Growing up Democratic : Does It Make a Difference? [1 ed.] 9781626375567, 9781626375192

Citation preview

Growing Up Democratic

The Global Barometers Series

The Global Barometers Series publishes work that draws on the Global Barometers Surveys. For information on the surveys, see http://www.globalbarometers.org.

Yun-han Chu, Asian Barometer Surveys Larry Diamond, Stanford University E. Gyimah-Boadi, Afrobarometer Christian W. Haerpfer, Eurasia Barometer Marta Lagos, Latinobarómetro Andrew J. Nathan (Chair), Asian Barometer Surveys Sandeep Shastri, South Asia Barometer Khalil Shikaki, Arab Barometer EDITORIAL BOARD

Growing Up Democratic Does It Make a Difference? edited by

David Denemark Robert Mattes Richard G. Niemi

b o u l d e r l o n d o n

Published in the United States of America in 2016 by Lynne Rienner Publishers, Inc. 1800 30th Street, Boulder, Colorado 80301 www.rienner.com

and in the United Kingdom by Lynne Rienner Publishers, Inc. 3 Henrietta Street, Covent Garden, London WC2E 8LU

© 2016 by Lynne Rienner Publishers, Inc. All rights reserved

Library of Congress Cataloging-in-Publication Data Names: Denemark, David, editor. | Mattes, Robert B., editor. | Niemi, Richard G., editor. Title: Growing up democratic : does it make a difference? / edited by David Denemark, Robert Mattes, and Richard G. Niemi. Description: Boulder, Colorado : Lynne Rienner Publishers, Inc., [2016] | Includes bibliographical references and index. Identifiers: LCCN 2016000719 | ISBN 9781626375192 (hardcover : alk. paper) Subjects: LCSH: Democracy—Cross-cultural studies. | Political socialization—Cross-cultural studies. | Political participation—Cross-cultural studies. Classification: LCC JC423 .G778 2016 | DDC 321.8—dc23 LC record available at http://lccn.loc.gov/2016000719

British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library.

Printed and bound in the United States of America

The paper used in this publication meets the requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992.

5

4

3

2

1

Contents

List of Tables and Figures Acknowledgments

1 Growing Up Democratic? Robert Mattes, David Denemark, and Richard G. Niemi

Part 1 Postauthoritarian Societies

2 Southern Europe: Elite-Led Culture Change Richard Gunther and José Ramón Montero 3 Latin America: The Modest Dividend of Growing Up Democratic Alejandro Moreno and Marta Lagos

4 East Asia: Variable Support for Democracy in a Diverse Region John Fuh-sheng Hsieh and Jih-wen Lin

5 South Asia: An Arm’s-Length Embrace of Democracy Sandeep Shastri, Reetika Syal, Suhas Palshikar, and Shreya Sarawgi

6 Eastern and Central Europe: Growing Up Communist, Learning to Be Democratic William Mishler, Richard Rose, and Natalia Matukhno v

vii xi 1

25 63 83

105 125

vi 7

Contents

Sub-Saharan Africa: The Positive Impact of Effective Democracy Robert Mattes

151

Part 2 Comparisons with Established Democracies and Nondemocracies

8 Advanced Democracies: The Erosion of Traditional Democratic Citizenship David Denemark, Todd Donovan, and Richard G. Niemi

181

10 China: The Impact of Modernization and Liberalization on Democratic Attitudes Min-hua Huang, Yun-han Chu, and Cao Yongrong

233

9 The Arab World: The Challenges of Political Islam Eleanor Gao

Part 3 Conclusion

11 Generational Change in Postauthoritarian Democracies? Robert Mattes, David Denemark, and Richard G. Niemi About the Online Appendix Bibliography The Contributors Index About the Book

207

263 283 285 307 311 319

Tables and Figures

Tables

1.1

2.1 2.2 2.3 2.4 2.5

2.6

2.7 3.1

3.2

4.1

4.2

4.3 4.4 4.5 5.1 5.2 5.3

Example Showing How to Identify Generations by Age at the Time of a Significant Political Event Political Generations in Greece, Italy, Portugal, and Spain Support for Democracy Support for Democracy by Political Generation Multivariate Analyses of Support for Democracy Multivariate Analyses of Support for Democracy in Spain, 1979 Multivariate Analyses of Frequency of Political Discussion with Friends Correlates and Predictors of Political Engagement Models of Support for Democracy, Satisfaction with Democracy, and Support for Liberal Democracy Effects of Democratic Cohort on Democratic Attitudes Within Individual Latin American Countries Years and Events Defining Political Generations in East Asian Countries Attitudes Toward Democracy in Countries Grouped by Type of Regime Attitudes Toward Democracy by Generations Attitudes Toward Democracy by Regime Types Attitudes Toward Democracy in Individual Countries Generational Cohorts by Regime Type in South Asia Support for Democracy Across Generations in South Asia Attitudes Toward Democracy in South Asian Countries vii

14 34 35 36 46

47

56 58

75

78

88

89 91 95 98 110 115 117

viii

Tables and Figures

6.1

Long Generations in Post-Communist Eastern and Central Europe Multilevel Model of the Effects of Generation, Age, and Time on Individual Attitudes Toward Democracy and Authoritarianism Multilevel Lifetime Learning Models of Individual Support for Democracy and Authoritarianism Changes in Demand for and Satisfaction with Democracy Across 16 Countries, 2002–2008 African Political Regimes Generational Differences in Demand for Democracy and Satisfaction with Democracy Across 20 Countries, 2008 Models of Demand for Democracy Across 20 Countries Models of Satisfaction with Democracy Across 20 Countries Political Generations in Advanced Democracies Support for Democracy and for Authoritarian Alternatives by Generation Eras Defining Political Generations in Arab Countries Predicted Levels of Support for Democracy and Political Islam Across Historical Eras Support for Democratic and Autocratic Regimes Overall and Across Generations Support for Political Islam Overall and Across Generations Impact of Generations on Support for Autocracy, Democracy, and Political Islam Predictions of Democratic Attitudes for Four Political Generations Democratic Attitudes by Generations in China: ANOVA Overall Samples, Pooled-Data Analysis Analyses of Generational Subsamples Democratic Values in Asia

6.2 6.3

7.1

7.2 7.3

7.4 7.5 8.1 8.2

9.1 9.2

9.3

9.4 9.5

10.1

10.2 10.3 10.4 10.5

Figures

2.1 2.2

2.3

2.4 3.1

Democracy Is Best, by Party Satisfaction with the Performance of Democracy and the Economy Support for Democracy, Authoritarianism, and Satisfaction with Democracy Internal Inefficacy Across Southern Europe, by Cohorts Support for and Satisfaction with Democracy in Latin America, 1995–2010

128 137

140

159 162

163 166 167 186

193 213

221

222 223

226

240 242 246 249 253 41

49

50 53

70

Tables and Figures

7.1

8.1a 8.1b 8.1c 8.2a 8.2b 8.2c 8.3 8.4a 8.4b 8.5a 8.5b 8.5c 8.6

Demand for and Satisfaction with Democracy Across 16 Countries, 2002–2008 Important to Live in a Democracy Having a Democratic System Is Good Democracy Is the Best Form of Government Having a Strong Leader Without Elections Is Good Having the Army Rule Is Good Having Experts Rule Is a Good Thing Satisfaction with Democracy Interest in Politics Voted in Last Election Signed a Petition Attended Peaceful Demonstrations Joined in Boycotts Use Internet/E-mail to Learn What’s Going On

ix 158 195 195 196 196 197 197 198 199 200 201 201 202 203

Acknowledgments

The research at the heart of this book saw its genesis in international collaborative efforts that were fortunate to receive financial and logistical support from several host universities and from the World Universities Network. A University of Western Australia Research Collaboration Award, cofunded by the University of Sydney, enabled a research workshop in Perth, “Democratic Participation in a Globalized World,” which brought together the three coeditors of this book for the first time. One outcome of that workshop was the publication of a special issue of the Australian Journal of Political Science, coedited by Denemark and Niemi and including Mattes’s research on attitudes toward democracy among South Africa’s first postapartheid generation, the “Born Frees.” It was the global implications of Mattes’s focus on generational patterns of attitudes toward democracy and authoritarianism in an emerging democracy that became the formative impetus for this book. It led us to build an international team of researchers, using regional data primarily from the Global Barometer Survey, to explore these dynamics around the world. The result, after three years of research, workshops, meetings, and presentations around the globe, is this volume, which addresses the question of whether growing up democratic makes a difference in citizens’ views of democracy in principle and in practice. We are grateful for the financial and practical support of the University of Western Australia, the University of Cape Town, and the University of Rochester, which has helped to sustain this project. We also thank the World Universities Network (WUN)—a global network of 19 research xi

xii

Acknowledgments

universities to which each of our universities belongs. WUN helped to underwrite this project’s collaboration in a follow-up workshop and supported several collaborative research meetings, which are vital to forging international teams of researchers. We acknowledge especially the Global Barometer Survey and its production of the excellent multinational databases utilized throughout this volume. The relatively new ability to use cross-national data—including the Global Barometer Survey but also the World Values Survey, the Comparative National Elections Project, the Comparative Study of Electoral Systems, and the International Social Survey Program—to analyze social, political, and economic questions of cross-national and global consequence is a profoundly important capability that should be undertaken more regularly. The rich variability and comparisons of this volume, we believe, confirm the value of cross-national analysis. Finally, we thank the authors of the individual chapters for their diligence in pursuing a largely common research design and for their patience as we formulated and revised our analytic framework and determined how it should be applied to countries with diverse histories, contemporary social and economic fortunes, and structures of governance. And we thank Thomas Isbell for constructing the index. —the Editors

1 Growing Up Democratic? Robert Mattes, David Denemark, and Richard G. Niemi

Beginning with the Portuguese coup of 1974, and escalating in frequency after the fall of the Berlin Wall in 1989, over 80 countries across the globe have successfully extricated themselves from various forms of autocratic rule and put in place either what Freedom House calls “electoral” or “liberal” democracy (Puddington 2013). Several dozen other countries have moved away from classic autocratic rule and implemented regular multiparty elections, though they have so far either failed to generate a sufficiently level electoral playing field or continued to impose such severe limits on democratic rights and liberties that they do not qualify as democracies (Schedler 2006; Levitsky and Way 2010). All in all, however, the world has become a fundamentally more democratic place over the past four decades. At least since the publication of Gabriel Almond and Sidney Verba’s (1963) classic, The Civic Culture, political scientists have seen public opinion and political culture as crucial forces in the consolidation and deepening of democracy. In particular, “congruence theory” would see the “third and fourth waves of democracy” that swept across the globe in the 30 years between 1975 and 20051 as the consequence of a disjuncture between the operating norms of the regimes and their constituent institutions, and those of the mass public (Eckstein 1961; Almond and Verba 1963). Thus, the key question motivating public opinion researchers of new democracies has been whether the value structures that questioned and delegitimated the former authoritarian and totalitarian regimes are sufficient to legitimate and consolidate new democracies. This sort of approach is evident in a number of high-profile studies that sought to understand this new democratic impetus. Beginning with 1

2

Robert Mattes, David Denemark, and Richard G. Niemi

analyses of third-wave democratization studies in southern Europe (Morlino and Montero 1995; Montero, Gunther, and Torcal 1997) and then Richard Rose and his colleagues’ studies of the recently democratized states of central and eastern Europe (Rose, Mishler, and Haerpfer 1998; Rose, Mishler, and Munro 2006), cross-national studies of public opinion have investigated the factors that lead people living in new, potentially fragile democracies, including Latin America (Lagos 2001; Moreno and Méndez 2002), Asia (Chu, Diamond, Nathan, and Shin 2008; deSouza, Palshikar, and Yadav 2008; Shin 2012a), and sub-Saharan Africa (Bratton, Mattes, and Gyimah-Boadi 2005; Bratton 2013), to prefer the new regime over the old or to accept and even support forms of authoritarian rule. Respondents in the surveys analyzed in these various studies, it should be noted, were almost all people who had lived some or even all of their adult lives under Marxist-Leninist one-party states, military dictatorships, or other forms of autocratic rule. Sensibly, therefore, researchers focused largely on discerning the balance of evaluations of economic and political performance, formal education and cognitive sophistication, and prodemocratic values that would lead them to see democracy as the preferable regime. Now, however, we are two decades or more beyond many of these countries’ transitions away from authoritarian rule. Memories of the old regime have dimmed, and postauthoritarian societies are populated with increasingly large proportions of young citizens who know only the new regime. And yet, as we will see, citizens’ engagement with democracy remains tentative in many cases, with support for democracy often illusory and less based on principle than a day-to-day confidence in the ability of democratic government to “deliver” on key priorities, such as economic and physical security. At the same time, the rejection of authoritarian rule is often halfhearted, with many citizens still tempted by strong leaders, experts, and the military and their perceived ability to “get things done.” The chapters in this volume present evidence about citizen attitudes toward democracy drawn from regional surveys of public opinion across the globe conducted between 2004 and 2009. They demonstrate that, far from being strong and ubiquitous (Inglehart and Welzel 2005; Norris 2011), popular support for democracy is equivocal and varies widely across and within regions. However, the data also demonstrate that the trajectories of public attitudes are moving in the “right” direction in all but one of the democratizing regions for which we have longitudinal data. Public support for democracy rose very rapidly in southern Europe in the 1980s, but support for democracy and rejection of authoritarian rule have also increased—though more modestly—in central and eastern Europe, Latin America, and sub-Saharan Africa. In East Asia, however,

Growing Up Democratic?

3

support for democracy declined in four of the six countries for which we have two rounds of data. And, in contrast to concerns by some analysts that unrealistically high popular expectations of what democracy is able to deliver will lead to steadily declining levels of satisfaction with its performance (O’Donnell, Schmitter, and Whitehead 1986; Przeworski et al. 1995), the public opinion time series reviewed in this volume reveal an over-time increase in Latin America in satisfaction with democracy (albeit at relatively low levels), trendless variation in sub-Saharan Africa and central and eastern Europe, and a slight decline in East Asia. Given the fact that a quarter of a century has passed since the fall of the Berlin Wall (the beginning of the “fourth wave of democracy”) and four decades have elapsed since Portugal’s “Carnation Revolution” (the start of the third wave), an apparently obvious explanation for the upward trends in public commitment to democracy is that mass publics are composed of increasingly large proportions of citizens who have “grown up democratic.” As older citizens who grew up under autocratic rule age and die, they have been replaced by younger people who have grown up after the democratic transition and have experienced freedom as well as democratic procedures and institutions as “normal politics”—that is, as part and parcel of the package of facts and repertoire of skills they have acquired about governance. At minimum, it is clear these new democrats have not had to “unlearn” political attitudes acquired under the old authoritarian regime.2 To the extent that growing up democratic does, in fact, produce these positive outcomes, democracy could be said, at least in part, to generate its own popular support. Thus, if the process of generational replacement is “normalizing democracy,” we would expect sharply distinctive attitudes toward democracy between the generations that have lived under authoritarian rule and those that have been raised within an emerging democracy. And yet, a number of interesting puzzles and questions pervade these dynamics. Far from being ubiquitous, as we suggest above, popular support for democracy in postauthoritarian societies is highly variable—a pattern that will be shown in the chapters that follow. The extent to which these societies are able to build a broader consensus supportive of democracy, especially as fewer and fewer people have explicit memories of the old regime, is one of the themes explored in this volume. Collectively, the authors ask whether there is evidence that the world’s new democracies are beginning to produce younger generations of citizens who view democracy differently from their parents or grandparents and who support democracy not simply because their country is more prosperous but because they have grown up democratic. Alternatively, we ask, are fledgling democrats the product of the younger generations’ exposure to increasingly universal public education, rising levels of affluence, free

4

Robert Mattes, David Denemark, and Richard G. Niemi

access to news and information, and other factors conducive to active involvement in politics and support for open and democratic governance? And what about those living under authoritarian regimes or in political systems that have only tentatively crossed the threshold into democracy? Are democratic institutions and principles seen as important for them? How strong is the impetus to build democracy if it must emerge from the ranks of those who have no direct experience with its procedures, institutions, and values? All told, explaining citizens’ support for democratic governance in emerging democracies and the role that generations, political and economic conditions, and education play in those sentiments remains a multifaceted analytic puzzle that has prompted the development of a number of perspectives designed to provide answers to the sorts of questions we ask above. The central impetus for this book is our desire to address this puzzle. We begin our analytic response by framing several alternative explanatory approaches, including the historically significant model of political socialization and generational learning, which we propose to test in the context of dozens of emerging democracies in the various regions of the world.

Adolescent Socialization and Generational Change

The argument that those who grew up under democratic governance see democracy differently than those who grew to adulthood under authoritarian rule has a distinguished precedent. Perhaps most famously, in his classic political anthropology of the early nineteenth-century United States, Alexis de Tocqueville (1873 [1835, 1840]) concluded that what made the Americans he observed so unique was that they had all grown up under conditions of freedom and equality and had not had to overthrow a despotic monarchy the way his fellow Frenchmen had. It set a theoretical stage on which a number of analyses have played important roles. Almost certainly, it is Karl Mannheim (1952 [1928]) who has provided the most influential reasoning about generations, beginning with his work in the 1920s (which became available in English only in the 1950s). Mannheim argued that societies were characterized not only by the class distinctions emphasized by the Marxist scholars of his day but also by important generational distinctions. Memberships in classes or generations endow the individuals sharing in them with a common location in the social and historical process, and thereby limit them to a specific range or

Growing Up Democratic?

5

potential experience, predisposing them for a certain characteristic mode of thought and experience, and a characteristic type of historically relevant action. Any given location, then, excludes a large number of possible modes of thought, experience, feeling, and action, and restricts the range of self-expression open to the individual to certain circumscribed possibilities. This negative delimitation, however, does not exhaust the matter. Inherent in a positive sense in every location is a tendency pointing towards certain definite modes of behaviour, feeling, and thought. (1952: 291) The fact that people are born at the same time, or that their youth, adulthood, and old age coincide, does not in itself involve similarity of location; what does create a similar location is that they are in a position to experience the same events and data etc., and especially that these experiences impinge upon a similarly “stratified” consciousness. . . . Only where contemporaries definitely are in a position to participate as an integrated group in certain common experiences can we rightly speak of community of location of a generation. (1952: 297–298)

Generations are defined by the events of early lifetime, Mannheim reasoned, because early events leave far more indelible impressions on people than later ones. Taking on what would later become known as the “online processing” and “lifetime learning” models, he argued: “Early impressions tend to coalesce into a natural view of the world. All later experiences then tend to receive their meaning from this original set, whether they appear as that set’s verification and fulfilment or as its negation and antithesis” (1952: 298). Mannheim’s arguments have been supported by a long line of empirical research in political science. In this paradigm of adolescent political socialization, people develop their fundamental beliefs during their “impressionable years” through learning from parents and siblings and by internalizing prevailing norms from friends, social organizations, and mass media (Campbell, Converse, Miller, and Stokes 1960; Jennings and Niemi 1974). They then tend to retain these attitudes as they age. Attitudes do change, especially in response to events, but basic values are established early in life and set the tone for understanding, interpreting, and evaluating situations, issues, and problems that individuals confront later in life (Kelley and De Graaf 1997; Myers 1996). Though, for a time, socialization researchers rejected the notion of long-term influence, a considerable amount of recent work has established the durability of early learning. This includes work such as Campbell’s (2006) study of early influences on voting habits; Prior’s (2010) work establishing the stability of people’s levels of political interest; Green, Palmquist, and Schickler’s (2002) work on the stability of partisanship in the United States; and a

6

Robert Mattes, David Denemark, and Richard G. Niemi

range of studies of attitudes at varying points in people’s lifetimes (e.g., Kroh 2014; Jennings, Stoker, and Bowers 2009; Zuckerman, Dasovi , and Fitzgerald 2007; Sears and Funk 1999; Dash and Niemi 1992; Niemi and Jennings 1991; Alwin, Cohen, and Newcomb 1991). With regard to new democracies, various scholars of post–World War II Europe and Japan found rapid and significant increases in prodemocratic values and attitudes in West Germany (Baker, Dalton, and Hildebrandt 1981), Austria (Muller 1984), Italy (Sani 1980), and Japan (Richardson 1974; Flanagan and Richardson 1984; Ikeda and Kohno 2008) that took hold especially among the young. While explanations for these transformations tended to focus on considered efforts in the schools, they also pointed to the important effect of generalized exposure to democracy. Dalton’s description of the Federal Republic of Germany is illustrative: Confronted by an uncertain public commitment to democracy, the government undertook a massive programme to re-educate the public. The schools, the media and political organizations were mobilized behind the effort. And the citizenry itself was changing—older generations raised under authoritarian regimes were being replaced by younger generations socialized during the postwar democratic era. These efforts created a political culture congruent with the new institutions and processes of the Federal Republic. The West German public also learned democratic norms by continued exposure to the new political system. As a result, a popular consensus slowly developed in support of the democratic political system. (1994a: 472, emphasis added)

The evidence of this kind of change in third- and fourth-wave democracies, however, is far more mixed. Richard Gunther and his colleagues found substantial increases in Spaniards’ support for democracy, particularly among the young, in the years after that country’s successful transition (Montero, Gunther, and Torcal 1997; Gunther, Sani, and Shabad 1986). And William Mishler and Richard Rose (2007) found small but significant and consistent generational differences across a 14-year time span in Russians’ attitudes toward their old and new regimes. However, a number of other studies have shown very little evidence of important generational differences in support for democracy (Chu, Diamond, Nathan, and Shin 2008; Bratton et al. 2005; Markowski 2005; Shin 1999).3 Indeed, in a recent global analysis, Norris (2011: 174) has found that democratic aspirations (measured as the extent to which people think it is important to live in a democracy, on a scale of 1 to 10) actually increase with age. At the same time, what she calls the “democratic deficit” (the perceived extent of democracy subtracted from democratic aspirations, both measured on a 1 to 10 scale) is highest among the young.

Growing Up Democratic?

7

Alternative Models of Support for Democracy

The primary task of the studies found in this book is to put the theory of adolescent learning and generational change to the test. Does growing up democratic matter? If this idea is correct, we should find that younger generations exhibit higher levels of support for democracy. Having learned about politics and government by experiencing it directly, they should better understand what democracy means and thus also reject nondemocratic alternatives more frequently than their older compatriots. However, the “direct learning” hypothesis of political socialization does not provide the only account of why mass publics may be becoming more supportive of democracy. An equally important alternative account is one of cognitive mobilization through the education and communication revolutions sweeping through the developing world, albeit at a different pace in different societies. As early as the pioneering work of Almond and Verba (1963) and other studies of that era, it was widely recognized that there were major attitudinal and behavioral differences across educational strata. Since then, education has repeatedly been shown to be an important predictor of voter turnout in the United States (Milligan, Moretti, and Oreopoulos 2004; Dee 2004; Verba, Schlozman, and Brady 1995; Wolfinger and Rosenstone 1980; Verba and Nie 1972) as well as in many other Western countries (Dalton 2013a; Norris 2002; Lipset 1960) though there are exceptions, often in countries with strong socialist parties that mobilize less educated working class voters (Milligan et al. 2004; Norris 2002; Powell 1986). Education has also been confirmed as a strong predictor of other citizen qualities such as interest in politics, newspaper readership, political knowledge, interpersonal trust, tolerance of political opponents, and a wide range of forms of political participation such as contacting elected leaders, joining community associations, attending community meetings, political Internet activism, and protest both in the United States (Dalton 2013a; Dee 2004; Burns, Schlozman, and Verba 2001; Brady et al. 1996; Nie, Junn, and Stehlik-Barry 1996; Verba et al. 1995; Putnam 1995; Bobo and Licari 1989; Hyman and Wright 1979; Hyman, Wright, and Reed 1978) and other democracies in Europe (Milligan et al. 2004; Dalton 2013a, b; Milner 2002). The impact of education is so regularly found that political scientist Philip Converse (1972) once called it the “universal solvent” of political participation. In a 1996 summary, Nie et al. (1996: 2) concluded: The notion that formal educational attainment is the primary mechanism behind citizenship characteristics is basically uncontested. . . . Formal education is almost without exception the strongest factor in explaining what citizens do in politics and how they think about politics.

8

Robert Mattes, David Denemark, and Richard G. Niemi

While more limited, the existing literature in developing societies also bears out the importance of education. Education has been shown to increase voter turnout and civic engagement in several different developing world contexts (Magalhães 2007; Lam and Kuan 2008; Anderson and Dodd 2006; Bellucci, Maraffi, and Segatti 2007; Verba, Nie, and Kim 1978). More importantly for present purposes, education has proved to be a very strong predictor of popular support for democracy in places like Korea, Chile, eastern Europe, Russia, and sub-Saharan Africa (Mattes and Bratton 2007; Rose et al. 2006; Bratton et al. 2005; Markowski 2005; Shin 1999; Rose et al. 1998). Overall, then, this literature points to education not only as having a significant, direct impact on support for democracy but, perhaps even more importantly, as playing an indirect, supportive role by imparting critical sensibilities, skills, and the perceived need for citizens to engage in the world of politics around them. If so, then the increasing proportions of democrats in postauthoritarian societies may simply reflect increasing levels of younger generations’ exposure to increasingly universal public education, or free access to news and information and other factors conducive to active involvement in politics and support for open and democratic governance. Thus, we need to distinguish between the generational impact of political socialization and of education and political information in the survey analyses carried out in each of the regions reported in this volume. As will be shown, this is done by utilizing multivariate tests of support for democracy and its authoritarian alternatives that control for respondents’ placement in various political socialization generations and their level of education, political interest, and access to political news and information, thus enabling us to distinguish between these two distinct if often covarying forms of generational effects. Beyond the general impact of education, a third body of research has focused on the special case of democracy and civic education. In many if not most established democracies, the general school curriculum (particularly in social studies) is designed to teach a broader set of values such as individualism, tolerance, and mutual respect. But democracies often go further and require students to take specific classes in civic education that inculcate students with the workings of democracy and government, as well as the rights and responsibilities of democratic citizenship (Norris 2011; Milner 2002; Torney-Purta et al. 2001; Torney-Purta, Schwille, and Amadeo 1999; Callan 1997; Chesney and Feinstein 1997). Whether through direct instruction about democratic norms or more indirectly through imparting political knowledge (Fesnic forthcoming), or through the effects of open classrooms (Torney-Purta 2002) and active learning strategies (Dassonneville et al. 2012), civic education has been shown to foster greater understanding of and appreciation for democratic government.

Growing Up Democratic?

9

While evidence of large-scale generational shifts has thus far been scant, other scholars have traced important effects among students exposed to small-scale experiments in school civic education programs in Poland and South Africa (Slomczynski and Shabad 1998; Finkel and Ernst 2005) as well as among adults participating in donor-supported adult civic education programs in the Dominican Republic, Poland, Senegal, and South Africa (Kuenzi 2005; Finkel 2002, 2003). At the same time, there is no evidence among the first generation of South Africans produced by South Africa’s postapartheid school curriculum of increases in democratic support (Mattes 2012). A fourth research perspective on the factors that promote or stymie change in political cultures focuses on citizens’ expectations for fundamental improvements in the new order and the ability of a new, democratic regime to divorce itself from past failings and to “deliver,” especially, economic and physical security that meets those expectations. In short, this perspective revolves around perceived governmental performance. Younger citizens in fledgling democracies may well encounter a number of performance-based realities that can be expected to erode or perhaps undermine their support for democracy. As the transition from autocracy to democracy is unlikely to be decisive in supplanting the old order’s social, economic, and political ways, new generations of citizens may find themselves living in a society with significant echoes of the autocratic past. Corruption and crime may continue to plague the new democracy, while attempts to effect economic modernization may continue to suffer from cronyism and inefficiencies long assured to be set aside under the new order (Diamond 2008). In short, the expectations placed on the new democratic system may regularly outpace the realities experienced by many citizens. If so, despite the attractions of the emerging democracy, citizens—and younger ones in particular—may become dissatisfied with the new democratic institutions (Norris 2011) or even come to view authoritarian alternatives of rule by a single party or the military as desirable. Thus, support for democracy in postauthoritarian societies may be constrained by citizens’ views about the new system’s ability to deliver, especially economic and physical security. However, the impact of democracy’s ability to deliver the goods may depend, at least to some extent, on the quality and age of that democracy. For instance, the combined effects of the regular holding of free and fair elections may, with the passage of time, lead people to lower their material expectations of democracy and develop a greater appreciation of the value of its procedures. Yet not all postauthoritarian countries have become full democracies, and people often experience imperfect, partial democracy. In many instances, while governments convene regular and even mostly free elections, political rights and civil liberties are routinely

10

Robert Mattes, David Denemark, and Richard G. Niemi

limited. In still others, regular elections even with multiparty systems coexist with such severe manipulations of electoral processes and news media that neither elections nor broader political competition and debate are free or fair at all (Levitsky and Way 2010; Schedler 2006). Young South Koreans, one would suppose, surely draw different conclusions about democracy than young Albanians or young Thais. Thus, even if new generations in such countries express more democratic attitudes than their predecessors in the same country, they may be less democratic than citizens in freer and more democratic countries. Likewise, citizens (both old and young) in new democracies may express less democratic views than those in long-time democracies. Finally, constituting a fifth perspective for our analysis, is Ronald Inglehart’s theory of postmaterialism, which accepts the “socialization” hypothesis, but rather than focusing on key historical events such as the fall of the Berlin Wall or Nelson Mandela’s release from jail, combines the socialization hypothesis with a “scarcity” hypothesis that people value that which is in least supply. Based on Abraham Maslow’s (1943) hierarchy of needs, the hypothesis predicts that people who come of age under conditions of relative abundance will value “higher order” needs such as democracy, self-expression, gender rights, and environmentalism, while people who grow up in destitution will not. Thus, while Inglehart would predict we will indeed uncover sharp generational differences, he would argue that these differences will be most evident in postauthoritarian societies that have undergone rapid economic and social modernization and where younger people have come of age among conditions of greater material welfare and physiological security than their parents or grandparents (Inglehart 1990; Abramson and Inglehart 1995; Inglehart and Welzel 2005). In short, this book explores a variety of factors promoting and suppressing support for democracy in countries experiencing democratic institutions and governance for the first time. Central among these is the role of adolescent socialization and generational change. However, as we emphasize above, generationally distinctive patterns of political norms and values may reflect, in turn, a number of different realities for citizens in these countries, including adolescent socialization, rising educational levels, and the realization of material security amid rising affluence and employment. The regional analyses presented in this book are designed as considered tests of these various approaches and utilize broadly comparable explanatory models and measures (detailed below) in order to afford cross-regional variations in the results yielded by these tests.

Growing Up Democratic?

11

The Definition of Generations

There seems to be little doubt that meaningful social or political generations exist—i.e., that there are groups of individuals of similar ages who have experienced a noteworthy historical event at the same time and who think or behave in a manner that is distinct from older (and perhaps younger) individuals.4 Researchers readily acknowledge, of course, that not all age differences should be called generational. The classic (political) case is turnout at elections; for many decades now, young people around the world have turned out at lower rates than older people, indicating that this phenomenon has to do with their youth, not the specific period (or country) in which they are observed. But often it seems clear that differences are not caused by age; rather, US citizens who grew up and served in World War II have different views about patriotism and sacrifice from those who matured during the Vietnam War era (as expressed, say, in a survey in 2000), and those differences are much more likely to be considered a product of their differing experiences of war than of their respective ages. For all its use, however, the term has never surrendered easily to precise definition. There are two crucial questions: What age groups form a generation? And what causes them to be distinctive? Consider first the question of who (which age group) is most likely to form a generation. At least since the 1950s, with the English translation of Mannheim’s (1952 [1928]) essay on “the problem of generations,” there has been widespread agreement that adolescence and young adulthood are “formative years” during which individuals form worldviews that in one way or another affect them for the rest of their lives. Mannheim’s work, as we noted above, perhaps most formatively laid out a number of vital assumptions about late adolescence and early adulthood as the key period for the formation of individuals’ social and political views. It is a perspective that has been broadly accepted by most analysts of political socialization and citizenship education since then (Jennings and Niemi 1974, 1981; Torney-Purta et al. 2001; Sherrod, Torney-Purta, and Flanagan 2010; Campbell, Levinson, and Hess 2012). Yet if we are to use this understanding in empirical analyses, we need a precise operationalization. Mannheim himself was aware of the difficulties of precise delineation of the relevant period. He noted that “the possibility of really questioning and reflecting on things only emerges at the point where personal experimentation with life begins—round about the age of 17, sometimes a little earlier and sometimes a little later” (1952: 300).5 As for the upper end, he said in a lengthy footnote that “it is difficult

12

Robert Mattes, David Denemark, and Richard G. Niemi

to decide just at what point this process is complete in an individual,” but he pointed to age 25, using as an analogy when “the spoken language and dialect does not change” (1952: 299–300). Subsequent researchers have largely echoed the ages identified by Mannheim, sometimes “rounding up” the lower bound to 18, we suspect, because this is commonly the legal age of adulthood and, coincidentally, the age of the youngest respondents in many surveys. Empirical work, however, as well as theoretical considerations, suggests that an earlier rather than later age may be a good starting point for the formative years. We know, for example, that young children learn about racial differences at a very early age and sometimes use that knowledge to guide their thoughts and actions (e.g., Sears and Levy 2003; Van Ausdale and Feagin 2001). Young children have a sense of nationality (Jahoda 1963a, 1963b), and they have a (sometimes highly implausible or mistaken) sense of political authority (e.g., Greenstein 1965; Carter and Teten 2002). Children develop loyalties relating to political parties as preadults (Campbell et al. 1960; Butler and Stokes 1969), and at least some recent evidence suggests very early formation of these impressions (Bartels and Jackman 2014). When asked to recall important national or world events over the past 50 years, Schuman and Scott’s (1989) subjects often cited events from when they were 15 and 16 or even younger. As to when youths develop adult modes of thinking, Adelson and O’Neil’s (1966) insightful study showed that youths develop a more abstract and community-centered as opposed to concrete and self-centered perspective at ages 13–15, or 2–4 years earlier than Mannheim’s date. Overall, these various strands of research suggest that for most youths meaningful political socialization is well under way by the midteens. Fourteen, rather than 17 or 18, seems to be the age at which the preponderance of young people begin to form their political selves. Some relevant ideas and knowledge develop before then, and of course there is individual variation in the speed of maturation. Nevertheless, when analysis requires a precise year, we think of 14 as the beginning of the formative years for political socialization. Defining the end point for the socialization process is a more difficult enterprise, not the least because to some degree change is a lifelong process.6 Again, then, the question is: What age shall we use when precision is required? When does the process of forming one’s political self become stable enough that one can think of an end to major change? Mannheim, we observed, cited 25 as a likely candidate, and others have followed suit (e.g., Grasso 2014). There is, however, little theoretical basis and scant empirical evidence for such an age; while perhaps “convenient” in that it is exactly in the mid-20s, there is nothing that is compelling

Growing Up Democratic?

13

about it. We, in fact, think that just as the age at which the socialization process begins should be lower than conventionally assumed, so too should the age at which it ends. In this volume, we utilize 22 as the cutoff age for the measurement of political socialization’s formative years, or the age at which the “attitudinal cement” begins to harden. By 22, most individuals have completed their education or at least have finished their initial years at university. Thus, they will have encountered a diversity of viewpoints that come with moving out of one’s family and the surroundings in which one grew up from childhood, whether that is in the workplace or in institutions of higher education. And, in terms of practical experience with the political process, by 22, citizens in democratic polities will have had time during late adolescence or adulthood to observe at least one, often two, elections, thus having the opportunity, at least, to contemplate political ideas and candidacies and to form and articulate preferences. Having specified beginning and ending points for the formative period, how do we distinguish one generation from another? For this, we need to confront the second crucial question: what single events or periods are politically relevant and sufficiently salient that they are likely to make a difference to individuals living through them? Here, of course, there is further room for ambiguity. Matters that are thought to distinguish one generation from another are not self-recognized and labeled, and they are rarely if ever precisely identified as to beginning and end. Consider, however, a relatively clear-cut case, the fall of communism. We might identify this as having happened (in the Soviet bloc) in 1989, the year the Berlin Wall fell. Using this date and the numbers above, we can say that those who were 18 prior to 1989 were socialized during the communist period, as they would have spent more than half of their formative years (14–18 versus 19–22) under that regime. And, of course, anyone who turned 18 in 1989 or later was a part of the postcommunist generation.7 If the crucial matter covers more than a single year, we can apply the same reasoning once we have identified beginning and ending years (see Table 1.1).8 The critical point, in this view, is the match between the years of the event and people’s formative years—most importantly, identification of those individuals who spent a majority of their formative years before, during, or after the event. Often, however, as a matter of convenience we “translate” how old a person was during the years of the event to the years in which they were born (see the right-hand column of Table 1.1). This is convenient mainly because birth years are invariant across survey years; generation XXXX ages over time, but their birth years never change. Using birth year also allows a shorthand expression (especially, a

Table 1.1 Example Showing How to Identify Generations by Age at the Time of a Significant Political Event Years of the XXXX political event Belongs to which generation?a

Age

Pre-event 1985

1986

1987

1988

1989

1990

1991

1992

Post-event 1993

Birth years

14

Turned 18 before the event—i.e., in 1985 or before

Pre-XXXX generation

1967 and before

Turned 18 during the event—i.e., in 1986–1992

Generation XXXX

1968–1974

Turned 18 after the event—i.e., in 1993 or after

Post-XXXX generation

1975 and after

a

Determined by whether an individual spent a majority of his or her formative years (14-22) before, during, or after the significant event.

Growing Up Democratic?

15

“birth cohort”—people born in years tttt through year uuuu).9 In addition, a generation is occasionally labeled by when its members were born, as with the Baby Boomer generation in the United States.10 Each of the chapters that follow employs generation as a central measurement concept, defining generations by the use of key sociopolitical periods in the nations covered. Each makes use of the idea of formative years in which individuals are most strongly influenced by political events. However, the precise operationalizations often vary in their details due to the author’s sense of exactly who might have been most affected by the event(s) in question, or by boundaries that have been previously established in academic, journalistic, or popular accounts. This variability, we feel, does not weaken the basic analytic measures nor undermine the idea of generational change as a vital and innovative approach to the analysis of support for democracy. It does suggest that more efforts should be made to define generations precisely and to explore the impact of various operationalizations. One final issue to consider is the fact that many transitions to democracy have been gradual and drawn out, and they have not always moved linearly in a progressive direction but often involved backsliding. And as we have indicated above, in many cases elements of authoritarian practice have endured and now coexist with multiparty democracy. Thus, the lack of a sharp, clean break between autocracy and democracy poses special challenges to identifying political generations. Yet in virtually every case, it is possible for our authors to identify a decisive moment when multiparty elections became entrenched, even if democracy did not, as well as other key thresholds between types of authoritarian regimes or qualitative breaks between lower- and higher-quality democracy.

Plan of the Book

In order to test our hypotheses about generational and other influences on views of democracy, the editors invited a team of experts familiar with the social, economic, and political dynamics of nations in a given region and asked them to develop analyses of the nations within their region. These studies were to be conducted in central and eastern Europe and Eurasia, Latin America, East Asia, and sub-Saharan Africa. However, even if we were to find consistent results in these regions, we would not be sure that they are the consequences of the democratization process unless we are able to demonstrate that they are not visible, or take different forms, in countries and regions that are either established democracies or have not yet democratized. Thus we also initiated studies of established

16

Robert Mattes, David Denemark, and Richard G. Niemi

democracies in North America and Western Europe and of regions or countries that have not yet democratized, including in the Middle East/North Africa region, parts of South Asia and Latin America, and China. In a sense, these chapters act as regional “controls” that allow us to draw sharper inferences and interpretations about our findings in the postauthoritarian regions. But which attitudes to democracy are the relevant ones? We see popular commitment to democracy as a necessary though clearly insufficient condition for a stable democracy. How people feel toward Democracy in the abstract—the D-word as we have come to call it— is a starting point. Thus, we began by asking authors of the regional studies to consider first whether there was public support for democracy in general. Where possible, we have asked authors to use the Global Barometer Surveys (see below) question that forces respondents to make choices, rather than the World Values Survey (WVS) Likert-style items that simply ask people how much they like various forms of government. The WVS items do not force respondents to make choices between competing regime models. For instance, in the 2005–2007 World Values Survey, 90 percent of South Africans told World Values Survey interviewers that democracy was a “very good” or “fairly good” “way of governing this country” (Norris 2011: 93). Yet just 67 percent told Afrobarometer interviewers in 2008 that “democracy is preferable to any other kind of government” in response to a question that also gave them the option of saying that “in some circumstances, a non-democratic government can be preferable,” or that “for someone like me it doesn’t matter what kind of government we have”—a full 23 percentage points lower.11 Indeed, across 57 postauthoritarian societies in which Global Barometer Surveys asked this question between 2004 and 2007, a quite modest average of just 55 percent said that “democracy is always preferable,” with a staggering variation of 55 percentage points, ranging from a high of 79 percent (in Uruguay) to a low of 24 percent (in Pakistan). And in 15 of the 57 societies, less than a majority said they always preferred democracy. But because people’s understandings of what democracy means tend to vary (see Bratton et al. 2005; Shi and Lu 2010; Mattes et al. 2015), at least some people who say they support or even prefer a “democratic” regime may see no conflict with practices such as one-party rule. Thus, we also asked authors to assess individuals’ rejection of nondemocratic, or authoritarian, forms of government. As a third step, we asked authors if possible to identify questions that probe an even deeper level of support for specific democratic institutions or that indicate an embrace of democratic or liberal values. Fourth, we asked them to explore how people evaluate the performance of democracy,

Growing Up Democratic?

17

using wherever possible the widely asked question on “satisfaction with democracy.” Most importantly, perhaps, we asked that each chapter contain a broad multivariate model that examines the potential impact of political generations, as defined individually with each country’s history in mind, along with a respondent’s level of attained education, while holding constant the respondents’ perceptions of socioeconomic conditions in their countries, their exposure to political information, their level of political interest, and several demographic factors. All told, then, authors have not only developed regionally or nationally specific demarcations of political generations but have also assembled measures of education, news media use, economic and political performance evaluations, and political interest, discussion and efficacy, plus standard controls for gender and age. At the same time, we asked the contributors to test the impact of any unique measures they felt had had special bearing on political attitudes in their regions. Without data from panel studies (which interview the same individuals across various points in time—and simply do not exist as yet in the developing world), these models will not fully sort out the statistical effects of age, period, and cohort, but they will go a long way toward testing whether respondents’ views vary meaningfully by generation even after taking account of changing education levels, perceptions of economic conditions, levels of education and political interest, exposure to political news and information, and so on. All told, our fundamental goal in this volume, then, was to have sufficient comparability in the survey analysis across the regions of the world to enable cross-national comparisons while having enough measurement flexibility to allow the contributors to draw out distinctive factors in their regions.12

The Data

As far as possible, we have asked our contributors to use data from the Global Barometer Surveys (GBS). GBS is a consortium of regional crossnational and longitudinal survey projects in postauthoritarian societies in the developing world and constitutes an important addition to the globalization of public opinion research that took place in the first decade of the twenty-first century (Norris 2009). Compared with another major survey project, the World Values Survey, GBS surveys are cast at a lower order of abstraction and focus on measuring how people relate to the actual regimes, institutions, governments, and incumbents they have in front of them (Mattes 2007). To use Richard Rose’s imagery (n.d.), they are

18

Robert Mattes, David Denemark, and Richard G. Niemi

“transformation” surveys that use “realist” measures of how people in transitional societies experience rapid political change and react to imperfect multiparty systems, rather than “destination” surveys that assess how far a country has progressed toward some ideal model of democracy. Thus, rather than asking people about democracy through Likert-type statements about democracy in the abstract, the GBS questionnaires ask people to choose between democracy and authoritarianism in general and between democracy and specific forms of autocratic rule with which respondents have some experience. The oldest of these projects, the Latinobarómetro, began in 1988 with a pilot survey in the Southern Cone countries of Argentina, Brazil, Uruguay, and Chile. In 1995 another round of surveys was conducted in eight countries, adding Peru, Venezuela, Mexico, and Paraguay. From 1996 to 2004, the project was expanded to include 17 countries (adding Bolivia, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, and Panama) covering the whole of continental Latin America. In 2005 the Dominican Republic was also included, bringing the total to 18 countries.13 Originally known as the New Democracies Barometer, the Eurasia Barometer dates back to 1991, with a first round of surveys in six postcommunist countries of central and eastern Europe (Bulgaria, Czechoslovakia, Hungary, Poland, Romania, and Slovenia) as well as a comparison survey of democratic Austria. Thereafter, and until 2005, six additional rounds of surveys corresponded to the split between the Czech Republic and Slovakia, and progressively added Croatia, Belarus, Estonia, Latvia, Lithuania, Montenegro, Russia, and Ukraine, bringing the total to 17 countries. From 2007, however, surveys came to an end in the new EU member states (which were now surveyed by Eurobarometer), but surveys continued in Belarus and Ukraine and added Azerbaijan, Armenia, Georgia, Kazakhstan, Kyrgyzstan, and Moldova. In addition, there is a time series of 20 separate national surveys of Russia (New Russia Barometer) conducted between 1992 and 2012.14 Afrobarometer conducted its first surveys in 12 countries in southern Africa (Botswana, Lesotho, Malawi, Namibia, South Africa, Zambia, and Zimbabwe), East Africa (Tanzania and Uganda), and West Africa (Ghana, Mali, and Nigeria) between 1999 and 2001. Round 2 was conducted in 16 countries between 2002 and 2004, adding Cabo Verde, Kenya, Mozambique, and Senegal. Round 3 was done in 18 countries in 2005–2006, adding Benin and Madagascar. Finally Round 4 was conducted in 20 countries in 2008–2009, adding Burkina Faso and Liberia.15 Originally called the East Asia Barometer survey, the Asian Barometer Survey began with surveys in eight countries (China, Hong Kong,

Growing Up Democratic?

19

Japan, Mongolia, Philippines, South Korea, Taiwan, and Thailand) in 2002–2003. A second round of surveys was conducted in 13 countries in 2006–2007 (adding Cambodia, Indonesia, Malaysia, Singapore, and Vietnam). And a third round of surveys was completed in the same set of countries between 2009 and 2013.16 As part of the Asian Barometer, a separate group of scholars also carried out a single wave of surveys in 2004–2005 in five countries on the subcontinent (Bangladesh, India, Nepal, Pakistan, and Sri Lanka) that together are known as the South Asia Barometer.17 Finally, the Arab Barometer carried out its first wave of surveys in seven North African and Middle East countries (Algeria, Jordan, Lebanon, Kuwait, Palestine, Tunisia, and Yemen) in 2006–2007 and a second round in 11countries in 2010–2011 (adding Egypt, Iraq, Saudi Arabia, and Sudan).18 Two of the chapters in our volume are on regions not covered by GBS projects. First, because the democratic transitions in Greece, Portugal, and Spain and the extensive democratization of Italy took place well before the start of these regional projects, Richard Gunther and José Ramón Montero’s chapter on southern Europe uses other national surveys from those countries that ask either identical or—as far as possible—conceptually equivalent questions on support for and satisfaction with democracy. Second, because there are no dedicated democracy-oriented projects in North America or western Europe that ask the full range of desired questions, the chapter by David Denemark, Todd Donovan, and Richard Niemi makes use of conceptually equivalent items from the World Values Survey (but taking heed of the limitations imposed by WVS question formats that we have discussed above) and the Comparative Study of Electoral Systems. Cases and Chapters While our chief goal is to examine opinion change in countries and regions that initiated transitions away from authoritarian rule since the beginning of the third wave in 1974, our use of data from Global Barometers and other related projects means that we are able to examine trends in public opinion in democracies such as Italy, Japan, and India whose origins stretch back to the late 1940s. Aside from these exceptions, Part 1 of this book presents a series of studies of democratization from the third and fourth waves in broadly chronological order. We begin with Portugal, Spain, Greece (and Italy) (Chapter 2), which democratized in the late 1970s, then move to Latin America (Chapter 3), where transitions occurred largely in the early to mid-1980s, and then on to East Asia (Chapter 4), and South Asia (Chapter 5), where political change came

20

Robert Mattes, David Denemark, and Richard G. Niemi

both before and after 1989. Chapters 6 (central and eastern Europe) and 7 (sub-Saharan Africa) consist only of transitions that occurred after the fall of the Berlin Wall in 1989. In Part 2, we take advantage of the availability of comparable data from the advanced, postindustrial democracies (Chapter 8) as well as the Arab world (Chapter 9) and China (Chapter 10) to determine whether findings from the democratizing world also occur in long-standing democracies or in countries that are clearly not democratic.

Conclusion

All told, this book was designed to explore patterns in support for democracy and opposition to authoritarianism from a truly global perspective. It does so first and foremost by putting the theory of adolescent political socialization and generational change to the test, and subsequently by investigating the impact of cognitive mobilization, or economic and political performance evaluations as complementary or competing explanations. The volume pursues this analysis across the various regions of the world, most of which are characterized by rising levels of support for democracy, amid ongoing patterns of support for old authoritarian systems, doubts about new governments’ ability to deliver economic and physical security, and fears that the principles of democratic rights and privileges will continue to be undermined by corruption and lawlessness. By using the Global Barometer Surveys, and because we have asked each team of authors to include comparable measures and analytic points of reference, the chapters enable the reader to gain insights not just from the factors unique to a region but also comparatively—across the regions as well as between old democracies, new democracies, and those nations yet to build democracy. The final chapter of the book attempts to distill those comparative insights into a set of the larger lessons that derive from the various chapters and, we hope, to provide at least some partial solutions—but also, perhaps, some new pieces—to the puzzle of democratization that inspired the book.

Notes 1. Samuel Huntington (1991) argued that the world was passing through the “third wave” of democratization, which had begun in 1974 with a coup that eventually led to the successful democratization of Portugal. While many scholars of democratization continue to refer to this entire period as the third wave, we are persuaded by scholars who have argued that the frequency, speed, and types of transitions since 1989 have been sufficiently different to label this period as the fourth wave of democratization.

Growing Up Democratic?

21

See Doreenspleet (2000), McFaul (2002), Berg-Schlosser (2007), and Markoff (2006). 2. We thank Russell Dalton for this point. 3. Finifter and Mickiewicz (1992) and Miller, Hesli, and Reisinger (1994) reported, respectively, that support for political change was greater among young people in the Soviet Union in late 1989 and in post-Soviet states in 1990–1992, but most of the individual items were not specifically about democracy. 4. The term cohort—especially birth cohort, as we will note later—is sometimes used as a way of distinguishing socially or politically defined generations from family (blood-related) generations. 5. He also wrote that experience “absorbed . . . in early youth [emphasis added]” is “not easily destabilized,” although it can be destabilized, more easily by youth than by “the older generation” (299–301). 6. We are not suggesting that change is limited to young people. Older people can and do change (Danigelis, Hardy, and Cutler 2007). What we are asserting is that attitudes learned early in life in many instances put a brake on later learning. 7. Given the inherent ambiguities, it seems pointless to push any further the question of exactly when an event occurred—i.e., what day and month. We implicitly assume that turning 18 in the year of the event means that an individual was 18 when the event occurred. 8. Assigning people to a generation based on the period in which they spent the majority of their formative years was used previously by Grasso (2014), though she used 15–25 as formative ages. 9. The term age cohort is sometimes used, but it is inherently ambiguous if one has surveys from more than one year because it could mean respondents of a given age at the time of the survey even though they might have been born in different years. 10. The terms cohort and generation are often used interchangeably, though generations are usually thought of as connected by some shared historical experience such as having grown up during the Great Depression or after the 9/11 attacks in the United States. The boundaries of such events are often imprecise; nevertheless, for purposes of analysis, generations are often operationalized in terms of specific birth years. 11. With which one of these statements are you most in agreement? Democracy is preferable to any other kind of government. In some circumstances, a nondemocratic government can be preferable. For someone like me, it doesn’t matter what kind of government we have. This question was originally designed by Leonardo Morlino and Juan Linz. 12. Fully sorting out age, period, and cohort/generational effects is an extraordinarily challenging task, one that ideally makes use of multiple kinds of data and complex statistical methodologies. While we cannot hope to do that in this volume, a number of steps described in the text go a long way toward testing whether respondents’ views vary meaningfully by generation even after taking account of changing education levels, perceptions of economic conditions, degrees of political interest, exposure to political news and information, and so on, and therefore testing the alternative hypotheses. These steps include the following: 1. Relying on a precise delineation of the generations in question: Each chapter is written by experts in the part of the world under consideration; in each case, the authors discuss briefly the history of each country in the area in order to determine the beginning and end of each type of governing structure. 2. Identifying precisely who (in terms of birth years) belongs to each generation: We specified one’s formative years (for political attitudes) as ages 14–22 and

22

Robert Mattes, David Denemark, and Richard G. Niemi

3.

4.

5.

6.

identified members of a given generation as those who experienced a majority of their formative years under a given type of governing structure. Identifying multiple democracy generations: While many countries have moved toward democracy quite recently, thus establishing a correlation between youthful age and being a member of a democratic generation, the fact that there were earlier periods of democracy in a number of countries reduces or eliminates this correlation. Identifying multiple types of generations: Almost every country’s history includes episodes of multiple kinds of authority structures—collectively including colonial rule, monarchy, military rule, indigenous one-person rule, communist party rule, multiparty elections, electoral democracy, and liberal democracy—thus allowing for the possibility that governing structures other than democracy lead citizens to adopt particular attitudes toward governance. Considering both democratic “demand” and “supply”: Each chapter uses questions about citizens’ evaluations of democracy versus various kinds of authoritarian alternatives but also about their satisfaction with the way democracy works in their country, recognizing that people of different ages and generations may respond differently to these two aspects of governing. Using multivariate statistical models: In each chapter, the authors use multivariate models (after showing simple differences across generations), including the respondent’s education, evaluation of the economy (the specific measure dependent on the survey question available), level of political interest and exposure to mass media news and information sources, and other measures, and sometimes age itself.

13. For more information, see www.latinobarometro.org. 14. For more information, see www.cspp.strath.ac.uk/catalog4_0.html and www .abdn.ac.uk/ecsr/research-projects. 15. For more information, see www.afrobarometer.org. 16. For more information, see www.asianbarometer.org. 17. For more information, see www.democracy-asia.org. 18. For more information, see www.arabbarometer.org.

Part 1 Postauthoritarian Societies

2 Southern Europe: Elite-Led Culture Change Richard Gunther and José Ramón Montero

The consolidation of new democratic regimes depends, to a considerable degree, on developing among citizens widespread support for democracy as the only legitimate framework for political conflict— that is, as “the only game in town.” Similarly, the vitality and proper functioning of democratic systems depend on an engaged democratic citizenry that, it has long been argued (e.g., Almond and Verba 1963), normally requires a sense of efficacy and competence. Finally, it is often argued that the legitimacy of democratic systems depends on widespread satisfaction with the performance of its political institutions. This exploration of generational change in postauthoritarian southern Europe focuses on the development of democratic support and its links to performance satisfaction, as well as to “internal political efficacy” and norms relevant to active engagement in democratic politics. As various chapters in this volume and our own earlier work (Montero, Gunther, and Torcal 1997; Gunther, Montero, and Torcal 2007) demonstrate, the most commonly used indicators of these democratic attitudes constitute three separate domains that are conceptually and empirically distinct from one another, with decidedly different antecedents and behavioral consequences. Two of these clusters of attitudes toward democracy are roughly similar to David Easton’s (1965) distinction between diffuse and specific support. Democratic support pertains to citizens’ beliefs that democratic politics and representative democratic institutions are the most appropriate (indeed, the only acceptable) framework for government (Linz 1988: 65). We regard this as the key attitudinal component of regime legitimacy and expect that such orientations will be stable over time. In contrast, 25

26

Richard Gunther and José Ramón Montero

satisfaction with the performance of democracy is based on “judgments about the day-to-day actions of political leaders and the operation of governmental institutions and processes” (Kornberg and Clarke 1992: 20). As such, we expect it to fluctuate in accord with government performance, the condition of the society and economy, and the performance of key political institutions. The following analysis, therefore, will separately analyze the correlates of democratic support and performance satisfaction. The third cluster of attitudes, political disaffection, is conceptually distinct from both of those described above. Political disaffection involves disengagement of members of a polity from both its core political institutions and, more generally, from politics (Di Palma 1970: 30; Montero et al. 1997). Due to the absence of systematic data on external efficacy in the surveys for this chapter,1 we limit our attention to one of disaffection’s key dimensions, “internal political efficacy”—or rather inefficacy: that is, the perception that politics is too complicated to understand (Campbell et al. 1960: 58). Internal efficacy has been shown in numerous studies to make a substantial contribution to, if not a prerequisite for, active engagement in democratic politics. The following analysis will be based primarily on two different methods. The first will be cohort analyses of these attitudes over time. This not only enables us to monitor the evolution of the attitudes of political generations in Greece, Italy, Portugal, and Spain over time but also provides more solid grounds for setting forth informed propositions about the causal processes underpinning the development of these democratic attitudes. The second will be multivariate analyses in which each of these orientations is, in turn, treated as a dependent variable, enabling us to explore in greater detail the formative processes that instill these attitudes among the electorates of these four countries. Our exploration of the likely origins of these democratic attitudes will address four specific hypotheses related to ideas set forth in the introductory chapter of this volume. First, informal childhood political socialization assumes that individuals’ political attitudes are affected by their informal observation of parents, siblings, and others, and tend to be stabilized by the time they have passed through their most impressionable period (i.e., ages 14 to 25). A cohort analysis over time should thus be able to identify distinctive generational markers, which would characterize a political generation from that age forward. With regard to attitudes such as basic support for democracy, we expect to see that older generations socialized entirely under displaced authoritarian regimes will exhibit lower levels of democratic support than the young, who acquired political attitudes within a democratic context.

Southern Europe

27

Closely related to this is the impact of formal socialization through a country’s education system. We test two different hypotheses linking education to the development of these fundamental political orientations. In the case of democratic support, we focus on the specific content of the regime’s formal socialization messages and hypothesize that those who were educated under the authoritarian regime—whose formal “civics training” was explicitly antidemocratic (e.g., in the case of Spain, see Fernández-Miranda 1965)—will be lower in their levels of support than younger generations who received formal education explicitly supportive of democracy. A different hypothesis relating to formal education focuses on “cognitive mobilization” and the development of participatory behavioral norms, rather than support for democratic regimes: instead of the explicit teaching about the preferability of one regime type over another, the crucial impact of education might be to develop cognitive and other skills that facilitate active citizenship later in life. In this case, it is not the specific propagandistic intent of the regime that would matter but rather the development of skills relevant to political activism and of broader orientations such as internal efficacy where we would find the most decisive impact of education. A third hypothesis abandons the focus on the putative formative years of one’s youth and instead emphasizes direct learning from observations of the political process that might occur at any age. A causal process of this kind would not leave behind a generational marker among cohorts, in which those individuals who pass through their formative years will display a distinctive set of attitudes for the rest of their lives. Instead, individuals of all ages might exhibit a “period effect,” in which all cohorts would be affected in the same way at the same time and would retain the resulting political orientations for the rest of their lives. A second difference between this direct learning process and the two previously described is that the fundamental formative agents are political actors at the macro level of the polity, rather than parents, siblings, peers, or civics textbooks. Finally, we explore whether support for democracy results from satisfaction with the performance of democratic political actors and institutions. If measures of performance satisfaction are found to fluctuate substantially over time, we would not expect to observe a lifelong impact in cohort analyses. Accordingly, testing this hypothesis will rely more on bivariate correlations and multivariate analysis of the links between these two sets of attitudes. Our analysis of these political attitudes in southern Europe will demonstrate that support for democracy is not dependent on satisfaction with the performance of democracy or the economy. Not only was democracy

28

Richard Gunther and José Ramón Montero

consolidated in two of these countries (Greece and Spain) at a time of economic crisis but support for democracy (at least in Spain) has not been shaken by the even more devastating “Great Recession” that began in 2008. Our cohort analysis will further demonstrate that support for democracy was not the product of formal political socialization: older southern Europeans who were educated under the former authoritarian regimes are no less supportive of democracy than younger respondents. Instead, support for democracy resulted from a process of direct learning from key political elites during crucial stages of the democratic transition. However, formal socialization does appear to be the principal cause of differing levels of internal efficacy and subsequent active engagement in democratic politics.

Political Generations in Southern Europe

The four southern European democracies—Greece, Italy, Portugal, and Spain—provide us with ideal case studies for exploring the extent and origins of core democratic norms and values. All four of them experienced periods of authoritarian rule, but by the 1980s all four were consolidated democratic systems. Despite these general similarities, their historical trajectories varied significantly, so we can examine the impact on the development of democratic attitudes of quite different socialization experiences among the various “political generations” that make up their contemporary electorates. Spain and Portugal suffered under authoritarian regimes that lasted several decades (1926–1974 in Portugal, 1939–1975 in Spain), while the so-called Colonels’ regime in Greece lasted only eight years (1967–1975). Accordingly, the great majority of the Spanish and Portuguese electorates underwent childhood socialization under authoritarian regimes, while the authoritarian interlude in Greece might be expected to have been so brief as not to leave a lasting imprint on the development of democratic norms and values among Greek citizens. And in Italy, the transition to democracy was so early (the mid- to late 1940s) that the overwhelming majority of respondents interviewed in our 1985 and 1996 surveys would have been socialized under the democratic regime of the so-called first republic (1946–1994). Similarly, the nature of these countries’ transitions to democracy also followed different trajectories. Benito Mussolini’s would-be totalitarian regime (1922–1943) came to an end in the face of impending defeat in World War II, and the establishment of Italy’s second-wave democracy was profoundly affected by the Cold War and the emergence of what were initially antisystem or semiloyal parties on both the left (the Italian

Southern Europe

29

Communist Party—PCI) and the right (the neofascist Italian Social Movement—MSI). The partial regime liberalization in Portugal under Marcelo Caetano (1968–1974) failed to forestall a full-scale social and political revolution that initially culminated in rule by a nondemocratic left-wing military junta and subsequently a long-term process of democratization and regime consolidation that was initiated only after a military countercoup in late 1975. In contrast with the protracted democratization process in Portugal, democratic transition and consolidation occurred quickly in Greece following the ouster of the colonels by the military high command in 1975 and the installation of a civilian government under Constantine Karamanlis. The four-decades-long authoritarian system of Spain ended with the death in 1975 of the elderly dictator Francisco Franco, but the emergence of a broad-based consensus in support of a new democratic regime was largely the product of negotiations among party leaders through the so-called politics of consensus (1977–1979). As will be seen, these various political trajectories make it possible to separate survey respondents into cohorts on the basis of clearly distinctive socialization experiences. Exploration of the development of democratic norms and values in southern Europe is greatly facilitated by the rich survey data resources at our disposal. All four of these countries are members of the Comparative National Elections Project (CNEP) (Gunther, Montero, and Puhle 2007; Gunther, Beck, et al. 2015), which includes strictly comparable measures of these democratic norms and values gathered through surveys conducted in the 1990s and 2000s. In addition, all four were included in the 1985 Four Nations surveys under the direction of Julián Santamaría and Giacomo Sani, making it possible for us to monitor the evolution of these norms and values over a period of two decades. In the case of Spain, our time frame is broadened considerably by two large surveys conducted in 1979 (less than two years after Spain’s first democratic elections) and 1982 (in conjunction with a dramatic party system realignment).2 In addition, our Spanish 1982–1988 panel study included a substantial in-depth interview segment that dealt explicitly with childhood socialization experiences and early involvement in politics under the Franco regime and during the democratic transition.3 The definition of cohorts to be used in this analysis is complicated by the fact that the periods of rapid economic development and societal modernization, transition to democracy, and democratic consolidation all differed substantially in terms of their timing, with Italy preceding the other three by as much as three decades. Moreover, as will be seen in the analysis that follows, we find that support for democracy is very much affected by events that unfolded during the transition, while internal

30

Richard Gunther and José Ramón Montero

political efficacy is strongly affected by the country’s education system, whose development is, in turn, closely linked to its level of affluence and socioeconomic modernization. Accordingly, no single breakdown by age could adequately capture these crucial formative influences in each of these four countries. Thus, we establish cohorts separately for each country, in accord with the most salient political developments that occurred at the time when the respondent was passing through his or her most crucial formative period. This conception of political generations is based on the idea that major and salient social and political developments can confer durable attitudes and behavioral norms on citizens (Mannheim 1952: 276), especially if these developments occur during the most formative period of adolescence or young adulthood. These lasting imprints may result from direct experience with the social or political events in question or from the ways in which they are constructed and interpreted within a polity’s collective memory (Schuman and Scott 1989). Our definition of each cohort as beginning with age 14 is consistent with other studies in this literature. Adelson and O’Neil (1966), for example, identified the crucial period of cognitive development when abstract notions like “community” become meaningful constructs as between age 13 and 15. Italy In the case of Italy, this logic supports the creation of five cohorts. The oldest political generation includes those who turned 14 in 1943 (when Mussolini was ousted from power and his fascist regime came to an end) or earlier. This included all of those respondents who were 67 or older at the time of the 1996 CNEP survey and those who were 56 or older at the time of the 1985 Four Nations survey. These individuals therefore experienced their most formative period of development under Mussolini’s fascist regime, within which formal socialization through the education system and state propaganda was explicitly antidemocratic. The next political generation (cohort 4 in the following tables) includes those respondents for whom late adolescence occurred between 1944 and 1952—that is, during the period when Italy’s democratic transition unfolded, the Cold War polarized politics, and its party system was consolidated. (It should be noted that Italy’s democratic regime was not consolidated—and even then, not fully consolidated—until sometime in the mid- to late 1970s.4) Cohort 3 includes those who came of age politically between 1953 and 1967, a period characterized by rapid economic growth under unstable coalition governments dominated by the Christian Democratic party (DC). The political generation making up cohort 2 would have been

Southern Europe

31

strongly influenced by events that unfolded between 1968 and 1980. This was a period in which the Communist Party abandoned its initial antisystem and semiloyal stands, and in the words of its leader, Enrico Berlinguer, recognized the “historically universal value of democracy” (Bosco 2001: 339). The formal ideological embrace of democracy (under the tenets of “Eurocommunism”) was accompanied by changes in the party’s policy stands (e.g., explicitly accepting NATO) and its behavior in local government and the national parliament (even to the point of giving parliamentary support to a minority DC government). But this time period was also one of growing frustration (especially on the left) with political stagnation and DC hegemony. The youngest cohort included those who turned 14 between 1981 and 1992, when the party system and the political elite of the governing coalition parties were increasingly discredited by widespread corruption and continuing political stagnation, resulting in plummeting levels of satisfaction with the performance of democracy. Revelations of massive corruption scandals at the end of this period culminated in the complete collapse and realignment of the Italian party system in 1994 (Sani and Segatti 2001). Portugal The first of the so-called third wave of democracies emerged in Portugal following the end of an authoritarian system that dated back to the overthrow of the democratic but anticlerical republic in 1926 by right-wing and religious military officers. The trajectories of both the predemocratic past and the transition to and consolidation of democracy in Portugal are rather complicated, leading to the creation of an additional cohort in an effort to capture possible influences of these developments. The first phase of relevance to this study is the regime of António de Oliveira Salazar, established in 1928 by the junta’s former finance minister. Salazar created a repressive authoritarian system that hid behind a democratic facade. Its elections were coercively manipulated by Salazar to guarantee the election of a president who would regularly reappoint him as prime minister. The oldest of the political generations used in our studies of Portugal included respondents who had reached the age of 14 prior to Salazar’s incapacitation by a stroke in 1968. The next political generation (cohort 5) includes individuals who passed through their most formative years under the authoritarian rule of another former university professor, Marcelo Caetano (1968–1974). Caetano initiated a partial liberalization of the regime that allowed for a slight enhancement of political pluralism. At this same time, however, Portugal was deeply divided by the terrible economic and human costs of wars in Angola and Mozambique, intended

32

Richard Gunther and José Ramón Montero

to preserve Portugal’s colonial empire. The Portuguese Revolution of 1974 to 1975 swept away all vestiges of the preceding right-wing authoritarian regime and undertook massive land seizures (creating collective farms) and nationalization of a substantial segment of the industrial and service economy (increasing the state’s share of gross fixed capital from 18 percent in 1973 to 46 percent in 1975). This chaotic economic and political regime transformation, however, did not move Portugal toward the establishment of representative democracy. That would occur only after a countercoup in late 1975 that culminated in the enactment of a new constitution establishing a democracy that was incomplete insofar as the military reserved for itself important political powers, and that included clauses that rigidly protected “the conquest of the Revolution” (thereby violating the definition of democracy set forth by Juan Linz and Alfred Stepan [1996]). Thus, political generation 4 includes individuals who reached the age of 14 between 1974 and 1981, at a time when the new Portuguese regime was not consolidated and, indeed, not fully democratic. The next step toward full democratization occurred in 1982 with the abolition of the nondemocratic Council of the Revolution, a militarydominated body with a variety of political powers including the authority to rule on the constitutionality of all legislation. Given those constitutional clauses that prohibited democratically elected governments from privatizing the revolution’s nationalized industries, however, the transition and consolidation of Portuguese democracy would not be completed until the constitutional reforms of 1989. Thus, Portuguese politics during the period when the fourth of our political generations came of age (1982–1986) was characterized by a still incomplete democratization process, as well as a high level of cabinet instability. A major change in the quality of Portuguese democracy occurred in 1987, when Portuguese voters elected to office the first of back-to-back single-party governments. The period of stable single-party government between 1987 and 1994 largely defines the context within which the next political generation (cohort 2) was socialized. Finally, after 1995, domination of party politics by the moderate center-left Socialist Party and the moderate center-right Social Democratic Party began to erode away, culminating in increased party system fragmentation. It is within this political context that our youngest generation of Portuguese voters was socialized. Greece Greece experienced two periods of authoritarian rule, the first of which was the dictatorship of Ioannis Metaxas, which lasted from 1936 until his death and the Nazi invasion of Greece in 1941. Following World War II,

Southern Europe

33

democracy was restored, but it was hindered by the violence and instability of the civil war (which ended in 1948). Accordingly, the oldest political generation includes those who came of age in 1949 or earlier. Over the following two decades, Greece had an illiberal democratic regime in which those who had supported or sympathized with the communist side in the civil war were deprived of fundamental political and civil liberties. This limited democratic experience dominated the political socialization of those who turned 14 between 1950 and 1966 (cohort 4). The third of our Greek cohorts includes those who experienced their most formative period under the right-wing and repressive Colonels’ regime of 1967 to 1974. The rapid transition to and consolidation of democracy that occurred between 1975 and 1981 dominated the political landscape and strongly influenced attitude development for the next political generation (cohort 2). Finally, the youngest of our cohorts would have been socialized during a period of democratic normalcy from 1982 to 1992. Spain The oldest of the political generations that we initially defined for Spain includes those who turned 14 in 1951 or earlier. Accordingly, the traumatic breakdown of the fully democratic but unstable Second Republic (1931–1936) and the ensuing civil war (1936–1939) would have exerted a powerful influence on the political socialization of the oldest respondents falling within this age group. They and those in this political generation born over the following decade also experienced the harsh years of political repression, economic deprivation, and international isolation during the first phase of the dictatorship of Francisco Franco. The next Spanish political generation (cohort 5) came of age during the period of stagnation and economic deprivation under the Franco regime between 1952 and 1966. These were followed by individuals who turned 14 between 1967 (when a partial liberalization of the Franco regime was initiated) and 1974, the last full year under Franco’s authoritarian rule. Political generation 3 includes Spaniards who lived their most formative years between Franco’s death in 1975 and 1981. This was a period of political uncertainty, as Spain transitioned from an authoritarian system and a rigidly centralized state to full democracy within a substantially decentralized state that acknowledged the self-government rights of important Basque and Catalan minorities. Although it soon culminated in a consolidated democratic system, it was a time of considerable political violence on the part of Basque terrorists, as well as an attempted military coup in 1981. But the hallmark of this crucial period was the highly successful “politics of consensus” that led to consensual support for the new democratic

34

Richard Gunther and José Ramón Montero

regime from all nationwide political parties, with only a minority of Basques not buying into the democratic consensus. Respondents included in cohort 2 passed through their most formative years during a time of political stability (marked by continuous government under the moderate socialist prime minister, Felipe González), rapid economic growth, and entry into the European Union. Finally, our youngest cohort includes those individuals who turned 14 between 1993 and 2000. In contrast with domination of Spanish politics by the moderate PSOE that characterized the previous time span, this was a period of balanced but increasingly polarizing competition between center-left and center-right parties at the national level. The age brackets that define political generations in these four Southern European countries are summarized in Table 2.1. Let us now turn our attention to the ways in which these authoritarian experiences and transitions to democracy affected varying levels of democratic support and participatory engagement in the democratic process.

Support for Democracy

We begin with a comparative examination of levels of support for democracy in these four southern European countries over time. We base this analysis on two different measures. The first asks respondents to choose among the following three statements: “Democracy is preferable to any other form of government”; “under some circumstances, an authoritarian regime, a dictatorship, is preferable to a democratic system”; and “for people like me, one regime is the same as another.” This questionnaire item was included in the 1985 Four Nations study of political attitudes and behavior in Greece, Italy, Portugal, and Spain, as well as in most Comparative

Table 2.1 Political Generations in Greece, Italy, Portugal, and Spain Year when respondent reached 14 years of age Cohort Youngest cohort Cohort 2 Cohort 3 Cohort 4 Cohort 5 Oldest cohort

Greece

Italy

Portugal

Spain

1982–1992 1975–1981 1967–1974 1950–1966 na Before 1950

1981–1992 1968–1980 1953–1967 1944–1952 na Before 1944

1995–2001 1987–1994 1982–1986 1974–1981 1969–1973 Before 1969

1993–2000 1982–1992 1975–1981 1967–1974 1952–1966 Before 1952

Note: “na” indicates not applicable.

Southern Europe

35

National Elections Project surveys conducted since the mid-1990s. However, a slightly different item was administered in Spain over a longer period among respondents in surveys that we conducted between 1979 and 2004 (as well as the 2005 Portuguese survey). This item asks respondents if they agree or disagree with the proposition that “Democracy is the best political system for a country like ours” (with “it depends” as an intermediary option). Given the additional analytical leverage we shall derive from analyzing these other Spanish data—particularly insofar as the 1979 survey predates what most observers regard as the consolidation of Spanish democracy—this alternative measure will be used in our indepth analysis of the Spanish case, as well as for Portugal in 2005. It should be noted, however, that there is no significant difference between the results obtained from the administration of these alternative measures, as can be seen in the two sets of Spanish data presented in Table 2.2. Democratic Support and Generations By the 1990s, democracy was consolidated in all four southern European countries, with democratic support particularly strong in Greece, Portugal, and Spain. Support for democracy was relatively weak in Italy in 1985 and continued to lag behind the other three countries in the mid-1990s, but even then the percentage of respondents agreeing that “democracy is preferable to any other form of government” was equal to that of Great Britain (76 percent), higher than levels of democratic support found in

Table 2.2 Support for Democracy (percent)

Democracy is preferable Greece Italy Portugal

Spain Democracy is preferable Democracy is best

1985

1988

1992

1996

2005

87 70 61

90 74 84

91 73 83

85 76 na

na na 87a

1979

1982

1993

2004

60 65

81 80

81 80

84 90

a “Democracy is best.” Sources: Data for Greece, Italy, and Portugal 1985, 1988, and 1992 from Montero et al. 1997, 129; 1996 data for Greece and Italy, 2005 data for Portugal, and 1993 and 2004 data for Spain from CNEP (http://u.osu.edu/cnep); Spanish data for “Democracy is preferable” from data archives of the Centro de Investigaciones Sociológicas (www.cis.es); and Spain 1979 and 1982, DATA, S.A. (Gunther, Sani, and Shabad 1986, and Linz and Montero 1986).

36

Richard Gunther and José Ramón Montero

Belgium (70 percent) and Ireland (63 percent), and substantially higher than in countries like Mexico (53 percent) and Brazil (50 percent) (Gunther et al. 2007: 35). It is also clear that democracy was not strongly consolidated in Spain in 1979 or Portugal in 1985. In both countries, however, support for democracy has strengthened substantially over time. These mass-level survey data corroborate claims based upon elite-level developments concerning the consolidation of democracy (as discussed in Endnote 5): accordingly, democracy can be said to have been consolidated in Greece by 1981; in Spain (except for the partial regional exception of some Basque nationalists) by 1982; in Portugal by the mid- to late 1980s; and in Italy by the mid- to late 1970s (Gunther, Diamandouros, and Puhle 1995: 1–32; Graham 1992; Cotta 1992). An examination of support for democracy broken down by cohorts helps us to address several of the hypotheses set forth above concerning the origins of democratic support. If formal socialization or informal childhood socialization processes were largely responsible for the acquisition of this important attitude, democratic support should be strongest among younger generations whose crucial formative period occurred under the current democratic system. Conversely, older respondents, especially in Portugal and Spain—whose late childhood and adolescence occurred under the Franco or Salazar/Caetano regimes when political indoctrination through formal education and state propaganda was explicitly antidemocratic—should exhibit lower levels of democratic support. In sharp contrast to these expectations, the data presented in Table 2.3 provide no support for these childhood socialization processes as the origins of democratic support. With the exception of Italy in 1996, there is no substantial link between generation and support for democracy, and in the Italian case this democratic attitude tends to be weakest among the youngest cohort.

Table 2.3 Support for Democracy, by Political Generation (percent) Percent preferring democracy over other options Cohort Youngest Cohort 2 Cohort 3 Cohort 4 Cohort 5 Oldest N

Spain 2004

Portugal 2005

Greece 1996

Italy 1996

88 91 92 95 92 86 2,890

83 89 91 93 92 88 2,705

84 86 82 86 na 86 959

69 78 81 75 na 79 2,503

Sources: See note, Table 2.2.

Southern Europe

37

This suggests that childhood and early adolescent political learning in the late 1980s and early 1990s—a period of extreme dissatisfaction with the behavior of virtually the entire Italian “political class” (ultimately culminating in the total breakdown and replacement of the party system of the “first republic”)—had some negative impact on the development of support for democracy among young people. But among the other cohorts, democratic support actually strengthened somewhat over the preceding decade, as we will see below. We continue our search for evidence of causal processes that might account for the development of attitudes supportive of democracy by examining support for democracy among these cohorts over time. In the case of Greece, both the 1985 and 1996 surveys reveal an overwhelming and stable consensus in support of democracy among all cohorts, which ranged between 85 and 92 percent among the four age cohorts in the 1985 survey, and between 82 and 86 percent in 1996. No significant agecohort effects are apparent (as can be seen in Online Appendix Figure A2.1). It is likely that this democratic consensus is the product of several factors. First, the Colonels’ regime was of such short duration that it lacked the impact on political culture of the more durable authoritarian regimes of Spain and Portugal. In addition, the Greek military junta never sought to mobilize popular support through the establishment of a political party or collaboration with previously existing parties, and therefore failed to resocialize the population in accord with its political values and bizarre “Helleno-Christian” worldview. Finally, and perhaps most importantly, the democratic transition that unfolded under the skillful leadership of the conservative Constantine Karamanlis succeeded in forging broad interparty support for the new democratic regime. He did so through major political reforms that brought about reconciliation between left and right, healing the deep rift that resulted from the Greek civil war of the 1940s, as well as the ostracism and persecution of the left under the flawed democracy of the 1950s and 1960s (Featherstone and Katsoudas 1987). In the case of Italy, there is little evidence that might provide clear insights into the origins of democratic support. In both 1985 and 1996, all but one of our cohorts were clustered tightly together and did not change substantially over the following decade, a slight strengthening of support for democracy notwithstanding: support for democracy ranged between 68 and 73 percent in 1985 and, with one exception, between 74 and 81 percent in 1996 (see Online Appendix Figure A2.2). That one exception is the appearance by 1996 of a new cohort of Italian citizens that departed somewhat from this pattern, insofar as the youngest respondents were somewhat less supportive of democracy than their predecessors. And as

38

Richard Gunther and José Ramón Montero

we will see in Table 2.4, when these cohorts were used to create “dummy variables” and entered into a multivariate equation with democratic support in 1996 as the dependent variable, only the coefficient for the youngest cohort achieved statistical significance. Other studies, however, provide some insights into the extent of regime support in earlier decades and suggest that mass media and political elites played a substantial role in generating legitimacy for Italy’s democratic regime. In the late 1950s, Joseph La Palombara and J. B. Waters (1961) found that there was weak support for democracy among the young, even though they had not been socialized under the fascist regime. A short time later, Gabriel Almond and Sidney Verba (1963: 402–404) painted a picture of an Italian political culture distinguished by alienation from the political system mixed with opposition to the democratic regime. But over the following decades, the mass media and political elites encouraged the development of democratic support among citizens. At the same time, these supportive attitudes were accompanied by extraordinarily high levels of dissatisfaction with the performance of democracy in Italy, as well as by widespread disaffection from politics (Sani 1980; Segatti 2006). The case of Spain, however, reveals very substantial change over the course of the 25 years between the new regime’s second democratic election (1979) and 2004. And the way in which these attitudes evolved among the various political generations provides clear insight into what kinds of political processes were at work—or not at work. Our earlier study (Montero, Gunther, and Torcal 1997) found some evidence of differences among cohorts in the early years of this study, with the youngest significantly more favorable toward democracy than older cohorts. This suggests that prodemocratic attitudes were beginning to emerge during the final years of the Franco regime—with exposure to news coverage of political developments in neighboring European countries serving as an important vehicle for socializing Spaniards into support for democracy (Gunther, Montero, and Wert 2000: 42). This initial generational effect reflected the evolution of Spanish society and its political system over the previous decades, including the downgrading of the Franco regime’s single party (the Falange) and its ideology in the 1950s, the progressive end of Spain’s international isolation, and, most importantly, the partial liberalization that occurred following enactment of the Press Law in 1966. Attitudes toward democracy, however, were relatively weak, inchoate, and somewhat incoherent (López Pintor 1981). Much more powerful than this generational effect (which quickly disappeared once the transition was well under way) was a period effect linked to a political variable to which we will turn our attention below. All cohorts in our survey of the period from 1979 to 2004 moved in the same direction at the same time:

Southern Europe

39

older cohorts at each point in time are not weaker in their support for democracy as the result of either informal socialization or formal education under the authoritarian regime of General Franco, and the youngest generations are not stronger in their support for democracy,5 despite having received “democracy education” in their youth within a fully democratic system. A period effect, consistent with a direct learning interpretation of these changes, however, is very much in evidence. While support for democracy increased incrementally after 1982, a dramatic increase in support for democracy occurred between the late 1970s and 1982, and this has left a lasting imprint on political values among all political generations. Party Elites and Democratic Support What processes brought about this change? Building on the “transition and consolidation effect” hypothesized by Mariano Torcal (1995), we have argued that the strategies and behavior of prominent political elites and organizations during particularly salient stages in the democratization process may have a major impact on the political attitudes of their respective sets of followers (Gunther et al. 2007). In a comparative analysis of cross-national variations in support for democracy emerging from several third-wave transitions, we found that in those instances where outgoing authoritarian elites played a positive and constructive role in the transition to a democratic regime, support for democracy did not emerge as a divisive cleavage separating supporters of one party from another, and a broad interparty consensus in support of democracy emerged. Conversely, when the dominant elite from the previous nondemocratic regime opposed democratization, as in Chile and Bulgaria, supporters of those parties with origins in the outgoing regime tended to be skeptical or hostile toward democracy and the new regime, and they retained those undemocratic attitudes over an extended period of time. Our Spanish survey data enable us to test the “transition and consolidation” hypothesis. Spain’s democratic regime was established through a process of interparty bargaining commonly referred to as “the politics of consensus.” With the exception of separatist Basques (a very small minority of the Spanish electorate), all parties of the center and left fully supported the new regime that came into being with the December 1978 constitutional referendum. Since these parties were backed by the vast majority of voters, this elite-level consensus largely explains the speed with which the regime was consolidated. Creation of this new democratic regime, however, was opposed by small parties on the extreme right, by a significant segment of the Basque population, and by half of the parliamentary delegation representing the right-wing Alianza Popular (AP,

40

Richard Gunther and José Ramón Montero

whose leaders had their origins in the more authoritarian and conservative factions of the Franco regime). Accordingly, the new democratic system was only partially consolidated by the time of our 1979 postelection survey, and levels of support for democracy closely corresponded to the stands taken by the leaders of the various parties. Among supporters of parties on the extreme right in 1979, for example, only 24 percent agreed with the proposition that “democracy is the best political system for a country like ours,” while 44 percent disagreed. With regard to supporters of the conservative AP, which underwent a schism in 1978 (with half of its deputies voting against the constitution, leading to their marginalization from the party, while the remainder followed party leader Manuel Fraga in supporting the constitution), an intermediate stance by the party’s supporters reflected this elite-level division of opinion: 48 percent of those who cast ballots for AP in 1979 agreed that democracy was best. In contrast, 70 percent of those who voted for more fully democratic political parties shared this democratic belief. The subsequent disappearance from active politics of the antidemocratic faction of the AP’s leadership, in combination with the complete discrediting of the antisystem extreme right wing in the aftermath of the 1981 coup attempt, led to a substantial increase in democratic support by the time of our 1982 survey. Over the long term, partisan differences in levels of support for democracy have disappeared. Replication of this analysis following the 1993 elections, for example, reveals no statistically significant relationship between partisan preference and support for democracy (Gunther and Montero 2001). We interpret this as the result of the substantial transformation of the Partido Popular (PP, successor of Alianza Popular), which had moved in the late 1980s and early 1990s toward the center of the political spectrum, fully embraced democracy, and undergone a dramatic demographic shift to a new generation of leaders (Montero 1987; García-Guereta 2001). This contrasted with the attitudes of AP voters in the 1979 election, which was held just three months after the schism divided the party’s founding elites. In short, despite Manuel Fraga’s firm commitment to the new democratic constitution, insufficient time had elapsed to allow the AP in that earlier election to demonstrate its new commitment to democracy and shed its links to the authoritarian past. One might suspect that this increase in support for democracy among AP/PP supporters over these three decades was the product of generational replacement—that the older generation of conservative Spaniards, socialized under the Franco regime to oppose democracy, progressively died off and were replaced by a new type of conservative voter more supportive of democracy. In order to test this childhood socialization hypothesis, we undertook a new round of cohort analyses in which the attitudes of AP/PP

Southern Europe

41

supporters can be compared with those of respondents who voted for other democratic parties.6 (The number of cohorts was reduced from six to three in order to have enough AP supporters to be meaningful.7) As can be seen in Figure 2.1, there is relatively weak evidence of a generational effect, and most of that is mediated by partisanship. To be sure, the oldest cohort of AP supporters exhibited by far the lowest level of support for democracy in 1979, followed by the second-oldest cohort. Both of the older cohorts of AP supporters, however, changed dramatically over the following two decades with regard to their levels of support for democracy. Agreement among AP voters born in 1937 or earlier that “democracy is the best form of government” increased from 45 percent in 1979 to 88 percent in 2004, while the second-oldest cohort shifted from 55 percent agreement in 1979 percent to 95 percent in 2004. By that time these two older generations of PP voters, 91 percent of whom expressed support for democracy, were no less supportive of democracy than those in the youngest cohort (which was socialized entirely under the new democratic system). Most importantly, there is very strong support for the period effect, in which AP voters of all ages progressively embraced democracy over time in accord with their party elites’ reorientation. Portugal also presents us with little evidence of a generational effect regarding support for democracy. There is, however, clear evidence of a period effect, as well as significant differences among partisan groups in the early period of Portuguese democracy. As we saw in Table 2.2, support for democracy in Portugal was weak in 1985 (with only 61 percent of all

Figure 2.1 Democracy Is Best, by Party (Spain, 1979–2004)

Source: See note, Table 2.2.

42

Richard Gunther and José Ramón Montero

respondents agreeing that democracy is the best form of government) but grew substantially over the following two decades. We attribute this initially low level of democratic support in part to the poor performance of Portugal’s government institutions: between 1976 and 1983 the average Portuguese government lasted just 10 months, giving party politicians a collective image of irresponsibility and undermining popular support for the new regime. Indeed, when asked which prime minister provided the best government for Portugal, fully 28 percent of respondents in a 1978 survey mentioned Marcelo Caetano (Bruneau 1981). The next-mostpopular leader was Socialist Party leader Mário Soares (9 percent), who narrowly surpassed two authoritarian figures, Vasco Gonçalves (8 percent) and António de Oliveira Salazar (7 percent). Significant differences in democratic support between Portuguese voters on the left and right provide evidence in support of a second causal factor, which is consistent with our transition and consolidation hypothesis. Only 59 percent of supporters of the two more conservative parties (the PSD and the CDS) agreed that “democracy is preferable,” while 70 percent of supporters of parties to the left of center (the socialist PS, the communist PCP and its allied APU) supported democracy. As mentioned in the historical overview presented earlier in this chapter, the left-wing Armed Forces Movement insisted on including in the 1976 constitution two institutional features, most strongly opposed by parties of the center and right, that violate the Linz and Stepan (1996) definition of democracy. The first was the Council of the Revolution—composed of military officers who, in conjunction with the president of the republic, were to rule on the constitutionality of legislation, play a crucial role in the appointment of governments and the dissolution of parliament, and function as supreme arbiter in all matters pertaining to the military. The second was the inclusion in the constitution of several clauses that made it impossible for duly elected and democratically accountable governments to privatize or streamline inefficient industries that had been nationalized during the revolution. The Council of the Revolution was abolished in August 1982, but these constitutional provisions could not be overturned until a second round of constitutional reforms in 1989. After that reform, democratic support strengthened considerably: support for democracy among the three age cohorts in 1985 ranged between 61 and 70 percent; by 2005 this had increased to between 83 and 93 percent (see Online Appendix Figure A2.3.) It is noteworthy that by 2005 there was no difference between parties of the left and right in this regard. Also noteworthy is the absence of a generational effect consistent with the childhood socialization hypotheses. Indeed, as we saw in Table 2.3, support for democracy in 2005 was weakest among the youngest cohort.

Southern Europe

43

In Greece, as well, there were some theoretically relevant differences in levels of support for democracy among the principal partisan groups. In 1985, 71 percent of supporters of the conservative New Democracy agreed with the statement that “democracy is always preferable,” significantly below the 92 percent level of support for democracy among PASOK voters. And even though agreement with this statement among ND voters increased to 76 percent over the following decade, their level of democratic support was lower than the 85 percent among the Greek electorate as a whole. In the absence of time series data prior to 1985, we can only speculate that these partisan differences resulted from a substantial reorientation of New Democracy under Karamanlis’s successor as party leader from 1981 through 1984, Evangelos Averoff. In a misguided effort to expand the party’s electoral base by appealing to the 7 percent of Greeks who had supported extreme right-wing political parties in 1977, Averoff co-opted several prominent leaders of the National Front party (EP) and shifted ND’s ideological appeals to the far right. While this strategy succeeded over the short term in capturing votes from former EP supporters, “this move effectively alienated many a liberal voter and cost ND its share of the traditional center” (Pappas 2001: 245). Thus, despite its origins under Karamanlis as a heterogeneous party of the center, ND was subsequently seen as a party of the right, and its electorate included a small far-right antidemocratic minority. Finally, the increase in support for democracy among all but the youngest cohort in Italy can also be attributed largely to changes in the official positions and public statements of two key sets of political elites. Initially, the party system of the first republic included antisystem parties on both the left (the Italian Communist Party—PCI) and the right (the neofascist MSI). The PCI had progressively abandoned its antisystem stance and behaved like a loyal democratic competitor by the 1980s, but its official repudiation of the traditional stance of communist parties and unequivocal embrace of democracy was not consummated until its rebirth as the PDS (Democratic Party of the Left) in 1991. This was followed by the formal abandonment of fascist nostalgia and antidemocratic sentiments by the MSI, which converted itself into the explicitly democratic (but right wing) Alleanza Nazionale (AN) in 1995. Our survey data clearly reflect these institutional and elite-level developments: support for democracy rose from 30 percent among MSI voters in the 1985 survey to 64 percent of AN voters in 1996, while preferences for authoritarian forms of government fell dramatically from 56 percent to 23 percent. Similarly, voters on the left reflected the change in their party’s formal stance as preferences for democracy rose from 70 percent among PCI voters in the 1985 survey to 89 percent in 1996. Indeed, virtually all of

44

Richard Gunther and José Ramón Montero

the strengthening of support for democracy that can be seen in Table 2.2 (and in Online Appendix Figure A2.2) can be attributed to the reorientation of these two segments of the electorate. To this point, we have argued that support for the new democracies of southern Europe is very much a function of mass-level responses to elite-level political developments, especially the roles played and the stands taken by prominent leaders of political parties at crucial stages of the democratization process. This was reflected most clearly in a powerful period effect in Spain and Portugal, in which citizens in all cohorts altered their attitudinal support for democracy in response to significant national-level political developments. Analysis of democratic support among political generations over time revealed only weak evidence of generational effects in Spain consistent with either the formal or informal processes of childhood socialization, and that was restricted to AP supporters. But can citizens’ satisfaction with the performance of democracy be seen as a determinant of basic support for democratic regimes?

Democratic Support and Satisfaction with the Performance of Democracy

Several scholars have asserted that there is a strong causal link between democratic support and performance satisfaction. They argue that the consolidation of new democracies is contingent upon their performance, and especially on how they manage their economies. Weatherford (1987: 13), for example, states that “over the long run, of course, legitimacy is wholly determined by policy performance.” Przeworski (1991: 95) flatly asserts, “As everyone agrees, the eventual survival of the new democracies [in post-Soviet Eastern Europe] will depend to a large extent on their economic performance.” Some (e.g., Fuchs and Klingemann 1995: 440) have even suggested that the legitimacy of established Western democracies is increasingly dependent on their performance. Several of our surveys include variables that enable us to analyze this relationship. We begin with a look at aggregate levels of satisfaction with the performance of democracy juxtaposed with the percentages of respondents in each country who believe “Democracy is the best political system” or “Democracy is preferable to any other form of government.” If democratic support were entirely dependent on satisfaction with the performance of democracy, only Spain would have a consolidated democratic regime. Levels of performance satisfaction were shockingly low in 1996 in Greece and Italy (where only 29 percent of respondents said that they were “very” or “fairly” satisfied with the performance of democracy), in

Southern Europe

45

sharp contrast with the very high level of support for democracy at that time in Greece (85 percent) and evidence of adequate consolidation in Italy, where 76 percent of respondents agreed that “Democracy is preferable to any other form of government” (see Online Appendix Table A2.1). A more detailed look at the relationship between democratic support and performance satisfaction reveals a link of only modest strength. The correlations within each country were statistically significant (at the .001 level or better) but ranged between weak and moderate: the Kendall’s tau-b statistic was .11 for Spain in 2004 and .14 in 1993, .21 in Portugal in 2005, and .22 in both Greece and Italy in 1996. This suggests that performance satisfaction can contribute to the development of attitudes supportive of democracy but is hardly deterministic. Indeed, the consolidation of democracy in Spain in 1981–1982 occurred in the immediate aftermath of a powerful wave of dissatisfaction (popularly referred to as desencanto) rooted in an unemployment rate of 16 percent, inflation in the cost of living ranging between 14 and 16 percent, and widespread disgust with the squabbling and disintegrating party in government (whose share of the vote fell from 35 percent in 1979 to 7 percent in 1982). The moderate impact of performance satisfaction on support for democracy can be more fully appreciated in the multivariate equations presented in Table 2.4, which include several standard control variables (tapping socioeconomic and demographic factors, as well as the respondents’ degree of political engagement), including cohorts converted into dummy variables—coded 1 for those included in a given cohort or zero for those excluded from that cohort (with the oldest cohort as the excluded category). In all four countries, performance satisfaction emerged as the strongest predictive variable, but this is largely because it is included in equations alongside other variables that have little impact on the development of democratic support. It is important to note that when performance satisfaction was entered as a second step following a previous equation containing all of the other variables, it increased the R2 (an indication of the predictive power of the equation, commonly understood as reflecting the percentage of the variance explained) by just .017 in Spain, by .017 in Portugal, by .031 in Greece, and by .041 in Italy. With one exception, the other relationships presented in this table are weak or statistically insignificant. The only other impact of an independent variable that is substantial is the aforementioned weakness of democratic support among the young in Italy, whose most salient period of political socialization would have corresponded with the corruption scandals of the early 1990s and the massive repudiation of virtually the entire political class in the 1994 election. In terms of political generations, the only

Table 2.4 Multivariate Analyses of Support for Democracy Spain 2004 b (se)

46

Youngest cohort .051 (.031) Cohort 2 .048 (.029) Cohort 3 .083 (.031) Cohort 4 .086 (.030) Cohort 5 .047 (.026) Education .007 (.004) Gender .032 (.016) Economic satisfaction –.017 (.005) Income .004 (.013) Interest in politics .015 (.011) Newspaper reader .024 (.019) Radio news listener .007 (.019) TV news viewer .077 (.025) Discuss politics .027 (.010) Satisfaction with democracy .081 (.012) R2 .05 N 2,755

beta — — .07** .07** — — .04* –.07*** — — — — .06** .05** .12***

Portugal 2005 b (se)

beta

–.144 (.039) –.024 (.034) –.021 (.040) .046 (.033) .043 (.041) .005 (.007) –.018 (.022) –.012 (.007) .028 (.012) .015 (.014) .007 (.027) –.005 (.024) .091 (.045) .046 (.015) .156 (.022) .06 1,998

–.09*** — — — — — — — .06* — — — .05* .07* .16***

Greece 1996 b (se)

Italy 1996 beta

–.088 –.005 –.066 –.001

(.072) (.076) (.072) (.060)

— — — —

.022 .048 .007 –.024 .005 .046 –.003 .008 .021 .158

(.012) (.042) (.012) (.027) (.022) (.047) (.047) (.086) (.044) (.029) .05 939

— — — — — — — — — .19***

b (se) –.295 –.134 –.008 –.057

beta

(.052) (.052) (.050) (.056)

–.19*** –.09* — —

.051 (.009) .011 (.027) .014 (.020)

.12*** — —

.106 (.029) .024 (.082) –.112 (.079) .067 (.018) .194 (.019) .11 2,404

.08*** — — .08*** .21***

a

a

Sources: See note, Table 2.2. a No cohort 5 in Greece and Italy (see Table 2.1). *p