Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment? 9781400885879

In this study of the breakdown of traditional party loyalties and voting patterns, prominent comparativists and country

129 24 29MB

English Pages 530 Year 2017

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Contents
List of Figures
List of Tables
Acknowledgments
One: Introduction
1. Electoral Change In Advanced Industrial Democracies
Two: Frameworks for Analysis
2. The Changing Structure of Political Cleavages in Western Society
3. Political Remobilization in Welfare States
Three: Patterns of Realignment
4. The West German Party System between Two Ages
5. The Dynamics of Issue Evolution: The United States
6. Electoral Change in Japan: A Study of Secular Realignment
7. Secular Trends and Partisan Realignment in Italy
Four: Patterns of Dealignment
8. The Dealignment Era in America
9. And the Walls Came Tumbling Down: Party Dealignment in The Netherlands
10. Dealignment and the Dynamics of Partisanship in Britain
11. Critical Electoral Change in Scandinavia
12. Continuity and Change in Spanish Politics
Five: Patterns of Stability
13. Canada: The Politics of Stable Dealignment
14. France: The Stalled Electorate
Six: Conclusion
15. Political Forces and Partisan Change
References
Index
Recommend Papers

Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment?
 9781400885879

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Electoral Change in Advanced Industrial Democracies

Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment? Editors RUSSELL J. DALTON SCOTT C. FLANAGAh PAUL ALLEN BECK Contributors JAMES ALT SAMUEL BARNES PAUL ALLEN BECK OLE BORRE EDWARD CARMINES RUSSELL J. DALTON KARL DITTRICH SCOTT FLANAGAN

RONALD INGLEHART GALEN IRWIN LAWRENCE LEDUC MICHAEL LEWIS-BECK ANTONIO LOPEZ ΡΙΝΑ PETER MCDONOUGH RISTO SANKIAHO JAMES STIMSON

PRINCETON UNIVERSITY PRESS

Copyright © 1984 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, Guildford, Surrey All Rights Reserved Library of Congress Cataloging in Publication Data will be found on the last printed page of this book ISBN 0-691-07675-8 ISBN 0-691-10165-5 (pbk.) This book has been composed in Linotron Times Roman Clothbound editions of Princeton University Press books are printed on acid-free paper, and binding materials are chosen for strength and durability. Paperbacks, although satisfactory for personal collections, are not usually suitable for library rebinding Printed in the United States of America by Princeton University Press, Princeton, New Jersey

Princeton Legacy Library edition 2017 Paperback ISBN: 978-0-691-61198-3 Hardcover ISBN: 978-0-691-65404-1

This book is dedicated to JACQUES-RENE RABIER for his invaluable contributions to comparative social research

CONTENTS

List of Figures List of Tables Acknowledgments

ix xi xv

ONE:

Introduction 1. Electoral Change in Advanced Industrial Democracies

3

RUSSELL J. DALTON, PAUL ALLEN BECK, and SCOTT C. FLANAGAN

TWO:

Frameworks for Analysis 2. The Changing Structure of Political Cleavages in Western Society

25

RONALD INGLEHART

3. Political Remobilization in Welfare States

70

RISTO SANKIAHO

THREE:

Patterns of Realignment 4. The West German Party System between Two Ages RUSSELL J. DALTON

104

5. The Dynamics of Issue Evolution: The United States

134

EDWARD G. CARMINES and JAMES A. STIMSON

6. Electoral Change in Japan: A Study of Secular Realignment

159

SCOTT C. FLANAGAN

7. Secular Trends and Partisan Realignment in Italy SAMUEL H. BARNES

205

viii

Contents

FOUR:

Patterns of Dealignment 8. The Dealignment Era in America

240

PAUL ALLEN BECK

9. And the Walls Came Tumbling Down: Party Dealignment in The Netherlands

267

GALEN IRWIN and KARL DITTRICH

10. Dealignment and the Dynamics of Partisanship in Britain

298

JAMES E. ALT

11. Critical Electoral Change in Scandinavia

330

OLE BORRE

12. Continuity and Change in Spanish Politics

365

PETER MCDONOUGH and ANTONIO LOPEZ ΡΙΝΑ

FIVE:

Patterns of Stability 13. Canada: The Politics of Stable Dealignment

402

LAWRENCE LEDUC

14. France: The Stalled Electorate

425

MICHAEL S. LEWIS-BECK

SIX:

Conclusion 15. Political Forces and Partisan Change

451

RUSSELL J. DALTON, SCOTT C. FLANAGAN, and PAUL ALLEN BECK

References Index

477 505

LIST OF FIGURES

2.1 The Trend in Class Voting in Four Western Democracies, 1948-1983 2.2 Economic Issue Positions and Left-Right Self-placement among Western European Publics, 1979 2.3 Noneconomic Issue Positions and Left-Right Self-placement among Western European Publics, 1979 2.4 Support for Nationalization of Industry and for Developing Nuclear Power among Electorates of Western European Political Parties, 1979 2.5 Support for Nationalization of Industry and for Developing Nuclear Power among Candidates to European Parliament, 1979 2.6 Correlations between Issue Positions Taken by Candidates of Given Parties, and Social Background Characteristics of Their Electorates 3.1 Easton's Model of Political Support 4.1 Party Images on the Left-Right Dimension, 1976-1983 4.2 Party Images on Cleavage Dimensions, 1976-1980 4.3 Old Politics and New Politics Dimensions, 1974-1980 4.4 The Placement of Social Classesin the Party Space, 1974-1980 4.5 The Placement of Educational Groups in the Party Space, 1974-1980 4.6 The Placement of Generations in the Party Space, 1974-1980 4.7 The Placement of Education Groups in the Party Space for Prewar and Postwar Generations, 1974-1980 5.1 Four Hypothetical Models of Issue Realignment 5.2 Mean Party Positions on Government Responsibility for Jobs and Standard of Living: A Reconstructed Time Series 5.3 The Decay of Interparty Polarization on the Jobs and Standard of Living Issue 5.4 The Desegregation Attitudes of Party Identifiers, 1945-1980 5.5 The Growth of Party Differentiation on Desegregation: Actual and Predicted 5.6 Growing Party Differentiation on Desegregation: Congress and the Mass Electorate 5.A Reconstructed Partisans by Reported Identification: A Panel Analysis

30 44 45 49 50 54 73 116 117 119 121 122 123 125 136 143 143 146 149 150 156

χ

5. Β Party Identifier Desegregation Attitudes in Cross Sections and by Reconstruction 6.1 A Path Analysis of Japanese Voting Behavior, 1976 6.2 Dimensions of Political Conflict in Japan, 1976 8.1 Partisanship of the American Electorate, 1952-1980 9.1 Mean Placement of Parties on Left-Right Scale, 1968-1982 9.2 Vote Switchers between KVP and Other Parties, 1967-1977 9.3 Vote Switchers between CDA and Other Parties, 1977 9.4 Vote Switchers between PvdA and Other Parties, 1967-1977 9.5 Vote Switchers between W D and Other Parties, 1967-1977 9.6 Vote Switchers between D'66 and Other Parties, 1967-1977 9.7 Percentage of Switching Voters, 1948-1982 10.1 Decline of Partisan Strength in Britain, 1964-1979 10.2 Beliefs of Labour Party Identifiers, 1964-1979 11.1 Electoral Strength of Social Democratic and Communist Parties in Scandinavia, 1947-1981 11.2 Relationship between Change in Support for the Norwegian Labor Party and Left-Wing Parties, 1953-1981 11.3 Electoral Strength of Scandinavian Conservative Parties, 19471981 11.4 Electoral Strength of Scandinavian Liberal Center Parties, 1947-1981 11.5 Change in the Social Democratic Vote and Party Identification in the Danish Election Surveys, 1971-1979 11.6 Trend in the Index of Class Voting in the Scandinavian Countries, 1955-1979 12.1 Selected Social and Political Indicators by Age, 1980 12.2 AID Analysis of the Determinants of Orientations toward Franco, 1978 12.3 AID Analysis of the Determinants of Orientations toward the Suarez Government, 1978 12.4 THAID Analysis of Partisan Preferences, 1980 13.1 Stability and Change in the 1979-1980 Elections 14.1 Party Identifiers, 1948-1981 14.2 The Distribution of Left-Right Ideology among the French Electorate, 1973, 1978, 1981

Figures

157 177 181 243 273 278 279 280 281 282 288 301 307 335 337 340 342 349 352 374 380 381 384 416 432 441

LIST OF TABLES

1.1 Fractionalization of Democratic Party Systems, 1955-1975 9 1.2 Volatility of Democratic Party Systems, 1948-1977 10 2.1 Factor Analysis of Issue Positions of Candidates to the European Parliament 35 2.2 Correlation between Left-Right Self-placement and Position on Specific Issues among Western Elites and Publics, 1979 39 2.3 Factor Analysis of Issue Preferences and Left-Right Selfplacement of Western Publics, 1979 41 2.4 Economic Development and Support for the Classic Economic Policies of the Left among Publics of European Community Countries, 1981 47 2.5 Correlation between Issue Positions of Candidates of a Given Party and Positions Taken by Electorate of That Party, 1979 52 2.6 Correlation between Policies Supported by Candidates of a Given Party and Characteristics of That Party's Electorate 55 2.7 Factor Analysis of Left-Right Group Sympathies and Establishment/Antiestablishment Attitudes in Eight Nations 60 2.8 Electoral Cleavages Based on Social Class, Religious Practice, and Personal Values, 1973-1979 64 2.9 Left-Right Self-placement According to Social Class, Religious Practice, and Personal Values, 1973-1979 65 2.10 Support for Social Change by Social Class, Religious Practice, and Personal Values, 1976-1979 66 2.11 Political Cleavages Based on Social Class, Religious Practice, and Personal Values, 1976-1979 67 3.1 The Establishment-Antiestablishment Dimension in Six Nations 79 3.2 The Distribution of Ideology Types 80 3.3 Political Alienation by Establishment Categories 81 3.4 Political Efficacy by Establishment Categories 82 3.5 Political Action by Establishment Categories 83 3.6 Ideological Thinking by Establishment Categories 85 3.7 Postmaterialist Values by Establishment Categories 86 3.8 Issue Orientations by Establishment Categories 88 3.9 Regression Analysis of the Establishment Dimension 89 4.1 Principal Components Analysis of Issue Salience, 1961-1983 110 4.2 The Distribution of Value Priorities, 1970-1982 111 4.3 Issue Orientations by Social Groups, 1961-1983 113 4.4 The Strength of Partisanship, 1972-1983 126

Tables

Xll

4.5 4.6 4.7 5.1 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 7.1 7.2 7.3 7.4 7.5 7.6 7.7 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 9.1

Class Voting Differences, 1953-1983 Generational Voting Differences, 1953-1983 Class Voting Indices by Generation, 1953-1983 The Evolution of Interparty Differences on Desegregation: A Dynamic Growth Model of Issue Realignment Vote by Party in the Postwar Japanese House of Representatives Elections, 1946-1980 Party Vote Shares in High and Low Density Prefectures, 1980 Party Support by Occupational Environment, 1976 Party Support by Community Integration, 1976 Party Support by Value Preferences, 1976 Index of Class Overrepresentation by Party Support, 1976 Party Cleavages on Ideology, Demographics, and Community Integration, 1976 Party Cleavages on Value Preferences, 1976 Party Cleavages on Five Issue Dimensions, 1976 Changes in the Percentages of Nonidentifiers and LDP Supporters, 1955-1981 A Comparison of Party Identification and Vote Discrepancies in the Late 1950s and 1970s Popular Vote Shares for the Chamber of Deputies Elections, 1946-1983 Partisan Preferences in the Italian Electorate, 1968-1981 Strength of Party Attachments, 1968-1981 Pattern of Vote Switching, 1972, 1975, 1979 Partisan Preference and Partisan Stability, 1972 Leftist Party Preferences among Cohorts of First Voters, 19681981 Partisan Preference and Support for Radical Social Change, 1972-1981 Stability in Party Identification, 1950s versus 1970s Patterns of Partisan Change, 1950s versus 1970s Partisanship by Age Cohort, 1952-1980 Patterns of Partisan Change by Age Cohort, 1950s versus 1970s Home Partisan Environment, Education, and Young-Adult Partisanship Transmission of Partisanship from Parent to Child, 1965 Partisan Change from Adolescence to Adulthood, 1965-1973 Antiparty Feeling among Young Adults and Their Parents, 1973 Popular Vote Shares for the Dutch Parliamentary Elections, 1946-1982

127 128 129 148 161 168 169 173 176 183 185 189 190 196 198 213 215 216 219 221 224 227 244 246 248 251 256 258 260 262 269

Tables 9.2 Percentages of Voters Who Express an Identification with Their Party, 1967-1982 9.3 Timing of the Vote Decision, 1971-1982 9.4 Timing of the 1971 Vote Decision and the Stability of the Vote, 1967-1977 9.5 Number of Preferences for Party (Panel Study, 1971-1977) 9.6 Number of Times Preferring D'66 by 1977 Vote Choice 9.7 Vote Preferences of Youngest Voters, 1971-1982 9.8 Indicators of Party Attachment by Vote Cohort, 1982 10.1 Expectations for the British Economy, October 1974 10.2 Expect Worse Off for Family, February 1974 10.3 Sorry Britain Is in Common Market, February 1974 10.4 Retrospective Evaluations of Performance, February 1974 10.5 Seven-Point Party Identification, June 1970 10.6 Seven-Point Party Identification, February 1974 10.7 Seven-Point Party Identification, October 1974 10.8 Predicted and Actual Dealignment in 1970 and 1974 11.1 The Scandinavian Parties Arranged into Partisan Blocs 11.2 Exchange of Social Democratic Votes with Left-Wing and Bourgeois Votes in the Scandinavian Countries, 1950-1979 11.3 Exchange of Conservative Votes with Liberal Center and Socialist Votes in the Scandinavian Countries, 1950-1979 11.4 Change in the Socialist Vote across Election Periods in the Scandinavian Countries, 1950-1968 and 1965-1979 11.5 Distribution of Party Identification in Denmark, Norway, and Sweden 11.6 Exchange among Danish Party Identifiers, 1971-1979 11.7 Age, Social Class, and Socialist Voting in Sweden, 1976 11.8 Issue Voting on Left-Right Items Appearing in Several Surveys in Denmark, Norway, and Sweden, 1960-1979 11.9 Issue Voting on the EEC Issue in Norway, 1969-1977 11.A Support for Four Partisan Blocs in Denmark, Norway, and Sweden at Parliamentary Elections, 1947-1979 11.Β Party Identifiers, Vote Defectors, and Nonidentifying Voters of the Five Largest Parties in Denmark, 1971-1979 12.1 Summary Results from Second and Third Spanish General Elections, 1979-1982 12.2 Government Performance and Popularity of Political Leaders by Recognition of Left-Right Scale, 1978 and 1980 12.3 Correlations between Four Indicators of Cleavages, 1980 12.4 Factor Analysis of Franco and Suarez Government Evaluations, 1978 and 1980 12.5 Distribution of Partisanship among Final THAID Groups, 1980

Xlll

285 286 287 290 291 294 295 314 315 317 319 320 322 324 327 333 338 343 345 347 351 353 357 360 363 364 366 372 376 378 385

XIV

12.6 Turnover Matrix of Partisan Identification, 1978-1980 12.7 Party Identification by Economic Situation in 1978 and Political Situation in 1980 13.1 Attributes of Partisanship in Canada as Measured by Three National Samples, 1974-1980 13.2 Multiple Regression Analysis of Canadian Voting Behavior, 1980 Federal Election 13.3 Voting Behavior of New Voters in the 1974, 1979, and 1980 Federal Elections 13.4 Selected Characteristics of Party Identification in Canada by Age Cohorts, 1980 13.5 Voting Behavior of Transient Voters in the 1974, 1979, and 1980 Federal Elections 13.6 Electoral Turnover, 1974-1979 13.7 Electoral Turnover, 1979-1980 13.8 Party Identification and Voting Behavior, Turnover in Three Elections, 1974, 1979, and 1980 14.1 French Legislative Election Results, 1958-1981 14.2 Social Class and Vote Intention, 1978 14.3 The Correlation of Traditional Factors with Vote Intention, 1958-1981 14.4 Religion and Vote Intention, 1978 14.5 The Distribution of Left-Right Ideology among the French Electorate, 1946-1981 14.6 Ideology and Vote Intention, 1978

Tables

386 389 405 408 412 413 415 418 418 421 427 435 436 437 440 442

ACKNOWLEDGMENTS

We see two lessons to be learned from this volume. First, the past decade has been a period of dramatic political change in most democratic party systems. New issues have come to the fore, traditional party coalitions are disappearing, and mass politics has shed all pretenses that the future holds an end to ideological conflict. This volume developed from a 1979 conference at Florida State University concerning electoral changes in the industrialized democracies. At the conference, the contributors to this volume first began to describe the changes now transforming these political systems. We were impressed by the similarity of our findings, even though we initially approached the topic from different perspectives. While the changes had unique expressions in each system, they nonetheless have enough in common to require truly comparative explanations. Indeed, these parallels led us to document the political changes now affecting advanced industrial democracies and to evaluate the causes and consequences in a more systematic fashion. The continuing discussions and revisions that followed this conference refined our understanding of the processes of partisan change, although we readily admit that our knowledge is still incomplete. However, during this period we had a haunting concern that these phenomena were the temporary manifestations of a passing age of affluence; many scholars have, of course, argued this point. But as the volume progressed, the pace of political change intensified—two new Dutch elections, a Socialist victory in France, a premature election in West Germany, and the emergence of new parties in Britain (SDP), West Germany (the Greens), and Finland and Austria (ecologists). As editors, we have learned the true meaning of the Chinese proverb "May you live in interesting times." The second lesson of this volume is a statement on the discipline of political science and comparative politics. Almost every contributor to this book is affiliated with a national election study team in his respective country of specialization. This means that the contributors bring a wealth of electoral and survey data to bear on the questions of this volume. More important, this volume stands as a tribute to the development of a truly international community of political scientists. Although we occasionally differed in theoretical interests or in the interpretation of data, we share a common scientific language and method that enables us to discuss and evaluate these theories in a comparative framework. Numerous debts have been accumulated in the years of compiling this volume. Of course, our greatest debt is to the individual chapter authors for

XVI

Acknowledgments

their outstanding contributions. We also would like to thank Florida State University, especially Dean Warren Mazek, of the College of Social Sciences, and Monte Palmer, Chairman of the Department of Political Science, for the financial support of the initial conference. Innumerable colleagues commented on various portions of the manuscript, and we want them to know their suggestions were appreciated. Russell Dalton would like to acknowledge the research fellowship of the Fulbright Commission in West Germany, which provided support for the initial drafting of several chapters. Scott Flanagan received generous support for his work from the National Science Foundation and the Bunko Hoso Foundation. Paul Allen Beck's work was supported by the Policy Sciences Program at Florida State University. The continued support of the Department of Political Science at Florida State University, especially the assistance of Susanne Miles and Betty Messer, was essential to the completion of the book. We are grateful to Suzanne Parker for her careful work in preparing the index. After considering numerous drafts, revisions, and editorial suggestions, it is impossible to separate the unique contribution of each editor. All three of us share equally in the substantive and intellectual efforts of this book, and each has carefully reviewed the entire volume. RJD SCF PAB Tallahassee, Florida

ONE

Introduction

1. Electoral Change in Advanced Industrial Democracies RUSSELL J. DALTON, PAUL ALLEN BECK S C O n C. FLANAGAN The Florida State University, Tallahassee, Florida

The First Waves of Change took many by surprise. Postwar economic recovery had ushered in unprecedented prosperity in the industrialized democracies. Many of the basic economic problems that historically plagued societies seemed on the verge of resolution. Theorists linked this dawning age of affluence to an era of consensual politics and an end to ideology (Bell, 1960; Lane, 1965; Lipset, 1964). However, student protests and political demonstrations in the 1960s challenged these views. The equal rights movements, Vietnam War, and environmental protection efforts combined to produce over a decade of unrest in American politics. The May Revolts in France brought students and intellectuals to the barricades, and the government to the brink of collapse. In Germany, the universities emptied in support of antiwar marches and social reform. The stability of postwar Japanese politics was jolted by student protests and citizen movements. Ethnic and regional conflicts intensified in Britain, Belgium, and Spain. Everywhere democracy seemed to be in crisis (Huntington, 1974; Crozier et al., 1975; Dahl, 1970; Habermas, 1973). Among virtually all the nations represented in this volume, the political unrest of the mid-sixties and early seventies stood in marked contrast to the halcyon politics of a decade earlier. Student demonstrations and political protests were visible and violent indicators of political change. They were the leading edge of broader and deeper changes that have been occurring in advanced industrial democracies. Evidence of these changes emerges from the schools, the workplace, the home— and the political system. The era of stability that seemed to be dawning less than two decades ago was quickly convulsed by massive changes in almost all areas of political life. One area involves the issues of popular concern (Inglehart, 1977; Baker et al., 1981: Chaps. 6 and 11). Industrial societies aimed at providing affluence

4

Introduction

and economic security. The success of advanced industrialism has fulfilled many basic economic needs for a sizable sector of the population. Thus concerns are shifting to new societal goals. Several of these new issues are common to advanced industrial democracies: social equality, environmental protection, the dangers of nuclear energy, sexual equality, and human rights. In some instances, historical conditions focused these general concerns on specific national problems; for example, racial equality in the United States or ethnolinguistic conflicts in Belgium, Canada, Spain, and Great Britain. However, all of these issues have been loosely integrated into a new political agenda, an agenda which has stimulated new political conflict over the past two decades. Another broad area of change involves the style of politics (Barnes, Kaase et al., 1979). Greater public participation in economic and political decisionmaking has become an important social goal. This development is closely tied to the spread of protest, citizen action groups, and unconventional political participation during this period; but it involves more. Citizens are less likely to be passive subjects and more likely to insist on being participants in the decisions that affect their lives (Dalton, forthcoming, Chaps. 2-4). While political change and social unrest now may be less visible than during the late 1960s, the process of change is continuing. The war in Vietnam is over, but protests over nuclear energy, women's rights, environmental quality, and disarmament are still with us. Moreover, attitudinal data indicate that many of the liberalizing trends begun in the 1960s have continued through the 1970s, even if their behavioral manifestations have decreased, or become less newsworthy (Yankelovitch, 1974; Inglehart, 1981). In addition, the 1970s and early 1980s increasingly have witnessed the emergence of a conservative counterattack around the issues raised by the New Left during the 1960s (Lipset, 1981b). A new morality has given rise to a new set of conservative social issues, and in response a New Right has surfaced in many countries to reassert traditional values as reflected in the movements against abortion, equal rights for women, gay rights, and lifestyle issues. In some countries, the New Right has mounted counter-socialization campaigns aimed at insulating the young from the new morality through the growth of private church schools and the censorship of public libraries. Many nations have witnessed mass assaults on the welfare state, usually through movements to reduce spending and taxation. What is new about the New Left and New Right is not only the issues that define their policy priorities but also the kinds of alignments and coalitions they are forging within national electorates (Ladd with Hadley, 1975; Miller and Levitin, 1976). The New Left has drawn disproportionate support from the new middle class, while the New Right has attracted various blue-collar and previously apolitical Fundamentalists. Moreover, the New Right is not simply antiliberal in the traditional sense, as one of its primary appeals has

Electoral Change in Advanced Industrial Democracies

5

been a reassertion of individualism and self-initiative in its fight against big government and bureaucratic regulations. Indeed, the New Right combines some of the elements of classical liberalism as exemplified by the antitax movements, efforts to limit (or reduce) the size of government, and attempts to return power to local levels of government that are more susceptible to citizen control. Even as recession weakened the economies of the industrialized nations in the early 1980s and economic issues again dominated the political agendas, the traditional coalitions of the industrial political order failed to reemerge. Unprecedented postwar unemployment in the United States, Germany, and Britain, for example, has not heightened dramatically the class basis of politics. It seems that several decades of prosperity have altered the social structure of these societies, protecting most citizens from the worst ravages of unemployment and increasing the size of the middle class. That a return to the old issues has not reproduced the Old Politics provides even more compelling confirmation of the proposition that change is the dominant feature of advanced industrial politics. While the initial shock waves of change that were set in motion in the 1960s have passed, the process of change is continuing. Furthermore, time is needed for these processes to run their course. It is too early to tell whether a new Left and Right will replace the old Left and Right in most advanced industrial societies. And if so, what realignments in electorates and party systems will such changes entail? Further distance will be necessary to place the events of the last two decades in their proper perspective. One intent of this research is to assemble the material necessary to begin such an evaluation. The theme of this volume is that fundamental changes are taking place in democratic political systems. Postwar generations reared in this new environment are bringing new concerns and skills to bear on the political process. These developments are introducing new tensions into democratic party systems. As a result, stable party alignments are fragmenting, and the traditional sociopsychological bonds between voters and parties are weakening. We admit at the outset that we cannot provide a simple prediction of what the future holds. Indeed, the chapters in this volume describe the rich variety of national responses to these forces of change. The unique institutional and political circumstances of each system should be expected to redirect these forces and adapt them to the national context. We will, however, search for a common pattern behind these trends and discuss the possible outcomes for democratic party systems. THE SOURCES OF CHANGE

Over the past generation, most of the nations in this volume reached at least the threshold of advanced industrialism. The social and economic changes

6

Introduction

involved in this transition often were of revolutionary proportions. In economics, for example, living standards rose to an unprecedented level of affluence. This growth was most dramatic in Italy, Germany, and Japan, where over the past twenty-five years (1953 to 1978) Gross Domestic Product per capita grew in real terms by 235, 265, and 595 percent, respectively (International Monetary Fund, 1979). In most Western democracies, contemporary income levels are two to four times greater than at any time in prewar history. Recent economic problems have slowed—but not reversed—this pattern. Affluence still exists, now tempered by slower growth rates. Concomitant with increasing affluence was a restructuring of the labor force. The farming sector virtually disappeared in most Western democracies, and the industrial sector remained stable or declined. With advanced industrialism came a marked shift in the labor force to the service sector. Several of the nations in this volume already have passed Daniel Bell's threshold for postindustrialism—half of the labor force employed in the tertiary sector (Bell, 1973; International Labour Organization, 1981). In addition, because of the expansion of national and local governments, public employment now constitutes a significant share of the labor force in most of these nations. Advanced industrialism is associated not only with changes in the relative size of the three principal industrial sectors, but also with changes in the context of the workplace and the residential neighborhood (Dahl and Tufte, 1973; Verba et al., 1978; Steiner et al., 1980). The continuing decline of rural populations and the expanding size of metropolitan centers stimulated changes in life expectations and life-styles. Urbanization meant a growing separation of the home from the workplace, a greater diversity of occupations and interests, an expanded range of career opportunities, and more geographic and social mobility. With these trends came changes in the forms of organization and interaction. Communal forms of organization were replaced by voluntary associations, which, in turn, became less institutionalized and more spontaneous in organization. These changes reflect the fact that communities are less bounded, that individuals are involved in increasingly complex and competing social networks that divide their loyalties, and that interpersonal and institutional attachments are becoming more fluid. Finally, this weakening of institutional loyalties and traditional social networks is associated with the undermining of traditional values and a growing volatility in political behavior. These economic and social changes were joined by an expansion of educational opportunities. Throughout the Western world, there has been a steady growth in compulsory education and a virtual explosion in university training. For example, over half of all American doctorates awarded in this century were earned in the past ten years. Similarly, over the past twenty-five years the proportion of the population attending colleges and universities increased by 347 percent in the United States, 472 percent in Britain, 503 percent in Germany, 815 percent in Sweden, 429 percent in Japan, and by almost equally

Electoral Change in Advanced Industrial Democracies

7

high rates in most other industrial democracies (Taylor and Jodice, 1983). More educational opportunities mean a growth in political skills and resources, producing the most sophisticated electorates in the history of democracies. And even this education level will look paltry in comparison to electorates in the year 2000. These increases in education and information-handling skills were accompanied by a parallel increase in information resources. The growth of the electronic media—especially television—was exceptional. Other information sources, such as book and journal publications, also increased. Even more revolutionary was the growth of electronic information processing—that is, computers, information retrieval and storage systems, word processing, and related technological innovations. More computers were produced in 1981 than in all previous years combined. Information is no longer a scarce commodity. The contemporary information problem is how to manage an evergrowing volume of complex and sophisticated knowledge. Moreover, new ideas are now diffused much more rapidly throughout mass publics, and as a result, new mass movements can emerge from obscurity to widespread popularity almost overnight. Thus the transformation of Western democratic societies is due to more than simply the politics of affluence. Indeed, if it were not, this book might be dated by the slowdown in economic growth rates since the OPEC oil shocks in the 1970s and the world recession of the 1980s. Changes in the occupational and social structure are altering life conditions and life-styles. The expanding political skills and resources of the electorate are changing political processes. And even though economic growth rates have slowed, the living standards of advanced industrial societies are still far better than a generation ago. Recent discussions of the politics of scarcity and economic decline are both premature and too narrow in addressing the changes that democratic societies are experiencing. The social, cultural, and economic trends of advanced industrialism have not regressed to prewar standards. Contemporary electorates remain fundamentally different from their predecessors.1 EVIDENCE OF PARTISAN CHANGE

Until recently, the prevailing theme in comparative party research was the persistence of democratic party systems. In addition to Lipset and Rokkan's treatise on the freezing of cleavage alignments, the empirical studies of Rose and Urwin concluded that the major question facing researchers was to explain 1

A number of analysts have predicted a dire economic future for industrial democracies. This would spell a fundamental reversal in the trends we have described if the prognostications become true (see Meadows et al., 1972; Forrester, 1971).

8

Introduction

the observed stability in democratic party systems (Lipset and Rokkan, 1967, Rose and Urwin, 1969, 1970) Clearly, something has changed dramatically in the last decade The parties are being presented with new demands and new challenges Partisan change— rather than partisan stability—is a common pattern in virtually all of the nations examined here This volume will focus on documenting and explaining these changes A major factor in the destabihzation of democratic party systems was the initial inability or unwillingness of the major established parties to respond fully to the new demands being placed upon them First, the agenda of advanced industnal politics is still evolving, and the political payoffs of adopting specific issue positions have been unclear Second, since contem­ porary political leaders were schooled in the earlier era, they are less re­ sponsive to new and unfamiliar demands In addition, many of these new issues cut across the traditional lines of party alignment Thus the larger parties often are internally divided on these issues (Inglehart, this volume Chap 2, Berger, 1979, Baker et al , 1981 Chap 12) Caution might be advisable in this situation However, this caution introduces new tensions into these party systems The immediate result of these trends is a decomposition of electoral align­ ments in many Western nations Parties are fragmenting, and the social and psychological bonds that traditionally link voters to specific parties are weak­ ening—party systems have entered a period of flux (Daalder and Mair, 1983) For example, in some instances smaller parties—like the PSU in France and the Center party in Sweden—have altered their positions to reflect changing political conditions In other cases, new parties have formed specifically to represent new political perspectives Such parties include the D'66 in The Netherlands, Ghstrup's tax party in Denmark, the Radicals in Italy, the recent ecologist parties in France and Germany, and the Australian Democrats In still other cases, a large proportion of the electorate has turned away from the entire party system, leading to a dealignment in Britain and the United States This decomposition of democratic party systems can be documented in several ways One aspect of decomposition involves the fractionalization of modern party systems Until recently, researchers argued that democratic party systems were gradually evolving toward large "catch-all" parties that would stabilize and unify the party system (Kirchheimer, 1966) Data on the frac­ tionalization of party votes find just the opposite (Table I I ) 2 Between 1955 2

The fractionalization scores are based on the party vote shares for the election closest to the time points given in Table 1 1 Fractionalization is computed as (Ν) ( N - l )

Electoral Change in Advanced Industrial

Democracies

TABLE 1.1

Fractionalization of Democratic Party Systems, 1955-1975 Country

1955

1965

7975

United Kingdom Japan France Denmark Norway Luxembourg Belgium United States Finland Sweden Switzerland Ireland Spain Austria Canada Germany Italy Netherlands Period average

.536 .750 .836 .737 .714 .676 .674 .494 .798 .703 .786 .646 — .597 .663 .697 .759 .765 .695

.588 .671 .755 .763 .743 .717 .748 .478 .808 .708 .822 .632 — .581 .699 .610 .758 .838 .701

.683 .761 .840 .827 .805 .769 .797 .518 .836 .723 .835 .643 .784 .567 .673 .583 .718 .749 .725"

1965-1975 Change + .095 + .090 + .085 + .064 + .062 + .052 + .049 + .040 + .028 + .015 + .013 + .011 — -.014 -.026 -.027 -.040 -.089

SOURCES: Fractionalization of vote shares data for 1965 and 1975 are from Taylor and Jodice (1983); 1955 data were compiled by the authors. * Period average for 1975 does not include Spain.

and 1965, party fractionalization is relatively constant for the nations examined in this volume, with increases in some nations balancing decreases in other nations. Beginning in the mid-sixties, the cumulative impact of social, economic, and political trends stimulates the introduction of new parties and the breakup of established parties. Consequently, the fractionalization of party vote shares increases between 1965 and 1975 for all but five of the nations we study. Another indicator of decomposition is the volatility of the party system; that is, fluctuations in voting results between elections. As old cleavages weaken and new concerns arise, interelection volatility should increase. Mogens Pedersen has documented the growth of aggregate party volatility during the past three decades (Table 1.2).3 The immediate postwar years were a time These data are drawn from Taylor and Jodice (1983), except for the 1955 time point, which was computed by the authors. For additional discussion of increasing fractionalization, see Mayer (1980), Wolinetz (1981). 3 Most data are drawn from Pedersen (1979b), except for Canada, Japan, Luxembourg and

Introduction

10 TABLE 1.2

Volatility of Democratic Party Systems, 1948-1977 Country Norway Denmark Netherlands Luxembourg United Kingdom Switzerland Sweden Finland United States Austria Japan Italy France Ireland West Germany Belgium Canada Period average

1948-1959

1960-1969

1970-1977

3.4 5.5 6.3 14.3 4.4 1.9 4.8 4.4 2.5 4.1 37.4 10.3 21.8 10.9 15.2 7.9 9.0 9.7

5.2 8.9 7.9 11.4 5.2 3.7 4.3 6.9 2.9 3.9 7.1 8.0 11.9 6.8 9.5 10.3 14.8 7.6

17.1 18.7 12.7 16.0 7.9 6.4 6.6 9.1 3.2 3.1 6.1 6.8 10.6 5.0 4.9 5.5 8.1 8.7

1960-1970 Change 11.9 9.8 4.8 4.6 2.7 2.7 2.3 2.1 0.3 -0.8 -1.0 -1.2 -1.3 -1.8 -4.6 -4.8 -6.7

SOURCES: Canada, Luxembourg, the United States, and Japan compiled by the authors; other nations are from Pedersen (1979b).

of substantial partisan volatility, largely due to party instability in the newly formed party systems of West Germany, Japan, and Italy. Interelection shifts in aggregate party support for all fourteen nations average 9.7 percent during this period. In the 1960s, party alignments stabilized, and party volatility decreased to 7.6 percent. This trend reverses in the 1970s as new issues challenge the existing party alignments, and party volatility increases to an average 8.7 percent change between elections. Moreover, these aggregate measures of party change undoubtedly underestimate the individual changes occurring within the electorate.4 the United States, which were computed by the authors. Partisan volatility is the sum of the vote share for new parties plus the percentage gained by parties that increased their vote share since the last election. 4 One problem with Pedersen's volatility index is that it measures partisan change at the aggregate level, and this cannot be used to infer the stability of individual partisan preferences. Most individual level data indicate substantially higher, and increasing, levels of partisan volatility (DeVries and Tarrance, 1972; Petersson, 1978; Barnes, this volume: Chap. 7; Crewe et al., 1977). Even in the West German case where aggregate party volatility is decreasing, individual voter volatility is increasing (Conradt, 1981: 130).

Electoral Change in Advanced Industrial Democracies

11

Half of our nations thus display a general pattern of increasing party volatility over the postwar period, but even the clearest deviations from this trend—Germany, Japan, Italy, and France—are notable exceptions. These nations instituted a democratic party system de novo following the collapse of wartime authoritarian regimes. In these cases, the postwar period was one of (re)instituting and stabilizing a democratic party system. However, the chapters in this volume will document the impact of advanced industrialism even in these nations. The aforementioned tables thus provide convincing evidence of the general decomposition of contemporary party systems. We cannot say how long these trends will continue, nor where they will lead. It is clear, however, that party fractionalization and volatility have increased. In addition, these data underscore the importance of unique national conditions in determining the course of party change. Parties, and entire party systems, may follow different courses in responding to advanced industrial politics. In one nation, parties might split on new issues, leading to increased fractionalization. In another nation, the parties may adopt ambiguous or flexible positions on the new issues, leading to greater volatility between election results. On the whole, these two dimensions of party system change are only weakly interrelated (r = .26). Thus, although the cumulative effects of advanced industrialism might provide the stimulus for partisan decomposition, the chapters in this volume describe how the unique institutional and political factors of each nation channel and direct these electoral forces. A TYPOLOGY OF ELECTORAL PERIODS

In describing the changes occurring in democratic party systems, the chapters in this volume frequently will rely on a typology derived from American electoral behavior. We distinguish between three general types of electoral periods—stable alignments, realignments, and dealignments. All three focus on the long-term bases of party support and are differentiated by the condition of party loyalties. The "normal" electoral period is one of stable alignments, marked by a constancy in party coalitions and aggregate partisan equilibrium (Campbell, 1966; Pomper, 1967). This surely does not imply complete stagnancy in the party system. Even in a time of unchanging partisan balance, the electorate is in a state of flux. What is different about periods of stability is that the coalitional basis of long-term support for the respective parties remains unaltered. Interelection differences represent only momentary defections from enduring partisan loyalties. These periods are thus conceptualized most appropriately as eras of dynamic equilibrium. Ever since the publication of The American Voter, the long-term party loyalties producing a stable alignment in American politics are defined as

12

Introduction

psychological identifications with a particular political party (Campbell et al., 1960; Converse, 1976). These loyalties are seen as analogous to religious or class identifications in their durability and meaning. Periods of electoral stability could be measured in terms of these underlying partisan attachments. For example, in spite of large Republican presidential victories in 1952 and 1956 in the United States, identifications with the two major parties were highly stable throughout the 1950s. In the wake of its success in explaining American electoral behavior, the party identification index, and (often only implicitly) its conceptual underpinnings, was exported to other party systems as a measure of enduring partisan loyalties. Almost from the beginning the American import seemed out of place.5 The independence from vote that justified a separate concept of enduring partisan loyalties in America was more difficult to find in other political systems. Rather, partisanship and vote seemed to travel together. Changes in voting patterns were relatively rare, but when they did occur, party loyalties often were altered to bring them into line. Furthermore, a close relationship was found between social group (especially religion and class) and party in other nations, leading investigators to question whether party loyalty was an independent force. The idiosyncrasies of American electoral politics probably lie behind the difficulties encountered in applying American conceptions of party loyalty to other democracies. What is challenged by the non-American results is the concept of party loyalty as a psychological identification with a party, not the idea of a standing partisan decision. The notion that party loyalties exist that reflect an enduring preference for a particular party and its candidates is widely accepted in comparative research (Budge and Farlie, 1977; Budge et al., 1976; Rose, 1974). What apparently differ are the reasons for party loyalty. In non-American settings, we often find that standing partisan commitments are explained by the social coalitions underlying party support. W. Phillips Shively argues that when strong social group identifications are matched by clear party positions on social cleavage issues, as they are in most European systems, then there is less need for voters to develop a party identification (Shively, 1972). In the United States, by contrast, partisan loyalties often develop an existence independent of their roots. In sum, party loyalties may reflect a sense of party identification or the party cues derived from social characteristics. When these long-term partisan commitments are widespread and relatively constant at both the aggregate and individual levels, we shall refer to this as a stable alignment period. 5 Differences in the meaning of partisanship for non-American electorates were first reported in Campbell and Valen (1966), and Butler and Stokes (1974) Several chapters in Budge et al (1976) have developed more fully the case against party identification For a somewhat different view of partisanship outside of the United States, see Baker et al (1981 Chap 8), and Cam and Ferejohn (1981)

Electoral Change in Advanced Industrial

Democracies

13

The previous section has shown that democratic systems are in a state of flux. Stable alignments are weakening, as is witnessed by the increasing volatility and fragmentation of these party systems. One interpretation of this trend is that it marks a temporary surge in short-term forces that are already beginning to recede. This scenario argues that partisan alignments will restabilize as these short-term forces recede, with few significant changes from the previous cleavage structure. However, if structural changes in industrialized societies are the driving force behind the recent decomposition of electoral alignments, then this cannot be considered a short-term force and the possibility of more fundamental partisan change exists. One potential pattern of change is for electoral systems to experience a partisan realignment. A realignment may be defined as a significant shift in the group bases of party coalitions, and usually in the distribution of popular support among the parties as a result.6 For example, the New Deal realignment in the United States is traced to, among other things, the entry of large numbers of blue-collar workers, Catholics, and blacks into the Democratic party coalition. In short, a realignment is a time during which the composition of party coalitions undergoes significant change, with many people who earlier would have been unaffiliated, or loyalists of one party, now affiliate with another. Realignments have been a regular feature of American electoral politics for well over a century and probably since the emergence of the first mass party coalitions around 1800 (Clubb et al., 1980; Sundquist, 1973). Similar historical realignments have occurred in European party systems (Butler and Stokes, 1974: 155-210; Robertson, 1976; Rose, 1974). However, the concept of partisan realignment can be traced to V. O. Key's work in the 1950s (Key, 1955, 1959). Key's examination of changing aggregate voting patterns uncovered two types of realignment in American electoral history. He distinguished between critical realignments caused by sharp massive partisan changes and slower evolving secular realignments. Key's work has been extended and refined by a number of authors. For example, sophisticated cohort analyses indicate that the shift in the overall balance of partisan loyalties in the 1930s was based primarily on the mobilization of new voters, rather than the conversion of voters with established party commitments (Campbell et al., 1960: 153-156; Andersen, 1979; Petrocik, 1981; cf. Erikson and Tedin, 1981). In addition, the pace of realignment (secular/critical) apparently depends on the nature of the realigning issue conflicts and the response of political elites. 6 Different definitions of realignment in the research literature make the concept somewhat unclear Our definition is the conventional one and possesses the virtue of conceptualizing electoral change in terms of the underlying party loyalties of the electorate, thereby separating the phenomenon to be explained from its causes and its effects For a discussion of the various definitions of realignment and a defense of the definition we use, see Petrocik (1981) Additional discussions of the meaning of realignment may be found in Campbell and Trilling (1979)

14

Introduction

Several of the authors in this volume will apply these concepts to determine whether contemporary party systems are undergoing a realignment process. A second pattern of possible electoral change is a dealignment. Strictly speaking, a dealignment is a period during which the party-affiliated portion of the electorate shrinks as the traditional party coalitions dissolve. Dealignment initially was considered to be a preliminary step in a realignment process— the weakening of old party loyalties to facilitate a new party alignment. Ronald Inglehart and Avram Hochstein were the first analysts to treat the decay as a separate phenomenon in describing the decline of American party identification in the late 1960s (Inglehart and Hochstein, 1972). Since their seminal article, party loyalties have continued to decay, leaving the American electorate significantly less partisan by the late 1970s than it was in the 1950s. With the development of the dealignment-type to characterize a period of weakening party loyalties has come the recognition that dealignment, like realignment, may be a regular feature of electoral politics (Beck, 1979; Burnham, 1970, 1978). But dealignment is difficult to detect without measures of partisanship at the individual level. Voting data record only vote choices, leaving little record of the underlying partisan loyalties that may (or may not) have guided them. Still, American voting records contain circumstantial evidence of partisan decay prior to each realignment since the 1850s. Prerealignment politics has been characterized by large interelection vote fluctuations, unusually successful minor party movements, and declines in turnout. Each of these phenomena is circumstantial evidence of the decay in partisan loyalties that defines a dealignment. When coupled with survey-based evidence of weak partisanship within the electorates of the 1920s and 1970s, the case for dealignment as a distinct electoral period is strengthened. Several of the case studies in this volume note a similar pattern in other democratic party systems. Dealignment and realignment are employed to describe processes of change, but they also may be used to characterize end states. We may speak of an electorate as dealigned or realigned to refer to the culmination of a change. For example, the American electorate of the 1950s was realigned relative to that of the 1920s as a result of the New Deal realignment. Or, the contemporary American electorate is dealigned relative to that of the 1950s, even if the process of partisan decay has ended. Extending the typology to embrace end states is especially useful for comparative analysis, because it enables us to contrast party systems in terms of the role played by partisan commitments (see Sarlvik and Crewe, 1983). How well can a typology of electoral periods—stable alignments, realignments, and dealignments—derived largely from American politics characterize partisan politics in other democratic nations? At first blush, generalizations from the American electoral experience would seem limited. Among the nations examined in this volume, the United States may well be the "most different system," especially where electoral politics is concerned. For in-

Electoral Change in Advanced Industrial Democracies

15

stance, because of the American two-party system, American theorists tend to equate a realignment in voting patterns with a clear shift in government control. However, in a multiparty system the linkage between voting patterns and government control is more tenuous (King, 1981; Downs, 1957). Voters in a multiparty system can shift their support among parties of the same tendance, without affecting control of the government (although government policy might be clearly affected). Similarly, government control might change because of the coalition decisions of party elites, independent of the voters' wishes to reward or punish specific parties. For example, in The Netherlands, citizens normally do not know the likely coalition patterns among the parties until after they vote. These and other institutional peculiarities of the United States should caution us about the unique context of American theorizing—especially in terms of the linkage between mass and elite change. However, our realignment/dealignment typology is used solely to describe partisan change among the mass public. The typology only requires that voters structure their decisions more or less in terms of enduring partisan loyalties. The concept of standing partisan commitments—either party identification or social group ties—is meaningful in modern democracies. Indeed, over-time continuity in voting patterns throughout the democratic world suggests that standing, rather than ad hoc, decisions are the norm. Whenever and wherever it is meaningful to speak of enduring partisan loyalties, our typology may be employed. Whether these loyalties are stable, changing, or in decline will determine the characteristics of electoral politics.

ADVANCED INDUSTRIALISM AND PARTISAN CHANGE

At the present time, it is difficult to predict whether democratic party systems will move toward realignment, dealignment, or a return to previous cleavage alignments. The chapters in this volume each contribute a fresh perspective on the problem. And in the concluding chapter we will try to draw together this evidence to speak with more conviction about the patterns of change. However, as an introduction we can examine briefly the causal processes that may translate the general trends in advanced industrial nations into more specific partisan changes. Previous research has postulated several theories that may provide a linkage mechanism: Embourgeoisement An early explanation of the political changes of modern electorates was based on economic factors. Dramatically increasing prosperity presumably produces a considerable overlap in the income and life-styles of the middle class and working class. European workers generally are not struggling to maintain subsistence incomes. They spend Saturdays washing their cars and

16

Introduction

Sundays driving into the countryside—clearly not the makings of a Marxian class struggle. With this narrowing of objective class differences, John Goldthorpe argued that a moderation of political conflict would occur as the affluent sector of the working class assumed the values of their middle-class life-styles (Goldthorpe et al., 1968). This thesis is often linked to the development of advanced industrialism and represents a "consensual" or "middle mass" model of politics. In apparent support of the embourgeoisement thesis, extensive research evidence suggests that class voting differences are declining in most democracies (Ladd with Hadley, 1975; Lipset, 1981a; Kemp, 1978; Borre, this volume: Chap. 11; Dalton, this volume: Chap. 4). Social Mobility Thesis A related theory explains the decline in class voting as a function of social and occupational mobility rather than the homogenizing effects of affluence. Virtually all of the nations in this volume experienced a decline of the agrarian sector and a more recent rise in the nonmanual service sector. In some nations, dramatic changes in the size of these sectors have occurred within a few short decades. Growing social mobility means that a child's ultimate social placement is increasingly different from his/her parents. For instance, many farmers' children who had conservative political upbringings have moved into unionized, leftist, working-class contexts in the cities, while many workingclass children from urban, leftist backgrounds have moved into conservative, white-collar occupations (Stephens, 1981: 175; Hamilton, 1967). These forces of urbanization and occupational mobility are blurring traditional class and economic alignments. Some socially mobile individuals will change their adult class identifications to conform to their new contexts, while others will not. To the extent that individuals adhere to their early political training, class voting will decline. Moreover, as Baker, Dalton, and Hildebrandt point out, the growth of the new middle class further obscures class lines (Baker et al., 1981; Kerr, forthcoming). The new middle class is in an ambiguous economic position. On the one hand, it is relatively affluent, but on the other, it shares some of the same problems as the working class and increasingly seeks security in unionization. This ambiguous class role has contributed to the decline in class voting in many advanced industrial societies. Thus social and occupational mobility also are weakening traditional class alignments. Mass Society Thesis A third theory of political change stresses the atomization of society, which accompanies advanced industrialism (Kornhauser, 1959). The rapid socioeconomic changes of the past decades presumably have eroded traditional group and institutional networks. The growth of the new middle class, for

Electoral Change in Advanced Industrial

Democracies

17

example, produced a large social stratum that is integrated into neither the bourgeois nor proletarian institutional structures. Geographic, social, and structural mobility have increased, and this has weakened primary group ties. The rapid expansion of the mass media has enticed the citizen away from personal networks for political information. These changes have been associated with the decline of institutional affiliations and loyalties. For example, in many advanced industrial societies the church and unions have been very important mobilizers of support for the Right and Left, respectively. There is abundant evidence from The Netherlands, Italy, and elsewhere that church attachments and church attendance are declining, and with them the ability of church organizations to mobilize support for confessional parties (Barnes, this volume: Chap. 7; Irwin and Dittrich, this volume: Chap. 9). At the same time, in many countries the church has retreated from active involvement in politics. The same phenomenon can be found among unions. Studies from both Western Europe and Japan have pointed to a weakening attachment to labor unions. The sense of belonging to a movement, the feeling of class solidarity, the dependence upon unions, and the commitment to unions all have decreased (Korpi, 1978a). Moreover, at least in some countries, traditional union-party attachments also have weakened.7 These forces of deconfessionalization, depillarization, and declining institutional loyalties are associated with partisan dealignment. In short, a traditional political style based on primary networks such as family, the village, the union, or the local church has become less relevant as these ties have eroded. This atomization of the individual should introduce considerable instability and volatility into the political system. Without the stabilizing framework of group ties, these newly independent voters are open to a variety of appeals and may be mobilized for a variety of causes. Community Integration A fourth approach focuses on the character of community settings. The social networks and contextual effects literature argue that individuals are likely to adopt the political views of those around them. Beginning with the early American voting studies of Lazarsfeld and his associates, researchers found that small groups tend toward partisan homogeneity and small group opinion leaders are important in transmitting political cues and increasing group uniformity in voting choices (Lazarsfeld et al., 1944; Berelson et al., 1954; Burstein, 1976; Wright, 1977). This model argues that important 7 In the Japanese case, for example, the largest national union federation, Sohyo, increasingly has taken a neutral position on party endorsements, as the Socialist party's monopoly on support from Sohyo's constituent unions was broken and the Communist party and other progressive parties won the support of the various affiliated unions In addition, while Sohyo is the oldest, largest, and most militant union, other more moderate and even politically neutral national federations have grown up over the last two decades

18

Introduction

political cues are derived from face-to-face contacts with family, friends, neighbors, and work colleagues. The contextual effects literature further assumes that these social networks are structured largely by class, ethnic, religious, and other social cleavages. Thus the more homogeneous the community social group context, the more effective will be communal group norms in influencing political attitudes, including partisan attachments. Advanced industrial societies are associated with high levels of urbanization and residential mobility, which produce a decline in community homogeneity and solidarity. The result has been an attenuation of the impact of residential context on voting decisions. For instance, the decline of community in Japan is directly associated with a decrease in support for the conservative LDP, while the same phenomena are associated with a decrease in support for the Socialist party in Sweden (Flanagan, this volume: Chap. 6; Stephens, 1981). In both cases, as communities became increasingly heterogeneous and transient, their capacity to mobilize votes for the dominant party weakened. In some cases, the community's associational networks assumed politically neutral positions to avoid offending elements in a more heterogeneous community. In other instances, social networks became less effective in integrating a growing number of new and socially diverse residents into the mainstream of community opinion. Therefore, changes in the character and composition of community settings are associated with a decline in the clarity of social and partisan alignments. Thus the theory predicts that as the partisan cues emanating from a voter's residential community become weaker and less clearly defined, political volatility should increase. Cognitive Mobilization A fifth approach to contemporary political change focuses on the qualitative increase in the political sophistication of the mass public (Allardt, 1968). Modern society has produced a tremendous expansion of secondary and higher education, as well as a diffusion of greater quantities of political information through the media, especially the electronic media. This, in turn, raised the political resources and cognitive skills of large segments of national populations. With this new level of political sophistication, mass democracy might reach the open, participatory, individualistic style that Rousseau, Locke, Tocqueville, and other philosophers considered essential to the preservation of democracy. For instance, sophisticated and well-informed voters need not depend on social cues or party identification to make their voting decisions; they can make their own decisions based on the issues and candidate positions (Shively, 1979; Borre and Katz, 1973; Dalton, 1984). Political activity is not limited to "elite-mobilized" participation such as campaign or party activity. With more developed political skills, citizens can initiate and focus activism through "elite-challenging" participation. The result is a weakening of traditional political structures and a shift in the style of political participation.

Electoral Change in Advanced Industrial Democracies

19

Cognitive mobilization, then, means that some voters have less functional need for party or social group attachments, and hence those attachments are weakening. On the other hand, this phenomenon also means that some previously unpoliticized and unmobilized constituents can be reached more effectively by political information and brought into the political process. This potential for expanding the politicized portion of the electorate is particularly evident in cases where the electronic media and computer mailings are used to target and mobilize previously apolitical constituents around new issues and movements. While the first type of voter is dealigned and not likely to develop an enduring partisan attachment, the second type presents political elites with a sizable pool of previously marginal participants that may become aligned with a party. Either of these processes, the weakening of old party ties or the mobilization of new attachments, may produce electoral change. And both phenomena seemingly are related to a growing plethora of singleissue movements and parties—from a women's liberation party in Japan, to an environmental party in Germany, to the right-to-life movement in the United States. Aging Party System A much different view of contemporary political change suggests a "lifecycle" approach to party systems (Clubb et al., 1980; Beck, this volume: Chap. 8). Allegiances to the present party systems often were born from the traumatic socialization experiences of the 1930s and 1940s, or the even earlier partisan contests as suffrage was being extended. With the passage of time, party systems may begin to "age." Some constituent elements of a party's support base may become alienated as its policy promises and slogans are enacted in government programs that fail to provide the expected solutions to group problems. The issues that initially structured party conflict may lose relevance over time, as these issues are resolved or new ones come to the fore. New voters in particular are likely to find little meaning in the old appeals that defined a party's support base. Paul Beck argues that there is often a cyclical pattern to the life of party systems, based on socialization experiences (Beck, 1974). This perspective suggests that the political changes of the past decades represent the strains of aging. A revitalization or realignment of these party systems would adjust them to contemporary political realities. Thus we may not be witnessing a revolutionary change in democratic politics, only a reoccurring "biological" process. Value Change A final approach to the study of electoral change posits a link between advanced industrialism and the values of the mass public. Basic value priorities presumably reflect the childhood environment when these values are first socialized, and they tend to persist through the life cycle. Thus the tremendous

20

Introduction

postwar changes in personal and societal conditions may be altering the value priorities of the mass public. Ronald Inglehart argues that the citizenries of industrial democracies are experiencing a shift from materialist to postmaterialist values (Inglehart, 1977, 1979a). Older generations who often experienced economic need or insecurity during their formative years maintain a relatively high priority for materialist, or Old Politics, values such as economic gains, domestic order, and social and military security. Postwar generations have been reared during a period of unprecedented affluence, economic well-being, and personal security. Consequently, many young people apparently are reaching a saturation level in regard to basic economic needs, and are shifting their attention toward jiostmaterial, or New Politics, goals. Scott Flanagan has conceptualized the process of value change in somewhat different terms—as a decline in respect for authority, conformity, religiosity, and the work ethic (Flanagan, 1982a, 1982b). In place of these more traditional values, he finds a growing emphasis on values that are instrumental for securing the goal of self-actualization—self-assertiveness, nonconformity, openness to new ideas, equality, the pursuit of leisure activities, a better quality of life, and a tolerance for a variety of life-styles. Regardless of how it is conceptualized, a process of value change clearly is occurring. Evidence of generational changes in value priorities is available for almost twenty industrial democracies—including most of those studied in this volume (Inglehart, 1981). New Politics values also are commonly found among other social groups identified with advanced industrial politics: the new middle class and the better-educated. To the extent that these value concerns gain in salience relative to Old Politics concerns, the emerging New Politics value cleavage potentially can restructure contemporary party politics. Each of the aforementioned models contributes to our understanding of the political changes occurring in industrial democracies. But each model also has its weaknesses. For example, the embourgeoisement model sees the working class as the major source of political change. In fact, advanced industrial politics appears linked primarily to change within the middle class. The decline of class voting, for example, is due mostly to the increasing liberalism of the middle class (Baker et al., 1981: Chap. 7; Ladd with Hadley, 1975: Chaps. 5-6; Kemp, 1978; Lipset, 1981a). Similarly, the universities—not the unions— are the institutional home of recent protest movements; the conservative Ivy League, Oxbridge, and Grandes Ecoles of the 1930s have become the antiestablishment bastions of the 1970s. The social mobility, mass society, community integration, cognitive mobilization, and aging theses also are compatible with many of the trends discussed in this volume. However, a weakness of all these models is that they offer no direction for the politics of advanced industrialism. Old patterns will break down, but there is no suggestion of what new patterns might develop

Electoral Change in Advanced Industrial Democracies

21

in their place. These theories are compatible with the rise of protest movements and political activism in the past two decades, but they fail to explain the similarities in the issues of protest across several nations. The observed commonalities between nations were a stimulus for this volume. Thus these theories need to be integrated into a larger model of political change. We believe that the significance and influence of advanced industrialism is not based on any one of these processes, but upon the effect of several reinforcing sources of change. Historical transformations in political systems are very infrequent experiences. Only when several forces overlap—as they do with advanced industrialism—can fundamental political change occur (Toffler, 1980). Thus only by combining several of the theories described previously can we accurately model the process leading to advanced industrial politics. The process begins with the weakening of traditional political alignments, following either the social mobility, mass society, community integration, or the aging party systems thesis. These eroding cleavages mean that many social groups are open to new political appeals and might be mobilized by new issues or a new ideology. If the parties can capture these new issues or ideologies, the widespread partisan mobilization of a realignment may result. Dealignment represents the absence, however temporary, of channeling partisan directions. The growth of these newly "independent" voters is naturally greatest among the groups identified with advanced industrialism: the young, the new middle class, and the highly educated. The postmaterialist values of these same groups provide them with new goals for citizen action. At the same time, the values and societal goals that these groups represent have created a new set of issue cleavages. Groups that feel threatened by the direction in which the New Left agenda is moving society are beginning to organize and formulate counter-ideologies around which a New Right may take shape. As a result, the Old Politics, structured largely on class cleavages, is being replaced by a New Politics based on a new set of societal cleavages— the new middle class versus the old middle class; affluent, skilled, blue-collar workers versus a poor, unskilled, and largely unemployable social substratum; the public sector versus the private sector; the young versus the old; traditional values versus the new morality; the technocrats versus the exponents of direct democracy (increased popular control). These cleavages are crosscutting, and hence it is difficult to predict from our present perspective how they may cumulate and what types of political changes they may stimulate. However, in several of the studies in this volume we can already witness ongoing partisan changes, even if the end states of those processes are still somewhat obscure. Finally, the processes of political change involve more than simply exchanging one set of issues and cleavage alignments for a new set. A fundamental change in social relations also is involved. Citizen direction of social and political life signifies a basic change in the role of the mass public and

22

Introduction

in the distribution of political power between masses and elites. Until recently, such a broadening of decision-making was probably unworkable, even if political elites would have tried. However, the revolutionary growth in cognitive mobilization is broadening the potential for citizen action. In addition, new technologies are increasing the potential for assessing mass preferences and perhaps more importantly are vastly expanding the capabilities of rising political elites to identify and mobilize new issue constituencies. Consequently, some citizen groups on both the Left and Right have struck out in new directions. A process of political change has begun. OVERVIEW

The remainder of this book is organized into five sections. The first section is comprised of two broad theoretical chapters that describe various aspects of the trends and processes that are taking place in advanced industrial societies. The following three sections are composed of single-nation case studies. These have been grouped according to whether the predominant process under investigation can be described as one of realignment, dealignment, or stable alignment. The processes of realignment and dealignment are not mutually exclusive, but generally a single process tends to dominate, which allows for an initial classification of cases. Each section is prefaced by an introduction that identifies the major themes developed in the section and compares and contrasts its constituent studies. The concluding section in the volume draws upon the findings of the various case studies and attempts to place them in a more coherent theoretical framework.

TWO

Frameworks for Analysis

2. The Changing Structure of Political Cleavages in Western Society RONALD INGLEHART Center for Political Studies, University of Michigan Ann Arbor, Michigan

The cleavage structures underlying politics in Western nations have changed profoundly during the past two decades. As a result, the textbook definitions of "Left" and "Right" (or, in the American context, of "liberal" and "conservative") are only partly valid today. Political cleavages are relatively stable patterns of polarization, in which given groups support given policies or parties, while other groups support opposing policies or parties. For almost a generation, the nature of (1) the groups and (2) the policy issues aligned with Left and Right have been changing. In the classic model of industrial society, political polarization was a direct reflection of social class conflict. The working class was considered the natural base of support for the Left—that is, of support for change in an egalitarian direction. And the key issue underlying the Left-Right polarization was conflict over ownership of the means of production and the distribution of income. As industrializing society gives way to advanced industrial society, there is a growing tendency for politics to polarize along a new dimension that cuts across this conventional Left-Right axis. Increasingly, support for social change comes from a postmateriahst base, largely middle class in origin. This group has raised a new set of issues that tend to dominate the contemporary political agenda. Today many of the most controversial issues and the most important poI am indebted to Jacques-Ren6 Rabier, my colleagues in the Political Action Study, my colleagues in the European Elections Study, and the editors of this volume for intellectual stimulation and helpful comments The Commission of the European Communities and the European Parliament supported the fieldwork on which most of this is based The analysis was supported by Grant SES-8208333 from the National Science Foundation The data analyzed here are available from the ICPSR survey data archive, the Belgian Archives for the Social Sciences, and the Zentralarchiv fur empinsche Sozialforschung

26

Frameworks for Analysis

litical movements polarize along a materialist-postmaterialist dimension. The environmentalist movement, the opposition to nuclear power, the peace movement, the women's movement, the limits to growth movement, the consumer advocacy movement—all are manifestations of conflict over an issue dimension that is only loosely related to conflict over ownership of the means of production and to traditional social class conflict. The fact that these movements have taken the center of the stage in contemporary politics reflects a long-term shift in the value priorities of Western publics (Inglehart, 1977, 1981). Thus far, this new axis of polarization has had only a limited impact on voting behavior. Long-established political party loyalties, reinforced by party organizations and institutional linkages with labor unions and churches, are highly resistant to change. People continue to vote for the parties prevailing in their milieu, which their parents or even grandparents may have supported. To a considerable degree, Lipset and Rokkan (1967) were correct in speaking of a "freezing of party alignments" dating back to the era when modern, mass-party systems were established. Although deep-rooted political party alignments continue to shape voting behavior in many countries, they no longer reflect the forces most likely to mobilize people to become politically active. Today the new axis of conflict is more apt to stimulate active protest and support for change than is the class-based axis that became institutionalized decades ago. This disparity between traditional political party alignments and the dynamics of contemporary issue-polarization places existing party systems under chronic stress. For extended periods of time, the traditional party systems may appear to be in business as usual—until suddenly, a basic restructuring occurs. Sometimes the change manifests itself in the emergence of new political parties, as in The Netherlands or Italy. But the capture of long-established parties by new elites is an even more promising avenue to success, for major political parties represent great psychological and institutional investments; established voting patterns are not lightly discarded. At the same time, this inertia means that party alignments can lag behind social change until the major ideological cleavage cuts squarely across established party spaces. When this happens, the alternatives are realignment or dealignment: the parties must either reorient themselves or risk being split—or suffer a gradual erosion of partisan loyalties. In many Western nations, from Great Britain and West Germany to the United States, that situation prevails. FROM CLASS-BASED TO VALUE-BASED POLITICAL POLARIZATION

The idea that politics is a struggle between rich and poor can be traced back to Plato. But unquestionably, the most influential modern version of this idea has been Karl Marx's argument that throughout industrial society, social

Changing Structure of Political Cleavages

27

class conflict is inevitably the central fact of political life. Marx's influence is reflected not only in a vast literature of social criticism, but also in the existence of an entire family of political parties that were inspired by his writings and, in varying degrees, purport to be guided by his analysis today. The idea that politics in industrial societies is a class struggle has received strong support in the findings of empirical social research. Thus in his classic and immensely influential work, Political Man, Seymour Martin Lipset (1960: 223-224) concludes, "The most important single fact about political party support is that in virtually every economically developed country the lower income groups vote mainly for the parties of the Left, while the higher income groups vote mainly for the parties of the Right." In another influential study based on data from four English-speaking democracies, Robert Alford (1963) found that in virtually every available survey, manual workers were more likely to vote for parties of the Left than nonmanual workers. Calculating a "class voting index" (obtained by subtracting the percentage of nonmanual respondents voting for the Left from the percentage of manual respondents voting for the Left), Alford found a mean index of + 16 for the United States, and one of +40 for Great Britain. More recent empirical analyses have demonstrated that religion is also a major factor, but confirmed that social class is one of the most powerful bases of political cleavage, towering above other variables, when it is not dominated by ethnic cleavages such as religion, language, or race (see e.g., Rose and Urwin, 1969; Lijphart, 1971, 1979; Rose, 1974). Nevertheless, there were grounds for believing that a paramount role for social class voting might not be an immutable fact of political life. Campbell et al. (1960) argued that class voting in the United States, to a considerable extent, reflected a cohort effect. It was most pronounced among the generation that came of age during the Great Depression, and weaker among both older and younger groups. They speculated that class voting may vary inversely with prosperity, with substantial time lags due to cohort effects. The present author (1971, 1977) carried this line of reasoning farther, presenting evidence of a pervasive intergenerational shift from materialist to postmaterialist value priorities among the publics of advanced industrial society. The postmaterialist outlook is linked with having spent one's formative years in conditions of economic and physical security. Hence, throughout Western society, it is far more prevalent among the postwar generation than among older cohorts, and tends to be concentrated among the more prosperous strata of any given age group. The political implications are significant and at first seem paradoxical. Postmaterialists give priority to such goals as a sense of community and the nonmaterial quality of life, but they live in societies that have traditionally emphasized economic gains above all. Though they tend to come from the most privileged and economically most favored strata of society, they tend

28

FrameworL· for Analysis

to be relatively dissatisfied with the kind of society in which they live and relatively favorable to social change. Though recruited from the higher-income groups that have traditionally supported the parties of the Right, they themselves tend to support the parties of the Left when they become politically engaged. Conversely, when postmaterialist issues (such as environmentalism, the women's movement, unilateral disarmament, opposition to nuclear power) become central, they may stimulate a materialist reaction in which much of the working class sides with the Right to reaffirm the traditional materialist emphasis on economic growth, military security, and domestic law and order. The rise of postmaterialist issues, therefore, tends to neutralize political polarization based on social class. Though long-established party loyalties and institutional ties link the working class to the Left and the middle class to the Right, the social basis of new support for the parties and policies of the Left tends to come disproportionately from middle-class sources. But at the same time, the Left parties become vulnerable to a potential split between their postmaterialist Left, intensely engaged by new issues, and their traditional materialist constituency. In 1972, this phenomenon temporarily shattered the Democratic party in the United States. In 1981, it contributed to a possibly more permanent division of the British Left, split between a Labour party that had been captured by a neo-Marxist and neutralist left wing, and a new Social Democratic party that won over much of the party's mass constituency. Throughout the past decade, a somewhat similar cleavage has threatened to split the German Social Democratic party, torn between a postmaterialist "Young Socialist" wing and the labor-oriented main body. Though the postmaterialist left was unable to take over the Social Democratic party, it did succeed in launching "Green" or environmentalist parties that had, by 1982, won seats in six of the eleven West German state parliaments. More important, the postmaterialist Left threatened to eliminate the Free Democrats from the federal parliament and take over their role as the party holding the balance of power at the national level. The postmaterialist basis or support for the environmentalist parties in Germany and France has been demonstrated by Buerklin (1981), Muller-Rommel (1982), and Fietkau et al. (forthcoming). In multiparty systems with straight proportional representation, the viability of new parties is greater than in the countries just discussed. Hence in The Netherlands, Scandinavia, and Italy, this phenomenon gave rise even earlier to small but influential postmaterialist parties (Lijphart, 1981). Leftist in policy orientation, their social base is largely middle class. After a lull in the middle 1970s, West European politics again shows widespread political upheaval. And despite the economic difficulties of the present period, postmaterialist issues continue to play a major role. Through-

Changing Structure of Political Cleavages

29

out the Western world, the most massive political demonstrations that have taken place in recent years have not been directed against unemployment or declining real income—on the contrary, the largest and most intense ones have been aimed at preventing the construction of nuclear power plants, highways, airports, military installations, hydroelectric dams, and other projects that might reduce unemployment. Now, as earlier, labor is concerned with unemployment, wages, and inflation, but political activism continues to reflect mainly postmaterialist concerns. Recent economic uncertainty seems to have slowed the growth of postmaterialism in Western Europe but has not stopped it. A postmaterialist value-type was more widespread at the end of the 1970s than at the start of that decade, and had shifted from being predominantly a student phenomenon, to being an important influence among elites (Inglehart, 1981, 1983). Our hypotheses concerning the emergence of a postmaterialist Left imply a long-term decline in social class voting. Has it taken place? Alford (1963:226) examined this possibility himself and concluded, "There had been no substantial shift in the class bases of American politics since the 1930s, despite the prosperity since World War II and despite the shifts to the Right during the Eisenhower era." Alford seems to have been correct in his interpretation of the evidence he examined; indeed, social class voting in the United States actually rose during the period he dealt with, reaching a peak about 1948 as the generation of the New Deal matured. But more recent studies by Glenn (1973), Abramson (1975, 1978), Books and Reynolds (1975), Inglehart (1977), Baker, Dalton, and Hildebrandt (1981), and Stephens (1981) support the conclusion that during the past few decades there has been a secular decline in social class voting, not only in the United States but throughout the Western world. This tendency is probabilistic, not deterministic. A variety of factors affect the voters' choice—among them, long-term party loyalties (sometimes transmitted from one generation to the next), religious and other group ties, the personalities of given candidates, the relative positions of the various parties on key issues, and the current economic situation. These factors can cause large fluctuations in class voting from one election to the next within a given nation, and help account for wide variations in class voting between countries. But a growing body of evidence points to the conclusion that underlying these fluctuations and cross-national differences, a long-term decline in class voting has taken place during the past thirty years. Thus in the revised edition of Political Man, Lipset (1981a) updates his own earlier conclusions about social class voting with a chapter that sums up several of the findings just cited. His graph, which I updated, is shown in Figure 2.1. The fluctuations we see in Figure 2.1 are sometimes dramatic, but the downward trend is unmistakable and seems to have continued into the 1970s, despite the economic setbacks of that decade.

30

Frameworks for

Analysis

FIGURE 2 1 The Trend in Class Voting in Four Western Democracies, 1948-19833 ALFORD INDEX OF CLASS VOTING

YEAR

SOURCE: Inglehart (1983). NOTE: Table entries are Alford indices of class voting

As Figure 2.1 demonstrates, class voting in the United States fell almost to zero in 1972, when the McGovernites captured the Democratic presidential nomination, mobilizing the postmaterialist constituency effectively, but also bringing a massive desertion of working-class voters. Many of the latter returned to their traditional party allegiances, under a centrist candidate in 1976. Still, class voting in the United States remains low—and even this modest level largely reflects its persistence among older voters: among the youngest American age cohorts, it is close to zero (Abramson, 1978). West European data show a similar pattern suggesting gradual decline, linked with generational differences. For the European community as a whole between 1976 and 1979, the class voting index for those more than fifty-four years old was +24; for those aged eighteen to thirty-four, it was only + 15. Class voting has declined. However, to grasp the implications of this phenomenon, we need to know why it has taken place. Does it reflect an intergenerational value change of the type we have hypothesized? If so, we can anticipate that it will continue, as younger, relatively postmaterialist age groups replace the oldest, most materialist-oriented age cohorts in the electorate. Or is the phenomenon a direct reflection of current economic conditions? In this case, we would expect a reversal of the downward trend in the present era, and a possible return to the politics of social class conflict that characterized the 1930s and 1940s. These are part of a set of questions that will be addressed in this chapter, for the decline of social class voting is only one aspect of a broader trans-

Changing Structure of Political Cleavages

31

formation of political polarization. And in some ways, class voting patterns give an understated impression of what has been going on. For voting behavior is shaped, to a considerable extent, by an internalized sense of political party identification in given individuals, and by institutional ties between given parties and given social networks, such as labor union or church. Insofar as this is true, voting behavior has a good deal of inertia; it does not necessarily respond to current conditions, but may continue to reflect old alignments long after the circumstances that gave rise to them have changed. Other indicators of political polarization may reflect the dynamics of contemporary politics more directly. For example, one of the standard questions in the Euro-Barometer surveys, sponsored by the European communities, asks whether the respondent favors revolutionary political change, gradual reform, or defense of the established order. One can use responses to this item to measure social class polarization by examining the differences between the responses of those with manual and nonmanual occupations. Because this item does not require the respondent to indicate a political party preference, it is less constrained by the influence of long-term political party loyalties than is the Alford index—and should reflect the decline (or rise) of social class polarization more immediately than does the latter. In an era of declining class polarization, this item should show smaller social class differences than those linked with party preference; in times of sharply rising social class conflict, it should show more class polarization than the Alford index. We will also use the respondent's self-placement on a Left-Right ideological scale as an indicator of political polarization. Previous research has demonstrated that this measure reflects a partisanship component, as well as an ideological component (Inglehart and Klingemann, 1976). Some respondents place themselves at a given point on this scale because that is where the party they support is conventionally located; placing oneself on the Left (or the Center, or Right) is more or less a surrogate for party identification. For these people, Left-Right self-placement would have much the same inertia as party identification itself. For many respondents, however, this scale taps one's overall ideological position: it seems to be a summary measure of one's stand on the most important current political issues. Insofar as this is true, our hypotheses imply that the political meaning of "Left" and "Right" (or of liberal and conservative, in the American sense) has been changing. With the rise of new issues, identification with the "Left" increasingly would come to connote support for new causes such as environmentalism, with a diminishing tendency to evoke the classic issues such as nationalization of industry. Similarly, selfplacement on the Left would have a declining linkage with working-class status.

32

Frameworks for Analysis

To some observers, this prospect seems almost inconceivable. Commenting on a surprisingly weak observed correlation between Left-Right self-placement and social class, Budge, Crewe, and Farlie (1976: 135) argue, "In our opinion the absence of a class influence on the Left-Right continuum is somewhat surprising and must raise doubts about its validity. For if the working class are not substantially located to the Left and the middle class to the Right, what meaning does the continuum have?" If we define the Left as that portion of the spectrum supported by the working class, then this finding does, of course, invalidate Left-Right selfplacement. But this is a circular and rather fruitless definition of the LeftRight dimension. If, as we argue, this dimension is a summary measure of one's overall ideological position—based on the issues that are most salient at a given time—then the relationship between the Left or Right and any given social group is an empirical question, and one that is subject to change over time. In West Germany, the sudden rise of new political movements has made the question "Is the meaning of Left and Right changing?" a subject of more than academic interest (Murphy et al., 1982; Klingemann, 1982; Buerklin, 1982). We will present evidence that the Left-Right ideological dimension does tend to assimilate whatever issues are most salient—and that its meaning has, to a surprising degree, already shifted to reflect the New Politics dimension. The two key hypotheses underlying this chapter can be summarized as follows: 1. An Issues Polarization hypothesis: A new issues dimension has attained a salience that now approaches that of the economic issues dimension that traditionally has been considered the basis of political polarization. The former dimension has arisen so recently that it has not yet been assimilated into one overarching Left-Right dimension. By contrast, the clerical/anticlerical split has largely become assimilated to a Left-Right partisanship dimension; in the long run, this may also happen to the new noneconomic issues dimension. However, since the issues underlying this dimension have not yet been resolved or institutionalized, they now constitute a more potent source of discontent and support for change than does the conventional Left-Right dimension. 2. A Group Polarization hypothesis: Closely linked with the rise of a new issues dimension has been the rise of a new axis of group polarization, alongside the familiar working class-middle class polarization. The growing salience of both this group polarization axis and the new issues dimension reflects a shift in the value priorities of Western peoples. The sources of these structural changes can be traced on two levels: at the individual level, the emergence of a politically active and articulate postmaterialist minority has had a major impact on both the issue agenda and the

Changing Structure of Political Cleavages

33

group basis of politics in Western nations. Placing greater importance on the social and aesthetic quality of life than on economic and physical security, the postmaterialists have emphasized new issues (such as environmentalism) or brought a new perspective to ageless ones (such as military expenditures). Not only are the postmaterialists themselves more apt to respond to these issues than to the classic labor-versus-management issues, but by bringing them to the center of the stage, they have sometimes engendered a materialist reaction. This reaction mobilizes segments of the working class, as well as the traditional middle class, in defense of materialist values—and in opposition to proposed social change. Though a minority, the postmaterialists now tend to control the issue agenda, and their impact tends to reshape patterns of group polarization. At the societal level, these shifts can be viewed as a logical response to changing circumstances. Economic issues are less urgent at a high level of economic development than at a low one. Economic growth is almost necessarily given priority by poor societies, once it is realized that it is possible to attain, and can bring an end to starvation. But at a high level of development, economic growth may no longer be a means to avoid hunger, but a means to provide the average family with a second car. This goal not only has less urgency, but may introduce elements of noise, pollution, and crowding that can become counterproductive to the maximization of human utilities. At both individual and societal levels, there tend to be significant time lags between economic change and its political consequences—which is why the New Politics began to emerge a number of years after the various postwar economic miracles. At the individual level, political change is linked with the process of intergenerational population replacement. Postmaterialism began to have a major impact only when the postwar generation reached a politically relevant age in the late 1960s. At the societal level, political change theoretically could take place rather quickly—except that it tends to be retarded by social networks and institutional ties that can be highly resistant to change. Political party identification, in particular, tends to resist changes in established political patterns, influencing an individual to remain loyal to whatever party he/she supported in the past, and even whatever party his/her parents supported. Hence if social class voting was strong in the past but has been weakened by relatively recent factors, it will be preserved most strongly among those who have relatively strong loyalties to established parties. Conversely, the impact of postmaterialism will be greatest on those political orientations that are least strongly linked with established party loyalties. Two FACES OF LEFT AND RIGHT

Our first hypothesis is that a new dimension of political conflict has become increasingly salient, reflecting a polarization between materialist and post-

34

Frameworks for Analysis

materialist issue preferences. In order to test this hypothesis, let us examine how Western elites and publics polarize in response to a battery of thirteen items included in surveys carried out in all nine nations belonging to the European community in spring 1979. Surveys were conducted simultaneously with: (1) representative national samples of the publics of each nation (as part of the Euro-Barometer surveys) and (2) a sample of 742 candidates running for seats in the European Parliament. The latter sample should give a reasonably good indication of the issue preferences of West European political elites. It includes politicians belonging to all of the important political parties in all nine nations. In social background, these respondents resemble the members of the respective national parliaments (in which many of them hold seats). Our battery of questions was designed to measure preferences on a wide range of issues: not only those that have become salient in recent years (such as nuclear power, terrorism, and abortion) but also such classic economic issues as nationalization of industry, redistribution of income, and the government role in the economy This battery was worded as follows: We'd like to hear your views on some important political issues. Could you tell me whether you agree or disagree with each of the following proposals? How strongly do you feel? (Show CARD) 1. Stronger public control should be exercised over the activities of multinational corporations. 2. Nuclear energy should be developed to meet future energy needs. 3. Greater effort should be made to reduce inequality of income. 4. More severe penalties should be introduced for acts of terrorism. 5. Public ownership of private industry should be expanded. 6. Government should play a greater role in the management of the economy. 7. Western Europe should make a stronger effort to provide adequate military defense. 8. Women should be free to decide for themselves in matters concerning abortion. 9. Employees should be given equal representation with shareholders on the governing boards of large companies. 10. Economic aid to Third World countries should be increased. 11. Stronger measures should be taken to protect the environment against pollution. 12. Stronger measures should be taken to protect the rights of individuals to express their own political views. 13. Economic aid to the less developed regions of the European community should be increased.

35

Changing Structure of Political Cleavages

The respondent was shown a card offering the following categories for response to each item: "Agree Strongly," "Agree," "Disagree," and "Disagree Strongly." Table 2.1 shows the results of a factor analysis with varimax rotation, based on responses to this battery of items, among candidates for the European Parliament. For reasons of space, only the results from a pooled sample of 742 candidates from all nine nations are shown here; separate nation-bynation analyses show essentially the same patterns, with minor variations. TABLE 2.1 Factor Analysis of Issue Positions of Candidates to the European Parliament 2. Noneconomic Left-Right (14%)

1. Economic Left-Right {37%) More government management of economy More public ownership of industry Reduce income inequality Public control of multinationals Equal representation for employees More aid to Third World

.764 .708 .642 .633

Stronger measures against terrorism Develop nuclear energy Stronger defense effort Women free to choose abortion More public ownership of industry

.776 .733 .727 -.574 -.451

.615 .372

SOURCE: Survey of Candidates for European Parliament conducted in spring 1979. For sampling details, see Inglehart et al. (1980). NOTE: Table entries are factor loadings from a factor analysis with varimax rotation. All factor loadings above .300 are shown.

The expected pattern emerges, with striking clarity. Our first factor is based on six items designed to tap the classic economic concerns; the most sensitive indicators of this dimension are one's attitude toward government management and ownership of the economy. The second factor shows a quite distinct content: its four highest-loading items are those designed to tap the New Politics. Nuclear energy and abortion are new issues—they literally did not exist as political issues a generation ago; terrorism has a long history, but its present form is new. Defense, obviously, is not a new issue—quite the contrary, it is probably the oldest concern of the state. But domestic opposition to one's own defense establishment took on new overtones during the war in Vietnam. Opposition to the war came to be motivated much less by traditional conservative reasons (above all, opposition to heavy government expenditures and higher taxes) than by a postmaterialist concern for the impact of the war on the purported enemy. Though the defense issue is ancient, both the mo-

36

FrameworL· for Analysis

tivations and social bases that underlie it have changed. A fifth item—concerning public ownership of industry—clearly does not fit our expectations, but it is by far the weakest-loading item. Its presence here signals the fact that this question plays a salient and pivotal role in the ideological structure of professional politicians—something that is not equally true of mass publics. As we will demonstrate shortly (see Table 2.3), the issue preferences of Western publics are structured in an almost identical fashion: similar analysis also reveals two dimensions, based on almost exactly the same items as those in Table 2.1—except that "public ownership of industry" does not load on the second factor. We hypothesize that the second dimension reflects a materialist/postmaterialist polarization, rather than traditional social class conflict. Whether or not this is true remains to be demonstrated. First, let us examine the degree to which we actually have two distinct dimensions. Varimax rotation can identify two or more independent components of an attitudinal structure even if the variables are only relatively distinct. And among the elites, these two dimensions are only relatively distinct. The mean correlation among the three highest-loading items on the first dimension is .50; the mean correlation among the three highest-loading items on the second dimension is .45; the mean correlation between the two sets of items is - .33. In other words, at the elite level we find two distinguishable issue clusters, but they are by no means unrelated. In a principal components analysis, all of these items show substantial loadings on what could be interpreted as an overarching Left-Right ideological dimension, or superissue. Nevertheless, it is meaningful to distinguish between these two issue clusters. Indeed, unless we do so, we lose sight of a major shift in the meaning and social bases of Left and Right. Moreover, though they tend to be integrated into an overarching Left-Right structure at the elite level, among the general public the two clusters are almost totally unrelated. To be specific: among European publics, the mean correlation among the three items concerning public ownership, public management, and income inequality is .28; the mean correlation among the items concerning terrorism, nuclear energy, and defense is .23; but the mean correlation between the two sets of items is — .05. At the public level, we are dealing with two completely independent dimensions. In part, this finding reflects a pronounced and pervasive tendency for mass publics to show less attitudinal constraint than elites. But it is also true, as we will see later, that the two issue clusters are fundamentally different in nature and antecedents. The fact that the two issue dimensions are distinct and relatively independent does not mean that they are unrelated to a broader Left-Right orientation, even among mass publics. For politics frequently demands a dichotomous choice: a politician must join or oppose a given coalition, or a voter must choose between Giscard and Mitterrand. The effort to build a winning coalition

Changing Structure of Political Cleavages

37

provides a powerful incentive to depict politics in bipolar terms that dichotomize the good guys and the bad guys. The Left-Right image is an oversimplification, but an almost inevitable one that in the long run tends to assimilate all important issues. We suggested earlier that the Left-Right dimension, as a political concept, is a higher-level abstraction used to summarize one's stand on the important political issues of the day. It serves the function of organizing and simplifying a complex political reality, providing an overall orientation toward a potentially limitless number of issues, political parties, and social groups. The pervasive use of the Left-Right concept through the years in Western political discourse testifies to its usefulness. Insofar as political reality can be reduced to one underlying dimension, then one can distinguish readily between friend and foe, and between the good and bad positions on given issues, in terms of relative distances from one's own position on this dimension. To be sure, social conflict is rarely if ever unidimensional. Thus, to speak in terms of Left and Right is always an oversimplification—but an extremely useful one. In order to describe individually the relationships between a mere dozen issues or parties, one would need to make sixty-six pairwise comparisons; fourteen issues or parties would require ninety-one comparisons. This degree of cognitive complexity is hopelessly unmanageable in practical politics. Ideologues and politicians almost inevitably tend to sum up the alternatives in terms of such all-embracing concepts as "Left" and "Right" that provide a relatively simple guideline for forming alliances or appealing for mass support. The core meaning of the Left-Right dimension, we believe, is whether one supports or opposes social change in an egalitarian direction. Typically, the Left (or, in America, the liberal side) supports change, while the Right opposes it (see Lipset et al., 1954). It is also important to specify the direction of desired change. While conservative movements may be content to defend the status quo, reactionary ones may seek change in the direction of greater inequality between classes, nationalities, or other groups. The utility of the Left-Right concept rests on the fact that through the years, and from one setting to another, the basic political conflicts quite often do reflect a polarization between those seeking social change and those opposing it. The concept is sufficiently general that as new issues arise, they usually can be fitted into the framework. The specific kinds of change may vary, but the question of more or less equality is usually involved, whether it be between social classes, nationalities, races, or sexes. Moreover, there is some continuity in which groups seek change. Generally, those who are least favorably situated in a given social order are most likely to support change. Hence over the years, certain social groups and political parties have come to be identified with either the "Left" or the "Right." Representative samples of the publics of the nine European community

38

Frameworks for Analysis

nations have repeatedly been asked, "In political matters people talk of 'the Left' and 'the Right.' How would you place your views on this scale?" When shown a scale with ten boxes ranging from "Left" to "Right," the overwhelming majority of respondents place themselves at some point on the scale, with little hesitation. These responses generally bear a coherent relationship to the respondents' other views. For example, in keeping with our concept of the core meaning of Left and Right, those who are most supportive of social change are likely to place themselves toward the Left end of this scale. A subjective sense of identification with the Left or the Right (or the Center) is widespread in Western Europe—but just what does it mean? It could, conceivably, be something similar to political party identification. In given countries, there is a consensus that given political parties are located at either the Left or the Right (or extreme Left, Center, or extreme Right). Originally, such images may have been based on the party's stand on salient issues, but over time they might become stereotypes that do not necessarily bear much relationship to current issues. How widespread is this phenomenon? A crossnational survey conducted in Britain, Germany, The Netherlands, Austria, and the United States in 1974 asked the question about Left-Right self-placement cited previously, and then followed it up with this open-ended question, "What does 'Left' mean to you?" . . . "What does 'Right' mean to you?" (For sampling information, see Barnes, Kaase et al., 1979.) In the four European countries, from one-fifth to one-half of the respective samples defined "Left" by referring to specific political parties; a slightly larger proportion defined "Right" in the same way. Only about 1 percent of the American public responded with party labels. The terms "Left" and "Right" have come into widespread use only recently in the United States and have not become generally accepted stereotypes for American political parties; but this does seem to be the case, to a considerable extent, in Western Europe. One component of the meaning of the Left-Right dimensions, then, seems to be the perception that given political parties are linked with specific points on the continuum. This might reflect an accurate summary of each party's current position on key issues—but it could also be a stereotype that persists long after the events that gave a given party a given image. It is clear, however, that (for a substantial share of the public, at least) the terms "Left" and "Right" have a meaning that goes beyond outdated stereotypes. In the 1974 surveys, in each country except Britain about half of the sample defined "Left" in terms of some ideology or with reference to more or less government, or to social or political change; and about 40 percent defined "Right" in similar broad, abstract terms. But what is the current meaning of Left and Right in terms of specific issues? In order to answer this, let us examine the correlations between LeftRight self-placement and the battery of items designed to measure preferences on both the classic economic issues, and some newer issues. To what extent

Changing Structure of Political Cleavages

39

have the new issues become assimilated to the Left-Right dimension? To what extent do the new and old issues give rise to separate axes of polarization? Table 2.2 provides an answer to the first question. One's stand on the traditional economic issues shows substantial correlations with Left-Right selfplacement in every case. For example, those who were most supportive of TABLE 2.2 Correlation between Left-Right Self-placement and Position on Specific Issues among Western Elites and Publics, 1979 1. Among Candidates for European Parliament Issue More public ownership of industry More government management of economy Stronger defense effort More control over multinationals Reduce income equality Women should be free to choose abortion More aid to Third World countries More severe antiterrorist measures Develop nuclear energy Employees equal representation on boards Stronger antipollution measures Protect freedom of expression More aid to poorer regions of Europe

2. Among Publics of Nine European Community Nations r = .617 .599 -.553 .519 .502 .474 .467 -.454 -.454 .342 .300 .262 .183

Issue Stronger defense effort Employees equal representation on boards Reduce income inequality More public ownership of industry Women should be free to choose abortion Develop nuclear energy More severe antiterrorist measures More government management of economy More control over multinationals Protect freedom of expression More aid to poorer regions of Europe More aid to Third World countries Stronger antipollution measures

r = -.355 .277 .271 .235 .200 -.200 -.198 .198 .197 .191 .176 .174 .126

SOURCES: 1979 Survey of candidates for the European Parliament, and EuroBarometer 11 (April 1979). NOTE: Table entries are Pearson product-moment correlations. Positive polarity indicates that those who support a given issue tend to place themselves on the Left. Based on pooled data from candidates from all nine nations, as well as publics, weighted according to population and (with the candidate data) according to party strength in the European Parliament.

40

Frameworks for Analysis

greater efforts to reduce income inequality showed a marked tendency to place themselves on the Left. But by 1979, the new issues were also integrated with the Left-Right orientation of both elites and mass publics to a truly impressive degree. The general pattern is similar among both elites and publics. At both levels, the items that correlate most strongly with Left-Right self-placement are the top-loading items on the economic Left-Right dimension and the noneconomic Left-Right dimension, respectively. In other words, our strongest indicators of both dimensions seem to have the greatest impact on whether an individual views himself/herself as located on the Left or the Right. But there are significant differences between elites and general publics. For one, these correlations are consistently stronger at the elite level than at the mass public level. On the average, one's Left-Right self-placement explains over four times as much variance in issue preferences at the elite level as at the mass level. Together with the evidence cited previously, this provides powerful support for Converse's viewpoint in the long-standing debate over the Nature of Belief Systems among Mass Publics (Converse, 1964, 1970, 1975a; Pierce and Rose, 1974; Achen, 1975; Nie, Verba, Petrocik, 1979; Sullivan et al., 1979), Constraint is much greater at the elite level, and since virtually identical questions were asked of both elites and general publics, the relatively low level of constraint among mass publics must be due to relatively low levels of political interest or other characteristics of the publics themselves, rather than to some artifact of the survey instrument, such as poor questionnaire construction. Of more immediate concern, however, is the fact that the economic issues are more closely linked with the elites' sense of Left-Right self-placement than are the noneconomic issues. Among the general publics, however, the situation is reversed: the noneconomic issues figure somewhat more prominently. Indeed, the strongest predictor of mass Left-Right self-placement— by a wide margin—is one's attitude toward national defense. The classic issues of government ownership and management of industry continue to define the terms "Left" and "Right" among political elites. These are the textbook examples of Left-Right issues, the stereotypes that figured prominently in the rhetoric and the literature on which the elites were socialized. But the mass public has not read the classic literature. Consequently, the textbook issues do not have the same resonance among mass publics as among elites; for the public, the connotations of Left and Right seem to be more influenced by current events than is true of the elites. Thus when we perform a factor analysis of issue orientations among the public, the same two dimensions emerge as among elites—but Left-Right self-placement tends to load on the New Politics dimension, rather than on the economic issues factor, as Table 2.3 demonstrates. Among the general public, the issue that showed the strongest correlation with Left-Right self-placement was sup-

Changing Structure of Political Cleavages

41

TABLE 2.3

Factor Analysis of Issue Preferences and Left-Right Self-placement of Western Publics, 1979 1. Economic Left-Right More economic aid to less developed regions Larger goverment role in managing economy Equal representation for employees and owners on boards Greater effort to reduce income inequality Stronger effort to protect free expression More economic aid to Third World countries More public ownership of industry Stronger antipollution measures Stronger public control over multinationals Self-placement on Left-Right scale

2. Noneconomic Left-Right .615 .583 .576 .559 .565 .553

Stronger military defense effort More severe penalties for terrorism Nuclear energy should be developed Women should be free to decide about abortion Equal representation for employees and owners on boards Self-placement on Left-Right scale

.694 .529 .516 -.346

-.300 .636

.514 .486 .452 -.260

SOURCE: Euro-Barometer 11 (April 1979). NOTE: Table entries are factor loadings from

a factor analysis with varimax rotation. All factor loadings above .250 are shown. port or opposition to a stronger military defense effort. One would not expect this a priori. At various times in the past, the relationship between Left and Right, and support for defense expenditures, seems to have fluctuated and even reversed polarity. In the first twenty-two months of World War II, for example, Western Communist parties opposed taking any part in the war, which was held to be a struggle among the ruling classes. After the Nazi invasion of the Soviet Union, partisans of the extreme Left reversed their positions and became some of the most ardent advocates of an all-out effort against Hitler. In the United States, before Pearl Harbor, Roosevelt and other liberals struggled desperately to build up the military preparedness of the United States and the West, against opposition that was particularly strong in conservative circles. Though the Japanese attack brought virtually unanimous support for the war effort, there was a reprise of the earlier situation after the war. Conservative Republicans, championed by Robert Taft, advocated reduced defense expend-

42

Frameworks for Analysis

itures and a withdrawal to Fortress America, while liberals supported a strong stand in the Cold War. This pattern seems to have persisted in the United States as recently as 1960, when Kennedy won victory over Nixon with a campaign that promised to close the "Missile Gap" and take a strong stand against Chinese threats to seize the islands of Quemoy and Matsu. Accordingly, in his analyses of the issue positions taken by Western political parties from 1957 to 1962, Janda's expectation was that the parties of the Left would be relatively favorable to higher allocations to the military. In the West as a whole, he found that support for higher military allocations turned out to be linked with the Right rather than the Left, but the association was very modest (Janda, 1970). It was probably the war in Vietnam that brought a clear and strong reversal of the earlier relationship. Opposition to the war became a major postmaterialist cause, linked with humanitarian (rather than economic) concerns, as well as opposition to the hierarchical authority patterns of industrial society. By the end of the 1970s, the military defense issue was— by a clear margin—the strongest correlate of Left-Right self-placement among Western publics. And such new issues as abortion, nuclear power, and measures against terrorism showed correlations with Left-Right self-placement that were as high as, or higher than, those of some classic welfare-state issues such as government management of the economy. The absorptive power of the Left-Right concept is all the more impressive in view of the fact that attitudes toward nuclear power and terrorism showed quite weak correlations with Left-Right self-placement among older respondents; nevertheless, the relatively strong correlations among the young brought the overall figures up to the levels shown in Table 2.2. To be specific: among respondents aged fifty-five and over, the correlation between Left-Right selfplacement and attitudes toward nuclear power and terrorism, respectively, were . 118 and .090. Among those aged fifteen to thirty-four, the figures were .246 and .265. Buerklin (1982) presents complementary evidence: among older Germans (over fifty years of age), the strongest predictor of Left-Right self-placement is whether or not one is a member of the working class. Among younger Germans (eighteen to twenty-nine years of age), Left-Right orientations are virtually unrelated to social class—but strongly related to attitudes toward New Politics issues such as nuclear power and the Green political party. It seems that the linkage between these issues and mass Left-Right orientations is recent, and so far has fully penetrated only the younger groups. For the classic economic issues, on the other hand, age makes no difference. The older groups show correlations between Left-Right self-placement and economic issues that are as strong as, or stronger than, those found among the young. The relationship between issue positions and Left-Right self-placement among mass publics tends to be somewhat curvilinear, which reduces the strength of the correlations in Table 2.2. Figures 2.2 and 2.3 show the mean

Changing Structure of Political

Cleavages

43

levels of support for seven key issues, among respondents placing themselves at each of the ten points on our Left-Right ideological scale.1 Again, to economize space we pool the data from all nine European community nations surveyed in 1979. Though interesting cross-national differences exist, the figures give a good idea of the general pattern. On these graphs, a mean score of 2.5 is the neutral point: a group with that score is evenly divided between support and opposition to the given proposal. A mean score of 3.0 could be obtained if 100 percent of the people in a given group were ' 'for'' the measure; or if 50 percent were "strongly for" and 50 percent were "against"; or by various other combinations. A score of 4.0 could be obtained only if everyone in a given group were "strongly for" that particular measure. With at least several hundred respondents located at each of the ten points on our LeftRight scale, we never obtain such extreme scores. The curvilinearity we observe results mainly from the fact that the respondents who identify with the extreme Right do not hold the most conservative positions on the various issues. This pattern is quite consistent. It is most pronounced in connection with attitudes toward expanding government regulation of the economy—where those placing themselves at the extreme Right (code 10 on our scale) take an issue position that is more "leftist" than all but the two extreme Left categories (codes 1 and 2 on the scale). This pattern (shown in Fig. 2.2) reflects the fact that the people located at both extremes of the scale tend to be dissatisfied with the way society and politics are functioning and feel that radical changes are needed. Furthermore, they are likely to see the government as the only possible instrument to achieve these changes. One implication is that, under extreme conditions, both the extreme Left and the extreme Right have a relatively high potential for totalitarianism. Government regulation is not totalitarian per se—but it may become so when carried to the point where government controls pervade every aspect of one's life. The two extremes of the political spectrum are relatively favorable to drastic change, and, therefore, to state intervention. Curvilinearity is less extreme in connection with other issues, and disappears almost completely in attitudes toward environmental protection. But the absence of curvilinearity in this case is linked with the fact that there is very little variation in attitudes toward protecting the environment: practically everyone is for it. Thus despite the absence of curvilinearity, it shows a very weak relationship with Left-Right self-placement. When this issue is reformulated to focus on the trade-off between environmental protection and economic growth, we do obtain substantial variance—and then we do find a relatively strong relationship with Left-Right self-placement. 1 Figures 2 2 and 2 3 include the three highest-loading issues on each of the two factors in Table 2 1 These key issues tend to be the strongest correlates of Left-Right self-placement For contrast, we also include one issue that shows a very weak relationship to one's sense of Left and Right protection of the environment against pollution

Frameworks for Analysis

44

FIGURE 2 2 Economic Issue Positions and Left-Right Selfplacement among Western European Publics, 1979 4.00

3.50

3.00

— 2.50

2.00

(.50 LEFT

SOURCE: Euro-Barometer 11 (April 1979)

The curvilinear relationship between Left-Right self-placement and support for government control of the economy is not a transient fluke. The finding is replicated in the November 1981 Euro-Barometer surveys (data not shown). Here again, only the two groups at the extreme Left (codes 1 and 2) are more favorable to increased government management of the economy than those who place themselves on the extreme Right (code 10). Attitudes toward public ownership also show a curvilinear tendency, though to a lesser degree; the extreme Right resembles those at the middle of the Left-Right continuum on this issue. This curvilinear pattern characterizes our samples from Britain, France, Germany, Italy, Belgium, The Netherlands, and Ireland, when they are analyzed separately. It does not apply to Denmark or Greece: in these two

Changing Structure

of Political

45

Cleavages

FIGURE 2 3 Noneconomic Issue Positions and Left-Right Self-placement among Western European Publics, 1979 40

35

ι

30

•-..

/

/^ / /

^ /s

/

/*

Λ\1

Α—τ-Α

More severe penalties for acts of terrorism

1 Λ

/ * // \\

1

^i—·'

.>:/

Λ ^

Develc ping nu clear en ergy

Γ

'

A stronger effort to provide military defense

/ jt

/

^25

s

/ / l'

8. CL

Ο

20

15

t

I LEFT

/ ,'

2

3

4

5

6

7

8

9

10 RIGHT

SOURCE Euro-Barometer 11 (April 1979)

countries, the extreme Right does take the extreme position on the classic economic issues. Interestingly enough, the two deviant cases fall at opposite extremes on the developmental spectrum; Greece is by far the poorest country in the European Community, and Denmark the richest. Curvilinearity in which the extreme Right turns back toward the Leftist position is visible in connection with both of the other two economic issues; but it is a particularly strong feature of attitudes toward public ownership of industry (where the extreme Right holds attitudes similar to those of the center Left), and a relatively weak one in connection with attitudes toward a more equal distribution of income. We suggested previously that support for change in an egalitarian direction was the core meaning of the Left-Right superissue. If this is true, it seems appropriate that attitudes toward reducing inequality

46

Frameworks for Analysis

of income have the strongest correlation with Left-Right self-placement of all the classic economic issues. A curvilinear tendency on the extreme Right also exists among the three issues depicted in Figure 2.3, though it is of negligible size in connection with attitudes toward terrorism. In the latter case, we again have a proposal that is backed by an overwhelming majority of the public. All segments of the Left-Right spectrum favor more severe penalties for acts of terrorism. There is a modest but perceptible curvilinearity at the extreme Right in attitudes toward military defense and the development of nuclear energy but the phenomena is less pronounced here than in connection with the classic economic issues. Though the most sensitive indicators of both the old and new issue dimensions show strong linkages with Left-Right self-placement, the position of the extreme Right on these issues is anomalous. Those who see themselves at the extreme Right do not hold views at the opposite extreme from the views of the extreme Left—in some ways, their views are relatively close. In part, Right extremism lies in the fact that its partisans feel extreme changes are needed—their position is not just a further extension of the conservative viewpoint. Turning to cross-national comparisons, the evidence suggests that at high levels of economic development, public support for the classic economic policies of the Left tends to diminish. As Table 2.4 demonstrates, Greece is by far the poorest country among the ten European community nations surveyed in 1981; and the Greek public has—by far—the highest level of support for both nationalization of industry and government management of the economy. At the opposite end of the developmental spectrum, Denmark is the richest country and the Danish public has the lowest level of support for these policies. By purely economic criteria, France should rank fifth, but it actually ranks third in support for these policies—a fact that may be related to a surge in support for such policies, linked with the remarkable electoral victories of the French Left in 1981. But this is the only anomaly. In all other respects, there is a perfect fit between economic developmental level and support for the classic economic policies of the Left. These findings are consistent within our theoretical framework: the principle of diminishing marginal utility seems to apply at the national level as well as the individual level. Greece is an economically underdeveloped country, with extreme contrasts between rich and poor. In such a context, it seems likely that the balance between rich and poor can be redressed only by strong government intervention. Denmark, on the other hand, is a relatively rich country that has long since developed some of the most advanced social welfare policies in the world—as well as one of the highest rates of taxation. Well over half of the Gross National Product is spent by the government;

Changing Structure of Political Cleavages

47

TABLE 2.4 Economic Development and Support for the Classic Economic Policies of the Left among Publics of European Community Countries, 1981 Rank, in Support far "Left" Economic Policies

Nation

Public Ownership

Government Management

Mean

Gross National Product/ Capita (1979)

1. 2. 3. 4. 5. 6. 7. 8. 9.

Greece Ireland France Italy United Kingdom Netherlands Belgium West Germany Denmark

1.60 2.27 2.58 2.95 2.57 2.94 2.92 2.72 3.20

1.25 1.85 1.95 1.85 2.25 1.98 2.21 2.47 2.45

1.43 2.06 2.27 2.40 2.41 2.46 2.57 2.60 2.83

$ 3,890 4,230 10,030 5,240 6,331 10,240 10,890 11,730 11,900

Support for:

SOURCES: Euro-Barometer 16 (November 1981); for sampling information, see ICPSR code book. Data for Luxembourg are omitted because of small sample size (N = 300).

Denmark is approaching the point where it becomes literally impossible to move very much farther in this direction. In Denmark, further redistribution by the government seems far less urgent than in Greece—and the costs of government intervention impinge on a very large share of the population. The incentives to press still farther with the traditional economic policies of the Left become relatively weak, and public resistance relatively strong. The neoconservative claim that these policies have failed seems inaccurate. Insofar as they succeed, they approach a point of diminishing returns. Figure 2.4 shows how the electorates of specific political parties stand on the two issue dimensions; Figure 2.5 gives the same information for the candidates of these parties. Rather than give a composite score based on combined results for several issues, we have selected two specific issues that convey the overall picture through a concrete illustration. We have chosen the respective groups' attitudes toward further nationalization of industry as an example of the classic economic issues; and we have chosen attitudes toward developing nuclear power as an example of the new noneconomic issues. Another reason why we chose these particular issues is because they lend themselves to graphic presentation, since they show something close to a 50:50 split, with about equal numbers of parties being predominantly "for" and "against" each proposal. These two issues convey a good idea of the relative positions of the respective parties on other key issues, since a given

48

Frameworks for Analysis

party's stand on each of these issues is strongly correlated with its stand on the other issues having high loadings on the same dimension. At the elite level, for example, we find a mean correlation of .76 between a given party's positions on nationalization of industry, government regulation, and redistribution of income. Similarly, there is a mean correlation of .84 between a given party's positions on nuclear power, defense, and measures against terrorism. But the linkage between these two clusters is weak: there is a mean correlation of only .16 between the two sets of issues. Accordingly, neither Figure 2.4 nor Figure 2.5 shows anything resembling a compact grouping of the parties along a Left-Right regression line (which would, theoretically, run from the upper-left corner to the lower-right corner of each figure). Quite the contrary, Figure 2.4 shows a scattering of party positions that is almost evenly distributed over the four quadrants. There is some tendency for the parties to be concentrated in the upper-left and lowerright quadrants, but to describe this pattern as a unidimensional Left-Right polarization would be a grotesque oversimplification. Turning to the elite level, Figure 2.5 comes somewhat closer to unidimensionality, for the lower left-hand quadrant is almost empty. Among candidates to the European Parliament, it is unusual to favor more nationalization of industry and the development of nuclear power. Nevertheless, two of Europe's largest parties and one smaller one do fall into that category, with the French Communists constituting an extreme deviation. Thus the overall pattern is far from unidimensional. A word of explanation is in order concerning the Belgian parties. Antagonism between the Flemish-speaking and French-speaking (or Walloon) segments of the population has led to the emergence of Flemish and Walloon nationalist parties, generally of modest size but with deep historical roots. Since the late 1960s, ethnic cleavages have become more pronounced, dividing all significant parties (except the Communists) into Flemish and Walloon sections that are only loosely allied. The parties are labeled accordingly. The Flemish and Walloon sections of the respective parties take quite similar positions on the two dimensions dealt with here. They are divided by ethnic factors that constitute a distinct and independent dimension of political cleavage. The only avowedly extreme-Right party with substantial numbers of voters—the Italian neo-Fascist party (MSI)—does not occupy the extreme lower right-hand corner of either figure. Clearly, it is located on the Right; but most of the European liberal parties are more conservative than the MSI, on at least one of the two dimensions. Nevertheless, it does hold an extreme position in another sense. Its electorate shows the highest level of political dissatisfaction among any of these parties, whether Left or Right, and one of the highest levels of support for revolutionary change. Figure 2.4 demonstrates some interesting cross-national contrasts. As one

Changing Structure of Political

Cleavages

49

FIGURE 2.4 Support for Nationalization of Industry and for Developing Nuclear Power among Electorates of Western European Political Parties, 1979

SOURCE: E u r o - B a r o m e t e r 11 (April 1979).

NOTE: Parties indicated by same symbol are members of same party federation in European Parliament.

would expect, the range of policy alternatives tends to be far greater in those countries having undiluted proportional representation, and a large number of parties. In particular, the Danish, Dutch, and Italian parties show a relatively extreme dispersion in space, offering radically different ideological positions. By contrast, the leading parties of both Britain and Germany are concentrated in a compact area, as would be expected of relatively large,

50

Frameworks for

Analysis

FIGURE 2.5 Support for Nationalization of Industry and for Developing Nuclear Power among Candidates to European Parliament, 1979

SOURCE: Europarliament Candidate Study, 1979. NOTE: Parties indicated by same symbol are members of same party federation in European Parliament.

catch-all parties. But there is a significant contrast between these two countries. The electorates of the German parties take an almost identical stand on the economic issue, and differ mainly on the noneconomic issue; while with the British parties, it is exactly the other way around. The German public polarizes over the New Politics; the British public still polarizes along class lines—and accordingly, it shows much higher class-voting indices than the Germans.

Changing Structure of Political Cleavages

51

Figure 2.5 shows far greater dispersion than Figure 2.4. In other words, the politicians of given parties take more extreme ideological positions than their electorates—confirming similar findings by previous investigators (Converse, 1975a). Figures 2.4 and 2.5 are presented in a way that actually understates the degree to which this is true. For the locations of the various electorates cluster relatively near the center of the issue space; in attitudes toward nuclear power, no electorate shows an aggregate score higher than approximately 3.75 or lower than 1.50, though the theoretically possible scores range from 1.00 to 4.00. Consequently, in order to depict the parties' relative positions, we have excluded the empty outer areas and enlarged the central area of Figure 2.4. Figure 2.5, on the other hand, depicts the entire possible issue space. The outer borders represent scores of 1.00 or 4.00, where all respondents are, respectively, "strongly against" or "strongly for" the given measure. The candidates of some parties actually are unanimously "strongly for" or "strongly against" given issues; some parties (the German "Greens" or ecologists, or the Danish Democratic Center, for example) unanimously take the most extreme possible position on both issues. The candidates of four different parties (three of them liberal parties) fall at the same point on the right-hand border of Figure 2.5. The attainment of these extreme scores is facilitated by the fact that some of these parties are represented by only two or three candidates. Still, some of the large parties, represented in our sample by many candidates, fall very near the extremes on one dimension or another: for example, the French Communists and Gaullists; the British Conservatives; and the Italian Christian Democrats. Thus while the electorates of the principal British parties are concentrated in a very small portion of the total issue space, the candidates of these parties are separated by relatively large distances. The split that later took place in Britain's Labour party, whereby the party elites lost the support of much of their electorate, seems to be foreshadowed by these data—for the elites take a position far to the Left of their electorate. While Conservative party elites are located to the Right of their electorate, the distance is considerably smaller. On the whole, there is a close relationship between the issue positions of the electorate and elites of a given party. Table 2.5 gives the precise correlations. Once again, we find that the most sensitive indicators of the two dimensions play key roles. The four strongest mass-elite correlations are based on the two highest-loading issues on the economic Left-Right dimension, and the two highest-loading issues on the noneconomic issues dimension shown in Table 2.1. These issue correlations are handsome indeed. They are all the more impressive because they are not intracranial correlations that might conceivably be attributed to some methodological artifact such as response set. The correlations are based on two completely independent data sets, measured at two different levels of the political system. Although they are aggregated to the

52

Frameworks for Analysis TABLE 2.5 Correlation between Issue Positions of Candidates of a Given Party and Positions Taken by Electorate of That Party, 1979 Issue Positions Develop nuclear power Stronger measures against terrorists More government management of economy More public ownership of industry Regional aid Stronger defense Codetermination Protection of individual rights Abortion should be available Reduce income inequality Antipollution measures Aid Third World Control multinationals

Correlation .645 .607 .547 .543 .541 .452 .447 .435 .400 .384 .384 .263 .251

SOURCE: Data set constructed using party as unit of analysis, based on aggregated responses of the given party's electorate (measured by reported voting intention) and aggregated responses of the given party's candidates to the European Parliament. Public was surveyed in April 1979 (EuroBarometer 11) and candidates in March-May 1979.

group level (which tends to reduce the measurement error perennially present in survey data), the correlations found with our best indicators of the two respective dimensions are impressive. The .645 correlation between mass and elite attitudes toward nuclear power, for example, could be interpreted as meaning that almost 42 percent of the variance in the party elite's stand on this issue can be attributed to constituency influences. In fact, we do not believe that the causal linkage is that simple. Part of this agreement may represent the elites' influence on their electorate, for example; or the electorate may support given parties because of the stand they take on key issues, without influencing their issue positions. Furthermore, the linkages may be based on cues concerning the two broader issue dimensions, rather than the specific indicators. We will not attempt to determine the specific causal connections that are involved here. For present purposes, our point is simply that strong linkages do exist between the positions that the electorates and candidates of given parties take on these issues. It is virtually inconceivable that this pattern could be due to chance alone. Either the politicians of given parties are influencing

Changing Structure of Political Cleavages

53

their electorates, or the electorates are influencing the candidates' positions on these issues; or the electorates are selectively recruited to support given parties, partly as a function of their stand on these issues; or candidates with appropriate views are more apt to be recruited. Each of these factors probably plays a role—and all on elite-level politics. The net result is a surprisingly close fit between the positions of the electorates and politicians of given parties on these issues. The two most sensitive indicators of the New Politics dimension, and the two most sensitive indicators of the economically based Left-Right dimension, are the issues that show the strongest elite-mass linkages. But the New Politics issues show even stronger linkages than do the economic issues. Although West European political elites still tend to describe Left-Right political polarization primarily in terms of the classic issues of state ownership and control of the means of production, the evidence suggests that the electorates select their party more on the basis of the new issues than the old. Issue preferences explain only part of the variance in political party choice, of course. An even larger proportion is probably due to long-term party affiliations and loyalties. But insofar as issues do influence party preferences, the new issues seem to have at least as much impact as the old. COMPARING CLEAVAGE DIMENSIONS

We hypothesized that the new noneconomic issues dimension reflects a materialist versus postmaterialist cleavage, rather than the social class and religious cleavages that gave rise to the conventional Left-Right dimension. Let us now test that hypothesis. Figure 2.6 shows the strength of the relationship between the issue positions taken by the candidates of given parties and the aggregate characteristics of the electorates of these parties. Table 2.6 gives the detailed statistics on which Figure 2.6 is based. Again, these are not intracranial correlations. They reflect elite-mass linkages, based on independent measurements at each level; and the characteristics examined here tend to be relatively enduring features of given electorates. Thus there are strong grounds for inferring that these mass characteristics have a causal impact on elite attitudes (or lead to selective recruitment of the candidates). In explaining the elite-mass issue correlations shown in Table 2.5, for example, one might plausibly argue that they simply reflect the tendency for the electorates of given parties to follow elite cues when they adopt their issue positions. But it is not plausible to argue that the electorates of given parties become predominantly working class, or adopt given religious or value orientations, because of elite cues. It seems far likelier that the elite issue positions are (in one or another of the ways outlined previously) influenced by their constituency's characteristics. The top half of Table 2.6 deals with three key economic issues. It dem-

Frameworks for Analysis

54

FIGURE 2.6 Correlations between Issue Positions Taken by Candidates of Given Parties, and Social Background Characteristics of Their Electorates CORRELATES OF ISSUE POSITIONS TAKEN BY CANDIDATES OF SIXTY-SIX PARTIES FROM NINE WEST EUROPEAN COUNTRIES I Economic Issues

223

SOCIAL CLASS Proportion of Manual Workers m Given Party s Electorate

Mean Church Attendance Rate Among Electorate

VALUES Ratio of Materialists to Post-Materialists Among Electorate

620

472

ν///λ

003

^¾¾

SOCIAL CLASS

RELIGION

VALUES

Source Issue Battery Included in Survey ot 7 4 2 Candidates to European Porhment Interviewed in Spring,1979 (See Inglehort et ol ,1980). and Data on Electorates From Cumulative results of Euro-Barometer Surveys 3 H 2

SOURCES: Europarhament Candidate Study, 1979, and data on electorates from cumulative results of Euro-Barometer Surveys 3-12

onstrates that among the three types of characteristics examined here, religiosity has the greatest impact. Not only has the religious factor been assimilated to the conventional Left-Right dimension—it actually outweighs social class as an influence on attitudes toward economic issues. When Rose and Urwin (1969) and Lijphart (1971) found that religion outweighs social class in its electoral impact in most Western nations, the finding seemed surprising in the light of prevailing social theory. The fact that religion outweighs social class in its impact even on attitudes toward specifically economic issues may seem more surprising still, but our data point to that conclusion. Materialist/postmaterialist values also seem to have a significant impact on economic issue attitudes, though their impact is weaker than that of religiosity. At the individual level, materialist/postmaterialist values have an ambiv-

55

Changing Structure of Political Cleavages TABLE 2.6 Correlation between Policies Supported by Candidates of a Given Party and Characteristics of That Party's Electorate I. Economic Issues

Characteristics of Electorate Percentage of manual workers among electorate Church attendance rate among electorate Ratio of materialists to postmaterialists among electorate

More Government Management of the Economy

More Public Ownership of Industry

Reduce Income Inequality

Mean Correlation, Three Issues

.325

.224

.120

.223

.460

.621

.480

.520

-.180

-.510

-.306

-.332

2. Ν uneconomic Issues

Percentage of manual workers among electorate Church attendance rate among electorate Ratio of materialists to postmaterialists among electorate

Develop Nuclear Power

Stronger Defense Effort

Stronger Measures against Terrorists

Mean Correlation, Three Issues

-.017

-.077

+ .086

-.003

-.554

-.414

-.448

-.472

.677

.512

.670

.620

SOURCE: Data set Constructed using party as unit of analysis. Public data are pooled results from Euro-Barometers 3-12; elite data are from 1979 Survey of Candidates for the European Parliament. NOTE: Table entries are Pearson product-moment correlations. alent relationship to the issues linked with the conventional economic LeftRight dimension. Postmaterialist respondents are only slightly more favorable to redistribution of income than are materialists—though the former are mark­ edly more favorable to increasing economic aid to Third World countries. This may be so because the two groups favor income redistribution for dif­ ferent reasons. Despite their relatively high-income levels, postmaterialists may favor redistribution for the sake of human solidarity. On the other hand,

56

Frameworks for Analysis

lower-income materialists may favor income redistribution because (to some extent) they are the ones who benefit from it. The situation is less ambiguous with regard to aid to Third World countries. Here neither group stands to gain material benefits—and the postmaterialists are much more favorable than the materialists. In keeping with their general tendency to support the positions of the Left, postmaterialists are somewhat likelier to favor nationalization of industry than are materialists. But postmaterialists are slightly less likely to favor a greater government role in managing the economy. For the Old Left, government ownership, regulation, and control of the economy was inherently good. Almost by definition, a larger government role was desirable in almost any situation. For the postmaterialist Left, big government is inherently dangerous. Like any large, hierarchical, bureaucratic organization, it tends to encroach on individual autonomy and expression. This fact poses a dilemma for the postmaterialist Left. Its adherents tend to favor social change; and almost any program of social change presupposes that the government, necessarily, will be the instrument to bring it about. But the postmaterialist Left—far more than the traditional Left—regards the state as a potential instrument of oppression and exploitation. Though they favor equality, they are reluctant to use the state to bring it about. One possible way out of this dilemma might be through decentralizing the state. Hence postmaterialists tend to favor regional autonomy. In the November 1981 EuroBarometer surveys, 36 percent of the materialists were "strongly for" greater regional autonomy, as compared with 51 percent of the postmaterialists in the ten-nation European community. Big government may be necessary to social change, but the postmaterialist Left is ambivalent toward it. On the other hand, the relationship between materialist/postmaterialist values and the noneconomic issues loading on the second dimension is clear and unequivocal. Materialists are more than twice as likely to favor a stronger defense effort than are postmaterialists, and they are almost twice as likely to favor developing nuclear power, or taking stronger measures against terrorism. The differences between materialists and postmaterialists on these issues are large and consistent, both from issue to issue and from nation to nation. When we turn to the lower half of both Figure 2.6 and Table 2.6, we find indications that these preferences have a political impact. Consistently and by a clear margin, the ratio of materialists to postmaterialists among the electorate is the strongest predictor of candidate attitudes toward the noneconomic issues. The value preferences of the electorate easily outweigh the impact of both the religious and social class indicators combined, partly because the latter effect is negligible. The linkage between the social class composition of a party's electorate and their candidate's stand on noneconomic

Changing Structure of Political Cleavages

57

issues is about as close to zero as one can get. Our religious indicator, on the other hand, does have a significant impact. The persistence of an apparently flourishing linkage between religious affiliations and political cleavages, coupled with a truly remarkable decline of social class cleavages, may seem paradoxical. But it is very much in keeping with the theoretical framework underlying this analysis. For reasons that we have discussed earlier in some detail (Inglehart, 1977: 217-222), the rise of postmaterialist politics has an inherent tendency to neutralize class-based political cleavages, but it does not have that impact on religious cleavages. On the contrary, it may even give them new life. Postmaterialists tend to be recruited from the more affluent strata that traditionally supported the Right, but they themselves tend to support the Left (and may engender a workingclass reaction that moves toward the Right). In other words, postmaterialism tends to reverse the polarity of the correlation between social class and the Left-Right dimension. But it does not have this impact on religion. Postmaterialists tend to be recruited from the nonpracticing segment of the religious spectrum, which traditionally has supported the Left, and continues to do so. Furthermore, postmaterialist support for cultural change may stimulate a conservative reaction on the part of those holding traditional religious values— reinforcing, rather than weakening, their alignment with the Right. The rise of a new kind of value-based politics may give relevance to much older valuebased cleavages, rooted in the preindustrial era (Pappi, 1977).

GROUP POLARIZATION AND PARTY ALIGNMENTS

We hypothesize that a process of value change has not only led to the emergence of a new issue dimension, but of a new axis of group polarization that cuts across the traditional working class-middle class or labor-management confrontation. As a result, support for political change increasingly comes from a new social base—though this phenomenon's impact on electoral behavior is retarded by the inertia of established party loyalties and group affiliations. In order to test this hypothesis, we need to distinguish between the traditional socioeconomic Left-Right group polarization on one hand, and support or opposition to the present social order and establishment groups, such as the bureaucracy and police. It is not an easy task, for the two tend to be lumped together: the term "Left" connotes support for sociopolitical change and support for various groups and parties. The polarization between workers and owners-managers traditionally constituted the social basis of the struggle between Left and Right. But it is not the only conceivable one. In societies in which the dominance of business interests has been reduced, or private ownership abolished altogether, political

58

Frameworks for Analysis

conflict continues—typically, between ordinary citizens and a ruling elite based on control of bureaucracy and police. The latter type of polarization seems likely to become more salient in advanced industrial society. Among the ideologically sensitive, support or opposition to the established order may be the core component of one's orientation toward Left and Right. Among those who are ideologically less aware, support for the "Left" or "Right" may be largely a matter of group and partisan loyalties, with little reference to support or opposition to current political change. The labels ' 'Left'' and "Right'' have been a staple part of political discourse for many decades, particularly in Europe. To some extent, support for a party of the "Left" undoubtedly does reflect an ideology of opposition to the existing sociopolitical order, which is what interests us at present. But insofar as the Left-Right labels are contaminated by a party identification component, they would not necessarily reflect one's attitude toward social change. Thus we must try to distinguish between two components of the excessively broad and inclusive Left-Right dimension, one based on long-term partisan loyalties and tiie other on one's attitude toward social change. We hypothesize that partisanship still tends to reflect the classic labor-management polarization, while support for social change is now mainly linked with a newer axis of group polarization that, in turn, reflects the conflict between materialist and postmaterialist priorities. For evidence on this score, let us turn to data from an eight-nation study of political action, fieldwork for which was carried out between 1974 and 1976 (for details, see Barnes, Kaase et al., 1979). Each person interviewed in the eight national samples was asked to indicate how friendly or unfriendly his/her feelings were toward various groups in his/her society, using a "feeling thermometer" that ranged from a score of zero (indicating extremely cold feelings) to 100 (extremely friendly feelings). Our objective was to develop indices of each respondent's position for or against the groups shown in Table 2.7. In doing so, an important concern was to remove a general "responseset" tendency from the thermometer ratings—that is, a pronounced tendency for some individuals to give all groups consistently high or low ratings. To remove this "response-set" tendency, the average rating across all groups was computed for each respondent. This mean score was then subtracted from the thermometer rating of each group. Thus for each respondent, some groups are evaluated negatively (below the respondent's average) while other groups are valued positively. These adjusted thermometer ratings were then used in constructing factor score indices. Factor analysis of these ratings revealed a consistent pattern, with two basic dimensions underlying responses in each of the eight nations studied. As Table 2.7 indicates, we find a LeftRight partisanship factor in each country, based on ratings of the most important conservative party in that nation, as well as the rating of the most

Changing Structure of Political

59

Cleavages

important party on the Left side of the political spectrum.2 Ratings of key institutions linked with the respective parties also have significant loadings on this factor, which simultaneously taps one's partisan loyalties and one's sympathies in the long-standing opposition between labor and management. In those countries with major Christian Democratic parties, one's rating of the clergy also loads on this factor, forming part of the cluster containing management and the most important political party of the Right. This holds true for The Netherlands, even though the "Right" party actually rated by the Dutch respondents was not church-related (they rated the Liberals instead). It does not hold true in Switzerland, though that country does have important church-linked parties. But this is not the only anomaly in the way group sympathies polarize in Switzerland: the dimension based on opposition between Left and Right, labor and management, which we find almost everywhere else, is strangely deformed in Switzerland. For one thing, the major Left party does not load on this factor. Although this is indeed anomalous, it seems to reflect the reality of Swiss politics quite accurately. For more than a generation, Switzerland has been governed by a cartel of the four major parties, with the Socialists part of a virtually permanent coalition that also includes the three major parties of the Right. If mass sympathies do not polarize between the Socialists and Radicals, it may be because the parties themselves do not polarize in their behavior. In Switzerland, Left-Right polarization is reduced to a truncated opposition between the Radical party and the labor unions (with a minor contribution from the women's movement). The cleavages between labor and management, and between religious and secular publics, have been part of the political scene for many decades; they seem to have been incorporated into the traditional Left-Right (or liberalconservative) dimension underlying Western politics, so that one's loyalties tend to lie with one set of groups or another in a predictable, and by now institutionalized, pattern. But new issues and new groups have become politically salient in recent 2

The specific parties rated, in each respective country, were as follows. Main Left Party Main Right Party Britain Labour party Conservative Germany Social Democratic party Christian Democratic party The Netherlands Labour party Liberal party Austria Socialist party People's party (former) United States Italy

Democratic party Communist party

Switzerland Finland

Socialist party Social Democratic party

Republican party Christian Democratic party Radical party National Coalition party

60

Frameworks for Analysis

TABLE 2.7 Factor Analysis of Left-Right Group Sympathies and Establishment/Antiestablishment Attitudes in Eight Nations United States

Ratings of: Left party Labor Unions Big business Right party Clergy Police Civil servants Small business Minority group 3 Women's movement Student protestors Revolutionary groups

Left/ Right .70 .58 -.37 -.64

.30

The Netherlands

Britain

Est./ Antiest.

.39

Left/ Right

Est./ Antiest.

.80 .75 -.54 -.76

.58 .72 .46 .37

.51 .61 .37 .36

-.57 -.76 -.71

-.44 -.70 -.67

Left/ Right .75 .66 -.59 -.78 -.52

Est./ Antiest. .32 .31 .39 .69 .54

Gem tany Left/ Right .75 .64 -.63 -.81 -.67 .76 .57 .48

-.34 -.31 -.68 -.45 -.69

Est./ Antiest.

.41 -.71 -.71

SOURCES: Eight-nation Political Action study, carried out between 1974 and 1976. For details offieldwork,see Zentralarchiv fiir empirische Sozialforschung, Political Action: An Eight Nation Study (Cologne: Zentralarchiv, 1979). Findings from five nations are reported in Barnes, Kaase et al. (1979). NOTE: The scores on the feeling thermometer were adjusted for response set by subtracting

years. One's reaction to these groups forms the basis of a second major dimension of sociopolitical cleavage. One tends to sympathize with the women's movement, student protestors, foreign workers, and revolutionary groups on one hand; or with the police, civil servants, and other elements of the established social order (including the clergy, in those countries where the church does not have formal ties with the dominant parties of the Right). One's position on this dimension is largely unrelated to one's partisan loyalties. The two dimensions seem to reflect: (1) the traditional socioeconomic Left-Right cleavage, with an infrastructure based on the polarization between labor and management (with religious cleavages also assimilated to this dimension, in some countries), and (2) an establishment-antiestablishment (or New Politics) dimension, based on one's reaction to groups that have become politically prominent much more recently than organized labor—and that, we suspect, today are more active carriers of support for social change. While the pattern is remarkably consistent cross-nationally, Italy's configuration is unique. On one hand, the polarization between the Communist and Christian Democratic parties appears not only on the Left-Right partisan dimension, but also on the establishment-antiestablishment dimension. In

Changing Structure of Political Cleavages

Austria

Ratings of: Left party Labor Unions Big business Right party Clergy Police Civil servants Small business Minority group" Women's movement Student protestors Revolutionary groups

Left/ Right

61

Italy

Est./ Antiest.

.81 .71 -.39 -.82 -.64

Left/ Right .71 .61 -.67 -.61

.65 .57 -.37 -.71 -.67

Switzerland Est./ Antiest.

Left/ Right

Est./ Antiest.

-.31 .79 .61 .31

.42 -.54 .49 .64 .52

.68 -.40 .56 -.46 -.40 -.60 -.69

.34

Finland Left/ Right

Est./ Antiest.

.49 .64 -.62 -.73

-.33

.53 .64 .34 .44 -.30

.46 -.62 -.61

-.63 -.69

the individual's mean rating for all groups from his rating of any given group. All loadings above .30 are shown. Factor analysis with varimax rotation was used. ""Foreign workers" in The Netherlands, Germany, Austria, and Switzerland; "workers from the South" in Italy; "blacks" in the United States; "colored immigrants" in Britain; "gypsies" in Finland. contrast with Switzerland (where neither dimension shows the full Left-Right party polarization), in Italy both dimensions reflect polarization according to party preference. While in Switzerland, the Left has permanently shared power, in Italy, the main party of the Left (the Italian Communist party) had been permanently excluded from power, with the Christian Democrats continuously in office from the end of World War II up to the time of our survey. Hence, in Italy, partisan conflict is not distinct from support for social change— on the contrary, partisan change may seem a prerequisite for any basic social change. Italy is unique, furthermore, in that attitudes toward the civil service are not linked with the establishment-antiestablishment factor (as everywhere else) but with the Left-Right partisan factor. Here the electorate reacts to the civil service as if it were linked with the Christian Democratic party—again, a rather accurate reflection of political reality (in this case, the colonization of the civil service). The establishment-antiestablishment factor shows a distinctive feature in one other country as well: The Netherlands. In addition to a perfectly normal polarization on the Left-Right partisanship dimension, both labor unions and big business show significant loadings on the establishment-antiestablishment

62

Frameworks for Analysis

dimension—and here they share the same polarity: both are seen as establishment groups. This may reflect The Netherlands' highly developed system of corporatist involvement of both labor and management organizations in the shaping of economic policy. In other respects, the polarization of group sympathies in The Netherlands conforms to the pattern that prevails among these eight nations. In six of the eight nations, political polarization corresponds closely to our expectations. We find a Left-Right partisanship dimension, intimately linked with one's attitudes toward labor and management (and the church, in those countries where the church has traditionally played a major role). Alongside this dimension, a second one taps support for radical and nonestablishment groups—or for the established authorities. In two countries, the structure of group sympathies deviates from the general pattern—in Switzerland, through an absence of partisan polarization; and in Italy, because party polarization is everywhere. The two group-sympathies dimensions are analyzed in greater detail in Inglehart (forthcoming). Let us briefly summarize two findings that are particularly relevant here: First, as one would expect, voting behavior is much more closely related to one's position on the first dimension, which taps labor-management sympathies and Left-Right partisanship, than with the second dimension, based on establishment group sympathies. But political activism is more closely linked with the second dimension. A protest potential scale, based on one's participation in and approval of an escalating series of protest actions, was included in the Political Action surveys (for a detailed description, see Barnes, Kaase et al., 1979: 65-81). While this protest potential scale showed significant correlations with both dimensions, its linkage with the establishmentantiestablishment dimension was clearly the stronger relationship. Contemporary political protest seems much less likely to spring from a prolabor stance than from alignment with other nonestablishment groups. Second, the social background variables linked with the two group-polarization dimensions show an almost perfect parallel with the pattern in Table 2.6. Like the economic issues dimension, polarization on the conventional Left-Right group sympathies dimension reflects one's religious background above all, with one's social class and value priorities making lesser but significant explanatory contributions. But—like the noneconomic issues dimension—one's position on the establishment-antiestablishment dimension reflects whether one has materialist or postmaterialist values. One's value priorities explain considerably more variance than religion and social class combined. As in the lower half of Table 2.6, the role of social class drops to a negligible level here. Insofar as there is any linkage, the lower economic strata tend to side with the establishment groups.

Changing Structure of Political Cleavages

63

THE IMPACT OF SOCIAL CLEAVAGES ON THREE TYPES OF POLITICAL POLARIZATION

We hypothesized that the impact of postmaterialism would be greatest on those forms of political polarization that are least strongly linked with established political party loyalties. Political party identification is reshaped only gradually by the rise of new values; hence it helps to preserve old patterns among those variables highly correlated with it. Where voting behavior is closely related to political party loyalties, it will polarize according to individual values only slightly more than party identification itself. Left-Right self-placement tends to reflect political party ties in part, but it also reflects one's reaction to current issues. Hence it is more likely to polarize according to one's values. Finally, support for social change is not necessarily tied to political party loyalties at all. Hence polarization over this basic superissue will reflect individual value priorities above all; being only minimally constrained by party ties, it will show a minimum level of social class polarization, under current conditions. Thus we would expect the impact of postmaterialism to be weakest on voting behavior—a relatively direct expression of party loyalties—and relatively strong on support for social change, a superissue that has no explicit linkage with party ties. Conversely, if it is true that social class conflict was a more important factor a generation ago than it is today, it would be preserved most fully in those forms of polarization most directly linked with the patternpreserving influence of long-term party loyalties. The impact of social class on voting should be far stronger than its impact on support for social change. As a hybrid variable reflecting both a partisanship component and an ideological component that sums up one's position on current issues, Left-Right self-placement should occupy an intermediate position between voting and support for social change. As the first step in testing these hypotheses, let us examine the absolute impact of social class, religion, and individual values on each of the three types of political polarization. Table 2.8 shows how each of the three social cleavages relates to voting patterns, in a pooled sample based on the Euro-Barometer surveys carried out in the middle and late 1970s, weighted according to population. This sample, of course, fails to convey the wealth of variation in the social background characteristics of the scores of parties in nine different countries. It groups parties together into two broad categories—the "Left" (the various Communist, socialist, and Social Democratic parties); and the "Right" (the various Christian Democratic, conservative, and some other parties). A minority (about 9 percent of those reporting a party preference) are not classified. The centrist and ethnic nationalist parties are excluded, for example, and the

64

Frameworks for Analysis

liberal parties are split between some that seem clearly part of the Right, and others that are considered centrist, and excluded from these tables Detailed information on how each party is coded appears in the ICPSR code books for the Euro-Barometer surveys Though it fails to convey a vast number of interesting details, this massive pooled sample does provide an exceptionally reliable data base for analysis of overall patterns of political polarization in Western Europe TABLE 2 8 Electoral Cleavages Based on Social Class, Religious Practice, Personal Values, 1973-1979 Percentage Voting for Left Social class voting Respondents from families headed by manual worker Respondents from families with head having nonmanual occupation Alford index of class voting

63%

Ν

(21,616)

45 + 18

(24,594)

Religious voting Respondent attends church at least once a week Respondent attends church at least a few times a year Respondent never attends church

27

(13,360)

49 65

(20,290) (10,174)

Value-based voting Respondent has materialist priorities Respondent has mixed priorities Respondent has postmatenahst priorities

43 53 71

(16,777) (23,180) (4,678)

SOURCES Pooled results from each of the nine-nation European community surveys carried out from 1973 through 1979 (Euro-Barometers 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12), weighted according to population of given nation NOTE Table entries are the percentage voting Left, 1973 1979 As Table 2 8 reveals, class voting was still fairly strong in Western Europe in the late 1970s the Alford index of class voting was + 18 for the European community as a whole While this is well below the levels reported for most West European countries in the 1950s, it is still an important feature of political cleavages in Western Europe It is clear from Table 2 8 that religious cleavages are also very strong, although there is some ambiguity about just where to draw the cutting line with church attendance Finally, Table 2 8 indicates that

Changing Structure of Political Cleavages

65

the voting behavior of the various value-types is quite distinctive. Indeed, if we were to compare simply the voting intentions of the pure materialist type (based on the original four-item values index), we would obtain a larger percentage difference than that found with social class. But our values in­ dicator is not dichotomous. Over half of those reporting a preference for parties grouped with the Left or the Right fall into the mixed values type; while the postmaterialist type shows quite distinctive voting behavior, the sheer percentage differences would convey an exaggerated impression of the impact of values on voting behavior. A multiple classification analysis (see Table 2.11) alleviates this problem, by calculating coefficients that are weighted according to the number of cases in each category of the independent variables. TABLE 2 9 Left-Right Self-placement According to Social Class, Religious Practice, and Personal Values, 1973-1979

By social class Manual head of family Nonmanual head of family By religious practice Respondent attends church at least once a week Respondent attends church a few times a year Respondent never attends church By value priorities Materialist Mixed Postmaterialist

Percentage Placing Selves on Left

Ν

61% 53

(23,498) (29,459)

42

(14,613)

52 63

(22,760) (11,529)

45 55 75

(20,891) (29,788) (6,296)

SOURCES: Pooled results of the 1973-1979 European community surveys. NOTE: Table entries are percentages placing themselves on Left half (codes 1-5) on Left-Right scale. Table 2.9 shows the percentage differences associated with self-placement on the Left-Right scale. While the differences linked with church attendance and value-type remain quite large, those associated with social class shrink to only 8 percentage points. Finally, Table 2.10 shows the relationship of social class, church attend­ ance, and value-type, with support for social change. With all three inde-

66

Frameworks for Analysis

TABLE 2.10 Support for Social Change by Social Class, Religious Practice, and Personal Values, 1976-1979. Respondent favors: Gradual Reform

Defense of Present Society

8% 7

62% 64

30% 29

100% 100

4

60

36

100

(9,865)

5 9

60 61

35 30

100 100

(16,473) (7,981)

4 8 17

57 62 69

38 30 14

99 100 100

(18,292) (26,694) (6,098)

Revolutionary Change By social class Manual Nonmanual By religious practice Attends church at least weekly Attends church a few times a year Never attends church By value priorities Materialist Mixed Postmaterialist

Ν (17,579) (24,025)

SOURCES: Pooled results from Euro-Barometers 6, 7, 8, 9, 10, 11, and 12, weighted according to population of given nation. pendent variables, a Reformist majority prevails. But even a cursory inspection of this table suggests that one's values are the dominant factor. There is literally no difference in the degree to which manual and nonmanual respond­ ents support social change. The former group is 1 percentage point more likely to favor revolutionary change, but it is also 1 point more likely to favor defense of the present society. Religion, on the other hand, does seem to have an appreciable impact on this orientation. Those who never attend church are about twice as likely to support revolutionary change as those who attend church weekly, and somewhat less likely to favor defense of the established social order. But the differences associated with value-type are substantially greater. Postmaterialists are four times as likely to favor revolutionary change as are materialists, and a great deal less likely to favor defending the present society. Now let us compare the impact of the three social background variables on each of the three types of political polarization. Table 2.11 gives the results of a multiple classification analysis based on these data. The eta coefficients show the relative strength of each predictor variable, weighted for the number

Changing Structure of Political Cleavages TABLE 2 11 Political Cleavages Based on Social Class, Religious Practice, and Personal Values, 1976-1979 Eta

Beta

Left-Right voting Church attendance Social class Value priorities

.264 .179 .141

.242 .162 .126

Left-Right self-placement Church attendance Value priorities Social class

.201 .188 .095

.195 .179 .084

Support for social change Value priorities Church attendance Social class

.185 .111 .032

.172 .094 .020

SOURCES: Pooled data from European community surveys carried out between 1973 and 1979 (1976 to 1979 with support for social change). NOTE: Table entries are results from multiple classification analysis.

of cases in each category, while the beta coefficients provide a similar statistic, controlling for the effects of each of the other two predictor variables. As we hypothesized, the net impact of social class on voting behavior outweighs that of value-type—though our religious indicator shows a stronger effect than either of them. But when we turn to support for social change, values are the strongest predictor by a wide margin; the impact of social class is negligible, while religious practice has a significant effect (though only about half as strong as that of values). Left-Right self-placement shows a pattern consistent with our characterization of it as a mixture of partisanship and ideological variables: religion has the strongest impact, followed by values and social class. The pattern obtained here is by now familiar. As was true in connection with issue polarization and the two group-sympathies dimensions, the evidence supports the interpretation that a new axis of political polarization has arisen in relatively recent years, which reflects an opposition between materialist and postmaterialist goals. Political party loyalties remain tied to the social class-based axis; but they seem to be out of phase with a New Politics cleavage. ,

68

Frameworks for Analysis

CONCLUSION

Building the welfare state and restoring economic growth were, understandably, the dominant political concerns of the 1930s and the postwar era. In the late 1960s and 1970s, postmatenalist forces captured the issue agenda of the West European Left and of American liberals. The effects of this coup have not yet been assimilated into either the rhetoric or the party alignments of Western nations. Protest has become divorced from partisanship, to a remarkable degree. Two main axes of political polarization exist side by side, with the leading parties aligned along the familiar Left-Right axis based on religion and social class, in uneasy coexistence with a largely independent polarization between materialists and postmaterialists. The latter continue to dominate the issue arena. Since the first OPEC oil shock in 1973, Western economies have functioned in an atmosphere of insecurity and diminishing expectations. Although everyone assumes that the issues of the 1980s will be economic issues, the ones that evoke political activism today are still largely postmaterialist ones. The rise of postmaterialism has placed existing party alignments under chronic stress. For in most countries, these alignments do not correspond to either the social bases of support for change or to polarization over the most heated issues. Kaase and Klingemann (1982) reach similar conclusions, using a quite different analytic approach: The tensions between party systems and value orientations in western democracies, which are alleged to exist by theorists of very different schools, could be unambiguously traced back to two different developmental tendencies in these societies. On one hand, on the left side of the ideological spectrum, there is an increasingly evident cleavage between the "old" and the "new" Left: the "old" Left is more materialist in the formulation of its political goals (economic growth, economic security); it is also more conservative in the selection of the means to attain these goals. In contrast, the "new" Left is postmaterialist and progressive in its forms of participation. This poses potentially overwhelming integrative tasks for parties of the Left. (385) The resulting stress can be resolved in various ways: (1) Dealignment: there may be a gradual decline of party loyalty and party identification, insofar as the most salient issues no longer provide an incentive that attaches new voters to existing parties; (2) Realignment: existing parties may split, or be taken over by reorienting elites. There is yet another possibility. The new axis of polarization based on materialist/postmaterialist values may decline in importance; or—more likely— it may be assimilated into a new synthesis. »

Changing Structure of Political Cleavages

69

A materialist consensus provided the rationale and the legitimating myth of industrial society. Its postmaterialist antithesis has not yet led to the emergence of a new synthesis. But it seems likely that the wave of the future is not undiluted acceptance of postmaterialist goals. The postmaterialists brought into the political arena a number of issues that had been largely ignored and neglected; in doing so, they help correct a course that tended to sacrifice the quality of life to one-sided economic considerations. But carried to an extreme, postmaterialism can be equally self-defeating. The antiindustrial outlook of some of the movement's ideologues could lead to a neglect of the economic base on which postmaterialism ultimately depends. In the long run, a new synthesis of materialist and postmaterialist orientations will emerge because it must.

3. Political Remobilization in Welfare States RISTO SANKIAHO Research Institute for Social Sciences University of Tampere, Tampere, Finland

The term "welfare state" has been used to refer to the Nordic and European states in which a Social Democratic party has participated in the government or where Social Democrats have been effective in politics. No distinct theoretical framework about the welfare state has been developed so far, even though it is typified by a group of general features (Bilski, 1976). German social legislation often is mentioned as the initial definition of a welfare state. Other explanations push its origins back to the Middle Ages and the beginning of the sixteenth century (Kuhnle, 1978). The concept of the welfare state developed from the revisionist branch of the Socialist movement. Karl Kautsky (1907) and Eduard Bernstein strongly implied that the essential influence and primary goal of contemporary society was to widen democracy by means of the general franchise in order to make it possible for the working class to participate in governing through cabinet representation (Gustafsson, 1969). This participation contributed to the formation of the welfare state, whose ideology consists of social safety and economic stability as every person's right of citizenship. This welfare state model can provide an analytic framework for many of the nations examined in this volume. THE SOCIAL DEMOCRATIC MOVEMENT: MAKER OF THE WELFARE STATE

Social democratic movements were efficient in the beginning of this century in mobilizing people into the working class and thereby into the political process (Korpi, 1978b; Hentila, 1979). After expansive fights had taken place, the labor movement managed to win civil rights for workers as well as defend This research was done at the Institute for Social Research, University of Michigan, under a grant from the American Council of Learned Societies. Many thanks to Samuel Barnes and Russell Dalton for a stimulating working environment and many suggestions for this chapter

Political Remobihzation

71

these rights against subsequent challenges. Citizens outside the working class participated directly in politics or were in some other way integrated into the political system (Pesonen and Sankiaho, 1979). As soon as universal male suffrage was enacted, labor parties were able to compete in elections. And in many countries their representatives participated in cabinets, especially during the 1920s. This, in turn, integrated the working-class parties into the system, a fact which gradually became a problem for these parties. The original goals of class struggle were partially put aside in the interest of "governing." Walter Korpi (1978b: 97) describes this process as follows: From conservative and liberal as well as leftist viewpoints, it has been argued that social democratic parties are deserting their ideological distinctiveness and are becoming incorporated as junior partners in capitalist society. Expanding on this theme, Finn Valentin (1978: 78) writes that while workers were weak in their essential position opposite capital, they were relatively strong at the parliamentary level where the social reforms were carried out. As a result, the structure that welfare reforms assumed was a clear reflection of this position of partial strength—a strong parliamentary position in a context of overall weakness vis-a-vis society as a whole. A need for a united front in the fight against fascism during the 1930s contributed to the integration of the working-class parties into their respective political systems. Significantly, many of these so-called people's front coalitions included Communists as well as Social Democrats. The integration of labor parties into the governing establishment placed the Social Democratic movement in a difficult position. These parties are heavily dependent on working-class support, but class voting and the dependence on class cues is diminishing. The internal structure of the working class is changing and noneconomic (or nonclass) issues are becoming more important as voting criteria. The leaders of the Social Democratic movement often are considered part of the establishment; they frequently are attacked by new political leaders with populist backgrounds who receive strong support among the working class (Nilsson, 1979; Esping-Andersen, 1978; Baker et al., 1981) The Social Democratic movement has attempted to develop the social programs of the welfare state for the benefit of the working class.1 At the 1 Most policy output studies have tended to stress the importance of economic development rather than political vanables in the development of welfare programs (Dye, 1966). On the importance of political vanables, however, see Lewis-Beck (1977) Moreover, a number of researchers have found clear relationships between support for Social Democratic parties and the development of the welfare state (Tompkins, 1975, Wnght, 1975, Korpi, 1979, Bilski, 1976, Esping-Andersen, 1979) The role of trade unions has also been stressed m studies of policy outputs (Wilensky, 1975, 1976, Peters and Hennessey, 1975) Finally, Social Democrats have been successful in Sweden in developing equality in social welfare, education, and income distribution according to Thernborn et al (1978) and Esping-Andersen (1979)

72

Frameworks for Analysis

same time, legislation has been enacted ensuring equality to all citizens. However, the development of the welfare state also has involved the expansion of power by authorities. Therefore, democratic politics has remained undeveloped insofar as decision-making has been concentrated in a centralized institution far away from the voter. The Social Democratic parties have not understood that society cannot be developed by mere state machinery. The smooth functioning of a society is more dependent on social relationships than on legislative norms. This is better realized in the Anglo-Saxon law based on precedents than in the norm legislation based on legal codes. The main problem for Finnish Social Democrats, as well as for their Nordic comrades, is the fact that they are identified with the state machinery and the leading bureaucracy. Strong dislike of lords and leaders is still living among the masses, and the Social Democrats in Finland and Sweden have had to pay a high price for their integration as elites into the political system (Leijon, 1974). To a certain extent, the political integration of the welfare state into the bourgeois state has resulted in an ideological integration as well. The capitalist concept of economic efficiency has been adopted by the welfare state, and this together with the influence of conservatives has accelerated industrial development (Korpi, 1978a). Production and employment have become the most important criteria when problems are dealt with. Economic policy, therefore, preempts other policy concerns. MODELS OF POLITICAL SUPPORT AND THE WELFARE STATE

The problems of the welfare state and the basis of its support can be derived from certain models of system support (Pesonen and Sankiaho, 1979). The most significant is David Easton's model, which distinguishes between specific and diffuse support (1965, 1975). He describes the factors affecting the type of support and the objects of support as depicted in Figure 3.1. We note that diffuse support is derived more from socialization experiences, while specific support depends more on policy output evaluations. Furthermore, diffuse support is directed more toward the political community and the regime, while specific support is directed more toward political authorities and policy outputs. Diffuse support, then, is an implicit, unquestioning support, while specific support is tentative and conditional. Others have presented similar typologies that are close in meaning to the distinction Easton is making. Richard L. Cole (1973) employs the terms "normative" and "practical trust." Normative trust is a general trust of those in power as the legitimate authorities, while practical trust is based on political outputs. Paul Sniderman (1978) classifies citizens into the categories of committed and supportive. The committed citizens believe that the system acts honestly and that they have real opportunities to influence it. Therefore, they

Political Remobdization

T3

FIGURE 3 1 Easton's Model of Political Support Sources of Support

Type of Support

childhood and adult socialization

diffuse support

Object of Support political community political regime

performance evalua tions

political authorities specific support

policy output evaluations

policy outputs

eschew protest behavior, adhere strictly to the system's laws, and uncritically support its actions. The supportive citizens evaluate the outputs of the system and employ a variety of means to maximize their benefits. Easton (1975) has stated that the difference between diffuse and specific support is indicated by the cognitive qualifications that voters have to evaluate a political process and specify their demands. Voters with specific support expect system outputs to satisfy their needs; thus the outputs of the political system are very important for these voters. As a result, they attempt to vote in a way that guarantees the satisfaction of their demands. The problems of the Social Democratic parties have been compounded by the fact that the increasing levels of education, information, and political sophistication are resulting in a dramatic shift in the basis of political support from diffuse to specific criteria. Thus the Social Democratic parties cannot assume that they will be able to hold onto their traditional constituencies simply because they are the traditional representatives of the working class or because of what they have done for workers in the past. Increasingly, they will have to respond to specific voter demands to maintain their support. The growing emphasis on specific rather than symbolic outputs is, on the one hand, threatening to bankrupt the welfare state and already has led to the taxrevolt backlash. On the other hand, the increasing output evaluation orientation of the voters is threatening to split the Social Democrats' traditional constituencies as the result of the rise of a new set of crosscutting issues. NEW CLEAVAGES IN WELFARE STATES

Initially, the ideology of the welfare state was based on revisionist social ideas that could be described as antiestablishment. Stage by stage, together with battles for increased political rights, the ideology and policies of the welfare state have been incorporated into the establishment. At the same time, however, the concerns of many citizens in advanced industrial societies have moved from a primary interest in economic programs and system outputs to

74

Frameworks for Analysis

an increasing concern with participation and quality of life issues. The Social Democratic movement, however, has not reflected this change. Rather, it has remained preoccupied with issues of economic welfare. Thus even in countries with Social Democratic governing parties, political rights have remained unchanged since the acceptance of universal suffrage. Only in exceptional cases have certain democratic participatory rights been expanded to economic life. Examples here would be the Mitbestimmung or cooperative decision-making in Germany, and the lontagarfonder or employee funds in Sweden (Partanen, 1978). Moreover, by and large, the Social Democratic movement in many countries has not been able to effect reforms regarding the democratization of the educational system. Nor have the Social Democratic parties been able to renovate their organizational machinery and mobilize their members. As a result, these parties are losing contact with their supporters, even among their traditional constituencies, and are turning from mass parties into cadre parties. Consequently, a new cleavage between establishment and antiestablishment forces has arisen in Western societies. Traditional Left-Right conflicts—such as class and religion—can still be seen in many West European democracies and can be regarded as the prevailing forms of conflict in capitalist systems (Rose and Urwin, 1970). As Korpi states: The main sources of power in capitalist, industrial societies are, on the one hand, control over capital, and on the other, the numbers of wage earners. The resources inherent in the numbers of the wage earners can be channeled through organizations for collective action, primarily unions and political parties. (1978b: 98) However, this conflict has diminished in importance for two reasons. First, the theoretical underpinnings of class conflict have been weakened, as pointed out by Ralf Dahrendorf (1959) and others. Second, conservatives have attempted to eliminate class-based political conflict. Paradoxically enough, the bourgeois parties have at the same time tried to benefit from this conflict in election tactics and speeches. In many countries, they have started to identify themselves as nonsocialists, thereby enabling people to channel their rancor against socialist countries in East Europe and socialism in general through these movements (Valentin, 1978: 80). The influence of trade unions as integrators of the working class also is declining since trade unionism is tied more to one's occupation than one's class. In general, Social Democratic and Communist supporters identify with the working class more than members of trade unions do. On the other hand, trade union members identify more with the working class than with the population as a whole (Matheson, 1979). Attacks against welfare states now come from either of two directions.

Political Remobilization

75

First, attacks based on the tax revolt originate from the Right, although populist mass movements also have taken this question as one of their major issues (Korpi, 1978b: 106; Nilsson, 1979: 17). A tendency to vote in accordance with the instructions received from one's wallet presently has turned people away from the Social Democrats. The rightist parties have gained a number of supporters who are dissatisfied with the trade unions because their wage raises have not kept up with inflation and progressive taxation. Increasing taxation may be the most pressing problem that welfare states have to solve in the future. Harold Wilensky, for example, has discussed the implications of the "Glistrup curve" for welfare states.2 According to Wilensky, the real reason for Glistrup's landslide victory in 1973 was the soaring tax burden, which went from 33 percent of Denmark's GNP in 1968 to 44 percent in 1971. The second attack on the welfare state is an antibureaucratic one. This attack comes from populist movements, although conservatives also have accused the welfare state of excessive bureaucracy. Increasing bureaucracy tends to centralize decision-making, which, in turn, makes even active citizens feel powerless and politically inefficacious. All of this, consequently, alienates voters from the political process (Elvander, 1972; Ferrara, 1979; Pesonen and Sankiaho, 1979). These attacks on the welfare state have increased dissatisfaction among the general public. People generally feel satisfied with the high level of government services, and so criticism often focuses on perceived waste or bureaucratic inefficiency. In some nations, the dissatisfaction has grown to such proportions that cuts in social welfare funding seem to have won popular acceptance (Ferrara, 1979; Butler and Stokes, 1974; Ladd et al., 1979). Dissatisfaction also involves controlling the functions of the welfare state. Voters do not like the state's role as a guardian, whereby it aims to create social and political norms by legislation. Moreover, the state machinery with its well-known leaders is a clear object for the voters to attack. It is much more difficult for the average citizen to focus his or her opposition on an abstract concept like capitalism, which is criticized by the extreme Left in obtuse and ideological terms. The question, however, is whether this discontent is a permanent or temporary phenomenon. Some researchers claim that dissatisfaction only results from the present economic situation. They think that people will not attempt 2 In Denmark, tax revenue as a percentage of GNP at market prices increased much more than in other countries from 1968 to 1971. So rapid was the increase in Denmark that there was a real slope to the plot of taxes on GNP by year, while in other countries the plot was a continuing straight line. Hence Wilensky's use of the term "Glistrup curve" to explain the conservative victory in Denmark. On the problem of rising taxes in welfare states, see Wilensky (1975 and 1976); Heclo (1977); Grauhan and Hickel (1978).

76

Frameworks for Analysis

a real attack on the welfare state and its programs since its general goals have been adopted even by conservatives (Anners, 1972:272-273; Wilensky, 1975: 40-41; Korpi, 1978b: 106; Ferrara, 1979: 9-10). REMOBILIZATION IN WELFARE STATES

The challenge facing welfare states is to reintegrate those sectors of the public that have been alienated by the present political situation—that is, to bring the disaffected back into the political process. This remobilization of the electorate is a necessary step if Social Democratic parties are to represent their supporters and if democratic political systems hope to regain public support. However, this remobilization may prove to be a difficult task. Remobilization of the electorate is clearly affected by the internal changes among the working class, since in an advanced industrial society, white-collar workers outnumber blue-collar workers. At the same time, workers in the welfare state have become more bourgeois-minded, and this often has led to decreasing support for the traditional working-class parties. The trade-union movement has succeeded in integrating people into the political system, but in many cases this integration has been superficial and has done little to increase the political competence of the workers. Remobilization may be based partly on a "silent revolution," that is, on a shift from materialist values to postmaterialist ones (Inglehart, 1977). However, this also poses a problem for the welfare state. The ideology of the welfare state loses some of its attraction as progress is made on its aims of material well-being and security. Unfortunately, the Social Democrats have failed to move from materialist goals to social goals; that is, they have not progressed from the standard of living to the quality of living. The Western democratic system naturally is based on a secure livelihood and minimum living standards that enable real democratic development, but this development has to be continued to democratize all structures in society. "New Politics" has been a related theme in many studies (White and Sjoberg, 1972; Hildebrandt and Dalton, 1978). This policy is related to postmaterialist thinking and its goals are more postmaterialist than materialist. In general, the New Politics has moved in an antiestablishment direction. Therefore, this new pro/antiestablishment cleavage might become a permanent aspect of politics and might affect future political developments. This cleavage is found not only in the political sphere but also inside many organizations. Thus this "silent revolution" actually could increase alienation from elites. The ecological movement and opposition to nuclear power, for instance, have developed among citizens who possess considerable resources, and such movements may produce a legitimacy crisis for democratic societies (Duverger, 1973; Inglehart, 1979a; Dalton, this volume: Chap. 4; Korpi, 1978a; Elvander, 1972; Pesonen and Sankiaho, 1979; Buerklin, 1981).

Political Remobilization

77

The present technocratic establishment is, according to many studies, incapable of solving the major social problems that lie ahead for welfare states. Bureaucratization of government is seen as a continuous development that is decreasing the public's feelings of political efficacy. Bureaucratic alienation and other factors could generate more support for antiestablishment views (Lindberg, 1976; Heclo, 1977; Levitt, 1976; Frey, 1979). The welfare state generally has succeeded in producing materialist services for its citizens, but to a certain extent it has failed to meet their social needs (Scharpf, 1977; Pesonen and Sankiaho, 1979). Many citizens no longer feel that the expanding legislation of welfare states is furthering the development of society or its social functions. Thus, for instance, Finnish young people who give priority to materialist values are less alienated from their society than those who place greater stress on postmaterialist concerns (Sankiaho, 1979b). The most integrated young are those whose parents had postmaterialist values while they had materialist values; in other words, those whose values developed in the opposite direction of the general trend. To summarize what has been said: (1) Social Democratic parties now are often identified with the state authorities, the bureaucratic state machinery, and the establishment in general. (2) The major issues of politics involve competition over the division of economic resources. The governing parties attempt to reach compromises on these issues in negotiations and to distribute the outputs among their supporters. (3) As a result of integration into the system and increased pragmatism, ideologies are pushed into the background. The governing parties lose their ideological consistency, and the everyday fight over voters makes them offer special "sales" on present issues. (4) Political activities within the parties are divided among small groups of specialists. Therefore, power in a political party is concentrated within a group of ministers and leading bureaucrats. The iron law of oligarchy is still valid as far as Social Democratic parties are concerned. (5) The governing parties draw themselves farther away from their supporters and from some of their members. Furthermore, they are unable to mobilize many elements of the population that their programs benefit because none of the organizations that they control have sufficient contacts with them. (6) Postmaterialist people cannot be mobilized by materialist outcomes and rewards. The goals of Social Democratic party programs are not valid for these voters. They partially share Social Democratic goals in the field of equality, but they also have many goals that are very different from labor parties' official lines, especially in ecological concerns.

TESTING THE MODEL

I have argued that the political tensions of welfare states are becoming organized along a pro/antiestablishment dimension, reflecting the voters' al-

78

Frameworks for

Analysis

ienation from the political system and a "silent revolution" of postmaterialist value change. In the future, this establishment dimension may be more im­ portant than traditional Left-Right cleavages, especially for remobilizing and reintegrating voters into the political system (Kaase and Barnes, 1979; Inglehart, 1977). The data to examine this thesis were collected in 1974-1975 in six industrial democracies, most of which (with the possible exception of the United States) can be labeled "welfare states." 3 The operationalization of the establishment dimension is based on the feeling thermometer ratings of various social and political groups. One of the first two dimensions of group evaluations in every nation is a factor distinguishing proestablishment and antiestablishment groups (Table 3.1). 4 The factor structure is not identical in each nation, but there are basic similarities. The police and civil servants consistently are seen as proestablishment groups, and often are accompanied by the church, small business, as well as big business. In The Netherlands, even the trade unions are regarded as an establishment group. In every case, student protestors and revolutionaries are the most antiestablishment groups, occasionally joined by women's groups or minorities. For examining the remobilization challenge facing Social Democrats, my analyses are based on a typology consisting of Left-Right party preferences and the pro/antiestablishment dimension (Table 3.2).5 This typology allows us to examine the relationship between these two cleavage dimensions. In general, supporters of leftist parties are more antiestablishment than supporters of the Right. The only exception is the United States, where virtually no relationship is found between the establishment dimension and party support. The future problem of support for welfare states will involve mobilization of antiestablishment-minded people. Because these antiestablishment groups are often leftists, this poses a special problem for Social Democratic parties. 3

This study has been conducted in eight industrialized countries, six of which are presented here The principal investigators were Alan Marsh and Mark Abrams (Great Britain), Samuel Barnes, Ronald Inglehart, and Μ Kent Jennings (United States), Max Kaase and Hans Klingemann (Germany), Philip Stouthard and Felix Heunks (The Netherlands), Leopold Rosenmayr (Austria), and Perth Pesonen and Risto Sankiaho (Finland) These data are available from the Inter-University Consortium for Political and Social Research, University of Michigan, or from the Zentralarchiv fur emptnsche Sozialforschung, Umversitat Koln Analyses from the first five nations and additional details about the study are presented in Barnes, Kaase et al (1979) Results from all eight nations (including Italy, Finland, and Switzerland) will appear in Pesonen, Kerr, and Klingemann, eds (forthcoming) 4 These analyses are based on factor analyses of the group thermometers, rotated to a vanmax solution For additional details on these scales, see Barnes, Kaase et al (1979 581-583) 5 The definition of the Left-Right party identification groups can be found in Barnes, Kaase et al (1979 575-577) In Finland, the Left was the Communists and Socialists, the Center was the Center party, Finnish Rural party, Liberal party, Swedish People's party, the Right was the Conservative party, Christian League, Constitutional party, and People's Unity party For more regarding the typology introduced here, see Sankiaho (1979b), Buerklin (1981)

Political Remobilization

79

TABLE 3.1 The Establishment-Antiestablishment Dimension in Six Nations

Police Clergy Civil servants Small business Big business Major Right party Unions Major Left party Minority group Women's movement Student protestors Revolutionary groups

Britain (GB)

Germany

Ψ)

Austria (A)

United States

Netherlands (NL)

Finland (SF)

.61 .51 .37 .37 .16 .18 -.08 .00 -.19 -.44 -.70 -.67

.76 .08 .57 .48 .13 .17 -.19 -.17 -.19 .02 -.71 -.71

.65 .23 .57 .25 .07 -.03 .15 .11 -.37 -.08 -.71 -.67

.72 .59 .46 .37 .39 .18 .03 -.01 -.24 -.57 -.76 -.71

.69 .39 .54 .18 .31 -.01 .32 .02 -.31 -.68 -.45 -.69

.64 .53 .34 .44 .38 .23 .12 .06 -.30 ί

-.63 -.69

SOURCE: Political Action Study, 1974-1975. NOTE: Table entries are loadings on the establishment-antiestablishment dimension obtained by factor analysing the group thermometers. ° Variable not included in study. The following sections examine the correlates of this new ideological typol­ ogy, thereby enabling us to understand the potential bases for mobilization. Patterns of Political Alienation One of the basic questions we face involves the degree of political support among antiestablishment groups. Table 3.3 displays the relationship between our ideological typology—combining the establishment dimension and LeftRight party preference—and two measures of political alienation (Pesonen and Sankiaho, 1979: 257-260; Citrin, 1977: 389). The first measure is a general index of political trust. Little difference is found among the supporters of the Center parties, but if we look respectively at the supporters of the Left and Right, in nearly every case those with a proestablishment orientation are more trusting than those with an antiestab­ lishment orientation. These differences are greatest in The Netherlands and Austria. We also find that in countries where the Social Democrats had been in power for some time prior to the 1974-1975 surveys, such as in Germany, Austria, and Finland, the political trust of leftists is higher than that of rightists. In the other countries we find the opposite pattern, where the conservatives are more trusting than the leftists. This is clearly an incumbency effect. The second measure in the table is an index of general dissatisfaction with

-6 583

PDI" (N)

-14 513

10 67 24 101

14 66 20 100

-6 662

A

D

-0 431

-4 859

5 545

13 65 22 100

14 68 18 100

19 67 14 100

SF

NL

USA

-2 203

15 68 17 100

GB

-7 113

11 72 18 101

D

Center"

21 300

30 62 9 101

SF

0 608

16 69 16 101

GB

11 726

19 73 8 100

D

16 176

21 74 5 100

A

3 387

16 72 13 101

USA

Right Party ID

11 458

19 73 8 100

NL

12 445

20 72 8 100

SF

Action Study, 1974-1975. scores for the establishment dimension were collapsed so that values less than —1.0 are antiestablishment, from —1.0 to +1.0 are neutral, and greater than +1.0 are proestablishment. * Only in three nations were there found a significant number of supporters of parties classified as ideologically in the Center. b This percentage difference index (PDI) is the difference between the pro- and antiestablishment categories.

13% 68 19 100

Proestablishment Neutral Antiestablishment

SOURCE: Political NOTE: The factor

GB

Orientation

Left Party ID

TABLE 3.2 The Distribution of Ideology Types

81

Political Remobihzation TABLE 3 3 Political Alienation by Establishment Categories Antiestablishment

Proestablishment

Neutral

L

C

R

L

C

R

L

C

R

Political trust" Britain Germany Austria United States Netherlands Finland

16 26 28 17 19 22

20 (2 7) — — — 23

23 23 23 20 20 (19)

17 29 31 19 24 23

20 29 — — — 24

26 23 25 21 23 24

18 31 32 19 27 24

19 (2 7) — — — 23

26 22 26 23 24 22

Political dissatisfaction Britain Germany Austria United States Netherlands Finland

32 30 26 32 32 34

28 (3 2) — — — 31

24 27 26 28 31 (3 4)

29 28 23 29 27 30

28 28 — — — 30

24 30 27 28 28 31

28 25 22 28 25 30

25 (2 7) — — — 30

23 30 24 26 26 29

SOURCE Political Action Study, 1974-1975 Table entries are mean scores on the scales Entries in parentheses are based on twenty cases or less a The trust index is coded (1) low trust to (4) high trust (Barnes, Kaase et al, 1979 575) b The political dissatisfaction index is based on government performance across a set of political issues (also see Table 3 8) The index is coded (1) satisfied to (5) dissatisfied (Barnes, Kaase et al, 1979 572) NOTE

government policies, close to Easton's concept of specific support General political dissatisfaction is obviously higher among antiestabhshment groups, especially for leftists This presents a problem for labor party governments How do they show these voters that support for the incumbent government is in their best interests'' In many cases, this dissatisfaction has led to opposition voting (Farah, Barnes, and Heunks, 1979, Pesonen and Sankiaho, 1979, Alt, this volume Chap 10) In summary, the high levels of distrust and dissatisfaction among antiestabhshment groups suggest that political parties may face severe difficulties in attempting to mobilize this sector and reintegrate these voters into the political system

82

Frameworks for Analysis TABLE 3 4

Political Efficacy by Establishment Categories Antiestablishment

Efficacy GB D A USA NL SF

Proestablishment

Neutral

L

C

R

L

C

R

L

C

R

21 25 20 23 23 22

24 (2 5) — — — 23

24 21 19 24 22 (19)

20 21 18 22 22 23

21 21 — — — 24

22 19 17 24 22 24

20 20 16 22 20 24

19 (19) — — — 23

22 18 16 23 20 20

SOURCE Political Action Study, 1974-1975 Political efficacy entries are the mean scores on the scale coded (1) not efficacious to (4) efficacious (Barnes, Kaase et al, 1979 573-574) NOTE

Patterns of Political Efficacy When we turn to the question of political efficacy (Table 3 4), we find an interesting contrast with the aforementioned pattern While the antiestabhshment categories for the Left and Right party supporters, respectively, are less trusting and satisfied with politics than the proestabhshment categories, these antiestabhshment categories, at the same time, score higher on the political efficacy index This pattern of higher efficacy among the antiestabhshment categories holds true for all countries except Finland Thus the antiestabhshment category has traits that often contribute to intense political opposition— alienation from the political system coupled with a sense of political efficacy Moreover, these orientations tend to be strongest among the antiestabhshment leftists whom the Social Democratic parties must remobihze Patterns of Political Action The ability to reintegrate antiestabhshment voters into the political system also may depend on the types of political involvement they pursue Table 3 5 presents the relationship between our ideological typology and several different modes of political action Conventional political participation is almost at the same level across the different establishment categories In general, leftists are more active than nghtists in the antiestabhshment category, but this difference is not large and tends to decline among the neutralists or proestabhshment-minded National differences often present inconsistent deviations from this pattern Protest potential—support for unconventional forms of political participation such as boycotts, demonstrations, and wildcat strikes—displays stronger

Political Remobilization

83

TABLE 3 5 Political Action by Establishment Categories Anti-

establishment L Conventional" political participation

Protest potential

Repression"

C

GB D A USA NL SF

19 20 33 (3 2) 22 — 32 — 24 — 26 15

GB D A USA NL SF

31 31 37 (3 4) 26 — 42 — 52 — 36 23

GB D A USA NL SF

16 12 10 (0 8) 16 — 07 — 01 — 08 21

Proestablishment

Neutral

R

L

C

R

L

13 22 19 20 16

16 23 20 26 17 20

18 20 — — — 19

18 20 20 30 20 19

16 19 21 (2 3) 23 — 28 — 14 — 19 18

20 23 15 24 34 22

22 21 — — — 15

18 16 17 24 22 19

15 22 15 15 (15) 12 13 — 13 17 — 17 24 — 17 13 12 12

20 18 20 20 07 17

22 21 — — — 24

27 25 25 22 14 22

21 29 26 (2 9) 22 — 26 — 10 — 23 28

(16)

20 19 21 30 32 (16)

22 19 16 14 06 (15)

C

R 19 18 15 29 18 25

28 32 31 29 22 30

SOURCE Political Action Study, 1974-1975 NOTE Table entries are mean scores on the scales Entries in parentheses are based on twenty cases or less a The conventional participation index is coded (0) inactive to (7) complete activist b The protest potential scale is coded (0) inactive to (7) complete protestor c The repression potential scale is coded (0) oppose repression to (5) support strong repression

and more consistent ideological differences The second panel in Table 3 5 indicates that protest potential is higher among leftists, with only a few exceptions Differences between the party groups are not so large among the proestablishment-minded However, party differences are very sharp among the antiestablishment category Antiestablishment leftists almost score twice as high on protest potential as the antiestablishment nghtists Moreover, all the antiestablishment categories are much higher on protest potential than are the proestabhshment categones

84

Frameworks for

Analysis

The third panel in Table 3.5 presents ideological patterns on an index of repression potential. Support for repressive action against political protestors is naturally much higher among proestablishment groups. In addition, the rightists in each establishment category are more eager than leftists to prevent protest action. In sum, antiestablishment leftists are the most active type of citizen in both conventional and unconventional modes of participation.6 Thus alienated Social Democratic voters appear especially willing to challenge political elites directly, thereby increasing the potential for political conflict in welfare states. Patterns of Ideological Thinking The political sophistication of the various establishment groups can be determined in several ways. The first panel in Table 3.6 presents the percentage in each typology group who do not use ideological concepts of any form in evaluating political parties—the lowest level of ideological sophistication (Klingemann, 1979: 281-284; Converse, 1964). The proestablishment groups have many more nonideologues than the antiestablishment groups, especially among leftists. The British case is a partial exception to this pattern in that antiestablishment rightists are more often nonideologues than proestablishment rightists. While on the whole there is less variation across establishment categories for rightists, in the other countries the pattern of lowest sophistication among the proestablishment groups holds. Another measure of ideological sophistication is based on the use of ideological concepts in connection with Left-Right self-placement.7 On the whole, we find that the proportion of Left ideologues is greater among the antiestablishment categories than the proestablishment groups. In the case of the percentages of Right ideologues, the differences across establishment categories are neither very great nor in a consistent direction. As is to be expected, those who identify themselves with the left side of an ideological self-placement scale do not always support a party of the Left. Right ideologues also occasionally defect from their logical ideological choices on the question of party support. It is interesting to note, however, that among the antiestablishment categories, more Left ideologues support parties of the Right than Right ideologues support parties of the Left. In the case of the proestablishment categories, we find the opposite pattern. This suggests that there is more movement away from the Social Democratic parties among the antiestablish6

Other tables (not shown) indicate that antiestablishment leftists also express an extremely high level of political interest, and score high on the action typology developed by Kaase and Marsh (1979 152-163) 7 Left ideologues are high or middle-level ideologues and score left of the overall mean on the Left-Right self-placement scale, Right ideologues also score high on the ideological sophistication scale but nght of the overall Left-Right mean For additional details, see Barnes, Kaase et al (1979 229, 563)

85

Political Remobilization TABLE 3.6 Ideological Thinking by Establishment Categories Antiestablishment

L Ideological thinking in politics Percentage nonideologues

Ideological orientation in politics Percentage Left ideologues

Percentage Right ideologues

Strength of party attachment Percentage very close

C

GB D A USA NL

65 40 63 21 (15) 28 4 8 — 5 0 4 4 — 4 7 21 — 28

69 35 49 64 41

GB D A USA NL

31 24 71 (65) 30 — 42 — 76 —

18 17 2 16 22

23 15 51 37 3 5 — 15 — 48 —

GB D A USA NL

4 32 8 (20) 12 — 14 — 3 —

33 55 48 36 50

7 14 16 20 11

GB D A USA NL

9 17 16 16 (15) 10 21 — 12 14 — 5 10 — 8 27 23 (11)

SF

Proestablishment

Neutral R

61 51 21 29 — 47 — 43 — 32

10 13 3 14 12

24 48 32 58 — 50 — 43 —57

19 12 12 10 24 — 18 — 10 — 14 13

14 12 18 11 10 11

79 63 49 39 (17) 31 50 — 53 66 — 49 38 — 39

15 50 40 11 54

4 1 (42) 8 — 2 — 18 — 14

3 27 50 11 (42) 61 10 — 45 23 — 33 9 — 47

33 23 35 25 20 17

13 30 (8) 23 — 22 — 22 — 23 23 36

SOURCE: Political Action Study, 1974-1975. NOTE: Entries in parentheses are based on twenty cases or less. See footnote 7 for descriptions of the ideology measures.

FrameworL· for Analysis

86

ment categories and more movement toward the Social Democratic parties among the proestablishment categories. The interplay between political sophistication and party commitment also can be measured by the strength of party attachments. The final panel in Table 3.6 shows that, on the whole, leftists are somewhat stronger identifiers than rightists. The more striking finding, however, is that the number of strong identifiers is much greater within the proestablishment category, for both the Left and Right. Thus an increase in antiestablishment voters may indeed weaken the bonds of the party system. Patterns of Value Priorities Materialist/postmaterialist values are an important aspect of the New Politics represented by the establishment dimension (Inglehart, this volume: Chap. 2). Table 3.7 describes the distribution of these value priorities across the ideological typology. Establishment categories show great differences in value orientations. The establishment groups are the most materialist, whether they are leftist or rightists. Postmaterialism is far more frequent within the antiestablishment category, and postmaterialists occasionally outnumber materialists among leftist antiestablishment voters. These value differences underscore the point I have been making: labor parties must adopt new policies in order to mobilize antiestablishment voters. Traditional, materialist, welfare state ideologies are not enough for cynical, but sophisticated, postmaterialist citizens. TABLE 3.7 Postmaterialist Values by Establishment Categories AndEstablishment

Values index Britain Germany Austria United States Netherlands Finland

ProEstablishment

Neutral

L

C

R

L

3 -1 -25 0 35 3

3 (0) — — — -8

-37 -57 -46 -16 -3 (-11)

-27 -49 -52 -41 4 -18

C

R 51 71 64 43 21 32

L

C

R

-41 -71 -61 -51 -36 -31

-43 (-75) — — — -35

-60 -79 -77 -64 -39 -56

SOURCE: Political Action Study, 1974-1975. NOTE: Table entries are the difference between the percentage of postmaterialists (codes 8-10) in each group minus the materialists (codes 1-3). Entries in parentheses are based on twenty cases or less.

Political Remobilization

87

Patterns of Issue Orientations General values and ideological beliefs also should be reflected in issue orientations. Table 3.8 compares orientations toward issues of individual wellbeing (the traditional concern of welfare states) and social equality issues (an example of New Politics concerns). All establishment categories are uniformly high in the belief that issues of individual well-being are an important responsibility of the nation-state. The biggest differences between establishment categories are in Germany and Austria, with proestablishment voters giving more importance to these issues. Agenda means are typically higher among leftists, independent of establishment category. However, these differences are weak in comparison to the similarities across ideological groups. Social equality issues are related to postmaterialist values (Farah, Barnes, and Heunks, 1979). In every nation, social equality issues are given a higher position on the political agenda by antiestablishment voters, especially among those who support leftist parties. Only in Austria and Finland is there little variation by establishment category. Leftists also are more likely than rightists to stress social equality issues in all establishment categories. Dissatisfaction with government policies follows much the same pattern for the individual well-being and social equality issues. In general, antiestablishment voters are more dissatisfied with government policies in both issue domains. Supporters of leftist parties usually are more dissatisfied than centrists or rightists. However, in the case of dissatisfaction with individual well-being, there is evidence of a government/opposition pattern. That is, voters in opposition are more dissatisfied even if they are rightists. The government-opposition cleavage is not so obvious for the social equality issues, perhaps because these issues tap concerns that governments have not identified themselves with clearly.

SUPPORT MODELS AND POLITICAL DEVELOPMENT

Based on the models of political support presented earlier, some obvious lines of development already are occurring and perhaps will continue in the future. In terms of Easton's model, specific support is increasing in comparison to diffuse support, and therefore citizen evaluations of the political process are becoming more important. The style of the "Old Politics" is based mainly on high levels of public trust in political party and government leaders, without a close and careful evaluation of political outputs. Our analysis of the establishment cleavage has shown that antiestablishment people are not so trustful of political authorities and are more likely to base their evaluations on performance criteria.

88

Frameworks for Analysis

TABLE 3.8 Issue Orientations by Establishment Categories Antiestablishment

Proestablishment

Neutral

L

C

R

L

C

R

L

C

R

Individual well-being agenda"

GB D A US NL SF

4.4 4.1 4.2 4.1 4.4 4.2

4.3 (4.0) — — — 4.1

4.4 4.0 4.1 3.5 4.2 (4.0)

4.5 4.2 4.2 3.8 4.3 4.1

4.4 4.2 — — — 4.0

4.2 4.1 4.0 3.5 4.2 3.8

4.4 4.4 4.4 3.9 4.3 4.0

4.4 (4.1) — — — 4.2

4.3 4.3 4.3 3.5 4.3 4.0

Social equality agenda"

GB D A US NL SF

3.2 3.8 3.0 3.5 4.0 3.1

3.1 (3.3) — — — 2.9

2.6 2.9 3.0 3.0 3.2 (2.4)

2.9 3.2 2.9 2.8 3.4 3.0

2.6 2.9 — — — 2.8

2.3 2.9 2.7 2.6 3.0 2.2

2.0 3.3 2.7 2.6 3.2 2.9

2.6 (2.6) — — — 2.9

2.1 2.8 2.8 2.2 2.6 2.5

GB Individual D well-being dissatisfaction* A

3.1 2.6 2.2 3.1 3.0 3.4

3.7 (2.7) — — — 3.1

2.1 2.3 2.2 2.8 2.7 (3.4)

2.6 2.3 1.9 2.8 2.4 2.8

2.5 2.4 — — — 2.8

2.2 2.5 2.2 2.6 2.5 2.9

2.6 2.0 1.7 2.7 2.2 2.7

2.2 (2.0) — — — 2.7

2.0 2.5 2.0 2.4 2.1 2.7

3.4 3.3 3.0 3.1 3.5 3.4

2.9 (3.4) — — — 3.2

2.8 2.9 2.9 2.9 3.2 (3.3)

3.2 3.1 2.7 2.8 3.0 3.2

3.1 3.1 — — — 3.0

2.8 3.1 2.9 2.7 3.0 3.0

3.0 2.9 2.6 2.9 2.8 3.3

2.9

2.9 3.1 2.6 2.7 2.9 2.9

US NL SF GB Social D equality dissatisfaction* A US NL SF

SOURCE: Political Action Study, 1974-1975. NOTE: Table entries are mean scores on the

(3-D — — — 3.0

issue agenda and dissatisfaction indices. Entries in parentheses are based on twenty cases or less. 1 Both issue agenda indices are coded: (1) issues are not important and the government is not responsible, to (5) issues are very important and they are an essential responsibility of government (Barnes, Kaase et al., 1979: 570). b Both issue dissatisfaction indices are coded: (1) issues are very important and government performance is very good, to (5) issues are very important but government is performing poorly (Barnes, Kaase et al., 1979: 572).

Political Remobilization

89

The overall characteristics of this antiestablishment category can be summarized with a series of multiple regression analyses, as shown in Table 3.9. The variables used in the previous descriptive analyses produce a multiple R ranging from .35 (Britain and Austria) to .47 (Germany). The best variable in predicting pro/antiestablishment views is postmaterialist values, with a significant beta coefficient in each nation. This variable alone is a major determinant of establishment views. Other variables that are significantly and consistently related to the establishment dimension are political trust, LeftRight self-placement, efficacy, and general political dissatisfaction. What distinguishes an antiestablishment orientation from a proestablishment orientation, then, is that the antiestablishment-type is more likely to have postmaterialist values, low trust, high efficacy, a leftist ideology, and a general dissatisfaction with politics in his or her country. TABLE 3.9 Regression Analysis of the Establishment Dimension GB Materialist/postmaterialist values Political trust Left-Right scale Political efficacy Political dissatisfaction Few big interests Interest in politics Ideological sophistication Parties only interested in votes Under/Privileged groups Multiple R

D

NL

A

US

SF

-.21 .14 .11 -.16 -.07 -.07* -.02* .01* .03* -.03*

-.27 .29 .19 -.11 -.05* -.16 -.10 .03* -.03* -.04*

-.24 .27 .10 -.14 -.16 -.16 -.04* .09 .07* -.02*

-.19 .07* .06* -.22 -.09 .00* -.06* -.02* .04* .04*

-.26 .21 .16 -.08 -.11 -.14 -.06* .06* .07 .02*

-.23 -.02* -.28 -.14 -.09 .04* -.05* — -.00* .07

.35

.43

.47

.35

.44

.45

SOURCE: Political Action Study, 1974-1975. NOTE: Table entries are partial standardized regression coefficients (betaweights). * Not significant at .01 level.

The future problem for Social Democratic parties and welfare states will be the mobilization of these antiestablishment citizens. The number of antiestablishment citizens also may grow, for education and other social effects have increased political efficacy at the same time political trust has declined tremendously (Pesonen and Sankiaho, 1979; Miller, 1974; Citrin, 1974). Growing political protest is also a typical development over the last decade (Marsh, 1977). In the future, mass political power may be located among antiestablishment people, and thus the establishment power structure will face greater problems in maintaining the system.

90

Frameworks for Analysis

In some cases, the present establishment has shown itself incapable of solving major problems involving the whole of society (Lindberg, 1976; Heclo, 1977; Anners, 1972). The development of the welfare state was relatively easy after World War II, a time in which there was consensus among highly industrial countries and social welfare values also were adopted by conservatives. However, the welfare state has not made many basic structural changes. According to Claus Offe: What appears in the welfare state are new elements within advanced capitalist societies, but no basic changes of these societies. That is, the welfare state has not changed political and economic power relationships. . . . [D]efense and space industries, agribusiness, the industrial users of government-guaranteed foreign loans and publicly financed research and development capture the lion's share of state welfare. (1972: 481-482) The increasing rates of taxation in welfare states have provoked tax revolts, which in California eroded the local authorities' tax collection powers (Attiyeh and Engle, 1979). Future materialist development may not match past growth rates. On the other hand, the New Politics is not demanding this type of progress. Rather, it calls for the development of a new society. CONCLUSION

I have argued that contemporary welfare states are going through a period of change. The present governing establishment—which often includes leftist parties, trade unions, and employee organizations—is labeled by Maurice Duverger (1973) a "technocratic democracy," whose only critics come from the antiestabhshment. Nils Elvander (1972) sees this development leading to technocratic solutions. Society then will lack political direction because the bureaucratic establishment can make only social-technocratic decisions. The present system also poses a problem because the pro/antiestablishment cleavage exists inside many social organizations, and many researchers regard the solution to this problem as one of the key issues ahead for social development (Duverger, 1973; Lindberg, 1976). If the present development toward increasing technocracy and bureaucracy cannot be changed, it may lead to an increasing concentration of authority and undemocratic developments. At the same time, value change is turning a portion of the electorate in the opposite direction. Perhaps the best example of the New Politics is the nuclear energy question. Social Democratic parties, as supporters of an economic development ideology, have been strong supporters of nuclear energy. In Sweden, this issue caused a change in government when bourgeois parties won the majority by a very small margin in the 1976 election (Petersson, 1977). Antiestabhshment

Political Remobilization

91

voters with postmaterialist values regard the nuclear energy question not only as an economic issue, but also as a threat to the future. Nuclear energy is connected to pollution and increasing government and technocratic control. The nuclear energy question will be even more important in the future because it is a life-and-death question to European youths. The postmaterialist, antiestablishment citizen is ready for alternative energy sources, even if it requires lowering the standard of living. The nuclear energy question also is an example of the difficulties Social Democratic parties will have in mobilizing antiestabhshment voters, and dealing with this cleavage within the parties. Some elections in the Western world offer obvious examples of the importance of mobilizing antiestabhshment citizens. In the United States, for example, Jimmy Carter as a Washington outsider could garner support from the alienated, distrustful, antiestabhshment voters in 1976—but no longer in 1980. Margaret Thatcher was also able to gain votes from antiestabhshment citizens who were ideologically on the Right or in the Center and perhaps even from some antiestabhshment leftists. As in the United States, the antiestablishment dimension in Britain may for some voters eclipse the traditional Left-Right dimension. In West Germany and Austria, the Social Democratic parties are in a different position, because in those countries people are more ideologically linked to politics. In Austria, the Social Democratic party is having great difficulties in gaining antiestabhshment participation in the government. The same could be said for West Germany until the recent change in government from the Left to the Right. One reason for this change in government was the organization of antiestabhshment voters as a "Green" party. A Green party also won parliamentary seats in the Finnish elections of 1983. In the future, Social Democratic parties may find themselves struggling with these Green parties for the loyalties of leftist antiestabhshment voters. Using rhetoric and political ploys to mobilize active antiestabhshment groups is not enough. That is why societies led by Social Democrats need not only pragmatic solutions, but also structural changes. Democratic steps should be taken especially in the economic field. Perhaps developing democracy in economics (codetermination) could be a guiding star in leading antiestabhshment groups back into the Social Democratic movement. The old and onesided standard-of-living thinking of the Social Democrats, which is heavily linked to individual welfare and a growth mentality, is not sufficient for mobilizing antiestabhshment, postmaterialist voters. The parties also may have to change their organizational structure and develop new types of participation channels, because active antiestabhshment groups with high protest potential are not satisfied with the present opportunities for participation within the parties. The social and political importance of the pro/antiestablishment cleavage

Frameworks for Analysis

92

is on the rise. While it will probably not replace traditional Left-Right cleavages, it is certain that this new cleavage is changing the traditional cleavage alignment.8 Political struggles in the future will be not only about bread-andbutter issues, but also about citizen influence, a better environment, and the quality of life. 8

Finnish Gallup surveys from 1972 to 1982 show that perceptions of the strength of the main social cleavage (capitalist-working class) have decreased: Capitalist Working-Class Cleavage Extremely strong Quite strong Quite weak Not at all Don't know

All Citizens

Social Democrats

1972

1978

1982

1972

1978

1982

56

44

36

68

51

43

27 10 2 5

36 11 2 7

42 14 2 6

22 6 4

36 9 0 4

40 11 2 3

100% (994)

100% (1171)

100¾ (1697)

100% (178)

100% (269)

100% (473)

Center Party

Communists

Conservatives

1972

1978

1982

1972

1978

1982

1971

1978

1982

78 12 5 1 3

64 26 8 1 1

52 35 7 2 4

46 38 10 2 4

42 31 13 5 8

28 47 13 2 10

46 37 13 2 2

29 45 16 3 7

25 47 22 3 3

100% (93)

100% (160)

100% (168)

100% (243)

100% (189)

100% (232)

100% (122)

100% (177)

100% (344)

THREE

Patterns of Realignment

Scoff C Flanagan

This section includes four case studies of realignment processes in Germany, the United States, Japan, and Italy. The focus on realignment does not imply that dealignment processes are necessarily absent from these cases. This should be evident from the fact that a second American case study is included in the dealignment section. The inclusion of these case studies in this section simply means that the focus of their analyses is on the processes of realignment. Definitions of electoral realignment normally are based on two factors: a change in the sources of party support, and a change in party vote shares. Both are important phenomena in their own right. A shift in the basis of party support may alter a party's policies and programs. A gain or loss in a party's vote share may be of major importance if it alters control of the government. The two aspects of electoral realignment normally occur in tandem; a realignment in group voting patterns generally produces a change in vote shares. Conceptually, however, these are distinct phenomena and, as will be seen, one can occur without the other. Either change, then, qualifies in our terminology as a realignment, albeit a different type of realignment. Another concern of research on realignment is the pace of partisan change— the distinction between critical and secular realignments. Critical realignments are a rarity. None of the authors in this volume present an example of a recent critical realignment, even though we examine the electoral histories of over a dozen democracies. Rather, the four studies in this section analyze more gradual secular realignments. These examples of secular realignment fall into two different patterns, which exemplify the two distinct aspects of realignment referred to previously. In two of the cases, Germany and the United States, we are observing a pattern of sectoral realignment, that is, changes in the base of support for one or more of the parties. In the German case, the SPD was able to gain new support, especially among the new middle class, without losing its traditional base among the working class. This sectoral realignment, therefore, was accompanied by a gradual rise in support for the SPD. In the American case, the sectoral realignment was preceded by an issue realignment at the elite level, which precipitated the movement of procivil rights Republicans into the Democratic party and racially conservative Democrats into the Republican party. If the sectoral realignment involves this kind of exchange, where the movement in one direction is roughly compensated by movement in the opposite direction, the realignment may have no impact on aggregate

96

Patterns of Realignment

vote distributions. Thus a shift in party constituencies does not necessitate a shift in party fortunes. In the remaining two cases, Japan and Italy, long-term changes in party vote shares reflect a process of ecological realignment, that is, changes in the relative size of one or more social groups or economic strata. In this case, the bases of party support do not necessarily undergo marked changes. Rather, the relative size of various demographic or attitudinal groups changes, bringing with it a shift in the fortunes of two or more parties. In both Japan and Italy, the agrarian sector has declined substantially over the last two to three decades, while the urban, industrial labor force and the service sector have grown. Moreover, in Japan there has been a reduction in the proportion of the electorate holding traditionalistic values, while in Italy religiosity has declined, especially among women. These changes have reduced the relative size of traditional conservative constituencies in both countries and resulted in falling levels of electoral support for the major conservative parties, the LDP in Japan and the DC in Italy. An important distinction should be made between our two major patterns of realignment. Ecological realignments are largely the product of social forces beyond the control of elites. Ecological realignments may induce elite responses, but are generally not directly caused by elite actions. In contrast, sectoral realignments often result from prior changes at the elite level. This latter form of realignment is generally a response to changes made by the parties themselves. One party may attempt to broaden its support by adding planks to its platform designed to appeal to groups outside its traditional constituency, another party may espouse new issues, or a third may reverse its position on a long-standing major issue. The process of sectoral realignment, then, draws attention to the relationship between changes in mass politics and actions at the elite level. Thus mass level changes cannot be described adequately without also considering the realignments that are taking place among the parties both in the parliamentary and party system contexts. REALIGNMENT OR DEALIGNMENT?

Most of our realignment case studies present a clearer picture of how old alignments are crumbling than they do of the form new alignments will take. This may be a product of our present time perspective. There may not have been sufficient time for the realigning forces to run their course. In some of these cases, therefore, we cannot be totally confident that we are witnessing realignments as opposed to a general weakening of party cleavages—that is, dealignment. For example, in the American case the form of the new realignment is most clearly spelled out. Yet the authors are describing an issue realignment rather than the more traditional, social group-based, partisan realignment. The parties have exchanged positions on a major issue and that

Patterns of Realignment

97

has precipitated a realignment of partisan identifications among segments of the population for whom the issue is highly salient. An issue realignment, however, is not incompatible with the model of a dealigning electorate in which issue voting increasingly takes precedence over traditional party loyalties. Issue voters may in fact be volatile identifiers whose party identifications travel with their shifting issue concerns (Flanagan and McDonald, 1979; Borre, this volume: Chap. 11). So, whereas a realignment among at least some strata does appear to be taking place in the United States, this new realignment may not be as stable or enduring as, say, the New Deal realignment. Similarly, the other three realignment cases display clear changes in voting patterns, but the underlying forces associated with these changes lead us to question the durability and definition of the new realignments. In the German, Japanese, and Italian cases, there are gradual long-term shifts in the party vote distributions over the past twenty to thirty years. In rough terms, all three of these shifts can be described as a gain in support for the progressive parties relative to the strength of the conservative parties. In each case, a dominant conservative party emerged shortly after the war. While in Japan and Italy these conservative parties consistently have held power, all three have suffered a long-term erosion of their support, at least relative to the parties to their Left. This shifting balance of power is derived from a number of common sources. But most of these causes are more closely associated with a blurring of the old alignments and cleavages than a clarification of the new. In Japan, the pattern of electoral change is closely associated with increased occupational and residential mobility, a decline of community integration, and a weakening of union loyalties. In Italy, electoral change is linked to the decline of the agrarian sector and deconfessionalization, while in Germany, the change is related to the weakening of class cleavages and the rise of the new middle class. Each case may be describing dealigning rather than realigning forces. Naturally, the processes of dealignment and realignment are intertwined, in the sense that a new pattern of partisan attachments cannot take shape until the old pattern breaks down. The functional model, however, which is associated with the new dealignment literature, is suggesting that contemporary electorates in the advanced industrial societies may never be realigned (Dalton et al., this volume: Chap. 15). This issue cannot be resolved from our present historical vantage point. But it is important to clarify to the reader that our realignment section is not composed of cases inevitably moving toward a clearly defined enduring realignment, just as the succeeding dealignment section is not composed of cases necessarily moving toward a permanent dealignment. Neither the sources of change nor the end points of change differentiate these two sections. What distinguishes them is that in the dealignment cases the pressures for change have resulted in increasing volatility

98

Patterns of Realignment

in voting patterns. In the realignment cases, however, many of the same pressures for change have resulted in long-term cumulative electoral realignments with potentially great implications for the party compositions of the governing administrations. As noted previously, the American study presents the special case of an issue realignment, but nonetheless serves to illustrate the potential for a new conservative coalition in the United States based on a number of such key issues. To account for the differing realignment patterns among our four cases, we must now turn to an analysis of the varying responses to the common pressures for change detailed in the introductory chapter. ELITE RESPONSES AND ELECTORAL REALIGNMENTS

In general terms, the traditional partisan alignments in all four cases in this section reflect a Left-Right polarization of the party system based on economic issues. This institutionalized cleavage alignment has become increasingly less representative of new cleavages and new issues. In country after country, the emergence of a variety of New Left issues, from environmental and antinuclear concerns to women's liberation and gay rights, is counterpoised against a Fundamentalist, moralist backlash (Lipset, 1981b). This cluster of realigning value cleavages is discussed at some length in the introductory section of this volume, and in this section it is best described in the chapters on Germany and Japan. These new issues and pressures for change have provided political elites with new problems and opportunities. In all four cases, the long-term direction of change resulting from these pressures would appear to be toward a decline in the power of the traditionally dominant party.1 This similarity, in one sense, however, masks great variability in elite responses to the pressures for change and the types of political realignments that have occurred. In the German and American cases, the elite response was a parliamentary realignment. There was no change in the party alternatives offered to the electorate, nor change in the party system. Rather, a parliamentary realignment involves a change in the policy priorities and/or coalition patterns among the existing parties. In the German case, affluence and welfare state policies reduced the intensity of the struggle over the distribution of wealth and income among economic strata. Party elites responded to the declining salience of the industrial cleavages by reducing the policy distance between the two major parties. The creation of the CDU-SPD Grand Coalition enabled the Socialists to gain legitimacy as a responsible, capable, governing party. The declining ideological polarization between the parties also created new alignment opportunities for the new middle class. As class issues declined in salience and New Left social and quality of life issues increased in salience, the new 1

The recent short-term gains by the traditionally dominant parties in Germany and Japan do not negate this conclusion or the importance of the electoral changes that have taken place in these two countries over the last three decades.

Patterns of

Realignment

99

middle class gradually shifted its alignment from dominant support for the CDU to dominant support for the SPD, as Dalton documents. In the German case, then, the pressures for change largely have been contained by the existing parties, although their traditional bases of support are shifting in important ways. The Old and New Politics dimensions of political cleavage, however, are not entirely compatible. Indeed, the New Politics concerns for quality of life and environmental issues conflict with labor's traditional concerns for jobs and industrial development. One sign that these incompatibilities may breach the existing German party system is the recent emergence of a small environmental party, the Green party, which appears to have gained new support largely at the expense of the SPD.2 Thus in the German case, it is still not clear how the emerging postindustrial cleavages will be integrated into the party system. In the American case, party elites responded to the growing salience of racial issues by realigning their positions on civil rights issues. So long as civil rights issues focused on access to public facilities and were limited in their effect to the South, the Republicans took the procivil rights lead while the Democrats remained divided and embarrassed by the issue. While Carmines and Stimson demonstrate that elite realignment on racial issues began in the latefiftiesand early sixties, the civilrightsmovement clearly accelerated and solidified that realignment. Once racial issues became defined in terms of welfare spending, affirmative action, job programs, busing to achieve racial balance, community action programs, and the like, the Democratic party, as the traditional party of the underprivileged, became clearly identified with the procivil rights side. In contrast, the Republican party lined up against many civil rights measures, being philosophically opposed to federal government intervention in local government affairs, expanding welfare programs and bureaucracies, and increasing governmental controls on the private sector via such weapons as guidelines for awarding government contracts and revoking tax-exempt status. The parliamentary realignment on the civilrightsissue in the United States that Carmines and Stimson document, however, has a limited realignment potential. While some sectoral realignment clearly has occurred, particularly among those groups for which racial issues are highly salient, the proportion of the electorate involved has been small, at least in comparison to the New Deal realignment. Moreover, the gains and losses in support for each party that can be attributed directly to this single-issue dimension so far have been relatively balanced. 2

The SPD is not the only Old Left party to suffer from the emergence of New Left issues In Sweden, the Social Democrats were so staunchly wedded to Old Left progrowth policies that some of the bourgeois parties were able to capitalize on such New Left issues as the opposition to nuclear power Such incompatibilities in the leftist camp resulted in the Social Democrats' first defeat in forty-four years in the 1976 election (Lipset, 1981a 27)

100

Patterns of Realignment

If, however, racial issues are viewed as simply one aspect of a New LeftNew Right cleavage in the United States, the potential for a much broader and deeper sectoral realignment becomes apparent. Indeed, Seymour M. Lipset (1981b) argues that minority rights and ethnic movements of all kinds are rising in importance in Western Europe and North America. These movements are closely aligned with the New Left side of emerging postindustrial cleavages that are associated with a host of new social and quality of life issues. Lipset also argues that the new life-styles and new morality exemplified by these new social and political currents in American society have triggered a Fundamentalist, moralist backlash among elements who feel dislocated by what they are increasingly coming to view as a cosmopolitan, internationalist attack on American values and sovereignty. The New Right in America is composed of a collage of programs and single-issue groups, from the political evangelicals of the Moral Majority to the John Birchers to the tax-revolt groups. The integrating elements running through these New Right groups are the traditional American values of God, country, and rugged individualism. These values are reflected in a set of policies designed to attack moral decay in America, beef up America's defenses to meet the Communist threat, and oppose the growing size and interventions of the federal government. In this context, the racial issues of forced busing, federal intervention in private schools to stop racial discrimination, and federal support for programs aimed at enhancing the prominence of black culture fit neatly into this broad New Left-New Right value cleavage. An emerging postindustrial cleavage in America, therefore, potentially may induce broader sectoral realignments than have been documented to date. Survey evidence has shown that postindustrial values are found primarily among the more affluent, better-educated strata, while evangelicals and other traditionalists are clustered disproportionately among the less-educated, lessprivileged, lower-middle- and working-class strata. To the extent that the Democratic party takes a clear stand for New Left issues and the Republican party courts the New Right, the racial realignment that Carmines and Stimson discuss may be extended to a broader set of issues involving wider segments of the public (Miller and Levitin, 1976; Ladd with Hadley, 1975). Could this sectoral realignment, if it were indeed to occur, enable the Republicans to replace the Democrats as the majority party? Lipset argues that the New Right primarily represents strata that are declining in size, have little political influence or sophistication, display little constraint from one New Right issue to another, and are typically apolitical and difficult to mobilize even with the new electronic technology (Lipset, 1981b). Nevertheless, the elements for packaging a new Republican majority are present. Whether or not a Republican leadership emerges that can assemble these disparate New Right elements into a cohesive majority coalition with the Old Right remains to be seen.

Patterns of Realignment

101

In Japan and Italy, elites have responded differently to the same kinds of mass pressures for change. In both countries, the magnitude of the Left-Right polarization and mutual distrust among the early postwar party elites was much greater than in either the German or American cases. Moreover, in both countries elements of the agrarian-industrial cleavage have remained active. Thus, while there has been a decline in the salience of the Left-Right ideological issues that define these party systems, in relative terms, industrial cleavage issues remain more politically salient in Japan and Italy. Perhaps for these reasons a parliamentary realignment failed to occur in these two countries. The ideological gap between the Left and Right simply remained too great for the elites to bridge. In the Japanese case, rather than a parliamentary realignment, the elite response has been a party system realignment, that is, the number or composition of the contending parties has changed as the result of the emergence of new parties, the disappearance of old parties, or splinterings and mergers among old parties. Since the established major parties, the LDP on the Right and the Socialists on the Left, were unable to shift their ideological positions sufficiently to reflect the changes that were taking place within the electorate, the opportunity arose for new parties to move into the Center. By 1976, the new Center parties, which had begun to appear in the 1960 election, had risen to gain 22 percent of the nationwide vote in the House of Representatives election. An even better estimate of the Center's national strength may be gleaned from the 1977 upper house national constituency race in which the combined Center party vote reached 31 percent.3 Indeed, in many urban constituencies, the Center parties presently represent the largest bloc of voters. One of the reasons that these Japanese Center parties were able to emerge so quickly was that three of the four largest Center parties (the DSP, NLC, and SDF) were the creations of factional defections from the Socialists or the conservative party, with candidate support networks already in place. In the Italian case, the Japanese response was not available because, in Sartori's terms, the Center was already occupied (Sartori, 1976). In Italy, the conservative DC is defined as a Center party because of the existence of fascist and monarchist parties to its Right. Moreover, whereas the Japanese system in 1955 resembled a two-party system, the Italian system consistently has had eight to ten contending parties. In highly fragmented multiparty systems, it is very difficult for new parties to emerge with a program and appeal that is distinct from all the others. The Italian case approaches an example of no elite response, a model of frustrated change. Although ideological polarization has declined somewhat, the major option on the Left is the Communist party, which has remained an 3

This is so because only in the upper house national constituency race are the voters in every part of the country presented with a full slate of party candidates, which includes all the smaller minor parties. See Chap. 6.

102

Patterns of Realignment

unacceptable coalition partner for the DC. This has ruled out any fundamental parliamentary realignment, although there has been some opening to the Left. As Barnes points out, the DC now needs the Socialists' active or passive support to sustain a parliamentary majority. With so many parties already competing, party system realignment also has been ruled out as an effective response to a changing electorate. What the Italian party elites have lacked is a charismatic figure like de Gaulle, who could aggregate forces around him and significantly reduce the party space. One of the reasons why ecological realignment is more pronounced in the Japanese and Italian cases than sectoral realignment is that the continued polarization between the leading parties in these systems has prevented the realignment of classes or other social groups. The new middle class in Italy has not had the option of moving to the Left, as it has in Germany, because communism has proven to be too much of an ideological barrier. In the Japanese case, the realignment of the new middle class has been limited by the Marxist-Leninist rhetoric of the Socialist party. Undoubtedly, there has been some movement of the new middle class from both the Left and Right to the Center in Japan, but nothing as pronounced or as class distinct a phenomenon as found in Germany. While there has been more of an elite response to pressures for change in Japan than in Italy, in neither country have these pressures resulted in any significant regime change to date. In fact, in neither system is there a viable alternative to the continuation of conservative-led governments. In Japan, the opposition parties represent a fragmented hodgepodge of ideologies from communism to neoconservatism and are incapable of forming a united coalition alternative to continued LDP rule. The public's unwillingness to entrust the government to a divided and untried opposition-led coalition is an important reason for the recent gains in LDP support after hovering on the brink of losing its slim parliamentary majorities throughout most of the 1970s. On the other hand, the emergence of a number of Center parties has taken much of the trauma out of the possibility of the LDP's losing its majority, by supplying a number of potential coalition partners with moderate views. In contrast, the prospects that realignment in Italy ultimately will be reflected in a change in government are much dimmer. Between 1948 and 1979, the combined vote for the Communist party and the parties to its Left rose from 31 percent to 46 percent. Yet these parties remain locked out of power. Sartori's model of polarized pluralism predicts that support will flee from an immobilized Center to antisystem parties on the Right and Left until the Center, and indeed the political system itself, collapses (Sartori, 1976). That has not happened in the Italian case because all of the flight has been in a leftward direction. The movement from the Center to the Left largely has been compensated for by movement from the Right to the Center. Opponents

Patterns of Realignment

103

of communism have rushed to the support of the DC as the strength of the Left grew. In sum, fundamental realignment of the Japanese and Italian electorates should be slower in coming than in the German and American cases. In the former examples, elements of New Left movements and postindustrial social forces are present, but not to the degree as in Germany and America. Moreover, the Italian and Japanese conservative parties have adopted rather flexible and pragmatic orientations throughout the postwar period and may well be able to diffuse the forces of realignment sufficiently to retain power. For example, the conservative ruling party in Japan has taken steps to remove some New Left issues from public debate by, for instance, legislating some of the toughest antipollution measures in the world. In Japan and Italy, therefore, significant shifts in party fortunes have taken place, but these changes are largely the product of ecological realignments. In both cases, regime change has been frustrated and a more fundamental New Left-New Right sectoral realignment of party constituencies lies even further down the road.

4. The West German Party System between Two Ages RUSSELL J DALTON The Florida State University, Tallahassee, Florida

Change has been a basic characteristic of the West German party system throughout the postwar period. The political parties have altered their policy programs, voter support, and governing coalitions (Loewenberg, 1978; Baker et al., 1981; Merkl, 1980). But while change has been a distinctive feature of the party system, the patterns of change have not been consistent throughout the past thirty years. Rather, the electoral history of the Federal Republic is characterized by two distinct periods. The first twenty years of the West German party system might be described as the growth decades. Social, economic, and political development progressed at a substantial pace. This was the era of the Economic Miracle with rapidly increasing affluence and the development of advanced industrialism. Not only was the rate of change exceptional, but socioeconomic trends reinforced one another, producing a cumulative pressure on the party system. This led to decreasing ideological conflict among the established parties, as consensus developed on the political goals of the nation. Consequently, during this period partisan change was almost predictable in its pattern. The party system stabilized as the number of parties in the Bundestag declined from ten in 1949 to only three in 1961. The system also displayed the effects of advanced industrialism; economic concerns and class voting steadily declined during this period. The SPD benefited from these trends with a nearly constant 3 percent increase in votes per election from 1953 to 1972. Since the late sixties, West German partisan politics apparently has entered a new phase. This period of transition reflects two separate, but interrelated, Portions of this chapter were written while on a Fulbnght Fellowship at the Lehrstuhl fuer pohtische Wissenschaft und international vergleichende Sozialforschung, University of Mannheim. I would like to thank Paul Beck, Wilhelm Buerkhn, Scott Flanagan, Peter Gluchowski, Kai Hildebrandt, and Max Kaase for commenting on an earlier version of this chapter

West German Party System

105

processes. First, the party system is experiencing the accumulated effects of the growth decades. Old partisan alignments, primarily economic, have eroded in importance while new issues such as the environment, social concerns, and life-style issues are increasing in salience. Second, while economic and social change is continuing, it now moves at a significantly slower pace. The rate of economic growth in the 1970s, for example, was one-third the 1950s' rate of increase. In the early 1980s, the economy has been virtually stagnant. Continuing affluence maintains support for new noneconomic issue concerns. However, with the economic pie expanding at a slower rate, conflict between materialist and nonmaterialist goals is increasing (Hildebrandt and Dalton, 1978; von Beyme, 1978; Inglehart, 1981). The political events of the early 1980s underscore this theme of political conflict and partisan volatility. In the late 1970s, the Social Democrats became internally divided on issues such as environmental quality, social reform, nuclear energy, and national defense. Then, following their success in the 1980 elections, conflict between the SPD and FDP on economic and social policies began almost immediately. These tensions were the major reason behind the downfall of the SPD-FDP coalition in late 1982, and the emergence of a new CDU/CSU-FDP government in Bonn. National elections in March 1983 were marked by unusually high levels of vote switching and ticket splitting—as well as the emergence of a new environmental party, the Greens, in the Bundestag. As the Christian-Liberal government begins to formulate a new conservative program, the same inherent conflicts over economic, social, and environmental policies are reemerging within both the government and the opposition. In short, the old party system is weakening—perhaps ending—but the new system is not yet clearly evident. Thus contemporary partisan politics has become more complex and less predictable. Long-term electoral trends have been replaced by short-term oscillations from election to election. This chapter offers an overview of the major trends in partisan politics over the past three decades. We focus on the changes leading to the present situation, and discuss the future possibilities for the German party system. THE GROWTH DECADES: THE WEAKENING OF TRADITIONAL CLEAVAGES

The postwar period has been a time of tremendous socioeconomic change in the Federal Republic. Most important for the electoral system is the development of an advanced industrial society. The most dramatic changes came in the economic domain (Baker et al., 1981: Chap. 4; Voss, 1980). For example, average incomes roughly tripled during the past thirty years, even after adjusting for inflation. Consumer durables, leisure time, and other indicators of economic well-being displayed rapid growth during the postwar years. By almost any economic indicator, the German population is now at

106

Patterns of Realignment

least three or four times more affluent than at any time in their prewar history. The composition of the labor force also shifted from the traditional agricultural and industrial sectors to the tertiary sector (service and technology). During the 1970s, the Federal Republic passed Daniel Bell's postindustrial threshold, with over half of the labor force employed in salaried white-collar {Angestellte) or civil-service (Beamte) positions (Bell, 1973; Voss, 1980). Similarly, the educational training of the population expanded to reflect and support the growth of the tertiary sector. In the early 1950s, the ratio of citizens with only primary schooling (Volksschule) versus higher secondary schooling (Arbiter) was 5 : 1 . This ratio now stands at virtual parity. Parallel trends have occurred in the expansion of the mass media, urbanization, and transportation. These developments have important consequences for German politics. One visible effect of advanced industrialism is the erosion of traditional group and institutional networks (Pappi, 1976,1977). Social ties based on the community were replaced by the cosmopolitan independence of urban life. Religious networks, especially the Catholic infrastructure, weakened as church attendance steadily declined during the postwar decades. Expanding access to political information also reduced reliance on social networks as sources of political cues and political information. The mass media—especially television—became the most common source of political information. Steady increases in education and political interest provided the cognitive skills necessary to evaluate and utilize the information now available from the media. These social changes, in turn, reduced the importance of traditional social cleavages as a basis for partisan politics. With increasing prosperity, the competition for economic rewards becomes less pressing to the electorate and the parties (Baker et al., 1981: chap. 4; Thomas, 1980). Similarly, the secularization of society apparently lowers the political salience of religious conflicts. Otto Kirchheimer argued that the weakening of traditional social cleavages leads to the development of catch-all parties (Volksparteien) and the end of traditional ideological conflicts (Kirchheimer, 1966). The parties actively promoted the decline of traditional cleavages during the growth decades. Early CDU/CSU governments followed economic policies that partially were intended to moderate the class cleavage. The Christian Democrats also bridged the historic religious cleavage by uniting Catholic and Protestant voters in a single party. The CDU/CSU was probably the first major catch-all party in central Europe. The Social Democrats also took a dramatic step to lessen ideological tensions with the Bad Godesberg Program in 1959. In a single act, the SPD renounced its Marxist economic policies and generally moved to the Center on domestic and foreign policies. SPD leaders sought and achieved a rapprochement with the Catholic church in the early 1960s. The culmination of this partisan convergence was the Grand

West German Party System

107

Coalition of 1966 to 1969. The CDU/CSU and SPD shared control of the government, and an "era of good feelings" characterized partisan politics. The declining emphasis on traditional social cleavages did not, of course, dismantle totally the social bases of the party system. The traditional LeftRight social alignment remains the primary—although weakening—basis of West German partisan politics. The CDU/CSU is still seen as the party of farmers, the old middle class, and Catholics; and the party depends on institutional support from these groups (Pappi, 1976). Moreover, these social groups still are oriented toward the traditional cleavage structure and continue to support the CDU/CSU with their votes. Similarly, the Social Democrats altered the manner in which they pursue working-class interests, but not the party's basic commitment to the workers. Thus workers and union members continue their strong support of the SPD. However, the moderation of social and partisan conflict did stimulate significant changes in the West German electorate. The lessening of partisan conflicts freed a substantial sector of society from the traditional social bases of the party system. New partisan ties were established by individuals who were only marginally tied to traditional social networks or who were strongly influenced by new political concerns. The most noticeable example of this partisan change is the new middle class of white-collar employees and civil servants (Lipset, 1964; Baker et al., 1981: Chap. 7; Dahrendorf, 1967). The new middle class exemplifies the shift of the labor force toward service, government, and technology. The political importance of the new middle class derives from its ambiguous position in the traditional bourgeois/proletariat class structure. The separation of management from capital creates a social stratum that neither owns nor produces capital. The new middle class finds itself with a position in the social structure that places it between the working class and the old middle class of shopkeepers, business owners, and free professionals. Consequently, the new middle class is integrated neither into the unionized working-class milieu nor the old middle-class milieu. This opens the new middle class to new political appeals. This independence from traditional cleavages is shared by better-educated voters (Allardt, 1968). Expanding educational opportunities broadened access to a university education, introducing diversity into the university milieu. In addition, the role of universities in an advanced industrial society changes from an enclave of the elite to a scientific and information center. As a result, the socialization experience in the West German university environment was transformed from a conservative bastion to a marketplace of alternative ideologies. Rapidly expanding university enrollments also increased the political skills of the West German public. Better-educated voters are more able to deal with

108

Patterns of Realignment

the complexities of politics and have less need for group or social cues to guide their political decisions (Shively, 1979; Dalton, 1984). Hence the bettereducated are more likely to change their partisan preferences in a period when traditional political alignments are weakening. A third source of change in the political system is new voters. Social cleavages were of decreasing relevance to political socialization in the depolarized environment of the growth decades (Baker, 1978). Increasing affluence and economic security reduced the salience of traditional class issues of economic rewards. The young are more secular and, therefore, less tied to religious networks. Thus another large group of voters has only weak links to traditional social cleavages and, therefore, the opportunity to independently determine new partisan ties. A TRANSITION PERIOD: THE EMERGENCE OF THE NEW POLITICS

The declining salience of traditional, Old Politics, cleavage issues was not the only change occurring during the postwar years. The ebb of traditional issues was accompanied by, and stimulated by, increasing attention to a new set of political issues. These new societal goals generally focused on noneconomic, qualitative, reformist, and life-style concerns—the issues of the New Politics (Hildebrandt and Dalton, 1978; Inglehart, 1977). The New Politics represents a combination of two processes. At one level, the problems of advanced industrialism gave rise to a new set of issue interests. Concerns about environmental quality, the complexities of a technological society, and increased international interdependence entered the agenda of politics. The New Politics implied a shift from issues involving the quantity of life (Old Politics goals such as economic security and growth) toward more emphasis on the quality of life. At a more basic level, the New Politics represents a shift in the value priorities of the German public. Value priorities are socialized early in life in reaction to the conditions of the "formative environment" (Inglehart, 1977, 1981; Dalton, 1981). After several decades of unprecedented growth and security, the electorate begins to take these accomplishments for granted, and priorities shift toward the nonmaterial goals embodied in the New Politics. Some goals—freedom of expression, participation, and personal freedom— represent a broadening of traditional, European, middle-class liberalism to a wider popular base. In addition, attitudes toward social relations, life-styles, and societal goals also are changing. Thus advanced industrialism stimulates a new social, economic, and political agenda, represented by the development of New Politics values. The popular meaning of the Old Politics and New Politics can be determined from a series of issue-salience questions asked in each West German election

West German Party

109

System

study since 1961.1 A principal components analysis was used to determine the structure of the public's issue interests. The first issue component in each study taps general political interest with all issue questions loading positively (data not shown). The second component consistently yields two sets of issues that generally conform to the Old Politics-New Politics dichotomy (Table 4.1). The issues of the traditional cleavage system are clearly seen in the positive loadings of the Old Politics cluster. The economic issues underlying class conflict are represented by several items: price stability, unemployment, taxes, old-age security, and economic growth. Personal and national security con­ cerns are tapped by such issues as law and order, protection from terrorists, and security from Russian attack. In short, the electorate sees these as inter­ related issues, stressing themes of security and economic well-being. The issues with negative loadings in Table 4.1 tap New Politics concerns. New Politics orientations are reflected in the foreign policy issues of European integration, foreign aid, and cooperation with other nations, including both the Western and Eastern bloc. New Politics concerns also surface in the domestic issues of improving educational opportunities, environmental issues, and guaranteeing the rights of foreign workers. The "democratization" of society is tied to the New Politics through worker participation in company management (codetermination) and citizen action groups (Buergeriniativen). The distribution of Old Politics and New Politics orientations also can be tracked over time, using a forced choice question developed by Ronald Inglehart (Table 4.2). 2 Old Politics adherents still outnumber those exclusively committed to New Politics goals; this should be expected since the conditions fostering the New Politics are still evolving. However, during the past decade there has been a slow but steady growth in New Politics values—despite the 1

The data utilized m these analyses were made available by the ICPSR, the Zentralarchiv fuer empinsche Sozialforschung, and the Zentrum fuer Umfragen, Methoden und Analysen Neither the original collectors of the data nor these institutions bear any responsibility for the analyses or interpretation presented here The following election studies were used July Sept Nov Sept June June Feb 2

1961 1965 1969 1972 1976 1980 1983

(ICPSR 7098 (ICPSR 7103 (ICPSR 7098 (ICPSR 7102 (ICPSR 7513 (ZA 1059)

- ZA 055) • ZA 556) • ZA 427) • ZA 635) • ZA 823)

Ν = 1679 N=1411 N = 763 Ν = 2052 Ν = 2076 Ν =1620 N=1197

Those respondents who believe that out of four political goals presented to them, " maintaining freedom of speech" and "increasing participation in important political decisions" are the two most important goals are labeled "New Politics" values Those who stress "fighting against rising prices" and "maintaining order in the nation" above the other two goals are considered to have "Old Politics" value references See Inglehart (1977 chap 2)

110

Patterns of Realignment

TABLE 4 1 Principal Components Analysis of Issue Salience, 1961-1983 1961

1969

Price stability Old-age security Economy Reduced taxes Reunification National security

69 58 47 42 13 13

Price stability Old-age security Reduce taxes Political honesty

70 58 29 23

Political honesty Short hours Housing

05 02 06

National security Church tax Reunification

05 00 - 06

— 18 — 25 — 26 — 27 — 42 — 51

Education USA relations USSR relations Recognize GDR Codetermination Redefine states

14 21 23 34 36 56

USSR relations Education USA relations European unity Nuclear questions Foreign aid

1976 Price stability Law and order Job security

67 52 45

Job training

-09

European unity Ostpohtik Citizen action

-20 -39 -42

1972 Law and order Price stability Old-age security Terrorists Health National security West relations Housing Agriculture Tax redistribution Environment

02 00 00 — 03 — 05

China relations East relations Education Abortion Foreign worker rights

— — — —

1980 Price stability Old-age security Law and order Job security Energy

USA relations Education Environment Citizen action East relations GDR relations

51 44 39 38 36

- 10 - 12 - 18 -35 -48 -49

61 55 39 29 24 21

28 28 31 38 44

1983 Unemployment Reduce federal deficit Price stability Old-age security Law and order

USA relations East relations Environment Citizen influence

36 36 30 28 28

- 11 -42 -56 -58

SOURCE West German Election Studies NOTE Numerical entries are loadings on the second dimension of principal components analyses The first dimension was a general issue salience dimension in all studies Analyses are based on hstwise deletion of missing data

West German Party System

111

TABLE 4 2

The Distribution of Value Priorities, 1970-1982

Old Politics Mixed values New Politics Mean" (N)

1970

1973

1976

1977

1978

1979

1980

1981

1982

45.6 43.7 10.7 1.68 (1,865)

MX 49.5 8.5 1.67 (1,953)

41.5 47.4 11.1 1.70 (891)

41.0 52.1 6.9 1.66 (893)

37.2 51.8 10.9 1.74 (944)

36.8 51.7 11.5 1.75 (944)

40.4 49.3 10.3 1.69 (934)

45.3 46.2 8.5 1.55 (897)

36.9 49.8 13.3 1.76 (1,195)

SOURCE: Euro-Barometer Series.

" Mean scores are based on the following coding: (1) Old Politics, (2) Mixed, (3) New Politics. ups and downs in economic and political conditions during this period. Beginning in 1970, the number of Old Politics adherents exceeded New Politics supporters by 35 percent; by 1982, this difference narrowed to 24 percent. Moreover, a plurality of West Germans now stress at least a mix of New Politics and Old Politics goals. If comparable data were available back to the 1950s, the longitudinal trend toward New Politics values undoubtedly would be even more dramatic (Heidenheimer and Kommers, 1975: 119). The significance of New Politics concerns is increased because these orientations hold special interest for the social groups that have been freed from traditional social ties—the new middle class, the better-educated, and postwar generations. Group differences on the Old Politics/New Politics continuum follow the same pattern whether measured by issue orientations or values priorities. Although values reflect more enduring beliefs, we will examine the issue orientations of these social groups because the issue time series extends back to the 1961 election.3 Class differences in issue orientations are presented in the top panel of Table 4.3. The new middle class displays the most interest in New Politics issues in each survey from 1961 (mean = - . 5 3 ) through 1983 (-.13). 4 Indeed, the new middle class has provided the bulk of support for citizen initiatives, environmental action groups, and social reform movements. Conversely, both the old middle class and the working class are relatively more concerned with the Old Politics issues that underlie the traditional bourgeois3 Issue orientations are measured by component scores derived from the analyses in Table 4 1 Negative scores represent greater than average interest in the New Politics, while positive scores signify Old Politics orientations Principal component scores are computed using hstwise deletion of missing data, see Baker et al (1981. 313) 4 It should be pointed out that comparisons should focus only on group differences within each survey Longitudinal comparisons of group scores are not meaningful since component scores are standardized measures based on a changing set of issues For longitudinal studies of value differences, see Hildebrandt and Dalton (1978)

112

Patterns of

Realignment

proletariat cleavage Thus members of the new middle class are neither workers in white-collar occupations nor an extension of the traditional middle class The new middle class' issue concerns distinguish it from both of these other two social strata Interest in New Politics issues is related to educational level (Table 4 3) Respondents with a Gymnasium or university education (Abitur or more) place considerably more emphasis on New Politics goals by a wide margin Education is a relatively strong determinant of issue orientations, as indicated by the wide range of mean score differences and by the high eta correlations Concern for New Politics issues also shows significant generational differences (Table 4 3 ) 5 Older generations—Wilhelmme, Weimar, and the Third Reich—consistently display greater interest in Old Politics issues Among the young, however, the salience of New Politics issues is significantly stronger For example, the "protest generation," which reached adulthood since the mid-1960s, shows clear New Politics orientations from 1976 (mean = — 38) to 1 9 8 3 ( - 37) In sum, the emergence of New Politics concerns is politically significant for several reasons The New Politics bnngs a new set of issues to the agenda of German politics Even though issues such as inflation, unemployment, and national security naturally continue to attract widespread public attention, many New Politics issues are gaining interest over time For instance, from the 1950s to the 1970s, public opinion polls generally show increasing concern for New Politics issues as the most important problem facing the nation Even the revival of economic concerns as a result of the 1970s oil shocks and 1980s recession have not removed New Politics issues from the political agenda Indeed, one might argue that economic challenges have intensified the commitment to New Politics issues among its basic supporters (e g , Buerklin, 1981, 1982, Baker et al , 1981 136-142, 268-269) The New Politics is important because of the attention it has attracted from the new middle class, the better-educated, and the postwar generations For these groups, New Politics values and issue interests apparently fill part of the political void left by the decline of Old Politics cleavages THE CONTEMPORARY PARTY SPACE

The emergence of the New Politics may signify a basic change in the structure of the West German party system, especially if we are correct in linking these issues to the ongoing effects of advanced industrialism Signs 5 The classification of generations is based on the historical period during which the respondent reached fifteen years of age The histoncal periods were defined as follows

Wilhelmme Weimar Third Reich

1918 and before 1919-1932 1933-1945

FRG postwar FRG affluence FRG protest

1946-1955 1956-1965 1966 and later

West German Party System

113

TABLE 4.3 Issue Orientations by Social Groups, 1961-1983 1961

1969

1972

1976

1980

1983

Social class New middle class Old middle class Workers Farmers ETA

-.53 .13 .16 .30 .16

-.08 .18 -.02 .34 .07

-.25 .02 .18 .21 .12

-.11 .07 .15 .54 .12

-.05 .00 .10 .12 .10

-.13 .09 .17 .39 .16

Education Primary Primary/technical Secondary Abitur ETA

.31 -.12 -.58 -1.11 .18

.06 .04 -.06 -.69 .11

.37 .12 -.71 -.96 .27

.30 .12 -.17 -.84 .25

.16 .07 -.14 -.57 .23

.17 .15 -.14 -.60 .26

-.38 -.29 .10 .14 .28 .18 .19

-.32 -.01 .02 .11 .16

-.37 -.02 .11 .05 .26

Generation FRG protest FRG affluence FRG postwar Third Reich Weimar Wilhelmine ETA

a a

-.15 .06 .05 .00 .04

a

-.42 -.06 .02 .23 .09 .12

a

-.68 -.32 .18 .41 .49 .27

a

.17

a

.22

SOURCE: West German Election Studies. NOTE: Numerical entries are the mean scores on

principal component indices of issue orientations. Negative values denote New Politics orientations; positive values denote Old Politics orientations. " Insufficient numbers of the cohort in the study. of new political conflict already are evident. The nuclear energy debate unites management and labor unions in opposition to liberal, middle-class environmentalists. Farmers and students combine to fight economic development projects. Even sectors of the religious movement have split over social issues and defense policies. The simple dichotomy between Left and Right is no longer adequate to describe the patterns of political conflict. The contemporary political space now is better described by two dimensions. One represents the traditional Old Politics axis, which historically has structured the party system. This LeftRight dimension is a fusion of several cleavages. The Old Left consists of the union, secular, and urban coalition, which traditionally has been the basis of SPD support. The Old Right unites business, religious, as well as agrarian

114

Patterns of Realignment

interests behind the CDU/CSU. When political issues tap these concerns— for example, wage settlements, employment programs, and social security benefits—the lines of political conflict are clearly linked to the party system. The New Politics has introduced a crosscutting dimension into the German political space. This dimension represents the conflict between proponents of New Politics goals and those who see their Old Politics goals threatened by New Politics proposals. For example, conflicts between environmentalists and union/management coalitions or debates between supporters and opponents of the women's movement typify this cleavage. The supporters of the New Politics have been described as the New Left, the Alternatives, or the "New Liberals." This group initially became visible in favoring progressive action on New Politics issues during the late 1960s and early 1970s. Gradually, a conservative response to the New Politics also has emerged. This "New Conservative'Orientation has not spawned new political groups, but largely draws off of existing interest groups. It primarily represents a reaction to the New Liberals in defense of the established order. The New Conservatives are distinguished from the Old Right not by their conservative ideological position, but by their ideological interests. They are reacting to the New Politics agenda (Flanagan, 1982a: 429-432). The traditional Left-Right dimension is still the primary basis of partisan conflict, and is likely to remain so for the immediate future. However, the emergence of a New Politics dimension can stimulate substantial change in this situation. As interest in the New Politics grows, one effect may be to weaken the salience and partisan impact of traditional social cleavages. The growth of New Politics values also may increase the potential for partisan realignment along the New Liberal-New Conservative dimension. As these political conflicts spread, the parties and voters may reorient themselves on the basis of this new cleavage dimension. Although the New Politics has the potential to stimulate a realignment in the West German party system, this clearly has not yet occurred. Several factors are necessary to explain this situation. While New Politics issues are growing in importance, most West Germans remain oriented toward Old Politics goals (see Table 4.2). The long-term growth of New Politics values will increase the potential for large-scale partisan realignment along the New Liberal-New Conservative dimension—especially among those groups most attuned to these concerns. Still, we will argue that even the present potential for partisan realignment on the New Politics dimension has been lessened by the party system's reaction to these concerns. So far, the established parties have limited the impact of the New Politics by adopting ambiguous positions on this dimension—at least in comparison to their images on traditional social cleavages. On some specific issues, the party positions have been clear. During the 1970s, for example, the SPDFDP coalition actively pursued policies of social reform and Ostpolitik that

West German Party System

115

typify the New Politics, and the CDU/CSU generally opposed these policies. In overall terms, however, party images on this dimension have been blurred. Substantial division exists within the major parties concerning how to address the New Politics because this dimension cuts across traditional party lines. Several recent New Politics issues—environmental protection, nuclear energy, foreign worker rights, and foreign policy—developed schisms within one or more of the major parties. Thus the battles between New Conservatives and New Liberals often are occurring within the parties, rather than between the parties. No longitudinal data are available that would allow us to document how party images and voter opinions have changed on these dimensions since the 1950s. However, we can describe the structure of the contemporary party space and position major social groups in this space. These findings at least can suggest how the shifting balance of Old Politics and New Politics concerns is influencing the party system. We have collected data from several sources in an attempt to map contemporary party images on both the Old Politics and New Politics dimensions. The most direct approach explicitly asks respondents to place the parties on several dimensions of political cleavage. The overall Left-Right dimension provides an overview of party images (Fig. 4.1). The SPD and CDU/CSU are clearly differentiated on this dimension, and party polarization has remained fairly constant over the past three elections (Dalton and Hildebrandt, 1983). Only images of the FDP change dramatically. The public perceives the party's shift to the Right following the 1982 change in government. The presence of the ecologist party, the Greens, on the far Left in 1980 and 1983 implies that the general terms "Left" and "Right" now contain at least a mix of Old Politics and New Politics components. The Old Politics images of the parties can be tested more directly on two indicators of economic cleavage available in the 1976 and 1980 surveys (Fig. 4.2). In 1976, voters place the SPD clearly to the Left on a scale of government control of the economy (mean = 3.74). The CDU (8.11) and CSU (8.50) are perceived as clear conservatives. A similar pattern emerges in 1980 on a measure of union influence. The clarity of these party cues may even strengthen the economic cleavage, since SPD and CDU/CSU supporters see their respective parties as more extreme than their own self-placement.6 The Old Politics conflict between church and state provides an even more distinct basis of partisan cleavage. Despite its successful rapprochement with 6 The following table describes the self-placement of partisans and their perceptions of their respective parties on the economic dimension in 1976

Self-placement Party placement

SPD

FDP

CDU

CM

4 89 4 10

6 31 6 37

7 19 8 11

6 76 7 91

116

Patterns of

Realignment

FIGURE 4.1 Party Images on the Left-Right Dimension, 1976-1983 SS

14

LEFT 1976

^

-

RIGHT 1976

%£ 9 j a.

vi

u_

— a.

J_l RIGHT 1980

LEFT 1980

4-A

LEFT 1983

3 ο

j_a

RIGHT 1983

SOURCES: 1976-1983 West German Election Studies NOTE: Table entries are mean scores for the perceptions of party locations on the Left-Right scale

the Catholic church, the SPD is still identified with the separation of church and state. Conversely, the CDU and CSU are seen as supporting church involvement in politics. Furthermore, party images are again more sharply drawn than the self-images of party supporters.7 Conflict along the New Politics dimension is represented by the dichotomy between "personal freedom" and "social order." The parties are seen as offering less clear choices on this dimension in 1976. The SPD is only slightly to the Left of the ideal party position, and the CDU is only slightly to the Right. Indeed, the SPD is perceived to be more conservative than its own supporters. Unfortunately, this dimension was not repeated in the 1980 study. More detailed data on party images are available from another source. In 1974 and 1980, separate national samples evaluated the political parties and a set of sociopolitical groups with a "feeling thermometer." 8 Although the 7

The following table describes the self-placement of partisans and their perceptions of their respective parties on the religious dimension in 1976

Self-placement Party placement

SPD

FDP

CDU

CSU

3 17 2 75

3 36 3 74

5 73 7 48

6 17 8 41

These data are drawn from the 1974 and 1980 Political Action surveys I would like to thank

West German Party System

117

FIGURE 4.2 Party Images on Cleavage Dimensions, 1976-1980

SOURCES: 1976 and 1980 West German Election Studies. NOTE: Table entries are mean scores for the perceptions of party locations on each scale.

selection of groups is not ideal for our purposes, these data should allow us to determine the structuring of major social groups in relation to the Old Politics and New Politics dimensions. While issue images can fluctuate conMax Kaase and Hans Klingemann for graciously allowing me access to these data. For additional information on the 1974 survey, see Barnes, Kaase et al. (1979).

118

Patterns of

Realignment

siderably from year to year, these group alignments should reflect the more fundamental structure of the party system Factor analyses indicate that two dimensions underlie these group evaluations in 1974 and 1980 (Fig 4 3 ) 9 The horizontal axis reflects the traditional political alignments of the Old Politics At one pole, the electorate perceives the Old Left coalition of the Social Democrats and the unions Voters who support the Old Left are also favorably disposed toward the FDP, undoubtedly because of its coalition with the SPD at the time of these surveys At the other extreme lies the traditional Old Right coalition Conservatives are well disposed to the CDU, CSU, the Catholic church, and big business The second dimension of group evaluations identifies the groups differentiated along the New Politics dimension Most organized New Liberal interest groups tend to be small and ideologically extremist Indeed, one of the problems of the New Liberals has been the failure to develop a moderate rallying point that could attract widespread popular recognition and support Consequently, the best contemporary example of the New Liberals may be the environmental party, the Greens (Buerkhn, 1981) This party is perceived at the New Liberal pole m 1980 Citizens who are positive toward the New Liberals also support student protestors and revolutionaries, reflecting the antiestabhshment element of the New Politics (Inglehart, 1979a, forthcoming) In addition, between 1974 and 1980 the women's movement develops a clearer New Liberal image New Conservative groups have not developed as separate entities in the German system Cntics of the women's movement, proponents of nuclear power, an antitax movement, or groups opposed to the influx of foreign workers eventually may provide a New Conservative focal point At the present, New Conservative opinions are organized by existing interest groups that oppose New Liberal policies Figure 4 3 indicates that New Conservatives are favorable toward small business, civil servants, police, and the courts Although these groups do not necessarily define the New Conservative coalition, they do reflect the establishment's opposition to the New Liberals Moreover, the 1974 to 1980 trend suggests that the New Conservative coalition is broadening, the Catholic church and big business are seen as shifting toward this area of the political space These dimensional analyses also suggest that party images are not clearly perceived on the New Politics axis In 1974, the SPD is seen as leaning slightly toward the New Liberals, and the CDU/CSU toward the New Conservatives However, these differences are minor when compared to party 9 Factor analyses were computed using the procedures described in Barnes, Kaase et al (1979 581-583) The two-factor spaces then were rotated to maximum fit using the OSIRIS program COMPARE For additional analyses of these factor results, see (Inglehart, forthcoming Sankiaho, this volume Chap 3)

West German Party System

119

FIGURE 4 3 Old Politics and New Politics Dimensions, 1974-1980 New Conservatives

Φ Ο

< ! Police

1980 1974

Courts Λ ^

Civil Servants

Ο Small Business •

Big \ ^ Business \^

CDU

Church X

>o

Unions

Α, 7*·\ Ο

Womens \ Movementx Α

Λ— ·

OH CSIJ

JO ~— Quest Workers

Green Party

o^ Student Protestors

V | Groups Revo

New Liberals

SOURCES: 1974 and 1980 Political Action Studies

polarization on the Old Politics dimension. Moreover, while the 1974 to 1980 trends suggest that several social groups are shifting toward a clearer New Politics alignment, perceptions of the SPD and CDU/CSU remain fairly stable on this dimension; in fact, they actually decrease slightly. If data from the 1983 election were available, it is unlikely that the New Politics images of the major parties would differ significantly from 1980 (Dalton and Baker, forthcoming). In sum, the alignment of the established parties remains primarily and almost exclusively oriented to the Old Politics dimension of class and religion. Consequently, the potential for a partisan realignment along the emerging New Politics dimension has been limited. Only if the parties offer clear choices

120

Patterns of

Realignment

can issues and value orientations be translated into partisan preferences (Page and Brody, 1972, Carmines and Stimson, this volume: Chap. 5).10 The absence of change by the parties does not preclude other political consequences of the New Politics dimension. Perceptions of group alignments are changing in reaction to the emergence of New Politics concerns. This realignment of social cleavages can be seen even more clearly if specific groups of voters are located in this two-dimensional space. Factor scores were computed for the party space in Figure 4.3. Group means on both scores were used to position social groups in the party space (cf. Buerklin, 1981). Social class often is used to imply positions along the Old Politics dimension. Both the old middle class and the working class emphasize Old Politics issues—although with different issue positions. The old middle class presumably favors entrepreneurial control of the economy and limited government. The working class historically has preferred government involvement in the economy and extensive economic security programs. The new middle class has no predetermined position on these issues; its position apparently is affected by the unique context of each national political system (Kerr, forthcoming). Different class orderings are expected on the New Politics dimension. The new middle class is more interested in the issues of the New Politics and should provide the bulk of the New Liberals. On the New Politics dimension, both the old middle class and workers should tend toward the New Conservative position, since they favor the goals of the present Old Politics system. Class positions in the contemporary political space are presented in Figure 4.4. Class differences are greatest on the Old Politics dimension. Farmers and the old middle class tend toward the Right on the Old Politics dimension, and display slightly conservative tendencies on the New Politics dimension. These views place both social strata closest to the CDU/CSU, which they disproportionately support with their votes. The working class clearly tends toward the Left of the Old Politics continuum, while displaying slightly conservative views on the New Politics. The working class is thus closer to the SPD in the overall space, and gives most of its votes to the Social Democrats. The new middle class positions itself near the Center of the space, with slight leftist tendencies on both the Old Politics and New Politics dimensions. Moreover, between 1974 and 1980 the new middle class moved toward a liberal position on the New Politics. We believe that the weakening of Old 10 A 1979 survey from the Konrad Adenauer Stiftung asked respondents to evaluate the issue positions of the CDU and SPD on a 7-pomt agree-disagree scale The parties are clearly differentiated on New Politics issues such as codetermination (2 21 difference) or abortion (2 05), but this is so because these issues also tap traditional economic and religious cleavages On a more distinct New Politics issue, environmental protection, the public sees virtually no difference (0.14) between the major parties.

121

West German Party System

FIGURE 4.4 The Placement of Social Classes in the Party Space, 1974-1980 New Conservatives



1980

Ο 1974

Farmers Old Middle

""J

^

New Middle

New Liberals

SOURCES: 1974 and 1980 Political Action Studies.

Politics cleavages has led the new middle class to break with the traditional middle-class position and adopt a centrist location on this dimension. In addition, the new middle class' interest in New Politics concerns undoubtedly contributes to this more moderate position on the Old Politics. However, the new middle class has yet to establish a clear position on the New Politics dimension. A clearer polarization on the New Politics dimension should emerge for education groups. Theorists have argued that the advanced industrial cleavage will match the technological elite against the technologically superfluous, the information-rich against the information-poor (Allardt, 1968; Bell, 1973). Social class is a crude measure of this cleavage since considerable variation exists within each stratum. Education, on the other hand, is a more reliable indicator of the technology-information cleavage. The highly educated are

122

Patterns of Realignment

likely to staff the ranks of the information-rich elite, while the less-educated are likely to be information-poor. Figure 4.5 presents the positions of education groups in the contemporary party space. There are only slight differences among education strata on the Old Politics dimension; these groups, instead, are aligned along the New Politics dimension (1974 eta = .19; 1980 eta = .21). The highly educated express strong New Liberal orientations, while the less-educated have New Conservative tendencies. Thus social status is an important determinant of New Politics orientations, but primarily through education rather than oc­ cupation. These class and education patterns underscore the complexity of present FIGURE 4 5 The Placement of Educational Groups in the Party Space, 1974-1980 New Conservatives

• 1980 Ο 1974

Primary

-7 >

New Liberals

SOURCES. 1974 and 1980 Political Action Studies

West German Party System

123

party coalitions. For instance, the SPD's bases of support—workers and the new middle class—are interested in different types of issues (Table 4.3). To appeal to the working class in terms of their Old Politics interests may alienate some better-educated New Liberals. Conversely, strong appeals to the New Politics interests of the better-educated may alienate sectors of the working class. Thus increased intraparty strains may be inevitable, despite the parties' hesitancy to confront the New Politics cleavage. The historical trends in the German party system also can be seen in the generational differences along the Old Politics and New Politics dimensions. Figure 4.6 presents the location of our six generations in the party space. In both 1974 and 1980, the generations are aligned along a diagonal axis comFIGURE 4 6 The Placement of Generations in the Party Space, 1974-1980 New Conservatives

New Liberals

SOVRCES 1974 and 1980 Political action Studies

124

Patterns of Realignment

bining the Old Politics and New Politics cleavages. The Wilhelmine cohort, for example, is conservative on both the Old Politics and New Politics dimensions. Conversely, the most recent Federal Republic generation (Protest) is liberal on both dimensions. Thus the postwar generations have shifted toward the Left, and generational change is not limited to either cleavage dimension. Overall, however, generational differences are greater on the New Politics dimension (1980 eta = .33) than on the Old Politics (eta = .20). The full complexity of generational change emerges when social status and generation are combined. Theory suggests an interaction between these two social characteristics. Among prewar generations concerned with the Old Politics, higher status groups should support the Old Right and lower status groups should support the Old Left. A much different pattern is expected among the younger generations. First, the young should be differentiated primarily along the New Politics dimension. Second, higher status youths should support New Liberal policies, while lower status groups should tend toward the New Conservatives (Inglehart, 1977; Hildebrandt and Dalton, 1978). These hypotheses can be tested by examining social status differences across generations. With six generations and several social status levels, the number of cases is quickly spread thin. Therefore, data were merged for the 1974 and 1980 studies, and generation was collapsed into two groups: pre-World War II and postwar generations. In addition, the two highest educational groups were combined because of their small size. The intergenerational pattern of educational differences provides evidence of an historic realignment of social status groups (Fig. 4.7). Older Germans are differentiated along the Old Politics dimension. The better-educated (those who have secondary education or more) are slightly more conservative, while the less-educated are more leftist. Among the young, there has been a shift to the Left by all education groups. However, the line of cleavage now is based on the New Politics dimension. Among the young, the better-educated are New Liberals, while the lesser-educated are New Conservatives—the traditional lines of Left-Right cleavage are reversed! In sum, these changes in social status and generational alignments are shifting the axis of ideological conflict—if not yet party conflict. In earlier times, "Left" and "Right" were synonymous with our "Old Left" and "Old Right" terms. While Left-Right self-placement is still strongly related to the Old Politics cleavage (1974, r = .52; 1980, r = .51), these beliefs are also related to the New Politics (1974, r = .26; 1980, r = .21)." West German politics now contains a mix of Old Politics and New Politics concerns (Buerklin, 1982). 1

' Ideological orientations were measured by a ten-point scale running from "Left'' to' 'Right '' For additional discussion of this ideological scale, see Khngemann (1979, 1982)

West German Party System

FIGURE 4 7 The Placement of Education Groups in the Party Space for Prewar and Postwar Generations, 1974-1980 New Conservatives

Ο

Prewar Generations

φ

Federal Republic

y°P Primary + w

Secondary

^

New Liberals

SOURCES. 1974 and 1980 Political Action Studies

THE EVIDENCE OF PARTISAN CHANGE

The traditional ideological conflicts along class and religious lines still provide the major basis of partisan alignment in West Germany. However, for large sectors of the electorate, these cleavages apparently are declining in importance. At the same time, value prionties and political interests are shifting toward New Politics issues. Our concern, like others in this volume, is whether these trends have produced a dealignment or realignment in party coalitions. Each of these possibilities will be considered. One consequence of recent political trends might be a dealignment in West German partisan attachments (Inglehart and Hochstein, 1972; Beck, this vol-

126

Patterns of Realignment

ume: Chap. 8). Social groups at the forefront of advanced industrialism—the new middle class, the better-educated, and the young—have been freed from the partisan ties of the Old Politics dimension, but the parties have not replaced this with a clear New Politics framework. In addition, these groups possess the political skills and resources that facilitate dealignment (Dalton, 1984). Thus we might expect to find a weakening of party ties, especially among New Politics groups. The dealignment hypothesis can be tested by tracking the strength of partisanship over time. German researchers disagree on whether such partisanship measures are based on party identification or social cues (Baker et al., 1981; Gluchowski, 1978, 1983). However, the dealignment thesis simply implies a general weakening of partisanship regardless of the source. Surveys over the past decade fail to detect evidence of dealignment among German voters (Table 4.4). From 1972 until 1983, between 40 and 50 percent of the electorate considered themselves to be strong or very strong partisans. Similarly, the number of independents is relatively constant over time. More extensive longitudinal analyses of an alternative measure of partisanship actually indicate a general increase in the strength of partisanship since 1961 (Baker et al., 1981: Chap. 8). Thus there is little evidence that changes in the German party system are prompting dealignment in partisanship. Partisan realignment is another possible consequence of recent political trends. The previous analyses indicated that social groups are changing their locations in the party space. Postwar generations have shifted to the Left on the Old Politics and New Politics dimensions, and social status alignments have experienced similar changes. Thus realignment should be visible if the voting behavior of these social groups is traced over time. TABLE 4.4 The Strength of Partisanship, 1972-1983 1972 Sept.

Very strong Strong Weak Independent Mean

1980

1976

1983

Oct.

Dec.

June

Sept.

Nov.

June

Sept.

Feb.

18 39 20

11 34 34

11 37 35

14 37 32

14 31 30

17 36 25 22

10 33 37 20

17% 32 21 30

23

17 40 20 23

100%

100%

100%

21 100%

17 100%

17 100%

25 100%

100%

100%

2.53

2.53

2.35

2.42

2.50

2.34

2.48

2.33

2.40

SOURCE: West German NOTE: Mean scores are

Strong, (4) Very strong.

Election Studies. based on the following coding: (1) Independent, (2) Weak, (3)

West German Party System

127

The voting behavior of social classes is presented in Table 4.5. A longterm shift in party coalitions is indicated by the steady decline in middleclass/working-class differences over the past three decades. In 1957, the Alford index of class voting was thirty-seven points; since 1969, it has averaged less than half this level. Closer inspection of the data shows that the decline in class voting primarily is due to change by the new middle class. This social stratum moved from the CDU/CSU camp in the 1950s to the Social Democrats in the 1970s and 1980s. Although the voting patterns of the working class and old middle class also have shifted slightly over time, change was greatest among the new middle class—the group most highly attuned to the New Politics. Moreover, this pattern continues even through the 1983 election when the CDU/CSU returned to power. Table 4.5 also tracks the voting pattern of educational strata during the past thirty years. The SPD garnered disproportionate support from the lessereducated in the 1950s. However, educational differences have narrowed considerably over time, largely due to an SPD shift among the better-educated. In fact, among the youngest voters these patterns are actually reversed, with higher leftist voting among better-educated youths (data not shown). Generational differences also have been a major source of partisan change. The early postwar elections witnessed only slight generational differences in voting behavior—with young and old alike supporting the Christian Democrats (Table 4.6). Since 1969, generational differences have increased. Although TABLE 4.5 Class Voting Differences, 1953-1983 July June July Sept. Sept. Sept. June June Feb. 1953 1957 1961 1965 1969 1972 1976 1980 1983 Social class Working class Middle class New middle Old middle Alford index8

59 29 34 19 30

61 24 26 20 37

57 29 31 24 28

58 32 38 15 26

59 47 55 25 12

70 53 57 38 17

58 42 47 23 16

67 51 54 40 16

59 49 52 29 10

Education Primary school Secondary Difference

44 19 25

48 23 25

42 29 13

45 38 7

51 49 2

60 51 9

50 46 4

60 52 8

52 50 2

SOURCE: West NOTE: Entries

German Election Studies. are the SPD share of the major two-party vote intention. a The Alford index is the differences in SPD vote between the middle class and working class.

Patterns of Realignment

128

most age groups have moved somewhat toward the SPD, younger voters accentuate this trend and now lean strongly toward the SPD. In the 1980 and 1983 elections, many young people went to further extremes, providing the bulk of support for the ecologist party, the Greens (Buerklin, 1981; Dalton and Baker, forthcoming). Thus again, change is concentrated among voters most interested in New Politics concerns. TABLE 4.6

Generational Voting Differences, 1953-1983 July July Sept. Sept. Sept. June June Feb. 1953 1961 1965 1969 1972 1976 1980 1983 FRG protest FRG affluence FRG postwar Third Reich Weimar Wilhelmine Range (High-Low)

a

a

a

a

46 46 44 40 6

48 43 39 32 16

a

46 47 44 44 36 11

a

57 49 52 50 46 11

71 63 59 61 57 44 27

54 55 50 45 47 32 23

61 49 58 48 47

68 58 62 52 49 a

19

i

14

SOURCE: West NOTE: Entries

German Election Studies. are the SPD share of the major two-party vote intention. a No members or insufficient numbers of the cohort in the study. These findings argue that the German party system has experienced a realignment during the postwar period. The 1969 election was a significant step in the process, since control of the government switched from the Christian Democrats to the SPD and FDP. However, shifts in the social bases of partisan support follow a long-term process of change—a secular realignment based on the new middle class and the young. The evidence for long-term realignment is considerably strengthened by the persistence of these voting trends even after the SPD lost power in 1982. Moreover, by the late 1970s these voting trends also were reflected in the changing composition of SPD activists (Feist et al., 1978). Realignment theory is based largely on the American experience, although an American realignment has not occurred since the development of survey research. These data thus offer a new, and empirically based, perspective on the realignment process, which can test some elements of past theorizing. Recent research argues that realignments primarily come from the mobilization of uncommitted voters, rather than conversion of voters with previous party commitments (Andersen, 1979; Carmines and Stimson, this volume: Chap. 5; Clubb et al., 1980). The West German case initially seems to undercut this position. Voter mobilization always has been high in the Federal Republic (turnout averages nearly 90 percent); thus the number of uninvolved adult

West German Party System

129

voters is limited However, a closer inspection of the class and generation patterns of realignment supports the mobilization hypothesis Partisan realignment has come primarily from young new voters Table 4 7 indicates that the long-term decline in class voting is almost exclusively due to change among postwar generations Class voting differences are large and relatively constant over time for the Wilhelmine, Weimar, and Third Reich generations Among Federal Republic generations, the Alford class indices begin at an initially low level, and decline dunng the 1970s l2 Thus the realignment of West German partisanship comes largely from the SPD's success in mobilizing young new voters TABLE 4 7

Class Voting Indices by Generation, 1953-1983"

FRG protest FRG affluence FRG postwar Third Reich Weimar Wilhelmine

1953

1961

1965

1969

1972

1976

1980

1983

— — 24 25 38

— 25 34 27 26

18 22 24 24 28

10 16 13 15 26

2 14 20 27 34

12 7 12 21 16 39

6 6 14 19 21

-14 10 12 19 36

a

a

SOURCE West German Election Studies NOTE Multiple surveys were combined when

available at each time point Entries are the Alford indices of class voting, measured as the middle class/working class difference in SPD vote a No members or insufficient numbers of the cohort in the study The West German case highlights another necessary revision in realignment theory A shift in party control of the government is one of the defining characteristics of Amencan realignments (Clubb et al , 1980 Chap 4, Campbell, 1966) A change in government control presumably offers the new incumbents an opportunity to implement the policy proposals that brought them to power, and to institutionalize their support However, the German experience indicates a weakness in this part of Amencan realignment theory Electoral realignment is decisive in the formation of Amencan governments because only two parties are competing In a parliamentary multiparty system, the relationship between election results and the formation of the government is often tenuous Bargaining among the political parties, rather than the election itself, may be decisive in determining government control For example, the 1969 change in government, which facilitated this realignment, is directly 12 Additional data (not shown) indicate that declining class voting among postwar generations is primarily due to an SPD shift by the middle class Thus realignment is clearly concentrated among the New Politics stronghold—the young middle class (see Baker et al , 1981 175-180)

130

Patterns of Realignment

due to a decision by FDP leaders to reverse the party's long-standing conservative image and ally itself with the Social Democrats. This coalition initially was not favored by the electorate as a whole or by SPD voters; the combined SPD and FDP vote had actually dropped between the 1965 and 1969 elections (Norpoth, 1980). Similarly, the 1982 change in government resulted from the FDP's decision to ally itself with the Christian Democrats, a coalition change that evoked sharp public criticism of the FDP. In a multiparty parliamentary system, a change in government control may arise from either mass or elite actions. CONCLUSION

Western European political systems have experienced a decade of exceptional political, social, and economic change in the 1970s. These changes have led to discussions of the transformation of European party systems. We have found that the basic structure of the West German party system has been remarkably unaffected by the events of the past decade. The pattern of contemporary party cleavages is not much different from what might have been found in the 1950s. The parties historically have been organized along the Old Politics lines of class and religion. Even if these cleavages have become less central to party programs, the group ties and institutional structure of the parties may perpetuate party images. Parties are, after all, still turning to the same interest groups and associations for the core of their support. Declining party emphasis on Old Politics issues in the 1950s and 1960s was not designed to throw out old supporters, but to attract new support. And unless there is a dramatic renunciation of group ties—which has not occurred in West Germany—traditional party images may linger on. Thus the public still perceives clear party differences on the Old Politics. While significant party differences may exist on some New Politics issues, only weak partisan polarization appears on the overall New Politics dimension. This finding suggests that just as old partisan images are easily perpetuated, new cleavages are difficult to integrate into a party system. Parties are naturally hesitant to take clear stances on new dimensions of conflict until the costs and benefits are clear, and the electoral consequences of the New Politics are not yet clear to see. First, the number of citizens exclusively concerned with New Politics issues is still quite small; thus their present electoral weight is limited. Second, the eventual distribution of voters along the New Politics dimension also is uncertain. The absence of group frameworks to organize New Politics concerns, especially on the Right, makes it difficult to predict the size and composition of the New Liberals and New Conservatives. For example, in West Germany the new middle class has developed a leftist orientation, while in Italy this class seems oriented toward the Right (Kerr,

West German Party System

131

forthcoming). In addition, the absence of a social group framework makes it difficult for the parties to institutionalize their ties to the New Left or New Right. Third, even if New Politics groups develop, to be integrated into the party system these groups must surmount the political hurdles facing any new political movement. Thus the potential exists for considerable partisan change based on New Politics issues, but this potential has not yet developed fully. Electoral politics has not, however, been totally without movement. Even though the party system has been relatively stable, significant shifts have occurred in group positions along these cleavage dimensions. Generation and education groups are now politically aligned along the New Politics dimension. However, the potential for partisan realignment on the New Politics dimension has been limited by the lack of clear party positions. The West German political space is therefore based on both the Old Politics and New Politics dimensions. However, the party space may be almost unidimensional since the established parties have minimized their differences on the New Politics dimension (Norpoth, 1979; Dalton and Hildebrandt, 1983). Shifting group positions provide the party system with a temporary outlet for new social and political forces, while leaving the system basically intact. Compared to the difficulty and uncertainty of changing the party system itself, group realignment is a more likely short-term response to pressure for change. But pressures for political change continue to exist, because New Politics views are still underrepresented by government policies. One cause and symptom of these unresolved political tensions are the factional divisions within the major parties. Originally organized along Old Politics lines, the parties often find themselves divided on New Politics issues. When the SPD was in power, the party was visibly split by the conflicting demands of its various supporters: unions, environmentalists, the new middle class, and the Young Socialists (Jusos). As Risto Sankiaho suggests (this volume: Chap. 3), the Social Democrats' commitment to social change was limited by the responsibilities of governing; and supporters were disappointed by the party's performance. Even in opposition the SPD may face these same policy tensions, as its reformist credentials are challenged by the Greens. Similar tensions between Old Politics and New Politics goals exist within the Christian Democratic Union, even if they have been less visible while the party has been in opposition. Now that the CDU/CSU has regained control of the federal government, these latent intraparty conflicts should escalate. These unresolved New Politics tensions have manifested themselves outside of the established parties. The growth of citizen action groups (Buergerinitativen), for example, reflects the new political style of advanced industrial societies as well as the existence of unrepresented political interests (Barnes, Kaase et al., 1979). Very often these unrepresented interests touch on New Politics concerns—environmental quality, life-style issues, social rights, and

132

Patterns of Realignment

women's liberation—because the parties have been unresponsive to these issues. Consequently, unconventional political behavior is more strongly related to New Politics views than to the Old Politics (Inglehart, 1979b). Another symptom of present party tensions can be seen in the founding of an ecological party, the Greens, in spring 1980. The party represents a heterogeneous collection of New Left and antiestablishment voters. The Greens differ so markedly from the established parties that one Green leader described her party as the "anti-party party." The Greens have challenged the established parties on a range of New Politics issues: nuclear energy, West Germany's participation in the arms race, environmental protection, and democratization of society. By 1982, the party had won seats in six state legislatures, followed by Bundestag representation after the 1983 federal elections. Now that the Greens are in the Bundestag, these issues will become a more visible component of national politics. This party may provide the stimulus for a broader New Politics realignment of the party system. More to the point, however, Green voters felt it was necessary to move outside of the established party system to have their views represented. In sum, we are witnessing the political strains and tensions of building pressure for partisan change. Although this pressure emanates from the public, the decisions made by political elites may be more important in determining how these pressures are eventually resolved. Thus, even though the eventual result of these pressures cannot be predicted from data on public attitudes, several options seem possible. One possibility is that the interest in the New Politics fades before producing a basic change in the party system. We do not believe that the New Politics was merely a fad of the sixties and seventies; its development reflects fundamental—and ongoing—socioeconomic transformations in West German politics. The growing economic problems of advanced industrial societies might threaten the social and economic conditions nurturing the New Politics. If the politics of affluence regresses to the politics of scarcity, the New Politics pressure for change indeed may slacken. However, a more likely scenario is that the West German economy will continue to develop, but at a slower pace than during the growth decades. The New Politics therefore should remain on the political agenda since affluence will maintain the New Politics base. At the same time, slower growth rates may exacerbate the conflicts between the Old Politics and New Politics cleavage dimensions. The West German party system will face increasing pressure to adjust to the New Politics cleavage. These forces of change may take many forms and many outlets. New parties such as the Greens may represent the New Politics dimension. Or, the FDP in its role as coalition partner may respond to the New Politics cleavage. The most basic change would have the SPD and the CDU/CSU adopting clearer positions on the New Politics dimension. The major parties can respond in varying degrees to the New Politics, moving the

West German Party System

133

axis of partisan cleavage closer to the diagonal axis of ideological cleavage. Such a shift by either or both of the major parties would inevitably change the composition of existing Left and Right coalitions, as well as undercut the support for new parties. To some extent, the SPD explored such a strategy to weaken support for the Greens in the 1983 election. The West German party system is now between two ages. The old party order is breaking down, but the shape of the new order is not evident. Considerable partisan change has occurred, but even more lies ahead.

5. The Dynamics of Issue Evolution: The United States EDWARD G CARMINES Indiana University, Bloomington, Indiana and JAMES A STIMSON University of Houston, Houston, Texas

The processes of partisan change are not well understood. Despite a number of rich historical accounts, there exists no satisfactory theory of change in the party system. The most widely accepted explanation is that provided by "critical election" realignment theory (Burnham, 1970; Clubb et al., 1980; Key, 1955; Sundquist, 1973). According to this perspective, new issues that cut across the normal development of the party system are the stimuli of longterm partisan transformations. These intense crosscutting issues, often arising from significant social and economic dislocations, lead not only to the dissolution of the existing partisan alignment, but to the formation of a new one built around the emerging issue cleavage. Thus the new majority party must formulate policy within the context of a fundamentally changed political environment as well as in response to a sharply altered issue agenda. While the critical election realignment perspective has contributed to an increased understanding of partisan transformations, it has a number of significant shortcomings. At best, it provides an incomplete account of the dynamics of partisan change. Why do some political events lead to permanent partisan alterations while others have only temporary effects? Do different partisan transformations follow a similar path, or do they differ greatly depending on the particular political issue and circumstances in each instance? Is partisan change abrupt or gradual in character? The aim of this chapter is to formulate and assess alternative models of partisan change. Stated more fully, our purpose is threefold: (1) to suggest Research for this chapter was supported by the National Science Foundation under Grant SOC7907543

Issue Evolution: The United States

135

competing processes by which change can occur in the party structure; (2) to show that these processes can be represented by appropriate time series intervention models; and (3) to examine these models with regard to the partisan evolution of particular issues in the American political system. Our ultimate objective is to develop a general theory of issue evolution, one that is limited neither to the American context nor to any specific political issue. Our conceptualization of issue evolution differs substantially from the conventional notion of realignment. The term "realignment" itself has been defined in a number of different and sometimes inconsistent ways: Burnham (1970, 1974) and more recently Clubb et al. (1980) often take it to mean a change in party pluralities. According to this perspective, the 1932 presidential election is referred to as a realigning election because it marked the termination of the Republican party as the majority party and inaugurated a period of Democratic party rule. The popular support accorded the parties shifted in a dramatic and permanent direction. This picture has a pleasing neatness, which must account for its staying power. For it is nowhere supported by empirical analyses that probe beneath the question of which party got more votes and took control of government. Focus on the alignment aspect of realignment is both an older (Key, 1955; Lipset and Rokkan, 1967) and newer (Petrocik, 1981; Beck, 1982) approach to the study of realignment. Turning away from election outcomes as a criterion of party system change, this emphasis looks to changes in the coalitional structure of the parties. Because election results may change suddenly but coalitions do not, the shift in criterion leads also to a lengthening time perspective. Those who look to shifting coalitions and new lines of cleavage typically expect to see change manifested over a period of years and decades, not overnight like the critical election realignment perspective. We draw from this ' 'coalitions and cleavages'' approach, but emphasize the cleavages, which we call "issue evolutions." The partisan evolution of a political issue occurs when there has been an altered division or heightened polarization along an issue cleavage that is manifested among partisan supporters. Issue evolutions thus lead to an alteration in the issue profile of the party coalitions. This process may result in a new majority party. Indeed, we believe that each of the classic partisan realignments in American political history can be understood within an issue evolution framework. But issue evolutions need not have dramatic impacts. They may realign the issue basis of the party system without leading to a major alteration in party pluralities. Thus issue evolutions can occur within the context of a relatively stable electoral system. The fundamental point is that while the evolution of political issues can result in traditional partisan realignments, it is more likely to lead to more complex and less spectacular transformations. So we shall use the term "realignment" to refer to issue

136

Patterns of Realignment

evolutions that lead to a rearrangement of the issue coalitions underlying the party system—a process that is necessary but not sufficient to bnng about a change in partisan majorities MODELS OF PARTISAN CHANGE

Imagine for the moment that we can measure polanzation between parties over a particular issue, candidate, or event A vanety of techniques might be employed, the one we shall later use is the simple difference between the aggregate (mean) issue attitudes of the party groups (1 e , party identifiers) over time Polanzation might or might not exist before the realignment process 1 begins, but it is the increase in polanzation that is important Thus we can presume without loss of generality a zero base line Models 1 through 4 in Figure 5 1 track four hypothetical responses (polanzation) of a party system to external shocks FIGURE 5 1 Four Hypothetical Models of Issue Realignment IMPULSE DECAY

CRITICAL ELECTION

SECULAR REALIGNMENT

DYNAMIC GROWTH

'

1

1

1

Γ

20

40

60

80

TIME

{IN HYPOTHETICAL YEARS)

Model 1 in Figure 5 1 is an illustration of an impulse-decay pattern of partisan change. The change can be dramatic Its effects, however, are tran­ sitory As soon as the temporary stimulus is removed, the system rapidly returns to its preexisting level of stability It is the temporary nature of this change rather than its abruptness that is its most distinctive charactenstic Many different kinds of political phenomena, especially presidential candi­ dates, presumably have this type of dramatic but momentary effect on the party system John Kennedy's nomination for the presidency by the Democrats in 1960, for example, led to a polanzation of the electorate on the religious 1 There are conditions under which one would want to model decreasing polarization as well The decay of old alignments should produce decaying polarization, as we will see later

Issue Evolution: The United States

137

issue (Converse et al., 1961). Catholics became more heavily Democratic, just as Protestants became more Republican. But this increased partisan differentiation between religious groups was transient. It is convenient to digress to our underlying conception of causality in partisan change to explain why this must be the case. We assume that the American party system exists in a state of equilibrium.2 Because it has powerful implications, the equilibrium assumption is not to be taken lightly. Equilibriums do not exist accidentally. Systems acquire that property when they have built-in causal forces that tend to restore them after deviation (Stokes and Iversen, 1966). We shall not speculate here—in part, to be candid, because we do not know—which particular mechanisms produce equilibrium, but one is hard pressed to account for the lengthy duration of any party system without recourse to the implicit causation of equilibrium. It follows naturally that any temporary disturbance must have temporary effects. The disturbance can drive the system away from equilibrium, but it cannot keep it there after its causal force is no longer present. There is an important corollary: it takes two changes to alter the equilibrium level, one to move the system from equilibrium to a new level and a second permanent redefinition of the grounds of party cleavage to keep it at a new level after the initial stimulus is no longer present. To the extent that most political issues affect partisan change at all, they are likely to be of the impulse-decay type. Occasionally, the public may become aroused about specific political issues (e.g., Vietnam and Watergate), even to the point of decisive electoral impact. These issues are linked typically to political events that are a major source of disturbance in the existing political environment. But while these issues can be important in a particular election, their effects are short term. They may influence system outcomes, but they do not change the system. They move the system away from its equilibrium level, but they do not keep it away. They have the important limitation of being unable to sustain themselves beyond the events that brought them into being. Thus as the events fade in public memory, the issues lose their salience, and with it their ability to influence party identification. The dramatic shortterm electoral importance of these issues is thus more than counterbalanced by their inconsequential long-term effects on the party system. It is only a matter of time before the effects of these issues decay, leaving no permanent mark on the system. Model 2 is a visual representation of a classic critical election realignment, the earliest and most simplistic account of party system change. The party system is in equilibrium prior to the impact of some unspecified event after 2 Equilibrium, more than any other aspect, probably differentiates the American party system from others discussed in this volume The equilibrium we presume is a short-term interelection phenomenon. Riker (1982) argues persuasively that disequilibrium in the long run is to be expected from party systems

138

Patterns of Realignment

which it shifts dramatically and permanently to a new equilibrium. A long period of stability is followed by a sudden burst of dramatic change that shifts the party system to a new level of stability. The party system is stationary before the critical election—an intervention that leads to a radical and profound alteration of the system manifested in a sharp and sudden increase in issue polarization. The scenario is a familiar one. A divisive political issue emerges that represents a major source of unresolved tension within the majority party's fragile electoral coalition. Despite its most strenuous efforts, the majority party is unable to keep the issue off the political agenda. Indeed, it eventually dominates the agenda. With the ascendance of the issue comes the inevitable and rapid collapse of the majority party fortunes; it can no longer command decisive support among the electorate. One party's misfortunes are another's opportunities as the minority party or a recently formed third party quickly assumes control of the major political institutions and forges a new direction in public policy. Critical election realignments can result from either or both of two particular mechanisms of partisan change. The sharply discontinuous and episodic character of change specified in critical realignment models is consistent only with massive individual partisan conversions or equally rapid partisan mobilization of new voters previously uninvolved in politics. Thus according to this perspective, abrupt and permanent transformations of the party system are caused by large numbers of voters discarding old party attachments in favor of new ones or massive numbers of newly active voters acquiring a distinctive partisan orientation. Thus the New Deal realignment, it is alternatively argued, came about because many Republicans came to identify with the Democratic party in response to economic depression (Sundquist, 1973; Burnham, 1970; Erikson and Tedin, 1981) or because the ranks of the Democratic party were swelled by the massive mobilization of previously inactive voters, especially immigrants and women (Andersen, 1976, 1979; see also Converse, 1975a; Salisbury and MacKuen, 1981; Clubb et al., 1980). Population replacement, as we will argue, is capable of profound alteration of the party system, but this mechanism is sharply inconsistent with critical election realignment models. The more obvious inconsistency is that replacement effects by themselves cannot be large enough to produce dramatic change in a single election. The less obvious inconsistency is the theoretical Achilles' heel of critical election theories: unless conversion or mobilization is total at the time of the critical election, population replacement should cause steady increases in issue polarization in the years immediately after the critical election. Prerealignment cohorts, those least polarized, will be successively replaced by newly eligible young voters who come of age when the alignment is at the peak of salience and should therefore be highly polarized (Beck, 1974, 1979). Replacement mechanisms entail a dynamic evolution to a new

Issue Evolution: The United States

139

equilibrium level; they are inconsistent with a one-time shift. And if replacement is ruled out, only overnight conversion or implausibly rapid mobilization is left. We are left to postulate an event of such magnitude that it can alter the party system in one election—and yet not be decisive enough to create distinctive party alignments of those coming of age in the years thereafter. Model 3 is a very different account of partisan change, a gradual transformation of the party system that takes place over an extended period. The change effected through this transformation is permanent; it leaves an imprint on the political landscape. And the change can be quite substantial, fundamentally altering the complexion of the party system. But the process is slow, gradual, incremental. This is a noncritical, evolutionary model of partisan change, a model that is consistent with Key's notion of secular realignment. As Key observed: A secular shift in party attachment may be regarded as a movement of the members of a population category from party to party that extends over several presidential elections and appears to be independent of the peculiar factors influencing the vote at individual elections . . . a movement that extends over a half century is a more persuasive indication of the phenomenon in mind than is one that lasts less than a decade. (1959: 199) Thus, Key viewed a secular realignment as a gradual shift in the partisan composition of the electorate. Though gradual, the shift is persistent and herein lies its significance. While critical realignments look to individual conversion or rapid mobilization to explain the dynamics of partisan change, secular realignments can be accounted for by the far more common mechanism of normal population replacement. They can result from such essentially nonpolitical forces as differential birthrates between the party coalitions, interregional migration patterns, or economic/technological transformations that gradually produce new political generations exposed to different political forces than their parents. Secular realignments do not depend upon large-scale changes in individual partisan attachment, a characteristic noted for its substantial level of stability (Converse and Markus, 1979), or upon massive partisan mobilizations of segments of the electorate. Instead, the gradual character of secular realignments means that they can be understood as a not untypical outcome of continuous population replacement. The most complex pattern of partisan change is seen in Model 4. It may be regarded as a dynamic evolutionary model of change in the party system. The model is dynamic in that it presumes that at some point the system moves from a fairly stationary steady-state period to a fairly dramatic change; the change is manifested by a "critical moment" in the time series. Significantly, however, the change—the dynamic growth—does not end with the critical moment. Instead, it continues over an extended period, albeit at a much slower

140

Patterns of

Realignment

pace It is this continued growth after the initial shock that defines the evolutionary character of the model The issue polarization, however, does not continue to grow indefinitely Instead, after an extended period of increase, the polarization decays at a gradual pace The pattern culminates in the establishment of a new equilibrium—one that shows a clear but lessened degree of polarization The dynamic growth and decay pattern of partisan change illustrated in Model 4 may be thought of as a synthesis of critical and secular realignment models The more pronounced the initial step compared to the eventual development of the issue polarization, the more closely the model follows a critical election realignment Conversely, if the critical moment is modest compared to the continuous long-term component of change, then the model approximates a pure secular realignment In either case, the main characteristic of the polarization is that it is not a one-shot, large-scale phenomenon but, rather, follows a discernible pattern over time The polarization not only sustains itself but grows larger for a lengthy period until it begins a modest decay that eventually leads to a new equilibrium 3 Thus, looked at over a sufficiently long period, it is the evolutionary character of the change, the fact that the electorate is undergoing small, incremental change after the critical moment, that is most impressive The initial shock merely sets in motion a pattern of change that continues into the future Dynamic evolution is consistent with a variety of causal mechanisms of change The initial step could be due partly to individual partisan conversions and partly to rapid partisan mobilization The gradual change over an extended period of time is probably the result of population replacement The gradual growth and decay in polarization that is a basic characteristic of dynamic issue evolutions is well accounted for by normal population replacement The growth in polarization comes about when older voters, who are relatively unaffected by the new issue cleavage due to their well-established partisan predispositions, inevitably leave the electorate and are replaced by newly eligible young voters whose weak partisan ties are easily influenced by the salience of the new issue This dual impact of population replacement leads to a predictable increase in issue polarization The later decay in polarization can be explained by the same mechanism With the passage of time, the salience of the issue gradually declines as does its influence on mass partisanship Its major influence is now concentrated among older voters— those whose partisanship was formed during the heat of political conflict Young voters, having had no direct experience with the issue, are least likely to be polanzed by it Thus both the gradual growth and partial decay in issue polarization are logical outcomes of normal population replacement Of the four models of partisan change we have presented, the first (lmpulse3 The process is bounded by an asymptotic equilibrium That new equilibrium occurs when the electorate becomes sufficiently polanzed around an issue that replacement can no longer add to the process

Issue Evolution: The United States

141

decay) is probably the most common issue impact, but it is not large or lengthy enough to be important in the evolution of party systems. The second, critical election realignment, is too simplistic to accord with the complexities of partisan change. Resting on assumed massive partisan conversions or equally massive (and sudden) partisan mobilizations, it overestimates the degree of change occurring at a single point and underestimates the amount of change occurring over time. Pure secular realignment, which might be an adequate account of the impact of demographic shifts and the like, cannot explain issue realignment, for it fails to account for the reason why the process begins in the first place. It is a fire with no spark. The dynamic growth and decay model is a plausible account of all realignments. Where critical election accounts are conversion with no replacement and secular realignment posits replacement but no conversion or (rapid) mobilization, the dynamic evolution model posits all three mechanisms; conversion and mobilization start the process and replacement sustains it. The "critical moment" of the dynamic growth model must be large enough to be visible—far less, however, than the convulsive change of the critical election. Long-term dynamic growth occurs when that "visible" shift is reinforced by recruitment (and derecruitment) that continues to emphasize the new cleavage in following years. And partial decay occurs when the issue loses its capacity to shape the partisan orientations of the newest members of the electorate. While we believe the dynamic evolution model provides a plausible account of how realignments work, it presents difficult problems of estimation. Most importantly, we lack the necessary information to track any single issue throughout its extended growth and partial decay. Instead, we will examine two political issues—each of which exemplifies a key component of the dynamic evolution model. We first turn to a limited analysis of an issue that represents the political conflict underlying the New Deal. This analysis clearly shows the decay of issue polarization between partisan supporters. Our more substantial analysis focuses on the most significant issue evolution occurring during the post-New Deal period in American politics—the partisan evolution of racial desegregation. While race no longer commands the public attention it once did, we will see that it is still in its dynamic growth phase of partisan development. Together these two issues illustrate the dynamic process of realignment. THE NEW DEAL IN DECAY: AN EMPIRICAL SKETCH

The "New Deal," we all agree, is the basis of the current party alignment. Analysts have predicted its demise for at least three decades, but it seems to persist. Few would disagree that the alignment is in decay. Indeed, the largescale abandonment of identification with and fidelity to parties in America is the subject of Chapter 8 in this volume. Decay has another face, the depolarization of attitudes that give policy

142

Patterns of

Realignment

content to the alignment For the New Deal was not only a party coalition, an appeal to the common man, it was also a set of programs that reached out to the working man out of work and said, "It is not your fault, you did not cause the Depression," and most important, "It is the government's responsibility to do something about it " Thus government responsibility was then and still is a controversial assertion More than anything else, it was the cleavage issue that separated partisans and still does Contemporary survey data allow the study of this party cleavage issue Recalled partisanship permits the reconstruction of parties of the past Decay of polanzation can be loosely estimated, which we will do, and modeled with some precision, which we will not do The reconstruction methodology (taken up later) demands more than we can reasonably assume Recovenng an event of the 1930s from respondents who lived on to be interviewed in the 1970s asks more than can be expected Pushed further back in time, our sample becomes ever smaller and progressively age-biased The technique requires memory of forty-year-old events, when Niemi et al (1980) have found four years problematic Thus we shall paint no portrait of the New Deal, but only a sketch No single issue portrays the New Deal realignment, it was a complex brew of issues, symbols, personalities, and coalitions But most would concur that if one issue had to be chosen, it would be the fundamental disagreement over the proper role of government intrusions into the marketplace to provide jobs for those who wanted to work We, in turn, have chosen to look at its best empincal manifestation in the Center for Political Studies survey item on this question 4 Issue alignment is too subtle a phenomenon to tolerate the considerable noise introduced by vanations in question wording over time Thus we must restnct our analysis to surveys where question wording is identical We have chosen the CPS cross sections of 1972, 1976, and 1978 for a reconstruction data base They are current and present no question wording artifact, but they are long removed in time from many of the events of interest To model the New Deal realignment is a tempting prospect And a senes can be reconstructed back to—perhaps even before—the event But the data are not good enough to sustain the modeling exercise 5 Consequently, we settle for the more modest goal of sketching issue depolanzation in progress The reconstructed attitude graphed in Figure 5 2 show the issue deahgnment we expected to see Ignonng the sharp upward changes at the end of the 4 The question is the familiar seven-point forced choice between "The government should see to it that every person has a job and a good standard of living' and ' 'Each person should get ahead on his own " 5 We are in a much stronger position for the desegregation analysis to come because the events of interest are contemporary We can measure alignment from a number of independent cross sections before and after the beginning of the issue evolution

Issue Evolution The United States FIGURE 5 2 Mean Party Positions on Government Responsibility for Jobs and Standard of Living A Reconstructed Time Series

REPUBLICANS DEMOCRATS

Most L beral

3•

—ι— 1932

~ι— 1940

— I — 1948

T 1956

196

— ι — 1972

YEAR

SOURCE SRC/CPS National Election Studies for 1972, 1976, and 1978

FIGURE 5 3 The Decay of Interparty Polarization on the Jobs and Standard of Living Issue

SOURCE

Recomputed from Figure 5 2

144

Patterns of Realignment

series—attributable to sampling fluctuation—the sketch has some surprising features. One is that the decay of cleavage is asymmetric; it is accounted for entirely by changes among Democratic identifiers. Republican identifiers show no evidence of shift on New Deal issues from Hoover through Carter. Nothing happens. Democrats are a more interesting lot. They, too, show no change over the last three decades, which means also that the New Deal cleavage has been stagnant for three decades. These probably conservative estimates show a decay of some 30 to 40 percent of the original party difference, all of which occurs in the 1930s and 1940s. This is unexpected in two regards. One is that the method is far more sensitive to recent than ancient changes. Yet Figures 5.2 and 5.3 portray remarkable stability for recent decades and change in the past. And the pattern of decay could be produced only if the old (in the 1970s samples) were considerably more liberal than the young, which contradicts our normal expectation. The pattern of the figures does make sense from another perspective. If it is assumed that the generation that experiences realignment is peculiarly sensitive to the party aligning issue, but passes on its partisanship to its children (and they to theirs) more successfully than its issue position, then the gross patterns of Figures 5.2 and 5.3 are what would be expected. Party alignments decay to a new equilibrium as the generation of realignment loses its numbers and its intensity over the years. Our sketch of the New Deal in decay is an unintended confirmation of Beck's (1974) "socialization theory of partisan realignment." It is consistent with the view that the socialization mechanism can pass on partisanship as identification but not the vivid emotional context in which the identification was forged. Where the ties between generations are imperfect, as they surely are, any alignment formed from idiosyncratic events must of necessity decay. THE POLARIZATION OF DESEGREGATION

Our discussion of the issue conflict underlying the New Deal suggests the properties that political issues must possess in order to undergo significant evolutionary change. Three such properties are these: 1. Issue preferences must be deeply felt and strongly symbolic in character. 2. Parties and candidates must take or become associated with different positions on the issues. 3. The issue must be on the political agenda for a long time. While numerous issues have crowded onto the political agenda since the formation of the New Deal and competed for public attention, only one— racial desegregation—clearly has all three attributes. Race has deep symbolic meaning in American political history and has touched a nerve in the body

Issue Evolution The United States

145

politic It has also been an issue on which the parties have taken relatively clear and distinct stands, at least since the 1964 presidential contest between Goldwater and Johnson The issue, finally, has a long political life cycle, it has been a recurring theme in American politics since our founding Most of that time it has been submerged, too contentious for the party system But it reemerged as a partisan conflict in the early 1960s and has remained since then Thus if a significant issue evolution is going on in Amencan politics, it is most likely to revolve around the issue of racial desegregation If an issue persists relatively unchanged for a lengthy period, we can get several cross-sectional estimates of the polarization of issue and party and from these varying readings estimate what a continuous time senes might have been if it had existed That is our method here We have used all the Amencan national election studies for presidential and off-year elections (1956 through 1980) to construct an "even-year" time senes of the desegregation attitudes of party identifiers Desegregation attitude scales are constructed for each cross section by summing attitude items6—each weighted equally—and transforming the scales obtained to a common metnc (mean 50, standard deviation 25) for all years The years 1945 to 1955 and odd-numbered years thereafter (except 1963) are then reconstructed from respondent reports of party identification change 7 The intensity and continuity of racial attitudes make us comfortable with this procedure But the need for intensity and continuity limits the method There are probably large numbers of issues of the impulse-decay type, for example, but we cannot study the decay of party alignments on them because questions cease to be asked when they are no longer topical Other issues may have a lengthy history, but since attitudes on them are not intense, cross-time measures are highly suspect from the intrusion of the varying cues of question wording Nierm, Katz, and Newman (1980) have demonstrated from panel data that recalled party identifications are frequently erroneous Since the reconstruction methodology rests in part—thankfully not in large part—on the quality of recall, it, too, may be erroneous Thus we provide more detailed exposition of the method, together with a modest validation of it, in Appendix 5 A Before getting on to the business of modeling issue realignment on race, we pause to lay out our expectations for the varying models of partisan change Polanzation is measured by the difference between the mean attitudes of the two-party groups over time Operationally, that entails simply computing party means for each year and subtracting one from the other Arbitranly, we subtract the Republican mean from the Democratic mean That leaves us with three time senes, one for each party and one for the difference between parties 6

Items usedfromthe 1978 study are school integration and desegregation (general) For other years see Carmines and Stimson (1981 116) 7 Party identifier attitudes for 1963 are estimated from a Hams survey of November 1963

146

Patterns of Realignment FIGURE 5.4 The Desegregation Attitudes of Party Identifiers, 1945-1980

— ι 1945

1 1950

1 1955

1 1960

1 1965

1 1970

1 1975

1— 1980

SOURCES: SRC/CPS National Election Study Series, 1956-1980, and Louis Harris and Associates, Inc. The party mean series are graphed in Figure 5.4 for a first visual demonstration of the phenomenon. Figure 5.4 shows a stable and very minor difference in the views of the two parties' identifiers up through 1962 followed by the appearance of growing polarization in 1963-1964. A cautious spectator of the civil rights movement before that time, in 1963, the Kennedy administration (and implicitly, the Democratic party) took up the cause of civil rights as its own. Whether that by itself would have been sufficient to create a partisan issue evolution we shall never know, for it was quickly followed by Barry Goldwater's aban­ donment of the traditional procivil rights Republican tradition the next year. Taken together these events and these two years are our critical moment. All of the four models of issue realignment may be expressed as variations of the first-order transfer function:

Y

' = ,_i and £,_ 2 ). Table 10.1(b) gives estimates with the intercept constrained to zero, "as if" a first-order process were truly at work. The estimate of b[ rises to .65, the fit deteriorates considerably, and bx and b2 do not sum to unity. Table 10.1(c) places no constraints on the intercept but requires bx and b2 to add 2 The data used in this chapter are drawn from three separate British panel studies, one of 1,114 respondents interviewed in 1969 and after the election of June 1970 (the "1969-1970 panel"), one of 1,096 respondents interviewed after the elections of 1970 and February 1974 (the "1970-February 1974 panel"), and one of 1,830 respondents interviewed after the elections of February and October 1974 (the "February-October panel") The panels do not overlap at all 3 The polytomous probit problem is nonlinear Estimates are required simultaneously of weights on the independent variables and category thresholds for the dependent variable to maximize the correct assignment of cases to fitted categories for the dependent variable Fiorina (1981 Ap­ pendix) gives a very accessible treatment of the logic of the estimation procedure

314

Patterns of Dealignment TABLE 10.1

Expectations for the British Economy, October 1974 Variable

Coefficient

t-statistic

(a) Constant October perception February expectation R2 = .17 (b) Constant October perception February expectation R2 = .08 (c) Constant October perception February expectation R2 = .15

.80 .45 .23 F = 210.9 suppressed .65 .43 F = 144.9 .26 .65 .35

14.4 12.9 10.4

obs = 1,520

SOURCE: British Election Study, February-October 1974 Panel (N = 1,830).

NOTE: Estimation of the equations was by OLS, un-

restricted in (a) and restricted in (b) and (c). Dependent variable was October response to the question "How well do you think Britain's economy will do in the next few years?," with response "better off" coded "2" in all cases, response "same" coded "1," and "worse" coded "0." up to one. Only the coefficient of b2 changes, which means that the constant term is not confounded with &, but rather with b2, the coefficient of £ , . , . Thus whatever omitted information the constant represents is more likely to arise from past expectations than with present perceptions. In any event, the intercept that is significantly estimated indicates that expectations were consistently more positive than the effects of (P, — E,_i) alone lead one to expect.4 Further evidence on adaptation can be found in the expectation of family well-being (Table 10.2). The adaptation is evident in the effect of having become better off, but there are also effects of the hypothetical view of how well Labour would have handled the Miners' strike, by the experience of 4 The result is still approximately a first-order adaptive process with an adaptation parameter of 0.6. This means that if one expected worse m February but thought things had gotten better by October, the October expectation would have improved to about "stay the same." If one expected better in February but thought things had gotten worse by October, the October expectation would come down to about "same." If February expectations (whatever they were) were confirmed by October perceptions, expectations would remain unaltered.

TABLE 10 2

Expect Worse Off for Family, February 1974 Constant Party identification 1970

150*

Very strong Conservative Fairly strong Conservative Not very strong Conservative Independent + Liberal Not very strong Labour Fairly strong Labour (Very strong Labour)

+ 019 -004 + 0.43 + 015 -010 + 010 —

Conservative handled prices 1974 Very well Fairly well Not very well (Not at all well)

-0.46 -000 -0.02 —

Labour would have handled prices 1974 Very well Fairly well Not very well (Not at all well)

- 0 38 + 013 + 006 —

Conservative handled Miners' strike 1974 Very well Fairly well Not very well (Not at all well)

-007 -009 -003 —

Labour would have handled Miners' strike 1974 Very well Fairly well Not very well (Not at all well)

-0.46* -043* + 0.01 —

Glad to be in Common Market 1974 (Sorry)

+ 0.23* —

Too many immigrants let in 1974 (Not too many)

-0.05 —

On three-day week 1974 (Not)

-008 —

Unemployed 1974 (Not)

-041* —

316

Patterns of Dealignment TABLE 10 2 (cont) Expect economy 1974 will Get Better Stay Same (Get Worse)

-0 57* -0 39*

Family situation in 1974 is Better off Same (Worse off)

- 0 28* +004

rho = 38 η = 663 Estimated thresholds Better off Same Worse off

1973

1968

1967

UDF

23 10 100%

121

21 23 2

1978

UDF

21 6 100%

1 19

] 38

16

1981

SOURCES: 1958-1968 from Jacques Chapsal, La Vie Politique en France depuis 1940, 3rd ed. (Paris: Presses Universitaires de France, 1972), 657-664; 1973 from Jacques Chapsal and Alain Lancelot, Le Vie Politique en France depuis 1940, 4th ed. (Paris: Presses Universitaires de France, 1975), 663-664; 1978 from Sondages 40, no. 1 (1978): 8; 1981 from Time (June 29, 1981), 27. NOTE: Table entries are the percentage of votes for the parties on the first ballot. a The following abbreviations are used: MPR, Mouvement Republicain Populaire; RI, Republicains Independants; PDM, Progres et Democratic Moderne; UDF, Union pour la Democratic Franqaise; CD, Centre democrate.

1958

Parties'

TABLE 14.1 French Legislative Election Results, 1958-1981

428

Patterns of Stability

it is possible to transfer power effectively from one to the other, in the face of electoral defeat. The 1981 presidential victory of Socialist Fran$ois Mitterrand, and his subsequent conquest in the legislative elections, illustrates that the opposition (always the Left, until now) can come to power and achieve the parliamentary majority necessary to govern. Given this transformation of political institutions and parties under the Fifth Republic, one would anticipate profound changes in French electoral behavior, the research focus of this study. Commonly, students of French politics assert that voters have developed relatively strong party loyalty, in contrast to their fragile attachment to party during the unstable Fourth Republic. Moreover, the appeal of the parties is alleged to be more pragmatic, more national, more "catch-all." Hence traditional divisions, such as class, religion, region, and ideology, are thought to be increasingly irrelevant for shaping the party preferences of the electorate. The foregoing characterization of French electoral politics, though widely subscribed to, is quite misleading, as will be demonstrated in this chapter. First, I look at the extent of party identification in the electorate and find it has experienced little change in over thirty years. Then, I examine the determinants of party choice across the Fifth Republic and discover that the traditional cleavages remain in force. The political attitudes and behavior of the French public appear, in a word, "stalled." THE EVOLUTION OF PARTY IDENTIFICATION

A review of the research on party identification in France suggests the following pattern of development. At the dawn of the Fifth Republic, in 1958, a relatively small portion of the public felt psychologically attached to a political party (Converse and Dupeux, 1962: 9). The widespread lack of party affiliation seemed to go far toward explaining the general instability of the multiparty system, in particular the appearance of "flash" parties like the Poujadists in 1956. However, as the institutions and forces of the Fifth Republic, especially Gaullism, took hold, there occurred a "swing into" partisanship (Cameron, 1972: 23; Inglehart and Hochstein, 1972: 353-355). The more extensive party identification observed in the late 1960s served as evidence of increasing political stability. While new (post 1960s) data have yet to be reported, it is still assumed that partisan attachment remains high. In sum, as Frank Wilson (1979: 83) has recently observed, during the Fifth Republic the French parties experienced a "revitalization." This scenario of the evaluation of parties in the French electorate, although widely believed, is in error. A more thorough reading of the available data indicates that, in fact, the percentage of party identifiers was high in the later years of the Fourth Republic, as well as the earlier years of the Fifth Republic. Further, this elevated level of party attachment in the electorate continued, with minor fluctuations, through the 1960s and 1970s.

France: The Stalled Electorate

429

One weakness of past research has been the exclusive reliance on a September 1958 survey for estimating French partisanship at the juncture of the Fourth and Fifth Republics. The essential strategy was to compare the 1958 results of Philip Converse and Georges Dupeux (1962: 9) to estimates from about ten years after (Cameron, 1972: 19-26; Inglehart and Hochstein, 1972: 352; Campbell, 1971: 33-34). It is difficult to exaggerate the influence of the Converse and Dupeux piece in shaping scholarly belief about the partisan attachments of the French public. Their study, the first systematic report on a French electoral survey to appear in an American academic journal, found only about 45 percent of the respondents identified themselves with a party or group, with another 10 to 15 percent aligning themselves to some political tendance. In contrast, later survey research consistently found party identification to be much higher, and thus concluded that great change had occurred among the French electorate. Ronald Inglehart and Avram Hochstein (1972: 352-353), for example, calculating party identification to be 76 percent in 1968, decide that "a process of rapid alignment was taking place in France over the decade." However, scrutiny of the data indicates this change was apparent, rather than real. The base-line estimate provided by the September 1958 survey seems to understate seriously the usual level of party identification among the French. As will be shown later, this September 1958 survey stands as an outlier in comparison to other estimates from 1958, and in comparison to other years generally. Therefore, when one uses it as a standard by which to evaluate other measures of party identification, one receives a spurious impression of change. This problem will be examined in more detail later, but first it is necessary to consider the measurement of party identification in the French context. Two kinds of survey items have been used: open-ended and closed-ended. A typical open-ended item is that from the September 1958 survey, which asked,' 'What is the political party to which you feel closest?'' with no prompts provided. An example of one of the closed-ended questions is, "Would you tell me to which political party you feel closest, with the help of this card [which lists the parties]?" (Inglehart and Hochstein, 1972: 371). These two types of measures yield wildly different estimates of the level of party identification in the French electorate. In 1968, with a closed-ended question, Inglehart found 76 percent of the French identified with a party, while Converse and Pierce, using an open-ended question, estimated only 50 percent were party identifiers in that year (Caldeira and Greenstein, 1978: 39-40). When the mere style of question spawns discrepancies of this magnitude, it is obvious that any evaluation of change in partisan identification over time requires that the same measure be used or, at the least, that measurement differences be controlled. For several reasons, I have some preference for consistent use of a closed-ended question. First, such items have received extensive trial in the assessment of party identification in the United States,

430

Patterns of Stability

for example, the standard question "Do you consider yourself to be a Democrat, a Republican, an independent, or what?" Second, a fixed-choice format helps the French voter cope with the whims of French political party nomenclature. The third reason, a practical one, is that more data points are available because more surveys have followed the closed-ended form. In the 1970s, for example, the Euro-Barometer had posed the following question once or twice a year: "Do you consider yourself a supporter of any particular political party? If so, do you feel yourself to be very involved in this party, fairly involved, or merely a sympathizer?" Unfortunately, by restricting oneself to closed-ended instruments, one leaves unmeasured the critical gap between 1948 and 1962. To fill in these missing years, the available open-ended responses were adjusted, in an attempt to generate estimates comparable to what closed-ended questions would have yielded. The approach I followed, a common one, was to consider that a "party identifier" is anyone who mentions a specific party, party formation, party leader, splinter group, or political tendance (Cameron, 1972: 24; Campbell, 1971: 33-34; Barnes and Pierce, 1971: 645). Such a definition undoubtedly strikes some readers as too generous; however, it is not illogical in a French context. Answering an open-ended question about party attachment can be rather taxing for the average French citizen, given the large number of parties and the frequency with which they change their names. The GauUist party, for example, has changed its label in virtually every election. The following are among its banners: in 1947, Rassemblement du Peuple Frangais (RPF); in 1958, Union pour la Nouvelle Republique (UNR); in 1962, the UNR and the Union Democratique du Travail (UDT); in 1967, Union Democratique pour la Cinquieme Republique (UDVe); in 1968, Union des Democrates pour la Republique (UDR); in 1973, Union des Republicans de Progres (URP); in 1978, Rassemblement pour la Republique (RPR). Clearly then, if an interviewee, in response to the question "Which party do you feel closest to?" simply answers "de Gaulle," it is not unreasonable to classify that person with the GauUist party, as currently named. By employing both open-ended and closed-ended items, a time series of fifteen observations on party identification in the Fourth and Fifth Republics was assembled (see Fig. 14.1). A glance reveals that outlier status of the September 1958 estimate (59 percent). Its deviation is especially curious, given that the other 1958 estimates (72 percent and 73 percent) evidence no departure from surrounding years. (Three 1958 estimates are possible, from September, November, or December, since the 1958 French Election Study took place in three separate waves, that is, before and after the referendum on the Constitution of the Fifth Republic, and after the legislative elections.) As noted earlier, Converse and Dupeux, and subsequent researchers, have focused on the September 1958 survey as their basis for evaluation of the degree of party identification in France. Because of its pivotal role, then, an

France: The Stalled Electorate

431

explanation for the outlying character of this observation must be sought. I would speculate that it represents a real, but short-lived, attitude shift. In September 1958, just before the establishment of the Fifth Republic, a portion of the French electorate naturally drew back momentarily from their parties, which stood for a troubled political past. Then, when constitutional order was restored, they rejoined them, as the increased party identification figure for November indicates. Having isolated this deviant September 1958 survey, let us look with more care at the broader pattern of party identification across time. The results from forgotten IFOP surveys of 1948 and 1955 are intriguing. While the wording of these early partisanship items clearly differs from the current format, the responses hint, surprisingly, that partisanship was at a fairly high level (around 70 percent) during much of the Fourth Republic. Further, we observe from the 1958 and 1962 surveys that at the onset of Gaullism, parties were well implanted in the electorate. These earlier polls provide evidence against the argument that the rise of Gaullism brought about an elevated level of party alignment. Even during the late 1960s, with Gaullist party strength at its peak, the rate of party identification does not really differ from this earlier period. Moreover, with the death of de Gaulle and the relative decline of his party in the 1970s, we fail to observe a lasting drop in overall party identification in the French electorate (e.g., by 1979, it had returned to 70 percent). In sum, it appears that Gaullism has had little impact on the extent of partisan feeling. Instead, party identification has maintained itself at a generally high level across the two Republics, ultimately impervious to the changing institutions, parties, and personalities of the postwar epoch. The conventional impression is that the French electorate is unstable in its partisan attachment. Yet the data of Figure 14.1 show this impression is incorrect. Over these thirty-four years, party identification has fluctuated within a narrow range (58 to 76 percent), around an elevated average rate (68 percent). Given the vagaries of sampling and measurement to which the series is subject, these are rather formidable signs of stability. Furthermore, careful inspection of the variations around the mean value fails to establish any clear pattern of change. (E.g., the volatility at the end of the curve cannot be taken as a trend because the observations cover only a few years.) These conclusions concerning the stability of party identification in France are reinforced by comparisons to the United States case. For the thirteen United States election surveys for 1952 to 1976, utilizing responses to the question "Generally speaking, do you usually think of yourself as a Republican, a Democrat, an independent, or what?'' the average rate of identification was found to be 70 percent, with a range of 60 to 76 percent (calculated from Inglehart, 1977: 308). These figures are almost identical to those from France. This comparable stability in the French electorate is particularly noteworthy when one considers that the United States results cover a shorter

432

Patterns of Stability

FIGURE 14.1 Party Identifiers, 1948-1981

90 -

β

Ο

Open end Format



Closed end Format

-Q72

70

60 -

50 40 -

1945

• 1950

• 1955

I 1960

I 1965

I 1970

I 1975

I 1980

Year

SOURCE: See Appendix 14.A

period, and are based on surveys using larger probability samples that always employed the same party identification item. If the French are as loyal to parties as Americans are, then why has the French party system appeared more volatile? Specifically, why do new parties continue to emerge, capturing sizable portions of the French electorate? An "elite" explanation for this phenomenon seems most plausible. French po­ litical leaders are willing, unlike their United States counterparts, to present the public with "third"-party alternatives. As the data have shown, for both countries there exists about the same portion of voters available for mobili­ zation, that is, the percentage of independents is about the same in France and in the United States. However, in the United States, disgruntled members of the political establishment seldom seek to draw these unattached voters to them by launching an alternative political party. Instead, they work to increase their support within one of the two established parties (e.g., witness Reagan) or, in the extreme, campaign for office without any party label (e.g., witness Anderson). By way of contrast, discontented political leaders in France, questing after electoral support, are not unwilling to change a party radically, or even form another. As Wilson (1977: 28-29) has observed, "The presence

France: The Stalled Electorate

433

or absence of capable leaders explains why party transformation took place so soon in some parties, so late in others, and not at all in still others." The readiness of French political leaders to pose new party alternatives provides the unattached French voter with opportunities of choice that the unattached American voter, who is offered only Democrats versus Republicans, simply does not have. Finally, the fact that French politicians do not behave in the same way as American politicians does not mean national character differences are at work. Rather, those behavioral differences appear due to the different incentives that the electoral systems provide, for example, the American parties can be taken over more easily by new forces because of the primary system. (For an explication of French election rules, see Pierce, 1968: Chap. 6.) Overall, the picture for France is one of partisan stability. However, the 1981 election year estimate (58 percent) is not so easily reconciled with this generalization. Does it signal the beginning of a dealignment, or is it merely a temporary dip, like the September 1958 estimate? In order to try to answer this question, it is necessary to reflect on the meaning of party identification. On the one hand, if one takes the traditional view of party identification, seeing it as a deep, enduring psychological attachment to a political object, then the 1981 drop probably represents a severance of ties that will not be mended quickly. On the other hand, according to the revisionist view currently gaining sway, party identification is much more endogenous than had been thought. That is, while it has an inertial component, it is also influenced in the short-run by issues, events, and personalities. (For an excellent treatment from this perspective, see Fiorina, 1981: Chap. 5; Alt, this volume: Chap. 10.) In this light, the April 1981 departure could be, in a fashion similar to the September 1958 deviation, a response to the unique conditions of the moment; specifically, the immediacy of a close presidential contest in which economic issues had an importance unparalleled in the Fifth Republic. Undoubtedly, these peculiar circumstances caused some to deny their party attachment. While it is not known whether this defection will persist, my guess is that it will not. Regardless of how the economy fares under leftist direction, I believe the overall level of partisanship in France will soon return to its normal high mark. This speculation stems from the notion of blaming the incumbent for hard times, coupled with the frequently observed differential attachment of Left and Right partisans in France. Thus, under a deteriorating economy, Right partisans who had defected to independence will tend to return to the fold, while Left partisans will tend to maintain their party identification. And" under an improving economy, Right partisans will tend to join the Left, while earlier Left partisans sustain their partisanship. With either scenario, then, party identification should rise.

434

Patterns of Stability

TRADITIONAL CLEAVAGES AND PARTY PREFERENCES

The traditional cleavages of French politics are well known, and include class, religion, region, and ideology (Cemy, 1972: 445). It is increasingly argued that the influences of these old divisions on party choice are diminishing. Roy Macridis (1975: 60-64, 91), in his important work about contemporary French politics, summarizes this view well, asserting that "religion is beginning to lose the influence it had upon the voters," and that there has been "the waning of class," "the waning of ideology," and the " 'nationalization' of the major political parties, in the sense of a more even geographic distribution of their vote.'' In other words, the standard guideposts of electoral decision for French citizens—how they make a living, whether they are religious, where they reside, if they place themselves to the Left or to the Right—have clearly declined in importance. This conclusion is incorrect, as I attempt to show next. The relationship of each of these factors to party preference will be examined in turn, beginning with class, moving on to religion, then region, and, finally, ideology. Social Class Regardless of their views on the changing role of social class, no analysts deny that it makes some electoral impact. Table 14.2, offering national survey data from the 1978 election year, illustrates this relationship. The social class indicator is occupation of the head of household, grouped into three categories: worker, peasant, middle class. The measure of vote intention comes from responses to this item: "If there were a general election tomorrow, which party would you support?" For purposes of presentation and analysis, these party responses have been ordered from Left to Right. At a glance, one sees that class is related to vote. The relationship is certainly not overwhelming, as with, say, British survey data. Still, the class differences are obvious. Robert Alford (1963) has developed an index to assess the strength of this tie, calculated by subtracting the middle-class percentage supporting the Left parties from the working-class percentage supporting the Left parties. Applying this formula to Table 14.2, the class voting index equals + 20. Clearly, the Left maintains its attractions for the blue-collar voter in France. The Alford index provides a helpful first look at the link between class and vote. However, as a summary measure, it is seriously flawed, for it fails to incorporate much of the available information. That is, it forces all the different parties on the Left into one category, at the same time ignoring the •other parties and occupational groupings. While such combinations and omissions may simplify the picture, they may also distort it. A more appropriate summary statistic would be a measure of association that merely correlates the two variables. For example, the Pearson product-moment correlation (r) between social class and vote, as measured and ordered in Table 14.2, is .23.

France: The Stalled Electorate

435

TABLE 14.2 Social Class and Vote Intention, 1978 Vote Intention Left parties Communist PSU Socialist Left Radical Center parties Center Right parties UDF Gaullist Total (N)

Social Class Worker

Peasant

Middle Class

25% 8 46 1

11% 5 30 0

8% 5 46 1

5

12

9

6 9 100% (248)

18 24 100% (54)

16 15 100% (332)

SOURCE: Euro-Barometer

10 (October-November 1978).

This correlation confirms what the percentage differences suggest, that is, there is a positive moderate association between class standing and party choice. In this chapter, I wish to test the hypothesis that the relationship between class and vote has declined. Therefore, the correlation of these two variables needs to be evaluated at several points in time. But before doing so, it is perhaps useful to clarify the placement of the peasant category in my ranking of French social classes. Given the size and importance of the peasantry in France, the frequent practice of excluding them from analysis, as the Alford index does, is unacceptable. Instead, they should be ranked between the workers and the middle class, both according to their relation to property and their political attitudes. With regard to their relation to property, it must be recalled that almost all French peasants own small farms that they work. Hence their class status is mixed, for they are both owners and workers. With regard to their political attitudes, they are a "supporting class" of the bourgeoisie, as Maurice Duverger (1960: xxii) called them. This supporting role of the peasantry vis-a-vis the bourgeoisie is suggested in Table 14.2, when vote intention for the Communist party is studied: whereas 25 percent of the working class backed the Communists, only 11 percent of the peasantry did so, just slightly more than the 8 percent support given them by the middle class. (For more on peasant politics, see Lewis-Beck, 1981.) For these reasons, then, the social class variable is ordered as follows: worker, peasant,

436

Patterns of Stability

middle class. (It is important to note that, even if the peasant category is not so ranked, and the three occupational categories are treated merely as nominal, the eta correlation of occupation and vote intention reveals the same pattern that is reported later in this chapter.) The top line in Table 14.3 shows the correlation (r) between these measures of class and vote for the five legislative election years for which national surveys are available. One observes a persistently positive moderate association of class standing and vote intention, which exhibits no downward trend over time. Rather, across this twenty-four-year period, the relationship is essentially stable, fluctuating within a narrow range, r = .13 to r = .30. Hence, contrary to the expectations of some, social class continues to shape party choice with undiminished force. TABLE 14.3 The Correlation of Traditional Factors with Vote Intention, 1958-1981 Election Year Factor Class Religion Region Ideology

1958

1967

1973

1978

1981

.13 — — —

.30 .39 — .57

.14 .43 .20 .73

.26 .40 .10 .74

.21 — .09 .62

SOURCES: The second and third waves of the 1958 French Election Study; 1967 French National Election Study; European Community Studies, 1973, 1978, 1981; the 1967 correlation for ideology was reported in Bruce Campbell (1974: 75). NOTE: The measures for the variables are described in the text. When no correlation is reported for a particular year, it means the independent variable was not measured in that survey. If there is more than one survey for the year, the correlations are averaged.

Religion As Guy Michelat and Michel Simon (1977: 159) note, "Most observers consider the tie between religion and politics a thing of the past." There are many reasons for this consideration. Following the pattern in other Western nations, church attendance has fallen off during the postwar period. Further, church-state relations have been settled, and the schools are no longer a live issue. Also, political parties linked to the church have disappeared. For ex-

437

France: The Stalled Electorate

ample, the church-supported Mouvement Republican Populaire (MRP), the biggest party in France in 1946, steadily lost ground and was disbanded in 1967. As well, the church has ceased its active campaign against parties on the Left. Indeed, there has been an increase in Catholic clergy and organizations that are sympathetic to leftist causes. Such important events make the conclusion of a diminished role for religion in electoral politics appear selfevident. However, on the basis of the available survey data, this conclusion appears false. In Table 14.4 religion is related to vote intention, measured as before. (For simplicity, the five-category religiosity variable is collapsed into those who attended church and those who did not.) One sees immediately that voters sharply divide themselves according to their religious practices. For instance, religion clearly separates Communist from non-Communist voters; that is, someone who does not attend church appears over four times as likely to vote Communist as someone who attends church. More generally, one observes that those who are irreligious have a decided preference for parties on the Left. To illustrate this, a sort of Alford index for religion and vote could be constructed, subtracting the percentage of churchgoers who vote Left from TABLE 14 4

Religion and Vote Intention, 1978 Vote Intention Left parties Communist PSU Socialist Left Radical Center parties Center Right parties UDF Gaullist Total (N)

Religion' Attend Church

Don't Attend Church

5% 4 38 2

22% 7 47 2

11

4

22 20 100% (444)

8 10 100% (436)

Euro-Barometer 10 (October-November 1978). The "attend church" category is composed of respondents who said they attended church "several times a week," "once a week," or "a few times a year." The "don't attend church" category is composed of those respondents who said either they had "no religion" or "never attended." SOURCE: a

438

Patterns of Stability

the percentage of nonchurchgoers who vote Left, which yields a percent difference of + 30. A more precise sense of the strength of this association comes from correlating vote intention with religious practice (measured with the uncollapsed, five-category variable stretching from "no religion" to attending religious services "several times a week"). For the 1978 data from Table 14.4, this correlation is rather strong, r = .39. Clearly, then, religion is an important factor in party choice among the French electorate. However, the more direct concern is whether the power of this association has declined, as is so often contended. In Table 14.3, line 2, this correlation is reported for the available election year surveys. (Unfortunately, the 1958 and 1981 surveys did not ask about religious practice.) The relationship is robust, mean r = .41, and exhibits no diminution over time. Evidently, across these years of the Fifth Republic, the allegiance of the faithful to the parties on the Right has been unswerving. A topic of enduring interest in the study of French politics is the relative importance of class and religion for voting. Roy Pierce (1980: 872) observes that there are no detailed investigations of this issue that yse data gathered after the 1960s, and speculates that the impact of class relative to religion may have increased since de Gaulle's death. However, the results of Table 14.3 echo the conclusion reached earlier by Richard Rose and Derek Urwin (1969) in their study of political parties: "Religion, not class, is the main social basis of parties in the Western world today." More precisely, religious practice appears twice as important as social class (the mean correlation for religion divided by the mean correlation for class = .41 / .21 s 2). Further, the dominance of religion over class as an influence on the vote has not dwindled across this time span, as the steady gap between the sets of correlation coefficients indicates. In the postwar period, the surface of French politics appears increasingly secular. Why, then, do we fail to observe a correspondent slackening of the impact of religious practice on vote intention? Certainly, the question admits of no simple answer. Still, Michelat and Simon (1977:181-182), who likewise find that this association remains strong, argue as follows. The connection between church attendance and vote for the Right is not an external, essentially spurious, link between two behaviors. Rather, it symbolizes a network of beliefs, images, and feelings that define a Catholic's political role. Historically, the church, directly and indirectly, has worked to instill a set of values supportive of the established order, of which the Left has not been a part. To be a good Catholic, then, virtually required one to be anti-Left. Because such a pattern of thought is fundamentally prerational, it has managed to survive, even in the face of a church that no longer seriously intervenes in politics and has shown signs of doctrinal liberalization, especially since Vatican II. This survival is not surprising, when one recalls Inglehart's works (1977: 216-229; this volume: Chap. 2), which indicate that "pre-industrial"

France: The Stalled Electorate

439

variables like religious attachment are transmitted from generation to generation with relative efficiency in Western publics. Further, he argues that religion remains relevant because it is related to the new value conflicts over life-style and moral issues. Region Historically, major regions of France have favored one party or tendance over another. For example, the Paris area has given disproportionate support to the Communists, the Northwest has been a bastion for parties on the Right. However, it is now often argued that, largely because of increased geographic mobility and mass communication, parties have developed a more national appeal and the regional voting distinctions are fading (F. Wilson, 1979: 95). Nevertheless, this argument, while attractive, has never been substantiated by actually looking at individual voters and relating their regional ties to their party preferences. According to the 1978 survey used in Table 14.4, some regional differences remain. For instance, voters from Paris appear about two and a half times as likely to say they will vote Communist as do voters from the conservative Northwest. In an attempt to get a more general picture of the relationship, we might examine the full bivariate table. Unfortunately, such a table is too cumbersome to interpret easily. In their 1978 survey, respondents were classified into seven regions: Northwest, Southwest, North, Paris region, Paris basin, East, Southeast. Along with the seven vote-intention categories, this yields an imposing forty-nine-cell table. Further, due to the particular geographic categories and their nominal nature, it is not obvious how best to group the data in order to make them more comprehensible. Rather, to get an overall view of the relationship, it is preferable simply to look at the correlation between the region and vote variables. For these 1978 data, a modest association is present, eta = . 14. Regional attachments are still of some importance for determining the vote in France. But how does the current magnitude of this association compare to previous years? Across the available surveys, regional ties maintained their tug on the French voter (see line 3, Table 14.3). For each sample, place of residence is found to be mildly associated with vote intention, the average eta = .13. The strength of the association varies little (from eta = .09 to eta = .20), manifesting no downward movement. Therefore, region appears to have sustained its traditional hold on the voter, facilitating the advance of some parties, while retarding the progress of others. Ideology Has the long-standing division of Left from Right begun to lose its force? Is it true, as Macridis (1975: 60) contends, that "the French public has broken off its old ideological mooring and is floating into unchartered waters"?

440

Patterns of Stability

According to an abundance of survey data, any declaration of an end to ideology among the French electorate is premature. On the contrary, the distribution and impact of ideological thinking appears remarkably stable throughout the post-World War II period. In this section, I first look at the distribution of ideological opinion in the French public. Then, I go on to assess the impact of these ideological differences on party preference. Because of the historic salience of the Left-Right dimension, self-placement on such a continuum is not a particularly difficult task for French voters. At different times, they have been presented with two basic types of survey questions aimed at assessing their ideology. The first asks respondents to locate themselves in one of three categories, for example, " D o you judge yourself to be of the 'Right,' of the 'Center,' or of the 'Left?' " The second is more precise, asking that they place themselves on a numerical scale, for example, "In political matters people talk of 'The Left' and 'The Right.' How would you place your views on this scale?" [Interviewee is shown a ten-point scale.] The answers to the second item can easily be made com­ parable to those from the first item by grouping the responses into three categories: Left, Center, and Right. With this adjustment, we can examine the distribution of ideology among the French public around election time during the Fourth and Fifth Republics, as measured in the surveys of Table 14.5. (Regrettably, a relevant survey for the 1958 legislative election could not be located.) One observes that the distribution of ideological opinion shows little change over time, with an absence of any trend. More specifically, contrary to the influential preliminary findings of Bruce Campbell (1976), the French electorate has not become more bipolar in its political thinking. That is, one does not find a depopulation of the Center TABLE 14.5 The Distribution of Left-Right Ideology among the French Electorate, 1946-1981

Ideology Left Center Right Total

1946'

196?

1966

1973

1978

1981

35% 39 26 100%

34% 49 _17 100%

35% 41 24 100%

32% 48 _20 100%

36% 44 21 ίθΤ%

30% 54 J6 100%

(mean = 34%) (mean = 46%) (mean = 21%)

SOURCE: See Appendix 14.B.

'These distributions are derived from items that ask respondents to place themselves in one of three categories. The other distributions are based on items that ask respondents to place themselves on a numerical scale running from Left to Right.

441

France: The Stalled Electorate

position as people gravitate to the Left and Right poles. Rather, the Center position reliably attracts about 46 percent of the electorate, while the Left and Right positions maintain their shares of about 34 percent and 21 percent, respectively. In fact, the distribution of political opinion, far from being bipolar (or bimodal), has continued to form an almost perfect unimodal normal curve. In Figure 14.2, detailed frequency distributions of the ideology variable are presented, using responses to the ten-point, Left-Right, self-placement scale administered in the legislative election years 1973, 1978, and 1981. Visually, the curves appear normal. This impression receives quantitative confirmation. The skewness statistics are, respectively, .22, - .04, .28, in­ dicating almost no departure from normality. Even assuming that the distribution of ideological opinion has changed little, a critic might argue that the impact of ideology on politics has lessened. For example, citizens' perceptions of themselves as occupying a Left, Center, or Right position could be increasingly less relevant in shaping their party preferences. Perhaps, then, voters are less responsive to their ideological commitments than they once were. Table 14.6 shows how those who placed themselves ideologically on a ten-point scale said they would vote in 1978. The results are rather remarkable. Fully 97 percent of those who perceived themselves as ideologically Left-leaning expressed a preference for one of FIGURE 14.2 The Distribution of Left-Right Ideology among the French Electorate, 1973, 1978, 1981a 35-

30-

25-

£ 20υ

/ /

Λ \ /

N

-l

15 -

\>"·

//

10-

\ s

\

5-

\

\ . \ \ . 1973 ^ O f c v < > 1981 ^ V ^ · · 1978 1 Left

2

3

4

5

6

7

8

Score on Left-Right Self Placement Scale

SOURCE: Euro-Barometer Series.

9

10 Right

442

Patterns of Stability TABLE 14.6 Ideology and Vote Intention, 1978 Vote Intention

Ideology" Left

Center

Right

Left parties Communist PSU Socialist Left Radical

24% 7 65 1

Center parties Center

1 2

16

5

1

23 20

39 42

100% (279)

100% (153)

Right parties UDF Gaullist

Total (N)

ϋ

100% (395)

5% 3 30 3

1% 2 10 0

SOURCE: Euro-Barometer 10 (October-November 1978). a Left ideology = those who placed themselves at 1, 2, 3, or 4 on the ten-point ideology scale; Center ideo­ logy = those who placed themselves at 5 or 6; Right ideology = those who placed themselves at 7, 8, 9, or 10. the Left parties, either Communist, PSU, Socialist, or Left Radical. In con­ trast, only a scattered few of them (3 percent) opted for one of the CenterRight parties. Obviously, in the contemporary French electorate, ideology is a potent structuring device for deciding how to vote. Indeed, for the French voter, ideology appears to serve as a "standing decision," functioning much as party identification does. (This possibility is explored further in LewisBeck, 1983.) The overall strength of the association is manifest when one examines the correlation between the ideology variable (with the full ten-point scale) and vote intention (measured as before). For this 1978 survey, r = .70. Moreover, across the time period for which data are available, this correlation has been consistently high, average r = .66 (see Table 14.3, line 4). Indeed, the association is markedly stronger than that of the other traditional forces eval­ uated here. For instance, it appears to have a much greater impact than religion, which has the second highest average correlation coefficient. Without doubt, then, ideological self-placement has been, and remains, a major force shaping the party choice of the French voter.

France: The Stalled Electorate

443

CONCLUSION

Under the Fifth Republic, the attitudes and behavior of the French electorate appear to have changed little, at least with regard to the critical variables studied here. The extent of party identification has seldom varied from Fourth Republic levels, which were found to be surprisingly high. This is clear evidence that partisan dealignment has yet to take place. Further, partisanship still seems determined by traditional forces to the same degree that it always was. That is, class, religion, region, and ideology have maintained their substantial influence on party preference. Moreover, these variables have even managed to sustain their importance relative to one another. By comparing the magnitudes of the correlation coefficients, we see that ideology is consistently first in importance, religion second, class third, and region fourth. Affirmation of the persistence of these old patterns of French electoral behavior is highly valuable in its own right. Furthermore, it has implications for a popular hypothesis about Fifth Republic politics. That is, the belief that a fundamental partisan realignment of the French electorate has occurred. According to Campbell (1974: 73), deciding whether or not there has been such a realignment is "the essential question" for contemporary French political studies. Several researchers have concluded that a realignment has taken place (Cameron, 1972: 23; Cameron and Hofferbert, 1973: 78, 93; Inglehart and Hochstein, 1972: 353). However, the results reported here indicate that one has not. In the United States case, testing an hypothesis of realignment is relatively simple. For instance, if one suspected that workers were detaching themselves from the Democrats and aligning with the Republicans, one could correlate class (measured by occupational category) with vote (Democrat or Republican) across several election surveys, and observe whether the strength of association took a downward path. In the French case, such a test for realignment is complicated by the larger number of parties, but the principle remains the same: correlate class with vote across several election surveys. This exercise is presented in Table 14.3, which was discussed earlier. No pattern of partisan realignment along class lines is detectable. While minor fluctuations from election to election are observable, one sees no lasting rise or fall in the predictability of partisanship from class standing. Throughout the Fifth Republic, the members of a particular social class have continued to gravitate toward like parties. For example, the fidelity of the working class to parties on the Left has hardly varied across this twenty-four-year span. Similarly, the cross-time correlations of vote intention with the other variables of religion, region, and ideology yield no evidence of realignment.1 1

A skeptic might doubt this conclusion because of the measurement of the dependent variable, which orders vote intention from Left to Right Recall, for example, that in the 1978 survey of Table 14 6 the variable is scored as follows. 1 = Communist party, 2 = Unified Socialists, 3 = Socialists, 4 = Left Radicals, 5 = Social Democratic Center, 6 = Union for French

444

Patterns of

Stability

It has been argued that with the 1981 election of Fran§ois Mitterrand to the presidency and the coming of the Left to power for the first time during the Fifth Republic, realignment has finally arrived in France. These events, it is commonly supposed, mark a turning point. The dominant view is summed up in the following comment, which appeared on the front page of the New York Times (June 14, 1981) the day after the first round of elections for the National Assembly: "Today's vote confirmed the major shift in France's politics that began with the election as President last month of the Socialist leader Fransois Mitterrand." Does this victory mean that the French electorate is no longer "stalled" in its old patterns? More specifically, does it represent an important and lasting change in partisan attachment from rightist to leftist parties? And was it caused by a voter realignment, such as a movement of the middle class to the Left, or of the religious away from the Right? To these queries, I must give negative answers. The 1981 returns can be explained within a traditional framework for the understanding of French politics. First, in terms of the popular vote, the actual departure from past trends was not dramatic. Second, the 1981 data in Tables 14.3 and 14.5 evidence no re­ alignment. Third, the Mitterrand triumph can be accounted for quickly, once one considers the usual response of French voters to adverse economic con­ ditions. Next, I develop each of these contentions, in turn. Socialist-led coalitions have governed France before, in the Third and Fourth Republics. While no leftist parties have previously held power under the Fifth Republic, they have always managed to capture a sizable share of the popular vote. In fact, by calculating the percentage of total vote in the legislative election (first ballot) going to all parties on the Left, large and Democracy, 7 = Rally for the Republic While the middle class, for instance, has continued to favor parties on the Right end of the scale, these parties have changed some over time, as Table 14 1 made clear Thus from one survey to the next, a different party may be assigned the same numerical score Since, then, the precise meaning of the scale scores vanes from time to time, the comparison of these correlations across surveys could be misleading A further objection involves the Left-Right scale itself One might disagree with my particular ranking, for example, maybe the Unified Socialists, rather than the Communists, should be scored " 1 " Or, more seriously, perhaps Left-Right is not the relevant dimension along which to array the parties In sum, therefore, if the parties had been ordered in some other way, the results might differ from Table 14 3 Is the pattern of stability observed in Table 14 3 brought about by the way the dependent variable is ordered'' To test this possibility, I turned to a measure of association that does not require this ordenng Cramer's V, a nominal measure of association, requires no ordenng of the vanables It simply indicates the degree to which categones on X are related to categones on Υ For example, relating social class and vote intention m the 1978 survey of Table 14 2, V = 21 (r = 23) This shows that social class is associated with vote intention, regardless of how the vote intention variable is scored or ordered If one substitutes a Cramer's V for each of the correlations in Table 14 3, one still finds an identical pattern That is, the V coefficients hardly vary from survey to survey, indicating that the impact of traditional forces on vote intention has remained virtually unchanged dunng the Fifth Republic

France: The Stalled Electorate

445

small, one discovers that their strength never falls below 41 percent, the figure for 1968 (see Lewis-Beck and Bellucci, 1982: 104-105). Actually, for some election years a parliamentary victory for the Left was even expected. Before the 1978 legislative elections, the pundits, the press, and the pollsters were generally convinced that the Union de la Gauche, an alliance led by the Socialists and the Communists, would receive a majority of the votes. For instance, the preelection surveys of the leading French polling organization, IFOP, consistently forecast a majority for the Union de Gauche. In sum, at different moments in the history of the Fifth Republic, the Left has been just short of electoral success. Mitterrand himself almost became president in 1974 when, running as the joint candidate of the Left, he garnered 49.2 percent of the vote in his second ballot race against Giscard. For 1981, in the same contest, he did win, with 51.8 percent of the total presidential vote. While this increase in vote share of 2.6 percent did bring him into office, it hardly represents a dramatic shift in partisanship at the mass level. On the contrary, what stands out is the relative stability of his electoral support across this seven-year period. Focusing on 1981 survey data, the continuity with past attitudes is striking. According to Table 14.5, no ideological shift to the Left accompanied the Mitterrand success. Rather, about 30 percent of the electorate placed themselves ideologically on the Left, a figure quite close to the 34 percent average for the entire thirty-six-year period. Furthermore, the correlations of traditional factors (class, region, ideology) with vote intention are almost identical to their mean values for Fifth Republic legislative elections in general (see Table 14.3). By these critical measures, then, realignment is completely absent from the 1981 election. What, therefore, caused the Mitterrand victory, which, in the words of Le Monde editor Jacques Fauvet (Wall Street Journal, May 13, 1981: 25), is not only "that of a new party but also that of all the left"? In my opinion, the most plausible explanation is a simple economic one. That is, adverse economic conditions, without precedent in the Fifth Republic, pushed people to vote Left. Until recently, explanations of voting have neglected the impact of economic conditions. There is now a good deal of work on this topic for the United States case, and research on France is also appearing. Studies at both the aggregate and individual levels of analysis consistently show that a worsening economy increases the vote share of leftist parties in legislative elections (Lewis-Beck, 1983; Lewis-Beck and Bellucci, 1982; Rosa and Amson, 1976). For example, four Euro-Barometer surveys (from 1973, 1975, 1976, and 1977) ask respondents how satisfied they are with their incomes. The correlations of this income satisfaction variable with the vote intention variable are, respectively, r = .20, r = .20, r = .21, r = .19. One observes that, as with the traditional forces examined in Table 14.3, this association appears very stable. As the national economy deteriorates, then, one would expect

446

Patterns of Stability

more individuals to express dissatisfaction over their incomes, and correspondingly more to vote for the Left. At the arrival of the 1981 presidential election, the economy was in worse shape than ever before under the Fifth Republic. While Giscard's government gave priority to the fight against inflation, it failed to slow the rate, which was running at 14 percent annually. Moreover, unemployment had reached 7 percent of the work force, a percentage well above normal for France. Inevitably, economic dissatisfaction was widespread, and this translated into additional votes for Mitterrand. Thus in the 1981 presidential election, economic malcontents delivered to Mitterrand the few extra votes he needed to make him a winner instead of a loser, as in 1974. In sum, the 1981 election results do not undercut the conclusion that the French electorate is "stalled" in past habits. The Mitterrand victory has not yet provided evidence of a lasting, dramatic shift in partisanship in favor of the Left. Further, it is not necessary to attribute the Mitterrand success to a process of realignment. Indeed, traditional factors such as class, region, ideology, as well as economic dissatisfaction continued to operate with undiminished force in determining voter choice. What did change from previous elections was the number of people who are economically dissatisfied. These are the people who brought Mitterrand to power. Will the French electorate eventually realign? Of course, it is difficult to say. What impresses me most about mass attitudes and behavior in France during the Fifth Republic is the persistence of the old political cleavages, despite all sorts of social, economic, and political change at the institutional and elite levels. Why such resistance, then, in the face of so much change? A clue comes from V. O. Key's (1959: 209) speculations on the United States experience, where he argues that realignments may have occurred because of "the non-doctrinaire quality of American politics." Perhaps, then, the "doctrinaire," or "ideological," quality of French politics has prevented any lasting electoral realignment. Certainly, the data examined here fully reflect the importance of ideology for the French public. For instance, French voters place themselves willingly on a Left-Right scale, with even more readiness than they express identification with a party. And, as has been shown, this self-perception is their primary tool for deciding among the parties. Indeed, ideological identity appears generally to serve as the French voter's compass, pointing a course across the turbulent sea of French politics.

APPENDIX 14.A

The following question wordings and sources were used for the estimation of party identification in Figure 14.1 1948 ' 'In your opinion, among the following political parties or groups, which is the one which, since the Liberation, has brought the most favorable policies to the affairs of the country?" [An Institut Frangais d'Opinion Publique (IFOP) survey reported in Sondages 10, no. 16 (October 15, 1948).] 1955 "What is the party or political group which seems to you the most capable of carrying out your wishes?" [An IFOP survey reported in Sondages 17, no. 4 (1955): 17.] 1958 "What is the political party to which you feel closest?" [The question was asked in September, November, and December, the three waves of The 1958 French Election Study.] 1962 "Which is the political party you feel closest to among the following?" [An IFOP survey reported in Sondages 25, no. 2 (1963): 59.] 1967 A first question was, "To which party do you usually feel the closest?" Then, a follow-up question was asked, "Do you feel that you are very close to this party, fairly close, or not very close?" [French National Election Study, 1967.] 1968 "Would you tell me to which political party you feel closest, with the help of this card [which lists the parties]?" [A survey reported in Inglehart and Hochstein, 1972: 352, 371.] 1975-"Do you consider yourself a supporter of any particular political party? 1981 If so, do you feel yourself to be very involved in this party, fairly involved, or merely a sympathizer?" [Euro-Barometer Series. In the years when two surveys were available, their results were averaged.]

APPENDIX 14.B

The following question wordings and sources were used for the estimation of the distribution of ideology in Table 14.5. 1946 "During the coming years, do you want government policy to be oriented more to the Left or more to the Right than it has been to date?" [An IFOP reported in Sondages 8, no. 14 (July 16, 1946): 166.] 1962 "Do you judge yourself to be of the 'Right,' of the 'Center,' or of the 'Left?' " [An IFOP survey reported in Sondages 25, no. 2 (1963): 109.] 1966 Respondents were asked to place themselves on a Left-Right continuum. [Surveys reported in Deutsch et al., 1966: 13, 106-107.] 1973-"In political matters people talk of the Left and the Right. How would 1981 you place your views on the scale?" [The respondent is shown a tenpoint scale running from "Left" to "Right."] [Euro-Barometer Series. In 1978, when two surveys were available, their results were averaged.]

SIX

Conclusion

15. Political Forces and Partisan Change RUSSELL J. DALTON, S C O n C. FLANAGAN, PAUL ALLEN BECK The Florida State University, Tallahassee, Florida

A Common Theme has run through most of the chapters in this volume. Electoral alignments are weakening, and party systems are experiencing increased fragmentation and electoral volatility. Moreover, the evidence suggests that the changes in all of these nations reflect more than short-term oscillations in party fortunes. This decomposition of electoral alignments often can be traced to shifts in the long-term bases of partisan support—party identification and social cleavages. Virtually everywhere among the industrialized democracies, the old order is crumbling. In many, perhaps most, of the case studies in this volume, evidence of both dealigning and realigning influences is presented. The complexities of electoral change and the overlapping nature of realignment and dealignment processes are most apparent in the American case, where separate chapters have presented evidence of both phenomena. Nevertheless, from the perspective of the early 1980s, a predominant pattern of electoral change can be identified often. Processes of realignment have been highlighted in Japan, West Germany, and Italy. In other cases—The Netherlands, Britain, Scandinavia, and Spain—party instability follows a pattern of at least temporary electoral dealignment. The frequent similarities in the patterns of electoral change are all the more striking given the wide variation in the historical, political, social, and economic conditions of the nations we have examined. We have studied electoral politics in one of the newest democracies, Spain, and in the oldest, Great Britain. These nations vary from the conservative hegemony in Italy to the leftist dominance in Scandinavia, from the complexity of the Dutch multiparty system to the American two-party system, from continuing economic growth in Japan to economic decline in Britain. In accounting for these partisan changes, we must walk a fine line. On the one hand, events can be explained by the unique political experiences of each nation. The 1968 May Revolts in France, for example, might be traced to

452

Conclusion

student demands for dormitory visiting privileges, and the government's overreaction to escalating conflict. One should not lose sight of these national idiosyncrasies. On the other hand, the common pattern of partisan volatility across most industrial democracies necessitates a search for general explanations. We therefore might discount the uniqueness of each event and search for common causal processes at work. From this perspective, the May Revolts in France might represent one expression of basic social and political tensions that were building in many industrialized nations, and the demonstrations at Nanterre only provided the match to set these tensions aflame (Inglehart, 1977: 262-290). Such a general approach to political change obviously must be less detailed and not as finely tuned to nuances of unique national experiences. However, if this general approach provides an adequate explanatory theory, we have learned much more than we would from the simple sum of case studies. Several commonalities can be observed among the nations considered here. For instance, the political agenda is broadening to include a new set of issues. A heightened concern for the environment is shared by Japanese, European, and North American publics. The same new life-styles and social concerns have emerged in a number of advanced industrial democracies. Discussions of women's rights, social equality, and a new world order are heard in Bonn, The Hague, London, and other capitals. Indeed, these strikingly parallel issue interests suggest that we are witnessing a cross-national process of change, rather than coincidental national similarities. Political change, however, has not been a one-way street. As new political movements have gained strength, they often have stimulated large and vocal opposition. The expansion of the welfare state has stimulated antitax sentiment; the women's movement has created a traditionalist reaction; even calls for greater citizen participation in decision-making have evoked concerns about the excesses of popular control. The rising tide of leftist movements seems to be matched by a new vehement Right—Thatcherism in Britain, neoconservatism in France and Italy, Reaganism in America. These conflicts are likely to be a continuing feature of industrial societies, as new social alignments emerge to represent and oppose the concerns of advanced industrialism. Just a decade after celebrating an end to ideology, ideological conflict seems to be consuming the industrialized world. The development of direct forms of political action is another force reshaping contemporary politics. As Americans surround Seabrook, Germans storm Brokdorf, Britons march on Windscale, and Japanese protest Minamata. Citizen action movements are now a common part of the democratic political process (Rochon, 1983). In addition, there has been a general drive to open the political process to greater citizen participation at all levels. There are also similarities in the patterns of electoral change described in this volume. The expansion of educational opportunities and the growth of

Political Forces and Partisan Change

453

the new middle class—two hallmarks of advanced industrialism—are eroding the class basis of contemporary politics. Furthermore, voters whose loyalties were formed in the cauldron of class strife during the 1920s and 1930s have virtually vanished from the electorate, leaving the stage to those whose political world views were formed in less class-polarized times. This trend provides one indicator of a general decomposition of electoral alignments in the last decade. Moreover, these changes are most noticeable among the postwar generation, which has been most exposed to the new issues and styles of democratic politics and least affected by the old order. In America and Britain, the young have contributed disproportionately to partisan dealignment; in West Germany, The Netherlands, Spain, and Scandinavia, the most pronounced decline in class voting has been among the youngest generations. As a result of these trends, the traditional distinctions between Left and Right, and even the meaning of these terms, are in a state of disarray. Indeed, there is evidence that a New Politics issue agenda is beginning to restructure both elite and mass political orientations (Inglehart, this volume: Chap. 2). Similarly, in an analysis of thirteen advanced industrial democracies, Dalton and Flanagan (1982) find that the public's Left-Right orientations are now almost equally related to social class and postmaterial values. Moreover, for the young and highly educated, the ideology scale was much more strongly associated with values than with class. In other words, the younger and more educated voters in advanced industrial democracies are beginning to define political competition more in terms of value cleavages than class cleavages. Along with these changes in mass perceptions and the shifts in the relative salience of the Old Politics and New Politics issue agendas have come some reordering of social group coalitions. Despite their differences, unions and management occasionally join to fight the opponents of nuclear energy. Farmers and students sometimes ally to oppose industrial development projects. Fundamentalist blue-collar and white-collar workers unite to oppose changes in moral codes. A New Left and New Right are emerging, which possibly may restructure social group alignments and party coalitions. Our position is that these national parallels are not all unique events, but often have common origins in the weakening of ties to the old political order as well as the interrelated changes stimulated by the social and economic forces described as advanced industrialism. Even if a process of political change has been set in motion, the long-term partisan consequences of this trend remain unclear. Part of this uncertainty arises because a distinct new basis of ideological cleavage has not yet emerged. Issue positions are not always consistently or coherently presented. The salience of the new issues is often limited and varies with socioeconomic conditions. At this point, the new social issues are neither red (Marxist) nor green (environmentalist), but a rainbow of alternative ideologies. This situation may change as new parties and social groups emerge to represent and articulate these new issue interests,

454

Conclusion

and integrate them into a coherent ideological program. At the present time, however, the broader ideological consequences of a changing issue agenda are unclear. Another uncertainty is how democratic party systems will respond to these new political forces. Two theories of party systems and partisan change have been intertwined in this volume, and it is best to reestablish the distinction. The social cleavage model attributes change in party systems to the rise and eclipse of social cleavages. The transition between electoral eras is marked by a realignment, as parties and the electorate adjust their positions along a new cleavage dimension. The functional model posits that recent electoral trends are the product of changes in the social and political roles served by parties and party identifications. This model implies that the historical decline in the relevance of parties to the political process and individual citizens may lead to continuing partisan dealignment. The evidence presented in this volume suggests that both processes of change are at work. From our present perspective, however, it is often difficult to distinguish between evidence of realignment and that of dealignment, since both predict at least a temporary decomposition of the party system as the old partisan order ends. It is not as yet clear, then, which of the two processes of change will be the predominant one. This is an important question as we look toward the future. The two models seem to suggest strikingly contrasting scenarios—a new stable realignment versus a permanent dealignment and perhaps a withering away of political parties as we know them. Thus in order to understand where electoral trends may be leading, we need to examine both scenarios. THE SOCIAL CLEAVAGE MODEL

One approach to party systems attempts to explain their development and persistence in terms of the historical conditions of national and socioeconomic development. Foremost in this tradition has been the work of Seymour Martin Lipset and Stein Rokkan (1967), who discuss political cleavages as the consequence of two successive revolutions in the modernization of Western Europe: the National Revolution and the Industrial Revolution. Each gave rise to different social cleavages. The National Revolution stimulated a center-periphery cleavage, between the dominant national culture and resisting ethnic, linguistic, or religious subject populations in the provinces and the peripheries. It also created a church-state cleavage, between the centralizing, standardizing, and mobilizing nation-state and the historically established corporate privileges of the church. The Industrial Revolution also produced two social cleavages. The land-industry conflict matched the economic and agricultural interests of landed elites against the economic concerns

Political Forces and Partisan Change

455

of the rising class of industrial entrepreneurs. The more recent owner-worker cleavage reflected the Marxian class struggle within the industrial sector. These four dimensions of cleavage were offered by Lipset and Rokkan as a framework for explaining the origins of party alignments in Western democracies. Social movements related to these cleavage dimensions developed during the nation-building period of the last several centuries. As these movements emerged, they sought and ultimately won access to the political decision-making process, prior to the democratization and mass mobilization of Western political systems. Consequently, the formation of political parties tended to institutionalize the existing elite coalitions, providing a framework for modern party systems. At the same time, Lipset and Rokkan further argued that the last of these four cleavages, the class cleavage, played the predominant role in structuring contemporary party alignments, because class issues were the most salient ones during the extension of the suffrage around 1900. With few exceptions, they maintained, the party systems of the 1960s still reflected the cleavage structures existing over a half century earlier. Although the Lipset-Rokkan model emphasized the institutionalization and freezing of cleavage alignments, the model also has dynamic properties. It views social alignments as emerging from the historical process of social and economic development. New alignments develop in response to major social transformations such as the National and Industrial revolutions. While the structure of cleavages is considered to be relatively fixed, the political salience of the various cleavages and patterns of party coalitions may fluctuate in reaction to contemporary events. The cleavage model suggests several ways in which the development of advanced industrialism may affect contemporary party systems. One method may be through the process of secondary mobilization described by Gordon Smith (1976). New social movements may realign or rejuvenate party systems without breaking down the primary cleavage structure. For example, Vincent McHale and Sandra Shaber (1976) maintain that increasing affluence and government planning have mobilized quiescent groups on the periphery of the French system who now demand access—thereby reviving center-periphery conflicts. A similar process may explain the resurgence of regionalism in Britain, Canada, and Belgium (Jacob, 1980). John Madeley (1977) claims that advanced industrial politics has led to a reintroduction of religious parties into Scandinavian politics. In Chapter 5 of this volume, Edward Carmines and James Stimson adopt a similar theme: conflict along racial issue lines reemerges in American politics as blacks are mobilized to full democratic citizenship. In short, new political movements might cause a resurgence or re-creation of old political cleavages. Several authors in this volume argue that advanced industrialism may produce a more fundamental change in the structure of party systems. A third

456

Conclusion

revolution, the Postindustrial, may create a new basis of social, and eventually partisan, cleavage (Bell, 1973; Inglehart, 1977). Lipset and Rokkan's work suggests that each previous social revolution has produced two new cleavage dimensions: between the old and the new order, and eventually among participants of the new order. We have seen evidence of the emergence of an industrial-postindustrial cleavage between proponents of the established industrial order and supporters of New Politics goals. For instance, issues of economic growth and nuclear energy find unions and management allied against environmental groups. The women's movement faces its strongest opposition from women supporting the traditional female role. In Chapter 2, Ronald Inglehart describes this cleavage as the conflict between material and postmaterial values, while in Chapter 3 Risto Sankiaho sees it as an establishment/antiestablishment battle. Other theorists have suggested possible social alignments that may structure the industrial-postindustrial cleavage. For example, Samuel Huntington (1974) predicts an intensification of the conflict between the declining manual workers and rising knowledge workers, the shrinking old middle class and expanding new middle class, and, finally, economically decaying central cities and affluent suburbs. The first of these cleavages may become the most explosive in the next two decades as the advent of robotics threatens to eliminate manual jobs faster than technical jobs can be created and workers retrained. Some analysts are predicting that in Japan, the leader in this new field, knowledge workers virtually will have replaced manual workers by the year 2000. While this prediction may be exaggerated, the potential for serious dislocations in the work force in the near future cannot be dismissed. Declining strata typically exercise a political clout beyond their numerical size or economic importance, due to their entrenched positions in the political power structure and their cohesion in the face of external threats. Thus these strata are likely to exercise vigorously their political influence and may constitute an important realigning force. Another emerging line of cleavage in advanced industrial societies runs between the public and private sectors. The number of white-collar workers on the public payroll has grown enormously at all levels of government. The struggle between increased wages for public employees and lower taxes for the citizenries is heating up in several countries. This conflict is further complicated because there is no objective standard of productivity in whitecollar work—debates over government efficiency are often based on philosophy rather than fact. As has been noted at several points within this volume, the New Left and Old Left are not necessarily compatible. In fact, they are defined by conflicting value priorities and represent different social group interests. Postindustrial values tend to oppose continued rapid industrial growth and emphasize quality of life issues. These New Left positions, however, strike at the heart of the

Political Forces and Partisan Change

457

Old Left's principal constituencies, labor and the poor, by inhibiting the growth of new blue-collar jobs and thwarting the ability of the poor to improve their condition through an expanding economic pie. As the New Left-Old Left cleavage becomes more clearly articulated, it will become more difficult to represent both preference schedules within the same party. Thus the likelihood of a partisan realignment will increase. A new set of issue cleavages within the rising postindustrial strata also may emerge if social change continues. These cleavages are as yet ill defined, but we can detect their outlines by identifying some of the contradictions inherent in postindustrial society. On the one hand, postindustrial values are associated with demands for the decentralization of decision-making, increased local autonomy, a curb on unrestrained development, and a belief that small is better. Postindustrial values stress psychic rather than industrial development, inner rather than outer growth. This thrust toward self-actualization is associated in part with efforts to gain control over one's destiny, either by a return to a simpler, more self-reliant life-style or by demanding more participation in the political and economic decisions that affect one's life. It is also associated with a growing trend toward self-indulgence, the pursuit of leisure activities, permissive life-styles, and an erosion of the work ethic. This has led to what Goldthorpe and his associates (1968) have called the paradox of the affluent worker. They find that affluent workers prefer "intrinsically satisfying work" but still choose "instrumental work" that provides higher incomes as the necessary price for maintaining their high consumption lifestyles. This paradox is likely to result in a declining commitment to work as well as a decline in the psychic gratifications achieved through work. The countercurrent to these value changes is the growing dependency of advanced industrial society on a rising class of technocrats who must adhere to the industrial values of diligence, efficiency, and achievement to advance their careers and maintain the system. Yasusuke Murakami (1982) argues that these industrial values will still be essential for specialization, investment, and technological innovation regardless of how automated and computerized postindustrial society becomes. Thus value cleavages may emerge between the administrative and technical elites in business and government as well as the affluent blue- and white-collar wage earners with high economic security and low administrative responsibility. This widening value cleavage may increase popular antipathy toward big business and big government, while the dynamics of advanced industrialism increasingly bureaucratize society. Even from our present historical vantage point, therefore, we can detect a potential for a number of new cleavages to emerge both between the rising and falling strata and among the rising strata in advanced industrial societies. Any of the new social cleavages we have discussed must cross many hurdles before becoming integrated into the political process. As Lipset and Rokkan noted, democratic party systems have displayed a surprising stability and

458

Conclusion

resistance to change. In outlining these hurdles, the social cleavage model sheds light on the preconditions for partisan change—the requisites for transforming new issues and social movements into partisan alignments. The model argues that social movements must develop an institutional base in order for a new social cleavage to become integrated into the party system. This base is important because it provides needed political resources: the foot soldiers and financing for political campaigns. An institutionalized social movement also is more likely to link specific issue interests into a structured ideological framework that can provide the programmatic basis of a new cleavage. Leadership is another critical factor for any new social movement. The success of the Social Democratic party in Britain, the New Liberal Club in Japan, and Glistrup's party in Denmark, for example, must be attributed partially to the personal appeal of their leaders. In some cases, a popular political figure may build an organizational base from the ground up. However, new social movements often are centered outside the circle of established political interests and therefore face difficulties in attracting well-regarded political leaders. It takes considerable time for a new movement to develop nationally recognized and respected leaders. For instance, a number of environmental parties throughout Europe have suffered from the lack of popularly accepted national spokespersons. The content of a new issue agenda also will affect the potential for partisan realignment. Walter Dean Burnham (1980) has argued that only position issues—with pro and con polarization—provide a potential base for partisan realignment. At best, valence issues can stimulate only temporary partisan surges (deviating elections) because there is a consensus on goals between political actors. However, a consensus on goals often hides a dissension on measures. In the context of crisis, a new means, such as the New Deal program, may precipitate a partisan realignment, particularly if it is successful in achieving a desperately desired end. Advanced industrialism generally has spawned new valence issues: environmental protection, equal rights, and citizen participation. Even when there has been a conflict between these and other goals such as economic growth, partisan conflict generally is mediated because both sides disagree only on the means to achieve both goals. Thus, Herbert Asher (1980) claims that the Democrats were able to defuse the "social issue" in the 1960s because it was a valence issue. For these valence issues to have a realigning impact, they must either spawn a crisis or become reframed into position issues. There are signs suggesting that some New Politics goals are coming closer to the requirements of position issues. Nuclear energy may be the best contemporary example. Established party and political elites (both liberal and conservative) often have been strong supporters of nuclear energy as a necessary cost of economic growth. In contrast, New Left groups frequently connect the nuclear

Political Forces and Partisan Change

459

energy question to pollution, nuclear defense issues, and increasing government and technocratic control—and oppose it on those grounds. The temporal pattern in which new issue conflicts develop is another factor affecting partisan change. Socioeconomic forces can produce considerable pressure for electoral change, but party systems also can display considerable resistance to new policies that would move them on to untested ground. It appears that a dramatic event or issue often is needed to unify public opinion and force the parties to chart a new course. For example, depressions are cited as the catalysts for the 1896 realignment in the United States and for the 1930s' realignments in Western Europe and the United States. The dramatic event to stimulate a New Politics realignment has not yet appeared in any nation. Such an event might occur—perhaps a major nuclear or environmental catastrophe—and swiftly translate the forces of advanced industrialism into clear electoral changes. This situation, however, can be neither expected nor predicted; otherwise, political elites would take steps to avoid catastrophe. While a critical issue is needed for a critical realignment, Chapter 5 by Carmines and Stimson points out that realignment does not require such a catalyst. Over a longer time span and with smaller incremental change, political forces gradually can produce a secular realignment in electoral forces. Such secular changes appear to have implemented the New Deal realignment in a number of American states (Key, 1959; Sundquist, 1973). Evidence of incremental change already is apparent in Japan, The Netherlands, West Germany, Italy, the United States, and several Scandinavian nations. Any new political movement faces the institutional and political hurdles we have just described. This is a major factor contributing to the longevity of present cleavage alignments. Lipset and Rokkan (1967) observed that the present four cleavage alignments were able to surmount these hurdles because they developed before and during the mobilization of mass electorates in Western societies. The development of a new cleavage in an already mobilized system is presumably much more difficult to achieve because political lines already are drawn and there are fewer untapped bases of electoral support. Still, a major partisan realignment may be possible. Advanced industrialism already has developed substantial pressures for political and partisan change, as described in this volume. In addition, the recent decomposition of electoral alignments provides a pool from which voters and elites can be mobilized, and may be a precursor to a realignment. These dealigned voters also are drawn more heavily from the groups most attuned to New Politics concerns— the young, the better-educated, and the new middle class. In short, one can see signs that the structure of democratic party systems, frozen for so much of our lifetime, is beginning to thaw. Several New Politics groups already have participated in elections and have won legislative representation. While still a minority, their political impact is growing. As democracies move further toward a truly postindustrial era, the force of change

460

Conclusion

may be sufficient to generate a new basis of partisan alignment throughout the democratic world. THE FUNCTIONAL MODEL

While the social cleavage model has considerable appeal, an entirely different framework can be used to explain recent changes in democratic party systems. The social cleavage model predicts a temporary partisan dealignment as a natural occurrence at the end of an electoral cycle. Scholars have observed evidence of weakening party roles at the end of every party system alignment in the United States, just before a surge in the importance of parties associated with the subsequent realignment (Beck, 1979; Clubb et al., 1980). The functional model, however, views the recent evidence of these phenomena in a different light. Rather than a lull between an old and a new party system, this alternate model sees recent voting trends as the beginnings of a fundamental change in electoral politics. The functional model examines parties and party systems in terms of the social, political, and electoral functions they perform, and focuses on how these functions are changing over time. This model has been applied at both the macro and micro levels. The macro-level research concentrates on the functions that are performed by the party system as a whole. The historical record often documents a decline in the social and political roles of parties during this century (Flanagan and Dalton, 1984; Epstein, 1980; Maisel and Cooper, 1978; Burnham, 1970). The decline has been clearest in the United States; the parties have lost their patronage to government social service agencies, and the introduction of open primaries and nonpartisan elections, as well as government financing for campaigns, have weakened the parties' dominance in electoral politics. In Europe, the declining membership of large mass-based parties has signaled a similar phenomenon. Other trends have had a more general effect on party systems in Europe, North America, and Japan. For instance, the growth of government and increasing bureaucratic autonomy have handicapped the ability of parties to control the governmental process (Katz, 1982). In addition, there has been a long-term decline in the parties' control of political information. Fewer and fewer newspapers remain affiliated with a specific political party. The development of the electronic media—especially television—has further eroded the parties' role as providers of information. The mass media also are assuming several other functions that have been handled traditionally by political parties—defining the issue agenda, oversight of government agencies, and criticism of government policies. These trends have been moderated in Europe because of government control of the electronic media; but inroads by the media should increase as new technologies (satellite and cable television)

Political Forces and Partisan Change

461

increase the media's independence and influence. Advances in communications and transportation also have undercut much of the need for party precinct work and are shifting campaigns toward personally managed contests. Perhaps the clearest example of these trends is the new Social Democratic party in Britain. The party lacked a constituency organization, and so it has adopted computerized mass-mailing to appeal directly to potential voters, and party contributions can be charged to bank credit cards. Even the parties' interest articulation function seems threatened by the rise of single-issue groups and citizen action committees. In sum, these trends contribute to the overall decline of parties as political institutions. Just as the role of parties at the macro level clearly has declined, advanced industrialism may be producing a parallel decline for individual partisanship. Politics is a complex undertaking, and most voters need a reliable source of political cues to guide their electoral decisions. As long as parties provide a framework for political debates and electoral politics, many voters will have a functional need for the political cues provided by long-term partisan predispositions. However, the need for such cues declines as the political skills and resources of the voters increase (Shively, 1979; Borre and Katz, 1973). The dramatic spread of education is increasing the sophistication of modern electorates, more voters are now able to deal with the complexity of politics and make their own political decisions without relying on external cues. In addition, the increasing volume and diversity of the information supplied by the mass media is substantially lessening the cost to the average voter of acquiring political information. In a much less constructive fashion, the media might facilitate greater elite emphasis on unmediated one-on-one contacts with voters. Despite its potential for facilitating more sophisticated citizen involvement, television also offers the possibility of a trivialized electoral politics in which video style outweighs substance in campaigning. Thus the functional value of long-term partisan predispositions may be decreasing for several reasons. The decline-of-parties thesis foresees a progressive decomposition of electoral alignments, leading to popular detachment from the parties and greater reliance on issues, candidate images, or ideology as a basis of voting decisions. The definition of dealignment also might be extended to include an expansion of citizen participation beyond the partisan contests of elections. The average citizen probably is better educated, better informed, and more sophisticated than most of the party activists who were entrusted with political decision-making in the early 1900s. At the same time, the expansion of the modern state means that politics has a more visible and direct influence on people's lives. In this context, it is difficult to argue that occasional opportunities to cast a single vote for a prepackaged party program provide an acceptable level of citizen input. One could more easily claim that by focusing

462

Conclusion

participation on voting, the parties are seeking to protect their privileged positions in the political process and actually limit citizen influence. A dealigned political system could provide the citizenry with additional, nonpartisan, participation opportunities more clearly related to specific issue and ideological concerns. Significantly, dealignment seems to be concentrated among groups whose greater political skills and resources enable them to be more functionally independent: the better-educated, the middle class, and the young (Dalton, 1984). For instance, in Chapter 10, James Alt notes that between 1964 and 1974 the weakening of British partisanship among citizens with advanced education was three times as great as among those with minimal education. Similar evidence was cited for Japan, the United States, Denmark, and several other nations. Moreover, the reinforcing effects of weakening partisanship and the institutional decay of the parties suggests a continuing process of partisan decline. It may be more than coincidental that the recent growth of independents is concomitant with an increase in direct forms of citizen action (Barnes, Kaase et al., 1979). These sophisticated independents are not constrained to party-organized forms of political action. Citizen initiatives, protests, and other forms of unconventional political behavior enable them to have a more direct influence on politics. In addition, these newly independent groups possess the necessary political skills to carry out more demanding forms of political action. The functional model thus yields a much different explanation of recent changes in democratic party systems. The model argues that parties and partisanship are becoming less relevant to the political system and the electorate. As institutions, the parties perform few of their former political functions. In addition, the growing sophistication of the mass public decreases the functional value of social or attitudinal partisanship. Most important, these are not temporary changes—as the cleavage model suggests—but preview a period of ongoing dealignment and direct citizen action. Still, it is premature to predict the withering away of parties even if many of their functions are curtailed and long-standing party identifications do weaken. Unless elections become purely contests of personalities, parties are likely to continue play an important role in structuring political choices, even in a purely dealigned and issue-oriented electorate. To move legislation, parliamentarians form policy factions that even in a very fluid system would constitute a basis for party labels. However, we can imagine a future where the institutional role of the party in the electorate withers as election campaigns increasingly are run as issue-oriented or candidate-centered campaigns that encourage voters to ignore party as a basis of choice. Then the development of parties and party systems will have run full circle, returning to the notion of "currents of opinion," which defined the essence of parties in seventeenthcentury Britain and the Federalist period in the United States.

Political Forces and Partisan Change

463

THE INSTITUTIONALIZATION OF PARTISAN CHANGE

Both of the aforementioned models are compatible with most of the partisan changes descnbed in this volume In the short-run, both models predict the present situation of partisan decomposition, rising partisan tensions, and distrust of the existing parties Individual chapter authors have varied as to which theoretical framework explicitly or implicitly underlies their analyses, and the evidence to decide between these two models is still inconclusive We can, however, continue to explore both models in terms of how political change should progress For both models, the source of change is based primarily on pressures emanating from the mass public The rate of social change, the organizational base of new issue groups, and the other contnbutory factors discussed in the preceding sections will influence the amount of pressure being exerted However, the responses of the party and political systems to mass pressures will be crucial factors in determining the final outcome of this process Elite and institutional actions are necessary to translate mass pressures into partisan change Within either theoretical framework, elite actions serve as intervening vanables—to direct, encourage, or inhibit partisan change The nature of these intervening vanables differs between the cleavage and functional models, and so it is best to discuss the models separately Variations in the Social Cleavage Process The social cleavage model predicts a penod of partisan decay and confusion as new issue interests are integrated into democratic party systems ' After integration, party systems generally will function as they did dunng previous stable alignment penods At question, however, is how (and whether) incompatibilities between new and old cleavages are resolved and how the new issue agendas are integrated into the existing party systems Perhaps the most important set of intervening vanables involves the willingness of the established parties to represent new issue conflicts The role of political parties in aggregating and articulating interests should lead new social movements to the parties, but party elites often have considerable latitude for action in deciding how to respond to these demands Party leaders must calculate the potential gain in new votes and the potential losses from their traditional constituency, the reaction of party activists and financial backers, and numerable other factors Leaders of a new social movement face similar calculations for example, the benefits of coalition versus independence, compatibility of goals, and compatibility with the existing party leaders The crucial choices, at least in the short term, are in the hands of political elites If the established parties embrace a new movement, through actions of the old leaders or take-over of the party by new leaders, then partisan 1

This section draws heavily upon the work of Charles Hauss and David Rayside (1978)

464

Conclusion

change will occur through realignment of the existing parties. However, when the established parties are not responsive, new movements may seek to form new parties to represent their views in the party system, or shun the party system altogether. In addition to calculations of electoral advantage and disadvantage, the responsiveness of the existing parties will be influenced by a series of structural constraints. For example, one characteristic affecting the responsiveness of the established parties is their ideological/pragmatic orientation. An ideological party (or party system) exists to represent specific political goals. Therefore, established ideological parties are not likely to be receptive to issue conflicts unrelated to their ideological groundings or to provide outlets for these new social movements. Indeed, some of the most hostile reactions to the New Left have come from the leadership of European Communist parties and traditional leftist ideologues. Conversely, the raison d'etre of pragmatic electorally oriented parties is to seek a broader base of voting support. These parties are therefore more likely than ideological parties to advocate new issue appeals that might add new elements to their coalition of constituencies. Thus a party system composed of pragmatic parties is more likely to represent new social movements within the existing party framework. A related but somewhat offsetting characteristic is party size. When we think of the prototype of the pragmatic party, the large catch-all parties in the United States, Britain, and Germany come to mind. However, party size presents barriers to the representation of new issues. While it is true that catch-all parties stand ready and willing to co-opt and absorb interests that have broad appeal and offer clear electoral advantages, they are generally unwilling to risk adopting minority positions as their own. Since many of the New Politics issues are narrow ones representing small constituencies, large parties are unlikely to take clear issue stands that may repel more voters than they attract. Smaller parties, if they are also pragmatically oriented, may find the strong support of a new movement or small constituency more attractive. Leadership roles are also more easily captured by new movements in the organizational setting of a small party. Thus party size may exert an independent influence on the incorporation of new issues by an established party. Another relevant characteristic is the distinction between cohesive hierarchical parties and loosely structured parties. The highly cohesive European parliamentary parties place a necessary premium on party unity and disciplined voting. This hardly provides a fertile ground for experimentation, independence, and political change. New social movements may have difficulties working within the established parties in such an environment. A likely result in this situation is the fragmentation of the established parties, with splinter parties representing the new issue interest—for example, the Radical party in The Netherlands, the PSU in France, the New People's party in Norway, and the New Liberal Club and the Social Democratic Federation in Japan. In

Political Forces and Partisan Change

465

contrast, loosely structured parties may be more receptive to new issues. Weak decentralized control encourages candidates to search out new voter groups, while still drawing on the institutional resources of the party. Dispersed polycentral party structures tolerate this intraparty diversity, and thereby encourage new interests to work within the established parties. When the established parties fail to provide an outlet for new social conflicts, these forces still may be integrated into the party system through the development of new parties. The chapters in this volume document a proliferation of new parties over the past decade. Glistrup's Progressive party's attack against the established parties in Denmark has had reverberations throughout Europe. In West Germany and France, new ecologist parties challenge the consensus on economic growth among the major parties. The emergence of D'66 in The Netherlands epitomizes the successful New Politics party; new conservative parties have been founded in Norway (Progressives) and in Canada (Social Credit). If the embryonic Social Democratic party of Britain survives, it will mark a breakthrough in one of the most stable European party systems. The potential for new parties is strongly affected by the openness of an electoral system to new contenders. An open system facilitates the formation of new parties to represent new issue conflicts. A closed system limits new actors; therefore, new conflicts that are unrepresented in the established parties are forced outside the existing party system. The openness of an electoral system generally is determined by institutional and legal factors. At one extreme, The Netherlands probably has the most open system; only 0.67 percent of the vote nationwide is needed to elect one member to the Dutch parliament. This system facilitates the representation of new issue interests through the formation of an entirely new political party, such as D'66. The Danish proportional representation system is somewhat less open, because it restricts parliamentary access to parties that receive at least 2 percent of the vote. However, this hurdle is sufficiently low that the Progressive party and Center Democratic party have been able to gain parliamentary representation for new political movements. The 5 percent hurdle of the West German proportional representation system is somewhat higher, and has excluded from recent Bundestags new parties on both the Right (NPD) and left (Die Gruenen until 1983). The Japanese electoral system presents new parties with a mixed representation threshold; 1 percent of the vote is generally sufficient for victory in upper house National Constituency races, but a minimum of 10 to 20 percent is needed to gain a seat in the more powerful lower house, due to its medium-sized multimember districts. At the upper extreme, majority rule systems based on single-member districts—as in the United States, Britain, and Canada—are relatively closed systems. The potential for new parties to win elections in these three nations is severely limited—unless they are regionally concentrated. The fate of Britain's Social

466

Conclusion

Democratic-Liberal Alliance in 1983 illustrates the difficulties "third" parties face in a first-past-the-post system. Thus political interests in these systems normally must seek representation within the established parties or find nonparty alternatives. Other institutional factors influence the openness of an electoral system. For example, in some cases party financing is publicly controlled and based on results of the preceding election. This handicaps new parties with a limited financial base. Similarly, access to government-controlled mass media may be based on prior election results that would restrain the growth of new parties. The qualifications for appearing on the ballot also may discriminate against new or minority parties. The initial difficulties faced by the 1968 Wallace candidacy and the 1980 Anderson candidacy in the United States highlight various means the established parties can use to block new contenders. These points suggest that the fewer the barriers (the greater the opportunities) for the incorporation and representation of new interests, the more rapidly new movements will become integrated into party systems that are experiencing realigning influences. A recent analysis of a number of party and electoral system characteristics across thirteen advanced industrial democracies found empirical evidence of a relationship between several of these characteristics and the polarization of partisan support along the emerging postmaterial value cleavage (Dalton and Flanagan, 1982). The single factor found to be most strongly associated with the integration of this new value cleavage into a nation's party system was fragmentation. Fragmented party systems provide new interests and movements with: (1) a larger number of parties in which they may try to gain representation and influence; (2) smaller parties that are more likely to espouse minority views; and (3) lower thresholds for the representation of new parties. Dalton and Flanagan also found that the integration of the value cleavage was associated with the importance of party voting and the type of political structure. Systems that require straight party voting provide fewer opportunities for experimentation and the representation of new interests. Conversely, a candidate-oriented electoral system encourages candidates to explore new issue appeals. Recent American presidential primaries provide excellent examples of this phenomenon. In addition, parliamentary systems apparently have incorporated the new value cleavage more readily than have presidential systems. Clearly, presidential systems exert more pressure to aggregate political interests to form majority coalitions than do parliamentary systems. Similarly, district size could be expected to facilitate the representation of new interests (Katz, 1980: 30-31). Smaller districts should mean stronger and more cohesive local interests. With fewer interests to aggregate and balance off against each other, candidates should be more responsive to local diversities, and hence provide an opportunity for representation to new movements organized around local pockets of strength. Finally, a federal system

Political Forces and Partisan Change

467

also should provide a greater variety of political arenas and opportunities for representation. A political movement first may win subnational elections where its strength is concentrated, and these bases can provide the infrastructure for further political action. For example, although there were no members of Die Gruenen in the West German federal parliament prior to the 1983 election, elected party representatives at the state and local levels provided leadership and visibility for the party as a whole. In the case of federalism, however, Dalton and Flanagan find that the impact of this factor is distorted by its relationship with party fragmentation. Most federal systems have low levels of fragmentation, which works to inhibit representation of the New Politics dimension. Also, many unitary states, such as Italy and Japan, have popularly elected local and/or regional governments that provide a secondary area for party competition. In sum, there are a number of intervening leadership and institutional variables that may mediate the manner in which realigning forces are integrated into the party systems of advanced industrial democracies. Nevertheless, if popular pressures on these party systems continue, these pressures may be resolved by some form of partisan realignment. In some instances, one or more of the established parties may take over new issues to prevent the emergence of new parties and add new elements to their coalition of constituencies. It is normally assumed that the New Left and New Right will align themselves with their traditional ideological counterparts. However, there are also cases where the traditional bourgeois and conservative parties have sought to represent new issues.2 Thus, as Lipset and Rokkan (1967) noted, crosscutting dimensions may generate a variety of alliance patterns, in part dependent on the unique political context of each nation. Realignment of the established parties might give the appearance of partisan stability, because of the continuity of party actors, particularly if the process is extended over a long period of time. A slow secular realignment initially may be difficult to identify against the backdrop of normal election-to-election variation in voting patterns. In the long run, however, the basis of support for the established parties, and therefore their policy objectives, would undergo substantial change as the party system accommodates itself to a new dimension of cleavage. The secular changes in party support described by Carmines and Stimson, Dalton, Barnes, and Flanagan perhaps illustrate the early stages of this process. A second alternative is that one or more new parties will be permanently added to contemporary party systems to represent new issue interests. New parties provide clearer evidence of the realignment process and probably clearer representation of a new cleavage dimension since these issues are not 2

For example, the Center party in Sweden has led the opposition to nuclear energy (Sarlvik, 1977), in Portugal, the Monarchists have been the strongest supporters of environmental protection

468

Conclusion

aggregated with other cleavage dimensions. However, new parties often are vulnerable to co-optation by the larger established parties. Thus although we have noted the growth of new parties in the past decade, it is too soon to determine whether they will provide the ultimate means of partisan change or merely an intermediate step. A third alternative is that political elites will resist mass pressures for partisan change long enough for them to dissipate. The established parties have an interest in maintaining the status quo for a variety of reasons. They may hope to ensure their own positions and influence, maintain representation for their clientele, or avoid the uncertainty of any fundamental change in the party system. We have discussed the ability of party elites to delay change in the short-run, but in the extreme the delay can be much longer. For example, American elites removed the race issue from the political agenda for several generations, while the basic problem remained unresolved; European party systems often have followed a similar tactic on regional or religious issues. We believe the pressure for political change will continue to grow among modern mass publics, so it will be increasingly difficult for elites to resist change. The pressures will be strengthened because the proponents of change come from highly articulate and highly educated sectors of the public. Thus the cleavage model suggests that eventual party realignment is likely, though not certain.3 In sum, a complex constellation of factors are relevant for explaining the patterns of partisan realignment. It therefore should come as no surprise that the actual patterns of partisan change described in this volume are equally complex. In some nations new parties have emerged, while in others the established parties are responding to new issue interests. The evidence further suggests that several forms of partisan realignment may occur simultaneously. For example, in West Germany, the environmental movement is working through both the new ecological party (Die Gruenen) and the SPD. In Japan, liberal social issues are represented by the neoconservative New Liberal Club as well as elements within the Socialist and Communist parties. Conservative Danish voters are attracted to Glistrup's Progressive party and the established bourgeois parties. Moreover, the process of partisan realignment apparently has begun in several nations, but in no case has it run its course. This complexity should not, however, detract us from our central theme. New issue interests are seeking influence in the political processes of advanced industrial democracies. The various intervening variables may lead them to explore several strategies for gaining influence. In the end, these strategies 3 A fourth alternative would be for the established parties to dissipate the pressures for a realignment by collectively endorsing an emerging issue This is only likely, however, if the issue becomes a valence issue that can be easily adopted, resolved, and dismissed See Sundquist (1973) on this and other realignment scenarios In this case, however, the New Politics issue agenda is likely to prove too divisive to be susceptible to such a convenient solution

Political Forces and Partisan

Change

469

are complementary paths leading to the result predicted by the social cleavage model—partisan realignment. Variations in the Dealignment Process The functional model argues that democratic political systems are experiencing a historical decline in the significance of political parties at both the macro and micro levels. The present decomposition of party systems is not, therefore, a prelude to another partisan realignment and a return to a new period of stable alignments in the view of the functional model. Even if some partisan realignment occurs in contemporary party systems, this is but a temporary respite from the long-term trend of party decline. Parties may never disappear entirely, but we have seen a decline in their political role over the past several decades and this trend may continue.4 Even if the functional thesis is correct, it will be difficult for citizens to adopt a style of dealigned politics if political systems remain organized in terms of a strict partisan framework. In order for mass pressures to alter the political order, a series of institutional changes are necessary (Rochon, 1983). Chief among them is a reduction of party control over electoral processes and political institutions. These changes would act as intervening variables to facilitate the development of dealigned politics. A focal point of this volume has been the declining role of parties in the electoral arena. This has been traced largely to the waning relevance of the old cleavages and the growth of new issue interests that cut across established party lines, as well as the increasing political sophistication of contemporary electorates. Moreover, the increased heterogeneity of advanced industrial societies suggests that issue interests are likely to remain diverse—and thus hinder attempts by large national parties to aggregate interests into unified policy programs. As issue interests increase in number and complexity, this should decrease the role of party as a cue for mass voting behavior. Nevertheless, some voters will continue to need the electoral cues of party labels, and all voters probably will continue to see party labels on the ballot. The question is, What does the label stand for? A trend toward dealigned politics would be enhanced if parties diversify internally, and do not impose a single-party ideology on all issues and all members. Party candidates would have greater freedom to represent the unique issue interests of their constituency, rather than the issue interests of their national party. Parties still would serve an organizational function for electoral and legislative politics, but party labels would not signify a single set of policy 4 Of course, the full logic of the functional model does not apply to a new democracy like Spain, where voters have not become so sophisticated that parties are no longer necessary In this case, the label "prealigned" may be more appropriate, see Converse (1969) Even so, the precarious state of Spanish democracy leaves open the possibility of a return to authoritarian rule and the resultant destruction of the party system—an aberrant form of dealignment

470

Conclusion

goals. Visible indicators of this development would be an increase in personalistic campaigns and a decline in party cohesion in the legislature. These characteristics of dealigned politics already are well developed in the United States, and similar patterns are emerging in non-American party systems (Epstein, 1980: 359-383). These may not be welcome trends if they contribute to the "crisis of governability" by absolving parties of policy responsibility. However, a continuance of these trends would lower the significance of party in structuring voting behavior. Alternatively, parties may proliferate and represent clear, but narrow, policy issues. For example, the Japanese party system recently has witnessed the formation of a number of single-issue mini-parties, such as a women's liberation party, a salaried workers' party, and an antinuclear peace party. In such cases, voters can select their preferred ideological mix through splitticket voting or by changing parties between elections. However, here, too, parties are not elected to govern but to represent specific issue interests. No psychological bond links voters to their parties and we would expect considerable volatility in such party systems. Another area of potential change involves new methods of citizen input that lessen the parties' role in interest articulation. For example, techniques of direct democracy—the initiative and referendum—have been available for centuries. These techniques seldom were used in democratic societies (except in Switzerland and the United States). When they were used, moreover, it often was to mobilize citizen support for government policy rather than to open decision-making to citizen input. Direct democracy may be more necessary and feasible in contemporary democracies. We have witnessed a qualitative improvement in the political sophistication of modern electorates. It is becoming more difficult to argue that a more active role in political decision-making cannot be entrusted to the citizenry—although some always will argue that position. In addition, citizens increasingly are concerned with specific government policies that affect their lives directly. Participation through political parties may not provide an efficient means of influencing specific policies, because of the parties' interest aggregation function. Therefore, direct democracy may be a more appropriate mode of participation in advanced industrial democracies. First, it increases the quantity of mass participation in step with the citizenries' developing political skills. Second, it increases the quality of participation by giving citizens more control over the agenda of participation. Some support for this position can be seen in the recent growth of direct democracy in many of the nations examined in this volume. For example, after six centuries of representative democracy, Britain held its first referendum in 1975 and followed with a second in 1979. Of the nations studied in this volume, only West Germany, Japan, The Netherlands, and the United

Political Forces and Partisan

Change

471

States have not had at least one national referendum in the postwar period— and in the latter case referendums are increasingly common at the subnational level. David Butler and Austin Ranney (1978) have documented a general growth in the incidence of national referendums in Western Europe during the postwar period. They conclude that referendums are almost certain to increase in number and importance in the years ahead. This trend can have only a negative impact on the influence of political parties and contribute to a dealigned politics. Another relevant aspect of direct democracy is that this participation mode often is linked to the new issue conflicts of advanced industrial democracies. Initiatives and referendums are a favored means of political influence where new issue groups find themselves aligned against the established order. American examples range from the conservative tax revolts spawned by California's Proposition 13, to the liberal environmental initiatives enacted in California, Oregon, and Washington. Even in West Germany, where Hitler's plebiscites of the 1930s discredited direct democracy methods, environmental and other New Left groups are calling for the revival of the referendum and initiative. One other feature of dealigned politics is the proliferation of citizen action groups and single-issue groups in recent years (Sharpe, 1979; Marsh, 1977; von Alemann, 1979). These groups provide a visible and direct method of interest articulation that circumvents the parties at least partially. Citizen involvement has long been commonplace for education and community-level policies. However, citizen action groups now are involved in a much wider range of issues—environment, energy, urban development, moral issues, social services—and the list is growing (McKean, 1981; Foster, 1980). In addition, the number of involved citizens is growing. In several nations, the membership in citizen action groups exceeds formal party membership. Citizen action groups now struggle to gain access to the policy-makingand policy-administration processes. The bureaucratized nature of the policy process tends to discourage input from ad hoc citizen groups. Citizen input often comes too late in the process, while parties and organized interest groups have the resources and contacts to monitor and participate in the policy process. To attain the "maximum feasible citizen participation'' would require a restructuring of the policy process.5 Legislation in some nations now is mandating the use of advisory committees to facilitate policy administration, and similar steps can be taken to increase citizen input at the beginning of the policy process (Bezold, 1978). Some of the ad hoc citizen groups also are seeking to enhance their influence through sustained organization and even 5 This term is identified with the Great Society programs initiated by the Johnson administration to increase the participation of minorities and low-income groups. Although the impetus in this case was from the federal government, these participation methods have been adopted by a range of self-motivated citizen groups in the United States and other democracies (see Sharpe, 1979).

472

Conclusion

institutionalization. To the extent that political institutions respond to these public demands, citizen action groups will contribute to the growth of dealigned politics. In sum, the functional model argues that citizens will continue to move away from party-oriented politics. To an important degree, however, the development of dealigned politics will depend on elite and institutional responsiveness to these changes. The willingness of political elites to institutionalize the features of dealigned politics—that is, direct democracy, citizen participation in the policy process, and a circumscribed electoral role for parties—would greatly influence the development of a dealigned political order. A FUTURE OF DIVERSITY

Two points stand out among the evidence collected in this volume. First, the recent increase in partisan instability has occurred in conditions that fall short of the revolutionary upheavals that formed the religious, regional, and class cleavages. The past decade has been one of exceptional political change for most democratic societies, but it hardly compares with the traumatic events—world wars, depressions, and civil wars—associated with previous periods of partisan realignment. However, we have suggested that contemporary electoral changes can be traced, at least partially, to silent generational and socioeconomic change. To the extent that these trends persist over time, then the processes of political change also should continue. Second, despite our emphasis on change, we are struck by the fact that no single author unequivocally states that a new political order has emerged. Most say that change is in the air; a few say that the old order is dead; but none announce the birth of a new political order. This finding alone cautions us to avoid the simple, overarching predictions that haunt the literature on political change. Indeed, if we have something concrete to conclude from our findings, it is that electoral change is not simple—either in its process or outcome. And in concluding we should strip away some of these simple predictions. One common tendency when studying partisan change is to search for, or anticipate, systematic liberal or conservative trends across democratic party systems. These expectations are unrealistic in the present situation because electoral change now involves more than a single liberal-conservative dimension. Indeed, as has been noted, the Left-Right political orientations of mass publics in these systems are in a state of transition, from being heavily defined in class terms to being increasingly defined by a value cleavage or multiple value cleavages. This multidimensional fragmenting of an overall ideological frame of reference is one reason behind the electoral instability we have observed. In addition, political forecasters overlook the complexities

Political Forces and Partisan Change

473

of the electoral process. For example, even if a New Left increases in size, its electoral impact might be overwhelmed by the mobilization of a reactionary New Right (Lipset, 1981b). There is also an interaction between mass values and party positions. Election results reflect the joint distribution of parties and voters in the political space, and interelection changes might be attributed to movement by either (Downs, 1957). For example, a decreasing vote share may signify that a party is moving away from its supporters, or the electorate is moving away from the party. Thus vote shares are a poor indicator of changes in mass values; we should not expect a systematic leftward or rightward trend to emerge as a consequence of advanced industrialism. Indeed, the evidence from advanced industrial societies fails to suggest a clear trend (Cameron, 1976). Another necessary point of clarification involves our two models of partisan change. It is crucial to distinguish between the social cleavage and functional models. They posit different theories of political change, different causal processes, different intervening variables, and different end states for democratic political systems. Yet, having drawn out these distinctions it would be a mistake to treat these processes as mutually exclusive, or to categorize an entire political system as experiencing only a single process of electoral change. Both of these forces are at work in virtually all advanced industrial democracies. Moreover, in some ways these two processes may not exert such disparate influences as have been suggested. Indeed, in their long-term impacts on voting patterns and party system, they may prove to be more reinforcing than contradictory. We already have noted several reasons why the most extreme scenario posited by the functional model, the withering away of parties, is not likely to occur. Similarly, it appears to be equally unlikely that any new realignment will be as stable, enduring, or expansive as the major realignments of earlier periods. In the past, realignments were based on the mobilization of social group interests. For these interests as well as the issues surrounding them to be successful in effecting a realignment, social groups must be mobilized to participate in politics as a bloc. Political cohesion is most likely if the constituents of a social group are: (1) highly organized into (2) exclusive and overlapping networks and associations (3) whose internal structures are personalistic and hierarchical (Zuckerman, 1982). In this regard, social groups based on the segmental cleavages of ethnicity, race, language, and religion have the greatest potential for achieving political cohesion, because such attributes generally are determined at birth and are highly visible, unambiguous, and immutable. Social groups based on class cleavages have a somewhat lower potential for political cohesion because of the gray areas between occupational categories and because of social mobility. Still, there are numerous historical examples where social class provided the basis of high levels of organization into exclusive, hierarchical networks.

474

Conclusion

However, the types of cleavages that divide advanced industrial societies are of a different form. We are moving from cleavages defined by social groups to value cleavages that identify only communities of like-minded individuals. The growing homogeneity, secularization, and embourgeoisement of society are weakening social group identifications. Increasing levels of urbanization and occupational/geographic mobility mitigate against the continued existence of exclusive social group networks (Stephens, 1981; Murakami, 1982). The revolutions in education and cognitive mobilization work against the survival of disciplined, hierarchical, and clientelistic associations. The new value cleavage is unlikely to provide a basis of mobilization into exclusive, cohesive associational networks. Indeed, postmaterial values are antithetical to such massive, disciplined organizations. Instead, we find an array of relatively small single-interest groups from the women's movement to peace organizations to environmental groups. Generally, these groups are loosely organized with ill-defined memberships that wax and wane. In sum, two kinds of changes are affecting the durability of a potential new realignment and its extensiveness (the proportion of the population affected). First, the shift from class politics to value politics marks a transformation from social-group cleavages to issue-group cleavages. Because issuegroup cleavages are more difficult to institutionalize or "freeze" via social group identifications linked to mass organizations, they may not be as enduring. Because many of the issues involved in these realignments directly concern only narrow-issue publics or involve questions of the relative priority of various positively valued social goals, the linkage between these cleavages and party support may remain less clear (Harmel and Janda, 1982: 121-125). Parties may adopt vague stands to avoid offending various narrow interests; they also can adjust priorities upward and downward much more easily than they can change sides in a debate between conflicting social group interests. A second change affects all types of cleavages. Various economic or ethnic issues may emerge that provoke some social group realignments. However, all forms of political mobilization in advanced industrial societies are subject to atomizing influences. Interest mobilization along any cleavage dimension necessarily will be characterized by more complex, overlapping, and crosscutting associational networks; more fluid institutional loyalties; and looser, more egalitarian organizational structures. This change necessarily affects the breadth, effectiveness, and stability of any realignment. In short, the forces of partisan realignment and partisan dealignment are in many ways having similar effects. Both may be viewed as increasing the amount of fluidity and volatility in party attachments and party systems. Realignment and dealignment, therefore, may be viewed as concurrent and even complementary processes. If both lead us to predict a future of diversity because of their common destabilizing properties, the concurrence of the two processes enhances diversity by providing citizens with alternate strategies of

Political Forces and Partisan Change

475

participation. Any new issue group seeks to maximize its influence in the political system. If it decides to work within the party system, that decision will promote realignment. If it opts for working outside the system, its political activities will promote deahgnment. However, since these two avenues of political action are merely alternative means of attaining the ultimate objective of policy influence, a new social movement is likely to explore both routes. For example, some sectors of the environmental movement form ecological parties or seek representation within the established parties; other environmentalists resort to citizen action groups and unconventional forms of political participation. Similarly, in some instances opposition to the welfare state has developed a partisan focus, while other groups have followed the nonpartisan procedures of dealigned politics. This suggests that realignment and deahgnment to some extent may reflect interchangeable means of initially gaining political influence. Thus the complexity of present electoral trends is to be expected if new issue groups pursue a variety of strategies for gaining access to decision-making—much as new political movements always have. However, what sets contemporary politics apart from earlier times is that dealigned politics now exists as a viable option in addition to political influence through the party system. In the long run, issue groups may gravitate toward a single means of influence. Past success in working through the party system encourages further partisan activity; success through citizen action movements encourages further emphasis on direct action methods. A dominant pattern may develop in the representation of the new issue interests. However, the development of a realignment or deahgnment pattern still may vary cross-nationally because of the intervening variables we have discussed. Institutional and structural constraints in the political system might encourage partisan realignment in one nation and deahgnment in another. For example, partisan realignment probably is accomplished more easily in the Dutch party system, while the American system appears more receptive to dealigned politics. If each political system eventually channels new issue movements along the course most consistent with prevailing political forms, then the same issue interests might find different outlets across the industrialized democracies. The characteristics of individual citizens are another likely source of diversity. Many citizens possess high levels of political skills and resources. However, these traits are not equally distributed throughout the electorate. Some individuals may be able to handle the complexities of politics without resorting to partisan or group ties; other citizens will continue to rely on political cues to guide their behavior (Shively, 1979; Dalton, 1984). The former group may orient itself to the methods of dealigned politics, while the latter group may respond to new issues in terms of the cleavage model. For example, sophisticated New Left groups may embrace direct action methods at the same time their opposition from traditional sectors of society focuses

476

Conclusion

on a partisan response. Thus even within nations, we may uncover diverse patterns of electoral change. In sum, the trends we have discussed here do not lend themselves to a single prediction of the future of democratic party systems. We have described theories that provide guideposts for understanding and interpreting patterns of electoral change, but we are less sanguine about the accuracy of any single theory for generally explaining the politics of advanced industrial societies. The dominant trends that emerged from the Reformation and the Industrial Revolution are lacking in the present situation. A clearer trend may emerge as the processes of political change develop further. However, there is growing evidence that diversity and the "demassification" of social patterns may be a defining characteristic of advanced industrial societies (Toffler, 1980). A search for single explanations of mass political behavior may be futile in a world where small is beautiful, power is decentralized, and individual freedoms are given greater latitude. The diversity we have documented here may be the best forecast. This diversity in itself constitutes a major departure from the structured partisan politics of the past.

REFERENCES

Abramson, P. R. 1975. Generational change in American politics. Lexington, Mass.: Lexington Books. . 1976. Generational change and the decline of party identification in America, 1952-1974. American Political Science Review 70: 469-478. . 1978. Class voting in the 1976 election. Journal of Politics 40: 1,066-1,072. . 1979. Developing party identification: A further examination of lifecycle, generational, and period effects. American Journal of Political Science 23: 78-96. Achen, C. H. 1975. Mass political attitudes and the survey response. American Political Science Review 69: 1,218-1,231. Alemann, Ulrich von, ed. 1979. Partizipation, Demokratisierung, Mitbestimmung. Opladen: Westdeutscher Verlag. Alford, R. R. 1963. Party and society: The Anglo-American democracies. Chicago: Rand McNally. Allardt, E. 1968. Past and emerging political cleavages. In Party organization and the politics of the new masses, ed. Otto Stammer, pp. 66-74. Berlin: Institute of Political Science at the Free University. AUinson, G. D. 1980. Opposition in the suburbs. In Political opposition and local politics in Japan, ed. K. Steiner et al., pp. 95-130. Princeton: Princeton University Press. Almerich, P. et al. 1975. Cambio social y religion en Espafia. Barcelona: Fontanella. Almond, G., and Verba, S. 1963. The civic culture. Princeton: Princeton University Press. Alt, J. 1979. The politics of economic decline. Cambridge: Cambridge University Press. Alt, J., and Chrystal, K. A. 1983. Political economics. Berkeley: University of California Press. Alt, J.; Crewe, I.; and Sarlvik, B. 1976. Partisanship and policy choice. British Journal of Political Science 6: 273-291. Alt, J., and Turner, J. 1982. The case of the silk-stocking socialists and the calculating children of the middle class. British Journal of Political Science 12: 239-248. Altares, P. 1982. Las ganas de vivir. El Pais (November 9): 9. Andersen, K. 1976. Generation, partisan shift and realignment: A glance back to the New Deal. In N. H. Nie, S. Verba and J. R. Petrocik, The changing American voter. Cambridge: Harvard University Press.

478

References

Andersen, Κ. 1979. The creation of a Democratic majority, 1928-1936, Chicago: University of Chicago Press. Andeweg, R. B. 1982. Dutch voters adrift: On explanations of electoral change, 1963-1977. Ph.D. dissertation, Leyden University. Andeweg, R. B.; Tak, Th. van der; and Dittrich, K. 1980. Government formation in The Netherlands. In The economy and politics of The Neth­ erlands since 1945, ed. R. T. Griffiths. The Hague: Martinus Nijhoff. Anners, E. 1972. Conservatism in Sweden. In Politics in the post-welfare state, ed. D. Hancock and G. Sjoberg. New York: Columbia University Press. Arian, Α., and Barnes, S. H. 1974. The dominant party system: A neglected model of democratic stability. Journal of Politics 36: 592-614. Arija, J. M. 1981. Gora Espafia askatuta. Cambio 16 (July 13): 18-23. Artola, M. 1974. Programas y partidos politicos, 1808-1936. Madrid: Aguilar. Asahi Nenkan. 1955-1981. Tokyo: Asahi Shimbunsha. Asher, H. 1980. Presidential elections and American politics. Homewood, 111.: Dorsey Press. Attiyeh, R., and Engle, R. 1979. Testing some propositions about Proposition 13. National Tax Journal 32: 131-146. Aviv, Α., and Aviv, I. 1981. The Madrid working class, the Spanish socialist party and the collapse of the Second Republic (1934-1936). Journal of Contemporary History 16: 229-250. Baena del Alcazar, M., and Garcia Madaria, J. M. 1979. Elite Franquista y burocracia en las Cortes actuales. Sistema 32: 3-37. Baker, K. 1978. Generational differences in the role of party identification in German political behavior. American Journal of Political Science 22: 106-129. Baker, K. L.; Dalton, R.; and Hildebrandt, K. 1981. Germany transformed. Cambridge: Harvard University Press. Balbus, I. D. 1982. Marxism and domination. Princeton: Princeton University Press. Barbagli, M., and Dei, M. 1969. Le vestali della classe media. Bologna: II Mulino. Barbagli, M. et al. 1979. Fluidita' elettorale e classi sociali in Italia. Bologna: II Mulino. Barnes, S. H. 1974. Italy: Religion and class in electoral behavior. In Electoral behavior: A comparative handbook, ed. R. Rose, pp. 171-225. New York: Free Press. . 1977. Representation in Italy: Institutionalized tradition and elec­ toral choice. Chicago: University of Chicago Press. Barnes, S. H., Kaase, M. et al. 1979. Political action: Mass participation in five Western democracies. Beverly Hills: Sage Publications.

References

479

Barnes, S. Η., and Pierce, R. 1971. Public opinion and political preferences in France and Italy. Midwest Journal of Political Science 15: 643-660. Beck, P. A. 1974. A socialization theory of partisan realignment. In R. G. Niemi and associates, The politics of future citizens, pp. 199-219. San Francisco: Jossey-Bass. . 1977. Partisan dealignment in the postwar South. American Political Science Review 71: 477-496. . 1979. The electoral cycle and patterns of American politics. British Journal of Political Science 9: 129-156. . 1982. Realignment begins? American Politics Quarterly 10: 421437. Beckman Stichting. 1951. Verkiezingen in Nederland. Amsterdam: Wiardi Beckman Stichting. Beer, S. 1978. The British political system. New York: Random House. . 1982. Britain against itself. New York: Norton. Bell, Daniel. 1960. The end of ideology. New York: Free Press. . 1973. The coming of post-industrial society. New York: Basic Books. Beltran, M. 1977. La elite burocratica espanola. Madrid: Ariel. Berelson, B.; Lazarsfeld, P.; and McPhee, W. 1954. Voting. Chicago: Uni­ versity of Chicago Press. Berger, S. 1979. Politics and anti-politics in Western Europe in the 1970s. Daedalus 108: 27-50. Berger, S., ed. 1981. Organizing interests in Western Europe. Cambridge: Cambridge University Press. Berger, S., and Piore, M. J. 1981. Dualism and discontinuity in industrial societies. Cambridge: Cambridge University Press. Berglund, S., and Lindstrom, U. 1978. The Scandinavian party system(s). Lund: Studentlitteratur. Beyme, Klaus von. 1978. The politics of limited pluralism. Government and Opposition 13: 265-287. Bezold, C , ed. 1978. Anticipatory democracy. New York: Random House. Bilski, R. 1976. Basic parameters of the welfare state. Social Indicator Re­ search. 3: 451^170. Blackmer, D. L. 1968. Unity in diversity: Italian communism and the Com­ munist world. Cambridge: Mass.: MIT Press. Books, J., and Reynolds, J. B. 1975. A note on class voting in Great Britain and the United States. Comparative Political Studies 8: 360-375 Borre, O. 1980. Electoral instability in four Nordic countries, 1950-77. Com­ parative Political Studies 13: 141-171. Borre, 0 . , and Katz, D. 1973. Party identification and its motivational base in a multiparty system. Scandinavian Political Studies 8: 69-111. Borre, O.; Nielsen, H.;Sauerberg, S.;andWorre, T. 1976. Vaelgerei70erne. Copenhagen: Akademisk Forlag.

480

References

Box, G.E.P., and Jenkins, G. M. 1976. Time series analysis: Forecasting and control. San Francisco: Holden-Day. Brinkgreve, C , and Korzec, M. 1978. Margriet weet raad: gevoel, gedrag, moraal in Nederland, 1938-1978. Utrecht: Uitgeverij Het Spectrum. Budge, I.; Crewe, I.; and Farlie, D., eds. 1976. Party identification and beyond. London and New York: Wiley. Budge, I., and Farlie, D. 1977. Voting and party competition. London: Wiley. Buerklin, W. 1981. Die Grunen und die 'Neue Politik.' Politische Vierteljahresschrift 22: 359-382. . 1982. Konzept und Fakten. Politische Vierteljahresschrift 23: 339345. Burnham, W. D. 1970. Critical elections and the mainsprings of American politics. New York: Norton. . 1973. Political immunization and political confessionalism: The United States and Weimar Germany. Journal of Interdisciplinary History 3: 325. . 1974. Theory and voting research: Some reflections on Converse's "Change in the American electorate." American Political Science Re­ view 68: 1,002-1,023. . 1975. American politics in the 1970s. In The future of political parties, ed. L. Maisel and P. Sacks, pp. 238-277. Beverly Hills: Sage Publications. . 1978. Great Britain: The death of the collectivist consensus? In Political parties: Development and decay, ed. L. Maisel and J. Cooper, pp. 267-308. Beverly Hills: Sage Publications. . 1980. American politics in the 1980s. Dissent: 149-160. Burstein, P. 1976. Social networks and voting. Social Forces 54: 833-848. Butler, D., and Ranney, Α., eds. 1978. Referendums. Washington, D.C.: American Enterprise Institute. Butler, D., and Stokes, D. 1974. Political change in Britain. New York: St. Martin's Press. Cain, B., and Ferejohn, J. 1981. Party identification in the United States and Great Britain. Comparative Political Studies 14: 31-48. Caldeira, G., and Greenstein, F. I. 1978. Partisan orientation and political socialization in Britain, France, and the United States. Political Science Quarterly 93: 35^19. Cameron, D. R. 1972. Stability and change in patterns of French partisanship. Public Opinion Quarterly 36: 19-30. . 1976. Postindustrial change and secular realignments. Ph.D. dis­ sertation, University of Michigan. . 1982. On the limits of the public economy. Annals of the American Academy of Political and Social Science 459: 46-62. Cameron, D., and Hofferbert, R. I. 1973. Continuity and change in Gaullism: The general's legacy. American Journal of Political Science 17: 77-98.

References

481

Campbell, A. 1966. A classification of the presidential elections. In A. Camp­ bell et al., Elections and the political order. New York: Wiley. Campbell, A. et al. 1960. The American voter. New York: Wiley. Campbell, Α., and Valen, H. 1966. Party identification in Norway and the United States. In A. Campbell et al., Elections and the political order. New York: Wiley. Campbell, Β. A. 1971. French elections and voter stability: A model of voting behavior for the French national electorate. Ph.D. dissertation, Univer­ sity of Michigan. . 1974. The future of the Gaullist majority: An analysis of French electoral politics. American Journal of Political Science 18: 67-94. . 1976. On the prospects of polarization in the French electorate. Comparative Politics 8: 272-290. . 1979. Realignment, party decomposition and issue voting. In Re­ alignment in American politics, ed. B. Campbell and R. Trilling, pp. 82-109. Austin: University of Texas Press. Campbell, B., and Trilling, R., eds. 1979. Realignment in American politics. Austin: University of Texas Press. Carlsson, G., and Karlsson, K. 1970. Age, cohorts, and the generation of generations. American Sociological Review 35: 710-718. Carmines, E. G.; Renten, S. H.; and Stimson, J. A. 1984. Events and align­ ments: The party image link. In Controversies in voting behavior, ed. R. Niemi and H. Weisberg. Washington, D.C.: Congressional Quarterly Press. Carmines, E. G., and Stimson, J. A. 1981. Issue evolution, population re­ placement, and normal partisan change. American Political Science Re­ view 75: 107-118. . 1983. The politics and policy of race in Congress. Paper presented at the Indiana Conference on Congress and Policy-making, Bloomington, Ind. Carr, R. 1980. Modern Spain, 1875-1980. London: Oxford University Press. Centro de Investigacion y Tecnicas Politicas. 1982. Segunda fase del panel politico. El Pais (October 14): 10-11. Cerny, P. G. 1972. Cleavage, aggregation and change in French politics. British Journal of Political Science 2: 443^455. Citrin, J. 1974. Comment. American Political Science Review 68: 973-988. . 1977. Political alienation as a social indicator. Social Indicators Research 4: 381^119. Citrin, J.; McCloskey, H.; Shanks, J. M.; and Sniderman, P. M. 1975. Personal and political sources of alienation. British Journal of Political Science 5: 1-31. Clark, R. P. 1981. Patterns of insurgent violence in Spain's Basque provinces. Paper delivered at the Southwestern Political Science Association Meet­ ings, Dallas, Tex., March 25-28.

482

References

Clarke, Η.; Jenson, J.; LeDuc, L.; and Pammett, J. 1979. Political choice in Canada. Toronto: McGraw-Hill Ryerson. Clubb, J. M.; Flanigan, W. H.; and Zingale, Ν. H. 1980. Partisan realign­ ment: Voters, parties and government in American history. Beverly Hills: Sage Publications. Cole, R. 1973. Toward a model of political trust. American Journal of Po­ litical Science 17: 809-817. Collier, D., ed. 1979. The new authoritarianism in Latin America. Princeton: Princeton University Press. Conradt, D. 1981. The German polity. 2nd ed. New York: Longman. Converse, P. E. 1964. The nature of belief systems among mass publics. In Ideology and discontent, ed. D. Apter, pp. 202-261. New York: Free Press. . 1969. Of time and partisan stability. Comparative Political Studies 2: 139-171. . 1970. Attitudes and non-attitudes: Continuation of a dialogue. In The quantitative analysis of social problems, ed. E. Tufte, pp. 168-190. Reading, Mass.: Addison-Wesley. . 1975a. Public opinion and voting behavior. In Handbook of political science, ed. F. Greenstein and N. W. Polsby, pp. 75-169. Reading, Mass.: Addison-Wesley. . 1975b. Some mass-elite contrasts in the perception of political space. Social Science Information 14: 49-83. . 1976. The dynamics of party support. Beverly Hills: Sage Publi­ cations. . 1979. Rejoinder to Abramson. American Journal of Political Science 23: 97-100. Converse, P. E.; Campbell, Α.; Miller, W. E.; and Stokes, D. E. 1961. Stability and change in 1960: A reinstating election. American Political Science Review 55: 269-280. Converse, P. E.; Clausen, A. R.; and Miller, W. E. 1965. Electoral myth and reality: The 1964 election. American Political Science Review 59: 321-336. Converse, P. E., and Dupeux, G. 1962. Politicization of the electorate in France and the United States. Public Opinion Quarterly 26: 1-23. Converse, P. E., and Markus, G. B. 1979. Plus qa change . . . : The new CPS election study panel. American Political Science Review 73: 3249. Crewe, I. 1980. Prospects for party realignment: An Anglo-American com­ parison. Comparative Politics 19: 379-400. . 1982a. Is Britain's two-party system really about to crumble? Elec­ toral Studies 1: 275-313. . 1982b. The Labour party and the electorate. In The politics of the Labour party, ed. D. Kavanagh. London: Allen and Unwin.

References

483

Crewe, I.; Sarlvik, B.; and Alt, J. 1977. Partisan dealignment in Britain, 1964-1974. British Journal of Political Science 7: 129-190. Crozier, M. et al. 1975. The crisis of democracy. New York: New York University Press. Curtice, J., and Steed, M. 1980. An analysis of the voting. In The British general election of 1979, ed. D. Butler and D. Kavanagh. London: Macmillan. Curtis, G. L. 1971. Election campaigning Japanese style. New York: Columbia University Press. Daalder, H. 1965-1966. De kleine politieke partijen—een voorlopige poging tot inventarisatie. Acta Politico 1: 172-196. . 1966. The Netherlands: Opposition in segmented society. In Political oppositions in Western democracies, ed. R. A. Dahl, pp. 188-236. New Haven: Yale University Press. . 1974. Politisering en lijdelijkheid in de Nederlandse politiek. Asser: Van Gorcum. Daalder, H., and Mair, P., eds. 1983. Western European party systems. Beverly Hills: Sage Publications. Dahl, R. 1970. After the revolution. New Haven: Yale University Press. . 1982. Dilemmas ofpluralist democracy. New Haven: Yale University Press. Dahl, R., and Tufte, E. 1973. Size and democracy. Stanford: Stanford University Press. Dahrendorf, R. 1959. Class and class conflict in industrial society. Stanford: Stanford University Press. . 1967. Society and democracy in Germany. New York: Doubleday. . 1979. Life chances: Approaches to social and political change. Chicago: University of Chicago Press. Dalton, R. 1981. The persistence of values and life cycle changes. Politische Vierteljahresschrift. Special issue: 189-207. . 1984. Cognitive mobilization and partisan dealignment in advanced industrial democracies. Journal of Politics 46: 264—284. . Forthcoming. Citizen politics in Western democracies: Public opinion and political parties. Chatham, N.J.: Chatham House. Dalton, R., and Baker, K. Forthcoming. The contours of West German opinion. In West German politics in the eighties, ed. P. Wallach and G. Romoser. New York: Praeger. Dalton, R., and Flanagan, S. 1982. The changing content of ideological beliefs in Western Europe, the United States, and Japan. Paper presented at the Annual Meetings of the American Political Science Association, Denver, Colo. Dalton, R., and Hildebrandt, K. 1983. Spannungslinie im Parteien System. In Wahlen und Politische Kultur, ed. M. Kaase and H. Klingemann. Opladen: Westdeutscher Verlag.

484

References

Daudt, Η. 1972. Constante kiezers, wisselaars en thuisblijvers. In De Nederlandse Kiezer '71. Meppel: Boom. De Esteban, J., and Lopez Guerra, L. 1977. La crisis del estado Franquista. Barcelona: Labor. Del Campo, S., and Navarro, M. 1976. Critica de la planificacion social espanola, 1964-1975. Madrid: Castellote. Del Campo, S.; Tezanos, J. F.; and Santin, W. 1983. The Spanish political elite: Permanence and change. In International yearbook of elite studies, ed. M. Czudnowski. DeKalb: Northern Illinois University Press. De Miguel, A. 1977. La piramide social espanola. Madrid: Ariel. De Nederlandse Kiezer. 1956. 'S-Gravenhage: Staatsdrukkerij en Uitgeversbedrijf. De Nederlandse Kiezer '71. 1972. Acta Politico 1. De Nederlandse Kiezer '72. 1973. Alphen a/d Rijn: Samsom. De Nederlandse Kiezer '77. 1977. Voorschoten: VAM. Dennis, J. 1975. Trends in public support for the American party system. British Journal of Political Science 5: 187-230. Deutsch, E.; London, D.; and Weill, P. 1966. Les families politiques aujourd'hui en France. Paris: Editions de Minuit. DeVries, W., and Tarrance, V. L. 1972. The ticket-splitter. Grand Rapids, Mich.: Eerdmans. Di Palma, G. 1977. Surviving without governing: The Italian parties in Par­ liament. Berkeley: University of California Press. . 1978. Political syncretism in Italy: Strategies and the present crisis. Berkeley: Institute of International Studies. Dittrich, K. 1975. De gevolgen van de veranderingen in de partijvoorkeur van de Nederlandse kiezers sinds 1966 voor de KVP en PvdA. M.A. thesis, Leyden University. Dobson, D., and St. Angelo, D. 1975. Party identification and the floating vote: Some dynamics. American Political Science Review 69: 481-490. Downs, A. 1957. An economic theory of democracy. New York: Harper and Row. Duverger, M. 1960. Introduction. In Lois electorates et inegalites de re­ presentation en France, 1936-1960, ed. J. M. Cotteret, C. Emeri, and P. Lalumiere. Paris: Foundation Nationale des Sciences Politiques, CahierNo. 107, A. Colin. . 1973. Modern democracies: Economic power versus political power. New York: Holt, Rinehart and Winston. Dye, T. 1966. Politics, economics and the public. Chicago: Rand McNally. Easton, D. 1965. A system analysis of political life. New York: Wiley. . 1975. A reassessment of the concept of political support. British Journal of Political Science 5: 435-457. Easton, D., and Dennis, J. 1969. Children in the political system. New York: McGraw-Hill.

References

485

Ehrmann, Η. W. 1983. Politics in France. 4th ed. Boston: Little, Brown. Eijk, C. van der, and Niemoller, B. 1979. Recall accuracy and its determi­ nants. Acta Politica 14: 289-342. —. 1983. Electoral change in The Netherlands. Amsterdam: C. T. Press. Elvander, N. 1972. Democracy and large organizations. In Politics in the post-welfare state, ed. D. Hancock and G. Sjoberg, pp. 301-324. New York: Columbia University Press. . 1980. Skandinavisk arbetarrorelse. Stockholm: Liber Forlag. Epstein, L. 1980. Political parties in Western democracies. New Brunswick, N.J.: Transaction Books. Equipo de Investigacion Sociologica. 1981. Ayuntamientos democraticos. Madrid: Fundacion Friedrich Ebert. Erickson, R.; Goldthorpe, J. H.; and Portocarero, L. 1982. Social fluidity in industrial nations. British Journal of Sociology 33: 1-34. Erikson, R. S.; Luttbeg, N.; and Tedin, K. 1980. American public opinion. New York: Wiley. Erikson, R. S., and Tedin, K. L. 1981. The 1928-1932 partisan realignment: The case for the conversion hypothesis. American Political Science Re­ view 75: 951-962. Esping-Andersen, G. 1978. Social class, social democracy and the state. Comparative Politics 11: 42-57. . 1979. Comparative social policy and political conflict in advanced welfare states. International Journal of Health Service 9: 269-293. Estudios Comerciales y de Opinion. 1979. Colectividad juvenil. Madrid: unpublished report. Fabris, G. 1977. II comportamento politico degli italiani. Milano: Angeli. Fair, R. 1978. The effect of economic events on votes for president. Review of Economics and Statistics 63: 159-173. Falconeri, G. R. 1976. The impact of rapid urban change on neighborhood solidarity. In Social change and community politics in Japan, ed. J. White and F. Munger, pp. 31-59. Chapel Hill, N.C.: Institute for Research in Social Science, University of North Carolina. Fanjul Sedeno, J. M. 1977-1981. Memoria elevada al gobierno. Madrid: Instituto Editorial Reus. Farah, B.; Barnes, S.; and Heunks, F. 1979. Political dissatisfaction. In S. Barnes, M. Kaase et al., Political action, pp. 403-447. Beverly Hills: Sage Publications. Farnsworth, Lee W. 1982. Japan in 1981: Meeting the challenges. Asian Survey 22: 56-68. Feist, U.; Guellner, M.; and Liepelt, K. 1978. Structural assimilation versus ideological polarization. In Elections and parties, ed. M. Kaase and K. von Beyme, pp. 171-190. Beverly Hills: Sage Publications. Fernandez de Castro, I., and Goytre, A. 1974. Clases sociales en Espana en el umbral de los anos sesenta. Madrid: Seculo 21.

486

References

Ferrara, Μ. 1979. Down with the welfare state. Paper presented at the ECPR workshop sessions, Florence, Italy. Fietkau, H.-J. et al. Forthcoming. Die Grune Zukunfi. Frankfurt: Campus Verlag. Finer, S. 1979. The changing British party system, 1945-1979. Washington, D.C.: American Enterprise Institute. Fiorina, M. P. 1981. Retrospective voting in American national elections. New Haven: Yale University Press. Fishman, R. M. 1982. The labor movement in Spain: From authoritarianism to democracy. Comparative Politics 14: 281-305. Flanagan, S. C. 1968. Voting behavior in Japan: The persistence of traditional patterns. Comparative Political Studies 1: 391—412. . 1979. Value change and partisan change in Japan. Comparative Politics 11: 253-278. . 1980a. National and local voting trends: Cross-level linkages and correlates of change. In Political opposition and local politics in Japan, ed. K. Steineretal.,pp. 131-184. Princeton: Princeton University Press. . 1980b. Cultural cleavages, social networks and voting behavior in Japan. Paper presented at the Annual Meeting of the Southern Political Science Association, Atlanta, Ga. . 1982a. Changing values in advanced industrial societies. Compar­ ative Political Studies. 14: 403^444. . 1982b. Measuring value change in advanced industrial societies. Comparative Political Studies 15: 99-128. Flanagan, S. C , and Dalton, R. 1984. Parties under stress. Western European Politics 7: 7-23. Flanagan, S. C , and McDonald, M. D. 1979. Party identification as a crossnational concept. Paper presented at the Annual Meeting of the American Political Science Association, Washington, D.C. Flanagan, S. C , and Richardson, Β. M. 1977. Japanese electoral behavior: Social cleavages, social networks and partisanship. Contemporary Po­ litical Sociology Series. London: Sage Publications. . 1979. Mass political behavior research in Japan. New York: Social Science Research Council. Forrester, J. 1971. World dynamics. Cambridge: Wright-Allen. Foster, C , ed. 1980. Comparative public policy and citizen participation. New York: Pergamon. Freeman, J. R., and Snidal, D. 1982. Diffusion, development and democ­ ratization: Enfranchisement in Western Europe. Canadian Journal of Political Science 15: 299-329. Frey, B. 1979. Economic policy by constitutional contract. Kyklos 32: 339362. Fuentes Quintana, E. 1980. La crisis economica espafiola. Papeles de Economia Espafiola 1: 31-67.

References

487

Fundacion Foessa. 1976. Estudios sociologicos sobre la situacion social de Espana, 1975. Madrid: Editorial Euramorica. Gagliani, G. 1981. How many working classes? American Journal of Soci­ ology 87: 259-285. Galli, G. et al. 1968. II comportamento elettorale in Italia. Bologna: II Mulino. Garcia Duran, J., and Puig Bastard, P. 1980. La calidad de vida en Espana: Ηacta un estudo de indicadores sociales. Madrid: Moneda y Credito. Garcia Ferrando, M. 1982. Regionalismo y autonomia en Espana, 19761979. Madrid: Centra de Investigaciones Sociologicas. Geller, D. S. 1982. Economic modernization and political instability in Latin America: A causal analysis of bureaucratic-authoritarianism. Western Political Science Quarterly 35: 33-49. Giner, S. 1971. Spain. In Contemporary Europe: Class, status and power, ed. M. S. Archer and S. Giner. New York: St. Martin's Press. . 1981. Political economy, cultural legitimation and the state in south­ ern Europe. Paper delivered at the Conference on Prospects for Democ­ racy, Wilson Center, Smithsonian Institution, Washington, D.C., June 4-7. Gladdish, K. R. 1972. Two-party vs. multi-party: The Netherlands and Brit­ ain. Acta Politica 7: 342-361. Glenn, N. D. 1972. Sources of the shift to political independence: Some evidence from a cohort analysis. Social Science Quarterly 53: 494—519. . 1973. Class and party support in the United States: Recent and emerging trends. Public Opinion Quarterly 33: 17-33. Gluchowski, P. 1978. Parteiidentifikation in politische System der Bundesrepublik Deutschland. In Waehlerverhalten in der Bundesrepublik Deutschland, ed. D. Oberndoerfer, pp. 265-323. Berlin: Duneker und Humbolt. . 1983, Wahlerfahrung und Parteiidentifikation. In Wahlen und Pol­ itische Kultur, ed. M. KaaseandH. Klingeman. Opladen: Westdeutscher Verlag. Goldstone, J. A. 1980. Theories of revolution: The third generation. World Politics 32: 425-453. Goldthorpe, J. et al. 1968. The affluent worker. Cambridge: Cambridge Uni­ versity Press. Grauhan, R., and Hickel, R. 1978. Krise des Steurstaats. Leviathan. Special issue. Graziano, L. 1979. Compromesso storico e democrazia consociativa: Verso una nuova democrazia? In La crisi italiana, ed. L. Graziano and S. Tarrow, 2: 719-723. Turin: Einaudi. Greenstein, F. I. 1965. Children and politics. New Haven: Yale University Press. Gunther, R. 1980. Public policy in a no-party state: Spanish planning and

488

References

budgeting in the twilight of the Franquist era. Berkeley: University of California Press. Gunther, R., and Blough, R. A. 1981. Religious conflict and consensus in Spain: A tale of two constitutions. World Affairs 143: 366-412. Gustafsson, B. 1969. Marxism och revisionism Eduard Bernsteins kritik av marxismen och dess idehistorika forutsattningar (Marxism and revision­ ism). Uppsala: Svenska Bokforlaget. Habermas, J. 1973. Legitimation crisis. Boston: Beacon. Hamilton, R. 1967. Affluence and the French worker in the Fourth Republic. Princeton: Princeton University Press. Harmel, R., and Janda, K. 1982. Parties and their environments. New York: Longman. Hauss, C , and Rayside, D. 1978. New political parties in Western democ­ racies. In Political parties: Development and decay, ed. L. Maisel and J. Cooper, pp. 31-58. Beverly Hills: Sage Publications. Heclo, H. 1977. Internal frontiers of social policy in Western Europe. In Comparing public policies, ed. J. Wiatr and R. Rose, pp. 199-219. Wroclaw: Polish Academy. Heidenheimer, Α., and Kommers, D. 1975. The governments of Germany. 4th ed. New York: Crowell. Hentila, S. 1979. Ruotsi Eiko koskaan sosialismia? (Sweden: will there never be socialism?) Kanava 7. Hibbs, D. 1982. Economic outcomes and political support for British gov­ ernments among occupational groups: A dynamic analysis. American Political Science Review 76: 259-279. Hildebrandt, K., and Dalton, R. 1978. The New Politics. In Elections and parties, ed. M. Kaase and K. von Beyme, pp. 69-96. Beverly Hills: Sage Publications. Holmberg, S. 1981. Svenska valjare. Stockholm: Liber Forlag. Holmberg, S.; Westerstahl, J.; and Branzen, K. 1979. Valjarna och Karnkraften. Stockholm: Liber Forlag. House, J. S., and Mason, W. M. 1975. Political alienation in America: 19511968. American Sociological Review 40: 123-147. Houska, J. J., Jr. 1979. The organizational connection: Elites, masses, and elections in Austria and The Netherlands. Unpublished Ph.D. disserta­ tion, Yale University. Howe, I., ed. 1982. Beyond the welfare state. New York: Schocken. Huntington, S. 1974. Post-industrial politics: How benign will it be? Com­ parative Politics 6: 163-191. Inchausti, Α., and Inchausti, J. A. 1977. Distribucion personal de la renta en Espana y en los paises de la OCDE. Hacienda Publico Espanola 47: 17-57. Inglehart, R. 1971. The silent revolution in Europe: Intergenerational change

References

489

in post-industrial societies. American Political Science Review 65: 9911,017. . 1977. The silent revolution: Changing values and political styles among Western publics. Princeton: Princeton University Press. . 1979a. Value priorities and socioeconomic change. In S. Barnes, M. Kaase et al., Political action, pp. 305-342. Beverly Hills: Sage Publications. . 1979b. Political action: The impact of values, cognitive level and social background. In S. Barnes, M. Kaase et al., Political action, pp. 343-380. Beverly Hills: Sage Publications. . 1981. Post-materialism in an environment of insecurity. American Political Science Review 75: 880-900. . 1983. The persistence of materialist and postmaterialist value orientations. European Journal of Political Research 11: 81-82. . Forthcoming. Group sympathies, party ties and support for change. In People and their polities, ed. H. Klingemann, H. Kerr, and P. Pesonen. Inglehart, R., and Hochstein, A. 1972. Alignment and dealignment of the electorate in France and the United States. Comparative Political Studies 5: 343-372. Inglehart, R., and Klingemann, H. D. 1976. Party identification, ideological preference and the left-right dimension among Western publics. In Party identification and beyond, ed. I. Budge, I. Crewe, and D. Farlie, pp. 243-273. London and New York: Wiley. Inglehart, R. et al. 1980. Broader powers for the European Parliament: The attitudes of candidates. European Journal of Political Research 8: 113— 132. Inoguchi, T. 1982. Review of political opposition and local politics in Japan. Journal of Japanese Studies 8: 191-198. International Labour Organization. 1981. ILO yearbook of labour statistics. Geneva: International Labour Organization. International Monetary Fund. 1979. International financial statistics yearbook. Washington, D.C.: International Monetary Fund. Irvine, W. 1974. Explaining the religious basis of Canadian partisan identity. Canadian Journal of Political Science 7: 560-565. Irwin, G. A. 1972. Political attitudes and compulsory voting in The Netherlands. Paper prepared for presentation at the Midwest Political Science Association Convention, Chicago, 111. . 1974. Compulsory voting legislation: Impact on vote turnout in The Netherlands. Comparative Political Studies 7: 292-315. . 1980. Patterns of voting behavior in The Netherlands. In The economy and politics of The Netherlands since 1945, ed. R. T. Griffiths. The Hague: Martinus Nijhoff.

490

References

Jackson, J. 1975. Issues, party choices, and presidential votes. American Journal of Political Science 19: 161-185. Jacob, J. 1980. Ethnic mobilization and the pursuit of post-industrial values. Tocqueville Review 2: 52-85. Janda, K. 1970. A conceptual framework for the comparative analysis of political parties. Beverly Hills: Sage Publications. Japan statistical yearbook. 1948-1981. Tokyo: Bureau of Statistics, Office of the Prime Minister. Jennings, Μ. K., and Niemi, R. G. 1968. Patterns of political learning. Harvard Educational Review 38: 443-467. . 1981. Generations and politics. Princeton: Princeton University Press. Jenson, J. 1975. Party loyalty in Canada: The question of party identification. Canadian Journal of Political Science 8: 543-553. Jenson, J.; Clarke, H.; LeDuc, L.; and Pammett, J. 1975. Patterns of par­ tisanship in Canada: Split identification and cross-time variation. Paper presented at the Annual Meeting of the American Political Science As­ sociation, San Francisco, Calif. Jong, J. J. de. 1956. Overheiden Onderdaan. Wageningen: Gebr. Zomeren Keunings. Kaase, Μ., and Barnes, S. 1979. In conclusion: The future of political protest in Western democracies. In S. Barnes, M. Kaase et al., Political action, pp. 523-536. Beverly Hills: Sage Publications. Kaase, M., and Klingemann, H.-D. 1982. Social structure, value-orientations and the party system. European Journal of Political Research 10: 367386. Kaase, M., and Marsh, A. 1979. Political action repertory. In S. Barnes, M. Kaase et al., Political action, pp. 133-166. Beverly Hills: Sage Publications. Katz, R. 1980. A theory of parties and electoral systems. Baltimore: Johns Hopkins University Press. . 1982. Party government: Some conceptual problems. Special paper of the European University Institute, Florence, Italy. Kautsky, K. 1907. Parlamentarismi, Kansanlainaadantoja sosialidemokratia (Parliamentarism, lawmaking and social democracy). Helsinki: Sosialistinen Aikakausilehti. Kawada, Η. 1973. Workers and their organizations. In Workers and employers in Japan, ed. K. Okochi, B. Karsh, and S. Levine, pp. 217-267. Tokyo: University of Tokyo Press. Keifer, C. W. 1976. Leadership, sociability and social change in a white collar danchi. In Social change and community politics in Japan, ed. J. White and F. Munger, pp. 15-30. Chapel Hill, N.C.: Institute for Research in Social Science, University of North Carolina.

References

491

Kemp, D. 1978. Society and electoral behavior in Australia. St. Lucia: Queensland University Press. Kerr, H. Forthcoming. Social class and party choice. In People and their polities, ed. H. Klingemann, H. Kerr, and P. Pesonen. Kesselman, M. 1980. Continuity and change on the French left: Revolutionary transformation or immobilism. Social Research ΑΠ: 93-113. Kessler, R., and Greenberg, D. 1981. Linear panel analysis. New York: Academic Press. Key, V. O. 1955. A theory of critical elections. Journal of Politics 17: 318. . 1959. Secular realignment and the party system. Journal of Politics 21: 198-210. . 1964. Politics, parties, and pressure groups. New York: Crowell. King, A. 1981. What do elections decide? In Democracy at the polls, E. D. Butler et al., pp. 293-324. Washington, D.C.: American Enterprise Institute. Kirchheimer, O. 1966. The transformation of Western European party sys­ tems. In Political parties and political development, ed. J. LaPalombara and M. Weiner, pp. 177-220. Princeton: Princeton University Press. Klingemann, H. 1979. Measuring ideological conceptualizations. In S. Barnes, M. Kaase et al., Political action, pp. 215-284. Beverly Hills: Sage Publications. . 1982. Fakten oder Programmatik? Politische Vierteljahresschrifi 23: 214-224. Kohei, S. 1974. Taisei sentaku no seiji ishiki. Bunken Geppo 24: 1-20. . 1978. Reversal of the balance of power thwarted. Japan Echo (special issue) 5: 109-120. Kornhauser, W. 1959. The politics of mass society. New York: Free Press. Korpi, W. 1978a. The working class in welfare capitalism. London: Routledge and Kegan Paul. . 1978b. Social democracy in welfare capitalism: Structural erosion, welfare backlash and incorporation. Acta Sociologica 21 (supplement): 97-111. . 1979. Valfardsstatens variationer: Forskningsproblem om socialpolitiska strategier i de kapitaliska demoktratierna. Sociologisk Forskning Argang 16: 3-18. Kosugi, T. 1982. New Liberal Club. In Japan in the 1980s, ed. R. Shiratori, pp. 258-272. Tokyo: Kodansha International. Kuhnle, S. 1978. The beginnings of the Nordic welfare states. Acta Socio­ logica 21 (supplement): 9-35. Kurth, J. R. 1979. The political consequences of the product cycle. Inter­ national Organization 33: 1-34.

492

References

Ladd, Ε. C. et al. 1979. The polls: Taxing and spending. Public Opinion Quarterly 43: 126-135. Ladd, E. C , with Hadley, C. 1975. Transformations of the American party system. New York: Norton. Lancaster, T. D. 1979. Toward an assessment of the Spanish social security system. Paper delivered at the Fourth European Studies Conference, Omaha, Nebr., October 11-13. Lane, R. 1965. The politics of consensus in an age of affluence. American Political Science Review 59: 874-895. Lange, P. 1979. II PCI e i possibili esiti della crisi italiana. In La crisi italiana, ed. L. Graziano and S. Tarrow, 2: 657-718. Turin: Einaudi. Lazarsfeld, P.; Berelson, B.; and Gaudet, H. 1944. The people's choice. New York: Columbia University Press. Leal, J. L. 1982. La democracia en la vida cotidiana. El Pais (January 29): 14. LeDuc, L. 1981. The dynamic properties of party identification: A four nation comparison. European Journal of Political Research 9: 257-268. Lehnen, R. G. and Koch, G. G. 1974. Analyzing panel data with uncontrolled attrition. Public Opinion Quarterly 38: 596-612. Leijon, A. 1974. Den svenska modellen (The Swedish model). Stockholm: Aldus. Levitt, T. 1976. Management and the "post-industrial" society. Public In­ terest 44: 69-103. Lewis-Beck, M. 1977. The relative importance of socioeconomic and political variables in public policy. American Political Science Review 71: 559566. . 1981. The electoral politics of the French peasantry: 1946-1978. Political Studies 29: 517-536. . 1983. Economics and the French voter. Public Opinion Quarterly 47: 347-360. Lewis-Beck, M., and Bellucci, P. 1982. Economic influences on legislative elections in multiparty systems: France and Italy. Political Behavior 4: 93-107. Lijphart, A. 1971. Class voting and religious voting in the European de­ mocracies: A preliminary report. Glasgow: University of Strathclyde. . 1974. The Netherlands: Continuity and change in voting behavior. In Electoral behavior: A comparative handbook, ed. R. Rose, pp. 227268. New York: Free Press. . 1975. The politics of accommodation: Pluralism and democracy in The Netherlands. Berkeley: University of California Press. . 1979. Religious vs. linguistic vs. class voting: The crucial experiment of comparing Belgium, Canada, South Africa and Switzerland. American Political Science Review 73: 442-458.

References

493

. 1981. Political parties: Ideologies and programs. In Democracy at the polls, ed. D. Butler et al., pp. 26-52. Washington, D.C.: American Enterprise Institute. Lindberg, L. 1976. Strategies and priorities for comparative research. In Politics and the future of industrial society, ed. L. Lindberg, pp. 222286. New York: McKay. Linz, J. J. 1964. An authoritarian regime: Spain. In Cleavages, ideologies, and party systems, ed.E. Allardt and Y. Littunen. Helsinki: Academic Bookstore. . 1978. II sistema partitico spagnolo. Revista italiana di scienza po­ litico 3: 365-414. . 1981a. La crisis de un estado unitario, nacionalismos perifericos y regionalismo. In La Espana de las autonomias, ed. R. Acosta et al. Madrid: Espasa Calpe. . 1981b. A century of politics and interests in Spain. In Organizing interests in Western Europe, ed. S. Berger. Cambridge: Cambridge Uni­ versity Press. Linz, J. J. etal. 1981. Informe sociologico sobre el cambiopolitico en Espana, 1975-1981. Madrid: Fundacion Foessa. Lipset, S. M. 1960. Political man: The social bases of politics. Garden City, N.Y.: Doubleday. . 1964. The changing class structure and contemporary European pol­ itics. Daedalus 93: 271-303. . 1981a. Political man: The social bases of politics. 2nd ed. Baltimore: Johns Hopkins University Press. . 1981b. The revolt against modernity. In Mobilization, center-pe­ riphery structures and nation-building, ed. P. Torsvik, pp. 451-500. Bergen: Universitetsforlaget. . 1982. No room for the ins: Elections around the world. Public Opinion 5: 41-43. Lipset, S. M. et al. 1954. The psychology of voting: An analysis of voting behavior. In Handbook of social psychology, ed. G. Lindzey, 2: 1,124— 1,175. Reading, Mass: Addison-Wesley. Lipset, S. M., and Rokkan, S. 1967. Cleavage structures, party systems and voter alignments. In Party systems and voter alignments, ed. Lipset and Rokkan, pp. 1-67. New York: Free Press. Loewenberg, G. 1978. The development of the German party system. In Germany at the polls, 1976, ed. K. Cerny, pp. 1-28. Washington, D.C.: American Enterprise Institute. Lopez Pina, Α., ed. 1977. La Espana democratica y Europa. Madrid: Edi­ torial Cambio 16. Lopez Pina, Α., and Aranguren, E. 1976. La cultura politico de la Espana de Franco. Madrid: Taurus.

494

References

Lopez Pintor, R. 1982. La opinion publica espanola: Del Franquismo a la democracia. Madrid: Centra de Investigaciones Sociologicas. Lopez Pintor, R., and Buceta, R. 1975. Los espanoles de los anos 70. Madrid: Editorial Tecnos. Luzzatto Fegiz, P. 1956. II volto sconosciuto dell'Italia: Died anni di sondaggi DOXA. Milan: Giuffre. . 1966. II volto sconosciuto dell'Italia: Died anni di sondaggi DOXA. 2nd ser. 1956—1965. Milan: Giuffre. Macridis, R. C. 1975. French politics in transition: The years after de Gaulle. Cambridge, Mass.: Winthrop. Madeley, J. 1977. Scandinavian Christian democracy. European Journal of Political Research 5: 267-286. Maier, C. S. 1977. The politics of productivity: Foundations of American international economic policy after World War II. International Organ­ ization 31: 607-633. Maisel, L., and Cooper, J., eds. 1978. Political parties: Development and decay. Beverly Hills: Sage Publications. Malefakis, Ε. E. 1970. Agrarian reform and peasant revolution in Spain. New Haven: Yale University Press. Mannheim, K. 1952. The problem of generations. In Essays on the sociology of knowledge, ed. P. Kecskemeti, pp. 276-320. London: Routledge and Kegan Paul. Maravall, J. M. 1981. La politica de la transicion, 1975-1980. Madrid: Taurus. Markus, G. B. 1983. Dynamic modeling of cohort change: The case of political partisanship. American Journal of Political Science 27: 717— 739. Markus, G. B., and Converse, P. 1983. A dynamic simultaneous equation model of electoral choice. American Political Science Review 73: 1,0551,070. Marsh, A. 1977. Protest and political consciousness. Beverly Hills: Sage Publications. Marshall, Τ. H. 1964. Class, citizenship, and social development. Garden City, N.Y.: Doubleday. Martinotti, G. 1978. Le tendenze deU'elettorato italiano. In La politica nell' Italia che cambia, ed. A. Martinelli and G. Pasquino. Milano. Mason, K. O.; Mason, W. M.; Winsborough, H. H.; and Poole, K. W. 1973. Some methodological issues in cohort analysis of archival data. American Sociological Review 38: 242-258. Matheson, D. 1979. Ideology, political action and the Finnish working class. Helsinki: Societas Scientiarium Fennica. Mayer, L. 1980. A note on the aggregation of party systems. In Western

References

495

European party systems, ed. P. Merkl, pp. 335-347. New York: Free Press. McDonough, P.; Lopez Pina, Α.; and Barnes, S. H. 1981. The Spanish public in political transition. British Journal of Political Science 11: 49-79. McDonough, P., and Lopez Pina, A. 1982a. Dynamics of the Spanish tran­ sition. Paper delivered at the IPSA World Congress, Rio de Janeiro, August 9-14. . 1982b. Perceptions of the post-authoritarian Cortes. Paper delivered at the Conference on Parliaments, Policy, and Regime Support, Duke University, Durham, N.C., December 2-5. McDonough, P.; Lopez Pina, Α.; and Barnes, S. H. 1983. Authority and autonomy: Spanish democracy in comparative perspective. Paper deliv­ ered at the Conference on the Spanish Elections, University of North Carolina, Chapel Hill, N.C., May 22-26. McHale, V., and Shaber, S. 1976. From aggressive to defensive Gaullism. Comparative Politics 8: 291-306. McKean, M. 1981. Environmental protest and citizen politics inJapan. Berkeley: University of California Press. McKelvey, R., and Zavoina, W. 1975. A statistical model for the analysis of ordinal level dependent variables. Journal of Mathematical Sociology 4: 103-120. McKeough, K. L., and Miller, J. D. 1979. The generational transmission of party identification: A different view of partisanship and transmission. Paper presented at the Annual Meeting of the Midwest Political Science Association, Chicago, 111. Meadows, D. et al. 1972. The limits to growth. Washington, D.C.: Potomac Associates. Meisel, J. 1975. Working papers in Canadian politics. Montreal: McGillQueen's Press. Merkl, P., ed. 1980. Western European party systems. New York: Free Press. Michelat, G., and Simon, M. 1977. Religion, class, and politics. Comparative Politics 10: 159-184. Miller, A. H. 1974. Political issues and trust in government: 1964-1970. American Political Science Review 68: 951-972. Miller, W., andLevitin, T. 1976. Leadershipandchange. Cambridge, Mass.: Winthrop. Miller, W., and Stouthard, P. 1975. Confessional attachment and electoral behavior. European Journal of Political Research 3: 219-258. Ministerio de Hacienda. 1980. El presupuesto espanol de ingressos. Boletin de Informacion de la Direction General de Tributos 3: 1-36. Miyake, I. 1981. Party support: Its stability and functions. Paper presented at the Conference on Japanese Voting Behavior, Kyoto, Japan.

496

References

Morgan, J. N., and Messenger, R. C. 1973. THAID: A sequential analysis program for the analysis of nominal scale dependent variables. Ann Arbor, Mich.: Institute for Social Research. Morris, C. B. 1972. Surrealism and Spain, 1920-1936. Cambridge: Cam­ bridge University Press. Mulaik, S. A. 1972. The foundations of factor analysis. New York: McGrawHill. Muller, Ε. N. 1982. The restoration of democracy after military rule. Paper delivered at the IPSA World Congress, Rio de Janeiro, August 8-14. Muller-Rommel, F. 1982. Parteien Neuentyps in Westeuropa: Eine vergleichende Analyse. Zeitschrifi fuer Parlamentsfragen 13: 367-391. Munoz, J.; Roldan, S.; and Serrano, S. 1978. La internacionalizacion del capital en Espana. Madrid: Cuadernos para el Di^logo. Murakami, Y. 1982. The age of new middle mass politics: The case of Japan. Journal of Japanese Studies 8: 29-72. Murphy, D. etal. 1982. Haben 'Links' and 'Rechts' noch Zukunft? Politische Vierteljahresschrift 22: 398^14. Negri, G., and Pappagallo, C. 1981. L'Italia negli anni settanta. Roma: Armando Armando. NHK Hoso Seron Chosajo. 1976. Dai 34-kai Shugiin senkyo zenkoku chosa kekkahyo. Tokyo: Nihon Hoso Kyokai. Nie, N. H.; Verba, S.; and Petrocik, J. 1979. The changing American voter. Cambridge: Harvard University Press. Nielsen, H. J. 1976. Politiske meninger og vaelgeradfaerd (Political opinions and voting behavior). In Vaelgere i 70'erne (Voters in the '70s), ed. O. Borre et al. Kobenhaven: Akademisk Forlag. Niemi, R. G. 1974. How family members perceive each other. New Haven: Yale University Press. Niemi, R.; Katz, R.; and Newman, D. 1980. Reconstructing past partisanship: The failure of the party identification recall questions. American Journal of Political Science 24: 633-651. Niemoller, B. 1980. Links en rechts in de zeventiger jaren. Mededelingen van de subfaculteit der algemene politieke en sociale wetenschappen nr. 6/7. Universiteit van Amsterdam. Nilsson, S.-S. 1979. Volatile voters, established parties. Paper presented at the ECPR workshop sessions, Florence, Italy. Norpoth, H. 1979. The parties come to order. American Political Science Review 73: 724-736. . 1980. Choosing a coalition partner. Comparative Political Studies 12:424-441. Norpoth, H., and Rusk, J. 1982. Partisan dealignment in the American elec­ torate. American Political Science Review 76: 522-537.

References

497

Offe, C. 1972. Advanced capitalism and the welfare state. Politics and Society 2: 48(M89. . 1982. Competitive party democracy and the Keynesian welfare state. Paper delivered at the IPSA World Congress, Rio de Janeiro, August 8-14. Orti, A. 1975. Estudio introductorio. In Oligarquia y caciquismo como la forma actual de gobierno en EspaAa, ed. J. Costa. Madrid: Ediciones de la Revista de Trabajo. Page, B., and Brody, R. 1972. Policy voting and the electoral process. American Political Science Review 66: 979-995. Page, B., and Jones, C. 1979. Reciprocal effects of policy preferences, party loyalties, and the vote. American Political Science Review 73: 1,0711,089. Pappi, F.-U. 1976. Sozialstruktur und politische Konfiikt in der Bundesrepublik (Habilitationsschrift). Cologne: University of Cologne. . 1977. Sozialstruktur, gesellschaftlicher Wertorientierung und Wahlabsicht. Politische Vierteljahresschrift 18: 195-229. Paramaio, L. 1982. The crisis of a magical discourse: The disenchantment of politics in post-Franco Spain. Paper delivered at the IPSA World Congress, Rio de Janeiro, August 8-14. Parkin, F. 1979. Marxism and class theory. New York: Columbia University Press. Partanen, J. 1978. Welfare, politics and a way of life. Acta Sociologica 21 (supplement): 59-72. Payne, S. G. 1981. Church politics, fascism, and Franquism in Spain. Unpublished manuscript. Pedersen, M. N., ed. 1979a. Danskpolitik i 1970 'erne. Copenhaven: Samfunds videnskabelight Forlag. . 1979b. The dynamics of European party systems. European Journal of Political Research 7: 1-26. Pempel, T. J. 1982. Policy and politics in Japan. Philadelphia: Temple University Press. Penniman, H. R., ed. 1977. Italy at the polls: The parliamentary elections of 1976. Washington, D.C.: American Enterprise Institute. . 1981. Italy at the polls: The parliamentary elections of 1979. Washington, D.C.: American Enterprise Institute. PeVez Dias, V. 1980. Close obrera, orden social y consciencia de close. Madrid: Fundacion del Institute Nacional de Industrie. Pesonen, P., and Sankiaho, R. 1979. Kansalaiset ja kansanvalta (Citizens and democracy). Porvoo-Helsinki-Juva: Werner Soderstrom Oy. Peters, G., and Hennessey, T. 1975. Political development and public policy in Sweden: 1865-1967. In Comparative public policy issues, theories

498

References

and methods, ed. C. Liske, W. Loehr, and J. McCamant, pp. 127-151. New York: Halsted Press. Petersson, O. 1977. Valundersokningar, Rapport 2, Valjarna och valet 1976. Stockholm: Statistiska Centralbyran. . 1978. The 1976 election: New trends in the Swedish electorate. Scandinavian Political Studies 1: 109-121. Petrocik, J. R. 1981. Party coalitions. Chicago: University of Chicago Press. Pierce, J. C , and Rose, D. 1974. Non-attitudes and American public opinion: The examination of a thesis. American Political Science Review 68: 626649. Pierce, R. 1968. French politics and political institutions. New York: Harper and Row. . 1980. Book review. American Political Science Review 74: 872. Plaza, R. 1983. Obispos contra el PSOE. Tiempo (August 29): 10-15. Poggi, G. 1968. Le preferenze politiche degli Italiani. Bologna: II Mulino. Pollack, B., and Taylor, J. 1983. Review article: The transition to democracy in Portugal and Spain. British Journal of Political Science 13: 209-242. Pomper, G. 1967. Classification of presidential elections. Journal of Politics 29: 535-566. Press, I. 1979. The city as context: Urbanism and behavioral constraints in Seville. Urbana: University of Illinois Press. Przeworski, A. 1975. Institutionalization of voting patterns, or is mobilization the source of decay? American Political Science Review 59: 49-67. . 1979. Some problems in the study of the transition to democracy. Paper delivered at the workshop on Prospects for Democracy, Wilson Center, Smithsonian Institution, Washington, D.C., September 25-26. . 1980. Material interests, class compromise, and the transition to socialism. Politics and Society 10: 125-154. Putnam, R. 1973. The beliefs of politicians: Ideology, conflict, and democracy in Britain and Italy. New Haven: Yale University Press. . 1977. Italian foreign policy: The emergent consensus. In Italy at the polls: The parliamentary elections of 1976, ed. H. R. Penniman, pp. 287-326. Washington, D.C.: American Enterprise Institute. Rae, D. 1971. The political consequences of electoral laws. New Haven: Yale University Press. Ranney, Α., and Sartori, G., eds. 1978. Eurocommunism: The Italian case. Washington, D.C.: American Enterprise Institute. Reiter, H. 1980. The perils of partisan recall. Public Opinion Quarterly 44: 385-388. Research Committee on the Study of the Japanese National Character. 1979. Kokuminsei no Kenkyu: Dai 6-kai zenkoku chosa. Tokyo: Tokei Suri Kenkyujo.

References

499

Richardson, Β. Μ. 1974. The political culture of Japan. Berkeley: University of California Press. . 1977. Stability and change in Japanese voting behavior, 1958-1972. Journal of Asian Studies 36: 675-693. Riker, W. 1982. Liberalism against populism. San Francisco: Freeman. Robertson, D. 1976. A theory of party competition. New York: Wiley. Rochon, T. 1983. Political change in ordered societies. Comparative Politics 15: 351-373. Rokkan, S. 1967. Geography, religion and social class: Crosscutting cleavages in Norwegian politics. In Party systems and voter alignments, ed. Lipset and Rokkan, pp. 367-444. New York: Free Press. Rosa, J.-J., and Amson, D. 1976. Conditions oconomiques et olections: Une analyse politico-6conom6trique (1920-1973). Revue Frangaise de Sci­ ence Politique 26: 1,101-1,124. Rose, R. 1980. Do parties make a difference? Chatham, N.J.: Chatham House. . 1982. From simple determinism to interactive models of voting. Comparative Political Studies 15: 145-169. Rose, R., ed. 1974. Comparative electoral behavior. New York: Free Press. Rose, R., and Urwin, D. 1969. Social cohesion, political parties, and strains in regime. Comparative Political Studies 2: 7-67. . 1970. Persistence and change in Western party systems since 1945. Political Studies 18: 287-319. Rusk, J. G., and Borre, O. 1974. The changing party space in Danish voter perceptions. European Journal of Political Research 3: 329-361. Sarlvik, B. 1977. Recent electoral trends in Sweden. In Scandinavia at the polls, ed. K. Cerny, pp. 73-130. Washington, D.C.: American Enter­ prise Institute. Sarlvik, B., and Crewe, I. 1983. Decade of dealignment. Cambridge: Cam­ bridge University Press. Salisbury, R. H., and MacKuen, M. 1981. On the study of party realignment. Journal of Politics 43: 523-530. Sanchez, J. M. 1964. The Spanish revolution: The politico-religious back­ ground of the Spanish Civil War. Chapel Hill: University of North Car­ olina Press. Sani, G. 1974. Determinants of party preference in Italy: Toward the inte­ gration of complementary models. American Journal of Political Science 18: 315-329. . 1975. Ricambio elettorale e identificazioni partitiche: Verso una egemonia delle sinistre? Rivista italiana di scienza politico 5: 515-544. . 1976a. Le elezioni degli anni settanta: Terremoto ο evoluzione? Rivista italiana di scienza politica 6: 261-288.

500

References

Sani, G. 1976b. Mass constraints on political realignment: Perceptions of antisystem parties in Italy. British Journal of Political Science 6: 1-32. . 1976c. L'immagine dei partiti neU'elettorato. In Un sistema politico allaprova, ed. M. Caciagli and A. Spreafisco, pp. 85-126. Bologna: II Mulino. . 1979. Ricambio electorate, mutamenti sociali, e preferenze politiche. In La crisi italiana, ed. L. Graziano and S. Tarrow. Turin: Einaudi. . 1980. The political culture of Italy: Continuity and change. In The civic culture revisited, ed. G. Almond and S. Verba, pp. 273-324. Boston: Little, Brown. . 1981a. Italian voters, 1976-1979. In Italy at the polls: The parlia­ mentary elections of 1979, ed. H. Penniman, pp. 34-63. Washington, D.C.: American Enterprise Institute. . 1981b. Partiti e atteggiamenti di massa in Spagna e Italia. Rivista italiana di scienza politico 11: 235-280. Sani, G., and Shabad, G. Forthcoming. Electoral change in Spain in 1982: A polarizing realignment. Revista de Derecho Politico. Sankiaho, R. 1979a. Kestaako hyvinvointivaltio eli quo vadis sosialidemokratia? Sosialistinen Aikakausilehti 3: 25-31. . 1979b. Political alienation of different generations in Finland. Paper presented at the ECPR workshop sessions, Florence, Italy. Sartori, G. 1976. Parties and party systems. New York: Cambridge University Press. Scalapino, R. Α., and Masumi, J. 1962. Parties and politics in contemporary Japan. Berkeley: University of California Press. Scharpf, F. 1977. Public organization and the waning of the welfare state. European Journal of Political Research 5: 339-362. Schmitter, P. C. 1981. Interest intermediation and regime govemability in contemporary Western Europe and North America. In Organizing in­ terests in Western Europe, ed. S. Berger. Cambridge: Cambridge Uni­ versity Press. Schumacher, E. F. 1973. Small is beautiful. New York: Harper and Row. Schwartz, M. 1974. Politics and territory. Montreal: McGill-Queen's Press. Seip, D. 1979. Niederlande. In Das Wahlrecht der Neun: Wahlsysteme in der Europaischen Gemeinschaft, ed. C. Sasse, J. Georgel, and G. J. Hand. Baden-Baden: Nomes Verlagsgesellschaft (in English: European electoral systems handbook. London: Butterworths). Shabad, G., and Gunther, R. 1982. Language, nationalism, and political conflict in Spain. Comparative Politics 14: 443-477. Share, D. 1983. Two transitions: Democratization and the evolution of the Spanish socialist left. Paper delivered at the Annual Meeting of the American Political Science Association, Chicago, 111., September 1-4.

References

501

Sharpe, L., ed. 1979. Decentralist trends in Western Democracies. Beverly Hills: Sage Publications. Shively, W. P. 1972. Voting stability and the nature of party attachments in the Weimar Republic. American Political Science Review 66: 1,2031,225. . 1977. Information costs and the partisan life cycle. Paper presented at the Annual Meeting of the American Political Science Association, Washington, D.C. . 1979. The development of party identification among adults. American Political Science Review 73: 1,039-1,054. . 1980. The nature of party identification: A review of recent developments. In The electorate reconsidered, ed. J. C. Pierce and J. L. Sullivan, pp. 219-236. Beverly Hills: Sage Publications. Skocpol, T. 1982. What makes peasants revolutionary? Comparative Politics 14: 351-375. Smith, G. 1976. Social movements and party systems in Western Europe. In Social and political movements in Western Europe, ed. M. Kolinsky and W. Patterson. New York: St. Martin's Press. Sniderman, P. 1978. The politics of faith. British Journal of Political Science 8: 21^14. Sniderman, P.; Forbes, H. D.; and Melzer, I. 1974. Party loyalty and electoral volatility. Canadian Journal of Political Science 7: 268-288. Sonquist, J. 1973. Searching for structure. Ann Arbor, Mich.: Institute for Social Research. Stapel, J. 1968. Wie en wat staan tussen links en rechts?—een klein opinie onderzoek. Acta Politico 4: 32-40. Statera, G. 1975. The death of a Utopia: The development and decline of student movements in Europe. Oxford: Oxford University Press. Steiner, K. 1965. Local government in Japan. Stanford: Stanford University Press. Steiner, K. et al., eds. 1980. Political opposition and local politics in Japan. Princeton: Princeton University Press. Stephens, J. D. 1981. The changing Swedish electorate: Class voting, contextual effects and voter volatility. Comparative Political Studies 14: 163-204. Stevens, J. D. 1979. The transition from capitalism to socialism. London: Macmillan. Stokes, D., and Iversen, G. 1966. On the existence of forces restoring party competition. In A. Campbell et al., Elections and the political order, pp. 180-193. New York: Wiley. Stouthard, P. 1971. De verkiezingen van maart 1970. Acta Politico 6: 1828.

502

References

Sullivan, J. L.; Piereson, J.; Marcus, G.; and Feldman, S. 1979. The more things change, the more they stay the same: The stability of mass belief systems. American Journal of Political Science 23: 176-186. Sundquist, J. L. 1973. Dynamics of the party system. Washington, D.C.: Brookings Institution. Suzuki, Y. 1981. Why is the performance of the Japanese economy so much better? Journal of Japanese Studies 7: 403^413. Tamames, R. 1983. El paro, la herencia y el futuro. El Pais (January 15): 36. Taylor, C , and Jodice H. 1983. World handbook III. New Haven: Yale University Press. Tezanos, J. F. 1978. Estructura de closes y conflictos de poder en la Espana posfranquista. Madrid: Cuadernos para el Dialogo. Thernborn, G. et al. 1978. Sweden before and after social democracy. Acta Sociologica 21: 44—53. Thijn, E. Van. 1967. Van Partijen naar Stembusaccoorden. In Partijvernieuwing? Open brief. E. Jurgens e. a. Amsterdam. Thomas, J. 1980. Ideological trends in Western political parties. In Western European party systems, ed. P. Merkl, pp. 348-366. New York: Free Press. Thomassen, J.J.A. 1976. Party identification as a cross-cultural concept: Its meaning in The Netherlands. In Party identification and beyond, ed. I. Budge, I. Crewe, and D. Farlie, pp. 63-80. London: Wiley. Toffler, A. 1980. The third wave. New York: Bantam. Tompkins, G. 1975. A causal model of state welfare expenditures. Journal of Politics 37: 392-416. Tsurutani, T. 1977. Political change in Japan. New York: McKay. Tusell Gomez, X. 1976. The popular front elections in Spain, 1936. In Politics and society in twentieth-century Spain, ed. S. G. Payne. New York: New Viewpoints. Ullman, J. C. 1968. The tragic week: A study of anticlericalism in Spain. Cambridge: Harvard University Press. Ullman, P. L. 1971. Mariano de Larra and Spanish political rhetoric. Mad­ ison: University of Wisconsin Press. Valen, H. 1981. Valg og Politikk. Oslo: NKS-Forlaget. Valen, H., and Katz, D. 1967. Political parties in Norway. Oslo: Universitetsforlaget. Valen, H., and Martinussen, W. 1972. Velgere og politiske frontlinjer: Oslo: Gyldendal Norsk forlag. Valen, Η., and Rokkan, S. 1974. The Storting election of 1973. Scandinavian Political Studies 9: 205-218. Valentin, F. 1978. Corporatism and the Danish welfare state. Acta Sociologica 21 (supplement): 73-95.

References

503

Verba, S., and Nie, Ν. H. 1972. Participation in America. New York: Harper and Row. Verba, S.; Nie, N.; and Kim, J. O. 1978. Participation and political equality. Cambridge: Cambridge University Press. Voss, W. 1980. Die Bundesrepublik Deutschland. Stuttgart: Kohlhammer. Wanat, J., and Burke, K. 1982. Estimating the degree of mobilization and conversion in the 1890s: An inquiry into the nature of electoral change. American Political Science Review 76: 360-370. Ward, R. 1959. The community and the political process. In Village Japan, ed. R. Beardsley et al., pp. 398-470. Chicago: University of Chicago Press. Watanuki, J. 1967. Patterns of politics in present-day Japan. In Party systems and voter alignments, ed. S. Lipset and S. Rokkan, pp. 447^66. New York: Free Press. Weisberg, H. F. 1980. A multidimensional conceptualization of party iden­ tification. Political Behavior 2: 33-60. White, J. W. 1970. The Sokagakkai and mass society. Stanford: Stanford University Press. White, 0., and Sjoberg, G. 1972. The emerging 'New Politics' in America. In Politics in the post-welfare state, ed. D. Hancock and G. Sjoberg. New York: Columbia University Press. Wilensky, H. 1975. The welfare state and equality. Berkeley: University of California Press. . 1976. The "New Corporatism," centralization and the welfare state. Beverly Hills: Sage Publications. Wilson, B. 1979. The return of the sacred. Journal for the Scientific Study of Religion 18: 268-280. Wilson, F. L. 1977. Sources of party transformation: The case of France. Paper prepared for delivery at the Midwest Political Science Association Annual Meeting, Chicago, 111. . 1979. The revitalization of French parties. Comparative Political Studies 12: 82-103. Wolinetz, S. 1981. The transformation of Western European party systems revisited. West European Politics 3: 4-28. Worre, T. 1978. Klassepartier og klassestemme i de nordiske lande (Class parties and class voting in the Nordic countries). Nordic Congress of Political Science, Bergen. Wright, E. O. et al. 1982. The American class structure. American Socio­ logical Review 47: 709-726. Wright, G. 1975. Interparty competition and state social welfare policy. Journal of Politics 37: 198-203. . 1977. Contextual models of electoral behavior. American Political Science Review 71: 497-509.

504

References

Yankelovitch, D. 1974. The new morality. New York: McGraw-Hill. Yoron Chosa Nenkan. 1970-1981. Tokyo: Naikaku Soridaijin Kanbo Kohoshitsu. Zirakzadeh, C. E. 1982. The puzzling "transformation" of the Spanish state. Paper delivered at the Conference of Europeanists, Washington, D.C., April 30. Zuckerman, A. 1982. New approaches to political cleavage. Comparative Political Studies 15: 131-144.

INDEX

Abramson, Ρ R , 29, 30, 216, 242, 247 Achen, C Η , 40 Agrarian Liberal party (Venstre), Denmark, 341, 342 Alemann, U von, 471 Alford, R R , 27, 29, 434 Alford index, 31, 64, 84, 127, 129, 351, 437 Ahanza Popular, Spain, 366, 382n, 385, 387 Allardt, Ε , 107, 121 Alhnson, G D , 172 Almench, Ρ , 381 Almond, G , 210 Alt, J , 10η, 81, 300, 301, 304, 305, 307, 309, 310, 313, 327, 328n, 433 Altares, Ρ , 368n Amson, D , 445 Anders Lange's party, 332 Andersen, Κ , 13, 128, 138, 244, 252, 275 Anderson, J , 466 Andeweg, R Β , 268, 276, 296 Aimers, Ε , 76, 90 Anti-Revolutionary party (ARP), Nether­ lands, 267, 274, 284 Aranguren, Ε , 373 Anan, A , 218 Anja, J Μ , 365 Artola, Μ , 368 Asahi Nenkan, 195, 196, 197, 198, 202 Asher, Η , 458 Attiyeh, R , 90 Aviv, A , 391n Aviv, I , 391n Baena del Alcazar, Μ , 368 Baker, Κ L , 3, 8, 12n, 16, 20, 29, 71, 104, 105, 106, 107, 108, l l l n , 112, 119, 126, 128, 129n Balbus, I D , 370 Barbagh, Μ , 214, 217, 218, 219, 220 Barnes, S Η , 4, 10n, 17, 38, 58, 62, 78, 81, 84, 87, 117n, 118n, 131, 207n, 210, 21 In, 218, 293, 370, 374, 383, 430, 462 Beck, Ρ A , 14, 19, 125, 135, 138, 208, 233, 240, 241, 246, 249n, 264, 265, 283, 460

Beer, S , 298, 301, 304 Bell, D , 3,4, 106, 121,456 Bellucci, Ρ , 400, 445 Beltran, Μ , 369 Berelson, Β , 17 Berger, S , 8, 380n, 394 Berglund, S , 330 Beyme, Κ von, 105 Bezold, C , 471 Bilski, R , 71n Blackmer, D L , 228 Blough, R A , 391n Books, J , 29 Borre, Ο , 16, 18, 97, 236, 330, 348, 461 Box, G Ε Ρ , 151η Bnnkgreve, C , 292 Brody, R , 120 Buceta, R , 373 Buddhist Clean Government party (CGP), Japan, 167, 181, 182, 184, 186, 187, 188, 189, 191, 192, 193, 197, 203 Budge, I , 12, 32, 211, 220 Buerklin, W , 28, 32, 41, 76, 78n, 112, 120, 124, 128 Burke, Κ , 244, 253 Burnham, W D , 14, 134, 135, 138, 458, 460 Burstein, Ρ , 17 Butler, D , 12n, 13, 75, 303, 470

Cain, Β , 12n Caldeira, G , 429 Cameron, D R , 370, 428, 429, 430, 443, 473 Campbell, A , 11, 12, 13, 129, 241n, 242, 257 Campbell, Β A , 13n, 429, 430, 440, 443 Carlsson, G , 208 Carmines, Ε G , 99, 100, 120, 128, 145n, 149n, 151n, 237, 240, 265, 455, 459 Carr, R , 392 Carter, J , 91, 260 Catholic People's party (KVP), Netherlands, 270, 272, 274, 277, 278, 279, 280, 281, 282, 284, 285, 292

506 Center Democratic Union (UCD), Spain, 365, 386, 387, 390 Center party, Scandinavia, 341, 353, 354, 360, 361 Cerny, Ρ G , 434 Christian Democratic Appeal (CDA), Nether­ lands, 270, 272, 274, 278, 281, 285, 286, 289, 290 Chnstian Democratic party (DC), Italy, 102, 103, 205, 206, 207, 209, 210, 214, 218, 220, 223, 225, 226, 228, 229, 230 Chnstian Democratic Union (CDU), West Germany, 98, 99, 105, 106, 114, 115, 120, 127, 130, 131, 132 Chnstian Histoncal Union (CHU), Nether­ lands, 270, 274, 278, 284 Chnstian party, Norway, 354, 360 Chnstian Social Union (CSU), West Ger­ many, 105, 106, 114, 115, 120, 127, 131, 132 Chrystal, Κ A , 300 church, 17, 26, 62, 116, 118, 130, 208, 223, 225, 254, 267, 268, 437, 438 citizen action groups, 131, 199, 471, 475 Citnn, J , 89 Clarke, Η , 403, 404n, 406, 409n, 410n, 41 In, 417n, 424 class voting, 29, 71, 74 See also Alford in­ dex cleavage agranan-industnal, 101, 160, class (social), 26-27, 53, 57, 60, 64, 68, 127, 355, 443, 455, generation, 127, 224, 293, 294, 374, 375, linguistic, 407, 408, party related, 239, 400, 453, 472, racial/ethnic, 100, 141, 153, regional (center/penphery), 355, 365, 377, 381, 382, 407, 408, 424, 439, 443, 454, religious, 64, 106, 115, 377, 380, 407, 408, 443, 454, social, 106, 107, 458, 463, value, 26-27, 64-65, 160, 376 Clubb, J Μ , 13, 19, 128, 129, 134, 135, 138, 206, 208, 460 codetermination (Mitbestimmung), 74, 91, 109 cognitive mobilization, 7, 18-19, 22, 474 Cole, R , 72 Collier, D , 393 Communist party (PCF), France, 442 Communist party (PCI), Italy, 205, 207, 208, 209, 210, 214, 218, 220, 222, 223, 225, 226, 227, 228, 229, 230

Index Communist party (CP), Japan, 159, 167, 168, 184, 185, 187, 188, 191, 192, 193, 197 Communist party (CPN), Netherlands, 273, 280 Communist party, Scandinavia, 332, 334, 336, 360 Communist party (PCE), Spain, 374, 386, 387, 390 community integration, 17-18, 97, 106, 169, 171, 177, 179, 187 Conradt, D , 10η conservative See Right Conservative party, Bntain, 299, 300, 304, 305, 306, 309, 316, 318, 321, 325, 326, 327 Conservative party, Scandinavia (Danish Konservative, Norwegian H0yre), 334, 340, 341, 346, 354, 359 Converse, Ρ Ε , 12, 13, 40, 84, 137, 138, 139, 216, 241η, 242, 244, 245, 247, 248, 249η, 254, 257, 305, 370, 428, 429, 430, 469η Cooper, J , 460 Craxi, Β , 206 Crewe, I , 10η, 12, 14, 32, 211, 301, 302, 304, 305, 307, 309, 310, 328η Crozier, Μ , 3 Curtice, J , 301 Curtis, G L , 163, 170 Daalder, Η , 8, 267, 268, 293 Dahl, R , 3, 6, 166, 370 Dahrendorf, R , 74, 107, 370 Dalton, R , 3, 4, 8, 12n, 16, 18, 20, 29, 71, 76, 97, 99, 104, 105, 106, 107, 108, l l l n , 119, 124, 126, 128, 129n, 131, 174, 460, 466, 475 Daudt, Η , 275, 288 dealignment (depillanzation), 11, 96, 125, 195, 211, 233, 271, 346, 399, 413, 414, 451, 474, 475, nature of, 8, 14, 68, 236, 240, 250, 253, 267, 293, 298, 299, 302, 306, 325, 329, 353 De Esteban, J , 370 de Gaulle, C , 102, 425, 426, 430, 431, 446 Dei, Μ , 214 Del Campo, S , 368, 369 De Miguel, A , 375n Democratic party, United States, 99, 100, 135, 136, 138, 145, 150, 155, 156, 157,

Index 234, 241, 242, 244, 245, 252, 253, 254, 260, 261,284, 306,433 Democratic Socialist party (DSP), Japan, 101, 163, 167, 181, 182, 184, 186, 187, 188, 189, 191, 193, 201, 203 Democratic Socialists '70 (DS'70), Nether­ lands, 270, 272, 278, 279, 280, 281, 282, 283, 293 Democrats '66 (D'66), Netherlands, 270, 271, 272, 274, 277, 278, 279, 280, 281, 284, 285, 289, 290, 292, 293, 465 Dennis, J , 258 Deutsch, Ε , 448 DeVnes, W , 10η Diefenbaker, J , 399, 402, 419 DiPalma, G , 207 direct democracy (initiative, referendum), 238, 470, 471 Dittnch, Κ , 17, 235, 236 Dobson, D , 241n Downs, A , 15, 473 Dupeux, G , 428, 429, 430 Duverger, Μ , 76, 90, 435 Dye, Τ , 71n Easton, D , 72, 73, 258 Ehrmann, Η W , 426 Eijk, C van der, 276, 277, 284, 288 electoral system, 166, 207, 234, 274, 426, 433, 465 See also proportional represen­ tation Elvander, Ν , 75, 76, 90, 336 embourgeoisement, 15, 20 Engle, R , 90 environmentahsm, 28, 43, 91, 99, 103, 105, 115, 131, 132, 465,474 Epstein, L , 460, 470 Enckson, R , 388n Enkson, R S , 13, 138, 154, 261 Esping-Anderson, G , 71 estabhshment/anti-estabhshment, 60, 61, 62, 73, 74, 78, 89, 90, 118, 181, 182, 275 Euro-Barometer, 31, 34, 44, 56, 63, 64, 210n, 211,216,226 European Economic Community (EEC, Common Market), 316, 318, 323, 337, 359, 360, 361, 368, 369, 370, 395, 426 Fabns, G , 218 Fair, R , 309 Falconen, G R , 172

507 Fanjul Sedeno, J Μ , 379n Farah, Β , 81, 87 Farhe, D , 12,32,211,220 Farmer's party (BP), Netherlands, 277, 279, 283 Farnsworth, L W , 203 Fascist party, Italy, 206, 207, 208 Feist, U , 128 Feldman, S , 40 Ferejohn, J , 12n Fernandez de Castro, I , 368 Ferrara, Μ , 75, 76 Fietkau, Η J , 28 Finer, S , 298, 301 Fiorina, Μ Ρ , 235, 308, 310, 311, 312, 433 Fishman, R Μ , 392 Flanagan, S C , 18, 20, 97, 114, 162, 163, 170, 171n, 173, 175, 178n, 197, 453, 460, 466 Flanigan, W Η , 13, 19, 128, 129, 134, 135, 138, 206, 208, 460 floating vote, 172, 236, 252, 287-288, 432 Ford, G , 264 Forrester, J , 7n Foster, C , 471 fractionahzation (fragmentation), 9, 164, 177, 191, 199, 200, 202, 203, 302, 339, 382, 451, 467 Franco (Franquist), 237, 365, 366, 368, 371, 372, 373, 375, 377, 378, 379, 380, 381, 382, 392, 393, 394 Free Democratic party (FDP), West Ger­ many, 105, 114, 115, 128, 130, 132 Freeman, J R , 369 Frey, Β , 77 Fuentes Quintana, Ε , 388 functional model, 97, 237, 239, 255, 256, 362, 460, 462, 469 Fundacion Foessa, 368 Gaghani, G , 388n Galh, G , 226 Garcia Duran, J , 368 Garcia Ferrando, Μ , 382 Garcia Madana, J Μ , 368 Gaudet, Η , 17 Gaulhst party, France, 426, 428, 430, 431 Geller, D S , 393 generations, 20, 208, 247, 250, 264, 292, 295, 303, 391, 453, generational differ-

508 generations (cont) ences in class voting, 124, 129, 304, 353, 453, generational replacement, 33, 108, 112, 123, 124, 233-234, 246, 252, 375, 399, 409 Giner, S , 367n, 388n Giscard d'Estaing, V , 36, 445, 446 Gladdish, Κ R , 271 Glenn, Ν D , 29, 216 Ghstrup's tax party, 8, 359, 458, 464, 468 Gluchowski, Ρ , 126 Goldstone, J A , 394 Goldthorpe, J Η , 16, 388n, 457 Goldwater, Β , 146 Gonzalez, F , 372 Goytre, A , 368 Grauhan, R , 75n Graziano, L , 222 Greenberg, D , 326n Green party (Greens), West Germany, 91, 99, 105, 115, 118, 128, 131, 132, 133, 465, 468 Greenstein, F I , 257, 429 Gunther, R , 370, 391n Gustafsson, Β , 70 Habermas, J , 3 Hadley, C , 4, 16, 20, 75, 100 Hamilton, R , 16 Harmel, R , 474 Hauss, C , 463n Heath, Ε , 316 Heclo, Η , 75n, 77, 90 Heidenheimer, A , 111 Hennessey, Τ , 7In Hentila, S , 70 Heunks, F , 81, 87 Hibbs, D , 299, 309, 310 Hickel, R , 75n Hildebrandt, Κ , 3, 8, 12n, 16, 20, 29, 71, 104, 105, 106, 107, 108, 11 In, 112, 124, 126, 129n, 131 Hochstein, A , 14, 125, 240, 428, 429, 447 Hofferbert, R I , 443 Holmberg, S , 330, 347, 353, 357 House, J S , 212 Houska, J J , Jr , 288 Howe, I , 370 Huntington, S , 3, 456 ideology See Left/Right dimension

Index Inchausti, A , 388n Inchausti, J A , 388n incumbent performance, 371, 433, 446 Inglehart, R , 3, 4, 8, 14, 20, 26, 29, 31, 57, 62, 76, 78, 86, 105, 108, 109n, 118, 124, 125, 132, 174, 240, 254, 292, 374n, 428, 429, 447, 452, 453, 456 Inoguchi, Τ , 162 International Labour Organization, 6 International Monetary Fund, 6 Inbarne, F , 382 Irvine, W , 403n, 423n Irwin, G A , 17, 235, 236, 274 issues, 34, 87, 109, 189, 452, economic, 5, 33, 44, 45, 46, 47, 109, 115, 190, 202, 358, 359, 361, 378, 379, 388, foreign policy, 41-42, 109, 190, moral, 191, 359, 387, New Left/New Right, 98, 103, New Politics, 116, quality of life, 4, 98, 105, 108, racial, 99, 100, social, 4, 105, 378 Iversen, G , 137 Jackson, J , 305 Jacob, J , 455 Janda, Κ , 4 1 , 474 Japan Statistical Yearbook, 161, 166, 168, 178, 198, 199 Japan Times Weekly, 203n Japan-U S Security Treaty, 190 Jenkins, G Μ , 151n Jennings, Μ Κ , 242, 254η, 259 Jenson, J , 403, 404n, 406, 409n, 410n, 41 In, 417n, 424 Jodice, Η , 7, 9n John Birchers, 100 Johnson, L Β , 260 Jones, C , 305, 328 Jong, J J de, 291 Kaase, Μ , 4, 38, 58, 62, 68, 78, 84n, 117n, 118n, 131, 293, 374,462 Karlsson, Κ , 208 Katz, D , 18, 348, 461 Katz, R , 145, 154, 218, 460, 466 Kautsky, Κ , 70 Kawada, Η , 162 Keifer, C W , 172 Kemp, D , 16, 20 Kennedy, J F , 42, 136, 146, 260 Kerr, Η , 16, 78n, 120, 130 Kesselman, Μ , 375

509

Index Kessler, R., 326n Key, V. O., 134, 135, 139, 265, 308, 446, 459 Kim, J.-O., 6, 346 King, Α., 15 Kirchheimer, O., 8, 106, 160 Klingemann, H., 31, 32, 68, 78n, 84, 117n, 124 Koch, G. G., 371n Kohei, S., 198, 200 Kommers, D., I l l Komhauser, W., 16 Κοφί, W., 17, 70, 71, 72, 74, 75, 76 Korzec, M., 292 Kosugi, T., 194 Kubota, Α., 178n Kuhnle, S., 70 Kuyper, Α., 267 Labor-Farm party, Japan, 159 Labor party (PvdA), Netherlands, 270, 272, 273, 277, 278, 279, 280, 282, 284, 285, 289, 290 Labor party, Scandinavia, 334, 336, 337, 338, 359, 360 Labour party, Britain, 298, 299, 300, 304, 305, 306, 307, 308, 309, 310, 314, 317, 318, 320, 321, 323, 326 Ladd, E. C.,4, 16,20,75, 100 LaMalfa, U., 206n Lancaster, T. D., 369 Lane, R., 3 Lange, P., 222 Lazarsfeld, P., 17 Leal, J. L., 382n LeDuc, L., 399, 403, 404, 406, 407n, 409n, 410n, 41 In, 417n, 424 Left (progressives, liberals), 222, 223, 226, 229; New Left (liberals), 4, 21, 100; Old Left, 113, 118 Left Radical party, France, 442 Left/Right dimension, 78, 84, 115, 185, 223, 273, 292, 355, 368, 429, 430, 440; correlates of, 32, 39-40, 42, 124, 376; distribution of, 36, 440, 441; meaning of, 25, 31, 37, 38; party positions on, 58, 59, 115, 116,273,274 Lehnen, R. G., 371n Leijon, Α., 72 Levitin, T., 4, 100 Levitt, T., 77

Lewis-Beck, Μ., 71n, 400, 435, 442, 445 Liberal Democratic party (LDP), Japan, 101, 102, 159, 160, 163, 164, 165, 167, 168, 172, 176, 178, 179, 180, 182, 184, 185, 186, 188, 189, 191, 193, 195, 197, 198, 199, 200, 201, 202, 203, 204 Liberal party, Britain, 302 Liberal party, Canada, 402, 404, 411, 414, 415, 417, 418, 419, 420, 422, 423, 424 Liberal party (PLI), Italy, 210, 218, 229 Liberal party ( W D ) , Netherlands, 272, 274, 277, 278, 279, 280, 282, 283, 285, 289, 293 life-cycle effects, 224, 247, 249, 303 Lijphart, Α., 27, 28, 54, 267, 291, 408 Lindberg, L., 77, 90 Lindstrom, U., 330 Linz, J. J., 365n, 368, 370, 382n Lipset, S. M., 3, 4, 7, 16, 20, 26, 27, 29, 37, 98, 99n, 100, 107, 135, 390, 424n, 454, 456, 459, 467, 473 Lockheed, 165, 191, 196 Loewenberg, G., 104 Lopez Guerra, L., 370 Lopez Pina, Α., 237, 370, 373, 382n, 383, 392 Lopez Pintor, R., 365n, 373 Luttbeg, N., 261 Luzzatto Fegiz, P., 210 McCloskey, H., 212 McDonald, M. D., 97, 163, 197 McDonough, P., 237, 370, 382n, 383 McHale, V.,455 McKean, M., 199, 471 McKelvey, R., 313 McKeough, K. L., 265n MacKuen, M., 138 McPhee, W., 17 Macridis, R. C , 434, 439 Madeley, J., 455 Maier, C. S., 370 Mair, P., 8 Maisel, L., 460 Malefakis, Ε. E., 368 Mannheim, K., 250 Maravall, J. Μ., 365n Markus, G. B., 139, 242n, 244, 248n Marsh, Α., 84, 89,471 Marshall, Τ. H., 369 Martinotti, G., 219

510 Marx, Κ , 26, 27, 106, 222, 226 Marxist-Leninist, 102, 453, 455 Mason, Κ Ο , 248η Mason, W Μ , 212, 248η mass society, 16-17 Masumi, J , 162 Matheson, D , 74 Mayer, L , 9n May revolts (France, 1968), 3, 425, 451, 452 Meadows, D , 7n media, 7, 106, 460, 461, 466 Meisel, J , 403n Merkl, Ρ , 104 Messenger, R C , 383n Michelat, G , 436, 438 Miller, Α Η , 89 Miller, J D , 265n Miller, W Ε , 4, 12, 13, 100, 241n, 242, 257 Muusteno de Hacienda, 370 Mitterand, F , 36, 400, 428, 444, 445 Miyake, I , 195 Moral Majority, 100 Morgan, J Ν , 383n Morns, C Β , 377n Mouvement Republican! Populaire (MRP), France, 426, 436 Mulaik, S A , 378n Muller, Ε Ν , 367η Muller-Rommel, F , 28 Munoz, J , 392 Murakami, Υ , 199, 457 Murphy, D , 32 National Democratic party (NPD), West Ger­ many, 465 Navarro, Μ , 369 Negri, G , 206n New Deal, 13, 97, 141, 158, 265 New Democratic party (NDP), Canada, 403 New Liberal Club (NLC), Japan, 101, 163, 167, 182, 184, 185, 186, 187, 188, 189, 191, 192, 193, 194, 203, 458, 464, 468 Newman, D , 145, 154, 218 New People's party, Norway, 464 New Politics/Old Politics dimension, 20, 50, 53, 99, 108, 114, 118, 120, 126, 130, 132, 180, 188, 194, 457, 473 new social movements, 19, 28, 33, 131, 132, 182, 471

Index Nie, Ν Η , 6, 40, 171, 216, 346, 371η Nielsen, Η J , 330, 357 Niemi, R G , 145, 154, 218, 254n, 255n, 259 Niemoller, Β , 276, 277, 284, 288 Nilsson, S S , 71, 75 Nixon, R Μ , 42, 260, 264 nonpartisans, 244, 245, 253, 254, 256, 257, 258, 261, 262, 264, 265, 303, 348, 349, 367, 384, 422 Norpoth, Η , 130, 131, 249 nuclear power, 4, 28, 29, 35, 42, 47, 51, 52, 90, 113, 115, 118, 132, 360, 458 Offe, C , 90, 370 Orti, A , 394 Pacifist Socialist party (PSP), Netherlands, 277, 279, 280 Page, Β , 120, 305, 328 Pammett, J , 403, 404n, 406, 409n, 410n, 411n, 417n, 424 Pappagallo, C , 206n Pappi, F U , 57, 106, 107 Paramaio, L , 395 Parkin, F , 382 Partanen, J , 74 parties bourgeois, 330, 338, 339, catchall, 8, 106, 160, 201, 203, 428, 464, center (moderates), 63, 101, 160, 163, 180, 182, 185, 188, 200, 201, 203, 229, 230, 330, 341, Christian democratic, 61, 63, 332, communist, 61, 63, 71, 74, 101, 180, 181, 182, 464, conservative, 97, 160, 331, 332, 340, 341, flash (short-lived), 167, 233, 428, labor, 71, 272, Left (lib­ eral), 63-64, 330, 331, 335, 336, liberal, 64, 272, major, 59, 163, 238, 300, 301, minor (or third), 59, 164, 167, 184, 186, 187, 192, 238, nationalist, 63, 235, 385, opposition, 197, 200, 220, progressive, 97, 334, Right (conservative), 63, 159, 330, social democratic, 70, 71, 73, 74, 77, 79, 89, 91, 162, 330, 331, 332, 334 Parti-Quebecois, Canada, 419 partisan change, 63, 138, 145, 243, 325 partisan conversion, 141, 350, 409, 410, 417 partisan mobilization, 138, 139, 141, 152, 339, 350 partisan replacement, 138, 140, 152, 252, 291,411,417

Index partisanship (party ID), 12n, 233, 240, 300, 305, 370, 400, 428, distribution of, 126, 210, 211, 214, 346, 406, 420, 431, 445, measurement of, 12, 126, 145, 155, 214, 244, 284, 285, 346, stability, 33, 245, 289, strength, 195, 212, 217, 218, 326, 327, 328, 348, 403 Payne, S G , 381 Pedersen, Μ Ν , 9, 235, 402 Pempel, Τ J , 162 Penniman, Η R , 212 Perez Dias, V , 392 period effects, 247, 250, 259, 261, 263, 264, 303 Pesonen, Ρ , 71, 72, 75, 76, 78n, 79, 81, 89 Peters, G , 71n Petersson, Ο , 10n, 90, 330, 348, 353 Petrocik, J R , 13, 40, 135, 153n, 216, 255n Pierce, J C , 40 Pierce, R , 430, 433, 438 Piereson, J , 40 Piore, Μ J , 388n Plaza, R , 383n Poggi, G , 210 political participation (conventional), 82, 170, 212, 346, 414 See also protest Pollack, Β , 367n Pomper, G , 11 Poole, Κ W , 248n Portocarero, L , 388n postindustnal society, 5, 6, 15, 21, 76, 105, 108 Press, I , 370 Progressive-Conservative party, Canada, 402, 404, 411, 414, 417, 418, 419, 420, 422, 423 Progress party, Scandinavia, 332, 340, 350, 354, 359 proportional representation, 28, 101 protest, 3, 29, 62, 83, 91, 132 Przeworski, A , 367n, 369, 394n Ping Bastard, Ρ , 368 Putnam, R , 208, 228 Radical Liberal party, Scandinavia, 332, 339, 341, 354 Radical Party, Italy, 226, 228 Radical Political party (PPR), Netherlands, 270, 272, 280, 282, 464 Rae, D , 274

511 Ranney, A , 222, 470 Rayside, D , 463n Reagan, R , 241,452 realignment, 11, 13, 14, 68, 95, 101, 102, 114, 119, 128, 132, 176, 183, 206, 217, 228, 399, 451, 474, 475, critical, 13, 95, 134, 137, 140, 141, 151, 265, 459, eco­ logical, 96, 103, 164, 201, 204, leader­ ship role in, 151, 361, parliamentary, 98, 99, 101, 130, 217, sectoral, 95, 96, 99, 100, 103, 164, 180, 194, 205, secular, 13, 128, 135, 139, 152, 265, 467 Reiter, Η , 158 religion See church Renten, S , 151n Republican party (PRI), Italy, 206, 207, 218, 226, 229, 230 Republican party (GOP), United States, 99, 100, 135, 137, 138, 145, 150, 155, 156, 157, 234, 241, 242, 243, 245, 254, 260, 261, 284, 306, 433 retrospective evaluations, voting, 235, 308, 309, 310, 311, 312, 313, 316, 318, 320, 327, 328 Reynolds, J Β , 29 Richardson, Β Μ , 163, 178η, 199 Right (conservatives), 170, 171, 208, New Right (conservatives), 4, 21, 100, Old Right, 113, 118 Riker, W , 137n, 305 Robertson, D , 13 Rochon, Τ , 452, 469 Rokkan, S , 7, 26, 135, 359, 424, 454, 456, 459, 467 Roldan, S , 392 Roman Catholic Party of The Netherlands (RKPN), 278 Roosevelt, F D , 41, 265 Rosa, J J , 445 Rose, D , 40 Rose, R , 7, 12, 13, 27, 54, 74, 304, 306, 307, 407, 438 Rusk, J G , 249 St Angelo, D , 241n Salisbury, R Η , 138 Sam, G , 208, 210, 212, 214, 217n, 219, 220, 222, 223, 225, 375n, 393 Sankiaho, R , 71, 72, 75, 76, 78n, 79, 81, 89, 118n, 131,456 Santin, W , 368

512 Sarlvik, Β , 10η, 14, 300, 301, 304, 307, 309, 310, 328n Sarton, G , 102, 222 Sauerberg, S , 330 Scalapino, R A , 162 Scharpf, F , 77 Schmitter, Ρ C , 394 Schwartz, Μ , 403n seculanzation (deconfessionalization), 97, 205, 270, 374, 391, 436, 474 Seip, D , 274 Serrano, S , 392 Shabad, G , 382n, 393n Shaber, S , 455 Shanks, J Μ , 212 Share, D , 390 Sharpe, L , 471 Shively, W Ρ , 12, 18, 108, 242n, 245, 303,461,475 Simon, Μ , 436, 438 Sjoberg, G , 76 Skocpol, Τ , 394 Smith, G , 455 Snidal, D , 369 Sniderman, Ρ Μ , 72, 212 social class, 6, 120, 130, 225, 300, 434, composition, 304, farmers (peasants), 96, 97, 354, 425, 435, middle class, 184, 226, 299, 353, new middle class, 95, 98, 102, 107, 111, 120, 130, 162, 184, 305, 453, working class, 184, 225, 299, 354, 425 social cleavage model, 454, 455, 456, 457, 473 Social Democratic Federation (SDF, also So­ cialist Citizen's League), Japan, 101, 160, 162, 163, 167, 200, 201, 203, 464 Social Democratic party (SDP), Bntam, 302, 458,461,465,466 Social Democratic party (PSDI), Italy, 209, 218, 225, 226, 229 Social Democratic party, Scandinavia, 332, 334, 335, 336, 337, 338, 339, 343, 345, 346, 348, 351, 353, 354, 357, 361 Social Democratic party (SPD), West Ger­ many, 98, 99, 104, 105, 113, 114, 115, 118, 120, 123, 127, 128, 130, 132, 133 Socialist Movement (MSI and MSI-DN), It­ aly, 210, 218 Socialist party (PS), France, 442, 444

Index Socialist party (PSI), Italy, 206, 209, 210, 218, 222, 225, 226, 229 Socialist party (SP), Japan, 101, 159, 167, 168, 180, 181, 182, 184, 185, 186, 188, 189, 193, 197, 201, 203 Socialist Party (PSOE), Spain, 366, 372, 374, 378, 385, 386, 387, 390, 392, 393, 395 Socialist People's party, Scandinavia, 332, 336, 354 Sohyo, 162, 201 Sonquist, J , 379n Spadohni, G , 206 stable alignment, 11, 13, 469 Stapel, J , 273 Statera, G , 214 Steed, Μ , 301 Sterner, Κ , 6, 170, 182n Stephens, J D , 16, 18, 29, 367n Stimson, J A , 99, 100, 120, 128, 145n, 149n, 151n, 237, 240, 265, 455, 459 Stokes, D Ε , 12, 13, 75, 137, 241n, 242, 257, 303 Stouthard, Ρ , 275 Suarez, A , 365, 366, 371, 372, 373, 378, 379, 380, 382 Sullivan, J L , 40 Sundquist, J L , 13, 134, 138, 459 support (trust, alienation), 79, 87, 89, 359, 365, 372, diffuse, 72, 79, 80, 90, specific, 72, 73, 79, 80, 81, 89 Suzuki, Υ , 203 Tart, R , 41 Tamames, R , 366 Tarrance, V L , 10η tax revolt, 73, 75, 90, 100 Taylor, C , 7, 9n Taylor, J , 367n technocratic control, 108, 392, 457, 459 Tedin, Κ L , 13, 138, 154, 261 terrorism, 205, 227 Tezanos, J F , 368 Thatcher, Μ , 299, 452 Thernborn, G , 71n Thijn, Ε van, 271, 272 Thomas, J , 106 Thomassen, J J A , 284 Toffler, A , 21, 476 Tompkins, G , 71n

513

Index Trilling, R., 13n Trudeau, P., 419, 423 Tsurutani, T., 193 Tufte, E., 6, 166 Turner, J., 305 Tusell Gomez, S., 368 Ullman, J. C , 377 Ullman, P. L., 377 Unified Socialist party (PSU), France, 442, 464 Union Democratique du Travail (UDT), France, 430 unions, 17, 26, 76, 90, 97, 118, 131, 193, 209, 225, 305, 306, 307 urbanization, 168, 171, 177, 178, 201, 225, 474 Urwin, 0 . , 7, 27, 54, 74, 438 Valen, Η., 12n, 330, 347, 354, 357, 358, 359, 360 Valentin, F., 71, 74 values, 63, 160, 178, 179, 474; authoritar­ ian/libertarian, 173, 174, 175, 192; material/postmaterial, 19-20, 21, 26, 27, 3436, 53, 68, 76, 77, 86, 89, 174, 254, 292, 456, 466; parochial/cosmopolitan, 174, 175; personalistic/universalistic, 174, 175, 192, 394; process of change, 173, 174, 175, 188, 199, 201, 457 Verba, S., 6, 171, 210, 216, 346, 371n Vietnam War, 3, 4, 42, 137, 254, 261, 265 volatility (partisan), 9, 10η, 11, 164, 201, 235, 291, 295, 301, 329, 330, 344, 356, 402, 422, 423, 451

Voss, W„ 105, 106 Vrije Universiteit, 292 Wallace, G., 466 Wanat, J., 244, 253 Ward, R., 170, 178n Watanuki, J., 178 Watergate, 137, 261, 265 Weisberg, H. F., 242 welfare state, 68, 70, 72, 73, 77, 78, 89, 90 White, J., 192 White, O., 76 Wiegel, H., 272 Wilensky, Η., 71n, 75n, 76 Wilson, B., 377 Wilson, F. L., 428, 432, 439 Winsborough, Η. Η., 248n Wolinetz, S., 9n women's movement, 114, 118, 132, 193, 474 World War II, 41, 124, 182, 249, 250, 332 Worre, T., 330 Wright, Ε. Ο., 388n Wright, G„ 17, 7ln

Yankelovitch, D., 4 Yoron Chosa Nenkan, 196

Zavonia, W., 313 Zingale, Ν. H., 13, 19, 128, 129, 134, 135, 138, 206, 208, 460 Zirakzadeh, C. E., 396 Zuckerman, Α., 473

Library of Congress Cataloging in Publication Data Main entry under title: Electoral change in advanced industrial democracies. Bibliography: p. Includes index. 1. Political parties—Addresses, essays, lectures. 2. Elections—Addresses, essays, lectures. 3. Comparative government—Addresses, essays, lectures, I. Dalton, Russell J. II. Flanagan, Scott C. III. Beck, Paul Allen. JF2011.E43 1985 324 84-42592 ISBN 0-691-07675-8 (alk. paper) ISBN 0-691-10165-6 (pbk.)