Faith in Numbers: Religion, Sectarianism, and Democracy 0197538010, 9780197538012

Table of contents :
Cover
Half Title Page
Title Page
Copyright Page
Dedication Page
Contents
Figures
Tables
Acknowledgments
1 Introduction
2 An Interest-Based Theory of Religion andSupport for Democracy
3 The Religious Experience Communal Prayer and Group Identity
4 Christians and Muslims in Lebanon Before the Syrian Civil War
Appendix 4A: Survey Questions Used
Appendix 4B: Supplementary Tables
5 After Syria
Appendix 5A: Description of Experiment
Appendix 5B: Description of Variables
6 Representation or Redistribution? Evidence from Iraq
Appendix 6A: Supplementary Tables and Figures
7 Conclusion
Appendix 7A: Additional Information, Cross-National
Notes
Bibliography
Index

Citation preview

Faith in Numbers

Faith in Numbers Religion, Sectarianism, and Democracy M IC HA E L HO F F M A N

1

3 Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America. © Oxford University Press 2021 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above. You must not circulate this work in any other form and you must impose this same condition on any acquirer. Library of Congress Control Number: 2020920844 ISBN 978–0–19–753801–2 DOI: 10.1093/oso/9780197538012.001.0001 1 3 5 7 9 8 6 4 2 Printed by Integrated Books International, United States of America

For Kathy

Contents Figures Tables Acknowledgments

ix xi xiii

1. Introduction

1

2. An Interest-Based Theory of Religion and Support for Democracy

11

3. The Religious Experience Communal Prayer and Group Identity

35

4. Christians and Muslims in Lebanon Before the Syrian Civil War

52

Appendix 4A: Survey Questions Used Appendix 4B: Supplementary Tables

5. After Syria: Communal Religion and Democracy in 2014 Lebanon Appendix 5A: Description of Experiment Appendix 5B: Description of Variables

6. Representation or Redistribution? Evidence from Iraq Appendix 6A:

Supplementary Tables and Figures

7. Conclusion: Implications for Religion and Politics Appendix 7A:

Notes Bibliography Index

Additional Information, Cross-National

74 75

78 114 115

117 140

145 157

159 173 187

Figures 4.1 Population and Parliamentary Numbers, by Sect

56

4.2 Sectarian Attitudes in 2005 Zogby Poll

59

4.3 Average Monthly Income, by Sect (in Lebanese Pounds)

64

4.4 Marginal Effects of Communal Prayer on Support for Democracy

70

4.5 Marginal Effects of Communal Prayer on Protest Participation

72

5.1 Support for Democracy, by Sect and Attendance (Observed)

91

5.2 Average Treatment Effects (Support for Democracy)

98

5.3 Frequency of Communal Prayer and Linked Fate

100

5.4 “Friday” Effect, by Weekly Attendance

102

5.5 Average Treatment Effects (Linked Fate)

104

5.6 Frequency of Communal Prayer and Perceived Benefits from Democracy

105

5.7 Frequency of Communal Prayer and Attitudes toward Ending Sectarian System

107

5.8 Frequency of Communal Prayer and Attitudes toward New Census

107

5.9 Frequency of Communal Prayer and Attitudes toward Economic Inequality

109

5.10 Frequency of Communal Prayer and Attitudes toward Redistributive Democracy

110

6.1 Average Income by Sect

126

6.2 Iraqi Districts, by Oil Reserves and Major Sects

127

6.3 Marginal Effects of Communal Prayer on Support for Democracy, by Sect

131

6.4 Marginal Effects of Communal Prayer on Support for Democracy, by Perceptions of Democracy

135

6.5 “Candidate Sect Important,” by Sect and Communal Prayer

136

6.6 Candidate Sect or Piety Important, by Sect and Communal Prayer

137

x figures 6A.1 Conditional Probabilities of Support for Democracy

144

7.1 Marginal Effects of Attendance, by Group Size

147

7.2 Conditional Effects by Levels of Democracy and Religious Polarization

149

7.3 Conditional Effects by Levels of Religiosity and Religious Affiliation

150

Tables 4.1 Demographics and Allocations, 1943 National Pact 4B.2 Logistic Regression Results, Support for Democracy (Christians) 4B.3 Logistic Regression Results, Support for Democracy (Muslims) 4B.4 Robustness Checks, Commitment to Democracy 4B.5 Propensity Score Matching Results 5.1 Baseline Results 5.2 Group Identification, Friday versus Other Days 5B.1 Summary Statistics 6A.1 Support for Democracy by Communal Prayer, Sunni 6A.2 Support for Democracy by Communal Prayer, Shi‘a 6A.3 Baseline Results, Support for Democracy 6A.4 Interactive Models, Support for Democracy 6A.5 Baseline Models, Candidate Sect Important 7A.1 Countries Included in Cross-National Tests

55 75 76 77 77 92 101 116 140 140 141 142 143 157

Acknowledgments Looking back at the process of writing this book, I am amazed by the generosity I have encountered from all kinds of sources. This project would not have been possible without help and support from countless people (certainly more than I can recount here) at every stage. First, I wish to thank the members of my dissertation committee, who have suffered through more versions of these chapters than anyone else. Amaney Jamal has been the most valuable mentor I could possibly have imagined. From the beginning of my time as a graduate student at Princeton—actually, even before then—she consistently treated me as a colleague and offered unwavering support and advice on all aspects of my professional development, all of which continues to this day. I know of no other advisor who is as generous with her students in providing research resources, professional connections, and opportunities for co-authorship. It goes without saying that I would not be in this position today if it were not for her enthusiasm, input, and support. Chris Achen has been a vital source of feedback since my second year at Princeton, reading far more of my writing than could be considered reasonable, and not letting me cut any corners—all of this despite the fact that I am an alumnus and an employee of the University of Notre Dame, the one-time rivals of his beloved Michigan Wolverines. Carles Boix encouraged me to look at this project from different perspectives, and has been exceedingly generous with his time and energy in working to make the book better. Bob Wuthnow provided a wonderful environment at the Center for the Study of Religion, allowing me to bring together insights about religion from many other disciplines. My understanding of religion has improved immeasurably as a result. At Princeton, I benefited from a large and vibrant intellectual community. I received valuable feedback on pieces of this project from the Comparative Politics Graduate Research Seminar, the Religion and Public Life Workshop at the Center for the Study of Religion, the Arab Political Development Workshop, and the Fellowship of Woodrow Wilson Scholars. There are far too many individuals to name among these groups, but special thanks

xiv acknowledgments go to Chantal Berman, John Chin, Colby Clabaugh, Sharan Grewal, Sam Jaroszewski, Kevin Mazur, Dan Tavana, Matt Tokeshi, and Alienor van den Bosch, just a few of the many grad school colleagues and friends who have offered advice. Liz Nugent deserves special acknowledgment, having endured an especially large number of chapter drafts and presentations. My teachers and advisors at earlier stages were also instrumental in making this book possible. My eighth-grade teacher, Frank Merk, inspired my interest in politics. While I was a student at Notre Dame, Fr. Bob Dowd took a chance on me as a research assistant and gave me the opportunity to dive head-first into serious political research. Fran Hagopian devoted a considerable amount of energy to making sure that I got into grad school, and helped me with the process of obtaining a National Science Foundation Graduate Research Fellowship that ended up funding a considerable portion of my dissertation research (I’d never heard of it). Most of all, Michael Coppedge transformed me from a political science student to a political scientist. Looking back, I find almost unthinkable the amount of time he spent with me working on my senior thesis, graduate school applications, and other projects. The award he received for undergraduate mentoring was well deserved. I’ve since had the pleasure of teaching alongside him, and he has continued to serve as an essential lifeline as I navigate the challenges of an academic career. In addition to those already mentioned, the community at Notre Dame has been extraordinarily supportive and inspiring, from my days as an undergrad through today. The friends and colleagues I’ve found at ND are far beyond anything I could ever have expected, and it has been an indescribable pleasure to work and live here for the past several years. Among many others, I have received valuable feedback, mentoring, and fellowship from Jaimie Bleck, Dave Campbell, Michael Coppedge, Darren Davis, Bob Dowd, Amitava Dutt, Hannah Early Bagdanov, Victoria Hui, Tahir Kilavuz, Karrie Koesel, Geoff Layman, Scott Mainwaring, Jim McAdams, Clare O’Hare, Joe Parent, Dan Philpott, Ben Radcliff, Emma Rosenberg, Luis Schiumerini, Tim Scully, Jazmin Sierra, Nate Sumaktoyo, Susanne Wengle, and Christina Wolbrecht. The book has also benefited from generous funding from the Kellogg Institute for International Studies, the Institute for Scholarship in the Liberal Arts, and Notre Dame Research. Other colleagues and friends in the discipline have likewise been abundantly generous with their time and insight, including Lisa Blaydes,

acknowledgments xv Steven Brooke, Melani Cammett, Youssef Chouhoud, Dan Corstange, Tarek Masoud, Christiana Parreira, Tom Pepinsky, and Mark Tessler. I have been blessed to be surrounded by such a supportive family. My parents, Polly and John, always made their children’s education a top priority, and have worked tirelessly to make it easy for us to obtain as much education as we wanted. My mom passed away before the publication of this book, but her fingerprints are present throughout its pages. My grandparents on both sides made major sacrifices to provide an education for my parents, without which I would never have reached this point. Through their pride and encouragement, my grandmother (“Mom-Mom”) and late grandfather (“Poppy”) inspired me to keep pushing through a sometimes frustrating process. My brother and sister-in-law, Bryan and Kathleen, have been vital sources of support and entertainment since well before I began this project, and their daughter Leah, who was born in 2018, has been an absolute joy for all of us. More than anyone, I wish to thank my wife, Kathy. The sheer volume of annoyance, frustration, and, frankly, nonsense that I have put her through during this process (and before) will surely put her on the fast track to canonization. The fact that she has endured all of this while often dealing with medical difficulties makes her sacrifice all the more impressive. She allowed me to forgo a respectable salary in favor of six years of graduate school, spend money from our own savings on a survey, and move her across the country just so that I could work in a job that I love. Most amazingly, she has done so with a smile on her face. Very few people would be willing to make those kinds of sacrifices, and that is not lost on me. Our wonderful son, Luke, came into our lives as I was finishing the final revisions to this book, and Kathy has (to no one’s surprise) proven herself to be the best mother imaginable. There is no acknowledgment that I could give that would adequately capture the love and support that she has shown both of us.

1 Introduction Is organized religion too inextricably bound to the status quo to save our nation and the world? —Martin Luther King, Jr., Letter from Birmingham Jail

In October 2015, foreign ministers from more than a dozen international powers issued an important statement on the ongoing Syrian conflict. In this statement, they declared that “Syria’s unity, independence, territorial integrity and secular character are fundamental.”1 The first three features are straightforward: for Syria to establish peace, stability, and freedom, it must be united, independent, and firmly in control of its territory. The last criterion is far more perplexing. Why must a free, peaceful Syria also be secular? This seemingly benign statement highlights an often unspoken assumption about the role of religion. For many public figures and casual observers, religion is inherently threatening to peace and freedom. As Karl Marx (1970 [1843]) famously declared, “Religion is the sigh of the oppressed creature. . .it is the opium of the people.” In other words, religion creates weak, passive citizens, unwilling and ill-equipped to resist domination at the hands of unjust rulers. Democracy, in particular, is said to depend on secularization. These claims are particularly common in reference to the Middle East. Islam, it is so often argued,2 stands in the way of democracy. Such arguments are typically built on cultural premises: Islam is incompatible with democracy because it does not allow for any notion of separation of religion and state, or perhaps because it does not contain built-in ideas of human rights and freedoms. Islam, and sometimes other religions, are said to shift believers’ focus away from the here and now and toward the hereafter. In these accounts, religion seems to be an obstacle to democracy, incapable of contributing anything to the democratic cause, and far more likely to breed submissive or even pro-authoritarian citizens. Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0001

 faith in numbers But even a casual glance at the historical evidence reveals that the picture is far more complicated than these theories imply. In many cases throughout the world, religion—including Islam—has provided enormous support for democratic movements. In Brazil, Chile, Poland, and many other countries, the Catholic Church played a key part in democratization (Philpott, 2007). In the Muslim world, religious factors have motivated and facilitated antiauthoritarian mobilization in Tunisia, Pakistan, and Malaysia, to name just a few examples (Nasr, 2005). These examples fly in the face of simplistic arguments linking religion with authoritarianism. A better explanation is needed in order to account for the variety of outcomes that countries have experienced: religion can, and often does, produce pro-authoritarian citizens, but it often does exactly the opposite.

Key Observations Religion has been a powerful force for authoritarian rule in many cases, and democracy in many others. What explains this puzzling variation? I will argue that recognizing a few simple facts will go a long way toward explaining the considerable ambiguity in religion’s role in regime politics, and particularly in the way individuals view democracy. The first key observation is that religious groups matter. In divided societies, people’s religious identities affect their political preferences; this relationship is perhaps uniquely strong in the Middle Eastern countries that have the sharpest sectarian divisions. Citizens do not assess regime possibilities—or other political issues, for that matter—through a purely individualistic lens. Rather, their preferences are motivated in part by the way in which political change would affect the members of their religious group in general. Accounts of regime preferences that only take into account individual characteristics (most commonly, income) are insufficient to explain attitudes in settings where identity groups are important social and political categories. For reasons outlined subsequently, religion may be a particularly powerful identity with regard to political preferences. Second, a comprehensive account of religion and democratic attitudes must acknowledge the uncomfortable reality that not all groups have reasons to favor democracy. In non-democratic societies divided by sect, some groups will have more power than others, and some groups will be larger than others. In many cases, the incentives faced by each group will point

introduction 3 in opposite directions. For large groups that are left out of power (for instance, Shi‘a Muslims in Saddam Hussein’s Iraq), there is considerable motivation to undermine the regime and push for democracy. This incentive derives from the simple arithmetic of electoral politics: larger groups will be more likely to win elections, so a transition to democracy would mean that the large, disenfranchised group would have a good chance of taking power. The other side of the coin, however, is that for privileged or smaller groups, democracy may be seen as a threat. For these sects, the prospect of democratization can create fears of a “tyranny of the majority,” where their religious freedom, property, and even lives might be in jeopardy. It is imperative, therefore, to consider what democracy means in a given context as well as how individuals respond to various possible implications of democracy. Despite the normative value of democracy, it does not represent the same possibilities for everyone. Third, religious behaviors affect group identity. In particular, communal religious practice tends to heighten sectarian identity in divided societies. The act of communal worship involves an experience shared, by definition, with members of the same sect. In some cases, the ways that such practices build sectarian identity are straightforward: religious sermons often contain group-centric or even explicitly political content, and congregants may discuss political or sectarian issues with others in the group. Even in the absence of such on-the-nose influences, however, communal worship can enhance sectarian identity. The shared experience of worship increases feelings of “closeness” with others in the group, and it is more than likely that this closeness will spill over into the realm of political preferences. A corollary of this claim is that individuals who frequently participate in communal worship will tend to view politics in a more sectarian way. Their preferences will be shaped more by group concerns than will those of individuals who do not participate in group worship. For issues of democracy, this change in viewpoint is of considerable importance: the more an individual attends communal worship services, the more closely her regime preferences will align with the interests of her sect. Taken together, these observations can help to account for the ambiguous relationship between religious behaviors and regime preferences. Communal religious practice pushes individuals’ political attitudes into closer alignment with the interests of their sects, but the direction of this effect on any given political issue is not initially clear. The direction of the effect depends on how each political possibility would affect each sect.

 faith in numbers For groups that would “win” in the event of democratization (typically large groups, underrepresented groups, or poor groups hoping for redistribution), communal practice would increase support for democracy. In such cases, the individual’s group has reasons to favor democratization, and the more the individual identifies with her sect, the more she will tend to espouse that view. On the other hand, for small, overrepresented, or wealthy groups, the exact opposite incentives will be present. In these cases, communal prayer will have an anti-democratic effect, due to precisely the same logic. This theory inevitably confronts an inferential problem. While I will argue that participation in communal worship heightens the salience of sectarian identity and, through this mechanism, promotes political preferences that are responsive to sectarian interests, it is possible that the causal arrow points the other way. In other words, it is possible that individuals who are already “more sectarian” choose to participate in communal prayer precisely because of this preexisting sectarian identification. A key focus of the empirical portions of this book—especially chapter 5—will be the selection issues that could undermine tests of this theory. Through a variety of statistical, experimental, quasi-experimental, and qualitative techniques, I will attempt to demonstrate that while selection effects might be present to a certain extent, there is a substantial independent effect of communal worship on support for democracy, and that this effect is channeled through a mechanism of sectarian identity. Importantly, the argument of this book is not based on culture. It is possible that distinct values of different religious groups affect regime attitudes in one way or another, but arguments of that kind tend to obscure more than they reveal. Rather, following a growing literature on rationalist or semi-rationalist approaches to religion and politics, I am suggesting a description of religion and regime preferences that is based on identities and interests. The reasoning described previously does not presume any particular political theology. Instead, it addresses how a common behavior (communal practice) shared by most of the world’s major religions can affect democratic preferences in a predictable way. It does so through a mechanism that does not distinguish between “pro-democratic” and “antidemocratic” religious groups from a theological perspective, but instead highlights religion’s role as an identity. Religious and democratic concepts may sometimes be intertwined in what seem to be the least likely environments. In 1979, during the Iranian Revolution, Ayatollah Ruhollah Khomeini began to invoke democratic

introduction 5 concepts in religious settings, at one point going so far as to declare during a sermon that democracy is an inherent feature of Islam, saying: “Democracy is incorporated into the Quran and people are free to express their opinions and to conduct their acts. Under the Islamic government, which is a democratic government, freedom of expression, opinion, and pen will be guaranteed for everyone” (quoted in Kian, 2014, p. 181). This statement, delivered by a figure considered in the West to be a stalwart autocrat, highlights the ways in which democracy and religion can interact for strategic purposes. In other settings, Khomeini vocally opposed democracy as a foreign, anti-Islamic form of government, but when he had instrumental reasons to promote democratic principles, he presented democracy as not only compatible with Islam, but an essential part of it. This type of democratic-religious shape-shifting is common. Religious leaders and ordinary religious individuals will often view the link between religion and politics through a filter of instrumental calculations. When democracy is desirable for one reason or another, religion tends to be used to support it; when democracy is a threat, however, religion is often presented as a reason to oppose democracy. Both extremes of this relationship can be present within the same religious tradition, indicating that theological differences are typically not to blame for pro- or anti-democratic sentiments. As the example of Khomeini demonstrates, even the same individual may use religion for seemingly pro-democratic purposes at one point, but argue against democracy on religious grounds at another point. At the height of the Islamic Revolution, Khomeini was a revolutionary; once his regime was consolidated, he was an autocrat. In both cases, religion served his political purposes.

Religion and Politics: Individuals, Identities, and Interests Despite the predictions of secularization theorists, there are more “traditionally religious” people on the planet than ever before, and they represent a growing share of the global population (Norris and Inglehart, 2004, p. 5). Moreover, the differences between religious and nonreligious individuals are in many ways sharper than previously realized.3 These twin trends make it more dangerous than ever to ignore the role of religion in public life. At the same time, however, there is little reason to treat religion as a purely “irrational” phenomenon, separate from the calculations of economic and

 faith in numbers political life. Although religion inevitably involves norms and values that cannot be understood through a lens of pure rationality, certain religious behaviors and preferences may nevertheless be related to self-interest.4 Such features are likely to operate unconsciously; that is, individuals themselves may not be aware that rational considerations are partially responsible for their behavior. Rationality may operate “beneath the surface” with regard to religion, but it can still provide an effective lens through which to analyze the consequences of religious beliefs and practices. Fortunately, recent work5 has begun to consider how traditionally “rational” considerations might interact with religion to determine political and social outcomes, and this book builds on the insights of those studies. In particular, a number of works highlight the ways in which rational calculations influence religion-state relations. Using the example of the Catholic Church in Europe, Kalyvas (1996) points out that religious groups do not approach democracy solely from an ideological perspective; they also consider how the prospect of democratization would affect their institutional interests. In Latin America, Gill (2001) argues, the Catholic Church responded to pressure for regime change differently based on the level of competition it faced from other groups.6 Toft et al. (2011) suggest that political theology alone does not explain the role that religious groups choose to play in democratization processes; their relationships with the state are also highly important. Importantly for this book, some studies have considered the links between religious and rational factors at the individual7 rather than group level. Scheve and Stasavage (2006) argue that religiosity influences attitudes toward state-provided social insurance through a rational mechanism: religion compensates individuals in the case of adverse life events, effectively substituting for state support. On the issue of left voting, De La O and Rodden (2008) suggest that religion “cross-cuts” self-interested preferences for redistribution, adding a “moral values” dimension to individuals’ beliefs (see also Stegmueller, 2013). Each of these works underscores an important point: religion and rationality are not necessarily separate spheres. Interest-based concerns are not always anathema to religion, but the types of interests valued by believers might be systematically different from those of nonbelievers or members of other groups. This book explores how religious behaviors heighten the importance of group-centric interests, translating spiritual practices into political preferences through a decidedly rational mechanism.

introduction 7 In this book, religion is treated as a multilevel phenomenon that interacts with politics in several ways. Most importantly, individual beliefs and behaviors influence political attitudes. The issue of individual regime preferences is a crucial one, and has fortunately begun to receive more attention in recent years. Traditionally, the attitudes of “the masses” have been regarded as relatively unimportant, particularly in developing countries where institutions are often ill equipped for (or downright hostile to) the representation of the interests of ordinary people. In the wake of substantial social movements across much of the developing world—most dramatically in the Middle East—we now know better. Recent mass protest activities across the world highlight the continued importance of the attitudes of ordinary citizens; views toward regimes are of particular importance. In addition to the much-discussed “Arab Spring,” widespread protests have threatened governments in Greece, Turkey, Brazil, and elsewhere. The uncertainty about the types of regimes that will succeed the now-deposed autocrats in several countries in the region makes understanding citizens’ attitudes about democracy all the more important: if the people were able to bring down entrenched dictators, what type of regime would they demand as a replacement? In many of these movements, religion has been at the center, either explicitly or implicitly—but little is known about how religion motivates pro- or anti-regime behavior. This book represents a rare attempt to explain systematically why religion sometimes bolsters democracy and sometimes undercuts it. These outcomes are important. In the Middle East and elsewhere, ordinary citizens were for too long presumed to be passive, powerless subjects incapable of effecting real political change. But recent experiences in Tunisia, Egypt, Libya, Algeria, Sudan, and elsewhere have shattered this assumption. It has become clear that the attitudes of the masses can have a decisive impact on the survival or destruction of existing leaders and institutions. For dictators across the region and beyond, the threat of revolution has become real. Preferences and demands for democracy serve as a constraint on regime behavior (Besley and Persson, 2019) and a signal to other citizens about the potential costs and benefits of mobilization (Welzel, 2007). Support for democracy and for the norms that go along with it is widespread, but far from universal (Wike and Fetterolf, 2018), and may be declining (Foa and Mounk, 2017). Moreover, recent research suggests that public support for democracy helps democracies to survive (Claassen, 2019a, 2019b). The mechanisms through which public support promotes democratization or

 faith in numbers protects democracy from threats are not always dramatic, but they are important in both non-democracies and the increasingly large category of at-risk democracies. If religious factors can explain variations in support for democracy, then they may also have important effects on macro-political outcomes.

Plan of the Book The remainder of this book will explore the issues described in this introduction, presenting and testing a theory of religion and democracy. It will do so in several sectarian contexts across a number of dimensions of political competition. Chapter 2 presents a new theory of religion, sectarian interests, and regime preferences. Religious behaviors shape regime preferences, and do so through a sectarian lens. Communal religious practice heightens the intensity of sectarian identity, and, in doing so, frames regime politics as a group issue. Depending on the interests of the group with respect to democracy (namely, the rights and privileges that a group would gain or lose in the event of democratization), communal prayer may have pro- or anti-democratic effects. A number of potential benefits and threats may accompany democracy; certain groups may gain or lose political voice, while others may benefit or suffer due to economic redistribution. In either case, group interests help to predict when religion will enhance support for democracy—and when it will do just the opposite. Chapter 3 describes the religious experience in sectarian environments as expressed by the participants themselves. This chapter provides essential texture to the analysis by allowing worshipers to speak for themselves. Using responses from open-ended interview questions in both Lebanon and Iraq, it reveals the ways in which communal worship promotes sectarian solidarity and group-centric political preferences. The Lebanese interviews illustrate the link between communal worship and political preferences. Distinct themes emerged between sects; while communal prayer heightened sectarian identity for all sects, each sect reported different political messages. For Christians, the emphasis was on preserving their community’s privileges in a changing political landscape. For Sunnis, the emphasis was on avoiding divisions imposed from outside. For Shi‘a, political messages stressed political and economic marginalization and called for an end to the sectarian

introduction 9 system. In each of these cases, interviews indicated that religious-political messaging in places of worship clearly reflects political circumstances and sectarian interests, a relationship explored in depth in subsequent chapters. Chapter 4 examines perhaps the most famous illustration of sectarian politics: Lebanon. It provides an account of Lebanon before the Syrian Civil War using data from the first wave of the Arab Barometer, conducted in 2007. This chapter considers communal prayer and support for democracy through a Christian-Muslim sectarian lens. Before the war in Syria—which pitted Sunnis and Shi‘a against each other in a more severe way than any time in the country’s recent history—it was possible for both Sunnis and Shi‘a to consider issues of representation through a Christian-Muslim perspective rather than through the Sunni-Shi‘a divide. Since both Muslim sects were underrepresented in politics and relatively poorer than Christians, their incentives with respect to democracy were aligned, despite their differences on other issues. Communal prayer therefore had a pro-democratic effect among Muslims, while the opposite was true for Christians, who sought to maintain their disproportionate influence in Lebanese politics as well as their greater wealth. Chapter 5 reveals how changing political circumstances—a new axis of political competition—dramatically altered the political interests of Lebanon’s sects and therefore adjusted the effect of communal prayer on regime preferences. Using an original, nationally representative survey of over 1,200 Lebanese respondents, I show that the Syrian conflict has centered political contestation around the Sunni-Shi‘a cleavage, with Christians divided on their relationship with each of these sects and on the Syrian conflict itself. Sunnis, relatively better represented and wealthier than Shi‘a, now have reason to fear Shi‘a ascendance in Lebanese politics. Since the new lines of political conflict have created a zero-sum situation between these two sects, democracy is no longer a palatable option for many Sunnis. Consequently, the effect of communal prayer on regime attitudes has shifted: for Shi‘a, mosque attendance continues to have a pro-democratic effect, but for Sunnis, this effect has reversed; fearing Shi‘a dominance, practicing Sunnis now tend to oppose democracy more than their less observant counterparts. Chapter 6 considers Iraq, a case in which majoritarian and redistributive understandings of democracy imply very different preferences for the main sects. When democracy is believed to be a fundamentally political arrangement (elections, freedom to criticize government, etc.), Shi‘a (the majority

 faith in numbers group) have reason to favor democracy, while the formerly powerful Sunni minority have reason to oppose it. Communal prayer pushes individuals’ regime preferences further in the direction of these sectarian interests; that is, mosque attendance increases support for democracy among Shi‘a but decreases such support among Sunnis. However, when democracy is considered in economic terms, namely, narrowing the gap between the rich and the poor, the interests of each sect change. Sunnis, now poorer than Shi‘a on average, have an incentive to support redistributive democracy, while Shi‘a have reason to fear democratic redistribution. The effect of communal prayer follows these same patterns: in this case, mosque attendances enhances support for democracy among Sunnis but has the opposite effect among Shi‘a. Evidence from the second wave of the Arab Barometer, conducted in 2011, is used to support each of these claims. Chapter 7 concludes, placing these findings in the context of religion and political behavior broadly considered. It tests my theory in a large sample of countries using the World Values Survey. Cross-national tests indicate that the pattern described in the preceding cases is evident in much of the world: the general trend is that for small sects, communal prayer decreases support for democracy, while the opposite is true for large groups. Larger groups can expect to benefit from free elections due to their sheer size, so increased salience of sectarian identity—such as that created by communal worship—should promote democratic attitudes; the reverse logic holds for smaller groups, who would be unlikely to win elections. Finally, the chapter uses the suggestive evidence from the World Values Survey to describe some of the conditions under which this theory should—and should not—apply.

2 An Interest-Based Theory of Religion and Support for Democracy Introduction Why does religiosity sometimes promote democratic attitudes and behaviors and sometimes have just the opposite effect? On the one hand, religious involvement is often argued to be positively associated with political engagement (Putnam, 2000) and mobilization against injustice (McAdam, 1982; Cone, 1986); but on the other hand, religious beliefs and practices are also cited as key tools for authoritarian stability (Huntington, 1996; Kalyvas, 2000). Even within the same religious tradition, the Janus-faced nature of religion is evident: while the Catholic Church maintained a strongly anti-democratic outlook for centuries in Western Europe and elsewhere, it eventually became a powerful tool for pro-democracy movements in Eastern Europe (Philpott, 2007). While some of these changes are explained by historical events unique to the contexts in which they developed, a broader theoretical explanation is possible (and important). In this chapter, I present a theory that focuses on the communal aspect of religion. The theory suggests that communal prayer can have either a pro- or anti-democratic effect on support for democracy depending on the sectarian interests of the individual’s group. Religious identity politics have become more prominent in many parts of the world. The spread of Islamist militant groups in much of the Arab World as well as parts of Africa has brought sectarian tensions to a new level in those regions. Concern about the possible fall of the ‘Alawite regime in Syria has brought fear to religious minorities in neighboring countries, including two countries examined in this book, Lebanon and Iraq. The changing power configurations in these countries has led to the rise of a new religious politics, in which regimes are increasingly viewed through a sectarian lens, and religion’s influence on democratic attitudes is especially strong. These developments require a new explanation. Why is it the case Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0002

 faith in numbers that in many of these settings, religion has served pro-democratic purposes, but in others, has done exactly the opposite?

Religion and Political Behavior: Remaining Ambiguities Despite growing awareness that “religious loyalties continue to structure political thought and action” (Grzymala-Busse, 2012b, 433), many comparativists have, until very recently, hesitated to study religion as an independent variable because of both theoretical and methodological limitations: as Bellin (2008, 316) writes, “The subfield has still failed to reckon with the power of religion as an independent variable, the noninstrumental aspect of religious behavior, and the malleability of religious ideas, as well as their differential appeal, persuasiveness, and political salience over time”. Part of the problem lies in the difficulty in determining at which level we should examine religion. On the one hand, religion is an individual-level phenomenon; individuals themselves choose whether or not to pray, to attend religious services, and to follow the prescriptions of their faith. At the same time, religion possesses important institutional structures. The majority of the world’s population profess one of only a handful of faiths, and believers in the same tradition should, presumably, resemble each other in significant ways. After decades of research, much of the link between religion in its various forms and individual political behavior remains unclear. Much of the existing scholarship focuses on denominational differences in political behavior rather than the effects of religious practices. Extending back at least to Weber (1922), this tradition generally examines the extent to which members of certain sects demonstrate meaningfully different political attitudes and behaviors compared to members of other sects, ignoring within-sect heterogeneity in religious beliefs and practices.1 As Layman (1997, p. 289) states, denominational affiliation “traditionally has been the only religious variable available in studies of political behavior” despite its very limited information content. In the subfield of comparative political behavior, studies of this kind have examined voting behavior (Lipset and Rokkan, 1967; Lijphart, 1979; Knutsen, 2004) and attitudes toward social issues (Kellstedt and Green, 1993; Adamczyk and Pitt, 2009), among other topics. A question of particular interest in this debate is whether Islam is compatible with liberal, tolerant, and democratic attitudes (Kedourie, 1994;

a theory of sectarian interests and democracy 13 Huntington, 1996; Kubba, 1996; Diamond et al., 2003; Hofmann, 2004). Unsurprisingly, these studies report a wide range of results and suggest no consensus on the link between Islamic affiliation and democratic attitudes. When we consider behavioral rather than denominational differences, even more ambiguity is present. On the issue of civic and political engagement, studies have suggested that religiosity (variously considered) can have a pro-civic effect (Smidt, 1999; Putnam and Campbell, 2010; Jamal, 2005), an anti-civic effect (Putnam, 1993; Schwadel, 2005), or both (Campbell, 2004; Driskell et al., 2008). Comparative studies of religion and attitudes toward the welfare state are similarly dissonant: some scholars suggest that religious engagement will have a consistently anti-redistributive effect;2 others state that there is no clear link between these variables (Pepinsky and Welborne, 2011); and still others argue that external factors (such as electoral concerns) determine the direction of this relationship (Amat and Wibbels, 2009).

Religion and Regime Attitudes Perhaps the clearest example of the uncertainty over religion’s influence in political behavior is found in attitudes toward democracy. Existing studies of the relationship between religion (particularly, personal religious behaviors) and democracy have usually focused primarily on individual-level characteristics or state-level factors without allowing for much interaction between the two.3 In accord with the works discussed previously, studies examining religiosity’s effect on regime attitudes are quite unclear. Scholars have suggested that religiosity depresses support for democracy in many settings, including Brazil (Geddes and Zaller, 1989), Iran (Tezcur et al., 2012), Bosnia (Valenta and Strabac, 2012), and Central Asia (Collins and Owen, 2012). These studies largely focus on the theological and attitudinal components of religion, which are associated with personality types that are said to be incompatible with democracy. Among these characteristics are authoritarian orientations, conservatism, susceptibility to propaganda, and a desire for religious involvement in government. Other studies have suggested that religiosity may actually enhance support for democracy. In the United States, Djupe and Calfano (2012) challenge the claim that religion is, at best, harmless with respect to democratic norms, arguing that religion can play an important role in promoting such values. Gu and Bomhoff (2012) report that religiosity increases support for

 faith in numbers democracy in both Catholic and Muslim countries, two religious groups that have commonly (at least until recently) been labeled as anti-democratic (Huntington, 1996). Similar results have been noted in India (Chhibber, 2014), Nigeria (Bratton, 2003), and much of the Arab world (Rizzo et al., 2007), among others. These studies cover a variety of religious traditions, and typically emphasize the ties between religious piety and concern for justice and accountability. Some authors have also suggested other mechanisms: increased civic and political engagement, political tolerance, and egalitarian attitudes. Meyer et al. (2008) make an important contribution to this literature by considering both individual-level religious variables and national context. At the individual level, they find that “people who find religion central to their lives and those who find it central to their politics think differently about democracy. The former favor democracy while the latter do not” (Meyer et al., 2008, p. 645). Their findings thus challenge many of the assumptions of a consistently negative relationship between religiosity and democratic attitudes.4 Still others claim that religiosity does not heavily influence regime attitudes in any way. Using survey data from the Arab world, Tessler (2002) suggests that Muslim piety has little effect on attitudes toward democracy (and less of an effect on political attitudes in general than is commonly assumed). Ciftci (2010) likewise argues that “religiosity and Islamic values poorly predict support for democracy in the Muslim world”; Kim (2008) reports similar findings in a cross-national setting. More recently, Ciftci (2013) finds that Arab citizens display high levels of support for both democracy and Islamic law, and generally do not see a conflict between the two. These studies suggest three possibilities: first, that religiosity simply has no effect on democratic attitudes. Second, it may be the case that one aspect of religiosity promotes democratic attitudes, while other aspects do the opposite. Finally, it is possible that religiosity enhances support for democracy for some people, but this effect is canceled out by an antidemocratic effect among other people. The latter two possibilities highlight the need to consider variation in the effects of religious behaviors on regime preferences. A final category of this literature argues that religiosity can either promote or depress support for democracy. Bloom and Arikan (2012) suggest that religion has a “double-edged” effect on democratic attitudes depending on whether individual belief or social behavior is used as the measure of religiosity. Bloom and Arikan (2013) expand on this claim, arguing that

a theory of sectarian interests and democracy 15 religious belief reduces support for democracy through traditional/survival values, while social religious behavior promotes democratic attitudes both directly and through mechanisms of political interest and trust. This line of thinking usefully suggests that different aspects of religious behavior can demonstrate different and even contradictory effects on regime attitudes. Each of these categories of literature reveals important information about the link between religious behaviors and political preferences. Several of these studies acknowledge the importance of addressing religion at multiple levels and in various forms, bringing greater refinement to the study of religion and politics. This book contributes to this important and growing literature by addressing group-level interests in addition to those at the individual and state levels. It builds on existing studies by demonstrating how sectarian interests can help to explain the often confusing relationship between religious behaviors and political preferences. What is clear from the extant research on this topic is that the relationship between religious behaviors and democracy is complicated. Philpott (2007, p. 510) observes that even within the “Third Wave” of democratization, religion has played a highly important role in promoting democracy in some places (e.g., Poland, Lithuania, and Indonesia) but very little role in others (e.g., Argentina and Senegal). He attributes these differences to divergent patterns in political theology and religion-state relations: the degree of fusion between religious and political institutions as well as the level of tension between them. The effect of religious practices on democratization therefore appears to be conditional not only on the contents and political orientations of religious groups themselves, but also on these groups’ orientations toward the state. My theory of religion, sectarian interest, and regime preferences considers individual and sectarian orientations toward the state in order to provide a coherent theory of the relationship between communal prayer and democratic attitudes.

A Theory of Religion, Sectarianism, and Democracy This book takes a significant step toward reducing the ambiguity of the effect of religious behavior on attitudes toward democracy, and explaining “the political ambivalence of religion” (Philpott, 2007; Appleby, 1999). I argue that the effect of one aspect of religiosity—communal prayer— depends heavily on both state-level and sectarian factors. Building on recent

 faith in numbers studies in religion and political behavior, I examine some of the political contextual factors that may influence this relationship at the level of the country, sect, or individual. I present a theory of why and how religious behaviors have certain effects in particular political-sectarian environments. I focus on one possible mechanism, identity, that is particularly important in religiously divided societies. While other factors (including information access, congregational monitoring, and others) undoubtedly play a part in linking religious practices with political preferences, this book explores the importance of communal prayer as an identity-strengthening aspect of the religious experience. A crucial—but often ignored—fact about religion as a social phenomenon is that it is multidimensional (Harris, 1994). I have argued that much of the confusion surrounding the relationship between religiosity and democratic values is caused by the literature’s tendency to treat “religiosity” as an aggregation of many different types of behaviors rather than considering each behavior on its own.5 Even in works that pay careful attention to the details of specific religious behaviors and practices, however, there is a tendency to assume that the effect of a given religious practice will be constant across space, time, and individuals. In subsequent pages, I will argue that the effect of one aspect of religiosity—communal prayer—depends heavily on both state-level and sectarian factors. Existing studies often discuss the effects of either sectarian affiliation or various types of religious behaviors without addressing higher-level factors. Some arguments acknowledge variation across countries; fewer address differences among sects, and almost none consider political contextual factors that may influence this relationship at the level of the country, sect, or individual.6 What is needed is a coherent, transportable theory of why and how religious behaviors have certain effects in particular political-sectarian environments. It is important not only to acknowledge that variation exists at each of these levels; we must also explain this variation from a theoretical standpoint. Some of the ambiguity remaining in existing work on this subject comes from confusion about what is meant by “religion.”7 Durkheim (2001) famously distinguishes between “beliefs” and “rites,” suggesting at least two dimensions to the concept of religion, an individual dimension and a group dimension. Indeed, Durkheim (2001, p. 42) identifies the communal aspect of religion as one of its key features:8 “Religious beliefs proper are always held by a defined collectivity that professes them and practices the rites that go with them. These beliefs are not only embraced by all the members

a theory of sectarian interests and democracy 17 of this collectivity as individuals, they belong to the group and unite it. The individuals who make up this group feel bound to one another by their common beliefs.” For Durkheim, this notion of “church” distinguishes religion from magic: “Religion must be something eminently collective” (2001, p. 46). Following Durkheim, I emphasize the importance of communal prayer, because it is in this part of the religious experience that individuals interact with co-sectarians. Since group concerns are central to issues of representation in much of the world, it is essential to account for the factors that influence sectarian identity and the political behaviors that result from such identity. I do not doubt that other aspects of individual religiosity influence regime attitudes and political behavior; however, since private religious experiences are, by definition, personal and solitary, these experiences will be much more individual-specific and therefore more difficult to observe and analyze. Further, private practices do not embody the same sectarian content as found in communal prayer. There is therefore less reason to believe that private religious behaviors will interact meaningfully with sectarian interests in a way that influences attitudes toward democracy. Even communal prayer, however, may have different effects on regime preferences for different individuals and sects. I consider how political context helps to determine when and where communal prayer will promote or inhibit democratic attitudes. In doing so, I analyze several levels: countries, sects, individuals, and features of democracy. I posit that the effect of communal prayer on attitudes toward political and redistributive policies depends on sectarian interests, which in turn are determined by the statelevel configuration of political power as well as the particular issues under contention. In this way, I do not assume that the effect of communal worship is constant across time periods, countries, and sects. Context is not only important; it is absolutely vital in predicting this effect, and the effect can reverse itself in response to changing political conditions, particularly those related to the sectarian environment. This book is distinct from earlier studies of religion and political attitudes in that it seeks to explain attitudes through the conditional effect of religious practice rather than through group identity, membership, or personal piety. While individual piety might demonstrate some effect on attitudes toward democracy independent of the communal aspect of religion, the theory presented here is about collective religion. In this sense, the emphasis of this book is on what Wald (1987) calls the “organizational” and

 faith in numbers “social interaction” dimensions of religion. The mechanisms outlined in this theory depend on collective religious practice, which in turn heightens the importance of group interests, thereby influencing attitudes toward democracy at the individual level. The theory presented in this book applies to religiously divided societies, and is most applicable to countries in which certain groups enjoy disproportionate political and/or economic power relative to others or relative to their share of the population. It is difficult to define sharp scope conditions on a theory of this kind, and its applicability is likely to be a matter of degree rather than a firm “applicable / not applicable” dichotomy. However, the key factors influencing the applicability of the theory to a given case are the level of democracy—because consolidated democracies are unlikely to witness this form of contention regarding the issue of democracy itself— and the political salience of religious divisions, because citizens of countries where religion does not constitute a political fault line are not likely to prioritize sectarian interests per se when evaluating regime types. Both of these boundary conditions will be explored empirically in chapter 7. I argue that the effect of communal prayer on regime preferences—and, through a similar logic, redistributive preferences—depends on group and state-level factors. The example of redistributive preferences illustrates this logic quite clearly: when an individual’s religious group would, on balance, benefit from further redistribution, communal prayer will tend to promote support for redistribution; if the group would be a “net payer” under redistributive policies, communal prayer will have the opposite effect. Importantly, these effects operate separately from an individual’s personal incentives with respect to redistribution. Wealthy members of poor sects may oppose redistribution, and poor members of wealthy groups may likewise favor it, as predicted by the canonical theories of redistributive preferences (Meltzer and Richard, 1981). My argument is not that personal interests do not influence redistributive attitudes; rather, I am suggesting that group interests factor separately into individuals’ preferences for redistribution. Thus, while wealthy members of poor sects might oppose redistribution, I expect that the most frequent attenders among this subset will be more likely to support redistribution than those who do not engage in communal prayer, holding constant their own income level. In other words, a wealthy member of a poor sect might be inclined to oppose redistribution on the grounds that she is wealthy. The more she cares about her sect, however, the more supportive of redistribution she will tend to be, because her group is poor.

a theory of sectarian interests and democracy 19 The basic intuition of this theory is as follows: in religiously divided societies, communal religious practice tends to promote political attitudes and behaviors in which individuals begin to think more along sectarian lines; that is, their political preferences reflect sectarian considerations more than they otherwise would. As a result, individuals who engage in communal religious practice are expected to support democracy more or less depending on the interests of their group. The terms “divided” and “religiously divided” are not meant to imply a dichotomy. Countries, whether generally or at a fixed point in time, cannot be separated neatly into the categories of divided and not divided. It is more appropriate to consider the extent of division rather than the presence or absence of it. For these purposes, it may be useful to use a modified version of Rabushka and Shepsle’s (1972, p. 21) concept of a plural society,9 which exists if the society “is culturally diverse and . . . its cultural sections are organized into cohesive political sections.” The arguments presented in this book are, all else equal, more applicable to more “divided” settings than to less divided ones. Unfortunately, this type of concept does not lend itself to easy measurement or classification, but the more that religious differences serve as an organizing force in political competition, the more applicable my theory will tend to be. In divided societies in which certain groups are substantially better off than others (whether politically, economically, or both), relatively disadvantaged groups that are large in size have an incentive to support more meaningful democratization, since a shift toward “rule by the people” would presumably increase the amount of power wielded by larger groups. A democratic transition would provide political privileges to the largest group. Thus, it is in the interests of these groups to support democracy. If a smaller group is in power in a non-democratic country, a democratic transition threatens not only their political influence but their guarantees of tolerance as well. If the majority takes power, as would be expected in a democracy, there may be no mechanisms to ensure that the rights of religious minorities will be protected; this is a particularly likely possibility if the minority group was previously politically dominant. By the same logic, an out-of-power majority group should prefer democracy, because by gaining control of the state, it can ensure its own religious rights. Cases like Syria and pre-2003 Iraq demonstrate that dictators from minority religious groups often fear the potential destabilizing effects of religious identity, and therefore seek to restrict the rights of other sects. The electoral and sectarian interests of a

 faith in numbers given group therefore point in the same direction: for large sects (especially those without political power), democracy is a preferable alternative; for small sects (especially those currently in power), democracy creates a fear of the tyranny of the majority.10 Similar reasoning can be applied to the economic realm, reflecting patterns in wealth distribution rather than demographic weight alone. A transition to democracy will likely lead citizens to expect an increase in the wealth of these relatively impoverished groups because democracies are often perceived to redistribute more than non-democracies (Boix, 2003; Acemoglu and Robinson, 2006). Consequently, poor sects will tend to have reasons to prefer democracy. The incentives for the better-off groups are exactly the opposite: since a transition to democracy would presumably decrease their share of power by dividing power more equally according to group size, they would lose from democratization. Accordingly, it is perhaps in their interest to prevent democratization. Consequently, this theory predicts that the positive effect of religious practice on support for democracy should not be present for privileged groups; indeed, religious practice should lessen support for democracy among these groups.11 In this theory, when I refer to “privileged” groups, I am describing groups who have significant political power, economic resources, or both, particularly if those privileges are disproportionate to their demographic weight. The comparison in this book between privileged and underprivileged groups is related to the notion of “between-group inequality” as described by Baldwin and Huber (2010). They suggest that when material conditions differ considerably between ethnic groups, public goods provision suffers. This finding highlights a crucial building block of the theory advanced in this chapter: economic (particularly, redistributive) policies are not always targeted at individuals simply based on their level of wealth; redistribution along the lines of identity groups is often an essential step in the political process.12 In this way, democracy (with its presumed redistribution) should signal different prospects to different groups, not just to relatively rich or poor individuals. Depending on what citizens believe democracy to mean, democratization could influence power structures through a multitude of political and/or economic channels. My theory emphasizes the interests of religious sects. Recent work on religion and democracy rightfully addresses the instrumental calculations made by religious actors with respect to the

a theory of sectarian interests and democracy 21 prospect of democracy.13 While doctrinal teachings and religious leaders undoubtedly influence believers’ perceptions of democracy,14 the group interest dimensions of religion’s relationship to the state deserve greater attention than they have received in most existing studies on this topic. According to this theory, communal religious practice increases the salience of group identity and therefore makes individuals more likely to espouse political beliefs and behaviors that promote the interest of their group, holding constant their individual interests. This is not to say that individual interests do not matter—on the contrary, many factors influence levels of support for democracy. Rather, this theory is claiming that holding constant the other factors that influence attitudes toward democracy, communal religious practice will tend to promote attitudes in line with the interest of an individual’s religious group. Likewise, this study does not arrive at any conclusions about the cumulative effect of all forms of religiosity; instead, it focuses on the particular phenomenon of communal religious practice as translated through sectarianism.

Sectarianism The way that sectarianism manifests itself in political and social life is neither fixed nor inevitable. The salience of sectarian identity waxes and wanes over time in every sectarian environment, and is a product of circumstances, not merely a cause of them. It is essential, therefore, not to reify the concepts of “sect” or “sectarianism” by presuming them to be infinitely historical or enduring. At the same time, however, the historical contingency of sectarianism does not diminish its importance. As Deeb (2017) explains in the context of Lebanon, “We can insist that sect is a constructed and contingent category and still understand it to hold meaning and affect daily life and interpersonal relationships.” It is well beyond the scope of this project to give a comprehensive explanation of the origins of sectarianism, but it is nevertheless important to highlight some factors that contribute to its rise. The mere presence of different religious groups in a country does not necessarily imply a sectarian environment. Sectarianism requires a certain salience to be attached to religious identity. Many religiously diverse countries would not qualify as “sectarian” to any significant extent because their religious divisions are not particularly politically relevant. In this sense, the religious identity politics

 faith in numbers underwritten by sectarianism is not an inevitable consequence of religious diversity. Why, then, do some countries experience sectarianism in much more extreme ways than other equally diverse countries? A full account of the causes of sectarianism is not possible, but a brief consideration of sectarianism’s possible origins will help to clarify the concept. Sectarian institutions often emerge—either organically or by imposition—in environments where sectarian identity is salient. Equally important, however, is the effect moving in the opposite direction: sectarian institutions in many ways create (or at least strengthen) sectarian citizens. A growing literature across the social sciences explores the ways in which identities are constructed,15 and sectarianism is no exception. In contrast to the “ancient hatreds” discourse favored by many Western pundits and policymakers, sectarianism is constantly evolving. It is produced and reproduced by a variety of factors, many of which are endogenous to the political environment. Rarely are sectarian identities entirely or mostly driven by simple doctrinal differences (Liechty and Clegg, 2001). In various environments, sectarianism has been traced to colonial history (Makdisi, 2000), social and political upheaval (Abdo, 2017), political and legal institutions (Weiss, 2010), social and political pressures (Brooke, 2017), and external intervention (Haddad, 2011), among others. The degree to which sectarian identities are internalized at the individual level and formalized at the institutional level is thus not a simple matter of theological difference. It is a product of circumstance, and often far more political than religious in nature. It would therefore be misguided to consider the origins of sectarianism in a political vacuum. This book will illustrate how political and economic power configurations can drastically alter the way that sectarianism is adopted and articulated. Intergroup differences in power are often far more salient than differences in theology, and changing political conditions can lead to significant shifts in the sectarian environment even in the absence of any meaningful religious change. Religion and politics, as is so often the case, cannot be artificially separated. Rather than seeking a one-size-fitsall theory of the origins of sectarianism, it is important to acknowledge the ways in which political, social, and economic context can lead to the rise or fall of sectarian identity. In other words, the origins of sectarianism—as well as its consequences—in many ways need to be evaluated on a case-by-case basis.

a theory of sectarian interests and democracy 23

Axis of Political Competition What determines how groups define their goals with respect to other sects? I argue that the axis of political competition is key. Depending on which groups view themselves to be in competition—and what they are competing over— the same group may experience vastly different trends with regard to religion and political behavior; this concept is similar to “situational salience” as described by Tajfel and Turner (1979). Importantly, for both theoretical purposes and for each of the empirical cases described in this book, the axis of political competition is not a constant, but can (and often does) fluctuate over time even within a given country. These fluctuations will be explored in the empirical chapters. Two features define the axis of political competition for the purposes of this book: which groups are in competition and the issues over which they are competing. First, it is important to consider which groups (in this case, sects) consider themselves to be in competition with one another. Sectarian political competition often tends to take on a zero-sum quality; if one group benefits from the changes brought by democracy, they presumably should come at the cost of privileges previously held by other groups.16 My theory is agnostic regarding when and how sectarian groups find themselves to be in competition with each other—and when this competition takes on heightened significance—but such competition is an important feature of the sectarian environment. Multiple groups could be democratic “winners” or “losers” depending on their size and the existing political-economic power configuration. It is difficult to determine a priori which groups will be in competition and how each group will view democracy without some understanding of the context, but identifying these conflicts and interests will help to predict when communal religion will enhance or depress support for democracy. Herberg (1983) has observed that the broadness or specificity of religious categories changes over time in response to evolving social and political circumstances. While it is beyond the scope of this book to explain the origins of religious categories, it is nevertheless important to recognize that these categories are shifting rather than fixed. In some environments, categories like “Muslim” or “Protestant” may be sufficiently specific; in other settings (or even the same place at another point in time) they may be so broad as to be nearly useless. Subsequent chapters will demonstrate that in Lebanon, the category of “Muslim” shifted from being a relatively sensible

 faith in numbers one for some purposes to eventually collapsing groups that were directly in conflict with one another into a single category. Relevant group boundaries thus require careful consideration rather than being presumed; the axis of political competition depends critically on who the key actors are and how they view themselves relative to other groups. Importantly, I do not assume that sects will always speak with a united voice. Religious leaders—even those who are nominally at the top of the religious hierarchy—may not always command the support or respect of their communities; Henley (2017) finds this to be especially true in the Lebanese case. When sectarian unity is largely absent, the relationship between communal prayer and political preferences will be much cloudier. Such circumstances are especially likely in the presence of competition for religious-political leadership. If different leaders (or institutions) within the same sect are unwilling to cooperate or express a shared vision, then sectarian “interests” are not clearly defined, and the effect of communal prayer will be ambiguous. Chapter 5 will illustrate this pattern with respect to Christians in Lebanon, who have become highly divided in the past several years with respect to key political issues. Rather than presenting a united front in favor of broader “Christian” interests, competing factions within the Christian community have allied themselves with either Sunni- or Shi‘a-led coalitions. Consequently, religious attendance does not have a clear effect on Christian political preferences in the way that it did prior to the emergence of these divisions. The second feature of the axis of political competition has to do with the issues under contention. In principle, any politically relevant issue can define the axis of competition, and could carry sectarian undertones. This book focuses on the issue of regime type, a highly salient issue in many countries that have not established or consolidated democracy. Since democracy creates winners and losers, it is in many settings subject to considerable contestation. Like group divisions, however, the issues motivating intergroup competition are not constant. As this book will explore, democracy involves multiple significant dimensions and thus implies different things to different people. In the cases explored in this book, many of the implications of democracy are fundamentally institutional: in Lebanon, democracy often connotes de-sectarianization, proportional representation, and an adjustment to (or removal of) existing sectarian quotas. In other cases, democracy might be perceived to be a mixture of institutional features and policy outcomes, and the relative weights placed on each of these characteristics affect

a theory of sectarian interests and democracy 25 individuals’ orientations toward the often abstract concept of democracy. In Iraq, for instance, the effects of communal religious practice on support for democracy depend not only on the sect to which an individual belongs, but also on what the individual perceives to be the most important feature(s) of democracy. The relevant characteristics of democracy may also change over time, with issues like electoral representation, human rights, or economic redistribution becoming more or less central concerns. It is not possible a priori to stipulate what features of democracy (or any other issue of political competition) are most relevant; rather, as the remainder of this book argues, context matters. In subsequent chapters, I will examine the importance of the axis of political competition in Lebanon and Iraq in a few ways. In Lebanon, I will argue, changing political circumstances shifted the relationship between communal prayer and regime attitudes because of a change in the configuration of competition among the groups. In Iraq, I will suggest that even at the same point in time, communal prayer’s effect on support for democracy depends on what the individual perceives democracy to mean. Both of these considerations are a part of the broader concept of the axis of political competition.

Mechanisms A number of mechanisms for the relationships I have described are possible. Many of these potential mechanisms have been discussed in existing social scientific accounts of religion. In his landmark sociological study of religion, Durkheim (2001, p. 171) argues that worship is not fundamentally about deities, but about the religious community; for Durkheim, deities are “merely the symbolic expression of society.”17 He further suggests that “by seeming to strengthen the ties that bind the worshiper and his god, [acts of worship] really strengthen the ties that bind the individual to his society.” Similarly, Beattie (1966) argues that religious rituals are frequently vehicles for instilling group values and clarifying the social order (similar claims are made by Djupe and Gilbert, 2008, among others). I will argue that the link between communal religious practice and sect-centric political attitudes is forged by sectarian solidarity. The more frequently an individual participates in communal worship, the closer she will feel to other members of her sect, and her political preferences will adjust accordingly.

 faith in numbers

Religion, Group Affect, and Solidarity Group solidarity plays an important role in translating religious practice into regime attitudes, and religious group identity may be particularly powerful. Communal religious practice would seem to be an especially potent force for building group solidarity: through structured, regular interaction with members of the same religious community, communal practice can build trust and norms of reciprocity that might, under the right circumstances, aid in the functioning of democracy. However, it is important to distinguish between religious belief (or private practices) and communal religious behaviors; as Putnam (2000, p. 74) notes, “Privatized religion may be morally compelling and psychically fulfilling, but it embodies less social capital.” How does this social-capital-producing process work? Face-to-face interaction with fellow congregants is an important factor, particularly since the congregation is the setting in which the overwhelming majority of religious interactions take place. These interactions enhance feelings of group solidarity and support. In a study on social support for the elderly, Krause (2002, p. S341) presents evidence suggesting that “greater congregational cohesiveness is associated with receiving more spiritual support and more emotional support from one’s fellow parishioners” and that congregants who attend services more frequently are more likely to view their congregations as highly cohesive. Psychological studies of mental health among Lebanese citizens have suggested that organizational religion influences mental health, while other aspects of religion do not, and that it does so through a mechanism of “social solidarity” (Chaaya et al., 2007). These findings indicate a significant social cohesion mechanism—the more citizens participate in communal worship, the more embedded they feel in their congregations. The solidarity mechanism demonstrated by communal prayer can work through at least two different channels: messages from religious leaders and exposure to co-sectarians. This dichotomy is closely related to the question of whether the political influence of religion comes “from the pulpit or the pews.”18 In many cases, sermons delivered by political leaders explicitly address political issues; these patterns will be discussed further in subsequent chapters. The appeals present in sermons are likely to be influenced considerably by the conditions of the religious community writ large; it is well established in other contexts that religious leaders are motivated by concern for the needs of their religious communities (Djupe and Gilbert, 2003). This motivation is particularly strong in cases where

a theory of sectarian interests and democracy 27 the religious group is underrepresented (Djupe and Gilbert, 2003; Djupe et al., 2016). Even in the absence of explicitly political content, however, sermons may have an indirect political influence. To the extent that sermons promote sectarian identity—which they frequently do in abundance in the cases studied in this book19 —worshipers’ political attitudes may be affected by sermons even without direct political appeals. Beyond formal sermons or exhortations from religious leaders, however, participation in communal worship increases feelings of group identity (Harris, 1994) and therefore promotes a sect-centric approach to political issues (cf. Wald et al., 1988). The very act of communal prayer may make individuals feel more closely united with their co-sectarians (Jamal, 2005). This theory builds on the insights of earlier works demonstrating a link between religious interactions and group-centric views (Wald et al., 1988; Jelen, 1993; Djupe and Gilbert, 2003). This effect is magnified by the fact that the composition of the worship community is by definition religiously homogeneous. Further, communal prayer provides an opportunity for members to interact before, during, or after worship rituals. This claim does not assume that worshipers meeting for coffee after religious ceremonies will always discuss political or sectarian matters; however, on a larger scale, such gatherings will frequently create fertile ground for political discussion and increased sectarian identity. Evolutionary biologists have suggested that “cultural evolutionary processes, driven by competition among groups, have exploited aspects of our evolved psychology, including certain cognitive byproducts, to gradually assemble packages of supernatural beliefs, devotions, and rituals that were increasingly effective at instilling deep commitment, galvanizing internal solidarity, and sustaining larger-scale cooperation” (Atran and Henrich, 2010, p. 5). Religious rituals play a particularly important role in facilitating cooperation. Experimental research has shown that “acting in synchrony with others [moving rhythmically together] can foster cooperation within groups by strengthening group cohesion” (Wiltermuth and Heath, 2009, p. 1). This type of behavior is abundant in religious services, complementing the more obvious solidarity-building exercises like praying, listening to sermons, bowing toward Mecca, and others. Since this type of behavior tends to promote solidarity only within the group, as discussed in more detail later, these arguments explain why a number of studies find that communal worship tends to promote out-group hostility as well.20 Participation in congregational worship increases feelings of cohesion with both the congregation and the broader sect. Importantly, the effect of

 faith in numbers communal worship on group identification does not depend on worshipers receiving explicit political or sectarian cues, though I will demonstrate in later chapters that such cues are common. Direct sectarian messages affect worshipers’ attitudes in predictable ways, but even much subtler cues can have a similar influence.21 Importantly, these social solidarity processes are not confined to the congregation alone. Communal prayer can also influence sectarian solidarity in noticeable ways. Participation in communal rituals whose history extends beyond the congregation may remind worshipers of their coreligionists in other areas; religious rituals can be particularly strong cultivators of “imagined” religious communities (Anderson, 1983). This process requires a certain degree of commonality between congregations: for single-congregation denominations (sects), there is no reason to expect communal rituals to enhance ties between members of the congregation and members of any other congregation. For the same reasons, the “imagined communities” effect will tend to be stronger within sects for whom rituals are perceived to be more similar across individual congregations. Wuthnow (2009, pp. 80–81) states that “Homogenization. . . such as dress and styles of worship or beliefs about the Holy Spirit, [is] also likely to increase the sense among widely scattered congregations that they share a common destiny and bear responsibility for one another.” The common rituals practiced by disparate congregations contribute to a mutually reinforcing relationship between collective memory and group identity (Brown et al., 2012). The sectarian groups discussed in this book all demonstrate a considerable amount of between-congregation linkage; the rituals, ideas, and leaders of these sects extend beyond single congregations to wider sectarian groups. It is important to note the particular form of group identification engendered by communal religious engagement. While it is possible that communal prayer might strengthen nonreligious identities, there are both theoretical and empirical reasons to suspect that these behaviors will have the most powerful effect on sectarian identity. The theoretical reasons for this expectation are clear. First, the content of communal prayer is religious in nature, and is therefore associated with sect, at least to some extent. Furthermore, and perhaps more importantly, communal worship involves engaging in a communal ritual with only members of one’s own sect. The sectarian homogeneity of worship environments ensures that while ethnic, national, or other identities may be strengthened by communal prayer, the primary identity effects will be found in the sectarian realm.22

a theory of sectarian interests and democracy 29 Experimental research has found substantial effects of religious primes on individual behavior related to group affect and social solidarity. One line of research23 suggests that religion tends to promote prosocial behavior, including generosity, honesty, and other socially desirable outcomes. However, another set of studies links religious primes to negative outcomes such as racism (Johnson et al., 2010), anxiety (Toburen and Meier, 2010), and submissiveness (Saroglou et al., 2009). The diversity of results coming from these studies may help to explain why religion appears to have an ambiguous effect on social capital and, by extension, support for democracy. More recent experiments have suggested that the prosocial effect of religion may be conditional: Johnson et al. (2012) find that religious priming promotes both in-group favoritism and out-group derogation, suggesting that these primes do not simply lead to greater prosocial behavior across the board. Blogowska and Saroglou (2011) find that religion promotes prosocial behavior only toward in-group members. These findings suggest an important consideration: religion appears to induce in-group favoritism, and should therefore lead to political attitudes more favorable toward in-groups than out-groups. These patterns are consistent with the claim that religion promotes only (or at least predominantly) within-group (or “bonding”) social capital (Putnam, 2000; Uslaner, 2002). Simply put, communal prayer makes citizens feel closer to members of their sect and care more about the well-being of those members. This change in group affect brings corresponding changes in political attitudes: since attenders care more than non-attenders about their co-sectarians’ wellbeing, they will tend to hold political preferences closer to those that would benefit their sect as a whole. Borrowing the terminology of rational choice theory, a person’s “utility” may be derived from both selfish and groupcentered interests (see Margolis, 1984). Consequently, wealthy individuals may prefer more redistributive policies, and poor individuals may prefer less. In this story, communal prayer does not affect the individual’s “selfish” interests; rather, it affects the relative weights placed on selfish and groupcentric concerns. To the extent that group-level concerns matter to the individual, her preferences are not “rational” in the narrow self-interested sense. Acemoglu and Robinson’s (2006) famous models of redistributive and regime preferences ignore group-level considerations that are not purely economic in nature. Since the groups they consider (poor, middle class, wealthy, etc.) are defined by their income, it is impossible for their individual interests

 faith in numbers to clash with their group interests. In reality, however, individuals also possess group identities, and these identities matter in determining their political preferences. In highly sectarian societies, religious affiliation is likely to matter quite a bit more than a vague concept of “class.” Since sects are not economically homogeneous, some group members will have personal financial incentives in conflict with those of the average group member. Existing models of redistributive preferences generally presume that individuals are exclusively interested in maximizing their own income, leaving aside any group-level concerns that might otherwise be important to them. The omission of identity factors (and, in particular, religion) in the classic models of redistributive democracy is a potential oversight in this line of research.24 Studies that do consider religion as a factor in redistributive preferences typically maintain the assumption that these preferences are ultimately reducible to the individual’s financial interests. Scheve and Stasavage (2006, p. 256) argue that religious individuals should be less supportive of welfare state policies because “religious involvement and social spending can both serve to insure individuals against the effects of adverse life events.” In their account, the religious community provides individuals with a substitute for welfare: in the event of an adverse life event, the individual’s coreligionists will step in to support her. In this case, the individual’s welfare preference is driven only by her financial well-being. What about her coreligionists? If indeed these coreligionists are willing to help the member of their community who was affected by an adverse life event, then their behavior must be driven by something other than selfinterest. In fact, the logic of this argument assumes that religion makes people care about their coreligionists, an assumption that implies that in a religious setting, preferences and behaviors are not motivated by self-interest alone. This account helps to explain the curious finding that, in contrast to the predictions of many of the canonical theoretical models of redistributive preferences, identity considerations in particular influence attitudes toward redistribution even in ways that may clash with the individual’s selfinterest.25 The failure to address identity concerns in theorizing political preferences helps to explain26 the empirical observation that redistributive preferences often do not derive only from income-based self-interest.27 For reasons described earlier, religion represents a particularly important identity, and the strength of this identity is partly a function of communal worship. For any given level of wealth or influence, communal prayer will

a theory of sectarian interests and democracy 31 tend to push citizens’ preferences more toward those indicated by their sect’s interests. Sometimes, as in the case of a wealthy member of a wealthy group, this shift in the relative weight of sectarian concerns will have little or no effect, since both types of concerns suggest a preference against redistribution. In other cases, however, an increase in the relative importance of group concerns can lead to a shift in preferences away from personal, selfregarding interests, and in the direction of the sect’s interests. In both cases, the theory implies that attenders should be more likely than non-attenders to hold political preferences consistent with their sect’s interests.

Direction of Causation Any test of this theory encounters a major obstacle to inference: religious attendance is not randomly assigned and may either be caused by high religious identification or jointly determined with religious identification by a third factor. In other words, it is entirely possible that people with stronger sectarian identities are more likely to engage in communal religious practice because of their sectarian identification, not the other way around. If this is the case, then the theory presented in this book wrongly reverses the causal arrow, and communal prayer is a consequence rather than a cause of sectarian identities and political preferences. The empirical chapters of this book will devote significant attention to addressing this very real concern. In subsequent chapters (especially in chapter 5), I will make use of innovative experimental and quasiexperimental techniques to reduce the dangers presented by the likely endogeneity of religious practice. I will argue that despite the potential for a two-directional relationship between religious practice and religious identity, the evidence suggests a powerful, independent effect of communal worship on the salience of sectarian identity and, by extension, political preferences.

Religious Identities In some ways, religion is distinct from other forms of identity.28 The various ways in which religion is different from ethnicity or other identity markers cannot all be accounted for in this chapter, but several illustrations demonstrate the importance of this distinction. For the most part, an

 faith in numbers individual’s religious identity is exclusive: a person can be a member of multiple ethnic groups due to mixed ancestry, but cannot be a member of multiple faiths. This exclusivity may make religion a less porous boundary, and therefore a potentially more intense identity (Lockwood, 1981). It is also the case that many individuals possess a sectarian identity without actually holding religious beliefs. In these cases, “religious” identity resembles ethnic identity much more closely. However, for reasons described in greater detail in later pages, there is reason to believe that individuals who do not participate in communal rites will tend to identify less strongly with their sectarian affiliation, even though they are nominally members of the group. An important distinction here is between identification and identity (see Hardin, 1997). Simply being a member of a given group (identity) is not sufficient to hold preferences in close alignment with the interests of that group. Rather, I have argued that identification (or, in my parlance, salience of group identity) is a major factor in determining individuals’ regime and redistributive preferences. The act of communal prayer—a feature unique to religion rather than other forms of identity—has a powerful effect on identification, and in turn plays a large part in directing citizens’ political preferences. While religion contains many elements that are similar to those found in ethnicity, class, gender, and other forms of identity, it also contains the uniquely powerful feature of communal prayer. The act of communal prayer heightens the importance of sectarian interests, promotes group solidarity, and increases an individual’s sense of “linked fate” with her religious group. Since weekly communal prayer is religiously mandated for at least half of the world’s citizens (Pew Forum, 2012), it is safe to say that communal religious practice is the most prevalent form of associational gathering in the world. The fact that these gatherings are by definition restricted to members of the same sect suggests that they will be especially potent galvanizers of group identity. While many other associations exist—some of which undoubtedly influence political attitudes—these associations are unlikely to possess the same influence afforded to gatherings that are regular, communal, and spiritual in nature. In many instances, religious identities have been found to be more salient than ethnic identities, though this is certainly not always the case. Jacobson (1997) argues that among British residents of Pakistani origin,29 religion is a more significant part of individuals’ identities than ethnicity, primarily because the former is seen as “the basis of an all-encompassing frame of reference which explains or establishes their place in the world,” a conception

a theory of sectarian interests and democracy 33 similar to the one provided by Horowitz (1985, p. 50), who notes that for many people, religion is “an inextricable component of their sense of peoplehood.” This distinction is related to the pattern described earlier, in which ethnic customs are perceived to be isolated to the group, while religious teachings are believed to be “universally applicable” (Jacobson, 1997, p. 240). Religion’s emphasis on value-enforcement marks one of its key distinctions from ethnicity. Religion has historically been “a particularly effective mechanism for transmitting whole suites of cultural norms” (Smaldino, 2014, p. 251), and religious practices facilitate cooperation and maintain group solidarity in stronger ways than other forms of identity (Sosis and Kiper, 2014).30 The salience of religious identity also endows religious groups with what Grzymala-Busse (2015) calls “moral authority”— a voice that allows them to weigh in on explicitly political issues beyond theological and moral debates. Religious groups may therefore be permitted more influence on important political issues than other types of social groups, and their authority may be more respected.31

Conclusion Religion remains a deeply misunderstood force in comparative politics, and in public discourse more broadly. For decades, secularization theorists believed—or perhaps hoped—that religion would largely disappear as a source of social and political influence. In reality, the exact opposite trend has emerged. Religion, in its various forms, remains as influential as ever, and in many cases, political circumstances have actually heightened the salience of religion as a set of identities and behaviors. The influence of religion—whatever its form—on political attitudes and behaviors is an empirical question rather than a normative one. While many scholars and casual observers have little interest in religion as a social phenomenon, I will show that, when considered carefully, religious factors can account for considerable variation in political attitudes that would be of interest to anyone who wishes to better understand the political world. This chapter has presented a theory of religion, group interest, and democracy that focuses on communal religion’s influence on regime preferences. While the existing research on religion and democracy tends to focus on individual behaviors or sectarian factors in isolation, my theory integrates these two features of religion and places them in the context of intra-state

 faith in numbers political competition. In doing so, it suggests that the various witnessed effects of religion on democratic attitudes can be accounted for in no small part by sectarian interests and political incentives. The effect of communal prayer is neither universally pro- or anti-democratic. The direction and magnitude of this relationship depend on state- and sect-level factors that determine the interests of group members. A number of factors condition the effect of communal prayer on regime and redistributive attitudes. First, merely being a member of a particular sect does not necessarily lead to political preferences in alignment with the interests of that group. Regular participation in communal worship has a substantial effect on the salience of group identity, and therefore makes believers much more likely to hold political positions consistent with their sectarian interests. Further, the arrangement of political and economic power in each country influences this relationship. For groups that enjoy special privileges under authoritarian regimes, communal prayer will tend to promote anti-democratic attitudes. This relationship is not due to an anti-democratic theology (indeed, it could apply to any privileged sect regardless of its theology), but rather due to group interests. Likewise, a group that would benefit from further democratization will demonstrate a pro-democratic effect of communal prayer. Finally, the axis of political competition matters: depending on who is competing with whom and what issues or resources they are competing over, the direction of these effects may switch. The most important implication of this theory is that communal prayer does not always have the same effect on democratic attitudes and behaviors. A fuller understanding of this relationship requires consideration of the interests of religious sects and the political arrangements in their countries. The reasoning applied in this chapter is intended to be portable, but not universal. While not all countries are characterized by salient religious divisions, a great number of them are, and these divisions are often particularly challenging for democratic development. In cases where religious affiliation constitutes a meaningful social identity, religious identities and interests will more than likely bear on political preferences in some way. The more closely interests map onto identities, the larger this influence will tend to be. The insights from this chapter can be used to explain why members of different religious groups approach democracy differently—regardless of theological differences—based on straightforward assessments of sectarian interests.

3 The Religious Experience Communal Prayer and Group Identity The theory developed in this book suggests that the practice of communal worship creates a sense of sectarian solidarity and, in doing so, promotes group-centric preferences toward political issues like democracy and redistribution. But what kinds of experiences do people have in churches and mosques in religiously divided societies? To provide a richer picture of the communal religious experience—and its effects on sectarian identities and political preferences—this chapter presents evidence from a series of large-scale interviews with Muslim and Christian citizens in both Lebanon and Iraq. These interviews gave respondents a chance to describe in some depth how the experience of praying at a mosque or a church affects their psychological state, their feelings toward their community, and their identity. The evidence provided in this chapter derives primarily from two sets of survey interviews conducted in Iraq and Lebanon. In each of these surveys, respondents were asked a series of open-ended questions about the religious experience. Naturally, a large number of respondents indicated that they did not engage in communal prayer, and thus could not provide much useful information about the experiences therein. However, among those who do participate in communal worship, these answers provide a uniquely rich look into the construction of sectarian identity in religious settings. The collection of this type of qualitative data has several advantages. First, it allows individuals to speak for themselves. It is often tempting for researchers, journalists, or casual observers to impose their own visions of these types of experiences on their subjects, and such impositions are problematic for obvious reasons. Allowing respondents the flexibility to describe their experiences without forcing their answers into discrete categories—or worse, describing their experiences for them—avoids these potential issues. Second, these types of interviews provide on-the-ground evidence of what the religious experience means to people, and what they view as its important features. By asking broadly open-ended questions, the researcher allows the Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0003

 faith in numbers interviewees to prioritize the key issues and provide as much detail as they see fit. Third, there are important practical advantages to using this approach. A large-scale qualitative survey aims to achieve some of the richness of open-ended interviews alongside the benefits of generalization present in large-N research, and goes a long way toward ensuring that the responses are largely representative of the population because the sample itself is representative. Practical considerations prevented me from conducting these interviews personally. In addition to security concerns in Iraq and parts of Lebanon, my position as a Western researcher would inevitably bias interviewees’ responses to sensitive questions about religion, sectarianism, and democracy. To remedy these issues, I chose alternative methods of data collection. In Iraq, responses were self-reported online, potentially yielding somewhat more superficial responses but avoiding the danger of interviewer effects and minimizing concerns about censorship, since internet censorship is relatively weak. In Lebanon, I employed professional survey enumerators to record citizens’ responses to these questions. Using local, trained enumerators is likely to improve the reliability of the responses offered by interviewees (see Cammett, 2013b, for a helpful explanation of some of the advantages of using this type of approach in places like Lebanon). These responses, embedded in a large-scale survey, admittedly do not possess the same depth that longer interviews would afford. However, allowing large and broadly representative samples of Lebanese and Iraqi citizens to use their own words to describe their religious experiences nevertheless yields helpful insights into the meaning of these experiences and how they translate into political and social attitudes. The interviews that provide the bulk of the information presented in this chapter were conducted in Lebanon between July and August 2018 and in Iraq in August 2018. In Lebanon, the interviews were part of a nationally representative in-person survey conducted in all of Lebanon’s 26 districts (qadaas), consisting of 1,500 total interviews split among the three major religious communities: Christians (of multiple denominations in proportion to their share of the Christian population), Sunnis, and Shi‘a. In Iraq, 500 Sunni, Shi‘a, and Christian respondents provided their responses online.1 Both sets of interviews were conducted in Arabic. In both cases, interviewees were asked to reflect on their experiences in churches/mosques (if applicable), how those experiences made them feel, and any political and/or sectarian messages they encountered in those settings.2

the religious experience 37

Why Attend? Two key questions for the analysis is this book have to do with who selects into religious attendance, and why they choose to do so. Individuals may have a number of different reasons for attending religious services. Many, if not most, of these reasons have little to do with politics or sectarianism. A person may participate in communal worship due to piety, habit, family pressure, guilt, or countless other personal considerations that are not politically motivated. Nevertheless, political cues, explicit or implicit, are abundant in the religious setting. While most worshipers are likely to attend for nonpolitical reasons, they are quite likely to be exposed to political content in the process. In Lebanon and Iraq, interviews revealed relatively similar motives for attending religious services; this was largely true regardless of sectarian affiliation. Many interviewees described communal prayer as a “duty” or a religious obligation, an opinion that is not universal, particularly in Islam. Others indicated that they hoped to receive blessings from God as a result of their attendance; they often characterized communal prayer as a “sacrifice” that would earn divine protection and lead God to watch over them during the following week. A number of Muslim interviewees referred to the common belief that praying at a mosque earns the worshiper a reward 27 times greater than praying at home3 (the number is sometimes quoted as 25). Needless to say, there was a widespread perception that communal prayer would yield substantial spiritual benefits. A high proportion of respondents in both Iraq and Lebanon—again, regardless of sect—indicated that they simply felt comfortable in communal worship settings. Many reported feeling close to God in a way that they could not experience in private prayer. They likewise indicated a sense of release from the world’s concerns: in a sentiment echoed by many others, a Lebanese respondent from Akkar described the experience of communal prayer as “feeling comfortable when you feel tired and worried.” Similarly, an Iraqi interviewee stated: “Psychological comfort and spiritual feeling are at the highest levels while praying in mosques. The place is more calm and makes you feel more secure. It is an opportunity for spiritual communication and distance from materialism. Religious feeling overwhelms when many people perform the same religious rituals. There is a sense of comfort and certainty in God.”

 faith in numbers Non-attenders provided a variety of explanations for their decision not to engage in communal worship. Most of these reasons were either spiritual or practical. Some non-attenders do not believe in God, while others do not consider themselves especially religious or do not pray at all. Some simply prefer praying alone, while others expressed frustration that people in their communities only attended church or mosque in order to “show off.” Others cited difficulties with leaving work, distance to the mosque/church, and weather challenges as obstacles to their attendance. Some expressed regret or guilt regarding their nonattendance and suggested the possibility that they might attend more regularly in the future. Among those who reported infrequent or nonattendance, descriptions of the mosque experience were often far less positive than those presented by attenders. Some Lebanese and, especially, Iraqis, felt that mosques had been compromised by sectarianism or political agendas: I usually pray at home and I rarely pray in the mosque. It is correct that prayer in the mosque is much better than praying at home. But since 2003, I have avoided praying in mosques when it became political and sectarian. I feel a lack of faith in prayer in the mosques. There are some sheikhs who incite sectarianism and accuse others of infidelity. —Iraqi, Baghdad Currently, praying in mosques is used for political purposes by authorities. In my personal opinion, it tears apart religious communities and weakens the religion. —Iraqi, Basra Groups who manage the mosque make you hate coming to it. —Lebanese, Tripoli

These responses—while not representative of the broader population— underscore a challenge present in most studies of the effects of religious practice: it is possible that people choose whether or not to attend religious services for political reasons. If this is the case, then the relationships observed between religious practices and political outcomes may be spurious. In the remainder of this book, I will use a variety of strategies to

the religious experience 39 help to demonstrate that the relationship between communal prayer and democratic attitudes is real and not driven solely be self-selection. Responses in this survey reflected the fact that, for the most part, participants choose whether or not to attend mosque or church for nonpolitical reasons. In general, interviewees were much more likely to emphasize the spiritual and emotional dimensions of communal prayer. Their responses often revealed a sense of duty (farida or wajib) or special spiritual benefits received by those who go to mosque. Others stated that they felt closer to God when praying at a house of worship compared to praying at home. It was very common for respondents to say that communal prayer was simply “better” than praying at home or, as one Iraqi respondent put it, “the best thing that a Muslim can do.” Many interviewees even provided specific numbers (often larger than the “multiplicative benefit” described earlier), stating that communal prayer is 100 or 1,000 times better than praying at home.

The Communal Experience The primary purpose of this chapter is to provide a sense of the experience of communal worship in these settings. In both Lebanon and Iraq, interviewees’ descriptions of their experiences at church or mosque centered on a handful of dominant themes.

Relief, Comfort, and Strength One of the most common terms brought up by respondents was relief : worshipers feel that the experience of communal worship lightens the load of everyday life and provides a sense of comfort that is unavailable elsewhere. A Sunni from northern Lebanon explained: “I feel that I am in the hands of God alongside those who are praying with me, and I forget my worries.” For many respondents, this sense of comfort transformed into a sense of safety; as a Lebanese interviewee from Baalbek stated, communal worship is “a great experience that gives serenity and safety.” For some respondents, particularly in Iraq, this sense of safety was more nostalgia than reflection of current reality:

 faith in numbers Praying in mosques creates a kind of humbleness, comfort, and a feeling of safety if both the atmosphere and the imam are suitable. Currently, you do not feel safety at the mosque because Shiite and Sunni mosques in Iraq are subject to the threat of bombings by extremist groups who do not observe the sanctity of the mosque. —Iraqi, Baghdad

Nevertheless, safety emerged as a common theme, even in areas particularly troubled by sectarian strife. Indeed, many respondents stated that attending mosque or church gave them a feeling of strength. An interviewee from Beirut summarized this sentiment: “Communal prayer makes you feel strong and proud. It brings believers closer together.” A respondent from Mount Lebanon reported a “sense of strength and sectarian cohesion,” and another from the Bekaa Valley stated, “It provides me with strength to endure the difficulties of life, and makes me feel a sense of solidarity.” In explaining this feeling of strength, some respondents referenced scriptural invocations to communal worship. For instance, a Greek Orthodox interviewee from the Bekaa Valley quoted the Gospel of Matthew, saying, “Community prayer gives us power. ‘If two persons come together in my name, I will be among them.”’ Other Iraqis reported similar experiences of strength in mosques: Praying in mosques is one of the most common things that makes people feel comfortable and reassured. When you pray with others, you feel that you are in a strong society and that you have a desire to help and do good deeds. It reminds you of your true role in this life. —Iraqi, Baghdad We feel that we are strong, as we have larger numbers in mosques. —Iraqi, Nineveh

Sectarian Unity and Solidarity While the topics of comfort and strength were common in interviewees’ descriptions of their religious experience, by far the most common themes to emerge in these interviews were sectarian unity and solidarity. Respondents overwhelmingly reported that communal worship made them feel closer

the religious experience 41 to their sect in a variety of ways. These reports illustrate the mechanisms proposed in this book: respondents who engage in communal prayer state that these experiences bring them closer to the members of their sect and induce a type of sectarian solidarity that likewise affects political preferences. Even without explicit political cues, respondents in both Lebanon and Iraq frequently indicated that mosque or church attendance heightened a sense of what scholars have called “linked fate” (Dawson, 1995) along with several other dimensions of sectarian solidarity. Respondents expressed these feelings in a number of different ways. A citizen from northern Lebanon simply stated that when she goes to church, she “feel[s] closeness and love toward members of [her] sect.” In general, interviewees described feeling “close,” “united,” or “loyal” to members of their sect in a general sense. Others use terms like “love” and “solidarity.” Still others describe the sectarian community—strengthened by communal worship—as a family. Many answers also expressed how communal worship softens the divides between members of the sect. A Sunni from Beirut explained that “rich and poor people pray together, which generates love between each other.” An Iraqi reported that “the sense of affinity and participation with others raises feelings of pride in religion. You see people of different groups and races performing the same rituals.” Shared worship also brings the concerns of co-sectarians to the center of worshipers’ minds: Prayer brings many people together in the same place. They meet and exchange greetings before praying and learn each other’s news and problems. —Lebanese, Mount Lebanon It [communal worship] also increases the interdependence of individuals and makes people more tolerant. It is also a good opportunity to meet and discuss matters of faith or plan to do humanitarian work and help the needy people. —Iraqi, Baghdad

Importantly, communal prayer often brings people together who might not meet each other under normal circumstances. Many respondents pointed to the importance of praying alongside co-sectarians whom they did not know. They often stated that forging these new connections heightened

 faith in numbers their sense of sectarian solidarity, as it extended their religious ties beyond their own family or neighborhood: Praying in mosques makes me feel more humble than praying alone at home. In the mosque, I find a great atmosphere. . .I usually perform these rituals with people who are unknown to me. —Iraqi, Erbil When I go to the mosque, I feel more comfortable than praying at home. I feel close to worshipers inside the mosque, even though I do not know most of them. —Iraqi, Nineveh It [communal prayer] is a reason for people to get to know each other and to sow love and intimacy with each other. —Lebanese, Beirut

For some people, the sectarian social capital created by communal prayer was presented as a source of sectarian hostility. A Sunni in northern Lebanon referred to “messages that provoke [sectarian] instincts and stir up hatred,” while a Maronite from Akkar plainly stated that church attendance “support[s] sectarianism.” In Iraq, a number of interviewees reported similar concerns, often more forcefully: [In mosques,] there is sectarianism and racism towards other Muslims or other religions. —Iraqi, Nineveh All mosques are sectarian. —Iraqi, Basra Sectarian messages inside the mosque are beginning to dismantle society and plant the seeds of discord and hypocrisy within it. —Iraqi, Maysan

the religious experience 43 In a number of cases, asking respondents about the mosque experience led to vivid expressions of religious chauvinism or accusations against other sects. These responses were particularly common in Iraq, regardless of the (presumed)4 sectarian affiliation of the respondent. These responses took a variety of forms: Shi‘a and infidels usually worship the family of the prophet Muhammad,5 peace and blessings be upon them. They hate Aisha6 and the companions of the Prophet. —Iraqi, Baghdad Concerning the Shi‘a community, there are messages of hatred and anger toward the Sunnis in Iraq. They issued a lot of messages of hatred and aggression against Sunnis. They may incite the killing of Sunnis. All this is issued within Shi‘a mosques. —Iraqi, Baghdad In Sunni mosques, there is hatred and incitement to kill apostates. Sunni Islam has produced terrorist organizations such as the Muslim Brotherhood, al-Qaeda, and ISIS. —Iraqi, Baghdad

Political Messages I have argued that the “sectarianizing” effect of religious attendance is not always as subtle as the generation of solidarity among co-sectarians. Indeed, in religiously divided settings, explicit invocation of sectarian political identities is common. In both Lebanon and Iraq, numerous interviewees reported experiencing clear political messaging from religious leaders. For example, a citizen from northern Lebanon described hearing that “you should not vote for a candidate from a different sect;”7 others reported similar injunctions, including a resident of Akkar who was told that “we should vote for the candidates belonging to our sect.” Reactions to these messages varied: It [the church] was a place for praying. It has become a political office. —Lebanese, Tripoli

 faith in numbers Mosques cannot be free from politics because it is impossible that Friday sermons could be free of politics. This is governed by the ruling Islamic parties in Iraq. It is impossible to hear the Friday sermon without talking about politics. Politics are impossible without sectarianism. —Iraqi, Nineveh

In some cases, respondents provided accounts of being told to vote for particular parties or candidates. These types of cues were less common than general calls for sectarian unity or a broad political agenda, but were nevertheless noticeable in a number of interviews. In our mosque, political discourse prevails because our imam is the preacher of a particular party that promotes politics within the mosque. —Iraqi, Duhok As for the political aspects, you find this every Friday, even in election season. Some mosques have used the mosque’s loudspeakers and amplifiers to urge people to elect a particular party. —Iraqi, Baghdad

Sectarian Themes in Lebanon The building of sectarian solidarity in religious settings transcends sect in these cases, as respondents from every major religious tradition reported feeling closer to their co-sectarians as a result of their participation in communal worship—and often described explicit calls for sectarian unity in these environments. However, the tone and content of sectarian and political messages varied between the different religious groups in important ways. This section explores some of the topics (and angles) taken up by religious leaders and communities across Lebanon’s8 three major sectarian groups. Consistent with the arguments presented about the Lebanese case in this book, distinct themes emerged in each sect’s discussions of religion and politics.

the religious experience 45

Christians: Unity and Preservation As I will describe in greater detail in chapter 4, Christians (and particularly the Maronite sect) have historically enjoyed considerable privileges in the Lebanese political and economic system. Consequently, it should perhaps come as no surprise that much of the religious-political discourse addressed in the interviews with Christian respondents centered around the preservation of Christian privilege. Specifically, interviewees regularly lamented the decrease in Christian power over the past few decades and frequently reported hearing messages in churches that called for the preservation of Christian political advantages. Some interviewees obliquely referred to the protection of Maronites in an increasingly hostile environment: The situation of Christians in the East, including Lebanon, has become unsatisfactory. We have to make efforts to get out of this situation. —Lebanese Maronite, Mount Lebanon

Other Christians were more explicit about the need to protect their political interests: We, the Maronites, are the ruling authority, and we should preserve this. —Lebanese Maronite, Mount Lebanon We should preserve the Maronite sect in Lebanon and call for coexistence. —Lebanese Maronite, Mount Lebanon

An important part of this discourse focused on the need for unity within the Christian community. The era following the outbreak of the Syrian Civil War has been marked by considerable divisions within the Christian community, with both politicians and ordinary Christian citizens split more or less evenly between the pro-Syrian regime (mostly Shi‘a) and anti-Syrian regime (mostly Sunni) camps. A number of Christian politicians (and even many religious leaders) have pointed to this division as a threat to Christian political power. They argue that the divide among Christians on the Syria issue has weakened the Christian political position in Lebanon, and have urged followers to protect Christian unity. Illustratively, a Maronite in South

 faith in numbers Lebanon stated that he heard messages in church regarding “the need for consensus among Christian politicians for the benefit of society.” In Mount Lebanon, another Maronite described a religious message asserting, “We need to preserve Christian unity.” These calls also included strategic political messages. Several Maronites (and other Christians) reported being warned against emigrating from Lebanon. The motivation behind these warnings is straightforward: the size of the Christian population has decreased dramatically since the National Pact of 1943, and this decrease is due in no small part to the disproportionate share of Christians among those who have left Lebanon (Salamey and Payne, 2008; Faour, 2007). Christian religious and political leaders unsurprisingly view this pattern as a threat to Christian political power in Lebanon; as the Christian share of the Lebanese population drops, it becomes increasingly difficult to justify the still-considerable privileges reserved for the Christian community in general and Maronites in particular. A Maronite in the Bekaa Valley described being told to “adher[e] to Christian values, land, and nonmigration.” Thus, in addition to vaguer calls for Christian unity and the protection of Christian interests in the country, many religious leaders have provided specific, strategic instructions, including discouraging members of their congregations from leaving the country. You are the foundation of the East. You must stay in this country and reduce the rate of immigration. —Lebanese Maronite, Mount Lebanon

Chapter 4 will present quantitative data that echo these patterns. Lebanese Christians, faced with the prospect of political reform, emphasize unity and the preservation of Christian political and economic advantages. Christians who go to church are particularly likely to view politics through this lens; they care more about their identity as Christians than non-attenders, and their political preferences reflect the interests of the sect.

Sunnis: Fabricated Divisions and External Threats Many of the explicitly political comments made by Sunni interviewees reflected or responded to recent developments in the region. Specifically, Sunnis were concerned with what they perceived as threats to Lebanon

the religious experience 47 coming from other parts of the Middle East. Such concerns are understandable. Lebanon has long been a setting in which external actors have engaged in various forms of proxy conflicts, and this pattern has been even more noticeable in recent years. Since the outbreak of the Syrian Civil War, Lebanon has, in many ways, come to resemble a microcosm of the region as a whole: divided, partially but not entirely, along sectarian lines with respect to the conflict. Various Lebanese factions have also been used as pawns in conflicts between larger regional powers, especially that between Saudi Arabia and Iran. In November 2017, these tensions flared up in dramatic fashion. Lebanese prime minister Saad Hariri was “summoned” to Saudi Arabia. Soon thereafter, he announced—from Saudi Arabia—that he intended to resign as prime minister, citing fears of assassination and particularly pointing to the threats posed by Hezbollah (a Shi‘a political party and militia) and its Iranian allies in Lebanon. Hariri remained in Saudi Arabia for nearly two weeks following this announcement, leading many observers (including Lebanese president Michel Aoun) to declare that he was being “held hostage” by the Saudis. While Hariri eventually returned to Lebanon and rescinded his resignation, the events of November 2017 serve as a strong symbol of regional interference in the domestic affairs of Lebanon and, particularly, its Sunni community. Consequently, it is not terribly surprising that Lebanese Sunnis would express concerns about external threats. As a Sunni respondent from Akkar expressed, “Saudi Arabia is exerting pressure on the prime minister.” This type of sentiment was shared by other Sunnis, along with some members of other sects as well. Some Sunnis described a general concern about involvement with other countries, including a 24-year-old from the Bekaa Valley, who emphasized the need for “coexistence and noninterference in foreign policy.” Other Sunnis referred to Iran: The discussion [in mosques] is linked to the political period or religious occasion. Politically and in general, the discussed issue is Iran. The message is related to current events and circumstances. —Lebanese Sunni, Saida

The second major theme expressed by Sunnis—and one that is closely tied to the issue of foreign interference—is the presence of artificial divisions within the Sunni community. Members of every sect stressed the importance of sectarian unity, and often suggested that communal religious practice

 faith in numbers furthered that goal, but Sunnis commonly referred to the particular presence of fabricated divisions. A surprising number of respondents used this specific term: Sunnis should unite against the fabricated policies that might lead to the fragmentation and division of Sunni Muslims. —Lebanese Sunni, Beirut There should not be a division among Sunnis; they should confront any problem that is being fabricated to divide Sunnis in Lebanon. —Lebanese Sunni, Beirut [Mosques] work toward the unity of Sunnis by confronting the fabricated political games. —Lebanese Sunni, Beirut

These remarks reflect both an emphasis on sectarian unity and a sensitivity to regional developments. Interviewees stated that foreign actors were conspiring to divide (and therefore weaken) the Sunni community. Such statements mirror the general Sunni concern of finding themselves on the losing side of national (and regional) conflict. Thus, as a Sunni from Hermel put it, “It is essential to unite to counter threats.” For Sunnis, it is evident that mosque attendance heightens awareness of regional developments and strengthens desires for sectarian strength and unity. Chapters 4 and 5 will use quantitative data to show that mosque attendance drives Sunni political preferences toward group-centric concerns, and that this effect can change as political conditions evolve.

Shi‘a: Deprivation and Anti-sectarianism The historical experience of Lebanese Shi‘a gives them a different perspective on sectarian-political issues in Lebanon. As subsequent chapters will explore, the history of Shi‘a in Lebanon has been marked, until very recently, by disenfranchisement, marginalization, and neglect by the state. In general, Shi‘a have had less voting power, less political influence, lower levels of socioeconomic development, and disproportionately poor access to state resources. Consequently, narratives of sectarian deprivation abound in

the religious experience 49 Shi‘a descriptions of their communal religious experiences. These messages focused on protecting the sect (as witnessed within other sects as well), but expressed a unique concern for righting the historical wrongs of state policy toward Shi‘a. Moreover, Shi‘a often conveyed feelings of their sect being threatened. A Shi‘a from South Lebanon stressed the importance of “maintaining the sect and not allowing it to be harmed.” Another Shi‘a from the same region described how communal prayer gave her a sense of “power and solidarity.” A third, from Nabatieh, reported that he heard messages in mosques “explaining [our] social status and raising awareness of its risks.” These sentiments reflect an understandable fear of mistreatment and an ambition for improving the sect’s political and economic status in Lebanon. [We must] protect the dignity and the rights of our sect, and stop underdevelopment. —Lebanese Shi‘a, Baalbek

Some of these comments addressed political underrepresentation and, in the words of a Shi‘a from Ba‘abda, “attempts to weaken Shi‘a.” Others emphasized economic neglect and underdevelopment; a Shi‘a from Nabatieh reported that he had been told in mosque that “the country neglects poor areas.”9 Still others presented a general sense of sectarian marginalization: Humiliation is a thing of the past. —Lebanese Shi‘a, Bint Jbeil I feel the unity and power of the sect. Humiliation is a thing of the past.10 —Lebanese Shi‘a, Beirut

A final major theme to emerge from Shi‘a reflections on the communal worship experience was an opposition to the sectarian political system. Straightforward demographic calculations give Shi‘a reasons to prefer the abolition of the sectarian system, which holds fixed their level of political representation in spite of their growing numbers in Lebanon. It is important to note that this “anti-sectarianism” does not mean an opposition to sect as a political identity or a lack of sectarian solidarity; as demonstrated previously, the messages presented in these interviews were often highly sectarian regardless of the sect of the interviewee. However, many Lebanese

 faith in numbers Shi‘a expressed anti-sectarian sentiments in the sense that they opposed sectarianism as a political system. In other words, they often favored a move toward an electoral system (and a political system more generally) that is not based on fixed sectarian quotas but rather reflects the changes in Lebanese demographics. For obvious reasons, such sentiments were virtually nonexistent among members of other sects. Many Lebanese Shi‘a, however, strongly emphasized the importance of moving away from the sectarian system as the 1989 Ta’if Accord had stipulated. In a response characteristic of the feelings of many Shi‘a, a respondent from Baalbek plainly called for “political reform and the abolition of sectarianism.” Others explicitly blamed the sectarian system for the poor experiences of their sect and the country more generally: Sectarian policy is the cause of Lebanon’s problems. —Lebanese Shi‘a, Zahle

As with Christians and Sunnis, Shi‘a descriptions of their religious experiences reflect a response to their political realities. Communal worship promotes sectarian solidarity, but its effects do not stop there. Such experiences promote attitudes in alignment with the interests of the worshiper’s sect, and, as explored in subsequent chapters, therefore have disparate effects on attitudes toward democracy, representation, and economic redistribution.

Conclusion This chapter has attempted to provide a broad sense of the religious experience of worshipers in Lebanon and Iraq as reflected in their own words. Responses in these two countries display many similarities. In both countries, mosques and churches are sites of either implicit or explicit sectarian solidarity. The process of religious identity-building sometimes works through overt channels such as political sermons. As many interviewees reported, however, this process often works more subtly. The communal religious experience brings worshipers closer together in a way that promotes sectarian solidarity and unity. The building of sectarian communities in these environments also serves to encourage sect-centric political attitudes and behaviors, where political outlooks are influenced heavily by the interests of the sect. These interests are where the effects

the religious experience 51 of communal prayer diverge. Sunnis, Shi‘a, and Christians in both Iraq and Lebanon all report that religious attendance heightens their sense of sectarian identity, but the implications of this identity for political attitudes and behaviors differ according to the interests of each sect, sometimes promoting support for democracy and/or redistribution but sometimes doing just the opposite. The next several chapters will use large-scale survey data to examine how these communal religious behaviors translate into sectarian political preferences. Chapters 4 and 5 provide quantitative evidence that echoes the themes in this chapter. As these chapters will show, the effect of religious attendance on attitudes toward democracy and redistribution is contingent on sectarian interests. Importantly, political context matters. The following two chapters examine the Lebanese case during two important periods in the country’s political development, and illustrate the ways in which the effect of religious attendance changes as the political context evolves.

4 Christians and Muslims in Lebanon Before the Syrian Civil War Introduction This chapter assesses my theory of communal prayer and democratic support in Lebanon in 2007. I argue that before the intensification of Sunni-Shi‘a tensions in 2008 and, especially, the Syrian Civil War, issues of electoral reform and economic redistribution placed Christians and Muslims at odds with each other on the issue of democratization. The evidence suggests that communal prayer had opposite effects on attitudes toward democracy: underrepresented and relatively poorer than Christians, Muslims had incentives to support democratic reforms, and communal prayer increased support for democracy. Among Christians, who had much to lose from both electoral reform and redistribution, the effect of communal prayer was just the opposite. Data from the first wave of the Arab Barometer support these claims. The Lebanese case is a useful one for the goals of this study because it is (a) a society that is clearly divided along religious lines, and (b) a state possessing institutions designed to mitigate intersectarian conflict, making the incentive structures for groups fairly clear for the purposes of this model. Despite its relatively peculiar set of institutions (not to mention its history), Lebanon presents several distinct advantages in helping us understand sectarian identities and institutions. First, and most obviously, Lebanon’s sectarian divisions ensure that religion plays a crucial role in politics; Harris (2009, p. 9) states that religion is “the primary characteristic” of Lebanon. Religion has a powerful effect on social and political life in Lebanon independently of its institutional effects. Second, Lebanon’s semidemocratic history provides two important assurances: (1) citizens have some experience of democratic politics (or something like it); but (2) democracy is not, in the words of Linz and Stepan (1996, p. 5), “the only game in town.” Consequently, there should be some genuine variation in Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0004

christians and muslims in lebanon 53 support for democracy in this setting. Existing surveys in Lebanon suggest that, indeed, while a majority of Lebanese citizens express support for liberal democracy, strong democratic attitudes are far from unanimous. Lebanon’s torturous political history creates a useful setting for analysis of attitudes toward democracy; its history of relatively liberal policies sets it apart from virtually all of its neighbors, but Lebanon has also experienced some of the most violent civil strife in the region, including a bloody and protracted civil war from 1975 to 1990. Despite these challenges, Fawaz (2009, p. 33) suggests that Lebanon “might be the only ray of light coming out of this region.” Each of these factors makes Lebanon a useful area to study the ambiguous link between religion and democracy. Importantly, Lebanon does not possess many other identity cleavages that could potentially muddy the sectarian waters. While the bulk of other religiously divided societies also possess ethnic, regional, linguistic, or other types of divisions that either overlap with or cross-cut religion, Lebanon is a small and largely ethnically homogeneous country (apart from sect, of course). The role of religious identity is so central in Lebanon1 that Corstange’s (2013) work refers to religious identities in Lebanon simply as “ethnicity.” The identity groupings in Lebanon are religious in nature and not exactly like nonreligious ethnic identities. These identity arrangements allow for an analysis of Lebanon that will ensure that any identified effects are due to religion specifically and not to other potentially intervening identities. In a study focusing on the relationship between religious identities and democracy, it is essential that the primary case demonstrate this rare characteristic. For these reasons, Lebanon is an ideal candidate for selection in this book.

Demographics, Representation, and Status in Lebanon Following independence from a French mandate in 1943, the leaders of Lebanon’s major sects (at the time, Maronite Christians and Sunni Muslims) agreed upon an unwritten National Pact, in which key political positions were divided among the sects, loosely based on the 1932 census (Trablousi, 2007). Under this system, parliamentary seats are contested only within sects; that is, each seat can only be occupied by a member of the designated sect. The motivation for this system is the assumption that allocating seats on the basis of sect will prevent intersectarian competition and instead

 faith in numbers force candidates and parties to compete with members of their own sects (Cammett and Issar, 2010, p. 382), thereby (theoretically) reducing the intensity of sectarian politics. Lijphart (1977, p. 149) argues that, at least prior to the civil war, this system “performed satisfactorily.” It is well beyond the scope of this book to provide an account of the Lebanese Civil War (1975–1990),2 but a few observations are important. First, the war was far from being a simple case of sectarian conflict: sharp divisions existed within each sectarian community, and the position of any particular group (or individual) could not be deduced merely from its sectarian affiliation. Moreover, alliances and interests shifted considerably during the war. Groups who were once allies often became bitter enemies, and groups who had fought bloody battles against each other sometimes reached agreements to end the fighting or even cooperate against other groups. As the Lebanese state broke down, militias filled in to provide basic services, leading to heavily segmented “cantons” controlled by armed groups. Finally, international factors, especially the Israeli and Syrian occupations of large parts of Lebanon, heavily influenced the exercise of political power in the country for many years. Importantly for the purposes of this book, however, sectarian grievances represented a major motivation for the conflict among many key groups. The 1989 Ta’if Accord, which formally ended the conflict, sought to remedy the intergroup inequality crystallized by the sectarian system, and these measures (discussed more subsequently) were seen as the most essential part of any peace agreement. However, postwar developments have led many observers to question the utility of the sectarian system itself. Political competition remains heavily sectarian, and sectarian violence is not uncommon. Some observers have questioned the fairness of a system in which many citizens cannot vote for any candidates from their own sect; in a number of districts, a substantial share of the population belongs to a sect that does not have any seats allocated to it in that particular district. Recent changes to the electoral law, culminating in the May 2018 parliamentary elections, have altered the details of how citizens vote (and which parties and candidates are chosen), but the fundamental principle of sectarian representation remains the same: competition is designed to take place within sects rather than between them, and no sect can hope to gain more seats (nor fear that it could lose seats) than it is allocated. This system, designed to keep sectarianism contained, actually perpetuates sectarian identity.

christians and muslims in lebanon 55 Importantly, no census has been conducted in Lebanon since 1932, largely due to fears of sectarian unrest if a census were to occur. Table 4.1 summarizes the allocations made by the National Pact as well as the results of the 1932 census. Since 1932, Lebanon has experienced significant demographic change. While it is difficult to determine the magnitude of this change with certainty because no census has been conducted, there is little doubt that the Christian population has decreased in relative size. The US State Department’s (2008) report on religious freedom in Lebanon cites recent statistics estimating that Sunnis and Shi‘a each constitute about 28% of the Lebanese population,3 with Christians representing less than 40% of the total and Maronites specifically only reaching about 22%. Even in the absence of official figures, Lebanese leaders and citizens are aware of the general demographic trends. Druze leader Walid Jumblatt suggests that “demography is not in [the Christians’] interest. . .they are now 30%.”4 Despite these demographic shifts, the Christian community remains centrally represented in nearly every facet of the national government. Although the Ta’if agreement of 1989, which ended the 15-year civil war, weakened the presidency and reallocated parliamentary seats in such a way as to guarantee 50% of seats for Muslims, the other 50% remained reserved for Christians.5 There have been few, if any, more recent measures to remedy the relative underrepresentation of Muslims in Lebanon. The fact that the Table 4.1 Demographics and Allocations, 1943 National Pact Confession

Percentage of Population, 1932

Political Allocation(s)

Maronite Greek Orthodox Greek Catholic Other Christian

28.8% 9.7% 5.9% 5.6%

President Deputy Speaker of Parliament

Total Christian

50.0%

6 parliamentary seats per 5 Muslims; control of army

Sunni Muslim Shi‘a Muslim Druze

22.4% 19.6% 6.8%

Prime Minister President of National Assembly

Total Muslim

48.8%

5 parliamentary seats per 6 Christians

Other

1.2%

 faith in numbers country has not conducted a census since 1932 provides further evidence of the Christian community’s ability to maintain its disproportionate share of formal representation despite changing demographic realities. Lebanon specialist Elias Muhanna (2013) reaches similar conclusions, estimating that for each Sunni seat in parliament, there are about 34,512 Sunni citizens, very close to the corresponding figure of 33,781 for Shi‘a.6 These numbers are vastly different from the representation level of Maronites, who have 20,655 voters per seat. Figure 4.1 displays the differences between population shares and parliamentary seat shares by sect. Christian leaders are aware of this disparity, but resistant to proposals to remedy it. When asked about the possibility of decreasing the proportion of Christian seats in parliament, Alain Aoun (a prominent member of parliament and nephew of president Michel Aoun) remarked: “I don’t think you will have any Christians accepting this.”7 A similar comment was made by Nadim Gemayel, a leading figure in the rival Christian Kataeb Party, saying, “Engaging any change with Ta’if could lead to an arrangement of power in which Christians pay the price, and I am not ready to do this.”8 In this environment, sectarian incentives with respect to democracy were relatively clear. Christians, whose political influence was weakened but

60

54

50

27

27 20

42

41

40

21

27 21

21 10

6 0 –6

–6

–12 –20 Maronite

Sunni

Shi‘a

% of Population Gap

Total Christian

% of Seats

Fig. 4.1 Population and Parliamentary Numbers, by Sect.

Total Muslim (excl. Druze & Alawi)

christians and muslims in lebanon 57 still disproportionately strong after the Ta’if Accord (and whose economic privilege remained considerable), had reasons to resist democracy. Muslims, who, whether Sunni or Shi‘a, could benefit at this point from more equal representation, faced pro-democratic incentives for the same reasons.

Christians and Muslims in 2007 Lebanon Viewed from a present-day perspective of sectarian relations in Lebanon, it might seem puzzling to consider Christian-Muslim divisions when the Lebanese Muslim community is so heavily divided in the current environment. However, there are both theoretical and practical reasons for placing Sunni and Shi‘a Lebanese in the same category for the purposes of this chapter: from a theoretical standpoint, both groups remain underrepresented in government and, perhaps more importantly, are much poorer on average than Christians. Further, in the period discussed in this chapter, tensions between Sunnis and Shi‘a had not reached nearly the same level of intensity that would emerge a few years later; it is therefore much more plausible to focus on the shared interests of the two sects at this stage of the country’s political development. From a practical standpoint, the available survey data simply do not contain enough respondents to separate Sunnis and Shi‘a and still retain enough power to detect meaningful relationships in the data (there are fewer than 200 non-missing observations from these groups even for the baseline models, compared to over 500 for Christians). The focus of this chapter is on issues of representation and economic equality (and it is here where Sunnis and Shi‘a are most alike to each other but markedly different from Christians). Data from the 2007 survey indicate that income and political representation differences between Sunnis and Shi‘a paled in comparison to the differences between Muslims in general and Christians in general during this period. With unlimited data availability, it would be ideal to consider these groups separately (and subsequent analyses will do so), but for the reasons just outlined, practical limitations render such an analysis unfeasible. Thus, this chapter considers them together as “Muslims” even though there are, without question, significant differences in the outlooks and interests of these groups.9 It is important to note that I am not assuming (nor suggesting) that Sunnis and Shi‘a are in any way “allies” in Lebanese politics; recent developments in the country have proven that this is most certainly not the case. However, their economic and political

 faith in numbers positions relative to Lebanese Christians placed their interest in democracy in the same camp during this period; it is more than reasonable to suspect that a full transition to democracy in Lebanon would transfer both political and economic resources away from Christians and toward both Sunnis and Shi‘a. Therefore, while the two sects are in many senses adversaries in Lebanese politics, their incentive structures with regard to democracy were highly similar in this context.10 In April 2005, Lebanon’s Sunni leaders achieved a major political victory. After 29 years of occupation, the Syrian army, whose presence in Lebanon was widely opposed by Sunnis, was forced to withdraw from the country. Syria had occupied Lebanon in 1976 during the early days of the Lebanese Civil War, ostensibly to protect the country (and the region) from Palestinian fighters. Troops remained in the country long after the end of the war, and Syria dominated Lebanese politics throughout that period. The spark that ultimately led to Syria’s withdrawal was the assassination of Rafiq Hariri, the former prime minister and political leader of Lebanon’s Sunnis. Syria was widely perceived to have been involved in the assassination, as Hariri had been a thorn in its side for years; massive public demonstrations along with considerable international pressure forced Syria to withdraw, marking the beginning of a new era in Lebanese politics. The Sunni community, victorious in its opposition to the Syrian occupation, had reason to be optimistic during this period. Elections held later that year brought a Sunni-led anti-Syrian coalition to power (the March 14 coalition). These elections—the first to be held after the Syrian withdrawal from Lebanon, and the first to allow the presence of international election observers—were viewed by many as part of a democratic revolution; the results indicated to the Sunni leaders that they could hold their own against other groups in open elections. The anti-Syrian alliance, led by the Sunni Future Movement, achieved a clear majority in parliament, and the Future Movement itself obtained more than twice as many seats as any other party. Despite their demographic disadvantage relative to Shi‘a, the Sunnis at this point had reason to believe that democracy might not inevitably lead to Shi‘a ascendance. In the next chapter, I will discuss the political changes that caused the Sunni leadership to change its course with respect to electoral reform; for now, it is sufficient to note that the political incentives faced by Lebanese leaders in the wake of the 2005 elections (including the time period addressed in this chapter) were meaningfully different from those in the post–Syrian Civil War setting. In this environment, on the issues of

christians and muslims in lebanon 59 electoral reform, both Sunnis and Shi‘a had incentives to support further democratization in order to shift the balance of power away from the overrepresented Christians. Indeed, sectarian patterns in public opinion reflect this division. A Zogby Poll11 conducted in 2005 asked respondents’ opinions about several issues relevant to the political environment of that period. Respondents were first asked about the use of a “one man, one vote” system to choose the next Lebanese president. Sunnis and Shi‘a reported similar levels of support for this proposal (Sunnis: 83% support, 12% oppose; Shi‘a: 79% support, 16% oppose). Strikingly given the seemingly relatively uncontroversial nature of the question, Maronites expressed far more opposition, with 49% supporting the proposal and 42% opposing it. Figure 4.2 displays the net levels of support by sect for the “one man, one vote” question along with several other illustrative questions. In 2005 (even after the assassination of Rafiq Hariri, when this survey was conducted), Sunnis and Shi‘a had similar attitudes on the prospect of changing the National Pact, the proper role of Shi‘a-dominated Hezbollah in Lebanon, and whether or not the United States should put more pressure on Syria to disarm Hezbollah. In each case, Christian responses to the question are

Change National Pact

One Man, One Vote

Hezbollah Role in Lebanon

US Pressure to Disarm Hezbollah

–100

–50

0

50

Net Support (Support–Oppose) Maronite Sunni

Fig. 4.2 Sectarian Attitudes in 2005 Zogby Poll.

Shi`a

100

 faith in numbers remarkably different from those of Muslims. It is striking that even on the role of Hezbollah, which would come to be one of the defining issues of the Sunni-Shi‘a divide in a few years, Sunnis and Shi‘a report relatively similar attitudes, in stark contrast to the attitudes of Christians. These attitudes reflect the interests of each sect at the time: prior to the 2008 Hezbollah– Future Movement conflict and, especially, the 2011 Syrian Civil War, Sunnis and Shi‘a were far closer to each other in terms of political preferences (and far more distant from Christians) than they would be in just a few years. The behavior of religious leaders during this period reflects the political interests of their sects as well. Reporting on the 2010 funeral of Lebanon’s preeminent Shi‘a religious figure, Muhammad Hussein Fadlallah, a writer for Hezbollah’s al-Manar news outlet suggested that the only absent Lebanese religious leader was “the small minded Maronite Cardinal Patriarch [Nasrallah Boutrous] Sfeir who took umbrage at Fadlallah’s positive views about one person-one vote democracy. Sayeed Fadlallah did not favor Lebanon being ruled by the current archaic French colonial legacy of parceling out political power based on the undemocratic confessional formulae of the 1943 Paris installed National Pact.”12 Fadlallah’s own words refer to “the Muslims who form the majority” (emphasis added) and stress “the legitimacy of the rotation of power,” indicating both support for democracy and a sense of broader Muslim political interest.13 Indeed, prior to 2008, the discourse of Fadlallah and other Muslim religious leaders was in general much more likely to emphasize Islamic cohesion rather than sectarianism; Fadlallah offered several public statements in 2006 and 2007 with the term “Islamic Unity” in the title; over the next few years, these titles shifted to refer to Sunni-Shi‘a issues with increasing frequency. The alignment of political interests within Muslim and Christian groups is evident among both political and religious leaders. A question remains as to the extent to which we can consider “Christians” to be a unified group in Lebanese politics. On many issues, this is most certainly not the case: issues such as Syria’s role in Lebanon have driven wedges between different Christian political actors, and Christians (even Maronites specifically) often disagree in high-profile political debates. However, on the fundamentals of Christian interest in the country—particularly in the political realm—relations between the otherwise conflicting Christian parties are remarkably harmonious. Negotiations regarding the electoral laws are illustrative. In 2011 (before Lebanon became heavily affected by the Syrian Civil War), the leaders of the Maronite Church organized a meeting

christians and muslims in lebanon 61 in which all of the major Maronite and Orthodox parties agreed to endorse an electoral law that would have turned the country into a single electoral district in which citizens voted exclusively for seats allocated to their own sects.14 Unsurprisingly, this arrangement was seen to benefit Christians at the expense of Muslims, and would have embedded sectarianism even more concretely into the electoral system. The major Christian leaders had demonstrated similar unity in 2010 when parliament considered lowering the voting age to 18, which would have diluted the (on average, older) Christian share of the voting population. On fundamental issues of regime and representation, Christians were—at this stage—able to come together despite major disagreements on other significant subjects. Likewise, while Sunnis and Shi‘a were often at odds with each other with respect to other political issues, it was clear that Christian attempts to reinforce their overrepresentation in government would disadvantage Muslims of both sects.15

What Democracy Means to Lebanese Citizens An important lingering question has to do with the meaning of democracy; that is, when citizens express opinions about democracy, what are they imagining the content of democracy to be? It is clearly not possible to provide a comprehensive account of the full meaning of democracy for the entire population. However, it is possible to sketch some broad suggestive patterns of what individuals consider to be the key features of democracy. To this end, the survey I conducted in Lebanon16 that was described in chapter 3 asked respondents to provide a few words or sentences to describe what comes to mind when they think about the meaning of democracy. These responses provide insights into what types of considerations factor into individual assessments of democracy, and generally suggest that most citizens view democracy in the expected ways. Respondents (1,500 in total) overwhelmingly emphasized core democratic values such as “freedom,” “equality,” and elections. These responses reflect both the electoral and liberal dimensions of democracy and suggest that ordinary Lebanese citizens generally understand democracy in ways that are relatively similar to scholarly conceptions of democracy. Furthermore, these perceptions did not noticeably differ by sect. More than half of responses17 contained the words “free” or “freedom,” with close to half of these responses explicitly mentioning freedom of expression. Nearly a third

 faith in numbers mentioned equality in some form. Almost 400 respondents referred to “the people.” In some ways, it is more useful to examine the second-tier responses, since the bulk of respondents’ descriptions focused on the basic features of democracy, equality, and rule by “the people.” Other themes were noticeably less common, but potentially reflect important perceptions beyond the basic characteristics of democracy. A number of respondents referred to electoral reform,18 the abolition of the sectarian system,19 or economic (often redistributive) policies.20 A number of respondents even explicitly mentioned “majority rule” or “the dictatorship of the majority.” These responses reflect the fact that citizens have a diverse set of views about what democracy means, but their responses generally fall into a set of relatively straightforward and predictable categories, suggesting that their answers to questions about “democracy” reflect understandings of democracy consistent with scholarly definitions of the concept. Importantly, almost none of the respondents referred to protections for minorities as a feature of democracy. If citizens viewed democracy as a means of ensuring safety and/or equality for minority groups, then the data would be somewhat stacked against the theory of this book, which suggests that minority groups often have reasons to fear democracy. However, it does not appear to be the case that respondents view democracy primarily as a means of protecting minority rights. Indeed, the word “minority” only appeared in two responses, one of which described democracy as “the majority governs the minority” and the other of which refers to democracy as “minority control.”21 Thus, citizens do not seem to perceive democracy as being defined essentially by its protection of minorities; the prospect of majoritarianism was far more likely to be cited than the ensuring of minority rights. A further question that is important for the purposes of this study is, how democratic is Lebanon? The country holds (mostly) regular elections22 and honors the results, but it would be a mistake to characterize Lebanon as anything close to a full democracy Freedom House (2011) plainly notes that “Lebanon is not an electoral democracy.” This characterization comes from a variety of factors in addition to the confessional system. In many cases, parties cut deals prior to election day in which they agree not to field candidates in certain districts in order to avoid “divisive campaigning.” Significant interference from foreign powers also calls into question Lebanon’s democratic credentials. Finally, vote-buying is rampant across

christians and muslims in lebanon 63 Lebanon: using survey data and new methods to elicit truthful answers to sensitive questions, Corstange (2012b) estimates that 55% of Lebanese sold their votes in the 2009 parliamentary elections. Taken together, it is reasonable to conclude that Lebanon falls far short of conventional standards of democracy.23 It should be noted that this understanding of the distribution of power in Lebanon is not undisputed. It is certainly the case that Lebanon has, since the end of the civil war, experienced some forms of liberalization, and these reforms have usually come at the expense of the previously dominant Christian groups. I have suggested (with some evidence from both the political and economic spheres) that Lebanon’s Christians remain more powerful than they are often claimed to be. However, to the extent that Lebanon has moved closer to representative democracy, Christians have usually been forced to cede some power to Sunnis and Shi‘a (Bahout, 2017). Thus, while I argue that such redistribution of power has not been substantial enough to genuinely undermine the Christian community’s privileged position in politics and society, one does not need to accept this account of Lebanese politics in order to accept the findings of this chapter. Two very different potential versions of the Lebanese case can both be reconciled with my theory. The first is the account I have presented, in which democratization has been limited and Christians have largely managed to preserve their power. The second is that Lebanon has democratized, and this transition has shifted power away from Christians and toward Muslims. A full discussion of the merits of each of these accounts is beyond the scope of this chapter, but either story is reconcilable with the theory presented in this book: in either case, democracy—whether potential or actual—would benefit Muslims disproportionately at the expense of Christians.

Economic Considerations Beyond the political realm, it is clear that considerable inequities exist between Christians and Muslims (and between smaller sects within these groups) in Lebanon. While the World Bank’s data on inequality are consistently absent for Lebanon (surely not a coincidence), Credit Suisse’s Global Wealth Databook (2012) covers countries omitted by other sources. Of the 166 countries in the sample, Lebanon ranks as the fifth most unequal country24 —and two of the four countries reported as more unequal than

 faith in numbers Lebanon (St. Kitts & Nevis and Belize) have populations of far less than 1 million people. Much of this inequality takes the form of intersectarian disparities. Figure 4.3 illustrates the considerable differences in average income across sects. These average income figures come from survey responses from the first wave of the Arab Barometer, and indicate that the average Christian in the sample earns over $6,000 more per year than the average Muslim, a 40% difference.25 Inequalities exist beyond wealth alone; Christians are more than twice as likely as Muslims (both Sunni and Shi‘a) to have attended private schools.26 Of the 178 countries for which World Bank data on education spending are available, Lebanon ranks fourth lowest; only the Central African Republic, Monaco, and Zambia spend less per capita on education. In the area of labor rights, Salloukh et al. (2015, ch. 5) argue that the country’s political elites have systematically “sectarianized” workers’ movements. Sectarian inequality of both wealth and opportunity is therefore reinforced by the state behaviors. The state has failed to invest in public goods and has suppressed labor mobilization; changes to these behaviors would tend to benefit Muslims, indicating that a regime change could improve the standing of these sects relative to the better-off Christians. Religious leaders have made note of these inequalities; Fadlallah stated in 2006 that “the wealth of the nation is being plundered wasted, and not distributed. A few people put their hands on it as if were private property. This constitutes a big injustice.”27

Average Monthly Income (In LBP)

2,500,000

2,000,000

1,500,000

1,000,000

500,000

0

Christian

Sunni

Fig. 4.3 Average Monthly Income, by Sect (in Lebanese Pounds).

Shi`a

christians and muslims in lebanon 65 This inequality is magnified further by perceptions of inequality: Christians have historically been perceived (by both themselves and by others) as the privileged group in Lebanon,28 while Muslims have historically had less access to both wealth and social resources (Mackey, 1989, pp. 13–14).29 Historical perceptions, of course, remain sticky—even if Muslim wealth were to increase relative to that of Christians, it would likely take some time before perceptions of inequality were updated completely. The prevalence of kinbased organizations as providers of social insurance likewise suggests that the economic fates of members of particular religious groups in Lebanon are linked together in significant ways (Baylouny, 2010). Since these kin ties overlap considerably with sect, members of the same sect (usually in smaller communities) frequently rely on each other for social services through such associations, which “play a crucial role in meeting the basic needs of the population” in Lebanon (Cammett and Issar, 2010, pp. 381–382). These structures provide a further incentive for Lebanese citizens to consider group interests rather than individual considerations alone. The group-centric view of economic issues manifests itself in political priorities as well. When Lebanese Arab Barometer respondents were asked about the most important problem facing the country today, Muslims were considerably more likely than Christians to point to the “economic situation” (over 61% compared to less than 50%). Much of the literature in the field of comparative democratization suggests that democracies are likely to redistribute wealth more than nondemocracies, and that wealthy elites therefore have incentives to prevent democratic transition, while the poor have incentives to favor it (Boix, 2003; Acemoglu and Robinson, 2006). While recent studies have cast doubt upon this relationship (Ross, 2006; Mulligan et al., 2003; Timmons, 2010), the key requirement for the logic presented here to hold is that citizens perceive democracies to be more redistributive than non-democracies. In the survey I conducted in Lebanon,30 most respondents believed that democracies tend to be more redistributive than non-democracies. When prompted with the statement: “Democracies redistribute income from the rich to the poor more than other types of governments,”31 61% of respondents agreed, while only 24% disagreed. The association between democracy and redistribution—in the minds of citizens, if not in cross-national data—suggests that citizens should believe that democracy benefits poor groups at the expense of rich groups. Thus, economic concerns also contribute to citizens’ regime

 faith in numbers preferences. In the case of Lebanon, Christian-Muslim (and later, SunniShi‘a) economic disparities serve as an additional motivation for democratic or non-democratic attitudes. Economic considerations also overlap with political forces in certain key ways. The Lebanese state’s effects on economic inequality are characterized much more by neglect or ineffectiveness than by active intervention. This neglect is not lost on Lebanese citizens; as Cammett (2015, p. S83) notes, “The phrase mafee dawla (there is no state) is a common refrain in everyday conversation.” This neglect uncoincidentally perpetuates existing structures of privilege and deprivation. As described earlier, Christians’ outsized political influence has, until recently, coincided with considerable economic privilege as well. Historically, Lebanese Christians have enjoyed greater access to state resources than their Muslim counterparts, and developed their own institutions for education, commerce, and social welfare earlier than Muslims (Cammett, 2014a, ch. 2). This was especially true for members of the Maronite Church, who had begun to invest heavily in educational institutions decades if not centuries before the creation of the modern Lebanese state. The physical protection of Christian communities— particularly the Maronites—by European states provided further opportunities not available to Muslims (Pearlman, 2013a, p. 111). While the effects of these historical intersectarian disparities have faded over time, they have not disappeared entirely in empirical reality or in collective memories and imaginations. Importantly, these disparities condition the way that many citizens view democracy. If people believe that democracy involves economic redistribution—and the evidence presented previously suggests that most Lebanese citizens do so—then sectarian economic inequality may provide an additional consideration for individuals’ regime preferences.

Communal Prayer and Sectarianism Regardless of its implications for political and economic attitudes, the salience of sectarian identity in Lebanon is scarcely disputed. Although the civil war from 1975 to 1990 was a multifaceted conflict that should not be reduced to simply sectarian strife, Lebanon’s modern history has largely been written along sectarian lines. The civil war only exacerbated these differences: as Salamé (1986) notes, “It is very difficult to distinguish a Druze from a Maronite or a Sunni from a Greek Orthodox in the street . . . the

christians and muslims in lebanon 67 system has rigidified and has become exceedingly compartmentalized into nearly non-communicating confessions.” Lebanon’s consociational institutions have been both a blessing and a curse. Lijphart’s (1977) classic work on consociational democracy cites Lebanon as something of a success because its institutions had, at the time of his writing, held the country together (for the most part) for 30 years. After the experience of the civil war, however, it became clear that Lebanon’s confessional system had proven too rigid to solve all of the country’s problems, and the post-civil war era has seen many calls for some form of secularization of the system. I have argued that the crystallization of sectarian identity in Lebanon— while driven considerably by political and institutional factors—is reinforced by communal religious behavior. Deeb’s (2006, pp. 105–106) ethnography of Shi‘a in the suburbs of Beirut observes that in Lebanon, “public prayer carries social capital.” She quotes a young volunteer: “Why is prayer in a group? Because you find everyone, poor, rich, educated, uneducated, all are standing in a single line praying together.” Communal prayer, it seems, creates a certain degree of within-group solidarity. By praying with others from the same confessional group, believers build relationships with members of the group and begin to view the world (including the political world) more through the eyes of the group.32 A byproduct of these linkages with members of the in-group is an increase in the salience of sectarian identity. The more citizens engage in communal prayer (which almost by definition involves congregating with members of their own sect and not others), the more they will tend to care about the political fates of their co-sectarians. Given the political and economic incentives already outlined, the linkages forged by communal worship should have effects on political preferences in alignment with sectarian interests. In total, the theory presented in this book and the details of the Lebanese case in the mid- to late 2000s suggest a few hypotheses: Hypothesis 1. Christians who engage in communal religious practice will tend to support democracy less than those who do not. Hypothesis 2. Muslims who engage in communal religious practice will tend to support democracy more than those who do not.

 faith in numbers

Data, Methods, and Findings I test these hypotheses using data from Lebanon coming from the first wave of the Arab Barometer survey, overseen by the Center for Strategic Studies at the University of Jordan. This survey was conducted by Statistics Lebanon in November 2007 in each of Lebanon’s six governorates (muhafazat). The survey includes 1,200 respondents representing each of Lebanon’s main sectarian groups. Without accounting for the key variable of communal prayer, Christians are actually noticeably (about 15%) more supportive of democracy than Muslims. The precise reasons for this first-glance difference are beyond the scope of this book, but are likely due to greater exposure to democratic countries, especially in Europe. To test the theory presented in this book, however, we must consider how communal religious practice affects support for democracy within each sect. The key explanatory variable (Practice) is a dummy variable indicating whether respondents pray at a mosque/church or not.33 A set of controls standard to survey-based regression analysis are added, as well as a handful of controls that might be important in the Lebanese case. Each model is fitted separately for Christians and Muslims in order to allow the effect of all of the explanatory variables to vary by sect; this is a more conservative test of the hypothesis than a pooled model with an interaction term between sect and communal practice (though such a model, of course, indicates similar results). The tests include a variety of specifications to account for several possible omitted variables. Table 4B.2 presents the results of the tests of Hypothesis 1, which predicts an anti-democratic effect of communal prayer among Christians. Model 1 is the baseline model; Model 2 controls for religious piety (measured by “importance of prayer as a quality in a spouse”); Model 3 controls for income (which is not included in the baseline model because of missing data); Model 4 controls for association membership to account for the possibility that religious attenders are simply more involved in group activities than are non-attenders. Cammett (2011) points out that in Lebanon, party affiliation and high-cost mobilization such as protest substantially affect access to social services and thus might be important omitted variables in this type of analysis. While the Arab Barometer did not ask Lebanese respondents about their partisan affiliations in the first wave of the survey,34 it did ask respondents about the importance of a candidate’s party in their voting decisions. Model 4 therefore also includes that question along with protest

christians and muslims in lebanon 69 participation as control variables. Model 5 includes all of the covariates included in Models 1–4.35 In these models, the dependent variable is a binary measure of (strong) support for democracy, in which respondents can agree or disagree that “Democracy may have its problems but is better than any other form of government.”36 In each of the models, the coefficient on communal religious practice is negative and significant. The magnitude of this effect is similar across models, indicating that communal religious practice has a robust association with support for democracy that is not simply due to demographic characteristics, personal religiosity, association membership, or partisan attachments and behaviors. All in all, the models support Hypothesis 1. Table 4B.3 presents the results for the same models as Table 4B.2, but for Muslims37 instead of Christians. In each of these models, the coefficient on communal religious practice is positive and statistically significant. In total, communal practice seems to have an effect on support for democracy among Muslims opposite to its effect among Christians. These results are consistent with Hypothesis 2. Since logistic regression coefficients are difficult to interpret substantively, Figure 4.4 provides a clearer depiction of the predictions of these models. This graph shows the difference in predicted probabilities (marginal effects) comparing the levels of support for democracy among those who participate in communal religious practice to support among those who do not. Each of the five models is run separately for Christians and Muslims. For these simulations, other covariates are held at their observed values (Hanmer and Kalkan, 2013). The graph shows that the estimates of these models are highly consistent across specifications and show strong effects of communal practice on support for democracy, but in opposite directions. For Muslims, communal religious practice is predicted to increase the probability of support for democracy by between 16 and 19 percentage points. For Christians, communal practice is expected to decrease the probability of support for democracy by between 13 and 18 percentage points. In each of these models, the coefficient on religious practice is statistically distinguishable from 0 at the .05 level or better. This figure suggests that the effect of communal religious practice on support for democracy is quite strong in Lebanon, but the direction of the effect is different for different religious groups. It is possible, of course, that these results are dependent on the type of model chosen. To address this concern, I use propensity score matching to improve covariate balance. As predictors of communal practice, I include

 faith in numbers .4

Marginal Effect of Communal Prayer

.2

0

–.2

Christians

Muslims

–.4 1

2

3

4

5

1

2

3

4

5

Model

Fig. 4.4 Marginal Effects of Communal Prayer on Support for Democracy.

age, education, gender, association membership, and religious belief, since presumably these factors might influence both tendency to engage in communal practice and attitudes toward democracy.38 Table 4B.5 presents the results of this procedure. The estimated treatment effects are actually slightly larger than they were found to be in the logit analysis, suggesting that the trends observed above are actually stronger among respondents for whom a close match can be found. For Christians, communal practice is predicted to reduce support for democracy by 20 percentage points, while for Muslims it increases support for democracy by 26 percentage points. Both of these differences are significant at the .05 level. These results provide further support for the claims presented in this section. I also test the theory using a broader measure of commitment to democracy. This measure incorporates three measures of democratic attitudes: it is an additive index of the “strong support” measure described along with two additional questions measuring democratic commitment.39 Scores of 0 indicate that the respondent did not give a strongly pro-democracy response to any of the items, while a 3 indicates that the respondent gave the most prodemocracy response for each of the questions; hence, this measure captures commitment to democracy.

christians and muslims in lebanon 71 Since this measure is ordinal, consisting of only four categories, I run a series of ordinal logistic regressions of the same form as used in the previous tests. Table 4B.4 displays the results of these tests. In each case, the coefficient on the communal prayer variable is in the expected direction and significant at the .1 level or better. Substantively, these effects are also considerable. Among Christians, praying at church is associated with a 5.7 percentage point decrease in the probability of reporting the most democratic attitudes. For Muslims, communal prayer increases the likelihood of the most prodemocracy response by 8.6 percentage points. It is evident that religious attendance in Lebanon is also an important determinant of a broader sense of commitment to democracy. A question remains as to whether or not these religious practices actually lead to different political behaviors among ordinary citizens. I address this question by examining whether church/mosque attenders differ from nonattenders in their propensity to participate in protest. Lebanon has a long history of political protest—sometimes, but not always, sectarian in nature— related to a variety of political issues. A recent example is the well-known “You Stink” movement of 2015, targeting the government’s failure to address a garbage removal crisis in Beirut. Protest is a relatively common practice. Nearly half of Lebanese respondents in the Arab Barometer 2007 survey reported having participated in a demonstration or protest march at least once, with nearly a third of respondents reporting having done so more than once. To determine whether or not attenders were more likely to protest compared to non-attenders, I fit an ordinal logistic regression (because the question allows respondents to respond “never,” “once,” or “more than once”) of the same form as the baseline model. Figure 4.5 displays the marginal effects of communal prayer on the likelihood of protest participation by sect. The effects are quite similar between the two sectarian groups: for both Christians and Muslims, attenders are about 10 percentage points more likely to have participated in protests at least once compared to non-attenders. These effects are considerable in size and are significant at around the .05 level (slightly lower for Muslims, likely due to the smaller sample size). Raw means, not controlling for other variables, show a similar pattern: for Christians, attenders are seven percentage points more likely to protest; for Muslims, the difference is over nine percentage points. The evidence therefore suggests that the effects of communal prayer are not simply limited to attitudes: participants in communal prayer also demonstrate higher levels of political activity, at least

Marginal Effect of Communal Prayer

 faith in numbers

.2

.15

.1

.05

0

Christian

Muslim

Fig. 4.5 Marginal Effects of Communal Prayer on Protest Participation.

in the form of protest. This finding presents an interesting contrast to the arguments (coming either from a Marxist or political culture perspective) that religion creates docile citizens. Among both Christians and Muslims in Lebanon, communal prayer actually makes citizens more politically engaged. Rather than focusing believers on the afterlife, religious practice seems to increase citizens’ interest in addressing worldly political issues.

Conclusion The results presented in this chapter suggest that the impact of religiosity— communal religious practice, in particular—is highly conditional on grouplevel factors. Perhaps surprisingly, this effect appears to be conditional on socioeconomic factors; namely, relative group impoverishment. These findings indicate that members of relatively impoverished religious groups (a) view democracy as a means of remedying the intergroup inequalities present in their countries, and (b) experience a heightened sense of “groupness” due to the act of worshiping with others from their same faith tradition. Importantly, this chapter described Lebanon during a particular phase of its political history. For many, the Syrian withdrawal meant a return to

christians and muslims in lebanon 73 Lebanese sovereignty and brought hopes for democratic reform. The SunniShi‘a tensions, while beginning to fester, had not reached anything close to what they would become in the next several years. Most crucially, the Syrian Civil War was several years away, and would have seemed unthinkable at the time. The Syrian conflict fundamentally reconfigured political alliances in Lebanon (as I will discuss in the next chapter) and hardened the boundaries between Sunnis and Shi‘a in an unprecedentedly severe way. Importantly, the conflict has altered the way that Lebanese leaders and citizens view their own state. Ali Hamdan, a leader for the largely Shi‘a Amal Movement, states, “This entity, this ‘Lebanon,’ is seriously at risk. In 2008, we didn’t used to think that the entity was at risk.”40 The major changes experienced by Lebanon as well as its neighboring countries in the years since 2007 mean that the Lebanon described in this chapter no longer exists. In only a few years, national and regional developments have led to a fundamentally different structure of political and sectarian alliances, shifting groups’ interests with respect to democracy as well as other political issues. The next chapter will discuss the implications of these shifts, and will highlight the need to consider variation in group interests over time in order to better understand the link between religious practices and democratic attitudes.

 faith in numbers

Appendix 4A: Survey Questions Used The following questions were used from the Arab Barometer: 202) Are you a member of any organization or formal groups? “Political parties, living cooperatives or local societies, religious organizations, sport and entertainment clubs, cultural organizations, associations or workers’ unions, farmer unions, professional unions or associations economic organizations or associations, entrepreneurial organizations, parent-teacher associations, or other voluntary organizations.” 1. Yes 2. No 215) Generally speaking, how interested would you say you are in politics? 1. Very interested 2. Interested

3. Little interested

4. Not interested

222) How often do you use the Internet? 1. Daily or almost daily 2. At least once a week 3. At least once a month 4. Several times a year 5. I do not use the Internet 232) To what extent do you agree/disagree with the following statements? 4. Democracy may have its problems but is better than any other form of government 1. Strongly agree 2. Agree 3. Disagree

4. Strongly disagree

502) When you consider what a suitable spouse is for your son or daughter, would you say that each of the following is very important, somewhat important, or not important? 1. S/he doesn’t pray 1. Very important important

2. Somewhat important

3. A little important

4. Not

701) Age . . . 702) Gender: 1. Male

2. Female

703) Level of education: 1. Illiterate 2. Elementary 3. Primary 5. Secondary 6. College Diploma—two years 7. BA 8. MA or higher 714) Do you pray at: 1. Mosque

2. Home

3. Both

4. Church

715) (Income includes all salaries, wages, and rent) Monthly income for individual in (local currency)

christians and muslims in lebanon 75

Appendix 4B: Supplementary Tables Table 4B.1 Descriptive Statistics Min

Max

Mean

Standard Deviation

0 0 18 1 1 1 1 1 0

1 1 82 7 2 5 4 10 1

0.48 0.64 38.47 4.37 1.43 3.83 2.06 5.36 0.18

0.50 0.48 13.17 1.48 0.50 1.64 1.15 2.88 0.38

(Strong) Support for Democracy Communal Prayer Age Education Female Internet Use Spouse Prayer Important Income (Deciles) Association Membership

Table 4B.2 Logistic Regression Results, Support for Democracy (Christians) Model 1 Communal Prayer Age Education Female Internet Use

−0.60∗∗∗ (0.23) 0.01 (0.01) 0.03 (0.07) −0.12 (0.18) 0.06 (0.06)

Spouse Prayer Important

Model 2 −0.55∗∗ (0.24) 0.00 (0.01) 0.02 (0.07) −0.13 (0.18) 0.07 (0.06) −0.18∗∗ (0.08)

Income

Model 3 −0.73∗∗ (0.31) 0.01 (0.01) −0.07 (0.09) 0.50∗ (0.26) 0.04 (0.07) 0.17∗∗∗ (0.05)

Association Membership Party Important Protest Constant Observations Pseudo R2 AIC

0.34 (0.63)

−0.16 (0.68)

−1.31 (0.82)

528 0.013 728.93

520 0.019 715.48

331 0.052 447.36

Note: Standard errors in parentheses. ∗ significant at p < .10; ∗∗ p < .05; ∗∗∗ p < .01.

Model 4 −0.62∗∗ (0.25) 0.00 (0.01) 0.01 (0.07) −0.24 (0.19) 0.06 (0.06)

−0.03 (0.25) −0.22∗∗∗ (0.07) 0.32∗∗∗ (0.12) 0.74 (0.69) 503 0.033 685.68

Model 5 −0.71∗∗ (0.33) 0.02 (0.01) −0.15 (0.10) 0.43 (0.27) 0.08 (0.08) −0.23∗∗ (0.11) 0.20∗∗∗ (0.05) 0.23 (0.34) −0.22∗∗ (0.10) 0.51∗∗∗ (0.16) −2.01∗∗ (0.99) 313 0.104 409.11

 faith in numbers Table 4B.3 Logistic Regression Results, Support for Democracy (Muslims) Model 1 Communal Prayer Age Education Female Internet Use

0.76∗∗∗ (0.28) 0.01 (0.01) −0.01 (0.08) 0.03 (0.29) −0.05 (0.08)

Spouse Prayer Important

Model 2 0.79∗∗∗ (0.29) 0.00 (0.01) 0.01 (0.09) −0.06 (0.29) −0.07 (0.08) −0.32∗∗∗ (0.10)

Income

Model 3 0.72∗∗ (0.34) −0.00 (0.01) 0.03 (0.11) −0.16 (0.40) 0.01 (0.10) 0.00 (0.06)

Association Membership

Model 4 0.77∗∗∗ (0.28) 0.00 (0.01) 0.03 (0.08) −0.04 (0.29) −0.05 (0.08)

−0.79∗∗∗ (0.30)

Party Important Protest Constant Observations Pseudo R2 AIC

−0.81 (0.88) 381 0.026 506.00

Note: Standard errors in parentheses. ∗ significant at p < .10; ∗∗ p < .05; ∗∗∗ p < .01.

−1.36 (0.94) 370 0.050 483.82

−0.78 (1.09) 239 0.027 318.89

−0.66 (0.89) 380 0.040 499.73

Model 5 0.73∗∗ (0.37) −0.01 (0.01) 0.09 (0.12) −0.34 (0.43) 0.01 (0.11) −0.31∗∗ (0.13) 0.04 (0.06) −0.23 (0.38) −0.33∗∗∗ (0.12) 0.19 (0.17) −1.16 (1.29) 227 0.092 293.51

christians and muslims in lebanon 77 Table 4B.4 Robustness Checks, Commitment to Democracy Model 1

Model 2

Model 3

Christians

Model 4

Muslims

−0.34∗ (0.21)

−0.63∗∗ (0.29)

Age

0.01 (0.01)

0.02 (0.01)

−0.00 (0.01)

−0.01 (0.01)

Education

0.03 (0.06)

−0.17∗ (0.09)

−0.01 (0.07)

0.03 (0.10)

Female

0.15 (0.17)

0.86∗∗∗ (0.24)

0.17 (0.26)

0.08 (0.37)

Internet Use

0.07 (0.05)

0.13∗ (0.07)

0.01 (0.07)

0.02 (0.09)

Communal Prayer

0.94∗∗∗ (0.26)

1.11∗∗∗ (0.33)

Association Membership

0.08 (0.30)

−0.13 (0.31)

Spouse Prayer Important

−0.20∗∗ (0.10)

−0.12 (0.11)

Income

0.20∗∗∗ (0.04)

Party Important

0.01 (0.09)

Protest

0.46∗∗∗ (0.14)

Observations Pseudo R2 AIC

502 0.005 1308.00

301 0.064 755.82

0.07 (0.06) −0.39∗∗∗ (0.11) 0.21 (0.14) 368 0.019 926.70

220 0.069 551.76

Note: Ordinal logistic regressions. Standard errors in parentheses. ∗ significant at p < .10; ∗∗ p < .05; ∗∗∗ p < .01.

Table 4B.5 Propensity Score Matching Results

Estimated effect (change in predicted probability) [ATT] Abadie-Imbens Standard error p-value Matched/Total Observations

Christians

Muslims

−0.195 0.095 0.041 452/508

0.257 0.105 0.014 151/369

5 After Syria Communal Religion and Democracy in 2014 Lebanon

In politics, you don’t have eternal enemies. —Mohammed Obeid, Amal Movement1

Introduction This chapter applies my theory of communal religion, group interest, and democracy to the case of Lebanon following the outbreak and spread of the Syrian Civil War beginning in 2011. Using data from an original nationally representative survey conducted in Lebanon between 2013 and 2014, it shows that the relationship between communal prayer and attitudes toward democracy can change drastically, even within a given sect, when the political and/or economic interests of that sect change. By shifting the axis of competition (in terms of both who is competing and what they are competing over), these changes can dramatically alter the effect of communal prayer on support for democracy or non-democracy. The findings in this chapter—both observational and experimental—suggest that the Syrian Civil War and its spillover effects in Lebanon have created an environment in which communal religion depresses support for democracy among Sunnis but enhances support for democracy among Shi‘a. For Christians, who are highly politically divided and largely “on the sidelines” with respect to the Syrian conflict (Hoffman and Nugent, 2015), communal practice no longer has a clear unidirectional effect on regime attitudes.2

Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0005

after syria 79

Religion in 2014 Lebanon Despite the emergence of a visible secular segment in Lebanon, the overwhelming majority of the Lebanese population remains “religious” in the traditional sense.3 The data4 indicate that 79% of Christians, 88% of Sunnis, and 87% of Shi‘a considered themselves “religious.” Over 97% of each sect indicated that religion was at least “somewhat important” in their lives, with a majority of respondents in each sect saying that religion was “very important.”5 Personal piety and self-identification as “religious” are very high across sects. Religious behaviors are also very common in Lebanon. As expected from the different tenets professed by each sect, the types of religious behavior performed by citizens vary across denomination. Muslims of either major sect are more likely to pray at home or at work; 70% of Sunnis and 77% of Shi‘a report praying daily (or more frequently), compared to 27% of Christians.6 Frequency of fasting in Ramadan or Lent is similar; 81% of Sunnis and 76% of Shi‘a report that they “always” observe this practice, compared to 43% of Christians (of course, fasting during Lent is not required in every Christian denomination). Communal prayer follows the opposite trend; 61% of Christians report weekly attendance, compared to 49% among Sunnis and 32% among Shi‘a. Patterns of communal practice also differ within sects. Importantly, gender differences might be present in relative rates of communal prayer. Christian women attend church weekly at a slightly higher rate than men (64% compared to 58%). A common—though mistaken, as shown later— claim is that Muslim women do not attend mosque services; this claim sometimes states that women are not permitted to pray at mosques, but otherwise simply states that women are not encouraged as strongly as men to attend communal prayer. Empirically, it is true that Muslim women engage in communal prayer less than Muslim men, but it is far from true that women never attend mosque services. While Sunni men are far more likely (62%) than Sunni women to attend mosque services on a weekly basis, over onethird of Sunni women report attending mosque at least weekly. Similarly, Shi‘a men report weekly attendance at a rate of 38%, compared to 23% of Shi‘a women. Thus, while it is most certainly the case that women attend mosques at lower rates than men in Lebanon, a considerable number of women from every sect participate in communal prayer regularly.

 faith in numbers An overwhelming majority of respondents from each of the major sects agreed that “religious conflict is the biggest problem facing Lebanon today” (90% of Christians, 83% of Sunnis, and 85% of Shi‘a; similar proportions agreed that “religion is the most important social division in Lebanon”). However, the fault lines defining these religious divisions are often in flux. Recent changes in political circumstances have substantially affected both the axis of political competition in Lebanon and the tone of political discourse. Deeb and Harb (2013, loc. 1076) state that “in 2005, after the assassination of former prime minister Rafiq Hariri, these [Sunni-Shi‘a] tensions came to the foreground and, for the first time in Lebanese history, began to take the form of Sunni-Shi‘i sectarian conflict” (emphasis added). These tensions, which were unprecedented prior to 2005, increased further with the May 2008 fighting between Hezbollah and the Sunni Future Movement, the first time that Hezbollah had used weapons against other Lebanese since the civil war. Furthermore, as I will demonstrate, the intensity of the Sunni-Shi‘a divide was brought to unprecedented heights after the onset of the Syrian Civil War.

Changes in Sectarian Conflict As discussed previously, no census has been conducted in Lebanon since 1932, largely due to fears of sectarian unrest if a census were to occur. The Lebanese Civil War (1975–1990), heavily influenced by concerns about sectarian representation in government, resulted in the Ta’if Accord, an agreement in which Christians and Muslims were each given 50% of parliamentary seats and Sunnis and Shi‘a were given equal shares of seats. While this reconfiguration of sectarian allotments came closer to representing demographic realities, it fell far short of matching the population shares of each sect. Sunnis remained underrepresented relative to their population share, but Shi‘a—who now make up around 40% of the population and are considerably more numerous than Sunnis—are particularly underrepresented. Since virtually every political resource is allocated (at least partially) along sectarian lines, this deprivation has remained a major narrative of Lebanese politics. A key feature of Lebanese sectarian politics in the post-Syria era and, especially, since 2008 has been a widening rift between the Sunni and Shi‘a communities. While the events surrounding the “Cedar Revolution” of 2005

after syria 81 exposed Sunni-Shi‘a divisions to a considerable extent, developments in the interim have brought these tensions to new heights, to the point that the key axis of sectarian conflict in the country is now along Sunni-Shi‘a lines. A relatively benign indication of this growing rift is pointed out by Cammett (2014a, ch. 6), who finds that between the 2005 and 2009 parliamentary elections, Hezbollah (a major vote-getter among predominantly Shi‘a parties) lost considerable support within the Sunni community and the Sunni-led Future Movement lost large numbers of Shi‘a voters. In 2005, Hezbollah (as well as the largely Shi‘a Amal Movement) won a majority of Sunni votes in several districts and the Future Movement won a majority of Shi‘a votes in others; by 2009, this phenomenon was quite rare, and cross-sectarian alliances were more likely to move across the Christian-Muslim divide than the Sunni-Shi‘a divide. Importantly, there is nothing inherent or inevitable about Sunni-Shi‘a tensions in Lebanon. As Cammett (2013a, p. 3) observes,7 “While ShiaSunni tensions are at a high point in Lebanon and in the region, they do not reflect essential and irreconcilable differences rooted in identity.” Rather, several recent developments are responsible for the heightening of the Sunni-Shi‘a division in contemporary Lebanese politics. The 2003 Iraq War unleashed a new wave of sectarian tensions in parts of the Middle East, and the 2005 assassination of former prime minister Rafiq Hariri amplified these tensions in Lebanon. However, the recent state of Sunni-Shi‘a relations was driven by three interrelated developments that have occurred since 2008: armed clashes between Shi‘a and Sunni groups in Beirut in 2008; the outbreak and sectarianization of the Syrian Civil War beginning in 2011; and the use of sectarian identities in broader regional conflicts. First, the series of clashes in May 2008 between Hezbollah and its allies on the one hand and the Future Movement on the other marked a major turning point in sectarian relations. The conflict originated from purely political concerns: the government attempted to shut down Hezbollah’s telecommunications network and remove a Hezbollah sympathizer as the head of security at Beirut International Airport. In response to these moves, Hezbollah fighters took over West Beirut, much of which served as an epicenter for Future Movement support. While the conflict did not, as many observers feared, spiral into a new civil war, it nevertheless marked an important symbolic inflection point: the first time in which Hezbollah used its weapons against fellow Lebanese. Regarding this crossing of a Lebanese Rubicon Salloukh, (2017 p. 67) writes:

 faith in numbers It was the first time in postwar Lebanon that the party had turned its firepower inward against fellow Lebanese, despite frequent promises by the party’s leadership that its weapons were aimed solely at Israel. This precedent left an indelible mark on Sunni-Shia relations in the country and the region, demonizing the party in the eyes of the pan-Arab and pan-Islamic audiences it had carefully cultivated.

Of course, this conflict, like others throughout Lebanon’s history, was not explicitly (or exclusively) sectarian. At face value, it was primarily a clash between two political parties. However, given the close linkages between these two parties and their respective sectarian communities, the reframing of these events as a case of sectarian violence was unsurprising. Mohammad Kabbani, a Future Movement politician, said, “The day Hezbollah took over the streets in Beirut in 2008 is now like Karbala8 for Sunnis. We were humiliated. May 7 created a wall between the Shi‘a and the Sunni” (quoted in Abdo, 2017, p. 101). This “sectarianization” of the conflict was exacerbated by media outlets linked to the participating parties, which “served to harden sectarian modes of identification at the individual level and mobilization at the communal level. It mirrored, but also fueled, social animosities between Lebanon’s sects” (Salloukh et al., 2015, p. 148). Thus, what could have been interpreted as a localized political skirmish took on symbolic meaning as a clash of competing sects, marking 2008 as a significant turning point in Sunni-Shi‘a relations in Lebanon. Even more than the 2008 clashes, however, the effects of the Syrian Civil War have hardened the Sunni-Shi‘a divide in a way that has shifted the axis of political competition in Lebanon since the conflict began. Political, economic, and cultural linkages between Lebanon and its much larger neighbor have remained strong even after the nearly 30-year-long Syrian occupation of Lebanon ended in 2005. The Syrian conflict has come to be seen in Lebanon through a highly sectarian lens, broadly pitting Sunnis versus Shi‘a, with Christians divided on the issue (but largely preferring to be uninvolved). Sunnis have largely supported the opposition, while Shi‘a have primarily sided with the Assad regime. Both sides have provided support in many forms, including sending fighters and financial support to the other side of the border. Lebanese Sunni involvement in Syria was initially claimed to be limited to a few radical groups, but subsequent developments suggested that Sunni involvement has likely been far deeper than previously believed (Asfura-Heim et al., 2013). Christians, divided between the two main blocs

after syria 83 (the primarily Sunni March 14 and the primarily Shi‘a March 8;9 see Salamey and Tabar, 2012), never took a united stance either for or against the regime. Nevertheless, the Syrian conflict has heavily influenced political discussions for each and every sect. The sects’ opposite positions in Syria have translated into political discord in Lebanon; prominent Maronite MP Alain Aoun has remarked that “since the Syrian crisis began, in 2011, Lebanon’s Sunnis and Shi’ites have been unable to speak to each other” (quoted in Muhanna, 2014). The pro/anti-Syrian cleavage in Lebanon can be traced back at least to the establishment of Lebanon as an independent state, but it took on new importance with the outbreak of the civil war in Syria and its subsequent sectarianization.10 As the uprising in Syria transformed into a full-blown war, several of the major political parties in Lebanon sided with either the regime (especially Hezbollah) or the opposition (especially the Future Movement). Since each of these parties, like most major parties in Lebanon, draws support heavily from a single sectarian group, their decision to take sides in the conflict had the downstream effect of “sectarianizing” the conflict from the Lebanese perspective, matching the growing emphasis on sectarian motivations across the border. As Salloukh (2017, p. 69) notes, “The spillover effects of the Syrian war on Lebanon had the most damaging effect on the country’s sectarian relations. The anti- and pro-regime mobilization along largely Sunni-Shia fault lines led to an increase in conflict and agitation.” Thus, while the Sunni-Shi‘a divide in Lebanon is certainly not new, the Syrian conflict “put the final seal on this divide” (Abdo, 2017, p. 100). The spillover of the Syrian Civil War into Lebanon must also be considered in the context of broader regional developments during the same period. The Arab uprisings, of which the Syrian conflict is one case, changed the tone of sectarian discourse in much of the Middle East. The division between Sunnis and Shi‘a in the region has “assumed a new urgency after the Arab Spring” (Salloukh et al., 2015, p. 155). Scholars have documented a rise in the usage of anti-Sunni and anti-Shi‘a hate speech since the beginning of the Syrian Civil War, a trend that has spread throughout the region (Siegel, 2017). Narratives of an Iran-led “Shi‘a coalition” versus a competing Sunni alliance led by Saudi Arabia, though certainly an oversimplification, have become abundant in the region, with Lebanon being no exception. Al-Akhbar (2014) reported that even the drawn-out Lebanese presidential election of 2014–2016, seemingly a domestic issue, was dictated heavily by competing visions of Lebanon’s role in the regional system; Sunnis (typically allied with Saudi Arabia) versus Shi‘a (typically allied with Iran). Thus, the heightening

 faith in numbers of Sunni-Shi‘a tensions in Lebanon has not happened in isolation, but rather as a part of broader regional trends11 in addition to domestic developments. A final feature of the Arab uprisings that has affected Lebanon’s sectarian landscape is the spread of jihadist groups in parts of the region. Prior to the uprisings, jihadism in Lebanon was often perceived to be “more of a nuisance than a problem” (Zelin, 2016, p. 51). However, the collapse of authoritarian regimes in parts of the region and, especially, the spread of conflict in Syria and Iraq opened the door for (mainly) Sunni jihadist groups, including ISIS, to emerge as a more serious threat. Such groups have fought actively against Hezbollah in Syria, and have intensified sectarian hostility elsewhere. While these groups certainly do not represent mainstream Sunni politics (and are relatively small in size), they nevertheless create a security dilemma in which intersectarian trust is compromised. Though not the main factor behind rising Sunni-Shi‘a conflict in Lebanon and the broader region, the increased prominence of these groups has nevertheless added a layer of sectarian tension. These changing circumstances suggest an important distinction from the previous political environment: as a result of several local and regional developments since the late 2000s, Sunnis and Shi‘a find their interests to be in direct conflict, in contrast to the previous alignment of interests in significant ways. Moreover, these tensions have taken on a new tone: as Bahout (2017, p. 152) observes, as the Syrian Civil War has drawn on, “The mechanisms of political identification have taken on a more existential dimension, characterized be a zero-sum approach to politics.” When presented in 2014 with the statement “Political differences between Muslims and Christians are stronger than differences between Sunnis and Shi‘a,”12 only a small minority agreed (less than 14% of Sunnis and just over 17% of Shi‘a; even Christians agreed with this statement at a rate of less than 27%). Christians, instead of being united in their opposition to electoral reform and redistribution, are divided on the key issue of the day, being split roughly evenly between pro-Syrian regime and anti-Syrian regime camps. While Christians still have reasons to oppose electoral reform and redistribution, the prospect of democratization per se is now less threatening, since Christian groups could be potentially very powerful coalition members in the absence of an outright majority. Al-Monitor (2014) quotes a Christian political leader in a border area as stating that since the Syrian conflict has erupted, “Everybody wants the Christians to be on their side.” By being a seemingly neutral group with respect to the conflict, Christians can

after syria 85 potentially ally with either side if political conditions change (see Hazran 2010 for a discussion of how Shi‘a leaders have altered their behavior based on the knowledge that an intersectarian coalition could overpower them). The experience of the Christian community in the post–Syrian Civil War environment provides an illustration of the potentially important role of leadership. The Christian religious establishment—most notably, the Maronite patriarchy—has frequently attempted to serve as a neutral arbiter bringing competing Christian parties together. These attempts have been largely unsuccessful. It is possible that these efforts have failed because the Maronite patriarchy as well as other institutions of religious leadership are not as powerful as they are often claimed to be; Henley (2017) argues that Lebanese religious leaders do not generally possess the amount of influence that they are typically presumed to hold. It is likely that the influence of sectarian political leaders overwhelms the effects of purely religious institutions. The Christian experience in this setting is a case in point. While in 2013–2014, Shi‘a political elites were largely united in their support of the Syrian regime and the Sunni elite was largely (though certainly not to the extent of the Shi‘a leadership) united in its opposition to the regime, Christian parties stood firmly on opposite sides. Leaders like the (Christian) Free Patriotic Movement’s Michel Aoun, once a harsh critic of the Assad regime, lined up with the Shi‘a Hezbollah-Amal alliance, while the Lebanese Forces’ Samir Geagea supported the regime’s opponents, drawn heavily from the Sunni community.13 These factors create a situation in which Sunnis and Shi‘a view the prospects of democratization through a Sunni-Shi‘a lens rather than a Christian-Muslim one. Consequently, the incentives of each group with respect to democracy are determined in large part by these alignments, reflecting a shift in the axis of political-sectarian competition. Although Sunnis are still underrepresented in parliament, they are significantly less underrepresented than Shi‘a, their current political adversaries.14 Although electoral reform in the direction of a more proportional system would benefit Sunnis in absolute terms (and relative to Christians), it would lessen their influence relative to Shi‘a. As a result, Sunni discourse has shifted away from electoral reform considerably. Ahmed Fatfat, a leading Sunni member of the Future Movement’s parliamentary bloc, now says that “Ta’if is a very good equilibrium,” referring to the 1989 Ta’if Agreement that reassigned parliamentary seats to increase the number of Muslim seats in parliament but still preserves the underrepresentation of Muslims. Further, he remarks

 faith in numbers that “we need to keep it [Ta’if] even though it can be non-democratic.”15 These comments reflect remarkable candor about the new incentives faced by Sunnis. Realizing that electoral reform would benefit Shi‘a at the expense of Sunnis, Sunni leaders have come to accept the current arrangement as being superior to the more democratic alternatives. Shi‘a are well aware of the implications of electoral reform under the new structure of political alliances and hostilities. As mentioned previously, political concerns have prevented Lebanon from conducting a census since 1932, but current estimates suggest that the Shi‘a are now more than likely the largest group in the country; common estimates place their proportion at 40% of the total population (Hazran, 2009, pp. 2–3; Shanahan, 2005, p. 118); some estimates place this proportion as high as 55% (Ahmed and Salas, 2005, p. 237). A proportional electoral system, therefore, would tend to benefit Shi‘a substantially, particularly relative to their current allotment of just 27% of seats in parliament. Lebanon’s voting age (21) also contributes to Shi‘a underrepresentation,16 since Shi‘a birth rates are higher and the proportion of their population under the age of 21 is larger than those of other sects (see Hazran, 2009); the largely Shi‘a Amal Movement estimates that 70% of citizens between the ages of 18 and 21 are Shi‘a (Shanahan, 2005, p. 121). In this sense, the Shi‘a are the underrepresented and underprivileged group and the group who would benefit the most from democratic reforms. Mohammed Obeid, a key figure in the Amal Movement and longtime advisor to Speaker of Parliament Nabih Berri, declares: “Let’s hope a miracle happens and they adopt a proportional law.”17 Economic incentives are also likely to lead to a pro-democracy effect of communal prayer among Shi‘a but the opposite among Sunnis. While Shi‘a have, without question, improved their economic standing in Lebanon in recent years, their economic opportunities are still limited compared to other sects, and their access to state resources is particularly lacking. Lebanon continues to be an immensely unequal country. As demonstrated in the previous chapter, Lebanese government spending indicates a remarkable lack of emphasis on, among other inequality-reducing policy areas,18 education. Greater access to public education for poorer citizens (and, by extension, poorer sects)19 would more than likely promote the development of skills that would enable such citizens to improve their economic standing. Since education is a key factor in upward mobility, the government’s inattention to the education sector reflects a lack of interest in reducing inequality, at least compared to other interests.

after syria 87 Much of Lebanon’s economic inequality follows sectarian lines. Since income is a less-than-ideal measure20 of group privilege in this setting, the survey uses a variety of other measures to capture this concept more accurately. A question about homeownership helps to distinguish “wealth” from income. Shi‘a relative deprivation is clear from this measure; compared to Sunnis, Shi‘a were about 12 percentage points less likely to own their homes. Of course, it is possible that this difference is simply due to demographic differences (specifically, average age) or settlement patterns. To address this possibility, I estimate the difference between groups while controlling for the respondent’s age and his/her qadaa21 (district). After including these controls, the difference between Sunnis and Shi‘a is actually larger (over 17 percentage points) and remains highly statistically significant (p-value < .001). It is important to note that wealth or poverty, as measured in the simplest ways, does not quite capture the most important dimensions of sectarian inequality for the purposes of this study. In a context of state weakness and neglect, it is not necessarily the case that poorer groups will always benefit more from redistribution than (relatively) wealthier groups. The relevant factor with respect to incentives toward democratic redistribution is access to state resources. For example, in the Lebanese case, historically impoverished Shi‘a areas have experienced considerable economic growth in the past two decades, largely due to the patronage of either nonstate actors or foreign actors (including the Iranian regime). The relative improvements in Shi‘a standards of living in much of the country are not an indication that the state has begun to devote more attention to these areas, nor are they a reason for Shi‘a to oppose (presumed) democratic redistribution. If the state were to pursue more activist redistributive policies, such policies would still benefit Shi‘a areas because they remain disproportionately overlooked by the state. Since redistribution is far more likely to take the form of infrastructure investments (roads, power grids, etc.) or human capital investments (like education), Shi‘a areas would still have quite a bit to gain from such policies despite their internally driven growth in recent years. Thus, poverty levels are only part of the story. In terms of the share of the population falling under the poverty line, the heavily Sunni North Governorate (mohafaza) is actually measured by some sources22 as the poorest region23 (followed by the mainly Shi‘a South and Bekaa Governorates, with the other regions reporting much lower poverty rates than any of these three). This region—which in the World Bank’s report (2009) includes the

 faith in numbers northernmost district of Akkar,24 , now its own mohafaza—has experienced high levels of poverty as well as government neglect in its own right, and has been rightfully described as “one of the most deprived regions” in Lebanon (UNHCR, 2015).25 Other sources point to Shi‘a regions as experiencing higher levels of poverty: Salti and Chaaban (2010, p. 642), for instance, report poverty rates of 38% in Bekaa, 37% in South Lebanon, and 46% in Nabatieh, compared to 31% in Sunni North Lebanon. However, simple poverty measures do not necessarily map linearly onto incentives toward government redistribution, especially in states that fail in the provision of public goods. For the purposes of assessing sectarian and regional incentives with respect to redistributive democracy, access to state-provided resources (or, at least, resources that should be provided by the state) is of particular importance. A frequently used measure of poverty in Lebanon is the number of hours per day in which an individual’s electricity is cut off (Corstange, 2012a, 2013). This measure is particularly useful because it captures an individual’s access to a resource that is supposed to be provided by the state. Shi‘a respondents experienced an average of about 37 more minutes without power per day than Sunnis.26 These differences are consistent with the findings of the World Bank’s study of regional inequalities in access to electricity, which it describes as “a clear case of inequitable distribution” (World Bank, 2009, p. 8). These inequities have forced households to fill in the gap left by the state, again falling noticeably along sectarian lines: nearly 80% of households in the mainly Shi‘a Bekaa Governorate use private generators to provide electricity, a difference of over 20 percentage points compared to the Sunni-majority North/Akkar Governorates, where electricity is on average available for an additional three hours per day compared to Bekaa (World Bank, 2009, pp. 9–10). The mostly Sunni northern governorates also have greater access to public network water on a daily basis (59% of households) compared to Bekaa (43%), South Lebanon (38%), and Nabatieh (20%).27 Further evidence of unequal state privileges is found in the area of employment. While the unemployment rate for Shi‘a was only slightly higher than that for Sunnis (a difference of 1.5 percentage points), the level of public sector employment differed vastly between the groups. While over 13% of Sunni respondents reported working in the public sector, less than 6% of Shi‘a respondents reported doing so. According to UNDP estimates (2015, p. 30), public sector employment rates are particularly high in Sunni North Lebanon, perhaps exceeding 20%, a figure much higher than in

after syria 89 South Lebanon or Nabatieh. These figures suggest a clear trend: to the extent that the Shi‘a community in Lebanon has enhanced its economic standing in recent years, it has done so with considerably less access to state largesse than its Sunni neighbors.28 Democratization—including, importantly, more proportional and equitable access to state resources— would therefore tend to benefit Shi‘a at the expense of other groups, both politically and economically.

Testing the Theory My theory of communal religion, group interest, and democracy as applied to the changing political environment in Lebanon suggests a new set of hypotheses. Broadly, we should expect that citizens who attend religious services frequently should be more likely to hold political attitudes in alignment with the interests of their sect than those who do not attend regularly. Hypothesis 3. Communal religious practice will tend to increase support for democracy among Shi‘a, decrease support for democracy among Sunnis, and have no significant effect on support for democracy among Maronites.29 To test this hypothesis, data are drawn from an original nationally representative survey in Lebanon conducted between late 2013 and early 2014.30 The survey used a multistage area probability sample covering all 26 of Lebanon’s districts in order to ensure proper representation. Surveys were conducted face to face using traditional pen-and-paper methods. The response rate for this survey was over 85%. The sample was stratified by sect such that 402 respondents were selected from each of the main sectarian groups: Christian,31 Sunni, and Shi‘a. The key dependent variable (support for democracy) is an index combining several different questions about attitudes toward regime types. For the observational section, this variable combines nine different items, all related to support for democracy; the full list of questions is available in appendix 5B. The dependent variables are rescaled to range from 0 to 1 for ease of interpretation. The key independent variable uses a question asking respondents how frequently they attended church or mosque.32 Variables included in the “baseline” set of controls are gender, age, education

 faith in numbers level, employment status, and income. Gender is perhaps a particularly important control variable since, as shown above, men and women differ considerably in their propensity to attend mosque.33 Subsequent models include measures of religiosity (“Are you a religious person?”) and qadaa (district) fixed effects. Importantly, later models also control for personal prayer.34 Inclusion of this variable should account for the possibility that communal prayer is not meaningfully different from other types of prayer.

Communal Prayer and Support for Democracy At first glance, the differences between sects in terms of support for democracy are minimal. After controlling for the demographic variables described above, each of the three main groups demonstrates similar levels of support for democracy (.62 for Maronites, .59 for Sunnis, and .61 for Shi‘a on a 0–1 scale). However, when taking into account the effects of communal worship, stark differences emerge within the Sunni and Shi‘a groups as predicted by the theory. As a first test of the sect-specific relationship between communal prayer and attitudes toward democracy, Figure 5.1 displays the mean levels of support for democracy by sect and frequency of religious attendance. These averages are observed from the data rather than being estimated from a model. The observed values generally support the hypotheses so far presented: for Sunnis, support for democracy decreases with more frequent attendance; for Shi‘a, support for democracy increases with attendance; and for Maronite Christians, no clear pattern is present. The Sunni and Shi‘a trends are both close to being monotonic, and the general trend is clear: without controlling for other variables, frequency of communal prayer is associated with democratic attitudes in closer alignment with sectarian interests for Muslims, while no trend is present for Maronites. Figure 5.1 displays the bivariate relationship between frequency of religious attendance and support for democracy within each of the three main sects. Table 5.1a presents the results of a series of ordinary least squares regressions of support for democracy on communal religious practice for Sunnis (Table 5.1aa), Shi‘a (Table 5.1ab), and Maronite Christians (Table 5.1ac). In each of these tables, the first five models use the continuous measure of frequency of communal prayer as described, while models 6–10 use indicator variables for each level of attendance in order to determine if there are meaningful nonlinearities in this relationship. Models 1 and 6 are

Average Support for Democracy (0−1 Scale)

after syria 91 .8 Maronite

.6

.6

Sunni

.64

.63

.58 .58 .51

Shi`a

.66 .61

.61 .59

.58

.65 .65

.53 .47

.4

.2

Ar

ou N n ev Ar d on e ou ce Ra r M r n p e or d o er l et m y ha nce on p n t on er h ce we e pe k rw ee k Ar ou N n ev Ar d on e ou ce Ra r M r n p or d o er ely et m ha nce on pe th n on r ce we pe ek rw ee k Ar ou N nd ev Ar on e ou ce Ra r M pe re n d l or e t on r m y ha ce on t p n on er h ce we pe ek rw ee k

0

Frequency of Communal Prayer

Fig. 5.1 Support for Democracy, by Sect and Attendance (Observed).

bivariate regressions that include no controls; the other models all include “baseline” controls as described previously. Models 3 and 8 control for selfreported religiosity; Models 4 and 9 control for personal prayer, and Models 5 and 10 include qadaa fixed effects.35 Additional tests (omitted here for the purposes of space) included a measure of political knowledge36 as a control to address the possibility that religious attenders may simply know more (or less) about politics, but this measure was never a significant predictor of support for democracy and never affected the substantive interpretation of the effects highlighted here. Since these results are from OLS models and both the dependent variable and the main independent variable are scaled to range from 0 to 1, the coefficients in the table can be interpreted as marginal effects or percentagepoint changes. For Sunnis, a one-unit increase in frequency of communal prayer (corresponding to an increase from “never” attends to “attends more than once a week”) is predicted to decrease support for democracy by between 15 and 22 percentage points; for Shi‘a, the corresponding effect is an increase of 9 to 11 percentage points. These effects are sizable, highly

402 .114 −249.98

(.03)

(.03)

402 .141 −252.15

X

−.17∗∗∗

−0.18∗∗∗

Note: Standard errors in parentheses. ∗ p < 0.1, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Baseline Controls Religiosity Prayer Qadaa FEs Observations R2 AIC

> Once/week

∼Once/week

∼Once/month

Rarely

Frequency

(2)

(1)

Table 5.1a Baseline Results

402 .147 −253.24

X X

(.03)

−.20∗∗∗

402 .157 −257.84

X

X

(.03)

−.22∗∗∗

X 402 .234 −258.35

X

(.03)

−.15∗∗∗

402 .141 −256.47

.03 (.03) −.01 (.05) −.04 (.03) −.15∗∗∗ (.03)

402 .166 −258.36

X

.04 (.03) −.02 (.05) −.04 (.03) −.14∗∗∗ (.03)

Communal Prayer and Support for Democracy (Sunnis) (3) (4) (5) (6) (7)

402 .167 −256.55

X X

.03 (.04) −.03 (.05) −.05 (.04) −.15∗∗∗ (.04)

(8)

402 .169 −257.79

X

X

.01 (.04) −.05 (.05) −.08∗ (.04) −.18∗∗∗ (.05)

(9)

X 402 0.261 −266.59

X

0.03 (0.03) 0.01 (0.05) −0.02 (0.03) −0.13∗∗∗ (0.03)

(10)

0.10∗∗∗ (0.02)

(1)

X X 402 0.048 −349.13

402 0.046 −350.48

0.09∗∗∗ (0.03)

402 0.048 −349.36

X

X

0.09∗∗∗ (0.03)

X 402 0.177 −369.67

X

0.11∗∗∗ (0.02)

402 0.058 −359.43

0.08∗∗∗ (0.02) 0.07∗ (0.04) 0.12∗∗∗ (0.03) 0.12∗∗∗ (0.03)

402 0.069 −354.00

X

0.09∗∗∗ (0.02) 0.07∗ (0.04) 0.13∗∗∗ (0.03) 0.13∗∗∗ (0.03)

Communal Prayer and Support for Democracy (Shi‘a) (3) (4) (5) (6) (7)

X

0.10∗∗∗ (0.02)

(2)

Note: Standard errors in parentheses. ∗ p < 0.1, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Baseline Controls Religiosity Prayer Qadaa FEs Observations 402 R2 0.040 AIC −357.69

> Once/week

∼Once/week

∼Once/month

Rarely

Frequency

Table 5.1b

402 0.070 −352.52

X X

0.10∗∗∗ (0.03) 0.08∗ (0.04) 0.14∗∗∗ (0.03) 0.14∗∗∗ (0.03)

(8)

402 0.076 −355.43

X

X

0.14∗∗∗ (0.04) 0.11∗∗ (0.05) 0.18∗∗∗ (0.04) 0.18∗∗∗ (0.04)

(9)

X 402 0.191 −370.56

X

0.08∗∗∗ (0.02) 0.08∗∗ (0.04) 0.12∗∗∗ (0.03) 0.13∗∗∗ (0.03)

(10)

222 0.006 −211.88

0.05 (0.04)

(1)

Note: Standard errors in parentheses. ∗ p < 0.1, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Baseline Controls Religiosity Prayer Qadaa FEs Observations R2 AIC

>Once/week

∼Once/week

∼Once/month

Rarely

Frequency

Table 5.1c

222 0.059 −214.03

X

0.06 (0.04)

(2)

222 0.061 −212.51

X X

0.05 (0.05)

222 0.064 −213.30

X

X

(0.04)

0.08∗

X 222 0.384 −258.33

X

0.03 (0.04)

222 0.072 −221.10

−0.02 (0.06) −0.02 (0.06) 0.04 (0.05) −0.09 (0.06)

222 0.106 −219.41

X

0.00 (0.06) −0.00 (0.06) 0.07 (0.05) −0.04 (0.07)

Communal Prayer and Support for Democracy (Maronites) (3) (4) (5) (6) (7)

222 0.106 −217.45

X X

0.00 (0.06) −0.00 (0.06) 0.06 (0.06) −0.05 (0.07)

(8)

222 0.107 −217.81

X

X

0.01 (0.06) 0.01 (0.06) 0.08 (0.06) −0.03 (0.07)

(9)

X 222 0.412 −262.34

X

−0.01 (0.05) −0.01 (0.06) 0.04 (0.05) −0.05 (0.06)

(10)

after syria 95 statistically significant (at the .01 level or better in every specification), and very stable across different model specifications. For both sects, frequency of attendance has a mostly monotonic association with democratic attitudes, as illustrated in models 6–10 in each table: frequent communal religious practice is robustly negatively correlated with support for democracy among Sunnis, but positively linked to support for democracy among Shi‘a. As expected, results for Maronite Christians are unclear and inconsistent; the political divisions within the political community over the issue of Syria make it difficult to identify what their “group interest” might be.37 In total, these results provide strong support for Hypothesis 3.

Alternative Hypotheses and Threats to Inference The analysis presented so far does not rule out the possibility of alternative explanations. One major competing explanation for the relationships identified is that religious attenders may simply be more tapped into clientelistic/distributive networks (Cammett and Issar, 2010; Cammett, 2015) or may be more politically engaged38 for nonsectarian reasons. This is particularly plausible in cases like Lebanon, where access to economic and social goods is closely linked to partisan-sectarian commitment (Corstange, 2017); in the Lebanese case, Cammett (2011, p. 70) finds that “political activism and a demonstrated commitment to a party are associated with access to social assistance.” It is possible, therefore, that the association between communal religious practice and sect-centric attitudes is due to the fact that religious attenders are more tightly embedded into these networks, and perhaps even select into communal prayer because of the material benefits that these networks offer (cf. Grzymala-Busse, 2012a). If this is the case, then connections to key political actors—most clearly, political parties—could be a key omitted variable. However, the evidence suggests that this is not the case. Refitting the baseline models reported in Table 5.1a and including controls for party membership have no appreciable effect on the size or statistical significance of communal prayer for any group. For both Sunnis and Shi‘a, the coefficient on communal prayer remains statistically significant at the p < .001 level when including party fixed effects, and the magnitude of the effect of communal prayer is actually slightly higher when partisanship is included as a control variable. Moreover, partisan effects themselves seem to be minimal:

 faith in numbers the only party affiliation that has any detectable association with support for democracy in either group is Sunni membership in Al-Jama‘ah Al-Islamiyah (The Islamic Group), whose members are marginally (p = .072) less supportive of democracy. Similar results are obtained when controlling for membership in charitable organizations, professional associations or trade unions, family or tribal associations, or unspecified civil society organizations. A related question concerns the role of religious leadership: it is possible that the effects presented here are mediated entirely by interactions with religious leaders. Again, however, this appears not to be the case. My survey asked respondents how frequently (if ever) they consulted religious leaders on political matters.39 Relatively few respondents reported consulting religious leaders frequently with regard to political issues: 75% reported that they never do so. Still, it is possible that this may be an important omitted variable for the subset of the population who does consult religious leaders on political matters. This does not appear to be a major threat to inference, however. Including this question as a control variable in the main model presented above does not eliminate the relationship between communal prayer and support for democracy despite the fact that it is correlated with communal prayer. For Sunnis, the effect of communal prayer is attenuated somewhat (the coefficient estimate decreases from −.17 to −.10), but the effect is still highly statistically significant. For Shi‘a, the coefficient estimate actually increases (from .10 to .12) when including this control, and the effect is even more statistically significant when this control is added. Thus, while engagement with religious leadership may play a part in determining attitudes toward democracy, it does not account for the effect of communal prayer on regime preferences. As described in chapter 5, my theory proposing that communal religious practice heightens the salience of sectarian identity and, consequently, pushes political preferences in a direction determined by sectarian interests, runs into a significant challenge: it may simply be the case that more “sectarian” individuals are more likely to participate in communal prayer precisely because they are more sectarian. If this is the case, then the causal arrow actually points in the opposite direction than my theory suggests. This very real possibility plagues much of the literature on religion and politics that relies on observational data alone, and is worthy of considerable attention. The remainder of this chapter uses experimental and quasi-experimental techniques to approximate the causal effect of communal prayer.

after syria 97 While it is impossible to “exogenize” communal prayer fully,40 I embedded a religious priming experiment in a nationally representative survey of Lebanon that aims to (unobtrusively) mimic the types of cues that individuals might hear in a communal worship environment. The goal of this type of experiment is to activate the types of religious/sectarian concepts that might be activated in a worship setting, randomly assigning respondents to a treatment group in order to eliminate the possibility of endogeneity. In this experiment, respondents were randomly “primed” with either religious or neutral concepts. Respondents were asked to choose the closest synonym to a handful of terms, either representing religious concepts (in the treatment group) or no clear theme (in the control group). These are a form of a “questions-as-treatment” experiment; the answers to the questions did not matter; the act of answering the questions was simply meant to prime a (communal) religious mindset. Of course, these primes cannot perfectly mirror the type of content that individuals experience at a church or mosque (particularly since this content will vary across both place of worship and time), but they nevertheless should invoke a certain type of “religious” mindset in respondents. Wenger (2004) states that these types of primes can trigger an “automatic” activation of religious concepts, and similar priming experiments have proven successful in studying the effect of religion (in various forms) on a number of different outcomes.41 In the treatment group, respondents were presented with the following terms: “sect,” “man of religion,”42 “shrine,” “sermon,” and “place of worship.” In the control group, where no particular concept was intended to be primed, respondents were presented with the terms “delicious,” “sleep,” “newspaper,” “work,” and “shirt.” Each respondent was then asked to choose the closest synonym to each word, aiming to invoke a communal religious mindset in the treatment group and no particular mindset in the control group. A full description of the experiment is available in appendix 5A, but it is important to note that the treatment was carefully designed to use communal religious words, rather than words that might activate notions of personal piety. The expectations for this experiment are analogous to those in the observational portion of the results; communal religious primes should shift respondents’ attitudes toward the interests of their group, much like communal prayer. Hypothesis 4. Communal religious primes will increase support for democracy among Shi‘a but decrease support for democracy among Sunnis.

 faith in numbers .1

Average Treatment Effect

0.5

0

–0.5

–1 Sunni

Shi‘a Sect

Fig. 5.2 Average Treatment Effects (Support for Democracy). Note: Figure displays the average treatment effect of the “communal” religious prime on support for democracy (scaled 0–1), with 95% confidence intervals.

Figure 5.2 displays the average treatment effects (by sect) of the communal religious primes on support for democracy.43 Among Sunnis, these primes decreased support for democracy by over six percentage points compared to the control group, while for Shi‘a, the prime increased support for democracy by about five percentage points. Both of these effects are statistically significant at the 5% level. It appears, therefore, that these primes had an effect similar to that demonstrated by mosque attendance; for Sunnis, communal religion in either form reduces support for democracy, while the opposite is true for Shi‘a. Importantly, these results are isolated to the communal primes; personal piety primes (as described in appendix 5A) had no such effects on any of the outcome variables.44

Causal Mechanisms I have suggested that the mechanism through which communal prayer translates into support for (or opposition to) democracy is as follows: mosque attendance heightens the salience of group identity, which in turn induces individuals to view political issues (e.g., electoral reform or redistribution) through a more sectarian lens, which influences their regime preferences

after syria 99 accordingly. While testing such mechanisms is difficult, particularly with observational data, it is nevertheless possible to examine the degree to which mosque attendance correlates with the expected mechanism outcomes as a “plausibility probe” (George and Bennett, 2005). To that end, I fit a series of regressions45 similar to the ones described previously (using the “baseline” control set), but with the outcome being more specific issue attitudes or measures of group affect rather than broader measures of regime attitudes. Communal prayer should increase an individual’s sense of “linked fate” (Dawson, 1995), making citizens more likely to believe that their fortunes are closely tied to those of their sect. To measure this concept, the survey asked respondents the degree to which they agreed that “policies that are good for my sect are good for me personally.”46 If the linked fate mechanism is at work, then members of either sect should be more likely to agree with this statement if they attend religious services regularly. This relationship is presented in Figure 5.3. Patterns are very similar between the two groups: 32%–33% of “never” attenders agree with the statement, compared to 54%–55% of the most frequent attenders. These relationships, significant at the .001 level, suggest that mosque attenders are indeed more likely to perceive a link between their own fates and the political fates of their sects. Here, again, the issue of party affiliation is a potential threat: it may be the case that highly “sectarian” individuals select into communal prayer because of the access it provides to welfare networks; if this is the case, then the effect of communal prayer is simply an artifact of a sectarianismparty affiliation link. However, when fitting these models of linked fate and adding party fixed effects as a control, the effects of communal prayer remain the same for both Sunnis and Shi‘a. In this case—unlike the findings for support for democracy—some direct party effects are noticeable: compared to unaffiliated Sunnis, members of the Future Movement and Al-Jama‘ah Al-Islamiyah (The Islamic Group) report somewhat higher levels of linked fate, while members of the Amal Movement are more likely than others to report linked fate among Shi‘a. Irrespective of individual partisan effects, the inclusion of partisan controls makes no appreciable difference in the relationship between communal prayer and linked fate, suggesting that the association between religious attendance and sectarian identification is not simply due to partisan attachments. As was the case with support for democracy, controlling for membership in charitable organizations, professional associations or trade unions, family or tribal associations, or unspecified civil society organizations does not affect the results, further bolstering the claim

 faith in numbers Sunnis

Pr (Agree)

.8 .6 .4 .2 Never

Rarely

Around once per month

Around once per week

More than once per week

Around once per week

More than once per week

Shi`a

Pr (Agree)

.8 .6 .4 .2 Never

Rarely

Around once per month

Frequency of Mosque Attendance

Fig. 5.3 Frequency of Communal Prayer and Linked Fate.

that religious attendance enhances sectarian identification independently from connections to other associations. It is possible, of course, that these results do not indicate causal mechanisms, but rather selection effects, in accordance with the endogeneity issue previously described. If high-identifiers are simply more likely to attend religious services, then my claim that mosque attendance enhances sectarian identification may not be true. To address this possibility, I use a “tough test” of the effect of mosque attendance on sectarian identity. If mosque attendance truly increases the salience of sectarian identity, then respondents interviewed on Friday (the Muslim day of communal prayer) should report higher levels of sectarian identification than those interviewed on other days.47 The logic of this test is as follows: if the experience of communal prayer heightens sectarian identification, then this effect should be noticeable on Fridays, the most common day for engaging in communal prayer and thus the day on which mosque attendance is the most “communal.” Furthermore, this effect should be isolated to attenders. Nonattenders provide a type of placebo test for this hypothesis: if respondents who do not participate in communal prayer report higher levels of sectarian

after syria 101 identification on Fridays than on other days, then the “Friday” effect is most likely due to factors other than mosque attendance, such as the mere awareness that it is Friday. Since the day of the week on which a respondent is interviewed is plausibly quasi-random, the Friday versus nonFriday comparison avoids the problems of endogeneity faced by the other observational analyses in this book. Non-attenders serve as a control group, not being exposed to the “treatment” (communal prayer) regardless of the day of the week on which they were interviewed. These expectations are summarized in Hypothesis 5: Hypothesis 5. Among mosque-attenders, respondents interviewed on Fridays will report higher levels of sectarian identification than attenders interviewed on other days. Among non-attenders, there should be no difference between respondents interviewed on Fridays versus other days. Table 5.2 displays the proportions of high-identifiers comparing Friday respondents to others, conditional on whether or not the respondent attends mosque services weekly. These proportions provide strong support for Hypothesis 5: among weekly attenders, respondents who were interviewed on Fridays were substantially more likely (a 12 percentage point difference) to report strong sectarian identification. Among non-attenders, no meaningful difference was found between respondents interviewed on Friday and those interviewed on other days; the minuscule difference that is present is actually in the opposite direction. These results indicate that Friday prayer has a substantial influence on sectarian identification among attenders. To ensure that these differences are not due to the fact that certain areas and types of respondents were more likely to be surveyed at particular times, I fit a logistic regression using the same dependent variable as in Figure 5.3 Table 5.2 Group Identification, Friday versus Other Days (a) Non-Attenders

Not Friday Friday Total

(b) Weekly Attenders

Low

High

Total

213 59% 69 60% 282 59%

150 41% 46 40% 196 41%

363 100% 115 100% 478 100%

Not Friday Friday Total

Low

High

Total

102 40% 19 28% 121 37%

156 60% 49 72% 205 63%

258 100% 68 100% 326 100%

 faith in numbers as well as the baseline set of covariates. To these covariates, I add a few additional controls to account for the possibility that the day of interview might not be independent of attendance or identification: governorate (mohafaza) indicators, week of interview indicators, and a sect indicator.48 Figure 5.4 displays the marginal effects of being interviewed on a Friday. The y-axis can be interpreted as the difference in the predicted probability of high sectarian identification between respondents interviewed on a Friday and respondents interviewed on any other day of the week. This figure indicates a substantial effect: among weekly attenders, Friday-interviewees were 14 percentage points more likely to report strong perceptions of linked fate compared to non-Friday-interviewees, and this difference has a p-value of .02. Among non-attenders, there is no meaningful difference between Friday-interviewees and others. These differences suggest that the act of attending mosque increases sectarian identification. A few caveats are necessary for interpreting these results. First, like the priming effects described previously, the “Friday” effect is inevitably shortlived. If the entire effect of mosque attendance survived long term, then there would be no detectable “Friday” effect because attenders would have built up a large stock of sectarian identification already. However, the observational .3

Marginal Effect (Friday)

.2

.1

0

–.1 Weekly Attendance

No Weekly Attendance Frequency of Attendance

Fig. 5.4 “Friday” Effect, by Weekly Attendance.

after syria 103 results pertaining to frequency of attendance indicate that long-term effects are present as well. The evidence suggests that there are both immediate and long-lasting effects of mosque attendance on sectarian identification. Much like certain medicines that require long-term administration to have their full effect but also provide an immediate “jolt” to the patient when they are administered, religious attendance enhances sectarian identification through both an immediate bump-up effect and a long-run accumulation effect. Second, the results presented in Figure 5.4 must be interpreted as average treatment effects on the treated (ATT) rather than overall average treatment effects. For non-attenders, we do not (and cannot) know the effect of the mosque attendance treatment precisely because they do not attend mosque. Their presence is useful in the model because it ensures that the “Friday” effect is not due to some other factor besides mosque attendance; however, it is not possible to determine how they would be affected by mosque attendance without highly questionable assumptions.49 . Nevertheless, the ATT is a highly useful quantity for this study. The fact that attenders are 14 percentage points more likely to report high sectarian identification on Fridays compared to other days is remarkable, and indicates that the act of mosque attendance has a sizable influence on the salience of sectarian identity among regular attenders. As a final test to address the possibility of self-selection, it is useful to consider the effect of the communal religious primes described previously on perceptions of linked fate. Using the same linked fate question, Figure 5.5 demonstrates that the “communal” primes increased respondents’ sense of linked fate, even though these questions were asked after the section on democracy, potentially allowing the effect of the primes to wear off. Among Sunnis, the communal treatment increased perceptions of linked fate by eight percentage points; the effect size was 5.4 percentage points among Shi‘a. The Sunni effect is significant at the .05 level, and the Shi‘a effect at the .1 level. These effect sizes and significance levels are fairly substantial considering how soft the primes were as well as the time in between the primes and the linked fate question. It appears, therefore, that these communal primes increased a sense of linked fate in a similar way as would communal practice, albeit on a smaller scale. It is also important to consider how these linked fate perceptions influence beliefs about regime types. For this purpose, the survey asked respondents how much they agree that “democracy is good for people like me.” As

 faith in numbers .2

Average Treatment Effect

.15

.1

.05

0

–.0.5

–.1 Sunni

Shi‘a Sect

Fig. 5.5 Average Treatment Effects (Linked Fate). Note: Figure displays the average treatment effect of the “communal” religious prime on support for democracy (scaled 0–1), with 95% confidence intervals.

expected, the effect of communal practice on this outcome diverges between the two groups; this relationship is depicted in Figure 5.6. Among Sunnis, 82% of non-attenders strongly believe that democracy is good for people like them, compared to only 43% of the most frequent attenders. Among Shi‘a, the opposite is true: less than 50% of non-attenders strongly agreed with the statement, compared to over 82% of very frequent attenders. In each of these models, the coefficient on the communal practice variable is significant at the .001 level. There is evidence, therefore, that the “linked fate” perceptions created by mosque attendance translate into political attitudes—by attending religious services, citizens become increasingly aware of how the prospect of democratic reforms would affect their sect, and are increasingly sensitive to such effects.

Which Features of Democracy? The preceding evidence suggests that communal prayer influences attitudes toward democracy in fairly consistent ways that can be predicted by the

after syria 105 Sunnis

Pr (Agree)

1 .8 .6 .4 Never

Rarely

Around once per month

Around once per week

More than once per week

Around once per week

More than once per week

Shi`a

Pr (Agree)

.9 .8 .7 .6 .5 .4 Never

Rarely

Around once per month

Frequency of Mosque Attendance

Fig. 5.6 Frequency of Communal Prayer and Perceived Benefits from Democracy.

interests of a citizen’s sect. However, it is unclear exactly which aspects of democracy are responsible for these attitudes. Respondents may be expressing support for concepts that are not necessarily features of democracy but are otherwise desirable features of a political system (Bratton, 2010; Canache, 2012; Jamal and Tessler, 2008).50 Citizens’ understandings of democracy—particularly in the Arab world—can largely be grouped into two categories: “political” and “economic” (Tessler et al., 2012). In the first case, democracy implies free elections, rotation of political power, freedom to criticize the government, and other “political” characteristics. In the latter case, democracy can imply greater economic equality, redistribution, greater provision of public goods, and similar “economic” considerations. In Lebanon, there is reason to expect that both of these categories will feature prominently in citizens’ understandings of democracy. As discussed in the previous chapter, whether or not democracies actually redistribute more than non-democracies is less of a concern than whether citizens perceive this claim to be true; the evidence presented in that chapter strongly suggests that most citizens do so.

 faith in numbers Further, the political and economic interests of each sect generally point in the same direction in the current political climate. Since Shi‘a are poorer and more underrepresented than Sunnis (their main political rivals), they have both political and economic incentives to support democracy. Of course, the opposite is true for Sunnis. To the extent that both political and economic concerns enter into citizens’ preferences regarding democracy, their responses to questions about these specific aspects of democracy should follow the same patterns as their responses to questions about democracy in general. To test this claim, the survey asked respondents separate questions about individual dimensions and interpretations of democracy in order to determine (1) which features of democracy are driving these results; and (2) whether respondents are simply expressing support for “democracy” without actually supporting any of democracy’s implications. First, respondents were asked about a concrete political dimension of democracy: electoral reform. Respondents were asked about their level of agreement with the following statement: “Lebanon should abandon its sectarian electoral system.”51 In the Lebanese political environment, the term “democracy” carries strong anti-sectarian undertones. Sectarian quotas in government are perhaps the clearest manifestation of undemocratic rule in the political system, and citizens are highly sensitive to this issue. Abandoning the sectarian electoral system would imply moving toward a more proportional system rather than allotting a fixed number of seats to each sect; as shown previously, this type of change would benefit Shi‘a substantially. Therefore, communal prayer— to the extent that it has the expected effects—should lead Sunnis to be less likely to support this statement, but increase Shi‘a support for abandoning the electoral system. Figure 5.7 provides evidence in support of this claim. Among Sunnis, 82% of never-attenders support the statement, compared to just 60% of the most frequent attenders. Conversely, 86% of never-attending Shi‘a support this statement, and 93% of the most frequent attenders do so. These differences are sizable and significant at the .001 (Sunni) or .05 (Shi‘a) level. A related question asks respondents whether they agreed that “Lebanon should conduct a new census.”52 As Figure 5.8 demonstrates, the effects of mosque attendance for this item were similar to the effects on attitudes toward the sectarian system. In both cases, respondents expressed high levels of support for this statement. A new census would indeed reveal that Sunnis are underrepresented in the current arrangement; however, it would

Sunnis

Pr (Agree)

.9 .8 .7 .6 .5 Never

Rarely

Around once per month

Around once per week

More than once per week

Around once per week

More than once per week

Shi`a

Pr (Agree)

.95 .9 .85 .8 Never

Rarely

Around once per month

Frequency of Mosque Attendance

Fig. 5.7 Frequency of Communal Prayer and Attitudes toward Ending Sectarian System. Sunnis

Pr (Agree)

.95 .9 .85 .8 .75 .7 Never

Rarely

Around once per month

Around once per week

More than once per week

Around once per week

More than once per week

Shi`a

Pr (Agree)

1 .9 .8 .7 .6 .5 Never

Rarely

Around once per month

Frequency of Mosque Attendance

Fig. 5.8 Frequency of Communal Prayer and Attitudes toward New Census.

 faith in numbers reveal that Shi‘a are substantially more underrepresented than Sunnis. To the extent that political competition is seen as a zero-sum game between Sunnis and Shi‘a (as is mostly the case since the outbreak of the Syrian Civil War), a net gain for Shi‘a could be seen to offset an absolute gain for Sunnis. For non-attending Sunnis, this competition appears not to be a concern: over 90% of respondents in this category agreed with the statement. However, frequent attenders—who possess higher levels of group solidarity and information about group interests—are 16 percentage points less likely to agree with this statement. For Shi‘a, the opposite pattern is once again present. Frequent attenders are 30 percentage points more likely to support this statement than are non-attenders. Never-attenders report support for this statement at a rate of 62%; the fact that this level is not higher is likely due to some combination of a lack of information about group size, lower levels of sectarian identification, and a lower likelihood of viewing politics as a Shi‘a-Sunni contest. In any case, the differences between attenders and nonattenders are clear for both sects: communal worship makes respondents far more likely to express attitudes toward the census in accordance with the interests of their sect. Economic interests will also condition the relationship between communal prayer and regime attitudes, so it is useful to consider whether or not attendance is correlated with economic attitudes consistent with group interests. First, respondents were asked if they agree that “economic inequality is a major problem in Lebanon.”53 If the mechanism proposed in this book is indeed present, then frequency of communal prayer should make Sunnis less likely to agree with this statement, but have the opposite effect for Shi‘a. Figure 5.9 illustrates that this is the case: the most frequently attending Sunnis are over eight percentage points less likely than the leastfrequent attenders to agree with this statement; the corresponding difference among Shi‘a is a (positive) 19 percentage points. These differences are substantial in size and are statistically significant at the .05 level for Sunnis and the .001 level for Shi‘a. It may be the case, however, that these attitudes do not translate directly into regime preferences. To address this possibility, the survey asked respondents about their attitudes toward democracy including a prime that stated that democracies tend to be redistributive. The question wording was as follows: “Democracy often involves redistribution of wealth from rich groups to poor groups. Do you support this type of policy?”54 Once again, if my theory is correct, communal prayer should increase agreement among

after syria 109 Sunnis

Pr (Agree)

.95 .9 .85 .8 .7.5 Never

Rarely

Around once per month

Around once per week

More than once per week

Around once per week

More than once per week

Shi`a

Pr (Agree)

.95 .9 .85 .8 .75 .7 Never

Rarely

Around once per month

Frequency of Mosque Attendance

Fig. 5.9 Frequency of Communal Prayer and Attitudes toward Economic Inequality.

Shi‘a but decrease agreement among Sunnis. Figure 5.10 shows that this is the case: 81% of non-attending Sunnis agreed, compared to 66% of the most frequent attenders; among Shi‘a, 76% of non-attenders agreed, compared to over 91% of the most frequent attenders. These effects are statistically significant at the .01 level or better.55

Sermons The results presented thus far suggest that communal religion has opposite effects on democratic attitudes among Sunnis and Shi‘a, the two groups engaged in the most antagonistic political competition at the time when my survey was fielded. At the individual level, citizens’ beliefs about political issues and regime types are influenced heavily by religious participation, but the direction of this influence depends crucially on the interests of their sects. How, exactly, does this process work? One channel through which communal prayer influences sectarian identity and political preferences is

 faith in numbers Sunnis

Pr (Agree)

.9 .8 .7 .6 Never

Rarely

Around once per month

Around once per week

More than once per week

Around once per week

More than once per week

Shi`a

Pr (Agree)

.95 .9 .85 .8 .75 .7 Never

Rarely

Around once per month

Frequency of Mosque Attendance

Fig. 5.10 Frequency of Communal Prayer and Attitudes toward Redistributive Democracy.

through explicit cues from religious leaders. While it is difficult to generalize about the behaviors of individual believers or religious leaders in mosques, it is nevertheless useful to consider citizens’ reports about the political content they experience in the mosque. To this end, the survey asked respondents whether they hear political sermons at church/mosque, and whether these sermons (if they answered “yes”) supported or opposed democracy. Given the social desirability issues present in the latter question, there would be every reason to expect that respondents would not report hearing anti-democratic sermons. However, this proved not to be the case among Sunnis, the group who would lose the most from democracy in the current environment. Among respondents who reported hearing political sermons discussing democracy (about 40% of attenders), over 38% of Sunnis reported hearing anti-democratic sermons, compared to less than 11% of Shi‘a. Thus, there is some evidence that the some of the pro- or antidemocratic sentiment cultivated in mosques may be coming “from the pulpit.”

after syria 111 Public examples of political rhetoric from mosque leaders are abundant in Lebanon. Among Sunnis, the tone of high-profile sermons tends to emphasize stability rather than reform, stressing that Sunnis would most likely be the losers from major political changes. The former grand mufti of Lebanon (the highest-ranking Sunni religious official in the country), Mohammed Rashid Qabbani, delivered a sermon at Lebanon’s largest mosque on Eid al-Adha in 2012, stating, “We will not allow the government to be toppled through street action.”56 He further remarked that “abandoning the constitution means the destruction of the republic and country.” These words call for the preservation of the current political arrangement— and, by implication, Sunni political privilege—regardless of public opinion. He had previously stated that “adherence to, rather than departure from, the Taif [Accord] should be renewed, because it is [our] foundation.”57 More recently, current Grand Mufti Abdul Latif Derian has said, “The Taif Accord ended the Civil War, after which the Lebanese were wholly engaged in rebuilding the state,” but that progress was being threatened by, among other things, “the marginalization of [Sunni] Muslims in public affairs.”58 These statements reflect a common theme in Sunni political discourse: that reform threatens to marginalize the Sunni community, who benefit from the non-democratic elements of the Ta’if Accord.59 Discourse on democracy among Shi‘a preachers generally follows the opposite pattern, emphasizing that moving away from the sectarian system and toward true democracy would benefit the Shi‘a community. These sermons stress the continuing political and economic disadvantages experienced by Lebanon’s Shi‘a, and suggest that electoral reform will help to remedy the existing inequality between the sects. Sheikh Ahmad Kabalan, a Shi‘a religious leader, has stated, “The Shiite sect has not made any political, economic or social gains and the Shiite community continues to compromise to preserve Lebanon as a model of coexistence, forgiveness and love.”60 Sayyid Ali Fadlallah, the son of (and successor to) Mohammed Fadlallah, perhaps Lebanon’s most influential Shi‘a cleric, has commented extensively on the recent political deadlock, calling for the election of a new president (as constitutionally required) and objecting to the parliament’s extension of its term.61 Mohammed Fadlallah himself frequently called for democratic reforms in Lebanon, even though he considered himself an Islamist. He viewed Islamic government in Lebanon as unrealistic, and his words suggest an acceptance of democracy as a reasonable compromise; he suggested that pluralistic democracy “helps to further the cause of

 faith in numbers Islam or secure total control by the will of the majority” (Sankari, 2005, p. 244). The reasons for Fadlallah’s reluctant acceptance of democracy reveal an important dimension of sectarian attitudes toward democracy: instrumental concerns are highly important in determining the relationship between religion and democratic preferences. Fadlallah himself was, at best, ambivalent toward democratic values per se, but favored democratic reforms because of the practical benefits they would bring to his community. Sermons on both sides of the sectarian divide emphasize the importance of communal prayer for group solidarity. Ali Fadlallah recently devoted a sermon to the importance of communal prayer. In this sermon, Fadlallah stated that communal prayer has a “psychological impact” on worshipers that “generates compassion, cohesion, and solidarity.”62 This “solidarity” discourse frequently takes on a distinctly sectarian character. In a recent sermon, Sheikh Salem al-Rafei, leader of the Sunni al-Taqwa Mosque in Tripoli, called for the creation of “a military council for the Sunni sect in Lebanon,” suggesting not only sectarian chauvinism, but also a severe concern that the Sunni sect is particularly threatened. Lebanon’s Sunni grand mufti, Sheikh Abdul Latif Derian, has begun distributing “suggested” sermons to all Sunni scholars and preachers, seeking to promote a unified message.63 This type of practice is common in Lebanon, and indicates that the content present in sermons is likely to be at least somewhat consistent across congregations, preaching a message of sectarian solidarity.

Conclusion This chapter has applied the theory of religion, group interest, and democracy to the changing Lebanese political context in the wake of the Syrian Civil War, which shifted the axis of competition focused on the Sunni-Shi‘a divide. The observational and experimental results presented in this chapter point to the same trend in 2014 Lebanon: communal religion, either in the form of regular attendance at religious services or experimental primes intended to mimic the effect of communal prayer, heightens the salience of sectarian identity and alters regime preferences accordingly. For Sunnis, communal religion has a consistently anti-democratic effect. For Shi‘a, the exact opposite trends are present. I have argued that the reason for these divergent effects is to be found in group interests. Electoral reform and

after syria 113 economic redistribution would both tend to benefit Shi‘a at the expense of Sunnis, and the ongoing conflict in Syria has drawn the lines of political competition rather clearly between these two groups, with Christians largely remaining on the sidelines. Statistical tests using data from an original survey, as well as evidence from religious and political discourse, support these claims. Communal religious practice drives support for democracy up for some groups, but down for others—and this difference can be explained by sectarian interests and changing political circumstances.

 faith in numbers

Appendix 5A: Description of Experiment In the religious priming experiment, respondents were randomly assigned to complete a “word association” task consisting of five items. In each case, the respondents were presented with five words and asked to select their closest match from three choices. The word groups were as follows:

Control Group Delicious: Sleep: Work: Newspaper: Shirt:

a) Hungry a) Tired a) Money a) Read a) Clothes

b) Tasty b) Rest b) Job b) Politics b) Pants

c) Good c) Bed c) Time c) Paper c) Shoes

Communal Treatment Group Sect: Man of religion: Shrine: Sermon: Place of worship:

a) Group a) Leader a) Altar a) Message a) Church/Mosque

b) People b) Preacher b) Holy Place b) Teaching b) Community

c) Community c) Teacher c) Saint c) Devotion c) Clergy

Personal Piety Treatment Group Faith: Prayer: Fasting: Heaven: Soul:

a) Belief a) God a) Duty a) Paradise a) Spirit

b) Religion b) Petition b) Obligation b) Earth b) Religion

c) Dogma c) Duty c) Sacrifice c) Eternity c) Heart

after syria 115

Appendix 5B: Description of Variables Support for Democracy, Observational Section For the observational models, the support for democracy measure consisted of an additive index of the following items (all questions are on a five-point Likert Scale): 1. Democracy is the best form of government. 2. I would support a political system governed by a strong authority which makes decisions without considering electoral results or the opinions of the opposition.∗ 3. I would support a political system governed by religious leaders.∗ 4. I would support a political system governed by the army.∗ 5. Democracy would be better for Lebanon than the current political system. 6. Democracy is better than any other form of government. 7. Lebanon needs to become more democratic than it is right now. 8. A non-democratic government is sometimes necessary for establishing stability.∗ 9. Democracy is always the best form of government. The experimental section used an additive index of items 1–4. In both cases, the support for democracy variable was rescaled to range from for ease of interpretation.

∗ Indicates that the question was inverted such that higher values indicate greater support for democracy.

 faith in numbers Table 5B.1 Summary Statistics Variable Support for Democracy (Observational) Support for Democracy (Experimental) Frequency of Communal Prayer Weekly Attendance Female Age Education Unemployed Monthly Income (USD) Religious Person Close to Members of Congregation Personal Prayer End Sectarian System Income Inequality Major Problem Government Must Reduce Sectarian Inequality Linked Fate Support for Democracy (Redistribution) Democracy Benefits People Like Me

Mean

Std. Dev.

Min.

Max.

.6 .57 3.05 .47 1.45 37.81 3.41 1.19 1,627.46 1.85 1.99 4.05 3.85 3.85 4.15 3.15 3.77 3.55

.17 .2 1.3 .5 .5 11.92 1.13 .4 1,045.66 .36 .66 1.7 .84 .93 .69 1.09 .91 .99

0 0 1 0 1 21 1 1 97 1 1 1 1 1 1 1 1 1

1 1 5 1 2 65 6 2 10,000 2 3 6 5 5 5 5 5 5

6 Representation or Redistribution? Evidence from Iraq Introduction This chapter tests my theory of religion, group interest, and democracy in the Iraqi context. I have suggested that group interests account for a great deal of the variation in the role played by religion in influencing democratic attitudes and behaviors. Broadly, communal religious practice (or communal prayer) tends to promote democratic attitudes when the individual’s sect would benefit from democracy. Conversely, when a citizen’s sect would tend to lose (power, rights, money, etc.) from democracy, communal prayer will tend to have the opposite effect. The structure of sectarian relations— and thus, the religious-sectarian patterns of support for democracy—are influenced by political, economic, and social factors that can and do change across time; it is a central theme of this book that the axis of political competition is a key determinant of how communal religion will tend to affect attitudes toward regimes. As Haddad (2011, p. 31) writes, “Political climate and geographical and class differences influence sectarian relations and the salience of sectarian identity.” I have proposed that communal prayer heightens the salience of sectarian identity, which in turn pushes believers’ political attitudes into closer alignment with the interests of their religious groups. This pattern helps to explain why religion sometimes serves as a major force for democracy, but sometimes undermines democracy altogether. While the Lebanese case is an important example of sectarian politics, its unique institutional configuration might limit the ability to extrapolate from its experience. This book therefore considers other cases in order to supplement the insights provided by Lebanon. The Iraqi case is another ideal setting in which to test the theory presented in chapter 2. First, and perhaps most importantly, Iraq is a society that is highly divided along religious lines; a casual glance at virtually any newspaper confirms the ongoing importance of sectarian identity in this country. Second, both staunch authoritarianism Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0006

 faith in numbers and attempts at democracy are in the recent memories of most Iraqi citizens. This feature of the Iraqi case is helpful insofar as it provides experiential contrast for citizens; when they are asked questions about democracy or authoritarianism, they are likely to have a richer understanding of the meanings and implications of these words than citizens of a country that has been steadfastly democratic or autocratic for many years. In other words, citizens are much more likely to have a clear picture of what it means to compare democracy to other types of regimes. Third, the centrality of the state in Iraq’s economy (particularly in the area of natural resources) means that the redistributive implications of regime type are immense. In cases where the state plays a minimal role in the economy, citizens might not have economic reasons to prefer one regime type over another; in a case like Iraq, where an overwhelming portion of the country’s revenues are controlled by the state, the economic ramifications of regime type could hardly be stronger. Sect plays a vital role in channeling demands (and policies) related to politics and redistribution in Iraq. In reference to the new state formation process undertaken following the US invasion, Jabar (2011, p. 21) notes that “within these grand blocs [sects] communal labels and unity at this stage were conceived of as an assured vehicle to reach out for power and fair distribution of national wealth” (emphasis added). Sectarian identities are thus an important part of virtually all political deliberations in Iraq. While Iraq has not conducted a complete census since 1987 and has delayed recent plans for a new census several times (Hiltermann, 2010), it is possible to characterize ethnic and sectarian demographic patterns broadly. The US State Department (2012) estimates that Shi‘a, most of whom are ethnically Arab, account for 60% to 65% of the population; 18% to 20% are Sunni Kurds,1 and 12% to 16% are Sunni Arabs. Regardless of the precise figures, it is clear that Arab Shi‘a constitute a numerical majority. Even counting all Sunnis (more than half of whom are non-Arab and could therefore scarcely be counted on as allies of Arab Sunnis), Sunnis are outnumbered by Arab Shi‘a. These demographic realities have made many Sunnis hostile to the new pseudo-democratic political arrangement, as they have lost the privileged status afforded to them under the prior autocratic regime; Jabar (2011, p. 21) writes that for Shi‘a, “demography is democracy,” but the main catchword for Sunnis has been “restoration.” The potential for heightened sectarian identity in Iraq is magnified by several factors. Iraq is home to several of the most important religious sites in Shi‘a Islam. Najaf is believed to be the burial site of ‘Ali ibn Abi Talib , the

representation or redistribution? 119 cousin and son-in-law of the prophet Muhammad and his rightful successor in the eyes of the Shi‘a. His tomb is considered a shrine and is a major site for pilgrimages from Shi‘a across the region and the world. Karbala, southwest of Baghdad, is the location of one of the most important battles (in AD 680) in Islamic history. The battle was a key event in the early days of the Sunni-Shi‘a division, and remains symbolically significant in the memories of Muslims. During this battle, ‘Ali’s son Husayn was killed along with several of ‘Ali’s other family members and supporters. Shi‘a continue to commemorate the death of Husayn by making pilgrimages to Karbala or holding ceremonies and processions on the day of ‘Ashura’, which often include public prayers, vocal displays of mourning, and self-flagellation. Iraq is also home to numerous key Shi‘a theological leaders and institutions. Historically, Iraq has been home to a number of influential maraji‘ (singular marja‘), the highest level of Shi‘a legal authorities, who are viewed as “sources of emulation” throughout the Shi‘a world. Individual Shi‘a in Iraq and elsewhere will often look to one marja‘ as a guide for living; many of the most popular maraji‘ have been—and continue to be—located in Iraq.2 Due to their frequent status as a minority community and their history of repression at the hands of Sunni regimes, Shi‘a religious figures have traditionally followed a practice of “quietism,” focusing on theological rather than political issues. This was not always the case in Iraq or elsewhere, and Sunni rulers of Iraq were often suspicious of these institutions, which were difficult to monitor and whose loyalty was not assured despite their public acquiescence to the regime’s rule. Saddam Hussein, the Sunni ruler of Iraq from the 1970s3 until his removal by American forces in 2003, viewed these religiously significant features of Iraqi society as potential rallying points for Shi‘a mobilization. His regime, though certainly not exclusively Sunni in composition, favored Sunnis heavily in a number of important ways. Continuing an earlier pattern, Saddam staffed the army’s officer corps predominantly with Sunnis despite their much smaller share of the population (al-Marashi and Salama, 2000). Saddam’s highest priority was loyalty; specifically, he believed that complete loyalty to the Ba‘ath Party4 was essential in order to prevent a coup (Sassoon, 2012). Ever mindful of the potentially conflicted allegiances of Iraqi Shi‘a, Saddam feared the possibility of an uprising inspired by the 1979 Shi‘a revolution in neighboring Iran. While the regime continued to recruit Shi‘a for the rank and file out of necessity, their loyalty was always subject to scrutiny. Furthermore, the regime’s paranoia about the potential

 faith in numbers Shi‘a threat drove it in many cases to repress the Shi‘a community on a wholesale basis. Blaydes (2018) argues that the regime’s tendency to repress the Shi‘a collectively may have actually had the counterproductive effect of heightening communal identity. Saddam’s favoritism toward Sunnis extended to the economic realm as well. He used his control of state resources to allocate more public funding to Sunnis and provided Sunni villages with access to more reliable electricity (Asquith, 2009, p. 5). Sunni tribes also received special access to land, money, and weapons (Karam, 2007). Many have argued that much of the regime’s anti-sectarian and anti-tribal rhetoric was meant to distract Shi‘a and Kurds from the realities of Sunni Arab rule. The Hussein regime tried to squash public sectarian identity, so it is difficult to find precise statistics on income or wealth levels by sect, but it is clear that Sunnis benefited disproportionately from the state’s largesse during this period. Ordinary Shi‘a held strong perceptions about the relative treatment of the sects during the Hussein regime; Haddad (2011, pp. 47–48) quotes a former Ba‘ath official as saying that the regime’s behavior “created an unmistakable belief amongst the Shi‘a that the state was against them.” The nominally secular (or at least cross-sectarian) nationalist ideology of Saddam’s Ba‘ath Party does not match the realities of social and religious policies in the country during this period. As Dodge (2012b, loc. 528) writes, although the Ba‘ath Party claimed to integrate Sunni and Shi‘a imagery into its ideology, “It was clearly more inclusive of Sunni symbolism than Shi‘a”; furthermore, Iraqi state schools taught Sunni, not Shi‘a, Islam, and a number of important Shi‘a religious practices were banned by the regime. Indeed, Cole (2003, p. 545) states that “the Ba‘ath massively persecuted the religious Shi‘a of the south” (emphasis added). These regime behaviors suggest that religion rather than mere sectarianism, in which religious identity functions in almost exclusively ethnic terms, held important political implications. On the other side of the equation, Shi‘a also clearly felt that religious considerations were of paramount importance. In April 2003, shortly after the fall of Saddam Hussein’s regime, millions of Shi‘a took part in Arba‘een, an important Shi‘a religious ritual that had been banned by Saddam (Jabar, 2003). This ritual only grew in scale, drawing over 20 million participants in 2015 (Christia, Dekeyser, and Knox, 2016a). Clearly, the link between political identity and religious behavior remains strong. A great deal of the violence of the post-Saddam era can be attributed to the changes that came with the new regime. Haddad (2017b) describes

representation or redistribution? 121 the post-Saddam state-building process as “Shi‘a-centric” and observes that this project has been broadly rejected by Sunnis. The US invasion removed state patronage, “the very life line” of many Sunni tribes (Jabar, 2011, p. 21). Since Sunni Arabs represent less than 20% of the country’s population, it is reasonable to expect that a would-be democratic regime would not present them with disproportionate influence or patronage. Dodge (2012b, loc. 726) suggests that the new regime went even further, creating “an exclusive elite bargain that consciously excluded, and indeed demonised, not only the old ruling elite, but also the whole Sunni section of Iraqi society from which that elite had largely come” (see also Dodge, 2012a). A number of Sunni groups have responded violently to their perceived unfair treatment. Karam (2007, p. 88) notes that about 90% of Iraq’s anti government insurgents are believed to be Sunni, and 85% of insurgent attacks following Saddam’s fall took place in the four provinces of the Sunni heartland (where only about 40% of the population lives). These patterns suggest that the loss of Sunni privileges in Iraq’s pseudo-democracy may account for the heightened propensity for violence in these communities; in concrete ways, democratic reforms have benefited Shi‘a at the expense of Sunnis. At the same time, it is possible that these grievances are only held by a small number of extremists; the “Sunni street” may not be as frustrated with the changes brought by the post-Saddam era. Attitudinal evidence from the Arab Barometer, however, suggests that this is not the case. In this survey, 47% of Shi‘a responded that the country’s economic situation was “good,” compared to only 26% of Sunnis. Sunnis were similarly disaffected in their political evaluations, scoring the government an average of 3.4 on a scale from 1 to 10, compared to 4.5 among Shi‘a. Sunnis also reported consistently lower levels of trust in Iraq’s major institutions when compared to Shi‘a; the sectarian difference is especially large for trust in the government and the army. These attitudes extend to tolerance of authoritarian regimes: Sunnis were over four times as likely to express support for an authoritarian system as Shi‘a.5 In total, it appears that Sunnis are on average much less satisfied with the changes brought by the current regime.

Religion and Sectarianism in Iraq Skeptics of the role of religion have argued—in many cases, fairly—that sectarianism is not really about religion at all. While there is no doubt

 faith in numbers that religious sects closely resemble other identities in certain ways, there is considerable evidence to suggest that the uniquely religious aspect of political behavior in Iraq is important over and above the more ethnic-style aspects of sectarian identity. The fact that Iraqi Muslims have, at various times during the post-Saddam era, organized for political purposes in large numbers after Friday prayers (a phenomenon witnessed elsewhere in the Muslim world as well) suggests that religion serves an important motivating purpose. It is, of course, possible that mosques serve as organizational centers only because they are places where members of the same sect regularly meet, and nothing more. However, in highly segmented sectarian societies such as Iraq, this explanation does not account for much of the importance of mosques. In Iraq (as well as many other countries experiencing high-level sectarian conflict), the bulk of civil society organizations are fairly homogeneous if not outright sectarian. If religion is simply a channel for intra group engagement, then mosques should be no more likely to organize or facilitate political behavior than any other organization, and this is not the case. It appears, then, that religion is a particularly powerful motivator of political attitudes and behavior in Iraq in a way that relates to—but does not completely overlap with—sectarianism. Moreover, if religion were simply a placeholder for social or economic grievances, the discourse of members of different faiths would reflect this reality. The rhetoric used by members of each sect following the US invasion, however, carries a distinctly religious inflection. While Shi‘a certainly referred to Sunnis as “Saddamists” in many cases, a large number of clearly religious labels were used; “Wahhabis” to indicate a strict conservative form of Islam, and nawasib, meaning those who reject Shi‘a imams and “hate” the ahl al-bayt, the family of the prophet Muhammad. To Sunnis, Shi‘a became al-rafidha, rejectors or deserters, a term used to denigrate Shi‘a6 for failing to acknowledge the early Islamic leaders believed by the Sunnis to be legitimate (Rosen, 2006). These terms (among others) have become increasingly common as Sunni-Shi‘a tensions have intensified throughout the region (Siegel, 2017; Zelin and Smyth, 2014) and perhaps especially so in Iraq (Haddad, 2013b). They suggest the perceived importance of the sects’ actual religious rather than merely political differences. The targets of political violence also reveal the particular importance of religion. Dodge (2012b, ch. 2) demonstrates that mosques and other religious institutions have been especially common targets. If mosques were just like any other sectarian civil society institutions, then they should not be any

representation or redistribution? 123 more likely to be targeted by attacks than any other civil society center. For instance, if a Sunni insurgent simply wanted to attack as many Shi‘a as possible, then any number of predominantly Shi‘a organizations could be targeted alongside mosques. Data on terrorist incidents in Iraq, however, suggest that mosques are frequently targeted. The Global Terrorism Database (2013) records data on terrorist incidents across the globe and contains thorough records of Iraqi incidents. Between 2003 and 2012, this database contains reports of 374 attacks against Iraqi religious figures/institutions, which only includes attacks on religious leaders (imams, priests, and others), and religious institutions, places, or objects (mosques, churches, shrines, relics). Thus, these several hundred attacks do not count the many other attacks on religious pilgrims, religious schools, and other possible religious targets. Of the 22 categories of attack targets used by the database, the relatively narrow category of religious figures/institutions ranks sixth in number of attacks in this period, behind only the very broad categories of business, government (general), police, military, and “private citizens and property,” a very broad category that includes attacks on religious pilgrims among other religiously motivated targets; this vague category accounts for over 35% of the total number of attacks. Even as it stands, considering only the narrow definition of religious figures/institutions, religious entities were targeted more than other seemingly common target categories like nongovernmental organizations, food/water supply, utilities, journalists/media, tourists, and nonstate militia. Furthermore, attacks on religious institutions/figures were on average far deadlier than other sorts of attacks, perhaps owing in part to the relative scarcity of targets and larger degree of planning required. Almost 14 people were injured in the average attack on a religious target, compared to 7.5 among non religious targets (the corresponding figures for deaths are 5.6 and 3.2). In terms of total number of wounded, religious targets actually outrank military targets (4,868 wounded compared to 4,261). Furthermore, these figures do not include attacks from 2013 and 2014, which casual observation suggests were even more likely to be targeted against mosques.7 In sum, it appears that religious institutions and figures—even narrowly defined—represent a significant portion of insurgent attacks and an even larger portion of injuries and deaths caused by such attacks. Accounts of religion in Iraq likewise suggest that mosques have historically played particularly important political roles, even in comparison to other civil society institutions. Jabar (2011, p. 22) states that Shi‘a mosques “acted more like a political agent, information, and ideological center,” while

 faith in numbers Sunni mosques became “a recruitment, mobilization and insurgent agency.” In other words, mosques in post-Saddam Iraq came to function similarly to places of worship across the world, providing information, ideology, and organization in the name of the religious group. Even such extreme examples as sectarian violence (on both sides) have frequently used mosques for organizational/motivational purposes; Dodge (2012b, loc. 1107) observes that radical Shi‘a cleric Muqtada al-Sadr posted orders and sermons for the military wing of his movement on mosque walls across Iraq. In cases like this, mosques are neither politically neutral nor simply filling in for other types of civil society institutions; rather, they serve as uniquely strong incubators for all sorts of political activity. Presumably, the politicizing function of religious behavior is why, as Jabar (2003, p. 18) observes, the Shi‘a religious resurgence has “startled Iraqi Sunnis, clerics and laymen.” For both Sunnis and Shi‘a, sectarian concerns remain central to political identities, preferences, and behaviors.

Politics or Economics? As a supplement to the Lebanese example, the Iraqi case provides an opportunity to distinguish rather clearly between the political and economic features of democracy and how each of these features affects sectarian interests. Concerns about representation in Iraq’s political institutions tell only part of the country’s sectarian story. Economic issues—particularly those surrounding government-led redistribution of resources—are also highly salient in the Iraqi context, and often take on a sectarian tone. Changes in sectarian wealth have become a central theme of political rhetoric, and these changes have had a dramatic effect on how Iraqis view the state, its institutions, and democracy in general. In short, it would be misguided to focus exclusively on issues of representation without acknowledging how sectarian economic concerns may influence regime preferences. Figure 6.1 displays the average income of each sect as derived from the nationally representative sample from the second wave of the Arab Barometer, conducted in 2010. Traditionally, Iraq’s Sunnis were considered wealthier—and had considerably more access to state resources—than Shi‘a because of their closer ties to the regime prior to the American invasion in 2003. Each of these figures, however, suggests that contrary to the assumptions of some observers, Shi‘a respondents are now wealthier than

representation or redistribution? 125 Sunnis on average. Including all ethnic groups, the average Shi‘a income is about 31% higher than that of the average Sunni; this difference is depicted in Figure 6.1a. This difference, however, may simply be due to the

(a) 800000 698,049

Mean Income, Dinars

600000 532,030

400000

200000

0

Shi‘a

Sunni

(b) 800000 698,908 600,933

Mean Income, Dinars

600000

400000

200000

0

Shi‘a

Fig. 6.1 Average Income by Sect.

Sunni

 faith in numbers fact that Kurds, who are overwhelmingly Sunni, are considerably poorer than Arabs on average. To address this possibility, Figure 6.1b restricts the calculation to Arab respondents. When limiting the sample only to Arab respondents, the difference between Sunnis and Shi‘a is smaller, but still substantial: the average Shi‘a Arab earns over 16% more than the average Sunni. The evidence suggests, therefore, that the previous perception of Sunnis as the wealthier sect in Iraq is no longer accurate, if indeed it ever was. The increase in relative Shi‘a wealth is likely a direct result of government redistributive policy. Whether or not the post-Saddam government has managed to make Shi‘a the wealthier sect, Shi‘a control over most of Iraq’s oil (and the government agencies responsible for it) makes such changes virtually inevitable, and this inevitability is not lost on Iraqi Sunnis. Blanchard (2011, p. 13) notes that “Sunni negotiators opposed Iraq’s new constitution in part because it empowers regions in oil production and revenue allocation policy.” This arrangement threatens Sunni interests in a clear way: most of Iraq’s proven oil reserves lie in heavily Shi‘a areas. The US Energy Information Administration reports that “Iraq’s resources are not evenly divided across sectarian-demographic lines. Most known hydrocarbon resources are concentrated in the Shiite areas of the south and the ethnically Kurdish region in the north, with few resources in control of the Sunni minority in central Iraq” (2013, p. 1). In a country with the fifth-largest proven crude oil reserves in the world, the redistributive implications of this inequality are substantial. Furthermore, the Ministry of Oil controls oil and gas production throughout Iraq with the exception of Kurdish territory, where production is controlled by the Kurdistan Regional Government (US Energy Information Administration, 2013, p. 2). The statecontrolled oil industry accounts for the bulk of the Iraqi economy; the World Bank estimates that nearly 78% of Iraq’s GDP in 2011 came from oil rents (2011). Figure 6.2 displays Iraq’s districts according to accessible oil reserves and sectarian (or ethnic) composition. Data for this map are taken from Berman et al. (2011). This figure confirms that the bulk of Iraq’s oil is concentrated in solidly Shi‘a or Kurdish districts. Only two solidly mainly districts fall above the lowest category (less than 1 billion accessible barrels), and even these districts remain in relatively lower categories of reserves compared to other districts, especially in the southern tip of the country, where several super-fields exist. Two factors combine to limit Sunni access to oil revenues:

representation or redistribution? 127

Legend Accessible Oil (barrels) Less than 1,000,000,000 1,000,000,000 – 3,000,000,000 3,000,000,000 – 4,500,000,000 4,500,000,000 – 6,000,000,000 6,000,000,000+ Primary Sect(s) Kurd Sunni Shi`a Shi`a & Kurd Sunni & Kurd Sunni & Shi`a

Fig. 6.2 Iraqi Districts, by Oil Reserves and Major Sects.

(1) Iraq’s federal structure allows some of the oil revenues to be distributed at the province, district, or local levels, so oil-rich areas tend to be able to keep a great deal of the money generated by these resources; and (2) the central government is heavily Shi‘a-led. Consequently, substantial sectarian inequality has emerged in the distribution of revenues produced by oil and other resources. Unsurprisingly, the oil sector has become an important focus for sectarian conflict in post-Saddam Iraq. The logic of this aspect of the conflict is straightforward: the fact that government rents account for such a large portion of the Iraqi economy means that who is in power is of the utmost importance, probably even more than in countries with more diverse or less state-centric economies. Since the state controls an especially large segment of the economy, its redistributive power is quite sizable. These realities will tend to heighten the stakes of political conflict regardless of the form that conflict takes. Whether political or economic in nature, the importance of sectarian concerns in the new political order is clear: Shi‘a have used the opportunity of political opening to amplify their voice and representation, and even to shift the distribution of economic resources in their favor. Religious-sectarian identities have provided an important lens through which citizens view these emergent issues.

 faith in numbers

Communal Prayer and Regime Preferences The political changes of the post-Saddam era8 have created differing incentives for Sunnis and Shi‘a to support or oppose democracy; however, a key argument of this book is that we must look for intra sectarian differences between citizens in order to understand the role played by religion more fully. Specifically, I have argued that individuals who engage in communal religious practice (in these cases, regular mosque attendance) will tend to hold attitudes toward democracy that are more closely aligned with the interests of their group. Before outlining the hypotheses of this chapter, however, it is important to consider how everyday religious behavior interacts with sectarianism in Iraq. Sectarian identity, though not always a source of conflict, has been politically relevant throughout Iraq’s modern history. Even under Saddam Hussein, whose regime desperately tried to push Iraq away from sectarian affiliations9 and toward secular nationalism, religious actors, organizations, and beliefs were almost constantly perceived as a potential threat to the regime. Religious organizations were relatively well insulated from the regime’s monitoring (not for the regime’s lack of trying), and some clerics used these groups to promote opposition to the status quo.10 In the 1970s, some Shi‘a religious leaders used the opportunity of large-scale Shi‘a religious rituals to launch protests against the regime; these protests were met with brutal repression, indicating the seriousness with which the regime viewed their potential. Access to—and engagement with—religious communities contributes significantly to the politicization of sectarian identity. The Hussein regime viewed Shi‘a communal activities, in the form of shrines, religious institutions, religious rituals, and many others, as challenging the stability of his regime. At the same time, as the ruler of a highly religious society, he needed to present himself as a pious Muslim, at times an awkward balance. Recognizing that he was a Sunni leader of a majority-Shi‘a country, he often moved back and forth between attempting to co-opt the majority sect on the one hand and brutally repressing its religious expression on the other. In the late 1970s, his regime invested millions of dollars in the Shi‘a holy sites of Najaf and Karbala, while almost simultaneously banning Ba‘ath Party members from participating in communal Shi‘a rituals and eventually banning these ceremonies altogether (Christia, Dekeyser, and Knox, 2016b, p. 17). Though the strategies for dealing with the dangers of Shi‘a religious

representation or redistribution? 129 mobilization changed over time, the recognition that religion posed a threat to the regime was constant. Accounts of sectarian political behavior in Iraq observe that mosques have continued to serve as loci for mobilization since the downfall of the Hussein regime. The most active and visible mobilization has come, of course, from Sunnis who feel that the Shi‘a-led government has discriminated against them. In April 2013, the BBC reported, “For months, thousands of Sunnis have been protesting after weekly Friday prayers” (Jabar, 2011). While all Sunnis have some reasons to feel threatened by democracy (or the current political arrangement), I have argued that these reasons are stronger among those who engage in communal religious practice than those who do not. As explored in previous chapters, communal prayer strengthens sectarian identity through several possible mechanisms. First, communal religious practice by definition involves meeting with others of the same sect, suggesting a potential to build “bonding” social capital (i.e., within the same sect; see Putnam, 2000). Second, communal prayer may serve an informational role: individuals who attend mosque may learn about the condition of their coreligionists (as well as their political interests) through interaction before, during, or after communal prayer. Third, messages received from religious leaders may increase the salience of sectarian interests. Through some combination of these mechanisms, individuals who participate in communal prayer will tend to develop attitudes toward democracy that are more closely tied to the interests of their group than those held by individuals who do not attend mosque, even within the same sect. Political conditions in Iraq have created an environment in which concerns of representation push Shi‘a to favor democracy but drive Sunnis to resist it. Numerically inferior and marginalized by the post-Saddam quasidemocratic reforms, many Sunnis feel disenchanted by democracy, or at least the partial form it has taken in post-invasion Iraq.11 Shi‘a, on the other hand, have maximized the benefits of majority rule. In the areas of representation and distributive politics, they have unquestionably elevated the status of their community, despite the challenges created by the security and economic situations. If, as I have argued, communal religious practice reinforces sect-centric views of democracy, religious attendance should further divide Sunnis and Shi‘a on the issue of regime preferences. Put simply, the logic of the argument posed by this book suggests a few hypotheses in the Iraqi case. The first of these relates to the general effects of mosque attendance on support for democracy within each sect:

 faith in numbers Hypothesis 6. Mosque attendance will tend to increase support for democracy among Shi‘a, but decrease support for democracy among Sunnis.

A Test of the Theory in Iraq To test Hypothesis 6, I conducted a series of statistical tests using the Iraq sample from the Arab Barometer. These tests aim to estimate the effect of communal prayer (mosque attendance) on support for democracy. The dependent variable is derived from a question asking respondents the extent to which they agree with the following statement: “A democratic system may have problems, yet it is better than other systems.” This question was chosen because it captures commitment to democracy more clearly than the other questions, which either ask about specific features of democracy or ask about conditional support for democracy. Accordingly, I have coded as 1 only those respondents who strongly agreed with this statement, since weak or conditional support for democracy is neither as conceptually meaningful nor as practically valuable as an outright commitment to democracy. The main independent variable is communal prayer, which is coded from a question asking respondents if they “attend Friday prayer.” Respondents are coded as 1 if they responded that they attend Friday prayer always or most of the time, and 0 otherwise. These models also include a handful of typical control variables (age, education level, income level), a control for individual religious behavior (Qur’an reading),12 and a few control variables necessary to capture the different experiences of the various sects in postSaddam Iraq (self-perceived safety, self-perceived equal treatment). These control variables help to ensure that the relationships observed between communal prayer and attitudes toward democracy are not spurious. Without accounting for the effects of communal prayer, differences between Arab13 Sunnis and Shi‘a in support for democracy are minimal: when accounting for demographic controls, the difference between Sunnis and Shi‘a in support for democracy is less than two percentage points, and is nowhere near statistical significance (p-value = .659). However, as in previous chapters, the key quantities of interest here are differences between attenders and non-attenders within each sect. A cursory look at the conditional distributions of the support for democracy variable according to sect and communal prayer suggests that, on balance, communal prayer has opposite effects on support for democracy between the two sects. Shi‘a who

representation or redistribution? 131 engage in communal prayer are about five percentage points more likely to support democracy than their coreligionists who do not pray at mosque; for Sunnis, the opposite effect is present: mosque-attenders are about 10 percentage points less likely to support democracy than those who do not attend. These distributions are displayed in Tables 6A.1 and 6A.2. Of course, these distributions do not control for any potentially confounding factors. It is important to consider whether or not these differences hold up in a multivariate setting, adjusting for the variables previously listed. Figure 6.3 displays the marginal effect of communal prayer on support for democracy for Sunnis and Shi‘a separately. Full results of the model are available in Table 6A.3. As this graph suggests, communal prayer has entirely opposite effects on support for democracy between the sects, consistent with Hypothesis 6. For Sunnis, communal prayer is associated with an 11.5 percentage point decrease in support for democracy, while for Shi‘a, communal prayer increases support for democracy by over eight percentage

Marginal Effect of Communal Prayer

.2

.1

0

–.1

–.2 Shi‘a

Sunni

Fig. 6.3 Marginal Effects of Communal Prayer on Support for Democracy, by Sect. Note: Figure displays changes in predicted probability of support for democracy comparing a respondent who engages in communal practice with a respondent who does not, i.e., Pr(Support|Practice)− Pr(Support| ∼ Practice). The predictions come from separate logistic regressions for each sect, controlling for the variables listed previously. Bars represent 95% confidence intervals. These calculations use the observed value approach for the control variables; see Hanmer and Kalkan 2013.

 faith in numbers points. These effects are sizable when we consider that only about 30% of respondents are coded as being strongly supportive of democracy according to the criteria used. I have suggested that the reason for this trend is that Sunnis, on balance, have a considerable amount to lose from true democratization in the area of representation, and their losses would be met with corresponding gains for Shi‘a. Consequently, communal religious practice has contradictory effects for these two groups. Sunnis who attend mosque become more attached to their sect, and therefore more closely attuned to the implications of democratization for their coreligionists in their country. The same process occurs for Shi‘a, but has the opposite effect: since Sunnis would presumably lose from democracy, communal prayer pushes Sunnis toward less democratic attitudes, while the opposite effect is present among Shi‘a, who might have much to gain from democratic reforms.

Representation or Redistribution? However, “democracy” means different things to different people. Perhaps the most obvious implication of democracy is free elections, which will tend to privilege larger groups at the expense of smaller ones. This understanding of democracy is the most commonly reported feature in Iraq, and the results presented previously are consistent with this characterization of regimes. Larger groups (in this case, Shi‘a) will generally benefit from free and fair elections, while smaller groups (here, Sunnis) may have reason to fear the development of a “tyranny of the majority” (Tocqueville, 2003 [1835]). This fear is especially likely in settings such as Iraq in which the minority group had enjoyed special privileges under the previous non-democratic regime. Majoritarian political concerns, however, are not the only aspect of democracy that citizens consider to be important. The axis of politicalsectarian competition, as described previously, is in part defined by what people consider the key implications of democracy to be. I have suggested that one of the main (perceived) features of democracy may be the redistribution of wealth from the rich to the poor, a claim whose history extends back at least to Aristotle, who argued that democracy “has in view the interest. . .of the needy” (Everson, 1996, p. 72). More recently, this claim has been modified by Boix (2003) and Acemoglu and Robinson (2006), among others, building upon the work of earlier models of redistributive preferences

representation or redistribution? 133 (Meltzer and Richard, 1981; Romer, 1975). Since the poor are more numerous than the rich, an electoral system that privileges larger groups will, in principle, lead to policies that benefit the poor at the expense of the rich. If this is the case, then democracy might not be entirely bad for Iraqi Sunnis; as demonstrated earlier, the average Sunni now appears to be poorer than the average Shi‘a, and the existing pseudo-democratic arrangement has tended to favor Shi‘a in the redistributive arena. Importantly, since my argument addresses political attitudes, perceptions are key: if respondents believe that democracies redistribute income from the rich to the poor, then it matters fairly little whether or not this belief is actually true.14 As demonstrated earlier, the post-Saddam era has reconfigured the patterns of sectarian economic distribution in Iraq in favor of the Shi‘a. The now-poorer Sunnis, left behind by this new arrangement, might have reason to prefer democracy if they believe that democracy is mainly a mechanism of redistribution from the rich to the poor. Shi‘a, on the other hand, might have exactly the opposite view: if they believe that democracy implies redistribution in favor of the poor (rather than in favor of the majority sect), they may be less inclined to support a democratic regime. If further democratization would imply more substantial redistribution from rich to poor, then Sunnis would, on balance, benefit from these transfers. However, if respondents do not believe that democracies tend to redistribute wealth in this way, then democracy would generally appear to be an almost universally pro-Shi‘a proposition. Thus, a more detailed understanding of the interests of these sects with respect to democracy suggests additional hypotheses: Hypothesis 7a. For Sunnis, mosque attendance will tend to increase support for democracy among respondents who believe that democracies are primarily characterized by redistribution, but decrease support for democracy among those who believe that democracies are primarily characterized by something else. Hypothesis 7b. For Shi‘a, mosque attendance will tend to decrease support for democracy among respondents who believe that democracies are primarily characterized by redistribution, but increase support for democracy among those who believe that democracies are primarily characterized by something else.

 faith in numbers Fortunately, the Arab Barometer asks respondents about their perceptions of what democracy means. Respondents were presented with the following prompt: “There is a difference in opinion among people regarding the most important features of democracy. If you had to choose one, which of the following features would you say is the most important?” The choices included “narrowing the gap between rich and poor” among several other options (elections, freedom to criticize the government, equality of political rights, etc.) The inclusion of this question allows for a direct test of the potentially conditional relationship between communal prayer and democratic support among Iraq’s sects. To allow for the possibility that this effect may vary depending on whether or not the respondent believes that democracies narrow the gap between the rich and the poor, I have coded respondents according to whether or not they named this criterion as the most important feature of democracy. I have then run logistic regressions identical to the ones presented earlier except that in this case, the “most important feature of democracy” variable is added and interacted with communal prayer. The results of these models are presented in Figure 6.4. This figure displays the marginal effect of communal prayer conditional on both sect and belief that democracy is primarily defined by narrowing the gap between the rich and the poor. These results are highly consistent with the expectations of the theory presented. For Sunnis, communal prayer has an anti-democratic effect (of about 14 percentage points) among respondents who do not believe that democracies narrow the gap between the rich and the poor, but has a large positive effect (over 29 percentage points) among those who do believe that this is the key feature defining democracy. These results are supportive of Hypothesis 7a. Once again, the opposite is true among Shi‘a, consistent with Hypothesis 7b: among those who do not believe that democracies are primarily redistributive, communal prayer increases support for democracy by over 12 percentage points, while among those who believe that democracies redistribute, communal prayer decreases support for democracy by close to 22 percentage points. Each of these marginal effects is significant at the 10% level or better.15 This evidence is thus consistent with the claim that communal religious practice tends to promote sect-centered attitudes toward democracy; since Sunnis would presumably benefit from redistributive democracy but would lose from majoritarian democracy, the effect of communal religious practice on support for democracy depends heavily on the respondent’s perception of

representation or redistribution? 135

.5 Marginal Effect of Communal Prayer

.4 .3 .2 .1 0 –.1 –.2 –.3 –.4 –.5 Not Redistributive

Redistributive Shi`a

Not Redistributive

Redistributive Sunni

Fig. 6.4 Marginal Effects of Communal Prayer on Support for Democracy, by Perceptions of Democracy. Note: Figure displays changes in predicted probability of support for democracy comparing a respondent who engages in communal practice with a respondent who does not, i.e., Pr(Support|Practice)− Pr(Support| ∼ Practice), according to both sect and belief that democracies narrow the gap between the rich and the poor. The predictions come from separate logistic regressions for each sect, controlling for the variables listed previously. Bars represent 95% confidence intervals. These calculations use the observed value approach for the control variables; see Hanmer and Kalkan 2013.

what actual changes democracy would bring. The same can be said for Shi‘a, but with the direction of the effects exactly flipped. These results suggest an interesting pattern: communal religious practice can have a strong proor anti-democratic effect for either sect, depending on what the respondent believes to be the primary characteristic of democratic government. As in previous chapters, I also consider the possibility that the relationship between communal religious practice and support for democracy is actually due to greater linkages with political parties among attenders. As Cammett (2011) observes, in divided societies where sectarian organizations (especially political parties) are major service providers, access to services is often a function of party affiliation and participation in high-cost mobilization such as protest. Accordingly, I fit models of the kind used in Figure 6.4 but with party fixed effects and an indicator of protest participation as controls. The results of these models are presented in Table 6A.4. These estimates

 faith in numbers show that including party affiliation and protest behavior as control variables actually increases the size and statistical significance of both the baseline effects of communal prayer and the interactive effects of communal prayer and beliefs about the nature of democracy. These findings lend strong support to the claim that the link between communal prayer and support for democracy is not simply a result of increased sectarian-political linkages between attenders and political parties. The nature of observational public opinion data makes mechanism testing difficult, but it is possible to conduct some basic tests to identify whether or not mosque attendance is, in fact, associated with higher levels of group consciousness; this is the main mechanism that I have suggested. Fortunately, the Arab Barometer contains an additional question in Iraq that can capture (at least to some extent) the political salience of the respondent’s sectarian identity. In this question, respondents were asked: “In general, to what extent is it important to you that the candidate belongs to your sect in deciding who to vote for in elections, whether the parliamentary, municipal or local elections?” For ease of interpretation, I have coded respondents as 1 if they responded “to a great extent” or “to a medium extent” and 0 otherwise. I then fit logistic regression models for each sect, exactly as in the tests previously described, but instead using the new dependent variable. The results displayed in Figure 6.5, are strong and remarkably consistent between the sects: for Sunnis, mosque attenders are 12 percentage points more likely to emphasize the importance of a candidate’s sect than non-attenders (p-value .02); for Shi‘a, the corresponding difference is 12.3 percentage points (p-value .01). Full results of these models are presented in Table 6A.5. This evidence supports the claim that mosque attendance increases the salience of sectarian identity, providing some insight into the channels through which religious participation is translated into views of democracy consistent with one’s group’s interests. To demonstrate the magnitude of the purely sectarian aspects of this mechanism, I fit the same model, but using as the dependent variable a question asking respondents whether a candidate’s piety is important in their voting decision. Presumably, communal prayer should have a large positive effect on this outcome: citizens who participate in communal worship should be considerably more likely to emphasize the importance of a candidate’s piety than citizens who do not participate in communal prayer. Figure 6.6 displays the “relative risks”16 of emphasizing each characteristic, comparing attenders and non-attenders within each sect. This

Marginal Effect of Communal Prayer

.25

.2

.15

.1

.05

0 Shi‘a

Sunni

Fig. 6.5 “Candidate Sect Important,” by Sect and Communal Prayer.

Relative Risk: Communal Prayer/No Communal Prayer

1.6

1.4

1.2

1

.8 Sunni, Sect

Sunni, Piety

Shi‘a, Sect

Shi‘a, Piety

Fig. 6.6 Candidate Sect or Piety Important, by Sect and Communal Prayer.

 faith in numbers figure demonstrates that communal prayer had a far stronger effect on emphasizing a candidate’s sect rather than the candidate’s piety.17 Among Sunnis, attenders were about 1.3 times as likely to emphasize a candidate’s sect as non-attenders; the corresponding ratio among Shi‘a is 1.36. The effect on emphasizing a candidate’s piety is far more muted. Among Shi‘a, attenders were 1.1 times as likely (or 10% more likely) to emphasize a candidate’s piety compared to non-attenders. Surprisingly, among Sunnis, no meaningful difference is detected between attenders and non-attenders on the issue of a candidate’s piety, despite the large difference in emphasis on the candidate’s sect. This evidence suggests that the influence of communal prayer on candidate preferences is more sectarian than religious: worship attendance might make citizens somewhat more concerned with a candidate’s piety,18 but it makes them much more interested in a candidate’s sect.

Conclusion This chapter has provided evidence in support of my theory of religion, group interest, and democracy in the Iraqi context. The evidence presented in this chapter suggests that the communal aspect of religion, in combination with sectarian concerns, has a powerful influence on democratic attitudes in Iraq, and does so according to identifiable and fairly consistent patterns. In contemporary Iraq, communal religious practice can either promote or undermine democratic support depending on the interests of the sect involved. For poorer and outnumbered Sunnis, communal prayer in general tends to promote undemocratic attitudes; however, the exact opposite effect is present among Sunnis who believe that the primary function of democracy is economic redistribution. Among Shi‘a, these patterns are reversed. Since Shi‘a are wealthier, numerically larger, and more powerful in the new political system, communal prayer tends to make Shi‘a more supportive of democracy; once again, however, the direction of this effect is reversed entirely among respondents who believe that democracies are primarily redistributive. As a general rule, majoritarian democracy would tend to benefit Shi‘a, but redistributive democracy would tend to benefit Sunnis. These trends, therefore, are consistent with the claim that communal prayer promotes or impedes support for democracy in accordance with the individual’s group interests, and varies depending on how the individual perceives democracy to function.

representation or redistribution? 139 It is also important to note that the data presented in this chapter were collected prior to the spread of the so-called Islamic State (or ISIS). The emergence of a heavily transnational conflict in which a Sunni extremist organization has taken over vast areas of the country has more than likely transformed the political attitudes and behaviors of members of both sects in considerable ways. In some ways, the behavior of ISIS reflects the logic presented in this chapter taken to an extreme degree: the call for a single state across the region attempts to capitalize on the Sunni numerical advantage in most of the Arab world, in resistance to the Shi‘a dominance of post-Saddam Iraq. For ordinary Sunnis, incentives may be unclear: neither the radical violence of ISIS nor the neglect of the Shi‘a regime is desirable. Likewise, Shi‘a interests with respect for democracy may have changed. Although they represent a solid majority in Iraq, Shi‘a are a fairly small minority when the Arab world is considered as a whole; consequently, majoritarian democracy looks quite different when the boundaries of states break down. This chapter has demonstrated the importance of examining individual contexts at a particular point in time. When the data used in this chapter were collected, representation and redistribution were two key questions facing Iraq’s emerging state. The chapter has shown how differing perceptions of these issues led to seemingly contradictory effects of religious attendance on attitudes toward democracy. In doing so, it has highlighted the importance of acknowledging democracy’s different aspects; the winners from free elections might nevertheless be the losers of redistribution.

 faith in numbers

Appendix 6A: Supplementary Tables and Figures

Table 6A.1 Support for Democracy by Communal Prayer, Sunni Strongly Supports Democracy Communal Prayer No Communal Prayer Communal Prayer Total

No

Yes

Total

117 (63%) 193 (73%) 310 (69%)

68 (37%) 70 (27%) 138 (31%)

185 (100%) 263 (100%) 448 (100%)

Table 6A.2 Support for Democracy by Communal Prayer, Shi‘a Strongly Supports Democracy Communal Prayer No Communal Prayer Communal Prayer Total

No

Yes

Total

198 (73%) 130 (68%) 328 (71%)

75 (27%) 61 (32%) 136 (29%)

273 (100%) 191 (100%) 464 (100%)

representation or redistribution? 141 Table 6A.3 Baseline Results, Support for Democracy (1) Sunni Communal Prayer Qur’an Reading

(2) Shi‘a

−.60∗∗ (.26)

.42∗ (.24)

.55∗∗ (.26)

.01 (.26)

Female

−.27 (.25)

.39∗ (.22)

Education Level

−.06 (.08)

−.05 (.08)

Income Level

.36∗∗∗ (.11)

.03 (.11)

Feels Safety Ensured

.75∗∗∗ (.29)

.46∗ (.25)

Kurd

.89∗∗∗ (.29)

.88 (1.12)

Age Group

.09 (.11)

−.06 (.10)

−.46 (.28)

−.62∗∗∗ (.23)

Constant

−1.71∗∗ (.69)

−1.60∗∗∗ (.61)

Observations Pseudo R2 AIC

424 .094 494.26

454 .028 543.70

Feels Treated Equally

Note: Standard errors in parentheses. ∗ p < .1, ∗∗ p < .05, ∗∗∗ p < .01.

 faith in numbers Table 6A.4 Interactive Models, Support for Democracy (1) Sunni

(2) Shi‘a

(3) Sunni

(4) Shi‘a

Communal Prayer

−.72∗∗∗ (.26)

.63∗∗ (.25)

−1.27∗∗∗ (.35)

.69∗∗∗ (.26)

Democracies Narrow Gap

−2.84∗∗∗ (1.01)

.39 (.48)

−3.94∗∗ (1.96)

.43 (.49)

Communal Prayer × Democracies Narrow Gap

2.91∗∗

−2.15∗∗

4.80∗∗

−2.83∗∗∗

(1.18)

(.83)

(2.14)

(1.02)

Qur’an Reading

.54∗∗ (.26)

.04 (.27)

.88∗∗ (.35)

.11 (.28)

Female

−.23 (.25)

.37 (.23)

.17 (.33)

.35 (.24)

Education Level

−.08 (.08)

−.02 (.09)

−.06 (.10)

−.05 (.09)

Income Level

.32∗∗∗ (.11)

−.01 (.11)

.39∗∗∗ (.15)

−.02 (.12)

Feels Safety Ensured

.80∗∗∗ (.29)

.33 (.26)

1.30∗∗∗ (.37)

−.00 (.28)

Kurd

.94∗∗∗ (.30)

.75 (1.11)

.83 (.51)

.22 (.99)

.10 (.11)

−.01 (.11)

.18 (.13)

.03 (.11)

−.62∗∗ (.29)

−.66∗∗∗ (.24)

−.63∗ (.38)

−.68∗∗ (.27)

2.14∗∗∗ (.36)

.16 (.40)

Age Group Feels Treated Equally Protest Constant Observations Pseudo R2 AIC

−1.49∗∗ (.69)

−1.63∗∗∗ (.62)

−4.13∗∗∗ (.97)

−1.46∗∗ (.65)

424 .113 488.52

454 .045 538.54

393 .363 364.60

416 .076 512.27

Note: Standard errors in parentheses. Models 3 and 4 include party fixed effects. ∗ p < .1, ∗∗ p < .05, ∗∗∗ p < .01.

representation or redistribution? 143 Table 6A.5 Baseline Models, Candidate Sect Important (1) Sunni

(2) Shi‘a

Communal Prayer

.52∗∗ (.23)

.57∗∗ (.23)

Qur’an Reading

.06 (.23)

−.99∗∗∗ (.25)

Female

−.36 (.23)

.07 (.22)

Education Level

−.08 (.07)

.16∗ (.09)

Income Level

−.06 (.09)

−.37∗∗∗ (.11)

Feels Safety Ensured

−1.14∗∗∗ (.28)

−.59∗∗ (.25)

Kurd Age Group

.44 (.29)

−.41 (.95)

−.13 (.09)

.01 (.10)

Feels Treated Equally

.60∗∗ (.26)

.50∗ (.26)

Constant

.98 (.64)

.50 (.58)

421 .064 556.10

417 .081 538.98

Observations Pseudo R2 AIC Note: Standard errors in parentheses. ∗ p < .1, ∗∗ p < .05, ∗∗∗ p < .01.

 faith in numbers Shi‘a

Sunni

Proportion (Strongly Support Democracy)

.4

.3

.2

.1

0 No Communal Communal Prayer Prayer Democracies Don't Narrow Gap

No Communal Communal Prayer Prayer

No Communal Communal Prayer Prayer

Democracies Narrow Gap

Democracies Don't Narrow Gap

No Communal Communal Prayer Prayer Democracies Narrow Gap

Fig. 6A.1 Conditional Probabilities of Support for Democracy. Note: These figures are conditional probabilities not controlling for any other variables (the graphical equivalent of a crosstab); subsequent models control for several other variables.

7 Conclusion Implications for Religion and Politics

Introduction To this point, I have shown that my theory of communal religious practice, sectarian interests, and democratic attitudes helps to explain the varying effects of religious practice on support for democracy in Lebanon and Iraq. It is possible, of course, that these cases are idiosyncratic, and that the theory only fits them well because of the peculiarities of the Arab region or the highly sectarian political systems found in these countries. I will suggest that the theory is not so limited. In this chapter, I demonstrate that the theory also finds support in countries outside of the Arab world that are not necessarily subjected to the same degree of sectarian tension found in that region. Using evidence from cross-national data, I show that the theory applies to a wide variety of settings. To test the theory discussed in previous chapters in a broader sample of countries, I rely on data from five waves of the World Values Survey.

Cross-National Tests My theory of communal prayer, sectarian interests, and democratic attitudes has found considerable support in Lebanon and Iraq. At this point, it is worth examining how well this theory “travels.” In this section, I provide— cautiously—broad tests of the theory in an attempt to give a “view from 30,000 feet.” It is not possible from a practical perspective to test the theory in every country in the way that I have done for the cases discussed in the preceding chapters; instead, I provide suggestive evidence that the effect of communal prayer varies according to country and sect in a way that is consistent with the sectarian interests logic that I have presented. While this evidence helps to address the issue of portability with respect Faith in Numbers. Michael Hoffman, Oxford University Press (2021). © Oxford University Press. DOI: 10.1093/oso/9780197538012.003.0007

 faith in numbers to the theory, it should nevertheless be interpreted with a good degree of caution, given the inevitable coarseness of the analysis of so many settings simultaneously. I use a series of simple tests to assess the theory in a cross-national setting. I use data from the first five waves of the World Values Survey (1981–2007). For each country in each wave, I use the data to estimate the shares of each sectarian group within the country, providing an approximation of its share in the population. The data set includes 87 countries across five waves, and a total of 1,158 country-wave-sect combinations. The main dependent variable is an index of support for democracy rescaled to range between 0 and 1. The quantity of interest is the effect of communal prayer on support for democracy; communal prayer is measured by asking respondents, “How often do you attend religious services?” and is likewise rescaled to range from 0 to 1. The main model is an ordinary least squares regression of support for democracy on the interaction between group size and frequency of communal prayer along with a set of control variables.1 The parameters of interest are the coefficient on communal worship and the interaction between communal worship and group size. The theory suggests that for larger groups (i.e., those who would be more competitive in democratic elections), communal prayer should, other things being equal, have a prodemocratic effect. For smaller groups, who would be less likely to win elections and whose interests might be threatened by majority rule, the opposite should be true. Figure 7.1 displays the marginal effect of communal prayer on support for democracy across the range of group population shares. The y-axis represents the difference in support for democracy comparing the most frequent attenders to the least frequent attenders. For sects that represent only a small portion of their country’s population, communal prayer has a negative and statistically significant effect on support for democracy. Conversely, for the largest groups, communal prayer increases support for democracy, and this difference is statistically significant. The effect sizes are quite modest (just over one percentage point at the highest and lowest group sizes), but it is worth noting that this type of test is rather difficult for this theory. The fact that significant differences emerge according to the pattern expected by my theory is encouraging, and suggests that it has at least some general explanatory power.

conclusion 147

Marginal Effect of Communal Prayer

.02

.01

0

–.01

–.02 0

.1

.2

.3

.4 .5 .6 Group Proportion

.7

.8

.9

1

Fig. 7.1 Marginal Effects of Attendance, by Group Size.

However, a test as broad as this one involves significant challenges in the area of interpretation. Perhaps the most difficult challenge emerges from the fact that these tests involve countries that possess characteristics that make the theory applicable as well as those that do not. Since many of the countries in the sample are religiously homogeneous, secular, and/or solidly democratic, many of the cases included in these models are environments where the theory has little applicability. It does not, for instance, seem sensible to apply the theory to Sweden, a consolidated democracy where only 9% of the population are regular church attenders and 42% are not affiliated with a religion at all.2 Thus, it is important to consider some scope conditions for the theory. Two factors are particularly likely to affect the applicability of this theory to particular countries: levels of democracy and the salience of religious identity. In countries where democracy is “the only game in town,” regime preferences are less likely to be affected by religious group interests because citizens have experienced democracy and, often, only democracy. In most of these cases, the idea of an autocracy in which religious groups are afforded certain privileges of inclusion while others are left out may not seem like a plausible option, and therefore may not enter into the mindsets of citizens.

 faith in numbers Likewise, it is intuitive that in countries where religion is not a salient social divide, either because the country is religiously homogeneous, few citizens identify with a religious tradition, or religious identities are simply not politically meaningful, this theory would perform poorly. If religious groups are not a significant source of identity, the theory presented in this book would not be applicable. Using existing data, it is possible to examine these scope conditions empirically, if somewhat superficially. Figure 7.2 attempts to do so by dividing countries according to their level of democracy and a proxy for the political salience of religious divisions. In the top two sub-graphs, countries are divided into “high” and “low” levels of democracy3 based on their scores on the Varieties of Democracy (V-Dem) Liberal Democracy Index (Coppedge et al., 2017). As these figures demonstrate, in “high democracy” countries, very little effect is present. Though the slope of the line is in the expected direction, communal prayer does not have a statistically significant effect at any level of group size, and the effect is not statistically distinguishable when comparing large groups to small ones. In non-democracies, however, the theory performs as expected: communal prayer substantially decreases support for democracy among small groups and increases support for democracy among large groups. For this group of countries, the interaction between group size and communal prayer is highly statistically significant (p < .001), and the magnitudes of the effects are much larger than they were for democracies: among the smallest groups, the effect size is more than six times as large in non-democracies as in non-democracies, and the effect size is nearly twice the size among large religious groups in non-democracies as compared to those in democracies. As expected, the theory appears to receive much more support in nondemocracies than in democracies. The salience of religious identity is much more difficult to measure, and the bottom two graphs should be interpreted with even greater caution than the top two. However, as a proxy for salient religious divisions, I use the measure of religious polarization4 introduced by Reynal-Querol (2002). This measure is intended to capture “latent religious conflict” and is a strong predictor of civil war (Reynal-Querol, 2002); it is therefore a better proxy for politically salient religious divisions than religious fractionalization, which only measures religious diversity. Here, a trend emerges similar to the one witnessed for democracy: in religiously polarized countries, communal prayer enhances support for democracy among large groups but suppresses

1

Marginal Effect of Communal Prayer

1

Marginal Effect of Communal Prayer

Marginal Effect of Communal Prayer

Marginal Effect of Communal Prayer

conclusion 149 High Level of Democracy .04 .02 0 –.02 –.04 –.06 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

High Level of Religious Polarization .04 .02 0 –.02 –.04 –.06 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

Low Level of Democracy .04 .02 0 –.02 –.04 –.06 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

1

Low Level of Religious Polarization .04 .02 0 –.02 –.04 –.06 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

1

Fig. 7.2 Conditional Effects by Levels of Democracy and Religious Polarization.

it in small groups (Group Size*Communal Prayer interaction term p-value = .001), but for countries with low levels of religious polarization, communal prayer’s effect is negligible and does not depend on group size (interaction term p-value = .851). Of course, religious polarization as measured by Reynal-Querol (2002) is not a perfect measure of the salience of religious identity because it is still based entirely on the sizes of religious groups. As an attempt to measure how much religious factors seem to matter in each country, I calculated the proportion of respondents in each country from the World Values Survey who (a) stated that they considered themselves “religious”; and (b) reported being a member of a religious group. Figure 7.3 reports the effects of communal prayer by group size when dividing countries according to average levels of religiosity and religious affiliation. As with the previous set of graphs, the trend is clear: in countries where large numbers of citizens are religious and/or are members of a religious denomination,5 the theory has some predictive power; in countries where relatively few people consider themselves religious and/or relatively few people are members of religious groups,6 the theory performs poorly. These findings should be interpreted with caution. My tests highlight the common trade-off between internal and external validity; this section

1

Marginal Effect of Communal Prayer

High Religiosity Countries

1

Marginal Effect of Communal Prayer

Marginal Effect of Communal Prayer

Marginal Effect of Communal Prayer

 faith in numbers

.06 .04 .02 0 –.02 –.04 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

High Religious Affiliation Countries .06 .04 .02 0 –.02 –.04 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

Low Religiosity Countries .06 .04 .02 0 –.02 –.04 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

1

Low Religious Affiliation Countries .06 .04 .02 0 –.02 –.04 0

.1

.2

.3 .4 .5 .6 .7 Group Proportion

.8

.9

1

Fig. 7.3 Conditional Effects by Levels of Religiosity and Religious Affiliation.

gains broad coverage at the expense of thoroughness. The pooled data used to estimate these models mask a considerable amount of heterogeneity, and while I have tried to account for as much of that heterogeneity as possible using control variables, there is no question that many factors remain unaddressed. Indeed, it has been a central argument of this book that the details of each case (both place and time) will play a major part in determining how religious attendance will affect attitudes toward democracy. Knowing which groups are involved in the political battles—and what they are battling over—is a vital part of a richer test of theories like this one. The case studies presented in earlier chapters provide a richer analysis of how these mechanisms operate in particular settings, and that richness simply cannot be provided in broad cross-national tests. It is nevertheless valuable to note that the theory has performed well in these cross-national tests despite the low resolution of the pictures they provide. The wide variety of peculiarities across these many cases is likely to wash out a considerable amount of the effect of communal prayer. The fact that any effects remain after so much dilution provides meaningful support for the theory. These findings indicate that the findings presented in earlier chapters are not simply artifacts of the few cases I have addressed in

conclusion 151 detail; the conditional relationship between communal prayer and support for democracy seems to be present in many parts of the world.

Findings of the Book This book has shown that the relationship between religious practice and democratic attitudes is complicated, but not random. Communal worship can either enhance or suppress support for democracy, and the direction of this relationship depends in large part on sectarian interests. For “winners” of democracy, communal prayer tends to heighten enthusiasm for democracy, while the exact opposite pattern is present among those whose groups would lose privileges in the event of democratization. Moreover, the book has provided evidence suggesting that these effects are not driven purely by self-selection. Experimental and quasi-experimental tests have shown that communal religion shifts individuals’ orientations toward their sect and, by extension, toward issues of democratization and redistribution. The effects of communal prayer are both substantial and real. In Lebanon, changing political circumstances have highlighted how this relationship is not static. Prior to the outbreak of the Syrian Civil War, and the intensification of Sunni-Shi‘a tensions that came with it, the prospect of democracy was more likely to be viewed as a Christian-Muslim issue. In this environment, the politically and economically favored Christian groups in the country had reasons to oppose democratization. Changes to the allocations of seats in political positions along sectarian lines—a key feature of the most likely form of democratization in Lebanon—would undoubtedly result in a decline in Christian influence to the benefit of Muslims, both Sunni and Shi‘a. My findings from 2007 demonstrate that communal religious practice, in accordance with these sectarian interests, increased support for democracy among Muslims in this environment, but had the opposite effect among Christians. The onset of the Syrian Civil War dramatically altered the political relationship between Sunnis and Shi‘a, removing any possibility of cooperation between the groups, at least in the short term. Christians, divided between the two main camps, found themselves largely on the sidelines of the conflict. In this setting, the main axis of political competition was within the Muslim community, pitting the poorer and disenfranchised Shi‘a against the (relatively) wealthier and better-represented Sunnis. Results from my

 faith in numbers original survey in 2014 indicate that for Shi‘a, communal religious practice has a powerful pro-democratic effect, but the opposite is true for Sunnis. This dramatic shift in the relationship between religious practice and regime preferences among Sunnis highlights the importance of political context: when the factors determining sectarian interests change, the link between religious behaviors and individual political preferences changes along with them. Evidence from Iraq provides an illustration of the multidimensionality of attitudes toward democracy. In Iraq, post-Saddam reforms have left Sunnis in a relatively underprivileged position, both politically and economically. Democratic and quasi-democratic measures have made it difficult for them to compete effectively in elections due to their smaller numbers, and they no longer enjoy the privileges allotted to them under the prior authoritarian regime. Consequently, many Iraqi Sunnis are skeptical toward democracy. Communal prayer heightens these sentiments among those with a political understanding of democracy. For Sunnis who believe that democracy is primarily about elections, protection of rights, or other political issues, communal prayer depresses support for democracy. Importantly, however, the opposite is the case among Sunnis who believe that democracy is primarily about economic redistribution. Since Sunnis are now poorer than Shi‘a, they would benefit disproportionately from redistribution. As a result, when individual preferences are pushed more in the direction of “Sunni interests,” Sunnis will tend to favor democracy if they believe it is mainly redistributive in nature. The exact opposite trend is present for Shi‘a. In short, communal worship increases support for democracy among Shi‘a who believe that democracy is political, but weakens democratic support among Shi‘a who perceive democracy to be primarily redistributive. Thus, an important implication of this chapter and the book more broadly is that the effects of religious practices on attitudes toward democracy depend not only on sectarian interests, but also on perceptions of what democracy means. Evidence from outside the Middle East provides broader coverage for the theory. This book has provided cross-national tests as a coarse indication that this relationship might be fairly widespread. Using survey data from several dozen countries, it shows that among large groups (i.e., those most likely to win competitive elections), communal prayer raises support for democracy, while for smaller groups, such behaviors lower support for democracy. The evidence suggests that these relationships are not isolated to a few countries, but rather, have considerable support in a number of different settings,

conclusion 153 though certainly not all of them. These effects appear to be present (only) in countries where religious divisions are salient and where democracy is not firmly established.

Implications Simplistic narratives regarding religion’s consistently pro- or antidemocratic effects have found little support. Theology per se—not examined in much detail here—may matter, but it does not seem to account for the bulk of the effect of religious factors on individual attitudes toward democracy. Readers interested in the ongoing debate regarding Islam and democracy should take note of these findings. While in some cases discussed in this book, Muslim religious practice is associated with lower levels of support for democracy, there are just as many cases in which communal prayer enhances democratic values among Muslims. In some cases, Sunnis in a given context witness effects the opposite of those experienced by Shi‘a in the same country. These divergent trends indicate that the reason for the differential effects is not theological; after all, Sunni-Shi‘a theological differences are quite mild compared to differences between Islam and Christianity or Hinduism. I have argued that the gap in our collective understanding of these differences can be filled—at least partially—by a greater acknowledgment of the importance of sectarian interests. In most cases, religious practices can serve either pro- or anti-democratic purposes irrespective of their theological content. A focus on the missing political considerations demonstrates why religion is often found to be politically ambiguous (Philpott, 2007). The cases discussed in this book are merely a sampling of the potential applications of this theory. The theory is likely to hold some explanatory power in countries where political and economic resources are distributed along sectarian lines, and where democracy is still a matter of dispute; that is, it is not “the only game in town” (Linz and Stepan, 1996). The applicability of the theory is certainly determined in part by the salience of sectarian identity. We should not expect the theory to fit very well in cases where sectarianism does not exist, or where it is not a strong political identity. However, as the results presented previously have shown, the conditioning of communal prayer’s effect on democratic attitudes by sectarian interests seems to be a fairly widespread phenomenon.

 faith in numbers There are several implications of these findings for both future scholarly work and policymaking. The first, most obvious, and perhaps most important is that we must avoid the temptation of dismissing religion. Other scholars have lamented the lack of attention paid to religion in comparative politics (Bellin, 2008; Grzymala-Busse, 2012b; Gill, 2001), and this study has shown that in the area of regime preferences, religious factors can have considerable effects. In a variety of settings, it is clear that religious variables (frequency of attendance, strength of identification, or even simple religious primes) account for enormous differences between individuals’ levels of support for democracy. Regardless of whether one thinks that religion should matter in this arena, the evidence suggests overwhelmingly that religion does matter. Second, as mentioned previously, theological explanations offer (at best) an incomplete picture of the link between religion and democratic attitudes. In different political and economic environments, the same religious tradition can have either a strong pro-democratic or an equally strong anti-democratic effect. The experience of politics in an individual’s country plays a large part in conditioning the direction and magnitude of the effect of religious practice on her attitudes toward democracy. The way that individuals think about democracy matters quite a bit. If democracy is thought of as an institutional configuration, group size and status are especially important; if it is assumed to imply economic redistribution, then group resource disparities might be more important. On some level, this finding should be encouraging for policymakers. While it is not realistic to change the theological content of religious groups, it is much more feasible to adjust their material and political incentives with respect to democracy. The evidence presented here suggests that if policymakers can manage to make democracy more appealing to religious groups, then religious practice can be effectively harnessed to build support for democracy. Of course, such adjustments come at a cost. Since democracy inevitably creates winners and losers, it will be necessary to ensure that the changes made to redistributive or political structures do not threaten the interests of other groups so much that it cancels out the pro-democratic effects witnessed by the beneficiary groups. Many of the challenges experienced during this process are similar to those found in other areas of democracy promotion—namely, creating more “winners” from democracy without also adding “losers”—but my findings provide a new avenue for democracy-promoters to explore in mobilizing support for democracy.

conclusion 155 Third, the influence of formal religious institutions—often held to be rigid and slow-moving—can change fairly rapidly. The experience of Lebanon demonstrates how for certain groups (most prominently, Sunnis), the effect of communal religious practice can shift dramatically over a period of only a few years. The political effects of communal prayer are highly responsive to changing political conditions, and these effects should not be assumed to be stable in the long term. Rather, attention must be paid to the ways in which changing political policies or the emergence of new issues might disrupt the existing link between religious practices and regime attitudes. There are some ways in which the theory presented here can be extended to other types of identities. When certain activities heighten the salience of politically relevant identities in divided countries, a similar logic might account for differences in attitudes toward democracy, redistribution, or other political issues. Such a pattern is especially likely when political resources and/or interests map closely onto seemingly non political group identities. However, religion is a natural starting point for a theory of this kind. The particular strength of religious identity, which is often stronger and less fungible than other identities (Grzymala-Busse, 2012b), makes it an especially likely candidate for considerable political influence. Furthermore, religious identity and practice are abundant throughout the world, and the sheer number of individuals who engage in communal worship suggests that religion should have a powerful effect on social identities and political preferences. Nevertheless, future work should examine when and how other forms of identities operate in similar ways, and what the implications of these identities are for democratic attitudes. Importantly, the findings presented in this book are bounded and contingent. Religion, undoubtedly, affects political preferences in numerous ways, including many channels addressed by the important existing literature discussed earlier in the book. When religious identity is not salient, where the population is religiously homogeneous, or where democracy is firmly established, the theory I have presented is likely to be of limited value. In these cases, better explanations of the role of religion in promoting regime preferences will be found elsewhere.

 faith in numbers

Conclusion Religion remains a complicated but deeply powerful force in politics in most parts of the world. Importantly, it remains poorly understood. The findings presented in this book suggest the need to move away from traditional understandings of religion as an irrational (and often unpredictable) influence on political attitudes and behaviors. This study has refocused discussions of religion and politics toward both affective and instrumental mechanisms. Religion’s influence on politics is about much more than theology; in fact, political factors themselves play a large role in how religious behaviors translate into political attitudes. The implications for democracy, as shown in this study, can run the gamut from threats to democracy to active mobilization in its favor. Recent estimates suggest that nearly 84% of the world’s population identifies with a religious group (Pew Forum, 2012). Three-fourths of people report that they receive comfort and strength from religion, and over twothirds of people state that religion is quite important in their lives.7 In some ways, the fact that the results of this study have indicated a major role of religion in political behavior should come as no surprise. What is perhaps more surprising is that this role is variegated, yet predictable. If, as scholars and policymakers, we choose to downplay the importance of religion, we do so at our own peril. The patterns presented in this book should give hope to those puzzled by religion. Rather than demonstrating a random, scatter shot series of effects on political attitudes and behaviors, religious practices often have systematic (and even rational) effects on individuals’ political beliefs. As long as religion remains one of the most potent drivers of political outcomes, it is far better to understand religion than to dismiss it. This book represents a modest step forward in that understanding.

conclusion 157

Appendix 7A: Additional Information, Cross-National

Table 7A.1 Countries Included in Cross-National Tests Albania Algeria Andorra Argentina Armenia Australia Azerbaijan Bangladesh Belarus Bosnia and Herzegovina Brazil Bulgaria Burkina Faso Canada Chile China Colombia Croatia Cyprus Czech Republic Dominican Republic Egypt El Salvador Estonia Ethiopia Finland France Georgia Germany

Ghana Great Britain Guatemala Hong Kong Hungary India Indonesia Iran Iraq Israel Italy Japan Jordan Kyrgyzstan Latvia Lithuania Macedonia Malaysia Mali Mexico Moldova Morocco Netherlands New Zealand Nigeria Norway Pakistan Peru Philippines

Poland Puerto Rico Romania Russian Federation Rwanda Saudi Arabia Serbia Serbia and Montenegro Singapore Slovakia Slovenia South Africa South Korea Spain Sweden Switzerland Taiwan Tanzania Thailand Trinidad and Tobago Turkey Uganda Ukraine United States Uruguay Venezuela Vietnam Zambia Zimbabwe

Notes Chapter 1 1. 2. 3. 4.

Vienna Communique on Syria, October 30, 2015. See, for example, Huntington (1991). See, for example, Putnam and Campbell (2010). See Iannaccone (1997) for a review of rational-choice approaches to the study of religion. 5. The works presented here are but a small sampling of the rich literature linking rational choice theory to religion. Other examples of this scholarship examine religious groups as strategic, interest-driven actors (Warner, 2000; Ekelund et al., 1996) or consider the rational reasons for the success of certain types of religious groups (Iannaccone, 1994; Atran, 2002), among other topics. 6. See also Trejo (2009). 7. Note that while the emphasis here is on the ways in which religious factors influence political outcomes, there may also be a reciprocal relationship at the individual level. See, for example, Karakoç and Ba¸skan (2012); Solt et al. (2011); Grewal et al., (2019); Stark and Finke (2000).

Chapter 2 1. Even The American Voter limits its consideration of within-denomination religious factors to “high” versus “low” identification; individual religious behaviors are not examined in their discussion of religious minorities, particularly Catholics and Jews (Campbell et al., 1960). 2. See Scheve and Stasavage (2006); McClendon and Riedl (2015); De La O and Rodden (2008). The most common argument for this relationship is that religious communities provide an insurance mechanism in the event of adversity; hence, religious individuals do not need to rely on the welfare state as much as other citizens (Dehejia et al., 2007). 3. Important exceptions include Meyer et al. (2008). 4. Meyer et al. (2008) also suggest that religious organizational structures might affect support for democracy, a claim worthy of further examination in cross-national settings. 5. Examples include Canetti-Nisim (2004); Meyer et al. (2008); Valenta and Strabac (2012).

 notes 6. Some existing studies (e.g., Smidt, 1999) allow for the possibility that behavioral and denominational effects may differ across countries, but ignore the possibility that religious behaviors may affect political outcomes differently between sects within the same country. Lijphart’s (1977) classic study considers the interaction between sect and religious behavior across countries, but does so in a largely descriptive rather than theoretical way. The overwhelming tendency in this line of research (both American and comparative) is either to presume similar effects across religious groups, or to offer little theoretical account of why these effects differ other than sui generis characteristics of particular sects (Verba et al., 1995; Campbell, 2004; Pettersson, 2009). 7. For William James (2012 [1902]), the fundamental features of religion are personal and private: his definition of religion is “the feelings, acts, and experiences of individual men in their solitude, so far as they apprehend themselves to stand in relation to whatever they may consider the divine.” In his account, the social aspects of religion (as well as theology) are secondary; they are byproducts of these solitary experiences. This definition is similar to Alfred North Whitehead’s: “Religion is what the individual does with his own solitariness.” 8. See Grzymala-Busse (2012a) for a discussion of some of the potential benefits of religious communities, including material and emotional support in addition to spiritual capital. 9. Like Rabushka and Shepsle (1972), I do not provide an explanation as to why some societies are more divided than others, but instead focus on the consequences of these divisions. 10. Note, however, that majoritarianism represents only one interpretation of democracy; citizens’ responses to alternative features of democracy will be explored in the empirical chapters. 11. These predictions align with Jamal and Tessler (2008), who argue that citizens in the Arab world often view democracy instrumentally, and support it insofar as it serves their interests. 12. Salti and Chaaban (2010) demonstrate that redistribution closely follows sectarian patterns in Lebanon, a pattern witnessed in many other countries as well. 13. Such works include Kalyvas and Van Kersbergen (2010); Toft et al. (2011); Driessen (2014). 14. See Grzymala-Busse (2016a, 2016b) for a discussion of how doctrine influences political outcomes in many settings. 15. See Chandra (2012, 2006); Posner (2005); among others. 16. It is important to note that violent conflict is not necessary for this pattern to emerge. The existence of political competition between groups should be more than sufficient to produce the type of behaviors suggested; Tajfel and Turner (1979) and others have found that even arbitrary group distinctions tend to lead to in-group bias and outgroup derogation. 17. For a general critique of the de-emphasis on deities, see Stark (2010). 18. See Brewer et al. (2003); Valenzuela (2011). 19. For examples, see Picard (1993); Ayoob (2004); Norton (2005); Davis (2005).

notes 161 20. For example, Iannaccone and Berman (2006); Ginges et al. (2009). 21. For instance, it is common for Lebanese preachers to refer to particular regions or neighborhoods as a shorthand for a given sect; a Shi‘a preacher’s reference to “the south” will often carry undertones implying that he is referring specifically to Shi‘a. 22. Empirical research supports these claims. Dana et al. (2011) observe that among Muslims in the United States, mosque attendance decreases identification with national origins, but increases identification as Muslims. Further, these authors observe that Muslims who are highly engaged in their mosques are four times as likely to believe that they have a great deal in common with other American Muslims as are the least engaged. Calhoun-Brown (1996) notes a similar trend among African American Christians, noting that church attendance is strongly negatively associated with racial identification. She suggests that “otherworldly concerns may prompt the respondent to think less about being black and more about being Christian” (Calhoun-Brown, 1996, p. 948). These examples indicate that communal prayer primarily builds denominational (sectarian) identity rather than nonreligious identities; in fact, this increase in sectarian identity may come at the expense of racial or ethnic identification. 23. See Shariff and Norenzayan (2007); Pichon et al. (2007); Randolph-Seng and Nielsen (2007); Carpenter and Marshall (2009). 24. A recent exception is Lindqvist and Östling (2013). 25. See Keely and Tan (2008); Klor and Shayo (2010). 26. There are certainly many other reasons for the lack of a one-to-one correspondence between income and preferences, but attention to identity concerns is of great importance. 27. See Corneo and Grüner (2002); Alesina and Giuliano (2009); Isaksson and Lindskog (2009). 28. See also Grzymala-Busse (2012b) for an excellent discussion of the unique features of religion. 29. See Smolicz (1981); Fox (2004); Harris (2007); Svensson (2007); Juergensmeyer (2008) for similar findings in other (sometimes comparative) contexts. 30. See also arguments describing religion as a particularly powerful group identity because it is a “hard-to-fake sign of commitment” and a “costly signal” (Irons, 2001; Sosis and Bressler, 2003; Atran, 2002). 31. Importantly, Grzymala-Busse (2015) argues that moral authority is highest where national and religious identities are fused together. While her study focuses on states, substate identities (national or otherwise) are likely to produce a similar logic. In sectarian societies, such as those addressed in this book, religious and social identities are tightly linked. Consequently, the moral authority of religious groups should be strong in these settings. (See also Grzymala-Busse, 2016c).

 notes

Chapter 3 1. Unlike the Lebanese sample, the Iraqi interviews are not strictly nationally representative. However, respondents were chosen to represent the demographic diversity of the country and provide a wide-reaching view of the variety of religious experiences among Iraqi citizens. 2. Some of the quotations in this chapter have been lightly edited for grammar. 3. Many Muslims believe that this multiplicative benefit is only present for men, while women should be encouraged to pray at home instead of the mosque. Subsequent chapters will demonstrate that, while women go to mosque in lower numbers than men, plenty of women nevertheless attend mosque regularly. 4. Due to the nature of the sample in Iraq, interviewees’ sectarian affiliations were not always identified. 5. This statement is a thinly veiled accusation of idolatry; Islamic theology emphasizes a firm monotheism and strongly discourages most forms of devotion to anyone or anything besides God. Sunnis often accuse Shi‘a of practicing a soft form of polytheism because of their particular reverence for Muhammad’s family. 6. Aisha was a wife of Muhammad who is often viewed negatively by Shi‘a due to her role in the original Sunni-Shi‘a split following Muhammad’s death. 7. This statement is particularly interesting because sectarian quotas in the Lebanese parliament are firmly established. The religious leader to whom the interviewee was referring in this case was discouraging worshipers from voting for candidates from other sects despite the fact that no sect can plausibly hope to gain any additional seats in any particular election. 8. As mentioned previously, because of the nature of the Iraqi sample, respondents’ sectarian affiliations were not always available. Consequently, this section only addresses the Lebanese case. 9. Historically, “poor areas” would be particularly likely to be Shi‘a areas. This heuristic is less true than it would have been in decades past, but the implication of this shorthand is still highly sectarian. 10. It is interesting, though not entirely surprising, that multiple respondents used this exact phrase.

Chapter 4 1. Cammett (2013b, p. 132) describes sectarian affiliation in Lebanon as “a virtually immutable characteristic inherited by blood.” 2. Fortunately, a number of excellent accounts of the war already exist. See, among others, Fisk (2001); Cooke (1987); Johnson (2001); Trablousi (2007); Mackey (1989). 3. Shi‘a almost certainly account for considerably more of the population than this figure indicates, suggesting that overall Muslim underrepresentation is even more severe than demonstrated by these numbers. 4. Remark to author, Beirut, March 2014.

notes 163 5. Furthermore, as Salloukh and Verheij (2017, p. 155) note, the Maronite presidency retained considerable executive powers even after the Ta’if Accord. 6. Again, as will be discussed in subsequent pages, Shi‘a are in reality even more underrepresented than these figures indicate. 7. Remark to author, Beirut, March 2014. 8. Remark to author, Beirut, March 2014. 9. The first wave of the Arab Barometer, the only survey of its kind in Lebanon during this period, contains relatively small numbers of Sunni and Shi‘a respondents. This small sample limits researchers’ abilities to consider the sects separately, but the survey itself is nevertheless an invaluable resource for studying historical political attitudes in a country where data on such attitudes are scarce, to say the least. 10. It is also important to recognize that these data were collected in 2007, before Hezbollah (which is overwhelmingly Shi‘a) had gained much of the political ground that it has in recent years. This is not to say that Hezbollah was an entirely marginalized player in 2007. However, its representation in government was not at the level it would reach in later years. Hezbollah’s representatives in the cabinet, included in the 2005 government as part of a cross-sectarian compromise, resigned in late 2006, and the party would not have any representation in the cabinet until the formation of a new government in 2008 after the Doha Accords. Thus, even if Sunnis might now fear democracy because it could imply Shi‘a dominance, this fear is certain to have been far less significant at the time when these data were collected. My analysis of data from 2014 (in a subsequent chapter) will separate Sunnis and Shi‘a because of the substantial widening of the rift between these groups during the period between the two surveys. For media discussion of the growing divisions between the sects, see “Lebanon’s Dangerous Sunni-Shiite Divide Widens,” al-Monitor, May 29, 2012. 11. Available at https://d3n8a8pro7vhmx.cloudfront.net/aai/pages/9759/attachments/ original/1431961485/UnityInLebanon_2005.pdf?1431961485. 12. “Mohammed Hussein Fadlallah Continues to Serve Lebanon,” al-Manar, February 12, 2012. 13. These and many other statements emphasizing both Islamic unity and democracy are found at the official website of his mosque: http://bayynat.org.lb/. 14. See “Maronite, Orthodox Electoral Law Draws Outrage,” Daily Star, December 21, 2011. 15. It is surely not an accident that even local media coverage of this issue tends to refer to “Muslims” rather than “Sunnis” and “Shi‘a.” 16. It should be noted that this survey was conducted in 2018, so it is possible that the perceptions of the meaning of democracy do not reflect the 2007 and/or 2014 environments. Thus, these responses should be interpreted cautiously. However, the findings from this survey nevertheless provide important insights into citizens’ general understandings of democracy. 17. It is worth noting that most responses were fairly short, so any mention of a given concept is significant. Many respondents chose to provide only one word or phrase, and none of the respondents provided especially long lists of characteristics, so the

 notes

18. 19.

20.

21.

22. 23.

24. 25.

26.

27. 28.

29.

30. 31.

choice to use a given word or phrase suggests that the respondent views it as an important feature. Examples include: “an electoral law that allows the representation of all people,” “an evolved electoral law,” “modern electoral laws,” “transparent electoral laws.” “Eliminate sectarianism,” “democracy relieves us from the problems of religious and sectarian quotas,” “democracy is not sectarian,” “democracy is that people deal with the politician or ruler in terms of his honesty and loyalty to the society, not in terms of his sect or religion.” ‘Equality between rich and poor people,” “secure livelihood and welfare for all citizens,” “providing job opportunities for all,” “improve electricity, water, and a clean environment,” “a fair social and economic life.” It is not clear whether the respondent meant minority rule or rule over the minority by this response. In any case, it is interesting (though probably coincidental) that both of these responses came from Maronites. Though, notably, not always on schedule. Importantly, Lebanon’s democratic underperformance does not indicate widespread satisfaction with the status quo. See Pearlman (2013b) for a discussion of how certain factors, including foreign remittances, help to sustain the political status quo even in the presence of widespread disenchantment. According to these data, the mean wealth per adult in Lebanon is over USD 33,000, but the median is just over 7,000. Following Horowitz (1985, p. 22), we may be able to think of the confessions in Lebanon as “ranked groups” since there is considerable overlap between class and group identity. Christians: 62%; Sunnis: 28%; Shi‘a: 29%. Source: Author’s survey. These figures reflect, in no small part, the historical legacies of Christian advantages in the development of educational institutions, which precede the modern state of Lebanon (see Cammett, 2014a, ch. 2). http://english.bayynat.org.lb/Stands/stand16052006.htm. The many reasons for this perception are beyond the scope of this project, but international factors contributed greatly to the view of Christians as a privileged group. Their unique access to European resources in particular gave them a commercial advantage over other sects, and higher levels of emigration and remittance-sending also advantaged the Christian community. See Pearlman (2013a) for a discussion of the material benefits obtained by Lebanese Christians through their historical patterns of emigration as well as the drawbacks of this practice. This inequality has often put Christians on the defensive; in 2011, Maronite labor minister Boutros Harb even submitted a draft law proposing to prohibit the sale of land across religions in response to a growing number of Iranian purchases of Christian property (see al-Arabiya, “Iran Possession of Lebanon Christian Property on Rise.” January 5, 2011). The details of this survey are discussed in the following chapter. Arabic wording:

notes 165 32. The mechanisms by which communal prayer translates into group solidarity and its corresponding political attitudes will be explored further in subsequent chapters, in which more detailed data are available. 33. A more detailed measure of frequency of attendance was not available in the first wave of the Arab Barometer. 34. I will explore this possibility more thoroughly in chapter 5, where more detailed data on partisan attachments are available. 35. Additional robustness checks controlling for other measures of religiosity yielded similar results. 36. Outcomes were similar when the models used an ordinal measure rather than the dichotomous variable presented here. The dichotomous variable is used in this chapter for clarity of presentation; results from the models using the ordinal measure are available upon request. 37. These results pool Sunnis and Shi‘a together as “Muslims” for the reasons discussed previously. Due to the small number of observations when separating these groups, virtually no estimated effects for any variable reach statistical significance. Reassuringly, however, the magnitude of the effect of religious practice is similar for Sunnis and Shi‘a, though when these groups are separated, the effect does not quite reach statistical significance at conventional levels due to the imprecision caused by the small sample sizes. These estimates are omitted for the purposes of space, but are available from the author upon request. 38. The propensity score matching method calculates a “propensity” to participate in communal practice based on these characteristics, matches observations based on this propensity score, and calculates an average treatment effect across matched groups. These models use one-to-one matching with replacement. These tests are performed using the Matching R package developed by Sekhon (2011), implementing the estimator proposed by Abadie and Imbens (2006). 39. The first additional question asks: “To what degree would you agree that the violation of human rights in [country name] is justifiable in the name of promoting security and stability?” and the second asks, “If ‘1’ means that democracy is completely unsuitable for Lebanon and ‘10’ means that it is completely suitable, where would you place your opinion about the degree to which democracy is suitable for Lebanon?” Since these questions are skewed toward more pro-democratic responses, I code each of these questions as 1 if the respondent gave the most pro-democratic response and 0 otherwise. The resulting index has four categories, with each level accounting for between 12% and 38% of responses. 40. Remark to author, Beirut, March 2014.

Chapter 5 1. Remark to author, Beirut, March 2014. 2. This description of sectarian interests is consistent with Salamey and Tabar (2012, p. 507), who find that Shi‘a generally prefer a proportional electoral system, Sunnis prefer a district-majority system, and Maronites are split between the two.

 notes 3. The findings presented in this section are broadly consistent with other surveys of religious behavior in Lebanon, such as Moaddel (2009) and Gärde (2012). These particular figures are taken from the author’s 2014 survey; the Arab Barometer and other earlier surveys did not contain as thorough a battery of religious questions as the 2014 survey, which was designed with this purpose in mind. 4. Source: author’s survey, described subsequently. 5. Similar distributions of responses are present for the question of whether respondents take comfort and strength from religion. 6. These differences are due in no small part to the fact that daily prayer is a religiously mandated behavior for Muslims, but not as clearly required for Christians. 7. Other scholars have assessed the level of Sunni-Shi‘a tensions similarly. Salem (2012, p. 4): “Sectarian tensions (particularly between Sunni and Shia communities) are close to an all-time high.” Byman (2014, p. 94): “Tension between the Shia and Sunni groups in Lebanon is worse than ever.” Abdo (2017, p. 149): “It is the sectarian fault line between the Sunni and Shi‘a and the identity politics that grows out of that division that are now the key mobilizing forces.” 8. Here, “Karbala” is a reference to a famous quasi-sectarian battle in the early days of Islam (AD 680) that is still mourned by Shi‘a today. 9. March 14 and March 8 are the main two political coalitions in Lebanon formed in 2005 after the assassination of former prime minister Rafiq Hariri. March 14, to whom most Sunnis are loyal, is anti-Syrian (specifically, anti-Syrian regime), and March 8, with whom most Shi‘a are affiliated, supports the Syrian regime. 10. It would be incorrect to characterize the Syrian Civil War as a primarily sectarian conflict; however, numerous observers have pointed out that the conflict took on an increasingly sectarian tone after its early days (Byman, 2014; Hashemi and Postel, 2017; Heydemann, 2013; Lynch et al., 2014; Salloukh, 2017; Wimmen, 2017). 11. It is possible that the uprisings in nearby countries could have contributed to an increased enthusiasm for democracy in Lebanon; however, such an increase would likely bias the effects presented in this chapter downward. Since the uprisings have generally occurred in Sunni-majority countries, it would perhaps be natural for Lebanese Sunnis to demonstrate an increase in support for democracy. However, the fact that communal religious practice changed from having a pro-democratic effect among Lebanese Sunnis in 2007 to an anti-democratic effect in 2014 suggests that if the regional effects of the Arab uprisings are present, they are dwarfed by other factors. 12. Arabic wording: 13. Note that these divisions no longer apply as neatly to the Lebanese case as they did just a few years ago. Since 2016, divisions within the Sunni community have intensified, while formerly hostile Christian parties have been increasingly able to compromise. However, these developments had not yet occurred when the survey used in this chapter was conducted, and were far from inevitable or even predictable. 14. As Tajfel and Turner (1979, p. 39) note, in settings of intergroup conflict, there is “a good deal of evidence that, within the pattern of responding in terms of in-group

notes 167

15. 16. 17. 18.

19.

20.

21. 22.

23. 24.

favoritism, maximum difference (M.D.) is more important than maximum in-group profit (M.I.P). Thus, they seem to be competing with the out-group, rather than following a strategy of simple economic gain for members of the in-group.” In the Lebanese setting, this behavior translates into Sunni hostility for democratic reforms; although Sunnis would benefit from democratization relative to Christians, Shi‘a would benefit even more than Sunnis. Since Shi‘a are now the relevant out-group with respect to Sunnis, Sunni preferences follow a “maximum difference” pattern comparing themselves to Shi‘a. Remarks to author, Beirut, March 27, 2014. Consequently, the leaders of both Amal and Hezbollah have explicitly called for Lebanon to lower the voting age to 18. Remarks to author, Beirut, March 2014. The negative effects of the state’s weakness in providing social services are magnified by the fact that access to such services has become highly politicized; see Cammett (2014b, 2015). Using a national survey, Salamey and Tabar (2012, p. 503) find that Shi‘a are much less likely to hold post-secondary degrees (43%) compared to Sunnis (52%) and Maronites (70%). They are also much more likely to live in areas where residential property values are worth less than USD 75,000 on average: 47% of Shi‘a live in such areas, compared to 26% of Sunnis and 20% of Maronites. The simple income measure is problematic as a measure of group privilege for several reasons. First, and perhaps most obviously, income alone does not indicate wealth, the latter of which is presumably more likely to be targeted by stateled redistribution. Second, respondents in this survey frequently had difficulty remembering or estimating their monthly income, so the measure is not entirely reliable. Finally, asking respondents about their monthly income does not capture the source of this income. While average income for both Sunnis and Shi‘a has increased in the past few decades, the degree to which the state has provided such income is highly uneven. Both groups benefit from external patronage and clientelistic political parties (Chen and Cammett, 2012), but Shi‘a have been much less likely to receive state benefits. Thus, even if certain groups report higher income, it is important to consider the source of that income: Shi‘a have increased their average income considerably, but very little of this increase is due to state intervention; they could benefit substantially from state redistributive policies despite their increasing income levels. The qadaa (district) is Lebanon’s second-level administrative unit; the country is divided into 26 qadaas. Sources vary on the rank-ordering of governorates by poverty, and even within a given study, ranks depend somewhat on the definition of poverty that is used (Central Administration of Statistics and World Bank, 2015). Source: World Bank (2009). The Akkar district, which is around two-thirds Sunni, has experienced relatively high levels of unemployment and low levels of education, exacerbated by the presence of large numbers of Syrian refugees in the district.

 notes 25. Interestingly, however, the youth population in North Lebanon reports somewhat higher levels of life satisfaction than other regions (UNDP, 2015, p. 45). 26. This difference remains statistically significant and of comparable size after including qadaa fixed effects, a test that is likely over-conservative since the state can (and likely does) choose to provide certain areas of the country with more or less reliable electricity based on the area’s sectarian composition. 27. World Bank (2009, p. 31). 28. Indeed, Cammett (2011, p. 82) has noted that Shi‘a citizens actually report relatively low levels of financial hardship, “which may arise from the community’s comparatively strong welfare institutions that have developed in recent decades.” The ground that Shi‘a have made up in the past few decades has been almost entirely internally driven; many Shi‘a areas remain disproportionately neglected by the state, providing incentives for nonstate actors to intervene, often successfully. 29. Since the Christian community is politically divided regarding the major political issue of the day (Syria), and has largely remained on the sidelines with respect to the Syrian conflict, it is unclear what to expect from communal religious practice among Christians. As a result, the remainder of this chapter will focus primarily on Sunnis and Shi‘a, for whom political interests are more easily identifiable. 30. The survey was designed by the author and implemented by Informational International, a Lebanese survey and consulting firm headquartered in Beirut. 31. Christian subgroups were sampled in proportion to Interior Ministry figures regarding their population shares, so the sample proportions of each Christian denomination are representative of their population proportions. 32. This question asked, “Other than weddings or funerals, how often do you attend religious services at a church or mosque?” Arabic wording:

33.

34.

35.

36.

Response categories included “never,” “rarely,” “around once per month,” “around once per week,” and “more than once per week.” Communal prayer and gender are both included as additive rather than interactive terms in these models because there is no theoretical reason to expect the effect of communal prayer to vary by gender; the omitted-variable problem could potentially come from leaving out the gender variable, since gender has been shown to be correlated with mosque attendance and could independently affect regime attitudes. As a robustness check, I have run models interacting these two terms, and the substantive results remain the same, albeit with less precision for women due to the smaller number of women attending mosque. In this question, respondents were asked: “How often do you pray at home or at work?” The qualification in this wording is important in order to distinguish personal/private from communal prayer. Results (omitted here for the purposes of space) are virtually identical when the models include a measure of association membership or include all of the previously mentioned control variables in a single model. This question measured knowledge about the sectarian system by asking respondents how many seats were reserved in parliament for Christians, Sunnis, and Shi‘a;

notes 169

37. 38. 39.

40. 41.

42. 43.

44. 45. 46.

“political knowledge” is measured by the total difference between the respondents’ answers and the true values, with lower scores indicating higher levels of knowledge. Political knowledge is also not correlated with perceptions of democracy such as democracy redistributing wealth from the rich to the poor (p-value = .147). Because of the lack of clear testable predictions for Maronites and other Christians, the further analyses presented subsequently mostly focus only on Muslims. Unfortunately, the survey used in this chapter did not include a question on high-risk political activism such as protest (Cammett, 2011). Additional questions asked about respondents’ frequency of consulting religious leaders on economic and family/personal matters. Including these questions as control variables does not affect the results. See Djupe and Calfano (2013, ch. 1) for an explanation of the feasibility—and importance—of experimental methods in the study of religion and politics. See Ahmed and Salas (2011); Djupe and Calfano (2013); Pichon et al. (2007); Randolph-Seng and Nielsen (2007); Saroglou et al. (2009); Shariff and Norenzayan (2007); and Van Cappellen et al. (2011). This term is familiar to Lebanese citizens and, importantly, not specific to one sect. Out of necessity, the experimental results use a smaller (four item) index because the effects of primes of this sort are fairly short-lasting. The four questions used for the experimental index described subsequently are a subset of the nine used in the observational section; originally, more items were included in the experimental portion of the questionnaire, but pilot testing revealed that the repetitiveness of these questions created considerable respondent fatigue. As a result, the regime questions were spread out in the final questionnaire, allowing them to be used in the observational section but not the experimental section. In any case, the results presented here are generally not sensitive to changes in the specification of the indices. Full results for the personal piety primes are available from the author upon request. Since the dependent variables in this section are each five-point Likert scales, these models use ordinal logistic regression. Arabic wording:

47. Unfortunately, time of interview was not recorded, so it is not possible to know whether respondents were interviewed before or after the jumu‘ah congregational prayers. However, the inability to account for the time of day should bias the results of the tests downward, since the data likely include respondents in the “treatment” group who had not yet received the actual treatment. 48. For this model, Sunnis and Shi‘a are pooled because of sample size considerations. Since the “treatment” group (attenders interviewed on a Friday) is so small within each sect (29 Sunnis and 39 Shi‘a), it is necessary to combine them to achieve a reasonable level of statistical power. When analyzed separately, the sects demonstrate similar patterns, albeit with noticeably larger standard errors. 49. The most obvious issue in assuming that the ATE would be the same as the ATT is that non-attenders might simply lack the preexisting piety or identification required

 notes for mosque attendance to have any effect. If this were the case, then the treatment effect for non-attenders (if it could be observed) would be quite different from the observed effect among attenders. 50. It is worth noting, however, that Braizat (2010) finds that Arab citizens widely understand democracy in the way it is understood elsewhere: civil liberties, political rights, and power rotation. 51. Arabic wording: 52. Arabic wording:

53. Arabic wording:

54. Arabic wording: 55. Using a similar question that does not specifically mention democracy but specifically refers to “sects” (“It is necessary for the government to work to improve income equality between different sects in Lebanon”) yielded similar results, significant at the .01 level or better. Arabic wording: 56. Sermon delivered at Mohammed al-Amin Mosque, Beirut, on October 26, 2012. Quoted in an-Nahar, October 26, 2012. 57. Sermon delivered at Mohammed al-Amin Mosque, Beirut, on June 10, 2011. Quoted in the Daily Star, June 11, 2011. 58. Sermon delivered on October 24, 2014. Cited in the Daily Star, October 24, 2014. 59. Rather interestingly, the Ta’if Accord explicitly calls for the gradual removal of formal sectarianism, but effectively had the opposite effect. 60. Quoted in the Daily Star, May 14, 2011. 61. Sermon entitled “The Silence Punished by God,” delivered September 5, 2014. Full text available at http://arabic.bayynat.org.lb/KhoutbePage.aspx?id=13657. 62. Sermon entitled “Congregational Prayer: A Virtue in Which We Need to Re-invest.” Full text available at http://arabic.bayynat.org.lb/KhoutbePage.aspx?id=12324. 63. See al-Shorfa: http://al-shorfa.com/en_GB/articles/meii/features/2014/10/27/feature-01.

Chapter 6 1. Kurds are an ethnic group representing a large share of the population in northern Iraq and neighboring areas. While they largely share a religious identity with Sunni Arabs, their ethnolinguistic identity sets them apart from Arabs, who speak a language that is entirely unrelated to Kurdish. The Kurdish people in Iraq and elsewhere have frequently called for the establishment of an independent Kurdish

notes 171

2. 3. 4.

5. 6. 7.

8.

9. 10. 11.

12. 13. 14.

state, and have been granted a certain degree of autonomy in post-Saddam Iraq. It is thus not entirely clear the extent to which we should consider their “sectarian interests” to be aligned with Sunni Arabs. Consequently, the analysis in this chapter focuses on Arabs, but includes Kurds where appropriate. Though Shi‘a are free to choose a marja‘ in any country, in practice, Iraqis tend to choose to emulate Iraqi maraji‘ (Christia, Dekeyser, and Knox, 2016b). Saddam Hussein did not formally become the president of Iraq until 1979, but had effectively been in charge for a number of years prior to his formal ascension. The Ba‘ath Party refers to a set of political parties established across the Arab world since the 1940s that call for various versions of Arab nationalism and socialism. Saddam Hussein used Iraq’s branch of the party throughout his rule as a tool to legitimize his position and downplay sectarian identities, with mixed success. See Hoffman (2012) for these results as well as others characterizing sectarian differences in Iraq. See Haddad (2017a) for a discussion of these terms as well as others related to sectarianism in the contemporary Middle East. Attacks on funeral processions and Eid celebrations were especially common (Ahlul Bayt News Agency, 2013). It is also important to note that mosque attacks were not limited to Sunni attackers; Sunni mosques have been targeted as well in deadly attacks (Raheem, 2013). These attacks even led Sunni leaders to announce that they would close down Baghdad’s Sunni mosques “indefinitely” in November 2013 (Muhanna, 2014). The origin of sectarianism in Iraq is a matter of considerable dispute, with some arguing rather simplistically that the sectarian divide stretches back to the early days of Islam, others arguing that this schism only emerged in the wake of the invasion, and still others taking positions seemingly everywhere in between these two. See Reidar (2008) for a review of this issue (and an argument for recent origins of Iraq’s sectarian divide); see also Haddad (2013a) for a discussion of how sectarian discourse has waxed and waned in Iraq. The age of this cleavage is not of great importance to this study, so I will sidestep this issue for the most part; it is sufficient to note that whatever its origins, sectarianism is important in contemporary Iraq. See Davis (2005) for a discussion of how Saddam not only tried to shift the narrative of the Iraqi state, but tried to fundamentally rewrite the country’s collective memory. See Blaydes (2018, ch. 9) for a discussion of the ways in which religion served as a basis for contentious politics during the Saddam era. Interestingly, however, Christia, Knox, and Al-Rikabi (2016) have found that Iraqi Sunnis have found better ways to access public resources than Shi‘a. This may be a compensation mechanism, whereby the relatively marginalized group develops more efficient methods of resource acquisition. See Hoffman and Jamal (2014) for a discussion of the political importance of Qur’an reading in the Arab world. Unsurprisingly for reasons described previously, Kurds are generally more supportive of democracy than Arabs. See Ross (2006) for a discussion of why this relationship might not hold in reality.

 notes 15. The confidence intervals around the “redistributive” estimates are wider due to the smaller number of respondents in this category. 16. I use relative risks here to account for the fact that certain candidate characteristics were more commonly emphasized than others. This quantity represents Pr(Important|Attender) Pr(Important|Non-Attender) . For example, the relative risk comparing attending Sunnis to non-attending Sunnis on the “candidate sect important” variable is 1.29, indicating an increase of 29% comparing attenders to non-attenders. This percentage increase corresponds to the 12 percentage point increase presented in Figure 6.5. 17. This finding is consistent with the results of a conjoint experiment conducted by Christia, Dekeyser, and Knox (2016b, p. 170), who find that among Shi‘a pilgrims, sect is more important than religiosity in determining preferences for the type of spouse a respondent’s child should marry. 18. Using a different question that asked the extent to which piety is an important quality for political leadership, I found no significant differences between attenders and nonattenders for either sect. This non-result bolsters the finding that communal prayer has, at best, a limited effect on demands for piety from political leaders.

Chapter 7 1. The control variables used in the main model are age, education level, income (country-specific), gender, marital status, religiosity (self-described), employment status, importance of God/religion, religion as a source of comfort and strength, and country and wave fixed effects. Minimal models or models including other relevant covariates yield comparable results. 2. Source: Pew Research Center (2018), “Being Christian in Western Europe.” https: //www.pewforum.org/2018/05/29/being-christian-in-western-europe/. 3. Here, the cutpoint between “high” and “low” levels of democracy is .3, though the results are not sensitive to this choice of threshold. 4. This index is calculated as follows: RelPol = 1 −

N 

(0.5 − πi )2 πi /0.25,

(7.1)

i=1

where πi is the proportion of each religion and N is the number of religious groups. 5. For both high religiosity and high religious affiliation, these models use a proportion of 0.7 as a cutpoint, but other cutpoints suggest similar results. 6. The confidence intervals in the “low religious affiliation” group are, of course, much larger than other groups because countries where relatively few people identify with a religious tradition do not generally have large religious groups. 7. Data from World Values Survey.

Bibliography Abadie, Alberto, and Imbens, Guido. 2006. “Large sample properties of matching estimators for average treatment effects.” Econometrica 74:1, pp. 235–267. Abdo, Geneive. 2017. The New Sectarianism: The Arab Uprisings and the Rebirth of the Shi’a-Sunni Divide. New York: Oxford University Press. Acemoglu, Daron and Robinson, James. 2006. Economic Origins of Dictatorship and Democracy. New York: Cambridge University Press. Adamczyk, Amy and Pitt, Cassady. 2009. “Shaping attitudes about homosexuality: The role of religion and cultural context.” Social Science Research 38:2, pp. 338–351. Ahlul Bayt News Agency. 2013. “Iraq: Blast at Imam Hussain Mosque killed 34 Shia Muslims.” http://www.abna.ir/data.asp?lang=3&Id=468047. Ahmed, Ali M. and Salas, Osvaldo. 2011. “Implicit influences of Christian religious representations on dictator and prisoner’s dilemma game decisions.” Journal of Socioeconomics 40:3, pp. 242–246. al-Marashi, Ibrahim and Salama, Sammy. 2008. Iraq’s Armed Forces: An Analytical History. New York: Routledge. Alagha, Joseph. 2005. “Hizballah after the Syrian withdrawal.” Issue number 237 (Winter 2005), pp. 34–39. al-Akhbar. 2014. “Regional developments to determine presidential election date.” April 2, 2014. http://english.al-akhbar.com/node/19256. Alesina, Alberto and Giuliano, Paola. 2009. “Preferences for redistribution.” Technical report, National Bureau of Economic Research. al-Monitor. 2014. “Christian villages on Lebanon-Syria border beef up Security.” Amat, Francesc and Wibbels, Erik. 2009. “Electoral incentives, group identity and preferences for redistribution.” Estudios/Working Papers no. 246,, Centro de Estudios Avanzados en Ciencias Sociales, p. 1. Anderson, Benedict. 1983. Imagined Communities: Reflections on the Origin and Spread of Nationalism. London: Verso. Appleby, Scott R. 1999. The Ambivalence of the Sacred: Religion, Violence, and Reconciliation. Lanham, MD: Rowman & Littlefield. Asfura-Heim, Patricio, Steinitz, Chris, and Schbley, Ghassan. 2013. “The specter of Sunni military mobilization in Lebanon.” Center for Naval Analysis. https://www.cna.org/CNA_files/PDF/DOP-2013-U-006349-Final.pdf Asquith, Christina. 2009. Sisters in War: A Story of Love, Family, and Survival in the New Iraq. New York: Random House. Atran, Scott. 2002. In Gods We Trust: The Evolutionary Landscape of Religion. New York: Oxford University Press. Atran, Scott and Henrich, Joseph. 2010. “The evolution of religion: How cognitive byproducts, adaptive learning heuristics, ritual displays, and group competition generate deep commitments to prosocial religion.” Biological Theory 5, pp. 18–30. Ayoob, Mohammed. 2004. “Political Islam: Image and reality.” World Policy Journal 21:3, pp. 1–14.

 bibliography Bahout, Joseph. 2017. “The unraveling of Taif: The limits of sect-based power-sharing in Lebanon.” In Beyond Sunni and Shia: The Roots of Sectarianism in a Changing Middle East. Ed. by Frederic, Wehrey. New York: Oxford University Press, pp. 135–156. Baldwin, Kate and Huber, John. 2010. “Economic versus cultural differences: Forms of ethnic diversity and public goods provision.” American Political Science Review 104:4, pp. 644–662. World Bank. 2009. Lebanon Social Impact Analysis: Electricity and Water Sectors. http://siteresources.worldbank.org/INTPSIA/Resources/490023-1120841262639/ Lebanon_electricity_water.pdf. Baylouny, Anne Marie. 2010. Privatizing Welfare in the Middle East: Kin Mutual Aid Associations in Jordan and Lebanon. Bloomington: Indiana University Press. Beattie, John. 1966. “Ritual and social change.” Man 1:1, pp. 60–74. Bellin, Eva. 2008. “Faith in politics: New trends in the study of religion and politics.” World Politics 60, pp. 315–347. Berman, Eli, Shapiro, Jacob, and Felter, Joseph. 2011. “Can hearts and minds be bought? The economics of counterinsurgency in Iraq.” Journal of Political Economy 119:4, pp. 766–819. Besley, Timothy and Persson, Torsten. 2019. “Democratic values and institutions.” American Economic Review: Insights 1:1, pp. 59–76. Blanchard, Christopher. 2011. Iraq: Oil and Gas Legislation, Revenue Sharing, and U.S. Policy. Darby, PA: DIANE Publishing. Blaydes, Lisa. 2018. State of Repression: Iraq under Saddam Hussein. Princeton: Princeton University Press. Blogowska, Joanna and Saroglou, Vassilis. 2011. “Religious fundamentalism and limited prosociality as a function of the target.” Journal for the Scientific Study of Religion 50:1, pp. 44–60. Bloom, Pazit Ben-Nun and Arikan, Gizem. 2012. “A two-edged sword: The differential effect of religious belief and religious social context on attitudes towards democracy.” Political Behavior 34:2, pp. 249–276. —2013. “Religion and support for democracy: A cross-national test of the mediating mechanisms.” British Journal of Political Science 43:2, pp. 375–397. Boix, Carles. 2003. Democracy and Redistribution. New York: Cambridge University Press. Braizat, Fares. 2010. “What Arabs think” Journal of Democracy 21:4, pp. 131–138. Bratton, Michael. 2003. “Briefing: Islam, democracy and public opinion in Africa.” African Affairs 102:408, pp. 493–501. —2010. “Anchoring the “D-word" in Africa” Journal of Democracy 21:4, pp. 106–113. Brewer, Mark, Kersh, Rogan, and Petersen, Eric. 2003. “Assessing conventional wisdom about religion and politics: A preliminary view from the pews.” Journal for the Scientific Study of Religion 42:1, pp. 125–136. Brooke, Steven. 2017. “Sectarianism and social conformity: Evidence from Egypt.” Political Research Quarterly 70:4, pp. 848–860. Brown, Adam D., Kouri, Nicole, and Hirst, William. 2012. “Memory’s malleability: Its role in shaping collective memory and social identity.” Frontiers in Psychology 3, pp. 1–3. Byman, Daniel. 2014. “Sectarianism afflicts the new Middle East.” Survival 56:1, pp. 79–100. Calhoun-Brown, Allison. 1996. “African American churches and political mobilization: The psychological impact of organizational resources.” Journal of Politics 58:4, pp. 935–953.

bibliography 175 Cammett, Melani. 2011. “Partisan activism and access to welfare in Lebanon.” Studies in Comparative International Development 46:1, pp. 70–97. —2013a. “The Syrian conflict’s impact on Lebanese politics.” United States Institute of Peace Peacebrief 158, pp. 1–4. —2013b. “Using proxy interviewing to address sensitive topics.” In Interview Research in Political Science. Ed. by Layna Mosley, pp. 125–143. Ithaca, NY: Cornell UP. —2014a. Compassionate Communalism: Welfare and Sectarianism in Lebanon. Ithaca: Cornell University Press. —2014b. “Sectarianism and the ambiguities of welfare in Lebanon.” In The Politics of Nonstate Social Welfare. Ed. by Melani Cammett and Lauren MacLean. Ithaca: Cornell University Press, pp. 137–156. —2015. “Sectarianism and the ambiguities of welfare in lebanon.” Current Anthropology 56:S11, S76–S87. Cammett, Melani and Issar, Sukriti. 2010. “Bricks and mortar clientelism: Sectarianism and the logics of welfare allocation in Lebanon.” World Politics 62:3, pp. 381–421. Campbell, Angus, Converse, Philip, Miller, Warren, and Stokes, Donald. 1960. The American Voter. Chicago: University of Chicago Press. Campbell, David. 2004. “Acts of faith: Churches and political engagement.” Political Behavior 26:2, pp. 155–180. Canache, Damarys. 2012. “Citizens’ conceptualizations of democracy: Structural complexity, substantive content, and political significance.” Comparative Political Studies 45:9, pp. 1132–1158. Canetti-Nisim, Daphna. 2004. “The effect of religiosity on endorsement of democratic values: The mediating influence of authoritarianism.” Political Behavior 26:4, pp. 377–398. Carpenter, Thomas and Marshall, Margaret. 2009. “An examination of religious priming and intrinsic religious motivation in the moral hypocrisy paradigm.” Journal for the Scientific Study of Religion 48:2, pp. 386–393. Central Administration of Statistics and World Bank. 2015. “Measuring poverty in Lebanon using 2011 HBS.” http://www.cas.gov.lb/images/Excel/Poverty/Measuring% 20poverty%20in%20Lebanon%20using%202011%20HBS_technical%20report.pdf. Chaaya, M., Sibai, A. M., Fayad, R., and El-Roueiheb, Z. 2007. “Religiosity and depression in older people: Evidence from underprivileged refugee and non-refugee communities in Lebanon.” Aging and Mental Health 11:1, pp. 37–44. Chandra, Kanchan. 2006. “What is ethnic identity and does it matter?” Annual Review of Political Science 9, pp. 397–424. —2012. Constructivist Theories of Ethnic Politics. New York: Oxford University Press. Chen, Bradley, and Cammett, Melani. 2012. “Informal politics and inequity of access to health care in Lebanon.” International Journal for Equity in Health 11:23, pp. 1–8. Chhibber, Pradeep. 2014. Religious Practice and Democracy in India. New York: Cambridge University Press. Christia, Fotini, Dekeyser, Elizabeth, and Knox, Dean. 2016a. “Mapping Shiite opinion: Insights from the Arba‘een pilgrimage.” Foreign Affairs, October. —2016b. “To Karbala: Surveying religious Shi‘a from Iran and Iraq.” Working paper. Christia, Fotini, Knox, Dean, and Al-Rikabi, Jaffar. 2016. “Networks of sectarianism: Experimental evidence on access to services in Baghdad.” Working paper.

 bibliography Ciftci, Sabri. 2010. “Modernization, Islam, or social capital: What explains attitudes toward democracy in the Muslim world?” Comparative Political Studies 43:11, pp. 1442–1470. —2013. “Secular-Islamist cleavage, values, and support for democracy and shari’a in the Arab world.” Political Research Quarterly 66:4, pp. 781–793. Claassen, Christopher. 2019a. “Does public support help democracy survive?” American Journal of Political Science. 64:1, pp. 118–134. —2019b. “Estimating smooth country-year panels of public opinion.” Political Analysis 27:1, pp. 1–20. Cole, Juan. 2003. “The United States and Shi’ite religious factions in post-Ba’athist Iraq.” International Peacekeeping 57:4, pp. 543–566. Collins, Kathleen and Owen, Erica. 2012. “Islamic religiosity and regime preferences: Explaining support for democracy and political Islam in Central Asia and the Caucasus.” Political Research Quarterly 65:3, pp. 499–515. Cone, James. 1986. A Black Theology of Liberation. Maryknoll, NY: Orbis. Cooke, Miriam. 1987. War’s Other Voices: Women Writers on the Lebanese Civil War, New York: Cambridge University Press. Coppedge, Michael. 2017. V-Dem Country-Year/Country-Date Dataset v7.1. https: //www.v-dem.net/en/data/data-version-7-1/. Corneo, Giacomo and Grüner, Hans. 2002. “Individual preferences for political redistribution.” Journal of Public Economics 83:1, pp. 83–107. Corstange, Daniel. 2012a. “Religion, pluralism, and iconography in the public sphere: Theory and evidence from Lebanon.” World Politics 64:1, pp. 116–160. —2012b. “Vote trafficking in Lebanon.” International Journal of Middle East Studies 44:3, pp. 483–505. —2013. “Ethnicity on the sleeve and class in the heart.” British Journal of Political Science 43:4, pp. 889–914. —2017. The Price of a Vote in the Middle East: Clientelism and Communal Politics in Lebanon and Yemen. New York: Cambridge University Press. Dana, Karam, Barreto, Matt, and Oskooii, Kassra. 2011. “Mosques as American institutions: Mosque attendance, religiosity and integration into the political system among American Muslims.” Religions 2:4, pp. 504–524. Davis, Eric. 2005. Memories of State: Politics, History, and Collective Identity in Modern Iraq. Berkeley: University of California Press. Dawson, Michael. 1995. Behind the Mule: Race and Class in African-American Politics. Princeton: Princeton University Press. Deeb, Lara. 2006. An Enchanted Modern: Gender and Public Piety in Shi’i Lebanon. Princeton: Princeton University Press. —2017. “Til sect do you part? On sectarianism and intermarriage in Lebanon.” Jadaliyya, September 14. https://www.jadaliyya.com/Details/34552. Deeb, Lara, and Harb, Mona. 2013. Leisurely Islam: Negotiating Geography and Morality in Shi’ite South Beirut. Kindle version. Princeton: Princeton University Press. Dehejia, Rajeev, DeLeire, Thomas, and Luttmer, Erzo. 2007. “Insuring consumption and happiness through religious organizations.” Journal of Public Economics 91:1, 259–279. De La O, Ana and Rodden, Jonathan. 2008. “Does religion distract the poor? Income and issue voting around the world.” Comparative Political Studies, 41:4–5, pp. 437–476. Diamond, Larry, Plattner, Marc, and Brumberg, Daniel. 2003. Islam and Democracy in the Middle East. Baltimore: Johns Hopkins University Press.

bibliography 177 Djupe, Paul A., Burge, Ryan P., and Calfano, Brian R. 2016. “The delegational pulpit? Clergy identifying as congregational political representatives.” Representation 52:1, pp. 43–69. Djupe, Paul A. and Calfano, Brian R. 2012. “The deliberative pulpit? The democratic norms and practices of the PCUSA.” Journal for the Scientific Study of Religion 51:1, pp. 90–109. Djupe, Paul and Calfano, Brian R. 2013. God Talk: Experimenting with the Religious Causes of Public Opinion. Philadelphia: Temple University Press. Djupe, Paul and Gilbert, Christopher P. 2003. The Prophetic Pulpit: Clergy, Churches, and Communities in American Politics. Lanham, MD: Rowman & Littlefield. Djupe, Paul and Gilbert, Christopher P. 2008. The Political Influence of Churches. New York: Cambridge University Press. Dodge, Toby. 2012a. “Enemy images, coercive socio-engineering and civil war in Iraq.” International Peacekeeping 19:4, pp. 461–477. —2012b. Iraq: From War to a New Authoritarianism. Kindle version. New York: Routledge. Driessen, Michael. 2014. Religion and Democratization: Framing Religious and Political Identities in Muslim and Catholic Societies. New York: Oxford University Press. Driskell, Robyn, Embry, Elizabeth, and Lyon, Larry. 2008. “Faith and politics: The influence of religious beliefs on political participation.” Social Science Quarterly 89:2, pp. 294–314. Durkheim, Émile. 2001. The Elementary Forms of Religious Life. New York: Oxford University Press. Ekelund, Robert B., Tollison, Robert D., Anderson, Gary M., Hébert, Robert F., and Davidson, Audrey B. 1996. Sacred Trust: The Medieval Church as an Economic Firm. New York: Oxford University Press. Everson, Stephen. 1996. Aristotle: The “Politics” and “The Constitution of Athens”. New York: Cambridge University Press. Faour, Muhammad A. 2007. “Religion, demography, and politics in Lebanon.” Middle Eastern Studies 43:6, pp. 909–921. Fawaz, Eli. 2009. “What makes Lebanon a distinctive country?” In Lebanon: Liberation, Conflict, and Crisis. Ed. by Barry Rubin. New York: Palgrave Macmillan, pp. 25–34. Fisk, Robert. 2001. Pity the Nation: Lebanon at War. New York: Oxford University Press. Foa, Roberto Stefan and Mounk, Yascha. 2017. “The signs of deconsolidation.” Journal of Democracy 28:1, pp. 5–15. Fox, Jonathan. 2004. “Religion and state failure: An examination of the extent and magnitude of religious conflict from 1950 to 1996.” International Political Science Review 25:1, pp. 55–76. Freedom House. 2011. Freedom in the World: Lebanon. Technical report. Gärde, Johan. 2012. “Patterns of religiosity among Sunnis, Shias, and Christians in Lebanon’s multi-confessional context.” Middle East Critique 21:3, pp. 291–308. Geddes, Barbara and Zaller, John. 1989. “Sources of popular support for authoritarian regimes.” American Journal of Political Science 33:2, pp. 319–347. George, Alexander, and Andrew, Bennett. 2005. Case Studies and Theory Development in the Social Sciences. Cambridge, MA: MIT Press. Gill, Anthony. 2001. “The Study of liberation theology: What’s next?” Journal for the Scientific Study of Religion 41:1, pp. 87–89.

 bibliography Ginges, Jeremy, Hansen, Ian, and Norenzayan, Ara. 2009. “Religion and support for suicide attacks.” Psychological Science 20:2, pp. 224–230. Grewal, Sharan, Jamal, Amaney A. Masoud, Tarek, and Nugent, Elizabeth R. 2019. “Poverty and divine rewards: The electoral advantage of Islamist political parties.” American Journal of Political Science 63:4, pp. 859–874. Grzymala-Busse, Anna. 2012a. “Good clubs and community support: Explaining the growth of strict religions.” Journal of Church and State 56:2, pp. 269–299. —2012b. “Why comparative politics should take religion (more) seriously.” Annual Review of Political Science 15, pp. 421–442. —2015. Nations under God: How Churches Use Moral Authority to Influence Policy. Princeton: Princeton University Press. —2016a. “Religion and European politics.” In The Oxford Handbook of Historical Institutionalism. Ed. by Karl Fioretos, Tulia Falleti, and Adam Sheingate. New York: Oxford University Press, pp. 467–485. —2016b. “The difficulty with doctrine: How religion can influence politics.” Government and Opposition 51:2, pp. 327–350. —2016c. “Weapons of the meek: How churches influence public policy.” World Politics 68:1, pp. 1–36. Gu, Man-Li and Bomhoff, Eduard. 2012. “Religion and support for democracy: A comparative study for Catholic and Muslim countries.” Politics and Religion 5:2, pp. 280–316. Haddad, Fanar. 2011. Sectarianism in Iraq: Antagonistic Visions of Unity. New York: Columbia University Press. —2013a. “Sectarian relations in Arab Iraq: Contextualizing the civil war of 2006–2007.” British Journal of Middle Eastern Studies 40:2, pp. 115–138. —2013b. “The language of anti-Shiism.” Foreign Policy 9. https://foreignpolicy.com/2013/ 08/09/the-language-of-anti-shiism/. —2017a. “‘Sectarianism’ and its discontents in the study of the Middle East.” Middle East Journal 71:3, pp. 363–382. —2017b. “Shia-centric state-building and Sunni rejection in post-2003 Iraq.” In Beyond Sunni and Shia: The Roots of Sectarianism in a Changing Middle East. Ed. by Frederic Wehrey. New York: Oxford University Press, pp. 115–134. Hanmer, Michael and Kalkan, Kerem. 2013. “Behind the curve: Clarifying the best approach to calculating predicted probabilities and marginal effects from limited dependent variable models.” American Journal of Political Science 57:1, pp. 263–277. Hardin, Russell. 1997. One for All: The Logic of Group Conflict. Princeton: Princeton University Press. Harris, Elizabeth. 2007. “The cost of peace: Buddhists and conflict transformation in Sri Lanka.” In Can Faiths Make Peace? Holy Wars and the Resolution of Religious Conflict. Ed. by P. Broadhead and D. Keown. London: I.B. Tauris, pp. 149–162. Harris, Frederick. 1994. “Something within: Religion as a mobilizer of African American political activism.” Journal of Politics, 56:1, pp. 42–68. Harris, William. 2009. “Reflections on Lebanon.” In Lebanon: Liberation, Conflict, and Crisis. Ed. by Barry Rubin. New York: Palgrave Macmillan, pp. 9–24. Hashemi, Nader and Postel, Danny. 2017. “Sectarianization: Mapping the new politics of the Middle East.” Review of Faith & International Affairs 15:3, pp. 1–13. Hazran, Yusri. 2009. “The Shiite community in Lebanon: From marginalization to ascendancy.” Crown Center for Middle East Studies, Middle East Brief 37.

bibliography 179 —2010. “The rise of politicized Shi’ite religiosity and the territorial state in Iraq and Lebanon.” Middle East Journal 64:4, pp. 521–541. Henley, Alexander. 2017. “Religious authority and sectarianism in Lebanon.” In Beyond Sunni and Shia: The Roots of Sectarianism in a Changing Middle East. Ed. by Frederic Wehrey. New York: Oxford University Press, pp. 283–302. Herberg, Will. 1983. Protestant–Catholic–Jew: An Essay in American Religious Sociology. Chicago: University of Chicago Press. Heydemann, Steven, 2013. “Tracking the ‘Arab Spring’: Syria and the future of authoritarianism.” Journal of Democracy 24:4, pp. 59–73. Hiltermann, Joost. 2010. “Thank goodness for Iraq’s census disaster.” Foreign Policy October 8. http://www.foreignpolicy.com/articles/2010/10/08/thank_goodness_for_iraq_s _census_disaster. Hoffman, Michael. 2012. “Iraq Country Report.” Arab Barometer. Hoffman, Michael and Jamal, Amaney. 2014. “Religion in the Arab Spring: Between two competing narratives.” Journal of Politics 76:3, pp. 593–606. Hoffman, Michael and Nugent, Elizabeth. 2015. “Communal religion and support for armed parties: Evidence from Lebanon.” Journal of Conflict Resolution 61:4, pp. 869–902. Hofmann, Steven Ryan. 2004. “Islam and democracy: Micro-level indications of compatibility.” Comparative Political Studies 37:6, pp. 652–676. Horowitz, Donald. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press. Huntington, Samuel. 1991. The Third Wave: Democratization in the Late Twentieth Century. Norman: University of Oklahoma Press. —1996. The Clash of Civilizations and the Remaking of World Order. New York: Simon and Schuster. Iannaccone, Laurence. 1994. “Why strict churches are strong.” American Journal of Sociology, 99:5, pp. 1180–1211. Iannaccone, Laurence R. 1997. “Rational choice: Framework for the scientific study of religion.” In Rational Choice Theory and Religion: Summary and Assessment. Ed. by Lawrence A. Young. New York: Routledge, pp. 25–44. Iannaccone, Laurence and Berman, Eli. 2006. “Religious extremism: The good, the bad, and the deadly.” Public Choice 128:1–2, pp. 109–129. Irons, William. 2001. “Religion as a hard-to-fake sign of commitment.” In Evolution and the Capacity for Commitment. Ed. by Randolph Nesse. New York: Russell Sage Foundation, pp. 292–309. Isaksson, Ann-Sofie and Lindskog, Annika. 2009. “Preferences for redistribution—a country comparison of fairness judgements.” Journal of Economic Behavior & Organization 72:3, pp. 884–902. Jabar, Faleh. 2003. “The worldly roots of religiosity in post-Saddam Iraq,” Middle East Report 227, pp. 12–18. —2011. “Religion, sect, ethnicity, and tribe: The uncertainties of identity politics in the new society.” In Iraq, Democracy and the Future of the Muslim World. Ed. by Ali Paya and John Esposito. New York: Routledge, pp. 19–35. Jaboori, Rafid. 2013. “Sectarian tensions stalk Iraq elections,” BBC News, April 20. http://www.bbc.co.uk/news/world-middle-east-22215632. Jacobson, Jessica. 1997. “Religion and ethnicity: Dual and alternative sources of identity among young British Pakistanis.” Ethnic and Racial Studies 20:2, pp. 238–256.

 bibliography Jamal, Amaney. 2005. “The political participation and engagement of Muslim Americans: Mosque involvement and group consciousness.” American Politics Research 33:4, pp. 521–544. Jamal, Amaney and Tessler, Mark. 2008. “Attitudes in the Arab world.” Journal of Democracy 19:1, pp. 97–110. James, William. 2012 [1902]. The Varieties of Religious Experience: A Study in Human Nature. New York: Oxford University Press. Jelen, Ted. 1993. “The political consequences of religious group attitudes.” Journal of Politics 55:1, pp. 178–190. Johnson, Megan, Rowatt, Wade, and LaBouff, Jordan. 2010. “Priming Christian religious concepts increases racial prejudice.” Social Psychological and Personality Science 1:2, pp. 119–126. —2012. “Religiosity and prejudice revisited: In-group favoritism, out-group derogation, or both?” Psychology of Religion and Spirituality 4:2, pp. 154–168. Johnson, Michael. 2001. All Honourable Men: The Social Origins of War in Lebanon. New York: I.B. Tauris. Juergensmeyer, Mark. 2008. Global Rebellion: Religious Challenges to the Secular State, from Christian Militias to al Qaeda. Berkeley: University of California Press. Kalyvas, Stathis. 1996. The Rise of Christian Democracy in Europe. Ithaca: Cornell University Press. —2000. “Commitment problems in emerging democracies: The case of religious parties.” Comparative Politics 32:4, pp. 379–398. Kalyvas, Stathis and Van Kersbergen, Kees. 2010. “Christian democracy.” Annual Review of Political Science 13, pp. 183–209. Karakoç, Ekrem and Ba¸skan, Birol. 2012. “Religion in politics: How does inequality affect public secularization?” Comparative Political Studies 45:12, pp. 1510–1541. Karam, Salam. 2007. “The multi-faced Sunni insurgency: A personal reflection.” Civil Wars 9:1, pp. 87–105. Kedourie, Elie. 1994. Democracy and Arab Political Culture. London: Frank Cass. Keely, Louise and Tan, Chih. 2008. “Understanding preferences for income redistribution.” Journal of Public Economics 92:5, pp. 944–961. Kellstedt, Lyman and Green, John. 1993. “Knowing God’s many people: Denominational preference and political behavior.” In Rediscovering the Religious Factor in American Politics. Ed. by David Leege and Lyman Kellstedt, Armonk, NY: ME Sharp, pp. 53–71. Kian, Azadeh. 2014. “Gendered Khomeini.” In A Critical Introduction to Khomeini. Ed. by Arshin Adib-Moghaddam. New York: Cambridge University Press, pp. 170–192. Kim, Myunghee. 2008. “Spiritual values, religious practices, and democratic attitudes.” Politics and Religion 1:2, pp. 216–236. Klor, Esteban F. and Shayo, Moses. 2010. “Social identity and preferences over redistribution.” Journal of Public Economics 94:3, pp. 269–278. Knutsen, Oddbjørn. 2004. “Religious denomination and party choice in Western Europe: A comparative longitudinal study from eight countries, 1970–97.” International Political Science Review 25:1, pp. 97–128. Krause, Neal. 2002. “Church-based social support and health in old age exploring variations by race.” Journals of Gerontology Series B: Psychological Sciences and Social Sciences 57:6, pp. S332–S347. Kubba, Laith. 1996. “Recognizing pluralism.” Journal of Democracy 7:2, pp. 86–89.

bibliography 181 Layman, Geoffrey. 1997. “Religion and political behavior in the United States: The impact of beliefs, affiliations, and commitment from 1980 to 1994.” Public Opinion Quarterly 61:2, pp. 288–316. Liechty, Joseph and Clegg, Cecelia. 2001. Moving Beyond Sectarianism: Religion, Conflict, and Reconciliation in Northern Ireland. Dublin: Columba Press. Lijphart, Arend. 1977. Democracy in Plural Societies: A Comparative Exploration. New Haven: Yale University Press. —1979. “Religious vs. linguistic vs. class voting: The “crucial experiment” of comparing Belgium, Canada, South Africa, and Switzerland.” American Political Science Review 73:2, pp. 442–458. Lindqvist, E. and Östling, R. 2013. “Identity and redistribution.” Public Choice 155:3–4, pp. 469–491. Linz, Juan and Stepan, Alfred. 1996. Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-Communist Europe. Baltimore: Johns Hopkins University Press. Lipset, Seymour Martin and Rokkan, Stein. 1967. Party Systems and Voter Alignments: Cross-National Perspectives. Toronto: Free Press. Lockwood, W. G. 1981. “Religion and language as criteria of ethnic identity: An exploratory comparison.” In Ethnicity and Nationalism in Southeastern Europe. Ed. by S. Beck and J. W. Cole. Amsterdam: Universiteit van Amsterdam, pp. 71–81. Lynch, Marc, Freelon, Deen, and Aday, Sean. 2014. “Syria in the Arab Spring: The integration of Syria’s conflict with the Arab uprisings, 2011–2013.” Research & Politics 1:3, pp. 1–7. Mackey, Sandra. 1989. Lebanon: A House Divided. New York: W.W. Norton. Makdisi, Ussama. 2000. The Culture of Sectarianism: Community, History, and Violence in Nineteenth-century Ottoman Lebanon. Berkeley: University of California Press. Margolis, Howard. 1984. Selfishness, Altruism, and Rationality. Chicago: University of Chicago Press. Marx, Karl. 1970 [1843]. Marx’s Critique of Hegel’s Philosophy of Right. New York: Cambridge University Press. McAdam, Doug. 1982. Political Process and the Development of Black Insurgency: 1930– 1970. Chicago: University of Chicago Press. McClendon, Gwyneth and Riedl, Rachel. 2015. “Religion, redistribution, and political participation: Evidence from an experiment in Nairobi, Kenya.” Journal of Politics 77:4, pp. 1045–1057 Meltzer, Allan and Richard, Scott. 1981. “A rational theory of the size of government.” Journal of Political Economy 89:5, pp. 914–927. Meyer, Katherine, Tope, Daniel, and Price, Anne. 2008. “Religion and support for democracy: A crossnational examination.” Sociological Spectrum 28:5, pp. 625–653. Moaddel, Mansoor. 2009. Ethnicity and Values Among the Lebanese Public: Findings from a Values Survey. Middle East Values Survey; https://mevs.org/wpcontent/uploads/2019/12/moaddel_lebanese_survey_pr_jan08.pdf. Muhanna, Elias. 2013. “Could the orthodox law be a boon for lebanon’s liberals?” http:// qifanabki.com/2013/01/15/could-the-orthodox-law-be-a-boon-for-lebanons-liberals/. New Yorker, February 21, 2014. “Lebanon’s war in Syria.” Mulligan, Casey, Sala-i Martin, Xavier, and Gil, Ricard. 2003. “Do democracies have different public policies than nondemocracies?” Technical report, National Bureau of Economic Research.

 bibliography Nasr, Vali. 2005. “The rise of ‘Muslim democracy’” Journal of Democracy 16:2, pp. 13–27. National Consortium for the Study of Terrorism and Responses to Terrorism (START). 2013. Global Terrorism Database.” http://www.start.umd.edu/gtd. Norris, Pippa and Inglehart, Ronald. 2004. Sacred and Secular: Religion and Politics Worldwide. New York: Cambridge University Press. Norton, Augustus Richard. 2005. “Ritual, blood, and Shiite identity: Ashura in Nabatiyya, Lebanon.” Drama Review 49:4, pp. 140–155. Pearlman, Wendy. 2013a. “Emigration and power: A study of sects in Lebanon, 1860–2010.” Politics & Society 41:1, pp. 103–133. —2013b. “Emigration and the resilience of politics in Lebanon.” Arab Studies Journal 21:1, pp. 191–213. Pepinsky, Thomas and Welborne, Bozena. 2011. “Piety and redistributive preferences in the Muslim world.” Political Research Quarterly 64:3, pp. 491–505. Pettersson, Thorleif. 2009. “Religious commitment and socio-political orientations.” In Routledge Handbook of Religion and Politics. Ed. by Jeffrey Haynes. New York: Routledge, pp. 246–270. Pew Forum. 2012. “The global religious landscape.” Philpott, Daniel. 2007. “Explaining the political ambivalence of religion.” American Political Science Review 101:3, pp. 505–525. Picard, Elizabeth. 1993. The Lebanese Shi’a and Political Violence. Geneva: United Nations Research Institute for Social Development. Pichon, Isabelle, Boccato, Giulio, and Saroglou, Vassilis. 2007. “Nonconscious influences of religion on prosociality: A priming study.” European Journal of Social Psychology 37, pp. 1032–1045. Posner, Daniel. 2005. Institutions and Ethnic Politics in Africa. New York: Cambridge University Press. Putnam, Robert. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press. —2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster. Putnam, Robert and Campbell, David. 2010. American Grace: How Religion Divides and Unites Us. New York: Simon and Schuster. Rabushka, Alvin and Shepsle, Kenneth A. 1972. Politics in Plural Societies. Stanford: Stanford University Press. Raheem, Kareem. 2013. “Bomb attacks on Sunni mosques in Iraq kill 23” Reuters. http://www.reuters.com/article/2013/07/13/us-iraq-violence-idUSBRE96C0CP20130713. Randolph-Seng, Brandon and Nielsen, Michael. 2007. “Honesty: One effect of primed religious representations.” International Journal for the Psychology of Religion 17:4, pp. 303–315. Reynal-Querol, Marta. 2002. “Ethnicity, political systems, and civil wars.” Journal of Conflict Resolution 46:1, pp. 29–54. Rizzo, Helen, Abdel-Latif, Abdel-Hamid, and Meyer, Katherine. 2007. “The relationship between gender equality and democracy: A comparison of Arab versus non-Arab Muslim societies.” Sociology 41:6, pp. 1151–1170. Romer, Thomas. 1975. “Individual welfare, majority voting, and the properties of a linear income tax.” Journal of Public Economics 4:2, pp. 163–185. Rosen, Nir. 2006. “Anatomy of a civil war: Iraq’s descent into chaos.” Boston Review.

bibliography 183 Ross, Michael. 2006. “Is democracy good for the poor?” American Journal of Political Science 50:4, pp. 860–874. Salamé, Ghassan. 1986. “Lebanon’s injured identities: Who represents whom during a civil war? In Papers on Lebanon. Oxford: Centre for Lebanese Studies. Salamey, Imad and Payne, Rhys. 2008. “Parliamentary consociationalism in Lebanon: Equal citizenry vs. quotated confessionalism.” Journal of Legislative Studies 14:4, pp. 451–473. Salamey, Imad, and Tabar, Paul. 2012. “Democratic transition and sectarian populism: The case of Lebanon.” Contemporary Arab Affairs 5:4, pp. 497–512. Salem, Paul. 2012. “Can Lebanon survive the Syrian crisis?” Carnegie Endowment for International Peace. Salloukh, Bassel. 2017. “The Syrian war: Spillover effects on Lebanon.” Middle East Policy 24:1, pp. 62–78. Salloukh, Bassel, Barakat, Rabie, Al-Habbal, Jinan, Khattab, Lara, and Mikaelian, Shoghig. 2015. The Politics of Sectarianism in Postwar Lebanon. London: Pluto Press. Salloukh, Bassel and Verheij, Renko. 2017. “Transforming power sharing: From corporate to hybrid consociation in postwar Lebanon.” Middle East Law and Governance 9:2, pp. 147–173. Salti, Nisreen and Chaaban, Jad. 2010. “The role of sectarianism in the allocation of public expenditure in postwar Lebanon.” International Journal of Middle East Studies 42:4, pp. 637–655. Sankari, Jamal. 2005. Fadlallah: The Making of a Radical Shi‘ite Leader. London: Saqi. Saroglou, Vassilis, Corneille, Olivier, and Van Cappellen, Patty. 2009. “Speak, Lord, your servant is listening: Religious priming activated submissive thoughts and behaviors.” International Journal for the Psychology of Religion 19, pp. 143–154. Sassoon, Joseph. 2012. Saddam Hussein’s Ba‘th Party: Inside an Authoritarian Regime. New York: Cambridge University Press. Scheve, Kenneth and Stasavage, David. 2006. “Religion and preferences for social insurance.” Quarterly Journal of Political Science 1:3, pp. 255–286. Schwadel, Philip. 2005. “Individual, congregational, and denominational effects on church members’ civic participation.” Journal for the Scientific Study of Religion 44:2, pp. 159–171. Sekhon, Jasjeet. 2011. “Multivariate and propensity score matching software with automated balance optimization: The matching package for R.” Journal of Statistical Software 42:7, pp. 1–52. Shanahan, Rodger. 2005. Shi’a of Lebanon: Clans, Parties and Clerics. New York: I.B. Tauris. Shariff, Azim and Norenzayan, Ara. 2007. “God is watching you: Priming God concepts increases prosocial behavior in an anonymous economic game.” Psychological Science 18:9, pp. 803–809. Siegel, Alexandra. 2017. “Twitter wars: Sunni-Shia conflict and cooperation in the digital age.” In: Beyond Sunni and Shia: The Roots of Sectarianism in a Changing Middle East. Ed. by Frederic Wehrey. New York: Oxford University Press, pp. 157–180. Smaldino, Paul. 2014. “The cultural evolution of emergent group-level traits.” Behavioral and Brain Sciences 37:3, pp. 243–254. Smidt, Corwin. 1999. “Religion and civic engagement: A comparative analysis.” Annals of the American Academy of Political and Social Science 565, pp. 176–192. Smolicz, Jerzy. 1981. “Core values and cultural identity.” Ethnic and Racial Studies 4:1, pp. 75–90.

 bibliography Solt, Frederick, Habel, Philip, and Grant, J. Tobin. 2011. “Economic inequality, relative power, and religiosity.” Social Science Quarterly 92:2, pp. 447–465. Sosis, Richard and Bressler, Eric. 2003. “Cooperation and commune longevity: A test of the costly signaling theory of religion.” Cross-Cultural Research 37:2, pp. 211–239. Sosis, Richard and Kiper, Jordan. 2014. “Why religion is better conceived as a complex system than a norm-enforcing institution.” Behavioral and Brain Sciences 37:3, pp. 275– 276. Stark, Rodney. 2010. “Returning God to the social scientific study of religion.” In Advanced Methodologies in the Scientific Study of Religion and Spirituality. Ed. by William Grassie. Philadelphia: Metanexus Institute, pp. 56–66. Stark, Rodney and Finke, Roger. 2000. Acts of faith: Explaining the Human Side of Religion. Berkeley: University of California Press. Stegmueller, Daniel. 2013. “Religion and redistributive voting in Western Europe.” Journal of Politics 75:4, pp. 1064–1076. Svensson, Isak. 2007. “Fighting with faith: Religion and conflict resolution in civil wars.” Journal of Conflict Resolution 51:6, pp. 930–949. Tajfel, Henri and Turner, John. 1979. “An integrative theory of intergroup conflict.” In The Social Psychology of Intergroup Relations. Ed. by William Austin and Stephen Worchel. Monterey, CA: Brooks/Cole, pp. 33–47. Tessler, Mark. 2002. “Islam and democracy in the Middle East: The impact of religious orientations on attitudes toward democracy in four Arab countries.” Comparative Politics 34:3, pp. 337–354. Tessler, Mark, Jamal, Amaney, and Robbins, Michael. 2012. “New findings on Arabs and democracy.” Journal of Democracy 23:4, pp. 89–103. Tezcur, Gunes Murat, Azadarmaki, Taghi, Bahar, Mehri, and Nayebi, Hooshang. 2012. “Support for democracy in Iran.” Political Research Quarterly 65:2, pp. 235–247. Tocqueville, Alexis de. 2003 [1835]. Democracy in America. Washington, DC: Regnery. Times, New York. 2013. “Sunnis close baghdad mosques to challenge religious attacks.” www.nytimes.com/2013/11/24/world/middleeast/sunnis-close-baghdad-mosquesto-challenge-religious-attacks.html. Timmons, Jeffrey. 2010. “Does democracy reduce economic inequality?” British Journal of Political Science 40:4, pp. 741–757. Toburen, Tina and Meier, Brian. 2010. “Priming God-related concepts increases anxiety and task persistence.” Journal of Social and Clinical Psychology 29:2, pp. 127–143. Toft, Monica Duffy, Philpott, Daniel, and Shah, Timothy Samuel. 2011. God’s Century: Resurgent Religion and Global Politics. New York: W.W. Norton. Trablousi, Fawwaz. 2007. A Modern History of Lebanon. London: Pluto Press. Trejo, Guillermo. 2009. “Religious competition and ethnic mobilization in Latin America: Why the Catholic Church promotes indigenous movements in Mexico.” American Political Science Review 103:3, pp. 323–342. United Nations Development Programme (UNDP)., 2015. Spotlight on Youth in Lebanon, http://www.lb.undp.org/content/dam/lebanon/docs/Poverty/Publications/Spotlight% 20on%20Youth%20Final%20(23-06-2016).pdf. United Nations High Commissioner for Refugees (UNHCR). 2015. Akkar Government Profile. https://data2.unhcr.org/fr/documents/download/44893. US Energy Information Administration. 2013. Iraq, Techical report http://www.eia.gov/ countries/analysisbriefs/Iraq/iraq.pdf.

bibliography 185 US State Department. 2008. “International religious freedom report: Lebanon.” http:// www.state.gov/g/drl/rls/irf/2008/108487.htm. US State Department. 2012. International Religious Freedom Report: Iraq. http://www. state.gov/documents/organization/208612.pdf. Uslaner, Eric. 2002. “Religion and civic engagement in Canada and the United States.” Journal for the Scientific Study of Religion 41:2, pp. 239–254. Valenta, Marko and Strabac, Zan. 2012. “Religion and support for democracy in bosniaherzegovina.” Politics and Religion 5:3, pp. 609–633. Valenzuela, Ali. 2011. From the pulpit or in the pews: Religious beliefs, clergy messages and social influence on latino politics.” Princeton University, Working Paper. Van Cappellen, Patty, Corneille, Olivier, Cols, Stephanie, and Saroglou, Vassilis. 2011. “Beyond mere compliance to authoritative figures: Religious priming increases conformity to informational influence among submissive people.” International Journal for the Psychology of Religion 21:2, pp. 97–105. Verba, Sidney, Schlozman, Kay, and Brady, Henry. 1995. Voice and Equality: Civic Voluntarism in American Politics. New York: Cambridge University Press. Visser, Reidar. 2008. “The western imposition of sectarianism on Iraqi politics.” Arab Studies Journal 15–16:2, pp. 83–99. Wald, Kenneth. 1987. Religion and Politics in the United States. New York: St. Martin’s Press. Wald, Kenneth, Owen, Dennis, and Hill, Samuel. 1988. “Churches as political communities.” American Political Science Review 82:2, pp. 531–548. Warner, Carolyn. 2000. Confessions of an Interest Group: The Catholic Church and Political Parties in Europe. Princeton: Princeton University Press. Weber, Max. 1922. The Sociology of Religion. Boston: Beacon Press. Weiss, Max. 2010. In the Shadow of Sectarianism: Law, Shi‘ism, and the Making of Modern Lebanon. Cambridge, MA: Harvard University Press. Welzel, Christian. 2007. “Are levels of democracy affected by mass attitudes? Testing attainment and sustainment effects on democracy.” International Political Science Review 28:4, pp. 397–424. Wenger, Jay. 2004. “The automatic activation of religious concepts: Implications for religious orientations.” International Journal for the Psychology of Religion 14:2, pp. 109–123. Wike, Richard and Fetterolf, Janell. 2018. “Liberal democracy’s crisis of confidence.” Journal of Democracy 29:4, pp. 136–150. Wiltermuth, Scott and Heath, Chip. 2009. “Synchrony and cooperation.” Psychological Science 20:1, pp. 1–5. Wimmen, Heiko. 2017. “The sectarianization of the Syrian war.” In Beyond Sunni and Shia: The Roots of Sectarianism in a Changing Middle East. Ed. by Frederic Wehrey. New York: Oxford University Press, pp. 61–86. World Bank. 2011. World Development Indicators. http://data.worldbank.org/indicator/ NY.GDP.PETR.RT.ZS. Wuthnow, Robert. 2009. Boundless Faith: The Global Outreach of American Churches. Berkeley: University of California Press.

 bibliography Zelin, Aaron Y. 2016. “Jihadism in Lebanon after the Syrian uprising.” In Lebanon and the Arab Uprisings: In the Eye of the Hurricane. Ed. by Maximilian Felschand and Martin Wählisch. New York: Routledge, pp. 50–69. Zelin, Aaron Y. and Smyth, Phillip. 2014. “The vocabulary of sectarianism.” Foreign Policy 29. https://foreignpolicy.com/2014/01/29/the-vocabulary-of-sectarianism/.

Index Page numbers in italics indicate tables, charts, and figures. Terms and names beginning with “Al-” are indexed under their main portions; thus Al-Monitor is listed in the Ms. Abadie, Alberto, 165n38 Acemoglu, Daron, 29, 132 African Americans, and racial/religious identity, 161n22 Aisha (wife of Muhammad), 43, 152n6 ‘Alawite regime in Syria, 111 Algeria, religion and democracy in, 7, 157 ‘Ali ibn Abi Talib, 118–119 Amal Movement (Shi‘a), Lebanon, 73, 78, 81, 85, 86, 99, 167n16 American Voter, 159n1 Aoun, Alain, 56, 83 Aoun, Michel, 47, 56, 85 Arab Barometer, 10 on Iraq, 121, 130, 134, 136 on Lebanon, 64, 68, 71, 74, 163n9, 165n33, 166n3 Arab Spring, 7, 83–84 Arba‘een, 120 Argentina, religion and democracy in, 15, 157 Arikan, Gizem, 14–15 Aristotle, 132 Atran, Scott, 27 average treatment effects (ATE), 97–98, 98, 169–170n49 average treatment effects on the treated (ATT), 77, 103, 169–170n49 axis of political competition, 23–25, 34, 117, 160n16 Ba‘ath party, Iraq, 119–120, 128, 171n4 Bahout, Joseph, 84 Beattie, John, 25 behavioral versus denominational differences, 12–13, 160n6 Bellin, Eva, 12

Berman, Eli, 126 Berri, Nabih, 86 Blaydes, Lisa, 120 Blogowska, Joanna, 29 Bloom, Pazit Ben-Nun, 14–15 Boix, Carles, 132 Bomhoff, Eduard, 13–14 Bosnia, religion and democracy in, 13, 157 Braizat, Fares, 170n50 Brazil, religion and democracy in, 2, 7, 13, 157 British residents of Pakistani origin, 32–33 Calfano, Brian R., 13 Calhoun-Brown, Allison, 161n22 Cammett, Melani, 66, 68, 81, 95, 135, 162n1, 168n28 Catholicism and democracy, 2, 6, 11, 14 causal direction. See direction of causation Cedar Revolution (2005), Lebanon, 80–81 Central Asia, religion and democracy in, 13 Chaaban, Jad, 88, 160n12 Chile, Catholicism and democracy in, 2, 157 Christia, Fotini, 171n11, 172n17 Christianity/Christians. See also Maronites Catholicism and democracy, 2, 6, 11, 14 communal religious experience of, 8, 40, 42, 45–46 daily prayer requirements for, 166n6 ethnic/national/racial versus religious identity, 161n22 Free Patriotic Movement, Lebanon, 85 Greek Catholics in Lebanon, 55

 index Greek Orthodox in Lebanon, 40, 55, 61 Lebanon, “Christian” as category in, 24, 60–61, 78 in Lebanon following Syrian Civil War, 9, 45, 78–80, 82–85, 90, 91, 95, 165n2, 166n13, 167n19, 168n29, 168n31 Christians in Lebanon before Syrian Civil War, 9 demographics, representation, and status, 53, 55, 55–57, 56, 164n26, 164nn28–29 economic inequalities, 63–66, 64 methodology, data, and findings, 68–72, 70, 72, 75 Muslim-Christian sectarian divisions, 57–61, 59 church/mosque attendance. See communal religious experience, group identity, and regime preference Ciftci, Sabri, 14 clientelistic/distributive networks, 95–96 Cole, Juan, 120 comfort, relief, and strength derived from communal religion, 39–40 communal religious experience, group identity, and regime preference, 3, 4, 8–9, 35–51, 145–157. See also direction of causation, and under specific entries at Lebanon of Christians, 8, 40, 42, 45–46 complex but non-random nature of relationship, 151–153 cross-national testing of, 145–151, 147, 149, 150, 152–153, 157, 172n1 economic inequalities and, 45, 46, 49 Friday mosque attendance and group identity, 100–104, 101, 102, 104 gender and, 79, 90, 162n3, 168n33 implications and applications of, 153–156 interest-based theory of religion/regime preferences and, 11, 15–16, 17, 33–34 in Iraq, 8, 35–36, 128–130 “linked fate” perceptions, 32, 41, 99, 100, 102–104, 104, 116

political messaging and, 43–44 private religious belief versus, 26, 160n7 reasons for attending/not attending religious services, 37–39 relief, comfort, and strength derived from, 39–40 religious identity and, 32 sectarian hostility and, 42–43, 162n5 of Shi‘a, 8–9, 40, 43, 48–50 solidarity of group and, 26–28, 40–43, 112 of Sunnis, 8, 40, 41, 42, 43, 46–48 survey interviews conducted for, 35–36 construction of identities, 22 Credit Suisse, Global Wealth Databook, 62 cross-national testing, 145–151, 147, 149, 150, 152–153, 157, 172n1 Dana, Karam, 161n22 De La O, Ana, 6 Deeb, Lara, 21, 67, 80 Dekeyser, Elizabeth, 172n17 democracy and democratization. See also religion, sectarianism, and democracy interest-based theory and, 20–21, 24–25 Iraq, experience of authoritarianism and democracy in, 117–118 meaning of democracy in Lebanon, 61–63, 163–164nn16–21 political versus economic understanding of, in Iraq, 132–138, 135, 137 political versus economic understanding of, in Lebanon, 104–109, 105, 107, 109, 110 redistribution, perception of democracy as more favorable to, 65 semi-democratic history and practice in Lebanon, 52–53, 62–63, 164n23 denominational versus behavioral differences, 12–13, 160n6 Derian, Abdul Latif, 111, 112 direction of causation average treatment effects, 97–98, 98 concept of, 4

index 189 interest-based theory and, 31, 159n7 in Lebanon following Syrian Civil War, 96–112, 98, 100–102, 104, 105, 107, 109, 110 mechanisms of, 98–104, 100–102, 104 political versus economic features of democracy and, 104–109, 105, 107, 109, 110 sermons and attitudes of religious leaders, 109–112 Djupe, Paul A., 13, 25 Dodge, Toby, 120, 121, 122, 124 Doha Accords, 106n10 Druze, 55 Durkheim, Emile, 16–17, 25 economic inequalities. See also redistribution clientelistic/distributive networks and, 95–96 communal religious experience and, 45, 46, 49 electricity, access to, as poverty measure, 88 income as measure of group privilege, 167n20 interest-based theory of religion/regime preference and, 18, 20, 22, 23, 29–30, 34 in Iraq, 124–127, 125, 127 in Lebanon before Syrian Civil War, 45, 46, 49, 57–58, 63–66, 64, 164n26, 164nn28–29 in Lebanon following Syrian Civil War, 10, 86–89, 108–109, 109, 167n19 perceptions of, 65 economic versus political understanding of democracy in Iraq, 132–138, 135, 137 in Lebanon, 105, 106–109, 107, 109, 110 economy of Iraq, state role in, 118 Egypt, religion and democracy in, 7, 157 electricity, access to, as poverty measure, 88 ethnic versus religious identity, 31–33, 121–124, 161n22 Europe, Catholicism and democracy in, 6, 11

Fadlallah, Muhammad Hussein, 60, 64, 111–112 Fadlallah, Sayyid Ali, 111 Fatfat, Ahmed, 85–86 FawanFawaz, Eli, 53 Free Patriotic Movement (Christian), Lebanon, 85 Freedom House, 62 Friday mosque attendance and group identity, 100–104, 101, 102, 104 Future Movement (Sunni), Lebanon, 58, 60, 80–83, 85, 99 Geagea, Samir, 85 Gemayel, Nadim, 56 gender and communal religious experience, 79, 90, 162n3, 168n33 Gilbert, Christopher, 25 Gill, Anthony, 6 Global Terrorism Database, 123 Greece, religion and democracy in, 7 Greek Catholics in Lebanon, 55 Greek Orthodox in Lebanon, 40, 55, 61 Grzymala-Busse, Anna, 12, 33, 160n8, 161n31 Gu, Man-Li, 13–14 Haddad, Fanar, 117, 120–121 Hamdan, Ali, 73 Harb, Boutros, 164n28 Harb, Mona, 80 Hariri, Rafiq, assassination of, 59, 80, 81, 166n9 Hariri, Saad, 47 Harris, Elizabeth, 52 Hazran, Yusri, 85 Henley, Alexander, 24, 85 Henrich, Joseph, 27 Herberg, Will, 23 Hezbollah communal religious experience and, 47 following Syrian Civil War, 80–82, 84, 85, 167n16 before Syrian Civil War, 59, 59–60, 163n10 Horowitz, Donald, 33 hostility, sectarian, 42–43, 83, 122, 162n5

 index Husayn ibn Ali, 119 Hussein, Saddam, 3, 119–120, 121, 128, 171n3, 171nn9–10 identification versus identity, 32 identity and interest. See communal religious experience, group identity, and regime preference; interest-based theory of religion and regime preferences “imagined” religious communities, 28 Imbens, Guido, 165n38 India, religiosity and democracy in, 14, 157 Indonesia, religion and democracy in, 15, 157 in-group favoritism, 29, 67, 166–167n14 interest-based theory of religion and regime preferences, 4, 6, 8, 11–34 ambiguity of religious/political nexus, 11–15 axis of political competition, 23–25, 34, 160n16 communal prayer as identity-strengthening activity, 11, 15–16, 17, 33–34 current studies on religion and political preference, 13–15, 30 definition of religion for purposes of, 16–17 democracy/democratization, concept of, 20–21, 24–25 direction of causation, 31, 159n7 distinguishing religious from ethnic or other identity markers, 31–33, 161n22 economic inequalities and, 18, 20, 22, 23, 29–30, 34 independent variable, studying religion as, 12–13 at individual and group levels, 6, 15, 17–18, 29–31, 65 mechanisms of, 25 multidimensional nature of religious behavior, 15, 16 rationalist analysis of, 6 in religiously divided societies, 18–21

sectarian unity, presence or absence of, 24 sectarianism, concept of, 21–22 solidarity of group, 26–31 state-level and sectarian factors, consideration of, 15–16, 18 statement of, 15–21 Iran, religion and democracy in, 4–5, 13, 57, 83, 157 Iraq, 9–10, 117–144, 152 Ba‘ath party in, 119–120, 128, 171n4 communal religious experience, group identity, and regime preference in, 8, 35–36, 128–130, 138 dictatorial fear of destabilizing effects of religious identity in, 19–20 economic inequalities in, 124–127, 125, 127 experience of authoritarianism and democracy in, 117–118 under Saddam Hussein, 3, 19, 119–120, 128, 171n3, 171nn9–10 hypotheses suggested by, 128–131, 133–134 Kurds in, 118, 120, 126, 127, 141–143, 170–171n1, 172n13 methodology, data, and findings, 130–132, 131, 140–144, 152 mosques as political foci in, 121–122, 123–124, 129, 171n7 oil wealth in, 118, 126–127, 127 political versus economic understandings of democracy in, 132–138, 135, 137 post-Saddam era in, 120–121, 129, 130, 139 public resources acquisition in, 171n11 religious nature of sectarian division in, 121–124 sectarian Sunni-Shi‘a division and identity in, 117–121, 171n8 state role in economy of, 118 Iraq War (2003), 81, 121 Islam. See Muslims Islamic State (ISIS), 84, 139 Israeli occupation of Lebanon, 54

index 191 Jabar, Faleh, 118, 123–124 Al-Jama‘ah Al-Islamiyah (The Islamic Group), Sunni membership in, 96, 99 Jamal, Amaney, 160n11 James, William, 160n6 jihadism in Lebanon, 84 Johnson, Megan, 29 Jumblatt, Walid, 55 Kabalan, Ahmad, 111 Kabbani Mohammad, 82 Kalyvas, Stathis, 6 Karam, Salam, 121 Karbala, battle of, 82, 119, 128, 166n8 Kataeb Party, Lebanon, 56 Khomeini, Ayatollah Ruhollah, 4–5 Kim, Myunghee, 14 King, Martin Luther, Jr., Letter from Birmingham Jail, 1 Knox, Dean, 171n11, 172n17 Krause, Neal, 26 Kurds, in Iraq, 118, 120, 126, 127, 141–143, 170–171n1, 172n13 Latin America, Catholicism and democracy in, 2, 6 Layman, Geoffrey, 12 Lebanese Civil War (1975-1990), 54, 66–67, 80, 111 Lebanon Cedar Revolution (2005), 80–81 Christian communal religious experience in, 45–46 communal religious experience, group identity, and regime preference in, 8, 35–36, 44–50, 52, 66–67 importance of religion in, 52 Israeli occupation of, 54 mental health and religion in, 26 “Muslim” and “Christian,” as categories in, 23–24, 60–61, 73, 78, 165n37 National Pact (1943), 46, 53, 55, 59, 59–60 as religiously divided state, 52 sectarian unity in, 24

semi-democratic history and practice of, 52–53, 62–63, 164n23 Shi‘a communal religious experience in, 48–50 Sunni communal religious experience in, 46–48 Syrian occupation of/withdrawal from, 54, 58, 72–73, 82 Ta’if Accord (1989), 50, 54, 55, 80, 85–86, 111, 163n5, 170n59 underrepresentation of Muslims generally in, 52, 57, 80, 85, 162n3 underrepresentation of Shi‘a in, 49, 80, 85, 86, 106, 108, 162n4, 163n6 underrepresentation of Sunnis in, 80, 85, 106 Lebanon before Syrian Civil War, 9, 52–76, 151. See also specific entries at Christians, Sunnis, and Shi‘a communal religious experience, effects of, 8, 35–36, 44–50, 52, 66–72, 70, 72 demographics, representation, and status in, 46, 53–57, 55, 56 economic inequalities in, 45, 46, 49, 57–58, 63–66, 64, 164n26, 164nn28–29 hypotheses suggested by, 67–69 meaning of democracy in, 61–63, 163–164nn16–21 methodology, data, and findings, 53–54, 68–72, 70, 72, 74, 75–76, 151, 165nn36–39 minorities, concept of protection for, 62 “Muslim” as category in, 23–24, 73, 163n10, 163n15, 165n37 political protests, engagement in, 71–72, 72 religious leaders and, 60 sectarian Christian/Muslim relations (in 2007), 57–61, 59 Lebanon following Syrian Civil War, 9, 78–116, 151–152. See also specific entries at Shi‘a and Sunnis alternative explanations and threats to inferences, 95–98, 98 Arab Spring and, 83–84

 index Christians in, 9, 45, 78–80, 82–85, 90, 91, 95, 165n2, 166n13, 167n19, 168n29, 168n31 communal religious experience, effects of, 89–95, 91–94 development of Sunni-Shi‘a political divide, 80–89 direction of causation in, 96–112, 98, 100–102, 104, 105, 107, 109, 110 economic inequalities in, 10, 86–89, 108–109, 109 foreign interference and effects, 47, 166n11 hypotheses suggested by, 89–95, 97, 100, 101 jihadism in, 84 methodology, data, and findings, 78, 90–104, 91–94, 98, 100–102, 104, 114–116, 151–152 political versus economic understanding of democracy in, 104–109, 105, 107, 109, 110 religious demographics in 2014 Lebanon, 79–80 religious leaders and, 85, 96, 109–112 sermons, effect of, 109–112 Libya, religion and democracy in, 7 Lijphart, Arend, 54, 67, 160n6 “linked fate” perceptions, 32, 41, 99, 100, 102–104, 104, 116 Linz, Juan, 52 Lithuania, religion and democracy in, 15, 157 majoritarianism, 9, 62, 132, 134, 138, 139, 160n10 Malaysia, religion and democracy in, 2, 157 maraji‘, 119, 171n2 March 8 (political coalition in Lebanon), 83, 166n9 March 14 (political coalition in Lebanon), 58, 83, 166n9 Maronites communal religious experience of, 42, 45–46 in Lebanon before Syrian Civil War, 53, 55, 56, 59, 60–61, 66, 163n5

in Lebanon following Syrian Civil War, 83, 85, 90, 91, 95, 165n2, 167n19 Marx, Karl, 1 Matching R package, 165n38 Meyer, Katherine, 14, 159n4 Al-Monitor, 84 moral authority and religious identity, 33, 161n31 mosque/church attendance. See communal religious experience, group identity, and regime preference mosques as political foci in Iraq, 121–122, 123–124, 129, 171n7 Muhanna, Elias, 56 Muslims. See also Shi‘a; Sunnis as category in Lebanon, 23–24, 73, 163n10, 163n15, 165n37 daily prayer requirements for, 166n6 democracy and, 1–2, 5, 12–13, 14 ethnic/national versus religious identity and, 161n22 sectarian Christian/Muslim relations (in 2007) in Lebanon, 57–61, 59 underrepresentation of, in Lebanon, 52, 57, 80, 85, 162n3 Najaf (Iraq), as holy site, 118–119, 128 National Pact (1943), Lebanon, 46, 53, 55, 59, 59–60 national versus religious identity, 31–33, 161n22 Nigeria, religiosity and democracy in, 14, 157 Obeid, Mohammed, 78, 86 oil wealth in Iraq, 118, 126–127, 127 Pakistan, religion and democracy in, 2, 157 Pakistani origin, British residents of, 32–33 Pearlman, Wendy, 164n28 Philpott, Daniel, 15 Poland, religion and democracy in, 2, 15, 157

index 193 “political knowledge,” measuring, 168–169n36 political protests, engagement in in Lebanon before Syrian Civil War, 71–72, 72 popular participation in, 7 political versus economic understanding of democracy. See economic versus political understanding of democracy privilege/size of religious group affecting political preferences, 2–3, 4, 18–20 Putnam, Robert, 26 Qabbani, Mohammed Rashid, 111 “questions-as-treatment” experiment, 97–98, 98 quietism, 119 Rabushka, Alvin, 19, 160n9 racial versus religious identity, 161n22 al-Rafei, Salem, 112 rationalist approaches to religion, 4, 5–8 redistribution, 9–10, 152, 154 interest-based theory of religion/regime preference and, 13, 17, 18, 29–30, 32, 34 in Iraq, 118, 126, 127, 132–134, 135, 138, 172n15 in Lebanon before Syrian Civil War, 62, 65 in Lebanon following Syrian Civil War, 87, 88, 108–109, 110, 167n20 perception of democracy as more favorable to, 65 regime preference and religion. See communal religious experience, group identity, and regime preference; democracy and democratization; majoritarianism; redistribution; religion, sectarianism, and democracy relief, comfort, and strength derived from communal religion, 39–40 religion, sectarianism, and democracy, 1–10. See also democracy and democratization

ambiguous relationship between, 1–5, 11–15 communal religious practice and, 3, 4, 8–9, 35–51 (see also communal religious experience, group identity, and regime preference) contextual analysis of, 9–10 (see also Iraq; specific entries at Lebanon) definition of religion, 16–17 differences between religious and nonreligious people, increase in, 5 direction of causation in, 4, 31, 159n7 global number of traditionally religious people, increase in, 5 at individual versus group levels, 6–7 interest-based theory of, 4, 6, 8, 11–34 (see also interest-based theory of religion and regime preferences) political preferences affected by religious identity, 2 popular support for democracy, importance of, 7–8 rationalist approaches to, 4, 5–8 secularization theory, confounding, 1–2, 5, 33 size/privilege of religious group affecting political preferences, 2–3, 4, 18–20 religious polarization measure, 148, 172n4 Reynal-Querol, Marta, 148 Al-Rikabi, Jaffar, 171n11 Robinson, James, 29, 132 Rodden, Jonathan, 6 al-Sadr, Muqtada, 124 Salamé, Ghassan, 66–67 Salamey, Imad, 165n2, 167n19 Salloukh, Bassel, 64, 81–82, 83 Salti, Nisreen, 88, 160n12 Saroglou, Vassilis, 29 Saudi Arabia in cross-national testing, 157 Lebanon and, 57, 83 Scheve, Kenneth, 6, 30 sectarianism, 21–22. See also religion, sectarianism, and democracy secularization theory, 1–2, 5, 33 Sekhon, Jasjeet, 165n38

 index Senegal, religion and democracy in, 15 sermons, 109–112 Shepsle, Kenneth A., 19, 160n9 Shi‘a. See also Hezbollah; Iraq Amal Movement, Lebanon, 73, 78, 81, 85, 86, 99, 167n16 communal religious experience and, 8–9, 40, 43, 48–50 geographic references to, 160n21 March 8 (political coalition in Lebanon), 83, 166n9 political versus theological divisions from Sunnis, 153 sectarian hostility toward, 43, 162n5 underrepresentation of, in Lebanon, 49, 80, 85, 86, 106, 108, 162n4, 163n6 Shi‘a in Lebanon before Syrian Civil War, 9 Christian/Muslim sectarian divisions, 57–61, 59 communal religious experience, group identity, and regime preferences, 67 demographics, representation, and status, 9, 52, 55, 56 economic inequities, 63–64, 64, 66 methodology, data, and findings, 73, 76 Shi‘a in Lebanon following Syrian Civil War, 9, 78, 112–113 alternative hypotheses and direction of causation, 95–112 average treatment effects, 97–98, 98 communal religious experience and support for democracy, 45, 90–95, 91 development of Sunni-Shi‘a divide, 80–89, 163n10, 166–167n14 economic inequalities and, 86–89, 108–109, 109, 167nn19–20, 168n28 “linked fate,” perception of, 99, 100 political versus economic features of democracy and, 105, 106–108, 107, 109, 110 regime preferences of, 165n2 religious conviction in 2014 Lebanon, 79–80 sermons and, 110–112 situational salience, 23

size/privilege of religious group affecting political preferences, 2–3, 4, 18–20 social effects of religion, 29 solidarity of group, 26–28, 40–43, 112 Stasavage, David, 6, 30 Statistics Lebanon, 68 Stepan, Alfred, 52 strength, comfort, and relief derived from communal religion, 39–40 Sudan, religion and democracy in, 7 Sunnis. See also Iraq communal religious experience and, 8, 40, 41, 42, 43, 46–48 fabricated divisions within, concerns about, 47–48 Future Movement, Lebanon, 58, 60, 80–83, 85, 99 Al-Jama‘ah Al-Islamiyah (The Islamic Group), membership in, 96, 99 Kurdish affiliation with, 126 March 14 (political coalition in Lebanon), 58, 83, 166n9 political versus theological divisions from Shi‘a, 153 sectarian hostility and, 43, 162n5 underrepresentation of, in Lebanon, 80, 85, 106 Sunnis in Lebanon before Syrian Civil War, 9, 52 Christian/Muslim sectarian divisions, 57–61, 59 demographics, representation, and status, 53, 55, 56 economic inequities, 63–64, 64, 66 methodology, data, and findings, 73, 76 Sunnis in Lebanon following Syrian Civil War, 9, 78, 112–113 alternative hypotheses and direction of causation, 95–112 average treatment effects, 97–98, 98 communal religious experience and support for democracy, 45, 47, 90–95, 91 development of Sunni-Shi‘a divide, 80–89, 163n10, 166–167n14 divisions within Sunni community, 166n13

index 195 economic inequalities and, 86–89, 108–109, 109, 167nn19–20 “linked fate,” perception of, 99, 100 political versus economic features of democracy and, 105, 106–109, 107, 109, 110 regme preferences of, 165n2 religious conviction in 2014 Lebanon, 79–80 sermons and, 110–112 Syria and Syrian conflict, 11. See also specific entries at Lebanon dictatorial fear of destabilizing effects of religious identity in, 19–20 foreign ministers’ statement (2015) on, 1 occupation of/withdrawal from Lebanon, 54, 58, 72–73, 82 sectarian nature of Syrian conflict, 166n10 Tabar, Paul, 165n2, 167n19 Ta’if Accord (1989), Lebanon, 50, 54, 55, 80, 85–86, 111, 163n5, 170n59 Tajfel, Henri, 23, 160n16, 166–167n14 Tessler, Mark, 14, 160n11 Toft, Monica, 6 Tunisia, religion and democracy in, 2, 7

Turkey, religion and democracy in, 7, 157 Turner, John, 23, 160n16, 166–167n14 UNDP, 88, 168n25 United States, religion and democracy in, 13, 157, 161n22 US State Department statistics, 55, 118 value-enforcement, religion’s emphasis on, 33 Varieties of Democracy (V-Dem) Liberal Democracy Index, 148 Wahhabism, 122 Weber, Max, 12 welfare state, religion and attitudes toward, 13, 159n2 Wenger, Jay, 97 Whitehead, Alfred North, 160n7 women, communal religious experience of, 79, 90, 162n3, 168n33 World Bank, 63, 64, 87–88, 126 World Values Survey, 10, 145, 146, 149 Wuthnow, Robert, 28 Zogby Polls, 59