(palgrave Studies In European Political Sociology) Jennifer Shore - The Welfare State And The Democratic Citizen-springer International Publishing_palgrave Macmillan (2019).pdf

Palgrave Studies in European Political Sociology

The Welfare State and the Democratic Citizen How Social Policies Shape Political Equality

Jennifer Shore

Palgrave Studies in European Political Sociology

Series Editors Carlo Ruzza Department of Sociology and Social Research University of Trento Trento, Italy Hans-Jörg Trenz Department of Media, Cognition & Communication University of Copenhagen Copenhagen, Denmark

Palgrave Studies in European Political Sociology addresses contemporary themes in the field of Political Sociology. Over recent years, attention has turned increasingly to processes of Europeanization and globalization and the social and political spaces that are opened by them. These processes comprise both institutional-constitutional change and new dynamics of social transnationalism. Europeanization and globalization are also about changing power relations as they affect people’s lives, social networks and forms of mobility. The Palgrave Studies in European Political Sociology series addresses linkages between regulation, institution building and the full range of societal repercussions at local, regional, national, European and global level, and will sharpen understanding of changing patterns of attitudes and behaviours of individuals and groups, the political use of new rights and opportunities by citizens, new conflict lines and coalitions, societal interactions and networking, and shifting loyalties and solidarity within and across the European space. We welcome proposals from across the spectrum of Political Sociology and Political Science, on dimensions of citizenship; political attitudes and values; political communication and public spheres; states, communities, governance structure and political institutions; forms of political participation; populism and the radical right; and democracy and democratization. More information about this series at http://www.palgrave.com/gp/series/14630

Jennifer Shore

The Welfare State and the Democratic Citizen How Social Policies Shape Political Equality

Jennifer Shore Mannheim Centre for European Social Research University of Mannheim Mannheim, Germany

Palgrave Studies in European Political Sociology ISBN 978-3-319-93960-5    ISBN 978-3-319-93961-2 (eBook) https://doi.org/10.1007/978-3-319-93961-2 Library of Congress Control Number: 2018946640 © The Editor(s) (if applicable) and The Author(s) under exclusive licence to Palgrave Macmillan, part of Springer Nature 2019 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover image © Blend Images - Hill Street Studios This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

This book is the culmination of what began as a graduate school application and eventually developed into a broader research project and would not have been possible without the support of my graduate advisors. I would, therefore, like to first thank Ellen Immergut who oversaw my dissertation project from day one. Her keen ability to see the forest from the trees helped me to focus on the big picture and not get bogged down in minutiae. I greatly benefited from her academic and professional experience as well as her feedback at all stages of the project. Bernhard Weßels took particular interest in my research even before joining as my second supervisor. Not only does he hold a wealth of knowledge on the subject matter of my dissertation, he also provided me with a great deal of guidance on the methodological and statistical aspects of my research. I am also indebted to many other experts who helped me along the way: Isabelle Stadelmann-Steffen and Staffan Kumlin—who happened to organize a weeklong workshop on policy feedback, democratic citizenship, and welfare states—provided me with many new sources of inspiration as well as detailed, constructive feedback on my project. I would also like to thank Markus Freitag, who, in addition to joining my committee, helped me get to where I am today. The Berlin Graduate School of Social Sciences (BGSS) at the Humboldt University of Berlin provided an ideal environment for my doctoral studies. I am grateful for the funding from the German Research Foundation v

vi  Preface

that allowed me the freedom to fully pursue my academic interests. The infrastructure and networks offered by the BGSS proved to be an invaluable source of support throughout my graduate studies. Martin Nagelschmidt, the academic manager at the BGSS, is an enthusiastic advocate for all the doctoral students, one who always encouraged our research and made sure the necessary resources were available. I would also like to thank the staff at the graduate school for their unwavering patience and help, in particular, Abenaa Adomako, Katrin Bauer, Sonja Grossmann, Christoph Raiser, Jana Rieckmann, and Thomas Wosnitza. The journey that the dissertation process entails can be a rather lonely one at times, and I certainly could not have done it without the support of my fellow graduate students at the BGSS, in particular, the members of my cohort. A few names deserve particular mention: Lisa Anders, Priska Daphi, Alexander Horn, Christoph Jindra,  Anne Krüger, Andreas Schäfer, Katharina Rathmann,  Anna Schlumbohm, and Jan Ullrich. In addition to the ideas, offices, lunches, and beers we shared, you helped me get through the more trying times of writing a dissertation. Though this book project has its roots in my doctoral work, I have benefited from the support and input from numerous colleagues well past turning in the dissertation, in particular, Jale Tosun, Felix Hörisch, and Bettina Schuck at Heidelberg University as well as my new ­colleagues at the Mannheim Centre for European Social Research. I am also grateful to the team at Palgrave Macmillan for their guidance and patience! To my dear friends scattered all across the globe: your support, much needed distractions, and friendship mean the world to me. Thank you, Allison, for your keen editorial eye. I am particularly grateful to René. Thank you for listening to me ramble about my research over the delicious dinners you made, never failing to make me laugh, and reminding me that there are things more important than university. Lastly, I am forever indebted to my family, which now spans both sides of the Atlantic. My parents, Susan and Kerry, instilled within me a sense of curiosity and determination that has been crucial to starting and finishing this research project. My sisters, Sarah and Traci, are both role models and

 Preface    

vii

wonderful friends. Although separated by an ocean, my family’s love and support have always been a source of strength and encouragement. Special thanks to the smallest (and forthcoming) family member for pushing me to get this done. I know you would have been the first to own a copy, so this book is dedicated to you, Dad. Mannheim, Germany

Jennifer Shore

Contents

1 Introduction   1 2 Democracies and Their Citizens   9 3 Why Context Matters: The Role of Public Policy  29 4 Empirical Approaches to the Study of Democratic Citizens and Welfare States  53 5 The Impact of Social Policy on Democratic Citizenship  75 6 Consequences of Unequal Citizenship 137 Appendix 153 Index 163

ix

List of Figures

Fig. 4.1 Social expenditures on working-age adults and families (% of GDP)64 Fig. 5.1 Income level and predicted turnout depending on welfare state spending93 Fig. 5.2 Political interest in 26 OECD countries 102 Fig. 5.3 Income level and predicted political interest depending on welfare state spending 109 Fig. 5.4 Trust in parliament across countries 114 Fig. 5.5 Income level and trust in parliament depending on welfare state spending 121 Fig. 5.6 Satisfaction with democracy across 22 countries 123 Fig. 5.7 Income level and satisfaction with democracy depending on welfare state spending 128

xi

List of Tables

Table 4.1

Disaggregated welfare state expenditures across countries (% of GDP) 60 Table 4.2 Welfare spending categories 61 Table 4.3 Demographic needs adjustments by welfare policy category 63 Table 4.4 Summary statistics working-age and family benefits (OECD) 64 Table 5.1 Countries, years, and survey sources included in the analyses of voting and welfare states 76 Table 5.2 Self-reported electoral participation and official turnout rates80 Table 5.3 Self-reported voter turnout by income group 81 Table 5.4 Multilevel analyses of voting 87 Table 5.5 Multilevel analyses of voting with cross-level interactions 90 Table 5.6 Marginal effects of welfare policy on voting by income category91 Table 5.7 Political interest by income group 103 Table 5.8 Multilevel analyses of political interest 105 Table 5.9 Multilevel analyses of political interest with cross-level interactions107 Table 5.10 Marginal effects of welfare policy on political interest by income category 108 Table 5.11 Political trust by income group 115 Table 5.12 Multilevel analyses of trust in parliament 117

xiii

xiv 

List of Tables

Table 5.13 Multilevel analyses of trust in parliament with cross-level interactions118 Table 5.14 Marginal effects of welfare policy on trust in parliament by income category 120 Table 5.15 Satisfaction with democracy by income group 124 Table 5.16 Multilevel analyses of satisfaction with democracy 126 Table 5.17 Marginal effects of welfare policy on satisfaction with democracy by income category 127 Table 5.18 Marginal effects on satisfaction with democracy, various welfare state indicators 129 Table 6.1 Marginal effects of welfare state spending on aspects of democratic citizenship 138 Table A1 Voting: Variables, descriptive statistics, sources, hypothesized direction of influence 153 Table A2 Interest in politics: Variables, descriptive statistics, sources, hypothesized direction of influence 156 Table A3 Trust in parliament: Variables, descriptive statistics, sources, hypothesized direction of influence 158 Table A4 Satisfaction with democracy: Variables, descriptive statistics, sources, hypothesized direction of influence 160 Table A5 Alternative welfare policy specifications and satisfaction with democracy 162

1 Introduction

Democracy, when boiled down to one person, one vote, explicitly demands equality. When examining who votes and otherwise participates, however, we all too frequently encounter an unequal reality: although modern democracies espouse the ideal of political equality, economic inequality severely constrains an egalitarian exercise of political rights. Both the level and the distribution of political participation and democratic engagement vary widely across countries, leaving us wondering how some countries have achieved greater political equality than others, and, more specifically, what role does the welfare state play in shaping democratic citizenship? Can the welfare state, in addressing income-based inequality, also alleviate the related political inequalities and thereby encourage political equality? These are the main questions I examine in this book. Beyond the importance of individual resources, personality characteristics, or political socialization, we must also consider the influence of public policy on political engagement. This idea, known as the policy feedback approach, emphasizes that policies impact recipients much like formal political institutions do—they too can transmit norms and impose rules and regulations—and can thereby transform politics itself. Why should we expect welfare policies to impact democratic citizenship? First, © The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2_1

1

2 

J. Shore

they confer people with material benefits, that is, greater resources enhance participatory capacity. Second, social spending can increase political saliency and relevancy of politics for people’s everyday lives. Third, strong welfare commitments send the message that citizens’ interests are protected and valued, thereby enhancing feelings of democratic support and political efficacy. In this book, I contribute to the extant literature on democratic political engagement by examining the ways in which welfare state policies impact individual political behaviors and attitudes. By looking at broader sets of policies over a large and diverse group of countries, this study aims to bridge the gap identified in the policy feedback literature regarding the effects of policies on mass publics. While there are numerous works that examine the effects of specific policies for smaller target groups in single countries, fewer studies have applied the policy feedback framework to larger populations. More specifically, I comparatively apply many of the ideas that have been investigated primarily in the United States to a diverse international setting. In doing so, I expand the literature that tends to quite narrowly focus on the role of resources for political engagement and systematically investigate the variety of ways in which the welfare state context may modify this relationship. I examine the varied responses democracies have developed to address social questions and how these responses shape the quality of democracy in terms of democratic citizenship. Although social stratification represents a mainstay in contemporary social science research, analyses of democratic citizenship and political inequality are often restricted to more theoretical and philosophical accounts and are not subject to systematic empirical analysis. And while many have examined the consequences of economic inequality on voter turnout, finding, for example, that a higher Gini coefficient—a measure of income inequality—is associated with lower turnout, particularly among the less advantaged, these studies tell us very little about the mechanisms behind these relationships. That resources play a role for political participation is by no means a new finding; people with more money or education or better networks participate far more frequently than people with fewer. But what is it about inequality that is so detrimental to political participation and democratic support? By turning the

 Introduction 

  3

attention to public policies, things with which people actually come into contact or have personal experiences, I argue that we can get closer to understanding just why inequality—or its mirror, a generous welfare state—affects political behaviors and attitudes, in terms of both the overall level and the equality of the distribution. In order for citizens to be interested and involved in democracy, they must witness that their governments actually do something to serve their interests. The fact that in many countries we find democratic citizenship to remain very much tied to socioeconomic status tells us that perhaps democracy could and should deliver more. This study aims to make two main contributions: first, it empirically examines the relationship between welfare states and democratic citizenship in advanced democracies. While the relationship between income inequality and political participation has been relatively well documented, far fewer studies approach political inequality from the perspective of democratic quality, comparing what democracies deliver to their citizens and how citizens respond to such policy offerings. Drawing on the policy feedback literature, I comparatively examine how specific policy foci impact the ways in which individuals behave toward and think about their governments. Second, this book goes beyond the procedural and purely institutional accounts of political participation and looks directly at what a democracy is able to deliver. At the end of their introduction to the edited volume Assessing the Quality of Democracy, Larry Diamond and Leonardo Morlino (2005) ask whether a quality democracy provides quality results, such as citizen satisfaction, and whether improvements in quality can bolster people’s attitudes toward and views of democracy. This book aims to provide some answers to questions such as these. The goal is not to be able to say what the perfect democracy looks like but rather to highlight how democratic citizenship can be fostered through public policy.

Overview of This Book In the chapter to follow, I introduce the concept of democratic citizenship, its determinants, and correlates. Next, I provide an overview of the literature on welfare states and their effects on individual behaviors and

4 

J. Shore

attitudes. The main theoretical frame—the policy feedback approach—is then introduced. This perspective states that policies can affect citizens much like formal institutions do—they too can shape identities and foster participation. The effects policies have on their citizens “feed back into political systems, producing spirals in which groups’ participatory and policy advantages (or disadvantages) accrue. Citizens’ relationships with government, and their experiences at the hand of government policy, help determine their participation levels and in turn, subsequent policy outcomes” (Campbell 2011, 2). That policies themselves can play a role in shaping the level and distribution of democratic engagement gives us good reason to study them, given their impact on the fundamental mechanisms of the democratic process. Furthermore, since the distribution of societal goods constitutes a basic function of democratic governments, who gets what, how much, and how, as determined by the design of policies, can weigh heavily on citizens’ capabilities and motivation for political participation. The empirical chapters begin with a discussion of the analytical, theoretical, and epistemological issues of welfare state measurement and a discussion of the statistical methods I employ. Due to the importance that measurements remain consistent with theoretical concepts, I posit that it is not appropriate to examine the welfare state through the lens of aggregate social expenditures. As many others have argued (e.g., Castles 2008; Esping-Andersen 1990), public social spending varies with regard to outcome depending on its object. For reasons I elaborate on in the empirical chapter, I look at spending on working-age adults and families when linking the welfare state to the democratic citizenship. In short, this category of spending is arguably most closely aligned to what EspingAndersen (1990) refers to as the theoretical core of the welfare state. After introducing the measurement of the welfare state employed in this study and discussing the use of multilevel regression as a tool for analyzing the effects of a policy context on the actions and attitudes of individuals in a given country, I move on to the empirical analyses of welfare spending on working-age adults and families on four central aspects of democratic citizenship: electoral participation, interest in politics, political trust, and satisfaction with democracy. The individual-level

 Introduction 

  5

data for these analyses come from the World Values Survey (WVS), the European Social Survey (ESS), and the Comparative Study of Electoral Systems (CSES). What is revealed from these quantitative analyses is that, controlling for both individual attributes and country-level factors, welfare spending on working-age adults and families is indeed associated with higher voter turnout, greater interest in politics, greater political trust, and satisfaction with democracy. The results of the initial multilevel models provide support for a direct effect of the welfare state on democratic citizenship, that is, we can observe higher levels of these democratic citizenship traits in countries, which place a larger policy focus on their working-age populations. Introducing cross-level interactions to the models tests whether this positive effect varies in strength or even direction depending on one’s income level. Although income is not the only predictor of democratic engagement, it is important to bear in mind that it is often the channel through which political influence travels. Moreover, it is these incomebased inequalities that the institutions and policies of generous welfare states seek to address. When looking into the ways in which welfare policies may be able to moderate the relationship between income and political participation and attitudes, however, the results require more differentiation. While the impact of additional welfare spending on the facets of democratic citizenship is positive regardless of which income group one belongs to, the effects differ in magnitude and degree of significance. For example, I find similarly positive and significant effects of social spending on the likelihood of voting for individuals in both the lowest- and middleincome quintiles; although a positive effect is also found for people in the highest-income quintiles, it is not significant. In terms of political interest, regardless of income category, increased welfare state spending on working-age adults and families exerts a positive and significant influence, with the largest and strongest influence being on the political interest of persons with the lowest incomes. This first set of results suggests that increased social spending lessens the importance of income for voting and interest in politics. In other words, although the levels increase across all income groups, people with the fewest resources experience the

6 

J. Shore

greatest boost from additional social expenditures. Those whose resources would already predict a high level of political activity experience a more moderate effect with regard to higher social spending, as is best illustrated with the voting analysis. In terms of political support attitudes—political trust and satisfaction with democracy—I find a seemingly contradictory set of results. Similar to the preceding analyses of political participation and interest, in countries where a greater amount is spent on working-age adults and families (as a percentage of the gross domestic product—GDP), individuals are more likely to be satisfied with the way democracy works as well as exhibit higher levels of trust in their national parliaments. In terms of political trust, however, I do not find that welfare state generosity exerts an equalizing effect; to the contrary, at the higher levels of spending, we find that the gap between those with the lowest incomes and those with the highest incomes actually widens in terms of their political trust. Although welfare spending positively shapes this form of democratic support, its relationship to income-based resources is not muted by increased social expenditures. The results from the final analysis of satisfaction with democracy, however, reveal the opposite pattern: not only are people in states where a larger policy foci is placed on the working-age segment of the population more likely to be satisfied with the functioning of democracy, but there are few discernible differences with regard to where one is located along the income distribution. Following the empirical analyses, I discuss the overall results, limitations, and conclusions of the study. I also touch on some of the directions future research may take based upon my findings, followed by a conclusion with some final remarks regarding the implications that the findings hold for the health and legitimacy of democratic states.

References Campbell, Andrea Louise. 2011. How Policies Make Citizens: Senior Political Activism and the American Welfare State. Princeton Studies in American Politics. Princeton: Princeton University Press.

 Introduction 

  7

Castles, Francis G. 2008. What Welfare States Do: A Disaggregated Expenditure Approach. Journal of Social Policy 38 (1): 45–62. https://doi.org/10.1017/ S0047279408002547. Diamond, Larry, and Leonardo Morlino, eds. 2005. Assessing the Quality of Democracy. Baltimore: Johns Hopkins University Press. Esping-Andersen, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Princeton: Princeton University Press.

2 Democracies and Their Citizens

Democracy and Inequality In his final book, On Political Equality, Dahl (2006) argues that despite the advancements and spread of democracy, political equality has not been achieved. Political equality is not only a desirable normative goal but one which is necessary for the survival of democracy. Although it can be argued that political equality is but an ideal, it is upon this ideal that democracy rests. Inequality, as it expresses itself across multiple dimensions of life, can therefore endanger democracy. Many of Dahl’s forerunners, including Plato, de Tocqueville, and James Madison, also noted the tenuous relationship between inequality and democratic stability (Anderson and Beramendi 2008). Recent studies have provided an empirical basis for the claim that inequality can erode democratic foundations (Acemoglu and Robinson 2006; Boix 2003). Not only is inequality bad for those at the very bottom of the income distribution but for democratic politics as well, as inequality affects “the choice of political regime, the selection of fiscal structures, parties’ mobilization strategies, and the decision to turn out to vote. Thus, inequality is political and institutional not only in its origins but also in its consequences” (Beramendi and Anderson 2008, 5). © The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2_2

9

10 

J. Shore

More broadly, inequality has many implications for the legitimacy of states. First of all, inequality is at odds with democratic norms of equality. Additionally, although nearly all members of a society may enjoy the procedural rights of democracy, if the costs of exercising those rights become too dear for more less advantaged citizens, political influence may be disproportionately limited only to those with greater resources, for as I discuss at greater length in the coming sections, economic disadvantage is very closely linked to political disadvantage in many societies. Moreover, this situation can create disincentives for policymakers to design policies with their poorer constituents in mind, for they are not the voters who could potentially punish them for not maintaining their relative advantage. High-income inequality also implies a more heterogeneous polity, one with a more diverse spectrum of interests, which may, in turn, exacerbate problems of accountability and agency. If it comes to be that only the interests of the better off are being served, political legitimacy will indeed be harmed by inequality. While societal heterogeneity need not necessarily imply negative consequences for a society, if there is nothing, no common or shared interest that binds the different groups together as a “demos,” we may witness decreased legitimacy as a consequence of an eroded sense of community. Eroded legitimacy may have the consequence of destabilizing a democracy, making it perhaps less attractive as a form of governance. Politics that cater to the rich and fail to address common interests is what Ferejohn (2009) refers to as “passive corruption.” On the other hand, if a state is able to include interests of all (or almost all) societal groups insofar as it is committed to universalistic policies, all people, not just the better off, have a stake in what happens in the political decision-making processes (Ringen 2007). Preserving the democratic ideal, therefore, requires much more than simply granting political rights. Democracy, in order to thrive and be considered as good, demands “the continuing responsiveness of the government to the preferences of its citizens, considered as political equals” (Dahl 1971, 1). Democratic citizens, in turn, “depend on democracy for freedom and well-being.” At the same time, “democracy depends equally on them— on their capacity to be citizens and on their beliefs about citizenship” (Ringen 2007, 2). In other words, democracies need citizen participation, and citizens, in order to participate, need democracy. As I discuss in

  Democracies and Their Citizens 

  11

the following section, not all democracies have been successful in guaranteeing political equality, as many have allowed an economic system to interfere with these mutually dependent relationships.

Democratic Quality Once preoccupied by the identification and categorization of authoritarian and democratic regimes, political scientists are now confronted with a new reality and, consequently, a new research agenda: democracy has, as Larry Diamond puts it, “gone global” (2010, 93). Of the 195 countries in the world, 88 can be labeled as a democracy of sort (Freedom House 2018). Democracy has witnessed a boom in attempts to conceptualize and measure it. Indices such as those developed by Polity, Freedom House, Economist Intelligence Unit, Democracy Barometer, and the Bertelsmann Foundation are among the more prominent. While such indices capture the more blatant differences between democracies and non-democracies, they tell us little about the qualitative differences found between the states with the now most common form of government in the world (the Democracy Barometer is the clear exception here). Although the autocracy versus democracy question garnered much attention during the 1990s, as has the question of procedural differences between democracies, until recently, the matter of quality, or how good a democracy actually is, has been notably overlooked. Much like parents wanting to avoid playing favorites, many simply point out that democracies have different ways of doing democracy and have different sets of strengths and weaknesses. Few, however, have taken the seemingly normative leap of identifying some democracies as better than others.1 Stein Ringen (2007), on the other hand, makes the impassioned plea that social scientists need also to be concerned with measuring the quality of democracies and not just their “democraticness” or procedural differences. This is not an altogether unproblematic undertaking, as it requires the researcher to consider “basic questions of norms and legitimacy.” The  This would of course first entail defining what is meant by better—a task often left to philosophers and the op-ed pages.

1

12 

J. Shore

analysis of democratic quality is thus a “value-laden and […] controversial subject” (Diamond and Morlino 2005b, ix). While it is not my intention to add another study on the measurement of democracy to the ever-growing body of literature, it is important to first examine how we conceive of democracy and to identify whether we are interested more in the procedural aspects or the outcomes—that is, what democracies actually deliver to their citizens. An interest in the latter provides the main framework for this study on welfare states and democratic citizenship. Abraham Lincoln’s Gettysburg Address is frequently cited as a means of identifying the different conceptions and purposes of democracy. In his call to ensure the survival of the young American democracy, he emphasized a “government of the people, by the people, and for the people.” In this formulation, one can identify different conceptions of democracy, ranging from a minimalist-elitist formulation (of the people) to a medium-­ participatory conceptualization (by the people) to a maximalist-social conception of democracy (for the people) (Bühlmann et  al. 2007). The minimalist, or procedural conception of democracy is often accredited to Robert Dahl, whose own formulation owes much to Joseph Schumpeter’s definition of democracy as “that institutional arrangement for arriving at political decisions in which individuals acquire the power to decide by means of a competitive struggle for the people’s vote” (Schumpeter 1942, 269). Drawing on this idea, Dahl’s (1971) concept of polyarchy requires inclusiveness (the right of most adults to take part in elections) and public contestation. From such a perspective, little is asked of the electorate, nor does such a conceptualization tell us much about the performance or quality of a democracy. The participatory conception of democracy, however, views citizen participation and political engagement to be crucial for a democracy, as such activities cultivate democratic attitudes and habits. In contrast to procedural democracy, the participatory concept of democracy requires a citizenry active in political life; nevertheless, like the minimal view, it requires little of the state for the people. The maximalist conceptualization of democracy, on the other hand, employs both the procedural and participatory aspects of the other types and adds a social dimension into the mix: not only are legal and civil rights of importance but fair procedures and opportunities as well. The maximalist conception is often neglected in democracy research, for defining social rights and equality remains a matter of contention (Bühlmann et al. 2007).

  Democracies and Their Citizens 

  13

Few would negate the statement “democracy is better than non-­ democracy” (Ringen 2007, 13), but are there some democracies that are better than other democracies? This question must be followed by another question: what are good democracies? If we accept a maximalist understanding of democracy, we argue that democracy is for the citizens (in addition to of and by them). Democracy, in operating in its citizens’ interests, must, therefore, provide them with something. Yet another question arises: what is it that citizens want? The most concise answer to this question would be a good life. In order to achieve this good life, people need the resources and opportunities to be able to decide on and pursue their visions of the good life. This turn toward the lives of people in democracy research parallels a paradigm shift in political science toward methodological individualism: “[A]ll social phenomena (their structure and their change) are in principle explicable only in terms of individuals—their properties, goals, and beliefs” (Elster 1982, 453). In other words, by only examining the system, we are overlooking many important aspects within the system. Ringen (2007) refers to this as “double bookkeeping”—that is, no single measure of democracy is sufficient. We should, therefore, not only be concerned with the differences between democracies and non- or semi-democracies (which, incidentally, have been well documented) but also with the differences between established democracies and the situations of the people living in them. If we rely on the conceptualization that democracy exists to serve the people, a democracy must, therefore, deliver something to its citizens in order to be considered good. In providing an answer to the question What Democracy Is For, Ringen (2007) identifies the security of freedom (how well each citizen’s freedom is ensured and protected) as the primary task of democracy. In advanced democracies, the protection of freedom means something very different than it would in a developing country. Although most people in established democracies need not worry about a military coup or whether they will survive childbirth, not all have the freedom to pursue their vision of the good life.2 In today’s democracies, we can observe marked differences in the quality of people’s lives. Another trend that can be easily observed is the dual presence of democracy and capitalism. In theory, both systems espouse  For a more detailed discussion of the “good life” and the state’s role in its pursuit, see Chap. 3.

2

14 

J. Shore

equality through their formal rules: markets require equal conditions for fair and efficient competition; democracy champions the ideal of one person, one vote. In reality, however, and perhaps through their coexistence,3 they have produced inequalities, both economic and political, and these inequalities can reinforce one another. With rising economic growth, we have also witnessed an increase in the clout of economic power. Although political power remains unchanged (one person, one vote), increasing wealth begets greater economic power. The bulk of this economic power, however, is concentrated in the hands of a few elites. A further development has been increased economic liberalization, which, consequently, has led to the increase of private capital in sectors once under public control (e.g., hospitals, transportation, education systems). Moreover, we have witnessed an increase in the weight of economic power in politics: major political campaigns are nearly impossible without sizeable private wealth or generous financial backing from private investors. Nevertheless, it is often taken for granted that capitalism and democracy go hand in hand. The challenge is, therefore, how to balance democracy and capitalism so as to promote economic efficiency without sacrificing economic democracy. The primary response on the side of nation-states to the incompatibilities that have arisen out of the partnership of democracy and capitalism has been to implement welfare policies. At the time when many welfare states came into existence (in post-war Europe), not only was poverty much more widespread but the risk of falling into poverty was also acute. Today, people are comparatively much better off than they were three-­ quarters of a century ago. While this is a partnership that is unlikely to be decoupled, some question whether political democracy can thrive without economic democracy: “If we have democracy in political life but not in economic life, and if the weight of economic power grows relative to political power, then the citizens might have reason to question how democratic their society really is and whether political democracy is really of much relevance” (Ringen 2007, 48). Such a pessimistic view of the relationship between democracy and capitalism predicts, among other things, a decline in political interest and a rise in indifference, not to  See Iversen (2008) for an in-depth discussion on the coexistence of capitalism and democracy.

3

  Democracies and Their Citizens 

  15

mention increased abstention from voting. Since high levels of political interest and voter turnout are often viewed as indicators of democratic vitality (Lijphart 2001), there are reasons to be concerned about the effects of economic inequality on democracy. As Diamond and Morlino (2005b, x) point out, when examining democratic quality, it is helpful to think about the various qualities as part of the same system, “in which improvement in one dimension can have diffuse benefits for others.” At the same time, deficiencies in a given dimension can reinforce deficiencies along others. For example, while all democracies must grant formal rights of political participation to their adult populations in order to be labeled democratic, good democracies are those that ensure that all citizens are able to exercise their political rights by ensuring equal social rights. This relationship, between what democracies are able to deliver to their citizens and how the citizens respond and are they themselves shaped by the offerings of their governments, is what I am interested in investigating in this book. While I certainly do not advocate we stop investigating the ways democracies elect their officials, how their legislation gets passed, or power is shared, I do, however, argue that if we are concerned about the quality of democracies (as partially indicated by the degree of democratic citizenship), it is also necessary to examine what democracies provide for their citizens and how they ensure freedom for all members: “Freedom and accountability, however they are understood, are necessarily linked to accountability and responsiveness” (Diamond and Morlino 2005b, xiii; Rueschemeyer 2004). All these dimensions4 along which quality of democracy can be measured—citizen participation, political equality, and government responsiveness—are intimately connected and cannot be viewed in isolation. When addressing the question as to why, for example, fewer and fewer are choosing to vote and citizen satisfaction with democracy is on the decline, which are also qualitative indicators of democracy, we must look beyond simplistic answers such as growing apathy or individualistic 4  This is not to say that these are the only dimensions. See, for example Bühlmann et al. (2007) or Diamond and Morlino (2005a) for more nuanced and in-depth discussions on the various dimensions of democratic quality.

16 

J. Shore

s­ocieties and examine the ways in which political and social structures subvert and constrain participation. Alienation from politics does not occur because people stop caring about politics; more likely is the explanation that some people have become alienated from the democratic process because they do not feel that they get anything out of political engagement—neither intrinsically nor materially. An elitist view sees citizens to blame: “Citizens are said to be indifferent, unwilling to participate, unrealistic in demands and expectations, or victims of new values, such as postmodern individualism” (Ringen 2007, 41). But perhaps this trend of turning away from the political has more to do with the quality of the democracy than with the quality of its citizenry. Low levels of political interest may simply indicate that there is less for the citizen to be interested in. If, however, states offer something to their citizens, here, in the form of generous family and employment-related policies, citizens, particularly those at the lower end of the income distribution, will not only have greater resources at their disposal for participation, but politics in general should have more meaning due to the saliency and relevance in their daily lives. Lastly, policies can serve as messengers, providing citizens with cues regarding their position in society and whether their interests are taken as valuable or rather ignored. How policies accomplish these things and what implications they have for the relationships between socioeconomic resources and political behaviors and attitudes are the topics of the chapters to follow.

Democratic Citizens For the remainder of this chapter as well as in the one that follows, I focus on the theoretical groundwork for the empirical chapters, investigating the relationships between aspects of democratic citizenship and the welfare state. Both in the analytical chapters and in the sections to follow, a particular focus is placed on the role of income, in terms of both how it affects people’s propensity to participate and how the welfare state can shape political behaviors and attitudes. As I illustrate, while one’s income is certainly not the only predictor of political engagement, it serves as a placeholder of sorts, and as discussed in the introduction, income

  Democracies and Their Citizens 

  17

i­ nequality and the types of unequal participation it can produce, is indeed a threat to democracies. While it is well established that income and other resources are related to political participation and democratic support (Brady et  al. 1995) and there are many studies showing that income inequality at the aggregate level depresses voter turnout (Dahl 2006; Lijphart 1997; Solt 2008), less is known about the relationships at the micro-level. In other words, although we know economic inequality depresses political participation, the discussion hitherto has centered largely on macro-level indicators such as the Gini index or measures of economic growth or stagnation and much less has been done in terms of public policies (which indeed play an important function in shaping overall levels of income inequality). Policies, as I argue, are what people come into contact with—not abstract indicators such as the Gini coefficient. How policies are designed and implemented are important for people’s feelings of political efficacy and thus for the ways in which individuals engage in and think about the state and democracy. Beginning with a brief overview of the historical and intellectual development of the concept of democratic citizenship, I then discuss the role of resources for political involvement. We then see that socioeconomic resources by no means exist in a deterministic relationship to democratic citizenship; rather, we must take the institutional and policy contexts within which citizens are located into consideration as well, for institutional settings and policy designs can play a major role in muting or amplifying the connections between resources and political behaviors. After discussing the theoretical origins of the welfare state and linking it to a broader consideration of the role states have in people’s lives, I introduce the policy feedback approach and discuss the types of effects public policy can have on people’s propensity to retreat from or get involved with politics.

Origins of Democratic Citizenship Stripped of all adjectives, citizenship itself is a complex concept with a multitude of meanings and temporally shifting ideals. The Aristotelian ideal held that “a citizen is one who permanently shares in the administration of

18 

J. Shore

justice and the holding of office” (Janowitz 1994). This model of “full-time citizenship” makes no distinction between the individual’s public and private identities. The Machiavellian approach, on the other hand, saw the ideal citizen as the citizen soldier, whose personal life was of no significance to the Prince. The traditional liberal approach to citizenship sees “part-time citizenship” as the ideal, with the focus on self-­realization and the individual pursuit of happiness. One’s private sphere trumps any sort of engagement in the political sphere (Hernes 1988). By contrast, the republican ideal comprises both a legal and an ethical dimension of citizenship. Classical republicans, from Rousseau to Tocqueville, recognized “something enriching about public life” (Dagger 2008, 147), citing civic virtue as the yardstick by which to measure the citizen. J.S. Mill even argued that only by becoming civically engaged would the individual be able to meet his5 full potential (Mill [1861] 1975). Many of these approaches to the ideal citizen are very much tied to normative understandings and set very high standards. Furthermore, although they diverge in terms of the scope of the ideal citizen’s activity, the classics tend to agree that “in order to be a citizen of a polis, in order to be able to participate fully in public life, one needed to be in a certain socio-economic position” (King and Waldron 1988). Aristotle, for example, recognized that a large gap between the richest and the poorest citizens would not be beneficial for democracy, for such an inequity would be destabilizing (Saunders 2000; King and Waldron 1988). Throughout the canon of classics, one repeatedly encounters the argument that inequality undermines civic life and is furthermore not conducive to the making of good citizens. More modern theorists, such as Lipset (1960), continue to underscore the link between economic security or growth and democratic orientations: “The more well-to-do a nation, the greater the chances that it will sustain democracy” (Lipset 1960, 131). This thesis, originally put forth by Lipset in his seminal work Political Man, succinctly gets at the core of his modernization theory. Drawing on empirical data from the 1950s, Lipset finds a clear link between stable democracy and economic development, as well as between democracy and levels of communication, education, industrialization, and urbanization. In other  Like his contemporaries, Mill’s ideal citizen was male.

5

  Democracies and Their Citizens 

  19

words, democracy is positively correlated to indicators of modernity. Aspects of modernization, such as increased education and industrialization, are associated with attitudes that support a democratic regime. Economic development also impacts on values: citizens in more economically advanced countries tend to exhibit “orientations conducive to democracy” (Lipset 1960). This is described as the “ego-enhancing” effect of development. A higher degree of modernization entails an increased sense of self-worth among the citizenry and higher levels of satisfaction, trust, and efficacy. To this idea, King and Waldron (1988, 428) add that economic disparity itself “blocks and interferes with the articulate and deliberative reflection” needed for active participation in civic life, ­echoing Maslow (1943), who posited that basic needs must be fulfilled before the individual is able to take part in activities based on self-actualization. While the debate about what makes a democratic citizen is as old as democracy itself, the concept has experienced varying levels of attention through a variety of different lenses over the past 50 years. The political culture movement of the 1960s (Almond and Verba 1963) saw political support as crucial for democratic systems: there must be a relatively high degree of public support of democracy and government in order for the public to accept the public policies and in order to guarantee the persistence of the system. For the civic culture scholars, dissatisfaction signaled instability and a weak democracy (Almond and Verba 1963; Norris 1999), a notion that would be challenged in the 1990s (Klingemann 1999; Norris 1999). Almond and Verba (1963) further emphasize the importance of informed citizens for the strength of democracies as well as the link between high levels of democratic citizenship and the strength of democracy. Participatory theorists of the 1980s, such as Pateman (2000) and Barber (1984), viewed political engagement as the defining aspect of the democratic citizen, which was in direct contrast to the classical liberal view of democracy.6 Similar to Nie et al. (1996) approach, the characteristics of democratic citizenship investigated in this study could also be divided into two complementary dimensions: political engagement and democratic ethos. 6  A democracy “could function effectively even in the absence of an especially virtuous citizenry by creating checks and balances” (Kymlicka 2002, 285).

20 

J. Shore

The first dimension, political engagement, encompasses those attributes of democratic citizenship having to do with political behaviors: voting in elections and interest in politics. The second dimension, democratic ethos, refers to citizenship attributes that reflect the “understanding of and adherence to norms and principles of democracy” (Nie et al. 1996, 5–6). The democratic ethos attributes investigated in this study are political trust and satisfaction with democracy. The authors do not see the citizen’s relationship to the state in terms of homo economicus (driven by individualism), nor as homo sociologicus (constructed purely by societal forces); instead, Nie et al. (1996, 4) view homo democraticus in terms of the transactions between the state and civil society, placing the most emphasis on the intersection of private versus public interests.

 nequal Citizenship: Participation U and the Role of Resources Drawing from both the classical and contemporary definitions of citizenship, it can thus be concluded that democratic citizenship refers to a set of citizenship characteristics that are frequently seen as essential to modern democratic systems. Although contemporary social scientists no longer are as concerned with defining their decade’s prototype of the ideal citizen, most would agree that “a well-ordered polity requires citizens with the appropriate knowledge, skills, and traits of character” (Galston 2001, 217). Additionally, “good citizens” are not born that way; rather, they are a product of their environments. It is not only academia, however, that has taken an interest in civic citizens. In the United States, civic education courses, which once part of the standard curriculum, have been reappearing in many high schools. Politicians, scholars, and the media alike frequently bemoan an apparent growth in political apathy in democracies all over the world. Although there are different interpretations of this trend and the normative character of democratic citizenship makes measuring and defining it somewhat tricky, one cannot help but wonder about the consequences of political disinterest for our societies and democracies. Galston (2001, 220) summarizes these concerns as

  Democracies and Their Citizens 

  21

f­ ollows, further emphasizing the deleterious effects of inequality of political engagement: Even if one rejects the philosophical proposition that active citizenship is essential to human flourishing, or self-regarding pursuits, it is hard to avoid the hypothesis that at some point the withdrawal from public engagement endangers the healthy functioning of democratic polities. At the very least, if the tendency to withdraw is asymmetrically distributed among population groups, then the outputs of the political system are likely to become increasingly unbalanced. And if those who withdraw the most are those who have the least, the system will become even less responsive to their political needs. Political engagement is not a sufficient condition for political effectiveness, but it is certainly necessary.

That political engagement and democratic support are unevenly distributed within societies, and given the implications of this imbalance for the legitimacy of democracies, the relevance of investigating the origins of democratic citizenship becomes even clearer. Nie et  al. (1996) outline two dimensions of how democratic citizenship is a necessary condition for the maintenance of democracy. First, political engagement signifies the capacity of individuals to engage in self-rule. Moreover, this type of activity is necessary to identify, define, and pursue individual interests and preferences. Second, political enlightenment demonstrates the individual’s understanding and acceptance of the rules and norms of democracy. Shklar (1991) describes the pursuit of democratic citizenship as the “quest for inclusion.” To be sure, just and legitimate institutions and regimes are a goal pursued by all democracies. As Janowitz (1991, 206) notes, “a modern democratic society is a society in which all strata have the potential of being democratic citizens.” In order for societies to reach this potential, their institutions also must be capable of assuring that all citizens truly have this opportunity. It could, therefore, be said that not all societies are modern in Janowitz’s terms, as democratic citizenship tends to be unevenly distributed across the social milieu. As I further expound upon in the chapters to follow, political activity tends to follow structural cleavages, which, in turn, has many implications for the representativeness and legitimacy of democratic governments.

22 

J. Shore

This conclusion has also been underscored by a group of prominent political scientists investigating the effects of inequality on the American democracy: “Citizens with lower or moderate incomes speak with a whisper that is lost on the ears of inattentive government officials, while the advantaged roar with a clarity and consistency that policy-makers readily hear and routinely follow” (American Political Science Association Task Force 2004, 651). That political participation and democratic values continue to be more widely practiced and held by the economically better off is by no means a new topic of research. We know, for example, that those with more education and higher incomes are also those who turn out to vote on a more frequent basis than their less well-off counterparts. We also know, however, that despite rising levels of education and wealth, voter participation has not kept pace (Berinsky and Lenz 2011; Persson 2010). A near-canonical set of explanations as to why people do not get involved in politics has been formulated by Sidney Verba and colleagues: “[B] ecause they can’t; because they don’t want to; or because nobody asked” (Verba et al. 1995, 265). The can’t is frequently given the most weight in explaining individual participation, indicating a lack of resources: “time to take part, money to contribute to campaigns and other political causes, and skills to use time and money effectively” (ibid., 16). The link between resources and participation is often thought to be related to the concept of political efficacy. Put simply, political efficacy refers to an individual’s perception of how much influence he or she can have politically. Lane (1959) was one of the first to view political efficacy as a twofold concept in terms of personal and communal influence, commonly referred to as internal and external political efficacy. Internal political efficacy refers to people’s beliefs regarding their own capabilities to understand and effectively participate in politics; external political ­efficacy is the feeling that government authorities and institutions are responsive to citizen demands. Political efficacy is generally understood as a description of citizens’ positive feelings about their own political capabilities and standing vis-à-vis the state; the absence of efficacy can be thought of as political alienation of subjective powerlessness (Hayes and Bean 1993). People who are politically efficacious are much more likely to participate in politics and be generally supportive of democratic norms and princi-

  Democracies and Their Citizens 

  23

ples, as they see themselves as being sufficiently capable and competent to understand politics and effectively participate. They hold the perception that their governments and authorities will be responsive to their demands and thus represent their interests and preferences. One of the most prominent findings with regard to political efficacy is that it is “disproportionately located among those with higher socio-economic status” (Hayes and Bean 1993, 269–70; Almond and Verba 1963). And while it is clear that education is certainly extremely beneficial to developing the cognitive abilities and confidence related to internal political efficacy, and even the feeling that government is responsive to one’s demands (Nie et  al. 1996), the reality in many modern democracies is that the better educated also tend to be disproportionately located among the better earners. Citizens with fewer resources, for example, in terms of income, education, or social networks, are also those who are the least likely to be engaged with politics. Those who lack socioeconomic, cultural, and cognitive capital are, therefore, at a much greater risk of experiencing disadvantages across various spheres of citizenship  (e.g., Stadelmann-Steffen 2012). The trap of poverty is thus not constrained to economic disadvantage but renders other forms of citizenship difficult to attain as well (Shore 2014). This trend is particularly acute among many immigrant groups who lack not only social citizenship rights but formal state membership as well. In many societies, however, there are groups who do have formal political rights (in the form of legal citizenship) but who remain at a definite disadvantage in terms of exercising those rights. The consequences of interlinked dimensions of inequality are numerous. If people from the upper socioeconomic echelons are those who are most likely to make their political voices heard, the outcomes of such unequal political input are indeed reason for concern. It is plausible to question whether this would result in unequal representation of interests in terms of both who gets elected and the types of policies that are implemented. As a result, those underrepresented individuals, those who lack social citizenship rights necessary to become active politically, will continue to be at a disadvantage: “Social citizenship becomes farther and farther out of reach; political citizenship as process becomes limited to fewer and fewer citizens. A vicious cycle has developed” (Simpson Bueker

24 

J. Shore

2009, 426). The consequences for democracy that can arise from inequality and exclusion are many, and while democratic institutions are certainly necessary to achieve political equality, they do not appear to be sufficient (Dahl 2006). While we continue to witness that people with greater resources, be it income, education, or, more frequently, the combination thereof (Verba and Nie 1972; Wolfinger and Rosenstone 1980), participate more than those with fewer resources, we are at the same time confronted with by the reality that aggregate voter turnout rates have declined over the past decades despite increasing economic growth and access to education. Moreover, not only does the level and degree of social stratification of political engagement vary across countries but the distributions of the resources known to correlates of participation also markedly vary (Alber and Kohler 2009; Lijphart 1997; Shore 2014; Solt 2008). Given this state of affairs, the next question to ask involves how aspects of democratic citizenship are affected by the context and content of the political environment in which one lives. Before, however, moving on to the theoretical and empirical policy feedback framework guiding this study of welfare states and democratic citizens, it is useful to first provide an overview of what exactly is meant by a welfare state, what its origins are, as well as some of the normative ideas it embodies.

References Acemoglu, Daron, and James A.  Robinson. 2006. Economic Origins of Dictatorship and Democracy. Cambridge: Cambridge University Press. Alber, Jens, and Ulrich Kohler. 2009. The Inequality of Electoral Participation in Europe and America and the Politically Integrative Functions of the Welfare State. In United in Diversity? Comparing Social Models in Europe and America, ed. Jens Alber and Neil Gilbert, 63–90. International Policy Exchange Series. New York and Oxford: Oxford University Press. Almond, Gabriel A., and Sidney Verba. 1963. The Civic Culture: Political Attitudes and Democracy in Five Nations. Newbury Park, CA: Sage Publications. American Political Science Association Task Force. 2004. American Democracy in an Age of Rising Inequality. Perspectives on Politics 2 (4): 651–666. https:// doi.org/10.1017/S153759270404040X.

  Democracies and Their Citizens 

  25

Anderson, Christopher J., and Pablo Beramendi. 2008. Income, Inequality, and Electoral Participation. In Democracy, Inequality, and Representation: A Comparative Perspective, ed. Pablo Beramendi and Christopher J. Anderson, 278–311. New York: Russell Sage Foundation. Barber, Benjamin R. 1984. Strong Democracy: Participatory Politics for a New Age. Berkeley: University of California Press. Beramendi, Pablo, and Christopher J. Anderson. 2008. Income Inequality and Democratic Representation. In Democracy, Inequality, and Representation: A Comparative Perspective, ed. Pablo Beramendi and Christopher J. Anderson, 3–24. New York: Russell Sage Foundation. Berinsky, Adam J., and Gabriel S.  Lenz. 2011. Education and Political Participation: Exploring the Causal Link. Political Behavior 33 (3): 357–373. https://doi.org/10.1007/s11109-010-9134-9. Boix, Carles. 2003. Democracy and Redistribution. Cambridge, UK and New York: Cambridge University Press. Brady, Henry E., Sidney Verba, and Kay Lehman Scholzman. 1995. Beyond SES: A Resource Model of Political Participation. American Political Science Review 89 (2): 271–294. Bühlmann, Marc, Wolfgang Merkel, Bernhard Wessels, and Lisa Müller. 2007. The Quality of Democracy: Democracy Barometer for Established Democracies. NCCR Working Paper (10). Accessed March 18, 2018. http:// opus.kobv.de/hsog/volltexte/2010/121/. Dagger, Richard. 2008. Republican Citizenship. In Handbook of Citizenship Studies, ed. Engin F. Isin and Bryan S. Turner. Repr., 145–157. Los Angeles, CA: Sage. Dahl, Robert A. 1971. Polyarchy: Participation and Opposition. New Haven, CT: Yale University Press. ———. 2006. On Political Equality. New Haven, CT: Yale University Press. Diamond, Larry. 2010. Why Are There No Arab Democracies. Journal of Democracy 21 (1): 93–104. Diamond, Larry, and Leonardo Morlino, eds. 2005a. Assessing the Quality of Democracy. Baltimore, MD: Johns Hopkins University Press. ———. 2005b. Introduction. In Assessing the Quality of Democracy, ed. Larry Diamond and Leonardo Morlino, ix–xliii. Baltimore, MD: Johns Hopkins University Press. Elster, Jon. 1982. The Case for Methodological Individualism. Theory and Society 11 (4): 453–482. http://www.jstor.org/stable/657101. Ferejohn, John. 2009. Is Inequality a Threat to Democracy? In The Unsustainable American State, ed. Lawrence R.  Jacobs and Desmond S.  King, 34–57. Oxford and New York: Oxford University Press.

26 

J. Shore

Freedom House. 2018. Freedom in the World 2018. Washington, DC. Accessed February 27, 2018. http://www.freedomhouse.org/reports. Galston, William A. 2001. Political Knowledge, Political Engagement, and Civic Education. Annual Review of Political Science 4: 217–234. Hayes, Bernadette C., and Clive S. Bean. 1993. Political Efficacy: A Comparative Study of the United States, West Germany, Great Britain and Australia. European Journal of Political Research 23 (3): 261–280. Hernes, Helga M. 1988. Scandinavian Citizenship. Acta Sociologica 31 (3): 199–215. Iversen, Torben. 2008. Capitalism and Democracy. In The Oxford Handbook of Political Economy, ed. Barry R. Weingast and Donald A. Wittman, 601–623. Oxford Handbooks of Political Science. Oxford: Oxford University Press. Janowitz, Morris. 1991. On Social Organization and Social Control. Chicago: University of Chicago Press. ———. 1994. Observations on the Sociology of Citizenship. In Citizenship: Critical Concepts, ed. Bryan S. Turner and Peter Hamilton, 42–62. London: Routledge. King, Desmond, and Jeremy Waldron. 1988. Citizenship, Social Citizenship and the Defence of Welfare Provision. British Journal of Political Science 18 (4): 415–443. Klingemann, Hans-Dieter. 1999. Mapping Political Support in the 1990s: A Global Analysis. In Critical Citizens: Global Support for Democratic Governance, ed. Pippa Norris, 31–56. Oxford: Oxford University Press. Kymlicka, Will. 2002. Contemporary Political Philosophy: An Introduction. Oxford: Oxford University Press. Lane, Robert Edwards. 1959. Political Life: Why People Get Involved in Politics. Glencoe, IL: The Free Press. Lijphart, Arend. 1997. Unequal Participation: Democracy’s Unresolved Dilemma—Presidential Address, American Political Science Association, 1996. American Political Science Review 91 (1): 1–14. ———. 2001. Democracy in the 21st Century: Can We Be Optimistic? European Review of Social Psychology 9 (2): 169–184. https://doi.org/10.1017/ S1062798701000163. Lipset, Seymour Martin. 1960. Political Man: The Social Bases of Politics. Garden City, NY: Anchor. Maslow, Abraham H. 1943. A Theory of Human Motivation. Psychological Review 50 (4): 370–396. Mill, John Stuart. [1861] 1975. Three Essays: On Liberty, Representative Government, the Subjection of Women. With the Assistance of R. (e.) Wollheim. Oxford: Oxford University Press.

  Democracies and Their Citizens 

  27

Nie, Norman H., Jane Junn, and Kenneth Stehlik-Barry. 1996. Education and Democratic Citizenship in America. Chicago: University of Chicago Press. Norris, Pippa, ed. 1999. Critical Citizens: Global Support for Democratic Governance. Oxford: Oxford University Press. Pateman, Carole. 2000. Participation and Democratic Theory. Reprint. Cambridge: Cambridge University Press. Persson, Mikael. 2010. An Empirical Test of the Relative Education Model in Sweden. Political Behavior 33 (3): 455–478. https://doi.org/10.1007/ s11109-010-9138-5. Ringen, Stein. 2007. What Democracy Is For: On Freedom and Moral Government. Princeton, NJ: Princeton University Press. Rueschemeyer, Dietrich. 2004. Addressing Inequality. Journal of Democracy 15: 76–90. Saunders, Trevor J. 2000. Politics: Books I and II. Reprinted. Oxford: Clarendon Press. Schumpeter, Joseph Alois. 1942. Socialism, Capitalism and Democracy. 3rd ed. New York: Harper and Bros. Shklar, Judith N. 1991. American Citizenship: The Quest for Inclusion. Cambridge, MA: Harvard University Press. Shore, Jennifer. 2014. How Welfare States Shape Participatory Patterns. In How Welfare States Shape the Democratic Public: Policy Feedback, Participation, Voting and Attitudes, ed. Staffan Kumlin and Isabelle Stadelmann-Steffen, 41–62. Cheltenham: Edward Elgar. Simpson Bueker, Catherine. 2009. The Limits of Political Citizenship. Society 46 (5): 423–428. Solt, Frederick. 2008. Economic Inequality and Democratic Political Engagement. American Journal of Political Science 52 (1): 48–60. Stadelmann-Steffen, Isabelle. 2012. A Remedy for Unequal Participation? How Welfare States Impact on Social and Political Engagement. In Professionalization and Individualized Collective Action: Analyzing New ‘Participatory’ Dimensions in Civil Society, ed. William Maloney and Jan W. van Deth, 139–156. Milton Park and New York: Routledge. Verba, Sidney, and Norman H.  Nie. 1972. Participation in America: Political Democracy and Social Equality. New York: Harper and Row. Verba, Sidney, Kay Lehman Scholzman, and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge: Harvard University Press. Wolfinger, Raymond E., and Steven J.  Rosenstone. 1980. Who Votes. New Haven, CT: Yale University Press.

3 Why Context Matters: The Role of Public Policy

The welfare state, because of its very existence, influences policies and politics. While many have studied welfare states in terms of their determinants or policy outcomes, a rapidly growing subfield of comparative politics is focusing increasingly on the ways in which welfare state institutions affect individual attitudes and behaviors (Crepaz and Damron 2009; Rapp et  al. 2017; Shore 2014; Kumlin and Stadelmann-Steffen 2014; Gingrich 2014; Kumlin 2002, 2004; Svallfors 2007). In this chapter, I examine how the welfare state can shape attitudes and behaviors. Starting with new institutionalist perspectives and touching theoretical accounts of the role of the state, the interconnectedness of rights is discussed with regard to the ways in which the welfare state can hinder or encourage political engagement. Drawing on the policy feedback approach, the chapter concludes with a more targeted account of how the welfare state can shape democratic citizenship.

© The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2_3

29

30 

J. Shore

New Institutionalist Approaches Do institutions matter? Not only is this question included in the titles of numerous studies employing institutionalist approaches, this question represents one of the more persistent quandaries in a variety of disciplines. Classical institutionalism has provided us with a preliminary answer to this question. As most would agree that, yes, institutions do matter, the question to follow must ask to what extent do they matter? While political institutions are a major focus in the field of political science, the term institution is rather ambiguous. Most agree that institutions embody the “rules of the game.” What one defines as rules, however, provides yet another source of ambiguity. The most common dichotomy is the distinction between formal and informal rules. Informal rules are less implicit and encompass non-formalized habits, routines, or procedures that have become the norm or custom in a given society (Rothstein 1998b; March and Olsen 1989). Formal rules, with regard to political institutions, are just that: “formal arrangements for aggregating individuals and regulating their behavior through the use of explicit rules and decision processes enforced by an actor or set of actors formally recognized as possessing such power” (Levi 1991, 405). Viewing political phenomena through the lens of institutions is by no means a recent approach. As Steinmo (2008, 118) points out: “Institutional theory is as old as the study of politics.” The works of the ancient Greek philosophers, such as Plato and Aristotle, are among the first documentations of how institutions affect political behavior, incentives, and normative values. The idea that political institutions are capable of shaping individual behaviors and attitudes finds its roots in new institutionalism; however, like its name implies, there is also a classical institutionalism. While the old institutionalism focused primarily on formal rules and constitutions, new institutionalism represents an explicitly empirical approach to analyzing the effectiveness and regulative power of political institutions. Classical institutionalism takes a narrower spectrum of varieties of democracy into consideration and frequently employs more or less static typologies in their classifications. Moreover, the mode of explanation is a structural one, looking at features of the institutional setting such as “size, demography, socio-economic development, cultural

  Why Context Matters: The Role of Public Policy 

  31

homogeneity, heritage from a former colonial empire or the place of a country in the international political and/or economic system” (Kaiser 1997, 422). New institutionalism, on the other hand, centers its explanations on actor-based models. With new institutionalism, we see a shift from the classical view of institutions as a reflection of structural conditions to a focus on actor-related explanations. New institutionalism begins with the assumption that “preferences can only be understood in the context of the institutionally generated incentives and institutionally available options that structure choice” (Grofman 1989, 1).1 How does the welfare state fit into this discussion of institutions? While the welfare state does not fall into the category of more traditional or primary political institutions (such as the electoral or party system, bicameralism or presidentialism), it can be considered as a secondary institution.2 Piven and Cloward (1977) point out that the welfare state as a secondary institution must serve the economy, the political system, as well as promote political legitimacy. Crepaz (2008, 137) notes that such institutions are the “result of political movements that shaped particular formal constitutional structures.” Once in place, the welfare state is capable of becoming an “instrument of establishing and constructing identities” (ibid.). The welfare state thus is able to influence individual attitudes and behaviors much like the primary institutional setting. In sum, the new institutionalist approach posits that the interactions one has with his or her social surroundings can have a substantial impact on individual choices, behaviors, and attitudes. Institutions encourage some behaviors by means of incentive mechanisms and discourage others through sanctions (Hall and Taylor 1996). With the new institutionalist approach, there is the inherent assumption that macro-level characteristics (institutional design) influence micro-level phenomena (individuals’ actions and attitudes)—that is, the political-institutional  The contemporary social sciences generally differentiate between three new institutionalisms: rational choice, sociological, and historical institutionalism (Hall and Taylor 1996). Referring specifically to historical institutionalism, although the same can be said of its other institutionalist counterparts, Steinmo (2008, 118) remarks that it is “neither a particular theory nor a specific method. It is best understood as an approach to studying politics.” 2  The term secondary institution should not be confused with Wilensky and Lebeaux’ (1958) conception of the term: their description of the welfare state as a secondary institution refers to its role relative to the family and market. 1

32 

J. Shore

context functions such that it structures individual attitudes and behaviors (Anderson and Singer 2008). Immergut (1998, 26), however, points out an important caveat that should be kept in mind when referring to the institutionalist approach: “Institutions do not determine behavior, they simply provide a context for action that helps us to understand why actors make the choices that they do.” In other words, when studying the effects of institutions, we might, for example, examine the ways in which they may intervene or shape or moderate relationships at the micro-level, such as the well-documented links between an individual’s resources and political behaviors and attitudes.

What Should a Welfare State Do? The type of effect the welfare state will have on democratic citizenship has much to do with how one conceptualizes the role of democracy and fairness. A generous, universal welfare state is thought to safeguard the rights of social citizenship by guaranteeing social benefits to all. The less extensive, minimalist welfare state represents a form of social insurance and grants assistance based on means testing. Rothstein (1998a) provides a detailed account of the competing views on the ways in which the universal welfare state affects individual autonomy. Put briefly, the positive view of the welfare state sees it as solidaristic and just, as it espouses equal treatment on the basis of equal rights and norms. Where coverage is broader, more needs are able to be met. Furthermore, the state is seen to foster individual autonomy, as it enables people to live their lives in the manner they see fit. The other view, however, characterizes the advanced welfare state as a sort of a “new Leviathan, a despotic master which orders citizens about at will and severely limits the prospects for individual freedom and self-determination” (Rothstein 1998a, 30). The welfare state, according to its critics, promotes an institutionalized helplessness through an unhealthy reliance of the citizens on social programs. Moreover, larger welfare states have higher levels of spending and, therefore, must tax their citizens at a high rate—something which some critics view as inhibitive to individual autonomy. Both of these views ultimately arise from the debate over what the proper role of the state is vis-à-vis its citizens.

  Why Context Matters: The Role of Public Policy 

  33

The liberal response is that the state must remain neutral in order to safeguard individual neutrality. Dworkin (1977, 180), drawing on the Rawlsian theory of justice,3 believes that the state must treat people with “equal concern and respect.” By this, he means that the state cannot impinge on one’s freedom to choose and pursue his or her life as he or she sees fit. Doing so would mean that the state is imposing its judgment or a set of moral values on the individual. Ultimately, the state would regard some “life projects” (Rothstein 1998a) as superior to others. This sentiment is reiterated by Kymlicka (1990, 207): “State neutrality is required to respect people’s self-determination.” Political liberalism champions the individual’s right to choose his or her conception of the “good life.” The counter-argument to this view, labeled by some as communitarianism (Rothstein 1998a), posits that a universal welfare state is also committed to equality but does not imply equal treatment. For example, some may require more medical treatment than others or choose to have more children and therefore make greater use of the public education system. Simply because someone receives social assistance does not mean his or her life’s plan is superior to another’s. If the needy individual did not receive the assistance he or she required in order to pursue his or her conception of the good life, this suffering would outweigh any sort of negative feelings that someone not in need of assistance would experience. The liberal view is also very wary of state intervention and protects people’s right to make mistakes, as state intervention would again imply the imposition of morals onto the people. The communitarian view, on the other hand, promotes a positive perspective of certain collective moral values and, consequently, does view some individual pursuits as superior to others. Furthermore, it is the state’s duty to further certain social norms through its actions and policies. Rothstein (1998a, 36) argues that the ideals of liberalism and communitarianism need not oppose one another: “[I]f the state is to remain neutral in respect to citizens’ choice of life project, it must also create equivalent opportunities for citizens to realize these varying choices.” Rothstein dismisses the notion that individual rights block communitarian projects aimed at solidarity and equality: 3  Stated very simply, John Rawls’ theory of justice (2003 [1971]) conceives of justice as fairness. In order for individual liberty and equality to be preserved, equal opportunities for all are necessary.

34 

J. Shore

“The political content of the concept of rights is determined […] by what kinds of rights the state guarantees to the individual” (ibid., 37). In other words, in order to safeguard individual rights for all, the state must take a stand, such as with the freedom of speech and expression, so that the resources with which one can realize his or her rights must be secured; otherwise, the risk arises that some individual goals are deemed to be better than others. A revised view of the communitarian perspective would then posit that while a state must represent a given set of values, they need not be based on notions of moral supremacy. Sandel (2008), for example, views the liberalist critique of communitarianism as misguided, as even in the liberal society, communities and collectivities play a role in crafting the individual personality. “If we accept the proposition that our individual personalities are created through the particular social and political contexts of which we are a part, then the political order upon which we should agree reflect a particular conception of the good life” (Rothstein 1998a, 39; emphasis in original). Communitarianism, even in its revised form, remains unable to adequately explain pluralist societies: as Rothstein (1998a) points out, simply because a group happens to be in the majority, it does not have the right to impose its values of what is right on others. To say that the state should promote citizen autonomy is not, however, to say that the state is morally indifferent. From the Rawlsian (1993) perspective, political institutions exist to ensure individuals have equal opportunities to pursue and fulfill their lives’ projects (Rothstein 1998a, 43). The Rawlsian goal is thus to devise just institutions that “shape the character and aims of the members of society, the kinds of person they are and want to be” (Rawls 1980, 538; Rothstein 1998a; Rorty 1993).4 In other words, just, democratic institutions create citizens who also value justice and democracy. From a normative stance, the welfare state is thus thought to be capable of imparting its citizens with the tools necessary to make well-informed decisions about their life choices—it does so  Rawls’ view has come under quite a bit of criticism, mainly by Amartya Sen (1982), who argues that even in a state dedicated to the provision of equal opportunities (i.e., justice), inequality (injustice) may still persist. An equal distribution of goods does not necessarily lead to equal opportunities, for there are some who are initially more disadvantaged than others and therefore require more in order to pursue his or her life project. Sen thus argues that the state must consider one’s basic capabilities as the basis for the provision of opportunities. 4

  Why Context Matters: The Role of Public Policy 

  35

by ensuring that all, regardless of their capabilities, have equal access to opportunity, certain basic resources, and, therefore, equal social rights. This conception of the good life is not limited only to the things that people want. In addition to the good life, people want their governments to do something—to protect from dangers and risks and to be attentive to grievances, all the while avoiding paternalism. People are more than their economic resources and generally desire to live in a “society where everyone has a reasonable chance to prosper and no one is consigned to a life of poverty” (Ferejohn 2009, 38). In the 1942 Beveridge Report, a comprehensive reform of social insurance was outlined which would provide coverage from the “cradle to the grave.” Its author, Sir William Beveridge, has consequently become known as the father of universality (Jacobs 2008). Universality, it is politically argued, provides the only secure basis for social protection, for all people included under its coverage and therefore all have “an interest in supporting the system” (Ringen 2007, 74). Today, over 70  years later, citizens most advanced democracies are no longer at such a great risk of becoming poor. The common men and women have comparatively many more resources—income, property, education—and therefore no longer require the same types of protections offered by the paternalistic welfare states that emerged in the post-war era. In most advanced democracies, people are much more capable of self-autonomy than they were during the mid-twentieth century. While we find ourselves today in an era of comparative affluence, this affluence has not been extended to all members. In a 1963 presidential speech, John F. Kennedy stated that a “rising tide lifts all boats.” What we have witnessed, however, is that a rising tide lifts some boats, while capsizing others. Although Beveridge dreamed that with the rise of affluence and universal welfare schemes poverty would be eradicated, this clearly has not occurred: “Advances in affluence to benefit the mainstream may simultaneously cause more people to fall outside of the mainstream” (Ringen 2007, 77). Not all are able to keep pace with the rapidly advancing knowledge-based economy; the need for social care has clearly not been eliminated. The problem with poverty is not just that the impoverished are without sufficient material power, but that this lack of economic resources so often translates into a lack of ­freedom to live one’s life as one wishes. Material resources are what

36 

J. Shore

enable individuals to “put power behind their demands” (ibid., 192). Furthermore, the state, should it systematically abandon its most needy members, serves to effectively exclude them politically as well.

The Interconnectedness of Rights Examining the role and effects of welfare states serves to bridge the debates on citizenship and the origins of democratic citizenship and its inequalities. Put simply, the general purpose of a welfare state is to reduce social risk and economic inequality. But does the reduction of economic inequality correspond to the reduction of other inequalities, such as inequalities of political participation? Furthermore, does a state that delivers more to its citizens in terms of generous social policy create more politically engaged citizens? Before looking into these questions, I briefly introduce the concept of the welfare state as well as its origins in order to provide a better understanding of why different welfare states may play a role in producing different sorts of democratic citizenship outcomes. While the modern welfare state began to take shape shortly after World War II, its origins can be traced back to much earlier times (Briggs 1969). For example, in his 1791 essay Rights of Man, Thomas Paine examined the linkages between poverty, criminality, and unemployment. He recognized the distorting effects of inequality and property and offers up a draft of sorts for a system of welfare (Pierson and Castles 2006, 6). An oft-cited example of a forerunner to contemporary welfare policies are the English Poor Laws. Stemming from late medieval and Tudor laws, the Poor Laws were a means of proving relief to the poorest members of English and Welsh society. It was not until 1948 that these laws were abolished with the passing of the National Assistance Act (Marshall 1950). These laws represented a recognition of inequality in society as well as an attempt to remedy it, or, at the very least, to assist those whose fates had been the most downtrodden. While these laws may have appeared to be altruistic, those who had no choice but to go to the poorhouses were stripped of their rights of citizenship. T.H. Marshall was highly critical of this divorce of social from political rights: “The Poor Law treated the claims of the poor, not as an integral

  Why Context Matters: The Role of Public Policy 

  37

part of the rights of the citizen, but as an alternative to them—as claims which would be met only if the claimants ceased to be citizens in any true sense of the word” (Marshall 1950, in Pierson and Castles 2006, 33). In Citizenship and Social Class (1950), Marshall presented a positive appraisal of the welfare state as a reaction to modern economic development. Alongside the civil rights (individual freedoms, liberty, justice) and political rights (rights to participate) that citizens possess, he introduced a third and equally important class of rights: social rights. Social rights encompass economic welfare and security as well as the “right to share to the full in the social heritage and to life the life of a civilized being according to the standards prevailing in the society” (Marshall 1950, in Pierson and Castles 2006, 30). Marshall saw state institutions, specifically, education and social services, as the best way of securing these social rights. In their safeguarding of social rights, such institutions have a socially integrative function and are capable of fostering a sense of solidarity and community. The objective of welfare institutions is, in the Marshallian view, not to equalize incomes but rather to reduce risks and insecurities and to perform some degree of equalization across more broader dimensions of disparity—for example, the healthy versus the sick, employed versus unemployed, or the fortunate versus the less fortunate. Social rights are indeed integral for exercising other types of rights, including political: “Without extending social rights to all, civil and political citizenship as we conceive of it today cannot be fully realized” (Simpson Bueker 2009, 423). Although Marshall (1950) outlined different spheres of citizenship, demonstrating that first civil rights were obtained, then political, followed lastly by social rights, he was highly critical of this sort of sequencing, for not all members of society would be able to take full advantage of all dimensions of citizenship. In other words, in the absence of rights of social citizenship, it is very difficult to take advantage of one’s political rights: “economic poverty translates into political poverty” (Simpson Bueker 2009, 426). For this reason, welfare state institutions are often seen as necessary in order to extend social citizenship rights to all members of a society. Not all welfare states, however, are equally capable of securing the rights of social citizenship. For example, while the social democratic model (found in the Scandinavian states) leads the way in reducing poverty and

38 

J. Shore

“­ provides the fullest development of social rights of citizenship” (Roche 2002, 77), liberal welfare states may in fact increase some inequalities by extolling “free markets, which reward values and skills that are unequally distributed in the population” (Schuck 2008, 140). These enduring inequalities can have a delegitimating effect for states, an effect which is multiplied when one considers that the most disadvantaged are also those who tend to be most disengaged from the political sphere. Before moving on to a more general discussion about the origins and effects of modern welfare states, it is important to once again underscore the interconnectedness of the different spheres of citizenship and how an inequity in one area is inherently linked to inequities in others: Political citizenship influences civil citizenship. Social citizenship influences civil and political citizenship. When individuals and groups lack one of these elements of citizenship, they essentially lack all three. Social citizenship, rather than being the third phase of membership, needs to come prior to or alongside political citizenship, otherwise political citizenship will be nothing more than a false promise. (Simpson Bueker 2009, 427)

What we in effect witness are numerous spillovers in terms of inequalities—that is, they do not remain confined to single areas of rights. The question that now arises is whether policies that are designed to fight economic inequality have a parallel spillover effect in terms of political inequality. Is a state that is responsive and secures the social rights of its citizens also one that is capable of rendering them, as Robert Dahl (1971) stated, “political equals?”

 he Impact of Public Policy: The Policy T Feedback Approach Beyond influences of individual resources, personality characteristics, or political socialization, we must also consider the role played by public policy in shaping democratic engagement. This line of thought, known as the policy feedback approach, can be traced to E.E.  Schattschneider’s (1935) hypothesis that “new policies create new politics.” In a later work analyzing the withdrawal of the American public from politics,

  Why Context Matters: The Role of Public Policy 

  39

Schattschneider (1960) confronted the prevailing sentiment that an apathetic and uninformed citizenry was to blame for its lack of participation, arguing that the reasons why the lower classes abstained from politics were not due to any sort of moral or civic deficiencies; rather, the political system itself was to blame, for it did not offer much to the public to mobilize, engage, or even encourage political participation (Mettler and Soss 2004). Other early works documenting policy feedback effects include the studies by Lowi (1964) and Wilson (1973), who drew on policy typologies to argue that political mobilization outputs were dependent on different types of policies. What Schattschneider, Lowi, Wilson, and their contemporaries emphasize is that policies can impact their recipients much like formal political institutions do—they too transmit norms and impose rules and regulations—and can thereby transform politics itself (Mettler 2002). When examining the links between citizens and how they think about and behave toward their states, we must recall what it is that actually affects their behavior. Although we know that political institutions structure and provide the rules of the game, it is social policies with which people often come into regular contact and have personal experiences. Citizens are all too frequently conceived of as “background actors in politics” (Mettler and Soss 2004, 55); sociological, psychological, and economic approaches to political behaviors and attitudes frequently neglect the influence policies can wield. From the policy feedback perspective, however, political attitudes and actions are conceived of as the result of the interaction between “state structures and institutions, political actions, communication flows, mobilization and demobilization, and the density and the patterning of political organizations” (ibid., 58). In this regard, political behavior can be seen as a form of policy feedback. Theda Skocpol (1992, 58) underscores two types of policy feedback effects: first, new policies can “transform or expand the capacities of the state”; and, second, “new policies affect the social identities, goals, and capabilities of groups.” In terms of the first set of feedback effects, Skocpol’s Protecting Soldiers and Mothers (1992) demonstrated how the development of extensive public pensions in the United States was stunted due to the corrupt pension system for Civil War veterans. Paul Pierson’s Dismantling the Welfare State? (1994) showed how the “design of

40 

J. Shore

the American Social Security System creates a barrier to privatized individual accounts because of the ‘double payment’ problem: today’s payroll tax receipts fund today’s benefits, and so there is no extra money to put aside in individual accounts to accrue for the future” (Campbell 2012, 335). Most of the research in the field of public policy focuses on this first type of effect (Pierson 1993); how public policies affect people’s attitudes and behaviors remains “remarkably undertheorized” (Mettler and Soss 2004, 60). By examining public policy’s effect on citizens’ political behavior and attitudes, two schools within the discipline of political science come together: public policy and political behavior. Not until recently have these two silos of the field been studied under a single research agenda. Scholars interested in political behavior tended to see public policy as something far off; something that eventually could be influenced by political action. Mettler and Soss (2004, 55) describe the division of public policy and political behavior as “opposite ends of the political process.” Campbell (2012) notes Schneider and Ingram’s (1993) Policy Design for Democracy as one of the first works to cross the once well cordoned off areas in political science. In their book, the authors argue that policy design wields substantial influence in terms of the “social construction of target populations.” Pierson (2006, 115) argues that we can study the effects of policies much like we study institutional effects: “policies are important instances of political institutions.” Policies, however, have received much less attention than the institutions from which they originate due mainly to their vast heterogeneity and their ephemeral character (unlike institutions, they are much more prone to change). There are important reasons to study policies as institutions, the first of which being their importance for individuals. Although some political scientists might try and tell us otherwise, policies are what people come into contact with on a daily basis— not the formal institutions: “If policies as institutions matter for political scientists, it is because the influence of policies on social actors—on who they are, on what they want, on how and with whom they organize—is such that it changes the way these actors engage in politics” (Pierson 2006, 116). At the macro-level, the policy regime literature has done much to examine policy effects—with one of the most prominent works

  Why Context Matters: The Role of Public Policy 

  41

being Esping-Andersen’s The Three Worlds of Welfare Capitalism (1990). At the micro-level, however, we are just beginning to examine the influence of policies on people’s behaviors and attitudes. Most of the works here have been limited to group-specific studies for the United States; the lack of comparative analyses incorporating policy feedback approaches on individual behavior indeed represents a palpable gap in the literature.

Effects of Policies We know that citizens’ participation can influence the outcomes of politics, but do the policy outcomes themselves have an impact on participation and the way people think about politics? Campbell (2012) argues they do: policies can have both negative and positive effects on participation: first, and most obviously, they may distribute (and redistribute) the resources needed for political involvement that might otherwise be lacking. Depending on the actual size of the benefits, political participation may be fostered due to an increase in the recipient’s capacity to participate (in terms of time and money). For example, a policy aimed at a certain group may supply that group with resources it needs for mobilization. Campbell (2011) finds that low-income pensioners—those most likely to receive social security benefits—in the United States are frequently voters and policy-based activists. This group of senior citizens was not always an active group: “In the 1950s, when Social Security benefits were modest and covered only a fraction of seniors, the elderly participation at lower rates than younger people” (Campbell  2011, 2). As mentioned, there have been numerous studies that look at the resource effect from the other side of public policies—that is, from the standpoint of policy outcomes, such as inequality indicators. Redistribution and resources, however, only tell us half the story. Public policies can do many things beyond redistribution. Closely related to the resource effect, policies can also motivate people to become interested in politics due to the personal stakes that policies can create for people’s lives. Policies may pique interest in certain issues or politics in general. For example, if a policy targets a program that directly affects the individual, he or she may be more likely to engage in political

42 

J. Shore

action due to one’s own personal stakes. An oft-cited example is the high voter turnout rates of farmers in the United States: they have a personal interest in the outcomes of agricultural policies and farm subsidies. Additionally, policies may create opportunities for people to learn civic skills (Marston 1993; Soss 1999) or foster one’s sense of obligation or responsibility to the state (Mettler 2002). Third, policies can have numerous cognitive effects on citizens (Pierson 1993). Policies may signal to the citizenry that they have rights to certain benefits, goods, or services; these policies also may foster the expectation that the government should, for example, address aspects of inequality (Mettler and Soss 2004; Piven and Cloward 1983). The messages policies send to people can relay information about where they stand in their community (Schneider and Ingram 1997) or “whether the government is responsive to their concerns, thereby encouraging engagement, passivity, or even alienation” (Wichowsky and Moynihan 2008, 909; see also Mettler 2002; Soss 1999). More concretely, the ways policies are designed influence people’s personal and everyday experiences with their governments, which in turn shapes the way they think and feel about them. For example, the responsiveness of the bureaucratic routines to the people’s demands shape how they come to view their government institutions (Kumlin 2004; Kumlin and Rothstein 2005; Moynihan and Soss 2014). Policy design sends certain messages to the citizenry: “By shaping citizens’ encounters with government, the design and implementation of public policy constitute important forces shaping citizens’ orientations toward the institutions and policies of government” (Mettler and Soss 2004, 62; see also Schneider and Ingram 1997). Examples from the United States reveal that “welfare” recipients (social assistance programs: Aid to Families with Dependent Children/ Temporary Assistance for Needy Families) feel stigmatized and socially isolated through the processes of means testing. These negative and degrading experiences lead to lower political efficacy among recipients (Soss 1999). The universal social security program in the United States, however, has the opposite effect: recipients report greater levels of political efficacy, as recipients are viewed as deserving of benefits with legitimate claims (Campbell 2011). In other words, how policies are designed and implemented can affect perceptions of both deservingness and societal stigma.

  Why Context Matters: The Role of Public Policy 

  43

Banting (1999) notes that the welfare state can also play a regionally and nationally integrative role. Using the example of the incorporation of Newfoundland into Canada in 1949, he describes how the extensive Canadian welfare system was one of the most attractive selling points for joining the Canadian state. Wanting to make their newest citizens feel welcome, the Canadian government assured that the Newfoundlanders received social benefits immediately after their incorporation. By granting social rights right away and effectively demonstrating that these new members of its society were on an equal level with the rest of the provinces, the Canadian welfare state was ultimately able to “make Canadians” (see Crepaz 2008, 138). This highly effective response by the Canadian government was able to integrate ethnically diverse groups. The most-­ cited and notorious example of the integrative power of the welfare state, however, can be found much further back in history: Otto von Bismarck’s introduction of social insurance during the 1880s not only served to quash the revolutionary potential of the socialists but also tied the “loyalties of the individual directly to the monarchy or central state authority” (Esping-Andersen 1990, 24). The class differences that were softened with the development of the welfare state was only one of the results of Bismarck’s actions; by centralizing the command over social insurance, the diverse ethnic groups comprising the German Reich were able to be brought together through their loyalty to the empire (Kuhnle and Sander 2010). Not only are welfare states designed to create cross-class solidarities, but across territories as well, integrating regionally or ethnically ­distinct territories into the borders of the nation state (see e.g., Moreno and McEwen 2005). As people who live in states with comprehensive and far-reaching welfare systems come into frequent direct contact with these welfare policies, they are bound to affect their perceptions. These perceptions are largely affected by what is known in the field of psychology as procedural justice. “This research has shown that people are concerned not only with the final results of personal contacts with public institutions. (e.g., ‘Did I get the allowance?’ ‘Was I convicted in court?’ ‘Am I satisfied with day care?’). They are frequently at least as interested in whether the process that eventually led to the final result was fair” (Kumlin and Rothstein 2005, 347). These questions raised by Kumlin and Rothstein hearken back to Dworkin’s conception of justice—that is, that people, when coming into

44 

J. Shore

contact with government institutions, must be treated with “equal concern and respect” (Dworkin 1977, 180). If people perceive that they are indeed treated with equal concern and respect, how does this affect their perceptions of their governments? Kumlin (2004) and Soss (1999), for example, find that a positive appraisal of procedural justice in terms of individual experiences with welfare state institutions fosters satisfaction with democracy. Testing these assumptions using Swedish survey data, Kumlin (2004) finds that the degree of citizen empowerment institutional arrangements are able to achieve is predictive of individual levels of satisfaction with democracy, trust in politicians, and support for state social programs. Even when controlling for individuals’ socioeconomic backgrounds, he finds that personal experiences with the different welfare state arrangements within Sweden are crucial to understanding people’s “broader political orientations” (ibid., 23). To conclude, there are many ways in which public policies may potentially impact people’s political attitudes and behaviors. Moreover, policies, depending on design or administration, have the potential to promote political equality or exacerbate exisiting inequalities (see e.g., Shore 2016). Most of the work drawing on policy feedback approaches, however, has been conducted for smaller groups within the United States. We lack studies of policy feedback that examine the effects of social policies on mass publics and ones that do so comparatively. Moreover, although we have plenty of evidence to demonstrate that income i­nequality and political inequality are fundamentally linked, the precise mechanisms into these relationships and how social policy may intervene have received less attention.

L inking the Welfare State and the Democratic Citizen From the literature, we know that political participation and democratic support are largely a function of political efficacy—the feeling that one can capably and competently take part in politics, and that representatives and authorities are responsive to citizen demands and preferences. We also know that political efficacy is largely a function of resources, with better-educated, well-connected, and higher-earning individuals more likely to feel politically confident and represented in many societies. This is a problem for democracy, for democracy not only requires citizen

  Why Context Matters: The Role of Public Policy 

  45

participation but must allow for equal citizen participation in order to be truly democratic. And although democracy is an increasingly common and sought-after form of government, trends of rising income inequality stand to threaten its legitimacy as such. Economic inequality, as numerous studies tell us, distorts political equality, and while we know that why people choose to participate or withdraw from politics is not always simply a function of one’s socioeconomic resources, a very unequal distribution of these resources is what structures the patterns of political behaviors and attitudes in many of the world’s most advanced democracies. Social policies, however, have the potential to assuage some of the incongruencies that have arisen out of the relationship between democracy and capitalism—both of which espouse equality of opportunity. Over the past decade or so, an interest in the feedback effects of public policies has emerged within political science scholarship, with most of the work conducted in the United States. T.H. Marshall (1965) could be considered a forerunner to contemporary policy feedback theorists: in his evaluation of the English Poor Laws he criticized the divorce of social from political rights. The emerging welfare states, responding to modern economic upheavals, were a means of unifying the different spheres of rights (civil, political, and social). Generous welfare structures and policies are thought to have an integrative effect, promoting the political incorporation of all social groups (Alber and Kohler 2009). It can, therefore, be assumed that welfare states strongly committed to the reduction of economic inequality across socioeconomic groups are also capable of promoting higher and more equitable levels of political participation because, in addition to the resources that they distribute and redistribute, they enhance efficacy and saliency through social integration. Internal efficacy is thought to be positively affected by the distribution and redistribution of resources needed for political participation. Some cash benefit schemes, for example, are able to hoist most people out of poverty, enabling them to engage in political activities, something which may have otherwise remained a luxury. External efficacy is fostered by generous welfare policies embodying norms of fairness and universality, which may convey the message that the government is attentive and responsive to the needs of the many and not just the better off. It can, therefore, be expected that in states where the welfare focus is placed on the reduction of economic inequalities, citizens will be more likely to take part in

46 

J. Shore

­ olitics—including voting in elections and being interested in politics. p People in these types of welfare states are also expected to be more satisfied with the way democracy works and exhibit trust in political institutions. Redistributive policies may very well provide greater incentives to some groups to participate than to others. For people whose socioeconomic resources would predict an already high propensity to participate, the effect of additional welfare state spending on their political behavior would be smaller than the effect for lower-income groups. While it is expected that generous welfare states will have an overall positive direct effect on political participation and attitudes toward democracy, conditional hypotheses regarding indirect effects can also be tested. Conditional hypotheses touch on the ways in which social policy, in addressing income inequality, can moderate political inequality. In contexts where welfare policies are aimed at tackling economic inequality, we can expect a weaker link between resources and democratic citizenship. Drawing on what we know from the empirical work on policy feedback in the US context, generous and universally designed social policies not only redistribute resources to poorer citizens, but they can also provide them with a personal stake in political outcomes—thereby fostering interest in politics—and have interpret effects, that is, the design or focus of a policy may signal to citizens that they are valued and deserving members of a society. Due to the ways in which these kinds of policies are expected to moderate the relationship between income and democratic citizenship, I hypothesize that they may be able to provide the greatest participatory benefits to citizens with the fewest resources, as government offerings may have a greater impact on their lives and well-­ being than those from the upper-income categories. Because of the increased relevance of politics accompanying more generous policies, poorer individuals may also stand to also benefit most from the integrative effects of welfare policies which convey to them that they too are valued and represented members of the polity. Such messages may be pivotal in mobilizing the vote, getting underrepresented groups to feel their voices can be heard in other arenas, sparking an interest in politics, as well as fostering positive attitudes toward democracy and the way it works. Based on the literature on welfare states and social policies and

  Why Context Matters: The Role of Public Policy 

  47

how income and political engagement are related, it is expected that the positive effect of welfare state generosity on democratic political engagement will be most pronounced for people with the fewest resources. These expectations take into account that even in less generous welfare state contexts, wealthier individuals already have a relative resource advantage compared to their less well-off counterparts. In other words, those who stand to benefit the most from increased resources, increased political stakes, and messages regarding universality and fairness are the members of society who have the least. In the chapter to follow, I put these expectations about the welfare state—regarding its direct effects and its indirect ones—to the empirical test in order to address the question whether more extensive welfare states, in addressing income inequality, are also able to produce greater political equality.

References Alber, Jens, and Ulrich Kohler. 2009. The Inequality of Electoral Participation in Europe and America and the Politically Integrative Functions of the Welfare State. In United in Diversity? Comparing Social Models in Europe and America, ed. Jens Alber and Neil Gilbert, 63–90. International Policy Exchange Series. New York and Oxford: Oxford University Press. Anderson, Christopher J., and Matthew M. Singer. 2008. The Sensitive Left and the Impervious Right: Multilevel Models and the Politics of Inequality, Ideology, and Legitimacy in Europe. Comparative Political Studies 41 (4/5): 564–599. Banting, Keith. 1999. Social Citizenship and the Multicultural Welfare State. In Citizenship, Diversity, and Pluralism: Canadian and Comparative Perspectives, ed. Alan C. Cairns, John C. Courtney, Peter Mackinnon, Hans J. Michelmann, and David E. Smith, 108–136. Quebec: McGill-Queen’s University Press. Briggs, Asa. 1969. The Welfare State in Historical Perspective. In The Welfare State, ed. Charles Schottland, 29–45. New York: Harper and Row. Campbell, Andrea Louise. 2011. How Policies Make Citizens: Senior Political Activism and the American Welfare State. Princeton Studies in American Politics. Princeton, NJ: Princeton University Press. ———. 2012. Policy Makes Mass Politics. Annual Review of Political Science 15 (1): 333–351. https://doi.org/10.1146/annurev-polisci-012610-135202.

48 

J. Shore

Crepaz, Markus M.L. 2008. Trust Beyond Borders: Immigration, the Welfare State, and Identity in Modern Societies. Ann Arbor, MI: University of Michigan Press. Crepaz, Markus M.L., and Regan Damron. 2009. Constructing Tolerance: How the Welfare State Shapes Attitudes About Immigrants. Comparative Political Studies 42 (3): 437–463. Dahl, Robert A. 1971. Polyarchy: Participation and Opposition. New Haven, CT: Yale University Press. Dworkin, Ronald. 1977. Taking Rights Seriously. London: Duckworth. Esping-Andersen, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press. Ferejohn, John. 2009. Is Inequality a Threat to Democracy? In The Unsustainable American State, ed. Lawrence R.  Jacobs and Desmond S.  King, 34–57. Oxford and New York: Oxford University Press. Gingrich, Jane. 2014. Visibility, Values, and Voters: The Informational Role of the Welfare State. The Journal of Politics 76 (2): 565–580. https://doi. org/10.1017/S0022381613001540. Grofman, Bernard. 1989. The Federalist Papers and the New Institutionalism: An Overview. In The Federalist Papers and the New Institutionalism, ed. Bernard Grofman and Donald Wittman, 1–6. New York: Agathon. Hall, Peter A., and Rosemary C.A. Taylor. 1996. Political Science and the Three New Institutionalisms. Political Studies 44 (5): 936–957. Immergut, Ellen M. 1998. The Theoretical Core of the New Institutionalism. Politics and Society 26 (1): 5–34. Jacobs, John. 2008. An Introduction to the Beveridge Report. In Understanding Health and Social Care: An Introductory Reader, ed. Julia Johnson and Corinne de Souza, 215–221. Los Angeles, CA and Milton Keynes: Sage and Open University. Kaiser, André. 1997. Types of Democracy: From Classical to New Institutionalism. Journal of Theoretical Politics 9 (4): 419–444. Kuhnle, Stein, and Anne Sander. 2010. The Emergence of the Western Welfare State. In The Oxford Handbook of the Welfare State, ed. Crancis G. Castles, Stephan Leibfried, Jane Lewis, Herbert Obinger, and Christopher Pierson, 61–80. Oxford: Oxford University Press. Kumlin, Staffan. 2002. Institutions-Experiences-Preferences: How Welfare State Design Affects Political Trust and Ideology. In Restructuring the Welfare State: Political Institutions and Policy Change, ed. Bo Rothstein and Sven Steinmo, 20–50. New York: Palgrave Macmillan.

  Why Context Matters: The Role of Public Policy 

  49

———. 2004. The Personal and the Political: How Personal Welfare State Experiences Affect Political Trust and Ideology. Political Evolution and Institutional Change. New York: Palgrave Macmillan. Kumlin, Staffan, and Bo Rothstein. 2005. Making and Breaking Social Capital: The Impact of Welfare-State Institutions. Comparative Political Studies 38 (4): 339–365. Kumlin, Staffan, and Isabelle Stadelmann-Steffen, eds. 2014. How Welfare States Shape the Democratic Public: Policy Feedback, Participation, Voting and Attitudes. Cheltenham: Edward Elgar. Kymlicka, Will. 1990. Contemporary Political Philosophy. Oxford: Clarendon Press. Levi, Margaret. 1991. A Logic of Constitutional Change. In The Limits of Rationality, ed. Karen S. Cook and Margaret Levi, 402–418. Chicago: The University of Chicago Press. Lowi, Theodore J. 1964. American Business, Public Policy, Case Studies, and Political Theory. World Politics 16 (4): 677–715. March, James G., and Johan P.  Olsen. 1989. Rediscovering Institutions: The Organizational Basic of Politics. New York: Free Press. Marshall, Thomas H. 1950. Citizenship and Social Class. London: Pluto. ———. 1965. Social Policy. London: Hutchinson. Marston, Sallie A. 1993. Citizen Action Programs and Participatory Politics in Tuscan. In Public Policy for Democracy, ed. Helen M.  Ingram and Steven R. Smith, 119–135. Washington, DC: Brookings Institution. Mettler, Suzanne. 2002. Bringing the State Back in to Civic Engagement: Policy Feedback Effects of the G.I. Bill for World War II Veterans. American Political Science Review 96 (2): 351–365. Mettler, Suzanne, and Joe Soss. 2004. The Consequences of Public Policy for Democratic Citizenship: Bridging Policy Studies and Mass Politics. Perspectives on Politics 2 (1): 55–73. Moreno, Luis, and Nicola McEwen. 2005. Exploring the Territorial Politics of Welfare. In The Territorial Politics of Welfare, ed. Nicola McEwen and Luis Moreno, 1–40. London: Routledge. Moynihan, Donald P., and Joe Soss. 2014. Policy Feedback and the Politics of Administration. Public Administration Review 74 (3): 320–332. https://doi. org/10.1111/puar.12200. Pierson, Christopher, and Francis G. Castles, eds. 2006. The Welfare State Reader. 2nd ed. Cambridge: Polity Press.

50 

J. Shore

Pierson, Paul. 1993. When Effect Becomes Cause: Policy Feedback and Political Change. World Politics 45 (4): 595–628. ———. 1994. Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. Cambridge Studies in Comparative Politics. Cambridge, UK and New York: Cambridge University Press. ———. 2006. Policies as Institutions. In Rethinking Political Institutions: The Art of the State, ed. Ian Shapiro, Stephen Skowronek, and Daniel Galvin, 114–132. New York: New York University Press. Piven, Frances Fox, and Richard A.  Cloward. 1977. Regulating the Poor: The Functions of Public Welfare. New York: Pantheon Books. ———. 1983. The American Road to Democratic Socialism. Democracy 3 (3): 58–69. Rapp, Carolin, Jennifer Shore, and Jale Tosun. 2017. Not So Risky Business? How Social Policies Shape the Perceived Feasibility of Self-employment. Journal of European Social Policy. https://doi.org/10.1177/09589287177 11973. Rawls, John. 1980. Kantian Constructivism in Moral Theory: The Dewey Lectures. The Journal of Philosophy 77 (9): 515–572. ———. 1993. Political Liberalism. New York: Columbia University Press. ———. 2003 [1971]. A Theory of Justice. Rev. ed., 5th–6th printing. Cambridge, MA: Belknap Press of Harvard University Press. Ringen, Stein. 2007. What Democracy Is For: On Freedom and Moral Government. Princeton, NJ: Princeton University Press. Roche, Maurice. 2002. Social Citizenship: Grounds of Social Change. In Handbook of Citizenship Studies, ed. Engin F.  Isin and Bryan S.  Turner, 69–86. London: Sage Publications. Rorty, Richard. 1993. Postmodernism and Democratic Theory. Philadelphia, PA: Temple University Press. Rothstein, Bo. 1998a. Just Institutions Matter: The Moral and Political Logic of the Universal Welfare State. Cambridge: Cambridge University Press. ———. 1998b. Political Institutions: An Overview. In A New Handbook of Political Science, ed. Robert E.  Goodin and Hans-Dieter Klingemann, 133–167. Oxford: Oxford University Press. Sandel, Michael J. 2008. Liberalism and the Limits of Justice, 2nd ed., 10th printing. Cambridge: Cambridge University Press. Schattschneider, Elmer E. 1935. Politics, Pressures and the Tariff. New  York: Prentice-Hall.

  Why Context Matters: The Role of Public Policy 

  51

———. 1960. The Semi-Sovereign People: A Study of Free Private Enterprise in Pressure Politics, as Shown in the 1929–1930 Revision of the Tariff. New York: Rinehart and Winston. Schneider, Anne, and Helen Ingram. 1993. Social Construction of Target Populations: Implications for Politics and Policy. American Political Science Review 87 (2): 334–347. https://doi.org/10.2307/2939044. ———. 1997. Policy Design for Democracy. Lawrence, KS: University Press of Kansas. Schuck, Peter H. 2008. Liberal Citizenship. In Handbook of Citizenship Studies, ed. Engin F. Isin and Bryan S. Turner, 131–144. London: Sage. Sen, Amartya. 1982. Choice, Welfare and Measurement. Cambridge, MA: MIT Press. Shore, Jennifer. 2014. How Welfare States Shape Participatory Patterns. In How Welfare States Shape the Democratic Public: Policy Feedback, Participation, Voting and Attitudes, ed. Staffan Kumlin and Isabelle Stadelmann-Steffen, 41–62. Cheltenham: Edward Elgar. ———. 2016. Political Inequality: Origins, Consequences, and Ways Ahead. In Understanding Inequality: Social Costs and Benefits, ed. Amanda Machin and Nico Stehr, 247–265. Wiesbaden: Springer VS. Simpson Bueker, Catherine. 2009. The Limits of Political Citizenship. Society 46 (5): 423–428. Skocpol, Theda. 1992. Protecting Soldiers and Mothers: The Political Origins of Social Policy in the United States. Cambridge, MA: Harvard University Press. Soss, Joe. 1999. Lessons of Welfare: Policy Design, Political Learning, and Political Action. American Political Science Review 93 (2): 363–380. Steinmo, Sven. 2008. Historical Institutionalism. In Approaches and Methodologies in the Social Sciences: A Pluralist Perspective, ed. Donatella Della Porta and Michael Keating, 118–138. Cambridge: Cambridge University Press. Svallfors, Stefan, ed. 2007. The Political Sociology of the Welfare State: Institutions, Social Cleavages, and Orientations. Stanford, CA: Stanford University Press. Wichowsky, Amber, and Donald P.  Moynihan. 2008. Measuring How Administration Shapes Citizenship: A Policy Feedback Perspective on Performance Management. Public Administration Review 68 (5): 908–920. https://doi.org/10.1111/j.1540-6210.2008.00931.x. Wilensky, Harold L., and Charles N.  Lebeaux. 1958. Industrial Society and Social Welfare. New York: Free Press. Wilson, James Q. 1973. Political Organizations. New York: Basic Books.

4 Empirical Approaches to the Study of Democratic Citizens and Welfare States

Before turning to the empirical analyses of the relationships between the welfare state and various aspects of democratic citizenship, this chapter provides essential information concerning the setup of the analyses to follow in Chap. 5. In order to investigate how the welfare state influences citizens’ individual political behaviors and attitudes, I employ multilevel models in order to take into account how different macro-level contexts affect the people living within them. Before moving on to the concrete analyses, I first describe the measurement of the welfare state used in the models as well as some of the issues surrounding the not altogether unproblematic task of measuring something so seemingly complex as the welfare state. In the discussion of the common research design framework used in all of the empirical analyses, I begin with a more general discussion of the method of multilevel modeling used in the analyses. In the chapter to follow (Chap. 5), I look at the ways in which the welfare state impacts the decision to vote, political interest, trust in parliament, and satisfaction with democracy. These aspects of democratic citizenship are exactly that—aspects of a rather broad concept. While these analyses by no means represent an exhaustive investigation of democratic citizenship, these © The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2_4

53

54 

J. Shore

individual attitudes and behaviors are arguably important components of democratic societies. Furthermore, these aspects are the ones for which we currently have the most data across a wide and varied sample of countries.

 perationalizing the Welfare State: O Theoretical and Epistemological Debates The main independent variable, the welfare state, is by no means easily operationalized, something which has been the subject of numerous theoretical and epistemological debates. The “dependent variable problem” has become a common theme as well as stumbling block for many studies examining welfare state change (Green-Pedersen 2004; Pierson 1994). As one would infer from its name, the dependent variable problem is about the definition of the phenomenon itself. Green-Pedersen (2004) contends that this problem has more to do with the theoretical conceptualization than the availability of data. In such a situation, one must first ask what should be measured, given the theoretical definition. Drawing on that answer, one can then move on to thinking about what can actually be measured: The first question is, however, more crucial in the sense that the question about the most appropriate data can only be answered when one knows exactly what to measure. The first question is theoretical since it can only be answered in light of each researcher’s theoretical perspective and research question. Thus, much of the disagreement about the dependent variable originates from scholars having different theoretical perspectives and not being always being fully aware of the implications of their theoretical approaches to welfare state retrenchment. Part of the problem also originates from disagreement about defining the welfare state notion in the first place. (Green-Pedersen 2004, 6)

How one defines the welfare state has clear consequences for its hypothesized effects. Before attempting to define this multifaceted and theoretically contentious concept, it is perhaps fruitful to discuss the

  Empirical Approaches to the Study of Democratic Citizens… 

  55

necessity of the close and careful alignment of concepts to theories, for only by doing so will I be able to avoid a potential independent variable problem. As Goertz (2006, 95) points out, “when designing numeric measures one needs to be concerned about how faithful the measure is to its theory (i.e., concept),” something which he refers to as “concept-measure consistency.” This idea, found in the opening paragraph of the chapter on concept-­measure consistency in his Social Science Concepts, is one which is particularly relevant for anyone who wishes to investigate the welfare state. To reiterate, I seek to comparatively examine the effects of welfare state generosity on democratic citizenship (i.e., individual political attitudes and behaviors). While this research agenda seems fairly straightforward, it is of utmost importance to utilize or develop an operationalization of the welfare state that remains true to the theorized functions and outcomes of the concept. Goertz underscores the importance that the numerical or quantitative measurement adequately reflects the basic structure of the concept. Without going into the details of his undertaking, what we can take away from his and others’ works is that the type of measurement we employ can have a significant impact on the outcomes at which we arrive. More specifically, the measurement can influence the variation of outcomes, may introduce a ceiling effect (e.g., when working with necessary conditions), or increase the gray zones. “Do the observations meaningfully capture the ideas contained in the concepts?” (Adcock and Collier 2001, 529). Inherent in this question is the idea of measurement validity. Unlike the validity of causal inference, which can be subdivided into internal and external validity, measurement validity has to do with connecting broader ideas to empirical facts. Similar to Goertz (2006), Adcock and Collier (2001) also map out the relationships between concepts and observations in four levels. The first level contains the very basic and broad “background concept.” Moving along the level of abstraction via the process of conceptualization, one arrives at the “systemized concept.” This is an explicit definition that is employed when referring to the broader concept. With the operationalization of this definition, one arrives at the third level—the indicator level. “This level includes any systematic scoring procedure, ranging from simple measures to complex aggregated indexes” (Adcock and Collier 2001,

56 

J. Shore

530). Finally, we have the scores for the individual cases at the fourth level. One should strive for continuity between the levels; in other words, the scores for the cases should be able to be meaningfully linked back to the systematized concept. The welfare state as a background concept covers a variety of ideas. As it is so far reaching, to speak simply of a welfare state does not provide us with many clues about what it is that we should examine. Encompassed in the background concept of welfare states are numerous systematized concepts—which systemized concept one ultimately uses is based on the aspects of the background concept that the research aims to address.

Measuring the Welfare State As Esping-Andersen (1990, 19) famously stated in his Worlds of Welfare, “not all spending counts equally.” For example, he points out that there are certain types of spending that do not really constitute part of the welfare effort—that is, social spending which is not aimed at or does not result in the reduction of inequalities. High levels of spending for civil servants’ benefits, “fiscal welfare for the middle class, disproportionate spending on means-tested social assistance” (Castles 2008, 45), or spending necessitated by high levels of unemployment are examples of such spending. These expenditures serve either to compensate people who are already relatively well off or “stigmatize those in receipt of benefits or are occasioned by the failure of the state to intervene in the first instance” (Castles 2008, 45). This category of spending does not reflect the core idea of the welfare state (Esping-Andersen 1990), which aims to guarantee equal rights of social citizenship for all. That being said, the aim is now to identify which types of spending are able to secure social rights for all members of a polity. Esping-Andersen’s (1990) remarks that unemployment spending and social assistance are clearly more aligned to the theoretical core of the welfare state constitute a clear critique of studies employing aggregate measures of social spending to reflect welfare state effort. His answer to this problem was the now-famous “de-commodification index” to measure the size and success of a welfare state. This index gives the most

  Empirical Approaches to the Study of Democratic Citizens… 

  57

weight to the manner in which welfare benefits are provided. In other words, how individuals are determined to be eligible recipients of welfare benefits is the most important aspect of the de-commodification approach. Generosity, or what people actually receive, comes in second, thereby taking spending levels into account. Esping-Andersen’s solution indeed comprised an innovative and more theoretically sound way to circumvent the problems inherent to aggregated measures of welfare efforts; however, as Castles (2008) points out, this critique was made at a time when disaggregated measures of welfare spending were simply not available. “If aggregation is the main problem, then the solution could be simply to avoid lumping together unlike categories in the first place or to group only those categories where one can demonstrate that they have significant things in common” (Castles 2008, 46). Today’s scholars have much more nuanced and detailed data on social expenditures at their fingertips. As a result, we are able to “ask meaningful questions about the consequences of spending of particular types” (ibid.). For example, the European System of Integrated Social Protection Statistics provided by Eurostat and the Organisation for Economic Co-operation and Development’s (OECD) Social Expenditure Database (SOCX) provide detailed program-level data that distinguish between cash benefits and services. What one sees when examining these types of disaggregated welfare spending data is that high spending in one area (be it health, age-­ related, or spending for working-age adults and families) does not necessarily mean that spending will be equally high across all welfare areas, thereby providing further support for the use of disaggregated measures. Castles (2008) is furthermore able to demonstrate that the determinants of spending vary according to spending type. It thus seems logical to assume that the effects of the different expenditures will also vary. “Different expenditure types will differ in the extent to which they are conducive to goals consonant with whatever is taken as being the ‘theoretical substance’ of welfare states” (Castles 2008, 57). Today, we are able to empirically test Esping-Andersen’s (1990, 19) statement “not all expenditure counts equally.” For example, some types of spending are more closely linked to distributional outcomes than others: neither age related nor health spending is significantly related to such outcomes.

58 

J. Shore

Spending on working-age adults and families, on the other hand, is most closely tied to distributional outcomes (e.g., in his analysis, Castles finds that the correlation coefficients for the Gini index and population and child poverty are over 0.8). Working-age spending even has a stronger relationship than the de-commodification index or public social expenditure tax incidence. Working-age spending thus “counts more toward alleviating poverty than other types of expenditures or other mechanisms of redistribution” (Castles 2008, 59). Although disaggregated measures of welfare state spending do constitute an improvement over aggregated ones, the approach is certainly not without its drawbacks. Drawing on Esping-Andersen’s de-­commodification index to create their own generosity index, Scruggs and Allan (2006) argue that only by looking specifically at the programmatic aspects of welfare states are we able to get a more complete picture of a state’s generosity and commitment to its citizens’ social rights. Moreover, looking at a single year of welfare state spending on working-age adults and families could be argued to be but a snapshot of larger processes of change and retrenchment. From the new institutionalist perspective, however, we should expect to see welfare state change within certain bounds (Pierson 2001, 415). That is, “while the notion of ‘path dependency’ allows for welfare spending/social rights data to vary up and down, the more important expectation is that each nation will remain within its existing cluster at each point in time” (Kühner 2007, 12). Despite the harsh criticism social expenditure data have been subjected to in the past, there are, nevertheless, many good reasons to use it. First of all, social spending has been shown to be highly related to redistribution; it is one of the biggest predictors of the redistributive capacity of advanced industrial nations (Huber et al. 2006). Moreover, as many studies show a depressing effect on political engagement by inequality, it may be exactly this redistributive aspect of welfare states that exerts the strongest overall effect on democratic citizenship. In using social expenditures as a predictor of democratic citizenship, I refrain from using a highly aggregated spending measure, such as total spending, for such an approach may include aspects which do not theoretically matter for many aspects of democratic citizenship. Using data on disaggregated social spending allows the researcher to both cover a broad geography and investigate recent trends in welfare states.

  Empirical Approaches to the Study of Democratic Citizens… 

  59

 lleviating Inequality: Welfare Policies Aimed A at Working-age Adults and Families In measuring the welfare state, I model my approach largely after Castles (2008), using a disaggregated measure of welfare spending, namely benefits aimed at working-age adults and families as a percentage of the GDP. This expenditure area also includes spending on Active Labor Market Policies (ALMPs). Although not part of the traditional conception of the welfare state per se, spending on such programs also constitute a commitment to working-age populations and the reduction of inequality through investments in employability and training programs. Countries vary extensively across most welfare areas (except for health spending): while states tend to spend more or less similar proportions of their GDPs on health care, the same cannot be said for expenditures aimed at pensioners and working-age adults and families. It is simply not true that countries which spend a lot in one area automatically also spend an equally high amount in others. There seems to be a trade-off in many countries with regard to spending on the elderly and working-age populations. These trade-offs can be seen in Table 4.1, which replicates Castles’ (2008) exploratory analysis for the countries and years used in the analyses to come. Kuitto (2011) tested Castles’ (2008) rather descriptive claims, finding that European welfare states do indeed tend to distinguish themselves not only in terms of the overall levels of social spending but also in their foci. Her cluster analysis reveals that while the continental European countries tend to focus on citizens who have stopped working, the Nordic countries exhibit comparatively higher levels of spending on the working-age population, reflecting a commitment to defamiliarization and de-­commodification. Additionally, Kuitto’s (2011, 361) analyses show that welfare efforts do not tend to be “bifurcated in cash transfers and services in the sense that these would build the main dimensions around which welfare regimes cluster”; rather, we see clear priorities and political responses to social risk at different points in the life cycle. For this reason, I also include spending on benefits in kind in addition to the cash-­ spending priorities Castles (2008) identifies. In sum, we see that some states spend considerably more than others or make a “greater welfare effort” (Wilensky 1975), and that different spending priorities clearly exist (Kuitto 2016).

60 

J. Shore

Table 4.1  Disaggregated welfare state expenditures across countries (% of GDP) Country

Total social expenditures

Health

Working age

Old age

Australia Austria Belgium Canada Czech Republic Denmark Estonia Finland France Germany Greece Hungary Ireland Israel Italy Netherlands New Zealand Norway Poland Portugal Slovak Republic Slovenia Spain Sweden Switzerland United Kingdom United States Mean Std. Dev. CV

16.40 26.30 26.00 16.80 18.10 26.50 12.70 24.70 29.70 25.10 21.60 23.00 16.70 15.50 24.70 21.10 18.60 20.50 19.70 22.70 15.70 19.50 21.30 27.30 18.50 20.40 16.30 20.94 4.30 20.56

5.70 6.70 7.00 7.00 5.60 6.70 3.80 5.90 8.70 7.70 5.90 5.20 5.90 4.00 6.60 7.00 7.00 5.60 4.50 6.70 5.20 5.60 6.10 6.50 5.60 6.90 7.40 6.17 1.10 17.83

5.40 6.50 9.20 2.90 5.00 11.00 3.60 8.80 7.10 6.00 2.70 7.10 6.30 5.50 3.90 7.10 6.20 7.80 4.30 5.00 3.90 3.60 6.40 10.30 5.50 6.30 2.40 5.92 2.22 37.55

4.90 12.60 9.00 4.10 7.30 7.30 5.20 9.20 12.90 10.60 12.10 9.80 3.90 5.10 14.10 5.50 4.30 6.40 10.70 10.80 6.20 9.70 8.40 9.50 6.70 5.90 6.00 8.08 2.95 36.53

Source: OECD SOCX Database. Public social spending (% of GDP) Notes: Figures do not add up to total listed due to the omission of “other” spending in this table (housing + other social policy areas). Old-age spending = Cash and benefits in kind for pensions and survivors; Health: Health expenditures; Working age: Cash and benefits in kind spending for incapacity, unemployment, family, and ALMPs. Italic fonts indicate above average spending in given column. CV: Coefficient of variation. Own calculations based on Castles (2008)

  Empirical Approaches to the Study of Democratic Citizens… 

  61

The welfare state data used in the analyses come from the OECD’s SOCX. The measure of the welfare state covers expenditures on working-­ age adults and families, measured as the percentage of the GDP spent on unemployment benefits, incapacity, families, and ALMPs. The OECD identifies nine branches of social policy. Of the four branches included in the welfare state measure (incapacity related, family, ALMP, and unemployment), two branches include both cash benefits and benefits in kind (incapacity related and family spending). Unemployment spending is cash only; ALMP is simply spending and no differentiation between cash or benefits in kind is made. Table 4.2 provides an overview of the expenditure types. In addition to taking the different spending priorities and effects into account, it is important to consider that the areas of spending included in the measure are often driven by various structural and demographic needs. In other words, it has been well documented that how much money is spent and on which groups has a lot to do with who is in need (Clayton and Pontusson 1998; Huber and Stephens 2001; Kangas and Palme 2007; Kuitto 2011; Saunders and Klau 1985; Siegel 2007). Just as a large share of retirees would likely result in greater spending on pensions, high levels of unemployment, for example, are also likely to set off increases in unemployment spending. Following Kuitto (2011), I control for the varying needs’ structures by adjusting the welfare spending areas by their respective targets. Unemployment benefits (cash and benefits in kind) as well as expenditures on ALMPs are weighted by the ­unemployment Table 4.2  Welfare spending categories Branch

Cash spending

OECD Social Expenditure Database (SOCX) Old age X Survivors X Incapacity related X Health Family X ALMP Unemployment X Housing Other social policy areas X Categories used in the primary analyses are in bold

Benefits in kind X X X X

X X

62 

J. Shore

rate; cash spending on families is adjusted by the proportion of the population under age 15 (Kangas and Palme 2007, 110); and family benefits in kind, which primarily comprises daycare services, are thus weighted by the share of the population aged 0–4.1 Kangas and Palme (2007) and Kuitto (2011) employ life expectancy as a proxy for a population’s general health status when looking at incapacity benefits; the idea here being that when populations are healthier, less money is spent on sick leave. This, however, is a very rough measure and, as Kuitto (2011, 353) points out, is “prone to counterfactual argumentation.” I, therefore, refrain from adjusting the incapacity spending. Although it is not altogether unproblematic to use a measure where some components have been weighted and others have not, in 2007, for example, the average amount spent on incapacity programs was less than 2 percent of the GDP in OECD countries. Moreover, this spending does not comprise a major benefit category and leaving it unweighted is likely to be less egregious than if one were to fail to adjust some of the other major social policy categories, such as unemployment. It must also be pointed out that all of the adjustments are proxies of dependent populations. It is simply not possible to identify the exact number of recipients or, in many cases, potential recipients of the different types of benefits. These measures and their adjustments are not, however, intended to capture precise numbers of recipients or ­beneficiaries of social expenditures; rather, as discussed in the preceding paragraphs, the measure is intended to capture the commitment a state makes to its working-age populations and reducing economic inequality. Table 4.3 provides an overview of the adjustments of the welfare policy spending areas included. The welfare state data refer to the expenditures on working-age adults and families as a percentage of the GDP in the year prior to when the respective social surveys were administered in order to ensure that cause 1  These population data are also available from the Organisation for Economic Co-operation and Development (OECD). The weight perhaps underestimates the actual daycare need for families, as in most countries, children have not entered into all-day school by the age of five. While this is certainly a weakness of the data, it can, however, be considered as a rather minor limitation. The adjustment for cash benefits (age 15) is also an approximation, as the age limits for child allowances also vary widely across countries. That being said, although these weights are approximations, failing to include any sort of adjustment of population and structural needs would lead to artificially high spending levels for many countries.

  Empirical Approaches to the Study of Democratic Citizens… 

  63

Table 4.3  Demographic needs adjustments by welfare policy category Branch

Cash spending

OECD Social Expenditure Database (SOCX) Incapacity related None Family % of population under age 15 Unemployment Unemployment rate ALMP

Benefits in kind None % of population age 0–4

Unemployment rate

indeed precedes effect (Davis 1985).2 For example, the WVS was ­conducted in 2006 in the United States; the welfare spending data for the United States, therefore, refer to the year 2005. Figure 4.1 provides an overview of the welfare state expenditures used in the analyses. As can be seen, Denmark leads the group of countries, spending an amount equal to 10.4  percent of its GDP on social expenditures for working-age adults and families. Not surprisingly, the other Scandinavian countries follow closely behind. At the bottom of the group is the United States, which spends only around 2.7 percent of its GDP on working-age populations. The mean expenditure across the 26 countries is 5.6  percent. Furthermore, the welfare state variable shows considerable variation, with a coefficient of variation of 38.5. Table 4.4 shows the summary statistics across all countries.

 tatistical Approaches: Bridging the Micro-­ S macro Divide “The frontier where critical new insights are most apt to be found now seems to be in the analysis of the linkages between macropolitical and micropolitical phenomena” (Inglehart 1983, 431). While Inglehart’s statement was certainly forward thinking and predictive of what was to come, it probably would have been difficult for him to imagine the statistical leaps that would accompany this paradigm shift (Anderson and Singer 2008, 587). One of the consequences of looking at individual 2  I return to the matter of causes and effects in the discussion of the results of the quantitative analyses.

64 

J. Shore

Fig. 4.1  Social expenditures on working-age adults and families (% of GDP). Source: OECD SOCX, various years. Country codes: AT: Austria; AU: Australia; BE: Belgium; CA: Canada: CH: Switzerland; CZ: Czech Republic; DE: Germany; DK: Denmark; EE: Estonia; ES: Spain; FI: Finland; FR: France; GB: United Kingdom; GR: Greece; HU: Hungary; IE: Ireland; IL: Israel; IT: Italy; NL: Netherlands; NO: Norway; NZ: New Zealand; PL: Poland; PT: Portugal; SE: Sweden; SI: Slovenia; SK: Slovakia; US: United States Table 4.4  Summary statistics working-age and family benefits (OECD) Mean

Standard deviation

Coefficient of variation

Range

5.61

2.16

38.50

2.24–10.47

patterns of behavior and attitudes through the institutional lens is the use of new methods of analysis. Contextual effects—that is, the effect of one’s environment—on behavior are nothing new to the field of comparative politics. Sociology and the behavioral sciences are also quite well versed in this approach: while they do not go so far as to replace country names with variable names, they do seek to understand the effects of social processes and structures on human behavior. Nevertheless, until the 1980s,

  Empirical Approaches to the Study of Democratic Citizens… 

  65

when large-scale cross-national survey projects began to emerge, political behavior tended to be studied more from the individual-level perspective than from the contextual. After this point, one could witness a shift in the focus in the social sciences from the micro-, behavioralist approach to an emphasis on macro-level differences and their implications for behavior. Equipped with the tools needed to measure micro-level phenomenon across a variety of macro-landscapes, new theories can be tested in a greater number of settings (Anderson and Singer 2008). The micro-macro relationship is rather simple and hearkens back to the traditional approach used in social psychology research: person × context (Lewin 1935). The basic intuition is that people are nested in identifiable contexts—that is, they form attitudes and make choices in variable macro-political (or other) environments (or contexts). These environments come in the form of formal institutional rules that govern people’s behavior or in the form of differential economic, social, and political conditions that shape people’s interpretations and actions. (Anderson and Singer 2008, 5)

While the institutions and structural conditions affect those who are nested within them, they may do so in different ways depending on the individual. In other words, “different kinds of people are affected differently by these environments” (ibid.). Therefore, we are faced with variations on two levels—on the individual and on the contextual level. The underlying assumption behind the idea, that different people interact with different contexts in different ways, is that the macro-environment is exogenous—“that is, external to the individual political behavior under investigation” (ibid., 6). Faced with heterogeneity across two levels of abstraction, the social sciences have witnessed many advances in statistical methods. Multilevel modeling, which got its start in education research, is a technique able to account for the hierarchical structure of such theories and data. As the research questions as well as the nature of the data imply hierarchical structures, I use multilevel analysis techniques to examine the effect a given context (welfare state) exerts on individual political dispositions and behaviors. Only by modeling the hierarchical structures in the

66 

J. Shore

data (each individual-level attribute corresponds to one of the countries in the sample) can both individual- and contextual-level aspects be taken into account, thereby avoiding a number of statistical problems one would encounter if the hierarchical nature of the data were to be ignored (Steenbergen and Jones 2002). Such an approach assumes that individual attitudes and behaviors vary across countries. In other words, it is assumed that these individual aspects vary according to welfare state context. Additionally, multilevel models enable specific macro-features, which also vary from country to country, to be accounted for as well. Surveys such as the ESS or WVS seem almost to have been designed with such an analytical approach in mind. “The basic intuition is that people are nested in identifiable contexts—that is, they form attitudes and make choices in variable macro-political (or other) environments (or contexts)” (Anderson and Singer 2008, 568).

 he Statistical and Theoretical Case for Multilevel T Analysis Why should one use multilevel models as opposed to pure macro-level techniques relying on aggregated data from the individual level? First, when individual-level data are aggregated, a great deal of information is lost: “[D]ifferent data values from sub-units are combined into fewer values for fewer high-level units” (Hox 2002, 3). Relying on aggregated analyses of democratic citizenship indicators and the welfare state would mean that we would have around 26 pieces of information on democratic citizens—one value for each country in the sample. Doing so would result in a substantial loss of information and, consequently, the statistical analyses would likely have very little power. If we were, however, to ignore the multilevel structure in the other direction, that is, disaggregate the data entirely by assigning all country-level values to individuals, “the result is that a few data values from a small number of super-units are ‘blown up’ into many more values for a much larger number of sub-­ units” (ibid.). In the disaggregated example, it is assumed that the data points (individuals) are independent of one another—something, which, conceptually, makes little sense when considering that would mean that

  Empirical Approaches to the Study of Democratic Citizens… 

  67

individuals are assigned welfare state values according to the country in which they live. The error terms for individuals in the same country are also likely to be correlated, thereby violating a basic assumption of multiple regression (Luke 2004, 7). Moreover, given the inflated sample size (with the N no longer referring to the number of countries included but the number of individuals), one is also more likely to find very significant results, that is, “significance tests that reject the null-hypothesis far more often than the nominal alpha level suggests” (Hox 2002, 3) would result. In other words, drawing on a wholly disaggregated sample would likely result in spurious findings. Theoretical and conceptual problems may also arise when ignoring the multilevel structure of the data. For one, the researcher may run into an ecological fallacy—that is, making conjectures about individuals based on aggregate data. For example, by making a cross-level inference based on macro-level data alone about individuals in various countries, we would be failing to consider the fact that people within countries may vary. On the other hand, when one draws conclusions about higher-level units (in this case, countries) based on individual-level findings, one is at risk of committing an atomistic fallacy, that is, we would be falsely assuming that there is no sort of national clustering of citizens, and that their policy environments play no role in shaping their political behavior. Multilevel analysis, however, does not define one level or another as the correct level of analysis; instead, this technique is able to acknowledge that reality often comprises multiple levels—be it school children in classrooms which are nested in schools or, in our case, individuals in national welfare state contexts. Multilevel analysis, therefore, aims to (1) examine how both higher-level and lower-level units are related to a given outcome as well as to (2) investigate “if the explanatory variables at the group level serve as moderators of individual-level relationships” (Hox 2002, 4).

Introduction to Multilevel Models We are interested in the effects of varying macro-conditions (the extent to which countries’ welfare policies reflect a commitment to working-age adults and families) on individual political behaviors and attitudes. The

68 

J. Shore

analyses in the following chapters make use of random intercept models, which work on the assumption that individual political behavior or attitudes vary across countries. These models are examples of basic two-level models, with individuals nested in different country contexts. The building block of any multilevel model is the level one model: Equation 1: Level 1 Model

Yij = b 0 j + b1 j X ij + eij ,



(4.1)

where Yij refers to a given outcome for individual i in context j. Unlike a normal regression model, this model assumes that each higher-level unit has a different intercept coefficient β0j and a different slope coefficient β1j. The residual errors eij are assumed as having a mean equal to zero as well as a variance. Both the intercept and slope can thus vary across higher-level units; in other words, we can consider both the intercepts and the slopes as outcomes of higher-level units, here countries (Luke 2004, 10). The first model shown in Eq. (4.1) shows us that the regression coefficients βj have a mean and variance. In the models below in Eq. (4.2), we can explain the variation in the coefficients β0j and β1j by adding predictor variables at the second level. The Level 2 model thus shows how the Level 1 predictors are related to the Level 2 predictors as well as the variability. Equation 2: Level 2 Model

b 0 j = g 00 + g 01 Z j + u0 j

b1 j = g 10 + g 11 Z j + u1 j

(4.2)

β0j is the Level 1 intercept in the second-level unit j; γ00 represents the mean value of the dependent variable at the first level when controlled for the Level 2 predictor Zj; .u0j is the residual error term at the second level, or the unmodeled variability for unit j. The second part of the Level 2 equation models the second-level effects on the slope of Xij, with β1j

  Empirical Approaches to the Study of Democratic Citizens… 

  69

r­ epresenting the slope of the Level 1 variable at the Level 2 unit j; γ10 is the mean value of the Level 1 slope when controlling for the second-level predictor Zj; γ11 represents the effect of the second-level predictor Zj; and u1j is the residual error term for second-level unit j. The models shown in Eqs. (4.1) and (4.2) can be expressed as a single equation by substituting the second set of equations into Eq. (4.1), resulting in the following multilevel model shown in Eq. (4.3): Equation 3: Multilevel Model

Yij = g 00 + g 10 X ij + g 01 Z j + g 11 X ij Z j + u1 j X ij + u0 j + eij



(4.3)

The first part of the model, γ00 + γ10Xij + γ01Zj + γ11XijZj, is the fixed part of the model; u1jXij + u0j + eij represents the random effects. The random effects can be considered as “additional error terms or sources of variability” (Luke 2004, 11). The starting point for most multilevel analyses is, however, the empty or null model, which contains no explanatory variables, neither at the first nor at the second level. The null model only contains an intercept and is used to produce an estimate of the intraclass correlation or the proportion of the variance in the dependent variable that can be attributed to differences between Level 2 units. Many scholars suggest building up a multilevel model, first adding lower-level predictor variables and then successively introducing Level 2 predictor variables, random slopes, and interaction terms (see, e.g., Luke 2004; Hox 2002).

Modeling Binary Data Many of the models in Chap. 5 examine attributes of democratic citizenship which are expressed dichotomously. For example, when a survey respondent is asked whether or not he or she cast a vote in the previous national election, only two answers are of interest—yes or no. Due to the binary nature of this kind of variable, modeling techniques must be chosen that fit to this kind of data. It would be inappropriate to treat a binary variable like a continuous one, as the regression assumption of normal ­ ichotomous, distribution cannot be fulfilled. As the dependent variable is d

70 

J. Shore

a generalized hierarchical linear model (GHLM) is needed: “GHLM works by including a necessary transformation and an appropriate error distribution for the dependent variable into the statistical model” (Luke 2004, 53). A dependent variable like voting is bounded by 0 and 1 and is, therefore, not normally distributed but rather binomially distributed with a mean of μ. The estimate of μ is p; p in this case would be the probability of voting. To transform a binomial model, a logit transformation is typically used: Equation 4: Logit Transformation



æ p ö logit ( p ) = ln ç ÷ è1- p ø

(4.4)

We use a logit transformation because although p is bounded between 0 and 1, the logit of p is unbounded. Moreover, the logit transformation gives us a density of p, which is much more normal. In terms of the multilevel logistic model, it looks quite similar to the linear model. The major difference, however, is that with logistic multilevel models, we do not have a term for the error variance at Level 1; instead, “variance is completely determined by the means, and thus is not a separate term to be estimated” (Luke 2004, 55). Equation 5: Logistic Multilevel Model

logit (p ij ) = g 00 + g 10 X ij + g 10 Z j + u0 j



(4.5)

Summary This very brief discussion of multilevel modeling provides but a brief look into the techniques that can be used to examine the relationships between individual political behaviors and attitudes and the country-level contexts in which they live. To review, as outlined in Chap. 3, we can expect that more generous welfare state contexts will have an overall positive

  Empirical Approaches to the Study of Democratic Citizens… 

  71

effect on democratic citizenship aspects. Moreover, income, a well-known predictor of democratic political engagement, should have a weaker relationship to political behaviors and attitudes in more generous welfare states, as welfare policies are assumed to moderate the strength of this relationship. So while we are likely to still observe a social gradient with regard to behaviors such as voting even in contexts where there is a strong welfare policy focus on working-age adults and families, the relationship between political participation and income is expected to be weaker; in other words, the slope of the line plotting political participation by income would be much flatter in a more generous welfare state. (I discuss this in more depth in the empirical analyses in the chapter to follow when introducing cross-level interactions between welfare policies and income.) Furthermore, as discussed, the intercepts for each country also vary. Where the intercept is located on the y-axis has to do with characteristics of the grouping variable—in this case, countries. Put differently, the level of political participation is determined partly by welfare state generosity as well as by other country-specific aspects, such as the electoral system. In the chapter to follow, wherein I analyze aspects of democratic citizenship across a variety of countries, the theoretical as well as mathematical reasoning behind the multilevel models should become clearer.

References Adcock, Robert, and David Collier. 2001. Measurement Validity: A Shared Standard for Qualitative and Quantitative Research. American Political Science Review 95 (3): 529–546. Anderson, Christopher J., and Matthew M. Singer. 2008. The Sensitive Left and the Impervious Right: Multilevel Models and the Politics of Inequality, Ideology, and Legitimacy in Europe. Comparative Political Studies 41 (4/5): 564–599. Castles, Francis G. 2008. What Welfare States Do: A Disaggregated Expenditure Approach. Journal of Social Policy 38 (1): 45–62. https://doi.org/10.1017/ S0047279408002547. Clayton, Richard, and Jonas G. Pontusson. 1998. Welfare-State Retrenchment Revisited: Entitlement Cuts, Public Sector Restructuring, and Inegalitarian Trends in Advanced Capitalist Societies. World Politics 51 (1): 67–98.

72 

J. Shore

Davis, James Allan. 1985. The Logic of Causal Order. Beverly Hills, CA: Sage Publications. Esping-Andersen, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press. Goertz, Gary. 2006. Social Science Concepts: A User’s Guide. Princeton, NJ: Princeton University Press. Green-Pedersen, Christoffer. 2004. The Dependent Variable Problem Within the Study of Welfare State Retrenchment: Defining the Problem and Looking for Solutions. Journal of Comparative Policy Analysis: Research and Practice 6 (1): 3–14. Hox, Joop. 2002. Multilevel Analysis. Techniques and Applications. New  York: Taylor and Francis Group. Huber, Evelyne, Francois Nielsen, Jenny Pribble, and John D. Stephens. 2006. Politics and Inequality in Latin America and the Caribbean. American Sociological Review 71 (6): 943–963. Huber, Evelyne, and John D.  Stephens. 2001. Development and Crisis of the Welfare State: Parties and Policies in Global Markets. Chicago: University of Chicago Press. Inglehart, Ronald. 1983. Changing Paradigms in Comparative Political Behavior. In Political Science: The State of the Discipline, ed. Ada W. Finifter, 429–469. Washington, DC: The American Political Science Association. Kangas, Olli, and Joakim Palme. 2007. Social Rights, Structural Needs and Social Expenditure: A Comparative Study of 18 OECD Countries 1960–2000. In Investigating Welfare State Change: The “Dependent Variable Problem” in Comparative Analysis, ed. Jochen Clasen and Nico A.  Siegel, 106–129. Cheltenham: Edward Elgar. Kühner, Stefan. 2007. Country-Level Comparisons of Welfare State Change Measures: Another Facet of the Dependent Variable Problem Within the Comparative Analysis of the Welfare State? Journal of European Social Policy 17 (1): 5–18. Kuitto, Kati. 2011. More Than Just Money: Patterns of Disaggregated Welfare Expenditure in the Enlarged Europe. Journal of Social Policy 21 (4): 348–364. ———. 2016. From Social Security to Social Investment? Compensating and Social Investment Welfare Policies in a Life-Course Perspective. Journal of European Social Policy 26 (5): 442–459. https://doi.org/10.1177/095892 8716664297. Lewin, Kurt. 1935. A Dynamic Theory of Personality. New York: McGraw-Hill. Luke, Douglas A. 2004. Multilevel Modeling. Thousand Oaks, CA: Sage.

  Empirical Approaches to the Study of Democratic Citizens… 

  73

Pierson, Paul. 1994. Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. Cambridge Studies in Comparative Politics. Cambridge, UK and New York: Cambridge University Press. ———. 2001. Coping with Permanent Austerity: Welfare State Restructuring in Affluent Democracies. In The New Politics of the Welfare State, ed. Paul Pierson, 410–456. Oxford: Oxford University Press. Saunders, Peter, and Friedrich Klau. 1985. The Role of the Public Sector: Causes and Consequences of the Growth of Government. Paris: Organisation for Economic Co-operation and Development. Scruggs, Lyle, and James Allan. 2006. Welfare-Decommodification in 18 OECD Countries: A Replication and Revision. European Journal of Social Policy 16 (1): 55–72. Siegel, Nico A. 2007. When (Only) Money Matters: The Pros and Cons of Expenditure Analysis. In Investigating Welfare State Change, ed. Jochen Clasen and Nico A. Siegel, 43–71. Cheltenham: Edward Elgar. Steenbergen, Marco R., and Bradford S. Jones. 2002. Modeling Multilevel Data Structures. American Journal of Political Science 46 (1): 218–237. Wilensky, Harold L. 1975. The Welfare State and Equality: Structural and Ideological Roots of Public Expenditures. Berkeley and London: University of California Press.

5 The Impact of Social Policy on Democratic Citizenship

In this chapter, the theoretical arguments regarding the influence of the welfare state on both the levels and distributions of democratic citizenship are put to the empirical test. The quantitative analyses draw on survey data from the European Social Survey (2006, 2008), Comparative Study of Electoral Systems (2015), and the World Values Survey (Inglehart et al. 2014). The first set of analyses examines electoral participation across welfare states and provides detailed information about the common research designs used in the subsequent analyses of political interest, political trust, and satisfaction with democracy.

Voting Of all the effects that the welfare state may have on democratic citizenship, voting could very well be considered the most important, as it is the primary channel through which citizens can exercise their political rights and influence politics in a democracy. Moreover, all conceptualizations of democracy, be it a very minimal proceduralist conceptualization or a maximalist participatory one, underscore the importance of elections in their formulations (refer to the discussion in Chap. 2). And because direct © The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2_5

75

76 

J. Shore

democratic institutions exist at the national level in only a few countries, voting in elections remains the act of political participation in contemporary democracies. In the analyses to follow, I test the arguments laid out in Chap. 3 on the ways in which welfare policies can impact citizens’ political behaviors and attitudes. Individuals are thought to be more likely to take part in elections in countries where there is a large welfare state policy focus on working-age adults and families. This positive effect of the welfare state on turnout is presumed to be largest for individuals with the fewest resources. In welfare states where there is a stronger focus on working-age populations, the salience of politics is greater, for public policy plays a greater role in people’s lives. This may be in the form of increased employment protection or greater family benefits. Where politics are of greater salience, we expect to see higher turnout. People’s life chances are more closely connected to policy outcomes, thereby creating personal political stakes. Beyond the higher likelihood of voting, more equal patterns of voting across socioeconomic groups should also be observed in countries with more generous welfare contexts (Solt 2008). The following analysis looks at the relationships between the welfare state and the individual propensity to vote as well as the ways in which the welfare state moderates the relationship between income and electoral participation. The analysis examines electoral participation in 26 OECD countries between the years 2005 and 2008. Table 5.1 provides an overview of the countries included in the analyses, the survey sources, as well as the year in which the individual data on voting were obtained. Some OECD members are not included due to reasons of data availability as well as comparability: Chile, Iceland, Japan, Luxembourg, Mexico, New Zealand, South Korea, and Turkey. Respondents in New Table 5.1  Countries, years, and survey sources included in the analyses of voting and welfare states Year

Countries

Survey

2005 2006 2008

Australia, Italy Austria, Belgium, Canada, Ireland, Slovakia, United States Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Israel, Netherlands, Norway, Poland, Portugal, Slovenia, Spain, Sweden, Switzerland, United Kingdom

WVS ESS/WVS ESS

  The Impact of Social Policy on Democratic Citizenship 

  77

Zealand were not asked whether or not they voted; other demographic information, such as employment status, is also missing in the case of New Zealand. While individual-level data on Iceland and Luxembourg are available from the European Values Surveys, the question on voting asks respondents about their intention to vote and not whether they participated in the last election. In terms of Mexico and Turkey, I opted to omit these countries because these states do not exhibit as high a degree of democracy as the other member states. Practically speaking, I have omitted countries from the sample that are not ranked by Freedom House as “free.” Turkey is right on the cusp of the “free”/“partially free” placement, and with a score of 2.5, scores considerably lower than the other countries—of which the vast majority have received the highest ranking of 1.0. Mexico is clearly not part of the group of “free” countries with its score of 3.0. Mexico and Turkey, while both electoral democracies, are categorized as partially free. Civil rights abuses, freedom of the press issues, corruption, and so on, are all factors that could potentially influence voting behavior and people’s attitudes toward democracy. The remaining non-Western omissions are based mainly on comparability issues as well as failures to conform to the theoretical expectations that resources foster political participation. For example, it has been shown that corruption can substantially decrease electoral participation (Stockemer et al. 2013), something which is indeed a problem in many less developed democracies. While these cases are certainly of interest, it is beyond the scope of this book to speculate as to why neither positive nor significant correlations between individual resources and voter turnout are found in these countries. Finally, although case selection is certainly guided to a large extent by data availability, there are, nevertheless, theoretical arguments for comparing a group of countries such as these. First, we find a wide range of variation of voter turnout. Second, all countries are members of the OECD. Despite their shared commitment to democracy and the market economy as OECD members, the states have produced very different responses to the social question. There are also reasons to restrict the investigation to advanced industrial states: the “paradox of the civil society argument,” as Michael Walzer (1992) calls it, refers to the mutually reinforcing relationship between civil society and democratic governance, wherein a “democratic civil society seems to

78 

J. Shore

require a democratic state, and a strong civil society seems to require a strong and responsive state” (Foley and Edwards 1996, 48).

The Decision to Vote Data for the dependent variable, voting (yes/no), as all individual-level variables come from various rounds of the ESS (2006 and 2008) and the WVS (2005). Although there are some drawbacks to using multiple survey sources, due mainly to differing sampling procedures, these surveys do, however, provide the advantage that they employ the exact same question (identical wording and response options) on voting: Did you vote in the last national election? While not all idiosyncrasies are eliminated with this common question, country-specific error terms in the models do, however, capture many of the differences that may result from issues arising from the different survey designs. The sample consists of 34,542 individuals of legal voting age1 or older in the 26 countries. The average reported voter turnout rate across all countries is 80.08  percent, ranging from 58.2  percent in the Czech Republic to 95.6 percent in Australia. It must be emphasized that these figures represent self-reported electoral participation and do not perfectly correspond to official national turnout rates. For example, the officially reported turnout rate in Germany for the 2005 parliamentary election was 77.7  percent, 64.9  percent in Canada for the 2006 election, and approximately 82 percent for the 2006 Swedish parliamentary elections.2 When comparing these rates to the average self-reported voting rates found in the surveys, we find notably higher figures: 83  percent in Germany, 74.7 percent in Canada, and 91.4 percent in Sweden. The fact that self-reported voting rates are considerably higher than the official rates is nothing new. Although voters are given the chance to answer no, 1  All individuals under the age of 18 have been omitted from the sample. While voting age in Austria has been lowered to 16, at the time the survey was conducted, the legal voting age was still 18. In Slovenia the voting age is 16 for employed persons; there are, however, no employed 16-year-­ olds in the sample. 2  Historical and cross-national official turnout rates are available from the International Institute for Democracy and Electoral Assistance (IDEA): https://www.idea.int/data-tools/data/voter-turnout (Accessed February11, 2017).

  The Impact of Social Policy on Democratic Citizenship 

  79

I did not vote, and many surveys attempt to frame the voting question in such a way so as to assure respondents that it is acceptable to admit to not having voted,3 survey data, nevertheless, tend to present overreported voter turnout rates; put simply: “many more people voted than possibly could have” (Highton 2005, 113). Inconsistencies such as these reflect what is known in survey research as problems of social desirability, wherein respondents report having voted because it presents them as a “good citizen” (Belli et al. 1999). Another possible reason surveys tend to overreport voting behavior has to do with memory: respondents simply fail to correctly recall whether or not they voted (Abelson et al. 1992). Despite downsides of using survey data to capture electoral participation, it is important to keep in mind that the correlates of voting do not tend to be as biased as reported turnout (Verba et al. 1995, 615–17), which is why scholars continue to depend on survey data to investigate the determinants of and changes in voting patterns. Moreover, as the focus of this study lies more on the distribution and less on the overall levels of electoral participation, the survey data provide an adequate means to analyze who votes. Table 5.2 provides an overview of the mean self-reported voter turnout as well as the actual official turnout rate. As the table shows, not all countries have overreported voter turnout; some even have underreported mean voter turnout. On the whole, however, voter turnout is overreported on average by around 8 percentage points. In addition to the differences we can see between the countries in terms of the overall level of turnout, there are also marked differences within countries with regard to turnout and income level. Furthermore, in many countries, there appear to be much smaller gaps between levels of participation across income groups than in others, as Table 5.3 illustrates. Across the 26 countries in the analysis, while individuals from the lowest-income quintile go to the polls at an average rate of 74.7 percent, their better-off counterparts in the highest quintile turn out to vote at an average rate of 87.0  percent—a turnout rate that is more than 12 ­percentage points higher than that of the lowest-income group. The middle-­income group (comprising quintiles of 2–4) votes at an average 3  In the European Social Survey (ESS), for example, before being asked about voting, respondents are first told, “Some people don’t vote nowadays for one reason or another” (European Social Survey 2010).

80 

J. Shore

Table 5.2  Self-reported electoral participation and official turnout rates Country

Self-reported (%)

Official (%)

% Difference

Australia Austria Belgium Canada Czech Republic Denmark Estonia Finland France Germany Greece Hungary Ireland Israel Italy Netherlands Norway Poland Portugal Slovakia Slovenia Spain Sweden Switzerland United Kingdom United States Mean

95.61 86.71 92.44 74.71 58.15 95.28 67.37 83.51 77.94 83.09 88.82 80.26 77.37 78.10 90.64 86.40 86.35 75.31 70.91 71.34 73.61 81.74 91.38 66.07 71.61 77.40 80.08

94.32 81.44 91.63 64.94 64.47 86.59 61.91 74.05 83.97 77.65 74.14 64.39 62.57 63.55 81.44 80.35 77.44 53.88 64.26 54.67 60.64 75.66 81.99 48.28 61.36 86.04 71.99

1.29 5.27 0.81 9.77 −6.32 8.69 5.46 9.46 −6.03 5.44 14.68 15.87 14.80 14.55 9.20 6.05 8.91 21.43 6.65 16.67 12.97 6.08 9.39 17.79 10.25 −8.64 8.10

Sources: Self-reported turnout: European Social Survey (rounds 3 and 4), World Values Survey. Official turnout: IDEA

rate of 77.7 percent. The largest difference between the voting rates of the highest- and lowest-income groups can be found in Germany (with a 21.4 percent difference), Switzerland (with a 19.6 percent difference), and in Poland (with a 19.4 percent difference). Countries with more equal levels of voter participation include Denmark (with a 3.4 percent difference), a group of countries with compulsory voting—Italy4 (with 4.7 percent difference), Belgium, and Australia (each with a difference of 4  As Lijphart (1997) speculated and Hooghe and Pelleriaux (1998) show for Belgium, compulsory voting is able to reduce some of the inequalities of voter turnout.

  The Impact of Social Policy on Democratic Citizenship 

  81

Table 5.3  Self-reported voter turnout by income group

Country

Turnout lowest quintile (%)

Turnout mid-level income (%)

Turnout highest quintile (%)

Difference (highest vs. lowest)

Australia Austria Belgium Canada Czech Republic Denmark Estonia Finland France Germany Greece Hungary Ireland Israel Italy Netherlands Norway Poland Portugal Slovakia Slovenia Spain Sweden Switzerland United Kingdom United States Mean

92.32 83.11 89.52 65.04 52.96 93.10 60.27 77.74 71.82 71.37 87.01 74.68 69.09 70.41 88.42 80.69 80.07 63.81 65.61 66.47 70.15 80.30 86.07 57.74 67.52 69.23 74.69

97.19 87.36 88.73 75.39 60.29 96.05 67.33 84.62 79.58 85.3 87.13 83.52 76.61 82.19 92.04 84.89 85.58 71.07 71.00 72.19 73.86 80.38 91.34 68.81 72.83 76.88 78.65

97.81 92.78 95.01 78.56 63.27 96.48 74.82 87.44 82.62 92.79 94.33 80.54 78.90 85.71 93.09 89.03 89.48 83.22 81.69 75.00 78.81 88.49 92.84 77.37 75.05 86.10 87.01

5.49 9.67 5.49 13.52 10.31 3.38 14.55 9.70 10.80 21.42 7.32 5.86 9.81 15.30 4.67 8.34 9.41 19.41 16.08 8.53 8.66 8.19 6.77 19.63 7.53 16.87 12.32

Sources: World Values Survey, wave 4; European Social Survey, rounds 3 and 4

5.5 percentage points between low- and high-income turnout)—and Hungary (with a 5.9 percent difference between the highest- and lowestincome groups’ self-reported electoral participation). After first testing to see whether welfare spending on working-age adults and families exerts an effect on people’s decisions to vote using logistic multilevel models with random intercepts, I then examine whether there are differential effects of welfare state generosity for different income groups. To do so, I introduce to the multilevel model cross-­ level interactions between one’s income category (high, medium, or low)

82 

J. Shore

and the level of welfare spending. The cross-level interactions themselves, however, are not able to tell us much about how welfare policies aimed at working-age adults and families affect the three income groups; in order to examine how the welfare state moderates or interacts with the relationship between income and voting, it is necessary to calculate the marginal effects of one variable when another changes (Brambor et al. 2006). After calculating the marginal effects, it will be possible to determine whether there are different effects for the different income groups. Income group-­ specific effects are, therefore, present if one or both of the following conditions are fulfilled: The marginal effect of welfare spending on working-age adults and families is significant for some income groups and not for others. In other words, we would find that such policies exert a substantive influence on the voting behavior of one or two of the income groups but do not significantly impact the other group or groups. Alternatively, we may speak of group-specific effects if we find that the marginal effect of welfare spending on working-age adults and families varies significantly across income groups. In order for this condition to be met, these welfare policies would need to either significantly impact the voting likelihood or all income groups or none of them. Given either one of these situations, the effect size would also need to be significantly greater or weaker for one income group.5 As is elaborated on in the section to follow, the cross-level interactions model and the marginal effects derived therefrom help us to identify the presence of group-specific policy effects. Whether political inequality can be diminished by welfare state spending on working-age adults and families is the subject of this chapter’s conclusion. Finally, by drawing on both individual-level and macro-level data, a multilevel model with cross-level interactions, this design is arguably less prone to the issues of endogeneity that tend to plague aggregate analyses of political participation and the welfare state or, more frequently, macro-level income inequality. For example, if one can observe that inequality in participation decreases between the income groups as welfare spending on working-age adults and families increases, we can no longer say that we are only observing a 5  See Stadelmann-Steffen’s (2011) article on female labor market participation and family policy for an excellent discussion on the identification and classification on group-specific policy effects.

  The Impact of Social Policy on Democratic Citizenship 

  83

level effect. Furthermore, arguments of endogeneity would be quite complex ones to make. I, however, return to this discussion in greater detail in the concluding chapter.

Multilevel Analysis of Voting I examine the effect of a given context (the welfare state) on the individual decision to participate in an election and, as the research question and the nature of the data (individuals nested in national contexts) imply hierarchical structures, multilevel analysis is the most appropriate analytical technique. As the dependent variable, whether or not the respondent voted, is dichotomous, logistic multilevel regression models are used. A number of individual as well as contextual factors are controlled for in the analysis. At the individual level, a person’s age,6 gender, level of education, employment status, area of residence (urban or not), and union membership are taken into consideration. These individual attributes, which mainly correspond to the resources for participation one has at his or her disposal, have been frequently identified in the extant literature as some of the most prominent covariates of voting (Brady et  al. 1995). Income, which has been identified in the theoretical sections as a key variable, is initially included in the models as a continuous variable (covering all ten income deciles) to first show its overall impact on the propensity to vote. The final two models include cross-level interactions between income and spending on working-age adults and families. In the first interaction model, income enters the interaction as a continuous variable to provide an overall picture of the ways in which income and welfare generosity interact; in the final model, I test for the group-specific effects by interacting the welfare spending variable with dummies for each of the income groups (low, medium, and high). In additional to the main context variable, the welfare state, I also include two additional institutional variables, as the design of the electoral system has also been shown to impact the calculus of voting. Most 6  In the analysis, age is included both as an absolute value (age in years) and as a quadratic term in order to allow for the possibility that the relationship between age and voting could be curvilinear in nature. Both terms are centered around the grand mean (Hox 2002, 56).

84 

J. Shore

prominently, numerous studies have found that highly disproportional electoral systems tend to exert a dampening effect on electoral participation (Blais 2006; Franklin 1997; Jackman 1987; Jackman and Miller 1995; Lijphart 1997). For example, it has been repeatedly found that turnout is lower in disproportional electoral systems where the translation of votes into legislative seats is highly distorted by the electoral rules. In such systems, smaller parties must accumulate many more votes in order to achieve legislative representation, which may cause their supporters to abstain from voting altogether, as they do not wish to waste their votes on nonviable parties. Disproportionality also exerts an influence on voters’ efficacy, as voters in highly disproportional systems are more likely to perceive the system to be unfair and less likely to feel that their vote matters (Karp and Banducci 2008). Second, I control for the effect of compulsory voting laws on the individual decision to vote. While the inclusion of such a variable may seem like a rather obvious choice, it has been demonstrated that compulsory voting rules only make a difference for voter turnout if penalties for nonvoting are both sufficiently severe and enforced (Blais and Carty 1990; Blais and Dobrzynska 1998; Franklin 1999; Panagopoulos 2008; Singh 2011). I, therefore, abstain from using a simple dichotomous measurement of the existence of compulsory voting and instead take cues from measurement schemes such as those proposed by Panagopoulos (2008) and Singh (2011), wherein both the severity and the enforcement of sanctions are taken into consideration. I further control for the possibility that the strength of the left parties could be both drivers of increased social spending and a mobilizing force for poor voters (Brady 2003; Hicks and Swank 1992; Ross 2000). Information about the strength of left come from the Comparative Political Data Set (Armingeon et  al. 2011), measuring social democratic and other left parties as a percentage of total cabinet posts, weighted by the number of days the government was in office in a given year. Finally, as we are dealing with a very diverse group of countries, not only in terms of their social policy foci and spending records but also with regard to their histories and experiences with democracy, I include

  The Impact of Social Policy on Democratic Citizenship 

  85

a control for the age of the current regime. Some of the democracies are much older than others, and despite the fact that all countries included in the analysis can be considered examples of advanced democracies, citizens in newer democracies have simply had shorter experiences with democracy. Despite the greater fluctuations in voter turnout experienced in post-communist countries, it can be observed that overall turnout in those countries tends to be lower, with many explanations centering on a less well-developed democratic political culture or citizen disenchantment due to “unfulfilled promises of democracy” (Dalton 2000, 930; Pacek et al. 2009). Tavits and Annus (2006), however, argue that voters in young democracies have to first learn about their new electoral systems and how to ensure they do not waste their votes. In line with what we know about the design of electoral systems and turnout, it could possibly be argued that older democracies may witness weaker voter turnout, as voting is no longer a novelty and, in highly disproportional systems, voters may have realized that voting may not be worth it if they prefer a smaller, less viable party. While these effects may not be as pronounced nearly 20 years after the dissolution of the Soviet Union, it is, nevertheless, important to account for the marked differences in terms of regime age found in the sample. At the same time, it can also be observed that turnout in the oldest democracies has declined more over the past decades than in others. In addition to the argument that voters in the oldest democracy may have quite simply grown tired of going to the polls, many have noted the processes of de-alignment that have occurred in many established democracies. That is, party identification has been on the decline in the past decades, and thus has the mobilizing role parties can have, particularly with regard to peripheral voters, has also been reduced. Moreover, the loosening of the ties between labor unions and labor parties has certainly led to increasing difficulties in mobilizing organizing interests and, consequently, to a drop in turnout (Gray and Caul 2016). The age of democracy could therefore exert either a positive or a negative influence on voting. A complete overview of all variables, their sources, operationalizations, and summary statistics can be found in the appendix.

86 

J. Shore

Empirical Findings: Voting In this section, I present the findings of the multilevel models.7 Beginning with an empty model (Model 1), the models are presented in a stepwise manner, with each successive model building on the last. Model 2 adds the individual-level explanatory variables; Model 3 introduces the welfare state variable; in Model 4, additional country-level predictors (compulsory voting and disproportionality of the electoral system) are included; and in Models 5 and 6 (in Table 5.9), I add cross-level interaction between income (first as a continuous variable and then as categories) and welfare state spending. The final models allow us to test whether welfare state policies aimed at working-age adults and families moderate the relationship between an individual’s income and his or her propensity to vote. From the model with the interaction terms between the income categories and welfare state spending, I then calculate marginal effects to investigate whether the effect of welfare state spending is uniform across all income groups. The first model (Model 1) is an empty, or null, model. Only a random intercept is included. The highly significant intercept and variance components at the country level lend support for multilevel analysis, as voting is indeed shown to vary significantly across countries. From the empty model (in Table 5.4), we can say that the log-odds of voting in an “average” country is estimated at 1.525. The intercept for country j with 1.525 + uoj. In substantive terms, if we calculate the odds ratios from the log-odds shown in Table  5.4, people are on average around 4.6 times more likely to have reported voting than not to have voted. The proportion of the variance in voting that comes from differences between countries is 0.125, that is, around 12.5  percent of the variance in voting originates from differences found at the country level. An intraclass correlation of this size provides further justification for continuing on to multilevel models. The second value listed is the estimated standard deviation of the random intercept, that is, the standard deviation at the 7  I describe and explain the models and procedures in much greater depth in the present analysis on voting than in the subsequent analyses, as the basic structure of the models remains more or less the same.

  The Impact of Social Policy on Democratic Citizenship 

  87

Table 5.4  Multilevel analyses of voting Voted—yes/no Constant Fixed effects Age (mean-centered) Age squared Union membership (ref. = non-member) Employed (ref. = not employed) Gender (ref. = male) Education level (0–5) Urban (ref. = non-urban) Income decile Welfare effort: working-age adults/families Compulsory voting Disproportionality Age of democracy Strength of left Random effects Rho Context variance N

Model 1

Model 2

Model 3

Model 4

1.525*** (0.135)

0.230 (0.142)

−0.463 (0.354)

−0.239 (0.379)

0.035*** (0.001) −0.000*** (0.000) 0.199*** (0.034) 0.096*** (0.036) −0.017 (0.029) 0.298*** (0.014) −0.072** (0.032) 0.090*** (0.007)

0.035*** (0.001) −0.000*** (0.000) 0.197*** (0.034) 0.096*** (0.036) −0.017 (0.029) 0.298*** (0.014) −0.071** (0.032) 0.090*** (0.007) 0.123**

0.035*** (0.001) −0.000*** (0.000) 0.196*** (0.034) 0.097*** (0.036) −0.017 (0.029) 0.298*** (0.014) −0.070** (0.032) 0.090*** (0.007) 0.098*

(0.058)

(0.053) 0.368*** (0.109) −0.025 (0.028) −0.000 (0.002) −0.002 (0.003)

0.125 0.119 0.103 0.069 0.685 0.668 0.615 0.495 34,542 (26) 34,542 (26) 34,542 (26) 34,542 (26)

Notes: Log-odds with standard errors in parentheses; Model 1: Empty model, intercept only; Model 2: Individual variables only; Model 3: Welfare state variable added; Model 4: Full model (individual and country-level predictors) ***p < 0.01; **p < 0.05; *p < 0.1

88 

J. Shore

c­ ountry level (Rabe-Hesketh and Skrondal 2008, 249). From the likelihood ratio statistic that tests the null hypothesis that øˆ = 0 , we can reject the null hypothesis. The test statistic of 2260.61 and its corresponding p-value that is less than 0.000 provides us with strong evidence that the between-country variance is non-zero. In Model 2, individual-level determinants of voting are added to the model. The first thing to be observed is the slight reduction in variance— both in the contextual-level variance and the proportion of the total variance which comes from differences between the higher-level units. Not surprisingly, all but one of the individual-level variables are highly significant. When dealing with such a large number of cases—34,542 individuals—it is quite likely to find highly significant relationships at this level. From this second model, the following conclusions can be drawn: the likelihood of voting increases with age, as seen by the positive and highly significant coefficient for age. The relationship between age and propensity to vote is, however, slightly U-shaped—that is, one’s likelihood of voting first increases and then tapers off slightly with increasing age. Members as well as former members of trade unions are also more likely to vote than non-members. This finding is a rather intuitive one, as union members are likely to be exposed to political information and experience mobilizing effects via membership. Increasing education also exerts a bolstering effect on one’s likelihood of voting. Compared to non-employed individuals, people who hold a job are more likely to vote. People who reside in urban areas are also more likely to be voters than people who live in rural locations. Another strong effect can be seen with regard to the relationship between income and electoral participation, thereby corroborating the hypothesis and volumes of academic work showing that resources matter for political participation. In terms of differences between men and women, Model 2 shows that men are slightly more likely to vote than women—a finding which echoes the extant literature on voter turnout. This gender difference, however, is not found to be statistically significant. Overall, we can say that people with greater resources are more likely to be voters. Models 3 and 4 in Table 5.4 include country-level predictors. In Model 3, the welfare state variable is added. In accordance with the hypothesized relationship between voting and welfare state spending, this variable is

  The Impact of Social Policy on Democratic Citizenship 

  89

shown to exert a positive and highly significant (p < 0.05) direct effect on the individual decision to vote: individuals are more likely to vote in countries where a greater percentage of the GDP is spent on working-age adults and families: one additional percentage point of spending amounts to an increase in the odds ratio (not shown) of voting by 1.13. Compared to the empty model, the proportion of the variance coming from the differences between voting is reduced by around 17.6 percent. When adding control variables (Model 4), the strength of the welfare state variable is notably reduced but continues to remain significant at the 10 percent level. While the effect of the level of disproportionality of a national electoral system works in the hypothesized direction—greater disproportionality is associated with a reduced likelihood of voting—its effect is negligible. Compulsory voting, on the other hand, is found to exert a positive and highly significant effect on one’s decision to vote. This result is in line with the hypothesized relationship; moreover, a quick glance at Table  5.6 shows that the countries with the most severe and enforced compulsory voting legislation (Australia and Belgium) also have very high rates of turnout. While the strength of left parties in government seems to have a negative effect on one’s propensity to vote, this effect is, however, negligible. The age of a democracy also does not appear to impact vote likelihood. Models 3 and 4 have shown that a welfare state policy focus on working adults and families positively and significantly impacts the probability that an individual decides to vote, even when controlling for other factors at both the individual and country levels. Compared to the empty model, Model 4 reduces the variance at the country level by around 28 percent and the proportion of the total variance, which comes from differences between countries by around 45 percent. Up until now, the models have shown that welfare state spending on working-age adults and families is able to contribute to the explanation of the differences between individual voter turnout in the 26 countries. Models 3 and 4 clearly show a direct effect of welfare state spending; however, in order to test whether this type of welfare state policy focus moderates the relationship between income and electoral participation as well as whether the effects are uniform across groups and thus able to promote political equality, it is necessary to include interaction terms

90 

J. Shore

between the income (both as a continuous variable and as categories) and the welfare state variable. The final models—Models 5 and 6  in Table  5.5—test whether there are differential effects of welfare state spending aimed at working-age adults and families across income groups by including cross-level interactions. The added value of including these interaction terms is that we are now able to take the micro-micro relationships between income and voting into account when looking at the micro-macro relationship between individual voter turnout and welfare state policies. Table 5.5  Multilevel analyses of voting with cross-level interactions Voted—yes/no Constant Fixed effects Age (mean-centered) Age squared Union membership (ref. = non-member) Employed (ref. = not employed) Gender (ref. = male) Education level (0–5) Urban (ref. = non-urban) Income (deciles)

Model 5

Model 6

−0.618 (0.402)

−0.243 (0.393)

0.035*** (0.001) −0.000*** (0.000) 0.198*** (0.034) 0.115*** (0.036) −0.021 (0.029) 0.306*** (0.014) −0.071** (0.032) 0.398*** (0.063)

0.035*** (0.001) −0.000*** (0.000) 0.196*** (0.034) 0.113*** (0.036) −0.020 (0.029) 0.307*** (0.014) −0.070** (0.032)

Income cat. (ref. = low) Mid-level

0.143**

0.423*** (0.113) 0.777*** (0.126) 0.124**

(0.057) 0.374*** (0.111)

(0.055) 0.374*** (0.112)

High Welfare effort: working-age adults/families Compulsory voting

(continued)

  The Impact of Social Policy on Democratic Citizenship 

  91

Table 5.5 (continued) Voted—yes/no Disproportionality Age of democracy Strength of left Welfare × Income

Model 5

Model 6

−0.026 (0.028) −0.000 (0.002) −0.002 (0.003) −0.019* (0.010)

−0.026 (0.029) −0.000 (0.002) −0.002 (0.003)

0.071 0.502 34,542 (26)

0.072 0.506 34542 (26)

Welfare × Mid-level income Welfare × High income Random effects Rho Context variance N

−0.014 (0.020) −0.036* (0.021)

Notes: Log-odds. Standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1

Table 5.6  Marginal effects of welfare policy on voting by income category Voted—yes/no

Spending on working-age adults and families

Low income

0.020** (0.009) 0.015** (0.007) 0.011 (0.007)

Mid-level income High income

Notes: Standard errors in parentheses. Based on Model 6 in Table 5.5; other values held at means **p < 0.05

The coefficients of the individual-level variables in Models 5 and 6 remain almost entirely unaffected, save for the coefficients of the income variables, which are part of the interaction terms. I refrain, however, from further discussion of the control variables in the model, as they are precisely that—controls. In Model 5, with the interaction between welfare spending and income as a continuous variable, we see that the overall

92 

J. Shore

impact of resources declines as welfare spending on working-age adults and families increases. Put another way, the importance of income for one’s propensity to vote increases where the state spends less on these groups. In Model 6, interaction between welfare spending and the income categories are included. The coefficient for the single welfare state term represents the effect of welfare state spending on voting for the lowest-income category (0.124). Here, we see a positive and significant effect at the 5  percent level. The interaction terms refer to the effects for the two higher-income categories. Interpreting the effects of interaction terms, however, is a notoriously thorny undertaking. When looking first to the significance of the interaction terms, we get an idea about whether the effect of welfare state spending on voting for the higher-income categories actually differs from the effect for the lowest category. Here, we can tentatively say that the effects of the welfare state on voting are not significantly different for the lowest group compared to the middle-income category. The effect of the welfare policy variable on voting for people in the highest-income category, however, does appear to be significantly different than the effects found for the lowest category, as indicated by the single star denoting significance at the 10 percent level. As Brambor et al. (2006) point out, although it is necessary to include the constitutive terms of an interaction in addition to the interaction itself, they should not be interpreted as unconditional marginal effects. In order to see whether there are significant interaction effects for the income categories, I, therefore, calculate the marginal effects for the respective groups. A marginal effect estimates the extent to which the dependent variable is expected to “increase or decrease for a unit change in an explanatory variable” (Buis 2010, 305). As mentioned, the single welfare state variable represents the impact of welfare state spending on working-age adults and families on the voting behavior of individuals from the lowest-­ income category. Table 5.6 shows the marginal effects of welfare state spending on voting for each of the three income groups. These figures show that the effect of welfare state spending on the propensity to vote does not affect all groups of individuals in the same manner, indicating the presence of the first condition—that the marginal effect of the welfare policy variable is

  The Impact of Social Policy on Democratic Citizenship 

  93

.85 .8 .75 .7

Probability of Voting

.9

significant for some income groups but not for others. Although we see positive effects of the welfare state on voting for all income groups, the magnitude as well as the significance of the effects are not uniform. Referring back to the regression Table 5.5, we see that although the effect of the welfare state on the middle-income group is not significantly different from the effect for the lowest-income group, clear differences in terms of effect size can be observed when comparing the highest- to the lowest-income group. The largest effect of additional welfare state spending on one’s propensity to vote is found for the lowest-income group. In other words, while we do observe an overall bolstering effect of the welfare state on voter turnout, this effect is the greatest for the lowest-income group, followed closely by the middle group. These differences are best observed graphically: Figure 5.1 illustrates that at the very lowest levels of spending on working-­age adults and families, electoral participation is highly stratified. This stratification is seen to level off considerably when the higher levels of spending are reached. As mentioned, the effects for the lowestand mid-level income groups are quite similar, as can be seen by the

2

3

4

5

6

7

8

9

10

11

% of GDP Spent on Working-Age Adults and Families Low Income

Mid-Level Income

High Income

Fig. 5.1  Income level and predicted turnout depending on welfare state spending. Notes: Based on model 6 in Table 5.5. All other values held at means

94 

J. Shore

similarity in steepness of the lines for these two income groups. The slope for the highest-income group, however, is clearly the least steep. The effect of additional welfare spending on voting is the smallest for individuals belonging to this highest-income category. In other words, those people who already have ample resources are not as affected by increases in working-­age spending than people with fewer resources. Nevertheless, we do still see an increase in the propensity to vote for individuals from the highest-income category as well (even though this effect does not reach conventional levels of significance). Despite the fact that electoral participation remains stratified even in the more egalitarian countries, the degree of stratification is far less severe than in countries that invest very little in their working-age populations. At the upper end of the welfare spending scale, individuals from the lowerincome groups seem to be catching up to their wealthier counterparts. These results do, however, require further testing. When dealing with a relatively small number of countries, single Level 2 units can quickly exert a large effect on country-­level effects. Further analyses not presented here, wherein Model 6 was run repeatedly, each time excluding one of the 26 countries, yielded results similar to those obtained with all countries. Nor is the model sensitive to the inclusion of Israel, whose status as an advanced democracy might be called into question. Moreover, when omitting the variables age of democracy and left government strength (separately and together), the model does not yield substantially different results.

Voting Patterns in Welfare States: Conclusions The previous analyses of individual voter turnout show two things: first, that the welfare state policies focused on working-age adults and families have a positive and significant direct effect on the individual likelihood of having voted in an election in the 26 OECD countries under consideration. This finding corroborates the hypothesis that increased welfare state spending has a bolstering effect on electoral participation due to the personal stakes such policies create and the increased overall salience of politics associated with a larger welfare state. While such a finding by

  The Impact of Social Policy on Democratic Citizenship 

  95

itself is of course an interesting one, it is perhaps even more interesting to examine the micro-micro relationships behind this micro-macro relationship. To do so, I examined the interaction between welfare state spending and income to investigate whether the welfare state moderates the link between monetary resources and one’s propensity to vote. In other words, do we find higher levels of turnout in countries with more generous welfare states because the wealthier are participating at overall higher levels or does the overall higher level reflect that “a rising tide lifts all boats?” The results of the logistic multilevel model with cross-level interactions between the income categories and the welfare state variable show the latter—not only does this bolstering effect span across social milieu, but it functions most strongly for the lowest-income group. That is, a strong commitment to working-age adults and families moderates the strength of the relationship between income and voting. For those individuals possessing greater socioeconomic resources, the effects of additional welfare spending are much less pronounced. It is, however, difficult to argue that we see this bolstering effect of welfare spending on the turnout of low- and middle-income earners because they are the recipients of such benefits, as in many societies, particularly the Nordic countries, cash allowances and benefits in kind are allotted on a universal and not a means-tested basis. For low- and middle-­ income individuals, however, welfare policies aimed at working-age adults and families are likely to have greater importance and relevance for their well-being daily lives, as it is precisely this type of spending that is likely to lessen some of the financial burden of raising a family or ­maintaining one’s standard of living during bouts of unemployment. While spending on family and unemployment can of course also be beneficial to high earners, it can be argued that it is less critical for their chances in life than such expenditures would be for the lower portion of the income distribution. If we believe that high levels of political participation are necessary for democratic health and vitality, not to mention matters of representation, these analyses are, in general, able to report something positive—that individuals in countries that invest in and spend more on working-age adults and families are more likely to participate in elections. More importantly, however, the results indicate that generous welfare states also

96 

J. Shore

have lower political inequality by reducing the dependence of political participation on individual economic resources. Despite this initial positive finding, we, nevertheless, see that resources, mainly in terms of income, continue to play a crucial role for the individual decision to vote. Although the participatory gap between rich and poor is narrower in countries with high levels of social spending on families and their working-­age populations, it is clear that more can be done to alleviate the political inequalities that all too often accompany economic inequalities. Whose voices get heard and whose remain quiet is indeed a cause for concern. Unequal participation and the resulting unequal representation have the potential to undermine the legitimacy of electoral and legislative outcomes. As the American Political Science Association’s Task Force on Inequality and American Democracy pointed out a decade and a half ago: “[c]itizens with lower or moderate incomes speak with a whisper that is lost on the ears of inattentive government officials, while the advantaged roar with a clarity and consistency that policy-makers readily hear and routinely follow” (American Political Science Association Task Force 2004, 651).

Political Interest What Is Political Interest and Why Does It Matter? Although we know that political interest is very important in explaining numerous aspects of democratic engagement, political scientists have shown relatively little interest in this topic (Macedo 2005, 35). Political interest, more so than education or political knowledge, can explain why people get involved in political activities (Verba et al. 1995). Moreover, it is extremely important in explaining one’s level of political sophistication (Luskin 1990) and has been shown as the main predictor of voting, party mobilization, and organizational involvement (Miller and Rahn 2002). In other words, political interest is an important precursor to a number of political activities. While it is fairly straightforward how education can enhance a person’s political interest—by enhancing one’s political efficacy or the ability to understand what can at times be a very

  The Impact of Social Policy on Democratic Citizenship 

  97

complicated political world (van Deth 1990)—the ways in which income is related to political interest and how welfare state policies may moderate this relationship do require further investigation. In this section, I empirically examine how welfare state policies can impact individuals’ interest in politics. The research design of the following analyses essentially mirrors the voting analysis: I again draw on logistic multilevel models with cross-­level interactions to examine how the context of a welfare state focus on working-age adults and families can moderate the relationships between income and political interest in 26 OECD countries. Political interest is frequently defined as “the degree to which politics arouse a citizen’s curiosity” (van Deth 1990, 278) and implies attentiveness as well as some sort of willingness to participate (van Deth 2000, 119). Political interest is, however, often overlooked empirically or taken as a given in the study of political behavior. The reasons to study political interest, however, are manifold, not in the least because it is seen as the necessary precursor to so many other forms of political engagement. At the same time, although political interest is thought to be a prerequisite to further, more demanding forms of political behavior, we cannot say that a mere interest in politics is equivalent to other more active forms of behavior. As Verba and Nie (1972) pointed out in their pioneering cross-­ national study of political behavior, political participation indeed implies something active beyond political attitudes and cognitive dispositions. In other words, “taking an interest in something” and “doing something” are clearly distinct activities (van Deth 1990; see also Macedo 2005; van Deth 2000). Moreover, not all who are interested actually participate. At the same time, voters may go to the polls on Election Day out of a sense of civic duty and not because they are particularly interested in politics (Campbell 2006). Nevertheless, political interest is an important ingredient for a healthy, stable democracy: “Without a minimum level of subjective political interest citizens would not even be aware of the political process or of the opportunities to defend their well-being, to contribute to collective actions, or to select a representative” (van Deth 2000, 119). In addition to political interest as curiosity or attentiveness, there exist concepts relating political interest to political saliency or politics’ relative importance: “Even if political phenomena arouse curiosity and concern among citizens, it is not obvious that politics is considered

98 

J. Shore

to be importance or even relevant for their lives” (ibid.). These two conceptions of political interest need not, however, be mutually exclusive; it is entirely conceivable that political interest could reflect both a curiosity in political matters and a feeling that politics are salient. What is known about political interest is that it is strongly linked to resources—both education and income (Brady et  al. 1995)—and that political socialization also plays a considerable role (e.g., Bennett and Bennett 1989; Hooghe 2004). Experiences with politics that people have prior to adulthood at home or in school partially shape how one thinks about and engages with politics (Luskin 1990). As, however, Macedo and his coauthors (2005, 35) point out, “pre-adult experiences” are not the only determinants of people’s political interest: between the 1992 and 1996 US presidential elections, the percentage of Americans reporting being “very interested” in politics dropped by 14 points; clearly, “preadult” experiences cannot account for such a dramatic shift in political interest. The political environment and the stimulation it provided (or lack thereof ), they argue, must have also played a role. This observation hearkens back to E.E. Schattschneider’s (1960) analysis of the American withdrawal from politics. He confronted the prevailing sentiment that an apathetic and uninformed citizenry was to blame for its lack of participation, arguing that the reasons why the lower classes abstained from politics were not due to any sort of moral or civic deficiencies; rather, the political system itself was to blame, for it did not offer much to the public to mobilize, engage, or even encourage political participation (Mettler and Soss 2004). In line with the discussions on policy feedback effects in Chap. 3, it can be assumed that people will be more likely to be interested in politics when politics actually has something to offer them. That is, when political outcomes are relevant and meaningful for people’s lives, they will be more likely to find politics to be salient and interesting. Generous welfare policies are clearly not only relevant for many people because of the implications for their own well-being and life chances, but they also have an overall redistributive effect. In his study of income inequality and political participation, Solt (2008, 49) argues that high levels of economic inequality reduce political engagement across all income strata but especially among poorer people due to power imbalances: “If a country’s

  The Impact of Social Policy on Democratic Citizenship 

  99

income and wealth are more concentrated, power within the country will be more concentrated; that is, where rich individuals are richer relative to poorer individuals, they will be more powerful relative to these poor individuals as well” (see also Goodin and Dryzek 1980). Not only should overall political interest be higher in countries with welfare policies clearly aimed at reducing economic inequality among working-age adults and families but the interest differences between income groups should also be reduced due to the fact that the issues affecting the less well-off citizens actually reach the political agenda. Generous welfare policies furthermore send a message to the citizenry that their interests are valued and protected, as such policy foci tend also to be universalistic and broadly inclusive in nature (e.g., Mettler and Soss 2004). If the interests of the less advantaged are not given space on the political docket and thus remain underrepresented, it can be assumed that poorer individuals may have little reason to be interested in politics. Moreover, we may also witness a depressing effect in terms of wealthier people’s political interest in societies where there is very little spent on policies aimed at alleviating and/or preventing distress among the age population: wealthier individuals may need to be less working-­ involved or interested in politics because they do not have to defend their interests against those of the lower socioeconomic status groups.8 In terms of the specific policies examined in the following analyses of political interest, I assume that a direct effect of welfare state spending may be found, as even the wealthier are (were) likely to be (or have been) recipients of some sort of family benefits. Poorer- and middle-class individuals are more likely to depend on these types of benefits or services and would, therefore, be interested in politics due mainly to importance and relevance such policies have for their everyday lives. If an indirect effect is to be found, further examination of the indirect effects (by means of cross-level interactions between income groups and welfare state spending) are needed. As mentioned, due to the already high levels of resources that wealthier individuals have at their disposal, their predicted 8  This line of argumentation is highly speculative; Schattschneider (1960, 105–107), however, added that even in the face of high inequality, the wealthier will continue to be motivated to participate in politics because conflicts of interest among the upper classes remain (see also Solt 2008).

100 

J. Shore

levels of political engagement are already relatively high. Additional welfare state spending, therefore, may not substantially affect their already high predicted levels of political interest. Individuals from the lower-­ income groups, however, are expected to be positively affected by increased welfare spending on working adults and families, as such policies are potentially very relevant and important for their lives and livelihoods. Moreover, as this type of policy focus represents a commitment to the social rights of all members of a society, it is assumed that politics will embody something that is more inclusive and engaging and less alienating of weaker members. Finally, we know from a growing body of works (e.g., Mau 2003; Rothstein 1998; Svallfors 2007), that depending on the design of welfare policy, “welfare systems embody and create norms about what is fair and just, and form citizens’ expectations and demands” (Oskarson 2010, 5). In sum, where policies intervene in meaningful ways in people’s everyday lives, we can expect greater political interest due to the increased saliency of politics; where these intervening policies embody norms of fairness and a commitment to alleviating economic-based inequality, we can expect that the positive impact on political interest will be the greatest for those who stand to benefit the most, that is, individuals at the lower portion of the income distribution.

Political Interest in 26 OECD Countries To investigate the impact of welfare state policy on individual political interest, I utilize the same basic research design as in the previous analyses of voting: the same 26 OECD countries are included in the analysis, I employ the same operationalization of the welfare state, and also rely on multilevel logistic models. I, therefore, refrain from going through the details of the research design again in this and in the subsequent analyses. Data on political interest as well as the individual-level attributes also come from the 2006 and 2008 waves of the ESS as well as from the 2005 WVS. The question gauging political interest, which is exactly the same in both surveys, is: how interested would you say you are in politics? Respondents are given four response options: 1—very interested; 2—somewhat interested; 3—not very interested; and 4—not at all interested. The

  The Impact of Social Policy on Democratic Citizenship 

  101

data have been recoded to form a binary variable, wherein people stating that they are very interested and those who are somewhat interested were assigned the value of 1; respondents who are not very interested and not at all interested received the value of 0. The sample consists of 36,274 adults in 26 democracies.9 When all respondents in all countries are taken together, interest in politics is nearly evenly split: around 50.1 percent responded that they are somewhat or very interested in politics; 49.1 percent tend toward political disinterest. These aggregated figures do not, however, tell much of a story. In Fig. 5.2, we see that the highest reported political interest can be found in Denmark, with around 75 percent of respondents stating an interest in politics. The Czechs appear to be the least interested, with only about 19  percent of respondents indicating political interest. Although political interest would at first appear to be evenly distributed across all ­respondents, Fig. 5.2 clearly shows that there are vast differences between countries with regard to the proportion of respondents who are somewhat or very interested in politics. There is a difference of around 56 percentage points between the share of respondents in the country with the highest reported political interest (Denmark) and the country with the fewest individuals interested in politics (Czech Republic). In Table 5.7, we can see that levels of political interest are considerably lower for individuals in the lowest-income quintile: 41.2 percent compared to 60.6 percent in the highest-income group. Portugal, the country with the third lowest overall levels of political interest, has the lowest level of political interest in the lowest-income category: only 14.2 percent of individuals reported being interested in politics. In all but one country— Slovakia—greater income is associated with greater interest in politics. The figures in the table and the comparisons of columns immediately reveal some interesting patterns: For example, low interest is not always associated with higher polarization between low- and high-income groups, nor is high overall political interest always accompanied by lower polarization.  Although the same data sources and countries are used for the voting and political analyses, the number of respondents differs due to the varying number of missing values on the two outcome variables.

9

102 

DK NL DE US CH AU SE GB AT FR SI CA BE FI NO IE EE IL PL IT SK HU GR PT ES CZ

J. Shore

19

27

31 30

40 39

47 47 46 45 43

53 53 51 50 50 50 49

63 62 61 59 59 58

68

75

Fig. 5.2  Political interest in 26 OECD countries. Notes: Percentage of respondents indicating that they are either very interested or somewhat interested in politics. Source: European Social Survey, rounds 3 and 4; World Values Survey, wave 4. Country codes: AT: Austria; AU: Australia; BE: Belgium; CA: Canada: CH: Switzerland; CZ: Czech Republic; DE: Germany; DK: Denmark; EE: Estonia; ES: Spain; FI: Finland; FR: France; GB: United Kingdom; GR: Greece; HU: Hungary; IE: Ireland; IL: Israel; IT: Italy; NL: Netherlands; NO: Norway; NZ: New Zealand; PL: Poland; PT: Portugal; SE: Sweden; SI: Slovenia; SK: Slovakia; US: United States

A series of multilevel logistic models will help illuminate the differences in political interest across countries and between income groups described earlier. As in the previous analysis, the first model, an empty model with only a random intercept, allows us to see how much of the total variance in political interest can be attributed to differences between the countries. The successive models will first introduce individual-level determinants of political interest, then the welfare state variable (public spending on working-age adults and families as a percentage of the GDP), further country-level variables, and finally, cross-level interactions between the welfare state variable and income (both as a continuous and as dummy-group variables) are introduced to test whether welfare policies moderate the relationship between income and the likelihood of being interested in politics.

  The Impact of Social Policy on Democratic Citizenship 

  103

Table 5.7  Political interest by income group

Country Australia Austria Belgium Canada Czech Republic Denmark Estonia Finland France Germany Greece Great Britain Hungary Ireland Israel Italy Netherlands Norway Poland Portugal Slovakia Slovenia Spain Sweden Switzerland United States Mean

Pol. % Reporting interest lowest political quintile interest

Pol. interest mid-level income

Pol. interest highest quintile

Difference (highest vs. lowest)

59.43 53.31 50.03 50.18 19.11

52.11 49.21 31.90 45.76 17.81

61.72 50.53 42.10 46.81 19.45

64.10 63.03 57.21 55.13 23.53

11.99 13.85 25.31 9.37 5.72

74.87 46.57 49.85 52.74 63.37 31.38 57.64

69.57 40.94 45.47 42.07 51.01 26.18 47.53

71.79 47.93 49.55 50.35 64.40 29.22 59.05

80.36 49.32 54.26 65.76 78.90 42.76 68.13

10.79 8.38 8.79 23.69 27.89 16.58 20.60

38.59 46.87 46.45 43.20 68.31 49.41 45.15 29.94 40.30 50.71 27.01 59.40 61.25 62.03

31.99 36.54 34.96 38.35 53.59 41.96 35.69 14.20 31.29 43.28 19.58 51.17 53.11 57.34

39.53 43.13 51.63 41.44 65.08 46.84 42.67 32.20 42.66 50.18 29.84 59.59 62.58 58.43

45.49 51.50 58.35 50.24 74.94 53.95 51.17 41.89 30.77 65.79 44.24 61.62 74.13 75.00

13.50 14.96 23.66 11.89 21.35 11.99 15.48 27.69 -0.52 22.51 24.66 10.45 21.02 17.66

50.01

41.16

47.69

60.61

18.99

Sources: World Values Survey, wave 4; European Social Survey, rounds 3 and 4

The individual-level variables included in the models are the same as those in the voting model: age, union membership, employment status, gender, educational level, area of residence, and income level. At the country level, I furthermore control for the disproportionality of the electoral system, age of the democracy (see previous chapter for discussion), and strength of left parties. In terms of the first country

104 

J. Shore

control—disproportionality—we can draw on the assumptions regarding its effect on voter turnout and apply them to political interest. It has been well documented that proportional representation (PR) systems produce higher turnout (e.g., Blais 2006; Jackman and Miller 1995) due to the enhancing effects that a more proportional system creates. efficacy-­ Proportionality has also been shown to strengthen party preferences. Because PR systems tend to produce a greater number of choices in terms of interest representation, interest in politics is assumed to be higher in countries with more proportional electoral systems (Karp and Banducci 2008). Including a country’s age as well as the strength of its left-governing parties allows us to control for further important differences that could indeed affect interest in a way similar to voting. A complete overview of these variables, their operationalizations, summary statistics, and sources can be found in the Appendix.

Political Interest in Welfare States: Empirical Results In this section, I present the findings of the logistic multilevel models. Beginning with the empty model (Model 1 in Table 5.8), we see that the proportion of the variance in the probability to be interested in politics that can be attributed to differences between countries is around 8.3 percent. Adding the individual-level variables (Model 2), the variance at the country level drops (from 0.546 to 0.480), as does the intraclass correlation—from 8.3 percent to 6.5 percent. The patterns of individual-level predictors of political interest are quite similar to those in the voting analysis. Political interest first increases with age and then drops slightly after a while, indicating a curvilinear relationship of age to political interest. Union members are not only more likely to be voters, but they are also much more likely to show some degree of political interest as compared to non-union members. Living in an urban area as well as increases in education are also strongly and positively linked to an increased likelihood of being interested in politics. Likewise, increasing income exerts a highly significant impact on political interest. The results of the individual-­ level model differ from the voting analysis in two main ways: first, we find a significant gender effect on political interest. Whereas gender was

  The Impact of Social Policy on Democratic Citizenship 

  105

Table 5.8  Multilevel analyses of political interest Political interest Constant Fixed effects Age (Mean-centered) Age squared Union membership (ref. = non-member) Employed (ref. = not employed) Gender (ref. = male) Education level (0–5) Urban (ref. = non-urban) Income decile

Model 1

Model 2

Model 3

−0.043 (0.108)

−1.367*** (0.104)

−1.646*** −1.815*** (0.269) (0.287)

0.022*** (0.001) −0.000*** (0.000) 0.207*** (0.027) −0.063** (0.029) −0.535*** (0.023) 0.228*** (0.025) 0.379*** (0.010) 0.064*** (0.005)

0.022*** (0.001) −0.000*** (0.000) 0.206*** (0.027) −0.063** (0.029) −0.535*** (0.023) 0.229*** (0.025) 0.379*** (0.010) 0.064*** (0.005) 0.050

0.022*** (0.001) −0.000*** (0.000) 0.206*** (0.027) −0.064** (0.029) −0.534*** (0.023) 0.229*** (0.025) 0.378*** (0.010) 0.064*** (0.005) 0.075*

(0.044)

(0.039) −0.019 (0.021) 0.000*** (0.000) −0.001 (0.002)

Welfare effort: working-age adults/families Disproportionality Age of democracy Strength of left Random effects Rho Context variance N

Model 4

0.083 0.065 0.062 0.041 0.546 0.480 0.468 0.377 36,274 (26) 36,274 (26) 36,274 (26) 36,274 (26)

Notes: Log-odds. Standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1

a non-significant factor in the individual-level voting model, being female is a clear and significant depressor of political interest. This gender gap in political interest is in line with previous findings that show it to be a rather persisting difference (Bennett and Bennett 1989; Burns 2002). The second major difference to the voting analysis regards the

106 

J. Shore

effect of employment. In terms of the propensity to vote, being employed was positively related; here, however, the status of being employed has significant negative effect on the tendency to be interested in politics. Adding the single welfare state variable to Model 3, we see slight decreases in the overall variance. The effect of spending on social policies benefiting working-age adults and families is positive, albeit not significant at traditional levels. Before, however, declaring that there is no substantial effect of welfare state spending on political interest, it is necessary to also test this variable against other country-level controls in order to account for the many differences found within this group of countries. As Model 4 depicts, when controlling for the level of disproportionality of the electoral system, the age of democracy, and the strength of left parties in government, the strength of the positive relationship between the welfare policy variable and the likelihood of being interested in politics increases. With these country-level attributes included in the model, we moreover see a moderately significant (p < 0.1) effect of social spending on working-age adults and families. Disproportionality exerts a negative effect on the probability of being politically interested and thus is in line with the theoretical expectations; this coefficient is, however, non-­ significant. The age of a democracy is furthermore positively and significantly related to the probability of being interested in politics. Strength of the left does not have any substantive effect on political interest. The final models (5 and 6 in Table 5.9) introduce cross-level interactions between the welfare state variable and income categories, first as a continuous variable and then with the income decile categories. As in the voting analyses, these cross-level interactions capture the micro-micro relationships: does welfare spending aimed at working-age adults and families moderate the strength of the relationship between economic resources and political interest? In Model 5, the single coefficient for the welfare state policy variable is significant for one’s propensity to be politically interested, as is the interaction with income (p < 0.1), albeit more modestly. The interaction coefficient is negative, indicating that the strength of the relationship between income and political interest is lower at high levels of welfare spending, and vice versa. Looking to Model 6, which tests whether the moderating effect of the welfare state on the relationship between income and political interest, we see that the single

Table 5.9  Multilevel analyses of political interest with cross-level interactions Political interest Constant Fixed effects Age (mean-centered) Age squared Union membership (ref. = non-member) Employed (ref. = not employed) Gender (ref. = male) Education level (0–5) Urban (ref. = non-urban) Income (deciles)

Model 5

Model 6

−1.976*** (0.305)

−1.756*** (0.296)

0.022*** (0.001) −0.000*** (0.000) 0.206*** (0.027) −0.047 (0.029) −0.538*** (0.023) 0.230*** (0.025) 0.385*** (0.010) 0.223*** (0.047)

0.022*** (0.001) −0.000*** (0.000) 0.206*** (0.027) −0.047 (0.029) −0.539*** (0.023) 0.229*** (0.025) 0.385*** (0.010)

Income (ref. = low) Mid-level High Welfare effort: working-age adults/families Disproportionality Age of democracy Strength of left Welfare × Income

0.090** (0.043) −0.020 (0.021) 0.000*** (0.000) −0.001 (0.002) −0.004* (0.002)

(0.041) −0.020 (0.021) 0.000*** (0.000) −0.001 (0.002)

0.042 0.381 36,274 (26)

0.042 0.381 36,274 (26)

Welfare × Mid-level income Welfare × High income Random effects Rho Context variance N

0.235*** (0.090) 0.451*** (0.095) 0.087**

Notes: Log-odds. Standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1

−0.009 (0.015) −0.010 (0.015)

108 

J. Shore

welfare state coefficient, which is the effect of welfare state spending on political interest for individuals in the very lowest-income category, is positive and significant at the 5 percent level. Compared to individuals in the lowest-income decile, the effect of additional welfare state spending on working-age adults and families on political interest is smaller for people in both the middle- and upper-income groups. The coefficient of the interaction term comprising welfare state spending and mid-level income is insignificant, indicating that the effect on the middle earners’ political interest is not significantly different from the effect found for people with the lowest incomes. The same pattern can be identified with regard to the effects of welfare spending on high earners’ political interest compared to their countrymen and women with lower incomes. In other words, we see from Model 6 that there do not appear to be income group-­specific effects concerning the moderating effect of welfare spending aimed at working-age adults and families on the propensity to be interested in politics. As in the previous analyses, I calculate the marginal effects of welfare state spending on political interest for each of the income groups, as the interaction terms cannot be interpreted as such. Recall that marginal effects estimate the extent to which the dependent variable political interest is expected to change, given a unit increase in the explanatory variable spending on working-age adults and families (Buis 2010). From Table 5.10, we can more clearly observe that there is a positive effect of social spending for all three income groups. Although Table 5.10  Marginal effects of welfare policy on political interest by income category Political interest—yes/no

Spending on working-age adults and families

Low income

0.087** (0.041) 0.078* (0.041) 0.077* (0.041)

Mid-level income High income

Notes: Standard errors in parentheses. Based on Model 6 in Table 5.9; all other values held at means ***p < 0.01; **p < 0.05; *p < 0.1

  The Impact of Social Policy on Democratic Citizenship 

  109

.6 .5 .4 .3

Probability of Political Interest

.7

we know from Model 6 that the effects are not substantively different from one another, it is clear that not only is the impact of welfare spending on political interest highest for people in the lowest-income group, this relationship is more significant (p < 0.05) among those between the middle- and high-income groups, respectively (p < 0.1). Figure 5.3 provides the clearest illustration of these effects. At the very lowest levels of welfare state spending, the probability of being interested in politics is also at its lowest. The lines for all three groups rise steadily as social spending increases. The gap between income groups also appears to narrow, ever so slightly, as can be seen from the minimally steeper slope for the line representing persons with the lowest incomes. In other words, while we do see a slight degree of convergence of the lines at the upper levels of spending, it is to a far lesser degree than we saw in terms of voting. Although political interest is less stratified in more generous welfare states and is overall greater, it remains clearly tied to income. These findings were furthermore confirmed when the models were re-run an additional 25 times, each time excluding one of the 26 countries in order to determine if the results were sensitive to a single case—something which

2

3

4

5

6

7

8

9

10

11

% of GDP Spent on Working-Age Adults and Families Low Income

Mid-Level Income

High Income

Fig. 5.3  Income level and predicted political interest depending on welfare state spending. Notes: Based on model 6 in Table 5.9; all other values held at means

110 

J. Shore

can easily occur with such a low number of Level 2 units. In particular, the exclusions of Israel and Slovakia (where a negative correlation between income and political interest exists) did not change the parameter estimates.

Political Interest: Conclusions This section examined a vital, yet oft-overlooked aspect of democratic citizenship: political interest. Political interest is considered a necessary antecedent to more active and demanding forms of political participation; at the same time, it is clearly something distinct and separate from activities such as voting. It can be thought of as a curiosity about or attentiveness to politics; it is the sentiment that politics represents something salient or relevant in one’s life. Welfare state policies aimed at working-­ age adults and families have been shown to have a positive direct effect on political interest. This seems to correspond to the literature wherein we find an increased relevance of politics in countries where the state is more prevalent in one’s daily life. Because of the important role of policies such as family allowances or unemployment benefits, people seem to more likely display some level of political interest, and this appears to hold true for people across all income levels. Although we do see higher levels of political interest at higher levels of social spending, this does not, however, result in a substantial reduction with regard to the socioeconomic stratification of interest in politics. While all income groups have higher levels of political interest in more generous welfare state settings, and the effect for the lowest earners is indeed the greatest and reaches the highest level of significance, we cannot really speak of an effect that is substantively different from the effects found for the middle- and highest-income earners. While much of the discussion here has been on the consequences of alienating less well-off citizens and the ways in which social policies may draw them into the political sphere, it is important to remember that wealthier individuals are also paying the largest share of the taxes that fund these programs, and, in universal systems, they too are the recipients and beneficiaries of generous family policies. Lastly, we can speak of policy feedback effects of a less tangible nature: policies that represent a clear

  The Impact of Social Policy on Democratic Citizenship 

  111

commitment to working-age adults and families, by reducing income inequality and representing the interests of society’s less well off, send messages to citizens. By embodying norms of fairness and universality, people are sent the message that the political arena is not just for the rich but rather something that all can be interested in, as political outcomes do in fact benefit all members of society.

Political Attitudes In addition to the various behaviors one can study when looking at democratic citizenship, numerous surveys have also supplied us with information on people’s attitudes and dispositions toward the state and democracy in general. Looking first at the ways in which the welfare state can shape people’s trust in parliament—the main decision-making institution in contemporary democracies—I analyze whether democracies that place a larger policy focus on working-age adults and families also have citizens who tend to exhibit greater political trust as well as whether trust has a more equitable distribution. Moving on to satisfaction with democracy, this analysis looks at whether and how the welfare state impacts the level and distribution of how satisfied people are with the way democracy in their country functions. As with the political participation sections, the investigations to follow do not constitute an exhaustive body of political attitudes but are two indicators of democratic support for which we have widely available data. Moreover, these two aspects on the attitudinal side of democratic citizenship tap into the ways in which people think and feel about democracy in their state. While not entirely unrelated, trust in political institutions and satisfaction with democracy reflect different types of evaluations and attitudes, as I explain in the following sections.

 emocratic Support: Trust in Parliament D and Satisfaction with Democracy More so than trust in other political institutions, trust in parliament has received a great deal of attention. A parliament is the central decision-­ making institution in a democracy, and as it is the body that creates the

112 

J. Shore

rules of the society, and since people are required to follow those rules, how much confidence citizens have in their parliament is indeed something worth studying. Because parliaments are long-standing and non-­ ephemeral parts of democracy, “[a]n erosion of confidence in the major institutions of society, especially those of representative democracy, is far more a serious threat to democracy than a loss of trust in other citizens or politicians” (Newton and Norris 2000, 52). In other words, by looking at trust in parliament, we may be able to get a good picture of how citizens generally feel about their democracies. Satisfaction with democracy, on the other hand, does not generally gauge whether people disagree with the principles of democracy but instead measures people’s support for how a democratic regime works. While trust has been found to tap into people’s ideas about whether democratic norms are desirable, satisfaction with democracy questions whether those norms are being fulfilled and whether political processes are aligned to the principles and norms of democracy. Satisfaction with democracy thus taps into people’s perceptions of regime performance—whether the practice of democracy, policy outputs, and their implementation function in accordance with the beliefs about democracy (Klingemann 1999, 55–56; Linde and Ekman 2003; Miller and Listhaug 1999, 205). And although people may not particularly like a party currently in government or the types of policies pursued, their trust in parliament and satisfaction with democracy correspond to what Easton (1975) termed “diffuse support”; that is, the normative expectations about democracy as a political system. And although one could think of trust in parliament as a long-term confidence in the political system and satisfaction with democracy as based more on short-­ term evaluations of regime performance in accordance with democratic principles and ideals, it can be argued that a persistent dissatisfaction with how democracy works in practice could very well chip away at support for democracy as a regime type and form of government (Klingemann 1999; see also Schäfer 2013). Thinking back to how we might define democracy, although many conceptualizations exist, at the core of even the most minimalist idea is a notion of political equality. Should, however, a democracy be perceived to not be representative of all citizens or fail to promote the democratic ideal of political equality, there is good reason to believe that support for

  The Impact of Social Policy on Democratic Citizenship 

  113

democracy will be lower, particularly among citizens with fewer resources, as it will be those citizens who will see the least representation in their regimes. The experience of poverty, or less severe economic hardship, is likely to depress confidence in political institutions and support of democracy, in general. If, however, policies exist that offer generous support to individuals and families who would otherwise be caught in the trap of poverty, the sentiment that a government is responsive to and representative of the needs of the many may arise, thereby increasing trust in democratic institutions and fostering the feeling that democratic norms exist and function. If constituents feel that the system outputs demonstrate that their voices are heard and their needs are addressed, they are more likely to have a supportive view of that system, with regard to both trust in parliament and satisfaction with democracy. Recall that the effect of greater welfare state spending on working-age adults and families is hypothesized to foster political support overall and, at the same time, provide the greatest boost to individuals whose relative economic position would otherwise predict less support, as economic disadvantage and its spillover into political disadvantage is likely to be conducive of less favorable views of both regime performance and, consequently, to declining support of a political system which may not appear to have the interests and needs of all at heart. The following analyses of political trust and satisfaction with democracy put these hypotheses to the empirical test.

 xamining the Effects of the Welfare State on Political E Trust Data on trust in parliament come from the ESS and includes 31,235 individuals in 22 European countries.10 Respondents were asked to indicate on a scale of 0–10 (with 0 indicating “no trust at all” and 10 “com The World Values Survey (WVS) asks a question about confidence in political institutions, including parliament. Not only is the wording of the question quite different, but the question also relies on an ordinal measurement scale. In order to ensure comparability across countries, I only include responses from the European Social Survey, as it covers the greatest number of countries with uniform question wording. 10

114 

J. Shore

DK FI CH NO SE NL AT ES IE DE BE FR SI SK GB IL EE GR PT CZ PL HU 0

2

4

6

8

10

Average level of trust in parliament

Fig. 5.4  Trust in parliament across countries. Notes: 0 = No trust at all; 10 = Trust completely. Source: European Social Survey, rounds 3 and 4. Country codes: AT: Austria; BE: Belgium; CH: Switzerland; CZ: Czech Republic; DE: Germany; DK: Denmark; EE: Estonia; ES: Spain; FI: Finland: FR: France: GB: Great Britain; GR: Greece; HU: Hungary; IE: Ireland; IL: Israel; NL: Netherlands; NO: Norway; PL: Poland; PT: Portugal; SE: Sweden; SI: Slovenia; SK: Slovakia

plete trust”) how much trust they have in their country’s national parliament. The average score across all countries on the 0–10 scale is relatively low—at 4.64. Looking to mean trust in parliament in individual countries, we see some familiar patterns, with the Danes and the Finns indicating the highest levels, with scores of 6.55 and 5.96, respectively, and the Hungarians and the Poles having the least trust in parliament (2.67 and 3.0, respectively). Figure 5.4 shows the mean political trust scores across all countries; Table 5.11 breaks down the average political trust by income group. From Table 5.11, we see that in all but one country (Ireland), average trust in parliament is higher among individuals with higher incomes than people with lower incomes. Despite this overall trend, the table does not reveal many discernible patterns with regard to income and political trust. In terms of variance, however, there is a greater spread of

  The Impact of Social Policy on Democratic Citizenship 

  115

Table 5.11  Political trust by income group

Country

Difference Political trust Political Average Political trust highest (highest political trust lowest mid-level vs. lowest) quintile income quintile trust

Austria Belgium Czech Republic Denmark Estonia Finland France Germany Greece Great Britain Hungary Ireland Israel Netherlands Norway Poland Portugal Slovakia Slovenia Spain Sweden Switzerland Mean

4.99 4.53 3.15 6.55 3.89 5.96 4.56 4.58 3.58 4.24 2.67 4.82 4.00 5.48 5.78 3.00 3.52 4.31 4.33 4.92 5.74 5.80 4.64

4.77 4.15 2.82 6.11 3.77 5.57 4.08 3.98 3.59 3.93 2.48 4.98 3.80 4.94 5.54 2.64 3.04 4.34 4.04 4.82 5.20 5.61 4.38

5.04 4.22 3.30 6.47 3.76 5.95 4.33 4.62 3.46 4.19 2.70 4.71 4.09 5.21 5.57 2.74 3.56 4.31 4.34 5.00 5.56 5.82 4.36

5.35 4.76 3.43 6.92 4.35 6.34 4.95 5.39 3.80 4.69 2.85 4.91 4.17 5.78 5.99 3.32 4.08 4.37 4.79 5.07 5.94 6.09 5.21

0.58 0.61 0.61 0.81 0.58 0.77 0.87 1.41 0.21 0.76 0.37 −0.07 0.37 0.84 0.45 0.68 1.04 0.03 0.75 0.25 0.74 0.48 0.83

Source: European Social Survey, rounds 3 and 4 Notes: Trust in national parliament (range 0–10; with 10 indicating complete trust)

responses within the lowest-income category compared to the highest: 6.43 versus 5.33. Whether these differences across countries and between income groups are systematic in nature is the subject of the following analyses.

Examining Political Trust in Europe As in the previous analyses, there are many individual attributes and demographic characteristics that should be taken into consideration when examining the impact of context on trust in parliament. In

116 

J. Shore

addition to the individual-level covariates of regime support, corresponding to the findings that both political trust and satisfaction with democracy increases with one’s level of resources, there are also institutional and contextual factors that have been shown to influence democratic support. First, the age of a regime has been shown to impact trust in parliament (van der Meer 2010). Increasing familiarity with democratic traditions has been shown to be conducive to trust and confidence in democratic institutions (McAllister 1999). Rose (1994) furthermore argues that it takes time for trust to develop (or redevelop) following experiences with non-democratic regimes, such as in the countries of the former Soviet Union. Second, I control for the effects of disproportionality of the electoral system. Although it might be argued that highly proportional ­systems are detrimental to political trust because they are the less accountable systems (Powell 2000) and the delineation of responsibility is less visible (Aarts and Thomassen 2008), studies have shown that political trust is in fact higher in PR systems than in majoritarian systems (Magalhaes 2006; van der Meer 2010). More proportional systems stimulate political trust because they are able to represent a larger share of the electorate and thus a broader spectrum of interests and political preferences (see also Karp and Banducci 2008). Lastly, I take the share of left parties in government into consideration, as not only do stronger left governments promote social spending on working-age adults and families, but, particularly when we are concerned with poorer constituents’ attitudes, it is important to control for the existence of parties that may be more aligned to their preferences. The measure of the welfare state—percentage of the GDP spent on working-age adults and families—remains the same. The analyses of trust in the parliament draw on hierarchical linear modeling techniques, as the trust variable, with responses from 0 to 10, does not need to undergo any transformation. The models in Table 5.12 provide us with an overall look at how welfare state spending influences political trust; Models 5 and 6 in Table 5.13 examine whether the relationship between income and political trust is muted by increased social spending on working-age adults and families.

  The Impact of Social Policy on Democratic Citizenship 

  117

Table 5.12  Multilevel analyses of trust in parliament Trust in parliament Constant

Model 1

Model 2

Model 3

Model 4

4.561*** (0.221)

3.358*** (0.141)

1.514*** (0.408)

2.077*** (0.525)

0.003*** (0.001) 0.000*** (0.000) 0.002 (0.029) 0.049 (0.033) −0.112*** (0.025) 0.196*** (0.011) 0.190*** (0.028) 0.079*** (0.006)

0.003*** (0.001) 0.000*** (0.000) −0.002 (0.029) 0.049 (0.032) −0.112*** (0.025) 0.195*** (0.011) 0.191*** (0.028) 0.078*** (0.006) 0.309***

0.003*** (0.001) 0.000*** (0.000) −0.002 (0.029) 0.049 (0.032) −0.112*** (0.025) 0.195*** (0.011) 0.191*** (0.028) 0.078*** (0.006) 0.215***

(0.064)

(0.079) −0.079** (0.039) −0.002 (0.004) 0.010** (0.004)

0.069 0.601 2.214 31,235 (22)

0.079 0.646 2.214 31,235 (22)

Fixed effects Age Age squared Union membership (ref. = non-member) Employed (ref. = not employed) Gender (ref. = male) Education level (0–5) Urban (ref. = non-urban Income (decile) Welfare effort: working-age adults/families Disproportionality Strength of left Age of democracy Random effects Rho Context variance Individual variance N

0.175 1.036 2.248 31,235 (22)

Notes: Standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1

0.072 0.616 2.214 31,235 (22)

118 

J. Shore

Table 5.13  Multilevel analyses of trust in parliament with cross-level interactions Trust in parliament Constant Fixed effects Age (mean-centered) Age squared Union membership (ref. = non-member) Employed (ref. = not employed) Gender (ref. = male) Education level (0–5) Urban (ref. = non-urban) Income (deciles)

Model 5

Model 6

2.466*** (0.557)

2.479*** (0.537)

0.003*** (0.001) 0.000*** (0.000) −0.004 (0.029) 0.045 (0.032) −0.114*** (0.025) 0.195*** (0.011) 0.194*** (0.028) 0.010 (0.017)

0.003*** (0.001) 0.000*** (0.000) −0.004 (0.029) 0.068** (0.032) −0.120*** (0.025) 0.204*** (0.011) 0.195*** (0.028)

Income (ref. = low) Mid-level High Welfare effort: working-age adults/families Disproportionality Age of democracy Strength of left Welfare × Income Welfare × Mid-level income Welfare × High income

0.151* (0.084) −0.079* (0.041) 0.010** (0.004) −0.002 (0.004) 0.011*** (0.003)

−0.088 (0.108) 0.105 (0.117) 0.179** (0.081) −0.079** (0.039) 0.010** (0.004) −0.003 (0.004)

0.044** (0.017) 0.059*** (0.017) (continued)

  The Impact of Social Policy on Democratic Citizenship 

  119

Table 5.13 (continued) Trust in parliament Random effects Rho Context variance Individual variance N

Model 5

Model 6

0.086 0.678 2.213 31,235 (22)

0.081 0.656 2.215 31,235 (22)

Notes: Standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1

Results: Political Trust We start out with a considerable amount of the total variance in political trust that can be attributed to differences between countries. Model 1, with an intraclass correlation of 0.175, reveals that 17.5% of the variance in individual political trust can be traced back to differences between countries. As in the previous analyses, we see the usual suspects in terms of the individual variables: those with more education, older individuals, employed people, urbanites, men, and people with greater levels of income have significantly greater trust in their national parliament. Although union members and employed persons also are found to exhibit greater political trust than non-members and non-employed, these effects are not significant. Adding the welfare state variable, we see that it too has a positive and highly significant impact on individuals’ confidence in parliament, which also holds when further country-level controls are added in Model 4 in Table 5.12. As predicted, higher levels of disproportionality exert a dampening effect on trust, and individuals are more trusting of their national decision-making body in older democracies. The strength of left parties in government does not add to the explanation of trust in parliament. Moving from the empty model to the full model results in a reduction in the proportion of the variance in political trust stemming from differences between countries from 17.5 percent to 7.9 percent. Models 5 and 6 in Table 5.13 introduce cross-level interactions between income (first as a continuous variable and then between the income categories—high, medium, and low) and welfare spending. In Model 5, we

120 

J. Shore

Table 5.14  Marginal effects of welfare policy on trust in parliament by income category Trust in parliament

Spending on working-age adults and families

Low income

0.179** (0.081) 0.223*** (0.080) 0.238*** (0.080)

Mid-level income High income

Notes: Standard errors in parentheses. Based on Model 6 in Table 5.13; all other values held at means ***p < 0.01; **p < 0.05

see that, contrary to the hypothesized conditional hypothesis, increased social spending on working-age adults and families interacts with income in a positive and highly significant manner. Here, we see that the strength of the trust-income relationship is stronger in countries where welfare spending on the aforementioned segment of the population is higher. Model 6 demonstrates that there are positive and significant effects for all individuals regardless of income quintile. This effect is smallest for persons in the first and second quintiles and increases, as shown in Model 5, with income category. The effects of social spending on trust in parliament furthermore differ according to income category, as indicated by the significant interaction terms. The marginal effects in Table 5.14 indeed reveal that increases in welfare state spending on working-age adults and families positively affect political trust for all income groups, with the largest effects found for people in the top-income quintile, followed by people in the middle-­ income group. These two marginal effects are significant at least at the 1 percent level; the marginal effect for people in the lowest-income group is the smallest. These findings deviate from the conditional hypothesis predicting the greatest boost in political trust for individuals with the lowest incomes. Despite the direct effect that welfare policies aimed at reducing economic inequality have on people’s trust in parliament, we cannot speak of an equalizing effect, as illustrated in Fig. 5.5. In Fig.  5.5, a modest fanning out can be observed: while at the lower end of the welfare state spending scale, the lines for individuals

  The Impact of Social Policy on Democratic Citizenship 

  121

Fig. 5.5  Income level and trust in parliament depending on welfare state spending. Notes: Based on model 6 in Table 5.13; all other values held at means

in the middle- and lowest-income groups converge, the lines for the highest and middle groups seem to increase at a fairly equal rate, with the line for the lowest-income earners lagging behind. Altogether, while greater welfare state spending increases individuals’ trust in parliament, it does so at slightly different rates depending on income, as observed by the fanning out at the highest levels of spending. We return to these seemingly counterintuitive results together with the findings from the analyses of satisfaction with democracy in the overall discussion of findings.

The Welfare State on Satisfaction with Democracy In analyzing satisfaction with democracy, I draw on individual-level data from CSES—Module 3, which allows one to examine micromacro links between political institutions and people’s perceptions and evaluations of democracy. The question gauging satisfaction with democracy is: “On the whole, are you very satisfied, fairly satisfied, not

122 

J. Shore

satisfied, or not at all satisfied with the way democracy works in [your country]?” While earlier versions of the ESS also include a question on democratic satisfaction, to avoid looking once again at only European states, CSES data are used in the following analyses. These questionnaires were administered directly, following elections between 2005 and 2011. The 18,443 respondents hail from the following 22 countries: Australia, Austria, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Iceland, Ireland, Israel, Netherlands, New Zealand, Norway, Poland, Portugal, Slovenia, Slovakia, Spain, Sweden, and Switzerland. The question on satisfaction with democracy has been recoded as a dummy variable, such that the value of 1 indicates that the respondent tends to be satisfied with democracy (including the “very satisfied” and “fairly satisfied” responses) and a 0 indicates a tendency of dissatisfaction (including the “not satisfied” and “not at all satisfied” response categories). Overall, 67.72 percent of respondents are fairly satisfied or very satisfied with the way democracy works in their country. Within countries, we again see some rather familiar patterns: 93.55 percent of Danes and 87.12 percent of Norwegians tend to be satisfied with democracy; only 21.11 percent of Slovenians and 26.95 percent of Greeks tend to be satisfied with democracy. Figure 5.6 furthermore illustrates the rather large differences in satisfaction with the functioning of democracy across countries. Furthermore, there are considerable differences within countries as well, as depicted in Table 5.15. In general, satisfaction with democracy is higher among people with higher incomes. Across all countries, the highest earners report being satisfied with democracy at a rate 11.72 percentage points higher than people in the bottom-income quintiles. A brief glance at the table also reveals that not only are there two cases where higher income is not associated with higher aggregate satisfaction but the middle income category does not always have a score between the lowest- and highestincome groups. In other words, unlike in the analysis of voting, the percentage of respondents indicating that they are satisfied with the functioning of democracy in their countries does not monotonically increase with income.

  The Impact of Social Policy on Democratic Citizenship 

DK NO AU CH SE ES IE NL FI IS AT NZ FR PL DE CZ EE PT SK IL SI GR

  123

93.55 87.12 86.75 81.85 81.77 81.55 80.71 80.39 75.59 71.51 70.78 66.67 65.03 58.82 54.39 49.30 48.63 41.29 37.37 34.22 27.11 26.95

% of respondents satisfied with how democracy functions

Fig. 5.6  Satisfaction with democracy across 22 countries. Source: CSES, Module 3. Country codes: AT: Austria; AU: Australia; CH: Switzerland; CZ: Czech Republic; DE: Germany; DK: Denmark; EE: Estonia; ES: Spain; FI: Finland; FR: France; GR: Greece; IE: Ireland; IL: Israel; IS: Iceland; NL: Netherlands; NO: Norway; NZ: New Zealand; PL: Poland; PT: Portugal; SE: Sweden; SI: Slovenia; SK: Slovakia

Satisfaction with Democracy in Welfare State As in the previous analyses, it is important to take individual-level characteristics and demographic attributes into consideration when examining satisfaction with democracy. In these analyses, I introduce an individual’s self-placement on a left-right scale of 0–10, with smaller values indicating a more leftist political leaning, and higher values a more rightist one, as those with more extreme political ideologies are likely to feel less represented and therefore less satisfied. The other individual-level variables mirror those in the previous analyses, with the assumption that individuals with greater resources tend to be more satisfied with how democracy works (Singh 2014). Due to the way income is coded in the original CSES dataset, income is only included as a categorical variable

124 

J. Shore

Table 5.15  Satisfaction with democracy by income group

Country Australia Austria Czech Republic Denmark Estonia Finland France Germany Greece Iceland Ireland Israel Netherlands New Zealand Norway Poland Portugal Slovakia Slovenia Spain Sweden Switzerland Mean

% Reporting satisfaction with dem.

Satisfaction with dem. lowest quintile

Satisfaction with dem. mid-level income

Satisfaction with dem. highest quintile

Difference (highest vs. lowest)

86.75 70.78 49.30

85.16 60.61 42.17

86.91 72.18 50.22

88.70 77.0 55.56

3.54 16.39 13.39

93.55 48.63 75.59 65.03 54.39 26.95 71.51 80.71 34.22 80.39 66.67

93.48 49.33 78.02 57.11 49.24 25.75 67.01 74.19 35.82 84.62 69.72

93.25 44.26 74.07 68.57 52.73 27.47 71.46 79.13 33.33 78.22 63.49

94.59 67.16 81.00 71.55 67.70 28.57 71.51 92.47 37.17 84.39 72.60

1.11 17.83 2.98 14.44 18.46 2.82 4.5 18.28 1.35 −0.23 2.88

87.12 58.82 41.29 37.37 21.11 81.55 81.77 81.85 67.72

80.95 50.24 35.09 33.33 20.83 86.54 81.01 76.56 63.12

88.07 59.32 42.70 39.52 27.51 81.71 80.96 85.37 67.64

87.12 66.33 26.32 33.33 43.48 75.47 85.21 78.03 74.84

6.17 16.09 −8.77 0 22.65 −11.07 4.2 1.47 11.72

Source: CSES, Module 3

(and not additionally as a continuous one). More information about the variables included in the analyses of satisfaction with democracy can be found in the Appendix. Furthermore, there has been a considerable amount of research about the effects of institutions on satisfaction with democracy. Lijphart (1999), for example, states that people in consensus democracies are much more satisfied than people living in majoritarian systems. The link between the type of democracy and satisfaction is thought to work through how people see their government in terms of representation and accountability (Anderson and Guillory 1997; Bernauer and Vatter 2012). Aarts and

  The Impact of Social Policy on Democratic Citizenship 

  125

Thomassen (2008) find that satisfaction with democracy is much more strongly linked to people’s views of representation than accountability, that is, in countries where people perceive their governments as representative, they were more likely to be satisfied with democracy. Representativeness, in turn, is a trait of proportional representation systems; majoritarian systems are seen as more accountable. For this analysis, it is, therefore, necessary to control for the effects of electoral disproportionality, as highly disproportional systems have been shown to produce lower levels of democratic satisfaction. Moreover, studies have found that the age of a democracy also matters for satisfaction with democracy, with representation perceived as lower in new democracies than in older ones (e.g., Aarts and Thomassen 2008). Furthermore, I control for the strength of the left, not only to account for its influence on social spending and constituency of poorer voters but, as Anderson and Singer (2008) find, left-leaning voters are much more dissatisfied in countries with high levels of income inequality.

Results: Satisfaction with Democracy From the empty model (Model 1 in Table 5.16) including only the random intercept, we first see that the proportion of the variance in satisfaction with democracy stemming from difference between countries is quite large—nearly 23 percent. After introducing individual-level explanatory factors in Model 2, this proportion of the variance drops only 0.4  percent. Of particular note are the large and significantly positive effects of income on satisfaction with democracy. Younger persons, women, the better educated, and people with more right-leaning political views are also more likely to be fairly or very satisfied with the way ­democracy works. While the employed and city dwellers also indicate a larger likelihood of satisfaction with democracy, these effects are not significant. When adding the single welfare state variable (Model 3), the variance proportion at the country level drops considerably—from 22.4 percent to 10.4 percent. The coefficient for the welfare state variable is moreover positive and highly significant. Of the country controls, only the age of

Table 5.16  Multilevel analyses of satisfaction with democracy Satisfaction with democracy Constant Fixed effects Age (mean-centered) Age squared Employed (ref. = not employed) Gender (ref. = male) Education level Urban (ref. = non-urban Political ideology Income (ref. = low) Mid-level High Welfare effort: working-age adults/families

Model 1

Model 2

Model 3

0.682*** (0.211)

−0.476** (0.229)

−2.750*** −2.920*** −3.015*** (0.436) (0.520) (0.533)

−0.001 (0.001) 0.000*** (0.000) 0.063 (0.043)

−0.001 (0.001) 0.000*** (0.000) 0.063 (0.043)

−0.001 (0.001) 0.000*** (0.000) 0.062 (0.043)

−0.001 (0.001) 0.000*** (0.000) 0.063 (0.043)

0.098*** (0.035) 0.080*** (0.017) 0.004 (0.037) 0.071*** (0.007)

0.097*** (0.035) 0.079*** (0.016) 0.005 (0.037) 0.071*** (0.007)

0.097*** (0.035) 0.077*** (0.017) 0.004 (0.037) 0.071*** (0.007)

0.097*** (0.035) 0.077*** (0.017) 0.004 (0.037) 0.071*** (0.007)

0.239*** (0.044) 0.418*** (0.063)

0.239*** (0.044) 0.419*** (0.063) 0.393***

0.241*** (0.044) 0.421*** (0.063) 0.356***

0.366** (0.160) 0.550** (0.227) 0.373***

(0.070)

(0.069) −0.003 (0.003) 0.007** (0.003) 0.019 (0.043)

(0.072) −0.003 (0.003) 0.007** (0.003) 0.018 (0.043) −0.022

Strength of left Age of democracy Disproportionality Welfare × Midlevel Income Welfare × High income

Model 4

Model 5

(0.027) −0.023 (0.038)

Random effects Rho Context variance N

0.228 0.224 0.104 0.083 0.083 0.985 0.975 0.619 0.546 0.546 18,443 (22) 18,443 (22) 18,443 (22) 18,443 (22) 18,443 (22)

Notes: Log-odds. Standard errors in parentheses ***p < 0.01; **p < 0.05; *p < 0.1

  The Impact of Social Policy on Democratic Citizenship 

  127

democracy shows any substantial effect, with citizens in older democracies exhibiting higher satisfaction than those in newer ones. Neither disproportionality nor strength of left parties seems to contribute much to the explanation of individual satisfaction with democracy. The cross-level interactions are introduced in Model 5 and test the conditional hypothesis that additional welfare spending on working-age adults and families are predicted to be the highest for individuals in the lowest-income category. The coefficient for the single welfare state variable reveals that there is a significantly positive effect to be found for people in the lowest-­ income quintiles with regard to their propensity to be satisfied with democracy. While the interaction terms both have negative coefficients, indicating that the effects for people with greater incomes are weaker than the effect for the poorest respondents, these effects cannot be considered to be substantially different than the aforementioned effect. The marginal effects in Table 5.17 present a similar picture: For individuals in all three income groups, satisfaction with democracy is positively impacted by increase in welfare spending. This relationship is highly significant at least in the 1 percent level. Although the marginal effects for each group are not substantively different from one another, the low-­ income group does, however, seem to reap the greatest benefits from increases in welfare generosity with regard to their satisfaction with the way democracy works. We also see in Fig. 5.7 that this slightly greater marginal effect for individuals from the lowest-income quintile results in an equalizing effect, Table 5.17  Marginal effects of welfare policy on satisfaction with democracy by income category Satisfaction with democracy

Spending on working-age adults and families

Low income

0.073*** (0.013) 0.066*** (0.013) 0.063*** (0.013)

Mid-level income High income

Notes: Standard errors in parentheses. Based on Model 5 in Table 5.16; all other values held at means ***p < 0.01

128 

J. Shore

Fig. 5.7  Income level and satisfaction with democracy depending on welfare state spending. Notes: Based on model 5  in Table 5.16; all other values held at means

with barely any discernible stratification by income for the probability of being satisfied with democracy in states where a greater percentage of the GDP is spent on working-age adults and families. Furthermore, we see much steeper slopes for all three income groups than in the previous analyses, as also indicated by the comparatively larger marginal effects.

Political Support and the Welfare State These findings, which conform to the hypothesized relationship between greater social expenditures on working-age adults and families and increased support for democracy, resonate with the literature on satisfaction with democracy and representativeness: in countries where we find a greater effort made to reduce economic inequality by means of more generous expenditures for working-age populations, we also find higher and more equitable levels of satisfaction with democracy. This could very well indicate that in those states people respond positively to the increased offerings by their governments, offerings which, in line with the policy

  The Impact of Social Policy on Democratic Citizenship 

  129

feedback literature, should have integrative effects. Not only is social citizenship expanded to include all social milieu, but by doing so, people feel their needs and concerns are being addressed by the state. Furthermore, as satisfaction with democracy also taps into people’s satisfaction with policy outputs, what these findings may very well indicate is that expenditures aimed at working-age adults and families make people more satisfied with regime performance. One might, nevertheless, argue that people in states tend to be more satisfied with democracy where overall levels of social spending is higher, such as in the Nordic countries, where we not only find a strong commitment to working-age populations but overall higher levels of social spending (see Table 4.1 in Chap. 4). For example, people might simply be more satisfied with policy outputs because of the general system of social welfare offered in generous universal welfare states. To test this possibility, I reran the final model (Model 5 in Table 5.16) with the cross-level interactions between income categories, substituting working-age and family expenditures with total social expenditures as a percentage of the GDP. I also tested the findings using health expenditures and old-age expenditures as other possible drivers of citizen satisfaction. As can be seen from the marginal effects in the table below, there does not seem to be empirical support for the argument that satisfaction with democracy is simply part of an overall large welfare state. Table  5.18 presents the marginal effects of different types of welfare spending on the satisfaction with democracy for individuals in the three income categories. Table 5.18  Marginal effects on satisfaction with democracy, various welfare state indicators Satisfaction with democracy Low income Mid-level income High income

Total social expenditures (% of GDP)

Health expenditures (% of GDP)

Old-age expenditures (% of GDP)

0.012 (0.007) 0.013* (0.007) 0.013** (0.007)

0.046 (0.037) 0.046 (0.035) 0.054 (0.034)

−0.015 (0.012) −0.009 (0.012) −0.008 (0.012)

Notes: Standard errors in parentheses **p < 0.05; *p < 0.1

130 

J. Shore

The findings in Table 5.18 resonate with what Kuitto (2011) in essence describes as worlds of spending. As discussed in Chap. 4, high spending in one area does not necessarily correspond to high spending across all areas. Moreover, this analysis suggests that there are differential outcomes of the different spending categories not only with regard to the reduction of economic inequality, but that these different pillars of the welfare state also bring about different effects in terms of satisfaction with the way democracy works in a given political context. While total social expenditures, which of course also include spending on working-age adults and families, do have a positive effect on satisfaction with democracy for all income groups, the total spending does not do anything to combat the social gradient. Furthermore, these marginal effects are weak at best. Health expenditures also have positive effects on satisfaction with democracy but are all insignificant. Lastly, expenditures related to old age even have a negative effect on satisfaction with democracy (albeit insignificant), perhaps providing further evidence to the argument that welfare states do indeed have clear spending priorities, and that there are often trade-offs with regard to age and where in the life cycle the social policy focus can be found. The additional models from which these marginal effects were derived can be found in the Appendix.

References Aarts, Kees, and Jacques Thomassen. 2008. Satisfaction with Democracy: Do Institutions Matter? Electoral Studies 27 (1): 5–18. https://doi.org/10.1016/j. electstud.2007.11.005. Abelson, Robert P., Elizabeth F.  Loftus, and Anthony Greenwald. 1992. Attempts to Improve the Accuracy of Self-reports of Voting. In Questions About Questions: Inquiries into the Cognitive Bases of Surveys, ed. Judith M. Tanur, 138–153. New York: Russell Sage Foundation. American Political Science Association Task Force. 2004. American Democracy in an Age of Rising Inequality. Perspectives on Politics 2 (4): 651–666. https:// doi.org/10.1017/S153759270404040X. Anderson, Christopher J., and Christine A. Guillory. 1997. Political Institutions and Satisfaction with Democracy: A Cross-National Analysis of Consensus and Majoritarian Systems. The American Political Science Review 91 (1): 66–81. https://doi.org/10.2307/2952259.

  The Impact of Social Policy on Democratic Citizenship 

  131

Anderson, Christopher J., and Matthew M. Singer. 2008. The Sensitive Left and the Impervious Right: Multilevel Models and the Politics of Inequality, Ideology, and Legitimacy in Europe. Comparative Political Studies 41 (4/5): 564–599. Armingeon, Klaus, David Weisstanner, Sarah Engler, Potolidis Panajotis, Gerber Marlène, and Philipp Leimgruber. 2011. Comparative Political Data Set 1960–2009. Berne: Institute of Political Science, University of Berne. Belli, Robert F., Michael W.  Traugott, Margaret Young, and Katherine A.  McGonagle. 1999. Reducing Vote Overreporting in Surveys: Social Desirability, Memory Failure, and Source Monitoring. The Public Opinion Quarterly 63 (1): 90–108. http://www.jstor.org/stable/2991270. Bennett, Linda, and Stephen E. Bennett. 1989. Enduring Gender Differences in Political Interest: The Impact of Socialization and Political Dispositions. American Politics Quarterly 17 (1): 105–122. Bernauer, Julian, and Adrian Vatter. 2012. Can’t Get No Satisfaction with the Westminster Model? Winners, Losers and the Effects of Consensual and Direct Democratic Institutions on Satisfaction with Democracy. European Journal of Political Research 51 (4): 435–468. https://doi. org/10.1111/j.1475-6765.2011.02007.x. Blais, Andre. 2006. What Affects Voter Turnout? Annual Review of Political Science 9: 111–125. Blais, Andre, and R. Kenneth Carty. 1990. Does Proportional Representation Foster Voter Turnout? European Journal of Political Research 18 (2): 167–181. https://doi.org/10.1111/j.1475-6765.1990.tb00227.x. Blais, Andre, and Agnieszka Dobrzynska. 1998. Turnout in Electoral Democracies. European Journal of Political Research 33 (2): 239–261. https:// doi.org/10.1111/1475-6765.00382. Brady, David. 2003. The Politics of Poverty: Left Political Institutions, the Welfare State, and Poverty. Social Forces 82 (2): 557–588. https://doi. org/10.1353/sof.2004.0004. Brady, Henry E., Sidney Verba, and Kay Lehman Scholzman. 1995. Beyond SES: A Resource Model of Political Participation. American Political Science Review 89 (2): 271–294. Brambor, Thomas, William Roberts Clark, and Matt Golder. 2006. Understanding Interaction Models: Improving Empirical Analyses. Political Analysis 14 (1): 63–82. Buis, Maarten L. 2010. Stata Tip 87: Interpretations of Interactions in Non-­ linear Models. The Stata Journal 10 (2): 305–308.

132 

J. Shore

Burns, Nancy. 2002. Gender: Public Opinion and Political Action. In Political Science: State of the Discipline, ed. Ira Katznelson and Helen V.  Milner, 462–487. New York: Norton. Campbell, David E. 2006. Why We Vote: How Schools and Communities Shape Our Civic Life. Princeton, NJ: Princeton University Press. Dalton, Russell J. 2000. Citizen Attitudes and Political Behavior. Comparative Political Studies 33 (6–7): 912–940. https://doi.org/10.1177/ 001041400003300609. Easton, David. 1975. A Re-Assessment of the Concept of Political Support. British Journal of Political Science 5 (4): 435–457. https://doi.org/10.1017/ S0007123400008309. ESS Round 3: European Social Survey Round 3 Data. 2006. Data file edition 3.6. NSD—Norwegian Centre for Research Data, Norway—Data Archive and distributor of ESS data for ESS ERIC. ESS Round 4: European Social Survey Round 4 Data. 2008. Data file edition 4.4. NSD—Norwegian Centre for Research Data, Norway—Data Archive and distributor of ESS data for ESS ERIC. European Social Survey. 2010. ESS Round 4 (2008/2009): Final Activity Report. London: Centre for Comparative Social Surveys: City University London. Foley, Michael W., and Bob Edwards. 1996. The Paradox of Civil Society. Journal of Democracy 7 (3): 38–52. Franklin, Mark N. 1997. Electoral Participation. In Comparing Democracies: Elections and Voting in Global Perspective, ed. Lawrence LeDuc, Richard G.  Niemi, and Pippa Norris, 216–235. Thousand Oaks, CA: Sage Publications. ———. 1999. Electoral Engineering and Cross-National Turnout Differences: What Role for Compulsory Voting? British Journal of Political Science 29 (1): 205–216. http://www.jstor.org/stable/194302. Goodin, Robert E., and John Dryzek. 1980. Rational Participation: The Politics of Relative Power. British Journal of Political Science 10 (3): 273–292. Gray, Mark, and Miki Caul. 2016. Declining Voter Turnout in Advanced Industrial Democracies, 1950 to 1997. Comparative Political Studies 33 (9): 1091–1122. https://doi.org/10.1177/0010414000033009001. Hicks, Alexander M., and Duane H.  Swank. 1992. Politics, Institutions, and Welfare Spending in Industrialized Democracies, 1960–82. The American Political Science Review 86 (3): 658–674. http://www.jstor.org/stable/1964129. Highton, Benjamin. 2005. Self-Reported Versus Proxy-Reported Voter Turnout in the Current Population Survey. Public Opinion Quarterly 69 (1): 113–123. https://doi.org/10.1093/poq/nfi003.

  The Impact of Social Policy on Democratic Citizenship 

  133

Hooghe, Marc. 2004. Political Socialization and the Future of Politics. Acta Politica 39 (4): 331–341. https://doi.org/10.1057/palgrave.ap.5500082. Hooghe, Marc, and Koen Pelleriaux. 1998. Compulsory Voting in Belgium: An Application of the Lijphart Thesis. Electoral Studies 17 (4): 419–424. Hox, Joop. 2002. Multilevel Analysis: Techniques and Applications. New  York: Taylor and Francis Group. Inglehart, R., C. Haerpfer, A. Moreno, C. Welzel, K. Kizilova, J. Diez-Medrano, M. Lagos, et al., eds. 2014. World Values Survey: Round Five—Country-­Pooled Datafile Version: www.worldvaluessurvey.org/WVSDocumentationWV5.jsp. Madrid: JD Systems Institute. Jackman, Robert W. 1987. Political Institutions and Voter Turnout in the Industrial Democracies. The American Political Science Review 81 (2): 405–424. https://doi.org/10.2307/1961959. Jackman, Robert W., and Ross A. Miller. 1995. Voter Turnout in the Industrial Democracies During the 1980s. Comparative Political Studies 27 (4): 467–492. Karp, Jeffrey A., and Susan A. Banducci. 2008. Political Efficacy and Participation in Twenty-Seven Democracies: How Electoral Systems Shape Political Behaviour. British Journal of Political Science 38 (02): 311–334. https://doi. org/10.1017/S0007123408000161. Klingemann, Hans-Dieter. 1999. Mapping Political Support in the 1990s: A Global Analysis. In Critical Citizens: Global Support for Democratic Governance, ed. Pippa Norris, 31–56. Oxford: Oxford University Press. Kuitto, Kati. 2011. More Than Just Money: Patterns of Disaggregated Welfare Expenditure in the Enlarged Europe. Journal of Social Policy 21 (4): 348–364. Lijphart, Arend. 1997. Unequal Participation: Democracy’s Unresolved Dilemma—Presidential Address, American Political Science Association, 1996. American Political Science Review 91 (1): 1–14. ———. 1999. Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries. New Haven, CT: Yale University Press. http://www.gbv. de/dms/sub-hamburg/254385370.pdf. Linde, Jonas, and Joakim Ekman. 2003. Satisfaction with Democracy: A Note on a Frequently Used Indicator in Comparative Politics. European Journal of Political Research 42 (3): 391–408. https://doi.org/10.1111/ 1475-6765.00089. Luskin, Robert C. 1990. Explaining Political Sophistication. Political Behavior 12 (4): 331–361. Macedo, Stephen. 2005. Democracy at Risk: How Political Choices Undermine Citizen Participation and What We Can Do About It. Washington, DC: Brookings Institution Press.

134 

J. Shore

Magalhaes, Pedro C. 2006. Confidence in Parliaments: Performance, Representation and Accountability. In Political Disaffection in Contemporary Democracies: Social Capital, Institutions and Politics, ed. Mariano Torcal and José R. Montero, 190–214. Routledge Research in Comparative Politics 13. London: Routledge. Mau, Steffen. 2003. The Moral Economy of Welfare States: Britain and Germany Compared. Routledge/EUI Studies in the Political Economy of Welfare 5. London and New  York: Routledge. http://site.ebrary.com/lib/alltitles/docDetail.action?docID=10097518. McAllister, Ian. 1999. The Economic Performance of Governments. In Critical Citizens: Global Support for Democratic Governance, ed. Pippa Norris, 188–203. Oxford: Oxford University Press. Mettler, Suzanne, and Joe Soss. 2004. The Consequences of Public Policy for Democratic Citizenship: Bridging Policy Studies and Mass Politics. Perspectives on Politics 2 (1): 55–73. Miller, Arthur, and Ola Listhaug. 1999. Political Performance and Institutional Trust. In Critical Citizens:  Global Support for Democratic Governance, ed. Pippa Norris, 204–216. Oxford: Oxford University Press. Miller, Joanne M., and Wendy Rahn. 2002. Identity-Based Feelings, Beliefs, and Actions: How Being Influences Doing. Presented at the annual meeting of the International Society of Political Psychology, Berlin, Germany. Newton, Kenneth, and Pippa Norris. 2000. Confidence in Public Institutions: Faith, Culture, or Performance? In Disaffected Democracies: What’s Troubling the Trilateral Countries? ed. Susan J. Pharr and Robert D. Putnam, 52–73. Princeton, NJ and Chichester: Princeton University Press. Oskarson, Maria. 2010. Social and Political Marginalization in Hard Times. APSA 2010 Annual Meeting Paper. Pacek, Alexander C., Grigore Pop-Eleches, and Joshua A.  Tucker. 2009. Disenchanted or Discerning: Voter Turnout in Post-Communist Countries. The Journal of Politics 71 (2): 473–491. https://doi.org/10.1017/ S0022381609090409. Panagopoulos, Costas. 2008. The Calculus of Voting in Compulsory Voting Systems. Political Behavior 30 (4): 455–467. https://doi.org/10.1007/ s11109-008-9058-9. Powell, G. Bingham. 2000. Elections as Instruments of Democracy: Majoritarian and Proportional Visions. New Haven, CT and London: Yale University Press. Rabe-Hesketh, Sophia, and Anders Skrondal. 2008. Multilevel and Longitudinal Modeling Using Stata. College Station, TX: Stata Press.

  The Impact of Social Policy on Democratic Citizenship 

  135

Rose, Richard. 1994. Postcommunism and the Problem of Trust. Journal of Democracy 5 (3): 18–30. https://doi.org/10.1353/jod.1994.0042. Ross, Fiona. 2000. ‘Beyond Left and Right’: The New Partisan Politics of Welfare. Governance 13 (2): 155–183. https://doi.org/10.1111/0952-1895.00127. Rothstein, Bo. 1998. Political Institutions: An Overview. In A New Handbook of Political Science, ed. Robert E.  Goodin and Hans-Dieter Klingemann, 133–167. Oxford: Oxford University Press. Schäfer, Armin. 2013. Liberalization, Inequality, and Democracy’s Discontent. In Politics in the Age of Austerity, ed. Armin Schäfer and Wolfgang Streeck, 169–195. Cambridge: Polity Press. Schattschneider, Elmer E. 1960. The Semi-Sovereign People: A Study of Free Private Enterprise in Pressure Politics, as Shown in the 1929–1930 Revision of the Tariff. New York: Rinehart and Winston. Singh, Shane. 2011. How Compelling is Compulsory Voting? A Multilevel Analysis of Turnout. Political Behavior 33 (1): 95–111. https://doi. org/10.1007/s11109-010-9107-z. ———. 2014. Not All Election Winners are Equal: Satisfaction with Democracy and the Nature of the Vote. European Journal of Political Research 53 (2): 308–327. https://doi.org/10.1111/1475-6765.12028. Solt, Frederick. 2008. Economic Inequality and Democratic Political Engagement. American Journal of Political Science 52 (1): 48–60. Stadelmann-Steffen, Isabelle. 2011. Dimensions of Family Policy and Female Labor Market Participation: Analyzing Group-Specific Policy Effects. Governance: An International Journal of Policy, Administration, and Institutions 24 (2): 331–357. Stockemer, D., Bernadette LaMontagne, and Lyle Scruggs. 2013. Bribes and Ballots: The Impact of Corruption on Voter Turnout in Democracies. International Political Science Review 34 (1): 74–90. https://doi.org/ 10.1177/0192512111419824. Svallfors, Stefan. 2007. Introduction. In The Political Sociology of the Welfare State: Institutions, Social Cleavages, and Orientations, ed. Stefan Svallfors, 1–29. Stanford, CA: Stanford University Press. Tavits, Margit, and Taavi Annus. 2006. Learning to Make Votes Count: The Role of Democratic Experience. Electoral Studies 25 (1): 72–90. https://doi. org/10.1016/j.electstud.2005.02.003. The Comparative Study of Electoral Systems. 2015. CSES Module 3 Full Release. [dataset]. December 15, 2015 version. https://doi.org/10.7804/cses. module3.2015-12-15.

136 

J. Shore

van der Meer, Tom. 2010. In What We Trust? A Multi-level Study into Trust in Parliament as an Evaluation of State Characteristics. International Review of Administrative Sciences 76 (3): 517–536. https://doi.org/10.1177/ 0020852310372450. van Deth, Jan W. 1990. Interest in Politics. In Continuities in Political Action: A Longitudinal Study of Political Orientations in Three Western Democracies, ed. M.K. Jennings, 275–312. Berlin: Walter de Gruyter. ———. 2000. Interesting But Irrelevant: Social Capital and the Saliency of Politics in Western Europe. European Journal of Political Research 37 (2): 115–147. https://doi.org/10.1111/1475-6765.00507. Verba, Sidney, and Norman H.  Nie. 1972. Participation in America: Political Democracy and Social Equality. New York: Harper and Row. Verba, Sidney, Kay Lehman Scholzman, and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge: Harvard University Press. Walzer, Michael. 1992. The Civil Society Argument. In Dimensions of Radical Democracy: Pluralism, Citizenship, Community, ed. Chantal Mouffe, 89–107. London: Verso.

6 Consequences of Unequal Citizenship

 he Effects of the Welfare State of Democratic T Citizenship The previous chapter’s analyses of democratic citizenship across OECD states clearly show a positive effect of welfare state spending on workingage adults and families. I tested the hypotheses that individuals who live in welfare states who spend more on working-age adults and families are more likely to vote, be politically interested, exhibit greater trust in their parliaments, and be more satisfied with the way democracy works in their country. Based on the policy feedback literature, I furthermore hypothesized that the positive effects of this type of social spending would be greatest for individuals with lower incomes due to the fact that social expenditures aimed at working-age adults and families have been found to be most closely aligned to reductions in economic inequality. Inequality and poverty, in turn, are not limited to economic consequences: in many countries, there is a very strong link between economic and political disadvantage. On the other hand, where the state is committed to social rights and offers its citizens generous social policies, I argue that citizens, all across the income distribution, would have greater reason to be involved in politics and support democracy. © The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2_6

137

138 

J. Shore

Across some of these aspects of democratic citizenship, equalizing effects of welfare state spending were found. The effects of additional social spending are greatest for individuals in the lowest-income quintile with regard to voting, political interest, and satisfaction with democracy. In terms of trust in parliament, while social spending is shown to increase the overall levels across all income categories, it does so while clearly preserving a pattern of social stratification. Table 6.1 provides an overview of the results from all analyses of democratic citizenship. The figures in bold-faced type in Table 6.1 show the effects that are significantly different from the effects found for incomes in the lowestincome category. These refer back to the regression tables with the crosslevel interactions between income groups, wherein significant interactions between the welfare state variable and middle and high income, respectively, indicated an effect for that group which is significantly different from the coefficient of the single welfare state variable—that is, the effect for individuals in the lowest-income quintile. The effects of expenditures aimed at working-age adults and families on the individual propensity to vote are the greatest for people with the least income, followed closely by the middle class. While people with the highest incomes are also positively affected by increased spending on working-age adults and families, this effect was not found to be significant. These individuals, people with the greatest incomes, are also those who are likely to have higher levels of Table 6.1  Marginal effects of welfare state spending on aspects of democratic citizenship

Low income Mid-level income High income

Voting

Interest in politics

Trust in parliament

Satisfaction with democracy

0.020** (0.009) 0.015** (0.007) 0.011 (0.007)

0.087** (0.041) 0.078* (0.041) 0.077* (0.041)

0.179** (0.081) 0.223*** (0080) 0.238*** (0.080)

0.072*** (0.013) 0.066*** (0.013) 0.062*** (0.013)

Notes: Based on the cross-level interaction models for the respective analyses; standard errors in parentheses. Bold-faced type indicates an effect that is significantly different from the effect found for individuals in the lowestincome category ***p < 0.01; **p < 0.05; *p < 0.1

  Consequences of Unequal Citizenship 

  139

education and better social networks; they tend to also have the loudest political voices and are not in any danger of underrepresentation. Their propensity to vote is already so high that additional social spending does not have a significant effect on their electoral participation. We find similar patterns regarding interest in politics and satisfaction with democracy, although here the effects for the top earners are also significant. So while the effects for the top earners are smaller, we continue to see a context effect for this group in terms of comparatively higher levels of democratic citizenship in countries that spend more on their working-age populations and families. This could, in part, be due to the fact that high-income earners are likely to also benefit from generous social policies aimed at working-age adults and families, as in the states where the greatest portion of the GDP is spent on this policy area, benefits tend to be universal in nature. With respect to the individuals with the lowest incomes, these are the people whose resources would predict a low likelihood of taking part in an election, low interest in politics, and less support of democracy in terms of trust in parliament and satisfaction with democracy. These are also the people who are most likely to benefit the most from working-age and family welfare policies for their own well-being. By creating this personal stake, these types of policies seem to have the ability to mobilize the people with more modest incomes to vote and pique their interest in politics. Satisfaction with democracy, which is a measure of support for democratic performance, is also greater in more generous welfare states. While this effect spans all income groups, it is largest for those with the lowest incomes. Furthermore, as higher levels of cash benefits and benefits in kind in this policy area tend to be universal, and not determined by means-tested criteria, poorer individuals in more generous welfare states are less likely to have experienced degrading and processes of applying for benefits. Targeted programs tend to be associated with greater bureaucratic hurdles and frequently convey stigma for recipients (Mettler and Stonecash 2008). Universal programs, however, have been shown to foster both political efficacy and user empowerment (Bruch et al. 2010; Kumlin 2004; Kumlin and Rothstein 2005; Wichowsky and Moynihan 2008; Moynihan and Soss 2014; Watson 2015). The feeling that one’s needs are being met are likely to increase satisfaction with democracy, as

140 

J. Shore

positive spillover effects can accrue from positive experiences with public services (Ariely 2013; Christensen and Laegreid 2005). Moreover, the redistributive effects of generous social policies may be able to give the poor the extra economic push needed to get out of the poverty trap, thereby enabling them engage with and participate in politics—something which may otherwise have remained an unattainable luxury. The messages of inclusiveness and fairness that policies aimed at working-age populations embody may reinforce the redistributive and saliency effects. By signaling that not only the interests of the wealthy get political attention, that the social rights of all members of society are valued and to be protected, politics becomes something of importance to all social milieu. A further pattern that we can see in Table  6.1 is the significance of welfare policies for the middle class with regard to democratic citizenship. We can make similar arguments with regard to the effects found for the poorest individuals. In this case, however, individuals in the middleincome group have greater resources at their disposal to facilitate their political engagement; for this reason, we tend to see slightly smaller effects for this group. Nevertheless, we continue to see a context effect for this group in terms of comparatively higher rates of participation and interest. It can thus be said that welfare states in which there is a clear focus on the well-being of working-age adults and families foster an overall inclusive political environment—one in which all citizens, regardless of economic advantage, go to the polls at a higher rate, are more interested in politics, have higher levels of trust in their national parliaments, and are more likely to be satisfied with the way democracy works. Before moving on to a discussion of some of the challenges faced by this study and other investigations of policy feedback effects, it remains necessary to attempt to explain the rather counterintuitive results from the analyses of trust in parliament. These results are particularly puzzling when compared to the results found for the other aspects of democratic citizenship, which would suggest not only a bolstering effect of increased spending on working-age adults and families but an equalizing one as well. Although individuals in countries that spend more on working-age adults and families have higher levels of trust, such a welfare state focus does not, however, promote more equitable levels of political trust. The

  Consequences of Unequal Citizenship 

  141

strongest effect on trust in parliament was found for the people with the highest incomes, followed by the effect for people in income quintiles of 2–4. While a positive effect was still found for people in the lowestincome group, it is both smaller and less statistically significant than the effects found for their wealthier counterparts. Also puzzling is why such large and equalizing effects are found with regard to satisfaction with democracy but not for trust. It would seem plausible that people who are satisfied with the performance of democracy would also be supportive of general democratic norms and principles. And while the results do show this to some extent, we must recall the fanning out of the lines at the highest levels of social spending, indicating that at the lowest levels of spending, while political trust is lower overall, it appears to be equally low across groups. Could these findings simply be an artifact of the data? Or perhaps the rather tepid level of political trust across all countries (even in the more generous welfare states) is more telling. Although the survey question on trust in institutions designed by social scientists is supposed to tap into people’s overall feelings about democracy, Grönlund and Setälä (2007, 405) point out that these institutions may lack significance for many people. Because the data available from these large cross-national surveys do not ask people about their specific personal experiences with various agencies and bureaucracies as they pertain to social welfare services and benefits, it may not be possible to really tap into the aspect of political trust drawing only on information about the national parliament—an institution which, although central to democracies, may be too far removed and abstract for many citizens.1 Moreover, by relying on but a cross-sectional snapshot, it is not certain whether the unequal levels of trust in the more generous welfare states have to do with a particular survey year or are indicative of long-term trends of decline and divide. Or it could perhaps be the case that although political trust remains the highest in countries where more is spent on working-age adults and 1  Staffan Kumlin’s (2004) approach in The Personal and the Political is much better suited to investigating political trust. Using survey data for Sweden, which inquired directly into people’s personal experiences, their opportunities to voice opinions, and their experiences of responsiveness of service agencies, he gets much closer to what political trust really means to people. Unfortunately, such data are not available for a large number of countries.

142 

J. Shore

f­ amilies, the downward trend in political trust could simply affect poorer individuals first, despite their overall satisfaction with how democracy is performing. That political trust has been on the decline in advanced democracies over the past decades is nothing new (Dalton 2004; Klingemann 1999; Norris 1999), and while support for democracy’s arguably most important institution is certainly desirable, a certain level of skepticism is also not necessarily a bad thing either. Critical citizens may demand more from their governments and keep a closer eye on decisions-makers (Norris 1999). Marien and Hooghe (2011) find that political trust does indeed matter, in particular, for law abidingness. People who indicated less trust in parliament (and thus less faith in the norms and principles of democracy) are also more likely to be permissive of illegal acts, such as tax evasion and social security fraud. Distrusters are more likely to be free riders (Tyler 2006), forcing the state to invest greater efforts into crucial activities such as collecting tax revenues and ensuring the social security payments are correct (Marien and Hooghe 2011, 283). For now, however, we can only say that political trust is significantly higher in welfare states that invest more in working-age adults and families. Furthermore, overall levels of political trust are rather unspectacular across all countries. In order to further get at what may be driving the less equal levels of trust in the more generous welfare states and whether perhaps changes in benefits or other social policies may be to blame or if we are simply looking at a small portion of a global trend, better survey data and longitudinal designs are needed.

Limitations, Challenges, and Ways Ahead There are numerous challenges for future studies of this kind to tackle. First of all, we may ask whether reductions in social spending affect all groups’ democratic citizenship in a similar manner? The intuitive answer would be that the most-at-risk groups in society would also be the most sensitive to cuts in benefits. But would reductions in social spending also adversely affect the ways the middle and upper classes also think about and engage with politics? Do people even notice such changes? Campbell

  Consequences of Unequal Citizenship 

  143

(2012, 347) argues that changing social and political conditions, such as a financial crisis, could “interrupt” the feedback effects that we would otherwise be able to document. Another challenge for future studies of democratic citizenship and social policy involves the role of mediating institutions. As we know, just as welfare state contexts differ across countries, so too do interest group organization, party structures, the media, and bureaucratic administration. Do these institutions matter for policy feedback mechanisms? Moreover, social policies aimed at working-age adults and families do not exist in a vacuum; they coexist with other social policies including pension, education, and health policies, as well as in conjunction with a wide spectrum of nonsocial policies. To make matters even more complicated, we see that individuals are nested not only in one single political context but are subject to policies made at multiple levels of government. How can we best capture what is most important and relevant to individuals? Is it the national level, as investigated in this study, or rather more proximate, local levels? In-depth qualitative interviews that would complement the use of large-scale cross-national surveys would help to address such questions. Moreover, experiments could help the researcher know if the differences observed between groups’ democratic citizenship are really due to policies and not to some other national idiosyncrasy. And while experiments could clear up some of the issues surrounding causal inference, they would not necessarily help us to determine the causal mechanisms of policy feedbacks. As Campbell (2012, 345) points out, “identifying the mechanisms that link public programs with individual attitudes and behaviors remains a stubborn challenge.” Research designs like the present, which draw on cross-national survey data are, at best, only able to speculate about the actual causal mechanisms involved in the relationships between aspects of democratic citizenship and welfare state policies. Although three main mechanisms are identified, involving the resources distributed and redistributed, the personal stakes created, and the messages and norms embodied by welfare policies aimed at working-age adults and families, we cannot be certain if different mechanisms affect different groups in various ways or even if welfare state policy relates to different aspects of democratic citizenship through various mechanisms.

144 

J. Shore

Although policy feedback effects are measured, one unfortunately cannot state with absolutely certainty the mechanisms through which it works to influence the aspects of democratic citizenship. In this case, in-depth qualitative analysis may be needed to go beyond mere speculation about the mechanisms at work. This is, however, a weakness of numerous studies of political behavior and policy feedback approaches. As Plutzer (2010) notes in his study of food stamp coverage and young adults’ political participation, he too is unable to say for certain if the positive effect of inclusive US State food stamp policies on youth political participation is due to the message sent by the policies or the tangible resources they confer. Moreover, the mechanisms are not entirely independent of one another, making the task all the more challenging. For example, generous benefits for families not only confer resources necessary for political participation, but in doing so, also increase the personal stakes and relevance of politics as well as send a message of inclusiveness. All of these mechanisms have a positive effect on people’s propensity to get involved with politics, and while it may not be possible to completely and definitively disentangle the mechanisms at work (nor may it be absolutely necessary to do so), mixed methods and experimental designs may help us to do so. Could it be argued that it is not generous welfare states that are causing high levels of political participation and engagement, but that high and equitable levels of participation were what influenced the institutionalization of such welfare states as we know them today? Despite transformations to the welfare state over the past decades, the core values of Scandinavian welfare states—universalism, solidarity, and de-commodification—have persisted (Cox 2004; Huber and Stephens 2002; Rothstein and Steinmo 2002). Broad class-cutting coalitions and alliances between labor and socialist parties seeking to enact generous social programs fought were guided by such values (Esping-Andersen 1990; Korpi 1989; Pierson 1993). Why these core values have remained largely intact and also why we continue to see more equitable levels of political engagement can be traced back to the path-dependent nature of welfare institutions. Owing to the historic mobilization of the working class and the influence that labor unions wielded during the development of the modern welfare state, we have witnessed high rates of participation among the working classes alongside the preservation of universally designed social

  Consequences of Unequal Citizenship 

  145

policies that serve wide swaths of the population. Mass public opinion can help explain this path dependency that, once set in motion, policies have the potential to generate feedback effects. Universal programs, for example, can bolster political legitimacy (Rothstein 1998), “suggesting that the vested interests that citizens develop in social democracies lead to greater preference for public over private provision” (Brooks and Manza 2006, 478). Because many people stand to benefit from such public services and benefits, program cutbacks and retrenchment are not without electoral risks for policymakers. Akin to the “lock-in effects” of welfare state development described by Pierson (1993), citizens also influence the preservation and continuity of social welfare policies. That being said, while the issue of reverse causality continues to loom large over many studies of political participation and welfare states and/ or income inequality, the design of the analyses in this contribution— accounting for variation both on the individual and on the country level and including interaction effects between individual income and social spending on working-age adults and families—provides some defense. The multilevel design with cross-level interactions allowed us to look beyond aggregate-level effects, by examining the ways in which welfare policy can intervene in the relationships between resources (here, income) and the aspects of democratic citizenship, the matter of equalizing effects could also be analyzed. In other words, for many of the dimensions, it would be rather difficult to make the argument that an endogenous relationship is at work, as group-specific effects were found and inequality of participation decreased. The relationships found here in terms of the way the welfare state can shape democratic citizenship are thus not simplelevel effects: greater social spending on the working-age segment of the population is indeed associated with a more equitable exercise of political voice. Lastly, the question of causality, however, remains a challenge for most studies of policy feedback on political behavior, particularly ones like the present that rely exclusively on cross-sectional research designs. While efforts have been made to make sure that cause does potentially precede effect (in measuring social spending prior to when the surveys were conducted), to more confidently rule out the possibility of endogeneity of welfare states to political behavior, a longitudinal design would be of

146 

J. Shore

some help. Even with such data (which we do not have for a long range across many aspects of democratic citizenship), policies, like institutions, are often slow to change. Pierson (1993, 610) posits that we would need a comparative case where lock in has not occurred in order to study the effects of policy feedback, most likely rendering a large N study drawing on a broad set of public policies (such as the present) less suitable. To conclude and as highlighted in the policy feedback literature, there are good reasons to believe that these are cases in which cause and effect exist in a mutually reinforcing relationship. Since there is evidence that the welfare state and democratic citizenship appear to interact, research designs that take time trends and change are needed, as are in-depth case studies that are better equipped to capture the nuanced effects of policy change on democratic political engagement and support.

Conclusions In asking the question about the quality of democracies, I have been able to show that citizens in advanced liberal democracies do indeed respond to what their governments offer them. Beyond the procedural and institutional accounts of the ways democracies do democracy, I have argued that democratic quality is intrinsically related to how citizens behave toward and think about their governments. Moreover, democratic quality is also about equality. In many states, however, democracy has been producing some very inequitable and possibly undemocratic outcomes. Rather than taking the view that an apathetic and uninterested citizenry is to blame for the trends of political withdrawal across many nations, I posited that in order for citizens to be interested and engaged with politics and support the ideals and principles of democracy, their democracies need to offer them something. The offerings I investigated were social welfare policies aimed at working-age adults and families. While many have documented the deleterious effects of income inequality on electoral participation, I approached this relationship in a different way, one that allows us to think more precisely about the causal mechanisms involved in the links between resources, inequality, and political participation. Although inequality is certainly a product of fiscal

  Consequences of Unequal Citizenship 

  147

and social policies, I argued that public policies are the things with which people actually have contact and personal experiences—not abstract macro-indicators such as the Gini index. Public policies, as advanced by policy feedback scholars, have the potential to produce feedback effects on mass publics. For example, much like formal institutions, policies can shape people’s incentives for political participation. Looking first at the well-documented links between resources and democratic citizenship, I then moved on to discuss the ways in which public policy can affect political attitudes and behaviors. For one, policies distribute and redistribute the resources needed for participation. They also can create personal stakes and pique one’s interest in politics. Lastly, public policies can serve as messengers: depending on design and generosity, policies can inform people about where they stand in society, whether they are seen as a deserving and integral part of the polity, or whether their status as a recipient signifies social stigma. I hypothesized that where social policies place a clear focus on reducing economic inequality, political inequality should be lessened as well. To test the assumptions of the policy feedback approach—which hitherto have tended to be confined to single countries (in particular, the United States) and also to smaller constituencies—I examined five aspects of political behaviors and attitudes in OECD countries: voting, alternative political participation, interest in politics, trust in parliament, and satisfaction with democracy. The policies I looked at furthermore constituted an improvement to previous studies drawing on either highly aggregated measures of welfare state spending or less tangible indicators of inequality. In particular, public policies that influence one’s economic well-being have a greater likelihood of stimulating or stunting democratic citizenship. For this reason, I chose to examine the effect of social policies aimed at working-age adults and families on the dimensions of democratic citizenship, showing that this group of policies is both theoretically and empirically most closely aligned to redistributive outcomes and the welfare state’s theoretical core. Not only do the overall levels of democratic citizenship tend to be higher in more generous welfare states, but the distributions of these aspects are also much more equitable. While wealthier individuals in all countries under consideration vote at higher rates than their less well-off

148 

J. Shore

fellow citizens, we can, however, witness a much smaller gap between the rich and the poor in welfare states that spend more on their working-age adults and families. Alternative forms of political participation also have a more frequent and equitable practice in more generous countries. Furthermore, political interest and satisfaction with democracy are not only more likely in nations where the state has a clear policy focus on the reduction of economic inequality, but they are also more equitably distributed across income groups. Finally, while trust in parliament—an indicator of one’s overall confidence in democratic political institutions— is higher in more generous welfare states, we cannot speak of a more equal incidence of political trust within the population. One possible reason for the continued dependence of trust on income could be the overall lower levels of political trust across countries. It could perhaps also be that what we are observing is part of a larger trend in declining trust levels, one that perhaps reaches the poorest citizens sooner and with greater force. “Citizens with lower or moderate incomes speak with a whisper that is lost on the ears of inattentive government officials, while the advantaged roar with a clarity and consistency that policy-makers readily hear and routinely follow” (American Political Science Association Task Force 2004, 651). The attention in this book has been on the feed part of policy feedback effects; the implication the previous statement, however, touches on the back element—that is, the policies that result from unequal participation and representation. While several studies have investigated how policies, their design, and their administration can affect political participation and democratic citizenship, the policy results or outcomes (i.e., what comes back out) have received far less empirical attention. Where propensity to vote is strongly tied to one’s socioeconomic resources, we are likely to encounter an overrepresentation of certain interests (Verba et  al. 1995, 493), leading to biased outcomes. The implications of unequal democratic citizenship are manifold, as all conceptions of democracy rest upon the ideal that all citizens should have equal opportunities to participate in political life. Furthermore, political inequality has many implications for the responsiveness, health, and ulti-

  Consequences of Unequal Citizenship 

  149

mately the legitimacy and stability of democratic states (Schäfer 2015; Gilens 2012). When, however, we witness that portions of the population are being left out of democratic citizenship, we must ask why this is so and, perhaps more importantly, what can be done to draw these groups back in? The present study offers further insights into the role of social policy for the distribution of democratic citizenship. Despite the drawbacks of a cross-sectional study, inroads have been forged into further examining and understanding the links between public policies, inequality, and political participation for mass publics across a diverse group of democracies. And while different types of designs are certainly needed to better identify what is feed and what is back in policy feedback effects, recognizing that in some countries we are confronted with virtuous cycles and in others vicious ones remain an important finding. If citizens with greater resources are consistently and actively engaging in political life while the less advantaged are withdrawing from it, the consequences of unequal political input are clear. As the policy feedback literature outlines, the outputs of previous democratic decision-making processes influence and shape the inputs of future processes. Recent studies examining the representativeness of public policy outputs or government composition have confirmed the suspicions regarding the effect of unequal input across a variety of states (e.g., Bartels 2008; Elsässer et al. 2017; Giger et al. 2012; Gilens 2005, 2012; Gilens and Page 2014; Peters and Ensink 2015; Rosset 2013). While we must also accept the fact that not everyone will or can be interested in and engaged with politics, the reality that political influence remains disproportionality concentrated in the hands of the better off is indeed cause for concern. As shown in this book, there are marked differences across countries in terms of whose voices are likely to get heard and whose may be ignored. Uncovering these disparities is, however, only half the story—what can be done to encourage political equality and more equal representation remains a challenging but vital task for contemporary democratic societies.

150 

J. Shore

References American Political Science Association Task Force. 2004. American Democracy in an Age of Rising Inequality. Perspectives on Politics 2 (04): 651–666. https://doi.org/10.1017/S153759270404040X. Ariely, Gal. 2013. Public Administration and Citizen Satisfaction with Democracy: Cross-national Evidence. International Review of Administrative Sciences 79 (4): 747–766. https://doi.org/10.1177/0020852313501432. Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age. New  York and Princeton, NJ: Russell Sage Foundation and Princeton University Press. Brooks, C., and J.  Manza. 2006. Social Policy Responsiveness in Developed Democracies. American Sociological Review 71 (3): 474–494. https://doi. org/10.1177/000312240607100306. Bruch, Sarah K., Myra M. Ferree, and Joe Soss. 2010. From Policy to Polity: Democracy, Paternalism, and the Incorporation of Disadvantaged Citizens. American Sociological Review 75 (2): 205–226. https://doi.org/10.1177/ 0003122410363563. Campbell, Andrea Louise. 2012. Policy Makes Mass Politics. Annual Review of Political Science 15 (1): 333–351. https://doi.org/10.1146/annurev-polisci012610-135202. Christensen, Tom, and Per Laegreid. 2005. Trust in Government: The Relative Importance of Service Satisfaction, Political Factors, and Demography. Public Performance & Management Review 28 (4): 487–511. https://doi.org/10.108 0/15309576.2005.11051848. Cox, Robert. 2004. The Path-Dependency of an Idea: Why Scandinavian Welfare States Remain Distinct. Social Policy & Administration 38 (2): 204–219. https://doi.org/10.1111/j.1467-9515.2004.00386.x. Dalton, Russell J.  2004. Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies. Oxford: Oxford University Press http://www.loc.gov/catdir/enhancements/fy0620/ 2004300752-t.html. Elsässer, Lea, Svenja Hense, and Armin Schäfer. 2017. “Dem Deutschen Volke”? Die ungleiche Responsivität des Bundestags. Zeitschrift für Politikwissenschaft 27 (2): 161–180. https://doi.org/10.1007/s41358-017-0097-9. Esping-Andersen, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Princeton: Princeton University Press.

  Consequences of Unequal Citizenship 

  151

Giger, Nathalie, Jan Rosset, and Julian Bernauer. 2012. The Poor Political Representation of the Poor in a Comparative Perspective. Representation 48 (1): 47–61. https://doi.org/10.1080/00344893.2012.653238. Gilens, Martin. 2005. Inequality and Democratic Responsiveness. Public Opinion Quarterly 69 (5): 778–796. https://doi.org/10.1093/poq/nfi058. ———. 2012. Affluence and Influence: Economic Inequality and Political Power in America. Princeton, NJ and New  York: Princeton University Press and Russell Sage Foundation. Gilens, Martin, and Benjamin I.  Page. 2014. Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens. Perspective on Politics 12 (3): 564–581. https://doi.org/10.1017/S1537592714001595. Grönlund, Kimmo, and Maija Setälä. 2007. Political Trust, Satisfaction and Voter Turnout. Comparative European Politics 5 (4): 400–422. https://doi. org/10.1057/palgrave.cep.6110113. Huber, Evelyne, and John D. Stephens. 2002. Globalisation, Competitiveness, and the Social Democratic Model. Social Policy and Society 1 (1): 47–57. https://doi.org/10.1017/S1474746402001070. Klingemann, Hans-Dieter. 1999. Mapping Political Support in the 1990s: A Global Analysis. In Critical Citizens: Global Support for Democratic Governance, ed. Pippa Norris, 31–56. Oxford: Oxford University Press. Korpi, Walter. 1989. Power, Politics, and State Autonomy in the Development of Social Citizenship: Social Rights During Sickness in Eighteen OECD Countries Since 1930. Reprint Series/Swedish Institute for Social Research 245. Stockholm: Stockholms University and Swedish Institute for Social Research. Kumlin, Staffan. 2004. The Personal and the Political: How Personal Welfare State Experiences Affect Political Trust and Ideology. Political Evolution and Institutional Change. New York: Palgrave Macmillan. Kumlin, Staffan, and Bo Rothstein. 2005. Making and Breaking Social Capital: The Impact of Welfare-State Institutions. Comparative Political Studies 38 (4): 339–365. Marien, Sofie, and Marc Hooghe. 2011. Does Political Trust Matter? An Empirical Investigation into the Relation Between Political Trust and Support for Law Compliance. European Journal of Political Research 50 (2): 267–291. https://doi.org/10.1111/j.1475-6765.2010.01930.x. Mettler, Suzanne, and Jeffrey M. Stonecash. 2008. Government Program Usage and Political Voice. Social Science Quarterly 89 (2): 273–293. https://doi. org/10.1111/j.1540-6237.2008.00532.x.

152 

J. Shore

Moynihan, Donald P., and Joe Soss. 2014. Policy Feedback and the Politics of Administration. Public Administration Review 74 (3): 320–332. https://doi. org/10.1111/puar.12200. Norris, Pippa, ed. 1999. Critical Citizens: Global Support for Democratic Governance. Oxford: Oxford University Press. Peters, Yvette, and Sander J.  Ensink. 2015. Differential Responsiveness in Europe: The Effects of Preference Difference and Electoral Participation. West European Politics 38 (3): 577–600. https://doi.org/10.1080/01402382.2014 .973260. Pierson, Paul. 1993. When Effect Becomes Cause: Policy Feedback and Political Change. World Politics 45 (4): 595–628. Plutzer, Eric. 2010. Do Highly Exclusive Social Welfare Programs Increase Political Inequality? A Comparative Analysis of the 50 US States. Social Science Research Center Berlin (WZB) Working Paper No. SP (2010-01). Rosset, Jan. 2013. Are the Policy Preferences of Relatively Poor Citizens Underrepresented in the Swiss Parliament? The Journal of Legislative Studies 19 (4): 490–504. https://doi.org/10.1080/13572334.2013.812363. Rothstein, Bo. 1998. Just Institutions Matter: The Moral and Political Logic of the Universal Welfare State. Cambridge: Cambridge University Press. Rothstein, Bo, and Sven Steinmo, eds. 2002. Restructuring the Welfare State: Political Institutions and Policy Change. New York: Palgrave Macmillan. Schäfer, Armin. 2015. Der Verlust politischer Gleichheit: Warum die sinkende Wahlbeteiligung der Demokratie schadet. Schriften aus dem Max-PlanckInstitut [MPI] für Gesellschaftsforschung, Köln Band 81. Frankfurt am Main: Campus Verlag. Tyler, Tom R. 2006. Why People Obey the Law. Princeton, NJ: Princeton University Press. Verba, Sidney, Kay Lehman Scholzman, and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge: Harvard University Press. Watson, Sara. 2015. Does Welfare Conditionality Reduce Democratic Participation? Comparative Political Studies 48 (5): 645–686. https://doi. org/10.1177/0010414014556043. Wichowsky, Amber, and Donald P.  Moynihan. 2008. Measuring How Administration Shapes Citizenship: A Policy Feedback Perspective on Performance Management. Public Administration Review 68 (5): 908–920.

Appendix

Table A1  Voting: Variables, descriptive statistics, sources, hypothesized direction of influence Dir. Dependent variable Electoral Shares: participation Yes: 80.19 No: 19.81

Individual-level variables Educational level Shares: 0: 2.46 1: 11.28 2: 21.54 3: 34.02 4: 7.51 5: 23.19 Gender Shares: Male: 46.48 Female: 53.52

Dummy: Voted in last national election 1 = Yes; 0 = No Source: WVS, wave 4; ESS, rounds 3 and 4. Completed level of education: 0 = Did not complete primary 1 = Primary or 1st stage of basic 2 = Lower secondary/ 2nd stage of basic 3 = Upper secondary 4 = Post-secondary, non-tertiary 5 = First stage of tertiary Dummy: 0 = male; 1 = female

+

−

(continued)

© The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2

153

154  Appendix Table A1 (continued) Dir. Union membership Shares: Members: 39.98 Nonmembers: 60.02 Age Mean: 49.30 SD: 17.50 Min.: 18 Max.: 105 Employment status Shares: Employed: 56.56 Unemployed: 43.44 Income level Shares: Low: 15.73 Mid: 65.79 High: 18.48 Country-level variables Mean: 5.83 Welfare state: SD: 2.10 Working-age Min.: 2.24 benefits Max.: 10.41

Dummy: Respondent is or was a member of a trade union 1 = Yes; 0 = No

+

Age in years of the respondent

+

Dummy: Respondent has reported doing paid work within the last seven days 1 = Yes; 0 = No

+

Reported income quintiles Low = first quintile; mid = second to fourth quintiles; high = fifth quintile

+

Total benefits for working-age adults plus expenditures for Active Labor Market Policies as a proportion of the GDP. Source: OECD Social Expenditure Database. Expenditures have been adjusted for structural and demographic pressures: Unemployment benefits and ALMP by the unemployment rate; family cash spending by the proportion of the population under age 15; and family benefits in kind spending by the proportion of the population ages 0–4. Source: OECD.

+

(continued)

 Appendix    

155

Table A1 (continued) Dir. Compulsory voting

Shares: 0: 86.96 1: 6.43 2: 0.00 3: 3.13 4: 3.48

Disproportionality of electoral system

Mean: 5.36 SD: 4.15 Min.: 0.35 Max.: 16.63 Mean: 61.37 SD: 46.56 Min.: 8 Max: 197

Age of democracy

Strength of left parties in government

Mean: 35.94 SD: 37.06 Min: 0 Max.: 100

Degree of compulsory voting with regard to severity and enforcement, ranging from 0 to 4: 0 = noncompulsory voting; 1 = low sanctions and enforcement; 2 = moderate sanctions; low enforcement; 3 = either high sanctions or high enforcement; 4 = both high sanctions and enforcement. Source: IDEA (International Institute for Democracy and Electoral Assistance) and own calculations based on Singh (2011) and Panagopoulos (2008). Gallagher Index of disproportionality. Source: Armingeon et al. (2011), Comparative Political Data Set.

+

−

+/− Age in years of present democratic regime at time of survey. Variable measures the number of years since the most recent regime change. Source: Armingeon et al. (2011), Comparative Political Data Set. Squared term included in the models. Social democratic and other left-wing + parties as a percentage of total cabinet posts, weighted by number of days government was in office in given year. Source: Armingeon et al. (2011), Comparative Political Data Set.

156  Appendix Table A2  Interest in politics: Variables, descriptive statistics, sources, hypothesized direction of influence Dir. Dependent variable Interest in politics Shares: Yes: 50.09 No: 49.91

Individual-level variables Educational level Shares: 0: 2.14 1: 12.18 2: 21.38 3: 34.70 4: 7.44 5: 24.30 Gender Shares: Male: 47.27 Female: 52.73 Union membership Shares: Members: 42.51 Nonmembers: 57.49 Age Mean: 48.65 SD: 17.41 Min.: 18 Max.: 105 Urban Shares: Urban: 35.06 Non-urban: 64.94 Employment status Shares: Employed: 58.06 Unemployed: 41.94

Respondent indicated some level of interest in politics. Recoded from original political interest variable with 4 categories: not at all interested and hardly interested = No (0); quite interested and very interested = Yes (1) Source: WVS, wave 4; ESS, rounds 3 and 4. Completed level of education: 0 = Did not complete primary 1 = Primary or 1st stage of basic 2 = Lower secondary/ 2nd stage of basic 3 = Upper secondary 4 = Post-secondary, non-tertiary 5 = First stage of tertiary Dummy: 0 = male; 1 = female

+

+

Respondent is a member or former member of a union. Dummy: 0 = Non-member; 1: Member

+

Age in years of the respondent

+

Dummy: Respondent lives in an urban area. 0 = No; 1 = Yes.

+

Respondent has reported doing paid work within the last seven days. Dummy: 1 = Yes; 0 = No

+

(continued)

 Appendix    

157

Table A2 (continued) Dir. Income level

Shares: Low: 15.83 Mid: 65.75 High: 18.41 Country-level variables Mean: 5.85 Welfare state: SD: 2.09 Working-age Min.: 2.24 benefits Max.: 10.41

Disproportionality of electoral system Age of democracy

Mean: 5.34 SD: 4.16 Min.: .35 Max: 16.63 Mean: 62.09 SD: 46.22 Min.: 8 Max: 197

Strength of the left Mean: 36.21 SD: 37.06 Min: 0 Max.: 100

Reported income quintiles Low = first quintile; mid = second to fourth quintiles; high = fifth quintile

+

Total benefits for working-age adults plus expenditures for Active Labor Market Policies as a proportion of the GDP. Source: OECD Social Expenditure Database. Expenditures have been adjusted for structural and demographic pressures: Cash and benefits in kind for unemployment by the unemployment rate; family cash spending by the proportion of the population under age 15; and family benefits in kind spending by the proportion of the population ages 0 to 4. Source: OECD. Gallagher Index of disproportionality. Source: Armingeon et al. (2011), Comparative Political Data Set.

+

Age in years of present democratic regime at time of survey. Variable measures the number of years since the most recent regime change. Source: Armingeon et al. (2011), Comparative Political Data Set. Squared term included in the models. Social democratic and other left-wing parties as a percentage of total cabinet posts, weighted by number of days government was in office in given year. Source: Armingeon et al. (2011), Comparative Political Data Set.

−

+/−

+

158  Appendix Table A3  Trust in parliament: Variables, descriptive statistics, sources, hypothesized direction of influence Dir. Dependent variable Trust in parliament Mean: 4.64 SD: 2.45 Min.: 0 Max.: 10

Individual-level variables Educational level Shares: 0: 2.41 1: 11.18 2: 19.48 3: 35.38 4: 6.46 5: 25.09 Gender Shares: Male: 47.41 Female: 52.59 Urban residence Shares: Urbanites: 33.02 Non-urban: 66.98 Age Mean: 48.86 SD: 17.53 Min.: 18 Max.: 105 Union membership Shares: Union: 45.46 Non-union: 54.54 Employment status Shares: Employed: 45.50 Unemployed: 54.50

Dummy: Respondent indicated participating in a noninstitutionalized political activity (contacting a politician and/or signing of a petition) 1 = Yes; 0 = No. Source: ESS, rounds 3 and 4 Completed level of education: 0 = Did not complete primary 1 = Primary or 1st stage of basic 2 = Lower secondary/ 2nd stage of basic 3 = Upper secondary 4 = Post-secondary, non-tertiary 5 = First stage of tertiary Dummy: 0 = male; 1 = female

+

+

Respondent lives in an urban area. Dummy: 0 = Rural/non-urban; 1: Urban

+

Age in years of the respondent

+

Dummy: Respondent is a past or present union member. 1 = Yes; 0 = No

+

Dummy: Respondent has reported doing paid work within the last seven days 1 = Yes; 0 = No

+

(continued)

 Appendix    

159

Table A3 (continued) Dir. Income level

Shares: Low: 15.54 Mid: 65.66 High: 18.80 Country-level variables Mean: 6.17 Welfare state: SD: 1.96 Working-age Min.: 2.45 benefits Max.: 10.41

Age of democracy

Mean: 53.17 SD: 38.85 Min.: 8 Max: 160

Strength of left parties

Mean: 41.35 SD: 36.82 Min.: 0 Max: 100

Disproportionality of electoral system

Mean: 5.06 SD: 4.24 Min.: .35 Max: 16.63

Reported income quintiles Low = first quintile; mid = second to fourth quintiles; high = fifth quintile

+

+ Total benefits for working-age adults plus expenditures for Active Labor Market Policies as a proportion of the GDP. Source: OECD Social Expenditure Database. Expenditures have been adjusted for structural and demographic pressures: Cash and benefits in kind for unemployment by the unemployment rate; family cash spending by the proportion of the population under age 15; and family benefits in kind spending by the proportion of the population ages 0 to 4. Source: OECD + Age in years of present democratic regime at time of survey. Variable measures the number of years since the most recent regime change. Source: Armingeon et al. (2011), Comparative Political Data Set. Squared term included in the models. Social democratic and other left-wing + parties as a percentage of total cabinet posts, weighted by number of days government was in office in given year. Source: Armingeon et al. (2011), Comparative Political Data Set. Gallagher Index of disproportionality. +/− Source: Armingeon et al. (2011), Comparative Political Data Set.

160  Appendix Table A4  Satisfaction with democracy: Variables, descriptive statistics, sources, hypothesized direction of influence Dir. Dependent variable Satisfaction with Shares: democracy Yes: 67.72 No: 32.28

Individual-level variables Educational level Shares: 1: 0.46 2: 1.91 3: 21.29 4: 15.25 5: 27.18 6: 33.91 Gender

Shares: Male: 51.14 Female: 48.86 Urban residence Shares: Urbanites: 51.22 NonUrban:48.78 Age Mean: 49.58 SD: 17.09 Min.: 18 Max.: 100 Union membership Shares: Union: 25.12 Non-union: 74.88 Employment status Shares: Employed: 55.72 Unemployed: 44.28

“On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the way democracy works in [COUNTRY]?” Fairly satisfied; very satisfied = yes (1); not very satisfied; not at all satisfied = no (0). Source: CSES, module 3 +

Education of respondent 1 = None 2 = Incomplete primary 3 = Primary completed 4 = Incomplete secondary 5 = Secondary completed 6 = Post-secondary; trade or vocational school completed Dummy: 0 = male; 1 = female

+

Respondent lives in an urban area. Dummy: 0 = Rural area or small village; small- or middle-sized town 1: Suburbs of large town or city; large town or city Age in years of the respondent

+

+

Dummy: Respondent is a union member 1 = Yes; 0 = No

+

Dummy: Current employment status of respondent 1 = Employed: Part-time or full-time 0 = Unemployed; not in labor force

+

(continued)

 Appendix    

161

Table A4 (continued) Dir. Left-right placement

Mean: 2.44 SD: 5.29 Min.: 0 Max.: 10

Income level

Shares: Low: 21.85 Mid: 63.33 High: 14.82 Country-level variables Mean: 5.95 Welfare state: SD: 1.77 Working-age Min.: 2.45 benefits Max.: 10.41

Age of democracy

Mean: 63.45 SD: 47.93 Min.: 8 Max: 160

Strength of left parties

Mean: 31.23 SD: 32.20 Min.: 0 Max: 100

Disproportionality of electoral system

Mean: 4.22 SD: 3.39 Min.: .35 Max: 13.69

“Where would you place yourself on this scale?” Scale ranges from 0 to 10, with lower values indicating a left-leaning position; higher values a right-leaning political position Reported income quintiles Low = first quintile; mid = second to fourth quintiles; high = fifth quintile

+

+

+ Total benefits for working-age adults and families + ALMP expenditures as % of GDP. Source: OECD SOCX. Expenditures adjusted for structural & demographic needs: Cash & benefits in kind for unemployment by unemployment rate; family cash spending by proportion of the population under age 15; and family benefits in kind spending by proportion of the population ages 0 to 4. Source: OECD + Age in years of present democratic regime at time of survey. Variable measures the number of years since the most recent regime change. Source: Armingeon et al. (2011), Comparative Political Data Set. Squared term included in the models. Social democratic and other left-wing + parties as a % of total cabinet posts, weighted by number of days government was in office in given year. Source: Armingeon et al. (2011), Comparative Political Data Set. Gallagher Index of disproportionality. +/− Source: Armingeon et al. (2011), Comparative Political Data Set.

162  Appendix Table A5  Alternative welfare policy specifications and satisfaction with democracy Satisfaction with democracy Constant Fixed effects Income (Ref. = low) Mid-level High Welfare expenditures Strength of left Age of democracy Disproportionality Welfare × Mid-level income Welfare × High income Random effects Rho Context variance N

Old-age

Health

Total social expenditures

−0.330 (0.719)

−2.177* (1.143)

−2.166** (0.919)

0.022 (0.118) 0.157 (0.162) −0.077 (0.064) −0.001 (0.005) 0.011** (0.005) −0.024 (0.063) 0.027

0.153 (0.256) −0.026 (0.353) 0.236 (0.195) −0.003 (0.005) 0.011** (0.005) −0.049 (0.061) 0.014

0.024 (0.207) 0.074 (0.296) 0.061 (0.195) −0.002 (0.005) 0.013*** (0.004) −0.040 (0.059) 0.010

(0.013)** 0.033 (0.019)*

(0.041) 0.073 (0.057)

(0.009) 0.016 (0.014)

0.165 0.807 18,443 (22)

0.159 0.790 18,443 (22)

0.151 0.764 18,443 (22)

Notes: Log-odds, standard errors in parentheses. All additional individual variables included in models but not shown. ***p < 0.01; **p < 0.05; *p < 0.1 Marginal effects for income groups and respective welfare indicators can be found in Table 5.18.

Index1

A

Age of democracy, 85, 87, 91, 94, 105–107, 117, 118, 125, 126 Almond, Gabriel, 19, 23 American Political Science Association Task Force, 22, 96, 148 C

Campbell, Andrea Louise, 4, 40–42, 97, 143 Castles, Francis, 4, 36, 37, 56–60 Causality, 145 Causal mechanisms, 143, 146

Comparative Study of Electoral Systems (CSES), 5, 121–124 Compulsory voting, 80, 80n4, 84, 86, 87, 89, 90 Context, 2, 4, 17, 24, 29–47, 65–68, 70, 71, 76, 83, 91, 97, 115, 130, 139, 140, 143 Cross-level interaction, 5, 71, 81–83, 86, 90–91, 95, 97, 99, 102, 106, 107, 118–119, 127, 129, 138, 145 D

Dahl, Robert, 9, 10, 12, 17, 24, 38 De-commodification, 56–59, 144

 Note: Page numbers followed by ‘n’ refer to notes.

1

© The Author(s) 2019 J. Shore, The Welfare State and the Democratic Citizen, Palgrave Studies in European Political Sociology, https://doi.org/10.1007/978-3-319-93961-2

163

164  Index

Democratic citizenship, 1–5, 12, 15–21, 24, 29, 32, 36, 46, 53–71, 75–130, 137–149 Democratic support, 2, 6, 17, 21, 44, 111–113, 116 Disproportionality, 84, 86, 87, 89, 91, 103–107, 116–119, 125–127, 149

Inequality, 1–3, 9–24, 36, 38, 41, 42, 44–47, 58–63, 82, 96, 98–100, 99n8, 111, 120, 125, 128, 130, 137, 145–149 Ingram, Helen, 40, 42 Interaction, 31, 39, 69, 83, 86, 89–92, 95, 106, 108, 120, 127, 138, 145

E

K

Egalitarian, 1, 94 Equality, 1, 3, 9, 10, 12, 14, 33, 33n3, 45, 112, 146, 149 Esping-Andersen, Gøsta, 4, 41, 43, 56–58, 144 Europe, 14, 115–119 European Social Survey (ESS), 5, 66, 76, 78, 79n3, 80, 81, 100, 102, 103, 113–115, 113n10, 122 Expenditures, 4, 6, 56–64, 95, 128–130, 137, 138

Kuitto, Kati, 59, 61, 62, 130 Kumlin, Staffan, 29, 42–44, 139, 141n1

H

M

Hox, Joop, 66, 67, 69, 83n6

Macro-level, 17, 31, 40, 53, 65–67, 82 Marginal effects, 82, 86, 91, 92, 108, 120, 127–130, 138 Marshall, T.H., 36, 37, 45 Measurement, 4, 12, 53, 55, 84, 113n10 Mettler, Suzanne, 39, 40, 42, 98, 99, 139 Multilevel analysis, 65–67, 69, 83–85, 90–91, 105, 107, 117–119, 126

I

Income, 1–3, 5, 6, 9, 10, 16, 17, 22–24, 35, 37, 41, 44–47, 71, 76, 79–83, 86–111, 114–130, 137–141, 145, 146, 148 Individual-level, 4, 65–67, 77, 78, 82, 86, 88, 91, 100, 102–105, 116, 121, 123

L

Legitimacy, 6, 10, 11, 21, 31, 45, 96, 145, 149 Lijphart, Arend, 15, 17, 24, 80n4, 84, 124 Logistic regression, 70, 81, 83, 95, 100, 102, 104

 Index    

165

New institutionalism, 30, 31, 31n1 Norris, Pippa, 19, 112, 142

Ringen, Stein, 10, 11, 13, 14, 16, 35 Rothstein, Bo, 30, 32–34, 42, 43, 100, 139, 144, 145

O

S

N

Organisation for Economic Co-operation and Development (OECD), 57, 60–64, 62n1, 76, 77, 94, 97, 100, 102, 137, 147

Quality of democracy, 2, 3, 11, 15, 146

Satisfaction with democracy, xiv, 4–6, 15, 20, 44, 53, 75, 111–113, 116, 121–130, 138, 139, 141, 147, 148 Schäfer, Armin, 112, 149 Schattschneider, Elmer E., 38, 39, 98, 99n8 Schneider, Anne, 40, 42 Social rights, 12, 15, 32, 35, 37, 38, 43, 56, 58, 100, 137, 140 Social spending, 2, 4–6, 56, 58–60, 84, 96, 106, 108–110, 116, 120, 125, 129, 137–139, 141, 142, 145 Socio-economic resources, 16, 17, 45, 46, 95, 148 Solt, Frederick, 17, 24, 76, 98, 99n8 Soss, Joe, 39, 40, 42, 44, 98, 99, 139 Stadelmann-Steffen, Isabelle, 23, 29, 82n5 Stakes, personal, 41, 42, 94, 143, 144, 147 Strength of the left, 84, 106, 125

R

T

P

Pierson, Paul, 37, 39, 40, 42, 54, 58, 144–146 Policy feedback, 1–4, 17, 24, 29, 38–41, 44–46, 98, 110, 129, 137, 140, 143–149 output, 112, 129, 149 Political efficacy, 2, 17, 22, 23, 42, 44, 45, 96, 139 interest, xiv, 5, 14–16, 53, 75, 96–111, 137–140, 147, 148 Q

Redistribution, 41, 45, 58 Representation, 23, 84, 95, 96, 104, 113, 124, 125, 148, 149

Trust, 4–6, 19, 20, 44, 46, 53, 75, 111–121, 137–142, 141n1, 147, 148

166  Index

Turnout, 2, 5, 15, 17, 24, 42, 76–81, 78n2, 80n4, 84, 85, 88–90, 93–95, 104 U

Unemployment, 36, 56, 60–63, 95, 110 United States, 2, 20, 39, 41, 42, 44–46, 60, 63, 64, 76, 80, 81, 98, 102, 103, 144, 147 Universalism, 144 V

van Deth, Jan, 97 Verba, Sidney, 19, 22–24, 79, 96, 97, 148 Vote, 1, 9, 12, 14, 15, 22, 46, 53, 69, 76–86, 79n3, 88, 89, 92–96, 106, 137–139, 148 Voting, xiv, 5, 6, 15, 20, 46, 70, 71, 75–80, 78n1, 79n3, 80n4, 82–97, 83n6, 86n7, 100, 101n9, 103–106, 109, 110, 122, 138, 147

W

Welfare policy, 1, 2, 5, 14, 36, 43, 45, 46, 59–63, 67, 71, 76, 82, 86, 89–92, 94, 95, 97–100, 102, 106, 108, 110, 120, 127, 139, 140, 143, 145, 146 priority, 59, 130 spending, 4–6, 32, 46, 56–59, 61–63, 81–83, 86, 88–95, 99, 100, 106, 108, 109, 113, 116, 119–121, 127–129, 137, 138, 147, 148 state, 1–6, 12, 14, 16, 17, 24, 29, 31–39, 31n2, 43–47, 53–71, 75, 76, 81–83, 86–90, 92–97, 99, 100, 102, 104–111, 113–115, 119–125, 127–130, 137–148 Working-age adults and families, 4–6, 57–59, 61–64, 67, 71, 76, 81–83, 86, 89, 90, 92–95, 97, 99, 102, 106, 108, 110, 111, 113, 116, 120, 127–130, 137–143, 145–148 World Values Survey (WVS), 5, 63, 66, 76, 78, 80, 81, 100, 102, 103, 113n10