement, mean (std. dev.) (scale: 1–10)

Socialist Communist FRAP Radicals Christian Democrats Liberals Conservatives Diff. between extreme left and right parties

2.7 (1) 1.7 (.7) 2.2 (.8) 4.1 (1.2) 4.15 (1.2) 6.2 (1.3) 7 (.6) 7–1.7 = 5.3

Table 2.4b.  Left–right self-placement and polarization of university students in pre-authoritarian Uruguay Uruguayan students Party with which student most identified Communist Socialist Christian Democrats Colorado Blanco Diff. between extreme left and right parties

Left–right self-placement, mean (std. dev.) (scale: 1–5) 2.2 (1) 2.8(.8) 3.1 (.9) 3.3 (.9) 3.6 (.8) 3.6–2.2 = 1.4*2 = 2.8

Source:  Author’s own construction based on two university student surveys reported in Lipset (1965) and available through the ROPER Center’s Latin American Databank .

system, with a difference of 5.3 points (on a 1–10 scale) between the most leftist (PC-Chi) and the most rightist partisan student group (Conservadores). In the Uruguayan case, the spread of self-identifications is significantly lower (1.4 on a 1–5 scale), even among university students, who are generally presumed to be more politicized than average citizens. This supports characterizations of the Chilean party system as having been more programmatically structured than that of Uruguay at the time. Organizationally, fractions were present in both traditional Uruguayan parties, and depending on their own relative electoral weight, were able to access state positions. Those positions, in turn, provided the basis for structuring clientelistic linkages with constituents. Fractions differed in their ideological and



Party–Voter Linkages in Chile and Uruguay

81

programmatic stances, although these stances do not seem to have been very salient in structuring party–voter linkages. Instead, fractions, usually through their highly influential leaders, provided the basis for strong identifications with a modicum of programmatic content (i.e., voters usually identified as Batllistas, Terristas, Herreristas, Blancos independientes, Wilsonistas, Ruralistas, etc.). Notwithstanding, party identifications were also very strong and were transmitted over generations (see e.g., Caetano et al.1987).11 Those identifications were rooted in the bloody and perennial civil wars of the nineteenth century (Devoto 1942). The much lower levels of electoral turnover in Uruguay compared to Chile point to the relative strength of Uruguayan partisan identities. The absence of successful outsiders in the system and the much lower incidence of new parties forming from splinter factions, were also consistent with the presence of strong partisan identities. In summary, fractions and agrupaciones in Uruguay established linkages to voters on the basis of a mix of clientelistic linkages and fractional identifications. The latter involved leadership-based appeals tied to particular ideological leanings and affinities. Parties, in turn, provided fractions with the capacity to pool resources (especially votes) with other fractions competing under the same lema. This subsequently translated into differential access to state office and resources useful for more clientelistic mobilization. At the aggregate level, political parties lacked programmatic coherence, and thus did not engage in programmatic mobilization. Yet, both traditional parties were able to mobilize constituents and, in particular, to organize fractional politics on the basis of strong and resilient partisan identifications. Fractions (and internal fractionalization) are thus important to understanding the logic of the system, but parties were also fundamental in shaping it (Buquet et al. 1998). Mobilization strategies were segmented along territorial and socioeconomic lines, but segmentation ran within and not between districts. In this regard, clientelism and partisan identifications were instrumental in mobilizing large and diverse electoral segments, but not clearly identified socioeconomic groups (cf. Rama 1971; Fa Robaina 1972; Aguiar 1984; Vernazza 1989). In general, it could be said that partisan identities as well as fractional leaderships were instrumental in mobilizing the upper and middle classes, especially in urban zones. In the countryside, rural caudillos organized around different programmatic platforms. Parties successfully engaged the middle and lower classes throughout the state via partisan identities as well as clientelistic linkages.

  11  At this time, the historical evolution and mid-century configuration of the Uruguayan party system clearly resembled that of Colombia. Hartlyn (1988) provides a similar account of that party system.

Segmented Representation

82

TRADITIONAL PARTY PID + DVS

Agrupación A Fraction A

Poor and Rural

Agrupación B Fraction B

MiddleUpper Class

Poor and Rural

MiddleUpper Class

Programmatic Candidate Traits Clientelism Party ID

Figure 2.2.  Territorial linkage segmentation in Uruguay (until 1973)

Uruguayan districts were large, encompassing different socioeconomic groups. The linkage strategies in pre-1973 Uruguay were therefore what I call the mixed type. Strategic harmonization at the party level was, in turn, minimal at best. Such minimal coordination was achieved by drawing on strong party identifications and institutional inducements, especially the DVS. Strategic harmonization was more extensive at the fractional level. Fractions pursued their clientelistic strategies through pervasive territorial organizations and decentralized clusters of political committees (clubes politicos) that played broker. They also provided essential campaign and communication tools in the pre-mass media era (Rama 1971). Fractions were also unified around powerful individual leaders, under whom pyramidal organizations of subaltern leaders and armies of political activists served. Political activists, who were often “owners” of neighborhood political clubs, could defect and switch to a different fraction or even another party, although they rarely did so. Switching was for the most part driven by the prospect of obtaining access to clientelistic side-payments under alternative fraction leaderships. Elections allowed party activists and fractions to signal their respective leverage in bringing votes to the fraction and to the party



Party–Voter Linkages in Chile and Uruguay

83

(“marcar el voto”). On this basis, they were granted access to state resources instrumental in clientelistic mobilization (Vernazza 1989). Figure  2.2 illustrates the patterns of territorial segmentation and strategic harmonization observed in pre-1973 Uruguay. The Historical Configuration of Party–Voter Linkages in Chile and Uruguay Concluding the stylized characterization of the pre-authoritarian party systems of Chile and Uruguay, Table 2.5 summarizes the narrative of each case. Chapter 6 returns to this characterization, and goes on to describe the comparative transformation of each system in the post-authoritarian period. The historical narrative presented in this chapter suggests that single linkage characterizations of parties’ electoral appeals risk misrepresenting how parties and party systems operated in the past. My narrative also suggests that the higher levels of social inequality observed in Chile (particularly in the countryside) were associated with greater levels of socioeconomic and territorial segmentation of party–voter linkages until 1964. When segmentation decreased due to increasing reliance on a single (programmatic) type of linkage for mobilizing different social classes, the system polarized and finally broke down in 1973. In Uruguay, the socioeconomic and territorial segmentation of electoral appeals was arguably lower. Yet, there too, both traditional parties and their fractions resorted to complex combinations of linkages for attracting voters. In both cases, parties were able to harmonize linkage strategies. Yet harmonization was higher in Chile, demonstrating that more linkage segmentation does not necessarily lead to greater difficulties in harmonizing appeals to distinct groups of voters. In Uruguay, the DSV and strong partisan identities contributed to generating some decentralized harmonization at the party level, but leadership-based harmonization was more efficient at the fraction level. Fraction leaders’ ability to access state resources was crucial for enabling them to achieve strategic harmonization.

PA RT Y– VO T E R L I N KAG E S I N P O S T- AU T HO R I TA R IA N C H I L E A N D U RU G UAY, 1 9 9 0 – 2 0 1 0 I now turn to a characterization of party–voter linkages since Chile and Uruguay re-democratized, in 1989 and 1984 respectively. I  first provide a

Table 2.5.  Comparative analysis of mobilization strategies in Chile and Uruguay (pre-1973)

Segmentation of linkages along socioeconomic lines Main linkage strategy and core constituency

Secondary linkage strategy and peripheral constituency

Chile until 1964

Chile 1964–1973

Uruguay until 1973 Uruguay until 1973 (parties as (fractions as relevant units) relevant units)

High

Low

Moderate to low

Moderate to low

Leader-based (sometimes programmatic appeals) in middle and upper classes

Party identification in all socioeconomic groups)

Programmatic (cleavage mobilization) in urban sectors/upper classes

Programmatic (cleavage mobilization) across classes and urban/ rural sectors regarding national-level politics Clientelism (“patronage, Clientelism (“patronage, pork, constituency pork, constituency service”) in rural sectors service”) in local-level and urban poor politics Between Within

Does segmentation occur within or between districts? Is the strategy harmonized? Yes Mechanisms for achieving Self-harmonized: harmonization programmatic mass party organization/ vertical patronage networks Segmentation/ Harmonized dual harminization outcome

Yes Self-harmonized: programmatic mass party organization/ vertical patronage networks Pure (harmonized)

Clientelism, in middle and lower classes and rural sectors Within

Within

Yes Harmonized by fraction leaders: clientelistic resources, candidate traits

No, or only minimally Decentralized harmonization: (achieved by default through party identification, and mechanical effects of DVS electoral system) Non-harmonized mixed

Harmonized mixed



Party–Voter Linkages in Chile and Uruguay

85

stylized analysis for the contemporary period equivalent to that offered in the previous section for the pre-1973 period. The final part of the present chapter also provides a historical narrative (parallel to the one offered in the appendix for the pre-1973 period), examining each case’s authoritarian experience and transition to democracy. This narrative is useful for contextualizing the analysis unfolding in the remainder of the book. Post-Transitional Party–Voter Linkages in Chile Contradicting early predictions that had dismissed the likelihood of a significant transformation of programmatic competition after the transition to democracy (Valenzuela and Scully 1997; Valenzuela and Valenzuela 1986; Valenzuela 1999; Scully 1992), the legacies of the 1964–1989 period radically reshaped the configuration of the post-transition party system. Available evidence suggests that a “political cleavage” structured from above and revolving around regime issues replaced the class cleavage along which parties competed and developed programmatic linkages to constituents in the pre-authoritarian period (Hagopian 2002; Torcal and Mainwaring 2003; Ortega 2003; Alcántara and Luna 2004). This substitution is also consistent with the processes of renovation and ideological moderation that Chilean political parties of the left and the right went through during the transition, as a result of both international and national events (Roberts 1992; Posner 1999; Siavelis 1999). In other words, in the context of shrinking opportunities to develop and compete along a socioeconomic divide, Chilean elites seem to have successfully activated a regime divide centered on the democratic-authoritarian question, mobilizing support on the basis of the deep political divide that emerged under the Unidad Popular (UP) and further consolidated under Pinochet. The new cleavage runs across class lines, partially blurring the patterns of class-based voting observed in pre-1973 Chile (Torcal and Mainwaring 2003; Ortega 2003). The regime divide still provides the basis for political identification with either the Concertación de Partidos por la Democracia (comprising, among others, the Partido Demócrata Cristiano, the Partido Socialista, the Partido por la Democracia, and the Partido Radical) or the Alianza Por Chile (dominated by Renovación Nacional and the Unión Demócrata Independiente). Those two partisan alliances consolidated as the dominant coalitions that structured party competition in the post-transitional period. The bi-polar logic of competition has been shaped and reinforced by the binominal electoral system, which provides strong incentives for maintaining two electoral pacts at the congressional level.

86

Segmented Representation

However, my evidence also qualifies that standard view of the post-transitional system. The priming of the regime divide has also blurred programmatic differences (and partisan differentiation) between and within the two electoral pacts, and has likely further reduced the capacity of individual parties to cultivate strong partisan identities. In this regard, the mobilization of the regime divide might partially explain the electoral weakening of Chilean parties.12 The evolution of programmatic and party-identity linkages at the national level has been coupled with increasing levels of linkage segmentation at the district level. Whereas programmatic alignments and strong partisan identifications centered on the authoritarian-democratic divide are still consequential in some electoral constituencies, congressional elections are more often centered on candidates’ personal reputations and non-programmatic linkages to their constituents. Although this personalism is not new, its interaction with shrinking programmatic differences and scarce resources for parties’ national–local brokerage networks reinforces the weakening of partisan organizations. Before 1973, patronage and clientelism used to reinforce partisan structures and subcultures and the programmatic stances of the parties’ rank and file, inducing high levels of party discipline and internal homogeneity (self-harmonization). In marked contrast, in the post-transitional period clientelism contributes to the opposite outcome:  the weakening of partisan organizations as cohesive and harmonious institutions. At the national level, presidential campaigns have also become increasingly personalized and less partisan. Post-Transitional Party–Voter Linkages in Uruguay In Uruguay, the authoritarian period did not result in a drastic reconfiguration of party politics. During the dictatorship virtually all significant parties and party fractions joined forces in opposing the authoritarian regime, thus inhibiting the configuration of a strong regime divide in the post-transitional   12  This is especially so in the aftermath of Pinochet’s detention in London in 1998 and subsequent court decisions against him in Chile. New evidence brought to light in 2002 and 2003 by the “Riggs” case, which connected the dictator’s personal economic fortune to privatization bribes and other serious irregularities, has further diminished the right’s willingness to identify with his legacy. Evidence of the financial corruption of the regime has made it virtually “indefensible,” even for those who tolerated and even praised Pinochet’s human rights violations as necessary for putting the country back on track after the “Marxist threat.”



Party–Voter Linkages in Chile and Uruguay

87

period. When democratic contestation resumed, the main features of the traditional system were restored. Political parties continued mobilizing supporters on the basis of strong partisan identifications, support for individual leaders, and the fractional brokerage of clientelistic side-payments and “pork.” Programmatic differentiation among fractions belonging to the same party was also maintained. The traditional parties nonetheless faced two challenges. First, they were challenged by a growing leftist opposition, which acquired a powerful partisan identity (Frenteamplismo) as a result of the political persecution of the party’s rank and file during the dictatorship.13 After 1990, when the FA won the municipal elections in Montevideo, home to almost half the country’s voters, the leftist challenge became particularly tangible and consequential. Second, governments headed by traditional parties had to confront increasing budget deficits and inflationary pressures. The fiscal crisis restricted clientelistic mobilization, while forcing governing parties to implement structural reforms. Despite the specter of a crumbling economy and increasing international pressure, Uruguay lacked a strong advocate of neoliberal reform. Once in office, however, the traditional parties embarked on gradual attempts at state reform. Given the omnipresent Batllista or statist/redistributive ideology and significant policy feedback from import-substitution industrialization (ISI), these reforms were, not surprisingly, extremely unpopular. In this context, the FA gave political expression to a “veto coalition” of ISI beneficiaries, while drawing on a reinterpretation of Batllismo to attract votes from sectors increasingly alienated from the traditional system. Ideologically, this coalition advanced a statist platform, advocating a rollback of reformist legislation and blocking it through direct democracy mechanisms (see Filgueira and Papadopulos 1996; Lanzaro 2004; Luna 2004; Castiglioni 2005; Altman 2011). As a result of these trends, programmatic mobilization emerged along the state–market divide, separating the two traditional parties from the FA. The latter enjoyed steadily increasing electoral support. It had appropriated the defense of the state-centric Batllismo, and was able to mobilize discontent against neoliberal reformers. The consolidation of the state-centric versus market reform divide yielded a bi-polar logic of competition among three major parties but between two “ideological families” (Moreira 2000a). The aging, but still present, traditional partisan identities differentiated Partido Nacional and Colorados within the traditional, now also “pro-market,” party family. Differentiation of identities was also present within each camp   13  Such persecution produced the disappearance of approximately 170 adults and three children, as well as massive exile or imprisonment of leftist activists and leaders.

88

Segmented Representation

due to persisting fractional politics. The bi-polar dynamic nonetheless endured, thanks in part to the constitutional reform of 1996, which established a runoff presidential election. In the run-offs of 1999 and 2009, the two traditional parties, despite their differences, joined forces against the FA’s presidential bid.14 Although very important, the strengthening of programmatic mobilization in the system, as well as the consolidation of Frenteamplismo and the relative weakening of Blanco and Colorado partisan identities within the electorate, cannot overshadow the persistence of other types of linkages. Budget constraints and state reforms significantly constrained the functioning of political machines brokering clientelistic linkages from the national to the local level. Yet, municipal politics continued to be the locus for crafting clientelistic and constituency–service linkages, especially with poor constituents. Those linkages were implemented in a more decentralized and personalized fashion than in the pre-1973 period. As in Chile, this transformation also fostered the personalization of politics, albeit to a lesser extent. Personalization in Uruguay is more limited to the municipal level, while in Chile, in part because of the electoral system, congressional candidates also compete more frequently as individuals than as party members. These dynamics also had implications for party fractions. Especially in the case of traditional parties, the effective number of fractions decreased significantly, converging to an average of two relevant fractions in each party (Piñeiro 2007). The 1996 constitutional reform’s restrictions on pooling votes among lists also reinforced this transformation, strengthening parties vis-à-vis fractions.15 In the FA, however, the effective number of fractions increased, reaching an average of five by the early 2000s (Piñeiro 2007; Piñeiro and Yaffé 2004). Yet, Frenteamplismo provided the leftist electoral coalition with a strong partisan identity, one arguably more important than its constituent parts (see Lanzaro 2004).

A P P LY I N G T H E A NA LY T IC A L F R A M EWO R K T O P O S T- T R A N SI T IO NA L PA RT Y SYS T E M S The configuration of party–voter linkages observed in both countries, after their transitions to democracy, clearly diverges from those observed in the   14  These efforts succeeded in 1999 and failed in 2009. In 2004, Tabaré Vázquez of the FA was elected with a majority in the first round.   15  Until this reform, each lower-chamber list could be nested under different senate lists (see Piñeiro 2007).



Party–Voter Linkages in Chile and Uruguay

89

Table 2.6.  Socioeconomic linkage segmentation in Chile (post-1989) Resources used in the appeal Agent making the Symbolic appeal Candidate Party

Candidate traits (all segments)*+ Partisan identity

Material: public policies

Material: private or club goods

Electoral platform (upper and middle classes)* Programmatic

Particularistic (popular sectors)+ Clientelistic

Key Bold: Predominant strategies in system. + Predominant strategy for peripheral constituency. * Predominant strategy for core constituency.

pre-1973 period. Table 2.6 represents the relative incidence of different types of linkages in mobilizing distinct socioeconomic constituencies in contemporary Chile. The table seeks to describe the strategies of five of the six parties analyzed in my fieldwork (the Renovación Nacional, the Partido Demócrata Cristiano, the Partido por la Democracia, and the Partido Socialista). Those parties currently compete on the basis of candidate-based appeals. Although the regime divide splits the Concertación and the Alianza, and also correlates with moderate differences in terms of public policy preferences, programmatic divergences among individual parties are marginal and have tended to decrease over time. These parties’ candidates have thus started to compete on the basis of individual strategies, which they adapt to the particular characteristics of each district. In general, upper- and upper-middle-class districts are mobilized through combinations of candidate-trait and candidate-platform appeals. Meanwhile, in popular districts, particularistic mobilization also enters the mix. In Chile, one crucial difference between the pre-1973 system and the contemporary one is that individual candidates displaced political parties (with the partial exception of the UDI) as the main agents pursuing electoral mobilization. The role of partisan organizations in diffusing programmatic stands, and in extracting and distributing state resources to the periphery, has diminished quite significantly. The emerging configuration has been very detrimental to parties for two reasons. On the one hand, the self-harmonizing linkages of the past have been replaced by new strategies that would ideally require a system of decentralized harmonization. On the other hand, parties have lost access to the symbolic and material resources that would enable such harmonization. The UDI, the exception to this rule, pursues a distinct strategy, which could be characterized as a highly segmented one. The UDI appeals to conservative groups and upper socioeconomic segments (the core constituency of the

Segmented Representation

90

party) on the basis of programmatic linkages. Those segments do not identify with the party very strongly, but consistently support the party and its leadership. The UDI has also developed a peripheral constituency, especially in the popular sectors of big metropolitan areas. UDI candidates appeal to these voters by combining particularistic mobilization (greater than other candidates usually pursue) and candidate-based traits. Moreover, the UDI also engages, as a party, in the clientelistic mobilization of its peripheral constituency, particularly when its candidates lack strength in their districts. The strategy of the UDI is fully described in Chapter 5. Regarding territorial segmentation, my evidence indicates that all Chilean parties implement linkage types that differ across districts, and that such differences reinforce the high socioeconomic segmentation of linkages observed across different social classes. I therefore characterize the contemporary system as one in which non-harmonized dual linkages predominate. Yet, in heterogeneous districts, segmentation within districts is also observed. Figure 2.3 presents the configuration of territorial segmentation observed in my fieldwork. In sum, the party-based appeals (whether programmatic or clientelistic) observed in Chile in the pre-1973 period have been replaced by candidate-based appeals, which make harmonization at the party level increasingly challenging

Chile Post-1990 Concertación

R E G I M E

Alianza

Party

Heterogeneous districts

Poor districts

Upper and uppermiddle class districts

Candidate traits Programmatic Partiularism

Figure 2.3.  Territorial linkage segmentation in Chile (post-1989)



Party–Voter Linkages in Chile and Uruguay

91

Table 2.7.  Socioeconomic linkage segmentation in Uruguay (post-1984) Resources used in the appeal Agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate

Candidate traits (all segments)+* Partisan identity*

Electoral platform

Particularistic

Programmatic (all segments)+*

Clientelistic (popular sectors)+

Party

Key Bold: Predominant strategies in system. + Predominant strategy for peripheral constituency. * Predominant strategy for core constituency.

for the party leadership. In terms of socioeconomic segmentation, the system has returned to the pre-1964 configuration, in which different types of appeals were deployed to mobilize different socioeconomic segments. This is also reflected territorially, with a more dual (and non-harmonized) pattern of linkage segmentation across districts re-emerging. Table  2.7 shows the linkage types that predominate in post-transitional Uruguay. In this case, programmatic representation has replaced clientelistic mobilization as the primary mobilization strategy. Parties’ positions regarding the state–market divide, along with the decline of traditional clientelistic linkages, have been very consequential in changing the competitive dynamic of the system. This change especially affects the political behavior of the middle classes and politically organized popular sectors. Yet, particularism and clientelistic mobilization at the local level, and leadership-based appeals, continue to be important in the system, especially for mobilizing less politically engaged citizens in the informal sector and across different socioeconomic segments of the electorate. As in Chile, it is important to distinguish system-level strategies from those implemented by the country’s most successful political party in recent elections (the FA). In this regard, the FA has been better able to segment its appeal to different social segments, whereas both traditional parties have confronted increasing constraints on maintaining their previously ever-present clientelistic machines. Unlike in the Chilean case, in Uruguay political parties remain the critical agents for structuring electoral appeals. Individual candidates compete as members of a collective organization, giving such membership primacy in their campaigns. In this way, Uruguayan parties self-harmonize their segmented electoral appeals.

92

Segmented Representation

Notwithstanding their capacity to self-harmonize linkage strategies, traditional parties’ capacity for pursuing a wide range of mobilization strategies (as they did in the pre-1973 era) has progressively diminished, as has their share of votes. The Blanco and the Colorados’s declining capacity to segment electoral appeals involves three detrimental developments, namely: both parties’ declining capacity to extract resources from the state to use in clientelistic campaigns; the erosion of the Blanco and the Colorado party-identification due to generational replacement and the growing hegemony of Frenteamplismo among the younger generations; and the emergence of the traditional party “family,” which was functional for pooling resources and competing against the FA, but which contributed to obscuring programmatic distinctions between the Blancos and the Colorados. In the case of the traditional parties, declining capacity to segment electoral appeals has also correlated with consolidation of more socioeconomically homogeneous social bases. In a nutshell, deprived of their former ability to use widespread PID-based mobilization or clientelistic side-payments, the Blancos and the Colorados lost their capacity to mobilize large numbers in the popular sectors. In comparison to Chilean parties, the traditional parties of Uruguay have less capacity to segment their appeals to different groups, and more capacity to harmonize. Joint consideration of the two cases of Chile and Uruguay suggests that neither segmentation nor harmonization alone is sufficient for electoral success. As argued in Chapter 5, the FA has recently proven better able to segment its electoral appeals, making significant electoral inroads into peripheral constituencies in recent years. The party mobilizes middle sectors and formal workers, as well as the cultural intelligentsia, on the basis of strong partisan identities and its historical programmatic platform. The FA has also been able to mobilize poor voters and the informal sectors by deploying a wide range of appeals, including: candidate-based appeals, particularistic linkages (usually structured around problem-solving initiatives at the local-community level), and in some cases differentiated programmatic appeals and policy platforms. In comparison to the UDI, the FA deploys a greater range of electoral appeals. It also mobilizes a more internally homogeneous social base in which the core and peripheral constituencies are less markedly differentiated than those of the UDI. Both parties have proven capable of effective harmonization, albeit by different means. While the purposive action of the party leadership is crucial in achieving harmonization in the case of the UDI, the FA relies more heavily on self-harmonizing mechanisms. Regarding territorial segmentation, Figure 2.4 shows my (abbreviated) characterization of the Uruguayan party system in the post-transitional period. Segmentation in the system has apparently increased, from the historically



Party–Voter Linkages in Chile and Uruguay Uruguary post-1990

93

Liberal vs Conservative

Traditional Parties

Frente Amplio State vs Market

Party

Agrupación B Fraction B equivalent to Agrupación A/FA

Agrupación A Fraction A

District 1

District 2

Particularism Programmatic PID Candidate Traits

Figure 2.4.  Territorial linkage segmentation in Uruguay (post-1984)

low levels observed before 1973. Yet, territorial segmentation still seems to be lower than in Chile, and is structured within districts, not between them. This configuration is consistent with a mixed-linkage scenario, in which politicians segment their linkage strategies to mobilize different constituent groups (usually pertaining to different socioeconomic groups) within a district. Differences across Uruguayan districts are not as sharp as those observed in Chile. Regarding harmonization, fractions continue to be important in the Uruguayan system. When fractions are considered the relevant unit, as in the pre-1973 era, harmonization is high. Yet fractions pertaining to the same party do not usually harmonize among themselves, and thus fail to completely coordinate their strategies at the partisan level. Therefore, if parties are considered the relevant unit, then even where the effective number of fractions within parties has decreased, harmonization is still lower at the party level than at the fraction level.

94

Segmented Representation

Table 2.8 summarizes the comparative analysis of the two systems outlined in this section. Chapters 3 and 4 present the empirical evidence justifying the coding of cases shown in Table 2.8. In Chapter 6, I return to each case’s coding, seeking to infer the causal dynamics explaining the variance observed over time and across cases. The remainder of this chapter presents a historical overview of the recent historical trajectory of each party system, as the online appendix does for the pre-1973 period.

R E C E N T EVO LU T IO N O F T H E T WO - PA RT Y SYS T E M S In this section, I  outline the main transformations that occurred in each country during the authoritarian period, during their transitions to democracy, and in the years afterward. Here I provide a linear historical narrative. Chapter 7 will use that narrative for formulating a causal argument about the transformation of party–voter linkages. Chile’s Bureaucratic-Authoritarian Regime (1973–1989) Chile’s authoritarian government is usually characterized as an instance of bureaucratic-authoritarianism, as defined by O’Donnell (1973) in his seminal work.16 Chile’s dictatorship, led by General Pinochet, fully expressed the “reactive” (repressive) and “foundational” phases characteristic of bureaucratic-authoritarian (BA) regimes (Garretón 1988). In this regard, from 1973 to 1989, Chile witnessed a socioeconomic revolution and a firm attempt to reshape traditional politics. The regime sought to do away with parties and political elites (“los señores politicos,” “politiquería”), evolving towards an exclusionary state-corporatist system of representation, and eventually to a “modern,” tutelary, and protected “democracy.” Although Pinochet failed to achieve his most ambitious political objectives, his dictatorship still significantly changed the historical configuration of Chilean politics and the country’s party system. The social bases of Chile’s historical sociopolitical arrangement were arguably dismantled, facilitating

  16  However, given its high degree of personalization in the figure of General Pinochet and the progressive consolidation of “one man rule,” the political structure of the regime more closely resembled previous (traditional) authoritarian governments in the region (Valenzuela and Valenzuela 1986; Garretón 1988).

Table 2.8.  Comparative analysis of mobilization strategies in Chile and Uruguay (post-transitional period) Level of linkage segmentation along socioeconomic lines Main linkage strategy Secondary linkage strategy and peripheral constituency Does segmentation occur within or between districts? Is the strategy harmonized? Mechanisms for achieving harmonization

Chile post-1990

UDI

Uruguay post-1990

FA

High

High

Moderate to low

Moderate

Particularistic, candidate Electoral platform traits Electoral platform, coalitional Clientelistic, candidate traits ID (regime divide)

Programmatic

Programmatic, PID

Candidate traits, particularism/ clientelism

Particularism, leadership, increasing PID

Between districts

Between districts

Within districts

Within districts

No

Yes

Yes, but more effectively at the fractional level Decentralized harmonization Harmonized by party Self-harmonized: required; minimum levels leaders: hierarchical leadership programmatic achieved on the basis of with access to centralized organizations/PID, regime divide inertia economic resources, “mystique” fraction leaders Segmentation/ Non-harmonized dual Harmonized dual Harmonized mixed/pure harmonization (harmonized) outcome

Yes, at the fractional and party levels Self-harmonized: programmatic mass-party organizations/PID, fraction, and national leaders Harmonized mixed

96

Segmented Representation

further transformations in the post-transitional period. In this regard, contemporary evidence supports Drake and Jaksic’s (1991) identification of the first post-authoritarian election in 1989 as a critical one. The electoral realignment of the Chilean system had begun before the dictatorship. According to Garretón (1988), the depth of the economic and political crisis under Allende intensified political polarization and mobilization. On the right and center, polarization and mobilization generated a tacit “consensus for termination” (of the UP government), which enjoyed relatively broad-based social and political support. Contemporary party system dynamics confirm the significant electoral realignment that had been taking place in the system before 1973. This relatively broad consensus moved the armed forces to center stage. According to Garretón (1988), it also had two significant consequences. First, it increased the range, depth, and duration of the repressive activities of the regime. Second, it strengthened the regime’s apparent mandate for reshaping and normalizing the “ill-conceived” body politic, and pursuing ambitious economic restructuring. The 1980 constitution, approved in a fraudulent plebiscite, represented the keystone of the “democracia tutelada” (protected democracy) project that the authoritarian regime envisioned for the country. The constitution sought to establish a “modern and protected democracy,” with the armed forces as the ultimate guarantors of the nation’s institutions (Valenzuela 1999). It also permanently banned leftist political groups and ideologies and set significant limits to political and civil liberties, empowering the National Security Council to replace any governmental authority they believed to threaten the institutional order or national security (Valenzuela 1999). Additionally, the constitution established the appointment of non-elected senators (making up a third of the chamber), created a strong presidency, and instituted an electoral system that significantly contributed to shaping the nature of party competition in the post-transitional period (see e.g., Siavelis 1999; Navia 2003). The 1980 constitution also set the pace and nature of Chile’s transition to democracy. The fragmentation and weakness of the political opposition to the regime and the relatively high support for the dictator also contributed to Pinochet’s capacity to follow his own timetable for democratization. The constitution “approved” in 1980 provided for a plebiscite in 1988, for which the commanders of the armed forces would pick a candidate to preside over the country for eight more years. In 1988, having withstood the recession and social protests of 1982–1983, and enjoying the benefits of sustained economic growth since 1985, Pinochet was “nominated” by the armed forces. Although a small, pro-democratic right reappeared in 1988, Pinochet, campaigning on



Party–Voter Linkages in Chile and Uruguay

97

Table 2.9.  Citizen’s justifications for their vote in the 1988 plebiscite Reasons for voting “No” Economic situation Human Rights Disapproval of Pinochet’s Government Return to democracy

72% 57% 39% 21%

Reasons for Voting “Yes” Order and tranquility Economic situation Pinochet Himself

49% 38% 30%

Anti-communism or anti-UP Democracy

16% 3%

Source: Reported by Varas (1991) on the basis of a public opinion poll.

economic growth, still enjoyed the substantial support of capitalist and business interest. Chile was then the only standing BA regime in the region, and was under international and internal pressures to celebrate a clean and fair plebiscite, as mandated by the 1980 constitution. Pinochet’s new term in office was therefore subject to a competitive plebiscite in late 1988. Catalyzed by the plebiscite, the democratic opposition to the regime finally coalesced, mounting an effective electoral campaign, and successfully defeated the regime on “its own terms” (Garretón 1988). With an electoral turnout of 97.22 percent, the “No” option supported by the opposition obtained 56 percent of the votes, while Pinochet obtained 44 percent. As shown by Varas (1991), whereas those who voted for the opposition did so on the basis of dissatisfaction with the economy and disapproval of Pinochet’s human rights violations and authoritarian rule, those who supported the dictator privileged the “order and tranquility brought by the regime,” as well as economic growth (see Table 2.9). Also, 30 percent of those who voted in favor of extending Pinochet’s rule manifested a personal allegiance to his leadership. This fact illuminates the significant realignment that took place in the Chilean system along pro- and anti-regime lines and the partial revitalization of the previously decaying right in the post-Pinochet era. The 1988 plebiscite results derailed the dictatorship’s plans for a delayed democratic transition, and cleared the way for negotiations between the regime and the opposition that led to the presidential and congressional elections of 1989 (Valenzuela 1999). In those negotiations, the democratic opposition (with the exception of the PC-Chi) “exchanged” acceptance of the legitimacy of the 1980 constitution for fifty-four constitutional amendments toning down its most salient anti-democratic features (Valenzuela 1999). These amendments notwithstanding, important “authoritarian enclaves” remained in place at the institutional level, conditioning the post-transitional system (Garretón 1988; Siavelis 1999; Navia 2003).

98

Segmented Representation

On December 14, 1989, presidential elections were held, and Patricio Aylwin of the Partido Demócrata Cristiano was elected with 55.2  percent of the popular vote. Aylwin represented the center-left Concertación de Partidos por la Democracia, which was created from sixteen political parties for opposing Pinochet in 1988. The dominant parties within the coalition were the Partido Demócrata Cristiano (DC), the Partido Socialista (PS), the Radical Party, and the new Partido por la Democracia (PPD) that had splintered from the PS. Former leaders of the then very influential MAPU (Movimiento de Acción Popular Unitaria) joined the PS and the PPD. Representing the rightist coalition dominated by the Renovación Nacional (RN) and the UDI, Hernán Büchi, a minister of finance under Pinochet, obtained 29.4  percent of the vote. Finally, Javier Errázuriz, an independent center-right candidate, garnered 15.4 percent. Although their electoral victory gave the opposition to Pinochet a majority in both chambers of Congress, it did not give the Concertación control of the Senate due to the presence of non-elected senators designated by the outgoing authoritarian regime. Despite apparent continuities, the party system that emerged from the BA regime diverged significantly from the pre-authoritarian one. As Drake and Jaksic put it, the political legacy of the authoritarian regime was “a stronger right, an enduring center, and a weaker [and divided] left” (1999, p.  14). Although later developments under the new democratic regime also contributed to reshaping the party-system and partisan competition in Chile, the different paths followed by each party “family” since the 1960s and 1970s and during the dictatorship also help to explain that outcome. On the right, the main party organizations had already merged in the 1960s into the National Party, which became progressively more authoritarian. Immediately after the military coup the National Party dissolved itself and adhered to the dictatorial regime. Both the Radical Democrats (a conservative offspring of the Radical Party) and the fascist Fatherland and Liberty Party followed the same trajectory, merging with the military regime (see Garretón 1988; Valenzuela and Valenzuela 1986; and Huneeus 2003). The “Group of the 24” constituted an exception to this rule. They chose not to participate in government and, together with members of the Partido Demócrata Cristiano, supported a constitution at odds with the one finally approved in 1980. Another group of rightist leaders constituted a “semi-opposition” within the regime, but were still willing to participate in government and state institutions (Valenzuela and Valenzuela 1986). Finally, in 1983 a conservative student movement born in the Universidad Católica (Gremialismo) allied with a group of “Chicago-boy” technocrats under the leadership of Jaime Guzmán to create the Unión Demócrata Independiente (UDI).



Party–Voter Linkages in Chile and Uruguay

99

In 1987, the more traditional rightist groups, with the exception of the National Party, coalesced to form a new party named Renovación Nacional. However, confrontations between Guzmán and the leadership of Renovación Nacional soon led the UDI to split back off from it. Renovación Nacional and the UDI would become the most popular parties within the right after the transition to democracy. The economic policies of the regime generally tended to fulfill the agenda of the right. Yet, the effects of the economic policies pursued under Pinochet at the same time contributed to reshaping the social bases of these parties, with financial interests and export-oriented producers after the transition newly overshadowing the traditional landowning elites predominant in the post-1973 right. In the center, once they were legalized under the political party law of 1983, the Partido Demócrata Cristiano’s most centrist faction resumed leadership of the party, seeking alliances with other centrist parties (like the Partido Radical) and advancing the center-left Concertación de Partidos por la Democracia (Garretón 1988). The situation of the party was somewhat uncomfortable. On the one hand, Demócrata Cristianos could bank on their relation to the Church and its role as the semi-legal public opposition to the regime. On the other hand, Demócrata Cristianos also carried the burden of having supported the coup as an unavoidable response to the UP, even though the party never accepted the authoritarian interpretation of the 1973 crisis as one of “regime and society” (Valenzuela and Valenzuela 1986; Valenzuela 1999). Suffering less political repression at the hands of the regime and less affected by the political exile of its members than parties of the left, the Partido Demócrata Cristiano never accepted the recess of party activities decreed by the government, and sought to maintain its partisan organization. It renovated its leadership, particularly after the death of Eduardo Frei in 1982 (Valenzuela and Valenzuela 1986). Indeed, even when the regime decided to dissolve all parties due to permanent violations of the recess, the Partido Demócrata Cristiano still managed to fill organizational spaces in civil society. In short, the party progressively acquired an oppositional role, which then translated into relatively high levels of popularity and organizational development (Garretón 1988). Two internal factions competed within the party. One opposed the “Christian” pursuit of the party’s “own [center-right] path” and promoted an alliance with the center-left. This faction prevailed, paving the way for the construction of the Concertación de Partidos por la Democracia and the consolidation of the Partido Demócrata Cristiano as the most organized and best-structured national party in the early post-transitional period (Garretón 1988). The left, in turn, constituted the focus of Pinochet’s repression, and suffered from the dispersion of the militants that managed to escape the regime and went into exile. The mass protests of 1983 signaled the resurrection of

100

Segmented Representation

social opposition to Pinochet. In the wake of these protests, the opposition movement (particularly made up of labor and shantytown dwellers) gained autonomy from the political parties. This autonomy, facilitated by the political repression and exile of leftist party leaders, contributed to breaking the “backbone” of the pre-1973 system and its traditional alliances (Garretón 1988; Drake and Jaksic 1991). The general mistrust of party politics and low degree of popular recognition of the need for parties and a multi-party system (see e.g., Aldunate 1985) also contributed to social movements’ autonomy and posed a fundamental challenge for partisan elites who were used to operating in a context of overlapping social and political alignments (see Garretón 1988 and 1989). Indeed, according to Garretón (1989), opposition parties faced three interrelated challenges. First, they needed to reconstitute their relationship with their social bases, which became more autonomous after the 1983 protests and fragmented, especially in the wake of the socioeconomic transformations described below. Second, they needed to reconstitute themselves internally after having been disarticulated as a result of repression, clandestine functioning, and political exile. Third, opposition parties needed to coordinate a unified front to oppose Pinochet. Internally divided and mistrustful of the Partido Demócrata Cristiano, meeting these challenges long eluded the parties of the left. This particularly applies to the PS, which lacked a unified national leadership and was broken into two mainstream factions (Almeyda and Altamirano). These factions were divided over a range of issues concerning the evaluation of the UP experience, the need for renovation and re-evaluation of political democracy, and the role of the Leninist strategy within the party. Different experiences during political exile also contributed to fostering divisions within the left (Funk 2004). These factions also differed in their degree of support for a “coalition for change,” which was eventually to be crafted through an alliance with the center (Garretón 1988; Roberts 1998). Various currents finally unified around two socialist parties, each named after its General Secretary. The Nuñez Partido Socialista (a heir of the Altamirano faction) had gone through a deeper renovation process and more fervently favored alliances with the center. Meanwhile, the Almeyda Partido Socialista took a more orthodox stance and favored the traditional alliance with the PC-Chi. With the 1988 plebiscite acting as a catalyst, the Almeyda Partido Socialista decided to abandon insurrectional strategies against Pinochet. Yet both factions pursued different legalization strategies. On the one hand, the Nuñez Partido Socialista created the Partido por la Democracia, in conjunction with other small leftist groups and some center and center-right figures and independents (Garretón 1988). Former MAPU leaders also played an important role in the creation of



Party–Voter Linkages in Chile and Uruguay

101

the PPD. This party, created as a way to bypass the ban on the PS, became very popular during the campaign for the plebiscite. Although it was dominated by the Nuñez Socialistas, it was clear that it had greater electoral potential than the PS. It gathered the allegiance of young people, renovated leftists and center-leftists, and former Demócrata Cristianos who saw their old party as too traditional and ideological (Garretón 1988). In turn, the Almeyda Socialistas joined the PC, the Izquierda Cristiana, and other smaller groups to create the Partido Amplio de Izquierda Socialista (PAIS). The PAIS registered as a political party by the end of 1988, illustrating the emergence of two lefts within the left: a renovated and an orthodox one (Roberts 1998). The PAIS finally broke down, and the Almeyda Socialistas joined the Concertación as the PS in 1989. In the party Congress of 1990, both factions of the Partido Socialista reunited, while the PPD continued to operate as an independent organization. Meanwhile, the PC-Chi developed a clandestine organization during the dictatorship and was better able to resist repression than other parties of the left, enjoying a greater insertion in the popular sectors that started to organize and mobilize against the regime in the early 1980s (Garretón 1988; Oxhorn 1995). However, a split also developed between its mostly traditional and pragmatic domestic leadership and the one living in exile, which came under the strong influence of Moscow (Roberts 1998). Until 1980, the party maintained its gradualist standpoint and opposed insurrectionary practices, blaming the UP for the coup and resulting BA regime (Garretón 1988). Yet, criticized by the USSR and its own exiled members, who supported a military strategy to confront Pinochet, the PC-Chi finally opted for using “all forms of struggle” to challenge the dictator (Garretón 1988; Valenzuela and Valenzuela 1986; Roberts 1998). In this context, the party drew closer to the MIR and created its own revolutionary militia, the Frente Patriotico Manuel Rodríguez (FPMR). Although this change moved the party away from its traditional constituency in the labor movement, it resonated well with the student and shantytown-dweller movements that were mobilizing and protesting in the streets against the regime (Garretón 1988). When the opposition decided to confront the regime from within its own institutional framework, a deeper division emerged within the PC-Chi. The party finally decided to support the “No” vote in the 1988 plebiscite. Yet, its dual nature as an insurrectionary and “systemic” party was accentuated after the plebiscite (Garretón 1988). The FPMR became autonomous, while small dissident groups with pro-renovation postures withdrew from the party (Roberts 1998). This produced a strong internal debate within the party that was not resolved by the creation of the PAIS. Ultimately, the PC-Chi did not join the Concertación and became increasingly isolated in the post-transitional

102

Segmented Representation

system, due to both the highly constraining electoral system and its own strategic choices (Roberts 1998; Funk 2004). The exclusion of the PC-Chi, along with the renovation (“socialdemocratization”) of the socialist camp, enabled the emergent Concertación de Partidos por la Democracia to reassure the political and economic right that the essentials of market reforms would remain in place. This reassurance was arguably important in advancing democratization. However, as Drake and Jaksic predicted, it also posed a fundamental dilemma for the upcoming Concertación governments, which were ever more strained by simultaneously seeking to “maintain the essentials of the free-market model to placate the upper class, and to address the accumulated social demands from the unprivileged majority” (1991, p. 7). To understand the role of business interests and the popular sectors in post-transitional Chile, it is important to keep in mind the economic transformations that occurred during the dictatorship. In the countryside, the process of agrarian reform initiated under Frei continued and deepened during Allende, destroying the old “latifundio” land system without creating a substitute (Martínez and Díaz 1996). This provided Pinochet with autonomy from the landed interests historically represented by the Conservative Party. The authoritarian regime also launched an extensive repression of the peasant movement. Together with destruction of the latifundio system, this repression resulted in demobilizing the main social actors in the countryside. With Pinochet’s only partial return of expropriated land to its historical owners and distribution of land to smallholders lacking access to credit and technology who were therefore quickly forced to sell, a new market for land emerged in the rural sector (Martínez and Díaz 1996). This in turn led to the consolidation of a new business-oriented class in the countryside. The semi-feudal system that had prevailed until 1964 was replaced by one of capitalist accumulation, enhancing the competitive capacity of medium-size agricultural businesses (Martínez and Díaz 1996). According to Silva (1996), the representation of distinct “capitalist coalitions” in the rural and urban sectors under Pinochet was more fundamental to upper sectors’ homogenous and consistent allegiance to the regime than their sense of “threat from below.” Although financial interests and export-oriented sectors predominated over local-market industrial producers in those capitalist coalitions, Pinochet was still better able than his counterparts in other countries to gain the allegiance of the traditional right, including rural capitalists in the south. As a result, Chile’s capitalist class and its allies on the political right consistently supported the dictator, significantly affecting the timing and nature of the democratic transition and its aftermath, and the founding the post-transitional right. Even when the economic crisis hit hard in 1983 and a



Party–Voter Linkages in Chile and Uruguay

103

significant share of Chilean enterprises went bankrupt, Pinochet’s recrafting of his policy-making coalition and the pursuit of more pragmatic neoliberal policies allowed him to maintain a firm grasp on the economic and political right. In this context, the Partido Demócrata Cristiano’s Alianza Democrática initiative, which attempted to attract support from the right for early democratization by allying with conservative sectors while excluding the left, went nowhere. A mix of political repression and economic and state reforms (privatization, decentralization, and market liberalization), as well as concurrent elite strategizing, led to the dismantling of national organizations representing the popular sectors. The economic transformations, as well as the harsh repression of the peasant movement, had the same effect on the organized popular sectors in the countryside (Kurtz 2004). The disarticulation and fragmentation of the peasant movement in the rural sector paralleled that of the union and popular movements in the cities and mining areas (Oxhorn 1995; Roberts 1998; Posner 2004; Posner 2008; De la Maza et al. 2004). Moreover, due to strategic motivations and the prevalence of “moderated” views within the Concertación’s leadership, the parties of the center-left decided not to mobilize their political grassroots in the popular sectors, further contributing to their organizational decay and demobilization (Posner 2004; Oxhorn 1995; Roberts 1998). Uruguay: The Bureaucratic-Authoritarian (BA) Regime and Transition to Democracy (1973–1984) In contrast to the Chilean case, the Uruguayan BA regime failed to pursue a “foundational project.” As a result, the authoritarian regime was less influential in shaping post-transitional politics and refurbishing the nature of partisan competition and the party-system (Rial 1984; Caetano and Rilla 2003; González 1991 and 1995). Accordingly, democratic elections resumed in 1984 (although with particular candidates periodically banned and with other restrictions) under the old constitution of 1967. The two-and-ahalf-party-system of 1971 was also “mostly restored” after the “authoritarian parenthesis” (González 1991). The recrafting of party–voter linkages observed in post-transitional Uruguay in fact owed much more to the political economy of the 1990s than to the 1973–1984 period. During its first years in office (1973–1975) the military’s general goal was to demobilize civil society and political parties, and especially to repress and exterminate the left. The regime was regarded as a “commissarial dictatorship,” and both the government and the opposition perceived it as an emergency

104

Segmented Representation

measure that was not aimed at radically altering the political system of the country (Caetano and Rilla 2003; González 1991). As with other significant programmatic divides in Uruguay, the regime divide ran through both traditional parties. The FA, in turn, uniformly opposed the regime. The most popular fraction of the Partido Colorado in 1971 (List 14 headed by Pacheco and Bordaberry) supported the coup and provided technical staff to the military regime. Meanwhile, its List 15 and Batllismo opposed the regime. Within the Partido Blanco, the situation was different. While the Blanco minority (led by General Aguerrondo) supported the regime, the Wilsonistas openly opposed it, forcefully denouncing human rights violations internationally. Given the repression of the left, the Partido Blanco and its exiled leader, Wilson Ferreira, thus became the regime’s most vocal opponent (González 1995). Once the situation was “under control,” a second stage of the regime began wherein the Junta attempted to create a new political society for the country (1975–1980). The decision to consolidate the regime led to increased repression from 1975 and 1978, ultimately making Uruguay the country with the most political prisoners per capita in the world (SERPAJ 1989). This foundational project sought to do away with traditional parties and looked forward to creating a party known as Partido del Proceso, representing the regime. According to Caetano and Rilla (2003), in so doing the military ignored the centrality of parties in Uruguayan society. However, the Junta did not overlook the fact that any significant attempt at political reform would need to be ratified through a popular plebiscite. Therefore, in a procedure unprecedented at the time, the military in 1980 submitted their foundational project to a popular vote. The plebiscite occurred in a context of “structural fraud,” given the lack of room for developing an opposition campaign. However, respecting the civic tradition of the country, the votes were counted fairly. The result was a spectacular defeat of the military’s project (Gillespie 1986; González 1991). Remarkably, while in 1971 the pro-regime fractions of the two traditional parties had gathered 37 percent of the vote, in the context of repression and structural fraud, the military in the 1980 plebiscite was able to obtain only 6 percent more. The electoral defeat of the military opened the way for a democratic transition. The political parties contributed to that result by mounting an underground campaign to counteract fear of repression. Unlike in Chile, all significant fractions took a stand on the plebiscite, forming an almost monolithic opposition to the regime. The results once again demonstrated the strength of the party system in society, and showed how impossible it would be to legitimize a system doing away with political parties. Acknowledging its failure, the military Junta initiated a transitional process (1980–1984), which ended with the elections of 1984. After its electoral defeat, the military



Party–Voter Linkages in Chile and Uruguay

105

took one year to sketch a transitional plan. It decided to hold concurrent primary elections for parties (excluding the FA) in 1982, and a national election in 1984. In contrast to Chile, the primary elections of 1982 were held in a context of a very deep economic crisis, which further boosted popular discontent with the authoritarian regime. Electoral turnout in the primary elections reached 60  percent of the electorate. In those elections, the fractions representing the opposition to the military regime gathered 70 percent of the vote within the Partido Colorado and 76  percent within the Partido Blanco (González 1995). The 43 percent support for the military obtained in the 1980 Plebiscite thereby plummeted to 20  percent by 1982. Meanwhile, the FA’s leadership asked for a blank vote in order for its supporters to “mark” its votes, but partially due to communication failures and tactical errors, this strategy proved inefficient. After the 1982 elections, formal negotiations between the elected party authorities and the Junta began in what came to be known as the Pacto del Club Naval. In this sense, Uruguay represents a clear instance of a pacted democratic transition. In the negotiations, a military Junta “weakened” by political defeat and economic chaos succeeded in obtaining just two conditions. First, they temporarily limited the capacity of the new civilian government to modify the top leadership of the armed forces. Second, they banned Wilson Ferreira’s participation in the election (Gillespie 1986). This banning of its main leader from the election triggered the Blancos’s withdrawal from the negotiations. Acknowledging the legitimacy problems that this withdrawal would entail for the transitional pact, the military was forced to accept the FA as a legitimate negotiating partner (Gillespie 1986). In exchange for legalizing the party, the military imposed the proscription of its most radical fractions and the banning of General Líber Seregni, the FA’s historical leader, from participation in the 1984 election. However, while Ferreira remained imprisoned until the election was held, Seregni was allowed to campaign actively. Once the results of the 1984 elections were known, both Seregni and Ferreira accepted the victory of the Partido Colorado. President Sanguinetti’s decision to grant amnesty to all political prisoners without exclusions symbolized the relatively rapid resumption of civil control over the country (González 1995). The 1984 election had a “restorative” character, epitomized by both the low volatility index observed between the elections of 1971 and 1984 and the internal defeat of pro-military fractions within both traditional parties (González 1995, pp. 155–7). Despite extensive market liberalization and the social effects of the 1982 crisis, including increased social inequality, the economic policies of the

106

Segmented Representation

Uruguayan BA regime did not produce the massive sociopolitical effects observed in Chile. As suggested by Castiglioni’s (2005) reconstruction of social policy making during the BA regime, in sharp contrasts to Pinochet’s achievements in Chile, substantial reforms in Uruguay were hindered by the dispersion of state power and by the ideological influence of Batllismo within the military apparatus. That influence crystallized into a variant of the National Security Doctrine that understood social development (instead of pure repression and economic restructuring) as the remedy for subversion, even while understanding the latter as also having been facilitated by “bad” and “corrupt” political leadership (Castiglioni 2005). The Uruguayan BA regime’s more impersonal system of rule by the military as an institution, which included a decentralized and dispersed allocation of authority, periodic rotation of generals appointed to the presidency, and the automatic retirement of older officers, also weakened the regime in the long run (Caetano and Rilla 2003; Castiglioni 2005).

P O ST- T R A N SI T IO NA L PA RT Y- SYS T E M S A N D P R O G R A M M AT IC PA RT Y– VO T E R L I N KAG E S Chile: 1989–2010 In a regional context where parties and party systems have experienced serious disruptions in several countries, the Chilean party system in the post-transitional years has been exceptionally robust and exhibited exceptionally high levels of electoral continuity (Mainwaring and Scully 1995; Hagopian 1998; Stokes 1999).17 Indeed, in line with their previous predictions, Valenzuela and Scully (1997) have argued for the system’s impressive continuity by comparing the local vote for major political parties in 1967 and 1992 (see also Valenzuela 1991; Scully 1992). Their findings in this regard appear particularly remarkable given the socioeconomic transformation of Chilean society, the strength of political repression directed against partisan organization during Pinochet’s regime, and the reach of the institutional reforms introduced in the 1980 constitution.

  17  The interpretation of these results, however, has been challenged by Montes et al. (2000) and by the recent evidence presented in Torcal and Mainwaring (2003), Ortega (2003), and Altman (2004), suggesting important discontinuities in the social base of Chilean political parties.



Party–Voter Linkages in Chile and Uruguay

107

Political parties in Chile therefore seem to have successfully adapted to the institutional and socio-structural constraints that have shaped electoral competition in the post-transitional period (Roberts 1992; Torcal and Mainwaring 2003), making this party-system a positive and noteworthy exception in Latin America (Mainwaring and Scully 1995; Hagopian 1998; Stokes 1999). This successful process of strategic adaptation and ideological moderation, many conclude, has fostered pragmatism and contributed to Chile’s “model” transition to democracy. According to this view, that transition combined structural reform with a consensual system of political negotiation and compromise, structured around two political coalitions:  a center-left (Concertación de Partidos por la Democracia) and a center-right (Alianza Por Chile). The strong incentives and restrictive conditions introduced by the Chilean “binomial” electoral system helped to keep these two coalitions in place.18 In contrast to these claims, a series of empirical evidence suggests that significant disruptions have taken place in the post-transitional party-system. As Montes et al. argue compellingly, “the dominant view that Chilean parties are strong has been overstated” (2000, p. 796).19 As shown in Graph 2.1, the weakening of Chilean parties has coincided with declining electoral turnout in both presidential and congressional elections. This phenomenon is significantly associated with age, with younger voters disproportionately unregistered (see Riquelme 1999; Toro 2007). In addition, blank and invalid voting has been on the rise. In Uruguay, these phenomena have been stable at significantly lower levels, always remaining below 5 percent with mandatory turnout exceeding 94 percent of registered voters in every election. In addition to the weakness of Chilean parties “in the electorate” (Luna 2007), analysts also now frequently note such disquieting emergent features as increasing levels of political apathy and social discontent with politics and parties (see e.g., Huneeus 1998; Valenzuela 1999; Posner 1999 and 2004).   18  The peculiar Chilean electoral system combines a district magnitude of two with a d’Hondt formula of proportional representation. As a result, to gain both seats in a district, a party or pact needs to receive twice as many votes as its closest competitor. As a result of the relative high number of parties, the system introduces strong incentives for parties to create and compete on the basis of electoral pacts. Additionally, although the center-right coalition had been forecast to receive around 40 percent of the vote, according to the level of support for regime in the 1980 Constitutional Plebiscite, the system guarantees equivalent Congressional representation for each coalition. This institutional system was secured by introducing a supra-majorities requirement for any significant reform. These rules fostered a system of consensus and compromise, and for more than a decade precluded the possibility of constitutional reform despite general dissatisfaction with the system. See Allamand (1999), Siavelis (1999), Fuentes (1999); and Navia (2003).   19  See also Luna and Altman (2011).

Segmented Representation

108 100 95 90 85 80 75 70 65

Chile Valid/VAP% 60 55

Chile Voters/VAP% Uruguay Valid/VAP% Uruguay Voters/VAP%

50 First Second Third Fourth Fifth Sixth Congressional Congressional Congressional Congressional Congressional Congressional Election Election Election Election Election Election

Graph 2.1.  Electoral participation rates in Chile and Uruguay (post-transitional period) Source: Author’s own construction on the basis of data from Servicio Electoral (Chile) and Corte Electoral (Uruguay).

Paradoxically, parties’ disconnect from increasingly fragmented and disorganized societal interests contributes to explaining both these weaknesses and the notable stability and institutionalization of the post-transitional system (Montes et al. 2000, p. 795; Luna and Altman 2011). Competing in this post-transitional context, following President Aylwin’s term, the Concertación was able to get three presidential candidates into office: Eduardo Frei (1994–2000) elected with 58 percent of the vote, Ricardo Lagos (2000–2006) elected with 51.31 percent in the runoff election of 2000 against Joaquín Lavin of the UDI (48.69  percent), and Michelle Bachelet (2006–2010) elected in the runoff election of 2006 with 53.5  percent of the votes against Sebastián Piñera of the Renovación Nacional (RN) (46.5 percent). In each case, the Concertación succeeded in selecting a candidate who promised change within a context of stability and moderation. The election of Eduardo Frei and his government was popularly perceived as being based on the promise of modernizing the state while maintaining the economic model. The election of Ricardo Lagos, in turn, marked the return to La Moneda of a Socialist president seeking to provide a “human face” to the Chilean economic model. Lagos’ presidency was marked by three legacies:  a strong emphasis on public and infrastructural works, the widening of the social policy system



Party–Voter Linkages in Chile and Uruguay

109

(Huber et  al. 2010), and a series of corruption scandals detrimental to the image of the Concertación (Rehren 2008).20 The nomination and election of Michelle Bachelet also signaled significant change within the governing coalition. Bachelet, the first woman elected president in Chile, was also a Socialist Party activist. Her government was characterized by a new deepening of the social protection matrix, including the universalization of basic pension benefits (Huber et al. 2010; Pribble 2013). Politically, Bachelet sought to implement a model of “citizens’ government,” distancing herself from political parties and traditional party leaders of the Concertación. Yet, that model proved short-lived. Soon after her appointment, Bachelet was obliged to resurrect the “elitist” government style of her predecessors. Soon after taking power, she confronted two very salient public opinion crises: one caused by the student strikes of 2006 (the so-called “Revolución pingüina”), and the other by failed implementation of a re-engineering of public transportation in the capital (the “Transantiago” crisis). Moreover, from 2008 on the country faced a decline in economic growth due to the international financial crisis. Bachelet nonetheless completed her term with high popularity ratings historically unprecedented in Chile. Notwithstanding, the Concertación did not benefit from Bachelet’s popularity in the elections of 2009 and 2010. In the 2010 runoff election, Sebastián Piñera of the RN defeated Eduardo Frei of the Concertación (48.4 percent) and became the first president elected by the Alianza pact, with 51.6 percent of the vote.21 Piñera’s election thus marked the return of the center-right to the presidency.22 Uruguay: 1984–2010 Until March 2005, when the FA won the presidency, three Colorado presidents alternated with one from the Blancos in Uruguay. Despite the continued electoral dominance of the Colorados and Blancos until 2004, the left’s electoral support grew continuously, while the traditional parties lost votes. This turning of fortunes was the most important change in the party-system since 1984, or even 1971. Graph 2.2 illustrates this phenomenon. The first president elected after the restoration of democracy was Julio María Sanguinetti, the winner of the 1984 elections within the Partido Colorado. His   20  See Salas (2011) for an analysis of media coverage of the most salient corruption scandals under the Concertación.   21  See Munck and Bosworth (1998) for a similar assessment.   22  Navia and Joignant (1999), Morales (2008), Luna and Toro (2011), and Luna and Mardones (2010) present a detailed description of the last three electoral cycles.

Segmented Representation

110 100 90 80 70

Traditional Parties

60

Challenger Parties

50 40 30 20 10

1942 1946 1950 1954 1958 1962 1966 1971 1984 1989 1994 1999 2004

0

Graph 2.2.  Electoral evolution of traditional and challenger parties in Uruguay (1942–2004) Source: Author’s own construction on the basis of data from Buquet (1999) and Corte Electoral.

government faced two basic tasks. First, he had to assure a peaceful transition to democracy and consolidate the democratic regime. Second, he needed to tackle the long-term socioeconomic problems troubling Uruguayan society. In the first respect, many analysts deemed his tenure successful, particularly after the approval of two amnesty laws that “resolved” the issue of human rights violations during the dictatorship: first for political prisoners held by the military regime, and second for the military officers involved in violations of human rights under the dictatorship (González 1995). The second of these laws was drafted by a group close to Blanco leader Wilson Ferreira, who was determined to support the government and provide governability. In 1987, with human rights groups floundering in their efforts to oppose the passing of the law, the left mobilized actively against the law granting immunity to human rights violators that had been approved by both traditional parties in 1986. The left participated in a petition campaign to force a binding referendum on whether the law should stand. Although the referendum results



Party–Voter Linkages in Chile and Uruguay

111

ultimately supported the law, the campaign benefited the left in two important ways. First, it helped complete the FA’s “traditionalization,” through which it progressively acquired the kind of strong partisan identity that the other two parties had inherited as a legacy from the armed confrontations of the nineteenth century (Caetano and Rilla 1995; Gallardo 1995; Queirolo 1999). Second, it proved that the instruments of direct democracy included in the constitution could be used for opposing legislation and effectively mobilizing popular support (Altman 2011). In terms of policy-making, Sanguinetti promoted a government of “national intonation” seeking to construct a broad alliance including not only the Blancos but also the FA in oversight roles. However, the honeymoon was brief. In the economic realm, inflationary pressures quickly resumed, with inflation reaching an annual average of 100  percent. According to data presented in González (1995), by the end of 1988 13 percent of Montevideo’s adult population thought the country was improving, 55 percent perceived it as stagnating, and 31 percent as declining. In turn, by 1989, presidential approval was below 20 percent. In this context, alternation between the Colorado and Blanco parties seemed the most likely outcome of that year’s election. However, in 1988 Wilson Ferreira died, creating a succession dispute. Luis Alberto Lacalle, Herrera’s grandson, quickly prevailed over Carlos Julio Pereira (from the MNR) and Alberto Zumarán, the party’s Wilsonista presidential candidate in 1984. Lacalle’s leadership was built around a renovated Herrerismo, a tendency that had pivoted ambivalently around Wilsonismo since 1985 (Caetano and Rilla 2003). Lacalle also built his career on the basis of continuous face-toface work in the interior, setting up a powerful political apparatus with local instances in every town and village. In 1989, leadership disputes over presidential candidacies triggered significant events within both the FA and the Partido Colorado’s List 15. In the FA, the most popular sector in the 1984 primary, the moderate List 99 (originally formed in 1971 by progressive splinters from List 15), abandoned the FA to create a new center-left party:  the Nuevo Espacio. According to González (1995), this confirmed the existence of two lefts: a radical one within which the Marxist-Leninist parties dominated, and a moderate one represented by Nuevo Espacio.23 Meanwhile, the remainder of the Partido Colorado’s original List 15 had competed since 1984 under the fraction Batllismo Unido. In 1989, Jorge Batlle challenged Sanguinetti’s nomination of his vice-president, Enrique Tarigo, for   23  See Altman et  al. (2008) for an argument on how these “two lefts” might persist today within the FA.

112

Segmented Representation

the 1989 election. This conflict led to a very heated primary that was won by Batlle. After this internal confrontation, Batllismo Unido split, yielding the Foro Batllista, led by President Sanguinetti, and the List 15, headed by Batlle. Programmatically, the Foro Batllista had a social democratic leaning, while the List 15 deepened its pro-market stance. Organizationally, while the List 15 was organized through informal networks coalescing around the figure of Jorge Batlle, Foro Batllista succeeded in developing a powerful political apparatus by distributing central and local state resources. The political apparatus of the Foristas, like that of the Herreristas, represented one of the best organized and most efficient political organizations for the reproduction of political support on the basis of “pork,” patronage, and clientelism. Such clientelist organizations gave both factions a great deal of political leverage in the post-transitional period, particularly until 1999. Despite being abandoned by List 99 in the elections of 1989, the FA was ultimately able to retain its 1984 share of support. Together with Nuevo Espacio, non-traditional parties gathered 30  percent of the vote. More importantly, the FA succeeded in winning the municipal government of Montevideo, the capital and traditional urban geographic base of the party, with 48 percent of electoral support. In the 1989 election, and during his ensuing term as mayor, Tabaré Vázquez consolidated his support base within FA, progressively challenging Liber Seregni, the leftist coalition’s historical leader, as well as other emergent leading figures, like those of Danilo Astori and Mariano Arana. Lacalle’s government started with a call for a Coincidencia Nacional (National Agreement), which once again proved short-lived. The inter-party agreement was vague and, unlike in 1984, did not include the opposition’s participation in ministerial positions. Growing inflation and declining real wages led to a sharp decline in the President’s popularity ratings. After just six months they were as low as Sanguinetti’s had been at the end of his term. During his first years in office, Lacalle promoted liberalizing and privatizing measures, which were crystallized in the Ley de Empresas Públicas approved in Congress. However, this law was challenged by a referendum promoted by state unions and the FA and then supported by Foro Batllista. In December of 1992, the citizenry rejected the law (79.1 percent against), leaving the government without one of its main policy planks. Additionally, during this period five unsuccessful proposals to reform the pension system were introduced. Meanwhile, thanks to his governing style in Montevideo, Tabaré Vázquez became a highly popular leader and a central player in the election of 1994. In those elections, Sanguinetti, from the Partido Colorado, succeeded in winning the presidency again, while the FA was re-elected to the municipal government of Montevideo. However, at the presidential level the electorate this time was divided in almost equal thirds, with the Partido Colorado obtaining



Party–Voter Linkages in Chile and Uruguay

113

32.3 percent, the Partido Blanco 31.2 percent, and the FA 30.6 percent. This outcome catalyzed the traditional parties’ creation of a stronger coalition after the election. The FA’s “almost victory” constituted a watershed in the country’s recent history, increasing the perceived costs of losing office for both traditional parties. While before 1994 losing office meant that the other traditional party would win, but without a majority, now the implications were far more drastic: both traditional parties would be left out of office. That prospect weakened the incentive structure for traditional parties to act as “free riders” (Filgueira and Filgueira 1997a). That strategy was conducive to frequent stalemates in decision making, and allowed the traditional party in opposition to benefit from the votes lost by the incumbent. Past co-participation practices had mostly been based on “pork” distribution and were frequently not tied, as they were this time, to an explicit reformist agenda. Interestingly, the traditional party coalition crystallized only when the growth of the left seriously called into question their capacity to maintain the presidential office. In this context, along with proposed reforms in other areas (social security, education, and health), both traditional parties promoted changing the electoral rules from a simple plurality election to a runoff electoral system that would continue to prevent the FA from entering into office. The constitutional reform enacting a runoff electoral system was approved in a plebiscite with extremely close results (50.45  percent in favor versus 49.55  percent against). The rationale behind the establishment of this new provision seems clear, as its most important justification was to prevent the victory of the left. Although counterfactual, it seems extremely plausible to think that, had the rules not been changed, the FA would have won the 1999 election, as it obtained almost 40 percent of the votes in the first round while the traditional Blanco and Colorado parties obtained only 22  percent and 33 percent respectively (Buquet 1999). The constitutional reform also called for party primaries to be held, a provision requiring unique candidacies in each party, and the elimination of vote pooling for congressional lists (Piñeiro 2007). These additional provisions in the reform were expected by some FA leaders to remove some of the sources of traditional parties’ flexibility and adaptation capacity. Danilo Astori and Liber Seregni from the FA therefore favored the reform, openly contesting Tabaré Vázquez’s opposition to it. This internal dissent within the party ultimately led to Seregni’s resignation from the party’s presidency and strengthened Tabaré Vázquez’s leadership. After approving the 1996 constitutional reform, as well as pension reform, the Blanco–Colorado coalition remained in place until the end of Sanguinetti’s term. That coalitional experience consolidated a bi-polar logic of programmatic competition structured around a traditional party

114

Segmented Representation

family (Blancos and Colorados) and a leftist family (FA and Nuevo Espacio).24 These two “ideological families” subsequently clustered together consistently in the face of critical events, especially the elections of 1999–2000 and 2000– 2004.25 In 2002, Nuevo Espacio merged with the FA.26 During this period, as noted by Chasquetti (2008), a typical presidential term was composed of two cycles, one of cooperation (in which an “automatic” majority existed in Congress due to the coalitional agreements struck by fraction leaders, usually outside Congress) and one of conflict (due to coalitional breakdown). Meanwhile, the left partially withdrew from the policy arena (channeling opposition by sponsoring direct democracy rebuttals of reformist legislation) and left the troubled coalition to pursue its timid reform attempts. In 1999, for the first time since the transition, the same party was able to maintain the presidency for two successive terms. Although the FA won the first electoral round, the coalition of Blancos and Colorados was able to defeat the left in the presidential runoff. This time, after successfully contesting the primary against Foristas, it was Batlle and List 15’s turn to get their candidate into office. Although Lacalle successfully won the Blanco primary thanks to his political apparatus (defeating an adversarial Juan Andrés Ramírez who competed on the basis of corruption charges against Lacalle and Volonté), the process of internal competition fatally wounded his candidacy. This led to a historical defeat of the Partido Blanco, which only obtained 22 percent of the vote. Within the FA, the primary between Vázquez and Astori once again confirmed the absolute primacy of the former in the fight for leading the leftist coalition. The Nuevo Espacio contested the election on its own, obtaining half of its original support, and quickly allied with the FA after the first round. Once inaugurated in 2000, President Batlle (like the Blanco Luis Alberto Lacalle in 1989) confronted the challenge of sticking to his reformist agenda while seeking to generate a congressional majority and reproduce electoral support using traditional means of clientelism and patronage. In 2002 and

  24  For further analysis of the coalitional logic that crystallized in the post-transitional period see Chasquetti (2008).   25  The only antecedent of cross-family alliances took place during the campaign leading to the Public Enterprises referendum (1992) in which the leading faction of the Partido Colorado opposed the Partido Blanco, establishing an implicit alliance with the FA. However, traditional parties were still free riding at the time, and the alliance was neither formally established nor informally recognized.   26  Julio Sanguinetti, the country’s President at that time, crafted the term “ideological family.” See Moreira (2000b) for a complete review of the debate and a concurring argument for conceptualizing party competition this way.



Party–Voter Linkages in Chile and Uruguay

115

2003 the country experienced its harshest economic crisis ever. Politically, the crisis paved the way for a comfortable victory by the FA. In November 2004, the FA was elected to presidential office in the first round, gathering 50.7 percent of the vote. Meanwhile, Jorge Larrañaga, who had triumphed over Lacalle in the Blanco primary, obtained 34.3  percent. The Partido Colorado hit an all-time low, obtaining only 10.5 percent of the national vote. In 2009, José Mujica of the FA was elected president in the runoff election, with 52.4 percent of the votes, defeating Luis Alberto Lacalle of the Partido Blanco (43.5 percent). Pedro Bordaberry, of the Partido Colorado, obtained a mere 17 percent of the vote in the first round that year.

SUM M I N G U P : T H E C O N T E M P O R A RY C O N F IG U R AT IO N O F PA RT Y– VO T E R L I N KAG E S I N C H I L E A N D U RU G UAY In Chile, the authoritarian regime was critical in redefining the nature of programmatic linkages observed in the system, as well as the overall levels of socioeconomic segmentation, after re-democratization. After the transition to democracy, the authoritarian–democratic divide consolidated as the primary competitive dimension, displacing the state–market and the liberal–conservative divides. Parties also gained autonomy from societal interests, and moderated their programmatic positions. In the meantime, the citizenry grew increasingly detached from political parties and politics. In Uruguay, the evolution of linkage patterns owed more to the political economy of the post-transitional period than to the authoritarian regime. The latter was crucial in solidifying the Frenteamplista partisan subculture, but polarization and competition around the state–market divide consolidated in the mid to late 1990s. Neither the regime nor the liberal–conservative divide were strongly politicized. The system realigned around two party families (traditional parties versus the FA) based on their respective political and economic positions, with the FA standing for opposition to the liberal market reforms attempted by the traditional parties. Since the transition, voters from Chile’s lower classes have shown lower levels of issue congruence and party-identification than their Uruguayan counterparts or upper classes in either country. Socioeconomic segmentation of programmatic linkages is therefore higher in Chile. Much the same can be said of linkages based on party identity. Chilean voters, particularly those with lower educational levels, also seemed to vote more on the basis of candidate traits.

116

Segmented Representation

In the Chapters 3–5 I provide evidence that justifies the stylized description of party–voter linkages in both cases. For each case, I  seek to illustrate the levels and types of observed linkage segmentation. I  also describe how different parties in each system compare regarding their capacity to harmonize segmented linkage strategies.

3 Socioeconomic Segmentation of Party–Voter Linkages in Post-Transitional Chile and Uruguay

I N T R O D U C T IO N This chapter traces the nature, evolution, and socioeconomic stratification of programmatic party–voter linkages in contemporary Chile and Uruguay. There are three reasons for this focus. First, according to the recent literature, Chile and Uruguay have the greatest levels of programmatic structuring in Latin America (Kitschelt et al. 2010; Luna and Zechmeister 2005). These findings suggest that programmatic linkages are central, if not prevalent, in both systems. Second, analyzing programmatic linkages and their socioeconomic differentiation is especially important for grasping the distributive consequences that segmented linkage strategies can produce. Finally, empirical information regarding other types of linkages is difficult to gather, particularly when working at the national level and seeking to cover a considerable time-span. Notwithstanding, the chapter also presents some evidence pertaining to other types of linkages, particularly party identification, and how they are socioeconomically stratified. In Chile candidate-based voting at the national level is more significant than in Uruguay, and is also examined in this regard. Finally, on the basis of expert survey results, I assess the extent of platform and clientelistic linkages in each system.

M E T HO D O L O G IC A L S T R AT E G Y:   P R O G R A M M AT IC L I N KAG E P R OX I E S Measuring and comparing the levels of programmatic structuration (i.e., the presence of programmatic linkages in a party system) is difficult and

118

Segmented Representation

controversial (Luna et al. n.d.; Kitschelt n.d.). No systematically applied technique for doing so is available, while multiple data limitations are present (Powell Jr. 2004). In response to these methodological difficulties, I have constructed empirical proxies for the nature and levels of programmatic structure. My main analytical strategy has been to use a series of case-specific confirmatory factor analyses to test the empirical configuration of three potential competitive divides: regime, moral/religious values, and state/market. These factors are then used to map the distribution of partisan groups and the internal coherence of partisan placements at both elite and voter levels. In this respect, my methodology is similar to that applied by, among others, Barnes and Kaase (1979), Inglehart (1984), Dalton (1988, ch. 7), Kitschelt (1994), Kitschelt et al. (1999), and Moreno (1999). Given my interest in the social distribution of linkage strategies, I examine the internal ideological cohesiveness of partisan groups with varying educational levels, using these as a proxy for socioeconomic status.1 Therefore, the analysis provides some insights on the current availability and equity of distribution of partisan ideological linkages when it comes to parties’ connections to socially unequal groups in each society. It is important to stress and clarify at the outset that, at the citizen level, the evidence draws on responses only from those who declare themselves to have voted for a given party. Especially in a context of high citizen alienation with parties, this cross-section might not be entirely representative of the full voting population. To provide a comparative and diachronic analysis of both systems, I analyze a series of confirmatory factor analyses performed on data available for congressional elites representing the 1993–1997 and 2001–2005 legislatures

  1  No standard measure of socioeconomic status is available across the different surveys I analyze. Of course, levels of education are not a perfect proxy for socioeconomic status or class, and they introduce some analytical complications, which I address below. Notwithstanding, I consider this to be a reasonable proxy for socioeconomic status in both cases. Each sample was split into three educational groups. Where a variable measuring the years of education attained by the interviewee was available, I constructed terciles. When such a variable was not available and educational levels were provided, three levels were kept: primary education (complete and incomplete), secondary education (complete and incomplete), and university education (complete and incomplete). For simplicity’s sake, in presenting my results I  show only the relative positioning of the first and third groups/terciles. The intermediate group/tercile, for which results are not shown, usually falls between the other two, displaying intermediate levels of programmatic structure. While in the case of the Projeto Cone Sul these levels were pre-defined, and it was not possible to reconstruct the original measure (the variable distinguishing between those who had finished or attended primary, secondary, or terciary education), terciles of the variable measuring the years of formal education for each individual were computed for the rest of the datasets. For the sake of consistency, I refer hereafter to “education groups.”



Socioeconomic Segmentation of Party–Voter Linkages

119

in Chile, and the 1995–1999 and 2005–2009 legislatures in Uruguay. For both countries, the data derives from the Encuesta de Elites Parlamentarias (PELA) of the Universidad de Salamanca, whose surveys were conducted in 1994 and 2006 in Chile, and in 1996 and 2005 in Uruguay. The same methodology was then replicated for electorates in 1988, 1996, and 2006. For these three years, a comparable survey was conducted in each country. The first of those datasets corresponds to the Projeto Cone Sul (Universidad de Campinhas), which surveyed political attitudes using an identical questionnaire in Argentina, Brazil, Chile, and Uruguay during 1988. The second dataset was constructed on the basis of the Third Wave of the World Value Survey (WVS), which was conducted in both countries during 1996. The last citizen waves for Chile and Uruguay correspond to the fifth round of the WVS, carried out in both cases in 2006. The six samples are representative of the urban population in each country and were selected on the basis of a random multi-level (clustered) sampling procedure. Finally, I apply a series of procrustes analyses, a variety of multidimensional scaling, to compute an issue congruence estimate for each case, cross-section, and time period.2 For each survey, a set of several questions covering each of the three divides was included, making this group of databases ideal for my analysis. In the WVS (1996 and 2006) the degree of preferred state intervention to provide for those in need (social policy) and to increase equality in society (as opposed to state withdrawal to promote competition) were imputed to represent the state/market divide. In the case of the Projeto Cone Sul, five items measuring the degree of preferred state intervention in the economy, education, health, public transport, and the financial system were included. A second set of questions in the WVS covers moral and religious issues, reflecting respondent’s preferences on abortion and legislation of divorce. In comparing results derived from the WVS to those from the Projeto Cone Sul surveys, which did not include these questions, items on the latter’s questionnaire regarding church attendance and assessment of the important of the Catholic Church in society were used. To measure the regime divide, two indicators were used:  one measuring democratic legitimacy and the other measuring the eventual trade-offs of democracy in terms of social and economic “disorder.” With the general exception of the PC-Chi, the number of observations available for every cross-section of party identifiers by educational group exceeded twenty in every survey.

  2  The online appendix includes a non-technical introduction to procrustes analysis and its proposed application to estimate issue-congruence.

Segmented Representation

120

Taking advantage of the fact that identical questionnaires were conducted in both Chile and Uruguay at the elite level, and in 1988, 1996, and 2006 at the citizen level, I computed case-specific confirmatory factor analyses. For each divide I ran a single factor analysis and generated factor scores for the first factor, which usually correlated significantly with each of the specific variables imputed. Mean scores and standard deviations were then obtained for each relevant group, constructed on the basis of partisan affiliation at the congressional level, declared vote-intention, and individual level of education.3 The analysis proceeds as follows. To assess the amount of partisan polarization along each competitive divide, I first present a series of ANOVA results. In the absence of independent salience measures, I  ran a series of ANOVA procedures, complemented through a Bonferroni test, to identify pairs of parties whose means were significantly different in a given factor. These measures will provide a proxy for the relative importance of each divide in every measure, and identify those factors in which parties present sufficient internal consistency and external divergence to produce significant results. To facilitate the display and interpretation of factor results, the online appendix presents matrixes that display the pair-wise ANOVA results based on Bonferroni tests at a 0.05 confidence level. I discuss the relevant results of these tests in the next section. Subsequently, I present programmatic mappings of parties’ congressional representatives’ mean position along the three divides, as well as equivalent mappings of the relative positioning of party voters (average positions, as well as mean positioning of the lower and upper educational group). Finally, I present the results of the procrustes analysis for each case and time period. The summary estimates generated through the procrustes analyses offer a glimpse into the degree of issue congruence observed in each case across time and for different socioeconomic groups.

T H E EVO LU T IO N A N D S O C IO E C O N OM IC SE G M E N TAT IO N O F P R O G R A M M AT IC L I N KAG E S I N C H I L E ( 1 9 8 9 – 2 0 1 0 ) A N D U RU G UAY ( 1 9 8 4 – 2 0 1 0 ) This comparative and diachronic design focusing on both voters and elites seeks to complement previous research restricted to elites (Hagopian 2002; Alcántara and Luna 2004), to one-shot studies (Moreno 1999; Zechmeister 2001; Rosas 2001; Luna and Zechmeister 2005, Kitschelt et al. 2010), and to  

3 

Documentation for these analyses is provided in the online appendix.



Socioeconomic Segmentation of Party–Voter Linkages

121

single-case studies (e.g., Torcal and Mainwaring 2003; Luna 2004; Luna 2008). The evidence I present here is consistent with previous more thorough analyses of issue congruence and programmatic mappings in both countries (Luna 2006).4 In the following sections I  present summary results for which full statistical documentation and technical notes are included in the online appendix. Additional analysis of left–right self-identifications and their substantive content are also included there. Here, I restrict my analysis to three potential programmatic divides mentioned above. However, because the sample sizes of those who declared having voted for a given party segmented by education groups are very small, this latter analysis should be considered exploratory. Significant Partisan Differences Across Programmatic Divides Through one-way ANOVA analyses and the application of Bonferroni post-hoc tests, one can assess the extent to which groups of voters and congressional representatives with different partisan affiliations overlap or diverge in statistically significant ways. Table A-3.1 in the online appendix presents the results obtained for the Chilean electorate, considering the 1988, 1996, and 2006 data. Table A-3.2 presents equivalent information for Uruguay. The diagonal of both tables displays the number of observations (declared voters) on which calculations were based for each party. In Chile, the regime divide is the only one along which significant partisan differences are observed, particularly in 1988 and 1996. Those differences consistently split the parties of the Concertación (PPD, PS, DC) from those of the Alianza (UDI, RN).5 In the Uruguayan case, all three divides at some point significantly split the FA from the traditional parties. Yet, the configuration of partisan differences over time is particularly telling. In 1988, no significant differences were observed. In 1996, significant differences emerge along all

  4  From a comparative point of view, at both the elite and citizen levels, Chilean parties are on average more pro-market, more religious, and less favorable to democracy than their Uruguayan counterparts. See Alcántara and Luna (2004) and Luna (2006) for comparative data on both cases, and Luna (2004) and Luna (2008) for programmatic mappings and evolution during this period for Uruguay and Chile respectively. Luna and Zechmeister (2005) and Kitschelt et  al. (2010) place both cases in a broader comparative perspective (applying measures of issue congruence and programmatic structure at the aggregate level). Luna (2006) presents estimates of issue congruence for both cases, both over time and for different socioeconomic cross-sections.   5  The Partido Demócrata Cristiano is represented by the acronym “DC”.

122

Segmented Representation

three divides. These differences then decline in 2006 to just one (liberal–conservative) divide. Tables A-3.3 and A-3.4 replicate the ANOVA analysis at the congressional elite level for the first and last round of the PELA database (Chile: 1993–1997 and 2006–2010; Uruguay: 1993–1997 and 2006–2010). In the case of Chile, all three divides produce significant splits in the 1993–1997 period. The regime divide, the only one significant at the citizens level, does not lead to significant splits at the elite level in the 2006–2010 period. The state and the moral divides, meanwhile, did continue to differentiate parties. Interestingly, whereas the divisions created by the state–market divide are consistent with the Alianza (pro-market) versus Concertación (pro-state) axis, the liberal–conservative divide created splits not only between the three Concertación parties and the UDI, but also within the Concertación (dividing the more conservative DC from the liberal PPD and PS). The PPD and the PS also differed significantly from the more conservative RN. In the Uruguayan case, the regime divide does not produce a significant split. Indeed, the “democracy is always preferable” question has no variance, with all respondents agreeing with that statement. A question asking respondents whether they thought parties were indispensable for democracy was therefore instead considered for the last measure. The only significant split in this case is between the FA, eight of whose congressmembers disagreed with the statement that parties were always needed for democracy, and the Blanco Party, whose congressmembers all moderately or totally agreed with it. In any case, the regime divide does not seem to be strong in the Uruguayan party system. This finding is consistent with the transitional dynamics observed in the country and described in Chapter 2. The state–market divide is the one that most strongly divides the FA (more statist) from the traditional parties (more pro-market), while between the traditional parties there is no significant difference. The moral divide also consistently splits the more liberal FA from the more conservative Partido Blanco and (only in 2005–2010) from the Partido Colorado. Relative Programmatic Placements of Partisan Electorates and Congressional Delegations I now present a series of mappings, showing the respective mean position of party voters and congressional delegations across the three programmatic divides. The analysis of the relative positioning of different parties (distance between points on each map) should be guided by the previous ANOVA results. In this regard, although potentially interesting, partisan placements



Socioeconomic Segmentation of Party–Voter Linkages

123

Electorates’ First Wave (circa 1988) State-Regime Crossnational

Authoritarian --> Democratic

0.6

FA

UY

PPD

0.4

PN PC

PS

DC

0.2

PC-Chi Chi

0 –0.5

–0.4

–0.3

–0.2

–0.1

0

0.1

0.2

0.3

0.4

0.5

–0.2

–0.4

–0.6

RN UDI

–0.8 State --> Market

Map 3.1.  Chilean and Uruguayan voters’ positions along the state–market and authoritarian–democratic divides (1988) Source: Author’s own construction on the basis of Projeto Cone Sul.

that do not relate to significant ANOVA results should be analyzed with care. For the sake of space and simplicity, complementary maps to which the text refers are included in the online appendix. I first analyze the 1988 citizen-level results. Map 3.1 gives a first comparative glimpse of the relative placement of voters in Chile and Uruguay. It should be noted that these placements have been calculated on the basis of country-specific factor scores on the state–market and authoritarian–democratic divides.6 The information displayed in Map 3.1 (as well as in Map A-3.1 in the online appendix) clearly suggests that, on average, Chilean voters were significantly less statist and more conservative in 1988 than their peers in Uruguay. This apparently applies even to the voters of the PC-Chi, which is included for the sake of illustration here despite the small number of cases for it. The PS is the only, partial exception in this regard. Moreover, on both divides, differences between voters of different parties within each country tend to be smaller than the cross-national ones. The one significant exception is that, as shown in Map 3.1, Chilean voters differ significantly by party along the regime divide. Whereas voters of the Concertación lie within the pro-democratic region (along with all Uruguayan parties), voters of the UDI and RN have more authoritarian views. In sum, whereas   6  Map A-1 in the online appendix presents the corresponding information for the liberal– conservative and state–market divides.

124

Segmented Representation

the regime divide is the only one that consistently splits voters of different parties into different camps in Chile, the Uruguayan electorate presents no trace of programmatic structure along this divide. Uruguayan voters have similar programmatic preferences irrespective of the party for which they voted. Maps 3.2 and 3.3 present the results obtained when this analytic strategy is replicated at the congressional level. The data here is derived from the first measurements available: those for the 1993–1997 legislature in Chile and the 1995–2000 legislature in Uruguay. The evidence appearing on the two maps points to both significant continuities and relevant changes. As shown in Map 3.2, since at least the mid-1990s at the congressional level parties have staked out more clearly their programmatic orientations, clustering around ideological groups instead of by national origin. The UDI, RN, and Uruguay’s traditional parties have been significantly more pro-market than the parties of either the Concertación in Chile or the FA in Uruguay. Parties’ relative positioning on the regime divide point to important differences between the two party systems. In the case of Chile, the parties of the Concertación and the Alianza align on a diagonal going from more pro-market and authoritarian views (the UDI and RN) to more statist and pro-democratic orientation (PS, PPD, and DC). In the Uruguayan case, the regime divide does not produce sizable divergences between parties, making party differences more one-dimensional. A similar conclusion could be reached on the basis of the liberal–conservative divide included in Map 3.3. Whereas the RN, the UDI, and the DC are significantly more conservative than the PS and the PPD, partisan differences in the Uruguayan case are less prominent. One might well ask how the programmatic orientations of partisan congressional delegations relate to those of their voters. Maps A-3.2, A-3.3, A-3.4, and A-3.5 in the online appendix display the relative positioning of partisan electorates along each divide. Taking advantage of marginally greater sample sizes than those available through the Projeto Cone Sul, these maps include estimates for each party, as well as partial estimates for the most and least educated voters of each party. At least descriptively, this analysis might serve as a proxy for the amount of programmatic heterogeneity within each party’s electorate. For the sake of simplicity, I present country-specific maps, labeling the averages for voters of different educational levels within each party by appending LE (low-education) and HE (high-education) to each party’s acronym. As shown in Map A-3.2 and in Map A-3.4 in the online appendix, with only marginal exceptions, the Chilean electorate was not programmatically polarized along the state–market divide. The highly educated voters of the UDI and RN and the voters of the PC-Chi have the most extreme positions on this divide. The richer cross-sections of the electorate (irrespective of the party for



Socioeconomic Segmentation of Party–Voter Linkages

125

Elites’ First Wave (circa 1996) State-Regime Crossnational 0.8 UDI

Democratic -->Authoritarian

0.6

0.4 RN

0.2 FA

–1

–0.8

PN

–0.6

–0.4

0

PC

–0.2

PS

0

0.2

0.4

0.6

0.8

–0.2 DC

PPD

–0.4 Market --> State

Map 3.2.  Chilean and Uruguayan congressional representatives’ positions along the state–market and authoritarian–democratic divides (1993/1997–1995–2000) Source: Author’s own construction on the basis of PELA. Elites’ First Wave (circa 1996) State-Moral Crossnational 1 0.8

PS

PPD

Conservative (Religious)--> Liberal

0.6 0.4 FA 0.2 0

PC

–1

–0.8

–0.6

–0.4 RN

PN

–0.2

0 –0.2

0.2

0.4

0.6

0.8

DC

–0.4 –0.6 UDI

–0.8 –1 Market --> State

Map 3.3.  Chilean and Uruguayan congressional representatives’ positions along the state–market and liberal–conservative divides (1993/1997–1995–2000) Source: Author’s own construction on the basis of PELA.

126

Segmented Representation

which they voted) meanwhile are relatively more pro-market than their less educated counterparts. However, average partisan placements and those of low-educated voters produce no meaningful alignments on this divide. Virtually the same conclusion could be reached for the regime divide, but there are small significant differences between voters of the Alianza (and particularly of the UDI) and those of the Concertación (in particular the PS and the PPD). Finally, regarding the liberal–conservative divide (Map A-3.4), the most educated voters of the PS and those of the PC-Chi are the only partisan groups differing significantly from others, with a more liberal orientation. The remaining partisan groups do not clearly align around this divide. In sharp contrast to the results obtained in 1988, partisan electorates in Uruguay occupy significantly different positions on the three divides. At the same time, they have substantial internal heterogeneity. Such heterogeneity is particularly visible in the case of the PC with respect to the state–market divide, on which its most educated voters have a markedly more pro-market view than the party’s average, and also in the case of the FA on both the democratic–authoritarian and the liberal–conservative divides, on which the most educated voters have a more clearly pro-democratic (Map A-3.3) and liberal orientation (Map A-3.5). Despite its internal heterogeneity, according to the ANOVA results reported above, the FA average is significantly more statist, less authoritarian, and more liberal than those of either traditional party. Closing this overview, I now turn to the last round of measurements available at each level (2006).7 Maps 3.4 and 3.5 present cross-national mappings, produced on the basis of country-specific factor analysis results, for the congressional delegations and voters of each party. Map 3.6 clearly reveals two partisan families. The first (the PPD, the PS, and the FA) is liberal and more pro-state, whereas the second (the UDI and Uruguay’s Partido Blanco) is more socially conservative and pro-market. The Chilean DC and RN, as well as the PC in Uruguay, are situated somewhat off the diagonal. Whereas they align “appropriately” on the state–market divide, they are more conservative (the DC) or relatively more liberal (RN and the PC) than the parties they usually ally with. This suggests that mobilizing the potentially salient moral divide might lead to partisan conflicts within coalitions. Those conflicts would be

  7  For the sake of space, the results obtained for the authoritarian–democratic divide, which has been losing strength over time, are not shown here. To maintain equivalence, I incorporated that variable into the regime factor for the Chilean case. In Chile, four respondents still replied that some alternatives might be better than a democratic regime. The results are nonetheless included in the online appendix.



Socioeconomic Segmentation of Party–Voter Linkages

127

Elites’ Second Wave (rcirca 2006) State-Moral Crossnational 1 0.8

PS

Conservative (Religious)--> Liberal

0.6 PPD 0.4

FA

0.2

–1.2

–1

PC

–0.8

–0.6

RN

0 –0.4

DC

0

–0.2

0.2

0.4

0.6

0.8

1

–0.2 –0.4 PN

–0.6 UDI

–0.8 –1 Market--> State

Map 3.4.  Chilean and Uruguayan partisan representatives on the state–market and liberal–conservative divides (2006/2010–2005/2010) Source: Author’s own construction on the basis of PELA.

particularly detrimental in the Chilean case, due to the institutionally driven need for keeping broad coalitions in place.8 Map 3.5 displays partisan mappings at the citizen level. Evidently, cross-national and national partisan differences on the state-marked divide have shrunk over time. In this regard, the voters of the parties of the Alianza (Chile) and those of the traditional parties in Uruguay appear marginally more pro-market than those of the Concertación and the FA. Yet, with the exception of the voters of the DC in Chile, the distance between these partisan groups declined over time. Cross-national differences are more prominent on the liberal–conservative divide. In general, the voters of Uruguayan parties are more liberal than their Chilean counterparts. Still, within countries significant differences persist. The voters of the UDI and the DC in Chile are more conservative than those of RN, the PPD, and the PS. Again, partisan differences on this divide are at odds with the Alianza versus Concertación alignment. In Uruguay, the voters of the FA are significantly more liberal than those of the traditional parties.

  8  On the possible conflicts triggered by the activation of this alternative divide see for instance Wilson (2012).

Segmented Representation

Religious--> Secular

128

Electorates’ Third Wave (circa 2006) State-Moral Crossnational 0.2 fa

0 –0.4

–0.35

–0.3

–0.25

–0.2

–0.15

–0.1

0

–0.05

0.05

0.1

–0.2

–0.4 pc-CHI pn

ps

–0.6 pc-UY rn

–0.8 ppd

dc

udi

–1 State --> Market

Map 3.5.  Chilean and Uruguayan partisan electorates on the state–market and liberal–conservative divides (2006) Source: Author’s own construction on the basis of WVS. Electorates’ Third Wave (circa 2006) State-Moral Chile

udihe

1.5 1

Religious--> Secular

pche

0.5 dche

–0.6

–0.5

–0.4

–0.3 pc

ppdhe udihe

0

–0.2psle –0.1 dcle

pcle

pshe

0

0.1

rnle udile

ps –0.5 ppdle dc ppd

–1

rnhe

0.2

0.3

udi

–1.5 State--> Market

Map 3.6.  Chilean partisan electorates on the state–market and liberal–conservative divides (2006) Source: Author’s own construction on the basis of WVS.

Finally, Maps 3.6. (Chile) and 3.7 (Uruguay) map partisan national electorates by educational levels. Positions within partisan electorates vary significantly depending on voters’ level of education. Interestingly, in the cases of the PPD, the UDI, RN, and the PC-Chi, differences are visible on



Socioeconomic Segmentation of Party–Voter Linkages

129

Electorates’ Third Wave (circa 2006) State-Moral Uruguay 0.8

Religious--> Secular

0.6 fahe

0.4

pnhe

0.2

fa

pche

0 –0.3

–0.2

–0.1

fale

–0.2

pnle

–0.4

0

0.1

0.2

0.3

0.4

pcle

–0.6 pn

pc

–0.8

State--> Market

Map 3.7.  Uruguayan partisan electorates on the state–market and liberal– conservative divides (2006) Source: Author’s own construction on the basis of WVS.

the state–market divide (low-educated voters are more statist than their highly educated co-partisans). Meanwhile, all partisan electorates are internally heterogeneous regarding the liberal–conservative divide, with lower-educated voters opting for more conservative views than those of highly educated ones. In the Uruguayan case, party differences on the state–market divide have moderated, producing visible overlaps among constituents of different parties (e.g., the low-educated voters of the Partido Blanco and the FA). At the same time, the Partido Blanco is very internally heterogeneous where that divide is concerned, with its most-educated voters being significantly more pro-market than its low-educated voters. In terms of the liberal–conservative divide, low-educated voters are uniformly more conservative than their most-educated co-partisans. Yet, such internal differences are especially prominent in the FA. Procrustes Analysis: Multidimensional Issue Congruence Although instructive, the preceding analysis might appear impressionistic. In this section, I  summarize the results obtained from procrustes analyses. These analyses were conducted for each country for two periods of time for

130

Segmented Representation

which data at both levels is available.9 These analyses estimate the fit between the factor scores displayed in the previous section for the elite (1993/1997 and 2006/2010 for Chile; 1995–2000 and 2005–2010 for Uruguay) and those for the electorate (in 1996 and 2006 for both cases). Graph 3.1 displays the results obtained. Three possible specifications for each time period and country are considered, yielding twenty-four procrustes statistics. First, I estimated three-dimensional solutions, including all of the three programmatic divides (state–market, regime, and moral). Second, I  estimated a set of two-dimensional solutions, restricting the analysis to those divides that continue to be active, at least at the elite level: the state–market and the liberal–conservative divides. Third, in contrast to the orthogonal rotations most commonly assumed in procrustes analyses, I  allowed for oblique rotations, in keeping with the assumption that the different programmatic dimensions might be correlated, instead of independent. That assumption is more realistic, particularly in light of the frequency with which the partisan alignments in the mappings above were configured diagonally. Finally, I estimated a first set of solutions considering all voters, and a second set in which the analysis was restricted to the least-educated voters of each party. The gap between the two solutions, if any, would indicate socioeconomic stratification (i.e., segmentation) of programmatic linkages. Graph 3.1 compares the results obtained across specifications, which are fairly robust. To facilitate assessment of these results, each bar in the graph represents the amount of issue congruence observed. Thus, higher bars indicate greater levels of issue congruence. I recoded each result as 1-(procrustes   9  Procrustes analysis is a multi-dimensional scaling technique that compares and maximizes the fit between objects described by two different, at least two-dimensional, matrices. One of the matrices is considered the source and the other the target matrix. Every object is represented by at least two coordinates in each matrix. If every object has the same coordinates in both matrixes, the overlap is complete. Unlike non-parametric techniques such as correspondence analysis, the procrustes technique computes fit statistics. The procrustes statistic is defined as the size of the residuals (the difference between objects in the target and source matrices) relative to the variation in the target variables. That statistic can be read as a 1-(R-square) measure, with results approaching 0 signaling the best fit between the target and the source matrices. For that reason, with procrustes statistics tending to zero, we can conclude that the relative position of a given set of objects in two multidimensional matrixes is fairly similar. The opposite becomes true as procrustes statistics approach 1. A classic example from geography helps illustrate this technique (see Cox and Cox 2001). Assume we want to compare the location of a set of towns on two maps: a modern map and one drawn in antique times. To analyze the accuracy of the antique map, we need to compare the scale of both maps, the relative (Euclidean) distance among towns, and the orientation of each map. To maximize the fit between both matrices (maps), procrustes analysis transforms the data, applying an optimal combination of three procedures: dilation (uniform rescaling), rotation (either orthogonal or oblique), and translation.



Socioeconomic Segmentation of Party–Voter Linkages

131

Levels of Issue Congruence = 1-(Procrustes Statistic)

1 0.9 0.8 0.7

Low education All

0.6

All oblique rotation

0.5

State–Moral LE State–Moral-All

0.4

Total if oblique rotation

0.3 0.2 0.1 0 Chile 1996–97

Chile 2006–06

Uruguay 1996–97

Uruguay 2006–05

Graph 3.1.  Summary of procrustes results obtained for different electoral cross-sections and model specifications (two/three dimensions, oblique/orthogonal rotations) Source: Author’s own construction on the basis of WVS and PELA data.

statistic). Chile obtains lower levels of issue congruence than Uruguay for both periods of time and across specifications. Yet, results in Uruguay were weaker in 2006 than in 1996, suggesting that issue congruence has declined over time. That decline was been especially marked for low-educated voters. In sum, segmentation and social stratification of programmatic linkages has increased in Uruguay. Segmentation has also increased in Chile, where moderate improvements in overall issue congruence (considering all voters) are matched by worsening of issue congruence for low-educated voters. Finally, comparing two- and three-dimensional specifications, both cases present greater levels of issue congruence when the regime divide is considered. Yet, that increase is marginal in the Uruguayan case, confirming that the regime divide does not significantly split party positions there. In Chile, the regime divide is particularly important when considering low-educated voters. When it is excluded, levels of issue congruence reach their nadir in that electoral segment. Across the board, oblique solutions produce greater levels of issue congruence, suggesting that, at least in these two cases where these three potential programmatic divides are considered, partisan positions are correlated across programmatic dimensions. Closing this section, Maps 3.8, 3.9, 3.10, and 3.11 graphically summarize the results obtained. Maps 3.8 and 3.9 compare the degree of issue congruence

Segmented Representation

132

Procrustes overlay plot stateel/regimeel 1

stateel/moralel PS

PS PS

DC

DC

DC

PPD

PPD

0

PPD

RN PPD

UDI

DC

PS

RN

UDI

RN

RN

UDI

UDI

–1 –1

0

1

regimeel/moralel 1 DC PPD DC

RN

0 UDI

PPD

PS

RN UDI

PS

–1 –1

0

1 Party Elites

Voters

Map 3.8.  Multidimensional issue congruence, Chile 2006 Source: Author’s own construction on the basis of WVS and PELA. Key: 1 = DC, 2 = RN, 3 = UDI, 4 = PPD, 5 = PS.

along the three programmatic divides in 2006. Also for 2006, Maps 3.10 and 3.11 display the bi-dimensional solutions (considering the state–market and liberal–conservative divides) obtained for low-educated voters in both systems. Coinciding with the results shown in Graph 3.1, the degree of observed issue congruence is lower in Maps 3.10 and 3.11 than in Maps 3.8 and 3.9, and in turn lower in Map 3.8 than in Map 3.9. The graphical display also identifies individual parties with higher or lower levels of issue congruence, which is taken as a proxy for measuring the scope of programmatic linkages in the system. Whereas in Uruguay distance between parties and their electoral bases seems to be relatively even across parties



Socioeconomic Segmentation of Party–Voter Linkages

133

Procrustes overlay plot stateel/moralel

stateel/regimeel

1 FA

FA

FA

FA

0 PN

PN PN

PN PC

PC PC

–1

PC

–1

0

1

moralel/regimeel 1 FA FA

0

PC

PC PN

PN

–1 –1

0

1 Party Elites

Voters

Map 3.9.  Multidimensional issue congruence, Uruguay 2006 Source: Author’s own construction on the basis of WVS and PELA. Key: 1 = PC, 2 = PN, 3 = FA.

(with the FA and the PC having marginally greater distances than the Partido Blanco), in the case of Chile parties differ in this regard. In particular, the UDI, but to some degree also the PPD and the PS, are characterized by highly incongruent programmatic profiles between the elite and the least-educated voters of each party. Cross-Validation: Interview and Expert Survey Evidence This section is based on interview data collected in both cases, as well as the results of an expert survey applied in both cases. The results of these

Segmented Representation

134 1

PS

0.5 DC DC

0

PPD

PS

PPD

RN UDI

–0.5

RN UDI

–1 –1

–0.5

0

Party Elites

0.5

1

Voters

Map 3.10.  Multidimensional issue congruence, Chile 2006: state–market/liberal– conservative (low-educated voters only) Source: Author’s own construction on the basis of WVS and PELA. Key: 1 = DC, 2 = RN, 3 = UDI, 4 = PPD, 5 = PS.

investigations are used to assess the validity of the descriptive portrait derived from my quantitative analysis. The quantitative analysis suggests that the economic model (represented by the state–market divide) is relatively uncontested in Chile. Meanwhile, the moral divide cuts across the two relevant partisan coalitions. This configuration deprives the moral divide of its potential for mobilizing popular support consistently over time. As a result, although partisan elite differences exist, the space for electoral competition on the basis of programmatic differentiation on either of these divides is constrained. The absence of significant partisan divides on salient economic issues in the era following transition to democracy has, in turn, reduced the importance of programmatic party–voter linkages in the system (Roberts 1998; Torcal and Mainwaring 2003; Hagopian forthcoming). This may well have produced programmatic destructuring at the mass-level, as well as the social stratification of programmatic linkages, making these more segmented in Chile. The following statements from two Concertación leaders seem to concur with this diagnosis: [...] in these [social policy] issues, I  never felt constrained by the right. We [the Concertación] do not want a public, state run, bankrupt system. And this is not only



Socioeconomic Segmentation of Party–Voter Linkages

135

1

0.5

FA

0

FA PN

PN PC

–0.5

PC

–1 –1

–0.5

0

Party Elites

0.5

1

Voters

Map 3.11.  Multidimensional issue congruence, Uruguay 2006: state–market/liberal– conservative (low-educated voters only) Source: Author’s own construction on the basis of WVS and PELA. Key: 1 = PC, 2 = PN, 3 = FA

due to financial aspects. Conceptually, many of us want to open spaces for the civil society [. ..] It is not that the right has blocked us, in some things it has, but in others we did not want things to change [. ..] Frankly, (if the right would not have control of the Senate) I do not believe things would have been different in the economic and social realms. In the political realm absolutely, we would have had another constitution, without designated senators [. ..] Look, I already stopped blaming it (the right), even though, within the context of a public speech, one might say “this is the right’s fault!” (Interviewed by Castiglioni 2005, p. 105) There are no ideological differences between us and the right; the economic model is the same. The only differences that remain are cultural; we just come from different political cultures. (PPD leader, Jorge Schaulsohn, personal interview, 2001)

Along the same lines, a UDI strategist explained some of the problems that this phenomenon of partisan collusion around market liberalism might have created for the right: The Concertación has defended the (economic) model so well that people who used to make money in Chile, are now making more money, and those who were sunk, are still sunk. The gap they wanted to bridge is still there, and they were able to keep the model in place without creating major social unrest. This is so clear that now I am afraid that

136

Segmented Representation

we will not be able to maintain economic support from business people. (Eugenio González, Lavin’s campaign advisor and UDI activist, personal interview, 2002)

Beyond programmatic collusion around the economic model and the existing constraints to mobilizing alternative issues (such as those related to liberal– conservative positions), the 1980 constitution is also the source of significant institutional constraints, especially on the views of more “orthodox” leftist leaders of the Concertación. The “culture of consensus” existing at the elite level may have led to further demobilization: The culture of consensus is so strong that every reform needs to be negotiated behind closed doors. And once you reach a consensus, then legislators raise their hands. This makes political conflict between parties with different positions invisible to the public, which translates into a lack of legitimacy of the system. Besides, the congressional tie produced by the binomial system, the extreme concentration of legislative initiative in the executive, and the need for qualified majorities for virtually every reform you want to carry through, prompts the executive to withhold or withdraw legislation for which it is known that consensus is not obtainable. This produces stalemate in the system, hinders government initiatives, and further contributes to delegitimizing politics and politicians. (Carolina Tohá, PPD congressional candidate, personal interview, 2001)

In the case of Uruguay, the emergence and later consolidation of the FA and its continuous electoral growth can be seen as fundamental causes for the overall increase in programmatic linkages in the system. On the one hand, lacking access to state resources (at least until 1990 with its arrival into the Mayoralty of Montevideo), the party needed to rely on linkage strategies other than clientelism to efficiently compete with the Blancos and the Colorados: We would not have any chance of campaigning on clientelism; we do not have much to offer to people. What we can do is to feed a hope that we need to construct among many of us. And that hope of collectively bringing change to our society is what we take to our meetings. It is undeniable that you always need to participate in meetings in which people are asking you to solve their particular problems; that’s ingrained in Uruguayan society. I won’t tell you that I do not receive those demands. And when I can, if it is reasonable, I try to help, because many times you have unnecessary unfulfilled needs, either due to bureaucratic problems or due to people’s lack of knowledge of formal procedures. But 80  percent of particular demands relate to the provision of jobs, and we cannot help with that. [. ..] And although they call you to address a specific problem, you always need to steal some minutes from them to talk about the common good, about general things that are happening in the country and how we, as a political party, are interpreting those things and what we are trying to do to improve the situation. (Victor Rossi, FA congressmember, personal interview, 2003) Clientelism comes with political paternalism and we need to break that up. Today, the “poor credential” is not helpful; today, I do not solve anything by handing out eight ceiling pieces per family. People need to organize collectively to have electricity, water,



Socioeconomic Segmentation of Party–Voter Linkages

137

and a better road. That way you build consciousness. (Artigas Melgarejo, FA congressmember, personal interview, 2003)

The FA’s traditionalization, control of crucial socialization vehicles in society, and ideological positioning as the defender of Batllismo against reformist attempts yielded both strong partisan identifications and a programmatic basis to compete with Blancos and Colorados. It also attracted increasingly dissatisfied voters to its ranks. (These processes are described in further detail in Chapter 5.) At the same time, the growing levels of electoral volatility in the system and the FA’s systematic growth in every election fostered increasing levels of cynicism regarding the virtues of clientelism, including in the eyes of traditional party leaders. These statements suggest that programmatic mobilization has, indeed, increased in the system, while traditional clientelism seems to be on the defensive: For traditional parties clientelism has been a total disaster. Every public employee seems to have been appointed by the Frente Amplio. During my political life, I might have placed, for different reasons, about six hundred people. Of those, I am sure that no one is voting for me. They are all leftist now. We put them there, and six months after they turn leftist. What a puzzle? Like every communist, they are traitors. They tell you they will support you, but you won’t ever see them back once they have gotten their post. (Hugo Cortis, political advisor of a Colorado congressmember, personal interview, 2003) People forget easily. The people I helped the most, the ones I put in the best places have abandoned me. The most needed, the ones you listened to but could not provide a definitive solution, are the most loyal ones. (Domingo Ramos, Colorado local activist, personal interview, 2003)

The 2009 expert survey on political linkages provides an additional way to assess the validity of the descriptive assessment of programmatic linkages offered in this chapter.10 The survey included five items to gauge experts’ opinions on the extent to which each party in their country sought to mobilize electoral support by pursuing each of five types of strategies: (1) featuring a party leader’s charismatic personality; (2)  emphasizing the attractiveness of the party’s positions on policy issues; (3) emphasizing the capacity of the party to deliver targeted material benefits to its electoral supporters; (4) invoking the party’s historical origins or the achievements of its historical leaders, as   10  The Latin American expert survey was implemented by a team headed by David Altman and Juan Pablo Luna in 2009. The survey was part of a worldwide effort to assess comparatively the incidence of different forms of electoral accountability in each country’s party system, which was designed and headed by Herbert Kitschelt at Duke University.

Segmented Representation

138 4 3.5 3 2.5

Programmatic Appeals Valence PID

2

Clientelistic Appeals Leader-Based Appeals

1.5 1 0.5 0 Chile

Uruguay

Graph 3.2.  Expert survey estimates of the use of different mobilization strategies in Chile and Uruguay Source: Author’s own construction on the basis of data from Altman et al. (2009) and the Democratic Accountability and Linkages Project (2010).

well as emphasizing party symbols and rituals to reinvigorate party identification; and (5) emphasizing the party’s general competence to govern and bring about or maintain economic, social, and political stability.11 Experts were asked to assign a number from 1 to 4 to each party, with 1 indicating that the party did not engage in that type of mobilization (“not at all”) and 4 indicating that the party made extensive use of that strategy (“to a great extent”). Graph 3.2 presents the obtained results for Chilean and Uruguayan parties. As shown in the graph, both party systems obtained an average above three when it came to programmatic appeals. Both party systems also rank high in terms of their use of “valence” mobilization, which emphasizes parties’ specific capacity to govern and deliver general public goods. The evidence also suggests that programmatic and valence mobilization are more central in Uruguay than in Chile, confirming the descriptive assessment offered in this chapter. The information displayed in the graph also conveys additional information. In Chile, leader-based appeals (“charismatic”) and clientelistic appeals (“delivery of material goods to voters”) are relatively more important than in Uruguay, where partisan identifications continue to provide a strong basis for crafting party–voter linkages. In line with the theoretical claim presented  

11 

The survey items corresponding to these strategies are E1, E2, E3, E4, and E5.



Socioeconomic Segmentation of Party–Voter Linkages

139

in Chapter 1, these results also strongly suggests that parties in both systems actively combine different kinds of linkages for mobilizing voters.

A LT E R NAT I V E T Y P E S O F L I N KAG E S A N D T H E I R PAT T E R N S O F S O C IO E C O N OM IC SE G M E N TAT IO N This section briefly presents available evidence concerning the patterns of socioeconomic segmentation of non-programmatic linkages that the expert survey suggest are relevant in each system, namely party-identification and (for Chile) candidate-based appeals.12. Available survey evidence for both cases (LAPOP 2008 and 2010, questions client1, client2, and client3) reports that significantly less than 10  percent of respondents were approached by parties engaging in clientelistic mobilization.13 The small reported incidence of clientelism in both systems limits my ability to estimate with validity its distribution across socioeconomic segments on the basis of LAPOP evidence. Chapter  4, however, presents other extensive evidence of the territorial and socioeconomic stratification of clientelism and particularism in both systems. Moreover, one of the stronger and more consistent findings in the literature on clientelism is that, when present, it is usually concentrated in the poorest social segments (Kitschelt and Wilkinson 2007). On this basis, it is reasonable to expect that in cases with a greater incidence of clientelistic mobilization, along with other types of linkages, linkage strategies are more likely to be socioeconomically segmented. The expert survey results thus suggest that linkage strategies in Chile are more socially segmented than in Uruguay. For linkages based on partisan identification, survey data supports the assessments based on the survey of experts. As shown in Graph 3.3, based on LAPOP (2010) results, party identifications are much stronger in Uruguay than in Chile. In both cases, a mild pattern of socioeconomic stratification of this type of linkage is observed, with the most-educated

  12  This information could not be replicated for Uruguay due to my lack of access to public post-election survey studies for a series of presidential elections.   13  In Chile, 93.6 percent of respondents reported that they had never been approached by a candidate offering material goods in exchange for their vote, and only 1.6 percent of the sample responded that they were frequently offered clientelistic handouts. The results obtained for Uruguay were 94.4 and 1.7 percent, respectively. Of course, these results could be distorted by the incidence of a social desirability bias that downplays the role of clientelistic mobilization in both systems.

Segmented Representation

140 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

Chile Uruguay

Primary

Secondary

University

Graph 3.3.  Socioeconomic segmentation of party identification in Chile and Uruguay Source: Author’s own construction on the basis of LAPOP (2010).

respondents identifying marginally more with particular parties than the least-educated ones. Finally, the information displayed in Table  3.1, based on post-election studies, seeks to gauge the relative importance of candidate traits (as opposed to ideological predispositions) in Chile, a case in which country experts consider “charismatic” mobilization to be relatively important. The table compares the percentage of respondents favoring the presidential candidacy of Joaquín Lavín (UDI) in 2001 to the percentages of respondents favoring the presidential candidacies of each of the four candidates who competed in the election of October 2005, in which Michelle Bachelet of the PS was elected. Different respondent categories were created by jointly considering respondent’s socioeconomic stratum (in three categories) and ideological leanings (left, center, right, and independent/none). The results obtained suggest that ideological leanings are significantly more important in determining vote-choice in the upper socioeconomic strata. This result is consistent with the analysis of the socioeconomic segmentation of programmatic linkages. In addition, Table  3.1 shows higher levels of ideological dealignment in lower socioeconomic strata, where significantly more independents and non-identifiers are found. Regarding candidate-based appeals, the information displayed in Table 3.1 suggests that, despite presidential candidates being more consistently supported by groups ideologically closer to them (Lavín by the right and



Socioeconomic Segmentation of Party–Voter Linkages

141

Table 3.1.  Preferred presidential candidate by social strata and ideological leaning (2001 and 2005)* Ideology Socioeconomic stratum

Preferred for president

Low (D–E)

Lavín 2001 Lavín 2005 Bachelet 2005 Piñera 2005 Hirsch 2005 Lavín 2001 Lavín 2005 Bachelet 2005 Piñera 2005 Hirsch 2005 Lavín 2001 Lavín 2005 Bachelet 2005 Piñera 2005 Hirsch 2005

Middle (C3)

Upper (ABC1–C2)

Left (%) Center (%) Right (%) Independent/ none (%) 4.2 0.8 21.8 2.5 0.7 4.0 0.7 26.2 4.4 3.3 2.0 0.6 20.4 2.5 2.5

5.9 2.8 8.2 3.8 0.3 6.2 2.2 7.5 4.8 0.4 3.4 4.3 7.4 4.3 0.6

19.2 11.5 4.6 9.5 0.2 21.8 14.5 3.7 13.2 0.0 28.2 11.1 1.9 30.2 0.0

12.4 7.0 14.6 6.7 0.7 7.9 3.3 8.4 5.3 0.4 4.7 2.5 5.6 5.6 0.0

N 371 145 312 143 11 207 95 211 127 21 58 31 57 69 5

Source: Author’s own construction on the basis of CEP Surveys . * Percentages do not add up to 100 because they are based on preference for all candidates mentioned in open survey item, including more minor ones not included here.

Bachelet by the left), the candidacies of Lavín in 2001 and Bachelet in 2005 were both very efficient in gathering substantial electoral support from more distant ideological identifiers. In other words, a sizable group of voters who supported the candidacy of the radical right in 2001 seem to have switched in 2005 to favoring the presidential candidacy of the center-left. This phenomenon is particularly strong in lower socioeconomic strata, where both candidates fared significantly better than their rivals in generating electoral support among independents and de-aligned voters. These findings confirm the increasing importance of candidate traits in mobilizing lower-class voters in presidential races.

C O N C LU SIO N Which competitive divides drive partisan competition in Chile and Uruguay? How has programmatic mobilization evolved in both countries during the

142

Segmented Representation

post-transitional period? Is programmatic representation segmented? And if so, is segmentation socially stratified? The evidence presented in this chapter suggests that the Chilean system continues to be structured around the regime divide. Yet, that divide has been losing strength over the years. This trend coincides with significant levels of ideological dealignment, and with decreasing levels of partisanship and electoral participation. Cross-sectionally, voters from the least educated groups have much less consistent programmatic positions than do the parties for which they vote. In turn, the Uruguayan system has moved from a situation in which parties and their electorates did not map out as significantly different units in terms of their programmatic stances to one in which the state–market divide consistently splits two party camps: both traditional parties (taken together, as a partisan “family”) and the FA. At the voter level, the two party-systems have moved in opposite directions during the post-transitional period. Whereas the Chilean system evolved from class-based ideological competition to a system where ideological divergence has shrunk (except along the regime divide), the Uruguayan system has evolved from a pattern of non-ideological competition during the first years of democratic recovery (and before 1973) to one in which the state–market divide became central in configuring two ideological families. Moreover, the evidence elaborated in this section points to high levels of social stratification (segmentation) of programmatic linkages in Chile, as well as increasing segmentation in Uruguay. Although partisan alignments configure two relatively well-defined partisan families in each system (the Concertación versus the Alianza in Chile, and the FA versus the traditional parties in Uruguay), the voters and congressional delegations of Uruguayan parties tend to present higher levels of internal coherence than their Chilean counterparts. The latter stand out for their significant levels of internal ideological divergence (particularly on the state–market and moral divides) among voters of different socioeconomic levels (measured through levels of education). My procrustes analyses confirm this pattern. Chilean parties vary significantly in the heterogeneity of the programmatic profiles of their social bases. The social bases of the UDI, in particular, and also to some extent those of the PS and PPD, are more heterogeneous in their programmatic orientations than are the social bases of other Chilean parties. Uruguayan parties are overall more homogeneous, exhibiting lower degrees of internal programmatic differences. There are two theoretically significant trends apparent in both cases. First, low-educated voters display varying degrees of programmatic structuring,



Socioeconomic Segmentation of Party–Voter Linkages

143

both across cases and over time. And second, the general strength of programmatic structuring varies also across cases and over time. Theoretically, this implies that the socioeconomic segmentation of programmatic linkages is not constant, but fluctuates over time. At least where programmatic linkages are considered, this finding contradicts the views of those explaining lower-class political attitudes in terms of “lack of political sophistication,” and consequent inability to read parties’ programs or interpret their programmatic structure (see e.g., Lupia et al. 2000). The greater levels of programmatic structure among low-educated voters in the mid-1990s in Uruguay coincided with political parties’ intensive mobilization of the state–market divide. This finding suggests that parties’ efforts to programmatically mobilize voters are pivotal in increasing (or decreasing) programmatic structuring in a system. Of the three political divides analyzed here, it is the state–market divide that most efficiently “aligns” less educated voters. In addition, as hypothesized by Kitschelt et al. (2010), programmatic structuring can decrease rapidly. Yet, it can also be built up in short periods of time if parties make particular efforts to mobilize voters programmatically and structural conditions permit them to do so. Regarding other types of linkages, my evidence suggests that partisan identification is very relevant in Uruguay, where it does not display strong levels of socioeconomic stratification. The opposite is true in Chile, where party-identification linkages are in short supply and have decreased steadily in recent years. Poorer voters in Chile were found to vote more than other social segments on the basis of candidate traits, more often shifting allegiance between candidates between elections irrespective of the candidates’ left–right positions.

4 Territorial Segmentation of Party–Voter Linkages in Post-Transitional Chile and Uruguay

I N T R O DU C T IO N A N D M E T HO D O L O G IC A L S T R AT E G Y This chapter turns to the local level and examines the territorial segmentation of party–voter linkages in Chile and Uruguay. I rely on qualitative evidence, proceeding from field research on seven electoral districts and nineteen municipalities in Chile and seven Uruguayan municipalities, which are also senatorial and lower-chamber districts. Interviews were conducted in three rounds, in 2001, 2002, and 2003. The different socioeconomic and political characteristics of these electoral districts are displayed in Tables 4.1 and 4.2. Due to the small size of Chilean districts, the sample for that country is largely urban and drawn from the metropolitan region. (Six of the seven districts and sixteen out of nineteen municipalities considered belong to the Greater Santiago; and only four of those nineteen municipalities belong to rural sectors.) This urban predominance was needed for efficiently covering relevant variance in district configurations. To control for the urban bias of my Chilean sample I conducted exploratory research in different rural locations within District 57 of the southern 10th Region of the country.1 Although I summarize the evidence obtained in those locations, the sample is not sufficiently large to make valid inferences about rural areas.2   1 Other towns and municipalities were not covered in my fieldwork (e.g., Maullín and Calbuco are also in District 57).   2  For instance, I cannot account for the nature of party–voter linkage in the northern region, where labor mobilization in the mining sector could ultimately translate into significant differences. That said, a second round of interviews pursued in 2008 and then again in 2010 and 2011 to survey the views of party leaders from different congressional districts in Chile largely confirms the configuration I describe here. The 2010–2011 sample consisted on more than fifty interviews with top party leaders from all parties with congressional representation. Leaders were selected for inclusion irrespective of the specific districts in which they competed. Fernando Rosenblatt conducted the interviews, and the results from that complementary research are publicly available (Luna and Rosenblatt 2012).



Territorial Segmentation of Party–Voter Linkages

145

Table 4.1.  Districts’ socioeconomic characteristics District

Chile

Extension Population Poverty (%) Human Human (sq km) (circa 2000) individuals development development (circa 2000) index (circa ranking 2002)

23 Las Condes 99.4 Vitacura 28.3 Lo Barnechea 1,023.70 26 La Florida 70.8 18 Quinta Normal 12.4 Cerro Navia 11.1 Lo Prado 6.7 24 Peñalolén 54.2 La Reina 23.4 27 La Cisterna 10 El Bosque 14.1 San Ramón 6.5 29 La Pintana 30.6 Pirque 445.3 Puente Alto 88.2 San José de 4994.8 Maipo 57 Puerto Montt 1,673.00 Cochamó 3,910.80 Calbuco 590.8 Maullín 860.8 Uruguay Montevideo 525 Canelones 4534 Salto 13,981 Tacuarembó 15,438 San José 4992 Paysandú 13992 Artigas     11918

232,814 86,419 73,895

0.23 N/A N/A

0.933 0.949 0.912

2 1 3

443,311

8.53

0.773

27

105,120 171,901 117,395

18.33 23.95 14.85

0.723 0.683 0.715

87 165 100

218,690 104,619

15.69 3.63

0.743 0.883

52 5

92,897 199,214 102,029

12.79 21.46 29.02

0.775 0.711 0.679

24 106 170

255,807 13,832 444,593 12,760

31.07 14.96 18.04 18.97

0.679 0.807 0.773 0.759

171 8 26 35

161,722 5,151 31,446 22,201 1,344,839 443,053 117,597 84,919 96,664 111,509 75,059

21.39 14.01 18.51 31.82 22.9 17 34 31 22 32     42

0.718 0.69 0.642 0.691 0.88 0.798 0.819 0.828 0.835 0.831         0.809

94 155 268 151 1 19 14 12 17 11        16

Source: Chile, SINIM and UNDP (2006); Uruguay, INE.

The configuration of Chilean congressional districts (usually comprised of more than one municipality) adds a significant complication to the analysis. In my characterization of district types I  identify two types of district configurations. In a first set of districts, all municipalities that belong to the same district (23, 18, 27, and 26) have similar socioeconomic characteristics. In a

Segmented Representation

146

Table 4.2.  Districts’ political trajectories Country District Chile

23 Las Condes Vitacura Lo Barnechea 26 La Florida 18

Magnitude1 Political trajectory, municipal level 2

2 2

Quinta Normal Cerro Navia Lo Prado 24 Peñalolén La Reina 27 La Cisterna El Bosque San Ramón 29 La Pintana Pirque Puente Alto

2 2

2

San José de Maipo 57 Puerto Montt Cochamó Calbuco Maullín Uruguay Montevideo Canelones Salto Tacuarembó San José Paysandú

44 14 3 2 2 3

Artigas

2

2

Hegemonic UDI Hegemonic RN Almost hegemonic RN Competitive DC-UDI-PS? Competitive DC-RN Almost hegemonic Concertación/PPD Almost hegemonic DC-PPD Competitive DC-RN Competitive Almost hegemonic DC-PPD Hegemonic PS Hegemonic DC Hegemonic PPD Competitive RN-PPD Almost competitive RN-Concertación Competitive PS-RN Almost hegemonic PS Competitive RN-PRSD Hegemonic PRSD Hegemonic DC Hegemonic FA Competitive to hegemonic FA Competitive Hegemonic PN Hegemonic PN Almost hegemonic PN to Competitive Hegemonic PC to Competitive

Political trajectory, legislative level Hegemonic Alianza

Almost competitive Concertación Almost hegemonic Concertación

Competitive Competitive

Competitive RN-PS

Competitive

Competitive Competitive Competitive Competitive Competitive Competitive Competitive

Source:  Author’s own construction on the basis of Servicio Electoral for Chile and Banco de Datos de la Facultad de Ciencias Sociales for Uruguay.



Territorial Segmentation of Party–Voter Linkages

147

second set of districts, municipalities with very different socioeconomic characteristics (29, 57, and 24)  are brought together into a single congressional district. The latter yields greater levels of internal differentiation regarding the linkages observed within districts. My description of district dynamics seeks to account for these nuances by introducing some basic distinctions among municipalities that belong to the last set of districts. The Uruguayan sample, in turn, is comprised of urban and rural sectors, which are both present in each of the districts I analyze. Due to their large size and social heterogeneity, I focused my research in the most populated districts on particular neighborhoods or towns, in order to cover the widest possible set of contextual conditions.3 The overlap between districts and municipalities simplifies the description of local politics in Uruguay. Moreover, the scope of the sample allows for broader generalizations than in the Chilean case.4 Unfortunately, my fieldwork was conducted during the harshest economic and financial crisis endured by the country since the 1930s. Likely, this provides greater salience to social assistance activities devised to cope with unemployment and the social emergency than in normal times. Thus, while the political dynamics I describe are structural, the centrality of linkages crafted around the distribution of social assistance might be amplified in the evidence I report for this party system. The evidence on which the analysis is based was collected through semi-structured interviews with relevant political and societal figures in each

  3  The focal locations for my fieldwork in each district were:  Malvín, Carrasco, Casabó, La Teja, and Cordón in Montevideo; Las Piedras, Ciudad de la Costa, and San Jacinto in Canelones; Salto (the provincial capital) in Salto, Artigas (the provincial capital) and Bella Unión in Artigas; Paysandú (the provincial capital) in Paysandú; San José (the provincial capital) and Rincón de la Bolsa in San José; and Tacuarembó (the provincial capital) in Tacuarembó.   4  Montevideo, the capital of the country, is one of those districts. It concentrates 43 percent of the total population (1.3 million people), 60 percent of the nation’s GDP, and elects almost 50  percent of all lower-chamber congressional representatives. Nonetheless, geographically Montevideo is the smallest political circumscription in the country. Canelones (a significant portion of whose territory is within the Montevideo metropolitan area) was also included in the sample. It is the second largest municipality, electing approximately 25 percent of lower-chamber congressmembers to represent 450,600 inhabitants. The remaining five districts selected were San José (also in the Montevideo metropolitan area, with a population of 96,200 people), the center-north district of Tacuarembó (the geographically largest district in the country, with a population of 84,600 people), the western districts of Salto (118,100 inhabitants) and Paysandú (110,000 inhabitants) which border Argentina, and the northern district of Artigas (bordering Brazil, with a total population of 75,100). As shown in Table 4.1, these districts display different configurations in terms of their overall socioeconomic indicators, economic production, financial standing, and weight of central state transfers in municipal income.

148

Segmented Representation

district, and in both countries. My sample included the following target categories: (a) successful and unsuccessful candidates that had contested recent elections; (b)  current and past legislators, mayors, and local council members, as well as salient political figures in the district; (c) leaders of community organizations; and (d)  key informants (party strategists and representatives from partisan think-tanks or congressional delegations, school teachers, priests, etc.).5 Comparative Notes on Party–Voter Linkages at the Local Level in Chile and Uruguay My analysis of party–voter linkages in the districts included in my sample suggests, first, that the level of personalization of party–voter linkages is greater in Chile than in Uruguay, especially at the congressional level. My description of Chilean districts therefore focused on both municipaland congressional-level figures, whereas my description of local politics in Uruguay is more clearly centered on mayors than on congressmembers. In Uruguay the latter are less pivotal as individuals in local party politics. That said, in smaller districts Uruguayan congressmembers are more visible and personalization is greater. Second, my findings suggest that Uruguayan districts and mayoralties have more electoral turnover. Moreover, politicians usually shift positions (from mayor to congressmember, or the other way around). In Chile, very strong incumbency advantages have instead taken root. Moreover, local politics there are marked by salient rivalries among mayors and congressmembers, even in cases where they belong to the same party. A more “individualized” description of local political careers is therefore warranted for Chilean districts and municipalities than for the Uruguayan ones. Last, but not least, I found greater levels of territorial segmentation in Chile. Segmentation there occurs both between districts and within them when they are socioeconomically heterogeneous. This segmentation produces clear-cut district types. In Uruguay, territorial segmentation is also present, but runs essentially within districts. Table 4.3 presents a summary of districts and linkage types for both cases. The classification identifies the predominant types of linkage strategies

  5  From 2003 to 2008 I  conducted a handful of additional interviews with salient political figures and key informants in Uruguay. Like the follow-up interviews conducted in Chile, these were used to check the inferences made for each Uruguayan district.



Territorial Segmentation of Party–Voter Linkages

149

Table 4.3.  Summary characterization of district types District

Observed linkage types

Homogeneous? If not, main distinction

18 23 24

Particularism, candidate based Programmatic, PID Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic

Yes Yes No. Peñalolén vs. La Reina.

26

27 29 57

Particularism, candidate based Particularism, candidate based Particularism, candidate based/PID

Montevideo

Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic, PID Particularism, candidate based/ programmatic, PID

Canelones San José Salto Paysandú Tacuarembó Artigas

Yes (although some distinctions between popular sectors vs. newly arrived upper-middle classes) Yes Yes No. Puerto Montt vs. Cochamó/ Río Puelo. No. Popular sectors vs. rest. No. Ciudad de la Costa, La Paz-Las Piedras vs. rest. No. Heterogeneous throughout. No. Heterogeneous throughout. No. Heterogeneous throughout. No. Heterogeneous throughout. No. Heterogeneous throughout, but with tenuous differences between Artigas City and Bella Unión.

observed in each district and the levels of territorial segmentation observed within districts. As observed in Table 4.3, more intra-district homogeneity, as well as inter-district differences regarding the types of linkages pursued are observed in Chile. If judged on the basis of the theoretical scenarios presented in Chapter 1, such configuration is consistent with a “dual pattern” of territorial segmentation. Instead, Uruguayan districts (as well as Districts 24, 29, and 57 in Chile), present greater levels of within-district segmentation of linkage strategies. When taken together, Uruguayan districts conform to a pattern of territorial segmentation that approximates a “mixed pattern,” as defined in Chapter 1. The remainder of this chapter uses the empirical information collected for each case to justify the classification of district configurations outlined in this introduction.

Segmented Representation

150

C H I L E :   OV E R A L L DY NA M IC S O F PA RT Y– VO T E R L I N KAG E S AT T H E D I S T R IC T   L EV E L Constituency-service, patronage, and brokerage networks are nothing new in Chilean politics. Indeed, plenty of convincing evidence shows that they played a vital role in institutionalizing the pre-authoritarian party-system (Valenzuela 1977; Garretón 1988; Valenzuela 1999; Borzutzky 2002). Notwithstanding, the articulation of these networks has gone through dramatic transformations in the post-transitional period. The evidence I  present here suggests that district-level politics in Chile have changed significantly, configuring a dual scenario. Programmatic and party-identity linkages have remained salient in upper sectors of society, whereas in the popular sectors linkages have become more personalized (instead of party-based), as well as increasingly based on constituency service and particularism. Electoral dealignment has also increased significantly. Meanwhile, the role of parties and their organizations in structuring the brokerage system (see Valenzuela 1977) has vanished. In this regard, the territorial networks that used to tie mayors (regidores), senators, deputies, and national executive figures to a system that distributed benefits to local communities and votes to the party have virtually disappeared. In this context, the capacity of political parties to collectively organize around these networks, which also provided a key mechanism for self-harmonization, has declined sharply. Summarizing the major trends identified in Chilean districts, upper sectors approach politics primarily through the media. These electoral constituencies are mobilized through the national currents represented by the Concertación and the Alianza. In socially organized low middle-class sectors, community service constitutes an important base for political support both for congressional candidates and municipal ones. Meanwhile, lower sectors present a combination of political alienation and particularism, structured at lower levels of interest aggregation. In these contexts, a high degree of personalization is observed. Therefore, while party leaders and highly politicized voters still identify on the regime divide, lower classes and younger generations are not strongly mobilized by this main divide structuring the party-system. This pattern triggers two different phenomena. Those who have a tradition of leftist political socialization, a history of political involvement, and a strong ideological background increasingly feel alienated from the Concertación and abstain from participating in party politics.6 In some cases they continue to support the  

6 

See Oxhorn (1995) and Posner (1999 and 2004) for similar evidence.



Territorial Segmentation of Party–Voter Linkages

151

PC-Chi, which gathers a significant amount of its small basis of support by canalizing systemic discontent (Siavelis 1999). Alternatively, even in sectors that suffered from intense repression under the military regime (usually land takeovers in poor neighborhoods), social “entrapment,” the social distance between residents and congressional candidates that come to the district during campaigns from the upper-income neighborhoods in which they reside, and political particularism usually reinforced by municipal politics all contribute to maintaining the status quo. In these cases, local political figures (typically mayors) with a real or perceived capacity to disburse basic subsistence goods discretionally in exchange for electoral support enjoy important incumbency advantages. In sum, poor voters who lack political socialization (particularly the youth), or a strong ideological imprint that sets them “culturally” apart from one partisan camp or the other, either also abstain from electoral participation or rely on non-programmatic linkages with parties and candidates. As a result, non-programmatic party– voter linkages became key in seducing the pivotal voto blando (“soft vote”), which began switching parties and pacts on the basis of candidate traits and/ or the provision of private goods. As one candidate explained his experience to me: Once, a group of ladies from Peñalolén invited me to a meeting to have tea with them. And they told me: “We know you, and we really esteem you a lot. But we are not voting for you. You know why? Because once the campaign is over, you will go. And if something happens here [e.g., a flood], we cannot go. We depend on them [the municipal government], and therefore, we need to take care of ourselves.” They stated it very clearly and in a straightforward way. And that showed me that fear still conditions their behavior and that clientelism is efficient in that context. (Carmén Lazo, former PS congressmember, personal interview, 2003)

In these social contexts, those parties that succeed in combining a national image-umbrella constructed through the media and popular candidacies with a system of local networks at the base (usually structured around municipal governments’ political machines or around those who administer national-level targeted social programs) obtain an important competitive edge in electoral campaigns. As a former congressional representative explained: You need to have a mix. You need to set up a mass-media story with some general projection, particularly for congressional or presidential elections. That’s what we had when Frei was elected. And then, you need to insert local leaders within your congressional candidate’s campaign showing them that you, as a congressmember, will get them connected to the political system, that you will be an efficient connection for them. And for sure, you need legitimate local networks for that. The mayors are essential in that respect. You cannot work outside those structures. For many people,

152

Segmented Representation

the municipality, their church, and their sport club are their only reality. That’s their life. And if you forget that, you are lost. Therefore, you need media, you need churches, you need clubs, and you need to connect them to municipal or governmental projects and programs. (Tomás Jocelyn-Holt, former DC congressmember, personal interview, 2003)

This portrait may in some ways appear to resemble that of the pre-authoritarian period. Two caveats apply, however. On the one hand, there is significant heterogeneity among districts in the system, making different formulas work better in different places. In this regard, socio-structural and territorial segmentation likely reinforce each other to create more cross-district divergence than before 1973. Comparing districts and municipalities with different socio-structural and political trajectories shows that there is considerable variation among districts in terms of how linkages within them are configured. Segmentation is so important for success today that even candidates working in socially heterogeneous districts clearly segment their approach to different constituencies within their districts. As a representative from District 24 containing both the more upper-class La Reina and the much more socioeconomically mixed Peñalolén explained, she has contrasting linkages to different groups of constituents: Poor people need you more frequently, at every moment. They need you every time they need to survive, because they have all doors closed to them. They don’t know where to go, how to do things. They don’t get the paperwork done; they need medical exams; they need to place a child in a given school [. . .] And that’s where we come in. Many times we do the same as the municipality, and obviously, they also ask the other congressmember to solve the problem too. But the important thing is to solve the problem, not who does it. The truth is that the greatest benefit from being a deputy is that you can pick up the phone and ask: Can we solve this? This should not be like this, but it is how the system works. And people come to us. [. . .] Peñalolén has the greatest concentration of everyday work for us. And it is the most enjoyable one, because there you realize that people need you. In La Reina, we don’t attract anybody’s attention. Nothing happens. Perhaps, if there is a specific issue they call you and want you to be there. (María Angélica Cristi, former Mayor of Peñalolén and RN congressmember, currently a UDI congressmember, personal interview, 2003)

On the other hand, the very logic of partisan competition in the system tends to hinder parties’ collective capacity to pursue segmentation through the implementation of specialized linkage strategies at the district and municipal levels while simultaneously achieving strategic harmonization. Differing from the pre-1973 period, parties as institutions now face important constraints when they seek to implement electoral strategies harmonized across districts so as to provide a national umbrella under which local candidates compete



Territorial Segmentation of Party–Voter Linkages

153

and cooperate. For instance, the decline of programmatic mobilization around issues able to mobilize broad constituencies, citizens’ alienation from traditional partisan figures, and the penetration of mass media have all increased the centrality of individual name recognition, and thereby media appearances enhancing candidates’ visibility and differentiation. As a result, congressional campaigns have become increasingly personalized. According to the participant observation reported in Díaz Rioseco et  al. (2006) and Giannini et  al. (2011), candidates frequently even choose colors and graphic designs for their campaign banners that hinder identification of the party sponsoring their candidacies. Also in this context, municipal machines provide local candidates with a crucial resource for advancing their career independently of their political party. Parties’ declining capacity to provide campaign funding to their candidates is supplemented by the increasing availability of non-partisan funding for individual candidacies, either via private donations or municipal-level resources that no longer hinge on access to the political center. In this new scenario, parties confront increasing challenges to maintaining the local– national brokerage networks historically vital for organizing their territorial apparatuses and achieving strategic harmonization at the national and congressional levels (Valenzuela 1977). In Chapter 5, I explore individual parties’ capacity to counteract these systemic traits. Linkage Segmentation in Chilean Districts: A Classification of District Types The different patterns of current party–voter linkages I classify in this section are closely related to how the socioeconomic characteristics of each district (e.g., levels of poverty and socioeconomic heterogeneity, availability of land for the development of new housing projects, etc.) have interacted with their historical political trajectories. I therefore first classify districts according to their socioeconomic and political characteristics, and then describe the type of party–voter linkages predominant in each of three types of district. Different linkage styles (or, paraphrasing Richard Fenno’s 1978 classic, “homestyles”) predominate in different types of district. The first type corresponds to higher-income municipalities, in which programmatic and party-identification linkages predominate and campaigns are fundamentally played out in the media. The second and third types correspond to medium-income, medium-low-income, and low-income municipalities historically identified with the left. These two types differ in terms of their current degrees of

154

Segmented Representation

social heterogeneity. The second type is less heterogeneous and has lower levels of social welfare along with higher levels of interest group fragmentation or dearticulation. In this type, non-programmatic linking predominates, and candidates from both partisan camps need to engage in extensive particularistic and personalistic mobilization to attract voters. For this reason, in spite of the leftist tradition of the districts, the political right (especially the UDI) has been able to make electoral inroads by fully engaging in this type of relationship with constituents. The third type of municipalities was historically similar to the second. However, in recent years these municipalities have received an inflow of middleand upper-middle-class residents, and have witnessed important development of their urban infrastructure. As a result, they also usually present higher levels of interest group organization and mobilization.7 They present a combination of the patterns seen in the two previous types, with middle and upper-middle sectors voting on programmatic appeals and popular sectors linked to candidates through particularism and personalism. Chile’s parties differ in how they approach this situation, and how successful they are at it. The parties of the Alianza combine public opinion and media campaigning to renew the support of the wealthiest electorate, while investing resources to personally connect with poor voters on the basis of particularistic mobilization and constituency service. The parties of the Concertación have remained successful when they have been able to draw on the leftist tradition present in some lower-class communities to craft new personalized and local leaderships working around a combination of constituency-service, leadership-based appeals, and particularism. This process is what I  have labeled a successful process of “linkage-substitution” by Concertación’s leaders.8 I now turn to a more specific characterization of district types. Upper-Income Sectors: “Rightist Ghettos” Higher-income districts (represented in the sample by District 23 and the Municipality of La Reina within District 24) are characterized by the absence of politicians developing fieldwork activities. Indeed, these districts are described by congressmembers as “public opinion” districts. Additionally, with the partial exception of La Reina at the municipal level (resulting from

  7  This is particularly so for the commune of La Florida, in Santiago.   8  The cases of Carlos Montes of the PS and Guido Girardi of the PPD represent successful substitution strategies. In the case of Montes, his strategy has been focused on investing



Territorial Segmentation of Party–Voter Linkages

155

the presence of a historical and very popular DC mayor until 2004), this group of municipalities present the highest levels of support for the Alianza in metropolitan Santiago, with this political pact being virtually always able to double the votes of the Concertación in the congressional election. This allows the pact to get both congressional seats while distributing the municipalities of Vitacura, Lo Barnechea, and Las Condes between the UDI and RN. National campaigns are played out in the media on the basis of partisan identifications and programmatic appeals. At the municipal level, mayors focus on the provision of public goods. As such, they engage in massive infrastructure improvements, development of green areas, public security, and highly visible cultural activities. Social programs in health and education are less salient in these districts, as the majority of the population has access and tends to rely on privatized services. The following statements by a UDI deputy and an opposition council member of Vitacura summarize the political dynamics at play in this type of district: The District [23] is a very peculiar one. It is likely the one in the country that is most heavily influenced by public opinion. That is, 90 percent of the people who live there are not expecting me to solve a specific problem for them. Nor are they expecting me to visit their home, give them something, or solve a social problem for them. What they expect is that I represent their opinions in the media. And that in Congress I vote like they would, if they were in my seat. Therefore, it is a district with almost no fieldwork activity. I only did fieldwork in the poor sectors that we still have there, like in Colón and Cerro 18 in Lo Barnechea. But that was it. And as an economist, what I do is to appear frequently on the media speaking about those topics that are interesting and important for my electors. And the rest comes from their identification with the UDI, which represents the hardcore of the Pinochet regime. These are the people that see the UDI as some kind of perpetuator of the military regime’s heritage. That’s the basis of our strength in this district. And I have the joy that in my district rightist supporters are many and that I don’t have to convince them, I don’t need to speak to them. (Julio Dittborn, UDI congressmember, personal interview, 2003) The right has political capital in this district that is unmovable. They can get 70 percent of the vote, they might get down to 68 percent or up to 71 percent, but that’s all. In the resources in the organization of civil society and constituency services. However, this strategy is less effective when relating to the lowest socioeconomic tier of society. Girardi, in turn, combines extremely frequent media addresses on contentious issues and an intensive fieldwork presence, in which he is accompanied by the municipal administration headed by his sister in Cerro Navia. When linkage substitution was not accomplished, new candidates, irrespective of their party affiliation, have been able to draw on national currents (e.g., the strength of the DC in the first years after the transition; the explosive growth of the PPD in 1997; the rise of the UDI since 1999; the popularity of Michelle Bachelet since 2007–2008) and personal resources to try to un-seat an increasingly unpopular incumbent.

156

Segmented Representation

Council we have always been five against one. And they don’t really care what happens here. There are some needy sectors in the municipality, like that of Los Castaños. They are not in extreme poverty, but they are poor. It is very funny! People complain: “They are constructing towers in front of our homes. How long are we going to stand this?” And then, the Mayor gets 69 percent of the vote! People think that because they live in Vitacura, they should vote for the right. It is absurd. We tried to talk to the people, we tried to organize them. However, electoral results are stable. They are fundamentalists [. . .] They don’t vote for individual candidates, they don’t vote for people, they vote for party labels. And even if they don’t actively engage in politics, they vote with extraordinary discipline. (Sergio Hernández, DC council member of the Municipality of Vitacura, personal interview, 2003)

Lower-Income Sectors Without Available Territories: “Leftist Ghettos” and Pinochet’s Successful Mayors in Congress Districts 18 and 27 present the lowest aggregate indexes of human development and the greatest percentage of people living below the poverty line in my sample. These districts and the municipalities they comprise (with the exception of Quinta Normal) originated at least partially from land takeovers under the auspices of leftist groups (predominantly the Movimiento de Izquierda Revolucionaria) and from the “relocation” plans pursued by Pinochet through which pobladores that inhabited higher-income municipalities were forcefully relocated to suburban areas of metropolitan Santiago. The leftist tradition of these districts is epitomized by the relatively high levels of electoral support obtained there by the UP in 1973, levels comparable only to those in La Florida. Unlike other municipalities with similar origins that are included in the sample (more prominently Peñalolén, but also La Florida in terms of its social structure in lower-income sectors), these municipalities (Cerro Navia, Lo Prado, Quinta Normal, La Cisterna, El Bosque, and San Ramón) lack significant space for the development of real estate and urbanization projects. Because of the lack of available territory for such projects, the social structure of the districts has changed less rapidly than that of Peñalolén and La Florida.9 Indeed, this kind of district did not see any major socio-structural developments in the era after the transition to democracy, beyond the general decline   9 Both Peñalolén and La Florida have recently received an inflow of middle- and upper-middle-class residents as a result of urban expansion and the construction of more affordable housing developments than in the traditional upper-class neighborhoods. The incoming population not only broadens the tax base, providing better resources to the municipalities, but also provides a more solid foundation for the electoral growth of the right.



Territorial Segmentation of Party–Voter Linkages

157

of poverty rates as a result of social policy expansion and targeting (see e.g., Pribble 2013). In this respect, this set of districts is similar to other popular municipalities not included in the sample, such as La Granja, San Miguel, and San Bernardo. In District 18, the Concertación has maintained a quasi-hegemonic position in congressional elections, winning both electoral seats available. The elections of 1989 and 2009 are the only exceptions to this rule. At the municipal level, the Concertación has also been nearly hegemonic, with the PPD and the DC obtaining the greatest electoral favor and RN twice (1996 and 2000) carrying the municipality of Quinta Normal. The leadership of Guido Girardi of the PPD (synergized with that of his sister, the Mayor of Cerro Navia, and later with that of his father and wife, who also won popular elections in the district) explains the extraordinary performance of the Concertación. Indeed, Girardi climbed from 42.7 percent in 1993 to 65.92 percent in 1997 (obtaining 72 percent in Cerro Navia). Although his electoral support declined in 2001 (to 58.4 percent at the district level), it was still sufficient to double the Alianza, despite the meager support obtained by the second most popular candidate within the Concertación (6.48 percent). As a result, Carlos Olivares from the DC was elected to Congress in 1997 and 2001 with less than 7 percent of the vote in each election, benefiting from a doblaje: the total vote for the district’s two Concertación candidates was over 66 percent thanks to Girardi’s overwhelming popularity, giving the pact both seats. Over time, however, Olivares was able to craft his own constituency, reaching 19.9 percent in 2005.10 In 2009, he left the Concertación pact to compete for the Partido Regionalista Independiente, and lost his seat. In 2005, Girardi moved to the Senate and successfully passed his electoral support on to his father, who was elected deputy for the district. In 2009, Girardi’s sister Cristina (formerly mayor) was elected deputy. Carlos Olivares’ exit from the Concertación contributed to the breaking up the doblaje, enabling RN candidate Nicolás Monckeberg to win the district’s second lower-chamber seat. In the meantime, Cristina Girardi’s mayoral seat was taken by Luis Plaza (RN). The Alianza had made important electoral inroads in this district before, climbing from 16.6 percent in the congressional elections of 1997 to 28.4 percent in those of 2001. Illustrating relatively high levels of electoral turnover and the weak relationship between elections at different levels of government,   10  During his tenure, Olivares built a swimming pool and a barbecue facility in the district, which was then lent to social organizations for their social gatherings. Swimming lessons were also provided to the public. “Olivares’ swimming pool” was, according to Olivares himself (personal interview 2002), the most electorally effective investment he made in the district.

158

Segmented Representation

Lavín, the Alianza’s presidential candidate, obtained 41.46  percent in 2000, against 58.54 percent for Concertación contender Lagos. In sum, linkages in the district are highly personalistic and not strongly related to partisan identification or ideological currents. Meanwhile, in District 27 each political pact has obtained one of the congressional district posts in all post-transitional elections, with the exception of that of 1989, in which the parties of Concertación (DC and PS) doubled those of the Alianza. Since the incursion of Iván Moreira—a former Pinochet-appointed Mayor of La Cisterna—as congressional candidate in 1993, each pact has obtained one legislator. While Moreira has continued to run, always obtaining a significant electoral share (28  percent in 1993, 37.4 percent in 1997, 42.4 percent in 2001, 30.4 percent in 2005, and 35.8 percent in 2009), parties from the Concertación have alternated in obtaining the other congressional seat. A PS congressmember was elected in 1989 and in 1993, but was replaced first by DC representative Eliana Caravall (1997 and 2001), and more recently by Tucapel Jimenez of the PPD (2005 and 2009). At the municipal level, Mayor Sadi Melo from the PS has been elected in El Bosque since 1992, subsequently obtaining 12.58 percent in 1992, 28.57 percent in 1996, 29.76  percent in 2000, 47  percent in 2004, and 46.5  percent in 2008. In San Ramón, DC’s Pedro Isla was also able to get continuously re-elected after coming into office in 1996 with 19.6 percent, replacing the first elected mayor, who in 1992 also belonged to the DC. In the last election (2008), Isla obtained 36.9 percent of the votes. Finally, in La Cisterna, an “older” municipality with less land takeovers and a greater influence of commercial (and some industrial) enterprises, more alternation in power was observed. In this case, the DC gained control of the municipal government in 1992 and 1996. In 2000, the UDI arrived into office, benefiting from an internal split among Concertación council members. However, in 2004 a PPD mayor recovered the municipality for the center-left pact and was re-elected in 2008 with 59 percent of the vote. Finally, in the presidential elections of 1999 Joaquín Lavín of the UDI obtained almost 46.96  percent of the valid vote, while Lagos obtained 53.04 percent. In the runoff presidential elections of 2005 and 2010, the candidate from the Concertación defeated that of the Alianza, Bachelet obtaining 56.01 percent against Piñera’s 43.99 percent in 2005, and Frei obtaining 50.65 percent against Piñera’s 49.35 percent in 2010. Unlike in higher-income municipalities, partisan voters are scarce in both districts, and electoral fortunes seem to depend on personalized leaderships pursuing particularistic linkages (intense fieldwork activities focused on low levels of interest aggregation). In short, given interest group fragmentation and pressing social needs, politicians have increasingly faced incentives to



Territorial Segmentation of Party–Voter Linkages

159

develop particularistic and problem-solving networks through which they exchange contingent political support for the satisfaction of voters’ immediate needs. Thus, politicians who are able to pay a household’s utility bills during the campaign period; distribute goods like TV sets, food boxes, and equipment for a neighborhood soccer club; or give away a cake for a bingo night organized by a centro de madres (mother’s center) or junta de vecinos (neighborhood council) on a regular basis are becoming increasingly successful in this type of district.11 During congressional campaigns, politicians also hire unemployed people to paint the neighborhood walls (and then protect the walls they painted from rival political groups) and pay around 60 dollars (in 2001) for each banner displayed on a household’s door. In sum, in this second type of district (low-income, homogeneous districts) politicians who engage in personal contact with members of poor communities develop a competitive advantage over more distant candidates. As suggested by the following testimonies, candidates who are able to be “in the field” offering the most in terms of satisfying immediate needs are the ones who increasingly tend to succeed in the polls:12 He [Iván Moreira] goes and visits people; he knocks at the door and says that the deputy wants to see how you are doing. He works in the district. He can miss a Congress appointment, but he does not miss any chance to be with the people here. He works for the vote and he is campaigning all day during the four years. That is what he does. He is not a good congressmember; he does not know anything about laws; he did not even finish school, but he is there. The evangelical sector is very important in the district, and he goes to every important ceremony at every church. The same happens with soccer clubs. He is everywhere kissing old ladies, and no other candidate does that. (Osvaldo Silva, congressional candidate from RN in District 27, personal interview, 2003) Today, a great proportion of Chilean politics resides in a group of personalities that are able to construct a special nexus with the community. In these municipalities, you don’t have organized political movements, you only have individual persons [. . .] the community is attracted to personalities and not to parties or programs. The politician that used to come to party meetings to talk about national issues no longer exists. Both Girardis [Guido and Cristina] are impressive political phenomena here. But that does not mean that my party, the PPD, is strong in this district. If they go from here, the PPD is done in this district. Votes are personally tied to them; they don’t belong to the PPD. And that is bad, it weakens the party base, the social network, which no longer exists. And the fragmentation and isolation it promotes, reinforces this   11  Figures A-4.1 and A-4.2 in the online appendix give examples of the types of interactions linking interest groups and politicians in this type of district.   12  Consider interview excerpts 4.1 and 4.2 in the online appendix.

160

Segmented Representation

political logic. [. . .] Council members are placed within a Chinese shoe here due to the new political culture that is emerging from the practice of giving away stuff. If I, as a council-member, don’t give you a cake for a bingo party, I am out. People now want you to give them stuff without any kind of effort or organizational counterpart on their side. They don’t get organized; they just come here and ask you to give them different things. And my party is falling into a sick paternalism from which there is no way out. Unfortunately, I now perceive that there are vast sectors of the left that feel good about this system. And therefore, people think we are the same thing. They are right; we are doing the same thing [as the right]. If we don’t get rid of mass-media guys and cosismo there is no future for the PPD. (Jorge Villar, president of the PPD’s district representation in District 18, personal interview, 2003)13

In this context, municipal governments have also become very powerful actors on the local political scene. Unlike in the pre-1973 era, they have gained autonomy from national-level politicians. At the municipal level, the same logic of competition applies, independently of the mayor’s political affiliation. She (the PPD mayor) is with the people. If there is a woman crying because something bad happened to her, she is there giving her a hug. She cares for the people and feels good helping them. She does a great job, not only taking care of people’s feelings, but also seeking solutions for their problems. She tries to provide solutions all the time, she is everywhere. (Isabel Mathus, DC council member of the Municipality of Cerro Navia, personal interview, 2003)

Finally, the importance of candidates’ familiarity with the district is similar in these municipalities to that observed in rural localities and cities of the interior, and constitutes a candidate trait that is important for linking to constituents. Indeed, although for different personal reasons, both the Girardi family and Iván Moreira are considered “locals” by these districts inhabitants.14 While Guido Girardi profits from continuous TV appearances (usually while leading specific popular protests in the streets or raising controversial topics in the media), his family has had personal ties with the district for fifty years. Additionally, he manages to be permanently present in the district, even when that means he needs to have “a siren on the car, to go from [the national Congress in] Valparaíso to the district, whenever that’s necessary.”15 Besides, his “media protest” style (sometimes used against the Concertación’s governmental decisions) leads people in the district to identify him as a “leftist.”

  13  Cosismo refers to the politics of giving away stuff (cosas), which epitomizes, in the eyes of Concertación’s activists, the political strategy of the UDI.   14 Personal interviews with PPD informants, as well as a personal interview with Guido Girardi (2003).   15  Personal interview with Guido Girardi (2003).



Territorial Segmentation of Party–Voter Linkages

161

While that identification resonates well with the district’s popular culture, it does not seem to conflict with the PPD’s doctrinaire commitments: “I don’t want to have a doctrine, because it boxes me in, it ties me down. I prefer having principles and values . . . I am in a party with principles and values, that has the capacity to be pragmatic.”16 Meanwhile, Iván Moreira does not appear so frequently in the media. He became identified with the district (especially in La Cisterna) during the military regime and has a personal (popular) style well suited to his tireless fieldwork activities. Indeed, despite his electoral success, top leaders of the UDI dislike Moreira for his style.17 At the same time, at the grassroots, his leadership is considered exemplary.18 Their personal ties to their districts and their particular political styles provide both Moreira and Guido Girardi with an important competitive edge. Especially in this type of district, politicians from all camps face strong incentives to compete on the basis of strongly personalistic leaderships and particularism for the support of an electorate that has grown increasingly dealigned, and that does not closely identify with political parties. In these competitive settings, politicians who do not deliver goods and services, or who do not think it is appropriate to engage in this new political style, get chastened at the polls. Interestingly, “public opinion” congressmembers, who are efficient in getting elected in upper social segment areas, face declining support rates in low-income districts. When interviewing former legislators who were voted out after one or two congressional elections, one systematically comes across similar statements.19 For instance: I was a public opinion congressmember. I was oriented towards the national public and took on technically very complex issues with zero electoral sex appeal. Therefore, I was not a low-profile congressmember, but I had a very different strategy from the one my district colleagues had. They relied on very extensive local networks. I  did not engage in constituency service. [. . .] I did not get into the “politics” of being a congressmember. I did not run a district office, nor did I use the district visiting week. What for? If people will not vote for me because I did not buy t-shirts for the soccer club that is their problem. [. . .] I was very skeptical of local politics. [. . .] Local networks and municipalities function like small feudal organizations, with increasing levels of corruption. And if you try to denounce those corrupt arrangements, you will obviously be in trouble. (Tomás Jocelyn-Holt, former DC congressmember, personal interview, 2003)

       

16 

Guido Girardi, cited in Plumb (1998). Personal interviews with UDI leaders. 18  Consider interview excerpt 4.3 in the online appendix. 19  Consider interview excerpt 4.4. 17 

162

Segmented Representation

Parties and party labels are much less salient, and much less valued than in the pre-1973 era, in this kind of district today. Indeed, while useful for securing a captive constituency in districts of the “upper social segments,” in lowerincome districts party labels are usually a hindrance. Particularly when politicians can still manage to provide for their constituents in spite of the lack of partisan affiliation, being perceived as “independent” provides an electoral edge over more partisan competitors. I am amazed. As an independent candidate, I am entering places where I had never been. People that used to throw stones at me and that dammed me as a UDI candidate are now calling on me to go and visit them. I don’t know if this will make me lose the support from strong supporters of Pinochet, that’s my only concern. [. . .] My work entails direct contact with people. I  need to be trusted, I  do not promise anything, I  just tell people the truth. My council stipend is spent on prizes for bingo games, sports tournaments, and the overall functioning of my office. Three times a week I go to the municipality, apart from other days on which you have to attend birthday parties or any type of social event. And there I have a team of four persons, well connected to the social and public works departments of the municipality, as well as judicial offices tied to the community. This way I can process demands, offer legal orientation, medical consultations, complaints regarding the municipal government, and so forth. (José Antonio Cavedo, former mayor and current council member for San Ramón, personal interview, 2003)

Middle and Lower-Income Districts with Available Territories: Heterogeneous Societies and Intra-District Segmentation Whereas the first two district types identified above constitute prototypical examples of linkage segmentation between districts, the third type I  identify in this section is characterized by within-district segmentation. A more detailed description of the configuration of each district is needed in these cases to provide an account of their political dynamics. I therefore analyze two municipalities (Peñalolén and La Florida) constituting prototypical cases of this third type of more socially mixed district. Both Peñalolén (a municipality that comprises District 24 along with La Reina) and La Florida (which alone comprises District 27)  used to be low-income and medium-to-low-income municipalities with a significant presence of rural settlements.20 During the 1960s and 1970s both   20  Both Maipú and San Bernardo, not included in my sample, have characteristics similar to these municipalities.



Territorial Segmentation of Party–Voter Linkages

163

municipalities witnessed extensive land takeovers, and the progressive expansion of middle-class housing resulting from the growth of metropolitan Santiago. In recent times, the majority of the remaining rural properties have been parceled out to construct real estate enterprises that progressively attracted professional and upper-middle classes to both municipalities. As a result, both municipalities present high degrees of social heterogeneity. In particular, La Florida is usually portrayed as a “sample of Chilean society” by politicians, composed of low-, middle-, and upper-middle-class segments.21 District 29, which also falls within this type, is similarly very heterogeneous. The district is composed of one of the most heavily populated and poorest municipalities in the country (Puente Alto), as well as three additional ones (La Pintana, Pirque, and San José de Maipo). Whereas La Pintana is structurally and politically similar to Districts 18 and 27 (comprised of Cerro Navia, Quinta Normal, Lo Prado, El Bosque, San Ramón, and La Cisterna), Puente Alto is similar in its social and political functioning to La Florida and Peñalolén. Meanwhile, Pirque and San José de Maipo are rural communities, with a distinctive political dynamic. Pirque, in particular, has some striking similarities to rural communities in the south (Cochamó and Río Puelo, which belong to District 57 along with the city and municipality of Puerto Montt, a large and socially heterogeneous urban center) where I conducted fieldwork. In all three communities (Pirque, Cochamó, and Río Puelo), memories from the past still shape strong partisan identifications at the popular level. Those memories vary from episodes of contestation around land reform in Pirque, to the construction of roads and bridges that significantly improved their connectivity to urban centers under the military dictatorship in the cases of Cochamó and Río Puelo. In each case, those memories perpetuate significant levels of electoral alignment. Particularly in the case of Río Puelo and Cochamó, where congressmembers usually do not visit the locality for years at a time, the label of the RN (the party most closely associated to the Pinochet legacy in the rural south)

  21  A former legislator from District 24 illustrates the political implications of the recent social changes in Peñalolén: “I entered politics when the ones that had political power were the ladies that went to the corner store to shop after a Catholic ceremony, carrying a rosary in their hands. Today, that power is in the hands of a middle-class guy, fat, who spends the weekend in the southern sector of Santiago, dissatisfied and angry, washing his car, with the cell phone ringing by his side, with a baby in the backseat and his wife asking him what time are they going to go to the mall. That guy is highly indebted, and the next day he will get up very early because he has to drive 40 minutes to work. The week will pass, and the next weekend, he will repeat the same routine. That guy is a neurotic” (Tomás Jocelyn Holt, former DC congressmember, personal interview, 2003).

164

Segmented Representation

provides a strong advantage to congressional candidacies. According to the secretary of the municipal office in that locality: Congressmembers do not come here; eventually they show up at election time. But we are very disciplined. Even if we don’t know her personally, the RN candidate will always have an edge [. . .]. People here are forever indebted to Pinochet. He built the Carretera Austral. Before, if someone was ill or needed something, he had to go to Puerto Montt by boat, which took long hours and could not be done during bad weather. And we know that Renovación Nacional represents those that supported our President at that time. (Omar Pérez, personal interview, 2002)

Returning to District 29, the rural areas of Pirque and San José de Maipo belong to the same district as the urban municipalities of Puente Alto and La Pintana. Those two rural municipalities have also observed a process of gentrification due to the arrival of middle- and upper-middle-class urban dwellers, segregating them from the urban areas. That process also contributes to setting them apart from Puente Alto and La Pintana, which continue to be the poorest municipalities in the district. In sum, the third type of district I  identify in Chile is characterized by higher levels of social heterogeneity. The political effects of this social heterogeneity are, however, mediated by the relative electoral size of each social segment in the district. For instance, due to its much greater electoral shares, Puerto Montt (in District 57) and Puente Alto (in District 29) set the tone, displacing the districts’ other municipalities when it comes to campaign and candidate activities. I now provide a closer look at the politics of two of the municipalities that most closely approximate the “socially heterogeneous/within-district segmentation” type: La Florida and Peñalolén. The conclusions drawn from my empirical analysis of these two municipalities have been cross-checked and largely confirmed by exploratory research in Districts 29 and 57, for which I offer a stylized characterization below. Historically, the Catholic Church (through the Holy Cross Congregation and the Vicaría de la Solidaridad) exerted a lasting influence in Peñalolén, contributing to social organization and mobilization, particularly during the transition to democracy. Those organizations provided an important base of support for the DC, especially during the first years of the post-transition to democracy era. The left also had significant strength, particularly in territories that were occupied through land takeovers in the 1960s and 1970s, and suffered intense repression during the dictatorship. In spite of this original configuration, the political trajectory of the municipality after the transition to democracy was shaped by Pinochet’s creation of the Municipality of Peñalolén, to which María Angélica Cristi was appointed.



Territorial Segmentation of Party–Voter Linkages

165

During the dictatorship, Cristi was able to draw on municipal resources to develop a very powerful clientelistic apparatus focused on the “maternalistic” provision of basic goods to the poorest sectors of the district. In this respect, her “horseback rides during a flood on winter nights to check if everything was ‘OK’ ” are still vividly remembered in Peñalolén’s poblaciones.22 On the basis of her popularity in Peñalolén, Cristi was able to get one of the two congressional seats for RN in the elections of 1989, obtaining 31.05  percent of Peñalolén’s votes (against 33.9 percent obtained by a candidate of the Partido Humanista). That electoral share was then expanded to 36.8 percent in 1993 and to 51.5 percent in 1997 on the basis of Cristi’s continuous presence in the district.23 In this sense, Cristi’s leadership is similar to that observed for other successful candidates operating in low-income communities. In 2001, Cristi’s support rate declined to 39 percent due to the UDI’s incursion into the district (riding on Joaquín Lavín’s increasing popularity) and the emergence of a new candidate (from the PPD) challenging the incumbent representatives, both of whom were suffering the effects of popular disenchantment during a sharp economic slow-down.24 In 2003, Cristi abandoned RN and joined the UDI. She was re-elected in 2005, and then again in 2009, obtaining 35.5 percent and then 31.26 percent in Peñalolén. The Concertación (particularly the DC) drew on its historical presence and mass organization in Peñalolén to maintain a very significant electoral share. In the elections of 1993 and 1997, Tomás Jocelyn-Holt obtained 31.6  percent and 21.27  percent of the vote running under the DC label.25 In those two elections, PS candidates also obtained significant showings at the polls (18.7 percent in 1993 and 11.6 percent in 1997). In 2001, Jocelyn-Holt’s electoral support decreased further, reaching 12.87  percent. His seat was taken by Enrique Accorsi, a PPD representative new to the district who obtained a 34 percent share of the vote in Peñalolén. His campaign focused on “renovation” within Concertación and rather intense fieldwork campaign activities. He was able to draw on strong name recognition as a former president of the Colegio Médico (the Chilean physicians’ professional association). His profession was also useful for addressing health-provision deficits in Peñalolén during the campaign. Accorsi has continued to be elected since then, obtaining   22  Interviews with Margarita Cofre and the youth group Buenos Días América.   23  In the district as a whole (which includes La Reina as well as Peñalolén), Cristi obtained 34.4 percent in 1989, 39.19 percent in 1993, 50.7 percent in 1997, and 38.52 percent in 2001.   24  This explanation was offered by Cristi herself and then cross-checked on the basis of personal interviews with Peñalolén’s political activists and candidates.   25  In the district (including results for La Reina), the leading candidate of the Concertación obtained 31.27 percent in 1993 and 21.47 in 1997.

166

Segmented Representation

31.17 percent in 2005 and 23.4 percent in 2009.26 In 2009, Accorsi successfully resisted a challenge by a DC candidate with strong support from Mayor Claudio Orrego. At the municipal level, the first elections were won by a DC candidate— Orrego—who triumphed over Carlos Alarcón of RN (Cristi’s Municipal Secretary during the dictatorship), obtaining 15.1  percent of the vote against 12.35 percent. However, after a corruption scandal in the municipality, Alarcón won the 1996 election against Orrego, obtaining 24.15  percent against 21.8 percent. Impressively, Alarcón was able to retain his seat in 2000 with 57.27 percent of the vote, defeating the DC candidate who obtained only 14.5  percent. Finally, in 2004, Orrego came back. Following a very aggressive campaign combining intensive fieldwork and media presence (including an internet campaign to raise funds among a “network of friends”), Orrego won the mayoral seat against Alarcón by a margin of 2 percent (48.6 percent against 45.99 percent). Orrego was then re-elected in 2008 with 58.36 percent. In spite of Alarcón’s close association with Cristi, my evidence points to the existence of two parallel and relatively independent political organizations (within the Alianza) in the district: the congressional office and the municipal government. This fact is consistent with an increasing personal rivalry between the two politicians that developed in light of Alarcón’s presumed congressional aspirations. This rivalry was reflected in the former mayor’s decision to stay within RN when Cristi exited to the UDI. Although similar to Peñalolén in other respects, La Florida is a municipality with a strong and active civil society, with communal organizations playing a greater role than in other districts. Additionally, this municipality had a stronger leftist tradition than Peñalolén, and constituted a harder place for military-appointed mayors to develop a political career after democratization. Moreover, intense military repression (especially against land takeovers) contributed to strengthening the leftist character of the district. In the aftermath of the transition to democracy, two strong leaderships consolidated in the district. According to a former DC congressmember, until 2000 “two political apparatuses dominated local politics, that of Carlos Montes (a PS congressmember) and that of Mayor Gonzalo Duarte (from the DC).”27 This fact made Duarte and Montes’ help crucial for other Concertación leaders to run successfully in La Florida. In 1989, Montes obtained 35.8 percent of the vote while an independent candidate received 21.2 percent. Meanwhile, in

  26  In the district (which also includes the vote share in La Reina) in 2001 Accorsi obtained 33.55 percent, in 2005 28.65 percent, and in 2009 23.19 percent.   27  Mariana Alwin. Personal interview 2003.



Territorial Segmentation of Party–Voter Linkages

167

the congressional elections of 1993 Montes obtained 32.6 percent of the vote and Mariana Alwin, from the DC, won 32.9 percent. This was possible for Alwin due to a series of combined effects:  Alwin’s name recognition (as the daughter of Patricio Alwin, the first president in the post-transitional period), the DC mayor’s support, and the national trend favoring the DC at that time.28 Yet, in 1997 a new leadership emerged on the right, that of Lily Pérez of RN. Pérez, a former council member elected in the municipal elections of 1992, succeeded in obtaining the second largest plurality (30.8 percent) in the district vote for Congress in 1997. Montes, in turn, climbed to 40.7 percent in 1997, further consolidating his leadership in 2001 when he succeeded in gathering 51 percent, followed by Pérez who was re-elected with 36.7 percent. In 2005, Montes climbed to 48.08 percent, while Pérez ran (and was defeated) in a senatorial race (in the eighth senatorial circumscription, which includes La Florida). Finally, in 2009 Montes obtained 50.44 percent, while Pérez ran a senatorial race in the Region of Valparaiso. Her seat was taken by a young UDI candidate, Gustavo Hasbún, who had worked under the auspices of Mayor Zalaquett and unsuccessfully contested the municipal election of 2008. At the municipal level, politics until 2000 were dominated by Gonzalo Duarte from the DC, the first elected mayor. While obtaining 29.94 percent when initially elected in 1992, he climbed to 32.36  percent in 1996 and to 37 percent in 2000. However, the 2000 increase was insufficient, and Duarte was replaced as mayor by Pablo Zalaquett of the UDI, who obtained 42.6 percent that year. Zalaquett would later regain municipal office in 2004 with 47.9  percent, with a PS candidate affiliated with Montes obtaining the second plurality (46.55 percent). In 2008, Zalaquett moved to the municipality of Santiago, where he successfully contested the mayoral seat. He was replaced as mayor by Jorge Gajardo, a PS figure associated with Montes and a popular television and cinema actor. According to one of Duarte’s aides, his defeat in 2000 can be explained by several factors related to erosion of the DC’s image, as well as to internal conflicts within the Concertación. Yet, he also stressed the fact that the electoral loss was caused by the arrival into the community of people “who do not directly depend on the municipality,” such as middle and upper-middle classes recently arrived in La Florida who have a “public opinion approach” to politics.29

   

28 

29 

Mariana Alwin. Personal interview 2003. Antonio Brandau, DC council member in La Florida, personal interview 2003.

Segmented Representation

168

Pérez clearly benefited from this change, too. Yet her strategy mixes personal leadership and intensive groundwork with social organizations. In her words: I always bet on middle strata, particularly those close to the DC, who are very receptive to a rightist but more liberal discourse. And some sectors of the left are also permeable to that platform. The strategy I used to reach those sectors was personal proximity. I did an immense door-to-door fieldwork during my campaign, which was very complicated because I was opposing the entire government, represented both by Duarte and Montes. I had many visits, many meetings with groups of people to discuss focused topics, particularly on social policy. [. . .] I have a district office, but I do not get involved in local problems as Montes does. I have seven people working with me in an open office in which we try to support social organizations. For instance, I taught training workshops for women, we support theater activities for children, and so forth. Montes, in turn, has tremendous support from the national executive. In 2001, a campaign year, the Interior Ministry approved projects for him that nearly amounted to $160,000 US, to invest in social programs and activities. I did not get one project approved. But I am happy, because people vote for me because they identify with me, not because I do a particular favor for them. I have a stable core of voters of around 30 percent, and they don’t vote for me on ideological grounds. We did a survey and they like me because they think I am intelligent, courageous, modern, and worked a lot. I think 98 percent of my vote is personal, because the right as a political current is really weak here. [. . .] Montes does not get that support only because he is a leftist; he also benefits from his personal ties with the district. His family always lived there, and at the beginning I was framed there as a “blond rich kid from Las Condes.” [. . .] I would say that social capital works here. The key here is to stay in the district during the four years, and that’s why you can have a less expensive campaign in the election year. You have to visit people and organizations all the time [. . .] That is, you don’t arrive in La Florida and win the election only with a good campaign in which you spend lots of money. It does not work like that. The UDI won the last municipal election (in 2000) because the erosion of Duarte’s image was too high at that point, and he suffered from a chastening vote, also tied to the national decline of the DC and the economic crisis. In that scenario, Pablo [Zalaquett], who was an unknown, came in as the candidate of Lavín, and he was able to win. (Lily Pérez, former RN council member and current congressmember of La Florida, personal interview, 2003)

Montes also implemented a mixed strategy, combining “socialist values” with personalized contact, and cultural promotion with intensive brokerage work between social organizations and the state’s social funds. This strategy is explicitly non-partisan, a fact that earned Montes harsh criticisms from the PS’s national leadership:30

 

30 

Carmen Lazo, personal interview 2002.



Territorial Segmentation of Party–Voter Linkages

169

In a world of vulnerability and ambivalence, people trust in a political style of closeness based on certain values. A style of dialogue is what explains my success. But it is also a style that allows people to grow, to develop; it is not clientelistic. It is based on making people organize around their collective objectives. And although that is difficult to accomplish, we try hard to do that. Here there are two political apparatuses, the municipality and my congressional office, where we run a cultural center, an NGO. We can’t compete with the municipality on resources because they even have their own TV channel, so we compete on quality. [. . .] Political parties have lost a lot, and they will not recover that in the short run. They lost the vehicles of socialization, and it takes time to reconstruct that. We cannot assume that people are already socialized and that we just need to invest in publicity. But it is difficult to get this message across. In the PS, they frequently interpret what I do as anti-party, because they think this is focused on me. And indeed, I am not a classical party-builder; I don’t invest in developing party activists and cadres. But I  am a modern party-builder, promoting cultural phenomena, and when possible, using that as an opportunity to increase political socialization and engagement. But cultural activities, social activities, should come first. This assumes an open mind. Political parties in Chile do not grow because they are tied to an elitist and old structure that has lost connection with the social reality. [. . .] Mine is not an apparatus; it is a social network. It is a group of people who share these values and try to maintain the network. This network informs, gathers information, and seeks solutions by talking to the government. So it is an enthusiastic social network, and it is a series of initiatives that tries to provide people access to knowledge, so they can think for themselves and find a solution. The resources we invest here come from my stipend as a congressmember, and our participation in the competitive funds offered by the presidency. And sometimes, we also try to engage the municipality, because this should be a public apparatus. I do not campaign during campaigns. That is what my everyday work during all my life grants me. I just continue my work. But people also get more motivated and work double, so if I usually have 400 or 500 people working with me permanently, that number is doubled during campaigns. And you spend more, in propaganda, in sound-equipment, etc. In that respect, we spend nearly four times more than normally. (Carlos Montes, current congressmember of La Florida, personal interview, 2003)

The value and “exceptionality” of Montes’ leadership is consistently recognized by other relevant political operators in the district, almost regardless of their partisan affiliation.31 However, while Montes has been able to develop a strong following in middle, low middle, and some sectors of the lower classes in La Florida, he faces important challenges when trying to engage some of the poorest poblaciones of the district with his political style.32 Those problems parallel those confronted by other politicians in trying to gather the support of alienated and socially fragmented lower classes.    

31 

32 

For instance, consider excerpt 4.5 in the online appendix. Consider excerpt 4.6 in the online appendix.

170

Segmented Representation

In this third set of districts, linkage segmentation is thereby observed within districts. This is particularly the case in District 24, which comprises the already heterogeneous Peñalolén and upper-income La Reina. Segmentation within districts is also observed in Districts 29 (Puente Alto, La Pintana, Pirque, and San José de Maipo) and 57 (Puerto Montt, Cochamó, and Río Puelo), both of which are made up of a heterogeneous set of municipalities. Moreover, organized civil society is more important in this third set of districts than in more homogeneously poor ones, in which civil society is more atomized and fragmented. As in more uniformly poor districts, however, partisan identifications and programmatic mobilization are again marginal. Therefore, personalized leadership is needed to engage in the systematic fieldwork activities necessary to advance electorally.33 Electoral turnover and alternation between candidates of the Alianza and the Concertación suggests that electoral allegiance is relatively volatile, and that it is more dependent on personal allegiances than on partisan traits. The second34 and third35 district types share those characteristics, which set them apart from the first district type.36 Unfortunately, my votes are my votes. At most, I  think some of my votes will stay within Concertación. It is sad, but it is like that. If I decided to run for a Senate seat, for instance, I do not think those votes would be transferable to another PS candidate, even though I have people working at the Council with me. But if, for example, my son were to run, independently of his quality as a candidate and politician he has an edge for being my son and carrying my last name. (Carlos Montes, current congressmember of La Florida, personal interview, 2003) Citizens are the clients and consumers of what the mayor does and the resources he distributes. Mayors relate to their communities and transform them into their own electoral and personal fief. I wish people were more active, with more chances to make decisions for themselves. But this does not happen here, because the same logic of the system reinforces that trend. From the mayor’s point of view, the best scenario is one in which people come one-by-one to ask for help in the municipality. (Nicanor Herrera, RN council member of the Municipality of La Florida, personal interview, 2003) The moment she [Cristi] quit RN to join UDI, RN disappeared in the district at the congressional level. We keep the vote at the municipal level, because we have the mayor,   33  Except, for instance, in rural segments or in specific neighborhoods in which political mobilization used to be strong in the past, such as locations in which land reform was massive in the 1960s and 1970s.   34  18: Cerro Navia, Lo Prado, Quinta Normal; 27: La Cisterna, El Bosque, San Ramón.   35  Peñalolén in District 24, 26: La Florida; 29: Puente Alto, La Pintana, Pirque, and San José de Maipo; and 57: Puerto Montt, Cochamó, and Río Puelo.   36  23: Las Condes, Vitacura, Lo Barnechea, as well as La Reina from District 24.



Territorial Segmentation of Party–Voter Linkages

171

and he has developed his own support base. [. . .] The vote here is personal; no one is voting for parties anymore. (Sergio Guerra, RN Council Member of the Municipality of Peñalolén, personal interview, 2003)

U RU G UAY:   OV E R A L L DY NA M IC S O F PA RT Y– VO T E R L I N KAG E S AT T H E D I S T R IC T   L EV E L Relative to those in Chile, the seven Uruguayan districts have greater levels of internal social heterogeneity. They are also more directly comparable to one another. The Human Development Index scores presented in Table  4.1 are correspondingly much more similar across districts in Uruguay than in Chile. Concurrently, Uruguay shows less territorially segmented patterns of political behavior than Chile, with national trends affecting local politics more directly. This is consistent with my classification of this case as one in which within-district segmentation predominates. For that reason, it is difficult to establish clear-cut distinctions among districts, and an analysis parallel to that presented for Chile would require a much more detailed analysis of specific zones within districts. Further complicating matters, electoral return data segmented by geographical zones within districts is incomplete and not as reliable as in Chile. I therefore do not classify Uruguayan districts by type as I  do for those in Chile. Instead, I  tentatively characterize districts’ political trajectories, while identifying a general set of linkage-type mutations observed since re-democratization. I thereby seek to identify distinct trends as well as relevant local political dynamics across and within districts to then identify recent transformations of electoral linkages at the local level. Where patterns are sufficiently clear and segmented to make valid inferences, basic distinctions are then drawn.

Ideological Families and the Recrafting of Local Party Systems: A Tentative Characterization of District Types in Uruguay As shown in Table 4.2, Uruguayan districts still present divergent and evolving patterns of local partisan competition. In some districts, one party has become nearly predominant (e.g., the FA in Montevideo and Canelones). In others, traditionally hegemonic forces have lost strength, producing competitive scenarios usually structured around a new bi-polar logic of competition between one traditional party (representing the traditional partisan family)

Segmented Representation

172

and the FA (e.g., in Salto and Paysandú). In still other districts (San José and Tacuarembó), the PN has been able to resist the FA’s sustained growth, maintaining and even consolidating its predominance at the expense of the Partido Colorado, which has virtually disappeared in both districts (especially in local elections). In these cases, the PN has emerged as the only viable representative of the traditional party family. The Partido Colorado was traditionally hegemonic in Artigas and Salto, and to a lesser extent in Montevideo (until 1989)  and Canelones (with the exception of 1989, when the national trend favoring the PN was too strong). These districts have recently witnessed significant alterations in their electoral patterns. These changes, in large part, have been connected to the growth of the FA. In Montevideo the FA became the hegemonic party, obtaining around 60 percent of the vote in both congressional and local elections. In the interior the FA grew, especially in more economically developed localities, urban settings, and areas marked by a greater presence of organized social interests and/or more “public opinion” voting (Paysandú, Salto, the Ciudad de la Costa and other metropolitan areas of Canelones, and Bella Unión in Artigas). In Canelones, the PC remained in office until 2004, despite increasing competitiveness (especially at the congressional level) punctuated by the FA’s sustained electoral growth. In 2004 and 2005, the PC lost more than 20 percent of its previous votes in national elections and more than 35 percent in municipal ones. Finally, in that year, it was defeated by the FA, which won a very sizable majority thanks in part to the internal (but “friendly”) competition of two popular PC candidates.37 The collapse of the PC in one of its traditional strongholds is explained both by national and local trends. Nationally, the decline of the party resulted from increasing discontent with government after the economic crisis of 2001–2002, and the national consolidation of the FA help explain this electoral result. Locally, the bankruptcy and inefficiency of the municipal government and persistent accusations of widespread corruption made by the opposition (including Lista 15 of the PC) as well as by the municipal labor union also contributed to this result. Indeed, the Mayor of Canelones in 2004 was the worst evaluated politician in the country, with less than 5 percent of the population approving his government (RADAR 2004). In Artigas, the PC declined for similar reasons. Indeed, by the mid-2000s these two municipalities had the highest levels of debt and relied the most heavily on direct patronage (i.e., municipal employment) as a prevalent electoral strategy.  

37 

Personal interview with FA congressmember José Mahía.



Territorial Segmentation of Party–Voter Linkages

173

While remaining stronger there than it was nationally, the PC also lost the district of Salto, another in which it used to be nearly hegemonic.38 In this case, economic problems were far less salient in the municipal government than they were in Artigas and Canelones, and the mayor was generally considered efficient. The national trend favoring the FA was catalyzed locally by the candidacy of a popular FA congressmember, and further strengthened by two local factors. The economic crisis hindered the articulation of the very extensive network of neighborhood organizations that the municipality (and List 1 of Foro Batllista) supported for surveying popular demands and redistributing goods to every location. Furthermore, the legal inability of a generally well-reputed mayor to run for a second re-election also created succession problems within List 1 of Foro Batllista and weakened support for the party in upper sectors of society. Traditional or recently consolidated Blanco (PN) strongholds, best exemplified by San José and Tacuarembó since 1958 and Paysandú since 1989, respectively, have been more resistant to change than their Colorado counterparts. The combination of strong local leaderships drawing on a segmented strategy (clientelistic and particularistic linkages in rural communities, platform and programmatic appeals in upper sectors of society) to generate and reproduce electoral support by municipal incumbents has proved decisive in transforming the character of partisan competition between Blanco incumbents and the left. Blanco administrations seem to have relied less than their Colorado counterparts on direct patronage, and tended to present lower levels of municipal debt by the mid-2000s (especially in San José and Paysandú). Yet, in Paysandú, a historical industrial pole in the country (currently depressed) with a significant presence of labor unions, the FA enjoyed better opportunities to grow and finally reached office in 2005. Towards National–Local Dealignment? The Characteristics of Recent Political Leaderships at the District Level Although more incipient and less clearly than in the Chilean case, there has been a trend towards denationalization of elections in Uruguay, particularly in 1999–2000, with local electoral contests presenting trends that have diverged from national elections in some districts. Specifically, whereas the Blanco and Colorado parties benefited from such a trend, the FA lost some of the electoral support it had received in the presidential election (Guerrini 2000; Magri  

38 

In the municipal elections of 2010, the PC regained the municipality of Salto.

174

Segmented Representation

2000).39 This can be explained by the split of national and local elections introduced in the constitutional reform of 1996, the consolidation of strong local leaderships, and increasing popular disenchantment with national ones. Under the new electoral law, mayors gained autonomy from national leaders and were better able to decide whether or not it was convenient for them to put their local political apparatuses into play for the national election. Such a trend was strengthened by the constitutional provision that established unique presidential candidacies for every party contesting the election. In cases where supporting a national presidential candidacy seemed overly inconvenient to a local leader, given the presidential candidate’s low likelihood of winning office and the mayor’s alignment with a different fraction from the one that had won the primary (and presidential nomination) of their party (e.g., as in the case of Larrañaga in Paysandú, Da Rosa in Tacuarembó, or Cerdeña in San José), mayors did not decisively help national leaders in their campaigns. After the first round, covert negotiations between local and national activists of both traditional parties also took place. For instance, in San José, Juan Chiruchí a former and extremely popular Herrerista mayor with a great personal following in the district, exchanged his support for the candidacy of Colorado Jorge Batlle in the presidential runoff for Lista 15’s support in his electoral battle against his former municipal secretary and incumbent Mayor Cerdeña.40 In 2005 however, likely as a result of the national wave in favor of the FA, the levels of national–local divergence were lower, granting the left eight mayoralties (five previously held by the Partido Blanco and three by the Partido Colorado), with first-time electoral victories in municipal contests outside Montevideo. In turn, while the Partido Colorado was only able to win (and maintain) one mayoralty, the Partido Blanco won ten. By the time I conducted my fieldwork, mostly between 2001 and 2003, different local competitive configurations were in place. In San José, one of the most consistent and durable local leaderships consolidated around Juan Chiruchí. This leadership is explained by a segmented strategy combining provision of public goods (including the lowest levels of municipal taxation in the country) particularly in the city of San José, attracting (international) private investment to the municipality, frequently communicating (and according to local sources, “exaggerating”) successful outcomes through public opinion appearances in the national media, and the   39 Such a trend was partially reversed in 2005. Yet, national–local dealignment became stronger again in 2010. Comparing the national election to the municipal ones and considering the nineteen Uruguayan electoral districts and municipalities, in 2010 the FA fared worse in thirteen of them at the municipal level than it did at the national level.   40  Interviews with Juan Chiruchí, Miguel Zunino, and Jorge Cerdeña in San José (2003).



Territorial Segmentation of Party–Voter Linkages

175

employment of traditional Herrerista politics (decentralized clientelism and brokerage) in rural and poor areas of the district. A similar strategy was traditionally applied in Salto by Lista 1 of the Foro Batllista, drawing on a decentralized structure of more than a hundred neighborhood councils. Those councils were defined as “non-partisan,” but were directly sponsored by the municipality, which relied on them to survey popular demands and distribute material benefits to each neighborhood, especially the neediest ones. At the same time, Lista 1 had also opened a permanent political club for receiving individual demands, which were then addressed by the municipality, a pattern very typical in municipal administrations.41 Finally, around 2000 Mayor Eduardo Malaquina consolidated his position as a national leader of the Foro Batllista, on the basis of public opinion appearances highlighting the superior efficiency of the municipal government of Salto. The two non-Herrerista Blanco mayoralties included in the sample (Paysandú and Tacuarembó) present similar levels of linkage strategy segmentation on the part of incumbents, with strong mayoral leaderships developing in both cases. On the one hand, the leadership style of Eber Da Rosa in Tacuarembó is similar to that of Chiruchí in San José. In Da Rosa’s case, however, the Mayor’s fraction is not hegemonic at the municipal level as is that of Chiruchí (Herrerismo) in San José, and as are those of Paysandú’s mayors (Alianza Nacional).42 Nonetheless, the intense electoral competition between two Blanco political machines (that of the Mayor and the one from Herrerismo consolidated around the leadership of former Mayor Sergio Chiesa) provided the party with a great electoral advantage over other contestants in Tacuarembo, as sympathizers of other parties, knowing that the outcome would depend on the internal dispute in the PN, voted strategically. Although similar to those leaderships, both of Paysandú’s recent mayors (Jorge Larrañaga and Diego Lamas) faced greater labor-union strength and conflict, especially with the sustained decay of industrial production witnessed in the district during the 1990s. Reflecting this trend, Paysandú is one of the districts that suffered the most from massive international emigration by the late 1990s. Confronting this scenario and the FA’s growing strength, both mayors opted for promoting a social pact strategy, cooperating with unions, providing FA activists with roles in the municipal executive, and vocally opposing the national government’s economic policy. Although this strategy effectively muted the FA’s growth at the municipal level in 2000, it felt short in 2005 when the electoral development of the left had reached its

   

41 

42 

Based on a series of interviews with district informants and political actors. Personal interview with Domingo Ramos and Hubaré Aliano (2003).

176

Segmented Representation

peak. Combined with national trends and districts’ particular characteristics, these governance styles have contributed to redefining the local party systems observed in each district. The Evolution of Linkage Types at the District Level Since the mid-1990s the traditional system of clientelistic mediation has suffered from three interrelated transformations. First, clientelism has declined globally. Second, clientelistic linkage strategies and the nature of the goods involved have changed. And finally, clientelism has become increasingly concentrated at the municipal level, where significant incumbency advantages have developed, moderately increasing the personalization of electoral appeals.43 Together, these trends have skewed the social distribution of representation in the system, with lower sectors of society engaging in clientelistic and particularistic linkages with parties, while middle and upper sectors vote programmatically. In this respect, the outcome has resembled that in Chile. Yet, when comparisons are drawn between the two partisan families the nationalization of electoral trends at different levels (presidential, congressional, and municipal) is significantly greater in Uruguay. This is clearly due to the FA’s efficient channeling of social discontent with both traditional parties in all social strata. The FA has effectively framed such discontent politically as a consequence of neoliberal reforms against Batllismo, thus creating a wider base for programmatic linking across all levels of the electorate. In Uruguay, the higher levels of interest group strength, particularly in middle and middle-low sectors (mainly around coalitions of ISI beneficiaries) has also contributed to aggregating interests and structuring more programmatic party–voter linkages in the system. Lower levels of popular alienation with politics and parties have then resulted as a positive externality. The persistence of strong political traditions in the country also provides parties with an important source of capital for avoiding the degree of personalization and partisan dealignment observed in the Chilean system. In this context, although both traditional parties have increasingly relied on partially paid political activists during campaigns, independents and “mercenaries” are still a rare specimen in Uruguay. Indeed, the case of the FA (discussed in Chapter 5) highlights the importance of having a very powerful and devoted   43 In each of these regards, my empirical findings concur with the theoretical claims of Filgueira et  al. (2003). I thus borrow from their labelling of each trend as the recession, the mutation, and the refuge of clientelism.



Territorial Segmentation of Party–Voter Linkages

177

network of partisan activists working in every locality of the country. In this respect, the party’s ability to “control” almost every channel of political socialization in society (including, very prominently, the educational system and the cultural intelligentsia) increases the salience of the Frenteamplista tradition. The competition between the traditional parties and the left has also provided the former with the opportunity to renew and reaffirm their political identities. Furthermore, the fiscal crisis of the Uruguayan state and the weaker development of the private sector in the economy decreased room for funding particularistic side-payments. Therefore, although probably growing, private sector donations for campaigns and their political effects are undoubtedly less important in Uruguay than in Chile. Territorial segmentation in Uruguay largely occurs within districts (as in Chile’s third, but unlike in its first and second district types). As a result, there has been less of a representation gap between voters of different districts in Uruguay than in Chile, particularly during the 1990s when intense programmatic mobilization developed in Uruguay.

The Receding of Clientelism Uruguay’s emerging party–voter linkage configuration can be compared to the one resulting from the development and consolidation of the Batllista sociopolitical arrangement (described in Chapter 2), in which partisan fractions, from both traditional parties and independently of their programmatic stances, competed by establishing vertical patronage and clientelism networks based on the exchange of electoral support for state-supplied goods (especially from the central state and public enterprises). Although organized interest groups were also active and effective in extracting state subsidies, it is worth mentioning that clientelistic networks included a true cross-section of Uruguayan society. Meanwhile, political careers and access to the state machine were contingent on each sectional partisan committee’s (agrupación), and each fraction’s electoral markup and their consequent leverage. In this context, very salient political traditions divided the two traditional parties, and sectional committees (clubs políticos) became the most significant agents of political socialization and brokerage, with a continuous presence in every locality. Today, the structure of party networks has changed. Political clubs of either traditional party are rare in Uruguayan neighborhoods, and they are only seen during electoral campaigns. As put by a Colorado congressmember: My party has lost its territorial organization. Leaders have not worried about maintaining a modern political structure for the party, which could allow a working framework

178

Segmented Representation

to directly connect to the people. We do not have a strong base structure; we just have a traditional organization that gets quickly organized during electoral periods. But we do not have something that helps us to connect on a regular basis with our party bases, with our activists. (Ronald Pais, Colorado congressmember, personal interview, 2003)

Political brokerage has atomized, and it is now exerted by local (and individual) players, usually tied to municipal governments or a congressmember. A successful Blanco congressmember working within the tradition of Herrerismo who has one of the best-developed political structures in Canelones states: Our day-to-day work is much harder than during campaigns. During campaigns you have fireworks, noise, barbecues, massive get-togethers, and emotional speeches. The day-to-day work is much more complicated, as you always face the temptation to lay back, to let days go by. And you cannot do that. I am always in contact with 190 or 200 local social players whom I have distributed all over the district, who thankfully are able to fulfill a social mission. With pride, I  can now say that in the neighborhood Artigas in Sauce there is one commission, and its president works for us; if there are people working on the CAIF Centers [targeted-community and NGO-sponsored pre-schooling and childcare for poor families], we have for sure someone from our organization inserted in that group; in school commissions, we have someone; in housing cooperatives, some members work for us. Therefore, we have people well inserted in the social network, and those are the ones I  work with. I  cannot go to Soca and tell them what is happening in their place. They need to tell me, and I need to seek solutions for them. That’s how it works. I am their “carrying-boy”; that’s your job if you are a young congressmember. And then we have fifty-two strictly political point people in the district. And fortunately, all of them are amateurs. I do not pay, although many people do. This is something you need to do with your heart; otherwise, politics would be death. Those fifty-two are distributed throughout the district, some working in family houses, others even paying to rent a small space. And I do fieldwork visits with them all the time. And if they have a problem, anytime, they have a direct line with me. We also have a group of friends who help us offer professional services that people cannot afford. We have two lawyers, two specialists in paperwork requirements for obtaining pensions, an agronomic-engineer who helps small producers, a psychologist. Finally, when the state has failed, we have also tried to substitute for it. In Paso Carrasco, a friend who is a physician proposed to build up a primary health clinic. Another friend who works in a pharmaceutical laboratory hands out free samples of medicine. We even had one activist who was a carpenter, and he crafted all the necessary furniture. And now, since the state clinic of Monterrey was shut down, we provide for the community with our own health clinic. (Luis Lacalle Pou, Blanco congressmember, personal interview, 2003)

Nonetheless, this type of personalized fieldwork structure is still rarely found within traditional parties, which cannot usually compete with the powerful territorial apparatus of the FA. Indeed, partially because the available stock of goods for establishing clientelistic linkages with constituents has been



Territorial Segmentation of Party–Voter Linkages

179

reduced, but also as a result of institutional incentives introduced in the 1996 constitution, both traditional parties have witnessed a process of bi-fractionalization, which has translated into a reduction of the high levels of internal competition historically observed (Piñeiro 2007).44 The penetration of mass media and survey research, in the context of increasing levels of programmatic competition, has made national leaders more able to bypass territorial partisan networks that, from an instrumental point of view, became increasingly expensive to maintain in the wake of the fiscal crisis of the state. As discussed below, this fact has important implications for the weakening of the partisan networks of both traditional parties and their decreasing capacity to offer a wide political menu in every electoral instance and at every territorial level. In turn, in line with the party’s ideological profile, the FA’s congressmembers (particularly those of Montevideo and metropolitan areas) usually highlight their role as party members, seeking to reduce the personalization of their political linkages with constituents. In addition, the FA members usually stress the role of social organizations and the importance of “promising a change,” instead of providing material incentives to people: We need to make a distinction. I think there is a clear difference between my party, the Socialist Party, and the traditional ones. They campaign during elections, and then each one of their candidates campaigns on a personal basis, setting up a political apparatus for herself. My party is not like that. We select candidates in a congress, democratically. And then, once we have the list, we work for the list, all together. This does not mean that we are not human beings and that we are [not] better known in certain social spheres in which we, as individuals, work more frequently [. . .] We help people, trying to provide answers to them, but what we need to try is getting close to those places [shantytowns] by stimulating people’s organization around their common problems. (Artigas Melgarejo, FA congressmember, personal interview, 2003)

Yet, the FA’s “distinctive approach” is also combined with more traditional practices. In this regard, all congressmembers interviewed point to job demands as the most pressing ones. In a context of state retrenchment, they frequently rely on their connections to private businessmen (including FA congressmembers) in order to seek employment for their constituents. Nonetheless, the job supply always falls short.45 The decline of resources to fund the clientelistic networks that tied national-level fractions to their local-level activists yields additional externalities. Given the characteristics of the Uruguayan electoral system, territorial structures    

44  45 

Consider excerpt 4.7 in the online appendix. Consider excerpt 4.8 in the online appendix.

180

Segmented Representation

are fundamental in campaigns. They, for instance, ensure that a fraction list will be available on Election Day at the voting booth. The number of lists is so high that in many cases the vote within a party and national fraction hinges on the availability and visibility of lists to voters, and/or the presence of those local point people in one list or the other. The erosion of national-level networks articulated around clientelistic linkages particularly impacted two traditional party fractions: the Foro Batllista of the Partido Colorado and the Herrerismo in the Blanco.46 As a local Colorado leader states: They have discouraged the little ants. We are little, but we are the ones gathering votes for them. Before, every weekend, I went out to the countryside, to little towns, to talk to the people. Today, I don’t do it anymore. I have maintained the friendship with the people, but we cannot sacrifice friends for politics. We cannot go out and promise what we don’t have. We don’t have anything now. So, we set up the list with a group of friends who were in a good economic situation, so we did not have to promise anything. If someone came and asked for something, we just told them that we did not have anything to give. The only thing we promised was to try to force an internal change in the party. (Hubaré Aliano, Colorado local activist, personal interview, 2003)

  46  Given the significance of party fractions in the Uruguayan context, it is important to provide a brief portrayal of the implications of this realignment for the most relevant party fractions of the two traditional parties (a parallel analysis of the FA is offered in Chapter 5). Within the Colorado Party, two main internal currents consolidated beginning in 1989: the Foro Batllista led by Sanguinetti, and the Lista 15 led by Jorge Batlle. The former represents continuity with Batllismo’s social-democratic imprint and has drawn, in the post-transition to democracy, on an important patronage and clientelistic network. The latter presents a liberal programmatic stance, and works around less formalized and stable support networks. As a Lista 15 leader himself put it:  “We have the same origin, and therefore our organizations are similar. But they [those in Foro Batllista] are a bit closer to communism; they have a social-democratic flavor. Therefore, they press a bit more with the use of the state, and that’s why they do so well in using municipal governments to maintain their political machines. They have not lost any municipality in recent times because they have a very well-developed political structure” (Hugo Cortis, political advisor of a Colorado congressmember, personal interview, 2003). Within the Blanco Party, Herrerismo, led by Lacalle, was during the 1990s able to overshadow the Wilsonista current, which has now been revitalized by the emergence of new leaderships in the interior of the country drawing on electoral successes at the municipal level. Those emerging leaderships have consolidated around a combination of public goods provision at the municipal level, and a resort to traditional clientelistic practices and the mobilization of the Blanco sub-culture. (The leadership of Jorge Larrañaga in Paysandú and that of Eber Da Rosa in Tacuarembó are the most prominent examples in my sample.) Those leaderships coalesced into the Alianza Nacional fraction, led by Jorge Larrañaga. In the meantime, Herrerismo represents the most consistently liberal party-fraction in the country. Yet, it has built important support bases working around networks of local key supporters with access to the state apparatus.



Territorial Segmentation of Party–Voter Linkages

181

Paradoxically, the weakening of parties’ territorial structures enhanced the need for individual congressmembers to stay personally in the field. This is particularly important for non-metropolitan congressmembers, who face specific constituent demands that cannot be addressed through the media. Along these lines, a Colorado congressmember points to the changes brought about by the restriction of clientelistic side-payments in the system: New opportunities have been created. You have to get to the ground from the heights and be close to the people. Talk to them, give them opinions, and inform them. Before, the congressmember visited every location once a year, at most. And then, people had to go to your office, stand on a line, and present their demands to you. Some congressmembers feel threatened by the impossibility of getting on the phone and solving people’s problems, but they don’t realize that they need to relate to people on new grounds. I came from a small village and when I was a child, you rarely saw a congressmember in the field. Today, everyone is there doing different things. (Jorge Duque, Colorado congressmember, personal interview, 2003)

Politically, the emergence, and later the consolidation, of the FA and its continuous electoral growth were fundamental causes for the decreasing overall importance of clientelistic linking in the system. On the one hand, lacking access to state resources (at least until 1989 with its arrival into the Mayoralty of Montevideo) the party needed to rely on alternative linkage strategies to efficiently compete with Blancos and Colorados. The FA’s traditionalization (Yaffé 2005), the party’s control of crucial socialization vehicles in society (Lanzaro 2004), and its ideological positioning as the defender of Batllismo against reformist attempts yielded both strong partisan identifications and a programmatic basis for competing with Blancos and Colorados, while increasingly attracting dissatisfied voters to its ranks (processes described in Chapter 5). On the other hand, the growing levels of electoral volatility in the system, and the FA’s systematic growth in every election, made traditional party leaders increasingly cynical about the virtues of clientelism: “we talk to everyone. We try to help everyone, but we know we are eating many frogs. You know they don’t vote for you. Still, we try to seek solutions for them” (Jorge Duque, Colorado congressmember, personal interview, 2003). In this context, the left and the traditional party family realigned around Batllismo, competing programmatically on the state–market divide. Both the increased programmatic tensions within the system and the increasing activism of pension beneficiaries, union members, and other ISI-beneficiary groups lobbying in Congress and using direct democracy mechanisms to promote or oppose legislation contributed to augmenting the level of interest aggregation in society.

182

Segmented Representation

Meanwhile, private sector workers and business interests have increasingly pushed for further market reforms, together with international financial institutions. Therefore, these trends have contributed to making clientelistic linkages less salient for a significant fraction of the Uruguayan population, particularly in middle and upper sectors of society. As a result, clientelistic linking has changed its nature and has increasingly sought refuge at the municipal level, where higher levels of personalism exist. Accordingly, local leaders operating around municipal machines or within decentralized networks of bureaucratic brokerage have gained centrality, displacing the traditional political committees of Blancos and Colorados. Clientelism has thus evolved into particularism at the local level, while becoming more targeted at poor constituents and less pervasive at the general level. The Mutation and (Municipal) Refuge of Clientelism and Constituency Service Uruguayan municipal governments frequently confronted important fiscal deficits and massive debts to other public institutions (very prominently the pension system and electricity, water, and phone providers) in the wake of the 2001–2002 financial crisis. However, they nonetheless retained relatively important degrees of autonomy, maintaining for instance the possibility of developing (“attenuated”) state patronage. Indeed, an FA congressmember characterizes the social role that municipalities needed to fulfill when confronting the greatest economic crisis in the country’s history and the financial collapse of the state, in these terms: Municipalities are facing severe budgetary cuts and the central government has not transferred resources to them. Hospitals are national, and only a few municipalities have primary health clinics. They cannot invest in infrastructure, which depends on the Ministerio de Transporte y Obras Públicas (Ministry of Transport and Public Works), which lacks resources. Before, the Banco Hipotecario [the state’s mortgage bank] and the Ministerio de Vivienda (Ministry of Housing) financed housing construction, and you also had housing cooperatives. Today, that’s all dead. So, I don’t know if I  would not be doing the same thing they are doing these days. Like the Mayor of Lavalleja, a great guy. He has cleaning jobs to hand out in the municipality and has created a rotation system in which one week the housewife, one week the husband, one week their children get those jobs. So, he has many families depending on the municipality. But you need to understand: in the interior you have basically three job sources. Productive facilities are closed. Therefore, either they go to the police, the army, or the municipality. So, although rotation is not as clean as we do it in Montevideo where we draft people, it is also a good way of generating some



Territorial Segmentation of Party–Voter Linkages

183

economic dynamism in the local economy. (Margarita Percovich, FA congressmember, personal interview, 2003)47

Especially when confronting atomized and impoverished social structures and low levels of electoral competitiveness, the relative autonomy that municipalities still maintain provides important leverage for developing encompassing clientelistic encapsulation. Indeed, municipal governments usually become “infernal (political) machines” in the context of political campaigns.48 This is illustrated in the dialogue between two Colorado local leaders of Tacuarembó: HA: With Sanguinetti, in 1985, we had the opportunity to gain the municipality. It was incredible; people came here to ask for the list. You did not even need to convince them. But we failed to select a good candidate here, and he did not work [. . .] And they [the Blancos] got the mayoralty, the hospital, everything. Ever since then, we have declined year after year [. . .] DR: Things are crystal clear. Even an idiot would have a floor of 2,000 votes if he runs the municipality, because at least 2,000 owe you favors [. . .] HA: Look at the last campaign. Within the Blanco Party, Alianza Nacional (the fraction of the mayor, Eber Da Rosa) defeated Herrerismo in the primaries of May 1999, by 4,000 votes. Then, Da Rosa renounced his position as mayor for the campaign, and the Vice Mayor gets his position. Right away he left Alianza Nacional and struck a deal with Herrerismo. Drawing on the municipal machine, in three months, they almost discounted the difference and were about to defeat Da Rosa in the election. . . DR: If the election had been a week later, they would have won it [. . .] HA: Do you know why they failed? Because they did not have a guy organizing the social service of the municipality, distributing building blocks, ceiling pieces, food, in an organized manner. Their logistics failed. They were distributing stuff overnight. It was like in the dictatorship, you heard trucks going up and down, all night. And that season, it rained a lot, and they could not deliver as many things as they have planned for. Eight-hundred social service orders were not distributed. And with those 800 orders, you can get at least a hundred votes. And they lost by a hundred. Otherwise, they would have won. Do you know what they did? They delivered building blocks as if they were birthday cakes [. . .] DR: Everyone does that. Tomorrow Pepito gets the mayoralty, and Pepito will do it. HA: Municipalities are vote-creating machines here. That’s it. (Hubaré Aliano and Domingo Ramos, Colorado local activists, personal interview, 2003).

  47  A similar system of rotation, using quincenas (rotating job contacts that lasted fifteen days each) was described as the key mechanism used by the Mayor of Artigas in several of the interviews I conducted in that district.   48  Interview with Hugo Cortis (2003).

184

Segmented Representation

In the context of pressing social needs, social service offices became the focus of municipal politics. In several instances, they also provided the basis for the development of new political careers in a district, especially during the economic crisis. This also applies more clearly to districts in the interior of the country, where municipal governments have greater local impact than the distant national government: This office has almost 400 employees and manages a total annual budget of 3.5 million dollars, not in money, but in terms of resources. We have different sections here that range from municipal garages to different workshops (carpenters, mechanics, and a municipal press). We also control garbage disposal, the potable water service, which is freely distributed in trucks to poor people, and so on. We also control the direction of hygiene and sanitary oversight of food. And finally, we have the social development department, through which we run extermination campaigns, cleaning campaigns against rats and other pests; the health service, which includes more than 70 primary health centers all around the district; food supply structured around a network of numerous municipal soup kitchens and childcare centers; and the distribution of food boxes, which are sent by INDA [the national food agency] and which we distribute here in cooperation with NGOs. [. . .] We never take into consideration the partisan affiliation of our beneficiaries. If they need help and if they qualify, they get what they are entitled to, even if we know that they are leftist. Of course, the fact of working in the neighborhoods every day, managing these resources, helps in that many people see me as an important figure. If people see that you are there, with your feet in the mud, they identify with you, and that creates a linkage. (Daniel Alcieri, Social Department Chair of Paysandu’s Municipality and Blanco activist, personal interview, 2003) We are now focused on resolving the emergency we are facing these days, and we have a team of people working permanently on these topics. We are taking care of primary health, with four decentralized clinics and the hospital. We have also implemented a decentralized service to carry out medical exams in the neighborhoods, so people do not have to move from their homes. And then we work with neighborhood commissions to try to satisfy the basic needs they have. Today everything is concentrated on food supplies. That’s what they most ask for. And we organize the distribution of food boxes, organize soup kitchens, and distribute clothes and milk. It is a very tough job. I  receive around twenty-five people a day asking for help, either because they are unemployed, ill, or face an extreme economic situation. In those cases, we send a social assistant and compile a priority list. And when things arrive, if something arrives, we distribute according to that list. Finally, if we have a flooding or a tornado, we ask the government for help and develop intensive fieldwork campaigns to survey people’s needs, trying to seek a solution for them. (Darley Bizcarra, Social Department Director and Colorado activist, Municipality of Artigas, personal interview, 2003)

Beyond strictly particularistic linkages, the mayor, given the centrality of his or her role in the district, can also bank on the provision of basic public



Territorial Segmentation of Party–Voter Linkages

185

goods for developing electoral support.49 Nonetheless, politicians who work in socially heterogeneous districts are able to draw finer distinctions between different localities and neighborhoods regarding the effectiveness of public goods provision at the local level in fostering incumbents’ electoral support. These distinctions clearly illustrate the segmentation of linkage strategies observed in these districts: In certain areas, you have people with higher cultural levels. You have an enormous presence [of such people] in Montevideo, and therefore, the opposition there makes a better showing. It is very difficult to deal with them. Indeed, many of the people who now live in Ciudad de la Costa came from Montevideo and are used to certain living standards in terms of urban development and basic infrastructural services. They came seeking peace and a better quality of life, but now have encountered important infrastructural problems. And no matter what you do, they are dissatisfied. In other places in the district, you build a road of 1,500 meters and everyone is happy. In Ciudad de la Costa, you can invest millions of dollars, and everyone is against you. They vote on the basis of what they see in the media. But, you just fix some holes in the road, and the people from rural areas or small towns in the interior will be thankful to you all their lives. (Jorge Duque, Colorado congressmember, personal interview, 2003)

In addition to the socio-structural characteristics of the population, the size of the private sector (and the corresponding weight of the municipal government in the local economy) also has an important impact on political behavior and the extent to which programmatic linkages are feasible. Even in a poor district like Artigas, politicians are able to distinguish clearly between the capital city and the second most important town (Bella Unión) on this basis: We analyzed this on the basis of Census data, and we reached the conclusion that in Artigas, 80  percent of the families depend on income coming from public sources, either national or municipal. In Bella Unión, the proportion was reversed, with only 20 percent living from direct or indirect state transfers. And political behavior correlates with that; the left does well where people are not dependent on the state. You cannot develop a classical leftist strategy where people are fearful of the power holder. And that happens when they depend on them for their jobs. Here, the municipality was not paying employer’s contributions to the pension system. That is money from the employees that was illegally appropriated by the municipality, hindering their future pension. In fact, they even lost access to credit because they showed up as debtors. We publicly denounced this situation, asking municipal employees to support a legal claim against the mayor. We have 1,200, 1,300 employees in this municipality. Do you know how many of them signed up? Eleven. In Bella Unión, people are more able to resist, and that facilitates our task of ideological and political formation. (Omar Alvez, FA local activist, personal interview, 2003)  

49 

Consider excerpt 4.9 in the online appendix.

186

Segmented Representation

Among Uruguayan districts, important differences exist between congressmembers from Montevideo and its metropolitan area and those of the interior of the country. As a Herrerista congressmember clearly explains, districts location and size conditions the different roles that legislators should play in their districts, as well as some of the implications that the system of lists blocked by fraction generates: I think that a deputy can be three things: a representative, a deputy, and a congressmember. And I think it is necessary to combine all three things. When parties are strong and coherent, those three functions can be fulfilled by different people, but when they are not, all three functions need to be partially fulfilled by each one of us. And representation today has two main forms of expression in this country. One thing is the deputy representing the department of Tacuarembó, and a very different thing is the one representing Montevideo or highly populated zones. The latter does not need to be in the district and does not engage in representing individuals, because you cannot represent all that mass of people individually. The only common topics you can grasp are national themes, and therefore you need to be on the media addressing those. So, the deputy from Montevideo is more a congressmember and less a territorial deputy, and I, as a congressmember from Montevideo, do not really know who my voters are. And although I could be vain enough to think that 2,000 or 3,000 voters vote for the list because they like me personally, I know they are voting the list. In the interior, the opposite is true. (Jaime Trobo, Blanco congressmember, personal interview, 2003)

In this regard, ceteris paribus, congressmembers elected in the smaller districts compete on the basis of more personalized linkages than those that are elected (usually by virtue of their inclusion at the tops of their fractions’ electoral list) in the most populated districts. Besides institutional incentives like the party’s control of blocked lists, party-centered campaign financing also plays an important role in helping to coordinate electoral campaigns. The Current Nature of Non-Programmatic Linkages in Uruguay: Overarching Implications With the emergence of programmatic linking in the system, clientelistic linkages between parties and individual voters (or with small locations or organizations) have arguably been affected by an important transformation of patrons and their political offers: Before, you had other things, like telephones, pensions, water connections, housing credits, etc. [. . .] We even had the police chief, so if people had to move from one house to another, we could use the police truck to assist them [. . .] Therefore, even if you did



Territorial Segmentation of Party–Voter Linkages

187

not have the local government, your presence was continuous in the district. Today, we don’t have anything like that. (Domingo Ramos, Colorado local activist, personal interview, 2003)

Some classic forms of patronage have, of course, endured. These include distribution of construction materials, bureaucratic brokerage in state agencies to obtain specific documentation, individual tax breaks, housing, state-bank loans, construction authorizations, rural roads, and connections to sewage and potable water. However, in a context of great fiscal restrictions and pressing social demands, new types of clientelistic goods have emerged: food packages, soup-kitchen services administered by partisan brokers, free legal consulting (usually preparing paperwork or providing information on how best to obtain bureaucratic or administrative documents, etc.), medical and dental check-ups, brokerage involving private sector enterprises (ranging from jobs to the provision of free bus tickets to urban centers, equipment for schools, or the provision of free lunches to a school delegation from the district), state contracts and concessions, and de jure or de facto tax breaks for “friendly” business enterprises. In general, the goods currently exchanged in clientelistic and personalistic linkages are less durable than were traditional ones like public employment or pensions. Political loyalty based on those types of linkages is therefore now weaker, and patrons are less able to engage in the monitoring of clientelistic exchanges.50 However, they find it difficult to develop successful higher-level interest aggregation strategies in the poorest sectors of society. The provision of public goods at the neighborhood or local level (constituency service) is a partial exception, as such provision has also gained salience in poor areas of the country. Examples of this sort of provision of public goods include development of legalization plans for shantytown settlements, connection of public utilities in irregular settlements, installation of primary-attention health clinics, and successful lobbying for nutritional services and schools in those areas. The operation of the FA in the periphery of Montevideo is at least partially based on this type of “social promotion.” Municipal-level particularism has also become more central, displacing national-level networks mediated by congressmembers and delivered through the traditional political clubs. The concentration of traditional, individualized clientelism at the municipal level correlates with the strengthening of mayors, who have progressively become the primary political figures in their districts. The existence of an “automatic majority” in municipal legislatures   50  According to one interviewee from a traditional party, an approximate 7 to 1 rate exists between favors extended and votes obtained. Interview with Hugo Cortis (2003).

188

Segmented Representation

(as elected mayors automatically enjoy legislative majorities on their municipal councils) reinforces the personalization of electoral politics at this level, promoting important incumbency advantages. In the context of relatively low levels of electoral competition, municipal governments tend to evolve towards a neo-feudal type of political leadership. In contrast, congressmembers have lost centrality in the provision of goods through non-programmatic linkages, given their declining capacity to extract resources from the central apparatus. Meanwhile, the executive directors of public utility companies able to invest directly in improving the infrastructure in their home districts have been able to promote their own political careers. Therefore, conforming to the trends observed in Chile, municipal governments have gained importance in forging clientelistic and personalistic linkages between (especially lower-class) voters and parties, with mayors dominating local politics and obtaining important incumbency advantages.51 As in Chile, mayors have therefore grown stronger vis-à-vis congressmembers. This being said, parties and fractions still play a much more central role in Uruguay than they do in Chile when it comes to pursuing and coordinating these types of linkages in the interest of a consistent, overall strategy. In sum, where civil society is poor and fragmented there are more non-programmatic linkages with municipal governments, especially clientelism at lower (usually individual) levels of interest aggregation, and there is, correspondingly, more municipal patronage (frequently in the form of temporary or seasonal hiring). The municipality of Artigas provides perhaps the clearest example of such a political logic. According to key informants, the frequent occurrence of flooding in the poorest neighborhoods of the city of Artigas provides a renewable source for this type of linking, with mayors capitalizing on their central role in distributing emergency funds and goods sent by the central government.52 In Artigas, the municipality also uses temporary fifteen-day jobs (popularly known as quincenas) alternating between different political supporters in order to provide for their constituents without completely overloading the system.53 Although important and effective in reproducing electoral support in depressed social contexts where state employment has become one of the only survival mechanisms available, these practices are less effective in areas with higher levels of socioeconomic development and civil society organization.54   51  In Uruguay, this advantage is to some extent limited by the fact that mayors cannot serve more than two consecutive terms in office.   52  Personal interviews with congressmembers from three major parties.   53  Interview with Sergio Arbiza (2003).   54  Interviews with council members Omar Alvez of the FA (2003) and Raúl Jiménez (2003) of the Blanco Party in Artigas.



Territorial Segmentation of Party–Voter Linkages

189

The city of Bella Unión exemplifies the limits of such electoral strategies. In contrast to Artigas, Bella Unión is a more dynamic industrial area where private companies and productive cooperatives used to industrially process sugar cane and other agricultural products, and in which economic dependence on public resources has been historically less important. Situated within the municipality of Artigas, the city’s electoral dynamics have been characterized by both greater support for the opposition (both Blanco and the FA) and greater electoral competitiveness. In the interior of Canelones, the configuration resembles the one in the city of Artigas. In this respect, after the municipal transition from a Colorado mayor who had been elected four times there (Tabaré Hackenbruck) to the recently elected the FA mayor, striking discoveries were made. For instance, in the locality of Santa Lucía, fifteen municipal employees (some with permanent and others with temporary contracts) had been hired to drive a single truck that had been inoperative for the last two years because it lacked tires. Moreover, in the small town of Tala three employees had been hired as “elevator operators” in a municipal office that lacked any type of elevator.55 In turn, the electoral behavior of metropolitan areas of Canelones are similar to those in neighboring Montevideo, and are more oriented towards programmatic linking. The comparison between Paysandú (another industrial pole) and Salto (a municipality in which the service sector has gained increasing centrality due to the development of tourism) on the one hand, and Artigas and the interior localities of Canelones on the other, further supports my claims. Non-programmatic linkages in Uruguay have thus incrementally moved from a system in which state patronage and pensions were central to one in which more specific and less permanent side-payments predominate, with political brokers exchanging small favors for votes. Military and police recruitment (in which congressmembers regularly extract positions from the corresponding ministries), temporary contracting in the central state apparatus, and “innovative” systems of municipal patronage are general exceptions to that trend. Private companies have begun to provide goods that congressmembers and municipal political brokers can then (seek to) exchange for votes. Yet, the size of the private sector in Uruguay is negligible when compared to that of Chile. Moreover, business interests in Uruguay do not function as a hegemonic group consistently pushing for a specific programmatic agenda as they do in Chile, where they can be seen as the de facto protectors of the status quo. Instead,  

55 

Based on a personal conversation with Martín Less, (2005).

190

Segmented Representation

they support political parties (usually companies support all major parties in the race) to enhance their opportunities for obtaining public concessions or protecting their particular business interests from damaging legislation. Therefore, it is possible to claim that a system for pursuing segmented linkages has also emerged in Uruguay on the basis of goods provided by business elites in exchange for interest representation, goods which are then supplied by political activists to poor voters. Yet, in this case, the system has not translated into skewed access to programmatic linkages between districts, as linkages cut transversally across all major political parties and are largely structured within districts.

5 Strategic Harmonization of Segmented Linkages: The UDI and FA in Comparative Perspective

I N T R O D U C T IO N This chapter moves from the systemic to the individual party level, analyzing the most and the least harmonized strategies in each party system. In Chile, the rightist UDI is the most strategically harmonized, in a context in which most parties segment (but are generally less able to harmonize) their linkage strategies. In Uruguay, the most simultaneously segmented and harmonized electoral strategy is that of the FA; other parties there have lost the capacity to segment their linkage strategies. Both the UDI and the FA draw on their particular resources to harmonize segmented party–voter linkage strategies. Their segmented but harmonized linkage strategies have paid off electorally, making these two parties the most electorally successful ones in their respective systems. The fact that both parties were in the opposition at the time of electoral growth might also have contributed significantly to their success. Yet, during the time it has been in government the FA has been able to maintain its strong electoral performance. The UDI entered the governing coalition in Chile in 2010. The parties of the Concertación and the RN in Chile have been less able to harmonize their appeals, while the traditional parties in Uruguay have been less able to segment their linkage strategies. The fact that the UDI and FA, while competing in two different systems, have both been able to successfully pursue segmented but harmonized linkage strategies while other parties in each system have been much less successful in doing so suggests that individual party characteristics, not just systemic or country-level traits, play a relevant role in causing the observed linkage strategies. In that regard, for instance, it is usually assumed that the effects of the binomial electoral system (along with other “authoritarian enclaves”) have

192

Segmented Representation

contributed to the UDI’s electoral development in Chile. Although those effects cannot be ignored, the primary focus of my analysis is the strategy developed by the UDI vis-à-vis other parties. After all, the electoral system itself cannot explain why parties competing within it adapt in different ways. Even if systematic biases favoring the Alianza por Chile (the rightist electoral pact) were present, the electoral system would still not explain why, until recently, the UDI was rapidly superseding the Renovación Nacional (RN) as the electorally more successful rightist party.1

SE G M E N TAT IO N A N D HA R M O N I Z AT IO N S T R AT E G I E S O F T H E U D I   A N D   FA One could wrongly think that harmonization is easiest when linkages are least segmented. The rest of this chapter presents evidence to the contrary. The UDI in Chile and the FA in Uruguay have implemented the most segmented strategies in their respective systems. Yet they have meanwhile achieved greater harmonization than their rivals. Such highly segmented but strategically harmonized electoral strategies have translated into leading electoral returns in recent years. Nonetheless, the types of linkages they deploy to mobilize their core and peripheral constituencies, as well as the mechanisms that each party employs to seek strategic harmonization, differ considerably. Table 5.1 characterizes the linkage strategies the UDI uses to mobilize different socioeconomic constituencies. The party mobilizes upper socioeconomic segments on the basis of programmatic (and interest) representation, as well as on the basis of the party’s identification with the economic model implemented by the dictatorship headed by General Pinochet. Meanwhile, the party also mobilizes voters in the poorest segments of Chilean society by combining candidate-trait and clientelistic strategies. The latter focuses on the distribution of material goods to individuals and small groups, as well as on the role of social workers providing medical or legal help. Table  5.2 characterizes the linkage strategies deployed by the FA in Uruguay. In this case, the range of linkage types observed is wider. The FA mobilizes its traditional social base (middle and upper-middle classes), drawing on strong partisan identities and the pursuit of programmatic mobilization. The party also draws on these strategies for mobilizing its new

  1  On the binominal electoral system see Zucco (2007), Huneeus (2006), and Navia (2005) for useful reviews and analyses.



Strategic Harmonization of Segmented Linkages

193

Table 5.1.  Socioeconomic linkage segmentation in the electoral strategy of the UDI Resources used in the appeal Agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate

Candidate traits(popular sectors)5 Partisan identity

Electoral Platform

Particularistic

Programmatic (upper classes)+

Clientelistic(popular sectors) 5

Party

Key BOLD: Predominant linkage strategies. + Links to core constituency. 5 Links to peripheral constituency.

Table 5.2.  Socioeconomic linkage segmentation in the electoral strategy of the FA Resources used in the appeal Agent making the appeal

Symbolic

Material: public policies

Material: private or club goods

Candidate Party

Candidate traits5 Partisan identity5+

Electoral platform Programmatic5+

Particularistic Clientelistic5

KEY BOLD: Predominant linkage strategies. + Links to core constituency. 5 Links to peripheral constituency.

constituency in the lower classes and informal sector. Yet, to mobilize this new constituency, the party relies more on the priming of candidate traits and the pursuit of clientelistic mobilization. The latter usually takes the form of “social promotion” activities and constituency service rather than individual “give-aways.” Figure 5.1 shows the UDI strategy of territorial segmentation and strategic harmonization. The party pursues a highly segmented strategy, including consolidating a dual pattern of territorial segmentation. Harmonization is achieved by deploying two specific resources otherwise scarce in the Chilean system: a party mystique imbuing party activists with a sense of mission, and the distribution of material resources. The party’s peculiar identity provides it with greater levels of internal coherence and discipline, helping it to resist Chile’s systemic trend towards atomization and declining centrality of parties. At the same time, the UDI’s capacity to obtain national party resources

Segmented Representation

194

Party

Votes

Party ID/ discipline Resources Policy

Peripheral Constituency: Poor districts

Resources

Core constituency: Upper classes

Key Leadership based Programmatic Particularistic

Figure 5.1.  The UDI’s linkage strategy

from its historical constituency (business and conservative interests) further enhances strategic harmonization and party discipline. The UDI is virtually the only party I observed that strategically allocates resources on the basis of a collective strategy. Figure 5.2 shows the territorial segmentation and strategic harmonization strategy of the FA, in Uruguay. The FA uses a mixed linkage strategy, combining different mobilization strategies within each district. The FA’s origin as an “externally mobilized party” (Shefter 1977)  over time provided a crucial resource:  a widespread partisan organization whose rank and file activism did not depend on the brokerage of clientelistic side-payments. The FA’s party apparatus enabled it to undertake social assistance and social organization activities, as well as to deliver constituency service once it took municipal office in Montevideo. The party has also benefited from the presence of strong partisan identifications in the population, which have progressively extended to the party’s new constituency. The FA also applies an old Uruguayan approach to segmenting its linkage strategies: different party fractions and leaders, while competing internally, also maximize electoral returns for the party by widening its capacity to supply everyone with an attractive electoral menu. Strong party identifications contribute to facilitating harmonization despite fractional competition. At the same time, the party’s mass organization and pursuit of party-centered linkages facilitates its self-harmonization.



Strategic Harmonization of Segmented Linkages

195

Party

FA’s Fractions

D1

D2

Figure 5.2.  The FA’s linkage strategy

SE T T I N G C OM PA R AT I V E PA R A M E T E R S :   L E S S SU C C E S SF U L PA RT I E S A N D S T R AT E G IC HA R M O N I Z AT IO N The ability of the UDI and FA to coordinate their linkage strategies is most striking when they are considered in relation to their systemic contexts and compared to their respective national-level competitors. The Chilean party system has recently been marked by three systemic trends: the weakening of parties as institutions, their declining social legitimacy, and the rise of independent-like candidacies within established parties. Party leaders of both the Concertación and RN customarily refer to the social devaluation of parties. Such devaluation is two-fold. On the one hand, parties have weakened as collective institutions within which like-minded individual leaderships pool resources and cooperate. Collectively available resources have diminished, and leaders increasingly seek to craft an independent image. In this context, parties became “archipelagos of individual personalities.”2 On the other hand, a schism seems to have consolidated between local and national leaders.3 Local leaders have more resources at their

  2 A PPD legislative advisor used this image to describe the internal situation of the Concertación’s parties in an informal conversation in Valparaíso (2003).   3  “Today mayors are the principal brokers in the system. And the deputy has been stripped of that. And if you analyze today the cases of corruption that have involved congressmembers, it is clear that what they were trying to do was to replicate the old logic of mediation. [ . . . ] I’ll give you an example. In the district, we have highly contaminating high-tension towers. We tried to organize people to ask for the removal of those towers. There are two ways of closing

196

Segmented Representation

disposal, but do not coordinate with national leaders or congressional candidates from their own party.4 The difficulties these tendencies present for parties were presented by Jorge Villar, a PPD president of district representation, and a former DC congressmember as follows: The moment you have strong figures, you are hurting the functioning of parties. They block those who can later challenge them. Here, we are in Girardi’s district, and we are PPD. But we don’t support him. We are old, we have a political trajectory, we are not at war with him, but we don’t agree and support what he does. He respects us and knows that we work in terms of a different logic. So, we just work independently from each other. He does not depend on us for the campaign; we don’t depend on him to support our humble communal committee. And obviously, we could not afford to support an alternative candidate. He has the same problem as other prominent figures. They are successful; they have a great image outside the party, but no one likes them inside the party. They are not people who are working for the party. (Jorge Villar, PPD president of district representation in district 18, personal interview, 2003) Party leaders do not care about the base. Exceptionally, they need them as electoral activists in a given moment. But, today, if you don’t have activists, you pay for them. Traditional activists have disappeared, and partisan networks are cut. [ . . . ] Local governments have strengthened. Yet, they are local machines. And congressmembers only go down to the district for campaigns. The other day, someone was telling me: “we need to rebuild the foundations of our political networks. And you can do that by sending resources from the state to the municipal level. But at the municipal level, it seems that the capacity to use those resources adequately is lacking. They control health and education, but they lack any type of technical capacity. And they function on the basis of very petty mentalities. (Ignacio Balbontín, former DC congressmember, personal interview, 2003)

Second, parties are also socially devalued. This, once again, creates incentives for candidates to invest in their personal career, which is divided from the party’s fortunes. In this context partisan structures have grown hollow (e.g., campaign personnel have to be hired), and have come to be perceived as dominated by closed “oligarchic” cliques with power (especially nominating negotiations with the electric company. Either you push, without concessions, until the towers are removed, or you dismantle the movement if the company gives you a hand with your campaign. That’s the difference. And in this competitive scenario we are heading towards the end of democracy and the rise of populism. And I would rather abstain from further political participation if those are the prevailing conditions” (Ignacio Balbontín, former DC congressmember, personal interview, 2003).   4  “I think parties have not understood that, unfortunately, this is not the time of parties [ . . . ] If you talk about parties, people don’t show up. They participate in open, plural activities [ . . . ]” (Carlos Montes, current congressmember of La Florida, personal interview, 2003).



Strategic Harmonization of Segmented Linkages

197

power when incumbents do not run for re-election), but with decreasing legitimacy and ascendance over the rank and file.5 As an RN municipal council member explained: Today I work as an independent with some ties to RN, but I am a community worker, I got rid of my political skin. People believe in me because of what I do for them, not because of what the party does. The deputy does not show up; she has her offices, her apparatus, and then comes for the campaign, but that’s it. Every party is facing enormous challenges here. (Nicanor Herrera, RN council member of the municipality of La Florida, personal interview, 2003)

In addition, fellow partisans compete among each other for access to resources such as media appearances and campaign funds, which are then used to promote their individual image or position on issues. The relationship between individual candidates, the media, and economic interests has also weakened the functioning of parties as collective bodies.6 Local leaders perceive reliance on private sector help, as a substitute for old partisan networks, as having become increasingly common: We need to be solving problems here, on a daily basis and with whichever resources to which we can have access. For instance, the other day a guy called me telling me that they had a death in the family and that they needed to get some cheap way of transporting the family to the cemetery. So, I called a guy who has a small van and works in an evangelical church, which I had helped before. And on that basis I was able to get them a very favorable fare to hire the van and solve their problem. (Isabel Mathus, DC council member of the municipality of Cerro Navia, personal interview, 2003)7

In sum, whereas before 1973 partisan networks worked out encompassing brokerage structures linking center and periphery, and were functional in reproducing organic partisan ties and popular loyalties, today those networks have atomized. This disintegration has hindered parties’ capacity for maintaining self-harmonization mechanisms. Instead, the emerging configuration feeds into increasing levels of cynicism, internal conflict, and anti-party attitudes at different levels: Before, candidacies were decided organically. You did not represent your desire and ambitions, but those of the party. Your merits within the party were the ones that merited your nomination, which was done by election on that basis. That happened at all

  5  See excerpt 5.1 in the online appendix.   6  See excerpt 5.2 in the online appendix.   7 Evangelical groups are extremely active at the local level, seeking benefits to improve their local establishments and to organize activities in the neighborhood. This differs from the Catholic Church, which essentially operates at the elite level and in relation to general issues.

198

Segmented Representation

levels. Today, congressmembers are the owners of their seats; they negotiate with other parties, but not within. From time to time, they have a meeting, get the bases together, and give some explanations. But they say exactly the contrary to what they end up doing. The truth is that the PS is today a party of managers dedicated exclusively to defending the President at all costs, in every possible respect. And as you don’t give the bases a voice, they are disenchanted and go away. (Carmén Lazo, former PS congressmember, personal interview, 2003) I think today many of these leaderships are beginning to work a bit as mercenaries. And parties do not have the capacity to take on issues. They don’t take issue with anything. You can see congressmembers taking on issues, but parties, as political parties, never. You don’t have a party making a proposal together in Congress, supporting a unified and collective motion. As politics is today so harshly oriented towards individual personality, a congressmember who has a good idea about something prefers to present it individually instead of sharing it with the party. [ . . . ] Parties today contain a variety of leaderships, some interesting, some with new proposals, some doing well . . . And that’s what the citizens perceive, more and more what we have are confrontational groups of leaders that hang together to defend their individual interests. (Carolina Tohá, PPD congressional candidate, personal interview, 2001)

The changing nature of party–voter linkages has also shifted power from the party elite to incumbents at the congressional and municipal level. While parties still retain their “nominating power,” which is pivotal under the binominal system (see Siavelis 2000), such power can only be effectively applied when districts have “open slots” (Navia 2008). Otherwise, the party is electorally better off granting the nomination to a successful incumbent (even if that incumbent does not always follow the party line and seeks to craft an “independent” career), than granting it to a new and more loyal party activist. Therefore, whereas decentralized harmonization by the party leadership would be required to achieve greater levels of strategic harmonization in the current context, party leaders in fact have less and less resources for implementing such a system. In Uruguay, strategic harmonization has suffered less discontinuity than in Chile. Nonetheless, there are some parallels to the Chilean situation: the decline of brokerage networks, the presence of a growing schism between local and national leaders, and the progressive development of more oligarchic party structures. National party leaders in Uruguay have been increasingly unable to deliver the goods needed to sustain their local structures. Local leaderships have instead become hegemonic. In this competitive context, partisan brokers from outside the faction of a locally hegemonic leader find themselves lacking resources to operate. They then either shut down their organizations (at least



Strategic Harmonization of Segmented Linkages

199

in between elections) or cross party lines to strike deals with local authorities willing to give them resources to keep their organizations alive. As one Blanco activist explained: The municipality [headed by a Colorado mayor] has given us some things. Their political style is to use the municipality to get our vote in the council, and we need to engage in that in order to get some stuff to maintain our organization. Other Blancos do not like that. They think we need to maintain a distance with the Colorados. I think we should do it; it is the way we can help people. If we do not win at the national level, if we do not have anything at the municipal level, that’s the only way we can help people and maintain our organization. (Juan Creceri, Blanco local activist, personal interview, 2003)

Both this response and that of shutting down organizations hamper articulation among traditional parties’ widely dispersed territorial organizations. Furthermore, the incentives introduced by the temporal separation of municipal and national elections (which provided local leaders with a greater capacity to detach themselves from locally unpopular national leaders) have particularly hindered the Blancos and Colorados. Within each party, there has been greater discontent among brokers within the fractions most traditionally reliant on extensive territorial networks: the Herreristas and the Foro Batllista. At the same time, facing increasing budget constraints and growing pressure from local brokers for material help, national leaders have favored “relatively cheap” local leaders with access to their own resources who are not dependent on national leaders’ brokerage or financial resources. They have thereby tended to lose contact with parts of their party networks. As a local Colorado Party activist explained: National leaders are important in various ways. First, they provide economic resources. Let’s look at the situation in this district. They are facing a very strong Blanco Party, very consolidated at the district level, which will be very difficult to strip from that position. So, within the Colorado Party, what you need to look for are very popular candidates, with a great popular following. That’s difficult here, because we lack those; the ones we had are gone. So, the national leaders of the party look at our department [the municipality of Tacuarembó], and they see that situation. And they wonder, “who is going to pay for this?” More now that campaigns cost a fortune, any senate or deputy list costs a lot. In Tacuarembó they have a person with great economic power, with a political career, with a political group functioning, so the answer is easy. They bet everything on that person and support her. She does not create problems for them, they don’t have to spend on her, and they already have a political base they can draw on. But now they are seeing that they will lose her congressional seat, because the Colorado Party is declining even further, as we and many other small groups are no longer working. So, they [the party’s national leaders] are coming back. Some senators came to meet us to see how they could solve this problem. And we

200

Segmented Representation

always came down to the same conclusion. We are willing to work, of course. We are more than willing to work. But, what are they willing to give us? As a result, what you have is a political conflict between small groups that collect the necessary votes for the congressmember to get elected, and that congressmember and the national leaders that support her. Those small groups are dispersed today. They don’t have an economic foundation for working, and they lack national support. Facing us, you have congressmember Montaner, with great economic power, continuous fieldwork throughout the district (because we recognize that she works a lot), and with power in Montevideo. Because if someone gives us something, she goes screaming at them complaining about that. And they don’t want to have to invest here, so they don’t want problems, and they continue to support her. It is a vicious circle from which we do not know how to get out. (Domingo Ramos, Colorado local activist, personal interview, 2003)

Blanco activists confronting similar strategic situations made similar arguments. For example: The political power in Montevideo is forgetting us. And that’s a terrible mistake. We cannot be connected only when they need us for the elections. There has to be a better way of staying in touch. That’s the tradition of Herrerismo; that’s why we were so strong. Now, national leaders have disappeared, and that hurts the party. This time they did not provide political offices to us. Let’s say two or three offices in the state, anything, somewhere. In the committee, you have people working all year round, and they are the ones keeping the party’s presence alive here. Then, when elections come, they benefit from that. But when we go there, they shut the door on your face. And if the party has no reciprocity with us, we cannot provide for the people. Then, how can I  go and ask them for their vote? (Juan Creceri, Blanco local activist, personal interview, 2003)

Yet, parties as collective institutions and social identities remain stronger in Uruguay than in Chile. This outcome is consistent with the pursuit of partycentered (instead of candidate-centered) linkages. Indeed, confronting a situation in which they lack ways of implementing other linkage strategies, traditional party candidates have sought to mobilize support by activating their political traditions and collective identities. The main drawback of that strategy has been that such traditions and identities became popularly associated with economic decline. Uruguay’s traditional parties historically had partisan offers characterized by high levels of internal diversity. One could say that they have gone through a process that has restricted that internal diversity, while deterring wider participatory processes within parties. This tendency atrophied traditional party networks and participatory dynamics in the popular sectors, opening space that the FA’s network of activists then stepped in to fill. As a local Colorado Party activist in one of Montevideo’s shantytowns lamented:



Strategic Harmonization of Segmented Linkages

201

Our congressmembers always make the same mistake. They come here four months before the election, promising things and giving away stuff. People say we steal and that we do not fulfill our promises, and I am afraid that is true. [ . . . ] Sometimes they send me a truck with stuff, and I just call the people and tell them to get it. But they [national leaders] are not here. [ . . . ] They come here wearing suits, and the kids look at them like they are saying: “I will steal everything from you.” [ . . . ] The Frente Amplio does it differently; they do not engage in clientelism. They are here all the time; they wear (clothes like we wear) and talk to people like we talk everyday. They engage the youth in music, in parties, in painting walls, in sports. My own children go with them, because they tell me it is fun. They get together, they play the drum, they drink wine, and they have fun. And I let them go because I prefer that to having them smoking pot in the corner all day. (Elida López, Colorado local activist, personal interview, 2003)

Concurrently, the partial elimination in 1996 of the DSV for the lower chamber increased institutional constraints on keeping the classic catch-all partisan offer open.8 In this context, and particularly in the midst of bad economic times, traditional party candidates were put in a “Chinese shoe.” While they faced ever-increasing demands, their traditional means of providing for the people on the basis of centralized brokerage networks was drastically reduced. At the same time, programmatically, they had to support the unpopular reforms implemented by their leaders in government: I started distributing telephones (telephone lines). And I do not know if that helped me or not for the elections, but what I can tell you today is that before I could respond to that demand. Today I face innumerable demands that I cannot respond to. We used to have pensions, water and electricity connections, public employment, free bus tickets . . . What happens is that Uruguayans are very clientelistic. And today, the problem is that we have many demands, and we lack the ability to respond. Everywhere in the world citizens demand more welfare, and there are two ways of interpreting that. You can see it as personal and individual welfare, or you can see it as greater collective welfare. The problem is that, today, we cannot provide either of those. We are in transition between a paternalistic state that provided for everyone and a liberal model in which everyone needs to be creative, inventive, and able to provide for himself. And in this transition, we have faced a huge economic crisis. So then what we have is a line of people who still think that we can provide for them, while we can’t, and at the same time, we cannot claim that the country is improving, because people are suffering a lot. (Jaime Trobo, Blanco congressmember, personal interview, 2003)

  8  This electoral reform did away with the “electoral cooperatives” in which district sub-lemas could accumulate votes by adding up the votes obtained by a great number of lists, each representing different local agrupaciones. Technically, the reform did away with what was known as the triple-simultaneous-vote. See Piñeiro (2007) for a detailed discussion of the reform and its implications.

202

Segmented Representation

As this last interviewee made clear, traditional parties lacked the capacity to sustain some linkage types that were central in the pursuit of their historically segmented strategies, and were found to have low levels of segmentation and declining levels of harmonization. Relying on party-centered appeals alone as the competitive context changed, those parties found themselves with virtually no credible appeals to make. They had neither a programmatic option that was compelling to the popular sectors (i.e., an alternative to unpopular liberalization measures), nor access to goods that could be centrally distributed to appeal to voters. In decline due to generational replacement and FA’s success in politically socializing more new voters, traditional parties were left with a modicum of party ID to rely upon. This scenario differs from that in Chile in at least two ways. First, candidates from the Uruguayan traditional parties have seemingly been unable to segment their linkage strategies, unlike Chilean candidates from every party. Second, higher segmentation and the pursuit of candidate-based appeals have translated in Chile into diminished harmonization (with the exception of the UDI, which is discussed below) and the weakening of parties as organizations. In Uruguay, in spite of increasing levels of personalization, linkages continue to be deployed by parties and fractions as institutions. Overall, this yields higher levels of harmonization than in Chilean parties. The Chilean parties “solved” the need to adapt to a new competitive context by taking advantage of highly segmented, candidate-centered appeals. They used those appeals to prop up their vote share in the absence of a compelling central party offer. Yet, this candidate-centered response also diminished their capacity for harmonization. As a result, Chilean parties were able to maintain their vote share over time (avoiding the decline observed in Uruguay for traditional parties). However, parties in Chile also “hollowed out,” increasingly becoming empty shells. This has been reflected by very low levels of party identification, especially by younger voters, and a gap between parties and social movements. Table 5.3 places each of the cases discussed here in comparative perspective, using the typology of harmonization mechanisms introduced in Chapter  1 and the evidence discussed in this section.

THE UDI IN CHILE This section explains why the rightist Unión Demócrata Independiente (UDI), a relatively new party with strong ties to the former military regime headed by General Augusto Pinochet is today the party with the most seats in Chile’s



Strategic Harmonization of Segmented Linkages

203

Table 5.3.  Mechanisms required for achieving harmonization on the basis of each party’s predominant linkage strategies Candidate centered appeals are used Party-centered appeals are used

Yes

No (or marginally)

Yes

Harmonization by party leadership needed: •  UDI

No

Decentralized harmonization needed (very low levels of harmonization observed in examples): •  RN •  PS •  PPD •  DC

Self-harmonized: •  FA •  Uruguay’s traditional parties (less harmonization observed)

Congress. I argue that the UDI’s involvement with Pinochet’s regime and its ties to business and socially conservative interests (such as the Opus Dei and the Legionnaires of Christ) help explain its very strong electoral performance since 1989 within its core constituency: the upper socioeconomic segments of Chilean society (Fontaine 2000). However, the UDI has also become more successful than other parties at recruiting voters among the poorest segments of Chilean society, including historical strongholds of the left. This strategy has enabled the Chilean right to show a “new face,” downplaying factors (i.e., its authoritarian past) that still hinder its electoral growth in non-core social groups (Garretón 2000). Regarding the UDI’s electoral growth, I argue that this party has pursued a dual representational strategy by extracting economic resources in exchange for ideological and interest representation from its “vote-poor/resource-rich” core constituency and using those resources to capture the vote of its non-core “vote-rich/resource-poor” constituency.9 This has been accomplished by developing a powerful grassroots network structuring non-programmatic linkages with poor constituents. During the dictatorship, the UDI was able to penetrate local politics and build a local political machine using state resources (Pollack 1999; Huneeus 2000a; Morales and Burgueño 2001; Klein 2004). By contrast, since the transition to democracy the party has relied on  

9 

See Kitschelt (2000) for a discussion of “dual-representation strategies” like this.

204

Segmented Representation

private donations to feed its local networks. The electoral growth in the party’s non-core constituency was also triggered by the emergence of a popular leader (and presidential candidate in 1999–2000 and 2005):  Joaquín Lavín. Lavín mobilized supporters, with “anti-politics” appeals differentiating him from traditional politicians. The UDI’s strong showing in the presidential elections of 1999/2000 and in the congressional elections of 2001 were therefore the cumulative result of both contextual and endogenous factors. Public discontent with the Concertación in the aftermath of the Asian financial crisis (1997–1998) provided a crucial opportunity for the electoral growth of oppositional parties. Drawing on its unique organizational matrix and strategic vision, as well as the development of a charismatic leadership and the availability of economic resources, the UDI was able to profit from this situation. In 2004 and 2009 the party still received the most votes at the congressional level. However, its presidential candidate in 2004 came in third behind the PS president-elect. In 2009, the party supported the candidacy of Sebastián Piñera of the RN. Erosion of the UDI’s leadership while the Concertacion and RN developed renewed presidential candidacies led to the UDI’s poor showing in the 2004 and 2009 presidential races. Despite its presidential misfortunes, the party’s grassroots political machine enabled it to perform surprisingly well in the 2004 and 2009 congressional races. Origins and Social Base Ironically, the UDI’s origins as a successful political party were tied to the emergence of an “anti-party” and “anti-political” movement in the late 1960s. The UDI originated as the political expression of the Movimiento Gremial, or Gremialismo. Strongly influenced by Spanish corporatism (Franquismo) the Movimiento Gremial was founded in the mid-1960s by Jaime Guzmán, a professor of law at Chile’s Universidad Católica (Cristi 2000). The Movimiento Gremial aimed to eradicate Marxism from Chile by creating a mechanism of comprehensive vertical representation. Such representation was to blur class and functional–organizational divides while depoliticizing society. Thus, Gremialismo sought to recraft politics by promoting the organization of specific interests in society while weakening (traditional) political parties, especially those of the left (Cristi 2000, pp. 23–44). The first nationally visible and explicitly political activities of the proto-party occurred in 1983, when, in reaction to popular mobilization against the military regime, the social network built by Gremialistas in Chile’s shantytowns organized a series of counter-protests in favor of



Strategic Harmonization of Segmented Linkages

205

the dictatorship. However, it was only in 1987, under the Ley de Partidos Políticos, that Gremialismo brought together the Frente Nacional del Trabajo and the Unión Nacional to form the RN. By the following year, however, internal strains within the RN had already led to the expulsion of Gremialistas from the party. In 1989, the UDI participated in elections for the first time as an autonomous party.10 Organizationally, the UDI originated as a homogenous, hierarchically organized movement of urban university students personally attached to Guzman’s leadership and to the Gremialista vision (Joignant and Navia 2003). When Pinochet came to power, Jaime Guzmán and a significant number of Gremialistas became close collaborators of his authoritarian regime. State retrenchment in the wake of neoliberal reforms created (strange) affinities between the “Chicago Boys” collaborating with Pinochet and Gremialistas advocating reduced state involvement in society, except to promote moral and religious values. Although initially conflictual, the relationship between Gremialismo and market-reformers strengthened and solidified under Pinochet, especially under the influence of high-level technocrats who were both Gremialistas and Chicago boys.11 The Gremialista movement thereby came to represent the dictatorship’s economic and political legacy, thus commanding special allegiance from business interests and conservative social groups such as Opus Dei and the Legionnaires of Christ (Huneeus 2007). Through their involvement in the authoritarian government, the Gremialistas pursued two central objectives. First, they sought to institutionalize the legacy of Pinochet’s rule. Guzman’s authorship of crucial segments of the 1980 constitution is the starkest example of this objective. Second, they worked to construct a new party that could become the main political force in the country in the event of re-democratization. The fulfillment of this second task, facilitated by institutional incentives in the 1980 constitution and the (binominal) electoral law introduced by Guzmán and his group, would be a crucial safeguard for Pinochet’s legacy after the inevitable (although, ideally limited) process of democratization that the UDI’s leaders were anticipating.12 The UDI’s characteristics and trajectory explain its unique ideological positioning in Chile’s party system. In the words of a former congressional   10  Guzman had nonetheless given birth to the party (“UDI por el Sí”) in 1983.   11  Miguel Kast, Gremialista chair of the National Planning Office (ODEPLAN), was especially influential in bridging the two groups (Huneeus 2001a).   12  For a more thorough and systematic account of the historical development of the party and its doctrine see Pollack (1999), Huneeus (2000a), Huneeus (2000b), Huneeus (2001a), Morales and Burgueño (2001), Joignant and Navia (2003), and San Francisco (2003).

206

Segmented Representation

Table 5.4.  Lower-chamber (1989–2001) and municipal election (2004) results per electoral pact and mainstream parties (%) Lower-chamber elections Main electoral pacts 1989 1993 1997 2001 2005 Concertación 51.50 55.40 Alianza 34.18 36.68 Main political parties Concertación DC 25.99 27.12 PS 10.40 11.93 PPD 11.84 Alianza UDI 9.82 12.11 RN 18.28 16.31 Other parties PC-Chi 4.38 4.99

Municipal elections 2009

1992 1996 2000 2004*

2008

50.51 47.90 51.77 40,4** 53.30 52.13 52.13 46.35** 28,271 36.26 44.27 38.7 39,58*** 29.67 32.47 40.09 38.12 38.35 22.98 18.92 20.78 12,95 11.05 10.00 10.02 9 12.55 12.73 15.44 11,56

28.93 26.03 21.62 20.90 8.53 10.70 11.28 11.41 9.21 11.71 11.41 8.23

15.97 10.25 7.73

14.45 25.18 22.34 21 16.77 13.77 14.12 16,22

10.19 3.36 15.97 19 13.44 13.60 15.54 14.97

17.58 14.66

6.88

5.22

5.14

6.55

5.09 3.24

3.93

Source: . Key * Average of electoral support in municipal council and mayoral elections which were held separately for the first time in 2004. ** Elections in which Concertación and Juntos Podemos united to form electoral alliances (“pacto por omission”). *** Election in which the Alianza pact was branded the “Coalición por el Cambio” as it included other minor parties.

candidate, “It is a popular party, it is a Christian party, and it is an economic [pro-market liberalism] party. And those three conditions together are lacking in the rest of the parties” (José Uriarte, UDI 2001 congressional candidate in Peñalolén, 2003). Although Guzman was murdered in 1991 by activists of the radical leftist group Frente Patriótico Manuel Rodríguez, the UDI managed to survive and prosper by following the guidance and example of its founder. Table 5.4 depicts the electoral trajectory of the party since re-democratization. While all other relevant parties in the system (with the exception of the PPD) lost electoral support in lower-chamber elections from 1989 to 2001, the UDI’s support base doubled over the same period. Since the presidential runoff of 2000 when the party’s presidential candidate obtained 47.7  percent of the vote, the UDI has significantly expanded its electoral share. In the 2001 lower-chamber elections the party became the most-voted party in Chile, obtaining 25.2 percent of the popular vote. In both 2000 and 2001, this electoral advance was crucial in reducing the historical gap favoring the Concertación over the Alianza.



Strategic Harmonization of Segmented Linkages

207

4 1997

3

2001

2

2005

1 0

Top 10%

Poorest 10% Richest 30–50%

–1 –2 –3 –4

Graph 5.1.  The Concertación’s social bases (lower-chamber elections) Source: Author’s own construction based on data from the Observatorio Político Electoral of the Universidad Diego Portales.

While declining slightly from its 2001 peak, the UDI remained the strongest party in the lower chamber in 2005 and 2009, obtaining 22.3 percent of the vote in 2005 and 21 percent in 2009. At the municipal level, the UDI has also experienced steady electoral gains, reaching almost 20 percent nationwide in 2004 and 20.05 percent in 2008. According to Hipsher’s (1996) and Altman’s (2004) reconstructions of each coalition’s electoral base, those who vote for the Alianza (and for the UDI, as the most-voted-for party within this pact) are located in the upper and lower socioeconomic strata. The Concertación, in contrast, obtains votes overwhelmingly from the middle classes. Evidence from recent lower-chamber elections (1997–2005) confirms this characterization. Graph 5.1 displays lower-chamber electoral support for Concertación parties in three sets of districts (poorest 10  percent, richest 10  percent, richest 30–50 percent).13 Bars for each election year represent the difference between the Concertación’s average national electoral support and the support it obtained in each social stratum. As the graph shows, the Concertación overperforms in the middle segment (richest 30–50 percent municipalities) and   13  Districts were classified in terms of their membership in poverty deciles (computed at the aggregate municipal level on the basis of the percent of households under the poverty line reported by the SINIM).

Segmented Representation

208 6 5

1997

4

2001

3

2005

2 1 0 –1 –2

Alianza Alianza Top 10%

UDI

Alianza

UDI

UDI Richest 30–50%

Poorest 10%

–3 –4 –5 –6 –7

Graph 5.2.  Alianza’s (and UDI’s) social bases (lower-chamber elections) Source: Author’s own construction based on data from the Observatorio Político Electoral of the Universidad Diego Portales.

underperforms in both the top and bottom poverty deciles (poorest 10 percent and richest 10 percent).14 As Graph 5.2 shows, the Alianza and the UDI exhibit a more complex pattern. With the marginal exception of the UDI in 2005, they had electoral surpluses only in the top 10 percent, with deficits in both the middle and lower segments. However, the UDI has achieved sharp electoral growth in the poorest 10 percent of districts, decreasing its electoral deficit in that group from 6 percent in 1997 to just 2 percent in 2005. In short, while it still obtained less than its national average in that social segment, the UDI has been the fastest growing political party in the country’s poorest districts. Further illustrating this point, Graph 5.3 displays the UDI’s electoral growth in districts of four different poverty deciles, taking electoral returns in 1997 as base 100. Although electoral growth (particularly in 2001) was evident across the board, the party’s biggest electoral gains were in the poorest two strata. There were also modest increases in areas associated with the historical base of the party (top 10 percent districts), where previous RN support has migrated to the UDI. The party has gained less in middle districts, however. Finally, Graph 5.4 shows the UDI’s performance in the 1999 and 2005 presidential elections under Lavin’s leadership, when the party performed   14  These analyses are drawn from data collected by the Observatorio Político Electoral of the Universidad Diego Portales.



Strategic Harmonization of Segmented Linkages 6 5

209

1999 2005

4 3 2 1 0 –1

Richest 10%

Richest 20–30%

Poorest 20–30%

Poorest 10%

–2 –3

Graph 5.3.  Electoral evolution of the UDI across social strata (lower-chamber elections, 1997 = 100) Source: Author’s own construction based on data from the Observatorio Político Electoral of the Universidad Diego Portales. 260 240

2001 2005

220 200 180 160 140 120 100 Total

Richest 10%

Richest 20–30%

Poorest 20–30%

Poorest 10%

Graph 5.4.  Votes for the UDI in presidential elections (first round: 1999 and 2005) Source: Author’s own construction based on data from the Observatorio Político Electoral of the Universidad Diego Portales.

much better in popular sectors than in its historical stronghold, suggesting that Lavín’s charismatic leadership greatly helped the party capture the popular vote. In sum, the UDI’s electoral performance has been most positively notable at the two extremes of the social ladder, maintaining a solid base in its core

Segmented Representation

210

constituency (the upper segments of Chilean society) while making steady inroads into a non-core electoral base (the popular sectors). The following sections seek to explain this electoral trajectory and social base configuration. The Core Constituency The UDI’s linkage to upper segments of society derives from the party’s historical trajectory as the perpetuator of Pinochet’s legacy of economic reform, social conservatism, and anti-Marxism. This ideology also crystallized in Gremialismo, which provided the basis for party-formation. The link between the party, business elites, and conservative sectors translates into a solid vote based on strong partisan and programmatic identifications frequently reinforced through media addresses by congressmembers and national candidates, as described in Chapter 3. The UDI’s linkage to upper segments of Chilean society is also based on interest representation in Congress, which party strategists see as a fundamental selling point for business elites who not only electorally support the party but also contribute resources for funding partisan activities elsewhere. Why would business elites contribute disproportionally to the UDI, particularly back when the party commanded a very small congressional contingent? According to a top party leader, “Our leaders convinced business elites that the party would be able to protect the market-oriented model introduced under Pinochet, aided by the special majority requirements that Jaime (Guzmán) included in the 1980 constitution” (Anonymous UDI national officer, personal interview, 2008). Recently available roll-call voting evidence shows the UDI to have been the most systematic defender of the market-oriented reforms introduced under the dictatorship. An analysis of 2006–2008 shows very high levels of internal discipline in 93  percent of the Alianza’s registered congressional votes. However, the parties’ respective congressional voting records in the remaining 7 percent (approximately ninety bills) shows the UDI to have been the most responsive to business elites’ interests. Using Visconti’s (2011) coding scheme and analysis, Graph 5.5 presents the mean partisan vote in favor (coded as 1)  of a series of legislative bills. The series is divided into eight categories, comprised of about five bills each.15 As Graph 5.5 shows, much more consistently than the RN, the UDI systematically opposed legislative packages potentially hurting business interests or  

15 

I thank Giancarlo Visconti for sharing these data and results.



Strategic Harmonization of Segmented Linkages

211

100% Percentage of Parties’ Congressional Representatives Voting in Favour of Law Initiatives

Renovación Nacional

UDI

90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Protections for Micro/Small Businesses

Minimum Wage Increases

Enhance Labour Regulation Popular Market and Control Election of Regulations (vs. Big Regional Companies) Councils

Greater State Subsidies in Education

Reforms Tax Reforms to the Pension System (vs. AFPs)

Graph 5.5.  Dissident roll-call votes from within the Alianza (2006–2008) Source: Author’s own construction based on data and coding from Visconti (2011).

redistributing resources to lower social strata, such as minimum wage increases, tighter market and labor regulation, state education subsidies, and tax and pension system reforms. In short, the UDI seems to have fulfilled its promise to protect business elites, protecting the legacy of the authoritarian regime in exchange for financial support. Campaign spending evidence suggests that the relationship between the party and business interests has been mutually beneficial. Graph 5.7 displays self-reported campaign spending averages for all lower-chamber candidates competing in 2005 races, under each party’s banner.16 The UDI spent an average of US$90,000 per campaign, US$20,000 more than the PPD, which came in second in spending. The UDI overspent its Alianza partner (RN) by an average of US$30,000 per campaign.17   16  I thank Rafael Piñeiro for providing access to his database on campaign spending. See also, Huneeus et al. (2007), Diaz et al. (2006).   17  Campaign finance legislation, which began being implemented in 2004, only distinguishes between private (anonymous and reserved) and public contributions. Therefore, there is no information publicly available on the private sources financing each party’s campaigns. However, the available information allows some informed speculation. Whereas anonymous contributions cannot exceed a maximum of US$800 and are therefore appropriate for individual contributions, reserved funds range from US$800 to US$22,500. A  party financed by business interests and wealthy donors should therefore receive more reserved donations, as these types of donors are

Segmented Representation

212 100000 90000 80000

Average Candidate Campaign Expenditures in USD for December 2005 (1US D=515 Chilean Pesos)

70000 60000 50000 40000 30000 20000 10000

I D U

PP D

PS

C

N

D

Pa r ic

al

R

ty

hi -C PC

H

R

ad

um an

is

tP ar

ty

0

Graph 5.6.  Average self-reported campaign expenditures by parties’ Lower-chamber candidates (2005) Source: Author’s own construction based on Rafael Piñeiro’s dataset and Servicio Electoral.

3500.0 Campaign Expenditures Origins of Campaing Funds in Millons of Chilean Pesos

Expenditures Private Donations

3000.0

Public Funding

2500.0 2000.0 1500.0 1000.0 500.0 0.0

UDI

RN

DC

PS

PPD

Graph 5.7.  Average self-reported campaign expenditures by parties’ municipal candidates (2008) Source: Author’s own construction based on data supplied by Proyecto Probidad, (2009).18

more likely to make larger donations. My analysis assumes, as is usual in the literature, that the gap between actual and reported expenditures is relatively equivalent across parties.   18  Proyecto Probidad, Financiamiento Político en Chile. Funded by the IADB and currently implemented by Chile Transparente (2009).



Strategic Harmonization of Segmented Linkages

213

In the case of municipal elections, according to a campaign spending report available for the 2004 municipal elections, the UDI spent most, outspending its closest rival (RN) by more than 100 percent.19 Furthermore, whereas reserved donations accounted for approximately 40 percent of the Concertación parties’ campaign expenditures, they accounted for more than 60 percent of the UDI’s.20 Available evidence for the 2008 municipal elections highlights once again the exceptionally heavy role of private donations in the financing of the UDI.21 As Graph 5.7 illustrates, the UDI was the party with the highest reported campaign expenditures. In comparison to those of other parties, the UDI’s campaigns were disproportionately funded through private donations. In short, the UDI has developed a financial edge over its competitors by extracting financial resources from its core constituency. Those resources are then strategically invested in crafting different types of linkages in poor communities, where the party has developed a non-core electoral base. The UDI’s Outreach to the Popular Sectors: Authoritarian Clientelism22 Immediately after the military coup, the dictatorship headed by Pinochet began implementing a municipal reform with lasting implications for local politics in Chile (Rehren 1999). This reform created a hierarchical, authoritarian structure in which mayors acted as the dictatorship’s local agents. Pinochet thereby sought to control the loci of political mobilization that had been most dynamic under the Unidad Popular (UP) government.23 At the same time, these appointed mayors gained significant autonomy from their local communities while receiving increasing amounts of public resources channeled through targeted social policy funds (i.e., housing and family subsidies or employment plans). At a time when poverty and economic malaise were prevalent, those funds presented an opportunity at the local level to resurrect patronage and

  19  Instituto Interamericano de Derechos Humanos 2005, p. 56.   20  Instituto Interamericano de Derechos Humanos 2005, p. 58.   21  Unfortunately, systematic analyses of campaign funding in congressional races are not yet available. However, for the UDI spending figures for congressional and municipal races coincide. There are therefore unlikely to be significant discrepancies between the proportion of campaign spending coming from private donations in the two sets of races.   22  For a more comprehensive account of the origins of the UDI’s electoral strategy under Pinochet, see Huneeus (2000a), Huneeus (2000b), and Klein (2004).   23  For a description of local politics before 1973, see Valenzuela (1977).

214

Segmented Representation

clientelism, features of the traditional party system that both the military regime and the Gremialistas rhetorically opposed (Klein 2004). Taking advantage of this opportunity, Gremialistas created a political stronghold in Chile’s shantytowns. To implement their political strategy, Guzmán and his group created an “apolitical” (although “anti-Marxist”) public service organization. Committed to stripping the left of its electoral base in the shantytowns, it successfully worked within the authoritarian regime to appropriate state resources that were then used to build support among popular sectors (Huneeus 2001a; Klein 2004). This strategy allowed UDI activists to craft an image differentiating them from “traditional politicians,” while at the same time creating a political structure that other political parties, banned and repressed at the time, lacked. The UDI’s work in the National Youth Secretariat (SNJ), in the Office for National Planning (ODEPLAN), and in the mayoralties of numerous local governments came to be essential for funding a powerful network that helped to expand Gremialismo among Chile’s poor (Pollack 1999; Huneeus 2001a; Morales and Burgueño 2001; Soto 2001; Klein 2004). The UDI’s progressive insertion into poblaciones (shantytowns) was achieved by organizing independent youth centers that collaborated with the SNJ in each area. Appointed mayors also organized social activities through which Gremialismo promoted its political project (Soto 2001) and selectively distributed much needed social assistance. Although pobladores previously mobilized by leftist groups initially resisted the UDI’s arrival into poblaciones, this resistance decreased steadily over time. Indeed, former local leaders of the Partido Demócrata Cristiano, and even some prominent leftist ones, were attracted to the party’s Departamento Poblacional (Soto 2001), which frequently organized social assistance, “Operativos de terreno,” while relying heavily on clientelistic transactions with poor constituencies (Hipsher 1996; Morales and Burgueño 2001; Barozet 2003). After democratization, the political dividends of this strategy became evident. Ten out of fourteen UDI deputies elected in 1989 had been appointed mayors under Pinochet in municipalities of the same congressional districts in which they ran for office (Morales and Burgueño 2001). The Political Opportunity Structure for Seducing the “Soft Vote” Two tactics allowed the UDI to profit from the post-transitional political opportunity while pursuing its overarching strategy. After the transition to democracy, Chilean partisan identities and traditional political apparatuses



Strategic Harmonization of Segmented Linkages

215

weakened, reducing traditional parties’ ability to mobilize support in popular communities (Oxhorn 1995; Roberts 1998; Posner 1999 and 2004). As a result, and in contrast to the linkage dynamic observed in wealthy neighborhoods, grassroots activities, constituency service, and particularistic exchanges became central elements for electoral mobilization. Successful politicians tended to be those who were able to pay a household’s utility bill during the campaign period, to offer legal or medical assistance, or to distribute TV sets, food boxes, optical lenses, equipment for a neighborhood soccer club, or cakes for bingo parties organized by community organizations.24 In sum, politicians who developed personal contact with members of poor communities—structuring efficient problem-solving networks or distributing cosas (“stuff ”)—came to enjoy a competitive advantage over more personally distant candidates. Given the decline of partisan identities and national partisan apparatuses, local networks have become structured around individual candidates and incumbent congressmembers and mayors, who have strategically chosen to downplay their partisan identity (see Chapter 4). In contrast, Renovación Nacional, the UDI’s partner in the Alianza, is a party of cadres lacking a well-developed and competitive partisan apparatus at the local level (Barozet and Aubry 2005). Two interrelated tactics proved important for reaching what UDI’s strategists describe as the “soft vote” (non-partisan and non-ideological voters) in Chile’s popular sectors. First, the UDI progressively expanded a disciplined party organization able both to capture more campaign resources than its competitors and to deploy those resources strategically through a pervasive grassroots political machine. Second, the party consolidated a national leadership able to moderate the party’s image while creating empathy among disillusioned voters. Among this group of voters, women were especially attracted to the party and to Lavin’s leadership.25 Political Organization and the UDI’s Partisan Apparatus After democratization, the UDI’s unique organizational penetration of society increased. Maintaining high levels of vertical integration, it attracted new activists while increasing internal discipline. The latter enabled party leaders to strategically target the resources that the party obtained from wealthy

  24  Personal interviews with local council-members and former and current congressmembers in 2002 and 2003.   25  For more information, see e.g., Moulián (2004), Altman (2004), and Morales (2008).

216

Segmented Representation

donors to specific congressional and municipal campaigns identified by survey research and centralized political marketing technology. Continuing the party’s distinctive tradition of promoting the engagement of young people in Gremialista social activities after democratization, the UDI persisted in Gremialismo’s aim of depoliticizing interest groups in a manner well adapted to the anti-partisan climate of the late 1990s. As a congressional candidate explained: We already had too much politics in Chile. Look what happened in the 1970s. Politics should be kept to a minimum [ . . . ] We do not want to be full of politicized masses, because then you have internal problems; people fight. What we want is to create is an environment in which youth who want to participate in society [by] solving problems for the people can do it, and that is also a way you can attract young people who are not interested in party politics [ . . . ] We send those interested to live with a host family in a given shanty town for two or three months. It is a way to let them grow acquainted with how poor people live in Chile and to realize what the problems are. (José Jara, UDI’s 2001 congressional candidate in La Florida, personal interview, 2003)

The UDI thereby started attracting young people from other parties (especially the RN), incorporating them as “social activists.” These (frequently non-partisan) social activists engaged in systematic fieldwork organized by independent youth groups that worked indirectly for the party.26 This strategy, coupled with religious appeals encouraging youth to volunteer in social service operations, made the party more successful than others at incorporating youth into its activities in a context of general apathy towards parties.27 As one youth leader recounted: Now we are receiving many youth from RN. Do you know how they teach youth about poverty in the RN? They show them movies. [ . . . ] We go and just get our feet in the mud. [ . . . ] If there is flooding, they [RN] usually are the ones that donate mattresses. But we are the ones helping to put those mattresses on the floor. And it is different. It feels different, and creates a different relationship with the people. Building a playground for poor children is great. You know that it will get destroyed in two days, [ . . . ] but the human relationship you create is worth it. (Rodrigo Bordachar, UDI’s youth leader in La Florida, personal interview, 2003)

  26 Social operatives organized by Gremialista student organizations (particularly at the Universidad Católica) during summer and winter holidays also work for this aim.   27  The involvement of Catholic networks in these social operations might also explain the remarkable enthusiasm, energy, and mystique characteristic of the operativos de terreno that UDI youths have implemented.



Strategic Harmonization of Segmented Linkages

217

At the same time, the UDI handpicked a small group for more political activity. The young activists selected by the party to play a political role share similar profiles and experiences. They are recruited from top private and Catholic universities (usually from law departments) and sent to pursue internships either in Washington DC’s Leadership Institute or at the Universidad de Navarra in Spain (José Jara and José Uriarte, personal interviews, 2003). Whenever possible, they run for Congress soon after achieving eligibility at age 21, in districts where they are very likely to lose and with virtually no financial help from the party. They are then expected to move on, either running with party support in more competitive districts or taking staff positions in one of the UDI’s municipal governments, and eventually in the national government. Their initial, unsuccessful campaigns have often spawned new youth groups, which have gone on to leadership development and social service activities, thereby expanding the activist base of the party. This account of his trajectory by a UDI youth leader is in many ways typical: I started working with Jarita, when he ran for Congress. We supported him, but we knew it was a lost campaign. Now, he has gone, but we created a youth group, the “Corporación Jóvenes de la Florida” through which we try to help Mayor Pablo Zalaquett and get more support for the party here [  .  .  .  ] We are now preparing fieldwork activity with university students who are in medical, dental, and law schools. We get together, ask our families and some friends that work in different companies to help us get stocks of food or medicine, and then we organize an operation in a shanty town. [ . . . ] As they do not identify us as a political group, we have even been able to use the political structure of Carlos Montes [a socialist deputy] to coordinate with the neighborhood organizations and get the logistics done [ . . . ] In the background on the medical receipt or on the legal file we just have a photograph of the mayor and the logo of our organization. (Rodrigo Bordachar, UDI’s youth leader in La Florida, personal interview, 2003)

This “commitment to public service” not only distinguishes UDI youth, it is also central in the day-to-day work of incumbent deputies, council members, and mayors who receive and process a wide range of demands from local organizations and individuals on a regular basis. Eugenio González, Lavín’s 2000 campaign strategist, explained: In the popular sectors we expect our council-members and mayors to go where the problems are. They need to keep their feet well into the mud. And in terms of territorial structure, we have realized that one has to maintain at least one activist in each zone to keep track of the problems and seek solutions. Then, when election time comes, you have a structure that is dormant there but that you can mobilize very easily. (Eugenio González, personal interview, 2003)

218

Segmented Representation

Although it is true that all council members receive many demands each day, the UDI’s council members usually receive the most requests and are seen as those who are most responsive and who “give away more.”28 In sum, the UDI’s involvement of young activists has played in important part in helping the party to extend its electoral appeal and organizational penetration of society to districts in which it previously lacked influence. This has also strengthened the party’s capacity to provide services to the community, generating a broad menu of social assistance initiatives as well as electoral clientelism (Hipsher 1996). The UDI’s intensive and systematic “fieldwork” in popular districts has complemented its simultaneous use of the media to cultivate an umbrella organization and national leadership. Given the absence of stringent campaign finance regulations in Chile (Villar 2006), the UDI’s unique ability to secure and administer financial resources through its private sector links has also given the party a major competitive advantage. This factor is particularly relevant now that state reform has diminished rival parties’ ability to rely on the state apparatus for sustaining patronage and clientelistic networks. A Demócrata Cristiano (DC) congressmember thus described the main difference between her and the UDI candidate this way: Their advantage is based on money. Once, he [the UDI congressmember] complained to me: “I got them used to this and now they are costing me a lot.” If he has to give out a reward for a lottery, he does not send a set of kitchen utensils, as I would do. He sends a brand new bicycle. And therefore every organization wants him to be their godfather, because he gives people better stuff. But that is not paid for out of his pocket. That’s paid by his friends’ donations, by businessmen, by people who benefited while he was the mayor. (Eliana Caravall, DC congressmember in District 27, personal interview, 2003)

Indeed, according to one former president of the PPD, the Alianza’s clear competitive edge led Concertación leaders to develop an “ideology of corruption” justifying use of public resources to finance Concertación campaign activities.29 At the local level, this ideology justified using municipal governments either directly as political machines or indirectly for generating campaign funds by illegally awarding municipal contracts to private companies paying kickbacks.30 This gave the UDI a double advantage. As an opposition

  28 Participant observation in congressional districts and interviews with Margarita Cofre (2003) and Osvaldo Silva (2003).   29  Jorge Schaulsohn, “Letter to the Tribunal Supremo of the Partido Por La Democracia” (Santiago, 2006).   30 Personal interviews with Osvaldo Silva (2003) and two officers of the División Municipalidades of the Contraloría General de la Nación (2003). This rationale was confirmed



Strategic Harmonization of Segmented Linkages

219

party, it was able to capitalize on public discontent with both the economic slowdown that lasted until 2002 and the corruption scandals first erupting in the late 1990s and continuing throughout the 2000s. At the same time, better access to private financing allowed the UDI to develop and “feed” an increasingly encompassing social network operating as a political machine for attracting non-core constituents. Internal discipline has been another key to the UDI’s success. Thanks to the cultural homogeneity and personal ties uniting the UDI’s national leadership, the party is much more internally disciplined and hierarchically organized than other Chilean political parties (Joignant and Navia 2003). Furthermore, the party’s religious commitment, linked to Guzman’s apostolic vision crystallized in Gremialismo, both supports adherence to a strict internal hierarchy and provides party activists and leaders with a mysticism and unity very useful for pursuing the Gremialista “crusade.” In the words of one young UDI party member and congressional candidate, “Within the UDI, we do not have factions [ . . . ] We have a democratic orientation, but we do not have elections. We trust our leaders. Guzmán, who was an exceptional human being and an exemplary Catholic, personally formed them” (José Uriarte, UDI’s 2001 congressional candidate in Peñalolén, personal interview, 2003). This internal discipline helps party leaders centralize decision making (usually assisted by survey analysis), avoiding internal conflicts and achieving vertical integration. As one party activist put it: My wife worked as a congressional candidate for six months, and then she was removed as a result of the pact with Angélica Cristi [a former RN congressmember who joined the UDI in 2001]. It was obviously good for the party, so my wife just came home silently, without making any public statements, as would be the case in other parties. Here, if you have to head home, you just say, “It is not my moment, I will go home.” If they call you, “It is my moment, I will go work.” We are extremely disciplined because we know that those who are at the top know their job and we should trust them. (Eugenio González, Lavín’s 2000 campaign strategist, personal interview, 2003)

This top-down decision-making style has allowed the party to allocate campaign resources more efficiently than rival parties, in which candidates seek their own campaign resources through personal networking with public officials, party leaders, and private donors. In the UDI, the national leadership decides where to allocate the resources it captures by strategically analyzing each candidate’s situation.31 Therefore, the UDI’s competitive financial edge is not only a product of the to me by several council-members and by Concertación congressmembers regularly complaining about the UDI’s superior private resources.   31 The description offered here is based on evidence collected during fieldwork in 2002 and 2003.

220

Segmented Representation

greater amount of resources available to the party, but also due to the UDI’s more efficient allocation of campaign funds. This greater harmonization also helps to unify and align local campaigns with national ones regarding tactics, issues, and marketing material.32 The higher degree of internal commitment that exists within UDI’s rank and file has also allowed the party to avoid the increasing split between local and national leaders that has weakened other Chilean parties. Indeed, UDI council members are frequently found working closely with their district’s congressmember. This harmonization has enabled the party to maintain a firm grasp at the grassroots level. The “Popular Party” and Lavin’s National Leadership Joaquín Lavín, the UDI’s 1999 and 2004 presidential candidate, guided the party’s successful crafting of a more moderate image, a strong national structure, and electoral growth. The UDI’s cleansing of its image involved the progressive detachment of its leadership from the “dark side” of Pinochet’s regime. Lavín was ahead of the curve in this regard, especially after Pinochet’s 1998 detention in London (Silva 2001). In particular, this new detachment meant refraining from publicly justifying the human rights violations committed by the dictatorship, while continuing to recognize “[Pinochet’s] success in transforming the Chilean economy and putting a halt to Marxism.”33 One of the central claims I heard at the UDI’s youth recruitment workshops was that the Gremialistas’ participation in that regime was crucial not only in enabling successful economic reform but also for “moderating” the human rights violations that took place.34 Additionally, after the economic crisis hit in 1998 the UDI’s more moderate stance also involved media addresses on social policy. At this particular conjuncture, party leaders insistently demanded that the government provide better social protection for the poor and unemployed. The UDI’s rhetorical swing in this respect was symbolized by its self-proclaimed identity as “The Popular Party.”35   32  Beyond this centralized allocation (mainly based on business and private donations to the party), each candidate is free to seek and spend his or her own resources.   33  Memo: “20 preguntas a un gremialista” and participant observation of a youth education seminar in La Florida, July 2003.   34  In a carefully documented historical account, Huneeus (2000a) fails to find evidence justifying the claim that Gremialistas actively sought to moderate human rights abuses under Pinochet.   35  The UDI adopted this new slogan in July 2002.



Strategic Harmonization of Segmented Linkages

221

In that year, the leadership of Joaquín Lavín gained center stage. Lavín’s leadership has been characterized alternately as neo-populist and technocratic (Silva 2001). It developed through an “anti-politics” campaign premised on “taking care of regular people’s everyday problems” by devising concrete solutions instead of discussing more far-reaching issues (Barozet 2003).36 Meanwhile, the UDI characterized the political style of the Concertación (and the RN) as elitist and therefore distant from people’s everyday lives, needs, and realities. This message meshed well with the mass social alienation from the Concertación’s governments that was then reaching its peak. Operationally, Lavín’s high visibility as the “innovative” Mayor of Las Condes, an affluent municipality in Chile’s capital city, favored the development of his leadership.37 Lavín was hailed as the creator of a new model of pragmatic and participatory political management. “People’s real problems” were broadly defined, and his response to them ranged from putting in place a private municipal security service and programs for youth, to fighting environmental pollution in Las Condes by “punching” clouds with an airplane to produce rain. Lavín’s “participatory style” referred to the use of large-scale public opinion surveys that informed policies and media addresses. In his previous political position as mayor of the municipality of Santiago Center, Lavín’s use of that approach had been less successful. The social problems in the poorer municipality of Santiago ran deeper. Lavín therefore devised some policies aimed specifically at highlighting the UDI’s response to less well-off citizens. The range of actions in this respect was wide, and included a municipal program providing daycare to poor families. During the summer, Lavín built a beach (popularly named “La Playa de Lavín”), giving poor people staying in Santiago the opportunity to enjoy the summer. During the winter, he then brought down snow from the Andes to create an artificial ski slope. The snow melted in a matter of hours, but the media coverage of Lavín’s initiative was long-lasting. Indeed, despite the general disdain shown by Concertación politicians, the “innovative” and “close to the people” style incarnated by Lavín and replicated across the country by UDI’s municipal governments garnered extensive media attention and publicity. For the presidential campaign of 1999–2000, the UDI organized an extensive fieldwork campaign, running a “town-by-town” tour in each Chilean

  36  While this was a traditional stance of Gremialismo, it gained further strength and visibility with the leadership of Lavín. See Pollack (1999) and Cristi (2000).   37  Joaquín Lavín was elected Mayor of Las Condes in 1992 with 31 percent of the vote. He was reelected in 1996 with 78 percent. In 2000 he ran for the mayoralty of Santiago Center, obtaining 61 percent of the vote.

222

Segmented Representation

province. This tour was intensely supported by modern marketing techniques, including targeted radio addresses and telephone calls in advance of Lavín’s visit to each town. The day before the UDI caravan was scheduled to visit, Lavín would make “personal” phone calls inviting locals to the central square to meet him in person to discuss their problems.38 His addresses tackled specific issues pertinent to that particular community according to surveys and key informants (party activists in the district). Operationally, Lavín’s crew created a registry of 3,000 local radio stations for implementing this strategy in a “segmented” and “low cost” way (Pollack 1999; Cristi 2000; Barozet 2003). Meanwhile, Lavín recorded 300 different personalized phone messages and shot and autographed an average of 2,000 Polaroid photographs per day.39 Lavín’s rising popularity helped the UDI develop an umbrella-like national leadership supporting its presidential candidate. This national umbrella went on to promote new candidates in congressional districts and municipalities where the UDI lacked a significant presence.40 Furthermore, the value of Lavin’s endorsement was used for negotiating better conditions with RN for the UDI’s own candidates, and increasing the party’s legislative share by attracting non-UDI candidates to its ranks. The “photo with Lavín” also became critical in local candidates’ competitive edge over RN. Indeed, former members of RN who wanted to enjoy the “advantages of being UDI” while escaping the internal strains that the UDI’s growing hegemony was generating within RN defected to Gremialismo.41 The importance of Lavín’s national leadership for the overall strategy of his party was confirmed by the UDI’s electoral decline in 2004 and 2005. The economic recovery and rising popularity of President Lagos (Morales 2008), along with the national government’s massive investments in targeted social policies and infrastructure all factored in Lavin’s declining popularity. The changing national context and erosion of Lavin’s image after six consecutive years of tenaciously opposing an increasingly popular government hindered his electoral bid. Meanwhile, the emergence within the Concertación of Michelle Bachelet, a “renovating” female leader benefiting from Lagos’ popularity while signaling a shift within the Concertación in favor of non-traditional politicians, also reduced Lavín’s chances.   38  See Barozet (2003).   39  Personal interviews with Eugenio González (2003) and José Luis Uriarte (2003).   40  The UDI’s expansion in the south of the country, a region traditionally dominated by the RN, has been a case in point.   41  Personal interview with Maria Angélica Cristi (2003). The internal strains between transfugas (rightist hardliners sympathetic to UDI) and the RN’s liberal wing represented by the party’s national leadership launched a pervasive internal crisis within the RN.



Strategic Harmonization of Segmented Linkages

223

90

Average popularity (in %)

80 70 60 Lagos Lavin Bachelet

50 40 30 20 10 0 2002

2003

2004

2005

Year

Graph 5.8.  Yearly average popularity ratings for selected presidential leaders (2002–2005) Source: Author’s own construction based on CEP survey.

Bachelet’s candidacy attracted politically disillusioned voters, including poor women and other electoral segments that had previously supported the extreme right. With Bachelet running, these segments switched to supporting the PS presidential candidate (Morales 2008, pp. 25–8). As Graph 5.8 shows, the popularity ratings of President Lagos and Joaquín Lavín mirrored one another in the 2002–2005 period, with Lavin becoming less popular as the election of 2005 neared and Lagos’ popularity rose. Meanwhile, although her popular support did diminish once she began being perceived as a serious presidential contender, Michelle Bachelet still boasted more popular support than either Lagos or Lavín. Nevertheless, in both 2005 and 2009 (and without even running a presidential candidacy in 2009), the UDI was Chile’s most electorally successful party, resisting the arguably strong coattail effects of Bachelet’s PS and Sebastián Piñera’s RN presidential candidacies in the respective years’ presidential elections. Corollary: The UDI’s Harmonized Dual Strategy Precisely considering the levels of hostility and suspicion that exist against the UDI in Chile, the party’s capture of a sizable non-core constituency is analytically puzzling. Contrary to conventional wisdom, UDI’s segmentation strategy did not lead to trade-offs, but rather to significant synergies between linkage types. The UDI’s strategic adaptation shared several features with other

224

Segmented Representation

successful instances of conservative mobilization in the region. The reliance on media-based appeals and the development of a strong anti-establishment and anti-politics leadership are among those features. Yet, the party also built up a powerful party machine at the grassroots, while investing resources in youth formation and party building. More internally coherent and disciplined than others in Chile, the UDI established itself as the most successful party within its national system, a feat also accomplished by the FA in Uruguay, albeit by different means.

T H E FA I N U RU G UAY In October 2004 the Uruguayan left, led by the FA, came into power, breaking 175 years of traditional party electoral dominance. The FA’s presidential triumph was coupled with attaining, for the first time since 1966, an absolute congressional majority (50.7 percent), marking a political watershed in the country’s history (Altman and Castiglioni 2006; Buquet 2005; Lanzaro 2004). The FA’s successful electoral bid was the result of the party’s complex transition from a predominantly urban, labor-based, center-left mass party to an increasingly catch-all, professional electoral organization. Operating within an opportunity structure marked by the crisis of Uruguay’s state-centric sociopolitical matrix (Cavarozzi et al. 2003), and the decreasing capacity of both traditional parties to satisfy and retain their historical constituents, the FA successfully moderated its platform while consistently opposing market reform. The party thereby mobilized its core constituency through programmatic appeals, defending the state-centric model and its stakeholders’ social interests. This programmatic stance allowed the party to maintain its core constituency in middle-class segments and organized labor. However, it was a complex intertwining of personality-based, community service linkages and localized efforts to enhance the social organization and political awareness of lower-class constituencies that enabled it progressively to win over an electorally critical peripheral constituency. The party also developed two powerful charismatic leaderships that helped it broaden its appeal: those of Tabaré Vázquez and José Mujica These leaders moderated and widened the FA’s electoral offer. The party’s territorial activists also took up brokerage roles and particularistic problem-solving activities. The FA’s capacity to develop and sustain a devoted network of partisan activists working in every locality of the country was pivotal in this last regard, especially when traditional partisan structures atrophied.



Strategic Harmonization of Segmented Linkages

225

From 1990 onward, the FA municipal government in Montevideo (home to approximately half the electorate) also played a decisive role in building the FA’s eminently successful 2004 electoral coalition. The municipal government promoted moderation, proved the party’s capacity to govern effectively, helped the FA penetrate strongholds of the traditional parties in Montevideo’s periphery, and catalyzed the charismatic leadership of the FA’s first elected president, Tabaré Vázquez. Widespread public discontent with both traditional parties and the policies they promoted during the 1990s, among both the party’s core and peripheral constituency, also significantly boosted the FA. This shared discontent, and the identification of the FA as a viable alternative to the traditional parties, promoted greater programmatic homogeneity across the two constituencies than in the case of the UDI. As a result of this trajectory, the FA currently has a double support base in Uruguayan society, comprised of its original constituency (urban middle sectors with ties to the old ISI model) and a new constituency (with voters in the informal labor market, the poor, and rural inhabitants). This new constituency gradually became disenchanted with the traditional party elite as a result of economic hardship and the decline of traditional clientelism. The FA was able to attract this new constituency, while maintaining a dense and very active internal organization and its programmatic ties to its core constituency. In short, as an opposition party the FA succeeded in articulating a successful electoral strategy that gradually allowed the party to grow in sectors of the population that had historically supported the traditional parties, without alienating its traditional leftist constituency. To do so, the party also embraced programmatic and electoral alliances with traditional party splinter groups and the centrist Nuevo Espacio (which eventually returned to FA), all the while preserving party discipline. The social and political opportunity structures that emerged in Uruguay during the 1990s helped the FA develop its strategy successfully. Institutionally, support for the two traditional parties slowly eroded, and no viable challenger emerged to the left of the FA. The FA capitalized on those enabling conditions by pursuing a complex pattern of partisan adaptation, both segmenting its linkage strategies to attract different constituencies and effectively unifying those segmented linkages. Origins and Social Base The FA was born in 1971 as a popular front of communists, socialists, Christian democrats, leftist independents, and splinters from both of Uruguay’s traditional parties (catch-all parties that had dominated Uruguayan politics since the

226

Segmented Representation

nineteenth century). Today the FA is comprised of twenty-nine fractions.42 The FA is today one of the region’s oldest and most institutionalized leftist parties, and the most resilient example of a leftist popular front in Latin America. After contesting the election of 1971, the party consolidated its social roots during the authoritarian regime, under which it—like other parties—was banned. Hardest hit by political prosecution and exile during the dictatorial period, Frenteamplismo quickly consolidated as a powerful political subculture that became associated with political resistance and the fight for democracy (Queirolo 1999). Whereas the Blanco and Colorado identities are rooted in the civil wars of the nineteenth century, the FA’s epic is based on popular opposition to the military regime (1973–1985). The FA’s appropriation and reinterpretation of Batllismo and its control of key vehicles of political socialization also helped to solidify a strong partisan subculture, which became instrumental in maintaining party unity as it became more heterogeneous internally (Lanzaro 2004). The high transmission rate of Frenteamplismo partisan identity from one generation to the next, as compared to traditional party identities, attests to its relative strength (see Moreira 2000a). The development of this strong political subculture facilitated the FA’s virtual hegemony on the left, while the lack of viable political alternatives to the left of the FA facilitated the party’s eventual shift to the ideological center.43 Table 5.5 illustrates the electoral trajectory of the FA. As shown in the table, especially in the 1990s, the FA was able to make continuous electoral inroads until obtaining an absolute majority in the election of 2004. However, when the performance of the party is analyzed across social segments groups, it becomes clear that voter support grew differentially and unevenly across distinct social strata.44 Graph 5.9 describes the FA’s historical performance in eighteen Montevideo neighborhoods (one in each of the city’s municipal zones). I further classify the neighborhoods into terciles reflecting different levels of social development.45 I then apply the same analytical strategy to

  42  The Effective Number of Fractions of the FA has fluctuated between 3.6 in 1971 and 4.2 in 2004, reaching its lowest level (2.3) in 1989 (Yaffé 2005).   43  In 2006, a radical leftist coalition (Asamblea Popular) emerged to the left of the FA, but it has not become a significant challenger to the party.   44  See Moreira (2000b and 2005) for additional evidence and a similar argument. The data presented in this section partially draws on hers.   45  Terciles were computed over a one-dimensional factor-analysis solution combining the percentage of population with at least some university education in 1999 (ranging between 1.5 percent and 28.1 percent) and the percentage of households living below the poverty line in 2004 (ranging between 3.2 percent and 48.3 percent). The two variables are correlated at a 0.98 level (Spearman’s coefficient).



Strategic Harmonization of Segmented Linkages

227

Table 5.5.  Electoral results 1942–1999 (%) Colorado Party 1942 1946 1950 1954 1958 1962 1966 1971 1984 1989 1994 1999 2004

57.2 46.3 52.3 50.5 37.7 44.5 49.3 41.0 30.3 30.2 32.3 32.7 10.4

Blanco Party 34.5 40.4 38.3 38.9 49.7 46.5 40.3 40.2 35.0 38.1 31.2 22.2 34.3

Catholics 4.3 5.2 4.4 5.0 3.7 3.0 3.0 2.4 0.2

Marxists 4.1 7.2 4.4 5.5 6.2 5.8 6.6

FA

18.3 21.3 21.2 30.6 40.3 50.7

Nuevo Espacio

9.0 5.2 4.6

Others 0 0.8 0.6 0 2.7 0.1 0.7 0.6 0.0 0.6 0.7 2.5

Source: Buquet (2005).

information on electoral behavior in the eighteen electoral districts of the country’s interior.46 Although incurring a potential ecological fallacy, this estimation strategy provides a glance into the evolution of electoral behavior by geographical units for which concurrent socio-demographic information is available. As Graph 5.9 shows, during the 1984–2004 period the FA grew significantly stronger in all three social strata and geographical units. Nonetheless, in Montevideo the FA’s electoral growth was especially concentrated in the lower social strata, where its emergent constituency was consolidated in the election of 1999.47 Importantly, the greatest surge in support from this social segment occurred not in 2004, when support from the other two segments rose, but in 1994, after the FA’s arrival to Montevideo’s municipal government in 1989. Quite notably, in 1994 the party was able to capture traditional clientelistic strongholds of the Colorado and Blanco parties in Montevideo’s periphery, as Mieres describes (1994). For instance, in zone 17 (Cerro, Casabó, Pajas Blancas, and Santa Catalina), where poverty affected 47.1 percent of households (and where 75.8 percent of residents voted for the FA in 2004), a 15 percent increase in electoral share in   46  As comparable education information was not available at the regional level, terciles were created only for estimated poverty levels.   47  Significant differences between lower strata and both middle and upper ones were obtained when comparing the FA’s electoral growth rates, taking 1984 and 1989 as a baseline, using one-way ANOVA.

Segmented Representation

228 80 70 60

Upper-Mvdeo.

50

Midlle-Mvdeo Low-Mvdeo

40

Upper-Interior Middle-Interior

30

Low-Interior

20 10 0 1984

1989

1994

1999

2004

Graph 5.9.  Electoral growth of the FA across socioeconomic and geographic segments Source: Author’s own construction on the basis of Moreira (2000a and 2005), Kaztman (1999), Mieres (1994) and data from the Corte Electoral 2004.

favor of the FA was observed between 1989 (42.3 percent) and 1994 (57.7 percent). This is further illustrated by a correlation analysis of vote shares over time in each zone, which demonstrates significant inconsistencies (lower correlations) between the vote shares of 2004 and those of 1984 (0.36) compared to those of 2004 and 1989 (0.79 sig=0.01).48 Since 1994, except for inertial electoral growth witnessed in each zone, relative electoral returns have been very stable, with highly significant and strong correlations obtained (1994 = 0.97 sig=0.01, 1999 = 0.96 sig=0.01). It is therefore clear that the most significant realignment of lower-sector support for the FA happened between 1984 and 1994. Although it began with a higher electoral support base there in 1984, the FA’s performance in the middle strata presents a growth trend similar to that observed in the lower strata, thus illustrating the party’s capacity to maintain and expand its hold on social sectors closer to its historical constituency, as well as on new ones. In 2004, in the aftermath of the harsh economic crisis of 2002, the FA grew almost evenly across all three social strata, making significant inroads in upper social sectors that were likely alienated by economic malaise. For instance, in zone 5 (Pocitos, Punta Gorda, Buceo) where only   48 Reported coefficients correspond to Spearman’s correlations of electoral returns in Montevideo’s eighteen zones. Source: raw data reported by Moreira (2005, pp. 37–41).



Strategic Harmonization of Segmented Linkages

229

3.4 percent of households were poor in 2004, and where the FA had increased its electoral share by only 6.7 percent in all previous elections (winning 33 percent in 1984, and 39.7 percent fifteen years later in 1999), support for the party grew by almost 10 percent in 2004 (to 49.4 percent). Figures from the country’s interior complement this portrait. Although electoral support for the FA was still lower in the interior than in the capital in 2004, still not even reaching the level of support it had achieved in Montevideo ten years earlier, the data do illustrate the progressive consolidation of the FA as a national party. In this regard, taking 1984 as a baseline, the party’s electoral growth rates were significantly higher in the interior than in Montevideo. In this case, in 1999 and coinciding with the first presidential and congressional elections that were held independently from municipal contests (and therefore with less influence from traditional local caudillos), the FA vote grew by an average of 16.4  percent, the single highest rise of any year. The FA’s performance in 2004 was relatively homogeneous across different social strata, with middle sectors presenting only slightly higher levels of support for the party than upper and lower strata. Although no significant differences of means were obtained, the evolution of the party’s social base in the interior appears to have been the opposite of that observed in Montevideo. Moreover, although in 1999 the FA grew more markedly in the lower strata than it had in 1984, 1989, or 1994, the absolute vote share the party obtained in this segment remained the smallest, a significant divergence from Montevideo, where the FA gained control of the municipal government. The FA’s constituency has therefore come together steadily but unevenly. Responding to different social-political processes, and to the party’s changing capacity to link to them by engaging in new types of mobilization strategies, different social segments were drawn towards the FA at different times, and the trend was more rapid with some social sectors than others. The party’s peripheral constituency was attracted to the FA later, and on the basis of a different set of linkage strategies than the party’s core constituency. The contrast between the capital and the interior also underscores the extent to which control of the local government contributed to shaping the composition of the party’s constituency in each district. In sum, the peripheral constituency was attracted to the party earlier in Montevideo than in the interior, suggesting that the FA’s control of the municipal government in the capital facilitated its electoral expansion. Although the FA has been treated thus far as a single actor, it should also be noted that individual party fractions perform differently in reaching the FA’s electoral constituency. Graph 5.10 presents the fractional distribution of the FA’s electoral vote in Montevideo in 1999. For simplicity’s sake, I present results for four major fractions of the party:  the Partido Comunista (PCU),

Segmented Representation

230 35 30 25 20

Low Mid High

15 10 5 0 PCU

PSU

AU

MPP

Graph 5.10.  Electoral performance of four FA fractions across social segments (1999) Source: Author’s own construction on the basis of Moreira (2000a), Mieres (1994), and Kaztman (1999).

the Partido Socialista (PSU), the Movimiento de Participación Popular (MPP), and the Asamblea Uruguay (AU).49 In the election of 2004, these fractions obtained nearly 70 percent of the overall vote captured by the FA. Moreover, the MPP (29.2 percent), AU (17.7 percent), and PSU (14.4 percent) were the three most-voted-for fractions within the FA at the congressional level. While the PSU and PCU represent traditional Marxist fractions, the MPP and AU are relatively new elements. The former originated from the Tupamaro urban-guerrilla movement of the late 1960s, which progressively transformed from a verticalist military organization to a democratic party apparatus after democratization. In contrast, the AU originated from a grouping of independent Frentistas and pro-renovation splinters of the PCU. As shown in Graph 5.10, the PCU and the PSU perform especially well among low and middle sectors. Additionally, in 1999 the PSU had already captured an important share of the lower sectors, partly because of the increasing popularity of Tabaré Vázquez’s (Montevideo’s first FA mayor and presidential candidate since 1994)  association with this fraction of the FA. Meanwhile the MPP, which had a cross-class constituency   49 This data was compiled on the basis of Corte Electoral reports containing electoral returns by vote-identification series. Those series were geo-referenced following Mieres (1994). A socio-structural index at the neighborhood level was created in terciles based on a one-dimensional factor-analysis solution generated from social indicators described in Kaztman (1999, pp. 329–40).



Strategic Harmonization of Segmented Linkages

231

in 1999, performed similarly within each population segment. Finally, the AU’s vote distribution was a virtual mirror image of those of the two Marxist fractions; its centrist profile attracted more support from high and middle strata. The AU is led by Danilo Astori and is heavily centered on his leadership, usually developed around a media-centered strategy. The AU consequently lacks the organizational density characterizing the other three fractions analyzed here. The FA’s fractions, therefore, exhibit varying degrees of “constituency specialization.” Such specialization has contributed to the segmentation of electoral appeals pursued by the party. As discussed in greater detail below, the ability of this internally heterogeneous coalition to remain coordinated is a crucial factor in explaining the FA’s electoral success. Taking advantage of the opportunity structure that emerged during the 1990s, the FA pursued a complex adaptation strategy that enabled it to use segmented linkages to capture a new, peripheral constituency without thereby alienating its traditional base, thus evading the trade-offs among linkages predicted in much of the literature. The FA’s Core Constituency The FA emerged in the context of an institutionalized, pluralist party system and the presence of an already autonomous labor movement. Nonetheless, the influence of the PCU within the FA led to the development of a Marxist-Leninist strategy in relation to labor, while the PSU retained its historic influence among the student movement and intellectual strata (Moreira 2000a). For these reasons, by 1990 the FA had evolved into a labor-oriented mass party with relatively weak capacity for capturing non-working-class subordinated sectors already co-opted by the clientelistic machines of the two traditional parties (Panizza 1990).50 As a result, by the time of the transition to democracy, the party’s core constituency was composed of organized labor, as well as a series of organized interest groups ranging from university students and the cultural intelligentsia. During the 1990s, the FA allied not only with labor unions but also with pension-beneficiary organizations, in efforts to derail market reforms in the country. Such early diversification aside, organized labor was clearly a key player within the party’s core constituency. The historical “brotherhood” of the FA and labor was reflected in the party’s adoption of the union confederation’s manifesto as the keystone of its

  50  This is a stylized characterization based on the predominant party–voter linkage strategies in the system. See Altman (2000) for a characterization of programmatic profiles of parties and fractions.

Segmented Representation

232 0.7 0.6 0.5 0.4

PCU-Senate

0.3

MPP-Senate

0.2

PCU-Deputies

PSU-Senate AU-Senate PSU-Deputies

0.1

MPP-Deputies AU-Deputies

0 1971

1984

1989

1994

1999

Graph 5.11.  Partisan affiliation of PIT-CNT’s directorate (2003 congress) Source: Doglio et al. 2004.

traditional (1971–1994) programmatic platform, in the continuous interaction between union and party leaders, and in the party’s reliance on labor as a transmission belt (Lanzaro 1986). The relationship between the party and the union movement is observed both in the partisan affiliation of the union movement’s leadership, as well as in the labor-union origin of the party’s congressional delegation. Graph 5.11 shows the partisan affiliation of the directorate of the labor-union peak organization (PIT-CNT) in the wake of the 2004 election. Although anarchists and “others” comprised a significant fraction of the directorate, communists, former-communists, and socialists made up the majority. Moreover, already in 2003 the MPP had also made inroads into the union movement, contesting the traditional position of the PSU and the PCU. Turning to the party’s congressional delegation, and following the analysis of Doglio et al. (2004), Graph 5.12 displays the proportion of each fraction’s senate and deputy electoral list that included union leaders in the top three senator and top six deputy slots. As shown in Graph 5.12, the (PSU) and (PCU) electoral lists have traditionally been headed by a significant proportion of union leaders. At least a third of AU’s senates lists are also composed of union leaders.51 Although the historical “brotherhood” of the party and the labor movement remains strong, the party had to deal with two transformations of unionized labor. During the 1990s the strength of organized labor declined steadily. Moreover, this decline was especially concentrated in the private sector, upsetting the balance between private and public sector unions within the labor confederation.  

51 

The MPP, in turn, has privileged historical MLN-T figures.



Strategic Harmonization of Segmented Linkages

233

PCU Former PCU MPP/MLN-T PSU Anarchists/PVP Other

Graph 5.12.  Proportion of union leaders in top positions on respective fractions’ senate and deputy electoral lists Source: Doglio et al. 2004.

The evolution of unionized labor reflected in Graph 5.13 challenged the FA in three ways. First, the decline of organized labor due to deindustrialization and the economic crisis of the early 2000s threatened to reduce the party’s core constituency, especially by weakening the role of labor unions as transmission belts for the FA. Second, the growing imbalance between more radical public sector unions (usually opposing any attempt at state reform) and those in the private sector, had the potential to create governance problems. Were the party to reach office, it would be forced to implement reforms that would likely have involved redistributing resources from the “remnants” of organized labor to other social segments (especially the informal sector and non-organized working class). Interestingly, the potential distributive struggle within the FA’s electoral constituency emerged because the party grew stronger precisely when stagnation began afflicting the economic model into which a significant share of its core constituency was politically and socially incorporated. As an FA congressmember observed: When I returned from exile in 1984, we still had powerful industries in the country, as well as very well articulated unions. Today we see them disappearing, the leather, the metallurgical, the textile, are all dead. And the most lively union is the banking and financial system one. It is obvious that these trends reflect the economic model that has been applied in Uruguay. The PIT-CNT [the central labor confederation] still breathes and has been maintained due to the power of the banking union and those of

Segmented Representation

234 250

Total Public Private

200

150

100

50

0 1987

1990

1993

1996

2001

2003

Graph 5.13.  Total union membership and public vs. private origin of union members (1987–2003) Source: Supervielle and Quiñones (2003).

the public administration [state employees]. But those unions are not typical unions; they are not working class unions. The working class here is destructured, atomized. And this also produces a decrease in political militancy. (Ernesto Agazzi, FA congressmember, personal interview, 2003).

The third challenge arose from the labor movement’s increasing autonomy from the FA fractions that had previously dominated the union confederation (the PCU and the PSU). Today, as shown above, several fractions are represented and compete within the labor movement. Especially since the PCU’s crisis in the aftermath of the fall of the Soviet Union, different fractions of the FA have competed to control a labor base that has gradually gained autonomy and become more radicalized even as the party has begun to moderate its programmatic stance and to assume governmental positions. For instance, just one year before the FA’s presidential victory, a long-lasting conflict between the municipal-workers’ union and the municipal government of Montevideo (headed by the FA) led to harsh confrontations between the labor movement and the party. Already ten years before that conflict, the leadership of the FA recognized that the historical relationship between organized labor and the party was being reshaped. As put by Reinaldo Gargano, a PSU leader: The Frente Amplio has had a tradition of relating to organized labor. Yet, fortunately, the labor movement is no longer the transmission belt of a given political party. They



Strategic Harmonization of Segmented Linkages

235

are more independent [ . . . ] Our party has to think not only about the interests of organized labor, but also about those of Uruguayan society.52

Despite unions’ greater autonomy from the party, and the potential conflicts that such autonomy and the differences of interest that the FA’s moderation and actions while in government might have produced, the “friendship” between the party and the labor unions persisted. One possible explanation for this outcome was the absence of viable partisan alternatives to the left of the FA. The very strength of Frenteamplismo contributed to dissuading potential splinters. Organized labor has historically been a fundamental part of the FA’s core constituency, and it continues to be so today. As in the case of the UDI, the core constituency of the party has two consequential characteristics. First, it is not sizable enough to constitute a winning electoral base on its own. Indeed, given the decline of organized labor in Uruguay, the core constituency of the FA was logically in danger of shrinking. Expanding the party’s electoral constituency to new social segments was therefore imperative. Second, as in the case of the UDI, the core constituency of the FA was able to provide resources that became fundamental in structuring new types of linkages to a peripheral (but much needed) electoral constituency. Yet, unlike in Chile, the core constituency of the FA did not provide enough financial resources to disproportionately favor the electoral bid of the party. Indeed, as shown in Graphs 5.14 and 5.15, the FA’s estimated campaign spending, although growing over time, was not larger than that of its most serious contender in the last electoral rounds (the Blanco Party). Graph 5.14 displays the estimated campaign expenditures for the elections of 1994 and 1999–2000, as estimated by Zamora (2005). In turn, Graph 5.15 displays self-reported expenditures on TV campaign advertising, as well as an independent estimate based on the IBOPE Media Uruguay’s counting of advertising seconds for each party. Instead of financial resources, the core constituency of the FA provides the party with a militant activist network significantly outnumbering those of traditional parties. The FA can therefore count on a decentralized network of supporters who are frequently willing volunteer for the party. Such a network played a vital role in enabling the FA to expand its electoral constituency to the poorest segments of Uruguayan society. Graph 5.16 shows the percentage of party sympathizers who also declared to have worked for a political party during the past electoral campaign. Close to 25 percent of the FA’s adherents

 

52 

Own translation of interview passage cited in Wettstein (1993, p. 44).

Segmented Representation

236 14

1994 1999–2000

12 10 8 6 4 2 0 PC

PN

FA

Graph 5.14.  Parties’ campaign expenditures: 1994 and 1999–2000 electoral cycles (estimated amounts expressed in 1995 US millions of dollars) Source: Estimated by Zamora (2005, Table 5.1) and reported in Chasquetti (2011).

1800000 1600000 1400000 1200000 Self-reported 1000000

Ibope Media Uruguay Expenditure Estimate based on TV ads seconds

800000 600000 400000 200000 0 PC

PN

FA

Graph 5.15.  Parties’ TV advertising expenditures: 2004 electoral Cycle (amounts reported in US dollars) Source: Chasquetti (2011), based on data reported in Rumeau (2005).



Strategic Harmonization of Segmented Linkages

237

30

25

20

FA

15

PC PN

10

5

0

% Worked

% Worked over total respondents

Graph 5.16.  Percentage of militant activists among political party sympathizers Source: Author’s own construction based on LAPOP 2006.

attested to having worked in the electoral campaign of 2004. Although the traditional parties’ proportion of activist sympathizers is not drastically lower, because they also had many fewer sympathizers the resulting estimated size of the FA’s activist network was more than triple that of its closest competitor (the Blanco Party). Although it has faced declining rates of organic participation in grassroots committees (Comités de Base), which form the territorial structure of the FA, the party still has a much more developed and efficient decentralized network than others. As a congressmember I interviewed claimed: We have hundreds of committees and several zone-wide coordinators. Therefore, you can easily connect to the people in a direct and personal way. We are in contact with people in every neighborhood, even the poorest ones. I am talking about problematic neighborhoods, ghettos. And we go there and talk to the people. And the bond is ideological, because we talk about general things. Of course, you have clientelism too, but that’s exceptional. The structure we have is highly profitable from an electoral point of view. So, it is not only about ideas; it is also about political efficiency. On the left, we do not have paid political activists. Many of my colleagues from the traditional parties have told me: “You do not know how much I envy your free labor.” That’s the importance of activism. But what happens is that, to have activists, you need to have political leaders who communicate ideals, visions. They need to know that people close to them, like them, are able to engage in the political arena and enjoy electoral victories or defeats. However, traditional leaders have concentrated things so much, and I am talking about Sanguinetti, Batlle, Lacalle, that they have discouraged their activists. They have extraordinarily well-prepared people, but they don’t let anyone speak out

238

Segmented Representation

because they are fearful of losing their positions. (Carlos Pita, FA congressmember, personal interview, 2003)

As Pita stresses here, symbolic and ideological factors have been very important for attracting and retaining FA activists. Their frequent engagement in opposition activities (like massive signature campaigns to push for plebiscites or referenda) have provided the party with a “cheap” way of maintaining mobilization, despite its lack of access to state resources (except in the municipality of Montevideo). In the capital, decentralization and pursuit of focused local social policy initiatives (installing health centers and soup kitchens, financing cultural activities, etc.) have generated significant synergy between the FA’s municipal governance and its electoral advance.53 Coupled with the relative collapse of traditional party structures, the FA’s superior activism is one of the general factors explaining its electoral growth, particularly in sectors of the electorate that previously related to parties on the basis of clientelistic side-payments. This crucial advantage is frequently acknowledged and admired by traditional party leaders. In the words of one Colorado congressmember: The sectors that we used to relate to in this way [the “clientelistic” vote] are now voting for the FA. Today we have lost our base organizations. We only have a traditional but intermittent organization that is quickly mobilized during elections [ . . . ] FA has a great activist base that is much cheaper to maintain. (Pais 2003)

Although similar figures for the Chilean case cannot be drawn (due to the fact that only about 3 percent of total respondents are activists), Graph 5.17 displays a proxy for the size of each party’s activist network by comparing the number of citizens officially registered as party supporters in the electoral registry. Surprisingly enough, although the UDI is the most-voted-for party in Chile, it actually has the lowest number of citizens enrolled as active members in the official party roster (Servicio Electoral, data available for 2011). In short, whereas in the case of the UDI the core constituency has access to financial resources and contributes them to the party, in the case of the FA its

  53 Decentralization in Montevideo basically involved a process of bureaucratic deconcentration, creating a decentralized network of communal centers, which, however, lack decision-making capacity. The communal centers are run by a secretary directly appointed by the mayor and zone-level councils, which work as advisory bodies and are directly elected by the people. Especially in poor zones of the capital, local council members are the most active political brokers in the area (particularly in the traditional parties), usually seeking to push solutions to their constituents’ needs through a very complicated bureaucratic structure. (Based on a series of interviews with local council-members, 2003. For a detailed analysis of decentralization in Montevideo, see Veneziano 2002.)



Strategic Harmonization of Segmented Linkages

239

120000

Number of Party Affiliates

100000 80000 60000 40000 20000 0 UDI

RN

PPD

PS

DC

Graph 5.17.  Number of party activists officially registered by the Servicio Electoral circa 2011 (Chile) Source: Author’s own construction on the basis of report by the Servicio Electoral.

core constituency is particularly important for providing volunteer activism. Each of these different core-constituency resources proved key for enabling successful outreach to a peripheral constituency through segmented linkages. The Political Opportunity Structure of the 1990s and 2000s This campaign [2004] is one of the easiest ones that we have had [ . . . ]. You just need to step on the corner and ask:  “Who is responsible for this mess?” Then you start distributing printed ballots in the street. [ . . . ] Today, there is the hope of the poor, of business people, and of the rural sector. (Eleuterio Fernández Huidobro, interviewed in Página 12, September 13, 2004)

In exchange for its core constituency’s unrelenting support, the FA led the opposition to market reforms proposed by both traditional parties. These efforts included attempts at privatizing public enterprises, and subsequent attempts at pension, education, and health reform. In a representative democracy escaping the patterns of hyper-presidentialism and delegation prevalent in the region in the 1990s (O’Donnell 1994), organizations of ISI beneficiaries allied with the FA to oppose attempts at market reform. In confronting reform attempts led by subsequent governing coalitions of the Blanco and Colorado fractions, this ISI coalition relied on congressional representation and direct democracy mechanisms to block reforms (Filgueira and Papadopulos 1996;

240

Segmented Representation

Castiglioni 2005). Indeed, the use of direct democracy mechanisms became a crucial part of the FA’s strategy for opposing reforms (Altman 2002). The presence of a veto coalition identified with the FA and opposed to reformist attempts synergized with the ideological hegemony of pro-statist Batllismo (even within fractions of the Colorado Party led by former President Sanguinetti), and became the focus for party competition in the 1990s. Yet, the presence of still strong traditional partisan subcultures and the gradual nature of the reforms (and economic decay) protected the country from electoral dealignment, and allowed for a programmatic realignment in the 1990s. The FA became the reinterpreter of Batllismo, while traditional parties (especially some fractions) promoted state reforms and economic liberalization. Accordingly, in 2004 the Uruguayan electorate was still the most “statist” in Latin America, with its citizens rating the state’s responsibility for people’s welfare at 6.02 (against a regional average of 5.2) on a ten-point scale. Likewise, Uruguayans were the least convinced of the market economy’s ability to promote development, with only 50 percent expressing confidence in the market, compared to the Latin American average of 63.1 percent. Moreover, whereas 60 percent of Latin Americans agreed that private enterprise was needed for economic development, only 49 percent of Uruguayans did (Latinobarómetro 2004). Since the 1990s, the FA has remained the party that has most closely represented the statist mood of Uruguayan public opinion. This fortuitous congruence arose from a favorable opportunity structure, to which the FA responded strategically through a segmented linkage strategy. As early as the elections of 1971, a strong correlation between age and party identification was already apparent. Younger voters consistently favored the FA, while older ones aligned with traditional parties. This correlation gave rise to a strictly sociological “demographic explanation” for the FA’s growth,54 which also became the most commonly accepted in Uruguayan political science (Buquet and De Armas 2004; see also Aguiar 2000; González and Queirolo 2000). According to this hypothesis, those strongly identifying with traditional parties would over time naturally exit the electorate, automatically decreasing their numbers among a renewed citizenry progressively more dominated by FA identifiers. Although the empirical evidence to some degree supports the existence of such a structural trend—explaining roughly 52  percent of the FA’s electoral growth between 1984 and 1999—the data also confirm the need for additional explanations (Buquet and De Armas 2004, p.  126). Moreover, the party’s

  54  César Aguiar proposed the so-called “demographic hypothesis” in 1984. See Buquet and De Armas (2004) for a review and a critical assessment of this explanation.



Strategic Harmonization of Segmented Linkages

241

electoral growth was not smoothly distributed over time and among socioeconomic segments of society, as a purely demographic explanation would predict. A more nuanced understanding of the complex causal process underlying the electoral result of 2004 is therefore needed. As stressed by Roberts (2002), the Uruguayan case presents an unusual combination. It has been an elite party system, but with relatively high levels of ISI peaking in the 1950s and 1960s and an expanding welfare state. This combination is crucial to understanding the emergence of a labor-mobilizing party that transformed the traditional party system in a way quite different from the transformations in other institutionalized “elite, two-party systems” (i.e., Venezuela, Colombia, and Costa Rica). As the crisis of the development model took hold and reform advanced in Uruguay, a system previously dominated by non-programmatic linkages became more programmatically oriented. In this context, the FA expanded its electoral support by defending the organizations of ISI beneficiaries, which then consolidated as its core constituency. At the same time, the FA began to provide an alternative leadership for those who felt alienated by economic malaise as well as those who were increasingly marginalized by the clientelistic machines of the traditional parties. This increasing marginalization occurred at the very time when a sizable fraction of the Uruguayan population was feeling the effects of a steady deindustrialization of the economy, of a sustained economic crisis in the agricultural sector, and of a rapid increase of unemployment (e.g., from 8.5 percent in 1990 to 18.3 percent in 2003) (INE 2005). Although significant reductions in poverty and inequality were achieved from 1990 to 1994, increasing social expenditure was insufficient for offsetting the social effects of trade liberalization and currency overvaluation. Indeed, despite these measures and the sustained economic growth of 1994–1997, by the start of the 2000s poverty and inequality had actually worsened.55 While social sectors with greater organizational capacity and the political leverage to veto reforms were better able to protect their interests, more vulnerable sectors of the population that had never been direct beneficiaries of the state-centered ISI model, or had been expelled from it during the crisis, became increasingly unable to bear the costs of adjustment. This disparity was evident in the uneven evolution of real wages and unemployment rates in the public sector compared to those in the private sector. A further illustration of this unequal capacity to ride out the storm was the real value of

  55  While the percentage of people under the poverty line was 25.5 percent in 1991, it reached 32.5 percent in 2002. Meanwhile, the Gini coefficient went from 41.05 in 1991 to 43.16 in 2000 (De León 2004).

242

Segmented Representation

pensions versus family allowances targeted to poor families with children. By the early 2000s, poverty rates for Uruguayan children were extremely high. Forty-two percent of children aged 0–5 lived in poverty, compared to 8 percent of Uruguayans aged 65 and older. Nonetheless, between 1991 and 1999 the real value of family allowances decreased by 47 percent, while pensions’ real value grew by 60 percent (Filgueira and Kaztman 2001). Against this socio-structural backdrop, both the original and emerging constituencies of the FA were consolidated and grew during the 1990s, helping to channel popular discontent while at the same time enhancing the level of interest aggregation in society and the significance of programmatic linkages for party competition. These trends resulted from the FA’s pursuit of a complex partisan adaptation process in a context where traditional party–voter linkages were in crisis. Reaching the FA’s Peripheral Constituency The FA’s electoral strategy enabled the party to adapt to these new circumstances and successfully expand beyond its core constituency. Several features of the party’s electoral strategy were important in this regard: ideological moderation, leadership renewal, and the pursuit of electoral alliances. Electoral strategizing at the fractional level was also fundamental to widening the electoral base of the party, as illustrated by the recent fortunes of its AU and MPP fractions. In different segments of the electorate, those fractions both contributed to the final push towards the election of 2004, attracting electoral support from social groups that had not traditionally voted for the left. In 2004, the MPP and AU obtained the two largest pluralities within the FA vote (29.2 percent and 17.7 percent respectively) by attracting non-traditional constituents such as the upper-middle classes in Montevideo and the interior (AU) and rural voters and supporters of traditional caudillos (MPP), as well as voters in the informal sector and in shantytowns. Moderation, Leadership Renewal, and Electoral Alliances56 Beginning in the early 1990s, the FA experienced a complex process of ideological and programmatic renovation, gradually replacing its original   56  This section draws on the careful historical reconstruction of Yaffé (2005). See also the works in Lanzaro (2004).



Strategic Harmonization of Segmented Linkages

243

“anti-imperialist” and “anti-oligarchic” platforms with a progressive programmatic stance centered on opposition to neoliberal reforms, and a reinterpretation and appropriation of Batllismo (Garcé and Yaffé 2004; Yaffé 2005). During this period the FA achieved two seemingly contradictory objectives, which came as unintended consequences of leadership disputes. First, it consolidated its image as the opposition against the traditional party family. Second, it moderated its program and expanded its electoral menu through fractionalization, becoming increasingly catch-all. Additionally, in spite of its increasing electoral orientation, the party managed to maintain a comparatively larger and more vibrant activist apparatus in the streets than those of the Blancos and Colorados. The FA accomplished this by mobilizing activists between elections in attempts to block pro-market legislation enacted or proposed by the governing coalition using direct democracy mechanisms. The FA’s renewal coincided with the fall of the Berlin Wall and the party’s election to the municipal government of Montevideo in 1990. Facing significant internal dissent led by the FA’s historical leader (Liber Seregni) and other conspicuous party figures (especially Danilo Astori), Tabaré Vázquez was elected mayor of the city. Prior to his term as mayor, Vázquez was a relatively unknown socialist activist. His vision for programmatic renovation included pursuing political alliances with centrist groups outside the party. Thus, between the national congressional elections of 1991 and 1994, the FA’s internal politics were characterized by a harsh confrontation between moderates (led by Vázquez with the support of Socialistas, Demócrata Cristianos, the Vertiente Artiguista, the pro-renovation fraction of the PCU, and PNU leaders from the Polo Progresista fraction) and radicals (of which Astori became the most vocal representative). Vázquez’s objective of crafting a macro-center-left coalition for the 1994 election crystallized that year with the proposal to create the Encuentro Progresista (EP), an alternative programmatic manifesto. Ultimately, the PCU supported the initiative in the 1994 Congress, breaking the internal deadlock that had existed since 1991. This enabled the FA moderates to seek alliances beyond the party and to craft the EP, eliminating the most radical proposals that they previously had failed to remove from the FA’s traditional platform. Following his electoral defeat in the 1994 presidential election, Vázquez had to face additional leadership challenges to consolidate his own position and push the process of moderation forward. The confrontation with Seregni and Astori over the constitutional reform of 1996 was crucial in this respect. While Vázquez opposed the reform, Seregni (still president of the FA) had personally negotiated the deal with Blanco and Colorado leaders and therefore supported the measure. Seregni, joined by Astori, claimed that beyond short-term hindrances to the electoral chances of the FA in 1999, the partial elimination of

244

Segmented Representation

the double simultaneous vote (DSV) and the introduction of one-time presidential candidacies were in line with the FA’s historical positions. Seregni’s resignation as president of the FA during the party’s twenty-fifth anniversary celebration in February 1996, and Vázquez’s subsequent ascent to that post, further consolidated the latter’s power within the party. Although the reform was popularly ratified in a plebiscite, Vázquez’s stance against it was supported by 49 percent of the electorate (almost 20 percent above the FA’s vote share in the 1994 election). Additionally, in the context of the constitutional reform debate, Vázquez’s refusal to adhere to an arrangement with the traditional parties consolidated his image as a harsh opposition leader. Internally, this granted Vázquez the support of radical groups that since 1990 had been aligned with Seregni and Astori, producing a crucial shift in the internal and external positioning of the FA leadership, and giving Astori a more moderate image for his willingness to enter into pacts with the traditional parties (Yaffé 2005). Thereafter, Vázquez received further compensation for his consistent opposition to the traditional parties. These included the integration of the MPP into his internal coalition in 1999 and the results he obtained in the presidential primary, where he defeated Astori by an uncontestable margin (82  percent versus 18  percent). Later that year, Astori’s fraction (AU) suffered further defeat at the polls, obtaining almost 50 percent fewer votes than the largest internal plurality obtained in the 1994 election (39 percent versus 20 percent). In spite of their personal confrontation, which after 1994 centered on whether it would be best to organize the opposition frontally or pragmatically, from 1995 onward Vázquez and Astori began to concur on the need to moderate the party’s program. The FA’s moderate stance was also tested during the financial collapse of 2002, and the party proved its commitment to democratic institutions by supporting the government at a time when it could have forced its removal by promoting street riots like those seen in Argentina.57 Also in 2002, the moderation strategy received another crucial push with the creation of the Nueva Mayoría (NM) electoral alliance, which enabled the return of the Nuevo Espacio fraction and the incorporation of other groups splitting from both traditional parties (Yaffé 2005). Although confrontations between Vázquez and Astori continued during Jorge Batlle’s presidential term (2000– 2005), they became progressively more centered on the consolidation of the   57  In January 2002 Vázquez stated that “the left does not like and does not promote social unrest” (cited in Paolillo 2004, p. 61). Later that year, in May 2002, Vázquez stated: “. . . if someone thinks that we are going to shut up, they are mistaken [ . . . ] if someone thinks we are going to strike, they are also mistaken” (cited in Paolillo 2004, p. 229). Own translation from the original.



Strategic Harmonization of Segmented Linkages

245

FA’s 2004 coalition.58 Whereas Vázquez’s harsh opposition tactic proved useful in retaining the support of the FA’s historical constituency, Astori’s pragmatism helped to reassure voters who were alienated from traditional parties, but still distrusting of the FA’s potential radicalism. In short, during the 1990s Astori switched his position to become more moderate. Meanwhile, Vázquez remained an ideologically moderate, pragmatic leader. However, the latter’s uncompromising (but accountable) opposition to the Colorado and Blanco governments won him the support of radical groups within the FA and alienated voters from the traditional parties. Vazquez’s anticipated nomination of Astori as his finance minister during a trip to Washington DC in the midst of the 2004 electoral campaign (following Astori’s decision not to contest the FA primary) symbolizes the ultimate convergence of both leaderships. That convergence enabled the party to maximize the electoral appeal of the FA across a broad electoral cross-section. The Municipal Government of Montevideo The FA’s fifteen years in municipal government in Montevideo was widely vindicated at the polls in 1995, 1999, and 2005, and, according to public opinion polls, enjoyed significantly higher approval levels than parallel administrations in the interior and at the national level led by Luis Lacalle, Julio Sanguinetti, and Jorge Batlle over the same period (Buquet and De Armas 2004). This governmental experience was instrumental in fostering the FA’s electoral growth in several regards. First, it contributed to the emergence and consolidation of the charismatic leadership of Tabaré Vázquez and, from 1995, that of Mariano Arana. Both mayors became highly popular figures, achieving visibility and popularity that went beyond Montevideo and was especially concentrated in the party’s emergent constituent segments. Second, it contributed to moderating the party, proving the FA’s capacity to rule effectively while at the same time providing its leaders with much needed government experience. In this regard, despite confronting well-documented budget discrimination by the national government, this new municipal role was developed in cooperation with civil society organizations and the private sector. Moreover, the implementation of a more progressive tax scheme,   58  Those confrontations reached their peak in 2003 during discussions of a law that sought to merge the state-owned oil refinery with a private partner. The law was struck down through a referendum promoted by the company’s union and sponsored by Vázquez, who once again defeated Astori and traditional party leaders with his popular rejection of the law.

246

Segmented Representation

a decrease in tax evasion, and a process of bureaucratic streamlining of the municipal apparatus also provided much needed financial resources (Schelotto 2004). Taking advantage of a national trend (Laurnaga and Guerrini 1994), the FA’s municipal administration went beyond traditional municipal roles to engage more actively in the promotion of social welfare (especially by providing decentralized social policy programming), economic development, and cultural policies (Schelotto 2004; Veneziano 2001). Third, although far less ambitious than the participatory budgets implemented in Brazil under the PT, the administrative deconcentration process that created eighteen decentralized municipal offices (Centros Comunales Zonales, or CCZs), along with the popular election of zonal councils, enabled the party to develop a “close to the people” administration. According to a CCZ director, in a marginalized neighborhood of Montevideo the Centro acts like a “big ear of the state, where before there was nothing.” And, “although often we cannot help [because solutions cannot be found at the municipal level], people leave with the feeling that someone listened to them” (García 2002). Most likely as a result of decentralization and the territorial penetration of the periphery, the FA was able to set up a powerful political apparatus that was instrumental in “getting close to the people, accompanying it in its anguish, and trying to help whenever possible [with decentralized social programs: i.e., health care and soup kitchens]” (Percovich 2002). The FA was thus able to strengthen its territorial hold by complementing its ongoing mobilization of partisan activists while at the same time organizing and helping former traditional party constituents. As two FA deputies state: [. . .] you always need to participate in meetings in which people are asking you to solve their particular problems, that’s very ingrained [  .  .  .  ] And when I  can, if it is reasonable, I  try to help, because many times you have unnecessarily unfulfilled needs, either due to bureaucratic problems or people’s lack of knowledge of formal procedures [ . . . ] Our political party has a neighborhood organization with which you regularly interrelate, either because they invite you or because you stop by during the monthly visits. And although they call you to address a specific problem, you always need to steal some minutes to talk to them about the common good, about [ . . . ] how we as a political party are interpreting national events and seeking improvements. (Rossi 2003) I think in Montevideo, decentralization was crucial. [  .  .  .  ] For the first time, particularly the people in the periphery felt that they had a voice, that their proposals could be heard. And that paid-off in electoral times. We don’t know how much that effect will last because now you have important degrees of disenchantment because we need to improve their autonomy and capacity of doing things. Still, we have an ethical credit. People see that we have been there and even those more resistant to change and those more critical of us are beginning to say: “These have not stolen yet [i.e., have



Strategic Harmonization of Segmented Linkages

247

not been in government]. Let’s try with these guys.” [ . . . ] even beyond administrative management, decentralization also provided the basis for organizing women groups, cultural activities, youth groups, which have a tremendous social richness. (Margarita Percovich, FA congressmember, personal interview, 2002)

In sum, the FA’s experience in Montevideo has been pivotal for both fostering partisan adaptation and increasing the party’s electoral dominance in the historical constituencies of both traditional parties. The comparison presented in Graph 5.9 between the social composition of the FA’s vote in Montevideo and that of the interior (where “municipal governments [continued to be] infernal vote-generating machines” dominated by traditional party mayors59 ) illustrates the importance of controlling the capital’s municipal government for expanding the electorate in lower segments of society. Reaching the FA’s Emerging Constituency: the AU and the MPP The FA’s success can also be credited to fractional tactics and their complementarities, which enabled them to reach different electoral cross-sections. Here I describe the organization and tactics of the most recent fractions that were effective in reaching the FA’s emerging constituency. Given its greater complexity and recent popularity, more space is devoted to explaining the strategy and tactics of the MPP. The AU’s tactics are structured around the figure of its leader, Danilo Astori, whose moderate positions have attracted previous supporters of both traditional parties in the country’s high and middle strata. In effect, Astori has reduced the “costs of entry” to the FA for previous traditional party supporters, who have been attracted by the classic “third-way” platform frequently promoted by the AU and its main leader in the national media. The AU has also developed a network of local leaders who operate much like local representatives of the traditional parties, coordinating problem-solving networks for particularistic needs: We have learned a lot from the traditional parties and we are currently doing things that we used to underrate. We now have a structure of local leaders, each one of them a caudillo in his place, working with a team [ . . . ] trying to help people to solve their most immediate needs. (Mahía 2003)

According to a local AU caudillo who ran as congressional candidate representing his fraction, and was a local council member when interviewed,  

59 

The expression was used by Hubaré Aliano, a Colorado caudillo from the interior (2003).

Segmented Representation

248

they “try to do everything” from helping someone to get a driver’s license, to pressuring the Food Institute (INDA) to establish a soup-kitchen, to helping organize a housing cooperative. After “you do something, you have to show up and tell them you did it.” Finally, when the fact of being an opposition party hinders the articulation of a solution (e.g., when they “request a job or just money”), “[we] tell them: ‘Look, we know we cannot get you this. This can be resolved by these guys [Colorado local officials] who are the ones in control. So, go with them and then you just vote for us.’ We are clear about that.”60 These structures have also allowed the AU to gather support from lower-class voters, particularly in the country’s interior, where the starkly leftist character of other fractions still alarms some citizens. The MPP recently emerged as a fashionable fraction and a gateway for all social strata to the FA. The characteristics of the MPP’s new electoral supporters, however, contrast sharply with those of MPP founders and current hardliner activists who bear the heritage of the MLN-T. Indeed, the self-critique of MPP’s leaders revolves around the difficulty of abandoning practices that are efficient in gathering electoral support but that contradict a bottom-up approach to representation and that might backfire in the future. The MPP’s main features can be grouped around three central notions (personal interviews with Nora Castro and Ernesto Agazzi, 2002). First, the MPP is organized around a bottom-up concept inspired by the MLN tradition, in which popular participation is the building block of political action. Accordingly, “winning elections is not winning power, because power is constructed from below” (Agazzi 2002). Second, the MPP is defined as a “movement,” without an organic structure. Different organizational manifestations ranging from NGOs and research institutes to rural unionization experiences, from the MLN-T to the Blanco and Colorado columns, converge and interact in this heterogeneous fraction. Finally, in tactical terms the MPP is extremely pragmatic and open, pursuing a cross-class electoral strategy as opposed to a classical Leninist approach. In the words of José Mujica, the most prominent figure of the MPP: “If it were necessary to grasp power, I would hug a snake.”61 Without a doubt, the fundamental reason for the MPP’s success is the charismatic leadership of José Mujica, who in 2009 was elected president of the country to succeed Tabaré Vázquez. Indeed, according to survey data 70 percent of MPP supporters feel essentially tied to his leadership, and only 9 percent of them would continue to vote for this fraction if Mujica left (RADAR

   

60 

Personal Interview with Artigas Reina, AU’s local activist in Artigas (2002).

61 .



Strategic Harmonization of Segmented Linkages

249

2004). The “Mujica phenomenon” is based on a series of personality and political traits that other leaders of the MPP also display. First, “he speaks clearly, like common people, in basic Spanish” and “he is able to interpret people’s feelings.”62 Second, he is “outstanding in terms of his austerity, his social and political commitment, and his moral convictions,” “has risked his life [as a member of the guerrilla movement and as a ransom political prisoner of the military] to pursue his ideals,” and in spite of being a senator, “he lives like the poor.”63 The political direction of the MPP has mixed emotions on the “Mujica phenomenon.” On the one hand, his leadership has brought people to the movement who would not have arrived otherwise. On the other hand, popular adherence to Mujica’s charismatic leadership contradicts the doctrine of the MPP. Therefore, whereas some MPP leaders consider Mujica’s leadership “a necessary evil” and a “double-edged sword,” others see the arrival of new voters as an opportunity to start developing a more consistent political organization by trying to modify “political cultures” inherited from a clientelistic past.64 In this context, the arrival of unemployed workers with a history of union militancy in Montevideo’s expanding shantytowns is also seen as an opportunity to contribute to organizing the lower classes. Indeed, according to data from INTEC reported in Álvarez (2012), the occupational background of shantytown dwellers in Montevideo had clearly changed over time. As shown in Graph 5.18, the proportion of unemployed shantytown dwellers felt drastically between 1984 and 1994. The arrival of more socially “integrated” inhabitants into the shantytowns, some carrying with them a tradition of communal organization or union-participation, is frequently referred to by FA activists both as a consequence of the economic deterioration of the country and as a facilitating condition for the party’s penetration of its peripheral constituency. Mujica’s leadership is also sustained by a continuous presence in the national media and an extensive network of local radio stations. This network has been essential in reaching the countryside, where a significant percentage of the emerging MPP constituency resides. Yet, these traits are complemented by the significant presence of local activists, which is enabled by the MPP’s powerful territorial structure.65 Today, the territorial structure of the MPP and the number of activists continuously outnumber those of other FA fractions, in

       

62 

Personal Interview with a group of MPP supporters in Bella Unión 2002. Personal Interview with a group of MPP supporters in Bella Unión 2002. 64  Personal Interviews with Congressmembers and leaders of the MPP. 65  Participant observation in Uruguayan districts during 2002–2003. 63 

Segmented Representation

250 80 70

1984

60

1994

50 40 50 20 10 0 Does not work

Private sector worker

Public sector worker

Self-employed

Other and no data

Graph 5.18.  Occupational background of shantytown dwellers Source: INTEC reported in Álvarez (2012).

virtually every community or social organization.66 Additionally, Mujica and other leaders continuously tour the country, holding mateadas (public gatherings to share a drink and talk about current issues). Attendance at mateadas ranges from 5 to 10 people in small villages, to 2,000 or more in the public squares of Montevideo. Programmatically, the MPP has focused on the idea of creating a “productive country” articulated around agricultural exports and industrialization. Symbolically, the MPP launched its electoral campaign in an empty building that was formerly home to a large industrial facility in a prototypical working-class neighborhood of Montevideo. The following extracts cite from a former Blanco local leader that recently adhered to the MPP illustrates how the programmatic stance of the MPP, the charismatic leadership of Mujica, and this fraction’s particular approach to politics come together in the everyday fieldwork activities of the MPP. The interview material also confirms the role of the economic crisis in catalyzing the fraction’s mobilization efforts: Here, you always had political parties throwing food at people, throwing everything to get people to vote. But now the country is bankrupt. Once every five years they [the   66  Personal Interviews with congressmembers and activists of the PSU, the PCU, the Vertiente Artiguista, and AU (2002).



Strategic Harmonization of Segmented Linkages

251

traditional parties] came to the interior and hug everybody. Everyone got drunk, big speeches, big promises that they never delivered. Frente Amplio does not do that. We go with honesty. When we go out with Mujica, in our tours, we sleep in humble houses, those of our compañeros, we eat humble foods. But we are not demagogues, we do that because that is the reality of our own lives and we think that in a bankrupt country, we as politicians are the ones that have to exemplify with our own lives how a government headed by FA will be. My presentation card is the following. Today I walk with fifty pesos in my pocket [one dollar and fifty cents at the time of the interview], I drive a 1992 car, I help everyone, I give free medical attention to the poor, and so on. Jorge Coronel [another Espacio 609 member] defends the poor pro-bono in Cerro Largo, many times having to pay himself. He also drives a 1992 or 1993 car, and we cannot update those cars. And we have no vices, we are honest, we do not get into political fights, we do not ask people who they are. That is the first thing Colorados and Blancos do, they ask people who they are. With Mujica and with Jorge Saravia [the leader of the Blanco columna in the Espacio 609, who in 2011 left the MPP] we sit on a stone, below a tree, and talk politics with people. We talk about the chance of changing this country, of making meat, wool, and leather the number one products in the country. Without having that, you cannot resurrect our industries. We had everything here, textiles, shoes, sugar, but they [the traditional parties] killed all that. (Aldorio Silveira, Espacio 609 MPP leader in Bella Unión, personal interview, 2003)

Another key element in explaining the MPP’s electoral growth is the pursuit of broad alliances with other sectors of the left and, even more importantly, with splinters from both traditional parties. In this context, Espacio 609 was created to accommodate a Blanco and a Colorado “column.”67 Those “columns” were not directly integrated into the MPP, but work within the movement as autonomous organizations. This political formation seeks to provide a “home” for those who decided to join the movement because they were alienated by the traditional parties and attracted by Mujica’s leadership. The columns provide a place in which their traditions and tactics are respected and maintained. According to the words of a Blanco leader who had joined the MPP by 2003, their switch was not merely tactical or pragmatic (i.e., caused by the declining resources available to local activists in the Blanco and Colorado brokerage networks), but inspired in their appreciation of the distinct qualities of the fraction and its leadership: I am a guy from the interior, and I think Mujica’s way of doing politics is very good. He demands a lot of the people who works with him, and many other guys are coming

  67 A columna is similar to the agrupaciones of traditional parties, and can be seen as a group within a fraction that presents independent lists to the lower house and for local elections. In the case of the MPP, the congressional list was unified, but the columns still represented independent political groups within the Espacio 609.

252

Segmented Representation

to the MPP [ . . . ] We have to get together [ . . . ] I think the FA is the only hope for us. We have demanded something of the FA, that they change their ways a bit; they need to be less extremist and change a bit without becoming dishonest [ . . . ] We need a new country in which people trust politicians, without corruption, without political vice, without patronage (acomodo) [ . . . ] and in return, we need the people who vote for us to act according to what the leaders of FA propose. (Aldorio Silveira, Espacio 609 MPP leader in Bella Unión, personal interview, 2003)

Through their integration into the columns, these traditional party leaders also think they can help to moderate the more “extremist” programmatic stances that still exist in other fractions of the FA.

The FA’s Harmonized Mixed Strategy As an opposition party, the FA successfully articulated a mixed linkage strategy that gradually allowed it to gain support among segments of the population that had historically supported the traditional parties, while avoiding alienating its traditional leftist constituency. In so doing, the FA changed from a Marxist mass party to an electoral-professional one, pursuing a segmented electoral strategy while still providing consistent opposition to neoliberal reforms, thereby collecting and channeling social discontent with the other parties in the system. Moreover, the FA’s programmatic realignment around Batllismo allowed it to increase the level of programmatic linkages between the party and poor constituents. This was based on a strategy of opposing incumbents (and their policy choices) in the media, while penetrating the clientelistic strongholds of traditional parties with its activist base. In addition, far from hindering the party’s electoral chances, its moderation and leadership disputes helped the party to widen its electoral menu through fractionalization, while the electoral supply of the two traditional parties became increasingly narrower. The FA’s administration of the municipal government of Montevideo, the development of a charismatic leadership, and its engagement in social policy provision and decentralization were also crucial to expanding its constituency in the periphery of the capitol. As a result, the party now draws support from a socially diverse constituency that can be schematically portrayed as composed of two groups having contrasting interests and diverging degrees of organization. The party’s historical constituency is comprised of groups that seek to defend the remnants of the state-centric model. This constituency is largely characterized by unions tied to inward-oriented production (where state employees are predominant)



Strategic Harmonization of Segmented Linkages

253

and pension beneficiaries. During the 1990s, relying on their organizational capacity and forging a tacit alliance with the FA, these social groups were better able than other social segments of the population to protect their interests from attempts at neoliberal reform. Meanwhile, the FA has also gradually consolidated its electoral grip in more vulnerable sectors of the population who did not directly benefit from the ISI model, or who suffered from downward social mobility during its crisis. Since 1994 in Montevideo, and more recently in the interior of the country, the FA has absorbed an emergent constituency that is comprised of informal sector workers, the unemployed, a middle class suffering downward mobility, rural workers, small merchants, agricultural producers, and even some elite sectors. Several conflicts of interest exist between the two constituencies, as well as among different fractions of the party, with privileged representation of given electoral segments (see Luna 2007). Such contradictions were also heartily felt by historical activists of the FA, in the wake of the party’s arrival into office: It is reasonable that people realize that the left is the political future of this country. Therefore, people are coming from everywhere. And to win the election, that’s fine. But what are we going to do afterwards? The militants that we have in the committees and who drive the internal politics of the party are totally worthless. They have been militants for thirty years, everyday, going to the committee, collecting signatures, campaigning. They now want some compensation and you will need to give them some positions. And that will be a mess. And the old sympathizers that we have in society are all against Frente Amplio’s current leadership, due to previous (internal) conflicts. Those people today are skeptical, unsure, uneasy. However, you need those. They are your pillars in society and they have technical capacity. But if you take them, the others will be infuriated. The same applies for electoral alliances. You have people who are now merging with you under the Nueva Mayoría, obviously seeking office. So, to gain at most an extra 3 percent of the electorate you are stripping your historical supporters of those positions and giving them away to a guy that the day before yesterday was your enemy. (David Rabinowitz, FA local activist, personal interview, 2003)

Programmatic differences were also prevalent among different FA fractions. During the first FA presidential term, those differences were clearly manifested on the debate about the convenience of signing a free-trade agreement with the US, which was promoted among others by Tabaré Vázquez and Danilo Astori, and which was opposed (victoriously) by some fractions of the FA (most prominently the PCU and several leaders of the PSU).68   68  Internal programmatic differences also extended to other issues. For instance, in contrast to the great majority of the FA and its labor base, Astori supported the association of strategic public enterprises with private capital, the promotion of macroeconomic policies to foster foreign investment, and an active engagement with international financial institutions to negotiate reforms on the basis of macroeconomic stability. Astori also publicly opposed the statement

254

Segmented Representation

Yet, the FA has been able to hold together, successfully claiming the presidential re-election in 2009. Its strong partisan identification, as well as the successful development (and alternation) of charismatic leaderships, has enabled the pursuit and maintenance of a self-harmonized electoral strategy to mobilize different segments of the Uruguayan population on the basis of segmented linkages. The UDI and FA in Comparative Perspective As argued in this chapter, the UDI and FA fared better than their electoral competitors in Chile and Uruguay. The UDI and FA differ in terms of their ideological platform, their internal organization (more coherent and hierarchical in the case of the UDI, much more diverse and confrontational in the case of the FA), and the opportunity structures they faced when competing in the Chilean and Uruguayan systems. In spite of such differences, both parties share a series of commonalities that are worth mentioning before closing this chapter. In particular, both parties were relatively better able to segment and at the same time coordinate their linkage strategies than their competitors. In the case of other Chilean parties, segmentation was possible, but harmonization was not achieved. In the case of traditional parties in Uruguay, segmentation was constrained by the erosion and restraint of clientelistic linkages and by the programmatic “cornering” of both parties in an unpopular pro-reform stance. Moreover, the scarcity of clientelistic side-payments, the erosion of both parties’ symbolic resources (i.e., the decay of previously very strong partisan identities), and the prospects of electoral defeat also hindered traditional parties’ capacity to self-coordinate their electoral strategies. In turn, besides differences caused by the systemic nature of linkage strategies (i.e., the UDI pursued a dual strategy and the FA a mixed one), both parties’ electoral performance hinged on their capacity to draw on the resources provided by its core constituencies to pursue electoral expansion. While the nature of those resources and the composition of each core constituency issued by the President of the Pension System Administration (Banco de Previsión Social) on the need to reverse the pension reform passed in 1996, which had partially privatized the “pay as you go system,” instituting individual accounts administered by private investment companies. In the same vein, while the PSU’s General Secretary stated that the FA needed to roll back capitalism and move towards socialism, José Mujica claimed that what Uruguay really needed was a “serious [uncorrupted] capitalism.” At the same time, however, the MPP has consistently advocated a return to “economic protectionism” to strengthen the economy’s internal market.



Strategic Harmonization of Segmented Linkages

255

differed markedly across cases (i.e., financial resources provided by business interests in the case of the UDI; an activists network and strong partisan identities provided by labor unions, the student and cooperative movement, and its historical middle-class electorate in the case of the FA), their effect in providing the party with a greater capacity to simultaneously segment and harmonize their linkage strategies is functionally equivalent. In both cases, the parties clove to their core constituencies. While in opposition they relied on the institutional resources they still had, such as their increasing congressional representation. The UDI also relied on the “authoritarian enclaves” of the Chilean constitution to ensure that the party could act as the keeper of market reforms even without commanding many congressional seats. The FA also relied on its capacity to mobilize large segments of the electorate around direct democracy initiatives to block disadvantageous legislation promoted by the governing parties. Moreover, drawing on their disproportional access to particular resources provided by their core constituencies and scarcely available to their competitors, both parties were able to carve out a peripheral constituency, penetrating previous strongholds of their rival parties. Finally, both parties were able to reduce the eventual electoral trade-offs and internal conflicts that their segmented linkage strategies might have produced. Contextual factors such as social segmentation and differences across districts (particularly Chile) were instrumental in reducing the risks of pursuing a segmented linkage strategy. Yet, partisan endowments and explicit strategies were also crucial for achieving strategic harmonization. Strong partisan identifications (especially in the case of the FA) and a sense of mission (especially in the case of the UDI), the development of charismatic leaderships that were able to cement an heterogeneous social and political coalition (particularly the FA), and the allure created by the prospects of winning presidential office proved pivotal in both cases.

PA R T I I : C AU S A L I N F E R E N C E

6 Causal Induction: Explaining Linkage Structures in Chile and Uruguay This chapter inductively develops a series of causal claims regarding the current configuration of party–voter linkages in Chile and Uruguay and their changes over time. The chapter introduces a series of explanatory propositions answering the following questions: Why is the socioeconomic segmentation of linkages greater in Chile than in Uruguay, and why are party–voter linkages segmented between districts in Chile and within districts in Uruguay? Why have programmatic linkages become more prevalent and clientelism declined more sharply in Uruguay than in Chile? Why have candidate-centered appeals trumped party-centered linkages in Chile, while party-centered linkages have remained crucial in Uruguay? Why are the Chilean UDI and Uruguayan FA so much better at simultaneously segmenting and harmonizing linkage strategies than other parties in their systems?

T H E VA R IA N C E T O B E E X P L A I N E D As observed in Chapters 2–5, between 1973 and 2010 the party systems of Chile and Uruguay witnessed significant transformations, reshaping the nature and societal distribution of party–voter linkages. Such changes challenge static and systemic views of these linkages. They also challenge the conventional characterization of Chile and Uruguay as cases with similar institutionalized and programmatic party systems, as well as explanations of those similarities stressing their long-term determinants (Kitschelt et al. 2010). Drawing on Tables 2.5 and 2.8, Table 6.1 summarizes the changes in linkage strategies noted over time and across cases. Table  6.1 focuses on the systemic-level trajectory of each case in terms of the following six dimensions:  (1)  level of socioeconomic segmentation; (2)  linkage types most relevant for mobilizing core constituencies; (3) linkage types most relevant for mobilizing peripheral constituencies; (4)  whether district-level territorial

Table 6.1.  Party–voter linkage change in Chile and Uruguay: a summary of the descriptive argument Dimension

Chile until 1964

Chile 1964–1973

Chile 1990s–2000s

Uruguay until 1973 (fractions)

Uruguay until 1973 (parties)

Uruguay1990s– 2000s

Socioeconomic segmentation Linkage to core constituency

High

Low

High

Moderate to low

Moderate to low

Moderate

Programmatic

Programmatic

Candidate-traits, PID programmatic fractional currents

Programmatic

Linkage to peripheral constituency

Clientelism Clientelism (patronage, pork, (declining constituency electoral role) service) Between districts Within districts (yet, close to pure linkage strategies) High High

Coalitional IDs (regime divide), weak programmatic. (Declining electoral role of core constituency.) Candidate traits/ particularism

Clientelism, particularism

Clientelism

Candidate-traits, particularism, PID

Between districts

Within districts

Within districts

Within districts

Low

High

Minimal to none

Moderate (parties), high (fractions)

Self-harmonization (less needed due to close to pure linkages)

Decentralized harmonization not achieved

Harmonization by fraction leaders

Pure linkages

Non-harmonized, dual

Harmonized, mixed

Decentralized Self-harmonization harmonization on the basis of PID and DVS electoral system Non-harmonized, Harmonized, mixed mixed

Patterns of territorial segmentation

Observed levels of strategic harmonization Harmonization Self-harmonization mechanism

Overall outcome

Harmonized, dual

Source: Author’s own construction on the basis of evidence presented in Chapters 2–5.



Causal Induction: Explaining Linkage Structures

261

segmentation is dual or mixed; (5)  level of strategic harmonization; and (6)  mechanisms and resources key for permitting harmonization where it occurs. On this basis, the last row of Table 6.1 identifies the overall linkage strategy observed in each case. Regarding the socioeconomic segmentation of party–voter linkages, Chile witnessed a decline from historically high levels of socioeconomic segmentation (between 1964 and 1973)  and then a return, after re-democratization, to its long-term pattern of high socioeconomic stratification. Chile’s current levels of socioeconomic segmentation are higher than in Uruguay, which has comparatively moderate levels of segmentation. Notwithstanding, Uruguay historically displayed even lower levels of socioeconomic segmentation due to its traditionally extensive, transversal patterns of both clientelism and party-identification. Declining programmatic mobilization of the popular sectors contributed to increasing segmentation in Chile between 1964 and 1973. In Uruguay, the decline and municipalization of clientelism, as well as the rise of programmatic mobilization in the post-transitional period, have also produced somewhat higher levels of segmentation than in the past, when clientelistic linkages extended to the middle classes. Whereas in Chile programmatic decline and social stratification increased the importance of linkages based on candidate-traits and particularism, in Uruguay, (segmented) programmatic linkages gained centrality and came to complement clientelistic and particularistic linkages, as well as the appeals based on party identifications and candidate traits especially important at the fractional and presidential levels. On that basis, I codified the evolution of parties’ main linkage to their core constituency, as well as parties’ electoral strategies to mobilize their peripheral electoral bases. One can conceptualize a party’s core constituency as a party’s more “captive” electorate. The core constituency of Chilean parties has decreased over time, making mobilization of a peripheral constituency more electorally crucial. Meanwhile, the parties’ core constituencies have likely remained constant or, particularly in the case of the FA, even increased in Uruguay. Mobilizing peripheral constituencies is still fundamental to parties’ strategizing in Uruguay, but the relative size of that constituency is likely smaller than in Chile. This is consistent with the presence in Uruguay of political parties (and partisan identities) penetrating the social fabric more deeply than in Chile. In Chile, both programmatic preferences, particularly regarding distributive issues or conservative/liberal policy stances, and partisan identities have continued to mobilize elites. However, these types of appeals are not central in mobilizing other voters, for whom they have been largely replaced by (or subsumed by) mobilization centered on a retrospective regime divide. These

262

Segmented Representation

strategies are the primary ones for mobilizing parties’ shrinking core constituencies. Constituency-service, particularism, and candidate-traits are in turn central for mobilizing parties’ peripheral, yet increasingly pivotal, constituencies. The hierarchical, party-centered brokerage networks that specialized in dealing with particularistic demands, as well as in providing pork and patronage at the local level during the pre-authoritarian period, have since been replaced by personalism and decentralized provision of particularistic resources. Candidate-based appeals have thus overtaken party-based linkages. In Uruguay, clientelistic linkages have receded as the primary mobilization device, and have been replaced by programmatic linkages centered on distributive preferences, often expressed as positions on the state–market divide. Such programmatic alignments have permeated voters’ preferences at both the elite and popular levels, becoming key to mobilizing parties’ core electors. Partisan identities and personalistic leaderships in turn probably now factor less in mobilizing parties’ peripheral constituencies, although both have remained important for electoral mobilization. Particularism has meanwhile generally receded to municipal politics. In terms of the nature of territorial linkage segmentation, Chile has evolved from a system with low segmentation that ran within districts to one with high segmentation structured, as it had been before 1964, between districts. Socioeconomically heterogeneous districts made up of different types of comunas are the exception to this rule, and present higher levels of within-district segmentation. Meanwhile, no such change has occurred in Uruguay, where segmentation has been and continues to be structured within districts. When it comes to strategic harmonization, the capacity of parties in Chile to harmonize linkage strategies across districts has sharply declined, in part because of changing linkage types. During the pre-1973 period, the presence of mass organizations specialized in the pursuit of programmatic and clientelistic linkages gave rise to powerful self-harmonizing mechanisms. In the contemporary period, however, the centrality of candidate-based appeals, as well as structural changes identified in this chapter has significantly reduced parties’ capacities to harmonize segmented appeals. In Uruguay, in turn, parties and especially party fractions have generally maintained their harmonizing capacity. Much like in pre-authoritarian Chile, party-based appeals play an important part in enabling parties to harmonize. However, such appeals are significantly less apparent in the case of the two traditional parties, which have lost access to state resources pivotal for the functioning of their mass organizations. Summing up, the last row of Table 6.1 contains the overall classification of linkage strategies observed in each case at the systemic level. In this regard,



Causal Induction: Explaining Linkage Structures

263

I classify Chile as a case that has evolved from a system in which harmonized mixed linkages predominated (between 1964 and 1973)  to one in which non-harmonized and dual linkages have consolidated in the post-1990 system. Uruguay has been more stable. The main transformation there has been a shift from clientelistic to programmatic linkages for mobilizing parties’ core constituencies, not changes in general strategies for segmenting and harmonizing different kinds of linkages.

T H E NAT U R E O F T H E C AU S A L A R G UM E N T:   T Y P E S O F C AU S AT IO N A N D C O D I N G O F E AC H   C A SE The subsequent sections of this chapter summarize the causal argument, one in which different causal factors interact to produce given results. For instance, state reforms contributed to the decline of programmatic mobilization in Chile, while contributing to programmatic structuring in Uruguay. The timing of those reforms (under authoritarianism in Chile, and under democratic contestation in Uruguay) explains the opposite trajectories. Similarly, institutional variables do not act in a vacuum, but interact with other clusters of variables (e.g., the spatial distribution of socioeconomic groups) in shaping strategies for segmenting linkages territorially. The same interactional effects can be seen at the individual party level, where constant and historic causes (Stinchcombe 1968)  interact in shaping party–voter linkages. For instance, while parties’ electoral strategies are driven by such structural factors as incentives deriving from particular electoral institutions and sociological configurations, they also hinge on factors arising from parties’ and party systems’ specific historical trajectories—for instance, the strong partisan identifications found in Uruguay, or the salience of the regime divide in Chile. At this level, too, multiple causal variables interact rather than operating in a simpler, purely additive manner. The results of my investigation reveal both “equifinal” and “multifinal” causal patterns—that is, patterns of causation in which distinct causal configurations produce the same results; and patterns of causation in which the same causal configuration is associated with different results. For instance, both the UDI and FA simultaneously segmented and harmonized their linkage strategies, thereby gaining an electoral edge over their competitors. However, structural, historical, and institutional differences at both the systemic and individual party levels led the FA and UDI each to achieve such segmentation and harmonization in different ways.

264

Segmented Representation E X P L A I N I N G T H E S C O P E O F S O C IO E C O N OM IC SE G M E N TAT IO N A N D P R E D OM I NA N T T Y P E S O F PA RT Y– VO T E R L I N KAG E S

In Chile, from 1973 to the 1990s and 2000s, party–voter linkages changed in four important ways for which my causal explanation seeks to account. First, their socioeconomic segmentation increased. Second, parties’ core constituencies shrank. Third, programmatic mobilization was focused on a retrospective regime divide. And finally, clientelistic linkages to peripheral constituencies declined in favor of linkages centered on candidate-traits and particularism. In Uruguay, meanwhile, one needs to explain the following developments from 1984 to the mid-1990s and 2000s. First, there has been a moderate increase in socioeconomic segmentation. Second, core constituencies have remained largely stable, or in the case of the FA, even grown. Third, linkages to core constituencies based on strong partisan identifications and candidate-traits (at the fraction level) have been largely replaced by programmatic mobilization. And finally, clientelism has become less important, and candidate-traits and particularism more so for mobilizing peripheral constituencies. Five main factors have influenced and interacted to produce these outcomes. Social Inequality High levels of social inequality tend to increase linkage segmentation while reducing programmatic mobilization. The nature of inequality in Chile and Uruguay differs. Income distribution is more heavily skewed (bimodal) in Chile. Uruguay has a more sizable middle-class, whereas in Chile there is a more significant gap between the top-income earners and the rest of society (see Torche 2005). Illustrating this situation, Graph 6.1 displays the distribution of income by ventiles in Chile and Uruguay. As a point of comparison, Graph 6.1 also includes values for the other four cases in which my argument is tested in Chapter 7. To facilitate the comparison of income distributions, each line shows the ratio between the income of each ventile and the income obtained by the top 5 percent (ventile 20). As shown in Graph 6.1, Bolivia and Brazil have the most skewed income distribution structures. Chile’s is similar to that of El Salvador, but with the



Causal Induction: Explaining Linkage Structures

265

0.7 0.6 0.5

Bolivia Brazil

0.4

India EI Salvador Chile

0.3

Uruguay Sweeden

0.2 0.1 0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19

Graph 6.1.  Income ventile distributions Source: Author’s own construction on the basis of Milanovic (2012). Standardized (ratio over income of ventile 20) income ventiles for 2002.

poorest ventiles marginally better off. Compared to the income distribution observed for Uruguay, a country with a larger middle-class, the Chilean distribution is more split between high- and low-income earners. The evolution of inequality in Chile and Uruguay is theoretically consistent with the evolution of programmatic mobilization and widespread linkage segmentation observed in each case. The two countries’ trajectories have been distinct when it comes to levels of social inequality. In the early 1970s, both societies had Gini coefficients lower than the regional mean of 0.53 (Chile = 0.5, Uruguay = 0.43). Since the 1970s, however, the two cases have diverged. Income inequality has clearly increased in Chile, reaching 0.6 around 1998–2000, against a regional mean that remained virtually constant at 0.54. Meanwhile, Uruguay’s Gini coefficient only marginally increased to 0.45 by 1998–2000.1 From the 1990s until at least the mid-2000s, Uruguay also displayed a gradual but consistent increase in inequality across different sets of indicators (Bucheli and Furtado 2005).   1  Income inequality estimates proceed from Behrman et al. (2006), and were based on data from World Development Indicators (2005), WIID2a (2005), Thomas, Fan, and Wang (2001), and Deininger and Squire (1996).

266

Segmented Representation

High inequality hinders parties’ attempts at mobilizing large electoral constituencies on the basis of redistributive preferences. Although a group of the electorate may be mobilized on that basis (Piñeiro and Rosenblatt 2011), high inequality generates distributive conflicts among different electoral constituencies with seemingly incompatible distributive preferences. Parties have an interest in reducing such conflicts internally, and thus have an incentive to make their programmatic stances less prominent, as both the FA and the UDI have done. Both parties have groups of constituents among whom distributive conflicts could emerge. The UDI faces the danger of conflicts between poor supporters and supporters among conservative business elites. Conflicts of interest between unionized workers and pension beneficiaries on the one hand, and poor, informal sector workers on the other, or between state and private sector employees, similarly present latent dangers to the unity of the FA. This type of configuration tends to encourage segmentation and reduce parties’ pursuit of encompassing programmatic appeals. In addition, high levels of inequality make segmentation more feasible. The traditional relationship between poverty and clientelism is well-documented among the incentives parties face in such contexts (Kitschelt and Wilkinson 2007). Inequality also shapes the nature of party–voter linkages through its interaction with other factors, as discussed further below. Chile and Uruguay also differ regarding the evolution of not only social inequality, but also other theoretically relevant socioeconomic indicators. Since the late 1980s, Chile has seen sustained economic growth. Chile’s maximum level of unemployment between 2000 and 2006 was 10 percent (in 2000). Since the 1990s, the country has also consistently reduced poverty, from around 20 percent in 2000 (and more than 40 percent in the wake of the transition to democracy) to about 14 percent in 2006 (Valenzuela and Duryea 2011). Uruguay, meanwhile, had the worst economic crisis in its history in 2002. Unemployment reached 17  percent, and poverty climbed up from 9.4  percent in 2000 to 18.8  percent in 2005 (Valenzuela and Duryea 2011). One might think that greater levels of economic development and lower levels of poverty would translate into more politically knowledgeable voters, encouraging more programmatic mobilization (Kitschelt and Wilkinson 2007). Following this logic, with rising poverty in Uruguay, one would thus expect a decline in programmatic linkages and an increase in clientelistic mobilization. One would also then expect to see rising linkage segmentation in Uruguay, not in Chile. In reality, however, precisely the opposite has occurred. Indeed, a dynamic analysis of the Uruguayan case suggests that increases in poverty caused by downward social mobility might not lead to an increase in non-programmatic mobilization, as could be expected on the basis of the strong empirical correlation between poverty and clientelism. For instance, new shantytown dwellers in post-crisis Uruguay brought with them high levels of



Causal Induction: Explaining Linkage Structures

267

political sophistication and organizational capacity that then became critical in relating to parties by articulating more programmatic demands. Privatization, Social Policy Provision, and Decentralization During the military dictatorship, Chile witnessed massive economic restructuring, with the liberalization of markets, privatization of most public enterprises, and overhauling of social policy programs (Martínez and Díaz 1996; Silva 1996; Castiglioni 2005). Education and health policy were decentralized to the municipal level, and a “dual” social policy matrix was consolidated. Coverage of social needs for the better off was privatized, while a system of targeted social policy that expanded during the 1990s and 2000s took care of basic social provision for the poor. Although it liberalized the economy, the Uruguayan military dictatorship did not attempt massive privatization and social policy reform (Castiglioni 2005). In this regard, the pre-authoritarian system remained virtually untouched, although significantly under-funded, in the post-transitional period. Graph 6.2, showing pension provision by income deciles, illustrates some of the distributional effects of each country’s trajectory. 60 50 40 30 20 10 0 I

II

III

IV

V

VI

VII

VIII

IX

X

Percentage of households receiving pensions in Chile Pensions’ participation in total household income (in%) in Chile Percentage of households receiving pensions in Uruguay Pensions’ participation in total household income (in%) in Uruguay

Graph 6.2.  Pension benefit structure by income decile in Chile (2003) and Uruguay (2005) Source: Valenzuela and Duryea (2011) on data from household surveys.

Total

268

Segmented Representation

As shown in Graph 6.2, public pension coverage and its contribution to households’ total income is significantly lower in Chile than in Uruguay. The public pension system therefore plays a more important role for citizens in Uruguay than in Chile. Graph 6.2 also suggests that distributive stakes regarding public pension provision differ markedly across cases. In Chile, public pensions are clearly targeted to the poorest deciles. They represent a significant share of the income received by households in the first income decile. For deciles 3 to 10, by contrast, public pension benefits represent less than 10 percent of households’ income. By contrast, post-transition Uruguay inherited a pay-as-you-go system consolidated under ISI. As in other such cases in the region, pension benefits are regressive (Huber et al. 2006). In this case, pension benefits are received by the richest deciles and represent a greater proportion of the income received by the middle and upper classes. By contrast, for those in the first two deciles, including a large segment of the informal sector, such benefits are virtually inconsequential. Distributive preferences mobilized through programmatic proposals centrally concern provision of public goods. Privatizing social services or other public goods tends to socially fragment interest in public goods because wealthier citizens are more likely to exit from public services in search of higher quality. The social coalition interested in the public provision of such goods then shrinks and weakens politically (Filgueira and Filgueira 1997a). In this regard, Graph 6.2 clearly suggests that the coalition of beneficiaries in favor of public pensions is politically weaker in Chile than in Uruguay, where a much larger proportion of citizens have a relevant stake in the system. The fragmentation of interest groups (except from business associations) in post-transitional Chile reflects the shrinking role of collective action around conflicts over provision of public goods. The labor legislation introduced under Pinochet made the firm the only lawful level for collective bargaining. Large corporations then fragmented labor organization through structures of ownership that turned every individual shop into a firm, and by making extensive use of subcontractors. As a result, within a single corporation it is common to find multiple and competing small unions (Palacios 2010). At the aggregate level, these patterns translate into a very fragmented labor movement unable to act at the national level. Within Chilean municipalities, there are similarly high levels of fragmentation and competition among small neighborhood organizations working around the same types of issues. Such fragmentation is catalyzed by competition for obtaining municipal and national-level funding for small projects. Moreover, territorial fragmentation is also high in Chile. According to the evidence collected by Castiglioni (2005, p. 39), breaking up large organized



Causal Induction: Explaining Linkage Structures

269

interests was a central political objective motivating the municipalization of education and health under Pinochet.2 The configuration of interest groups in Uruguay is radically different. In Uruguay, integrated, national-level organizations represent such relevant interest groups as business, labor, and rural producers. Unions, cooperative movements, and pension-beneficiary groups also operate at high levels of aggregation. As a result, political parties in Uruguay can mobilize large interests groups through programmatic, distributive appeals. In Chile, by contrast, parties face functionally and territorially fragmented interest groups. They therefore have more difficulty developing encompassing programmatic messages. Instead, Chilean parties and candidates therefore tend to link to interest groups through other types of appeals, often entailing one-by-one delivery of goods and services (clientelism or particularism). Despite the fact that Chile and Uruguay are still two of the most centralized polities in Latin America (Eaton 2004), decentralization has also been an important driver of linkage change in each case. In contemporary political science, “decentralization” refers to at least three different processes: fiscal decentralization, administrative decentralization, and political decentralization (see Falleti 2010). Two kinds of decentralization processes seem to have contributed to transforming linkage patterns in Chile and Uruguay. First, even in the absence of fiscal decentralization both countries have seen the administrative decentralization of services, such as health and education, previously provided by central state agencies. Second, despite their generally limited political decentralization, the split between local (municipal) and national elections in both Uruguay and Chile has contributed to reducing the local electoral impact of presidential and congressional election results, while strengthening mayors. When privatization and decentralization are jointly pursued, provision of public goods becomes not only socially segmented, but also territorially fragmented. Actors mobilized around redistribution of public goods then need to coordinate across districts to form a more powerful coalition. Alternatively, they can enter into one-to-one relationships with their local provider, be it a municipal government, or a private agency. Particularly in the popular sectors, local mobilization is, at least in the short run, usually more cost-effective than collective action requiring cross-district coordination. My analysis of these two cases therefore shows that state reforms and decentralization contribute to fragmenting collective action around provision of public goods and reducing its political salience. These developments in turn hinder parties’ use of programmatic linkages for mobilizing large  

2 

Interview with Mercedes Cifuentes cited in Castiglioni (2005).

270

Segmented Representation

constituencies. Parties then usually segment their linkage strategies, mobilizing one electoral constituency programmatically, and others with different linkages. While significantly reducing parties’ scope for mobilizing programmatic distributive preferences, state reforms also contribute to reshaping non-programmatic linkages, such as clientelism, particularism, and candidate-based mobilization. Privatization and decentralization impede parties’ access to state resources for feeding clientelistic and patronage machines instrumental in maintaining parties’ territorial organizations (see Harbers 2010; Morgan 2011; Dargent and Muñoz 2011). To be sure, local political machines may flourish at the local level, but national-level party structures are weakened. Parties’ weakening brokerage role and augmented local-level resources in turn strengthen local political figures vis-à-vis national leaders. To solve problems, voters increasingly target municipal-level politicians, reinforcing the “localization” of party–voter linkages. In the absence of stringent campaign finance legislation, privatization also reduces the role of state-provided goods in political campaigns. This outcome may be positive in terms of fiscal responsibility and the rule of law. However, it also empowers private actors able to fund campaigns. These include business actors as well as other interest groups ranging from conservative groups to illegal organizations. The importance of private financing in Chile illustrates this effect. Timing of Market and State Reforms Once accomplished, state and market reforms introduce limitations to the pursuit of programmatic linkages. However, if attempted under conditions permitting democratic contestation, they can have the opposite effect. Indeed, increased programmatic linking in Uruguay arose largely from the politics of state reforms during the 1990s. The findings of Hagopian et  al.’s (2009) examination of Brazil and Kitschelt and Kselman’s (2011) large-N analysis are similar. The latter finds that economic crises create opportunities for parties to mobilize around policy options addressing the crisis, thus increasing the salience of programmatic politics. By contrast, in cases where drastic market and state reforms were introduced under authoritarianism, this “positive” effect was lost. For instance, between 1973 and 1980 455 of a total of 479 public enterprises were privatized in Chile, while health and education were municipalized. Those radical reforms did not produce lasting programmatic alignments that could later be mobilized after democratization (Vergara 1986, p. 90).



Causal Induction: Explaining Linkage Structures

271

Authoritarian Legacies and the Politics of the Transition to Democracy As described in Chapters  2 and 3, the politics of each case’s authoritarian breakdown and transition to democracy created enduring alignments in the party system. It also changed the valence of strong partisan identities and subcultures, eroding them in Chile and consolidating them in Uruguay. Both cases illustrate how a violent past splitting actors along lines of political identity can freeze citizens’ political allegiances for a significant period of time. In Chile, the coup experience of 1973 and the alignments emerging during the transition to democracy gave rise to the Alianza and the Concertación party camps. The incorporation of the Partido Demócrata Cristiano within the Concertación, as well as the exclusion of the PC-Chi, decisively shaped the conflicts and cleavages mobilized in the post-transitional period. The socialist renovation triggered by the traumatic experience of the military coup and “political learning” during the European exile (Roberts 1992; Funk 2004) also contributed to shaping the Concertación’s moderated programmatic stand. As a result, the center-left coalition downplayed distributive issues, as well as moral ones, making the authoritarian/democratic cleavage electorally predominant. In Uruguay, no regime divide split the parties between different camps. However, resistance against the dictatorship contributed to solidifying citizens’ traditional party identifications and helped define the FA’s identity. The relative size of parties’ core constituencies and strength of parties’ links to them based on party identity therefore also turned on the politics of the transition to democracy and its aftermath. Institutional Factors The binominal electoral system of Chile overdetermines neither the partisan composition nor the mobilization strategies of the two main electoral pacts. It has, however, definitely contributed to freezing the alignments crystallized during Chile’s transition to democracy. In Uruguay, the constitutional reform of 1996 that introduced the presidential runoff also helped keep electoral competition structured around two partisan families. More importantly, that reform also contributed to increasing local–national dealignment by introducing a split between municipal and national-level elections. Chile’s more personalized electoral formula with its smaller district magnitudes and open lists has facilitated candidate-centered appeals, weakening party-based linkages. Uruguay’s electoral system, with its larger district

272

Segmented Representation

magnitudes, PR, and blocked lists, has meanwhile encouraged party-centered linkage strategies. These findings are in line with standing predictions in the literature (i.e., Carey and Shugart 1995). Moreover, my analysis suggests that other institutional features can also become pivotal for reinforcing particular linkage strategies. For instance, in Uruguay the FA and ISI-beneficiary groups coalesced around a strategy of opposing state reforms, using direct democracy mechanisms included in the Uruguayan constitution (Altman 2011). Political and electoral institutions thus also play a significant role in helping to crystallize and reinforce the causal effects identified above.

E X P L A I N I N G PAT T E R N S O F T E R R I T O R IA L SE G M E N TAT IO N The patterns of territorial segmentation in Chile differ significantly from those in Uruguay, in large part because of differences in districting. The solely urban and the purely rural districts of Chile are smaller and more socially homogeneous, and district boundaries typically coincide with territorial divisions between socioeconomic groups.3 This pattern has correlated with increasing levels of between-district segmentation. In more heterogeneous districts comprised of both urban and rural settlements, segmentation has instead run within districts. In Uruguay, districts are larger with boundaries cutting across socioeconomic divides. There, segmentation has been largely structured within districts. Maps 6.1 and 6.2 show the overlap between socioeconomic and district demarcations in Chile and the social heterogeneity of Uruguayan districts. Map 6.1 shows the geographic distribution of municipal human development indexes computed for the metropolitan region of Santiago de Chile (circa 2000). Map 6.2, in turn, displays the geographic distribution of schooling rates (a proxy for socioeconomic development) by neighborhood in Uruguay’s capital, Montevideo. Both cities are sociogeographically fragmented, with people of similar socioeconomic backgrounds living close together and segregated from those of different backgrounds. Let us now consider district magnitude and district sizes (and borders). The dashed lines in Map 6.1 represent the borders of lower chamber congressional districts in Santiago. Each of these districts elects two representatives to the lower house, independently of their respective populations. Meanwhile, the city of Montevideo and its surrounding   3  This is not completely exogenous to political will. Rather, it was sought in the gerrymandering pursued under the military dictatorship (see Navia 2005 and Londregan 2000).



Causal Induction: Explaining Linkage Structures

273 N W

E S

Municipal Human-Development Index Very High = HDI > .85 High = .80 < HDI > .85 Intermediate = .75 < HDI > .80 Low = .70 < HDI > .75 Very Low HDI < .70 District limits Municipal limits

0

15

30

60

90 Kilometers

Map 6.1.  Social fragmentation and district boundaries (each elects two congressional representatives) in the metropolitan region of Santiago Source: Author’s own construction on the basis of Municipal Human Development Index (UNDP 2000).

rural area comprise a single congressional district, which elected forty-four lower house representatives in 2009. In sum, electoral institutions and socioterritorial segmentation interact and are together crucial to determining whether segmentation occurs within or between districts. When district boundaries cut across the territorial distribution of socioeconomic (or ethnic, racial or religious) groups, segmentation within districts then becomes more attractive. Larger districts also make such intra-district segmentation easier to hide from constituents and thus more feasible. Conversely, where district boundaries reinforce territorial segregation, segmentation between districts becomes more likely. The nature of interest group coordination across districts also shapes territorial segmentation. If an interest group is able to organize across districts, it can relate to the party system at the aggregate/national level. In such cases, segmentation between districts should be lower than where interest groups are organized only locally. The difference between evangelical groups and the Catholic Church in Chile provides a good illustration. Whereas the former are territorially fragmented and typically seek particularistic benefits

274

Segmented Representation N W

E S

Percentage of households in neighborhood with low education climate Up to 15% Between 15% and 30% Between 30% and 50% Between 50% and 70% More than 70% District limits

RÍo de la Plata 0

1,25 2,5

5

7.5

10 Kilometers

Map 6.2.  Social fragmentation in the district of Montevideo (elects forty-four congressional representatives) Source: Author’s own construction on the basis of information from the Servicio de Información Geográfica of the Intendencia Municipal de Montevideo. Households with “Low Education Climate” are defined as those in which the average schooling rate of adult inhabitants is six years or lower.

through municipal-level mobilization, the latter has much more national-level influence.

E X P L A I N I N G T Y P E S O F S T R AT E G IC HA R M O N I Z AT IO N Chile’s levels of strategic harmonization were high in the pre-1973 period but very low after 1990. Uruguay had only minimal levels of strategic harmonization at the party level before 1973, although harmonization was then much greater at the fractional level. In post-transitional Uruguay, party-level harmonization has moderately increased. The two party systems have thus had opposite trajectories in this regard. Several of the factors already identified above as important in determining the structure of linkage segmentation also



Causal Induction: Explaining Linkage Structures

275

play an important role in shaping the types of strategic harmonization parties tend to pursue in each case. Notwithstanding these national-level effects, as argued in Chapter 1, individual parties’ particular symbolic and material endowments and the linkage types they pursue are also critical for shaping their harmonization strategies. When a party competes on the basis of candidate-centered appeals, a system of decentralized harmonization is needed. This is of course difficult to achieve if the party and its leaders do not have massive symbolic and/or material resources that are relatively unavailable to individual candidates acting alone. If candidate-based linkages pay off electorally, parties as collective institutions are thereby progressively weakened. Political parties in contemporary Chile have suffered such effects. Conversely, when parties compete on the basis of party-centered appeals, a series of self-harmonizing mechanisms come to bear, reducing the need to purposely seek strategic harmonization. Typical party-based linkages are clientelism, PID, and programmatic appeals. Self-harmonizing, party-centered mechanisms reinforce the value of parties and their candidate endorsements, thus reducing incentives for individual candidates to act “independently” at the district level. For instance, when a party has a popular national-level leader, the party label becomes more valuable, creating greater incentives for individual candidates to converge to the national-level strategy, in exchange for the substantial benefits of riding on the national leader’s coattails.4 The same logic applies to programmatic stances, or parties’ reputed ability to distribute material resources to their voters. Here too, candidates seek to associate themselves with the party label, thus further enabling party leaders to strategically harmonize the party’s electoral strategy. The cases of Chilean parties in the pre-1973 era, as well as those of Uruguayan parties in the post-transitional period, approximate this scenario. A last, intermediate scenario arises when parties implement both partyand candidate-based linkages. In this case, party-based linkages still provide partisan elites with useful resources for strategically harmonizing campaigns through district-level party leaders, even if campaigns at the district level are structured around candidate-based appeals. The case of the UDI in contemporary Chile is consistent with this type of scenario. Additional, relatively exogenous factors can also provide partisan elites with material, symbolic, and institutional resources facilitating harmonization.

  4  As argued in Chapter 4, the cases of Joaquín Lavín and Michelle Bachelet in Chile, as well as those of Tabaré Vázquez and José Mujica in Uruguay (along with national-level fractional leaders) exemplify this mechanism.

276

Segmented Representation

Such exogenous factors are more likely to be important where selfharmonizing mechanisms are absent. In short, at the systemic level electoral institutions, and institutional factors more broadly, are critical. Privatization also conditions parties’ harmonizing capacity at the systemic level. Meanwhile, at the individual party level parties’ trajectories within the system, as well as their differential access to symbolic and material resources, explains their relative capacity to strategically harmonize segmented linkage strategies. District Magnitude The number of candidates from the same party competing for congressional seats within a district is a function of district magnitude. Greater district magnitude should increase the impact of party-level harmonization. Such harmonization, in turn, is apt to increase segmentation within districts. It should also augment the influence of party strategy on district campaigns, therefore limiting segmentation between districts. Conversely, if district magnitudes are small campaigns will tend to be more personalized. The prevalence of candidate-based linkages in turn makes party-level harmonization more challenging by obstructing self-harmonizing mechanisms. These effects of small district magnitude can be observed in the case of Chile. Even within Uruguay, where districts are larger, there is more personalization in districts with smaller magnitudes than in districts with larger ones. Montevideo and Canelones, for instance, display less personalization than districts in the interior with magnitudes of 2 and 3. Separation of Municipal and National-Level Elections The split between local- or municipal- and national-level elections also diminishes strategic harmonization. It increases the disjuncture between nationaland municipal-level candidates’ electoral fortunes, thereby setting the stage for more internal party conflict as candidates from the same party come to compete more and cooperate less at the district level. This scenario is illustrated by the increasing levels of “localization” observed in both Uruguay (since the 1996 constitutional reform) and in Chile (since 1992, when the first post-authoritarian municipal elections were held).



Causal Induction: Explaining Linkage Structures

277

Nomination Procedures As often argued in the institutionalist literature, candidate-nomination procedures also seem to affect levels of harmonization at the party level (Carey and Shugart 1995; Siavelis and Morgenstern 2008; Samuels 1999; Hazan and Rahat 2010). PR and blocked party-lists, as observed in Uruguay, should facilitate partisan harmonization. The binominal system and informal, intra-pact bargaining regarding district-level nominations should also centralize nominating power at the elite level (Carey and Siavelis 2003). Such mechanisms should enhance party leaders’ power to enforce high levels of strategic harmonization (Siavelis and Morgenstern 2008; Navia 2008; Altman 2008). Yet, my evidence, as well as that discussed by Navia (2008), suggests that such nominating power is only exerted when the incumbent decides to exit the district and an open congressional seat becomes available. Meanwhile, strong incumbency advantages have emerged from candidates’ ability to craft strong personal relationships with their local constituents. Over time those advantages have shifted power from partisan elites to congressional incumbents, thereby reducing the role of nomination as an instrument for achieving strategic harmonization. Campaign Finance Regulations Campaign finance legislation also shapes parties’ capacity to harmonize linkage strategies. Funding structured at the party level makes harmonization more feasible than where individual candidates raise funds in a decentralized way. In Chile, congressional campaigns largely depend on individual candidates’ capacity to provide their own funding. Indeed, when partisan elites select candidates for an open congressional slot, potential candidates’ financial capacity and name recognition in the district usually determine the nomination (Luna and Rosenblatt 2012). Moreover, after the election public funding is allocated to candidates on a decentralized and personal basis, depending on the votes they received in their district. In Uruguay, the public electoral subsidy also provides ex post funding based on the number of votes received. However, the subsidy is proportionally distributed to each party’s individual senate and deputy lists. Therefore, at least at the fractional level, the potential for strategic harmonization is greater. Although individual candidates with individual access to funding could eventually “climb up” on electoral lists, fraction leaders still have significant control over campaign contributions and public funding.

278

Segmented Representation Privatization and State Reform

The diachronic analysis of the Chilean case and of the traditional parties in Uruguay also suggests that privatization, as well as state reform, reduced parties’ access to state resources that were previously instrumental in achieving strategic harmonization through self-harmonizing mechanisms. In both cases, parties had relied on pervasive mass organizations funded through local–national brokerage networks that later declined. In Uruguay, the rise of programmatic mobilization in the post-transitional period, as well as the resilience of party identities, compensated for the decline of brokerage networks. As a result, self-harmonizing mechanisms continued to be available to political parties. In Chile, the increasing pursuit of candidate-based appeals further weakened parties’ capacity to strategically harmonize their electoral campaigns. Overall, the decline of parties’ roles as brokers of central state goods was less detrimental to externally mobilized parties (such as the FA in Uruguay) and new ones (such as the UDI or the PPD in Chile) that had not historically relied on that type of brokerage network to maintain a territorial apparatus. Moreover, privatization also increases the potential role of private actors in providing material resources useful in the crafting of party–voter linkages. Access to such resources can be skewed towards one party (as in the case of the UDI in Chile) or be more evenly distributed across the system (as in Uruguay). In the first case, and particularly if campaign finance legislation is favorable or non-enforced, access to private financing introduces significant biases in favor of particular parties. As in the case of public financing, private donations are more pivotal for achieving harmonization when they are received and administered by parties, rather than individual candidates. Individual Parties’ Access to Material and Symbolic Resources The Chilean and Uruguayan cases also illustrate the influence of non-institutional factors on harmonization. Such non-institutional factors interact with parties’ individual traits and strategic positioning in the party system and are most helpful for explaining individual party differences in achieving harmonization. Despite important differences in the types of resources and mechanisms involved, as discussed in Chapter 5, both the UDI’s and the FA’s harmonization of their electoral strategies illustrates this claim. Explanatory proposition 8 (in the Conclusion to this chapter), draws on that discussion and contradicts expectations of complete contagion of partisan strategies.



Causal Induction: Explaining Linkage Structures

279

C O N C LU SIO N Particular analytical payoffs of the kind of small-N comparative case study method used in Part I of this book include both improving descriptive inference and generating new, researchable hypotheses. Complementing this book’s primary objective of improving descriptive inference, this chapter derives a series of general propositions from the Chilean and Uruguayan comparative case studies presented in the preceding chapters. In closing this chapter, I now summarize the inferred causal arguments identified above, formalizing a set of seventeen explanatory propositions: • Explanatory proposition 1: Higher levels of social inequality increase linkage segmentation. • Explanatory proposition 2: Higher levels of social inequality reduce programmatic linkages. • Explanatory proposition 3: Interest groups organized across districts and at higher levels of aggregation facilitate programmatic linkages and reduce linkage segmentation. • Explanatory proposition 4: More extensive public good provision increases opportunities for crafting more extensive, less socioeconomically segmented programmatic linkages. • Explanatory proposition 5: Increasing political decentralization facilitates territorially segmented linkages. • Explanatory proposition 6: Privatization reduces use of state resources for funding non-programmatic linkages, while enhancing opportunities for private interests to gain influence through political financing. • Explanatory proposition 7: Pursuit of market reforms in a context of democratic contestation increases programmatic linkages. • Explanatory proposition 8:  The strength of party identities and the predominant programmatic alignments in a party system depend on salient past events and are therefore path-dependent. Individual parties’ trajectories in the party system condition their relative access to material and symbolic resources for segmenting and harmonizing their linkage strategies. • Explanatory proposition 9: More personalized electoral formulas encourage candidate-based linkage strategies. • Explanatory proposition 10: Small district sizes combined with socioeconomic territorial segregation (which is generally associated with socioeconomic inequality) favor a dual pattern of linkage segmentation. • Explanatory proposition 11: When interest groups coordinate across district boundaries, a dual pattern of linkage segmentation becomes less likely.

280

Segmented Representation

• Explanatory proposition 12: Pursuing party-based linkages improves parties’ capacity for strategically harmonizing segmented linkage strategies. • Explanatory proposition 13: Smaller district magnitudes should decrease parties’ strategic harmonization capacity. • Explanatory proposition 14:  Greater levels of political decentralization should reduce parties’ levels of strategic harmonization. • Explanatory proposition 15a: Institutional rules enhancing party leaders’ power to nominate candidates facilitate strategic harmonization. • Explanatory proposition 15b: Incumbency advantages decrease party leaders’ capacity for strategic harmonization, thereby wholly or partially offsetting the impact of the institutional rules in 15a. • Explanatory proposition 16: Campaign funding systems distributing funds at the party level (or fraction level) are more conducive to strategic harmonization by partisan elites. • Explanatory proposition 17:  Privatization and state reform reduce the stock of material goods available to parties for the pursuit of strategic harmonization. Drawing on available large-N data, as well as on secondary sources on four shadow cases, Chapter 7 develops a series of preliminary tests of key descriptive and causal claims advanced in this work.

7 Plausibility and Scope: Out-of-Sample Tests This chapter seeks to test the external validity of the causal propositions inferred from the two case studies developed in previous chapters. To do so, I  draw on two methodological strategies. First, I  pursue a series of large-N analyses, drawing on the results from the comparative dataset of the Democratic Accountability and Linkages Project (2010); hereafter, the expert survey.1 However, if the causes shaping party–voter linkage strategies are in fact equifinal and multifinal, as suggested in Chapter 6, then attributing precise causes for observed outcomes on the basis of standard, large-N comparative analysis is bound to be very difficult. A  comprehensive, large-N test of the full argument is difficult to specify for a large sample given the data currently available. For that reason, this chapter only provides large-N tests for some general, testable implications of my argument taken individually. For the sake of consistency between the theoretical model and empirical exploration, I moreover draw more heavily on set theory for assessing causal sufficiency and necessity than on more standard, linear statistical models (see Goertz and Mahoney 2012). Second, I present four shadow cases, describing the electoral strategies of political parties recently obtaining consistent pluralities: the PT in Brazil, the MAS in Bolivia, the ARENA in El Salvador, and the BJP in India. These parties all operate in very unequal societies, but the contexts in which they compete otherwise differ in terms of levels of socioeconomic development, the pace and nature of market reforms, their electoral systems, levels of federalism and presidentialism, and the presence or absence of salient political divides other than distributive ones (e.g., ethnic and religious divisions). The parties analyzed also differ in terms of their longevity and trajectory in the party system and their ideological leanings. Whereas the case of the PT resembles that of the FA, and that of ARENA shares important similarities to the case of the UDI, the structural and institutional conditions within which these pairs of parties compete differ significantly. Moreover, both the BJP and MAS differ

 

1 

See .

282

Segmented Representation

quite substantially from either the FA or UDI. In addition to permitting testing specific propositions identified in Chapter 6, the four shadow cases also support a crucial causal claim made in the book:  in unequal societies, parties can prosper by simultaneously segmenting and strategically harmonizing diversified appeals to socially distinct constituencies. This book aims above all to improve descriptive inference, an analytic task too often neglected. Before analyzing the external validity of my explanatory propositions, I  therefore test two fundamental descriptive claims advanced in Chapter 1. First, parties most often seek to improve their electoral returns by segmenting their linkage strategies. Second, exclusive pursuit of programmatic or other single types of linkages is the exception, not the rule, in contemporary party politics. Using data from the expert survey, I  test both of these propositions. The expert survey was of course not designed for testing my propositions, but several indirect measures relevant for doing so can be derived from it. For some elements of my framework, including patterns of territorial segmentation, degree of strategic harmonization and its determinants, and candidate versus party-based linkage strategies, the expert survey does not include satisfactory proxies. Other elements, including parties’ nominating and funding procedures, organization of salient interest groups, and parties’ relations to interest groups, do not readily lend themselves to comparative operationalizing across a large number of cases. My use of the expert survey data therefore does not extend to assessing claims involving those dimensions. Instead, I use the survey’s large-N results to explore the plausibility of the two descriptive propositions and seven of the seventeen explanatory propositions introduced in Chapter 6, testing each argument at the systemic level. I then use the four out-of-sample shadow cases to test implications of the arguments at the individual party level.

O P E R AT IO NA L I Z AT IO N :   I D E N T I F Y I N G N O N - SE G M E N T E D, SE G M E N T E D A N D HA R M O N I Z E D, A N D SE G M E N T E D A N D N O N - HA R M O N I Z E D L I N KAG E S T R AT E G I E S Five questions included in the expert survey, already introduced in Chapter 3, were used to operationalize the dependent variables explored in this chapter. Those questions asked country experts to rank the extent to which relevant parties sought to mobilize support by engaging in five types of strategies: “featuring a party leader’s charismatic personality,” “emphasizing the attractiveness



Plausibility Tests: Out-of-Sample Cases

283

of the party’s positions on policy issues,” “emphasizing the capacity of the party to deliver targeted benefits to its electoral supporters,” “drawing on and appealing to voter’s long-term partisan loyalty” (party identification), and “emphasizing their general competence to govern and bring about or maintain economic, social and political stability.” On each item, experts were asked to assess parties’ reliance on this set of strategies on a four-point scale ranging from 1 (“not at all”) to 4 (“to a great extent”). The country database produced on the basis of the survey includes an entry for each system’s weighted average on each item. Weighting is based on parties’ respective showing in the last two congressional elections before 2008.2 Quite obviously, these five measures provide proxies for only some of the linkage strategies identified in Chapter  1. Moreover, the last item (valence) could conflate several types of linkages. It could apply to either individual candidates or parties, and it is debatable whether or not it should be considered a “soft” form of programmatic linkage. I therefore decided to use only the first four items, except when examining my shadow case studies. On this basis, I  computed each case’s membership in four different conceptual sets:  the set of cases in which programmatic mobilization is high, the set of cases in which leadership-based mobilization is high, the set of cases in which clientelistic mobilization is high, and the set of cases in which partisan-identification mobilization is high. Together, these four sets could be considered as linkage-type sets. Membership in each of these four sets was calibrated. To calibrate each variable I relied on the direct method presented in Ragin (2008).3 The direct method relies on the researcher’s identification of three anchors:  a minimum threshold for set membership (below which cases are definitely considered not to be members of the set), a crossover point (above which cases are more likely than not to be members of the set), and a maximum threshold (above which cases are definitely considered set members). Although specific criteria could be used to calibrate each variable, for the sake of transparency I decided to set these variable’s anchors using three percentiles: minimum threshold = 25 percent, crossover point = 75 percent, and maximum threshold = 90 percent. My criteria for including cases in any given linkage set was thus relatively stringent, to make my operationalization more consistent with the literature’s assumption of significant trade-offs among alternative linkage strategies. Making it more difficult for a case to obtain a high membership score in any

  2  I thank Herbert Kitschelt for granting access to this dataset at an early stage of the research project. I also thank Yi-Ting Wang for providing assistance in working with the database.   3  The calibration was pursued in Stata 12 with the fuzzy.ado module.

284

Segmented Representation

given linkage strategy set makes it less likely for cases to qualify for membership in more than one set, and thus sets the bar higher when it comes to confirming one of the book’s central theoretical premises: that parties usually rely on several different linkage strategies.4 I was, however, ultimately more interested in whether parties pursued single or multiple types of linkages than on which types of linkages they pursued. I therefore then used the four linkage sets I had created to derive three overall linkage strategy sets: non-segmented (hereafter, NS), segmented and non-harmonized (hereafter, SnH), and segmented and harmonized (hereafter, SH). The online appendix to this chapter explains in detail the operationalization procedure leading to identification of those three linkage strategy sets. Table 7.1 summarizes the resulting classification of cases. The NS set includes as positive instances cases in which only one linkage strategy predominates. The SH set includes cases with both linkage segmentation and likely high levels of harmonization. These are cases with parties regarding whose strategies experts concur in their judgments. The SnH set, in turn, is comprised of cases where there is a lack of consensus among experts as to what parties’ predominant strategies are, suggesting that the party system is apt to be one in which parties’ strategies are segmented but non-harmonized. I justify this claim based on the following premise: these are cases in which country experts could not consensually identify a linkage or set of linkages as the predominant one(s) in the party system they observed. Such a result is not logically inconsistent with a scenario in which parties pursue several linkage strategies in a segmented way, without being able to strategically harmonize those mobilization attempts into a well-crystallized party strategy. Such a scenario could present experts with a confusing panorama, in which the relative importance or “priority” of different linkage types is unclear because parties themselves have failed to establish any clear ordering or overall organization of multiple, segmented but non-harmonized and potentially conflicting types of linkages. There are, however, also some plausible alternatives to this interpretation. As discussed in the online appendix, precisely identifying SnH strategies is particularly difficult. Three alternative operationalizations (broad, intermediate, and restrictive) for SnH strategies are therefore presented in the remainder of this chapter. These operationalizations differ in the thresholds imposed for including cases in the SnH set.

  4  To be sure, the standard measures (weighted averages on a 1–4 scale) are fairly highly correlated to my fuzzy-set or calibrated operationalizations. The lowest correlation is 0.77 (sig < 0.000; obs = 88) for the original and calibrated measures of party-identification.

Table 7.1.  Operationalization of fuzzy sets representing different segmentation and harmonization scenarios

Criteria for set membership = 1 Criteria for set membership = 0.5

NS (not segmented) Typical member: Sweden, Zambia

SnH broad (segmented, non-harmonized) Typical member: Chile

SH (segmented & harmonized) Typical member: Uruguay

One and only one linkage set displaying set membership scores greater than 0.75 (membership scores in other linkage sets should be less than 0.5) One and only one linkage set displaying set membership scores lower than 0.75 but greater than 0.5 (membership scores in other linkage sets should be less than 0.5)

Membership scores are lower than 0.5 in all of the four linkage sets

At least two linkage sets displaying set membership scores greater than 0.75

Subset 1

OP (only programmatic): the linkage type displaying a 0.75 or greater membership score is the programmatic set Typical member: Sweden

Subset 2

OnP (only non-programmatic): the linkage type displaying a 0.75 or greater membership score is other than the programmatic set (i.e., clientelistic, leadership, or PID)Typical member: Zambia

Sets to be used in subsequent analyses

NS, OP, OnP

More than one set membership At least one linkage set displaying set greater than 0.4, but lower than 0.5 membership scores greater than 0.75, and at least an additional one displaying membership scores greater than 0.5 (yet lower than 0.75) SnH intermediate (cases are excluded None from the set if they display membership scores lower than 0.5 in the high inequality set and membership scores lower than 0.5 but greater than 0.4 in at least one linkage type set) SnH restrictive (cases are excluded None from the set if they display membership scores lower than 0.5 in the high inequality set or membership scores lower than 0.5 but greater than 0.4 in at least one linkage type set) SnH Broad, SnH Intermediate, SnH SH Restrictive

286

Segmented Representation

An assumption central to political economy models of the impact of inequality on redistribution is that linkages are non-segmented and programmatic. Two subsets within the NS set are therefore identified:  those with only programmatic (OP) linkages and those whose only linkage type is non-programmatic (OnP). The latter group is comprised of cases in which only leadership, clientelistic, or PID linkage strategies are observed. Table 7.1 summarizes the characteristics of each of the sets analyzed in the remainder of the chapter. To facilitate appraising the classifications obtained using these criteria, Table 7.2 lists the cases with full (= 1) and partial (= 0.5) membership in each of the sets. For cases with NS linkages, I identify in brackets the predominant mobilization strategy (P  =  programmatic, L  =  leadership, C  =  clientelistic, PID = party identification). A full listing of set membership scores for each case, in each of the five linkage-based fuzzy sets (i.e., including membership in the set of cases with valence mobilization) is available in the online appendix. The shadow cases presented later in this chapter vary in terms of the sets to which they belong. Two of the four are borderline cases when it comes to their membership in one or more sets. With membership scores below 0.5 in Table 7.2.  Party-system’s membership into four fuzzy sets NS

SnH

SH

Austria OP

Argentina

Australia OP Belgium OP Costa Rica OP Czech Republic OP Denmark OP Germany OP Netherlands OP New Zeeland OP

Bolivia Brazil Bulgaria (mc) Canada (c, mc) Chile Croatia (c, mc) Estonia (c, mc) El Salvador (partial 0.5 membership) Finland (c, mc) France (mc) Greece (c, mc) Indonesia (mc) Italy (c, mc) Japan (mc) Kenya South Korea (c, mc) Latvia Lithuania (mc)

Dominican Republic (0.5 membership) Ecuador Ghana Guatemala (0.5 membership) Hungary Jamaica Lebanon Macedonia Mauritius

Norway OP Slovenia OP Sweden OP Switzerland OP United States OP Spain (0.5 membership) OP Israel (0.5 membership) OP Ireland (0.5 membership) OP Paraguay OnP: C Angola OnP: C

Namibia Nicaragua (0.5 membership) Niger (0.5 membership) Panama Philipines Senegal South Africa Tanzania Uruguay Venezuela (Continued)



Plausibility Tests: Out-of-Sample Cases

287

Table 7.2.  (Continued) NS

SnH

Colombia (0.5 membership) OnP: C Egypt OnP: C Mongolia (0.5 membership) OnP: C Zambia OnP: C Thailand OnP: C Albania OnP: L Benin OnP: L Georgia OnP: L Peru (0.5 membership) OnP: L Portugal OnP: L Taiwan (0.5 membership) OnP: L Turkey OnP: L Ukraine (0.5 membership) OnP: L India (0.5 membership) OnP: L Botswana OnP: L Mozambique OnP: L

Malaysia (mc)

Full set 31 cases (35 %), Intermediate Version set 21 cases (24 %), Restrictive Version set 12 cases (14 %) Full membership 29 cases (33 %) Full membership 15 cases (17 %)

SH

Mali Mexico Moldova (c, mc) Morocco Nigeria Pakistan Poland (mc) Romania Russia (mc) Slovakia (c, mc) United Kingdom (c, mc)

Serbia and Montenegro OnP: L Bangladesh OnP: L Honduras OnP: L 38 cases (43 %)OP = 17 cases (19 %)OnP = 21 cases (24 %) 19 cases (21 %)

Full membership 31 cases (34 %)

Source: Author’s own construction on the basis of Democratic Accountability and Linkages Project. CODA:  P “programmatic linkage,” C “clientelistic linkage,” L “leadership-based linkage,” c “case excluded from conservative version of the set,” mc “case excluded from the most conservative version of the set,” c, mc “case excluded from both the conservative and the most conservative version of the set.”

all linkage sets, El Salvador is classified as a case in which SnH strategies predominate. Yet, for three linkage-type sets (leadership, clientelistic, and PID), the case displays membership likelihoods equal to or greater than 0.4. Under a slightly more lenient operationalization than the one applied here, El Salvador could thus be considered a member of the SH set.

Segmented Representation

288

Table 7.3.  Linkage and strategy classification of six case studies Case

Leadership set

Programmatic set

Clientelistic set

PID set

Overall strategy set SnH SnH SnH SnH/borderline SH Partial NS, OnP (leadership)/ borderline SH SH

Bolivia Brazil Chile El Salvador India

0.22 0.001 0.31 0.44 0.52

0.11 0.09 0.11 0.13 0.12

0.16 0.08 0.05 0.45 0.37

0.03 0.000 0.004 0.4 0.2

Uruguay

0.06

0.86

0.001

0.75

Source: Author’s own construction on the basis of Democratic Accountability and Linkages Project.

India, in turn, is classified as a partial member of the NS set (and OnP subset). Its likelihood of membership in the leadership set was 0.52, and it did not surpass the crossover point for any of the other three linkage sets. Yet, the case barely passes the threshold for membership in the NS set and, had that threshold been set just slightly higher, it would instead have been categorized as another SnH case. Bolivia and Brazil are, in turn, typical members of the SnH set. They do not have high membership scores in any of the four linkage strategy sets considered. Returning to the book’s main cases, Chile is classified as SnH. It has low membership likelihoods for all sets, with the marginal exception of leadership, for which it obtains a set membership score of 0.31. Meanwhile, Uruguay is a member of the SH set, due to its high probability of membership in two linkage type sets: programmatic (0.86) and PID (0.75). Table 7.3 provides a detailed characterization of these six cases.

Strategy for Operationalizing Other Variables Two of the propositions inferred from my analysis of Chile and Uruguay in Chapter  6, and tested in this chapter (propositions 9 and 10), relate to whether parties pursue candidate-based or party-based appeals. To construct a proxy for the extent to which parties rely on candidate-based appeals, I used three of the five measures of linkage types included in the expert survey: the extent to which parties compete by emphasizing leaders’ charisma, programmatic linkages, and material exchanges with voters (i.e., clientelism). I treated personality-based linkages as epitomizing candidate-based appeals. Programmatic and clientelistic appeals, by contrast, are more clearly



Plausibility Tests: Out-of-Sample Cases

289

structured by parties.5 Using systems’ scores on these measures, I derived two possible proxies. The first proxy, which I call “candidate-based appeals,” was obtained by dividing each case’s score on the candidate-based measure by the sum of its scores on the programmatic and clientelistic measures. On this basis, I computed each case’s fuzzy-set membership likelihood measure, once again drawing on the direct calibration measure and three anchors set at 0.25, 0.75, and 0.90. The second proxy is derived from each case’s membership score in the set of cases with high leadership-based mobilization. The two proxies were significantly correlated at 0.5 (sig  0.5.

linkages. Moreover, the recent rise of programmatic mobilization and polarization might have biased experts’ overall assessments of the nature of linkages observed in the US party system, which has been traditionally described as one in which not only issue positions and programmatic platforms but also PID, candidate traits, constituency service, and pork have played a significant role in shaping voting behavior (see e.g., Miller et al. 1960; Mayhew 1974; Fenno 1978; Miller and Shanks 1996; Evans 2004). Finally, Costa Rica, classified by country experts as one with OP linkages, is a member of neither the low inequality nor the high GDP per capita sets. This case suggests that neither of these causal conditions is logically necessary for a case to develop OP linkages. That said, the recent evolution of the Costa Rican party system may call into question the sustainability of an OP linkage structure in the context of a relatively unequal democracy with low GDP (Lehoucq 2005; Vargas Cullell 2007). In sum, the evidence and configuration described in Figure 7.1 suggest that high GDP and low inequality are neither individually nor jointly sufficient causes of non-segmented, programmatic linkages. Yet, especially while taken together, the presence of both conditions creates an unambiguously favorable context for OP linkages. The two conditions together thus represent one possible (and usual) path to OP linkages.



Plausibility Tests: Out-of-Sample Cases

295

My analysis also suggests that OP linkage structures are observed largely in a very specific subset of cases (Western European countries), while they are very scarce outside that region. This finding thus suggests that usual assumptions regarding the presence of trade-offs between different linkage strategies and the preeminence of programmatic mobilization over other alternatives apply only within a restricted scope. To more fully assess proposition 2, Graph 7.3 provides complementary evidence regarding the distribution of programmatic linkages across cases, irrespective of the specific patterns of segmentation observed. The graph presents three scatter plots, displaying the relationship between membership in the high GDP per capita set, the high inequality set, and an alternative low inequality set based on the income share of the bottom two income deciles in each case. The three scatter plots show significant correlations, with magnitudes ranging from 0.25 (for income share of the bottom 20 percent) to 0.58 (for GDP per capita). Similar results were obtained when a citizen-based measure (described in the online appendix) was substituted for the expert survey measure of whether programmatic linkages were present. Using the citizen-based measure, the correlations were −0.49 (sig=0.001) for Gini inequality, 0.33 (sig=0.01) for income share of the bottom 20 percent, and 0.41 (sig=0.001) for GDP per capita. Interestingly, based on the citizen-based measure, inequality appears to trump GDP per capita as a significant predictor of the presence of programmatic linkages if a multiple regression model is specified. Yet, when interpreted from a fuzzy-set perspective, the scatter plots in Graph 7.3 also reveal a distribution of cases that is close to a “triangular” pattern. As described in Ragin (2008), such patterns reveal important information regarding causal necessity and sufficiency. In the case of GDP per capita, most cases lie in a right-hand-side triangle, with Macedonia, Tanzania, South Africa, Costa Rica, and Uruguay as the cases deviating most from it. A very similar pattern is obtained in the figure plotting the presence of programmatic linkages against the fuzzy-set distribution of income share of the poorest 20 percent. Due to the negative relationship observed between inequality (as measured by the Gini coefficient) and the presence of programmatic linkages, in the third scatter plot most cases fall within a left-hand-side triangle. Outliers in this case are found in the top-right corner. Those outliers (i.e., South Africa, the US, Costa Rica, Uruguay, and Tanzania) suggest that in a small subset of cases programmatic politics mattered, despite these being among the cases with the poorest distribution of income to the lowest 20 percent of the population. The presence of cases outside each triangle reflects a lack of strict causal necessity or sufficiency. Yet, the observed patterns reveal that low inequality, as well as high socioeconomic development, behave as almost necessary

Graph 7.3.  GDP per capita, inequality, and programmatic linkages Source: Author’s own construction on the basis of Democratic Accountability and Linkages Project and World Development Indicators (World Bank).



Plausibility Tests: Out-of-Sample Cases

297

conditions for programmatic linkages. Cases with high membership scores in the independent variable, but low scores in the programmatic linkages fuzzy set, in turn reveal that these causal conditions are not sufficient to cause the outcome. Large-N statistical testing also supports proposition 4:  that more extensive public good provision provides more opportunities for crafting more extensive and less socioeconomically segmented programmatic linkages. For testing, I  use three indicators for public goods provision in society:  public social expenditures in 1998, public education expenditures in 2001, and public health expenditures in 2001. Controlling for the effects of GDP per capita, each is quantified in terms of percentage of GDP. The education and health expenditure measures are available for more cases (34 and 45 respectively). Social policy provision of course does not exhaust the possible range of public goods provision in society. However, an important advantage of using a social policy expenditure measure is that it is available for a large set of cases on the basis of comparative estimates. It therefore provides a relatively satisfactory proxy for my independent variable. For the dependent variable, I rely once again on the set membership measure related to pursuit of programmatic linkages in a given party system. The three measures for the independent variable are significantly correlated to the expert survey measure on the pursuit of programmatic linkages (social expenditures = 0.48 sig=0.05, education expenditures = 0.34 sig=0.05, and health expenditures  =  0.65 sig=0.001). Moreover, the citizen-based measure, for which the number of cases available is significantly lower, is also positively correlated to the three expenditure measures, as well as significantly correlated to health expenditures (r = 0.37 sig=0.05. Graph 7.4 displays the relationship between health expenditures and the expert survey-based estimate on pursuit of programmatic linkages. The distribution of cases in the scatter plot reveals a “triangular” pattern (Ragin 2008) or a “perfect predictor” (Goertz and Mahoney 2012). Accordingly, public goods provision in society, at least as operationalized here in terms of health expenditures over GDP, behaves as a close to necessary (although insufficient) cause for high levels of programmatic linking in a given society. This evidence is consistent with proposition 4. Chapter  6 also identifies institutional variables as potentially causing, or at least reinforcing, particular types of linkages and segmentation structures. According to proposition 9, more personalized electoral formulas promote the pursuit of candidate-based linkage strategies. And according to proposition 10, the interaction between small district sizes and high socioeconomic inequality favors dual or “inter-district” types of linkage segmentation. Drawing once again in the expert survey results allows for preliminary testing

298

Segmented Representation

Graph 7.4.  Health expenditures as % of GDP and programmatic linkages Source: Author’s own construction on the basis of Democratic Accountability and Linkages Project and World Development Indicators (World Bank).

of these institutional propositions. Because no direct measures for either candidate-centered appeals or patterns of territorial segmentation are available, these tests of propositions 9 and 10 are tentative. Specific territorial patterns of segmentation cannot be measured directly at the cross-national level with the information at hand. I therefore instead only tested whether different institutional incentives had elective affinities with different levels of linkage segmentation. If more segmentation is observed when certain identifiable institutional incentives are present, this correlation at least suggests that institutional rules might play a significant role in shaping different segmentation structures. I draw on the Democracy Cross-National Data 3.0 (Norris 2009) to operationalize relevant independent variables. The independent variables in this data set include a range of relevant institutional characteristics, such as district size and magnitude, type of electoral system,6 use of open or blocked electoral lists, and polities’ federal versus unitary configuration. I also included a   6  Dummy variables are used for proportional representation, majoritarian, and mixed electoral formulas.



Plausibility Tests: Out-of-Sample Cases

299

Table 7.5.  Observed linear correlations between institutional incentives, candidate-based linkage types, and linkage segmentation outcomes Membership in set with. . . High candidate-centered appeals High incidence of leadership appeals High district size High district magnitude Low district magnitude*high inequality Open list PR electoral system Majoritarian electoral system Mixed electoral system Unitary system Sub-national expenditures Administrative Decentralization Political Decentralization

Candidate-based Leadership appeals mobilization

NS

0.5*** 0.5***

NS-OP

NS-OnP

−0.09

0.28***

−0.33***

0.27***

0.26 −0.01 0.08

0.14 0.11 0.15

−0.16 0.12 −0.1

−0.21 0.03 −0.27

−0.01 0.08 0.22

0.05 0.02 0.00 −0.02 0.10 −0.19 −0.08

−0.21 −0.06 −0.07 0.13 0.09 −0.32* −0.14

0.03 0.15 0.03 −0.2 −0.17 0.27 −0.28

0.10 0.1 0.02 −0.13 −0.16 0.21 −0.14

−0.21 −0.2 0.05 −0.18 −0.03 0.03 −0.03

−0.32

−0.29

0.09

0.08

−0.26

Source:  Author’s own construction on the basis of Democratic Accountability and Linkages Project and Norris (2009).

series of measures covering administrative, political, and fiscal dimensions of decentralization. Table 7.5 summarizes the relevant linear correlations obtained. Regarding proposition 9, only one of the institutional variables significantly correlates (negatively) to the pursuit of partisan leadership-based mobilization:  level of sub-national expenditures, an indicator of fiscal decentralization. More fiscally decentralized polities are thus less likely to provide incentives for leadership-based mobilization. A possible mechanism explaining this negative correlation is that leaders with access to centralized resources can more readily develop that type of linkage. Graph 7.5 once again reveals a triangular distribution of data-points. Only Turkey, and to a much lesser extent Peru and India, massively fall outside the triangle, displaying high decentralization and strong leadership-based mobilization. The scatter plot suggests that low levels of fiscal decentralization are not sufficient to cause leadership-based appeals; yet, such appeals very rarely occur where spending is not highly centralized. In general, though, this evidence suggests that institutional variables do not play a strong causal role in determining the types of linkages pursued and the

300

Segmented Representation

Graph 7.5.  Levels of fiscal decentralization and leadership-based appeals Source: Author’s own construction on the basis of Democratic Accountability and Linkages Project and Norris (2009).

scope of segmentation of such linkages. This result is consistent with those reported by Kitschelt et al. (2010) and Kitschelt et al. (n.d.). S T R AT E G IC HA R M O N I Z AT IO N In this section I provide preliminary tests for two propositions regarding strategic harmonization: that smaller district magnitudes decrease parties’ capacity for strategic harmonization (proposition 13), and that levels of strategic harmonization are lower in more decentralized polities (proposition 14). Here again, no direct measures of the dependent variable are available at the cross-national level, and those available through the process of operationalization proposed here are very tentative. For testing propositions 13 and 14, I draw on the four fuzzy sets described above, which were created on the basis of responses to the expert survey. For the independent variables, I rely on the same indicators used above for testing the effects of electoral institutions and different decentralization arrangements on harmonization. Table 7.6 presents the results regarding potential causes of harmonization. In general variables related to electoral institutions do not seem to influence harmonization outcomes, although some did correlate to segmentation. Yet,



Plausibility Tests: Out-of-Sample Cases

301

Table 7.6.  Observed linear correlations between linkage types, institutional incentives, and linkage harmonization outcomes. Membership in set with: High district size High district magnitude Low district magnitude/high inequality PR electoral system Majoritarian electoral system Mixed electoral system Unitary system Open list Sub-national expenditures Administrative decentralization Political decentralization

SH

SnH

SnH SnH (intermediate) (restrictive)

−0.19 −0.35 0.07

0.19 0.43 −0.15

0.26 0.21 −0.12

0.06 −0.19 0.065 0.19 −0.08 −0.44* −0.30 −0.1

−0.13 0.21 0.0 −0.005 0.16 .51** 0.17 0.09

−0.12 −0.14 0.03 0.13 −0.05 0.02 0.47*** 0.19

0.14 0.22 0.03 −0.03 0.06 −0.17 0.06 0.00 −0.07 0.31 0.00

Source: Author’s own construction on the basis of data from the Democratic Accountability and Linkages Project and Norris (2009).

though not high enough to be of statistically significance (sig. = 0.06, n = 19), the information in the table does suggest that greater district magnitudes might eventually contribute to diminishing harmonization. The presence of multiple candidates from the same party pursuing segmented linkage strategies within a single district might eventually create more harmonization problems than when a single candidate represents the party in the district and relates more directly to the party’s “commanding heights.” In sum, although it cannot be discarded on the basis of this test, proposition 13 is not strongly supported by the data. My evidence lends stronger support to proposition 14. Greater levels of fiscal decentralization are negatively correlated to the presence of harmonized linkage strategies. Moreover, administrative decentralization is positively and significantly correlated to absence of strategic harmonization. Although not statistically significant, the measures of other dimensions of decentralization are consistently signed. These findings are consistent with recent evidence on the negative externalities that decentralization can create for the survival of encompassing party organizations (Tanaka 2005; Harbers 2010; Dargent and Muñoz 2011; Avellaneda and Escobar-Lemnon 2012). Tables  7.7 and 7.8 explore the nature of the possible causal relationship between decentralization and strategic harmonization from the point of view of causal sufficiency and/or necessity. Given that the dependent variable is dichotomous, I decided to dichotomize the independent variable using a 0.5

302

Segmented Representation Table 7.7.  Fiscal decentralization as a necessary/ sufficient cause of segmented and harmonized linkage strategies? SH? Fiscally decentralized? No Yes Total

No 10 10 20

Yes 5 Empty 5

Source: Author’s own construction on the basis of data from the Democratic Accountability and Linkages Project and Norris (2009).

Table 7.8.  Fiscal decentralization as a necessary/ sufficient cause of segmented and non-harmonized linkage strategies? SH? Fiscally decentralized? No Yes Total

No 6 Empty 6

Yes 9 10 19

Source: Author’s own construction on the basis of data from the Democratic Accountability and Linkages Project (2010) and Norris (2009).

threshold. The resulting 2 × 2 contingency tables are functionally equivalent to the scatterplots presented above for continuous variables. As in the rest of the section, the analysis was restricted to cases with segmentation. Information on levels of fiscal decentralization is available for twenty-five of those cases. Both contingency tables display the presence of association, with chi square statistics significant at the 0.05 level. Yet, from the point of view of sufficiency and necessity, the distribution of cases across cells is of particular interest. In each table, one of the four cells is empty, with the location of the empty cell depending on the sign of the association observed. The significance of this finding is equivalent to that of a triangular scatterplot for continuous variables. As seen in Table 7.7, cases that are fiscally decentralized and also present segmented and harmonized linkage strategies do not exist. Likewise, as seen in Table 7.8, cases that are fiscally decentralized and escape a pattern of segmentation and non-harmonization are also non-existent. In sum, in fiscally decentralized polities parties that pursue segmented linkage strategies face strong constraints when seeking harmonization. In other words, high levels of fiscal decentralization might act as a sufficient condition for observing lower levels of harmonization of segmented linkage strategies.



Plausibility Tests: Out-of-Sample Cases

303

I N D I V I DUA L LY SU C C E S SF U L PA RT I E S I N H IG H LY U N E Q UA L S O C I E T I E S The cases of the BJP in India, the MAS in Bolivia, the PT in Brazil, and ARENA in El Salvador constitute instances of recently successful parties competing in highly unequal societies. In the 1990s and 2000s, each of these parties was the most electorally successful within its respective party system. Both ARENA (since 2009) and the BJP (since 2004) have witnessed a decline in their electoral returns. Yet, from the early 1990s to 2009 and 2004 respectively, each led the government for a sustained period of time. In El Salvador, ARENA also garnered a plurality in the 2012 congressional elections (obtaining thirty-three seats). These four cases, like those of the FA and the UDI, also constitute instances of externally mobilized parties of relatively recent creation. Whereas ARENA and the MAS rapidly ascended to a dominant position in their respective party systems, the PT and BJP are examples of more long-term party building from the opposition. Before one can analyze the configurations of linkages in these cases and the relation of those configurations to other factors, one first needs to understand their respective political trajectories. The Bharatiya Janata Party (BJP) The BJP’s origins go back to the founding of the Jana Sangh in 1951. In 1977 and 1980, the BJP competed electorally as a member of the Janata Party’s coalition, which in 1977, in the aftermath of the State of Emergency, was the first to defeat the dominant Congress Party. Yet, until 1989 the long history of the BJP in the Indian party system did not translate into an impressive electoral showing. Having obtained only two seats in the Lok Sabha (the Indian parliament) in the 1984–1985 elections, in 1989 the BJP was able to claim eighty-five seats. Subsequently, in 1996, it became the most-voted-for party, obtaining 161 seats in the chamber. In 1998, it further climbed to 182 seats.7 The party lost its majority in Congress in the election of 2004, when it fell to 138 seats and was replaced by the Indian National Congress Party (with 145 seats) as leader of the majority coalition. The case of the BJP is interesting not only because of its recent electoral success, but also due to the particular ideological and ethnic background of the party. Considered the party best representing India’s two major upper-caste  

7 

Election Commission of India.

304

Segmented Representation

groups, analysts of the Indian party system predicted that it would never be able to appeal to lower castes and poor voters (Graham 1990). The BJP’s adherence to Brahminical ideology and Hindu nationalism was key to explaining why the party could not be expected to attract the vote of backward castes, or those most disfavored by the traditional caste system. The party’s direct involvement in the 1992 demolition of the Babri Masjid in Ayodhya (Uttar Pradesh), which led to more than 2,000 deaths in the riots following the demolition, further contributed to the consolidation of the BJP’s radical Hindu-nationalist positions. When the catch-all Congress Party strategy began to falter, and ethnic and religious mobilization crystallized as the emergent cleavage in Indian society (see Chhibber 2001), the BJP therefore was not in a comfortable position for electorally mobilizing the “scheduled” or backward castes (Brass 1993). Indeed, the BJP’s agitation tactics in the late 1980s and early 1990s proved electorally inefficient for mobilizing poor voters (Thachil 2011). Parties such as the Bahujan Samaj Party (BSP) or the Janata Dal (JD), by contrast, enjoyed competitive advantages when it came to mobilizing “backward castes” on the basis of ethnic/religious identities (Chandra 2007). The BJP’s affinity with the upper castes also became evident in the wake of the 1990 Mandal Commission Report, which promoted affirmative action to improve access of scheduled castes to government posts. Although party alignments varied widely across Indian states, the BJP reacted against the implementation of the report’s affirmative action recommendations, and thus in favor of upper castes’ socially privileged positions. In contrast, the BSP openly favored such new affirmative action measures and campaigned on their contribution to “social justice” in favor of the most backward castes (Chandra 2007, p. 205). According to Thachil (2011), both BJP elites and the party’s organization were extremely reluctant to open slots for lower-caste candidates on the party’s lists, and subsequently refrained from supporting affirmative action proposals to avoid internal disputes (Thachil 2011, p. 440). An electoral strategy based on the kind of “ethnic head-count” encouraged by the BSP (Chandra 2007)  was therefore much less feasible for the BJP. Mobilizing poor voters on the basis of class, as the Indian Communist Party had done in states such as Kerala and West Bengal (Kohli 1991), was also unfeasible for the BJP, as it would have alienated the party’s core constituency. Nor could the BJP hope to expand its support from India’s lower classes by campaigning on state–market issues (Chhibber 2001). The BJP favored the pursuit of market reforms, and had traditionally opposed the Congress Party’s political system. That system was structured around pyramidal patronage and clientelism networks through which national elites co-opt state-level elites,



Plausibility Tests: Out-of-Sample Cases

305

which in turn had access to extensive patronage resources for mobilizing voters electorally at the local level (Brass 1965; Weiner 1967; Frankel 1978; Kohli 1991). Although the BJP’s pro-market position could attract middle-class and business support, it also reinforced the party’s upper-class profile. Finally, as an externally mobilized party, the BJP did not, at least at the national level, enjoy the possibility of using the spoils of office electorally. In sum, the BJP was unable to mobilize backward castes and poor voters on the basis of programmatic or ethnic/religious-based appeals, and was also unable to access central state resources for clientelistic mobilization of poor voters. In line with the BJP’s ideological position, in the elections of 1993 and 1996 its core constituency was a coalition of Hindi nationalists and pro-market (middle-class) voters (Chhibber 2001, pp.  172–5). The party’s strong showing among Hindus and forward castes was consistent with the BJP’s historical (1967) electoral base. Hindu nationalism and religious attitudes were correspondingly prominent within the party’s activist base (Chhibber 2001, pp.  162–4). Yet, according to post-election survey results, in both 1993 and 1996 new BJP voters came disproportionately from among educated, white-collar, private sector workers. In sum, the BJP consolidated its programmatic positioning in the system, distinctively combining religious nationalism with a pro-market agenda. This gave the party a competitive advantage against the Congress Party, whose statist policies and secularism became progressively less efficient in mobilizing large cross-sections of the population across Indian states (Chhibber 2001). Chibber’s account of the rise of the BJP constitutes a persuasive analysis of the party’s evolving relation to its core constituency. Notwithstanding, such an account cannot explain the BJP’s most recent inroads within India’s poorest electoral constituencies, in which, against all odds, the BJP has proved increasingly able to succeed. In the elections of 2004, the BJP in seven of India’s major states obtained more than 30  percent of the vote from India’s poorest groups: Dalits (the “untouchable” castes) and Adivasis (indigenous tribal groups) (Thachil 2011, p. 438). According to Thachil (2011), the BJP developed a specific electoral strategy for penetrating its peripheral constituency. That strategy was based on provision of social services (education and health) in poor communities, implemented by a network of religious activists affiliated to the BJP. This strategy was particularly efficient in poor states, where public infrastructure was inadequate and state-provided social services were scattered. In these contexts, services provided by religious activists benefited users more than state provision and were judged more efficient, even by non-religious users (Thachil 2011, p. 448). Working in social service provision and explicitly assuming a non-partisan and “apolitical” posture, members of the upper castes were able to gain access

306

Segmented Representation

to poor and distrustful communities (Thachil 2011, p.  448). The growth of the BJP’s affiliated network of “apolitical” service provision projects correlates to the party’s improving electoral performance during the 1990s and 2000s (Thachil 2011, p. 455). Such a network allowed the BJP to persuade poor voters to switch to supporting the party. Moreover, service organizations also provided the BJP with an efficient vehicle for selecting and recruiting to the party a small but influential cadre of local leaders, which could then run for office counting on their popularity in each locality (Thachil 2011, p. 450). According to the framework proposed in this book, the BJP thus pursued a highly segmented strategy, appealing to distinct socioeconomic segments and ethnic groups through markedly different linkage strategies. While the BJP appealed to its core constituency (upper castes, Hindu-nationalists, and pro-state middle classes) on the basis of programmatic appeals, the party progressively penetrated a peripheral constituency by engaging in the particularistic mobilization of backward caste voters by providing social services at the local level. The latter strategy was developed through a network of religious activists, who identified neither as party members nor as political agents. Service provision was privately financed, and did not rely on state-provided goods. As in the other cases I analyze below, the specific patterns of territorial segmentation used by the party are difficult to infer based on secondary sources. Yet available evidence, in particular Thachil’s (2011) district-level evidence, suggests that territorial segmentation is high in this case. High territorial segmentation is also likely given the rising importance of state-specific party systems on electoral outcomes at the national level that has developed since the era of Congress Party dominance (Chhibber 2001). The BJP also seems to have been able to achieve high levels of strategic harmonization, through centralized control by party elites, who have controlled key symbolic and material resources for funding and motivating the party’s “non-political” activists. Moreover, the party has also been able to grow electorally by pursuing electoral and coalitional alliances with state-level parties (Chhibber 2001). India’s federalism and parliamentarism have further provided a favorable institutional context within which to pursue such strategies for further electoral expansion without compromising individual parties’ strategic harmonization and unity. The Alianza Republicana Nacional (ARENA) The ARENA Party was founded in 1981 by Roberto d’Aubuisson, an officer of the Salvadoran intelligence unit with close ties to paramilitary organizations informally allied to the country’s authoritarian regime (Stanley 1996; Wood



Plausibility Tests: Out-of-Sample Cases

307

2000).8 ARENA originated as an explicitly anti-communist and anti-reformist coalition, in reaction to the guerrilla insurgency led by the Frente Farabundo Martí de Liberación Nacional (FMLN), but also in reaction to reformist policies by the military Junta that had governed the country since 1979. Particularly in its early stages, the party built its territorial apparatus by drawing on two pre-existing organizations engaged in paramilitary activities: the FA Nacionalista and ORDEN (Martín Baro 1982, p. 457; Chacón 1984, p. 19; Koivumaeki 2014; Loxton 2014). By the late 1970s, the economic elite had become disappointed by the inability of the traditional Conservative Party (Partido de Conciliación Nacional, PCN) to tackle the country’s economic and political crisis. Moreover, after the 1979 coup a reformist Junta promoted extensive land reform and the nationalization of the financial system (Chacón 1984, p. 20). In the wake of PCN’s failure, and in reaction to the leftist threat posed by the Junta and the insurgency led by the FMLN, the Salvadorian economic elite began to perceive ARENA as an efficient vehicle for protecting its political and economic interests (Koivumaeki forthcoming). Eventually, ARENA forged a coalition between landowning elites reacting against agrarian reform and military officers involved in counter-insurgent and repressive activities. In 1983, in coalition with the PCN, d’Aubuisson presided over the Constitutional Assembly in charge of drafting the Salvadoran constitution. Yet, in 1984 d’Aubuisson lost his presidential bid against José Napoleón Duarte of the Christian Democratic Party, which governed the country until 1989 through a sharp economic crisis and the last years of the civil war. Notwithstanding, ARENA’s rapid electoral rise in the elections of 1984 and 1985 were pivotal in its gaining decisive support from the country’s elite (Koivumaeki forthcoming). D’Aubuisson’s resignation of the party’s presidency in favor of Alfredo Cristiani, a well-known businessman, also played a key role in consolidating ARENA as the political representative of the business elite (Artiga 2002; Albiac 1999). As documented by Eguizábal Mendoza (1989), successful businessmen as well as the country’s traditional landed oligarchy became increasingly active in the party’s organization. In 1989, Cristiani defeated Fidel Chávez from the Christian Democratic Party, becoming ARENA’s first elected national president. From then until 1992, the country witnessed escalating violence, triggered by the military   8 The Comisión de la Verdad para El Salvador (the Salvadoran Truth Commission), established in 1992 by the Chapultepec agreements putting an end to the country’s 1980–1992 civil war, accused d’Aubuisson of being responsible for the death of Oscar Romero, a Catholic bishop involved in the protection of human rights. ARENA’s founder was also linked to organizing and running Salvadoran death squads, beginning in the 1970s.

308

Segmented Representation

offensive of the left-wing insurrectionary FMLN and the army’s heavy-handed strategy of counter-insurgency. ARENA leaders were linked to human rights violations among innocent civilians, as well as the murder of six priests in the Universidad Centro-Americana in the early days of Cristiani’s government Cristiani’s presidency was also associated with economic reform and privatization, further reinforcing ARENA’s pro-business orientation, as well as its linkages to such key elite organizations as the Asociación Nacional de la Empresa Privada (ANEP). In ARENA’s subsequent presidential terms, the ANEP was pivotal in providing the party with technocratic staff for filling key positions related to economics and finance, as well as in providing ARENA with links to international financial institutions (Koivumaeki 2009). In 1994, ARENA won the presidency with Armando Calderón, and in 1999 with Francisco Flores. In 2004, Antonio Saca was elected president and governed until 2009. In 2009, Rodrigo Ávila of ARENA was defeated by Mauricio Funes of the FMLN. Starting in the mid-1990s, ARENA’s programmatic platform combined a pro-market stance with an increasingly salient mano dura program to fight El Salvador’s increasing levels of criminality (Holland forthcoming). Opinion polls showed that insecurity had become the most important issue for El Salvador’s public (Córdova and Cruz 2007). Competing on the public security issue helped ARENA to offset anti-incumbency’s gradual erosion of its electoral performance (Holland forthcoming). A favorable political opportunity structure permitted ARENA to make these political gains despite its clearly pro-elite orientation. After 1989 the political center, represented by the Christian Democratic Party, weakened significantly, ultimately becoming inconsequential by the early 2000s. Moreover, Christian Democratic congressmembers, as well as those of the PCN, usually provided congressional support to the governments headed by ARENA, which contributed to obscuring the specificity of their party’s profile. Meanwhile, the FMLN insisted on a radical political agenda, which it did not moderate until the 2009 election. Until then, the FMLN nominated former guerrilla leaders as presidential nominees, facilitating ARENA’s anti-communist “scare” campaigns. That strategy was instrumental in winning support not only at the elite level, where the party already had its stronghold, but also among the rural poor, a segment formerly affected by political violence and leftist mobilization in the countryside during the 1970s and 1980s. Although its economic policies were detrimental to both the rural poor and peasants, ARENA was more effective in electorally mobilizing them than the urban informal sector using such tactics (Wood 2000). On top of its strategic positioning in the party system, ARENA also enjoyed disproportionate access to material and symbolic resources. ARENA’s campaign spending averaged three times that of the FMLN (Koivumaeki 2009).



Plausibility Tests: Out-of-Sample Cases

309

The party’s privileged relationship to landowning and business elites, and notably to media companies, explain the party’s clear financial advantage. That advantage became all the more pivotal in the absence of stringent campaign finance regulations (Artiga 2004). At the same time, ARENA was also able to cement strong partisan identities around the civil war divide (Artiga 2001). Its strong partisan identity in turn contributed to fostering a dense party organization network penetrating much of the country’s territory (Artiga 2004; Zamora 1998). Moreover, strong partisan identification at the elite level also contributed to fostering high levels of congressional discipline. Last but not least, the party inherited an organizational infrastructure from the paramilitary groups upon which it was founded (Loxton forthcoming). Despite its non-democratic origins, this organizational infrastructure proved pivotal electorally (Loxton forthcoming). It was used in “get out the vote efforts” (Wantchekon 1999, p. 819), as well as for implementing clientelistic strategies (Middlebrook 2000, p. 40; Stanley 1996, p. 241; Wood 2000, pp. 247–8). Drawing on these material and symbolic resources the party expanded, capturing a peripheral constituency through a combination of programmatic and valence mobilizing on non-distributive issues like anti-crime policy (Holland forthcoming), outspending the FMLM on campaign advertising (Koivumaeki 2009), and pursuing clientelistic and particularistic mobilization (especially in the rural and informal sectors) (Koivumaeki forthcoming; Artiga 2004). ARENA thereby pursued a segmented strategy, appealing to distinct socioeconomic constituencies on the basis of different linkage strategies and differentiated programmatic appeals. The party’s core constituency of landowning elites and business interests was mobilized through programmatic linkages and interest representation regarding the state–market divide. The party also relied on strong partisan identities, crafted during the civil war, to connect to its core constituency as well as to mobilize rural segments and small-business owners attracted to the party by its anti-communist and anti-insurgent roots. Meanwhile, when it came to mobilizing lower-class constituents, and those most hurt by economic austerity measures, the party primed its anti-crime profile to programmatically mobilize voters. Finally, the party also relied on clientelism (particularly targeted at small, local-level organizations in the informal and rural social sectors) to link to lower-class constituents. The resulting dualism of ARENA’s social base is reflected in Artiga’s reconstruction of Salvadoran parties’ social bases using IUDOP public opinion surveys. From 1997–2003, ARENA’s electoral base was concentrated at the two extremes of the social ladder, as well as among peasants. By 2003, 62.5 percent of those identified as upper-class supported ARENA, whereas only 12.5 percent of that class voted for the FMLN. Meanwhile, ARENA obtained 39.5 percent support from among “marginal” sectors, within which the FMLN actually

310

Segmented Representation

trailed with only 25 percent. In the middle sectors, within which the FMLN was gaining strength, ARENA’s support was lower, at 25  percent. Finally, among rural voters ARENA had 23.4 percent support, while for the FMLN support was declared by only 4.6  percent of those surveyed (Artiga 2004, pp. 172 and 190).9 How ARENA’s linkage strategies may have been segmented territorially is difficult to determine from available secondary sources, but the literature does clearly distinguish ARENA’s approach to capturing rural and urban segments, as well as its different approaches to socioeconomic segments in lower-class municipalities and in upper-class ones (Artiga 2004; Wood 2000). I therefore tentatively classify the case as one in which territorial segmentation seems to reinforce socioeconomically stratified strategies. The party is also described as having a very centralized internal structure, with internal rules concentrating power in the Consejo Ejecutivo Nacional (Artiga 2004, p.  185). This kind of organizational structure is one that should normally be conducive to harmonization. The Partido dos Trabalhadores (PT) The Workers’ Party (PT) of Brazil was founded in 1980. After more than two decades in the opposition, the party came to power in the presidential election of 2002, and was subsequently re-elected in 2006 and again in 2010. In the course of its transition from opposition to office, the party transformed from a radical “mass-organic” party inspired by democratic socialism to an “electoral-professional” organization (Hunter 2010). The PT was founded in the wake of Brazil’s transition to democracy, through an alliance between independent trade-unionists in the industrial poles of Brazil’s “ABC region,” the leftist urban intelligentsia, and Catholic organizations inspired by liberation theology. The PT was decisively shaped by the consolidation of an independent labor movement, which had already emerged in the strikes against the military government in the late 1970s. The independent trade movement was able to break with corporatist ties to the state, as well as with more traditional forces of the left, such as the Brazilian Communist Party, that still advocated Leninist models of labor organization. Instead, the emerging labor movement sought to emulate the strategy of the Polish movement Solidarity, and therefore created its own party: the PT (Keck 1992). Subsequently, the PT and its leadership were directly involved in the  

9 

Because no data is available for the 2003 elections, this figure for the rural sector is for 1999.



Plausibility Tests: Out-of-Sample Cases

311

founding, in 1983, of Brazil’s new umbrella labor organization (the Central Única dos Trabhaladores, CUT). In 1984, the party also took on a direct role in the campaign for direct elections (Direitas Ja!), which formally ended in 1988 with approval of a new constitution, closing the country’s protracted transition to democracy. After democratization, the PT competed in the context of an inchoate party-system characterized by the presence of catch-all and highly personalistic parties (Mainwaring 1999; Ames 2001). A typical externally mobilized party, the PT was clearly distinguished by its strong programmatic profile (democratic socialism) and its participatory and organic internal structure (Meneghello 1989; Keck 1992; Hunter 1997). Programmatically, the PT advocated radical redistributive reforms, including significant land reform. Engaging closely with social movements was also in line with the PT’s focus on basismo, the grassroots and participatory style through which the party sought to deepen Brazilian democracy (Keck 1992; Hunte 2010). On this basis, for instance, the party developed a close linkage to the landless movement (MST), which was formally founded in 1984 and had begun occupying land in the early 1980s (Ondetti 2008). The party’s innovative approach to democracy was later realized in the development of new forms of local governance, such as the participatory budgets that the party first implemented at the municipal level (Goldfrank 2006). Competing in a context of political corruption and weak party organizations, the PT developed a very salient anti-corruption stand and implemented highly routinized processes for selecting and recruiting party members into a very cohesive and disciplined organization. At the local level, transparency and a participatory style characterized the party’s innovative approach to municipal governance, while the PT’s congressional delegation stood out for its party discipline and strong programmatic commitments (Hunter 2010). From its early years, Luiz Inácio Lula da Silva (popularly known as Lula), was the party’s most prominent leader. Although he developed a personal following based on his charisma, as a historic union leader and as a political figure closely tied to the opposition to Brazil’s military regime, Lula also quite directly symbolized the party’s commitment to radical change and democracy. Lula first ran for office in 1982, as an unsuccessful candidate for the state government of São Paulo. Notwithstanding, in 1986 he obtained the largest single plurality in the country’s congressional elections, and was thereby elected to Congress. In 1989, in the country’s first direct elections for president since the 1960s, Lula ran for the PT and was defeated in the runoff by Fernando Collor de Mello. Collor’s government ended prematurely, through a congressional impeachment in the wake of a corruption scandal and hyperinflation. In 1990, Lula opted not to seek re-election to Congress, and instead devoted himself

312

Segmented Representation

to working “internally” for the territorial expansion of the PT. The latter was needed, as the party’s electoral base was very heavily skewed towards the south of the country and concentrated in Brazil’s industrial poles. Especially in the northeast, where poverty was highest in the country, the political landscape was dominated by personalistic and clientelistic electoral vehicles of the Brazilian landowning elite (Montero 2012). In the early 1990s, Fernando Henrique Cardoso, President José Sarney’s finance minister, successfully implemented a stabilization plan (the Plan Real) to curb hyperinflation. He thereby developed a strong following on the center-right. Running with the Social Democratic Party of Brazil (PSDB), Cardoso was elected president in 1994 and successfully re-elected in 1998. In both elections, Cardoso defeated Lula and the PT. Although it had consolidated as a programmatic party with a strong and committed core constituency, and had developed a very popular leadership, the PT’s electoral growth had seemingly reached its limits by the mid-1990s. Economic growth after successful economic stabilization and market reforms also imposed limits on radical leftist programmatic platforms (Baker 2009; Hunter 2010). Confronting this strategic situation, the PT embraced moderation. According to Hunter (2010), the party not only de-emphasized its socialist programmatic platform and detached itself from the social movements (especially from the radical MST), but also adopted modern political marketing technology, such as frequent polling. The PT’s governing experience at the municipal level, as in the case of the FA in Uruguay, also served a dual purpose. While contributing to pragmatism and political learning within the party (see Samuels 2004), it also signaled moderation to voters. While in opposition, the PT’s unrelenting resistance against market reform proposals in Congress, especially its opposition to the pension reform proposal submitted by President Cardoso, was crucial in signaling the party’s programmatic commitments to its core constituency (Hunter 2010). The PT’s masterful combination of moderation and commitment to change became crucial for its electoral success in the presidential election of 2002. This combination also helps explain why, during the 1990s, the PT’s image moderation did not produce significant internal crises as it later would during the party’s first term in office (Samuels 2004).10 In the aftermath of the Asian crisis (1998) and devaluation of Brazil’s currency, presidential approval for Cardoso declined. In this context, José Serra,   10  In 2003, four long-time congressmembers of the PT were expelled from the party after opposing its proposal for pension reform. In 2005, 112 members of PT’s most leftist faction abandoned the party, making public a manifesto in which they characterized the party as an “instrument of the status quo.”



Plausibility Tests: Out-of-Sample Cases

313

Cardoso’s minister (first of planning and subsequently of health), ran for the PSDB against Lula, who responded to the public’s desire for change, especially for the worse off. This time, however, Lula was also prepared to signal that such change had to be a moderate one, working within the context of what would continue to be a market society (Hunter 2010). Running on this platform, and counting on the PT’s unmatched territorial apparatus, Lula won 46.4 percent in the first round of the presidential elections of 2002, and reaching 61.3 percent in the runoff election against Serra (Baker et al. 2006). Lula’s bid for re-election in 2006 produced similar aggregate results:  he obtained 48.6 percent in the first round and 60.8 percent in the runoff election. The similar aggregate electoral results of the 2002 and 2006 elections mask significant changes in the PT’s electoral base. As shown by Zucco’s (2008) municipal and precinct-level analyses, Lula’s electoral constituency in 2006 was concentrated in the northeast, the traditional electoral bastion of Brazilian conservative parties (Mainwaring et  al. 1999; Montero 2012). At the same time, the PT witnessed moderate declines in its historical electoral strongholds of the south and in the ABC region. The PT’s growth in northeastern Brazil coincided with implementation of Bolsa Familia, a conditional cash transfer targeting the poorest segments of the Brazilian population. Available evidence points to the significant electoral effect of the program (see Zucco 2008). Although the electoral effects of Bolsa Familia might be interpreted as a consequence of “clientelism,” current analyses suggest that the program was technically designed and properly implemented, and that its socioeconomic targeting has been apt (Zucco 2008; Hunter and Borges Sugiyama 2011). In this regard, the vote gains for the PT in the northeast seem to represent more a case of retrospective programmatic voting, with voters responding favorably to a public policy that was implemented well (Zucco 2008). The weakening of traditional clientelistic and patronage machines in the region, and the bypassing of local officials for implementing the conditional cash transfer, have also helped create a favorable opportunity structure for such programmatically driven electoral realignment (Montero 2012). Notwithstanding, Lula’s charisma and high levels of popularity may also have contributed to consolidating this new peripheral constituency for the PT. Since arriving in presidential office, the PT has in contrast had a harder time satisfying its core constituency. Lula’s orthodox macroeconomic policies, lack of significant reforms in health or education, the shyness towards land reform, and the lack of pro-participatory style initiatives related to national-level government have failed to bear out the PT’s historic programmatic platform, leaving sectors of the party’s core constituency feeling betrayed (Hunter 2010). Moreover, the 2005 passage of a pension reform seeking to reduce

314

Segmented Representation

the structural deficit of Brazil’s pension system directly conflicted with the interests of public servants, a relevant interest group within the party’s core constituency (Hunter and Power 2005). Finally, the eruption of highly visible corruption scandals involving historical leaders of the PT were also received as a betrayal of the party’s programmatic commitment to clean government (Hunter 2010). Nonetheless, and despite suffering some setbacks within its core constituency, the PT was able to retain a significant share of its original, core voters. During the PT’s tenure in office, Brazil witnessed impressive economic growth, as well as an unprecedented reduction of poverty and inequality (Hunter 2011). Although it did so through non-socialist means, the PT has therefore achieved some of its key policy goals. From a strategic point of view, the absence of viable alternatives to the left of the PT may well also have contributed to the party’s ability to retain much of its core constituency. Notwithstanding, the importance of alternative linkage strategies should also be emphasized. Over the years, the PT had developed strong partisan identifications, becoming a powerful partisan subculture in the context of the Brazilian party system. According to Samuels and Zucco (2011), identification with the PT, as well as negative identifications with the party, became an important anchor for voting attitudes in the system. The strength of the PT’s identity likely provided the party with a “captive” portion of its electoral base, one voting for the party on the basis of non-contingent party–voter linkage exchanges. This in turn created opportunities for the PT to court new electoral bases through more contingent linkage exchanges based on pragmatic and non-traditional public policy initiatives. The role of Lula and his candidate traits also cannot be dismissed. The PT thus pursued a socioeconomically segmented linkage strategy, similar to that of the FA in Uruguay. The PT had a strong programmatic core constituency, which expanded slowly through the party’s years in the opposition. The party’s moderation and increasing levels of pragmatism (see Samuels 2004), the development of highly popular leaderships at the national and municipal levels, and increasing electoral discontent with incumbent parties, facilitated the PT’s eventual arrival into presidential office. Especially after arriving in office, the PT developed a peripheral constituency by implementing a highly effective social policy program targeted at the country’s extreme poor. Although betraying several of its historical programmatic stances, based on which it has originally developed programmatic linkages to its core constituency, the PT was nonetheless also able to hold on to a large part of its original social base. Strong PID, as well as Lula’s popularity, seem to have successfully made up for the weakening of programmatic linkages to the core constituency. Although very tentatively, the PT’s linkage strategies to its core and peripheral constituencies



Plausibility Tests: Out-of-Sample Cases

315

could be characterized as territorially segmented, due to the presence of clear geographic patterns in the vote for the party that overlap with socioeconomic demarcations (Zucco 2008). Finally, despite visible internal conflicts (Hunter and Power 2005) the party has seemingly been able to strategically harmonize its linkage strategies, through the self-harmonizing mechanisms characteristic of mass-parties pursuing party-centered linkage strategies. The Movimiento al Socialismo (MAS) Originating as the “political instrument” of the rural coca-producing peasant groups in the Chapare region of Cochabamba, the MAS (Movement to Socialism) rapidly transformed, from a social movement focused on resisting coca-eradication policies into Bolivia’s largest party.11 This transformation took no more than a decade. In 1994, the new Ley de Participación Popular (Popular Participation Law) opened new institutional spaces at the local level, facilitating greater political mobilization of indigenous peoples (Van Cott 2005). In this context, peasant and indigenous unions initially focused on defending their interests against the “criminalization” of coca triggered by anti-drug initiatives transformed themselves into an electoral party (Zuazo 2009). The resulting “political instrument” gained rapid access to local office, showcasing the electoral channel as a new and viable way of reaching power (Zuazo 2009). Early electoral results were also crucial for the internal politics of the proto-party (Komadina and Geffroy 2007). Initial success in the local elections contributed to consolidating the cocalero unions in the Chapare region as the leading faction of the indigenous and peasant movement (Zuazo 2009). The results also contributed to consolidating Evo Morales’ leadership in the movement, displacing other historical leaders such as Felipe Quispe, who initially underestimated this electoral path to power (Zuazo 2009).12   11  The idea underpinning the “political instrument” is the “self-representation” of the masses. This idea was developed theoretically by vice-president García Linera. The social organizations that comprise the MAS’s core constituency reject the “party” classification on the grounds that parties in Bolivia have tended to divide popular sectors. They argue that they are merely spokespeople for their constituencies, as opposed to representatives. Given its self-understanding as the “self-representation of the masses,” the MAS does not intend to build a conventional party; it aims rather at strengthening itself as a decentralized movement.   12  Quispe was never a member of the MAS. In the national elections of 1997, competing as Izquierda Unida, the MAS obtained only 3.7 percent of the vote. Yet, the party climbed to nearly 17 percent in Cochabamba, winning four congressional seats. In this election, Morales obtained the largest uninominal plurality in the country, being elected to Congress with 70 percent of the vote in the Chapare and Carrasco circumscription (Zuazo 2009).

316

Segmented Representation

Although it originally emerged at the local level and in rural communities, the MAS rapidly grew into a national-level political party. This involved a series of internal transformations: from being a rural movement to becoming a national movement with both rural and urban bases; from being a movement of agricultural producers to becoming a poly-class movement; and from being a predominantly indigenous movement to becoming one incorporating other groups. The “nationalization” of the MAS was intrinsically related to social mobilizations against market reforms, which the party seized as strategic opportunities for empathizing with social discontent, even while not taking direct action in terms of initiating and leading strikes (Silva 2009; Lazar 2006). In this context, by incorporating the demands and adopting the discourse of protesters against market reforms, the MAS and its leadership were able to grow their following, while moving to the national political arena (Webber 2012; Anría forthcoming). Three subsequent waves of protest were pivotal in that regard: the “Water War” in Cochabamba during 2000; the “Gas War” in 2003, leading to the resignation of President Gonzalo Sánchez de Losada; and the “May–June” 2005 protests (known as the “second Gas War”), leading to the resignation of President Carlos Mesa and the call for elections that allowed the MAS to win the presidency (Silva 2009). As a result of this process, the MAS became able to mobilize a large and diverse constituency, incorporating not only indigenous peoples, but also mestizo voters, union leaders, and a large group of politically disenchanted voters (Madrid 2012). Although it failed to win the presidency in 2002, the party won twenty-seven congressional seats. In 2006, MAS leader Evo Morales won the Bolivian presidency, while the MAS continued to accrue institutional positions thanks to its exponential electoral growth.13 After successfully reforming the Bolivian constitution through a popularly elected Constitutional Assembly, Morales was re-elected with an unprecedented 64  percent of the vote in December 2009. Conventional accounts of the MAS’s rise to power center on the “populist” leadership of Evo Morales and how it drew together a very heterogeneous social constituency (Madrid 2012). Indigenous issues are then described as having provided the framing for articulating a large and diverse social coalition, around what Madrid calls “ethno-populism” (Albro 2005; Madrid 2012, p. 242).

  13  The MAS won the 2006 constituent assembly elections, the 2008 recall referendum, the 2009 constitutional referendum, and the 2010 municipal/regional elections. Moreover, in the 2009 general election the MAS won 114 (out of 163)  seats in Congress (now the Asamblea Legislativa Plurinacional).



Plausibility Tests: Out-of-Sample Cases

317

In reality, however, although important, Morales’ leadership only partly explains the rise of the MAS. The party at different stages actually brought together a diverse social constituency with varying levels of attachment to the MAS on the basis of several distinct strategies (Anría forthcoming). Levels of split voting (for the MAS at the presidential level, for other parties or candidates at the municipal level) have therefore differed significantly between urban and rural areas representing distinct social bases, for example (Zuazo 2009). It is also illustrated by the composition of Morales’ presidential cabinet. The over-representation of leaders from urban organizations such as the FEJUVE of El Alto, and the relative under-representation of the MAS’s indigenous and peasant original base, is quite telling in this regard (Zuazo 2009). In a very schematic way, the MAS could be described as having two different electoral bases. The core constituency of the party is in the rural sector, and especially in peasant organizations that unified while resisting market reforms and state-repression and prosecution in the context of the war on drugs (Anría forthcoming). In this segment, the MAS is essentially conceived of as a “political instrument” of peasant social organizations. Such organizations distrust party politics, and organize around a series of decentralized and horizontal coordinating institutional mechanisms. In terms of its ethnic composition, the core constituency of the MAS, the Chapare cocaleros, tend to be Quechua. If cocaleros are “colonos” (settlers)—as many of them are—they tend to come from the mining communities of Oruro, and to a lesser extent Potosí. The latter are bearers of the Trotskyist ideological legacy prevalent in the mining areas (Anría forthcoming). The peripheral constituency of the MAS emerged in the early 2000s and is composed of urban voters, residing in particular in the cities of La Paz and El Alto (Anría forthcoming). Political discontent with the traditional party system and the rapid emergence of the MAS as an alternative to established parties sped the MAS’s electoral inroads in the two cities. Those inroads were facilitated by the pursuit of segmented electoral strategies. Individual leadership was also essential. In El Alto, Felipe Quispe, the Aymara leader of the Unique Confederation of Rural Laborers of Bolivia (CSUTCB) was the key player, spreading the Water War (originating in Cochabamba) to the city (Espósito and Arteaga 2006). Although Quispe later founded the MIP to compete (unsuccessfully) with the MAS, his actions were fundamental in mobilizing discontent against the traditional party system, and thus in opening space for the MAS’s growth in the city (Van Cott 2005; Anría forthcoming). Interestingly, according to Madrid (2008) the MIP and the Liberation Movement Tupac Katari (MRTKL) did so poorly relative to the MAS due to their “exclusionary” indigenous mobilization, in contrast to the MAS’s much more inclusive appeal.

318

Segmented Representation

El Alto is characterized as a rapidly growing city, populated through massive immigration from rural Aymara communities. The city is also overwhelmingly poor. According to Anría’s (forthcoming) characterization, 64  percent of El Alto inhabitants are poor, while an additional 17 percent live in conditions of extreme poverty. Moreover, 75 percent of El Alto’s inhabitants identify as Aymara (Anría forthcoming, Table 1). In La Paz, poverty and extreme poverty are significantly lower (33 percent and 9 percent), while indigenous self-identifications are less homogeneous, with 49  percent identifying as Aymara, 10 percent as Quechua, and a sizable 39 percent not identifying with any particular indigenous group (Anría forthcoming, Table 1). In El Alto, the MAS has attempted to grow electorally by pursuing a “corporatist” strategy of co-opting base organizations by distributing resources and patronage (Zuazo 2009; Anría forthcoming). According to Zuazo’s (2009) interview evidence, MAS leaders in the city point to their capacity to deliver jobs and benefits to their base organizations (particularly non-indigenous organizations) as pivotal for their electoral ascendance (Zuazo 2009). The internal functioning of the party thus also seems to have become dualistic. In relating to its core constituency, MAS has been shaped by bottom-up, participatory processes. Yet, in El Alto in particular top-down co-optation is more frequently observed (Zuazo 2009; Anría forthcoming). Moreover, as clearly illustrated by Anría’s interview evidence, the MAS pursued an infiltration strategy, first seeking control over base organizations, then competing for control of larger “umbrella-organizations.” The electoral payoffs of this strategy have been significant. In the general elections of 2002 the party obtained 27.9  percent in El Alto, subsequently climbing to 77.1 percent in 2005, and to an astonishing 87.5 percent in 2009. Yet, as reported by Anría (forthcoming, Table 3), significant levels of split voting persist in municipal elections, suggesting that El Alto’s support for the MAS could still be “conditional.” The MAS’s strategy in La Paz was different. In La Paz, traditional parties collapsed less rapidly than in El Alto (where the non-traditional CONDEPA Party had already made significant electoral inroads in the 1990s). Moreover, in 1999 the center-left Movement without Fear (hereafter MSM) was elected to the mayoralty of La Paz. The MSM and its most conspicuous leader, the elected mayor of La Paz, Juan del Granado, had split off from the MIR (a leftist traditional party). Del Granado was re-elected as mayor in 2004. Although the MAS had insisted for the 2004 election on a “zero alliances” strategy, which it defended as necessary for avoiding contaminating the political instrument by including traditional politicians, the MAS nonetheless celebrated an informal alliance with the MSM (Anría forthcoming). The alliance was key to successfully engaging the electorate from La Paz, and allowed the MAS to increase its electoral ascendancy in the



Plausibility Tests: Out-of-Sample Cases

319

presidential race, obtaining 63 percent in the 2009 election, while consolidating the MSM’s municipal strength (the MSM garnered 48.5 percent of the vote in the municipal elections of 2010). Moreover, some leaders of the MSM were allowed to run as MAS candidates in the general elections of 2005 and 2009, and they have occupied important positions in both MAS governments. Thus, in seeking to compete in urban and middle-class settings, the MAS allied with traditional center-left figures with more broad appeal in those sectors. This strategy moderated the MAS’s image among urban, middle-class voters from its earlier image as a more radical, indigenous party (Madrid 2012). Overall, the MAS thereby pursued a highly segmented strategy for linking to its core and peripheral constituencies. The MAS’s different linkage strategies were also associated with markedly different organizational strategies:  bottom-up and pluralist vis-à-vis its core constituency in the rural sectors, top-down and co-optative vis-à-vis popular urban sectors, and coalition-making vis-à-vis urban middle-class segments. The party has related to its core constituency, as well as to middle-urban segments, through programmatic and interest representation. Candidate-traits and platform appeals also appear pivotal in mobilizing middle sectors in La Paz. In turn, when seeking to mobilize the urban poor, clientelistic mobilization through distribution of patronage and service provision to particular organizations became the party’s primary linkage strategy. The overlap between socioeconomic, ethnic, and district boundaries suggests that a dual pattern of territorial segmentation consolidated. Meanwhile, the MAS’s capacity, as an organization, for strategically harmonizing its segmented linkage appeals seems to have decreased. Centralization of power in the president and the party’s dependence on state resources for seeking harmonization both point to important weaknesses in this regard (Madrid 2012). Moreover, even while the distribution of state resources and patronage has facilitated neo-corporatist, co-optative strategies with social organizations, the MAS has had to confront increasing levels of discontent both in its core constituency and among its middle-class allies (Anría forthcoming). The crowding out of participatory dynamics in the MAS’s relation to its popular base, the Gasolinazo protests of 2011, and the dissolution of the MSM-MAS alliance in La Paz, have revealed such tensions (Anría forthcoming).

C OM PA R AT I V E A NA LYSI S O F T H E SHA D OW C A SE S’ C AU S A L C O N F IG U R AT IO N S From a descriptive point of view, the four shadow case vignettes presented above seek to illustrate the ubiquity and electoral promise of socioeconomically

320

Segmented Representation

segmented linkage strategies in the context of unequal societies. In this regard, the cases of the BJP, the MAS, ARENA, and the PT lend support to a fundamental claim developed in this book: especially when competing in the context of unequal societies, political parties face incentives for segmenting their linkage strategies to appeal to (socioeconomically) different electoral constituencies. Moreover, parties that succeed at implementing segmented but harmonious strategies can thereby become more electorally successful than rivals that fail to simultaneously segment and harmonize their electoral strategies. In all four cases, programmatic appeals tied to distributive preferences were also present. The relative salience of the programmatic appeals of the parties analyzed here could be attributed to their all being externally mobilized parties. With the partial exception of ARENA, all of these cases, as well as those of the FA and the UDI, are relatively “new entrants” into their respective party systems. Yet, despite the centrality of programmatic mobilization for externally mobilized parties, the four shadow cases also illustrate the pervasiveness of complementary linkage strategies, as an effective way to mobilize other electoral segments. Those strategies range from leader-based appeals (in the cases of the MAS and the PT), developing strong partisan identifications (in the case of ARENA and the PT), focusing on alternative programmatic divides or such valence issues as economic growth, innovative government styles, or crime prevention (in the cases of the PT and ARENA), and cultivating clientelistic linkages (in the case of the MAS under neo-corporatist arrangements and of ARENA), and particularistic service provision (with the BJP). In sum, programmatic appeals structured around distributive interests were pivotal for targeting only a specific segment of each of those parties’ electoral constituencies. Usually it is the core constituency of a political party that is mobilized through that type of linkage. However, in the case of the PT, programmatic representation seems to have shifted from the party’s core constituency to the peripheral one after the party’s arrival in office. The four shadow cases also allow for assessing the external validity of some additional propositions from Chapter 6. Regarding the configuration of programmatic appeals, in all four cases the pursuit of state and market reforms became consequential for parties’ linkages to their constituents. Such empirical regularity seems consistent with explanatory proposition 7 (that pursuit of market reforms in a context of democratic contestation leads to greater levels of programmatic linkages), at least for the contemporary period. Both of the leftist parties (the PT and the MAS) developed their programmatic stances by fiercely opposing incumbents’ market reform attempts. Somewhat more covertly, both the BJP and ARENA consolidated their core constituencies by promoting market reforms. State retrenchment and privatization favored pivotal



Plausibility Tests: Out-of-Sample Cases

321

segments of both of their electoral coalitions. Consistent with proposition 7, state and market reforms in turn provided a competitive edge to parties with privileged ties to business interests, through private sector finance. The latter applies specifically to the two rightist shadow cases (the BJP and ARENA), as well as to the UDI in Chile. The causal configurations observed in the four shadow cases are also consistent with the explanatory propositions that privatization reduces the role of state resources in funding non-programmatic linkages, while creating greater opportunities for private interests to engage in party financing (proposition 6), and that privatization and state reform reduce the stock of material goods available to parties for the pursuit of strategic harmonization (proposition 17). In three of the shadow cases, implementation of state reforms also correlated to established parties’ declining capacities to retain a segment of their social base (i.e., the segment dependent on state patronage and clientelism). This was particularly clear in the cases of traditional parties’ grip in northeastern Brazil and the Congress Party in India, but also applies to the Bolivian traditional party system. In El Salvador, the civil war and relatively recent timing of democratization (at the moment of ARENA’s consolidation) prevent the testing of these two propositions. As in both Chile and Uruguay, established parties’ capacity for strategic harmonization in Brazil, Bolivia, and, most clearly of all, India also seem to have declined in the wake of privatization and state reform. In India, state reforms even fostered the fragmentation of the traditional, hegemonic party system at the state-level (Chhibber 2001). Structural party system traits, as well as programmatic alignments, also contributed to generating favorable political opportunity structures for the new entrants to each party system examined. Albeit in very case-specific ways, and thus general terms, this is consistent with explanatory proposition 8, that predominant programmatic alignments in the party-system, as well as the presence of strong partisan identities are contingent (path-dependent) on salient events from the past. Yet, depending on their trajectory in the party system, individual parties enjoy differential access to material and symbolic resources that jointly facilitate linkage segmentation and strategic harmonization. In Brazil, for instance, the PT accrued pivotal symbolic resources by using the “logic of difference” to compete with established parties in the system (Keck 1992). This strategy was crucial in fostering development of a strong partisan subculture, as well as unusually high levels of programmatic discipline. In India, the BJP was able to court its core constituency thanks in part to a context in which the traditional Congress Party system started to decline, opening opportunities for the mobilization of ethnic and religious cleavages. On the basis of its pro-market reform stand the party was also thereby able to mobilize the petite bourgeoisie, which opposed a system based on state-funded

322

Segmented Representation

patronage. In turn, in El Salvador, ARENA could bank on the strong identifications inherited from the country’s civil war, on the basis of which it could link to lower-class voters whom it could not represent through programmatic linkages around distributive preferences. The party’s close association with paramilitary groups in the 1970s and 1980s was also central in political campaigns of the 2000s focusing on the party’s ability to fight crime firmly and effectively. In the wake of the traditional party system’s collapse in Bolivia, the MAS was able to expand beyond its very specific original core constituency by taking advantage of the resulting political opening, and presenting itself as the most viable option for radical political renewal. In such a context, the party was able very rapidly to expand its electoral coalition using segmented linkage strategies without confronting significant strategic trade-offs. In three of the four cases, party-based linkage strategies were used for appealing to at least one electoral segment. The exception in this regard is the MAS, which was explicitly founded as a social movement. That party was also the one least successful in harmonizing its linkage strategy. It was highly dependent on both access to state resources and the figure of Evo Morales to achieve even a modicum of strategic harmonization. In the other three cases, party-centered linkages provided resources that simplified strategic harmonization. These findings are in agreement with explanatory proposition 12: that the pursuit of party-based linkages provides parties with more capacity to succeed at strategically harmonizing segmented linkage strategies. Both ARENA and the BJP have pursued harmonization through hierarchical organizations. In both cases, the parties’ top hierarchy was able to centralize resources (i.e., access to nomination and thus to the “party label” and to material resources for mobilizing voters). Both parties were thereby able to achieve high levels of strategic harmonization. The cases of ARENA and the BJP resemble that of the Chilean UDI. In each of these cases, party elites commanded access to resources and strategically allocated them through a hierarchical organization, facilitating the pursuit of a highly segmented yet harmonized strategy. In turn, the PT achieved strategic harmonization through self-harmonizing mechanisms typical of mass party organizations pursuing programmatic linkages. Gradually, the party also benefited from the emergence of a strong partisan subculture. Different types of harmonization were thus observed for right-wing parties than for the left-wing parties. The shadow cases also point to interesting differences regarding the types of linkage segmentation pursued by parties of different ideological leanings. In very schematic ways, the parties of the right analyzed here, like the UDI in Chile, seem to specialize in obtaining votes at both extremes of the socioeconomic ladder. To succeed in doing so, rightist parties have a greater need to hide their programmatic commitment to protecting the material interests of



Plausibility Tests: Out-of-Sample Cases

323

upper classes when mobilizing lower-class voters. Social inequality and territorial segmentation facilitate doing so, while parties’ privileged relation to upper socioeconomic segments provide resources that are pivotal for making such rightist parties electorally competitive given their pro-market and non-redistributive orientation. Conversely, the leftist parties analyzed here seem to specialize in crafting middle-class coalitions, albeit internally very heterogeneous ones. These parties’ distributive programmatic appeals are more palatable to other electoral segments within their constituency. Leftist parties can thus openly promote their programmatic preferences, especially when in opposition. At the same time, however, those general programmatic preferences usually conceal distributive conflicts among different segments of their electoral base. For example, in the cases of the PT and FA, they conceal interest conflict between civil servants and pensioners on the one hand, and poor and informal sector workers on the other. In the case of the MAS, there are comparable “hidden” conflicts between urban and rural supporters. Once in office, leftist parties thus face the challenge of figuring out how to mediate these concealed conflicts of interest when translating their general programmatic commitments into specific policy proposals. Otherwise, they risk rising internal strife and party breakdown.

C O N C LU SIO N This chapter has provided preliminary tests for both descriptive and explanatory propositions presented in previous chapters.14 In general, the evidence presented here supports two fundamental descriptive claims about linkage strategies in highly unequal societies. In such contexts parties tend to segment their linkage strategies, and programmatic mobilization, while present, is restricted to a particular electoral segment. This chapter’s analysis of the causal propositions inferred from my two case studies lends preliminary support to some of them, while suggesting that the external validity of others is restricted, even if one takes the relatively more lenient and flexible approach to causation advocated here (that of accepting causal complexity). In particular, political-economic factors, as well as

  14  Lack of information regarding several dimensions included in my framework unfortunately does not allow for testing of the other propositions from Chapter 6 (propositions 3, 5, 11, and 16).

324

Segmented Representation

each case’s historical trajectory, seem to figure importantly when it comes to explaining the types of linkage strategies observed. The evidence presented in this chapter also indicates that the presence of non-segmented and programmatic linkages is especially prevalent within a specific group of cases: highly economically developed and egalitarian societies. Moreover, the presence of large (encompassing) public goods in society seems to act as an almost necessary condition for observing programmatic linkages. When such causal conditions are absent, the likelihood of observing segmented linkage strategies increases dramatically. In turn, electoral-institutional rules, with the clear and important exception of political and administrative decentralization, do not seem to play a univocal and autonomous role in shaping parties’ linkage strategies. This result is not only reflected in the large-N analysis, but is also apparent in the patterns observed across the set of shadow cases. For instance, in Brazil and India (as well as pre-2006 Uruguay), parties’ linkage strategies changed significantly without any corresponding electoral-institutional reforms. In the case of Bolivia, the Popular Participation Law did favorably influence the development of the MAS as a national political force. However, it did so in ways that did not coincide with electoral-institutional arguments focusing on the incentives particular electoral formulas introduce for inducing different linkage strategies. The centrality of institutional arguments for explaining types of party–voter linkages, levels of segmentation, and effective harmonization should therefore be reconsidered. Based on the cases studies presented in this chapter, it is also possible to conclude that parties on the right as well as those on the left use segmented linkages to recruit voters with distinct distributive stakes. To do so, however, they engage in different strategies. In general, parties on the left appeal to middle and lower classes on the basis of general distributive appeals. Yet, those appeals have different “meanings” for different segments of their electoral constituency, and are implemented through distinct and segmented linkage strategies. Such differentiation is needed for leftist parties to successfully downplay their core and peripheral constituencies’ conflicting distributive stakes, or reconcile them while in office. In general, parties on the right recruit voters at the two extremes of the socioeconomic ladder. This requires more linkage segmentation, and usually also entails “obscuring” tactics to hide the programmatic stance of the party (linked to the interests of its upper-class core constituency), when seeking to mobilize poor voters.

8 Conclusion We usually have someone [a traditional caudillo with a local political network] acting as a bridge, and then we go and try, very slowly, to talk to the people. We reach the Blancos with a “ruralist” and “Artiguist” discourse. And they also like our rebellious past as “Tupamaros,” because that is the root of Blanco identity. Meanwhile, we reach the Colorados by talking about the old Batlle. However, if you tell them about Marx and Lenin, or about Frenteamplismo, forget it. Lucía Topolanski, FA MPP national leader, personal interview, 2003

The epigraph describes the appeals designed by the MPP, a fraction of the FA, for reaching out to new voters in traditional party strongholds. Those appeals have differed from the ones the fraction and the party have used for linking to their core constituency. As with other strategies described in this book, the MPP and the FA have identified different electoral segments and drawn on specific linkages for targeting each of them. In targeting particular groups, the party’s strategy has involved stressing some linkages while obscuring others pivotal for mobilizing other constituencies. By harmonizing their development of these segmented strategies, the MPP and FA were able to reduce potential trade-offs between them. In this book, I have argued that the presence of segmented party–voter linkages constitutes a long-term trait of party systems functioning in the context of unequal societies. Although linkage segmentation has been historically ubiquitous in unequal societies, my findings suggest that major changes can nonetheless occur over time within particular party systems. The predominant types of linkages that parties pursue in a system, as well as their capacity to harmonize linkage strategies by drawing on specific symbolic and material resources, can change significantly over time. They also vary greatly across cases. For instance, programmatic linkages have become more important in Uruguay since the mid-1990s, while becoming less central since 1990 in Chile. And while Chilean parties (with the exception of the UDI) have suffered eroding capacity for harmonizing their segmented strategies, parties’ capacity for harmonization has remained relatively constant in Uruguay. Which factors explain such variance over time and space? Regarding the types of party–voter linkages observed at the systemic level, long-term

326

Segmented Representation

and structural conditions as well as institutional variables are influential in shaping linkage patterns. That said, they are neither necessary nor sufficient for particular patterns of linkages to develop. These factors operate through complex interactions with other causal conditions. For instance, it is the interaction between patterns of territorial and socioeconomic segmentation observed in a given society, and the presence of specific electoral rules (i.e., district size, district magnitude, and the overlap between district and socioeconomic boundaries), that ultimately determines whether dual or mixed linkage strategies will predominate at the district level. Moreover, the short-term political economy of market and state reforms, pursuit of political and administrative decentralization, and the structure (and relative endurance) of the main political alignments emerging from each country’s transition to democracy were particularly pivotal in causing the types and evolution of party–voter linkages observed in Chile and Uruguay after democratization. Moving to the individual party level, the evolution and political trajectory of particular parties within a party system, as well as their patterns of linkages to their core constituencies, account for significant differences in how able they are to use segmented and harmonized strategies to expand electorally. Confronting a given opportunity structure for electoral expansion, parties’ differential access to material and/or symbolic resources critically influence their relative ability to reap increasing electoral returns through harmonized segmentation. Moreover, whether parties rely on candidate-based or party-based linkage strategies affects their capacities as collective organizations over time. Party-based strategies, like those predominant in contemporary Uruguay, reinforce parties as institutions and enable them better to constrain the behavior of political entrepreneurs. Over time, such strategies enhance parties’ capacities for harmonization. By contrast, candidate-based strategies, like those predominant in contemporary Chile, have the opposite effect. To be sure, not every process of electoral expansion results from successful segmentation and harmonization of linkage strategies. Kirchheimer’s classic notion of the “catch-all party,” for instance, conveys a different process, in which a party diffuses its ideological profile in an effort to expand its electoral constituency. However, although segmentation and harmonization are not the only way parties can prosper, my evidence shows that such a strategy has been pivotal in explaining the remarkable electoral expansion of many recently successful political parties. As suggested in Chapter 7, individual parties’ ideological profiles may correlate to the structure of their electoral constituencies. Whereas rightist parties seem to recruit voters at the two extremes of the social ladder, leftist parties

Conclusion

327

tend to have a more middle-class (yet internally heterogeneous) social constituency. Notwithstanding, both rightist and leftist parties usually diversify their linkage strategies to seduce different electoral segments. They therefore often rely on different types of linkages to relate to different social groups. In other instances, the same general type of linkage may be deployed differently, or given different content, in targeting one group than in targeting another. For instance, a party may segment its appeals by priming one programmatic message in one social constituency and a different one when targeting another group, or segment clientelistic linkages by distributing handouts to poor voters and state rents to wealthy ones. These findings call into question the scope and validity of the usual assumption that inherent trade-offs necessarily or inevitably emerge among different types of linkages, an assumption mainly derived from analysis of West European party systems. The literature’s assumption in that regard is not borne out by my analysis of Latin American cases (or of the BJP in India). Calling the strict assumption of inherent trade-offs among linkage types into question has important theoretical, substantive, and methodological implications.

M E T HO D O L O G IC A L I M P L IC AT IO N S Relaxing the trade-off assumption has important implications for the basic process of descriptive inference. It entails the need to assess the relative incidence of different types of linkage strategies in a given party-system, rather than taking the presence of one linkage type as evidence of other types’ absence. It also entails the need to unpack the different strategies that individual parties deploy in relating to different electoral segments. For instance, one cannot assume that the dominant linkage strategy observed in a party’s interaction with its core social constituency is indicative of how it in turn relates to other, peripheral social constituencies. As likely, the kinds of linkages on which it relies in each case may be very different. In fact, for parties operating in very unequal social contexts, the analyst should begin by presuming that this is most likely the case, revising that presumption only if and when the empirical evidence ultimately so allows. This type of analysis of linkages requires a labor-intensive, multi-method, and multi-level analytical strategy. One could argue that the payoffs of such an analytic strategy in terms of improved descriptive inference are insufficient to outweigh its obvious demands. Why then devote such efforts to improving conceptualization and calibrating descriptive inferences? The resulting, more precise characterization

328

Segmented Representation

of cases is admittedly less parsimonious, and may not always lend itself to computing aggregate scores on different variables that can then be readily used for large-N analysis. For instance, it may not lead to an index of programmatic structure, or scales measuring degree of clientelism or charismatic leadership observed across a large number of cases. One might then ask why such a demanding methodological effort is worth the investment. Ultimately, the answer is that valid descriptive inferences are indispensable as a basis for pursuing the kind of valid and meaningful causal inference undeniably highly valued in political science. Even the most sophisticated process of causal inference is bound to yield often inaccurate or misleading results if based on a thin or casual initial characterization of the object of study. More accurate, detailed, and precise characterization of individual cases is ultimately needed to improve our understanding of political representation and related political phenomena in contemporary societies. Turning to causal inference, explaining segmented linkage strategies on the basis of more thorough characterization of individual cases requires dealing with patterns of conjunctural causation across time and space. Such causal complexity and equifinality inevitably complicate the precise assessment and attribution of individual factors’ contribution to causal outcomes. Moreover, the role of case-specific factors often limits the scope of theoretical generalization possible on the basis of particular case studies of this kind. Yet, ignoring the types of conjunctural causation that drive the nature of linkage strategies comes at a cost.

SU B S TA N T I V E I M P L IC AT IO N S Linkage Strategies’ Redistributive Impacts The party systems of Chile and Uruguay are usually lumped together in comparative analysis. It is often claimed that both systems display similar levels of party system institutionalization and programmatic structure. Whereas these two party systems are indeed similar, compared to other Latin American cases where massive partisan turmoil has occurred, there are nonetheless significant differences in the nature and evolution of their respective party–voter linkage strategies. Those differences have consequentially affected normatively important public policy outcomes. Pinpointing the precise causal mechanisms connecting kinds of party– voter linkages and distributive results would require a very fine-grained analysis of the policy-making process in each case. For instance, one would need

Conclusion

329

to analyze legislative coalition-making and the effects of different institutional settings and veto-players. However, even without undertaking a full analysis of this kind, one can note a consistent pattern in the nature of distributive public policy initiatives in Chile and Uruguay. One can also see how that pattern is theoretically consistent with what one would expect given the biases in political representation arising from the structure of party–voter linkages observed in each case. A useful way to compare policy outcomes in the two cases is to analyze the key distributive policy initiatives implemented in each country under the presidencies headed by center-left parties (the PS in Chile from 2000 to 2010, and the FA in Uruguay from 2004 to 2012). Recent analyses of the “turn to the left” in Latin America place both countries’ center-left governments among the “social democratic” or “moderate left” type (see, e.g., Weyland et al. 2010; Cameron and Hershberg 2010; Levitsky and Roberts 2011). In spite of the alleged similarities between the two cases, the comparison of policy outcomes in each clearly reflects significant differences (Pribble 2013; Huber and Stephens 2012). While confronting increasing levels of social inequality, the socialist-headed governments of the Concertación in Chile have avoided including progressive tax reform in their policy agenda (Fairfield 2010). In Uruguay, a progressive income-tax reform was implemented by the FA after only two years in office (Castiglioni 2010). More recently, a system of differential consumer-tax was embedded in Uruguay’s CCT policy, also seeking to increase redistribution to the poor. Moreover, while the FA government resurrected collective wage bargaining immediately after taking office, further strengthening the labor movement (Castiglioni 2010), the Concertación’s agenda has been more oriented towards further liberalizing labor policies (Palacios 2010). The divergent trajectory of these two political forces could also explain why the FA, after a process of internal conflict and debate involving confrontation with the labor movement, decided not to pursue a free-trade agreement with the United States, while the Chilean government, under the socialist-led presidency of Ricardo Lagos, signed one without any significant opposition from civil society or party activists. If one looks at social policy reform initiatives, relevant differences also emerge. Both the content and political process behind tax and social policy reform have corresponded to the different strategies through which parties mobilize political support in each country. Interestingly, all good things do not seem to go together here. In Chile, the Concertación’s detachment from the popular sectors and the union movement, as well as its reliance on non-programmatic party–voter linkages, gave center-left governments greater autonomy for introducing innovative and technically efficient reforms from above. Those reforms were very effective in decreasing poverty and addressing

330

Segmented Representation

the needs of the most vulnerable segments in society. The government thereby effectively targeted and gradually incorporated universalistic benefits for the poorest forty percent of the population. As a result, social policy provision is, for instance, less age-biased in favor of pensioners who are better off than poor children in Uruguay (Filgueira and Kaztman 2001; CEPAL 2010). The social policy biases observed in Uruguay correspond to the political clout of pension beneficiaries and other powerful corporate interests, such as state employees. Yet, Chile’s reforms were less progressive than Uruguay’s when it came to redistribution of resources among different social classes, in part due to the skewed access to interest representation that the private sector and the wealthy segments of society enjoy in Chile. Such greater access to programmatic representation by the wealthy is accompanied by a reliance on non-programmatic and non-partisan linkages to appeal to the popular sectors. Moreover, even if inequality has become recently politicized in Chile as a result of the increasing levels of social mobilization observed since 2011, established political parties and especially the center-left have found it difficult to programmatically channel emergent social demands in a context in which personalistic mobilization has gained increasing ground. In Uruguay, the FA was less autonomous from the social and union movement, and could not introduce reforms from above. To be viable, reforms had to accommodate the interests of the party’s core constituency, even if that reinforced “corporatist” benefits against those of poorer and more vulnerable sectors. Efficient targeting of the most needy was thus hindered. As a result, the more redistributive reforms implemented in Uruguay were distorted by corporatist interests, diminishing their overall impact. In the end, however, their redistributive effect was still greater than that of reforms in Chile. The latter result is consistent with a party system in which programmatic and interest representation is not predominantly skewed in favor of wealthy constituents, as in Chile. The pattern of policy initiatives implemented in health, pensions, and education also seems consistent with these representational differences. In Chile, the Concertación incorporated “universalistic” pillars into the health and pension systems, through the AUGE Plan in health implemented under Lagos and the universal pension system implemented under Bachelet (Pribble 2013; Huber and Stephens 2012; Ewig and Kay 2011; Ewig 2008; Dávila 2005). Yet, both reforms were implemented as a supplement to the existing dual system (with private, high-quality coverage for upper sectors, and public and lower-quality coverage for poorer sectors). The reforms were crucial for improving social provision for the poor, but did not challenge the system’s overall regressive biases in quality and coverage. From a political standpoint, both reforms were implemented technocratically from above, while protecting the interests of

Conclusion

331

powerful business interests tied to the private insurance system covering the wealthy (Pribble 2013). In Uruguay, the pension system combined a private pillar with a public one that was broader and provided greater coverage than the Chilean system. Introduced in 1995, it was not structurally reformed under the FA governments. Yet, pensions’ scope and real value increased significantly, thanks to initiatives like integrating health coverage for pensioners. Collective bargaining and better labor code and tax enforcement also contributed to sharply increasing the number of workers eligible for pension coverage. The strength and political articulation of the pension-beneficiary movement likely inhibited major reform attempts, while stimulating the increase in pensions’ real value. In terms of health policy, however, the FA did overhaul Uruguay’s care system. Public providers were strengthened, thus reducing private–public segmentation and gaps in quality, and private providers (mutualistas) received bailouts. Creation of the national health fund (FONASA) also significantly expanded formal health coverage (Castiglioni 2010). Politically, the health system reform process was long and heavily negotiated, with the medical union, the labor movement, and private providers all brought to the table (Chasquetti 2007). Here again, privileged representation of organized corporatist interests contributed significantly to shaping social policy, with distributive implications, in Uruguay. Interestingly, both countries’ educational systems are considered “in crisis” (Mizala and Romaguera 2005; Cox 2003; Arellano 2001; Castiglioni 2010; Bentancur 2010). Education spending increased significantly under the governments of both the Concertación and FA.1 However, structural education reforms were not introduced in either country, for different reasons. In Chile, the influence of private actors (especially education providers), as well as the public sector’s municipal-level decentralization, seems to have obstructed structural educational reforms (Castiglioni 2005). Meanwhile, no political party had or developed linkages to the student movement, leading to   1  In Chile, under President Lagos, a new loan (known as the Crédito con Aval del Estado) for financing education expenditures was created, seeking to expand tertiary education coverage. This loan, however, did not change the logic of the financial system and benefitted private financial institutions through the expansion of the credit market. A few years later in Chile, President Bachelet sent a series of reform initiatives to Congress (most of which were rejected by the legislative branch). An important exception to this was the successful replacement of the LOCE (Ley Orgánica Constitucional de Educación) with the LEGE (Ley General de Educación) in 2009. The latter sought to respond to the student riots of 2006. Yet, the student movement progressively came to regard the LEGE as a fiasco, which once again did not alter the structural characteristics of the educational system, later in part triggering the student riots of 2011 (Jackson 2013).

332

Segmented Representation

the riots of 2006. In Uruguay, the government of Tabaré Vázquez considered education a priority, and called for a national dialogue on educational reform (Chasquetti 2007). That dialogue was designed to be as inclusive and decentralized as possible, and it produced a series of reform proposals clearly aligned with the interests of the teachers’ union. They were also perceived as technically inadequate by the FA’s government. Implementing reforms preferred by the government, however, would have led to a political crisis involving the party’s core constituency. The reform process therefore stalled, and continues to be protracted (Pribble 2013; Huber and Stephens 2012; Chasquetti 2007; Castiglioni 2010). Given the importance of patterns of territorial segmentation of party–voter linkages, the presence of distributive biases and their consistency with the linkage strategies observed in each case should also be further analyzed. In Uruguay, where mixed linkage strategies predominate, territorial distributive biases are more blurred and less significant because cross-district disparities are not observable. By contrast, in Chile, where dual-linkage strategies are prevalent and high territorial segregation is present, there are more notable distributive inequalities across different types of districts. Segmenting party– voter linkages between districts, as in Chile, reduces politicization of territorialized inequality and social fragmentation, making it easier for them to flourish. In this type of context, social inequality does not seem to be corrected but rather reinforced by biases of political representation related to segmented linkage strategies. To provide an illustration on the extent of territorialized inequality of social spending in Chile, let us compare data from two municipalities within districts without intra-district linkage segmentation: Cerro Navia (from District 18)  and Vitacura (from District 23). The municipality of Cerro Navia was controlled until 2008 by a PPD leader (Cristina Girardi). From 2001 to 2010 the Concertación received double the Alianza vote there at the district level, obtaining both congressional seats as a result of the binominal system. During the same period, the municipal government of Vitacura was led by the RN, where the Alianza won double the Concertación vote in the district, obtaining both congressional seats. Whereas programmatic mobilization predominated in District 23, particularism and candidate-based appeals were predominant in District 18. Cerro Navia is a very densely populated municipality, with a relatively high incidence of poverty. Its annual per capita spending is just 17  percent of Vitacura’s (SINIM 2012). Furthermore, 87  percent of what Cerro Navia spends is provided through progressive redistribution from the Municipal Common Fund, which redistributes resources from wealthy municipalities like Vitacura to most municipal governments in the country. Even after

Conclusion

333

Table 8.1.  Cerro Navia and Vitacura: educational policy over time (2001–2010) Municipality Indicator/year Municipal education coverage (% of students in municipal system) Total education expenditures (M$) Total educational expenditure per student (based on the average number of students enrolled in municipal education during the year)* % of students scoring above 450 on PSUn** % gap between municipal and publicly subsidized schools’ percentage of students with PSUn scores over 450 (Psubv.—Municipal)**

Cerro Navia

Vitacura

2001 29.3 percent

2010 32.8 percent

2001 17.7 percent

2010 16.2 percent

4,856.513

5,782.650

1,918.062

2,928.530

331.502

546.151

577.000

1.197762

18.3

16.4

75.1

91.7

1.3

18.4

−15.5

−2.1

Source: SINIM (2012). Key * Available estimation for Cerro Navia corresponds to 2002. ** Available estimation for Cerro Navia and Vitacura correspond to 2003.

receiving those transfers, Cerro Navia has 17 percent of the per capita budget of Vitacura to administer and run four municipal health clinics (compared to one in Vitacura) and twenty-four municipal schools (compared to two in Vitacura). These comparative figures reflect the levels of territorialized inequality that characterize Chilean municipalities and electoral districts. Yet, they cannot tell us much about how the dynamics of territorial inequality play out over time or their association with the types of linkage strategies described in this book. Table 8.1 seeks to provide a more dynamic view, by focusing on the evolution of educational policy in each of these two municipalities over the last decade, under the Concertación presidencies of Lagos and Bachelet. Educational policy in Chile is structured around a three-tier system composed of municipal schools (public schools), publicly subsidized schools (private schools in which students pay a relatively low-cost tuition to attend, and the state contributes a fixed amount to the school for each student enrolled), and fully private schools. Both municipal and subsidized schools receive funds from the Ministry of Education, which are allocated (primarily) on the basis of the number of students enrolled in each school.

Segmented Representation

334

1992–2010 Evolution Municipal Ed. Enrollment 2010 Municipal Ed. Enrollment % of Municipal Students Scoring +450 points in PSUn

100

2010 Poverty Rate 80

60

40

20

o

EI Alto Bo s C er que ro N Sa av i n R a am ó La n Pi nt an a

ad Pu

en

Pr

Lo

te

a

ue

rn te

rq La

C

ls

ai po

Pi

ol en

M

Pe ñ

SJ

de

or Fl

al

al id a

a

m or

La

he

N

ec

a

rn

nt ui

–40

Q

ra

ei na

R

La

Lo

Ba

cu ta

de

La

s

Vi

C

on

–20

s

0

Graph 8.1.  Poverty estimates, municipal enrollment, and PSUn scores in Greater Santiago municipalities included in the electoral district sample (Chapter 4). Source: SINIM and Fundación Sol (2011).

As shown in Table 8.1, Cerro Navia spends significantly more on education than Vitacura, where fully private school enrollment predominates. Cerro Navia has to cover the educational needs of a much greater percentage of the population. Although the net amounts spent per student enrolled in municipal education increased over time in Cerro Navia between 2002 and 2010, educational expenditures there grew more slowly than in Vitacura. As a result, while in 2001 Cerro Navia spent 57 percent of what Vitacura spent on education per student, by 2010 that percentage had fallen to 45 percent. In short, although redistributive transfers among municipalities are significant, they remained largely insufficient for reducing inequality over these ten years. Table 8.1 also shows comparative educational results, noting the percentage of students with scores of 450 points or more on the university selection test (PSUn).2 Looking at the trend on this indicator over time allows one to control for the massive impact that socioeconomic background plays on PSUn scores. Whereas Vitacura’s results improved by 15 percent from 2003 to 2010, those

 

2 

A score of 450 is the minimum considered sufficient for entering a university.

Conclusion

335

of Cerro Navia marginally worsened. Moreover, the gap between municipal schools and subsidized schools increased by almost 17 percent in Cerro Navia. The poorest in the community (those unable to afford even the very modest tuition of a subsidized school) are increasingly trapped in “ghettos of failure.” By contrast, in Vitacura municipal school students actually did better on the PSUn test than those from publicly subsidized schools. Graph 8.1 puts both municipalities in comparative context, providing data for all the municipalities that comprise the electoral districts of the metropolitan region analyzed in this work (see Chapter 4). Municipalities are sorted from left to right according to their poverty levels. The remaining bars in Graph 8.1 display the percentage of municipal education enrollment in 2010, the evolution of such enrollment between 1992 and 2010, and the percentage of PSUn scores of at least 450 points. Although specific outliers exist (e.g., La Cisterna regarding PSUn performance, Puente Alto regarding 2010 municipal enrollment), the stylized trends described for Vitacura and Cerro Navia hold. In that regard, poorer municipalities display much weaker PSUn performances, as well as a more marked reduction in municipal education enrollment. In sum, this cursory comparison illustrates how, even during a period when social investment increased significantly in Chile (see Pribble 2013), territorialized inequality was maintained, and even reinforced, in ways consistent with the expected distributive effects of dual-linkage strategies. The example of Cerro Navia, in which particularistic and candidate-based appeals predominate, also illustrates the presence of deep patterns of social fragmentation among the very poor. For instance, while public school attendance has decreased sharply in the country (reaching less than 40 percent in 2012), attendance of publicly subsidized private schools by children from the lower-income deciles has increased markedly (Fundación Sol 2011). That result has been driven by the attempts by the poor to socially segregate their children from the poorest in the neighborhood. Such segregation can be obtained through a relatively small monthly co-payment and, over time, contributes to fragmenting and deteriorating the quality of public goods provided to the community. Such a processes of social fragmentation also contributes to reinforcing the regressive biases of public provision. It translates into conflicts of interest regarding collective action, even among groups very close to one another in income. Social fragmentation among the poor thereby constrains their political agency and reinforces the fragmented provision of public goods. Such fragmentation then further undermines potential political coalitions that could eventually push for the provision of better public education through collective action and programmatic partisan mobilization (Filgueira and Filgueira 1997b). In this way, over time, “effects become cause” (Pierson 1993).

Segmented Representation

336 Chile

Uruguay

10 9 8 7 6 5 4 3 2 1 Actual politicians

Ideal politicians

Graph 8.2.  Ideal vs. perceived role of politicians in terms of favoring the “poor” or the “rich” circa 2012 (Chile and Uruguay) Source: LAPOP 2012. Coda: 1 = favors the rich, 10 = favors the poor.

Moreover, when seeking to mobilize voters in this social context, parties face increasing difficulties structuring programmatic appeals to them. In Uruguay, in turn, the strength of interest groups and their capacity to coordinate across districts over time has reinforced parties’ capacity for programmatically mobilizing larger societal groups. The greater levels of social heterogeneity observed within Uruguayan districts gives parties’ greater incentives to pursue mixed linkage strategies. It has also contributed to taming the emergence of different patterns of distributive public policy across districts.

I N T E R- T E M P O R A L PA RT Y A N D PA RT Y- SYS T E M DY NA M IC S , A N D S TAT E S’ BU R E AU C R AT IC C A PAC I T Y The different structures of political linkage I  have described for Chile and Uruguay also have other noteworthy effects. First, they have significant effects on politicians’ social legitimacy. Second, they ultimately affect states’

Conclusion

337

autonomy and technical capacity. The differences between Chile and Uruguay in these regards are relevant for understanding significant political differences between the two cases. The kinds of linkages parties pursue play a role in determining their future collective capacities and relative resource endowments. For instance, parties crafting candidate-based appeals risk eroding their capacity later to compete on the basis of party-based appeals. They may also face increasing difficulty harmonizing segmented linkage strategies. Reliance on party-based appeals has the opposite effects. This pattern is borne out by the evolution of party identities in Chile and Uruguay. The results of a recent poll (LAPOP 2012) taken in both countries suggests that the types of linkage structures on which parties rely can affect their social legitimacy, as well as citizens’ perceptions of politicians’ role in promoting social equality. Graph 8.2 presents citizens’ responses to two survey questions. The first question asked citizens about the degree to which “ideal politicians” should contribute to redistributing resources from the rich to the poor in their country. The second asked the degree to which “actual politicians” in their country did in fact redistribute resources from the rich to the poor. Whereas there was a gap between the “ideal” and “actual” role of politicians in advancing social equality in both countries, the gap was much greater in Chile than in Uruguay. That result is consistent with Chile having more of a “dual-linkage strategy” pattern, as well as a more socially “uprooted” party system (Luna and Altman 2011). The recent emergence of inequality as a central political issue in Chilean society has contributed to making “dual-linkage strategies” and their distributive effects more visible to the public, and hence more politically consequential. The appearance of inequality on the social agenda is not the by-product of partisan programmatic mobilization, but rather the result of societal discontent with established parties and social elites. In that context, established parties (which still continue to downplay programmatic mobilization) and social elites have become increasingly illegitimate, opening the space for anti-systemic social mobilization giving voice to popular dissatisfaction regarding socioeconomic inequality and its multiple manifestations across societal and political arenas. Meanwhile, despite corporatist logics biasing redistributive policies to already (relatively) privileged segments of society, the Uruguayan party system was perceived as generating comparatively less socially regressive results. This finding is consistent with the evolution of public policy described above, with the presence of mixed linkage strategies that incorporate a modicum of programmatic representation for different societal groups, and with a party system that is more “rooted” in society thanks to pursuit of more party-centered linkage strategies than in Chile.

338

Segmented Representation

The types of party–voter linkages and states’ bureaucratic and technocratic capacities have often been causally associated (Grindle 2012; Stein et al. 2006; Kitschelt 2000; Domínguez 1997; Shefter 1994; Geddes 1994). My comparative findings confirm the relevance of such causal connections. They also suggest that, beyond the classic distinction between clientelistic and programmatic systems, the specific structure and societal distribution of party–voter linkages shapes important dimensions of state bureaucracy. The Chilean bureaucracy is often praised for its high levels of technocratic capacity and political autonomy (Stein et  al. 2006; Mainwaring and Scully 2010). Those features enhance the country’s relative advantage in implementing sound public policy, especially regarding macroeconomic and budget management, investment promotion, and social policy. The Uruguayan bureaucracy, in turn, has historically been characterized as colonized by partisan actors (parties and their fractions) and lacking technocratic capacity (Panizza 1990; Garcé 2009). The legacy of state patronage has been seen as an important root of these weaknesses (Filgueira et al. 2003). In this regard, the two cases have different historical and enduring patterns that persist today. Such patterns, in my view, are closely linked to the types of party–voter linkages that parties have pursued in each system. These historical characteristics of each state seem to have been maintained, and likely reinforced, through parties’ reliance on different linkage strategies during recent years. In Chile, for instance, independent figures dominated cabinet appointments under both the Bachelet and Piñera presidencies, although the ideological and technical profiles of cabinet appointments under the two presidents differed. Whereas members of Expansiva (a liberal and politically unaffiliated think tank) obtained key positions under Bachelet, successful businessmen without previous political careers were extensively nominated under Piñera. In Uruguay, Tabaré Vásquez instead sought to promote “governability” by appointing leaders from the FA’s main fractions to his cabinet (Chasquetti 2008, pp. 394–5). This move was seen as necessary for aligning FA fractions in Congress and the executive branch. Although Vásquez also sought to strengthen technical coordination and capacity by empowering the Office for Planning and Budget (the OPP), this effort was hindered by fractional disputes across and within ministries.3 Moreover, under Mujica, the OPP was once again weakened in favor of political appointees, this time pertaining to the president’s fraction and its main political allies.

  3  On the basis of personal communication with a top officer leading this effort. Montevideo, September 2011.

Conclusion

339

In other words, the types of party–voter linkages observed in Chile contribute to isolating political leaders from society and from their own political organizations. Higher levels of personalization, as well as the endurance of top-down political mobilization anchored in non-programmatic linkages, give Chilean politicians more political autonomy from their electoral bases. In Uruguay, partisan factors drive cabinet and bureaucratic appointments. In this context, contrary to what has been increasingly observed in Chile, a politically unaffiliated “technocrat” without experience as a party activist would be highly unlikely to be appointed to a Uruguayan government cabinet. The types and segmentation structure of party–voter linkages observed in each case help cast light on why the two cases are also so different in this regard.

T H E O R E T IC A L I M P L IC AT IO N S The analysis developed in this book has broader theoretical implications for our understanding of the relation between social inequality and political democracy. Patterns of inequality and their interaction with relevant historical and structural features, electoral-institutional factors, and party endowments shape the nature of political representation observed in different polities. In particular, as my evidence has shown, they provide incentives for the pursuit of different party–voter linkages and segmentation strategies. Different kinds of linkages and their patterns of socioeconomic and territorial segmentation then, in turn, mediate how social inequalities are politically translated and influence policy outcomes. Representation biases may in fact be essential to explaining both the durability of democracy and its distributive outcomes. Here lies a possible key to the long-standing puzzle of why democracy often has not effaced social inequalities as political economy formal models based on Meltzer and Richard’s (1981) work would have us expect. And here too lies a key to understanding why, in contexts of enduring social inequality, democracy itself has often proven surprisingly stable. In short, assuming that the widespread and enduring presence of segmented linkage strategies is impossible and therefore must not exist risks blinding analysts to important features of both political representation and democracy, especially in the context of unequal and territorially fragmented civil societies. In addition, the analysis of patterns of linkage segmentation has the potential to illuminate specific distributive patterns observed across different cases. Available evidence suggests that during the 2000s, Latin American countries have “unambiguously” reduced their levels of social

340

Segmented Representation

inequality (Lustig et al. 2013a and Lustig et al. 2013b; see also Huber and Stephens 2012). Yet, the evidence also suggests that significant variance exists across cases regarding the degree and pace of inequality reduction. Moreover, cross-nationally equivalent inequality drops respond to case-specific trajectories regarding different types of redistributive policies (Lustig et al. 2013b) and exogenous economic factors (Lustig et al. 2013a). Finally, similar levels of redistribution in the aggregate might also obscure significant cross-sectional biases in redistribution (e.g., the age-bias long-documented for Uruguay; see also Lynch 2006). In sum, more fully incorporating linkage segmentation patterns into distributive policy analyses can eventually contribute to a more nuanced understanding of the relationship between democratic representation and redistribution. Finally, the analysis pursued in the book may also have broader implications for the analysis of political representation in contemporary societies. The characterization and explanation of parties’ and party systems’ features have recently been dominated by neo-institutional approaches. While contributing to creating a wealth of comparative indicators and relevant causal arguments about the drivers of party and party-system variance, the predominance of neo-institutional approaches has developed at the expense of more traditional comparative historical and political sociology analyses of party and party-system change (e.g., Lipset and Rokkan 1967; Collier and Collier 1991; Luebbert 1991; Rueschemeyer et al. 1992). This overshadowing of comparative historical and political sociology frameworks for understanding party and party-system change has also been catalyzed by the expansion of political economy models of political representation. In my view, such models too often and too heavily draw on excessively stylized accounts of party strategizing. While theoretically elegant, such accounts are often very distant from real political life. And more often than not, such analyses mischaracterize agents’ strategic choices, minimizing the theoretical relevance of key contextual conditions. Against this backdrop, the analysis pursued in this book has sought to combine a political sociological and comparative historical approach to the analysis of party and party system change, with relevant insights from the neo-institutional and political economy literatures. To do so, I have applied a fieldwork-intensive, mixed-methods, and multi-level methodological strategy. The study has thereby sought to recover relevant inter-temporal, cross-national, cross-sectional, and individual party variance. The end result represents an effort to strike an elusive middle ground between complementary, but unfortunately frequently divorced, approaches in contemporary political science. In my view, this middle ground is also a singularly fertile terrain for the study of complex and socially relevant political phenomena.

Bibliography Acemoglu, Daron, and James Robinson. 2006. Economic Origins of Dictatorship and Democracy. New York: Cambridge University Press. Agazzi, Ernesto. 2002. Personal Interview. FA and MPP Congressmember for Canelones. Aguiar, Cesar A. 1984. Hipótesis preliminares para una discusión de las perspectivas de democratización en el Uruguay actual. Montevideo: CIEDUR. Aguiar, César A. 2000. La Historia y la historia: Opinión Pública y opinión pública en el Uruguay. Prisma (15): 7–45. Ahumada, Jaime. 2003. Personal Interview. Former Municipal Council Candidate for the PS in Cerro Navia. Albiac, M. Dolores. 1999. Los ricos más ricos de El Salvador. ECA 54: 852–3. Albro, Robert. 2005. Indigenous in the Plural in Bolivian Oppositional Politics. Bulletin of Latin American Research 24: 433–53. Alcántara S., Manuel, and Juan Pablo Luna. 2004. Ideología y competencia partidaria en dos post-transiciones:  Chile y Uruguay en perspectiva comparada. Revista de Ciencia Política XXIV (1): 128–68. Alcieri, Daniel. 2003. Personal Interview. Social Department Chair of Paysandu’s Municipality. Aldrich, John. 1995. Why Parties? Chicago, IL: Chicago University Press. Aldunate, Adolfo. 1985. Antecedentes Socioeconómicos y Resultados Electorales. In Estudio sobre el Sistema de Partidos en Chile, edited by FLACSO Chile. Santiago: FLACSO. Aliano, Hubaré. 2003. Personal Interview. Colorado Local Activist in Tacuarembó. Allamand, Andrés. 1999. La travesía del desierto. Santiago: Aguilar. Altman, David. 2000. The Politics of Coalition Formation and Survival in Multiparty Presidential Democracies:  The Case of Uruguay 1989–1999. Party Politics 6 (3): 259–83. Altman, David. 2002. Popular Initiatives in Uruguay: Confidence Votes on Government or Political Loyalties? Electoral Studies 21 (4): 617–30. Altman, David. 2004. Redibujando el mapa electoral chileno: Incidencia de factores socioeconómicos y género en las urnas. Revista de Ciencia Política XXIV (2): 49–66. Altman, David. 2008. Political Recruitment and Candidate Selection in Chile (1990– 2003): The Executive Branch. In Pathway to Power in Latin America, edited by P. Siavelis and S. Morgenstern. University Park: Penn State University Press. Altman, David. 2011. Direct Democracy Worldwide. New  York:  Cambridge University Press. Altman, David, and Rossana Castiglioni. 2006. The 2004 Uruguayan Elections: A Political Earthquake Foretold. Electoral Studies 25 (1): 147–54. Altman, David, Rossana Castiglioni, and Juan Pablo Luna. 2008. Uruguay:  A  Role Model for the Left? In Leftovers:  Tales of the Latin American Left, edited by J. G. Castañeda and M. A. Morales. London and New York: Routledge.

342

Bibliography

Altman, David, Juan Pablo Luna, Sergio Toro, and Rafael Piñeiro. 2009. Partidos y Sistemas de Partidos en América Latina:  Aproximaciones desde la Encuesta a Expertos 2009. Revista de Ciencia Política 29 (3): 775–98. Álvarez Rivadulla, M. J. 2012. Clientelism or Something Else? Squatter Politics in Montevideo. Latin American Politics and Society 54 (1): 37–67. Alvez, Omar. 2003. Personal Interview. FA’s Local Activist in Artigas. Alwin, Mariana. 2003. Personal Interview. Former DC Congressmember of La Florida. América, Buenos Días. 2003. Personal Interview. Community-Based Youth Group. Ames, Barry. 2001. The Deadlock of Democracy in Brazil. Ann Arbor: University of Michigan Press. Anonymous. 2008. Personal Interview. Anonymous UDI National Officer. Anría, Santiago. Forthcoming. Social Movements, Party Organization, and Populism:  A  Study of the Bolivian Movement Toward Socialism (MAS). Latin American Politics and Society. Ansolabehere, Stephen, John M. De Figueiredo, and James M. Snyder. 2003. Why Is There So Little Money in Politics? Cambridge:  National Bureau of Economic Research. Arbiza, Sergio. 2003. Personal Interview. Former Blanco congressional and mayoral candidate for Artigas. Arellano, José P. 2001. La Reforma Educacional Chilena. Revista de la CEPAL 73 (April): 83–94. Artiga González, Álvaro. 2001. El Salvador. In Partidos políticos de América Latina:  Centroamérica, México y República Dominicana, edited by M. Alcántara Sáez and F. Freidenberg. Salamanca: Ediciones Universidad de Salamanca. Artiga González, Álvaro. 2002. Viejos y nuevos partidos políticos en El Salvador. Estudios Centroamericanos 57 (641/642): 253–74. Artiga González, Álvaro. 2004. El Salvador:  Maremoto electoral en 2004. Nueva Sociedad 192 (Julio-Agosto): 12–22. Auyero, Javier. 2001. Poor People’s Politics. Durham: Duke University Press. Avellaneda, Claudia N., and Maria Escobar-Lemmon. 2012. All By Myself: The Role of Personal Qualifications Versus Party Affiliation in Colombia Mayoral Elections. Latin American Politics and Society 54 (2): 109–32. Baker, Andy. 2009. The Market and the Masses in Latin America: Policy Reform and Consumption in Liberalizing Economies. New York: Cambridge University Press. Baker, Andy, Barry Ames, and Lucio Renno. 2006. Social Context and Campaign Volatility in New Democracies:  Networks and Neighborhoods in Brazil’s 2002 Elections. American Journal of Political Science 50 (2): 382–99. Balbontín, Ignacio. 2003. Personal Interview. Former DC Congressional Representative. Banco de Datos de la Facultad de Ciencias Sociales-UDELAR. Multiple years. Montevideo:  Area de Política y Relaciones Internacionales-FCS. accessed December 2013. Barnes, Samuel, and Max Kaase. 1979. Political Action:  Mass Participation in Five Western Democracies. Beverly Hills, CA: Sage.

Bibliography

343

Barozet, Emmanuelle. 2003. Movilización de recursos y redes sociales en los neopopulismos: hipótesis de trabajo para el caso chileno. Revista de Ciencia Política XXIII (1): 39–54. Barozet, Emmanuelle, and Marcel Aubry. 2005. De las reformas internas a la candidatura presidencial autónoma: Los nuevos caminos institucionales de Renovación Nacional. Revista Política 45: 165–97. Behrman, J. R., Nancy Birdsall, and Gunilla Pettersson. 2006. Inequality and Financial Reform in Latin America. Washington, DC: Center for Global Development. Bentancur, Nicolás. 2010. Aporte para una topografía de las políticas educativas en Uruguay:  Instituciones, ideas y actores. Revista uruguaya de ciencia política 21 (1): 65–92. Bizcarra, Darley. 2003. Personal Interview. Social Department Chair, Municipality of Artigas. Blofield, Merike, and Juan Pablo Luna. 2011. Public Opinion on Income Inequalities in Latin America. In The Great Gap: The Politics of Income Inequality in Latin America, edited by M. Blofield. University Park: Penn State University Press. Boas, Taylor C. 2010. Varieties of Electioneering: Success Contagion and Presidential Campaigns in Latin America. World Politics 62 (4): 636–75. Boix, Carles. 2003. Democracy and Redistribution. New  York:  Cambridge University Press. Bordachar, Rodrigo. 2003. Personal Interview, UDI Activists in La Florida. Borzutzky, Silvia. 2002. Vital Connections:  Politics, Social Security, and Inequality in Chile. South Bend, IN: Notre Dame University Press. Brass, Paul R. 1965. Factional Politics in an Indian State. Berkeley:  University of California Press. Brass, Paul R. 1993. The Rise of the BJP and the Future of Party Politics in Uttar Pradesh. In India Votes. The Quest for Consensus, 1989 and 1991, edited by Harold A. Gould and Sumit Ganguly. Boulder: Westview Press. Bruhn, Kathleen. 1997. Taking Goliath:  The Emergence of a New Left Party and the Struggle for Democracy in Mexico. University Park: Penn State University Press. Bucheli, Marisa, and Magdalena Furtado. 2005. Uruguay 1998–2002:  Income Distribution During the Crisis. CEPAL Review 86 (August): 161–75. Buquet, Daniel. 1999. Los cambios que aseguraron la continuidad: coalición, reforma, elecciones y después. In Elecciones 1999–2000, edited by Instituto de Ciencia Política. Montevideo: Ediciones de la Banda Oriental. Buquet, Daniel, ed. 2005. Las Claves del Cambio:  Ciclo electoral y nuevo gobierno 2004/2005. Montevideo: Banda Oriental, Instituto de Ciencia Política. Buquet, Daniel, Daniel Chasquetti, and Juan Andrés Moraes. 1998. Fragmentación política y gobierno en Uruguay: ¿un enfermo imaginario? Montevideo: Facultad de Ciencias Sociales. Buquet, Daniel, and Gustavo De Armas. 2004. Claves de la evolución electoral de la izquierda: crecimiento demográfico y moderación ideológica. In La izquierda uruguaya entre la oposición y el gobierno, edited by Jorge Lanzaro. Montevideo: Editorial Fin de Siglo.

344

Bibliography

Burgess, Katrina, and Steven Levitsky. 2003. Explaining Populist Party Adaptation in Latin America: Environmental and Organizational Determinants of Party Change in Argentina, Mexico, Peru, and Venezuela. Comparative Political Studies 36 (8): 881–911. Caetano, Gerardo, and José Rilla. 1995. Izquierda y tradición: Un problema y su versión en Uruguay. In La izquierda Uruguaya. Tradición, innovación y política, edited by Gerardo Caetano, José Rilla, and Javier Gallardo. Montevideo: Trilce. Caetano, Gerardo, and José Rilla. 2003. Los partidos políticos uruguayos en el siglo XX. In El Uruguay del siglo XX, edited by Benjamín Nahum. Montevideo: Ediciones de la Banda Oriental. Calvo, Ernesto, and Marcelo Escolar. 2005. La nueva política de partidos en la Argentina:  Crisis política, realineamientos partidarios y reforma electoral. Buenos Aires: Prometeo. Calvo, Ernesto, and Maria Victoria Murillo. 2007. How Many Clients Does it Take to Win an Election? Estimating the Size and Structure of Political Networks in Argentina and Chile. Paper read at the Elections and Distribution Workshop. New Haven: Yale University. Calvo, Ernesto, and Maria Victoria Murillo. n.d. Responsiveness and Responsibility. Book manuscript in progress. Cameron, Maxwell A., and Eric Hershberg. 2010. Latin America’s Left Turns: Politics, Policies, and Trajectories of Change. Boulder: Lynne Rienner. Campbell, Angus, Phillip Converse, Warren Miller, and Donald Stokes. 1960. The American Voter. New York: John Wiley and Sons. Campbell, Angus, Philip E. Converse, Warren Edward Miller, and Donald E. Stokes. 1980. The American Voter. Chicago: University of Chicago Press. Campello, Daniela. 2011. The Politics of Redistribution in Less Developed Democracies:  Evidence from Brazil, Ecuador, and Venezuela. In The Great Gap:  Inequality and the Politics of Redistribution in Latin America, edited by M. Blofield. University Park: Penn State Press. Caramani, Daniele. 2006. The Nationalization of Politics:  The Formation of National Electorates and Party Systems in Western Europe. New  York:  Cambridge University Press. Caravall, Eliana. 2003. Personal Interview, DC Congressmember in District 27. Carey, John, and Matthew S. Shugart. 1995. Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas. Electoral Studies 14 (4): 417–39. Carey, John M., and Peter Siavelis. 2003. El “seguro” para los subcampeones electorales y la sobrevivencia de la Concertación. Estudios Públicos 90 (Otoño): 5–27. Castiglioni, Rossana. 2005. The Politics of Social Policy Change in Chile and Uruguay:  Retrenchment versus Maintenance 1973–1998. New  York & London: Routledge. Castiglioni, Rossana. 2010. Las políticas sociales de la nueva (vieja) izquierda uruguaya. Woodrow Wilson Center Update on the Americas 6, pp. 1–12. Castro, Nora. 2002. Personal Interview, FA’s and MPP’s Congressmember for Montevideo.

Bibliography

345

Cavarozzi, Marcelo, Peter Cleaves, Manuel Antonio Garretón, Gary Gerreffi, and Jonathan Hartlyn. 2003. Latin America in the Twenty-First Century: Toward a New Sociopolitical Matrix. Boulder: Lyne Rienner. Cavedo, José Antonio. 2003. Personal Interview. Former Designated Mayor and Current Councilmember of San Ramón. CEPAL. 2010. Panorama Social de América Latina. Santiago de Chile: CEPAL. Cerdeña, Jorge. 2003. Personal Interview. Former Blanco Mayor of San José and Candidate of the Nueva Mayoría. Chacón, Ricardo. 1984. Las campañas de los partidos. ECA 426–427: 229–52. Chandra, Kanchan. 2007. Why Ethnic Parties Succeed:  Patronage and Ethnic Head Counts in India. New York: Cambridge University Press. Chasquetti, Daniel. 2007. Uruguay 2006:  Éxitos y dilemas del gobierno de izquierda. Revista de Ciencia Política Anuario Político de América Latina 2006 (Número Especial): 249–63. Chasquetti, Daniel. 2008. Uruguay 2007: El complejo año de las reformas. Revista de Ciencia Política 28 (1): 385–403. Chasquetti, Daniel. 2011. Financiamiento Político en Uruguay. In Financiamiento de los Partidos Políticos en América Latina, edited by Daniel Zovatto and Pablo Gutiérrez. México, DF: IDEA-OEA-UNAM. Chhibber, Pradeep. 2001. Democracy Without Associations. Transformation of the Party System and Social Cleavages in India. Ann Arbor: University of Michigan Press. Chhibber, Pradeep, and Ken Kollman. 2004. The Formation of National Party Systems: Federalism and Party Competition in Canada, Great Britain, India, and the United States. Princeton, NJ: Princeton University Press. Chile Transparente. 2009. Proyecto probidad:  Financiamiento político en Chile. Washington, DC: Inter-American Development Bank. Chiruchi, Juan. 2003. Personal Interview. Blanco Mayor of San José. Cofre, Margarita. 2003. Personal Interview. Pobladora from Lo Hermida in Peñalolén. Collier, David, and Ruth Berins Collier. 1991. Shaping the Political Arena. Princeton: Princeton University Press. Collier, David, and Steven Levitsky. 1997. Democracy with Adjectives:  Finding Conceptual Order in Recent Comparative Research. World Politics 49 (3): 430–51. Coppedge, Michael. 1998. The Dynamic Diversity of Latin American Party Systems. Party Politics 4 (4): 547–68. Córdova, Ricardo, and José Miguel Cruz. 2007. The Political Culture of Democracy in El Salvador. San Salvador: IUDOP-FUNDAUNGO-Vanderbilt University. Corte Electoral. Multiple Years. Montevideo:  Uruguay. accessed December 2013. Cortis, Hugo. 2003. Personal Interview. Political Advisor to a Colorado Congressmember. Cox, Gary W. 1987. Making Votes Count: Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press. Cox, Cristián. 2003. Las políticas educacionales de Chile en las últimas dos décadas del siglo XX. In Políticas educacionales en el cambio de siglo. Santiago:  Editorial Universitaria.

346

Bibliography

Cox, Trevor F., and Michael A. A. Cox. 2001. Multidimensional Scaling. Monographs on Statistics and Applied Probabilities Vol. 88. Boca Raton: CRC Press. Creceri, Juan. 2003. Personal Interview. Blanco Local Activists in Artigas. Cristi, María Angélica. 2003. Personal Interview. Former Mayor of Peñalolén and UDI Congressmember, Currently RN Congressmember. Cristi, Renato. 2000. El pensamiento político de Jaime Guzmán. Santiago:  LOM Ediciones. Dalton, Russell. 1988. Citizen Politics in Western Democracies. London: Chatham House. Dalton, Russell J., and Steven Weldon. 2007. Partisanship and Party System Institutionalization. Party Politics 13 (2): 179–96. Dargent, Eduardo, and Paula Muñoz. 2011. Democracy Against Parties? Party System Deinstitutionalization in Colombia. Journal of Politics in Latin America 3 (2): 43–71. Dávila, Mireya. 2005. Health Reform in Contemporary Chile: Does Politics Matter? MA Thesis. Department of Political Science, University of North Carolina at Chapel Hill. De la Maza, Gonzalo, Carlos Ochsenius, and Claudia Robles. 2004. Trayectorias, redes y poder: sociedad civil y política en Chile, 1990–2003. Lima: Departamento de Ciencias Sociales, Pontificia Universidad Católica del Perú. De León, Eduardo. 2004. Apuntes para un Balance de las Políticas Sociales en Uruguay. Montevideo: FESUR. Deininger, Klaus W., and Lyn Squire. 1996. A New Data Set Measuring Income Inequality. The World Bank Economic Review 10 (3): 565–91. Democratic Accountability and Linkages Project. 2010. Durham:  Duke University. accessed December 2013. Díaz Rioseco, Diego, Pilar Gianini, Juan Pablo Luna, and Rodrigo Nuñez. 2006. El Secreto de mi Éxito. Seis Caminos Para Llegar y Permanecer en Valparaiso. Revista de Ciencia Política 26 (1): 169–90. Diaz-Cayeros, Alberto, Federico Estévez, and Beatriz Magaloni. (n.d.) Poverty and Vote Buying: The Politics of Social Transfers in Mexico, 1989–2006. Book manuscript in progress. Stanford University. Dittborn, Julio. 2003. Personal Interview. UDI Congressmember. Dix, Robert. 1989. Cleavage Structures and Party Systems in Latin America. Comparative Politics, 22 (1): 23–37. Dixit, Avinash, and John Londregan. 1996. The determinants of success of special interests in redistributive politics. Journal of Politics, 58 (4): 1132–55. Doglio, Natalia, Luis Senatore, and Jaime Yaffé. 2004. Izquierda política y sindicatos en Uruguay. In La izquierda uruguaya entre la oposición y el gobierno, edited by J. Lanzaro. Montevideo: Fin de Siglo/FESUR. Domínguez, Jorge I. 1997. Technopols: Freeing Politics and Markets in Latin America in the 1990s. University Park: Penn State Press. Downs, Anthony. 1957. An Economic Theory of Democracy. New  York:  Harper and Row. Drake, Paul, and Iván Jaksic. 1991. The Struggle for Democracy in Chile: 1982–1990. Lincoln: University of Nebraska Press.

Bibliography

347

Drake, Paul, and Iván Jaksic. 1999. El modelo chileno. Democracia y desarrollo en los noventa. Santiago: LOM Ediciones. Duque, Jorge. 2003. Personal Interview. Colorado Congressmember for Canelones. Duverger, Maurice. 1951. The Influence of Electoral Systems on Political Life. International Social Science Bulletin 3: 314–52. Eaton, Kent. 2004. Politics Beyond the Capital: The Design of Subnational Institutions in South America. Stanford, CA: Stanford University Press. Eguizábal Mendoza, Cristina. 1989. Partidos, programas y proyectos políticos en El Salvador. Contribuciones, Universidad de Costa Rica, Facultad de Ciencias Sociales, Instituto de Investigaciones Sociales. Election Commission of India. Multiple years. accessed December 2013. Erikson, Robert, Michael Mackuen, and James Stimson. 2002. The Macro-Polity. New York: Cambridge University Press. Errandonea, Alfredo. 1989. Notas Sobre la Caracterización del Sistema de Partidos en el Uruguay. In Los partidos de cara al 90, edited by Instituto de Ciencia Política. Montevideo: FESUR-Fundación de Cultura Universitaria. Espinoza, Vicente. 1998. Historia social de la acción colectiva urbana: Los pobladores de Santiago, 1957–1987. EURE 24: 71–84. Espósito, Carla, and Walter Arteaga. 2006. Movimientos sociales urbano-populares en Bolivia: Una lucha contra la exclusion social, politica, y económica. La Paz: Fundación Carolina. Evans, Diana. 2004. Greasing the Wheels: Using Pork Barrel Projects to Build Majority Coalitions in Congress. New York: Cambridge University Press. Ewig, Christina. 2008. Reproduction, Reform and the Reconfigured State: Feminists and Neo-Liberal Health Reforms in Chile. In Beyond States and Markets: The Challenge of Social Reproduction, edited by I. Bakker and R. Silvey. New York: Routledge. Ewig, Christina, and Stephen J. Kay. 2011. Postentrenchment Politics: Policy Feedback in Chile’s Health and Pension Reforms. Latin American Politics and Society 53 (4): 67–99. Fa Robaina, Juan Carlos. 1972. Cartas a un diputado. Montevideo: Editorial Alfa S.A. Fairfield, Tasha. 2010. Business Power and Tax Reform: Taxing Income and Profits in Chile and Argentina. Latin American Politics and Society 52 (2): 37–71. Faleti, Tulia. 2010. Decentralization and Subnational Politics in Latin America. New York: Cambridge University Press. Fenno, Richard. 1978. Home Style: House Members in their Districts. New York: Longman. Filgueira, Carlos, and Fernando Filgueira. 1997a. Taming Market Reform: The Politics of Social State Reform in Uruguay. Paper read at Social Policies for the Urban Poor in Latin America. Welfare Reform in a Democratic Context, at Kellogg Institute, University of Notre Dame, September 12–14. Filgueira, Carlos, and Fernando Filgueira. 1997b. Coaliciones Reticentes:  Sistema Electoral, Partidos y Reforma en el Uruguay de Fin de Siglo. Cuadernos del CLAEH, 78–9: 325–47. Filgueira, Fernando. 2005. Welfare and Democracy in Latin America:  The Development, Crises, and Aftermath of Universal, Dual, and Exclusionary Social States. In UNRISD Project on Social Policy and Democracy. Geneva: UNRISD.

348

Bibliography

Filgueira, Fernando, Adolfo Garcé, Conrado Ramos, and Jaime Yaffé. 2003. Los dos ciclos del estado uruguayo en el siglo XX. In El Uruguay del siglo XX. La política. Montevideo: Ediciones de la Banda Oriental/Instituto de Ciencia Política. Filgueira, Fernando, and Rubén Kaztman. 2001. Panorama de la Infancia y la Familia. Montevideo: Universidad Católica del Uruguay. Filgueira, Fernando, and Jorge Papadopulos. 1996. ¿Sacando partido del conservadurismo? Crisis de larga duración y alternativas vedadas en Uruguay. Prisma (6): 109–46. Finch, Martin, Henry John, Benjamín Nahum, and José de Torres Wilson. 1980. Historia económica del Uruguay contemporáneo. Vol. 74. Montevideo: Ediciones de la Banda Oriental. Fiorina, Morris. 1981. Retrospective Voting in American Presidential Elections. New Haven, CT: Yale University Press. Fontaine Talavera, Arturo. 2000. Chile’s Elections: The New Face of the New Right. Journal of Democracy 11 (2): 70–7. Frankel, Francine R. 1978. India’s Political Economy, 1947–1977:  The Gradual Revolution. Princeton: Princeton University Press. Fuentes, Claudio. 1999. Partidos y Coaliciones en el Chile de los 90: Entre Pactos y Proyectos. In El modelo chileno. Democracia y desarrollo en los noventa, edited by P. Drake and I. Jaskic. Santiago de Chile: LOM Ediciones. Fundación Sol. 2011. El desalojo de la educación pública. In Ideas para el Buen Vivir:  Area Institucionalidad y Desarrollo, edited by M. Kremerman. Santiago de Chile: Fundación Sol. Funk, Robert. 2004. Renovation and Continuity:  The Transition to Democracy in Chile Revisited. PhD dissertation. The London School of Economics and Political Science, Department of Government, London. Gallardo, Javier. 1995. La izquierda uruguaya. La parábola de los “zorros” y los “leones”. In La izquierda Uruguaya. Tradición, innovación y política, edited by Gerardo Caetano, José Rilla, and Javier Gallardo. Montevideo: Trilce. Garcé, Adolfo. 2009. Panorama de la relación entre Think Tanks y partidos políticos en América Latina. In Dime a quién escuchas. Think Tanks y partidos políticos en América Latina, edited by Enrique Mendizabal and Kristen Sample. Buenos Aires: IDEA-ODI. Garcé, Adolfo, and Jaime Yaffé. 2004. La Era Progresista. Montevideo:  Editorial Fin de Siglo. García, “Turco.” 2002. Interview by Cecilia Blumetto, Clara Caponi, and Denisse Gelber. Executive Director of Centro Comunal Zonal, PCU (FA) activist. Montevideo, May 12. Garretón, Manuel Antonio. 1988. El Plebiscito y la Democracia. Santiago de Chile: FLACSO. Garretón, Manuel Antonio. 1989. The Chilean Political Process. Winchester: UNWIN HYMAN. Garretón, Manuel Antonio. 2000. Atavism and Democratic Ambiguity in the Chilean Right. In Conservative Parties, the Right, and Democracy in Latin America, edited by K.J. Middlebrook. Baltimore, MD: Johns Hopkins University Press.

Bibliography

349

Geddes, Barbara. 1994. Politicians’ Dilemma: Building State Capacity in Latin America. Berkeley: University of California Press. Giannini, Pilar, Hector Bahamonde, Juan Pablo Luna, Rodolfo López, and Gonzalo Recart. 2011. El secreto de mi éxito parte II:  Los caminos a Valparaíso en 2009. Revista de Ciencia Política 31 (2): 285–310. Gibson, Edward L. 1992. Conservative Electoral Movements and Democratic Politics: Core Constituencies, Coalition Building, and the Latin American Electoral Right. In The Right and Democracy in Latin America, edited by Douglas A. Chalmers, María Carmo Campbello de Souza, and Atilio A. Borón. New York: Praeger: 13–42. Gibson, Edward L. 1996. Class and Conservative Parties: Argentina in Comparative Perspective. Baltimore, MD: Johns Hopkins University Press. Gibson, Edward L. 1997. The Populist Road to Market Reform: Policy and Electoral Coalitions in Mexico and Argentina. World Politics 49 (3): 339–70. Gibson, Edward L. 2005. Boundary Control. World Politics 58 (1): 101–32. Gibson, Edward L., and Ernesto Calvo. 2001. Federalismo y sobre representación: la dimensión territorial de la reforma económica en Argentina. In El federalismo electoral argentino. Sobre representación, reforma política y gobierno dividido en Argentina, edited by Ernesto Calvo and Juan Manuel Abal Medina. Buenos Aires: INAP-Eudeba. Gibson, Edward L., and Julieta Suarez-Cao. 2010. Federalized Party Systems and Subnational Party Competition: Theory and an Empirical Application to Argentina. Comparative Politics 43 (1): 21–39. Gil, Federico. 1966. The Political System of Chile. Boston: Houghton Mifflin. Gillespie, Charles G. 1986. Uruguay’s Transition from Collegial Military-Technocratic Rule. In Transitions from Authoritarian Rule:  Prospects for Democracy, edited by Guillermo O’Donnell, Philippe Schmitter, and Laurence Whitehead. Baltimore, MD: Johns Hopkins University Press. Girardi, Guido. 2003. Personal Interview. PPD Senator. Goertz, Gary, and James Mahoney. 2012. A Tale of Two Cultures:  Qualitative and Quantitative Research in the Social Sciences. Princeton: Princeton University Press. Goldfrank, Benjamin. 2006. Los Procesos de “Presupuesto Participativo” en América Latina: Exito, Fracaso y Cambio. Revista de Ciencia Política XXVI (2): 3–28. González, Eugenio. 2003. Personal Interview. Lavin’s Campaign Advisor and UDI Activist. González, Luis E. 1991. Political Structures and Democracy in Uruguay. Notre Dame: University of Notre Dame Press. González, Luis E. 1995. Continuity and Change in the Uruguayan Party System. In Building Democratic Institutions:  Party Systems in Latin America, edited by Scott Mainwaring and Timothy Scully, Stanford: Stanford University Press. González, Luis E., and Rosario Queirolo. 2000. Las elecciones nacionales del 2004: Posibles escenarios. In Elecciones 1999–2000, edited by Instituto de Ciencia Política. Montevideo: Ediciones de la Banda Oriental. González-Ocantos, Ezequiel, Chad Kiewiet de Jonge, Carlos Meléndez, Javier Osorio, and David W. Nickerson. 2011. Vote Buying and Social Desirability

350

Bibliography

Bias: Experimental Evidence from Nicaragua. American Journal of Political Science 56 (6): 2002–217. Goodin, Robert, and John Dryzek. 1980. Rational Participation: The Politics of Relative Power. British Journal of Political Science 10 (3): 273–92. Graham, Bruce Desmond. 1990. Hindu Nationalism and Indian Politics: The Origins and Development of the Bharatiya Jana Sangh. Cambridge: Cambridge University Press. Grindle, Merilee Serrill. 2012. Jobs for the Boys: Patronage and the State in Comparative Perspective. Cambridge, MA: Harvard University Press. Grupo RADAR. 2004. El fenómeno Mujica. accessed May 2006. Grzymala-Busse, Anna M. 2002. Redeeming the Communist Past: The Regeneration of Communist Parties in East Central Europe. New York: Cambridge University Press. Guerra, Sergio. 2003. Personal Interview. RN council member in Peñalolén. Guerrini, Aldo. 2000. El peso politico de los intendentes del interior:  Apuntes para una discusion a la luz de las ultimas consultas electorales. In Elecciones 1999–2000, edited by Instituto de Ciencia Política. Montevideo: Ediciones de la Banda Oriental. Guha, Ramachandra. 2008. India After Gandhi:  A  History of the World’s Largest Democracy. New York: Harper Collins. Hagopian, Frances. 1998. Democracy and Political Representation in Latin America in the 1990s:  Pause, Reorganization, or Decline? In Fault Lines of Democracy in Post-transitional Latin America, edited by Felipe Agüero and Jeffrey Stark. Miami: North-South Center Press, University of Miami. Hagopian, Frances. 2002. Economic Liberalization, Political Competition, and Political Representation in Latin America. Paper read at Mellon Foundation Sawyer Seminar, Crises, Choices, and Change: The Micro-Foundations of the Neo-liberal Turn in Latin America, December 2, Tulane University. Hagopian, Frances. 2009. Religious Pluralism, Democracy, and the Catholic Church in Latin America. South Bend, IN: University of Notre Dame Press. Hagopian, Frances (n.d.). Reorganizing Political Representation in Latin America: Parties, Program, and Patronage in Argentina, Brazil, Chile, and Mexico. Book manuscript in progress. Harvard University. Hagopian, Frances, Carlos Gervasoni, and Juan Andrés Moraes. 2009. From Patronage to Program:  The Emergence of Party-Oriented Legislators in Brazil. Comparative Political Studies 42 (3): 360–91. Harbers, Imke. 2010. Political Organization in Multi-level Settings: Mexican and Latin American Parties and Party Systems after Decentralization. Faculty of Social and Behavioural Sciences, Leiden University. Hartlyn, Jonathan. 1988. The Politics of Coalition Rule in Colombia, Cambridge Latin American Studies 66. Cambridge and New York: Cambridge University Press. Hawkins, Kirk A. 2010. Venezuela’s Chavismo and Populism in Comparative Perspective. New York: Cambridge University Press. Hazan, Reuven Y., and Gideon Rahat. 2010. Democracy Within Parties:  Candidate Selection Methods and their Political Consequences. New  York:  Oxford University Press.

Bibliography

351

Hernández, Sergio. 2003. Personal Interview. DC Councilmember of Vitacura. Herrera, Nicanor. 2003. Personal Interview. RN Councilmember in La Florida. Hicken, Allen. 2009. Building Party Systems in Developing Democracies. New York: Cambridge University Press. Hinnich, Melvin, and Michael Munger. 1994. Ideology and the Theory of Political Choice. Ann Arbor: Michigan University Press. Hipsher, Patricia. 1996. The New Electoral Right in Chile and the Poor:  Strange Bedfellows. South Eastern Latin Americanist 39 (3/4): 17–34. Holland, Alisha. Forthcoming. Right on Crime? Conservative Party Politics and Mano Dura Policies in El Salvador. Latin American Research Review. Huber, Evelyne, Francois Nielsen, Jenny Pribble, and John D. Stephens. 2006. Politics and Inequality in Latin America and the Caribbean. American Sociological Review 71(6): 943–63. Huber, Evelyne, Jennifer Pribble, and John D. Stephens. 2010. The Chilean Left in Power. In Leftist Governments in Latin America: Successes and Shortcomings, edited by Kurt Weyland, Raúl L. Madrid, and Wendy Hunter. New  York:  Cambridge University Press. Huber, Evelyne, and John D. Stephens. 2012. Democracy and the Left: Social Policy and Inequality in Latin America: Chicago: University of Chicago Press. Huneeus, Carlos. 1998. Malestar y desencanto en Chile. Legados del autoritarismo y costos de la transición. Santiago: Papeles de Trabajo-Programa de Estudios Prospectivos. Huneeus, Carlos. 2000a. El Régimen de Pinochet. Santiago: Editorial Sudamericana. Huneeus, Carlos. 2000b. Technocrats and Politicians in an Authoritarian Regime: The “ODEPLAN Boys” and the “Gremialists” in Pinochet’s Chile. Journal of Latin American Studies 32 (2): 461–501. Huneeus, Carlos. 2001a. La Derecha en el Chile Después de Pinochet: El Caso de la Unión Demócrata Independiente. In The Kellogg Institute Working Papers. South Bend, IN: Notre Dame University. Huneeus, Carlos. 2001b. Organización y liderazgo en los partidos políticos. Un análisis comparado de Alemania, España y Chile. Revista de Ciencia Política XXI (1): 107–34. Huneeus, Carlos. 2003. A Highly Institutionalized Political Party: Christian Democracy in Chile. In Christian Democracy in Latin America: Electoral Competition and Regime Conflicts, edited by Scott Mainwaring and Timothy R. Scully. Stanford:  Stanford University Press. Huneeus, Carlos. 2006. La reforma al sistema binominal en Chile: Una contribución al debate. Santiago: Konrad Adenauer Stiftung. Huneeus, Carlos. 2007. The Pinochet Regime. Boulder, CO: Lynne Rienner Publishers. Huneeus, Carlos, Fabiola Berríos, and Ricardo Gamboa. 2007. Las elecciones chilenas de 2005: Partidos, coaliciones y votantes en transición. Santiago: Catalonia. Hunter, Wendy. 1997. Eroding Military Influence in Brazil: Politicians Against Soldiers. Chapel Hill: University of North Carolina Press. Hunter, Wendy. 2010. The Transformation of the Workers’ Party in Brazil, 1989–2009. Cambridge: Cambridge University Press.

352

Bibliography

Hunter, Wendy. 2011. Brazil: The PT in Power. In The Resurgence of the Latin American Left, edited by Steven Levitsky and Kenneth M. Roberts. Baltimore, MD:  Johns Hopkins University Press. Hunter, Wendy, and Natasha Borges Sugiyama. 2011. Documenting Citizenship:  Contemporary Efforts Toward Social Inclusion in Brazil. Anual Meeting Paper, APSA 2011. Hunter, Wendy, and Timothy Power. 2005. Lula’s Brazil at Midterm. Journal of Democracy 16 (3): 127–39. INE (Instituto Nacional de Estadísticas). Multiple years. Santiago de Chile. Inglehart, Ronald. 1984. The Changing Structure of Political Cleavages in Western Society. In Electoral Change in Advanced Industrial Democracies:  Realignment or Dealignment? edited by Russell Dalton, Scott Flanagan, and Paul A. Beck. Princeton: Princeton University Press. Instituto Interamericano de Derechos Humanos. 2005. Paper read at seminar, Evaluación de la puesta en marcha de la Ley sobre Transparencia, Control, y Límite al Gasto Electoral, Santiago de Chile. Jackson, Giorgio. 2013. El país que soñamos. Santiago de Chile: Feria Chilena del Libro. Jara, José. 2003. Personal Interview. UDI’s 2001 Congressional Candidate in La Florida. Jiménez, Raúl. 2003. Personal Interview, Blanco Local Activist in Artigas. Jocelyn-Holt, Tomás. 2003. Personal Interview. Former DC Congressmember. Joignant, Alfredo, and Patricio Navia. 2003. De la Política de Individuos a los Hombres del Partido. Socialización, Competencia Política y Penetración Electoral de la UDI (1989–2001). Estudios Públicos 89 (Verano): 129–71. Jones, Mark. 2005. The Role of Parties and Party Systems in the Policymaking Process. Manuscript. Washington, DC: Inter-American Development Bank. Katz, Richard S., and Peter Mair. 1995. Changing Models of Party Organization and Party Democracy:  The Emergence of the Cartel Party. Party Politics 1 (1): 5–28. Kaufman, Robert R. 2009. The Political Effects of Inequality in Latin America: Some Inconvenient Facts. Comparative Politics 41 (3): 359–79. Kaztman, Rubén. 1999. Activos y estructura de oportunidades: estudios sobre las raíces de la vulnerabilidad social en Uruguay. Montevideo: CEPAL-UNDP. Keck, Margaret E. 1992. The Worker’s Party and Democratization in Brazil. New Haven, CT: Yale University Press. Keefer, Philip. 2008. Insurgency and Credible Commitment in Autocracies and Democracies. The World Bank Economic Review 22 (1): 33–61. Kitschelt, Herbert. 1989. The Internal Politics of Parties:  The Law of Curvilinear Disparity Revisited. Political Studies XXXVII: 400–21. Kitschelt, Herbert. 1994. The Transformation of European Social Democracy. New York: Cambridge University Press. Kitschelt, Herbert. 2000. Linkages between Citizens and Politicians in Democratic Polities. Comparative Political Studies 33 (6–7): 845–79. Kitschelt, Herbert. et  al. n.d. Programmatic Parties and Party Systems. Manuscript. Stockholm: International IDEA.

Bibliography

353

Kitschelt, Herbert, Kirk Hawkins, Juan Pablo Luna, Guillermo Rosas, and Elizabeth-Jean Zechmeister. 2010. Latin American Party Systems. New  York:  Cambridge University Press. Kitschelt, Herbert, Zdenka Mansfeldova, Radoslaw Markowski, and Gábor Tóka. 1999. Post-Communist Party Systems: Competition, Representation, and Inter-Party Cooperation. Cambridge: Cambridge University Press. Kitschelt, Herbert, and Daniel M. Kselman. 2011. Do Institutions Matter For Parties’ Electoral Linkage Strategies? Annual Meeting Paper APSA 2011. Kitschelt, Herbert, and Matthew Singer. 2011. Do Everything (DoE) Parties:  When Can Politicians Combine Clientelistic and Programmatic Strategies. Annual Meeting Paper APSA 2011. Kitschelt, Herbert, and Steven Wilkinson. 2007. Patrons, Clients and Policies: Patterns of Democratic Accountability and Political Competition. New  York:  Cambridge University Press. Klein, Marcus. 2004. The Unión Demócrata Independiente and the Poor (1983– 1992):  The Survival of Clientelistic Traditions in Chilean Politics. Jahrbuch fur Geschichte Lateinamerikas(JbLA) 41: 301–24. Kohli, Atul. 1991. Democracy and Discontent: India’s Growing Crisis of Governability. Cambridge: Cambridge University Press. Koivumaeki, Riitta-Ilona. 2009. Business, Economic Experts, and Conservative Party Building in Latin America:  The Case of El Salvador. Journal of Politics in Latin America 2 (1): 79–106. Koivumaeki, Riitta-Ilona. 2014. Societal Cleavages and Strategic Elites: Explaining the Success of the Right in El Salvador. In The Resilience of the Latin American Right, edited by Juan Pablo Luna and Cristóbal Rovira-Kaltwasser. Baltimore, MD: Johns Hopkins University Press. Komadina, Jorge, and Céline Geffroy Komadina. 2007. El poder del movimiento político:  Estrategia, tramas organizativas e identidad del MAS en Cochabamba (1999–2005). Vol. 5. La Paz: Fundacion Pieb. Kurtz, Marcus J. 2004. The Dilemmas of Democracy in the Open Economy: Lessons from Latin America. World Politics 56 (2): 262–302. Lacalle Pou, Luis. 2003. Personal Interview. Blanco Congressmember for Canelones. Lanzaro, Jorge. 1986. Sindicatos y sistema político. Relaciones corporativas en el Uruguay. 1940–1985. Montevideo: Fundación de Cultura Universitaria. Lanzaro, Jorge. ed. 2004. La Izquierda Uruguaya:  Entra la Oposición y el Gobierno. Montevideo: Editorial Fin de Siglo. LAPOP Survey. Multiple Years. Nashville, TN:  Vanderbilt UniversityAmericasBarometer. accessed December 2013. Latinobarómetro. 2004. Santiago: Corporación Latinobarómetro. accessed December 2013. Laurnaga, Maria Elena, and Aldo U. Guerrini. 1994. Del “Buen Vecino” and “Intendente Emprendedor”: El Rol de los Intendentes Departamentales en la Reforma del Estado. Revista Uruguaya de Ciencia Política 7: 83–97.

354

Bibliography

Lazar, Sian. 2006. El Alto, Ciudad Rebelde: Organisational Bases for Revolt. Bulletin of Latin American Research 25 (2): 183–99. Lazarsfeld, Paul F., Bernard Berelson, and Hazel Gaudet. 1944. The People’s Choice: How the Voter Makes Up His Mind in a Presidential Campaign. New  York:  Columbia University Press. Lazo, Carmén. 2003. Personal Interview. Former PS Congressmember. Lehoucq, Fabrice. 2005. Costa Rica:  Paradise in Doubt. Journal of Democracy 16 (3): 140–54. Less, Martin. 2005. Personal Interview. Personal Secretary of Mayor Carámbula, Canelones. Levitsky, Steven. 2001. Inside the Black Box: Recent Studies in Latin American Party Organizations. Studies in Comparative International Development 36 (2): 92–110. Levitsky, Steven. 2003. Transforming Labor-based Parties in Latin America. New York: Cambridge University Press. Levitsky, Steven, and Kenneth M. Roberts. 2011. The Resurgence of the Latin American Left. Baltimore, MD: Johns Hopkins University Press. Lipset, Seymour M. 1965. Estudiantes universitarios y politica en el tercer mundo. Montevideo: Editorial Alfa. Lindhal, Göran. 1962. Uruguay’s New Path:  A  Study in Politics During the First Colegiado, 1919–33. Doctoral Dissertation. Stockholm: University of Stockholm. Lipset, Seymour M., and Stein Rokkan, eds. 1967. Party Systems and Voter Alignments: Cross-national Perspectives. New York: Free Press. Londregan, John B. 2000. Legislative Institutions and Ideology in Chile. New York: Cambridge University Press. López, Elida. 2003. Personal Interview. Colorado Local Activist in Montevideo. López-Calva, Luis F., and Nora Lustig. 2010. Declining Inequality in Latin America: A Decade of Progress? Baltimore, MD: Brookings Institute Press. Loveman, Brian. 1976. Struggle in the Countryside: Politics and Rural Labor in Chile, 1919–1973. Bloomington: Indiana University Press. Loxton, James. 2014. The Authoritarian Roots of New Right Party Success in Latin America. In The Resilience of the Latin American Right, edited by Juan Pablo Luna and Cristóbal Rovira-Kaltwasser. Baltimore, MD: Johns Hopkins University Press. Luebbert, Gregory M. 1991. Liberalism, Fascism, or Social Democracy:  Social Classes and the Political Origins of Regimes in Interwar Europe. New York: Oxford University Press. Luna, Juan Pablo. 2004. De familias y parentescos políticos. Ideología y competencia electoral en el Uruguay contemporáneo. In La Izquierda Uruguaya: Entre el gobierno y la oposición, edited by Jorge Lanzaro. Montevideo: Editorial Fin de Siglo. Luna, Juan Pablo, and Elizabeth J. Zechmeister. 2005. Political Representation in Latin America: A Study of Elite-Mass Congruence in Nine Countries. Comparative Political Studies 38 (4): 388–416. Luna, Juan Pablo. 2006. Programmatic and Non-programmatic Linkages in Two Institutionalized Party-systems:  Chile and Uruguay in Comparative Perspective. Doctoral Dissertation. Chapel Hill: University of North Carolina at Chapel Hill.

Bibliography

355

Luna, Juan Pablo. 2007. Frente Amplio and the Crafting of a Social-Democratic Alternative in Uruguay:  Electoral Success and Forthcoming Challenges. Latin American Politics and Society 49 (4): 1–30. Luna, Juan Pablo. 2008. Partidos Políticos y Sociedad en Chile: Trayectoria Histórica y Mutaciones Recientes. In Reforma de los Partidos Políticos en Chile, edited by Arturo Fontaine, Ignacio Walter, Jorge Navarrete and Cristián Larroulet. Santiago: PNUD-Chile. Luna, Juan Pablo. 2010. Segmented Party Voter Linkages in Latin America: The Case of the UDI. Journal of Latin American Studies 42 (2): 325–56. Luna, Juan Pablo, and David Altman. 2011. Uprooted But Stable:  Chilean Parties and the Concept of Party System Institutionalization. Latin American Politics and Society 53 (2): 1–28. Luna, Juan Pablo, and Rodrigo Mardones. 2010. Chile: Are the Parties Over? Journal of Democracy, 21 (3): 107–21. Luna, Juan Pablo, and Fernando Rosenblatt. 2012. ¿Notas para una autopsia? Los partidos políticos en el Chile actual. In Democracia con partidos. Informe para la reforma de los partidos políticos en Chile, edited by Francisco J. Díaz and Lucas Sierra. Santiago: CEP-Cieplan. Luna, Juan Pablo, Fernando Rosenblatt, and Sergio Toro. n.d. Desk Review on Programmatic Parties. Mimeo, Stockholm: IDEA-International. Luna, Juan Pablo, and Sergio Toro. 2011. The Chilean elections of December 2009 and January 2010. Electoral Studies 30 (1): 226–30. Lupia, Arthur, Mathew D. McCubbins, and Samuel L. Popkin, eds. 2000. Elements of Reason: Cognition, Choice, and the Bounds of Rationality. Cambridge: Cambridge University Press. Lupu, Noam. 2010. Brand Dilution, Partisanship, and the Breakdown of Political Parties in Latin America. Presented at the conference Political Parties in the Developing World. Princeton University. Lupu, Noam, and Jonas Pontusson. 2011. The Structure of Inequality and the Politics of Redistribution. American Political Science Review 105 (2): 316–36. Lustig, Nora, Luis F. Lopez-Clava, and Eduardo Ortiz-Juarez. 2013a. Deconstructing the Decline in Inequality in Latin America. In Working Papers, Tulane University Department of Economics. Lustig, Nora, Carola Pessino, and John Scott. 2013b. The Impact of Taxes and Social Security Spending on Inequality and Poverty in Argentina, Bolivia, Brazil, Mexico, Peru, and Uruguay: An Overview. In CEQ Working Paper: Commitment to Equity. Lynch, Julia. 2006. Age in the Welfare State: The Origins of Social Spending on Pensioners, Workers, and Children. New York: Cambridge University Press. Lyne, Mona M. 2008. The Voter’s Dilemma and Democratic Accountability:  Latin America and Beyond. University Park: Penn State University Press. Madrid, Raul. 2008. The Rise of Ethnopopulism in Latin America. World Politics 60 (3): 475–508. Madrid, Raul. 2012. The Rise of Ethnic Politics in Latin America. New York: Cambridge University Press.

356

Bibliography

Magaloni, Beatriz, Alberto Diaz-Cayeros, and Federico Estevez. 2007. Clientelism and Portfolio Diversification: A Model of Electoral Investment with Applications to Mexico. In Patrons, Clients, and Policies: Patterns of Democratic Accountability, edited by Herbert Kitschelt and Steven Wilkinson. New York: Cambridge University Press. Magri, Altair. 2000. La llave del triunfo electoral en Mayo 2000: Reelección, continuidad y “por la vuelta.” In Elecciones 1999–2000, edited by Instituto de Ciencia Política. Montevideo: Ediciones de la Banda Oriental. Mahía, José. 2003. Personal Interview. FA Congressmember for Canelones. Mainwaring, Scott. 1999. Rethinking Party Systems in the Third Wave of Democratization: The Case of Brazil. Stanford: Stanford University Press. Mainwaring, Scott, Ana María Bejarano, and Eduardo Pizarro. 2006. The Crisis of Democratic Representation in the Andes. Stanford: Stanford University Press. Mainwaring, Scott, and Mark Jones. 2003. The Nationalization of Parties and Party Systems. Party Politics 9 (2): 139–66. Mainwaring, Scott, Rachel Meneguello, and Timothy Joseph Power. 1999. Conservative Parties, Democracy, and Economic Reform in Contemporary Brazil. Hellen Kellogg Institute for International Studies, Working Paper 264. Mainwaring, Scott, and Timothy R. Scully, eds. 1995. Building Democratic Institutions. Stanford: Stanford University Press. Mainwaring, Scott, and Timothy R. Scully, eds. 2010. Democratic Governance in Latin America. Stanford: Stanford University Press. Mainwaring, Scott, and Edurne Zocco. 2007. Political Sequences and the Stabilization of Interparty Competition. Party Politics 13 (2): 155–78. Mair, Peter, ed. 1990. The West European Party Systems. Oxford: Oxford University Press. Martín Baro, Ignacio. 1982. El llamado de la extrema derecha. ECA 37: 403–404/453–466. Martínez, Javier, and Alvaro H. Díaz. 1996. Chile:  The Great Transformation. Washington, DC: Brookings Institution/UNRISD. Marx, Karl. 1978 (1844). On the Jewish Question. In The Marx-Engels Reader, edited by R.C. Tucker. New York: Norton. Mathus, Isabel. 2003. Personal Interview. DC Councilmember in Cerro Navia. Mayhew, David. 1974. Congress:  The Electoral Connection. New Haven, CT:  Yale University Press. McGuire, James W. 1997. Peronism Without Perón: Unions, Parties, and Democracy in Argentina. Stanford: Stanford University Press. Melgarejo, Artigas. 2003. Personal Interview. FA Congressmember for Canelones. Meltzer, Allan, and Scott Richard. 1981. A Rational Theory of the Size of Government. Journal of Political Economy 89: 914–27. Meneghello, Rachel. 1989. PT: A Formação de um Partido. Rio de Janeiro: Paz e Terra. Mesa-Lago, Carmelo. 1997. Social Welfare Reform in the Context of Economic-Political Liberalization: Latin American Cases. World Development 25 (4): 497–517. Middlebrook, Kevin J., ed. 2000. Conservative Parties, the Right and Democracy in Latin America. Baltimore, MD: The Johns Hopkins University Press. Mieres, Pablo. 1994. El Voto: Desobediencia y Lealtad Ciudadana. Montevideo: Fin de Siglo-CLAEH.

Bibliography

357

Miller, George A, Eugene Galanter, and Karl Pribram. 1960. Plans and the Structure of Behavior. New York: Holt, Rinehart and Winston. Miller, Warren Edward, and J. Merrill Shanks. 1996. The New American Voter. Cambridge: Harvard University Press. Mizala, Alejandra, and Pilar Romaguera. 2005. Teacher’s Salary Structure and Incentives in Chile. Centro de Economía Aplicada, Universidad de Chile. Documentos de Trabajo 193. Montero, Alfred P. 2012. A Reversal of Political Fortune: The Transitional Dynamics of Conservative Rule in the Brazilian Northeast. Latin American Politics and Society 54 (1): 1–36. Montes, Carlos. 2003. Personal Interview. Current Congressmember of La Florida. Montes, Juan Esteban, Scott Mainwaring, and Eugenio Ortega. 2000. Rethinking the Chilean Party System. Journal of Latin American Studies 32 (3): 795–824. Morales, Mauricio. 2008. La primera mujer presidenta de Chile: ¿Qué explicó el triunfo de Michelle Bachelet en las elecciones de 2005–2006? Latin American Research Review 43 (1): 7–32. Morales, Mauricio, and Rodrigo Burgueño. 2001. La UDI como expresión de la Nueva Derecha en Chile. Estudios Sociales 107: 215–48. Moreira, Constanza. 2000a. La Izquierda en Uruguay y Brasil:  Cultura Política y Desarrollo Político-Partidario. In La Larga Espera. Itinerarios de las Izquierdas en Argentina, Brasil y Uruguay, edited by Susana Mallo and Constanza Moreira. Montevideo: Ediciones de la Banda Oriental. Moreira, Constanza. 2000b. Las paradojales a del fin de siglo uruguayo: Comportamiento electoral y cultura política. In Elecciones 1999–2000, edited by Instituto de Ciencia Política. Montevideo: Ediciones de la Banda Oriental. Moreira, Constanza. 2005. El voto moderno y el voto clasista revisado:  Explicando el desempaño electoral de la izquierda en las elecciones de 2004 en Uruguay. In Las Claves del Cambio: Ciclo electoral y nuevo gobierno 2004/2005, edited by Daniel Buquet. Montevideo: Ediciones de la Banda Oriental, Instituto de Ciencia Política. Moreno, Alejandro. 1999. Political Cleavages: Issues, Parties, and the Consolidation of Democracy. Boulder, CO: Westview Press. Morgan, Jana. 2011. Bankrupt Representation and Party System Collapse. University Park: Penn State University Press. Morgenstern, Scott, and Richard F. Potthoff. 2005. The Components of Elections: District Heterogeneity, District-Time Effects, and Volatility. Electoral Studies 24 (1): 17–40 Moulián, Tomás. 1984. Tensiones y crisis política: Análisis de la década del sesenta. Centro de Estudios del Desarrollo Working Paper 17: 35. Moulián, Tomás. 1985. Violencia, gradualismo y reformas en el desarrollo político chileno. In Estudios sobre el sistema de partidos en Chile, edited by Angel Flisfisch, Tomás Moulián, and Adolfo Aldunate B. Santiago: FLACSO. Moulián, Tomás. 2004. De la política letrada a la política analfabeta:  la crisis de la política en el Chile actual y el “lavinismo.” Santiago: Lom Ediciones. Munck, Gerardo and Jeffrey Bosworth. 1998. Patterns of Representation and Competition:  Parties and Democracy in Post-Pinochet Chile. Party Politics 4 (4): 471–93.

358

Bibliography

Navia, Patricio. 2003. You Select the Rules of the Game and You Lose? Advantages and Constraints When Choosing Electoral Rules: The Case of Chile. Doctoral Dissertation, Political Science, New York University, New York. Navia, Patricio. 2005. Transformando votos en escaños:  leyes electorales en Chile, 1833–2003. Política y Gobierno XII (2): 233–76. Navia, Patricio. 2008. Political Recruitment and Candidate Selection in Chile (1990– 2003):  The Executive Branch. In Pathways to Power:  Political Recruitment and Candidate Selection in Latin America, edited by Peter Siavelis and Scott Morgenstern. New York: Cambridge University Press. Navia, Patricio and Alfredo Joignant. 1999. Las elecciones presidenciales de 1999: la participación electoral y el nuevo votante chileno. In Chile 1999–2000. Nuevo Gobierno: desafíos de la reconciliación. Santiago: FLACSO. Nichter, Simeon. 2008. Vote Buying or Turnout Buying? Machine Politics and the Secret Ballot. American Political Science Review 102 (1): 19–31. Norris, Pippa. 2009. Democracy Crossnational Data, Version 3.0. Observatorio Político Electoral de la Universidad Diego Portales. Multiple years.

accessed December 2013. O’Donnell, Guillermo. 1973. Modernization and Bureaucratic-Authoritarianism. Berkeley, CA: University of California Press. O’Donnell, Guillermo. 1994. Delegative Democracy. Journal of Democracy 5 (1): 55–69. Olivares, Carlos. 2003. Personal Interview. DC Congressmember. Ondetti, Gabriel A. 2008. Land, Protest, and Politics: The Landless Movement and the Struggle for Agrarian Reform in Brazil. University Park: Penn State University Press. Ortega, Eugenio. 2003. Los partidos políticos chilenos: cambio y estabilidad en el comportamiento electoral 1990–2000. Revista de Ciencia Política 23 (2): 109–47. Ostiguy, Pierre. 2009. Argentina’s Double Political Spectrum: Party System, Political Identities, and Strategies 1944–2007. Kellogg Institute Working Paper 361. Oxhorn, Phillip. 1995. Organizing Civil Society: The Popular Sectors and the Struggle for Democracy in Chile. University Park: Penn State University Press. Pais, Ronald. 2003. Personal Interview. Colorado Congressmember for Montevideo. Palacios, Indira. 2010. From Militancy to Clientelism:  Labor Union Strategies and Membership Trajectories in Contemporary Chile. Latin American Politics and Society 52 (2): 73–102. Panebianco, Angelo. 1988. Political Parties:  Organization and Power. Cambridge: Cambridge University Press. Panizza, Francisco. 1990. Uruguay, Batllismo y Después: Pacheco, Militares y Tupamaros en la Crisis del Uruguay Batllista. Montevideo: Ediciones de la Banda Oriental. Paolillo, Claudio. 2004. Con los días contados. Montevideo: Editorial Fin de Siglo. Pasotti, Eleonora. 2010. Political Branding in Cities: The Decline of Machine Politics in Bogota, Naples, and Chicago. New York: Cambridge University Press. PELA. Multiple years. Salamanca: Observatorio de Elites Parlamentarias en América Latina, Instituto de Iberoamérica, Universidad de Salamanca. accessed December 2013. Percovich, Margarita. 2002. Personal Interview. FA’s Congressmember.

Bibliography

359

Pérez Antón, Romeo. 1988. Cuatro antagonismos sucesivos: La concreta instauración de la democracia uruguaya. Revista de Ciencia Política 2: 41–52. Pérez, Lily. 2003. Personal Interview. Former Council Member and Current Congressmember of La Florida. Perez, Omar. 2002. Personal Interview. Municipal Secretary of Río Puelo. Piattoni, Simona, ed. 2001. Clientelism, Interests, and Democratic Representation: The European Experience in Historical and Comparative Perspective. Cambridge: Cambridge University Press. Pierson, Paul. 1993. When Effect Becomes Cause. World Politics 42 (4): 595–628. Piñeiro, Rafael. 2007. El sueño de la lista propia: los dilemas de coordinación electoral post-reforma de 1997. Revista Uruguaya de Ciencia Política 16: 51–71. Piñeiro, Rafael, and Fernando Rosenblatt. 2011. The Missing Link: Politics and Political Interest in Unequal Societies. Manuscript. AmericasBarometer. Piñeiro, Rafael, and Jaime Yaffé. 2004. El Frente Amplio por dentro. Las fracciones frenteamplistas 1971–1999. In La Izquierda Uruguaya entre la oposición y el gobierno, edited by Jorge Lanzaro. Montevideo: Editorial Fin de Siglo. Pita, Carlos. 2003. Personal Interview. FA’s Congressmember from Montevideo. Plumb, David. 1998. El Partido por la Democracia: The Birth of Chile’s Postmaterialist Catch-All Left. Party Politics 4 (January): 93–106. Pollack, Marcelo. 1999. The New Right in Chile 1973–1997. London: Macmillan. Posner, Paul. 1999. Popular Representation and Political Disattisfaction in Chile´s New Democracy. Journal of Interamerican Studies and World Affairs 41 (1): 59–85. Posner, Paul. 2004. Local Democracy and the Transformation of Popular Participation in Chile. Latin American Politics and Society 46 (3): 55–81. Posner, Paul. 2008. State, Market, and Democracy in Chile: The Constraint of Popular Participation. New York: Palgrave Macmillan. Powell, G. Bingham Jr. 2004. Political Representation in Comparative Politics. Annual Review of Political Science 7: 273–96. Pribble, Jennifer. 2013. Welfare and Party Politics in Latin America. New York: Cambridge University Press. Przeworski, Adam. 2010. Democracy and the Limits of Self-government. Vol. 9. New York: Cambridge University Press. Przeworski, Adam, and John Sprague. 1986. Paper Stones:  A  History of Electoral Socialism. Chicago: Chicago University Press. Przeworski, Adam, and Michael Wallerstein. 1988. Structural Dependence of the State on Capital. American Political Science Review 82 (1): 11–29. Queirolo, Rosario. 1999. La tradicionalización del Frente Amplio. La conflictividad del proceso de cambio. In Los partidos políticos uruguayos en tiempos de cambios, edited by Luis Eduardo González. Montevideo:  Universidad Católica del Uruguay-Fundación Bank-Boston-Fundación de Cultura Universitaria. Rabinowitz, David. 2003. Personal Interview. FA Local Activist in San José. RADAR, Grupo. 2004. Continúa el predominio del MPP en la interna del EP-FA 2004. accessed October 2005. Ragin, Charles. 1987. The Comparative Method:  Moving beyond Qualitative and Quantitative Strategies. Berkeley: University of California Press.

360

Bibliography

Ragin, Charles. 2008. Redesigning Social Inquiry:  Fuzzy Sets and Beyond. Chicago: Chicago University Press. Rama, Germán. 1971. El club político. Montevideo: Arca. Ramos, Domingo. 2003. Personal Interview. Colorado Local Activist in Tacuarembó. Real de Azúa, Carlos. 1971. Uruguay, ¿una sociedad amortiguadora? Montevideo: Ediciones de la Banda Oriental. Rehren, Alfredo. 1999. Clientelismo político, corrupción y ética en la nueva democracia chilena. In El modelo chileno. Democracia y desarrollo en los noventa, edited by P. Drake and I. Jaskic. Santiago de Chile: LOM Ediciones. Rehren, Alfredo. 2008. La evolución de la agenda de transparencia en los gobiernos de la Concertación, edited by Dirección de Asuntos Públicos. Santiago: Pontificia Universidad Católica de Chile. Reina, Artigas. 2003. Personal Interview. FA and AU Local Activist in Bella Unión. Rial, Juan. 1984. Partidos políticos, democracia y autoritarismo. Montevideo: CIESU, Ediciones de la Banda Oriental. Riquelme, Alfredo. 1999. ¿Quiénes y por qué “no están ni ahí”? Marginación y/o automarginación en la democracia transicional. Chile. 1988–1997. In El modelo chileno. Democracia y desarrollo en los noventa, edited by P. Drake and I. Jaskic. Santiago de Chile: LOM Ediciones. Roberts, Kenneth. 1992. In Search of a New Identity:  Dictatorship, Democracy, and the Evolution of the Left in Chile and Peru. Doctoral Dissertation, Department of Political Science, Stanford University. Roberts, Kenneth. 1998. Deepening Democracy? The Modern Left and Social Movements in Chile and Peru. Stanford: Stanford University Press. Roberts, Kenneth. 2002. Social Inequalities Without Class Cleavages in Latin America’s Neoliberal Era. Studies in Comparative International Development 1 (36): 3–33. Roberts, Kenneth. n.d. Changing Course: Party Systems in Latin America’s Neoliberal Era. Book manuscript in progress. Cornell University. Rodden, Jonathan. 2010. The Geographic Distribution of Political Preferences. Review of Political Science 13: 321–40. Rosas, Guillermo. 2001. Policy Preferences, Ideology, and Political Competition in Latin American Legislatures. Durham, NC: Duke University Press. Rosenblatt, Fernando. 2013. How to Party? Static and Dynamic Party Survival in Consolidated Latin American Democracies. Doctoral Dissertation, Instituto de Ciencia Política, Pontificia Universidad Católica de Chile, Santiago. Rossi, Victor. 2003. Personal Interview. FA’s Congressmember for Canelones. Rueschemeyer, Dietrich, Evelyne Huber Stephens, and John D. Stephens. 1992. Capitalist Development and Democracy. Chicago: Chicago University Press. Rumeau, Dominique. 2005. Financiamiento de las campañas electorales en Uruguay. IV Jornadas de Investigación de la Facultad de Ciencias Sociales, Universidad de la República. Salas, Valentina. 2011. Accountability Social y Accountability Horizontal en Chile: Medios Pasivos, Agencias Estatales Reactivas. Seminario de Grado, Instituto de Ciencia Política, Pontificia Universidad Católica, Santiago de Chile.

Bibliography

361

Samuels, David 1999. Incentives to Cultivate a Party Vote in Candidate-centric Electoral Systems: Evidence from Brazil. Comparative Political Studies 32 (4): 487–518. Samuels, David. 2004. From Socialism to Social Democracy. Comparative Political Studies 37 (9): 999–1024. Samuels, David, and Cesar Zucco. 2011. The Strength of Party Labels in Brazil: Evidence from Survey Experiments. APSA 2011. San Francisco, Alejandro. 2003. Political Parties and Democracy Transition in Chile 1989–2001: The Case of the Unión Demócrata Independiente (UDI). Paper read at the meeting of the ECPR, Edinburgh. Sartori, Giovanni. 1970. Concept Misformation in Comparative Politics. American Political Science Review 64 (4): 1033–46. Sartori, Giovanni. 1976. Parties and Party Systems:  A  Framework for Analysis, New York: Cambridge University Press. Schaffer, Frederic. 2007. Elections for Sale: The Causes and Consequences of Vote Buying. Boulder, CO: Lynne Rienner. Schaulsohn, Jorge. 2001. Personal Interview. Former PPD Leader. Schelotto, Salvador. 2004. Por el ojo de una cerradura. Una mirada sobre la experiencia de gobierno municipal de la izquierda en Montevideo (1990–2004). In La izquierda uruguaya:  entre la oposición y el gobierno, edited by Jorge Lanzaro. Montevideo: Editorial Fin de Siglo. Scully, Timothy R. 1992. Rethinking the Center:  Party Politics in Nineteenth and Twentieth-Century Chile. Stanford: Stanford University Press. SERPAJ. 1989. Informe de Uruguay Nunca Más. Montevideo: Servicio de Paz y Justicia. Servicio Electoral. Multiple years. Santiago: Gobierno de Chile. accessed December 2013. Shapiro, Ian. 2002. Why the Poor Don’t Soak the Rich. Deadalus 131 (1): 118–28. Shefter, Martin. 1977. Party and Patronage: Germany, England, and Italy. Politics and Society 7 (December): 403–51. Shefter, Martin. 1994. Political Parties and the State: The American Historical Experience. Princeton: Princeton University Press. Siavelis, Peter. 1999. Continuidad y transformación del sistema de partidos en una transición “Modelo.” In El modelo chileno: Democracia y desarrollo en los noventa, edited by P. Drake and I. Jaskic. Santiago de Chile: LOM Ediciones. Siavelis, Peter. 2000. The President and Congress in Postauthoritarian Chile:  Institutional Constraints to Democratic Consolidation. University Park: Penn State University Press. Siavelis, Peter M., and Scott Morgenstern, eds. 2008. Pathways to Power:  Political Recruitment and Candidate Selection in Latin America. University Park: Penn State University Press. Silva, Eduardo. 1996. The State and Capital in Chile: Business Elites, Technocrats, and Market Economics. Boulder, CO: Westview Press. Silva, Eduardo. 2009. Challenging Neoliberalism in Latin America. Cambridge: Cambridge University Press. Silva, Osvaldo. 2003. Personal Interview. RN Congressional Candidate in 2001.

362

Bibliography

Silva, Patricio. 2001. Towards Technocratic Mass Politics in Chile? The 1999– 2000 Elections and the “Lavín Phenomenon.” Revista Europea de Estudios Latinoamericanos y del Caribe, 70 (April): 25–39. Silveira, Aldorio. 2003. Personal Interview. Espacio 609 MPP Leader in Bella Unión. SINIM (Sistema de Nacional de Información Municipal). Multiple years. accessed December 2013. Soto, Angel. 2001. La irrupción de la UDI en las poblaciones, 1983–1987. Paper read at Latin American Studies Association, Washington, DC. Stanley, William. 1996. The Protection Racket State: Elite Politics, Military Extortion, and Civil War in El Salvador. Philadelphia, PA: Temple University Press. Stein, Ernesto, Mariano Tommasi, Koldo Echebarría, Eduardo Lora, and Mark Payne. 2006. La política de las políticas públicas. Washington, DC: Inter-American Development Bank. Stinchcombe, Arthur L. 1968. Constructing Social Theories. New York: Harcourt Brace & World. Stokes, Susan C. 1999. What Do Policy Switches Tell Us About Democracy? In Democracy, Accountability, and Representation, edited by Adam Przeworski, Susan C. Stokes, and Bernard Manin. New York: Cambridge University Press. Stokes, Susan C. 2005. Perverse Accountability: A Formal Model of Machine Politics with Evidence from Argentina. American Political Science Review 99 (3): 315–25. Stokes, Susan C. 2009. Pork, By Any Other Name: Building a Conceptual Scheme of Distributive Politics. Paper read at the Annual Meeting of the American Political Science Association, Toronto. Supervielle, Marcos, and Mariela Quiñones. 2003. Las nuevas funciones del sindicalismo en el cambio del milenio. Montevideo: Facultad de Ciencias Sociales, Departamento de Sociología. Tanaka, Martín. 2005. Los estudios políticos en el Perú: ausencias, desconexión de la realidad, y la necesidad de la ciencia política como disciplina. Revista de Ciencia Política 25 (1): 222–31. Tavits, Margit, and Natalia Letki. 2009. When Left Is Right: Party Ideology and Policy in Post-Communist Europe. American Political Science Review 103 (4): 555–69. Taylor-Robinson, Michelle M. 2010. Do the Poor Count?: Democratic Institutions and Accountability in a Context of Poverty. University Park: Penn State University Press. Thachil, Tariq. 2011. Embedded Mobilization: Nonstate Service Provision as Electoral Strategy in India. World Politics 63 (3): 434–69. The Economist. 2012. Gini Back in the Bottle. (October 13). Thomas, Vinod, Xibo Fan, and Yan Wang. 2001. Measuring Educational Inequality: Gini Coefficients of Education. World Bank Working Paper 25. Tohá, Carolina. 2001. Personal Interview. PPD Congressional Candidate. Topolanski, Lucía. 2003. Personal Interview. FA and MPP Congressmember for Montevideo. Torcal, Mariano, and Scott Mainwaring. 2003. The Political Recrafting of Social Bases of Party Competition: Chile, 1973–95. British Journal of Political Science 33 (1): 55–84. Torche, Florencia. 2005. Unequal but Fluid: Social Mobility in Chile in Comparative Perspective. American Sociological Review 70 (3): 422–50.

Bibliography

363

Toro, Sergio. 2007. La inscripción electoral de los jóvenes en Chile. Factores de incidencia y aproximaciones al debate. In Modernización del régimen electoral chileno, edited by Arturo Fontaine, Cristián Larroulet, José Antonio Viera Gallo, and Ignacio Walker. Santiago: PNUD. Trobo, Jaime. 2003. Personal Interview. Blanco Congressmember for Canelones. Uriarte, José. 2003. Personal Interview. UDI 2001 Congressional Candidate in Peñalolén. Valenzuela, Arturo. 1977. Political Brokers in Chile: Local Government in a Centralized Polity. Durham, NC: Duke University Press. Valenzuela, Arturo. 1978. The Breakdown of Democratic Regimes:  Chile. Baltimore, MD: Johns Hopkins University Press. Valenzuela, Arturo. 1991. The Military in Power: The Consolidation of One-Man Rule. In The Struggle for Democracy in Chile, 1982–1990, edited by Paul Drake and Ivan Jaksic. Lincoln: University of Nebraska Press. Valenzuela, Arturo. 1995. Orígenes y transformaciones del sistema de partidos en Chile. Estudios Públicos 58 (Otoño): 6–77. Valenzuela, Arturo. 1999. Chile:  Origins and Consolidation of a Latin American Democracy. In Democracy in Developing Countries:  Latin America, edited by Larry Diamond, Jonathan Hartlyn, Juan Linz, and Seymour M. Lipset. Boulder, CO: Lynne Rienner. Valenzuela, Arturo, and Samuel Valenzuela. 1986. Military Rule in Chile: Dictatorship and Oppositions. Baltimore, MD: Johns Hopkins University Press. Valenzuela, Juan Pablo, and Suzanne Duryea. 2011. Examinando la prominente posición de Chile a nivel mundial en cuanto a desigualdad de ingresos:  Comparaciones regionales. Estudios de Economia 38 (1): 259–93. Valenzuela, Samuel J., and Timothy R. Scully. 1997. Electoral Choices and the Party System in Chile:  Continuities and Changes at the Recovery of Democracy. Comparative Politics 29 (4): 511–27. Van Cott, Donna Lee. 2005. From Movements to Parties in Latin America: The Evolution of Ethnic Politics. New York: Cambridge University Press. Varas, Augusto. 1991. The Crisis of Legitimacy of Military Rule in the 1980s. In The Struggle for Democracy in Chile: 1982–1990, edited by Paul Drake and Iván Jaksic. Lincoln: University of Nebraska Press. Vargas Cullell, Jorge. 2007. Costa Rica:  Fin de una era política. Revista de Ciencia Política 27 (Anuario Político de América Latina): 113–28. Veneziano, Alicia. 2002. La participación en la descentralización de Montevideo: 10 años de gestión de izquierda y reflexiones para América Latina. Revista de Política y Gestión 3. Vergara, Pilar. 1986. Changes in the Economic Functions of the Chilean State under the Military Regime. In Military Rule in Chile:  Dictatorship and Oppositions, edited by A. Valenzuela and S. Valenzuela. Baltimore, MD:  John Hopkins University Press. Vernazza, Francisco. 1989. Minoristas, mayoristas y generalistas en el sistema electoral uruguayo. Revista Uruguaya de Ciencia Política 3: 35–71.

364

Bibliography

Villar, Andrés. 2006. Financiamiento político en Chile: Una batalla incierta. In Desafíos Democráticos, edited by Claudio Fuentes and Andrés Villar. Santiago:  FLACSO Chile-LOM Ediciones. Villar, Jorge. 2003. Personal Interview. President of PPD’s Distrital Representation in District 18. Visconti, Giancarlo. 2011. Comportamiento diacrónico del Congreso de Chile. Revista de Ciencia Política 31 (1): 91–115. Wantchekon, Leonard. 1999. Why Do Resource Abundant Countries Have Authoritarian Governments? Leitner Working Article 11. Yale University. Webber, Jeffery R. 2012. Revolution Against “Progress”:  The TIPNIS Struggle and Class Contradictions in Bolivia. International Socialism 133: 28. Weiner, Myron. 1967. Party Building in a New Nation: The Indian National Congress. Chicago: University of Chicago Press. Weitz-Shapiro, Rebecca. 2012. What Wins Votes: Why Some Politicians Opt Out of Clientelism. American Journal of Political Science 56: 568–83. Wettstein, Germán. 1993. El Frente Amplio: en el umbral del gobierno nacional, tomos I a VI. La República. Montevideo. Weyland, Kurt, Raúl L. Madrid, and Wendy Hunter, eds. 2010. Leftist Governments in Latin America: Successes and Shortcomings. New York: Cambridge University Press. WIID2a. 2005. World Income Inequality Database. Helsinki:  United Nations University, World Institute for Development Economic Research. accessed December 2012. Wilson, José Miguel. 2012. Aborto terapéutico agudiza debate en el Senado tras dichos de Von Baer. La Tercera. accessed March 2012. Wood, Elisabeth J. 2000. Civil War and the Transformation of Elite Representation in El Salvador. In Conservative Parties, the Right, and Democracy in Latin America, edited by Kevin J. Middlebrook. Baltimore, MD: The Johns Hopkins University Press. World Development Indicators. Multiple years. Washington, DC:  World Bank. accessed December 2012. World Values Survey. Multiple years. Stockholm:  World Values Survey Association. accessed December 2012. Yaffé, Jaime. 2005. Al centro y adentro: La renovación de la izquierda y el triunfo del Frente Amplio en Uruguay. Montevideo: Linardi y Risso. Zamora, Kevin Casas. 2005. Paying for Democracy: Political Finance and State Funding for Parties. Essex: ECPR Press. Zamora, Rubén. 1998. El Salvador, heridas que no cierran: los partidos políticos en la post-guerra. San Salvador: FLACSO. Zechmeister, Elizabeth-Jean. 2001. “Left” and “Right” in Latin America. Paper read at the Annual Meeting of the American Political Science Association, Washington, DC. Zuazo, Moira. 2009. ¿Cómo nació el MAS? La ruralización de la política en Bolivia. La Paz: Frederick Ebert Stiftung. Zucco Cesar Jr. 2007. Where’s the Bias? A  Reassessment of the Chilean Electoral System. Electoral Studies 26 (2): 303–14. Zucco, Cesar Jr. 2008. The President’s “New” Constituency: Lula and the Pragmatic Vote in Brazil’s 2006 Presidential Elections. Journal of Latin American Studies 40 (1): 29–49.

Name Index Accorsi, Enrique, 165, 166 Acemoglu, Daron, 5, 9–11 Agazzi, Ernesto, 234 Aguiar, César, 75, 81, 240 Alarcón, Carlos, 166 Albiac, M. Dolores, 307 Albro, Robert, 316 Alcántara, Manuel, xx, 85, 120, 121 Alcieri, Daniel, 184 Aldrich, John, 28, 34, 35, 41, 63 Aliano, Hubaré, 175, 180, 183, 247 Allamand, Andrés, 107 Allende, Salvador, 77, 96, 102 Altman, David, xxi, xxii, xxiii, 87, 106, 107, 108, 111, 137, 138, 207, 215, 224, 231, 240, 272, 277, 337 Álvarez Rivadulla, M. José., xxi, 249, 250 Alvez, Omar, 185, 188 Ambedkar, Bhimrao Ramji, 8 Ames, Barry, 55, 311 Anría, Santiago, xxii, xxiii, 316–19 Ansolabehere, Stephen, 9, Arana, Mariano, 112, 245 Arbiza, Sergio, 188 Arellano, José P., 331 Artiga González, Álvaro, 307, 309, 310 Astori, Danilo, 112–14, 231, 243–5, 247, 253 Aubry, Marcel, 215 Auyero, Javier, 24–6, 45 Avellaneda, Claudia, 301 Aylwin, Mariana, 166, 167 Aylwin, Patricio, 98, 108, Bachelet, Michelle, 27, 108, 109, 140, 141, 155, 158, 222, 223, 275, 330, 331, 333, 338 Baker, Andy, 312, 313 Balbontín, Ignacio, 196

Barnes, Samuel, 48, 118 Barozet, Emmanuelle, 214, 215, 221, 222 Batlle, Jorge, 111, 112, 114, 175, 180, 237, 244, 245, 325 Behrman, Jere R.  265 Bentancur, Nicolás, 331 Bizcarra, Darley, 184 Blofield, Merike, xxii, 60 Boas, Taylor, 45, 46 Boix, Carles, 5, 9–11 Bordachar, Rodrigo, 216, 217 Borges Sugiyama, Natasha, 313 Borzutzky, Silvia, 64, 72, 76, 150 Bosworth, Jeffrey , 109 Brandau, Antonio, 167 Brass, Paul, 304, 305 Bruhn, Kathleen, 65 Bucheli, Marisa, 265 Büchi, Hernán, 98 Buquet, Daniel, 74, 81, 110, 113, 224, 227, 240, 245 Burgess, Katrina, 58 Bugueño, Rodrigo, 203, 205, 214 Caetano, Gerardo, xxi, 81, 103, 104, 106, 111 Calderón, Armando, 308 Calvo, Ernesto, 6, 21, 28, 45, 50 Cameron, Maxwell, 329 Campbell, Angus, 28, 56, 57 Campello, Daniela, xxiii, 5 Caramani, Daniele, 41 Caravall, Eliana, 158, 218 Cardoso, Fernando Henrique, 312, 313 Carey, John, 54–6, 272, 277 Castiglioni, Rossana, xxi-xxiii, 87, 106, 135, 224, 240, 267, 268, 269, 329, 331, 332 Castro, Nora, 248

366

Name Index

Cavarozzi, Marcelo, 224 Cavedo, Antonio, 162 Cerdeña, Jorge, 174 Chacón, Ricardo, 307 Chandra, Kanchan, 8, 56, 58, 304 Chasquetti, Daniel, 114, 236, 331, 332, 338 Chávez, Fidel, 307 Chhibber, Pradeep, 41, 304, 305, 306, 321 Chiesa, Sergio, 175 Chiruchí, Juan, 174, 175 Cifuentes, Mercedes, 269 Cofre, Margarita, xxiv, 165, 218 Collier, David, xxii, 24, 340 Collier, Ruth Berins, 340 Collor de Mello, Fernando, 311 Coppedge, Michael, 6, 30 Córdova, Ricardo, 308 Cortis, Hugo, 137, 180, 183, 187 Cox, Cristián, 331 Cox, Gary, 41, 55, 56 Cox, Michael, 130 Cox, Trevor, 130 Creceri, Juan, 199, 200 Cristi, María Angélica, 152, 164, 165, 166, 170, 219, 222 Cristi, Renato, 204, 221, 222 Cristiani, Alfredo, 307, 308 Cruz, José Miguel, 308 d'Aubuisson, Roberto, 306, 307 Da Rosa, Eber, 174, 175, 180, 183 Dalton, Russell, 49, 55, 56, 118 Dargent, Eduardo, 270, 301 Dávila, Mireya, xxi, xxii, 330 De Armas, Gustavo, 240 De la Maza, Gonzalo  103 De León, Eduardo, 241 Deininger, Klaus, 265 del Granado, Juan, 318 Díaz, Alvaro, 102, 211, 267 Díaz-Cayeros, Alberto, 23, 45 Díaz Rioseco, Diego, 153 Dittborn, Julio, 155

Dixit, Avenash, 25 Doglio, Natalia, 232, 233 Domínguez, Jorge, xxiii, 338 Downs, Anthony, 34 Drake, Paul, 96, 98, 100, 102 Dryzek, John, 9 Duarte, Gonzalo, 166–8, Duarte, José Napoleón, 307 Duque, Jorge, 181, 185 Duryea, Suzanne, 266, 267 Duverger, Maurice, 41, 45 Eaton, Kent, xxiii, 55, 56, 269 Eguizábal Mendoza, Cristina, 307 Erikson, Robert, 57 Errandonea, Alfredo, 74 Errázuriz, Javier, 98 Escobar-Lemnon, María, 301 Escolar, Marcelo  21 Espinoza, Vicente, 77 Espósito, Carla, 317 Evans, Diana, 24, 294 Ewig, Christina, xxi, 330 Fa Robaina, Juan Carlos, 75, 81 Fairfield, Tasha, 11, 329 Falleti, Tulia, 269 Fan, Xibo, 265 Fenno, Richard, 24, 153, 294 Ferreira, Wilson, 104, 105, 110, 111 Filgueira, Carlos, 62, 113, 268, 335 Filgueira, Fernando, xx, 5, 62, 87, 113, 176, 239, 242, 268, 330, 335, 338, Finch, Martin Henry, 74 Fiorina, Morris, 56, 57 Flores, Francisco, 308 Fontaine, Arturo, 203 Frankel, Francine, 305 Frei Montalva, Eduardo, 77, 102 Frei Ruiz-Tagle, Eduardo, 99, 108, 109, 151, 158 Fuentes, Claudio, xxii, 107 Funk, Robert, xxii, xxiii, 100, 102, 271



Name Index

Gajardo, Jorge, 167 Gallardo, Javier, 111 Garcé, Adolfo, 243, 338 García Linera, Álvaro, 315 García, "Turco", 246 Gargano, Reinaldo, 234 Garretón, Manuel Antonio, xxiii, 71, 94, 96–101, 150, 203 Geddes, Barbara, 338 Geffroy, Céline, 315 Giannini, Pilar, xxiii, 153 Gibson, Edward, 5, 6, 21, 22, 30 Gil, Federico, 71, 77, 104, 105 Gillespie, Charles, 104, 105 Girardi, Cristina, 157, 159, 160, 332 Girardi, Guido, 154, 155, 157, 159, 160, 161, 196 Goertz, Gary, xxii, 281, 297 Goldfrank, Benjamin, 311 González-Ocantos, Ezequiel, 50 González, Eugenio, 136, 217, 219, 222 González, Luis  64, 73, 74, 103, 104, 105, 110, 111, 240 Goodin, Robert, 9 Graham, Bruce Desmond, 304 Grindle, Merilee Serrill, 338 Gryzmala-Busse, Anna M., 26 Guerra, Sergio, 171 Guerrini, Aldo, 173, 246 Guha, Ramachandra  8 Guzmán, Jaime, 98, 99, 204–6, 210, 214, 219 Hackenbruck, Tabaré, 189 Hagopian, Frances, xxiii, 6, 21, 61, 85, 106, 107, 120, 134, 270 Harbers, Imke, 54–6, 64, 270, 301 Hartlyn, Jonathan, xx, 28, 81 Hasbún, Gustavo, 167 Hawkins, Kirk, xxi, 45, 56, 57 Hazan, Reuven, 277 Hernández, Sergio, xx, 156 Herrera, Nicanor, 111, 170, 197

367

Hershberg, Eric, 328 Hicken, Allen, 41 Hinnich, Melvin, 28 Hipsher, Patricia, 207, 214, 218 Holland, Alisha, 308, 309 Huber, Evelyne, xix, xxiii, 5, 9, 109, 268, 329, 330, 332, 340 Huidobro, Eleuterio Fernández, 239 Huneeus, Carlos, 98, 107, 192, 203, 205, 211, 213, 214, 220 Hunter, Wendy, 45, 58, 310–15 Inglehart, Ronald, 49, 118 Isla, Pedro, 158 Jackson, Giorgio, 331 Jaksic, Ivan, 96, 98, 100, 102 Jara, José, 216, 217 Jiménez, Raúl, 188 Jiménez, Tucapel, 158 Jocelyn-Holt, Tomás, 152, 161, 163, 165 Joignant, Alfredo, xxi, 109, 205, 219 Jones, Mark, 12, 41 Kaase, Max, 49, 118 Katz, Richard, 58 Kaufman, Robert, xxi, 9 Kay, Stephen, 330 Kaztman, Rubén, xxii, 228, 230, 242, 330 Keck, Margaret, 310, 311, 321 Keefer, Philip, 56, 57 Kitschelt, Herbert, xxi, 6, 9–13, 16, 22–5, 28–30, 34, 45, 47, 48, 53, 56–61, 66, 117–21, 137, 139, 143, 203, 259, 266, 270, 283, 293, 300, 338 Klein, Marcus, 203, 213, 214 Kohli, Atul, 304, 305 Koivumaeki, Riitta-Ilona, 307–9 Kollman, Ken, 41 Komadina, Jorge, 315 Kselman, Daniel, 59, 270 Kurtz, Marcus, 61, 103

368

Name Index

Lacalle Pou, Luis, 178, 180, 237, 245 Lacalle, Luis Alberto, 111–15, Lagos, Ricardo, 108, 158, 222, 223, 329–31, 333 Lamas, Diego, 175 Lanzaro, Jorge, 87, 88, 181, 224, 226, 232, 242 Larrañaga, Jorge, 115, 174, 175, 180 Laurnaga, María Elena, 246 Lavín, Joaquín, 27, 108, 136, 140, 141, 158, 165, 168, 204, 208, 209, 215, 217, 219–23, 275 Lazar, Sian, 316 Lazarsfeld, Paul, 56, 57 Lazo, Carmen, 151, 168, 198 Lehoucq, Fabrice, 294 Less, Martín, 189 Letki, Natalia, 293 Levitsky, Steven, xxiii, 6, 24, 30, 34, 58, 65, 66, 329 Lindhal, Göran, 74 Lipset, Seymour Martin, 80, 340 Londregan, John, 25, 272 López-Calva, Luis, 9 López, Elida, 201 López, Rodolfo, xxiii, 153 Loveman, Brian, 77 Loxton, James, 307, 309 Luebbert, Gregory, 340 Lula da Silva, Luiz Inácio, 311–14, Lupia, Arthur, 143 Lupu, Noam, xxiii, 7, 11 Lustig, Nora, 9, 340 Lynch, Julia, 340 Lyne, Mona, 23, 25, 31, 56, 58 Madrid, Raul, 316, 317, 319 Magri, Altair, 173 Mahía, José, 172, 247 Mahoney, James, xxiii, 281, 297 Mainwaring, Scott, 11, 12, 16, 21, 41, 55, 56, 57, 65, 71, 85, 106, 107, 121, 134, 311, 313, 338 Mair, Peter, 58, 63

Malaquina, Eduardo, 175 Mardones, Rodrigo, xxiii, 109 Martín Baro, Ignacio, 307 Martínez Lamas, Julio, 74 Martínez, Javier, 102, 267 Mathus, Isabel, 160, 197 Mayhew, David, 294 McGuire, James, 65 Melgarejo, Artigas, 137, 179 Melo, Sadi, 154 Meltzer, Allan  9, 11, 339 Meneghello, Rachel, 311 Mesa-Lago, Carmelo  5 Mesa, Carlos, 316 Middlebrook, Kevin, 309 Miller, George, 294 Miller, Warren Edward, 294 Mizala, Alejandra, 331 Monckeberg, Nicolás, 157 Montero, Alfred, 312, 313 Montes, Carlos, 154, 166–70, 196, 217 Montes, Juan Esteban, xxi, 71, 76, 106, 107, 108 Morales, Evo, 315–17, 322 Morales, Mauricio, 109, 203, 205, 214, 215, 222, 223 Moreira, Constanza, xxi, 87, 114, 226, 228, 230, 231 Moreira, Iván, 158–61 Moreno, Alejandro, 49, 118, 120 Morgan, Jana, 6, 21, 30, 56, 57, 270 Morgenstern, Scott, 41, 277 Moulián, Tomás, xxi, 71, 71, 77, 215 Mujíca, José, 37, 115, 224, 248–51, 254, 275, 338 Munck, Gerardo, xxii, 109 Munger, Michael, 28 Murillo, María Victoria, xxii, 6, 28, 45, 50 Navia, Patricio, 48, 69, 96, 97, 107, 109, 192, 198, 205, 219, 272, 277 Nichter, Simeon, 25 Norris, Pippa, 298–302



Name Index

O'Donnell, Guillermo, xxiii, 94, 239 Olivares, Carlos, 157 Ondetti, Gabriel A., 311 Orrego, Claudio, 166 Ortega, Eugenio, 85, 106 Ostiguy, Pierre, xxiii, 24, 30 Oxhorn, Phillip, 101, 103, 215 Pais, Ronald, 178, 238 Palacios, Indira, xxi, 268, 329 Panebianco, Angelo, 34, 58, 66 Panizza, Francisco, 231, 338 Paolillo, Claudio, 244 Papadopulos, Jorge, 87, 239 Pasotti, Eleonora, 61 Percovich, Margarita, 183, 246, 247 Pereira, Carlos Julio, 111 Pérez Antón, Romeo, 74 Pérez, Lily, 167, 168 Perez, Omar, 164 Piattoni, Simona, 23, 56, 58 Pierson, Paul, 335 Piñeiro, Rafael, xxiii, 60, 73, 88, 113, 179, 201, 211, 212, 266 Piñera, Sebastián, 108, 109, 141, 158, 204, 223, 338 Pinochet, Augusto, 96–8, 164, 192 PIT-CNT, 232, 233 Pita, Carlos, 238 Plaza, Luis, 157 Plumb, David, 161 Pollack, Marcelo, 203, 205, 214, 221, 222 Pontusson, Jonas, 7 Posner, Paul, 85, 103, 107, 150, 215 Potthoff, Richard, 41 Powell, G. Bingham, Jr., 49, 118 Pribble, Jennifer, xxi, xxii, xxiii, 11, 109, 157, 329–32, 335 Przeworski, Adam, 5, 8, 30, 58 Queirolo, Rosario, xxi, 111, 226, 240 Quispe, Felipe, 315, 317 Rabinowitz, David, 253

369

Ragin, Charles, 66, 283, 295 Rahat, Gideon, 277 Rama, Germán, 75, 81, 82 Ramírez, Juan Andrés, 114 Ramos, Domingo, 137, 175, 183, 187, 200 Real de Azúa, Carlos, 74, 75 Rehren, Alfredo, 109, 213 Reina, Artigas, 248 Rial, Juan, 103 Richard, Scott, 9, 11, 339 Rilla, José, 103, 104, 106, 111 Riquelme, Alfredo, 107 Roberts, Kenneth, xxii, 10, 45, 61, 66, 71, 85, 100–103, 107, 134, 215, 241, 271, 329 Robinson, James, 5, 9, 10, 11 Rodden, Jonathan, 33 Rokkan, Stein, 340 Romaguera, Pilar, 331 Romero, Oscar (Mons.), 307 Rosas, Guillermo, xxi, 120 Rosenblatt, Fernando, xxiii, 26, 60, 145, 266, 277 Rossi, Victor, 136, 246 Rueschemeyer, Dietrich, 340 Salas, Valentina, xxiii, 109 Samuels, David, 277, 312, 314 Sánchez de Losada, Gonzalo, 316 Sanguinetti, Julio María, 105, 109, 111, 180, 245 Sarney, José, 312 Sartori, Giovanni, 24, 74 Schaffer, Frederic, 24 Schaulsohn, Jorge, 135, 218 Schelotto, Salvador, 246 Scully, Timothy, 12, 16, 71, 72, 76, 77, 85, 106, 107, 338 Seregni, Líber, 105, 112, 243 Serra, José, 312, 313 Shanks, J. Merrill, 294 Shapiro, Ian, 60 Shefter, Martin, 45, 56, 57, 194, 338

370

Name Index

Shugart, Matthew, 54–6, 272, 277 Siavelis, Peter, 71, 85, 96, 97, 107, 151, 198, 277 Silva, Eduardo, 102, 267, 316 Silva, Osvaldo, 159, 218 Silva, Patricio, 220, 221 Silveira, Aldorio, 251, 252 Singer, Matthew, 30 Soto, Angel, 214 Sprague, John, 5, 30, 58 Squire, Lyn, 265 Stanley, William, 306, 309 Stein, Ernesto, 338 Stephens, John, xxiii, 9, 329, 330, 332, 340 Stinchcombe, Arthur, 263 Stokes, Susan, 23–5, 28, 45, 56, 58, 106, 107 Suarez-Cao, Julieta, xxiii, 21 Supervielle, Marcos, 234 Tanaka, Martín, 301 Tarigo, Enrique, 111 Tavits, Margit, 293 Taylor-Robinson, Michelle, 6, 7, 30, 56, 58 Thachil, Tariq, 304, 305, 306 Thomas, Vinod, 263 Tohá, Carolina, 136, 198 Topolanski, Lucía, 325 Torcal, Mariano, 85, 106, 107, 121, 134 Torche, Florencia, 264 Toro, Sergio, xxiii, 107, 109 Trobo, Jaime, 186, 201 Uriarte, José, 206, 217, 219, 222 Valenzuela, Arturo, xxiii, 6, 7, 30, 64, 71, 72, 76, 77, 85, 94–101, 106, 107, 150, 153, 213

Valenzuela, Juan Pablo, 266, 267 Valenzuela, Samuel, 71, 85, 94, 85, 94, 98, 99, 101, 106, 266 Van Cott, Donna Lee, 315, 317 Varas, Augusto, 97 Vargas Cullell, Jorge, 294 Vázquez, Tabaré, 88, 112–15, 224, 230, 243, 248, 332, 338 Veneziano, Alicia, 238, 246 Vergara, Pilar, 270 Vernazza, Francisco, 75, 81, 82 Villar, Andrés, 218 Villar, Jorge, 160, 196 Visconti, Giancarlo, xxiii, 210, 211 Wallerstein, Michael, 5 Wang, Yan, 265, 283 Wantchekon, Leonard, 309 Webber, Jeffery, 316 Weiner, Myron, 305 Weitz-Shapiro, Rebecca, 10, 30, 32, 44 Weldon, Steven, 55, 56 Wettstein, Germán, 235 Weyland, Kurt, 30, 329 Wilkinson, Steven, 22–5, 56, 58, 60, 139, 266 Wilson, José Miguel, 127, 306 Wood, Elisabeth, 306–10 Yaffé, Jaime, 88, 181, 226, 242–4 Zalaquett, Pablo, 167, 168, 217 Zamora, Kevin Casas, 235, 236 Zamora, Rubén, 309 Zechmeister, Elizabeth-Jean, xxi, 16, 117, 120, 121 Zuazo, Moira, 315, 317, 318 Zucco, Cesar, 21, 192, 313–15 Zumarán, Alberto, 111 Zunino, Miguel, 174

Subject Index Alianza por Chile, 67, 85, 107, 121 Alianza Republicana Nacionalista (ARENA), 4, 16, 280, 306–10, 322 Argentina, 21, 24, 25, 119, 147, 244, 286 Asian financial crisis (1988), 312 Bahujan Samaj Party (BSP), 304 Batllismo, 87, 106, 112, 137, 176, 180, 199, 226, 243, 252 Bharatiya Janata Party (BJP), 4, 16, 280, 303–6, 322 Bolivia, 4, 16, 264, 288 Gas War, 316 Gasolinazo protests, 319 Ley de Participación Popular, 315 Water War, 316 Brazil, 4, 16, 65, 246, 264, 288 1988 Constitution, 311 Bolsa Familia, 313 land reform, 313 bureaucratic-authoritarian (BA) regime, 94, 103, 106 campaign finance, 30, 31, 66, 211, 218, 236, 270, 277, 280 Catholic Church, 199, 164, 197, 273 liberation theology, 310 causal inference, 12, 259, 328 Chile 1973 coup, 72 1980 Constitution, 96, 136, 210 1988 plebescite, 96, 97 authoritarian enclaves, 97, 191, 255 Chicago-boys, 98, 205 consensual politics, 107, 126, 136 dictatorship, 3, 67, 85, 86, 94, 106, 155, 164, 202, 205, 267, 269 districts, 144–6, 154–71 education reforms, 331

electoral mobilization strategies, 94 government autonomy healthcare, 330 ideological thirds, 71 latifundio system, 102 National Security Council, 96 party system, 12, 68, 198, 288, 328 realigment, 96, 98 pension system, 330 PSU, 334 regime divide, 85, 86, 90, 115, 118, 122, 131, 142, 150, 261, 271 student protests, 109, 330, 331 Transantiago, 109 cleavage mobilization, 76, 85 Colombia, 81, 241, 287 conceptual stretching, 24 Concertación de Partidos por la Democracia, 65, 85, 98, 107, 121, 134–6, 329 conditional cash transfers (CCTs), 329 Congress Party, 303 core constituency, 22, 203, 210, 231, 252, 254, 255, 261, 304, 314, 315, 320, 324 cosismo, 160, 215 Costa Rica, 241, 294 decentralization, 62, 64, 66, 246, 269, 279, 301, 302, 324 Democracy Cross-National Data 3.0, 298 Democratic Accountability and Linkages Project, 53, 281, 292, 298 descriptive inference, 12, 59 bias, 6, 279, 281, 328 direct democracy, 272 education levels, 143, 272 El Salvador, 4, 16, 264, 287 civil war, 321, 322

372

Subject Index

El Salvador (Cont.) Comisión para la Verdad para El Salvador, 307 death squads, 307 human rights violations, 307 military junta, 307 electoral appeals, see party-voter linkages electoral coordination, see strategic harmonization electoral dealignment, 12, 57, 72, 81, 107, 150, 157, 170, 202 electoral rules, 59, 62, 68, 107, 276, 279, 297, 300 binominal electoral system, 107, 157, 191, 198, 271, 277, 332 concurrent elections, 276 d'Hondt formula, 107 double simultaneous vote (DSV), 73, 82, 201, 260 proportional representation, 73 triple-simultaneous-vote, 201 electoral segmentation, 32 candidate-based appeals, 139–41, 148, 159, 168, 261 clientelistic linkage, 28, 58, 176–80 definition, 4 dilemma, 21 dual linkages, 33, 68, 77, 223, 262 ideal types, 27, 41–4 mixed linkages, 33, 68, 162, 252, 262, 290 particularism, 182 partisan identity (PID) linkage, 28 programmatic linkage, 28, 117, 137, 154, 261 two-dimensional typology, 24 types of resources, 26 Encuesta de Elites Parlamentarias de América Latina (PELA), 15, 119, 122, 125, 127, 131, 132, 133, 135, 137 equifinality, 66, 263, 328 external validity, 16, 52, 281, 289

Finland, 286, 293 France, 14, 286, 293 Frente Amplio (FA), 4, 21, 26, 37, 65, 87, 105, 109, 115, 121, 225–55, 290, 323 cross-party alliances, 251 demographic explanation, 240 moderation, 244, 252 territorial structure, 249 Frente Farabundo Martí para la Liberación Nacional (FMLN), 307 moderation, 307 Frente Patriotica Manuel Rodríguez (FPMR), 101, 206 Greece, 14, 58, 286 Gremialismo, 98, 204 gross domestic product (GDP), 291, 295 import-substitution industrialization (ISI), 87, 176, 181, 239, 268, 272 incumbency advantage, 151, 188, 277, 280 India, 4, 16, 288, 299 castes, 304, 305 nationalism, 305 partisan activists, 305 Mandal Commission Report, 304 Indian Communist Party, 304 inequality, 1 cognitive effects, 60 effects on democratic representation, 60, 291 effects on democratic stability, 339 effects on electoral mobilization, 266 effects on electoral segmentation, 7, 279, 320, 323, 324 redistribution, 117, 264, 289, 329, 330, 332, 339 socioeconomic, 2, 32, 77, 82, 131, 133, 261 territorial, 2, 32, 77, 82, 144, 272, 306 interest group aggregation, 172, 269, 279, 336 fragmentation, 1, 268, 158, 335



Subject Index

issue-congruence, 49, 129, 134 Italy, 14, 58, 286, 293 LAPOP survey, xxiv, 12, 15, 139, 140, 237, 336, 337 Latinobarómetro, 240 market reforms, 61, 66, 87, 103, 105, 182, 239, 263, 270, 278, 279, 304, 316, 320, 325 Marxism, 204, 210, 220 Mexico, 21, 25, 287 mobilization strategies, see party-voter linkages Montevideo, 87, 111, 136, 147, 171, 179, 186, 226, 234, 252, 274 Movimiento al Socialismo (MAS), 4, 16, 280, 315–19, 322 expansion, 316 moderation, 318 Movimiento de Acción Popular Unida (MAPU), 98, 100 Movimiento de Izquierda Revolucionaria (MIR), 77, 101, 156 Movimiento de Participación Popular, (MPP), 37, 230 Movimiento Sin Miedo (MSM), 318 multi-finality, 66, 263 neo-institutionalism, 340 Nuevo Espacio, 111, 225 Opus Dei, 203, 205 Partido Amplio de Izquierda Socialista (PAIS), 101 Partido Blanco, 72, 88, 92, 104, 109 Partido Colorado, 72, 88, 92, 104, 109, 111 Partido Comunista (Chile), 77, 97 Partido Comunista (Uruguay), 229 Partido de Conciliación Nacional, 307 Partido Demócrata Cristiano (Chile), 85, 98 Partido Demócrata Cristiano (El Salvador), 307

373

Partido dos Trabalhadores (PT), 4, 16, 280, 310–15, 323 moderation, 312 Partido Justicialista, 65 Partido Nacional (Uruguay) see Partido Blanco Partido Radical (Chile), 98 Partido Regionalista Independiente (PRI, Chile), 157 Partido Revolucionario Institucional (PRI, Mexico), 65 Partido Socialista (Chile), 98 Partido Socialista (Uruguay), 230 party system decline, 57, 195, 275, 322, 325, 337 party-voter linkages framework, 22, 25 aggregate level analysis, 13 clientelism, 24, 75 effects on party system, 86 definitions, 23 disaggregate analysis, 16 distributive results, 328 limitations, 23 state capacity, 338 trade-offs, 2, 30, 36, 37, 47, 63, 289, 295, 327 Western-European benchmarks, 14 path dependence, 279, 321 patronage networks, 64, 72, 80, 88, 112, 150, 176, 179, 270, 304, 313, 318 pensions, 267 peripheral constituency, 22, 203, 213, 215, 261, 305, 314, 320, 324 Peru, 187, 299 Poland, Solidarity, 310 polarization, 60, 71, 79, 82, 96 political parties definition, 34 partisan adaptation, 58 socialization, 57, 169, 177, 202 populism, 196, 316 Portugal, 287, 293 poverty, 242, 249, 266, 318, 332 privatization, 62, 64 programmatic structuring, 49, 57

374

Subject Index

Projeto Cone Sul, 15, 119 public goods, 2, 297, 324 Renovación Nacional (RN), 65, 85, 98, 215 Research design ANOVA, 120, 227 Bonferroni test, 120 comparative analysis, 14, 16, 52, 191, 289, 291 comparative historical approach, 341 confirmatory factor analysis, 118 cross-sectional analysis, 14 ethnographic fieldwork, 15, 47, 51, 144 factor analysis, 49 fuzzy-set analysis  16, 53, 284, 295, 300 large-N analysis, 14, 16, 280, 297 measurement of dependent variable, 54 multi-level analysis, 14, 46 political sociological approach, 341 procrustees analysis, 49, 119, 129 procrustees analysis, description of, 130 proxies, 118 public opinion surveys, 47 selection on the dependent variable, 12 small-N analysis, 47, 279 statistical analysis, 15 Santiago de Chile, 1, 38, 221, 273 soft vote, see peripheral constituency South Africa, 295 split voting, 318 state retrenchment, 179, 320 strategic harmonization decentralized harmonization, 37, 83, 89, 198, 203 harmonization of segmented linkages  63, 191, 192, 254 harmonized segmented appeals, 2, 36, 83, 92, 220, 262, 284 resources, 29 resources, candidate-based, 36–41, 193 196, 326 resources, party-based, 2, 35, 89, 193, 326, 337

self-harmonization, 37, 86, 89, 203 structural determinism, 7 systemic contagion, 46, 65, 278 Tanzania, 295 The Economist, 9 Turkey, 287, 299 Unidad Popular (UP), 85, 96, 99, 213 Unión Demócrata Independiente (UDI), 3, 21, 26, 65, 85, 97, 204–24, 290, 322 territorial structure, 214–217 United States of America (US), 293 Uruguay 1955 economic crisis, 75 1973 military coup, 75 1980 plebescite, 104 1996 electoral reform, 113, 201, 271 2002 economic crisis, 51, 115, 143, 241 commissarial dictatorship, 103 districts, 147–9, 171–5 education reforms, 331 governability, 338 Ley de Empresas Públicas, 112 military junta, 103, 267 Pacto del Club Naval, 105 partisan activists, 177, 224, 235–7, 249 party system, 12, 68, 198, 288, 328 fractions, 74, 93, 104, 180, 194, 229, 243, 252 pension system, 330 regime divide, 104, 120 state-market divide, 93, 115, 122, 129, 141 USSR, 101, 234 Uttar Pradesh, India, 304 valence mobilization, 138, 286 Venezuela, 241, 286 World Values Survey (WVS), xxiv 15, 119, 128, 129, 132, 133, 135 Zambia, 285, 287