Beyond Ethnic Politics in Africa 1107171490, 9781107171497

Table of contents :
Cover
Half-title page
Title page
Copyright page
Dedication
Contents
List of Figures
List of Maps
List of Tables
Acknowledgments
List of Abbreviations
1 The Challenge of Winning Votes and Ethnic Politics in Africa
2 A Theory of Social Ties and Electoral Politics
3 Social Structure and Its Origins
4 Mobilization Strategies and Electoral Outcomes in Senegal and Benin
5 Intermediaries in Urban and Rural Settings
6 Social Structure and Ethnic Politics in Africa and Beyond
7 Conclusion
Appendix A Socially Salient Identities in Senegal and Benin
Appendix B The Role of Social and Political Leaders in Senegal and Benin
Appendix C Additional Electoral Data for Senegal and Benin
Appendix D Incumbent Support in Urban and Rural Areas in Senegal
References
Index

Citation preview

Beyond Ethnic Politics in Africa

Why do ethnic politics emerge in some ethnically diverse societies but not others? Focusing on sub-Saharan Africa, Dominika Koter argues that the prevailing social structures of a country play a central role in how politicians attempt to mobilize voters. In particular, politicians consider the strength of local leaders, such as chiefs or religious dignitaries, who have historically played a crucial role in many parts of rural Africa. Local leaders can change the electoral dynamics by helping politicians secure votes among people of different ethnicities. Ethnic politics thus can be avoided where there are local leaders who can serve as credible electoral intermediaries between voters and politicians. Koter shows that there is widespread variation in the standing of local leaders across Africa, as a result of long-term historical trends, which has meant that politicians have mobilized voters in qualitatively different ways, resulting in different levels of ethnic politics across the continent. Dominika Koter is Assistant Professor of Political Science at Colgate University. She received the Gregory Luebbert Award for Best Article in Comparative Politics and the African Politics Conference Groups’ award for best article published on African politics in 2013.

Beyond Ethnic Politics in Africa

DOMINIKA KOTER Colgate University, New York

University Printing House, Cambridge CB2 8BS, United Kingdom Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781107171497 © Dominika Koter 2016 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2016 A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Koter, Dominika, author. Beyond ethnic politics in Africa / Dominika Koter. Cambridge University Press : New York, 2016. | Includes bibliographical references. LCCN 2016026620 | ISBN 9781107171497 LCSH: Political campaigns – Africa, Sub-Saharan. | Politics, Practical – Africa, Sub-Saharan. | Elections – Africa, Sub-Saharan. | Social structure – Political aspects – Africa, Sub-Saharan. | Local government – Africa, Sub-Saharan. | Africa, Sub-Saharan – Ethnic relations – Political aspects. | Africa, Sub-Saharan – Politics and government. LCC JQ1879.A5 K68 2016 | DDC 324.720967–dc23 LC record available at https://lccn.loc.gov/2016026620 ISBN 978-1-107-17149-7 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party Internet Websites referred to in this publication and does not guarantee that any content on such Websites is, or will remain, accurate or appropriate.

For my family and in memory of Bill Foltz

Contents

List of Figures List of Maps

page viii ix

List of Tables Acknowledgments

x xii

List of Abbreviations

xiv

2

The Challenge of Winning Votes and Ethnic Politics in Africa A Theory of Social Ties and Electoral Politics

1 31

3

Social Structure and Its Origins

57

4

Mobilization Strategies and Electoral Outcomes in Senegal and Benin

81

1

5

Intermediaries in Urban and Rural Settings

128

6 7

Social Structure and Ethnic Politics in Africa and Beyond Conclusion

146 166

Appendix A Socially Salient Identities in Senegal and Benin

175

Appendix B The Role of Social and Political Leaders in Senegal and Benin Appendix C Additional Electoral Data for Senegal and Benin

177 178

Appendix D Incumbent Support in Urban and Rural Areas in Senegal References

184 186

Index

199 vii

Figures

1.1 1.2 1.3 2.1 2.2 3.1 6.1

Level of Ethnic Politics in Africa Summary of the Argument Politicians’ Mobilization Strategies Predicted Dominant Modes of Electoral Mobilization Mobilization Modes and Electoral Patterns The Sufi Hierarchy Trust in Local Leaders and the Level of Ethnic Politics

viii

page 7 18 19 40 50 62 162

Maps

1.1 Main and Secondary Cases 3.1 Map of Senegal 3.2 Map of Benin

page 24 61 72

ix

Tables

2.1 Summary of Measures of Ethnic Politics page 55 3.1 Caste Stratification 64 3.2 Balance of Power between Traditional and Religious Authorities in Senegal 70 3.3 Local Leaders in Senegal 70 3.4 Local Leaders in Benin 75 3.5 Reliance on Traditional and Religious Leaders 78 4.1 1960 General Council Elections 92 4.2 Ethnic Composition of the Electorates of the Main Presidential Candidates in Benin 94 4.3 Ethnic Composition of the Electorates of the Main Presidential Candidates in Senegal 123 4.4 Religious and Brotherhood Composition of the Electorates of the Main Presidential Candidates in Senegal 123 5.1 Incumbent Vote Share in Presidential Elections in Senegal 137 6.1 Ethnic Composition of the Electorates of the Main Parties in Mali 152 A.1 Identity Repertoire in Senegal 175 A.2 Identity Repertoire in Benin 176 B.1 Reliance on Different Types of Leaders in Senegal and Benin 177 C.1 1970 Presidential Election in Benin 179 C.2 Ethnic Composition of the Electorate of the Main Presidential Candidates in Benin (2008) 180 C.3 Ethnic Composition of the Electorate of the Main Presidential Candidates in Benin (2011) 180 C.4 Extent of Ethnic Politics in Senegal and Benin 181 x

List of Tables C.5 Ethnic Composition of the Electorate of the Main Parties in Legislative Elections in Senegal C.6 Religious and Brotherhood Composition of the Electorate of Main Parties in Legislative Elections in Senegal C.7 Ethnic Composition of the Main Candidates/Parties in Senegal (2005) C.8 Ethnic Composition of the Main Candidates/Parties in Senegal (2008) C.9 Ethnic Composition of the Main Candidates/Parties in Senegal (2013) D.1 Determinants of Incumbent Support in Senegal

xi

182 182 182 183 183 185

Acknowledgments

While writing this book I became indebted to many individuals. This project would not have happened without the generosity of all the people whom I met in the field. I am grateful to everyone who opened their doors to me and was willing to answer my countless questions: to politicians who made time for formal interviews as well as numerous food vendors, taxi drivers and other regular people in the street who all taught me something about politics. I wish to thank Professors Momar Coumba Diop and Ousmane Sène and their colleagues at the West African Research Center in Dakar for providing a research base and for sharing their expertise. I am also grateful to Professor Adolphe Dansou and his colleagues at the University Gaston-Berger; to Ibrahima Faye, Hamidou Bâ and Malick Gakou in Senegal; and to Moïse Houngnikpo, Mathias Hounkpe and Jean-Baptise Hounkpe in Benin. The Thiam family welcomed me into their Dakar home for over half a year and served as my adopted family and was an insightful source of knowledge about Senegalese society. Many friends in the field provided much needed encouragement and companionship. Yale University provided an inspiring and supportive environment for the dissertation research that led to this book. I want to thank members of my dissertation committee: Ellen Lust, Sue Stokes and Keith Darden, for their invaluable feedback, encouragement, patience and friendship. I could not have asked for a better set of mentors. Thad Dunning, Adria Lawrence, Steven Wilkinson, Kwame Onoma, David Simon and many others provided much valuable feedback at different stages. I am particularly grateful to Bill Foltz, who inspired me to study French West Africa. Bill was there from the very beginning of this project and I wish xii

Acknowledgments

xiii

that he could see the final product. My friends at Yale provided a wonderful working and social environment. In particular, I want to thank Abbey Steele, Laia Balcells, Steve Kaplan, Harris Mylonas, Ryan Sheely and Nathaniel Cogly for their feedback and support. Much of the writing happened at the University of Chicago where I spent two stimulating years as a postdoc. I will never be able to thank Dan Slater enough for his insightful feedback on numerous drafts and his continued support. I also received valuable suggestions from John McCormick, Erica Simmons, Alberto Simpser, Paul Staniland, Lisa Wedeen, Susanne Wengle, Stan Markus and the participants of the Comparative Politics workshop. Colgate University provided a very supportive environment in which I was able to complete the manuscript. I am particularly grateful to Jeff Herbst, Michael Johnston, Illan Nam, Bruce Rutherford, Ed Fogarty and Jon Hyslop for their feedback. Albert Naïm provided excellent research assistance and Justin LoScalzo and Jeff Potts helped create figures and maps. Many others provided useful comments at many different stages. I am particularly grateful to Nic van de Walle, Evan Lieberman, Leo Arriola and Dan Posner. I graciously acknowledge the generous financial support for field research in Senegal and Benin from Fulbright-Hays, Yale Program on Democracy, the Yale MacMillan Center and Colgate University. Parts of the manuscript previously appeared as “King Makers: Local Leaders and Ethnic Politics in Africa,” 2013, World Politics 65, 2: 187–232. An earlier version of Chapter 5 was published as “Urban and Rural Voting Patterns in Senegal: The Spatial Aspects of Incumbency, c.1978–2012,” 2013, The Journal of Modern African Studies 51, 4. The material is reprinted with permission. Finally, I want to thank my family: my parents, Marek and Graz˙yna Koter, my husband, Rob Becker, my parents-in-law, Phil and Cindy Becker, and my children, Alassan and Natalia. This book is for them.

Abbreviations

ABT ADD AFP AJ/PADS APR BDP BDS BJP BNF BPP CENA CPP CVELI ELF FARD-Alafia FCBE KANU LD/MPT MADEP MFDC NDC NPP PDG PDS PIT PND

Alliance pour un Bénin Triomphant Alliance pour la Démocratie et le Développement Alliance des Forces de Progrès And-Jëf/Parti Africain pour la Démocratie et le Socialisme Alliance pour la République Botswana Democratic Party Bloc Démocratique Sénégalais Bharatiya Janata Party Botswana National Front Botswana People’s Party National Electoral Commission (Benin) Convention People’s Party Cramer’s V Ethno-linguistic Voting Index Ethnolinguistic Fractionalization Index Front d’Action pour le Renouveau et le Développement Forces Cauris pour un Bénin Emergent Kenya African National Union Ligue Démocratique/Mouvement pour le Parti du Travail Mouvement Africain pour la Développement et le Progrès Mouvement des Forces Démocratiques de la Casamance National Democratic Congress New Patriotic Party Parti Démocratique de Guinée Parti Démocratique Sénégalais Parti de l’Indépendence et du Travail Parti des Nationalistes Dahoméens xiv

List of Abbreviations PNS PRD PREG PS PSB PSD PSP PVD RB RDA RDD RND RP RPG TDS UBF UDD UFDG UN UPS US

Party Nationalization Scores Parti du Renouveau Démocratique Politically Relevant Ethnic Group Index Parti Socialiste Parti Socialiste du Bénin Parti Social Démocrate Parti Soudanais Progressiste Parti de la Vérité pour le Développement Renaissance du Bénin Rassemblement Démocratique Africain Rassemblement Démocratique Dahoméen Rassemblement National Démocratique Rassemblement du Peuple Rassemblement du Peuple de Guinée Takku Defaraat Sénégal Union pour le Bénin du Futur Union Démocratique Dahoméenne Union des Forces Démocratiques de Guinée Union fait la Nation Union Progressiste Sénégalaise Union Soudanaise

xv

1 The Challenge of Winning Votes and Ethnic Politics in Africa

The challenge for politicians to forge linkages with voters is a substantial one in all democracies. This is particularly true in the relatively young, immature, weakly institutionalized democracies of sub-Saharan Africa. Along these lines, conventional wisdom and a great deal of scholarship argue that ethnic bonds are the most important link, the easiest default option, in the African political arena. In this book, I challenge that portrait by highlighting that even in some ethnically diverse societies politicians often forge very different types of personalistic links. I show that the use of ethnicity1 is just one possible strategic choice on the part of calculating politicians. We now have a sufficient historical record to demonstrate that African voters and African politicians are far more dynamic in their choices and behaviors than has been previously recognized. In fact, we can predict when politicians are more or less likely to use ethnic appeals given the resources available to them. In contrast to much of the existing scholarship, I argue that in crafting their mobilization strategies, politicians don’t only look at demographics, such as the ethnic composition of the electorate, but pay attention to 1

The term “ethnic identity” is conventionally used by comparative political scientists to denote identities based not only on ethnicity per se but also on language, race, religion, caste and tribe (Horowitz 1985, Varshney 2002, Chandra 2004, Wilkinson 2004, Posner 2005). Chandra (2006) provides the clearest definition that captures the conventional meaning of the term. It states simply that ethnic identity is a subset of identity categories that is based on descent. Chandra aptly points out that we should rid the definition of characteristics of ethnic groups, such as common culture or myth of origin, that are variable rather than intrinsic. Throughout this book, I follow the convention and use the term “ethnic identity” to refer to identity categories, such as religion, tribe, language in addition to ethnicity in a strict sense.

1

2

The Challenge of Winning Votes in Africa

patterns of influence and dependence. Social structure, in which electoral politics take place, plays a central role in politicians’ decisions of how to mobilize voters. In particular, politicians are likely to consider the strength of local leaders, such as chiefs, or religious dignitaries, who have historically played a crucial role in many parts of rural Africa. The existing literature acknowledges the importance of local leaders in rural Africa, but it fails to consider how they alter electoral mobilization. Local leaders are pivotal, if undertheorized, actors. They change the electoral dynamics because they can help politicians secure votes among non-coethnics. Ethnic politics thus can be avoided where there are local leaders who can serve as credible electoral intermediaries between voters and politicians. Yet, politicians can’t mold social structure to their liking in the short term; it takes years to build trust and following that are necessary to command authority on the ground. Instead, in crafting their strategies, politicians are responding to preexisting conditions, which are historically contingent. Because of widespread variation in the standing of local leaders across Africa – an artifact of precolonial and colonial legacy, exogenous to the earliest mass elections – politicians have mobilized voters in qualitatively different ways, resulting in strikingly different levels of ethnic mobilization across the continent. As the main cases in this book will show, politicians made varying use of ethnic appeals. In Senegal, where local leaders have been historically strong and remain influential to this day, several generations of politicians built ethnically diverse clientelistic networks through local leaders, avoiding electoral mobilization of ethnic groups. In contrast, in Benin, where local authority figures were severely undermined during the colonial period, politicians consistently resorted to appeals to their coethnics, generating ethnic electoral patterns.

why ethnic politics? the problem of winning votes in africa Ethnic politics are often considered the norm in Africa, largely due to the difficulties of forging ties with voters. Underdeveloped media, linguistic fragmentation and a poor communication infrastructure make it hard for politicians to connect with voters. Whereas political choices in developed democracies are to a large degree determined by ideology and programmatic differences, this framework is widely viewed as less suitable to developing ethnically diverse democracies. Indeed, limited ideological or programmatic differences are some of the trademark features

Why Ethnic Politics? The Problem of Winning Votes in Africa

3

of political competition in African inchoate democracies (van de Walle 2003, 2007). Most African parties do not present voters with competing strategies of developing and running their countries. As van de Walle notes, “[I]deological differences have been minor across parties, and debates about specific policy issues have been virtually non-existent” (2007: 62). Case studies of individual elections consistently confirm this view over time. Describing political competition in West Africa in the 1950s, Thompson (1963) notes that parties were not programmatic and that one should rather view them as “cliques around personalities.” In his study of Ghanaian elections in the 1990s, Jeffries (1998) stressed the virtually nonexistent programmatic difference between parties and candidates. Joseph highlights that ideological content is often present in exchanges between party enthusiasts, but not in the actual recruitment of supporters (1987: 36). The Far Left or Communist parties, which have a consistent platform, have been a notable historical exception to this trend, but they are electorally marginal. Recent studies add nuance to this view: They show that African voters are not indifferent to what their governments are doing and to the qualities of their candidates. People notice economic progress or lack thereof (Posner and Simon 2002, Bratton, Bhavnani and Chen 2011), and they often judge their candidates perhaps not on their platforms but on their character, qualifications or their career achievements. Several studies highlight the importance of evaluative voting behavior (Lindberg and Morrison 2008, Hoffman and Long 2013, Weghorst and Lindberg 2013). Some prominent candidates in recent years campaigned as “technocrats,” even if they did not fit neatly into an ideological category. Other politicians, such as the late president of Zambia, Michael Sata, ran populist campaigns (Resnick 2011, 2014). The discourse on the campaign trails throughout Africa is not devoid of discussions of problems and challenges. Politicians do refer to issues important to voters, such as unemployment or the scarcity of basic services, but their pronouncements are best described as valance rather than position issues; politicians promise to tackle social problems, but they do not articulate competing solutions to these challenges (Bleck and van de Walle 2011, 2013). Yet, despite a growing variety of concerns of African voters and their demand for solutions, few parties in Africa present a distinct policy platform, and policy debates during electoral campaigns remain scarce. Some scholars indicate that this low salience of ideology contributes to the prominent role that ethnicity plays in electoral politics. As Ottaway

4

The Challenge of Winning Votes in Africa

has pointed out, “[T]he absence of ideological or programmatic differences left ethnicity as the major characteristic by which the various parties could differentiate themselves” (1998: 311, quoted in van de Walle 2007: 63). In contrast to weak political structures or underdeveloped political platforms, ethnic identities are a very tangible aspect of life in Africa. Ethnic identities are socially salient: People define themselves and others in ethnic terms, and one’s ethnicity is usually easy to decipher. Voters do not need much information to know a given politician’s identity. Ethnicity acts as an easily available “organizing principle” (Ajulu 2002). In addition to the limited role of programmatic politics, the primacy of clientelist access to resources is the principal reason why ethnicity seems to be an important factor in electoral politics. Distribution of resources is one of the most tangible stakes of electoral competition. Voters try to affect how resources are allocated and make sure that they are not left out. In many developing countries, especially in Africa, a substantial part of resources is spread through clientelism or redistribution targeting specific communities (Lemarchand 1972, 1988, Bratton and van de Walle 1997, van de Walle 2001, 2007, Lindberg 2003). The rush for spoils, or what Bayart (1987) evocatively called the “politics of the belly,” has been an important dynamic in African politics since the first mass elections. The conditional benefits offered to African voters range from cash, small consumer goods, bureaucratic intermediation to collective goods for communities, including wells, roads, school buildings or water pumps. Given the high centralization of power in most African countries (van de Walle 2001), insufficient government transparence and a relatively weak private sector, access to elected office has very important implications for ordinary people’s lives. It is worth highlighting that in contrast to many cases of clientelism in Latin America that are dominated by individual benefits to voters,2 collective communal benefits play a significant role in African political competition. African politicians distribute gifts and cash to individuals during political campaigns, but they also make promises to provide important infrastructure to voters’ villages or neighborhoods, conditional on voters’ electoral behavior. The competition over resources thus should not be equated merely with individual vote buying; it takes a much broader range of contingent transactions.3 2 3

See, for example, Auyero (2001). The contingent nature of provision of public goods makes this practice clientelistic. This view is consistent with Kitschelt and Wilkinson (2007).

Why Ethnic Politics? The Problem of Winning Votes in Africa

5

Although expressive motivations for ethnic voting should not be discounted,4 most scholars believe that voters choose candidates from the same group because of expectations of material gains. Bates argues that “ethnic groups persist largely because of their capacity to extract goods and services from the modern sector and thereby satisfy the demands of their members for the components of modernity” (1974: 471). Likewise, Kasfir (1979) suggests that groups use ethnicity to advance their goals, improve their own share of economic rewards and avoid domination by others.5 For Joseph (1987), political competition between ethnic groups is about the division of what Nigerians call “the national cake.”6 More recent studies support the connection between competition over resources and ethnic politics, and they articulate more clearly the specific mechanism through which ethnicity helps advance voters’ material goals. Posner argues that in situations of information scarcity, such as lack of credible policy platforms, ethnic affiliation gives voters credible information about which groups will benefit, if a given party or candidate wins the election (Posner 2005: 104). Similarly, Chandra (2004) claims that ethnic cues act as information shortcuts about who will benefit from a given politician’s policies. Arguably, voters believe that politicians from the same ethnic group will favor their group more than a non-coethnic would. As van de Walle puts it, “[C]itizens may feel that only a member of their own ethnic group may end up defending the interests of the ethnic group as a whole, and that voting for another ethnic group will certainly not do so” (2007: 65). Yet, despite widespread assumptions about the importance of ethnicity in African elections,7 there is significant empirical variation in the extent to which ethnicity plays a role in politics. The primacy of ethnicity in electoral politics manifests itself in ethnic voting, namely, voting for a coethnic politician, and the existence of ethnic parties or candidates,

4 5

6

7

See, for example, Horowitz (1985). See also Skinner (1985) who argues that ethnic groups compete for material goods and the resources of the state. Joseph further argues that this competition for material goods, and hence the control of the state which governs access to them, further accelerated the “ethnicizing of Nigerian society” (1987: 49). Similarly, Young highlights “the importance of scarcity of resources and competition for status in crystallization of contemporary identities” (1982: 89). See, for example, Dresang (1974) and Posner (2003, 2005) on Zambia, Kaspin (1995) on Malawi, Chazan (1982) on Ghana, Ferree (2004) on South Africa, Young (1976) on Congo and Ndegwa (1997), Ajulu (2002) on Kenya.

6

The Challenge of Winning Votes in Africa

i.e. those that garner most of their support from their coethnics.8 We can think about ethnic candidates and parties as the opposite of national candidates and parties. National parties and national candidates draw support from all groups in society, and the composition of their electorate is broadly similar to the composition of the general public. Ethnic parties and ethnic candidates, in contrast, are much less representative of the electorate because they largely rely on votes from one or two allied ethnic groups.9 Importantly, there is a gap between ethnic diversity and ethnic politics. For example, Daniel Posner’s Politically Relevant Ethnic Group (PREG) Index, which measures the politicization of ethnic cleavages, shows that there are several countries in Africa with high levels of ethnic fragmentation but low levels of politicized ethnicity (2004: 856). Other studies of political dimensions of ethnicity, such as Wimmer, Cederman and Min (2009) and Cheeseman and Ford (2007), further indicate that diversity is not automatically translated into ethnic politics. This gap raises an important question, namely, why is ethnicity politicized in some contexts but not others, and when does ethnic diversity lead to ethnic politics? There are countries in Africa without ethnic electoral patterns in otherwise similar environments. Consider the case of Senegal. Like most countries in Africa, it is a very diverse society, not only in terms of ethnicity but also religion. Its parties do not present substantially different programs; policy debates are rare or nonexistent; and clientelism, or the proffering of material goods in return for electoral support,10 is and always has been an important component of political competition.11 Yet, none of the major parties or candidates has an ethnic or religious base. Instead, each party’s electorate is as diverse as the electorate as a whole. Senegal is an illustrative example, but it is not an isolated phenomenon; there are other African countries where we see clientelist, nonprogrammatic competition and yet no ethnic politics. An index, developed by Dowd and Driessen, which measures the association between ethnic identity and vote choice, provides a good illustration of this variation (see Figure 1.1). The values of the index can be interpreted as the percentage of vote choice that can be predicted by voters’ ethnic 8

9

10

Horowitz defines an ethnically based party as a party which “derives its support overwhelmingly from an identifiable ethnic group (or a cluster of groups) and serves the interests of that group” (1985: 291). Given the domination of African politics by individuals, rather than parties, throughout the book, I study candidates and parties, rather than parties alone. 11 Definition of clientelism from Stokes (2007). See, for example, Beck (2008).

Why Ethnic Politics? The Problem of Winning Votes in Africa

7

C

Le s ap oth e o T Bo ow ts n wa S n M en a oz e am ga bi l qu e Ta Ma nz li an ia M Gh ad a n a a So ga ut sca h A r Zi fri m ca ba bw N e ig e U r ia ga nd M a al N awi am ib i Ke a ny Za a m bi Be a ni n

0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

figure 1.1 Level of Ethnic Politics in Africa

identity. The graph shows that while ethnicity is a good predictor of vote in some places, such as Benin, Zambia or Kenya, where it can account for over a third of vote choice, it has little explanatory value in Botswana, Senegal or Mali (under 20 percent). Why then do ethnic electoral blocs emerge in some countries but not in others? The case of Senegal also helps highlight other electoral anomalies, which are difficult to understand by focusing on ethnic identity. In the most recent presidential election in 2012, President Abdoulaye Wade enjoyed very high levels of support in his hometown of Kébémer. Winning by a high margin in one’s home region is not surprising, but what is more remarkable is that the inhabitants of Kébémer did not start voting for their “favorite son” until he became president. When Wade was a challenger, his town preferred to side with the then incumbent, Abdou Diouf, who was from a different ethnic background and who had no personal connections to the area. Kébémer was not the only area in Senegal that exhibited strikingly different support toward presidential candidates, depending on their incumbency status. Indeed, the vast majority of rural areas throughout Senegal changed their electoral allegiance. As a challenger, Abdoulaye Wade had sparse support in the countryside, but as an incumbent, the rural areas voted overwhelmingly for him. This shift was so dramatic that between 2000 and 2012, Wade’s electorate changed its composition from a largely urban one to one dominated by rural voters. In some rural areas, the incumbent increased his support by over 50 percentage points in just a few years’ time. Why does the same candidate have an urban base with little rural support as a challenger but builds a predominantly rural base as an incumbent? Theories that focus exclusively on ethnic ties

8

The Challenge of Winning Votes in Africa

neglect the importance of incumbency advantage and the spatial dynamics of electoral support in urban and rural areas. In this book, I engage with these questions, helping paint a more complete understanding of electoral politics in Africa.

existing theories of variation in ethnic politics Despite puzzling empirical variation, there is a shortage of answers as to why ethnic politics emerge in some ethnically diverse settings but not others. In fact, seeing ethnic mobilization in Africa as almost inevitable most scholars over the past several decades tended to ask why one particular dimension of ethnicity becomes salient, and not why any dimensions of ethnicity become salient at all (Laitin 1986, Posner 2005). Although several recent studies highlight that, despite the conventional wisdom, ethnicity is not a perfect determinant of vote choice in Africa (Bratton and Kimenyi 2008, Hoffman and Long 2013), most of the works have focused on why, in countries with widespread ethnic politics, some individuals vote ethnically whereas others do not, rather than why there are significantly different levels of ethnic voting in different societies. In an experiment in Uganda, Conroy-Krutz (2012) finds that as voters gain more information, especially negative, about their coethnic politicians, they are less likely to support them. Ichino and Nathan (2013) provide a compelling argument, with evidence from Ghana, that when voters are an ethnic minority in a district they are less likely to vote for their coethnic politician. These studies elucidate important individual or local variation in the propensity to vote for coethnics, but they do not address the question of why entire countries or regions with similar socioeconomic characteristics have surprisingly different levels of ethnic politics. Why is ethnic politics rampant in Benin but visibly absent in Senegal? There are no convincing explanations for the divergent electoral patterns in Senegal and Benin, the key set of cases examined in this book. First, it is important to point out that the absence of ethnic politics in Senegal does not result from a lack of social salience of ethnicity. Ethnic categories in both Senegal and Benin are regularly used by people to describe themselves and others; they also feature in official documents, such as censuses. Ethnic labels have social meaning and markers, such as names, rituals or stereotypes attached to different identities.12 As Diouf 12

Certain family names in Senegal, as in Benin, are associated with different ethnic groups. For example, Bâ or Diallo would be immediately identified as Peul.

Existing Theories of Variation in Ethnic Politics

9

points out, among most Senegalese there exists a “certain dose of ethnocentrism: one has a very flattering auto-portrait of one’s own ethnic group,” and one paints portraits of other ethnic groups “made of prejudices” (1994: 61). A study commissioned by UNESCO found that, for example, the Wolof view the Tukulor as overly conservative, whereas the Tukulor and other ethnic groups describe the Wolof as loud (gueulard), proud, materialistic and deceitful (Diouf 1994: 57).13 The Diola frequently describe the Wolof as “impolite and disrespectful” (Lambert 1998: 597). These ethnic stereotypes are an important corrective to the commonly held views by outsiders about the Wolof hegemony, namely, the dominance of the Wolof culture. McLaughlin (1995) shows that the spread of the Wolof culture in Senegal has generated resistance, especially among the Peul community. Wolofization has not eliminated ethnic differences. Nor has it privileged the Wolof in political competition. Only one president, Abdoulaye Wade, out of the four Senegalese presidents since independence, was Wolof. But perhaps it is in politicians’ interest to mobilize different identities instead of ethnicity? Advocates of institutionalist arguments (e.g. Posner 2005, Chandra 2004) suggest that politicians will mobilize along a cleavage, which creates groups closest in size to the minimum winning coalition at a given level of competition, national in the case of former French colonies. Based on these theories, one could hypothesize that Senegalese politicians do not mobilize ethnic identities because they are better off activating a different ascriptive identity, one which gets them closer to the minimum winning coalition. Yet, this is not what Senegalese politicians do, even though they could. Senegalese politicians do have other socially salient identity cleavages that they could activate, just like in Benin. In addition to ethnicity, political entrepreneurs could feasibly mobilize voters based on religion and brotherhood affiliation. Brotherhood affiliation in Senegal is highly 13

See also Smith (2006) for other ethnic stereotypes and McLaughlin (1995) on the manifestations of Haalpulaar (Peul and Tukulor) identity. Some additional stereotypes are based on the most common occupations of different ethnic groups: The Wolof are traditionally merchants, the Peul are pastoralists, the Tukulor are sedentary agriculturalists and the Serer and Lebou are fishermen. Moreover, ethnic groups have their ethnic homelands. Just like in Benin Abomey and its surroundings are the Fon heartland, or the Borgu is Bariba territory, the Senegal River Valley is home to the Tukulor, Casamance is considered the “Diola country” (pays Diola), Sine-Saloum is the heart of the Serer homeland and much of central Senegal is pays Wolof. See Diouf (1994: 33–40) for detailed data on ethnic composition of different regions.

10

The Challenge of Winning Votes in Africa

socially salient. There is a marked variation in religious practice and celebration, with distinct pilgrimages and holy places for different Sufi orders.14 People display allegiance to their brotherhoods by hanging pictures of religious leaders, or the founders of the respective brotherhoods. Many of my interviewees invoked common stereotypes about members of the two largest brotherhoods. For example, the Tijanis describe the Mourides as loud, boisterous and ostentatious, while portraying themselves as calm, discreet and modest. Given the plethora of social markers, politicians could in principle easily appeal to brotherhood affiliation in their political campaigns, just as their counterparts do across Africa. Yet, as the Senegalese electoral data indicate, none of the three most socially salient identities – ethnicity, brotherhood affiliation or religion – structures political competition, demonstrated later in Chapter 4. Nor do we find coalitions of different ethnic groups. It is thus not the case that Senegalese politicians don’t mobilize along ethnic lines because they choose to activate a different identity cleavage, even though they could feasibly do so. This outcome is inconsistent and even runs counter to the expectations of leading institutionalist theories (Chandra 2004, Posner 2005), in particular because the size of ethnic groups in Senegal and Benin is similar. While no ethnic group or brotherhood constitutes an outright majority, ethnic or religious groups in Senegal are no more fragmented than in Benin: The Wolof in Senegal constitute around 43 percent of the population, similar in size to the Fon in Benin, whereas brotherhood divisions create the largest group close to the desired 50 percent. The very similar sizes of the largest ethnic groups in Senegal and Benin allow us to account for Elischer’s (2013) important alternative explanation that countries with a majority (core) ethnic group are less likely to have ethnic parties than more ethnically fragmented states. The difference between Senegal and Benin cannot be ascribed to electoral demography and group size. It is also worth pointing out that electoral strategies in Benin do not follow the predictions of the minimum-winning coalition theory, as articulated by Posner (2005) and Chandra (2004). Based on their theories, it would be more advantageous for southern Beninese politicians to activate the “Southern” identity, a label that is very socially salient,15 rather than a Fon, Adja or Yoruba ethnic identity. If they followed this logic,

14

15

Touba is the holy place of the Mouridiyya, whereas Tivaouane is the holy place of the Tijaniyya. The most important pilgrimage for the Mourides is the Magal, whereas for Tijanis it is the Gammu. Banégas (2003: 8–9) provides a good description of the meaning of Southern identity.

Existing Theories of Variation in Ethnic Politics

11

Southerners could have a permanent hold on the presidency, as they constitute about two-thirds of the electorate. Instead, a southern politician has served only a single five-year presidential term (Soglo 1991–1996) since the introduction of multiparty elections 25 years ago.16 Other institutionalist theories, such as Ferree (2010), provide a compelling argument that could explain why ethnic census-type elections persist in Benin. Just as Ferree found in South Africa, Beninese politicians work hard to portray their adversaries as unlikely to represent noncoethnics, thereby fueling ethnic voting. But it is difficult to apply this theory to shed light on the Senegalese outcome, where politicians have never engaged in such behavior, even though they could. The varying levels of ethnic politics cannot be explained by electoral rules either. There are good reasons to expect that differences in electoral systems would affect politicians’ strategies of voter mobilization and would thus shape the ensuing landscape of parties (e.g. Brambor, Clark and Golder 2007). The main cases in this book were carefully matched to allow me to account for this factor. Senegal and Benin share key characteristics, such as multimember districts, that could affect electoral outcomes in legislative elections.17 More importantly, they have an identical electoral system in place for presidential elections, the “top prize” in African politics (Mozaffar and Scarritt 2005: 415). Both countries use a two-round system with a run-off when no candidate secures an outright majority in the first round. The divergent electoral outcomes are thus not a result of different electoral systems. What about other explanations? Some scholars propose that the existence of crosscutting cleavages can dampen ethnic polarization in the electoral system.18 They hypothesize that while overlapping cleavages can reinforce conflict, the existence of crosscutting cleavages diminishes the importance of any one identity dimension. Both Senegal and Benin have some crosscutting ties. In Senegal, an ethnic Wolof could be a member of several different Sufi brotherhoods. Likewise, in Benin, most ethnic groups are diverse with respect to religion. For example, an

16

17

18

In Chapter 2 I discuss why politicians might not want to follow the minimum coalition logic. The only institutional difference is that Benin has a pure PR system whereas Senegal uses a mix of PR and majoritarianism. The two countries also have an almost identical value of ethnolinguistic fractionalization (ELF), another factor that Brambor, Clark and Golder (2007) suggested can interact with the electoral system to shape the party system. Examples include Dahl (1956), Lipset and Rokkan (1967), Dunning and Harrison (2010).

12

The Challenge of Winning Votes in Africa

ethnic Fon could be a member of different Christian denominations or practice indigenous religion. Vodoun, the main indigenous practice, spans across ethnic and geographic lines. The existence of overlapping cleavages in general does not help us account for the variation in the extent of ethnic politics in our cases, but what about specific crosscutting ties? For example, Dunning and Harrison (2010), using a highly innovative experiment in Mali, show that crosscutting ties based on an informal institution of joking kinship, or cousinage, can counterbalance ethnic ties. Cousinage also exists in Senegal, but its importance in Senegalese politics should not be overstated. I witnessed how joking relations are routinely invoked in everyday interactions in Senegal, but direct evidence that they feature prominently in national electoral campaigns is lacking. For example, whereas appeals to religious leaders or other intermediaries, or ethnic appeals in Benin, are chronicled extensively in the press, references to politicians invoking cousinage ties with voters are strikingly absent. Even when nominal joking kinship exists, whether it is used may depend on local cultural and political entrepreneurs. Galvan suggests that joking kinship is not invoked uniformly and that “generalizations about its [joking kinship’s] causal impact across circumstances, time and space is nearly impossible” (2006: 824). He further explains that the use of cousinage ties is itself endogenous to existing alliances. Namely, people practice cousinage rituals with their existing allies but not their adversaries. For example, in some instances interethnic Serer–Wolof joking kinships have not emerged on any meaningful scale because of the ongoing political and economic competition between the Serer and Wolof.19 As Galvan argues, “[I]t is certainly the case that when individuals who belong to distinct, rival clientelistic networks enter into conflict, clientelistic loyalties often neutralize the cooperative effects of joking kinships that may bind the same individuals” (Galvan 2006: 823). In other words, the nominal ties afforded by joking kinship create an opportunity to counterbalance the salience of ethnicity, but this opportunity is not always exploited. The examples of ethnic tensions in Senegal’s 19

Galvan (2006: 818–819). As Galvan further explains, “[Among the Serer] redeployment of inter-familial joking kinship to include Wolofs seems to hinge on the degree of personalism, of face-to-face contact linking a Wolof Faye and the Serer Diouf.” Only when “such a relationship exists between people who live in the same community, are part of the same occupational or patronage network, or otherwise have cause to come into fairly regular face-to-face contact with one another,” actors have actively used joking kinships between the Serer and the Wolof (Galvan 2006: 819).

Existing Theories of Variation in Ethnic Politics

13

Casamance region, the site of an ethnic rebellion, and ethnic voting in the latest, and the country’s first, free multiparty elections in Guinea, both parts of West Africa where the institution of cousinage is present, show that this informal institution is not by itself a sufficient condition for interethnic cooperation.20 Another factor that has been suggested by scholars as potentially affecting the salience of ethnicity in the political arena is the extent of nation-building after independence. Miguel (2004) highlights the difference in, and subsequent consequences of, nation-building policies in Kenya and Tanzania. He notes that despite their shared geography, history and colonial institutional legacy, the two countries pursued qualitatively different nation-building policies (2004: 327). The respective independence leaders, Jomo Kenyatta and Julius Nyerere, had different philosophies regarding the role of national and ethnic identities. Whereas Tanzania focused on fostering national identity through the adoption of Swahili as an official language and medium of education, post-independence Kenyan leadership only inflamed ethnic rivalries. Miguel argues that nation-building is important because it creates a “taste for ethnic cooperation.” Although Miguel examined the effects of nation-building on public goods provision, one could plausibly expect that the same mechanism could apply to the development of cross-ethnic parties, predicting that such parties would be more likely where there was extensive nationbuilding. Yet, while the argument about the positive effects of nationbuilding is very compelling, the case of Tanzania is fairly idiosyncratic. Tanzania is frequently cited as an example of successful nation-building precisely because it is so rare. Neither of the West African cases examined here (nor other cases in Africa) experienced policies similar to the Tanzanian-style programs. Even though the use of Wolof is very widespread in Senegal, it is not an official language, and primary education is conducted in different vernacular languages. Wolof is used very successfully as a second language, but it has not eliminated other languages. Furthermore, a mere existence of a widespread language does not seem sufficient to thwart ethnic parties. Fon is virtually a lingua franca in 20

See Galvan (2006), de Jong (2005) and Villalón (1995: 54–56) on cousinage between Diola and other ethnic groups and ethnic rebellion in Casamance, and Dunning and Harrison (2010: 37) for the example of cousinage in Guinea. For the ethnic character of Guinean elections, see BBC (2012) and Jeune Afrique (2010). Incidentally, Guinea is also the country that witnessed under Sékou Touré the most extensive postcolonial destruction of traditional authority. I further discuss this in Chapter 6.

14

The Challenge of Winning Votes in Africa

southern Benin and it still did not prevent the creation of three distinct ethnic voting blocs in the South. More importantly, ethnic or nonethnic voting patterns emerged already before independence, prior to the era of nation-building. Thus we cannot view the post-independence leadership as a cause of ethnic politics, even though it certainly could have dampened or further inflamed the existing tendencies. A growing body of work focuses on the period of authoritarian rule to understand multiparty politics of the last two decades (Bratton and van de Walle 1997, Mozaffar and Scarritt 2005, LeBas 2011, Riedl 2014). In their landmark study, Bratton and van de Walle (1997) argue that the nature of prior authoritarian regime affected the prospects of a country’s transition to democracy. LeBas (2011) claims that the treatment of organized labor by the authoritarian leadership, namely, whether it was repressed or not, had important implications for the strength of the opposition under multiparty politics. Riedl (2014) argues that the authoritarian strategies of rule impacted institutionalization of democratic party systems. Mozaffar and Scarritt (2005) suggest that political restrictions under authoritarianism compelled politicians to use ethnicity as a source of strategic coordination. These works offer very compelling answers to their specific questions, but their arguments cannot be easily repurposed to explain the divergence in ethnic politics. First, while the authoritarian period in Africa was influential in many ways, the tradition of using or eschewing ethnic politics predates it. Wide cross-country variation in the degree of ethnic politics emerged already at the time of mass politics in the 1950s, preceding the time of authoritarian one-party politics in Africa. Moreover, the contemporary ethnic cleavages and electoral patterns are very similar to those early dynamics, indicating that the era of oneparty politics did not originally cause this divergence. Second, while Mozaffar and Scarritt (2005) provide a logical explanation of the continued usefulness of ethnic cleavages in African politics, it remains unclear why countries, such as Mali or Burkina Faso, with a history of restrictive authoritarian regimes have very low levels of ethnic politics in the multiparty era. Finally, it bears highlighting that the question of politicians’ ability to build ethnically diverse electorates is distinct from the issue of multiethnic coalitions. An impressive recent work by Leonardo Arriola (2013) explains why politicians in some African countries are able to construct electoral coalitions across ethnic divides at the time of presidential elections. Arriola convincingly argues that where business is autonomous

Existing Theories of Variation in Ethnic Politics

15

from state-controlled capital, as is the case in Kenya, opposition candidates can access the necessary resources to buy support from their potential competitors representing different ethnic groups, and thus stitch together multiethnic coalitions in presidential elections. While on a seemingly similar topic, Arriola’s study asks a different question and focuses on a distinct dependent variable. Multiethnic coalitions and multiethnic electorates differ with respect to the composition of politicians’ support and their degree of inclusiveness. While coalitions entail cobbling together a temporary alliance of ethnic groups A and B against groups C and D, politicians with diverse national electorates would have relatively similar levels of support among all those groups. The case of Kenya provides a good illustration of how ad hoc multiethnic coalitions differ from multiethnic electorates. In 2002 Mwai Kibaki was able to stitch together a coalition that included several ethnic groups, including the Luo delivered by Raila Odinga, but his own electoral base remained dominated by his fellow Kikuyus. Yet, in the following election, in 2007, Kenya saw large-scale ethnic clashes between the Kikuyus supporting President Kibaki and Raila Odinga’s Luo base. Unlike Kibaki who does not enjoy electoral support among the Luo, or Odinga who struggles among the Kikuyus, Senegalese presidential contenders, if you recall, garner support from all ethnic and religious groups. Kibaki’s once successful multiethnic coalition undoubtedly contributed to the first alternation in power since the reintroduction of multiparty politics, but it did little to eliminate ethnic voting, or to avoid ethnic conflict. Sadly, the 2007–2008 post-electoral violence, which claimed the lives of around 1,500 Kenyans and resulted in the displacement of further 200,000, was not the first such incidence in Kenya’s history.21 Previous ethnic clashes associated with electoral contests took place throughout the 1990s (Boone 2011, Branch 2012). In fact, many analysts cite Kenya as one of the quintessential examples of ethnic politics (Ndegwa 1997, Ajulu 2002). The individual-level data from the Afrobarometer indicate that Kenya is one of the countries with the strongest association between voters’ ethnic identity and vote choice, with around one-third of vote choice correctly predicted by ethnic identity alone. In the Afrobarometer’s sample of 18 countries, ethnicity was a better predictor of vote choice only in Benin.

21

Figures obtained from the BBC, April 12, 2008.

16

The Challenge of Winning Votes in Africa

a new theory of social structure and ethnic politics In this book, I argue that direct ethnic mobilization is not as unavoidable as commonly thought, despite widespread clientelism in Africa. This is because nominal ethnic ties may be less amenable to clientelism than it is expected. Ethnic ties are supposed to matter because they can serve as channels of clientelism and yet they do not inherently possess many of the features that are typically viewed as necessary for clientelism, such as personal relationships. Both classic and more contemporary studies of clientelism place a great deal of emphasis on personal ties as the basis for distributing resources.22 They highlight “personalized reciprocal relations between an inferior and a superior” and “personal problemsolving networks.”23 Ethnicity alone does not guarantee such relations between individuals and is therefore not as useful to voters as expected. There is in fact empirical evidence that brings into question the material benefits enjoyed by coethnics of officeholders. For example, Kasara shows that in Kenya peasants from the president’s ethnic group are actually taxed at a higher rate than other ethnic groups (Kasara 2007). Wrong argues that ordinary Kikuyus have “little to show for being from the president’s tribe” (2009: 52–53). Similarly, van de Walle finds no ethnic favoritism in Uganda (2007: 65). More recent studies find some evidence of ethnic favoritism, but they also highlight that such effects are variable across countries and that the conclusions one reaches depend on specific goods analyzed.24 Often, ethnic support is based more on perceptions than concrete patterns of redistribution.25 It rests on the hope of gaining access to resources rather than actual benefits. Why is there such a gap between persistent perceptions of ethnic favoritism and the much more mixed empirical record? The scarcity of information and the difficulty in analyzing the disbursement of money can largely account for this situation. Ordinary voters do not have the skills and access to sophisticated econometric techniques, as do economists and political scientists, to determine whether their politicians favor specific

22

23 24

25

For examples of classic studies of clientelism, see Scott (1972), Lemarchand and Legg (1972). More contemporary examples include Stokes (2007) and Brusco, Nazareno and Stokes (2007). Lemarchand and Legg (1972) and Brusco, Nazareno and Stokes (2007), respectively. These studies include Franck and Rainer (2012), Kramon and Posner (2013), Kramon and Posner (2012). See also Posner (2005, Chapter 4).

A New Theory of Social Structure and Ethnic Politics

17

groups. In such a setting, rumors can be more influential than concrete data. Perceptions of mass ethnic favoritism can also result from reports of elite ethnic favoritism.26 Politicians often have a stake in sustaining such perceptions of ethnic favoritism. Stoking fear of exclusion or mistreatment by non-coethnic politicians is a powerful electoral strategy (e.g. Ferree 2010). Because ethnic politics in reality bring uncertain material benefits, voting for a coethnic politician or campaigning through appeals to ethnicity may not be the best strategy, especially in political environments with alternative problem-solving networks. In other words, ethnic clientelism is not the only way of obtaining or distributing resources. On the contrary, politicians actually select different clientelist strategies in response to the social structure in which politics take place, and these strategies, in turn, create variation in electoral patterns. What I mean by social structure is not the demographic composition of a society but its level of stratification in the form of ties of dependence between local authority figures and their followers, what I call hierarchical ties. By local leaders I mean customary, typically unelected, authority figures, such as chiefs, royals or religious dignitaries. Their degree of actual authority, namely how much people listen to them, can range from extremely influential to rather ceremonial and insignificant. First, I suggest that politicians pursue two distinct modes of nonprogrammatic electoral mobilization: (1) by directly relying on the support of voters from one’s own ethnic background, and (2) by indirectly working through electoral intermediaries: local leaders who command moral authority, control resources and can influence the electoral behavior of their dependents. Yet, because the power of local leaders varies greatly, the option to use electoral intermediaries is not available in all settings. In environments without strong local leaders, politicians default to ethnic mobilization as the dominant non-programmatic strategy.

26

Wrong (2009) reports that Kenyan newspapers are full of articles about important contracts or appointments (to boards of trustees of public companies, etc.). Many of them in recent years have been awarded to the Kikuyu, feeding the perception of preferential treatment of President Kibaki’s group (Wrong 2009). Similarly, many southern Beninese complain that under Yayi Boni most prominent competitions (concours) for state jobs have been won by the president’s fellow Northerners. As Banégas explains, “Every week, ordinary people listen to announcements of administrative appointments on the radio and complain about ‘northern favoritism’” (2014: 453–454). As a result, Southerners became “convinced of the existence of a plan to concentrate power in the hands of northerners” (Banégas 2014: 453).

18

The Challenge of Winning Votes in Africa Strength of Hierarchical Ties:

Strong Local Leaders

Weak Local Leaders

Dominant Mode of Mobilization:

Indirect mobilization via intermediaries

Direct mobilization of ethnic groups

Electoral Outcomes:

Diverse nonethnic electoral patterns

Ethnic electoral patterns

figure 1.2 Summary of the Argument

In turn, in environments with powerful local elites, politicians can enlist them to mobilize voters. Finally, the choice of electoral mobilization is neither trivial nor cosmetic: It affects national electoral outcomes (see Figure 1.2). By severing the direct link between politicians and voters, intermediaries reduce a campaign’s reliance on shared identity between voters and politicians.27 Rather than being just another link in electoral mobilization, chiefs, religious leaders or ethnic elites open up the possibility for cross-ethnic electoral patterns. When politicians rely on ethnic solidarity with voters, they tend to create ethnic blocs. Yet politicians who enlist the help of local leaders may be able to access a far broader base of support. Intermediaries are motivated by material rewards for themselves and their community. It is thus in their interest to work with a politician most able to provide resources, rather than a coethnic, unless their coethnics are the ones with the most resources. By virtue of representing blocs of voters, intermediaries are in much better structural position than individual voters to forge deals with different politicians. Whereas ordinary voters cannot trust noncoethnic politicians to deliver resources (Posner 2005), influential local

27

I regard mobilization as ethnic when politicians search for votes among their own coethnics. Mobilization of voters for non-coethnic politicians is considered nonethnic regardless of the intermediaries’ identity. In other words, if voters and intermediaries have the same identity, but votes are delivered to politicians of a different background, then this is nonethnic mobilization. Voters and intermediaries very often, but not always, share the same identity. It should not be surprising that religious leaders have the same religion as their followers. Because many villages are mono-ethnic, in such circumstances village chiefs would naturally have the same ethnic identity as other villagers. The key difference is that ethnic mobilization links politicians with voters from their own ethnic group, whereas nonethnic mobilization links politicians with a broad range of voters from diverse backgrounds.

A New Theory of Social Structure and Ethnic Politics

19

Politicians’ Mobilization Strategies

Non-Programmatic/ Clientelist

Direct Mobilization of Ethnic Groups

Indirect Mobilization via Intermediaries

Programmatic

Policy, Technocratic & Populist Appeals

figure 1.3 Politicians’ Mobilization Strategies

leaders can credibly enforce deals with coethnic and non-coethnic politicians alike. Not all electoral politics in Africa is clientelist (see Figure 1.3), but in order to understand the difference between societies in the degree of ethnic voting, one has to study clientelist strategies. Burgeoning economic voting can coexist with entrenched clientelist practices, but it is the variation in clientelist strategies that explains whether or not we get ethnic voting patterns. The existing literature suggests that ethnic groups vote differently from one another, not because they have ideological differences but because they all want access to limited state resources (Melson and Wolpe 1970).28 Most non-clientelist campaigns in Africa invoke very general and inclusive messages of change, development or technocratic appeals, which do not cater to specific groups (Resnick 2014). For these reasons, it is the clientelist competition that is most responsible for ethnic voting. However, as I argue in this book, some clientelist strategies are much less likely to lead to ethnic voting than others. 28

New research by Lieberman and McClendon (2013) challenges the common assumption that ethnic groups have similar preferences regarding the distribution of state resources. They find that in most countries, preferences vary based on ethnic group membership. This is a very welcome correction to the conventional wisdom. However, even if ethnic groups differ on their most pressing needs or their choices of how resources should be spent, the characterization that political competition is about the distribution of resources to different groups remains valid.

20

The Challenge of Winning Votes in Africa

empirical approach An argument positing that social structure affects the type of electoral mobilization and the ensuing electoral patterns requires a historically grounded empirical approach. First, studying the interplay of social authority and electoral strategies only at present would be very problematic. We would not be able to eliminate the possibility that social structure is itself endogenous to political competition. It is reasonable to expect that decades of political competition affected the standing of local authority figures. To avoid this problem, it is thus essential to measure the standing of local elites prior to the onset of mass politics.29 Fortunately, in the case of West Africa, the onset of mass politics is easily identifiable; it happened in the 1950s, after a rapid and massive expansion of the franchise. I thus employ a comparative historical approach to explain why at the time of first mass elections some areas had robust local leadership, whereas other areas had weak local authority, and to document these differences. To this end, I study the precolonial variation in the strength of local leaders, as well as the impact of colonial rule on their position, using a rich collection of histories and ethnographies of various groups. This allows me to establish that the strength of local leaders was exogenous to electoral politics. Studying the historical moment of the introduction of mass politics offers additional leverage over alternative strategies.30 There are striking similarities between current electoral patterns and those from the first rounds of mass elections, despite an interlude of almost three decades of one-party dictatorship in all but a handful of countries on the continent. Virtually every African country that currently shows a high degree of ethnic politics already experienced this problem in the 1950s and 1960s. For example, contemporary ethnic cleavages in Kenya and Benin existed in a similar form a few decades earlier. Conversely, Botswana, Senegal and Mali, which do not have marked ethnic politics, did not experience them during the onset of mass politics either. The continuity between the electoral patterns in the 29

30

Henceforth, I will use mass politics as shorthand for mass electoral politics. It is important to note that unlike many other parts of the world, West Africa did not witness other types of mass mobilization, such as mass anticolonial movements. Mass electoral politics were thus the first form of mass political participation. One of the main benefits of the comparative-historical approach, especially when studying areas of the world where quantitative data is often limited or unreliable, is that it offers systematic and contextualized comparison, concern with causal analysis and emphasis on process (Mahoney and Reuschemeyer 2003: 6). In addition to providing rich narrative and process-tracing, carefully crafted controlled comparisons help generate internal and external validity (Slater and Ziblatt 2013).

Empirical Approach

21

1950s/1960s and the 1990s/2000s suggests a high degree of path dependence in electoral politics. Studies that consider current electoral patterns in a historic vacuum, risk overlooking critical junctures, which played a role in shaping electoral politics.31 The importance and the effect of the onset of mass politics should not be surprising to political scientists. Many scholars of electoral politics around the world highlight the remarkable durability of early electoral cleavages. Lipset and Rokkan (1967) most famously drew attention to the importance of founding elections and the “freezing” of original cleavages produced by these electoral contests. Wittenberg (2006) shows a striking continuity between pre-Communist and post-Communist support for rightist parties in Hungary, despite a half-century of dictatorial oneparty rule. Darden (2010) documents the persistence of political cleavages throughout Central and Eastern Europe, resulting from different imperial legacies. Some of the most prominent works on European electoral cleavages trace contemporary cross-country differences to the early years of political competition. Bartolini (2000) argues that the mobilization at the time of the extension of the suffrage to the lower classes in nineteenthcentury Europe explains the presence or absence of Socialist parties in contemporary politics. Likewise, Kalyvas (1996) successfully employs a comparative historical analysis to understand why there are Christian Democratic parties in some European countries but not in others. Scholars of US electoral politics also highlight the effect of critical junctures on voting patterns (e.g. Aldrich 1995). There are thus important theoretical reasons to believe that studying the founding elections is vital for our understanding of contemporary electoral dynamics. But why do electoral patterns persist? The early elections are important because all subsequent electoral strategies are devised in response to the existing patterns. Both winners and losers try to emulate successful campaigns. These effective dominant strategies then become “natural” ways to win votes.32 Once politicians become convinced that ethnic appeals win them votes, they will be reluctant to try alternative strategies. Conversely, when politicians originally win without making ethnic appeals, ethnic mobilization does not become the obvious, go-to 31

32

This book thus follows very well-crafted recent studies by MacLean (2010) and Riedl (2014) that make a convincing case that a historical perspective is necessary to elucidate contemporary political outcomes in Africa – different conceptions of citizenship and reciprocity for MacLean and different levels of party system institutionalization for Riedl. Riedl (2014) shows how emulation plays an important role in reproducing patterns of party politics.

22

The Challenge of Winning Votes in Africa

strategy. Once politicians win with the help of intermediaries, they become convinced that these vote-getters are essential for their success. Few politicians are likely to abandon winning tactics. The original strategies and patterns thus can create a dominant narrative about how politicians should campaign to achieve electoral success.33 As the empirical chapters will make clear, politicians and pundits in my case studies voice distinct “common wisdoms” about viable electoral strategies, be it through ethnic appeals, or with the help of intermediaries. The first elections thus provide templates for political actors that are then replicated over time. Electoral strategies and patterns are persistent and path-dependent but not immutable. My argument is that electoral strategies employed by politicians are conditioned by the landscape of authority at the local level. Significant changes to the social structure, such as the loss of authority by local leaders, should thus alter political strategies. However, I expect such changes to be gradual rather than abrupt and we should not see drastic differences in between consecutive elections. For example, urbanization can undermine the power of local leaders over the course of a couple of decades, but it will not happen overnight. In sum, the iterative nature of elections contributes to the stickiness of electoral patterns, but social structure continues to play a role.34 In the empirical chapters, I consider how local leaders were treated since independence to understand any continuity or change in their role in electoral politics and voting patterns. Social organization and electoral practices are not static, but empirically they have been fairly sticky, following the critical juncture of the first mass elections in the 1950s. Using contemporary survey data from the Afrobarometer and additional secondary sources, I demonstrate the high degree of continuity in the strength of local leaders since independence, which helps explain the durability of the electoral patterns. To establish the link between the strength of local leaders and the electoral strategies selected by politicians, I rely on process tracing. Since a crucial part of my argument postulates that politicians pursue distinct

33

34

This is akin to David Laitin’s (1986) argument about the hegemonic effect of the initial politicization of an ethnic cleavage. In the case of the Yoruba, once the ancestral city, rather than religion, was mobilized, it became the “natural” and obvious way of doing politics for decades to come. This approach is similar to MacLean (2010), albeit in a different context, where path dependence has to be bolstered by repeated interactions.

Empirical Approach

23

ethnic and nonethnic modes of political mobilization, gathering detailed data on electoral strategies and campaigns was vital for this project. For the first mass elections, I rely on monographs and other secondary sources on electoral campaigns. To study electoral strategies in the contemporary period, I conducted over 11 months of fieldwork (between 2005 and 2013) in Senegal and Benin. To gain an understanding and evidence of different modes of electoral mobilization, I attended campaign rallies, candidates’ and parties’ press conferences and various other campaign events. I conducted original, in-depth elite interviews with politicians from all major political formations in both countries, as well as with journalists, academics and pundits, in addition to focus groups with voters. I also studied over 18 months of press coverage from several different newspapers of the last six elections (presidential and legislative) in each country. Finally, I use disaggregated electoral data for both early and contemporary elections, much of it collected directly from electoral commissions in the field. This eclectic range of data sources, from historical records to ethnographic material from the field, was essential since this project aims to establish the process and mechanisms of electoral politics, rather than a mere correlation between different variables. Most of the data essential for this project could not be taken off the shelf but had to be gathered in the field or compiled from historical sources. Because this process was labor intensive and it required much local knowledge, it limited the number of cases that I could study. To achieve both depth and breadth, I combine a detailed structured comparison of the two West African countries of Senegal and Benin, at cross and subnational levels, which serve as the empirical core of the book, with additional African countries, both Anglophone and Francophone, which provide empirical extensions (see Map 1.1). I test the argument with two additional matched sets of countries (Guinea and Mali, and Kenya and Botswana) and I assess the generalizability of the argument with data from 13 African countries. Why were Senegal and Benin selected for the controlled comparison? West Africa is one of the most ethnically diverse and economically underdeveloped regions of the world. This is precisely the place where we would expect ethnicity to play an important role in electoral politics, not only because ethnicity is very socially salient but also due to widespread clientelism, and the limited role of programmatic politics. Yet, as the graph representing the variation in ethnic politics highlighted, the two countries fall on the opposite ends of the spectrum, with Benin being the quintessential example of ethnic politics, whereas Senegal is one of the cases

The Challenge of Winning Votes in Africa

24

SENEGAL

MALI

GUINEA

BENIN KENYA

Main Cases Secondary Cases

BOTSWANA

map 1.1 Main and Secondary Cases

where ethnicity seems to matter the least, despite salient social identities, including ethnicity, religion, caste and region.35 Although even the most carefully crafted structured comparison cannot control for everything, this pairing allows us to address several potential

35

Senegal is a more instructive case than some of the other countries in that sample with a low level of ethnic politics. Cape Verde and Lesotho are both tiny, homogenous countries, thus making a comparison with an ethnically diverse country more contentious. One of the most commonly used measures of ethnic diversity is the ELF index, which measures the probability that two random individuals in a given country will speak a different language (with a value between 0 and 1). On this measure, Senegal is even slightly more diverse than Benin, with an ELF value of 0.72, compared to 0.62 in Benin. ELF is not without its flaws; see, for example, Posner (2004a) for a thorough review.

Normative Implications of Electoral Politics

25

rival hypotheses, by holding a certain number of factors constant. Senegal and Benin are former French colonies, parts of the same administrative unit, French West Africa, with similar legacies and with important institutional commonalities, such as nationally centered political competition with a closed-list voting system and multimember districts. They both entered the era of mass politics at the same time, with the dramatic expansion of the franchise in the 1950s, when politicians for the first time faced the task of mobilizing a largely rural and illiterate mass electorate. These countries also had the two most educated elites in all of French West Africa. Whereas Senegal was the seat of French West Africa, with a high concentration of administrators and educational opportunities, Benin provided very large numbers of clerks and officials for the entire region, earning the country the nickname of African Quartier Latin. At present, they are the two most robust Francophone democracies in West Africa, with vibrant independent media and relatively good electoral data by the continent’s standards.

the normative foundations of an analysis of african electoral politics Why does it matter whether or not ethnicity is central to politics in various countries? While some studies find that ethnic diversity produces negative social and economic outcomes, such as lower economic growth or underprovision of non-excludable public goods (Easterly and Levine 1997, Kimenyi 2006), many scholars believe that it is the politicization of ethnicity, and not diversity per se, that is harmful to societies. Posner (2004a) points out that ethnic diversity should inhibit economic growth if ethnic divisions are actually politicized. He finds that his PREG index, which measures politicized ethnic divisions, has a much more negative impact on growth than mere ethnic diversity, as measured by ELF index (Posner 2004a: 860). Lieberman (2007, 2009) shows that ethnic politics lead to diminished expenditures on HIV/AIDS.36 Wimmer, Cederman and Min argue that it is ethnic politics, as measured by their Ethnic Power Relations index, and not mere diversity that is likely to lead to violent conflict (2009). Lieberman and Singh (2012) convincingly demonstrate with cases from southern Africa that institutionalized 36

Lieberman (2009) uses ELF as an independent variable and finds that ethnic fractionalization reduced AIDS-related expenditures, but he also finds that countries with relatively high ethnic diversity, but little politicized ethnicity, escape this pattern.

26

The Challenge of Winning Votes in Africa

ethnic categories have the tendency to lead to ethnic conflict. These studies indicate that scholars should pay attention to ethnic politics, which are a clear manifestation of the political salience of ethnicity, rather than just to ethnic diversity. Political competition along ethnic lines is commonly regarded as dangerous, both on normative and empirical grounds.37 Ishiyama (2011) argues that ethnic parties do not seek integration into broader national identities, but draw boundaries between ethnic friends and foes. Ethnic voting and the existence of ethnic parties limit voters’ choices, as politicians focus only on their own group. The primacy of ethnicity in elections also reduces incentives for parties or candidates to boost their credentials, apart from burnishing their identity profile. Political competition along ethnic lines encourages parties to abandon national platforms and to cater to specific groups (Wantchekon 2003). In other words, it reduces the inclusiveness of parties and candidates. It also creates a zero-sum game: An ethnic group as a whole is either in power or not. Ethnic groups who lose might be marginalized and left without resources. Bratton and Kimenyi (2008) argue that ethnic voting often typifies fear of exclusion from other parties. Lake and Rothchild (1998) further claim that ethnic competition for resources has negative economic consequences. When each group calls for group-specific benefits, they typically distort the economy and can reduce national wealth in the long run (1998: 10). Ethnic politics can also reduce politicians’ accountability. Some studies suggest that when ethnic voting blocs exist, politicians have fewer reasons to be accountable to their electorate and voters will tolerate more abuse from politicians (Padró i Miquel 2007). Dowd and Driessen (2008) also show that the degree to which party systems are ethnically dominated has negative effect on several measures of the quality of democracy. Their analysis indicates that in ethnically dominated party systems, people are less likely to perceive elections as free and they have lower satisfaction with their government (2008: 11). Finally, close electoral contests between ethnic parties raise the specter of possible violence. This is a scenario most recently materialized with deadly consequences during the 2007 presidential election in Kenya and in 2010–2011 in Côte d’Ivoire. This is not to say that ethnic politics are

37

Chandra (2005) is a notable exception to this view. She argues that ethnic politics can have both malignant as well as benign forms, depending on the effect of the institutional context on the fixity of ethnic identities (Chandra 2005: 245). She argues that when ethnic identities are not fixed they should not have a destabilizing effect on politics.

Normative Implications of Electoral Politics

27

inherently violent, but that they provide a ready-made script, should postelectoral violence occur. The problem of ethnic politics is particularly acute under multiparty politics. As Joseph argued, it is “competitive politics [that] encourages recourse to sectional identities” (1987: 43). The increase in competitive multiparty politics in Africa over the last two decades has thus brought renewed concern about ethnic politics. The salience of ethnicity never disappeared under one-party rule, which was prevalent in all but a few African states between the mid-1960s to the end of the 1980s, but it was very carefully managed. Ethnic accommodation was easier because leaders had to placate ethnic elites rather than the masses. Given the absence of legal opposition parties, situations in which one group, associated with a specific party, was pitted against another group, affiliated with a different party, were unlikely to happen. While ethnicity played a role at the local level (Posner 2005), for example in the selection of party candidates, in the absence of competitive politics at the national level, voters simply did not have the option to back ethnic candidates. Instead, the single party typically tried to bring to the fold representatives of all ethnic groups. As Rothchild (1985) highlights, co-optation of ethnic elites, as representatives of the ethno-regional periphery, was a widespread practice. Admittedly, several leaders privileged their own group over others,38 but ethnic balancing through informal grand coalitions of elites was the dominant strategy.39 Indeed, many post-independence leaders justified the rationale for single-party rule on the basis of national unity and avoidance of ethnic divisions (Foltz 1969, Wallerstein 1965). For example, Nyerere (1967: 196) expressed fears that multiparty elections would encourage ethnic factionalism. As one Tanzanian official put it, “African countries wanted a system of democracy which would avoid the divisive tendencies apparent in many democratic systems” (Cliffe 1967). General Soglo justified his military takeover in Dahomey (present-day Benin) by “the fear that the elections scheduled for 1966 might crystalize the North–South cleavage and result in disorder similar to that which prevailed among the Yoruba of neighboring West Nigeria during recent elections” (Zolberg 1968: 80). While it is tempting to dismiss the threats of ethnic politics

38

39

Some notable examples include Presidents Moi of Kenya, Eyadema of Togo or Banda of Malawi. Notable examples include Kenya under Kenyatta, Zambia or Ivory Coast (Rothchild 1985: 77).

28

The Challenge of Winning Votes in Africa

under multiparty politics voiced by African presidents as self-serving, they contained a kernel of truth. For example, Diamond (1988) argues that ethnic polarization was indeed one of the reasons for the failure of the Nigerian First Republic.

plan of the book The subsequent chapter outlines in greater detail the central theory of the book that social structure, and the strength of local traditional and religious leadership, affects the ways politicians mobilize voters, namely whether they make appeals to their own ethnic groups or craft ethnically diverse electorates with the help of electoral intermediaries. It explains why in the context of weak states, underdeveloped parties and limited media, local traditional and religious leaders can serve as electoral intermediaries between national politicians and voters. It stresses that in order to be able to play an intermediary role, local leaders need to have substantial authority and strong ties with their followers. It reviews ethnographic studies and contemporary survey research to highlight that the power of local leaders varies dramatically across and within African countries. This variation is crucial because politicians can use electoral intermediaries to mobilize non-coethnic voters only where strong local leaders exist. The chapter further outlines why mobilization via intermediaries does not lead to ethnic electoral patterns, in contrast to direct appeals to ethnic groups. Electoral intermediaries are motivated by material gains rather than ethnic affinity. Hence they have a preference for a politician/party most able to provide material resources. Intermediaries with a demonstrated following are in a strong negotiating position and they are able to make deals with non-coethnic politicians and deliver blocs of votes across ethnic lines, resulting in diverse voting patterns. Chapters 3 and 4 apply this theory to the cases of Senegal and Benin, focusing both on cross-national and subnational variation. Chapter 3 studies the variation in social structure in different regions of Senegal and Benin, using historical and ethnographic sources. It explains how precolonial variation in social organization and the transformative impact of colonial rule produced varying strength of local leadership throughout these countries by the eve of first mass elections in the 1950s. Whereas French colonizers destroyed most of the traditional leadership in Benin, Senegalese local leaders managed to preserve their authority in most parts of the country. The chapter then discusses why social structure sustained relatively little change throughout the post-independence era, and it

Plan of the Book

29

documents the persistent differences in the strength of contemporary local leadership, using recent survey data from the Afrobarometer. Chapter 4 then shows that this variation in social structure impacted how voters were mobilized and why ethnic voting blocs emerged in Benin but not in Senegal. The chapter employs original interview material from several dozen politicians and political experts in these two countries, media coverage, disaggregated electoral data from the 1950s to the present, in addition to various secondary sources. It first documents the two distinct modes of mobilization: Whereas Beninese politicians have been consistently making ethnic appeals, both in the 1950s and at present, their Senegalese counterparts have always mobilized voters through traditional and religious leaders. The chapter then studies the behavior of electoral intermediaries, demonstrating that in over 50 years they have not displayed a preference for coethnic politicians but rather for the bestendowed candidate (usually the incumbent) across the ethnic spectrum. The chapter also provides electoral data to show that where politicians mobilize through intermediaries, we find ethnically diverse electorates, in contrast to ethnic blocs, which emerge when politicians rely on ethnic appeals to voters. Chapter 5 exploits additional subnational variation to test the intuition that the stronger the tie between local leaders and their followers, the more effective the intermediaries are at influencing voters. Using survey data and rich secondary sources, I document the widespread and persistent difference in social structure and the role of local leaders in urban and rural areas in Senegal. These data show that rural voters are much more dependent on their local leaders than their urban counterparts. Consequently, they are more likely to follow those leaders’ voting suggestions, as the data indicate. I then test an observable implication of this argument. Since intermediaries tend to support the richest candidate, which in most African countries tends to be the incumbent president, we would expect incumbents to win in rural areas by higher margins. I use electoral data from multiple electoral cycles to substantiate this argument. In addition to data from Senegal, which provides the backbone of this chapter, I use additional data from other African countries. This chapter provides an important insight into understanding the urban/ rural cleavage, which is widespread throughout Africa. Chapter 6 tests the main argument of this book on additional cases. It engages in two different structured comparisons. First, it applies the theory to a set of neighboring countries in French West Africa, Guinea and Mali. Although the two cases had similar social structures in the

30

The Challenge of Winning Votes in Africa

precolonial period, and they share some of the same ethnic groups, traditional leadership at the time of the founding elections in the 1950s was much weaker in Guinea than in Mali, and it was weakened even further in the post-independence period. Contemporary electoral data show that while electoral politics in Guinea are marred by ethnic cleavages, Malian parties have ethnically diverse electorates. The second comparison moves beyond French West Africa to the former British Empire. It contrasts the state of traditional leadership in Kenya and Botswana, and their divergent electoral patterns. Kenya had extremely weak traditional leadership at the time of the founding elections and it has a long history of ethnic politics. Botswana, in turn, had some of the most robust traditional leadership, well-documented use of local leaders as traditional intermediaries and the absence of ethnic politics. The book concludes by discussing further applicability of the theory and its implications for understanding other phenomena, such as differences in the provision of public goods and distinct types of redistributive politics in Africa and beyond.

2 A Theory of Social Ties and Electoral Politics

introduction Ethnic politics violate important democratic ideals, such as accountability, voter choice and inclusiveness. In more serious instances, politicized cleavages can translate into conflicts between different groups. Despite these stakes, there is insufficient understanding of conditions under which politicians mobilize voters along ethnic lines. In this book I suggest that the strength of local leadership in different societies influences politicians’ choice whether or not to pursue ethnic mobilization. This chapter elaborates on the different elements of this argument in greater detail. I explain why local leaders provide an appealing alternative to ethnic mobilization and why they can act as intermediaries between politicians and voters. I outline how politicians’ ability to use intermediaries is contingent on the existing social structure. I make the case that social structure is exogenous to the earliest mass elections, and I discuss the sources of variation in the strength of local leadership across Africa. Finally, the last section elaborates on the implications of different electoral strategies pursued by politicians for the degree of ethnic politics.

alternatives to ethnic voting African politicians’ strategies of mobilizing voters have always been restricted, making ethnic mobilization an easy option and a common strategy. Nonethnic programmatic strategies could be a clear alternative, but their usefulness is limited, in particular in rural areas where the majority of the electorate lives. As Resnick (2014) notes, even populist 31

32

A Theory of Social Ties and Electoral Politics

strategies serve politicians well only in cities, making winning the majority of the electorate difficult. She demonstrates that even those politicians who use urban populism are forced to combine it with clientelist strategies in rural areas. To this day, the majority of African voters live in rural areas,1 despite rapid urbanization in recent years. The urban population was even smaller at the onset of mass politics in the 1950s, when no country in Africa was more than a quarter urbanized and the average urbanization level for the continent was below 20 percent (OECD, UN).2 Thus while the potential for using programmatic strategies is nonnegligible, African politicians still have to use clientelist strategies to win elections. But could they use clientelist strategies that do not cater to ethnic groups? Why don’t politicians, for example, build party machines to pursue clientelist strategies? The weakness of political parties and the personalistic nature of African electoral politics (van de Walle 2003) make this potential strategy considerably less suitable. Most African parties do not have much inherent value independent of their party leader, such as party labels or partisan attachments. The majority of parties are seen as extensions of their leaders, or in some cases even as “one-man shows.” Many African voters do not recognize party names and they refer to parties by the names of their leaders. A given party’s prospects without the current leader are viewed as highly uncertain. For example, the national coordinator of the ruling party in Benin acknowledged that he wasn’t sure how much of his party would remain once the incumbent president retires after his two terms in office.3 Likewise, in Senegal, at the height of President Wade’s rule in 2006, a journalist covering the campaign told me that within the ruling party, “the only constant factor is [President] Wade.”4 These views indicate how the fate of a party becomes highly uncertain without its leader. Because national politicians don’t draw many benefits from parties, they have few incentives to invest in them. Political actors are better off acting as independent agents, and there is a primacy of personal credit claiming. Party leaders often contribute to the image that the party would be nothing without them, whereas their internal rivals often threaten that they will leave the party and take their “share” with them. Many political

1

2 3 4

The urbanization levels in Benin and Senegal are approximately 43 percent (World Bank 2013 figures). In 2000, the figure for all of West Africa was 31 percent (OECD). The 1950 figure was 7 percent for Benin and 18 percent for Senegal (OECD). Author interview with Eugène Azatassou, Cotonou, Benin, January 4, 2013. Author interview with El Hajj Saïdou Nourou Dia, Dakar, November 1, 2006.

Mobilizing Through Intermediaries

33

actors believe that you don’t count in politics unless “you have a mass behind you.” This imperative to document how many people are loyal directly to you further encourages individualistic behavior. These conditions, which make classic machine politics impractical, make mobilization through intermediaries much more suitable. In the next section, I explain why electoral intermediaries are an appealing strategy of mobilizing voters in this context.

mobilizing through intermediaries The Importance of Local Leaders In their quest to win votes, politicians may employ traditional chiefs, religious dignitaries or other local leaders as electoral intermediaries. In doing so, they take advantage of preexisting relations of dependence. The strength of local leaders rests on their degree of authority and provision of material assistance. In the context of weak states, local leaders who enjoy the trust of their community play a very important role in people’s lives. In many developing countries, the scarcity of public goods and the difficulties of obtaining basic services from the state are striking. Krishna (2011) has shown in the context of India that most poor people cannot interact on their own with the state and that they are unable to secure independently necessary services. At the same time, political parties or local governments are not the most popular channels for gaining access to desired goods (Krishna 2011: 109). Such an environment creates demand for intermediation by local leaders. These individuals often act as social mediators, control access to resources and provide valuable goods and services. In sum, they provide essential safety nets to their dependents in an environment of poverty and unmet needs. Admittedly, the relationship between local leaders and their followers is complex; it can be based both on reciprocity as well as some degree of exploitation. Voters can trust and rely on their leaders but also feel trapped in their subordinate position. Traditional leaders,5 such as chiefs or royals, and religious dignitaries have been the most influential local leaders in many parts of Africa (e.g. Sklar 1993). As Miller describes:

5

The term “traditional leader” can be contentious. Without further qualification, it could imply that the position is unchanged and somehow sheltered from modernity. Miller (1968) provides much valuable nuance to this understanding. As he points out, “traditional” leadership is in fact syncretistic, as it combines “the opposing forces of

34

A Theory of Social Ties and Electoral Politics

Viewed from below, from the inner recess of the village, the leader is a man of authority; a man who used wealth, heredity, or personal magnetism to gain a position of influence. As seen by nation builders and development experts, the rural leader is tacitly pointed to as the key to success. It is he who can mobilize the people. It is through him that more energy will be expended, more muscles used, and more attitudes changed. (1968: 183).

Miller further explains that in rural Africa traditional leaders provide “the vital linkage between the government and the people” and many of them have “continued to dictate local policies and shape major decisions” (1968: 183). The alien nature of the postcolonial state and its continued lack of legitimacy contributed to the very real need of both peasants and government to have a rural intermediary with a degree of authority and influence (Miller 1968: 185). Since independence, local leaders, such as the marabout in Senegal, had the power to make or break government projects (e.g. Cruise O’Brien 1971, 1975). Recent studies also show the importance of local leaders, such as the Zambian chiefs, for the success of government development projects (Baldwin 2013). Logan (2011) highlights that traditional elites still play an important role in many parts of Africa because they represent their communities, help manage and resolve conflicts and are accessible to ordinary people. The continued importance of traditional and religious leaders in the lives of many rural Africans makes them valuable partners for national politicians. Local Leaders as Electoral Intermediaries In the realm of electoral politics, local leaders can be deployed as intermediaries, allowing national politicians to access an electorate where the latter otherwise have no reservoir of trust, as is the case among noncoethnics. Local leaders can affect the voting preferences of their dependents and assist them to access goods and services. By “bundling” votes, intermediaries gain leverage to acquire promises of resources on behalf of their followers. For voters, the logic of following an intermediary’s suggestion is simple. Voters want to find their champion, someone who can provide assistance to them and their community. But that champion need not be a politician from their ethnic group. For voters, knowing a patron with government connections may be more advantageous than relying on

traditionalism and modernism” (183–184). I use the term “traditional” with this nuance in mind. This term is used in common parlance by Africans and most people have intuitive grasp of its meaning.

Mobilizing Through Intermediaries

35

the campaign promises of redistribution to their ethnic group by someone they cannot easily reach. When voters depend more on local leaders than national politicians for most of their needs, they are more likely to be swayed by the former when casting their ballot. If a voter’s benefactor profits from the election of a particular candidate or party, such a voter can expect to experience diffuse benefits. In following an intermediary, voters are thus both repaying him for past and ongoing benefits, and are also trying to secure future advantages. When electoral politics is clientelist in nature, as is the case in much of Africa, reliance on intermediaries presents an appealing mobilization strategy for politicians. Instead of reaching out to the public on the basis of impersonal ties, politicians can create constituencies through personal connections existing between voters and their leaders. In doing so, they are taking advantage of the actual networks that people use in their daily lives. Personal relationships between local leaders and voters are valuable to politicians because of the difficulty in monitoring voters’ behavior. As van de Walle (2007: 53, 64) points out, African parties have little capacity to enforce vote-counting. Vicente and Wantchekon (2009: 294) highlight that in the context of a secret ballot, there is no clear enforcement mechanism for vote-buying from individuals. Politicians could give out gifts or cash before the election, but they have no effective strategy to ensure that voters will indeed support them.6 Engaging in direct votebuying on a mass scale is thus not a viable option for politicians or parties. African parties further lack their own agents who could assist with voter mobilization, as might be the case with more institutionalized parties, for example in Latin America. Instead, candidates or parties can subcontract mobilization to established local leaders with strong ties to the population. Ties of dependence between local authority figures and their followers, what I call hierarchical ties,7 are particularly suitable for the creation of clientelist networks and voter mobilization because they provide the necessary “social cement,” to borrow Susan Stokes’s (2007a) term. Clientelist systems thrive on deeply rooted organizational structures, based on face-to-face interactions, 6

7

Banégas (2003: 430) discusses this issue in the context of Benin and highlights that many voters take money from multiple parties but vote according to other criteria, mostly ethnic identity. Incidentally, many anti vote-buying campaigners do not discourage people from accepting gifts from politicians, but rather tell them to take what is offered but to vote according to their conscience. This approach illustrates politicians’ inability to effectively monitor individual voters. For lexical variety, I use “hierarchical ties” and “ties of dependence” interchangeably.

36

A Theory of Social Ties and Electoral Politics

personal loyalty and the social embeddedness of local powerbrokers.8 Most importantly, while it is very difficult, if not impossible, to monitor individual voters, it is easy to observe the electoral behavior of blocs of voters controlled by local intermediaries. In many countries, electoral results are available for each voting bureau, with only several hundred voters in each one, allowing politicians to reward or punish a whole unit. For example, in Senegal, electoral results are publicly displayed for each polling station, which ranges in size between 400 and 800 registered voters, typically one per village.9 The proposition that politicians can mobilize through intermediaries, instead of appealing directly to voters on the basis of identity, is not just a theoretical idea. Lemarchand documents that there are many instances of “traditional authority figures capitalizing on their social status to deliver votes of their constituents” (1988: 152). Already under colonial rule, many West African candidates to the French National Assembly in the 1950s depended on local intermediaries for their electoral support. For example, Joseph Conombo from Upper Volta (present Burkina Faso) was elected through a network of Mossi chiefs (Schachter Morgenthau 1964: 345). Another representative to the French Assembly, Hubert Maga from northern Benin, relied on the consent of “grand electors,” such as merchants, ruling families and Fulani chiefs (Staniland 1973: 307). After independence, the ruling Convention People’s Party (CPP) in Ghana used Dagomba chiefs as intermediaries, since they had a tight grip over their subjects (Staniland 1975, Ladouceur 1979 cited in Boone 2003). As Boone notes, the CPP concluded that “votes could be mobilized en masse by chiefs” (2003: 174). More recently, such strategy was observed after the return of multiparty rule in Ghana. In 1992, the incumbent president, Jerry Rawlings, made a concerted effort to cultivate the support of chiefs and elders and appealed to them to “advise” their people how to vote (Jeffries and Thomas 1993: 341). Similarly, Baldwin (2013) documents how politicians in Zambia courted chiefs in expectation of gaining the ballots of their dependents and that these chiefs did in fact have influence on the political choices of their followers. The use of intermediaries extends beyond Africa to other areas with low development. Writing in the 1970s, Scott argued that the use of local

8

9

These properties were highlighted by Stokes (2007), Schaffer and Schedler (2007), Powell (1970). Author interview with Babacar Kanté, Vice President of the Constitutional Council (Conseil Constitutionnel), a body overseeing elections, Dakar, Senegal, February 2, 2007.

When Do Politicians Use Intermediaries?

37

patrons as intermediaries in electoral politics in Southeast Asia was “as characteristic of the area’s contemporary politics as of its traditional politics” (1972: 101). As he explains, “[A] party succeeded best at the polls by securing the adhesion of the important local patrons, who would deliver their clients as a matter of course” (1972: 110). Among numerous Southeast Asian examples, Indonesian parties in the 1950s were forging links with voters in Java through local patrons, as they “struggled with one another for influence with the bupatis, the wedanas, and the tjamats” (Feith and Geertz quoted in Scott 1972: 110). Such a strategy was also used in nineteenth-century Europe, before the era of mass literacy and before the benefits of development accrued to the lower classes. Lehoucq (2007) describes the use of landlords as intermediaries in the nineteenth-century English countryside. Similarly, he notes that in Imperial Germany the Catholic hierarchy, landlords and factory owners were used to influence voters in the countryside (Lehoucq 2007: 36–37).

when do politicians use intermediaries? While the option of using intermediaries may be appealing to politicians, the viability of this approach depends on the strength of local leaders, because, as I argue, only strong leaders can act as credible intermediaries. Yet, empirically the power of local leaders has been highly variable across Africa. This variation in social structure thus had important implications for electoral politics as it presented politicians with different mobilization options. The robustness of ties binding local leaders and their dependents is crucial. In the absence of effective mechanisms of monitoring individual voters, intermediaries have to rely on their social clout to induce compliance. Put simply, leaders need strong ties with their dependents to be able to influence their vote. Moreover, robust hierarchical ties are also essential for intermediaries to be able to strike deals with politicians. A demonstrated position of authority in a community is necessary to convince politicians that one is capable of delivering votes, even across ethnic and confessional lines. In short, the strength of the ties determines voters’ compliance. This is precisely why intermediaries’ functions can only be performed by local leaders with strong hierarchical ties to the population. Voters are likely to follow intermediaries when (1) they depend on them financially and (2) they trust them to make decisions on behalf of

38

A Theory of Social Ties and Electoral Politics

their community.10 To have electoral sway, especially in the context of free and fair elections with a secret ballot, local leaders require both dimensions of power; they need to affect the well-being of those around them and enjoy a substantial amount of deference. Material dependence is a sine qua non condition because voters who are trying to maximize their access to resources are unlikely to follow poor leaders who cannot offer important services. Yet, local leaders who only have material power but no authority have serious shortcomings as intermediaries. Given parties’ and candidates’ weak capacity to monitor voting, intermediaries’ moral clout helps enforce voting in a given candidate’s favor. Since intermediaries cannot always know how individuals voted, and thus cannot successfully threaten to withdraw assistance, voters’ voluntary compliance significantly boosts intermediaries’ effectiveness.11 For example, unpopular landowners could coerce voters to vote according to their wishes, but their ability to do so when the secret ballot is respected is significantly diminished. The usefulness of moral authority was already appreciated during the colonial period by French officials, who found that popular and respected local authority figures were more effective at fostering cooperation. In sum, while material dependence is indispensable, moral authority strengthens intermediaries’ clout. Yet, not all communities have meaningful hierarchical ties and not all chiefs or religious leaders are suited to become electoral intermediaries.12 Even though religious and customary authorities exist across Africa, their prominence varies greatly. Boone (2003: 29) notes that historically there have been marked differences in communal structures across Africa. For example, while the northern Senoufo of Côte d’Ivoire have tangible authority figures, the “rural social structures across the South [of Côte d’Ivoire] shared a common social-structural feature: the extreme weakness of indigenous political hierarchy” (Boone 2003: 181). Nowadays, the standing of local leaders is not uniform either. Logan points out that, while over three quarters of Malians trust their traditional leaders, less

10

11

12

Various studies of local authorities highlight these two elements (Scott 1972, Herbst 2000, Boone 2003). This does not mean that leaders without moral authority can never deliver blocs of votes but that they are less effective, because they have to rely on coercion rather than voluntary compliance. As one senior Senegalese politician pointed out, many people are posturing as potential intermediaries. Yet, politicians are skeptical that such “upstarts” have any influence over voting behavior, when compared to more established powerbrokers, such as the marabouts. Author interview with Mbaye-Jacques Diop, Dakar, March 1, 2007.

Implications of Variation in Social Structure

39

than a third of Nigerians do so (Logan 2008). The power of local leaders can vary within countries as well. For example, in Ghana only 37 percent of the southern Akan group trusts their traditional leaders, but this figure is much higher among the northern Dagomba, standing at 70 percent (Afrobarometer Round 4, 2005).

implications of variation in social structure This varying strength of local elites is consequential, given that politicians can use intermediaries only if they have powerful local leaders to work with. As much as it might be tempting for political actors to manufacture intermediaries where they were hitherto lacking, such efforts are likely to fail. Strong relationships between local leaders and their followers are not built overnight. In any given election, politicians thus have to work with a preexisting set of leaders, rather than try to create them de novo. This does not preclude change in the intermediary landscape in the long term. Politicians’ actions might influence the standing of local leaders over a period of time. I expect politicians to adjust their strategies based on changes in social structure. I do suggest, however, that while one could swiftly undermine local leaders, it is much harder to create authority in a short period of time. Because of the iterative nature of electoral politics, the landscape of authority at the onset of mass politics was particularly influential for electoral outcomes, generating a substantial degree of path dependence. In the absence of influential local leaders, politicians had few viable alternatives to direct mobilization of their coethnics. In turn, where powerful local leaders existed, politicians were able to use intermediaries to make inroads among their non-coethnics. While there are clear benefits to relying on intermediaries, one might still question whether politicians will actually use them. Perhaps they could forgo the benefits of using intermediaries and prefer to appeal directly to voters. Yet, where there are strong local leaders, bypassing these actors carries potential costs. If local leaders are able to mobilize support in someone’s favor, they are just as capable of demobilizing voters for a given politician, either by discouraging turnout or by encouraging voters to support someone else. When facing the choice of using intermediaries, politicians thus are presented both with “carrots” and “sticks.” Given the combined effect of the benefits of using local leaders and the costs of bypassing them, politicians have strong incentives to employ them as electoral intermediaries, in particular because most politicians are pragmatic.

40

A Theory of Social Ties and Electoral Politics

Strength of Hierarchical Ties

Dominant Modes of Electoral Mobilization

Strong hierarchical ties

Indirect mobilization via intermediaries

Weak hierarchical ties

Direct mobilization of ethnic groups

figure 2.1 Predicted Dominant Modes of Electoral Mobilization

Independence-era politicians had different ideological proclivities, with more “reactionary” politicians ideologically opposed to chiefs and other local leaders, viewing them as an obstacle to progress, and “conservative” politicians, more willing to work with the existing power structures (Alexander 1970a). Yet, even those politicians who despised the chiefs frequently worked with them, because they were not always in a position to defeat them. Politicians could crush weak or unpopular chiefs, but not highly popular ones. For example, Kwame Nkrumah, who was known to be fiercely opposed to chiefs, fought those chiefs whom he could defeat, but had to work, albeit reluctantly, with the chiefs that he deemed too difficult to dislodge (Rathbone 2000).13 In this sense, while ideological proclivities might not be irrelevant, pragmatic politicians will nonetheless respond to the existing social structure in crafting their approach to local leaders and selecting their electoral strategies.14 In sum, in settings characterized by strong hierarchical ties, I expect indirect mobilization through intermediaries to become the dominant strategy, whereas direct mobilization of ethnic groups will prevail in contexts characterized by weak hierarchical ties (see Figure 2.1). These two modes of mobilization are not mutually exclusive, but it can be difficult to combine them in practice. I thus expect one strategy to become dominant in a given setting. In principle, politicians could rely on both strategies simultaneously. From any individual politician’s point of view, such an approach would make perfect sense: one appeals to ethnic ties among coethnics and enlists the help of intermediaries among noncoethnics. However, all politicians cannot successfully execute this strategy at the same time. First, if there are powerful local leaders in their ethnic homeland, politicians are unlikely to appeal directly to their own ethnic

13 14

See also Boone (2003: 146) on Nkrumah’s use of chiefs. This argument is in contrast to Riedl (2014) who considers politicians’ treatment of local elite to be exogenous.

Implications of Variation in Social Structure

41

groups for fear of paying the price of alienating those leaders. Second, politicians cannot take for granted the support of coethnic local leaders; this support is not free and in an effort to secure it, native politicians might get outbid by non-coethnic politicians. In conceptualizing politicians’ behavior, my theory shares some of the same assumptions as many other works on electoral politics. Like most studies, I assume that politicians’ main concern is winning votes. However, in contrast to some previous works (Chandra 2004, Posner 2004b, 2005), I expect politicians to maximize votes, rather than try to achieve a minimum winning coalition. Proponents of the minimumwinning coalition approach argue that since winning votes is costly, due to the resources that have to be distributed or promised to voters, politicians want to gain as few votes as necessary to take them past the electoral threshold ensuring victory (50 percent + 1 vote under majority systems; possibly even fewer under plurality). I do not expect politicians to follow this logic because of the benefits of a large majority and the potential pitfalls of the minimum-winning coalition strategy. First, there are substantial advantages to winning by a large margin. Politicians with high vote scores can claim to have an important mandate from the people. In parliaments, having a large majority makes passing desired legislation much easier and a super-majority allows constitutional changes. Leaders with sizable support have an important edge over their rivals. Indeed, a large victory can have an intimidating effect on political opponents.15 I acknowledge the fact that winning votes is costly and that any benefits of additional support have to be weighed by the costs of gaining that support. However, empirically, politicians have variable resources at their disposal. For some politicians, especially those who use state coffers to fund their campaigns, costs might not be severely limiting. I thus expect politicians to try to maximize votes, as far as they can afford it, rather than win by the narrowest margin possible. A second concern about the minimum-winning coalition logic is that there might be a danger for politicians of “cutting it too close.” By wanting to win by a slim margin, politicians could miscalculate. In environments with poor demographic data, “doing the math” accurately can be a challenge. The ease of obtaining precise data on ethnic composition of different districts should not be overestimated. Of course, there are instances when it is a well-established fact that a certain group has the majority. Yet, in other cases, where there are two groups of roughly 15

See, for example, Magaloni (2006).

42

A Theory of Social Ties and Electoral Politics

similar sizes, determining which one is the larger one might be quite difficult. Many developing countries do not conduct their national census at a frequent enough interval. For example, during the 2007 election in Senegal, the most recent publicly available census was from 1988. A group count that is almost 20 years old in a country with a very high population growth rate, where the majority of the population is under 18, is bound to be grossly inaccurate. Regardless of the accuracy of demographic data, placing so much importance on electoral math exaggerates the ability of politicians to actually mobilize every member of the group. Group size only tells you how many people are part of a given group but it doesn’t tell you if they are all going to vote for you. We don’t have a simple way of knowing the defection rate of members of an ethnic group from their coethnic candidate’s base, and if different groups have different defection rates, then the groups’ original sizes are not very informative.16 Furthermore, the minimum-winning coalition logic makes an implicit assumption that mobilizing a group smaller than the minimum-winning coalition is irrational and should be avoided. Yet, numerous candidates across Africa pursue this strategy nonetheless. Why might that be the case? Perhaps scholars take too narrow a view of electoral competition. Winning matters a lot, but it is not the only motivation for an individual who is casting a ballot, or for the politician competing in an election. While in highly programmatic systems, we regard voting for a candidate who is bound to lose as serving an expressive function, in nonprogrammatic systems, we tend to view such acts as irrational. Yet, this might not be the case. Voting for a doomed candidate can be strategic if voters can gain resources, even if the candidate loses. Voters can reap rewards for displaying their loyalty to a patron. Likewise, candidates who are set to fail might have instrumental reasons for running. A large part of politics is showing “your worth,” namely how many people you can mobilize. Commanding a sizable bloc of voters can allow losing candidates to extract important resources or concessions from the winning candidate in exchange for one’s support (see also Arriola 2013). Many African politicians have made a career of losing elections. For example, Bruno Amoussou in Benin has lost four presidential elections in a row, but he was able to leverage his support among his Adja group to secure prominent political positions, including the speaker (président) of the National Assembly. 16

Chandra (2004) acknowledges some of these limitations.

Accounting for Variation in Social Structure

43

While the size of a politician’s ethnic group can be a limiting factor as it places a cap on how many voters can be mobilized, in the case of mobilization via intermediaries, I expect politicians to try to win support of as many intermediaries as possible, given the resources at their disposal.

accounting for variation in social structure An argument about the effect of social structure on electoral politics raises potential fears of endogeneity. Fortunately, variation in social structure can be considered exogenous to political competition, because substantial differences in the strength of local leaders existed already before mass electoral politics, which began in Africa in the 1950s. But why was it the case that on the eve of mass electoral politics some societies had strong local leaders whereas others had weak authority structures? Social structure varied for two main reasons – due to precolonial variation and the impact of colonial occupation. First, precolonial societies had uneven patterns of authority and hierarchy. Although some had clearly defined structures of authority, others had no hierarchies above the village level or even beyond an extended family compound. Whereas the Wolof, the Fon or the Baganda are known for their elaborate social hierarchies and precolonial states, many other groups, such as the Diola or the Somba were acephalous (stateless) societies with weak social organization. Anthropologists conventionally describe the former type as hierarchical and the latter as “egalitarian,” indicating an absence of status differentials between members of society.17 Ethnographies of the precolonial period document and measure the variation in the degree of hierarchy of groups encountered by colonial administrators and ethnographers (see Murdoch 1967, van Rouveroy van Nieuwaal 1987). This variation can be regarded as completely exogenous; there are no clear patterns in the distribution of hierarchical and nonhierarchical societies. The open savanna of West Africa probably had a higher concentration of stratified societies than the forest region, but there are too many welldocumented hierarchical precolonial societies in the forest zone to view terrain and climate as a determinant of social organization. The second factor that affected social structure before the onset of mass politics was the impact of colonialism. Not all societies with strong 17

The term can be misleading because of its contemporary connotations. It does not reflect the treatment of minorities or women, but only the absence of delegation of power to higher authority.

44

A Theory of Social Ties and Electoral Politics

precolonial hierarchies retained their local authority throughout the colonial period. The colonial conquest and occupation had an uneven impact on traditional authority.18 In some cases, the colonizers preserved existing social structures, whereas in some instances they destroyed them. First, the impact of colonial intervention on traditional authority varied because not all societies were colonized at the same time and they did not occupy the same importance in the eyes of colonial officials. It is useful to think about different areas as either core or periphery. Boone (2003) showed how colonial administration in West Africa varied according to the economic importance of a given area. French colonial officials frequently divided their colonial possessions, between those areas that were useful (pays utile) and those that were not (pays inutile).19 Areas that were not of prime economic or strategic concern were governed with a much smaller administrative footprint, and consequently there was less disruption to traditional patterns of authority. As Alexandre explains, “[I]n the regions where there had been little economic or social development, generally those farthest from the coast, the legitimate customary chiefs, even the conservative if not reactionary ones, managed until very recently to keep their authority intact” (1970b: 42). In areas deemed strategically and economically important, compliance with the colonial administration was imperative. Many such areas experienced military conquest, rather than an establishment of administrative presence with scarce personnel. The French, in particular, did not hesitate to defeat, abolish, exile or replace traditional and religious elites that posed resistance. Whereas elites in the colonial backwater had a better chance of survival, as they were frequently left to their own devices, elites in core areas were doomed if they did not cooperate with the colonial powers. Their only chance of survival was to reach an accommodation with the colonizer; they had to avoid outright resistance and prove useful enough to the colonizer to avoid destruction, and also retain some autonomy in order to maintain their following and not to appear as an extension of the colonial state. It is important to note here that, much to their later regret, the colonial powers were much more successful at destroying existing authority than at remaking it to their liking. As the case studies

18

19

MacLean (2010) shows that the colonial state had a profound effect on village social institutions in West Africa. For example, in the context of Benin, the French considered only the southern part, between the coast and the town of Savalou to the north, to be “le pays utile” (Glélé 1969: 35).

Effects on Electoral Patterns

45

in subsequent chapters will show, local leaders created by the colonizers were seen as stooges and they rarely commanded authority. Changes since the Onset of Mass Politics These two factors, precolonial and colonial variation, affected social structure on the eve of mass politics in the 1950s – the critical juncture for the development of electoral politics. Yet, there is no reason to expect that social structure should remain immutable during the five to six decades following independence. Post-independence leaders had the time and the possibility to reshape the power of local authorities. Their actions could either consolidate or undermine the power that local leaders enjoyed around independence. One of the reasons why independence leaders could try to curb the power of local authorities is if they saw the latter as their competitors, a threat or even an obstacle to their country’s modernization. Political leaders could take actions to replace traditional leaders and revoke their privileges, such as allocation of land, as happened in Guinea under Sékou Touré (discussed in Chapter 6) or they could bolster traditional leaders’ authority by delegating more power and resources to them (discussed in Chapter 7). Riedl (2014) documents different strategies that politicians used vis-à-vis traditional authorities, distinguishing between those approaches that were supportive (incorporation) and destructive (substitution). Such actions created important feedback loops, potentially altering the strength of local leaders, with implications for future electoral contests.

effects on electoral patterns Different modes of electoral mobilization, pursued by politicians in response to variable social structure in which they were operating, have crucial implications for the ensuing electoral outcomes. Not surprisingly, the direct mobilization of shared ethnicity results in ethnic voting blocs. In contrast, when politicians use intermediaries, both actors have incentives to work across ethnic lines. For intermediaries, the advantages of engaging in political mobilization are straightforward: They can convert their social clout into material gains, “cashing in” on their involvement in the community. Support for a successful candidate may translate into wells, roads or other public goods for their locality. In addition, intermediaries may also expect personal goods, including cash, political positions or gifts in kind. For politicians, intermediaries are valuable because

46

A Theory of Social Ties and Electoral Politics

of their ability to deliver votes, which is particularly important when parties are weak, as is the case in most of Africa and much of the developing world. Politicians are willing to negotiate deals with intermediaries in which they exchange resources in return for blocs of votes. In the context of weak and non-programmatic parties, transactions between politicians and intermediaries resemble a free market, where neither set of actors has the incentives to limit themselves to working with members of the same group.20 Politicians want to find the most efficient intermediaries, whereas intermediaries favor those politicians best positioned to provide resources. Narrowing the range of acceptable patrons only to coethnics would limit intermediaries’ prospects of material rewards. Rather, it is in intermediaries’ interest to keep their options open to land the most lucrative offer possible. I argue that intermediaries’ choice of a political patron depends largely on material benefits. As a result, I do not expect intermediaries to be always loyal to one party or to consistently support their coethnic politicians. Intermediaries do not want to be unconditionally locked in an alliance with a particular party, knowing that as soon as that party loses, they would be better off in a different alliance. The weakness of political parties further reduces intermediaries’ incentives to stay loyal to a particular party. But wouldn’t intermediaries prefer to strike deals with coethnic politicians? For example, intermediaries might have an easier time convincing their dependents to vote for a coethnic politician, because of voters’ preexisting affinity for politicians from their background. Intermediaries themselves might also have such affinity for their coethnics. I would expect this preference to play a role if all political candidates had equal financial resources. Intermediaries would most likely support their coethnics, if they were offered the exact same benefits by coethnics and non-coethnics alike. However, I do argue that if intermediaries can exact higher benefits from non-coethnics, they will choose the material reward over ethnic affinity. I therefore expect that intermediaries will deal with the highest bidder from across the ethnic spectrum. If intermediaries can get higher rewards for themselves and their community from non-coethnics, they will support those candidates over

20

In settings in which ideology is more salient, there will be inevitably more constraints on brokers. In Argentina, for example, a Peronist broker would not be able to switch to the Radical Party and convince his dependents to follow him. I further discuss this in Chapter 7.

Effects on Electoral Patterns

47

coethnics. Intermediaries’ drive to maximize benefits might undermine politicians’ efforts to invoke ethnic solidarity. In sum, intermediaries have a preference for a politician who can offer them the best deal. Intermediaries might be also sensitive to what other intermediaries are doing, using their behavior as cues about different candidates’ chances of winning, or the depth of their pockets.21 A candidate who manages to enlist a lot of intermediaries, in a setting where such alliances are highly publicized, can convince other “vote getters” to join his team. One side effect of intermediaries’ preference for the highest bidder is a possible “bandwagon effect,” with a majority of vote carriers flocking to the best endowed candidate when one candidate or party has a substantial material advantage over others.22 .

Inequality of Resources and Incumbency Advantage It is a rare scenario when all politicians have comparable resources at their disposal and can make equally appealing offers to intermediaries. On average, incumbent presidents and their parties23 have an advantage over challengers. There are two types of rewards that politicians can offer to intermediaries: direct payments or gifts in kind during electoral campaigns and promises of future rewards contingent on electoral results (see also Arriola 2013, Chapter 2). Actors mix pre- and postelectoral payoffs because both types of rewards play important but different functions. On the one hand, rewards given before the election are disadvantageous to politicians because there is no clear enforcement mechanism.24 An intermediary could take the reward and not even try to deliver. Yet, while politicians might prefer to distribute benefits after the election to ensure enforcement and to reward actual intermediary support, promises of post-electoral transfers might not seem credible to intermediaries. This is a particular hurdle for challengers, who do not have an established track record and whose likelihood of winning might be

21 22 23

24

I am grateful to one of the reviewers for helping formulate this point. See Scott (1972: 110) on the bandwagon effect. When talking about incumbency advantage, especially in the context of French West Africa, it makes most sense to talk about the incumbent president and his party rather than incumbent MPs. As van de Walle (2003) points out, most African systems exhibit excessive presidentialism: most power and control of the state coffers rest with the president and his entourage. Individual incumbent MPs have little power. See, for example, Vicente and Wantchekon (2009: 294) on this point.

48

A Theory of Social Ties and Electoral Politics

perceived as relatively low, given the traditionally high reelection rates in Africa. Vicente and Wantchekon (2009) argue that promises of postelectoral rewards on the part of the incumbent seem more credible (2009: 293–295, 300–301).25 Magaloni (2006: 9) also claims that an established track record and uncertainty about the opposition’s likely behavior in office are important components of the incumbency advantage. Given the uncertainty of post-electoral payoffs, intermediaries expect, or hope, to receive some rewards before the polls. In this respect, gifts distributed prior to electoral contests serve an important function: By channeling resources, a given politician presents himself as a viable, serious and generous candidate. Such payoffs might be necessary to establish oneself as a contender and to signal the strength of one’s campaign. Because of the trade-offs between the two types of rewards for the two sides, actors try to combine both pre- and post-electoral payoffs. In general, these conditions on average favor the incumbent and his party: They can give more credible promises about post-electoral redistribution, as highlighted above, and they can draw on their access to state coffers to give out resources before the election. State coffers might not always be vast but they are not negligible. Even in countries lacking mineral wealth or oil, such as Senegal, incumbents can draw on considerable state resources, and thus they tend to have a financial advantage over the opposition. For example, in her study of the 1996 local elections in Senegal, Patterson (2002) found that the ruling party was able to distribute much more funds on the eve of the election than the opposition. Several opposition leaders whom I interviewed during the 2007 presidential election in Senegal, including a presidential candidate Ousmane Tanor Dieng, decried the vast size of President Wade’s war chest. This is consistent with Vicente and Wantchekon who also state that incumbents have a “pre-election disproportionate control of public resources and allocations” (2009: 294). Admittedly, if all resources were given out prior to elections, perhaps incumbents’ dominance would not be as substantial, depending on the war chest of their opponents. Challengers’ personal wealth or other sources of campaign funds (e.g. from the diaspora) could mitigate to some extent the incumbency advantage. Yet, even wealthy challengers are hampered to some extent by their difficulty to make credible promises of post-electoral rewards, as explained earlier (see also Arriola 2013, Chapter 2).

25

For detailed discussion, see Vicente and Wantchekon (2009: 293–295, 300–301).

Effects on Electoral Patterns

49

The incumbent’s probability of winning can also have an effect on the behavior of electoral intermediaries. Magaloni (2006) showed in the case of Mexico that voters’ support for the ruling party is sensitive to the changing perceptions of incumbents’ invincibility. This is because in addition to any gifts distributed during a campaign, voters and electoral intermediaries expect post-electoral rewards in exchange for their support. When there are doubts about an incumbent’s victory, electoral intermediaries are reluctant to publicly back him. The perception of an incumbent’s invincibility in a given election thus affects the intermediaries’ incentives to flock to his side. These conditions thus on average create a pro-incumbency bias. Importantly, the issue of incumbency advantage should not affect our predictions about the relationship between the use of intermediaries and less-pronounced ethnic politics. As I postulate, intermediaries have a preference for the candidate who can distribute the most resources. In most cases in Africa, it is likely to be the incumbent. However, the argument that intermediaries will favor the highest bidder rather than a coethnic, unless the two overlap, holds regardless of whether that candidate is an incumbent or a challenger. However, incumbents’ perpetual reelection is not inevitable: presidential term limits or acute economic crises, which reduce resources available to incumbents, can create openings for the opposition. In addition to the primacy of material benefits, the absence of significant programmatic or ideological differences is another factor that allows intermediaries to switch their support between different political formations. If voters had strong party attachments, or if one party was, for example, seen as friendlier to peasants than its rivals, intermediaries would have a harder time persuading voters to support various parties. Intermediaries not only have the incentives but also have the means to forge ties with non-coethnic politicians. While an individual voter might only trust a coethnic to deliver, intermediaries are in a structurally better position to drive a hard bargain with any politician. Intermediaries with a substantial base have bargaining chips: They can credibly threaten to work against a given political actor, either by defecting to a different candidate or party, or by discouraging turnout. Once in office, incumbents are also ill-advised to renege on promises made to intermediaries, because the latter can support someone else in future elections. Intermediaries who have a track record of delivering votes to politicians can exact a high price because they seem reliable. They can point to their past results to prove

50

A Theory of Social Ties and Electoral Politics Dominant Mode of Mobilization

Electoral Electoral Patterns Pattern

Mobilization via intermediaries

Cross-ethnic electoral patterns

Mobilization along ethnic identities

Ethnic electoral blocs

figure 2.2 Mobilization Modes and Electoral Patterns

their mobilization prowess.26 It is the reputation of effective vote getters that enables intermediaries to make lucrative deals with politicians. Since intermediaries have both the incentives and the means to work across the spectrum of ethnic identity, this form of mobilization leads to more diverse electoral patterns, to the exclusion of ethnic voting blocs (see Figure 2.2). It is worth stressing that relative state weakness and underdevelopment are important boundary conditions for my argument, as well as for many other works on electoral politics in Africa. These two factors are relevant in several different ways. First, economic underdevelopment is a key, albeit not only, factor enabling clientelism (Stokes 2007a). As incomes rise, the demand for, or reliance on, clientelist handouts decreases, and clientelist payments become prohibitively expensive to continue as a mass mobilization strategy. Economic development does not eliminate clientelism outright, but it makes it less appealing to politicians. Relatedly, economic development also matters because a wealthy developmentalist state could undercut clientelist politics and the reliance on local leaders or ethnic ties by forging links and bringing benefits directly to the people. Recent evidence from Brazil indicates that incumbents can benefit electorally from nonclientelist cash transfer programs that benefit the poor, such as the famous Bolsa Familia (Zucco 2008, 2013). If the vast majority of the electorate could access important safety nets, such as pensions, insurance or health benefits, directly from the state, following clear and transparent criteria,27 they would become much less dependent on social ties. Yet, such programs are expensive and are unlikely to be developed on a large scale in Africa in

26

27

Even in first-time elections intermediaries can usually point to success in other mobilization efforts: for example, some local leaders in French West Africa helped recruit wartime volunteers to serve in the French army. Transparency of access criteria, such as income thresholds, for example, is crucial. In the absence of transparency, brokers continue to mediate voters’ access to benefits. Indeed, many studies show that opaque eligibility criteria for accessing benefits are important drivers of clientelism.

Additional Implications of the Argument

51

the near future. While poverty and insecurity increase voters’ need for clientelist benefits, material resources and state capacity put a limit on how successfully the ruling class could build ties with the electorate. State strength is also important to my argument because it affects politicians’ standing vis-à-vis local leaders. A strong state might not need the assistance of local leaders to implement laws and development projects or help boost local cooperation with government policies. The need for local leaders could be reduced if state capacity, as well as trust in the central state, rises. Bureaucratic capacity also limits what politicians can do to shape the standing of local elites. Politicians bent on undermining or destroying local elites need high capacity to execute their plans. It is not enough to be hostile to traditional authorities to actually undermine their standing on the ground. Similarly, it takes substantial resources to prop up authorities that are weak. State capacity thus limits politicians’ ability to reshape social structure.

additional implications of the argument The main theoretical proposition of this chapter is that local leaders with a demonstrated following provide an attractive mobilization strategy for politicians and an alternative to ethnic appeals. The effects of the variation in social structure on the type of mobilization and the subsequent voting patterns can be tested both at the subnational and cross-national level. Most countries have substantial variation in the strength of local leaders in different regions, but these subnational differences also produce aggregate variation between states. The argument also has more subtle implications. Any variation over time and space in social structure and the role of local leaders should affect electoral outcomes. Apart from differences between various groups, urban and rural areas will also differ in their social structure. Rural local leaders across Africa are viewed as much more influential than their urban counterparts. We would thus expect rural voters to be more influenced in their voting behavior by electoral intermediaries, a proposition which I test in Chapter 5. For the sake of clarity, I presented the difference in social structure as one between weak and strong local leaders, but in many cases this difference will be more gradual and we should expect any incremental variation in the power of local leaders to have empirical implications. Local leaders should have greater electoral influence, the tighter the tie between them and their followers. Thus groups with higher levels of dependence on local leaders should be more likely to follow intermediaries’ lead.

52

A Theory of Social Ties and Electoral Politics

conceptualizing variables Hierarchical Ties Hierarchical ties, the key independent variable in this book, embody relations of dependence between local leaders and their followers. But how do we establish in practice whether there are local leaders who wield substantial influence over their subordinates? Strong local leaders are those individuals who control resources and command authority. In assessing the role of local leaders, I therefore look for indicators that reflect these two dimensions of power. Material Control. The power of local leaders rests on their control of vital resources, which create relations of dependence. One of the most important resources in agrarian societies, in Africa and elsewhere, is land. Writing about Southeast Asia in the 1970s, Scott argued that “reliance on direct control of real property” is a common means of building a clientele (1972: 98). In large parts of Africa, there are no private property laws and land is controlled by traditional authorities who grant “use rights” to cultivate land, or to access pasture (e.g. Herbst 2000: 173). Herbst argues that the role of local elites in the distribution of land is critical to their power by giving them autonomy from the state (2000: 173). In contrast, individual property rights, where they are implemented, reduce the role of traditional chiefs (Lombard 1967: 66).28 Both Herbst (2000) and Boone (2003) use custodianship of land as a proxy for the power of local elites. It is important to consider data on actual land-use patterns, as gleaned from ethnographies, over legal status of land from government documents, because states frequently fail to enforce official land tenure policy at the local level (e.g. Galvan 2004). Land is an important but not the only resource that local authorities can regulate. They may also control trade, labor, water rights and other common pool resources, as well as state resources such as subsidized fertilizer or seeds. Local elites might maintain material influence by giving out loans, distributing food and other important goods and services to people in their community. They are the ones that people turn to in search of material help and who provide vital safety nets. Moral Authority. The second element is local leaders’ moral authority, or command of their dependents’ respect and trust. Admittedly, moral

28

See also Onoma (2009) on the interplay of property rights and the power of local authorities.

Conceptualizing Variables

53

authority is a hard concept to measure. In analyzing my cases, I ask whether local elites have higher social status, based on a traditional or religious position, and whether there is evidence that their subordinates respect their position and listen and defer to them. While there is no single indicator to measure moral authority, fortunately, many ethnographies, or historical accounts, describe the perceptions of local elites’ stature (see a literature review by van Rouveroy van Nieuwaal 1987). Researchers, as well as colonial officials before them, have been attuned to whether local leaders seem widely respected or whether their position is weak or merely symbolic. In addition to secondary historical sources, public opinion data can provide a valuable additional piece of evidence for the contemporary period.29 The Afrobarometer contains several questions that try to measure moral authority and material clout of traditional leaders by asking respondents how much they trust leaders, how important a role do they play in their community and how often do respondents turn to leaders for help. Unfortunately, such public opinion data regarding people’s perceptions of local leaders is not available for the beginning of mass politics. Ethnic Politics How do we measure the extent of ethnic politics? Put simply, we seek to determine whether ethnicity is a good predictor of electoral choices. In his classic book Ethnic Groups in Conflict, Horowitz (1985) explains that ethnic politics are based on two intertwined phenomena: ethnic voting, namely voting for a candidate or party of the same ethnic background, and the existence of ethnic parties, or ethnic candidates, namely those that garner a disproportionate share of their electoral support from their coethnics.30 Ethnic candidates and parties are the opposite of national political actors with broadly representative, multiethnic electorates. To operationalize the concept of ethnic politics, we need to look at electoral data on parties’ or candidates’ vote distribution among the electorate. There are several existing indices that quantify this notion of ethnic

29

30

Recent research by Gottlieb (2014) also shows the potential of employing behavioral games to measure the clout of local authority figures. Her results from Senegal confirm high degree of leaders’ influence over the behavior of their subordinates. I prefer this definition to some of its alternatives, such as Chandra and Metz’s (2002) articulation, which requires overt appeals to ethnicity to classify a party as ethnic. Given that the formation of parties based on ethnic identities is outlawed in many African countries, this definition is too restrictive. Parties can cater to ethnic constituencies without making “overt” appeals.

54

A Theory of Social Ties and Electoral Politics

politics. One of them is Cramer’s V Ethno-linguistic Voting Index (CVELI), developed by Dowd and Driessen (2008), which gauges the association between vote choice and ethnicity. CVELI measures the percentage of vote choice that can be predicted by individuals’ identity, aggregated at a country level, as explained in Chapter 1. Other indices include Cheeseman and Ford’s (2007) ethnic polarization and ethnic diversity of political parties. Elischer (2013) calculates Party Nationalization Scores (PNS), namely the spread of the party’s support across the polity, with higher values indicating that the party’s support can be seen as national. In this book, I combine all available measures to paint the most complete picture of the presence or lack of ethnic politics. I use a range of existing indicators and raw electoral data rather than a single measure to assess the degree of ethnic politics because no single indicator in itself is definitive or offers exhaustive evidence and some of them are available only for a specific point in time. For example, some of the indicators can only consider one dimension of ethnic identity and evaluate the association between that particular identity and vote choice. Yet, empirically, we know that in most heterogeneous societies ethnic identity is multidimensional (Laitin 1986, Posner 2005). A unidimensional index can thus attest to the existence of ethnic parties but not certify its absence, because electoral results could be strongly associated with another dimension of identity. Thus in the case of Senegal, I look at all the socially relevant identities that could be in principle mobilized politically and check whether any of them is associated with vote choice, something that no single index can accomplish. When should we regard politicians as ethnic candidates? I suggest that when three quarters of a candidate’s (or party’s) electorate come from coethnics, we should view such a candidate as an ethnic candidate.31 This benchmark represents a relatively high threshold given that most ethnic groups constitute a minority of the electorate and very few candidates or parties in Africa garner support exclusively among coethnics (Cheeseman and Ford 2007).

31

There is some variation in the literature on what the “appropriate” threshold for an ethnic candidate or party should be. On the high end of the spectrum, Horowitz (1985) suggests 85 percent. Given the diversity and ethnic fragmentation of most African electorates, this threshold is high. Such coding would ignore key variation since the majority of the candidates would get coded as nonethnic. Scarritt (2006) suggests adding additional thresholds and categories: 66.6–85 percent indicates “potentially ethnic parties” and 50–66.6 percent means “multi-ethnic party with a majority ethnic group” (cited in Cheesemand and Ford 2007: 12). I seek to suggest a reasonable, middle-ground threshold.

Conceptualizing Variables

55

Electoral data also shed light on how representative a given candidate’s or party’s electorate is of the general public. Few parties have an electorate identical to the country’s population. The relevant question is not whether the electorate is the same as the population, but how much does it deviate from it and whether it appears to be reasonably representative of the public at large. For example, when a candidate receives twice as many, or half as many, votes from a given group as we would expect based on the group’s size, such fluctuations are indicative of skewed ethnic appeal and lower representativeness. When a candidate’s electorate is broadly similar to the makeup of the country, we can consider such a politician a national candidate. In sum, a case with high level of ethnic politics corresponds to high association between ethnicity and vote choice (high CVELI32), the existence of ethnic candidates and electorates that deviate from the general public and low nationalization scores for parties (PNS).33 A case of low ethnic politics has few or no ethnic candidates, candidates’ electorates are broadly representative of the national electorate, and the association between ethnicity and vote choice is low (see summary in Table 2.1). table 2.1 Summary of Measures of Ethnic Politics High level of ethnic politics High association between ethnicity and vote choice: • CVELI over 30% Existence of ethnic candidates and/or ethnic parties Candidates’/parties’ electorates markedly dissimilar from the general public: • Some groups markedly underor overrepresented in the parties’/candidates’ electorate • Low PNS score (Elischer 2013)

32

33

Low level of ethnic politics Low association between ethnicity and vote choice: • CVELI under 20% Absence of ethnic candidates and/or ethnic parties Candidates’/parties’ electorates similar to the general public: • No or very few groups markedly under- or overrepresented in the parties’/candidates’ electorate • High PNS score (Elischer 2013)

I deem the values as low when they fall below 20 percent, medium for the 20–30 percent range and high for levels above 30 percent. These values correspond to the distribution and range of outcomes depicted in Figure 1.1. Elischer (2013) finds that the average value of PNS in Africa is 0.68. I consider the values above this number as high and those below it as low.

56

A Theory of Social Ties and Electoral Politics

conclusion This chapter explained the importance of social structure, conceived here as the landscape of authority at the local level, for the type of mobilization pursued by politicians and the resulting electoral patterns. Politicians’ electoral choices were constrained by the social environment in which they were operating. Where politicians encountered influential local leaders, such as chiefs or religious dignitaries, they could use them as effective means of garnering electoral support among coethnic and non-coethnic voters. In the absence of such leaders, politicians defaulted to relying on the support of their coethnics. Crucially, social structure differed throughout Africa already on the eve of first mass elections, as a result of precolonial variation and the varied impact of the colonial rule. In the chapters that follow, I provide empirical evidence for this theory. Chapter 3 examines in detail the origin and the state of local leadership throughout Senegal and Benin and documents the much stronger hierarchical ties between local leaders and their followers in Senegal than in Benin.

3 Social Structure and Its Origins

introduction Why do we get ethnic voting blocs in some polities but not in others? Thus far I argued that we can understand this variation by looking at the social organization of different polities. Social structure matters because it provides politicians with different options to mobilize the electorate. Analysts have conventionally assumed that the dominant social ties used for political mobilization are based on common membership in an ethnic group. But a significant component of social organization in Africa also consists of a second type of linkage, namely hierarchical ties, which encompass relations of inequality between local leaders, such as traditional chiefs or religious leaders, and their dependents. Empirically, the existence of hierarchical ties varies. While some societies have strong hierarchical ties, with local leaders who have authority over their followers, such ties are considerably weaker in other settings. As the previous chapter highlighted, this variation is crucial because it presents politicians with different mobilization strategies. When there are strong hierarchical ties, politicians can eschew ethnic appeals because they are able to access voters via intermediaries. When such ties are weak, politicians resort to direct appeals to ethnic groups. This chapter analyzes the variation in social structure in Senegal and Benin, showing how the different subnational units aggregate to create a different landscape of social ties at the national level in these two countries, while Chapter 4 will analyze how this variation affected mobilization strategies used by politicians, resulting in different voting patterns. While both Benin and Senegal have multiple ethnic identities, they 57

58

Social Structure and Its Origins

have very different social organization. At the onset of mass politics in the 1950s, there were strong hierarchical ties with influential local leaders in most parts of Senegal, whereas such ties were considerably weaker in Benin. Careful process tracing shows that the origins of the relevant social hierarchies long predate, and are not a product of, political competition. The chapter is organized as follows: The first section describes the timeline of the development of mass politics in French West Africa. The following section discusses why by the 1950s the authority of local leaders varied across Africa. It then compares social organization in Senegal and Benin at the onset of mass politics, and traces the origin of this variation. The final section concludes and paves the way for Chapter 4, which will show how the varying social organization determined the nature of electoral mobilization.

onset of mass politics in west africa The era of mass politics in French West Africa started in the 1950s. The onset of mass political competition, or the so-called founding elections, is particularly interesting to political scientists. All subsequent contests are to some extent affected by the original cleavages, what Lipset and Rokkan (1967) characterized as “freezing” of the political system. That is not to say that electoral dynamics are immutable, especially in response to institutional changes, but there is a substantial amount of path dependence, once certain electoral patterns are put in place. As the next chapter will show, current electoral results bear a striking resemblance to the founding elections. While Africans in French colonies had been participating in elections to the French National Assembly for many years, as early as 1875 in Senegal,1 those with the right to vote represented a tiny fraction of the inhabitants of the colonies, under 10 percent of the population.2 Up until the 1950s, the African electorate was a very small, mainly urban group, limited by requirements of literacy in French or Arabic. This situation did not change until after World War II. The participation of numerous African troops in the war effort created a particularly strong push for electoral reform and an expansion of the franchise. Proponents of reform

1

2

At the time, voting was limited only to inhabitants of four coastal communes who were French citizens. Between 1914 and 1936 the size of the electorate ranged from 8,000 to just over 20,000, considerably fewer than 10 percent of the colony’s population (Robinson 1960: 308).

Social Organization in Senegal and Benin

59

advanced a moral argument that if Africans have been sacrificing their lives in defense of France, they should at least be given the right to vote. In response to these demands, in 1951 the French parliament had given the vote to three new categories of Africans – pensioners, the heads of families who paid the personal tax and mothers of at least two children (Law of March 21, 1951) (Thompson 1963: 174–175). This law brought about a dramatic expansion of the size of the electorate. In Dahomey, presentday Benin, the number of eligible voters increased six-fold (from 54,208 to 334,435) and represented 22 percent of the total population of the colony (1963). In Senegal, the same reform more than tripled the electorate from 200,000 to 655,000 (Cruise O’Brien 1971: 266). The extension of the suffrage changed dramatically not only the size of the electorate but also its composition, shifting the balance from mainly urban electoral core to a largely illiterate rural one. A few years later, in 1956 the French parliament passed Loi cadre (dubbed in English “The Enabling Act”), which introduced universal suffrage for African voters and it gave all of France’s African territories an embryonic structure of parliamentary government, in the form of territorial assemblies (Thompson 1963: 178). As a result of these reforms, African electoral contests changed within just a few years’ time from a privilege of the urban elite to a popular contest for office. The elections of the 1950s for the first time involved a mass electorate. At the time of these crucial electoral contests, politicians faced decisions about how to access the vastly illiterate, newly enfranchised rural electorate. This was particularly a daunting task, as party organization was uniformly weak and channels of communication, such as media, were severely underdeveloped. Yet, the social structure in which electoral politics were about to take place varied dramatically and it thus presented politicians with different mobilization options.

social organization in senegal and benin Senegal and Benin both have a rich identity repertoire, creating groups of various sizes that could be in principle mobilized by a political entrepreneur, but their social organization differs considerably. While most segments of the Senegalese society have strong hierarchical ties, such ties have always been much weaker in most parts of Benin. At the onset of mass politics, Senegal had strong authority figures that played an important role in their followers’ lives and could be used as electoral intermediaries, whereas Benin did not. These distinct social organizations thus presented politicians with different ways to access the electorate. The following

60

Social Structure and Its Origins

sections will analyze social structure among different groups in Senegal and Benin in greater detail. It will discuss the subnational differences in hierarchical ties and how they aggregate to create different social structure “profiles” at the national level. Looking at the subnational units is essential, if painstaking, to account for the national variation. In the narrative I also trace the origin of social ties. To understand why some places did or didn’t have strong hierarchical ties, I investigate when these structures came into existence and whether they survived until the onset of mass politics. Senegal At the start of mass politics, Senegal had strong hierarchical ties among most segments of the population, with influential authority figures who played an important role in the lives of their dependents. Historically, there were two types of powerful elites: religious clerics from Sufi brotherhoods, known as marabouts, and traditional, caste-based elites. The main ethnic groups, the Wolof, the Haalpulaar (Tukulor and Peul) and the Serer, constituting over three quarters of the population, had such strong leaders, be it religious or traditional (Pélissier 1966: 107, 197, Diop 1981, Diouf 1994: 47, Beck 2008). The only exception are the Diola and Balant groups who make up together approximately 5 percent of the population (see Map 3.1). They have virtually no hierarchical ties above family level: they lack a customary political class, such as noble families (Beck 2008: 161), they do not submit themselves to chiefs’ authority and do not accept castes (Pélissier 1966: 28, 593, 682), and with much lower rates of Islamization, Sufi clerics do not play an important role. The two sets of hierarchies present among the major groups – Sufi orders and caste-based elites– generated powerful local leaders, as the following paragraphs will explain. Religious Elites Sufi Islam, widespread throughout Senegal and large parts of West Africa, generates powerful religious leaders, with strong hierarchical ties to their dependents. The key characteristic of this religious organization is that an individual becomes a member of a Sufi order by declaring oneself a follower and taking a pledge of allegiance to a religious leader, a marabout (Villalón 1995: 64). A disciple thus is connected to the order through his personal religious leader and other leaders above him. Among Senegalese Muslims, virtually everyone has his marabout. The brotherhood hierarchy can be schematically depicted as follows (see Figure 3.1).

Social Organization in Senegal and Benin

61

O c e a n

M AU R I TA N I A Sen

egal Riv e

Saint Louis

r V al l

ey

Louga P

E

U

L -

W

Thies Dakar

Touba

O

L

T U

Groundnut Basin

O

K U

F

L O

Kaolack

A t l a n t i c

MALI

R

Tambacounda SERE

R

GAMBIA

DI

OLA

Casamance Kolda

G U I N E A - B I S S AU GUINEA

0 15 30

60

90

120 Miles

map 3.1 Map of Senegal

The religious leaders, both shaikhs and khalifas, are known by a generic French term marabout. They play several important spiritual, social and material functions in their followers’ lives. In the religious realm, they are the essential conduit of faith and a link with God. They are widely credited with having magical and supernatural powers, and followers believe that the marabouts will help them attain paradise (Villalón 1995: 117). The relationship with a marabout also confers certain social and material benefits. First, being a follower of a marabout places an individual inside an important network, which has consequences for his advancement and well-being. Marabouts can facilitate the acquisition of essential goods and services (Villalón 1995: 124) and they often provide an individual with a form of social insurance (Coulon 1981: 115). Disciples have a sense that should they be afflicted by misfortune, such as illness or drought, they can turn to their marabout for help. This help often extends to finding employment, or even a spouse

62

Social Structure and Its Origins

Khalifa General (head of a brotherhood)

Khalifas of different lineages

Shaikhs (religious leaders with disciples)

Talibés (disciples of a shaikh)

figure 3.1 The Sufi Hierarchy

(Coulon 1981: 112). Lastly, marabouts redistribute some of their wealth among their followers in more regular and mundane ways, such as providing meals to followers who come to ask for assistance (Villalón 1995: 191). In particular, during religious holidays, significant amounts are spent on feasts in which, inevitably, a whole base will participate. In brief, the marabouts “serve as conduits for redistribution,” which allows them to maintain their following and position (Villalón 1995: 187). In many cases, the marabouts also control scarce resources such as land, especially in the so-called Groundnut Basin of central Senegal. It is an area on the fringe of the Ferlo desert, outside the traditional zones of agricultural production. The predominantly Mouride marabouts attracted settlers into these areas and they became the pioneers of cultivation of these “new lands” (terres neuves). In these areas the marabouts controlled access to land and cultivation. They acted as custodians of land, the chefs de terre (Coulon 1981: 112). This role was particularly important in the institutional context of lack of land titles,3 where peasants were both formally landless and without a guaranteed right to use 3

The prevailing land tenure throughout Senegal was based on allocation of land by traditional authorities.

Social Organization in Senegal and Benin

63

land. Only being linked to a marabout gave one the right to cultivation (Cross 1968 cited in Coulon 1981: 113). The marabouts developed a network of collective farms (daaras), worked by young men in the service of a marabout (Villalón 1995: 118–119). The young men would provide unremunerated work in a daara for several years. When they were released from servitude they were granted a field with no title for their personal cultivation, and sometimes even a wife. Thus gaining and retaining land was dependent on one’s ties with a marabout. In fact, the entire organization of agricultural production was built around the figure of the marabout. The marabouts managed to build and retain substantial authority among the masses as they played a significant role in the lives of their followers. They procured many valuable services to their disciples, both material and religious: They provided “a guarantee of economic security through material aid” and by virtue of Sufi beliefs, they assured “psychological gratification through the promise of paradise in the hereafter” (Cruise O’Brien 1971: 9–10). One cannot stress enough the significance of personal, and not just nominal, ties with a marabout and not mere affiliation with a religious order. As Villalón claims, “[E]ffective exploitation of the advantages of clientelism can only come by developing personal ties to the marabouts of that order” (1995: 125). He further elaborates that adherence to an order per se has limited sociopolitical consequences. Instead, it is “the nature and extent of personal relationship with a marabout which has a direct impact on an individual’s opportunities, and his or her behavior” (1995: 145). Traditional Elites: Caste-Based Aristocracy Apart from hierarchical linkages with marabouts, most of the Senegalese ethnic groups (including the Wolof, Serer and Tukulor), with the exception of the Diola, have another form of hierarchical ties: a caste system, with a group of royals, aristocracy or other hereditary princely lineages at the top of the hierarchy (see Table 3.1). This stratification engenders clear ideas about superior and subordinate status of one group versus the other and caste confers significant prestige irrespective of wealth (Tamari 1991: 225). This differentiation has important consequences for the distribution of power in society. Authority is inherent in the higher social categories, and the structures of hierarchy are accompanied by suppositions about the dependent status of individuals in the lower positions (Diop 1981, Villalón 1995: 58–59). It is the aristocratic groups which typically control access to valuable

Social Structure and Its Origins

64

table 3.1 Caste Stratification Main category Freemen

Caste people Slaves

Subgroups Royals, Notables, Commoners Blacksmiths, bard, leatherworkers, artisan Serfs

Note: The specific subcategories vary between ethnic groups.

resources, such as land for cultivation or pasture. This is the case to this day among the Tukulor elites, the toorobe,4 in the Senegal River Valley and among some Serer in Siin. In the Senegal River Valley, the toorobe, stemming from Tukulor castebased aristocratic lineages, acted as customary landowners regulating agricultural production, by providing access to fertile land in exchange for rents. This “feudal” relationship persisted throughout the colonial period, and at the onset of mass politics the Tukulor masses were still locked in a relation of dependence with the toorobe nobles (Beck 2008: 125). The authority of the toorobe and their influence over lower Tukulor strata were based on the convention of higher status and reinforced by control of resources. Such relations of dependence gave the toorobe elite significant influence over their subordinates. The Serer have an analogous system of caste stratification, with aristocrats at the top of the social hierarchy. Galvan (2004) describes in his study of the Serer of Siin, one of the two main precolonial Serer kingdoms, that the aristocrats not only enjoyed higher status but they also controlled access to land. The old rural aristocrats, the lamans, were the customary land holders. They had “holistic custodianship of land – allocating land, coordinating field rotation, maintaining fallow and pasture” (2004: 133). They also served as arbiters in land-related disputes as well as social conflicts. In the context of lack of codified land tenure, the lamans acted as “owners” of land, which belonged to a lineage. They did not cultivate all of the land under their control, but allocated it to their subordinates. In this system, the lamans had the “title” to land and they were merely granting “use rights” to the Serer peasants, commoners and low-caste 4

The singular form of toorobe is toorodo.

Social Organization in Senegal and Benin

65

people, in exchange for regular payments (Galvan 2004: 106–107). Thus a Serer peasant’s livelihood depends on his superiors, the Serer aristocrats. To reiterate, the three main groups in Senegal, which comprised over three quarters of the population, all had powerful local leaders, be it religious leaders or caste-based aristocrats, who exercised considerable influence over their followers. Origin of Social Structure It is essential to describe briefly the origin of the hierarchical bonds discussed thus far since the reader might wonder whether they are exogenous to electoral politics. The hierarchical structures described earlier clearly predate, and are not the product of, mass political competition. Sufi Islam arrived in West Africa in late eighteenth century, spreading from North Africa and the Middle East (Cruise O’Brien 1971: 25). Caste structures appeared among the Malinké no later than 1300 and independently among the Soninké and the Wolof no later than 1500, eventually spreading to 15 West African ethnic groups, inhabiting at least 14 states (Tamari 1991: 221). Caste structures among all these ethnic groups can be traced to these three sources of origin and their current distribution can be explained by the process of migration (Tamari 1991: 249). While the original hierarchical structures precede modern political competition, not all of them survived until the onset of mass politics. Two events in particular, the spread of Islam and French colonization, transformed the social structure, by destroying some of the hierarchical orders. The spread of Islam created a potential power struggle between the two sets of elites: traditional and religious. Where traditional elites posed an obstacle to the ambitions of the religious clerics, a conflict was likely to emerge between the two, leaving one set of elite weakened. In other instances, traditional elites embraced Islam and prospered. The second shock to the landscape of authority came from colonization. In some instances, the French colonizers faced elites that posed fierce resistance and thus had to be eliminated. In other cases, the local elites were more compliant and the French co-opted them. These responses varied not only between colonies but also within them, creating different constellations of ties in different regions. Tracing Hierarchical Ties at the Subnational Level in Senegal The clashes between the religious and traditional elites, as well as their interactions with the French, account for the regional differences in the strength of hierarchical ties. While in every part of Senegal outside

66

Social Structure and Its Origins

Casamance we find strong hierarchies and influential authority figures, the balance of power between the two sets of elites, namely religious and traditional, varies between different regions. As the following paragraphs will retrace, in the central parts of the country, inhabited mainly by the Wolof and Serer, religious leaders have eclipsed the traditional elites. On the other hand, in Senegal River Valley in northern Senegal, castebased elites are more influential than religious clerics. Central Senegal: The Primacy of Marabouts In the Senegalese heartland, the once powerful traditional Wolof elites were severely weakened, first by warring Muslim leaders, then by the French. While in the nineteenth century the Wolof masses converted, or were converting at a fairly rapid pace, to Islam, their traditional leaders did not. The Wolof leaders were seen as heretics and an obstacle to the spread of Islam and were consequently fought by warrior marabouts. The marabouts’ effect was very destructive as they managed to weaken the traditional Wolof social order (Cruise O’Brien 1971: 30). The second front in the attack against the Wolof elites came from the French colonizers. For three decades since the 1850s, the French have been waging a military campaign against the Wolof leaders who resisted colonial annexation (Cruise O’Brien 1971: 13). Their final military victory in 1886 was an important step in the destruction of the Wolof social order. Subsequently, the power of Wolof chiefs was severely limited. The Wolof chiefs lost their ability to allocate land (decree of 1906), and they lost their power of arbitrage and punishment (decrees of 1903, 1912 and 1924) (Cruise O’Brien 2002a: 19). Following the colonial conquest, the French set out to reorganize the traditional order to their liking. Since 1886 nominations of Senegalese chiefs in the conquered Wolof territory depended exclusively on the French authorities (Cruise O’Brien 2002a: 20). As Cruise O’Brien notes, “the chiefs were now to be chosen by the colonial rulers in consideration of their submissiveness and their literacy in French, rather than their traditional qualifications” and consequently they were “reduced to a subordinate status” (1971: 14). Beck offers a similar view as she argues that “the colonial state has erased, or at least limited, political responsibility of chiefs, first by weakening the authority of the indigenous chiefs, then by intervening in succession” (2002: 532). The chiefs’ new role was to serve French interests rather than those of their former followers. They were given unpopular tasks, such as tax collection and recruitment, while at the same time they were deprived of their traditional role (Klein

Social Organization in Senegal and Benin

67

1968: 202). The chiefs quickly became disliked and discredited in their new role and their authority was weakened by the colonial rule. At the same time, the religious authorities from the Sufi brotherhoods directly benefited from the weakening of the traditional Wolof elites. While they were initially a partial cause of the destruction of the traditional Wolof order, they soon became a response to it. Whereas the spread of Islam among the Wolof masses was the impetus for the clash between traditional and religious elites, the defeat of the former only accelerated the consolidation of Sufi Islam among the Wolof. Villalón argues that the collapse of the social organization following the colonial conquest gave impetus to widespread conversions among the Wolof as Islam filled the “imperatives of reconstructing social orders” (1995: 61). The French created a void of authority that the religious leaders filled. By the late 1800s, the marabouts replaced the chiefs as the predominant rural authority in central Senegal (Creevey 2006). As Klein observes, “without intending to do so, the French aided the replacement of the chief by the marabout” (1968: 202). The further expansion of the Sufi brotherhoods, especially the home-grown Mouride tariqa,5 can be attributed largely to the fact that “it provided a means of reconstructing the old social order on a new religious basis” (Cruise O’Brien 1971: 15). As Cruise O’Brien further notes, the religious authorities “took over the status and many of the functions of the discredited chiefs” (1971). This period marked the end of the chefferie Wolof and a switch to marabout control (Diouf 1992: 240). Levtzion and Pouwels note that the marabouts replaced the defeated Wolof aristocrats as intermediaries with the colonial state (2000: 13). The marabouts of central Senegal, especially the Mourides, prospered under the colonial rule, albeit after some early clashes (Robinson 1999). They built substantial authority among the masses by playing a significant role in the lives of their followers and, by cooperating with the French, avoided their own destruction. The key to the marabouts’ success and survival was their ability to find an accommodation with the French, while retaining a high degree of independence. As Robinson describes, this “accommodation permitted the marabouts and brotherhoods to develop considerable autonomy in the religious, economic and social spheres while surrendering the political and administrative domain to the French” (1999: 193). Substantial autonomy from the French was essential for 5

Whereas the Qadriyya and Tijaniyya orders spread to Senegal from North Africa and the Middle East, the Mouride brotherhood originated in Senegal in the 1880s in direct response to the void of authority left by the French colonial conquest (Cruise O’Brien 1971).

68

Social Structure and Its Origins

retaining credibility among the marabouts’ followers. As Diop and Diouf argue, “[T]he autonomy of marabouts vis-à-vis the colonial administration made them look as if they were working only for the peasant’s benefits.” The marabouts softened the most coercive aspects of colonial domination, for example by helping peasants to defer payments of taxes. On the other hand, they helped ensure substantial compliance with the French colonial edicts, while retaining control over social, juridical and land disputes at the village level. They managed with great “efficiency” the interaction between peasants and the colonial state (Cruise O’Brien 2002b: 86). As a 1904 French report recognized, the marabouts helped solve significant problems of agricultural production, labor supply and social control, and substantially increased the yields of the main crop, the groundnut (Robinson 1999: 207).6 The marabouts succeeded in a delicate balancing act: They convinced both the colonial administration and the peasants that they were acting on their behalf. The contrast between the trajectory of the Wolof chiefs and the marabouts of central Senegal is stark. Senegalese historian, Mamadou Diouf, notes that while the French authorities left Wolof chiefs no possibilities to legitimize their authority, the opposite was true of the marabouts (1992: 24). Whereas the French weakened the power of the traditional kings and nobles, they recognized the marabouts’ utility and reached an accommodation with them (Behrman 1970, Creevey 2006). Cruise O’Brien (2002a) similarly argues that the marabouts managed to avoid the fate of the chiefs because unlike the chiefs, the marabouts were not nominated by the French but retained control over their order. By the time of World War II, the marabouts had a growing popular support, whereas the chiefs “lost prestige and popular support” and “were more hated than ever” (Cruise O’Brien 2002a: 26–27). Diouf concludes that by the onset of mass politics the marabouts were the only leaders with access to the population of central Senegal (1992: 241). Northern Senegal: The Primacy of Traditional Elites In contrast to central Senegal, the balance of power between traditional and religious elites played out differently in the Senegal River Valley in the North. The traditional leaders of the Takrur Empire converted to Islam already in the eleventh century and coexisted with religious clerics. When a challenge to their rule came in the mid-eighteenth century from an 6

The marabouts also helped increase tax payments and recruitment for the colonial army (Robinson 1999: 210).

Social Organization in Senegal and Benin

69

Islamic reformist movement, known as the Toorodo revolution, the traditional elite was able to survive, albeit in a changed form (Beck 2008: 125). After suffering some military challenges from Moor slave raiders, the revolutionary religious elites were forced to incorporate into their ranks members of the old traditional elites, who were keen to preserve their status and power. The new toorodo ruling class thus embodied a fusion of religious and traditional elites. After the previously centralized state was divided into provinces, power and social status was further concentrated among newly emerged ruling dynasties, known as the grandes familles.7 Despite the initial fusion of religious and traditional authority, over the course of the nineteenth century some separation of the two functions emerged, with some dynasties controlling religious positions, while others monopolized positions of village chiefs. The clash with French colonizers in the Fouta Toro also produced a different outcome than in the Groundnut Basin. When the French colonizers set out to establish direct control in the 1850s, after years of commercial presence, the main resistance encountered was from Tijani marabouts led by Oumar Tall, rather than the traditional elite. While French soldiers fought and defeated Tall and his followers, they embraced the toorobe families who were willing to cooperate with them. Hence they preserved the existing caste-based social hierarchies. As Beck explains: [Under French rule] the “toorobe families who were co-operative allies maintained or even enhanced their power in Tukulor society. Although collaboration with the French may have tarnished their reputations, it did not adversely affect their social status. Those associated with resistance to colonial rule, on the other hand, found themselves excluded from access to colonial economic and political resources, which ultimately undermined both their power and status” (2008: 127).

Tracing the clashes between religious and traditional elites, as well as their different responses (resistance versus accommodation) to the French incursions, helps account for the varying dominance of one set of elite or the relative balance of power between them (see Table 3.2). Regardless of these regional differences described above, Senegal as a whole, with the exception of Lower Casamance, has significant social hierarchies among the vast majority of its population, stemming from Sufi orders, caste and aristocratic lineages. Within these structures, there are natural local leaders whose power rests on status and control or provision 7

As Beck explains, the centralized state collapsed in 1806. The Fouta Toro was divided into provinces (domains) and power was devolved to the new ruling dynasties who were to run them (2008: 125).

Social Structure and Its Origins

70

table 3.2 Balance of Power between Traditional and Religious Authorities in Senegal

Central Senegal Northern Senegal

Traditional elites

Religious elites

Rebel & lose Adapt & win

Adapt & win Rebel & lose

table 3.3 Local Leaders in Senegal

Ethnic group

Share of population

Wolof

44%

Tukulor/Peul

23%

Serer

15%

Diola and Balant

6%

Strong local leaders? Traditional elites

Religious elites

Yes – the marabouts No. Originally yes, but won the power weakened by the time of struggle with the onset of mass politics – Wolof chiefs and Wolof chiefs have been they weathered the undermined by warrior colonial conquest; marabouts and the they control vast French colonial areas of land, serve conquest as social mediators Yes, but weaker than Yes, there are powerful the traditional elite traditional elites, especially the Toorobe in the Senegal River Valley control land and enjoy traditional legitimacy Yes, traditionally stratified Yes, have Sufi orders, but not as heavily by caste, some traditional implanted as among authorities have the Wolof customary land tenure No traditional authority, No, low levels of Islamization, very anomic society, no caste weak implantation stratification of Sufi brotherhoods

Sources: Beck (2008), Boone (2003), Coulon (1981), Cruise O’Brien (1971), Diop (1981), Galvan (2004), Murdoch (1981), Pélissier (1966), Tamari (1991) and Villalón (1995).

of material resources, such as land and welfare safety nets. Senegal as a whole has strong hierarchical ties, since in most parts of the country there are influential leaders, be it traditional or religious (see Table 3.3 for summary).

Social Organization in Senegal and Benin

71

Benin In comparison to Senegal, at the onset of mass politics, Benin had few influential local authority figures. Traditionally, Benin had important social hierarchies in the form of royalty. The South of the territory was dominated by two precolonial kingdoms: Abomey (Danhomé) and Porto-Novo (Adjacé), whereas in the North there were several smaller Bariba kingdoms of Nikki, Kounandé and Djougou (Glélé 1969: 19, see Map 3.2). However, in southern Benin, where two-thirds of the population lived, the traditional royalty and chiefs were severely undermined by the French colonial conquest. The most important traditional elite, the Fon royalty, fiercely resisted colonial annexation, just like the Wolof rulers. Consequently, as they often did whenever they encountered resistance, French administrators replaced the legitimate traditional royalty with more compliant individuals of their own choosing. When the territory of Benin was placed under direct rule, the French “revolutionized the organization of the ancient kingdoms” and produced new “customary authorities” (Banégas 2003: 320). Yet, by appointing their stooges to traditional positions, they failed to provide the would-be authority figures with legitimacy, undermining hierarchical structures. However, unlike in central Senegal where the Sufi brotherhoods presented alternative forms of social hierarchies, such structures were absent in southern Benin, which is predominantly Christian and animist. Wherever the royalty posed a direct obstacle and threat to the French colonial expansion, they were fought and undermined by the colonizer. In the case of the Abomey kingdom, the French deposed king Gbehanzin in 1900, after he refused to surrender to Colonel Dodds. He was then sent into exile in Gabon, officially under the pretext of being punished for making human sacrifices (Glélé 1969: 21). Glélé notes that the French administration wanted to “get rid of a king who knew how to use all his ancestral, legitimate rights and who became inconvenient for the French presence” (1969: 21). At the same time, French colonizers installed kings who would be more pliable. In 1894 they made Agoli-Agbo king, who reorganized the monarchy “according to the wishes of the occupant” (Banégas 2003: 323). Yet, even this officially sanctioned king did not hold his position for very long. He was eventually deposed and the monarchy abolished. In other instances, the French undermined the traditional system by preventing succession. After the death of king Toffa of Porto-Novo, the French reduced the role of his son and would-be successor, Prince Tolli Toffa, to “superior chief” (Glélé 1969: 21). As a result of French actions,

Social Structure and Its Origins

72

0

15

30

60

90

120 Miles

NIGER

B U R K I N O FA S O

B

A

R

IB

A

Natitingou

Parakou

NIGERIA

TO G O

YO R

U

B A

Abomey

F O N A ADJ

GHANA

Lokossa

Porto-Novo

Ouidah Cotonou

map 3.2 Map of Benin

the ancient monarchs were “marginalized [and] reduced to an honorary role” (Banégas 2003: 320). The remaking of hierarchical ties to suit French interests extended beyond kings, to all types of customary chiefs. The colonial power turned

Social Organization in Senegal and Benin

73

customary chiefs into colonial administrative aides, chefs de cantons. This policy required changing the role of legitimate chiefs, replacing “uncooperative” ones, and at times even “inventing” new chiefs where necessary. One crucial aspect of this policy was that the French colonizers were not concerned at the time with issues of legitimacy of those they placed in power. Crowder and Ikime (1970) note that the French colonial officials rarely hesitated to replace an inefficient, illiterate, legitimate chief by an old soldier or retired clerk, who they felt would understand what was required of him. As a result, “a whole group of people were made chiefs who in precolonial times would have had no right whatsoever to such posts” (Crowder and Ikime 1970). Yet, this blanket policy of replacing anti-French traditional leaders with those appointed by the colonial power had an unintended consequence; it led directly to the weakening of the hierarchical ties between authority figures and their dependents. These newly created hierarchies were seen as illegitimate and consequently the new chiefs had little power over their subjects. Ronen notes that the imposed chiefs were viewed as artificial and abusive (1974: 59–61). Many of them were appointed, and therefore they disrupted tradition. They did not represent people as the customary chiefs had in the past and overall the masses were against them. Ronen (1974) sums up the situation in southern Benin as crise de chefferies, where the masses were in opposition to the chiefs.8 Banégas describes the new hierarchy of chiefs, the chefferie de canton, as “invented,” while the traditional authorities saw their influence decline (2003: 326). Similarly, Thompson observes a “widespread rejection of authority in the South” (1963: 173). The French colonizers did not change their understanding of the role of chiefs until the second decade of the twentieth century. As Alexandre notes, the era of “ethnologist administrators” began with French Governor General Joost van Vollenhoven, in office between 1917 and 1918 (1970a: 4). There were important changes in French official thinking at the time, as the colonizers became aware of the issue of legitimacy. Colonial officials finally understood that chiefs needed to be respected by the people in order to command their authority. As Governor General van Vollenhoven admits, “[W]e are forced to conclude that we need a chief who is willingly accepted and wanted by the people” (quoted in Alexandre 1970b: 64). Yet, in southern Benin, where the French reshaped traditional authority 20 years earlier, the new colonial thinking 8

See also Ronen (1975).

74

Social Structure and Its Origins

came too late to undo the damage. In 1917 Governor General van Vollenhoven observed that the direct administration “led to a disaster, by creating the loss of authority of the chiefs traditionally listened to” (quoted in Banégas 2003: 321). Another Governor General, Joseph Jules Brévié, who served in this function between 1930 and 1936, subsequently noted the ineffectiveness of the imposed chiefs. As he observed, “[C]hiefs coming from the outside – even if they belong to the same race as the people under their authority – are not easily accepted or tolerated by them” (Brévié 1932, quoted in Alexandre 1970b: 67).9 Realizing the weakness of the new system, the French let it atrophy: whenever an old chief died he was not replaced (Banégas 2003: 326). As Glélé notes, around the time of independence the chefs de cantons had no role in territorial administration. At [the time of] their death none of them were replaced. Besides, everybody was liberated from their tutelage, the yoke of chefs de cantons who were nothing but a relic, the kinglets without treasury, without subjects, having nothing but moral authority and the memories of their former glory. (1969: 247, quoted in Banégas 2003: 326).

The remaining chiefs did not play an important role in people’s lives because they lacked two important sources of power of a local leader: legitimacy derived from a traditional role and control of resources. Hierarchical structures in southern Benin had been undermined and the chiefs weakened long before the onset of mass politics. In contrast to the more populous South, patterns of social organization are entirely different in the northern part of the country. In the northwestern Atacora mountains, the mainly Somba population was dispersed, living in extended family compounds rather than multifamily villages (Thompson 1963: 201). Hence there was very little social stratification and lack of obvious hierarchical ties. In the northeast, where there were a few small precolonial Bariba kingdoms, hierarchy was more pronounced. Jacques Lombard, one of the foremost ethnographers of the North, describes the Bariba as “strictly hierarchical” and he sums up their social organization as a “system of domination imposed by the [Bariba] nobility” (1965: 37). Members of various groups differentiated themselves on the basis of their social status and they occupied a different place in the social hierarchy (Lombard 1965: 34). At the same time, the French had a less disruptive effect on social organization in the North.

9

Alexandre (1970b) provides multiple examples of French failure to force chiefs onto people throughout French West Africa.

Endurance of Social Structure

75

table 3.4 Local Leaders in Benin Group

Share of population

Strong local leaders?

Fon (South)

39%

Adja (S)

15%

Yoruba (S)

12%

No, originally strong royalty, but destroyed during the French colonial occupation No, traditional authority severely undermined during the French colonial occupation No, traditional authority severely undermined during the French colonial occupation Moderately strong leaders from the Bariba nobility; traditional authority in the North was less disrupted by the French (due to much later and sparser presence)

Bariba (North)

9%

Sources: Banégas (2003), Glélé (1969), Lombard (1965), Ronen (1974) and Thompson (1963).

They didn’t try to establish effective control over the northern territory until almost the turn of the twentieth century, much later than in the South. By that time, the colonial boundaries were firmly in place and the French did not face threats from the traditional Bariba authorities. As they didn’t have to fight local elites, the French were able to graft their institutions onto traditional structures, turning traditional princes into chefs de cantons, instead of replacing them. As Glélé observes, in the eyes of the population, “the chefs de cantons of the regions of north and center, incarnated royal and traditional authority” and were thus more legitimate (1969: 22). However, these few remaining hierarchical structures in parts of the North affected less than 10 percent of the population. Despite a few powerful local leaders in the northeast, the social organization in Benin as a whole in the 1950s was characterized by few and weak hierarchical ties (see Table 3.4).

endurance of social structure Thus far, we have considered the social structure that politicians inherited in the 1950s. In the case of Benin, where traditional authority was severely undermined already by the early twentieth century, the weakness of

76

Social Structure and Its Origins

contemporary traditional leaders will come as no surprise. Two decades of Mathieu Kérékou’s Marxist dictatorship, with its typically Marxist hostility to traditional rulers, further destroyed any remaining power of traditional chiefs or chefs de village.10 As Banégas notes: It seems that the cadaver of traditional authorities still moved too much for the military regime which took power in 1972. . . .The government engaged itself in a violent policy of eradication of “obscurantist forces.” A violent repression fell upon customary chiefs, royal dynasties, and above all on representatives of traditional religions, all grouped under the rubric of retrograde antirevolutionary forces. (Banégas 2003: 326).

In 1974 the various royalties were abolished, even though they only played a ceremonial function. Chiefs were replaced by new local structures, in the form of Local Revolutionary Committees (Comités locaux révolutionnaires). The regime created “anti-feudal” committees at the local level, which cut down sacred trees, closed convents and tried to control all types of ceremonies, including weddings, funerals and baptisms (Banégas 2003: 327). In 1976 anti-feudal law was passed, which abolished royal processions and forbid enthronization of new kings and chiefs. What might be more surprising is how well the traditional and religious elites in Senegal preserved their power for over half a century. Indeed, as Cruise O’Brien notes, the symbiotic arrangements between Senegalese national politicians and local leaders, religious dignitaries in particular, “have been constantly negotiated and re-negotiated in the post-colonial period” (2002: 11). Three complimentary explanations of elites’ resilience are worth highlighting. First, local elites continue to be useful for local populations and in many instances they remain the sole providers of essential services. The Senegalese postcolonial state, as many other African states, did not bring much development to rural areas and it failed to provide many essential safety nets to its citizens. As long as people are lacking the most basic services and local elites are in a position to satisfy some of these needs, the elites will continue to be perceived as important by their communities. Second, politicians’ demand for intermediary services has remained high throughout the post-independence period. As many observers note, the postcolonial state suffered from a serious legitimacy deficit.

10

Author interview with Dr. Djibril Debourou, a historian and MP, Cotonou, April 3, 2007. See also Iroko (2003) and Bako-Arifari and LeMeur (2003) on this topic.

Contemporary Strength of Local Leaders

77

The shortcomings of the postcolonial state include not only its results, namely its failure to foster development, but also its form. Many citizens view the state and its formal institutions as alien. As Diop and Diouf highlight, “state institutions in rural areas are seen as formal and empty” (2002a: 37). Moreover, officials encountered the population’s mistrust of all types of administration (Diop and Diouf 2002a: 42). To the extent that the state and its national political elites lack acceptance or popularity, they continuously need local elites to help legitimize their rule. Not coincidentally, the Senegalese frequently call electoral support gained through intermediaries as vote légitimiste. Since local leaders were very useful for the governing elite, national politicians had few incentives to undermine the position of those who were helping them stay in power. Both religious and traditional elites have supported and cooperated with the ruling party consistently over the years. The symbiotic relationship between those in power and Senegalese local leaders became known as the “Senegalese social contract,” following Cruise O’Brien’s widely used term (1992). Other scholars talk about symbiosis, reciprocity, interdependence or even “mutually reinforcing coexistence” between the state and local leaders (Diop and Diouf 2002a). Finally, the weakness of the Senegalese state meant that any attempts to wrest power from local leaders would be difficult to execute in practice. Several state interventions, which tried to reduce the prerogatives of local elites, proved strikingly unsuccessful. For example, the National Domain Law of 1974, which in theory placed most land under the control of the state, in practice failed to curtail local elites’ power to allocate land (Galvan 2004). Several other institutional reforms failed due to lack of support from local leaders (Diop and Diouf 2002a: 38–39). Senegalese national elites also tried to establish direct ties with the population but encountered resistance from local authorities and ultimately failed in their undertaking (Diouf 1992: 250). Thus the relative weakness of the state combined with politicians’ continued dependence on local elites created a propitious environment for local leaders to solidify their position.

contemporary strength of local leaders The different strength of traditional and religious leaders in Senegal and Benin described at the time of the onset of mass politics has been fairly persistent and it is evident in contemporary survey research. To this day,

78

Social Structure and Its Origins table 3.5 Reliance on Traditional and Religious Leaders

Religious leader Traditional leader

Senegal

Benin

22.9% 15.5%

5.0% 5.3%

Note: The values denote the percentage of men saying that during the course of last year they often turned for help to the following type of leader. Source: Question 32, Senegal and Benin, Afrobarometer, Round 3 (2005).

local leaders play a greater role in Senegal, in comparison to Benin. Rounds 3 and 4 of the Afrobarometer survey, conducted in 2005 and 2008, respectively, contain three questions that serve as good proxies for the power of local leaders. One question asks respondents to state how often during the course of last year did they turn for help, or to give their point of view, to different types of leaders, including MPs, government councilors, political party officials and religious and traditional rulers.11 This question is a good indicator because hierarchical ties and the power of local leaders are based on relationships of dependence that encompass all aspects of a person’s life – material, social and political. Local leaders can exert influence over their dependents if they affect their well-being and have a real impact on their followers’ lives. The Afrobarometer question captures well the relative level of such dependence. The above Table 3.5 contrasts the role of religious and traditional leaders in Senegal and Benin. The cells show the percentage of adult men stating that they often turn for help to a given leader. I choose to present the results for men, because in this cultural context it would be typical for men, rather than women, to ask for help on behalf of their family.12 The different importance of religious and traditional leaders in these countries is quite telling. While the Beninese rarely turn for help to their local leaders, the Senegalese are three to four times more likely to do so. This is a very substantial difference and it captures the varying strength 11 12

Question 32, Round 3, 2005. When men and women are grouped together the direction of effects remains the same; the percentages are just lower across the board, undervaluing the strength of leaders. An even more “culturally” appropriate measure would be to look specifically at heads of families. While this is possible, it would reduce dramatically the number of respondents, especially given the large size of an average family.

Conclusion

79

of hierarchical ties in these countries. It is worth noting that it is not the case that the role played by religious and traditional elites in Senegal is played by different actors in Benin. The Afrobarometer survey asked about reliance on six different types of leaders. In Benin none of them approached even remotely the role played by traditional and religious leaders in Senegal, as the highest level of reliance did not surpass 5.5 percent (see Appendix B). Two other survey questions paint a similar picture. Question 49i (in Round 4) asks respondents how much they trust traditional leaders. Again, the difference between Senegal and Benin is substantial. While 66 percent of Senegalese say that they trust traditional leaders a lot, this percentage is much lower in Benin, standing at 33 percent.13 In Senegal, traditional leaders enjoy much higher levels of trust than the political actors, including the president (30 percent), the National Assembly (21 percent) or the ruling party (20 percent).14 In contrast, in Benin, traditional leaders are less trusted than the president (33 percent versus 45 percent). Finally, the Afrobarometer survey asks respondents about traditional leaders’ influence in governing their community.15 As in the other two questions, traditional leaders in Benin appear less influential. While 21 percent of Senegalese say that traditional leaders influence their community “a great deal,” at 12 percent, this number is lower in Benin.

conclusion The goal of this chapter was to establish that at the onset of mass politics in the 1950s social structure differed considerably in Senegal and Benin. Specifically, there was a striking variation in the role of local leaders among various groups in these two countries. While the groups comprising the majority of the Senegalese population (Wolof, Tukulor/Peul and Serer) had strong hierarchical ties between leaders, either traditional or

13

14

Answers to this question also confirm the subnational variation between different groups. In Senegal, the hierarchical Wolof and Halpulaar trust leaders more than the Diola, the single group that had no traditional or religious hierarchy. Trust levels are 69 percent and 66 percent among the Wolof and the Halpulaar, respectively, and 51 percent among the Diola. Likewise, in Benin, the main southern groups, Fon and Adja, whose traditional leadership was destroyed, have much lower trust levels than the Bariba, the only group with traditional leaders undisturbed by the French. The rates are 24 percent and 17 percent for the two southern groups, respectively, and 60 percent for the northern Bariba. 15 Questions 49a, b and e, Round 4, 2008. Question 65, Round 4, 2008.

80

Social Structure and Its Origins

religious, and their followers, that was not the case in Benin. Traditional leaders in the South of Benin, where the majority of the population lived, were all but destroyed by the colonial occupation. While there remained some fairly influential traditional leaders in the North, they controlled only a small segment of the national population. This observed difference in the strength of hierarchical ties between local leaders and their dependents has important political implications, as it presented politicians with different mobilization options. As the following chapter will show, in Benin, where hierarchical ties were limited, politicians continually relied on shared ethnic identities to win electoral support. They did so in the first phase of political competition as well as in recent elections after the reintroduction of competitive politics. In contrast, in Senegal, where there were powerful local leaders, they were used by politicians as electoral intermediaries. Chapter 4 documents these two distinct mobilization styles and shows how these strategies result in different electoral patterns, with the presence or absence of ethnic voting blocs.

4 Mobilization Strategies and Electoral Outcomes in Senegal and Benin

introduction This chapter analyzes the electoral consequences of social structure in which politics takes place. First, it shows that in Benin, where hierarchical ties were limited, politicians relied on shared ethnic identities to win electoral support. They did so in the first phase of political competition as well as in recent elections after the reintroduction of competitive politics. In contrast, in Senegal, where hierarchical ties were strong at the onset of mass political competition, politicians mobilized voters via intermediaries, instead of appealing to ethnic identities. The choice of the dominant mobilization strategy, in turn, has consequences for electoral outcomes. If politicians rely on shared identity to win support, ethnic voting patterns will emerge. Not surprisingly, Beninese electoral politics were and are dominated by ethnic voting blocs. In contrast, when politicians rely on intermediaries they are more likely to be able to win voters across the ethnic spectrum since intermediaries have an incentive to work for politicians most able to deliver resources, rather than their coethnics. As the Senegalese case illustrates, intermediaries negotiated to deliver their blocs of voters to politicians from different backgrounds, often switching allegiance over time. Intermediaries’ flexibility in building electoral alliances created diverse electoral patterns, with no fixed ethnic voting blocs. This chapter proceeds as follows: For each country, I first provide a brief outline of the context of political competition before discussing the dominant modes of electoral mobilization and the ensuing electoral outcomes. In the Senegalese case, I further discuss the behavior of intermediaries to illuminate their logic of supporting different candidates, which explains the absence of ethnic voting patterns. 81

82

Mobilization Strategies and Electoral Outcomes

the context of elections in benin Benin followed a typical political pattern of many African countries – first a brief period of competitive politics immediately before and after independence, a prolonged one-party rule in the 1970s and 1980s and redemocratization in the early 1990s. Instead of fostering democracy, the first mass electoral contests caused significant instability in the country leading to four military coups. In the first 12 years of independence Benin experienced a repeated cycle of elections, gridlock, coups and further elections.1 The country’s early experiment with competitive electoral politics was finally stopped by a decisive coup in 1972, led by Mathieu Kérékou. In the wake of the coup, Kérékou established a Marxist one-party military dictatorship, which lasted 19 years. Following significant protests and calls for democratization at the end of the Cold War, Kérékou called a National Conference to negotiate with the opposition and important civil society actors. The Beninese National Conference, Africa’s first, later became a template for other countries across the continent. Following the proceedings of the National Conference, oneparty rule was disbanded and an interim leadership was established ahead of the multiparty elections in 1991. Kérékou lost, and conceded, the election to Nicéphore Soglo. Since the transition, Benin held five presidential elections and has been consistently ranked as one of Africa’s most democratic states. The elections have been largely free and fair, and very competitive, resulting in several alternations in power. Despite high competitiveness of elections, the number of viable political actors has been limited. Both periods of competitive politics – the early post-independence years and the post-democratization period – have been dominated by a few big personalities and there has been high level of what Africans often call “recycling” of political elites. All elections before the 1972 coup can be seen as contests between three politicians – Justin Ahomadegbe, Sourou Apithy and Hubert Maga. In the post-democratization period, elections also have been dominated by a handful of actors: the former dictator, Mathieu Kérékou, who later won two democratic elections and returned to the presidency between 1996 and 2006, Nicéphore Soglo, the first president after the transition, Adrien Houngbedji and Bruno Amoussou, two perpetually unsuccessful 1

Importantly, ethnic patterns described in subsequent sections, which emerged already in the first set of mass elections, preceded the first military coup. The ethnic cleavages in later elections thus are not a product of the coup. Rather, the coups were brought about in part by the divisive electoral results.

Electoral Mobilization in Benin

83

presidential candidates and Yayi Boni,2 in power between 2006 and 2016. This focus on a few figures is not accidental. They have been the key candidates in all presidential elections and the heads of most important parties or alliances in legislative elections. As in many African countries, politics in Benin is very personalized.3 Voters rarely refer to, or even know, parties’ names. Even in legislative elections, when asked who they voted for, people typically mention the name of the head of the party rather than the party itself. Parties appear to be mere extensions of important political leaders. They are also almost completely dormant in between elections. For example, parties’ offices are closed and there are very few party activities. As is also typical in Africa, there are well over a hundred registered parties in Benin. Most of them are very small, almost one-man operations. Only the few leaders mentioned above run substantial organizations and are capable of mobilizing a significant number of voters. Since parties are secondary to the few established politicians, political competition in Benin is best understood as contests between several politicians, representing different segments of Beninese society. Despite Benin’s high democratic ranking, the actual political campaigns can be rather disappointing for scholars of democracy. The election period is mostly devoid of important debates and instead is dominated by “campaign folklore,” namely song and dance performances at rallies, and ubiquitous distribution of gifts and money. In addition to the large rallies, most politicians conduct what they call proximity campaigns (campagnes de proximité), when they personally visit voters.

electoral mobilization in benin In the absence of strong hierarchical ties in Benin, local leaders lack the influence to act as effective intermediaries between politicians and the populace. As expected in such cases, politicians relied extensively on appeals to shared ethnic identities to win voters’ support. The following paragraphs will show that the use of ethnic references has a long pedigree in Beninese politics from the first mass election to the most recent polls. As Seely describes, ethnic affiliations have been an important means of political mobilization since independence (2007: 197). 2

3

The President’s full name is Thomas Boni Yayi, but in Benin he is typically called Yayi Boni. Reversing the order of first and last names is very common in everyday speech in Benin. See van de Walle (2003) for a review of personalism in African politics.

84

Mobilization Strategies and Electoral Outcomes

Observers of the early rounds of elections note that appeals to ethnic allegiance were common. Decalo dates the rise of ethno-regionalist4 parties in Benin to the 1951 elections to the French National Assembly, first poll to involve a mass electorate (1973: 453, also Banégas 2003: 31). Glélé notes that this cleavage was a product of ethnic campaign manipulations as political leaders mobilized tribal sentiment (1969: Chapter 10). Within a few years’ time, there was a “routinization of political appeals to primordial loyalties” (Decalo 1973: 453). This politicization of ethno-regional cleavages by politically ambitious elites led to the coalescence of regional sentiments and the emergence of “the tribal fiefs” of the three main political leaders (Decalo 1970: 450–452). In these fiefs, Ahomadegbe represented the Fon of the former kingdom of Abomey, Apithy the Yoruba and Goun of the former kingdom of Porto-Novo and Maga the Bariba of Borgou and other Northerners. Electoral campaigns involved the language of intense regionalism and ethnic exclusiveness. All three main political leaders played on historical rivalries to consolidate their electoral blocs. For example, Maga “appealed to regional and ethnic sentiments” and “exploited northern grievances against southern domination” (Decalo 1970: 453). In turn, during the 1963 campaign, Southerners were mobilized against Maga with hurtful ethnic stereotypes and chants such as “a guy from the North cannot govern us” (Banégas 2003: 8). In the 1970 election “appeals to trial loyalties resurfaced rapidly” as the three main leaders continued to use them (Decalo 1970: 454–456). At the same time, efforts to create intertribal parties were unsuccessful. In the 1970 election there was only one candidate, Emile Zinsou, who campaigned as a nontribal, nonregional alternative but he only received 3 percent of the vote. Despite a 20-year-long suspension of multiparty politics from 1972 to 1991, reliance on ethnic mobilization resurfaced in Beninese politics after the democratic transition and restoration of competitive elections. Media accounts of recent elections commonly contain overt references to ethnic factors. Writing about the 1995 legislative elections, a leading daily notes that “the parties got their score according to the population of the ethnic group of the party leader” and that “the ethnic group of the party leader is the major asset for political parties.”5 Kouassi Degboé, an oft-cited observer of contemporary politics, argues that “in the 1995 legislative 4

5

The terms ethnic, regional and ethno-regional are often used interchangeably in Benin, because of the tight overlap between ethnicity and region. Le Matinal, January 26, 1996.

Electoral Mobilization in Benin

85

elections ethnic group affiliation was the fundamental basis of electoral expression. Ethnicity seems the focal point around which the electorate is constituted.”6 Media descriptions of Beninese elections liken those contests to an ethnic census: “Elections reserve little surprise: attachment to the son of the soil (fils du terroir) and an unconditional vote in his favor constitute a strongly implanted reality in the Beninese tradition and democratic culture.”7 A journalist covering the 2007 legislative campaign for the Beninese national television ORTB explained that, in his view, money and ethnicity are the only things that matter and that “the vote is purely ethnic.”8 The dominant view in the country is that a party wins votes based on the origin of its head, together with the ethnic identities of the party’s candidates. This “double-ethnic strategy,” as it is commonly known in Benin, was often brought up by politicians whom I interviewed. For example, Edgard Capo Chichi, a campaign coordinator for the presidential alliance, Forces Cauris pour un Bénin Emergent (FCBE), in the 15th electoral district of Cotonou, confirmed that ethnic considerations constitute the main electoral logic of political campaigns, including the choice of candidates.9 FCBE’s national coordinator, Eugène Azatassou, argued that “identity voting (réflexe identitaire) has always been very strong in Benin.”10 Dr. Djibril Debourou, a Member of Parliament (MP) elected from the FCBE list in the seventh district, concurs that the two sine qua non conditions for electoral success are the origin of the candidate and the identity of the party leader.11 Such views were repeated almost verbatim by several other politicians. Professor Théodore Holo from Renaissance du Bénin (RB) argues that voters first consider the origin of the party leader and then the origin of the candidates in their district.12 Ali Houdou from Parti Socialiste du Bénin explains that “the ethnic factor dominates” and “candidates are primarily viewed through the party leader.”13 Aurelien Housso, a former minister under President Soglo (RB), also stressed ethnic solidarity ties as the building blocks of electoral strategies.14 In his view, parties are “ethnicized” (ethnicisé) and ethnicity is the principal electoral 6 8 9 10 11 12 13 14

7 Le Matinal, January 26, 2006. Le Matinal, February 22, 1996. Author interview, Cotonou, March 11, 2006. Author interview, Cotonou, March 27, 2007. Author interview, Cotonou, January 4, 2013. Author interview, Cotonou, April 3, 2007. Author interview, Abomey-Calavi, April 3, 2007. Author interview, Cotonou, April 4, 2007. Author interview, Cotonou, April 5, 2007.

86

Mobilization Strategies and Electoral Outcomes

variable. Reliance on electoral intermediaries is not very common, with some minor use of traditional chiefs and healers (guerriseurs), and appeals are made to ethnic groups as a whole. As he elaborates, in presidential elections, each group usually has its own candidate and ethnic solidarity is the key electoral strategy. Despite initial hopes that Yayi Boni’s election in 2006 would end the era of ethnic politics, ethno-regional tensions have not waned in recent years. President Yayi’s mixed ethnic parentage – his father is a Nago (a subgroup of Yoruba) and his mother a Bariba15 – made him distinct from all his rivals, who were ethnically and regionally anchored. Commentators talked about his “transculturality” and speculated that he would appeal across ethnic lines. Indeed, his 2006 campaign stood out for its technocratic character and its emphasis on change. Yayi was the sole candidate who placed political advertisements in newspapers and he portrayed himself as a “development professional” (Mayrargue 2006: 163). He also accentuated his political “virginity’’ since he did not belong to any political establishment (Mayrargue 2006: 164). Despite pursuing this highly unusual campaign as a challenger in 2006, during his tenure in office and over the course of his reelection bid in 2011, President Yayi revived ethnic appeals. During the 2011 campaign, the leader of the main opposition alliance, Union fait la Nation (UN), decried “the regionalist character of the Yayi regime.”16 Similarly, another important political adversary, Abdoulaye Bio Tchané, said that he “regretted that this [2011] election was marked to such extent by regionalism.”17 Accusations of ethno-regional appeals weren’t voiced solely by political opponents. As Professor Emmanuel Ahlinvi argues, “Yayi does a lot to highlight ethnic cleavages.”18 In his view, regionalism has increased in Benin since 2006. Three different employees of the National Electoral Commission (CENA) have commented to me on the use of regional cleavages by the president.19 They argued that the president’s “transculturality” was just an appearance. While President Yayi was seen at first as a technocrat, people now call him a “regionalist.” The fact that the president surrounds himself primarily by Northerners and allegedly invests a lot of state resources in the North is cited as an example of his

15 16 18 19

See Mayrargue (2006: 164) for a description of President Yayi’s background. 17 Le Matinal, February 16, 2011. Le Matinal, March 16, 2011. Author interview, Parakou, January 9, 2013. Author’s interviews with Guy Yarou, Romouald Goubiyi and Britanicus Ogoubiyi, Cotonou, January 4, 2013.

Electoral Mobilization in Benin

87

regionalism. Banégas (2014) expresses a similar judgment of ethnic favoritism during Yayi’s presidency. He notes that seven years into Yayi’s time in office “issues of ethnicity are now discussed openly and are increasingly used to interpret the political and social situation” while “the regionalist mode of thinking has now become central to political thought and activity” (Banégas 2014: 453, 457). Politicians bear significant responsibility for invoking historical animosities and making differences between regions highly salient.20 For example, in northern politicians’ campaigns, “the South is often portrayed as a threat.”21 In one version of this tactic, northern voters are told that “their misery is caused by the South.”22 Cyprien Koboudé argues that President Yayi reinforced the idea that “it would be dangerous for people in the North to have a Southerner in power.”23 According to him, this “regionalist discourse” affects not only voters, but also potential candidates. Southern parties struggle to recruit Northerners as candidates because “[P]resident Yayi would tell people that a Northern guy who works for a Southern party is a ‘traitor’.” A northern politician, Mayor Soulé Alagbé of Parakou, accuses his southern counterparts of the same tactics. According to him, southern politicians “want to be regionalists, they divided the country in two [as] they try to get rid of a Northerner.” He accuses them of telling voters that “the North already had power, now it is the South’s turn.”24 Even the national coordinator of the presidential alliance FCBE, Eugène Azatassou, admits that unfortunately politicians play on ethnic sentiments.25 As he laments, “[W]hen you don’t have programs (projets de societé), you rely on ethnic voting (réflexe identitaire).” As he explains, when politicians campaign in their home areas they say: “I am your son (fils), will you take an outsider to govern you?” As a Southerner himself, he says that one frequently hears politicians in the South say “we are governed by a guy from the North, do you want that to continue?” While ethnic references are widespread on the campaign trail, Azatassou points out that politicians usually make them in their home regions “outside the presence of the press.” As he explains, in case you occasionally slip up in 20

21 22 23 24 25

This dynamic is akin to Ferree’s (2010) argument that politicians in South Africa actively portray other parties in racially exclusive terms to gain electoral advantage. Author’s interview with Aurelien Housso, Cotonou, April 5, 2007. Author’s interview with Djibril Debourou, Cotonou, April 3, 2007. Author interview, Cotonou, January 4, 2013. Author interview, Parakou, January 9, 2013. Author’s interview with Eugène Azatassou, Cotonou, January 4, 2013.

88

Mobilization Strategies and Electoral Outcomes

front of a journalist, you ask the press to “censor your remarks,” namely, to leave out the ethnic references. Beyond mere recriminations between politicians, ethnic electoral considerations are reflected in where parties and candidates choose to campaign. It is very common among Beninese politicians to completely ignore entire regions where they do not have coethnics. As one politician explains this logic: “[P]arties campaign where they have a chance of winning and they don’t want to waste their energy where they won’t win votes.”26 For example, one northern politician admitted that his party didn’t campaign in the South because it didn’t think it had a chance.27 Another politician pointed out the fact that in the 2011 presidential election, Adrien Houngbedji, the main southern candidate, hardly campaigned in the North, because “it wasn’t worth it.”28 Houngbedji did not make a trip to the North until one week into a two-week-long official presidential campaign period and he only made one stop, in the North’s largest city, Parakou. Cyprien Koboudé, a politician from UN, an electoral alliance that supported Houngbedji’s 2011 presidential bid, claims that concentrating campaign efforts in the South is entirely logical, because it is difficult for a Southerner to go and campaign in the North.29 This view is echoed by Professor Ahlinvi, who argues that “people in the North are not open to supporting southern politicians.”30 Likewise, Salomon Biokou notes that “the North was always inaccessible” for southern politicians due to their “resistance to Southerners.”31 The choice of avoiding campaigning outside one’s ethnic homeland is based on a belief that voters are not open to supporting non-coethnic politicians. As Biokou argues, talking about his hometown of Porto-Novo, people “won’t take a guy from the North, they don’t want outsiders.”32 In conversations about ethnicity and voting, two words, “brother” and “blood,” come up repeatedly. Coethnic politicians are often explicitly described as bound by ties of blood with members of their group. A court clerk in Parakou sums up popular perceptions about the logic of

26 27 28

29 30 31 32

Author’s interview with Innocent Attanon, Cotonou, March 27, 2007. Author’s interview with Ali Houdou, Cotonou, April 4, 2007. Author’s interview with Eugène Azatassou, Cotonou, January 4, 2013. The exact same judgment was made to me by Emmanuel Ahlinvi, a Professor of Political Science at the University of Parakou, January 9, 2013. Author interview, Cotonou, January 4, 2013. Author interview, Parakou, January 9, 2013. Author interview, Porto-Novo, January 5, 2013. Author interview, Porto-Novo, January 5, 2013.

Electoral Mobilization in Benin

89

voting the following way: “[H]ere people vote according to their blood” (ici on vote le sang).33 Talking about his northern region, he adds that “people will vote for a Northerner even if he is not great, but he is from the North.” Professor Ahlinvi explains that people are convinced that you have to support “your brother” (frère), i.e. a coethnic. They believe that only “your brother” will help you and would understand you. He goes on to say that people don’t understand that, in reality, you don’t have to have someone from your own ethnic group to help your area, but politicians take advantage of people’s ignorance.34 A southern politician, Cyprien Koboudé, shows a similar understanding of vote choice. According to him, people say to themselves: “I need to have my brother there [in office], so that I can find work, so that I can prosper.”35 He adds that sometimes even if your coethnic politicians don’t give you anything, you still support them, because of the perception that only one’s own would help you. People believe that “you could knock on your coethnic’s door” and ask for help. The perception that it would be easier to turn to a coethnic, if such a need arises, is a common theme in everyday conversations about politics. Many voters explicitly mention that the potential ease of contact is a very important reason why people prefer to vote for coethnics.36 As some electors explained, it wouldn’t be easy to contact a politician from another area. Others expressed the belief that “non-coethnics would be reluctant to see you.” In contrast, people believe that coethnic politicians would “fix their problems” and bring money to their area. An Adja voter from Mono discussed with me the allegiance of Adja voters to their coethnic and perennial presidential candidate, Bruno Amoussou. He explained that people in Mono vote for Amoussou “[b]ecause he is their brother.” “He has money. They know it and they know him.” Over the years Amoussou provided very concrete goods for his area – he built a hospital, extended access to potable water and paid school fees for many children. As this voter concluded referring to the choice between his coethnic, Amoussou, and the president, “I’m from Mono, I can’t just get up one day and support a northern politician.” In contrast to ubiquitous references to the political use of ethnicity by politicians, journalists and voters, references to intermediaries are almost

33 34 35 36

Author interview, Parakou, January 8, 2013. Author interview, Parakou, January 9, 2013. Author interview, Cotonou, January 4, 2013. Interviews conducted in Cotonou, January 2013. In contrast to interviews with elites, I withhold the names of individual voters.

90

Mobilization Strategies and Electoral Outcomes

nonexistent. While my Senegalese interlocutors brought up the usefulness of intermediaries almost in every interview, such statements were lacking during my interviews with Beninese politicians. When asked specific questions whether they would consider using intermediaries, most Beninese politicians did not understand the question. While in Senegal there are several French synonyms to describe intermediaries, all of them well understood by the public, none of these terms were used in Benin. The term “opinion leader” (leader d’opinion) is the only expression remotely approximating an intermediary. However, the frequency of this term and the perceptions of the potential usefulness of such an actor were low. While studying months of newspaper coverage and interviewing journalists, the absence of references to intermediaries became striking. Subnational Differences in Benin Although in Benin on the whole intermediaries played a marginal role, there are some differences between the North and the South, which are worth noting. Examples of electoral intermediaries in the South are hard to find, but they are more frequent in the North, as we would expect, given that hierarchical ties are more pronounced in the North. This regional contrast was manifest from the earliest elections in the 1950s, when Beninese representatives to the French Assembly employed electoral strategies that mirrored their regional social structures. Sourou-Migan Apithy, a Southerner, commanded an explicit and direct ethnic vote, whereas Hubert Maga, a Northerner, relied on the chiefs to organize electoral support (Staniland 1973: 306). In order to gain votes, Maga needed the consent of a few grands électeurs, including several ruling families, Fulani chiefs and merchants, such as the most famous trader from Djougou, Paul Darboux, known as “the notable” (Glélé 1969: 121, Staniland 1973: 307). As Decalo pointed out, “social cohesion around Hubert Maga is enforced by chiefs and royal princes” (1973: 451). Maga relied heavily on political brokers in small villages in the North, typically the village chief or the council of elders, who mobilized on his behalf (Decalo 1973: 453). Maurice Glélé portrays well the degree of Maga’s dependence on electoral intermediaries and observes that Maga entered his political career as a “prisoner,” linked to “certain grands électeurs from the North, who brought him their clientele at the price of substantial political and material advantages” (1969: 121).37 37

For example, one of these intermediaries was rewarded with the position of Minister of Economy and Commerce (Glélé 1969: 247).

Electoral Outcomes in Benin

91

Some use of intermediaries in the North is also evident during the more recent elections after the restoration of multiparty politics, although it remains a rare strategy in Benin on the whole. Describing the inner workings of President Mathieu Kérékou’s reelection campaigns in 2001, Bierschenk reports that Kérékou’s coordinator in the northern city of Parakou “did not forget the traditional peddlers of political influence, in particular the city’s numerous mosques and their imams, who all received more or less significant sums of money from him during the election” (2006: 560). During the 2008 communal elections in the region of Parakou, an incumbent MP was visiting notables and Muslim religious leaders in search of support (Bokoma 2011: 12). There were also instances of Muslim religious associations staging prayer sessions for the incumbent president (Mama 2008: 21). It needs to be stressed that the instances mentioned earlier when intermediaries were used are an exception to an otherwise widespread strategy of mobilization of ethnic groups, but the existing exceptions come from the one part of Benin where some hierarchical ties remained.

electoral outcomes in benin In Benin, where there were no suitable intermediaries and electoral mobilization took place along ethnic identities, political competition resulted in a substantial degree of ethnic politics. Parties demonstrate distinct ethnic compositions, and most of them receive the majority of their support from a single ethnic group. Ethno-regional electoral parties emerged in Benin almost instantly after the introduction of mass politics. Already in the 1950s, while Benin was still under French colonial rule, there was an ethno-regional party system composed of three regional blocs (Decalo 1970, 1973, Staniland 1973). In this three-party system, Sourou-Migan Apithy controlled the Goun and Yoruba of Porto-Novo area, which is dominated by Houngbedji’s Parti du Renouveau Démocratique (PRD) in the current era, Justin Ahomadegbe represented the Fon-dominated Southwest, now Soglo’s base, and Hubert Maga controlled the North. Accordingly, some of the earliest post-independence political contests show ethnically based electoral results. Unfortunately, we do not have individual-level data for that time period, as surveys of electoral preferences were virtually nonexistent. Any analysis of voting patterns thus has to grapple with ecological inference problems. Despite this limitation, we can still draw some conclusions about the electoral patterns. In the 1960 General Council elections, Apithy’s Parti des Nationalistes Dahoméens

Mobilization Strategies and Electoral Outcomes

92

table 4.1 1960 General Council Elections PND (Apithy) District Southeast South Southwest Center Northeast Northwest Total

UDD (Ahomadegbe)

RDD (Maga)

Total seats

Votes

Seats

Votes

Seats

Votes

Seats

45 35 30 45 30 40 225

53,912 30,121 24,884 24,971 0 0 133,888

45 20 9 10 0 0 84

13,432 38,690 33,210 66,931 0 6,214 158,477

0 15 21 35 0 0 71

0 0 0 0 43,184 49,970 87,154

0 0 0 0 30 40 70

Source: Decalo (1973: 454).

(PND) won all 45 seats in the Goun and Yoruba–dominated Southeast, whereas Ahomadegbe’s Union Démocratique Dahoméenne (UDD) won the majority of seats in the Fon-dominated Center and Southwest. Neither party won any seats or more than a handful of votes in the North (see Table 4.1). Most strikingly, Maga’s Rassemblement Démocratique Dahoméen (RDD) won all of the seats in the North without contesting any constituencies in the South or Center of the country. While we lack the fine-grained data to precisely map ethnic identity onto voting patterns that would allow us to prove conclusively that the PND drew the majority of its support from the Goun and UDD from the Fon, we can ascertain that their electoral bases differed in ethnic terms from the RDD. Given that the North and South of Benin have completely different ethnic composition, a few facts emerge. RDD had no electoral support from either Fon or Adja, the country’s two largest ethnic groups. RDD won votes only in the North, where there are hardly any Fon or Adja, under 3 percent and 1 percent of the northern population, respectively.38 Similarly, we can establish that the PND had no support from the Bariba, Dendi or Ottomari, who together constitute around 17 percent of the national population. These groups reside almost exclusively in the North – they total less than 1 percent in the rest of Benin39 – and the PND did not win a single vote in the North.

38 39

Institut National de la Statistique et de l’Analyse Économique (INSAE), Cotonou, Benin. INSAE.

Electoral Outcomes in Benin

93

The same pattern emerged during the 1960 presidential election and then was repeated ten years later in the 1970 election. Decalo argues that the ethnic-oriented results of the 1970 election “vividly confirmed the tribal-regional basis of power” (1973: 470). These ethno-regional electoral patterns were remarkably sticky, as the percentage of votes each candidate obtained in his core area in 1970 was virtually the same as in 1960 (see Table C.1 in Appendix C). The 1972 coup, led by Mathieu Kérékou, marked the suspension of multiparty politics and the beginning of almost 20 years of Marxist dictatorship. Electoral political competition did not resume until 1991 when Benin was one of the first African countries to enter the “third wave” of democratization. Remarkably, the first multiparty elections following Kérékou’s dictatorship bore a striking resemblance to the pre1972 ethnic political dynamics. Once again, a system of ethno-regional parties built around regional leaders emerged (Bierschenk 2006). This is despite the new Political Party Charter’s stipulation that parties could not be formed on ethnic principles, but must have a “national character.”40 Each of the ethno-regional blocs has its presidential candidate. Just as a Goun (Porto-Novo) bloc coalesced around Apithy in the 1950s and 1960s, it reemerged in the 1990s, controlled first by Albert Tevoedjre and then by Adrien Houngbedji. Likewise, one can see Soglo’s Fon base as the continuation of Ahomadegbé’s southwestern base in the 1960s. Finally, the North remains the “residual” bloc, never having been incorporated into any of the southern ethnic blocs. The only change is the increased number of ethno-regional blocs from three to four, after the Adja in Southwest broke away from the Fon bloc under the leadership of Bruno Amoussou’s, Parti Social Démocrate (PSD). Banégas concludes that after the transition to democracy, Benin went from a national single party to regional single parties (2003: 243). These observations are borne out by the Afrobarometer surveys, which provide individual-level data on voters’ identity and their political allegiance. For example, the 2005 data show that the PSD and its leader, Amoussou, has a predominantly (79 percent) Adja base, RB and Soglo have mainly Fon supporters (86 percent) and the PRD and Houngbedji have a mostly (80 percent) Fon electorate, but mainly from the distinct Goun subgroup41 (see Table 4.2). 40 41

Article 4, Law No. 90-025, August 13, 1990 (cited in Battle and Seely 2007: 4–5). See also Afrobarometer Briefing Paper No. 17, 2005: 7. The Beninese census does not differentiate between Goun and Fon. Therefore, PRD’s supporters are (misleadingly)

Mobilization Strategies and Electoral Outcomes

94

table 4.2 Ethnic Composition of the Electorates of the Main Presidential Candidates in Benin Candidates Ethnic group Fon Adja Bariba Yoruba Ditamari Peul

Group’s share of Soglo Houngbedji the population (RB) (PRD) 42% 16% 10% 12% 7% 5%

86% 8% 6% 0% 0% 0%

80% 8% 2% 8% 1% 0%

Amoussou Yayi (PSD) (Unaffiliated) 11% 79% 4% 2% 0% 0%

34% 4% 22% 19% 4% 3%

Note: Respondents were asked “if the presidential elections were held tomorrow, who (a candidate from which party) would you vote for?” (Question 99). I include candidates who received over 5% of the vote in the 2006 presidential election. Together the four candidates received over 80% of the vote. Please note that since the Afrobarometer survey preceded the election by about ten months, around 40% of respondents stated that they did not know how they would vote and further 16% indicated that they wanted to vote for the then incumbent, Mathieu Kérékou, who was barred from running for reelection due to presidential term limits. Ethnic group membership is based on Q79. Group’s share of the population size is based on the group’s share of the sample (rather than the population as a whole). Since it is an equal probability sample, it should very closely approximate groups’ share of the population. Source: Afrobarometer, Round 3 (2005). N = 1198.

The results from the 2005 survey are by no means idiosyncratic. Round 4 of the Afrobarometer survey, conducted in 2008, as well as Round 5, conducted in 2011, broadly confirm the electoral patterns (see Tables C.2 and C.3 in Appendix C). Adrien Houngbedji, who came second in the 2006 and 2011 elections and who is currently Benin’s most wellknown southern politician, has virtually no support among ethnic groups in northern Benin. In Round 5 of the Afrobarometer survey, not a single Bariba respondent, from the North’s largest ethnic group, expressed a willingness to vote for Houngbedji. Yayi Boni is the only major national politician in the multiparty era that has a more varied, albeit not fully representative, electorate.42

42

classified as Fon. Battle and Seely (2007: 16) cross-checked PRD’s supporters’ geographic location with fine-grained identity data for the relevant localities and they confirmed that the vast majority of them were indeed Goun. It is worth pointing out that, in contrast to other candidates, Yayi Boni benefits from a mixed ethnic background (Mayrargue 2006: 164).

Electoral Outcomes in Benin

95

President Yayi retains respectable levels of support among the country’s major ethnic groups, but his coethnic Bariba are nonetheless much more likely to back him. In round 5 of the Afrobarometer survey, 72 percent of Bariba, but only 37 percent of Fon, the country’s largest group, selected President Yayi as their favorite candidate (Round 5, 2011). Other studies confirm the ethnic character of Beninese electoral patterns. An index that measures the association between ethnic identity and party preference (CVELI), calculated on the basis of the individual-level data from the Afrobarometer, indicates that 41 percent of vote choice in Benin can be predicted by ethnicity (Dowd and Driessen 2008). Using original individual surveys, Basedau and Stroh (2011) measure parties’ degree of ethnicization, namely their deviation from the general ethnic composition of society. They find that among Benin’s main parties, only FCBE has a low level of ethnicization, whereas the PRD has medium level and both the PSD and RB are ranked as strongly ethnicized. When they aggregate these results at the party system level, they find that the Beninese system is 44 percent ethnicized. In addition, the analysis shows that all main Beninese parties have either a medium or strong level of regionalization, namely that they draw a large share of their electoral support from a specific region (see Table C.4 in Appendix C for summary). It is crucial to consider whether the electoral patterns among the Bariba in the Northeast, where local leadership remained stronger than elsewhere in Benin, are any different from the rest of the country.43 On the one hand, in presidential elections, the Bariba seem more amenable to support noncoethnics than some of the other relatively small groups, such as the Adja or the Goun. In the first three presidential elections since the return to multiparty democracy, the Bariba did not promote a Bariba candidate. This might not seem surprising, as it is a small group with little electoral weight, but other small groups, such as the Adja, have supported ethnic candidates. The Bariba voted for the incumbent or the eventual winner in all contests, but in all cases this also happened to be a Northerner, making it impossible to verify whether they would be willing to support a Southerner.

43

One complication in assessing how the Bariba should have behaved in presidential elections is that, unlike in the case of Senegal, my theory and the affinity voting theories would make similar predictions for four out of the five elections.

96

Mobilization Strategies and Electoral Outcomes

the context of elections in senegal Senegal also followed the familiar pattern of an early phase of competitive elections, a period of one-party rule and a return to multiparty elections. However, Senegal did not experience any coups; its oneparty rule was less repressive than in Benin and its transition to democracy was more gradual – starting earlier than in Benin, but taking longer to accomplish. Following the early competitive elections, Senegal’s first president, Léopold Sédar Senghor began to restrict political competition in 1963, culminating in the creation of a one-party rule under Union Progressiste Sénégalaise (UPS), later renamed Parti Socialiste (PS).44 Amidst growing social pressures, Senghor disbanded the one-party state in 1976, allowing limited and carefully managed multiparty competition. The state recognized three official political currents – socialist, liberal and MarxistCommunist. It restricted political competition to just three parties, each one forced to run under one of the three different currents.45 These restrictions were lifted when Senghor’s successor, Abdou Diouf, came to power in 1981. Multiparty elections continued throughout Diouf’s rule, but Senegal was regarded as a semi-democracy rather than a full democracy, because of remaining political strictures, an unlevel playing field and a lack of alternation of power.46 Diouf undertook significant political reforms in the 1990s, culminating in the first political alternation in the country’s history when the incumbent lost the 2000 election and gracefully conceded to Abdoulaye Wade. As in Benin, Senegalese politics have been dominated by several big personalities. Senghor and his chief adversary, Lamine Guèye, dominated the early years of mass politics. The post-one-party period saw repeated contests between Senegal’s most notable opponent, Abdoulaye Wade, and the two presidents, first Senghor and then Diouf. Wade ran against the incumbent presidents in five consecutive elections, starting in 1978 and finally winning in 2000. In the post-alternation period, often called époque Wade, the new incumbent president contested all subsequent 44 45

46

The name was changed in 1976. The UPS chose to label itself as a socialist party, while Abdoulaye Wade’s Parti Démocratique Sénégalais (PDS) was forced to define itself as liberal democratic and the Parti Africain de l’Indépendence (PAI) accepted the “Marxist or communist label.” Beyond these three categories, other parties, such as Cheikh Anta Diop’s Rassemblement National Démocratique, were barred from participation. See, for example, Villalón (1994) and Coulon (1988) for a description of Senegal as a semi-democracy.

The Context of Elections in Senegal

97

elections, including a controversial, and ultimately unsuccessful, bid to win a third term in 2012. Wade competed against a couple of perennial opponents, a former and a current member of the PS, Moustapha Niasse and Ousmane Tanor Dieng, and two younger-generation politicians, Idrissa Seck and Macky Sall. Both Seck, who came second in the 2007 election, and Sall, who won the 2012 election, were former Prime Ministers from Wade’s Parti Démocratique Sénégalais (PDS), and Wade’s putative successors, until each of them had a very public falling out with their boss. These figures dominated electoral competition as well as party politics. As in Benin, there are hundreds of registered parties, but only three or four viable parties and alliances, led by the main presidential candidates. Although Senegalese parties are more institutionalized than their Beninese counterparts (Riedl 2014), they are nonetheless personalized and very weak by the standards of established democracies. This characterization includes even the seemingly robust PS, which ran the country for 40 years, and the PDS, which was the chief opposition for two decades and was subsequently in power for 12 years. Both parties saw a mass exodus of their members, including incumbent MPs, following their loss of the presidency. Indeed, the extremely high number of MPs switching parties, a practice called transhumance, literally meaning a seasonal migration of cattle, is one of the most salient features of Senegalese politics. While transhumance occurs across Africa, Senegalese MPs more often leave their party to join the new incumbent party, rather than to form their own. Because of the direction of switching toward the incumbent, political commentators often talk about the phenomenon of phagocytosis (phagocytose), where the incumbent “ingests” the opposition. It will become clear in subsequent sections that the high incidence of phagocytosis is aided by the absence of ethnic cleavages, as it gives politicians the latitude to switch allegiance at will. As in Benin, there are also very significant amounts of money disbursed during the campaign period, marring observers’ perception about the quality of the democratic practice. Senegalese elections are also short on specific debates. Parties do not have distinct programs. Instead, elections are characterized by a strong regime cleavage. Voters are either for the president and his party, or against him. Opposition candidates uniformly run on an anti-incumbent platform of change. All opposition candidates compete to be perceived as the most credible agents for change.47 47

This is consistent with the dynamics described by Bleck and van de Walle (2013) in several other African countries.

98

Mobilization Strategies and Electoral Outcomes

electoral mobilization in senegal In contrast to Benin, appeals to ethnic identities are strikingly absent on the campaign trail, in media coverage and in academic works on Senegalese elections. As Beck highlights, “neither the PS [the former ruling party], nor the opposition ever mobilized ethnicity, even though communal identities are important” (2002: 530). On the other hand, the role of intermediaries in Senegalese politics is highly salient. Their presence is apparent to regular citizens and scholars alike. The concept of an intermediary, someone who serves as a link between politicians and the masses, is very well understood and frequently used in any discussion of politics. Indeed, it is so common that the Senegalese have several words to refer to people serving such functions: porteurs de voix (literally “vote carriers”), the most frequently used term, but also “electoral relays” (relais électoraux), vectors (vecteurs) and intermediaries (intermédiaires). Scholars of Senegalese politics have extensively documented the political involvement of communal brokers, such as the marabouts, or the toorodo oligarchy in the Fouta region.48 The only part of Senegal where politicians historically struggled to find suitable intermediaries, given the weakness of the traditional leadership, was Casamance (Lambert 1998, Diop and Diouf 2002a, 2002b, Boone 2003: 109, 116, Beck 2008 ). A rich body of work documents the continuous use of intermediaries by politicians and even the French administration since the onset of mass politics in the 1950s. Cruise O’Brien (1971) gives many examples of intermediary involvement in historic elections, most notably to vote against independence from France in 1958. That was the primary electoral strategy of the two main parties in the 1950s: the Bloc Démocratique Sénégalais (BDS), led by Léopold Sédar Senghor, and the French Socialist Party (SFIO), led by Lamine Guèye. As Boone notes in her case study of the Groundnut Basin, “the BDS and SFIO tried to outdo each other in courting Mouride and Tijan leaders” (2003: 62). After independence, ties with intermediaries, especially the religious leaders, were crucial for President Senghor and his UPS, which succeeded the former BDS (Schaffer 1998: 107). Boone argues that Senghor’s electoral success relied on his use of “established rural powerbrokers” (2003: 60). This tactic was certainly not 48

The most prominent works on the topic include Beck (1997, 2001, 2008), Galvan (2001b), Foltz (1969), Diop (2002), Diop, Diouf and Diaw (2000), Cruise O’Brien (1971, 1975), Coulon (1988), Young and Kanté (1992), Villalón (1994, 1995), Boone (2003). Additional important works on clientelism in Senegal more broadly include Fatton (1986) and Schumacher (1975).

Electoral Mobilization in Senegal

99

unusual since, as Boone notes, no party in Senegal has ever adopted a different strategy for mobilizing electoral support (2003: 60). The same practice was continued in the 1980s and 1990s by Senghor’s successor, Abdou Diouf (e.g. Zuccarelli 1988: 165, Schaffer 1998: 107). Young and Kanté noted that at the basis of Diouf’s electoral success “lay networks of clientelistic linkages joining the president and party to civil society through the intermediation of a host of marabouts and local patrons” (1991: 72). Beck paints a similar picture and, argues that “clientelist networks remain for Senegalese politicians the most efficacious method of garnering political support, also permitting communal leaders to access resources which allow them to reproduce social hierarchies and their status inside those hierarchies” (2002: 543). Newspaper coverage of electoral campaigns is full of politicians’ visits to prominent religious leaders in search of political support. A headline in Walfadjri, “Everyone has a turn with the marabout,” on the eve of the 2001 legislative election, aptly illustrates politicians’ electoral strategies. As the article further explains, “After the campaign of Sopi [change, in Wolof], headed by Wade, which met with all imams from the community on April 17, now it is the turn of the PS to lead a real trek (randonnée) to all the marabouts of all religious localities.”49 During the run up to the February 2007 election, the presence of intermediaries was very noticeable. My interviewees indicated that intermediaries are still used frequently in electoral politics. Members of the ruling PDS talked freely about their party’s reliance on local powerbrokers, especially religious leaders.50 Similarly, on the opposition’s side, a spokesman for a presidential candidate Idrissa Seck also admitted that his boss acquired some support of local leaders, which, as he added, wasn’t free.51 Samir Abourizk, a leader of a small party, Démocratie Citoyenne, who eventually joined the president’s electoral alliance, argued that “all parties during elections are obligated to look for support among religious leaders.”52 Talking specifically about the 2007 campaign, he added that in his view, “all candidates cultivate links with religious chiefs.” An opposition MP from Ligue Démocratique / Mouvement pour le Parti du Travail (LD/MPT), Opa Diallo, likewise believes that all parties

49 50

51 52

Walfadjri, April 23, 2001. Author’s interviews with Babacar Gaye, Dakar, December 1, 2006 and Moustapha Diakhaté, Dakar, January 31, 2007. Author’s interview with Mr. Bocoum, Dakar, February 14, 2007. Author interview, Rufisque, October 16, 2006.

100

Mobilization Strategies and Electoral Outcomes

use intermediaries, whereas many other opposition politicians insisted that it is mainly a strategy of the incumbent and decried it as influence trafficking.53 Sémou Pathé Guèye, a politician from one of the very rare programmatic but electorally irrelevant parties, Parti de l’Indépendence et du Travail (PIT), criticized president Wade for wanting to “control lower classes by local intermediaries.”54 He added that “Wade analyzed very well the country’s network of influence,” conceding that the president and his entourage are very skilled at this practice. The search for intermediaries was also freely discussed in the media by those in President Wade’s entourage. The way Thierno Ousmane Sy (son of Minister of Justice, Cheikh Tidiane Sy) described the quest for intermediaries on behalf of President Wade is highly illustrative: We [those in power] ended up identifying all the vote carriers (porteurs de voix) in the country, all who were likely to make us win in all localities and in all categories and social classes of the society. These vote carriers, at this point, we studied them [and] we know all of them. Their habits, their tastes, their strengths and their weaknesses.55

Similarly, the press chronicled how President Wade, in 2012, relied on traditional authority and frequently visited religious leaders, soliciting their support. As one daily put it, “Wade believes in his chances if he manages to render effective the support of certain religious dignitaries.”56 Other analysts concurred and argued that Wade’s strategy included “seducing the grand electors – marabouts, village chiefs and various religious dignitaries – with ever more resources.”57 Karim Diagne, head of Wade’s electoral alliance in Santhiaba, evocatively explained: “The plan is simply that the President will give us money and we will go see notables, religious leaders and we will convince them and ask them to talk to young people; in exchange we will leave them some banknotes.”58 There is a remarkable consensus among politicians from different parties on who are the best intermediaries in different parts of the country. As Cheikh Seye, from PS, explains, “we know who are the vote carriers, we know who can make the masses vote, it is very objective, we know who weighs heavily.”59 Seye indicates that politicians across the political spectrum tend to analyze social structure, with its networks of dependence, in 53

54 56 58

Author’s interviews with Opa Diallo, Dakar, November 2006, Madior Diouf, Dakar, March 1, 2007, Mamadou Ly and Ibrahima Sene, Dakar, February 13, 2007. 55 Author interviews, Dakar, March 1 and 3, 2007. Le Quotidien, March 3, 2007. 57 Walfadjri, February 6, 2012. Jeune Afrique, March 8, 2012. Sud Quotidien, March 12, 2012. 59 Author interview, Dakar, December 18, 2006.

Electoral Mobilization in Senegal

101

similar ways, and they identify the same people as potentially influential. The case of politicians in the department of Mbour can illustrate this point. The week before the presidential election, I interviewed in the town of Mbour local politicians from all major political forces.60 All of them identified religious leaders as the main players, followed by village chiefs. More importantly, those in search of intermediaries were trying to acquire them from across the identity spectrum. It is not the case that President Wade, a Mouride, seeks support only in the Mouride brotherhood. As the PDS’s campaign chief in the department of Mbour, Magatte Diop, summed up, President “Wade has tentacles in all religious families” in Senegal.61 Diop is well positioned to talk about PDS’s selection of intermediaries. He was personally courting intermediaries in the Mbour area, on behalf of his party’s candidate, President Wade. He was visiting local religious notables and other influential local leaders and “asking for their prayers.” Diop explained that “prayers” are a favorite campaign euphemism describing electoral support. After reviewing weeks of press coverage from several most popular newspapers for the 2007 and 2012 campaigns, it became clear that all major candidates visited religious leaders from all brotherhoods. The religious capitals were bustling with visitors in the run-up to the elections, with different candidates often crossing paths as they criss-crossed the Senegalese religious establishment. The 2007 election saw political involvement of intermediaries holding various social functions and at different levels of renown, ranging from the local to the national level. Religious leaders remain the clearest examples of intermediaries,62 but politicians also rely on customary chiefs or village chiefs. Among religious leaders, politicians worked both with local marabouts as well as those at the top of the hierarchy. Ousmane Tanor Dieng, the presidential candidate from the second largest party, the PS, lamented after his lost presidential bid that President Wade managed to acquire all levels of marabouts as intermediaries, much to the detriment of other candidates.63

60

61

62 63

Author’s interviews with Ibrahima Faye (LD/MPT), Modou Diop (LD/MPT), Tidiane Diop (LD/MPT), Magatte Diop (PDS) and Ousmane Sow (TDS), Mbour, February 20, 2007. Author interview, Mbour, February 20, 2007. A similar point also finds confirmation in Riedl (2014). One of her interviewees, a constituency representative in Podor, stated that during the single-party era the PS was “open to all traditional chiefly families and all (religious) brotherhoods” (2014: 123). Author’s interview with Mbaye-Jacques Diop, Rufisque, March 1, 2007. Author interview, Dakar, April 14, 2007.

102

Mobilization Strategies and Electoral Outcomes

Arguably, the single most important intermediary in 2007 was the Khalifa Général of the Mouride brotherhood. According to many, including a well-known Mouride politician Abbas Bâ, on the eve of the 2007 election, the Khalifa Général sent a coded voting order (ndigal implicite) to his followers (taalibés) in support of President Wade.64 The style of ndigal65 has changed much over the years. While voting orders used to be very public and obvious until the 1980s, in recent years they have become much more discreet.66 Religious leaders do not articulate an explicit voting order, but rather they show their unambiguous support for a particular presidential contender. In 2007, the Khalifa’s support for President Wade became clear during an announcement of the date for the annual Mouride pilgrimage, the Grand Magal. In 2007, the Magal was supposed to take place shortly after the presidential election. During the announcement, the Khalifa stated that President Wade provided all the necessary funds for the Magal and that right after the pilgrimage President Wade would personally oversee great state infrastructure projects (grands travaux) in the holy city of Touba, the center of the brotherhood. The fact that the great projects associated with the Magal were supposed to take place after the election strongly indicated that the president needed to be reelected in order to carry out his work. While to an outside observer this might seem cryptic, or even convoluted, the Mouride voters whom I interviewed had an unambiguous interpretation of this message.67 The use of intermediaries by politicians and parties is accompanied by a belief that intermediaries indeed have an impact on voters’ electoral choices. Mbaye-Jacques Diop, the fourth person in command in Senegal,68 explains that “when people vote, the national figure matters, but you vote above all for the local figure” (le personage local).69 A member of the presidential candidate’s, Idrissa Seck’s, Rewmi party from Thiès observed that when it comes to voting “people listen to 64 65

66

67 68

69

Author interview, Dakar, February 26, 2007. The religious order ndigal also appears in the literature spelled ndigel; throughout this text I will use the first spelling, as it appears in the influential work of Villalón (1995) and Schaffer (1998). One of the reasons behind the changing nature of ndigal over the years is growing social objection to voting orders. Many activists have portrayed ndigal as antidemocratic, making overt voting orders less socially acceptable. Author interviews, Touba, March 2007 and Palmarin, February 2007. President of the Conseil de la République, an institution akin to the upper house of parliament. Author interview, Rufisque, March 1, 2007.

Electoral Mobilization in Senegal

103

those who feed them.”70 Babacar Gaye from the ruling PDS explained that intermediaries are useful because they can influence scores of people.71 Local leaders of LD/MPT in Mbour reported that they work with religious dignitaries, customary and village chiefs to get votes because in their impoverished area “everyone is dependent” and local leaders often influence how people vote.72 There is even evidence that after the 2007 election, President Wade attributed his high score in the department of Mbacké, including the capital of the Mouride brotherhood, the holy city of Touba, to the Khalifa’s backing. Wade arrived in Touba within 24 hours of the electoral results indicating his victory to thank the religious establishment for their support. In a meeting with the Khalifa Général of the Mouride brotherhood, Wade proclaimed: “This victory is yours. I come to show you my gratefulness.”73 The president then went on to reassure the Khalifa Général that he will uphold his end of the deal: “Inchallah [God willing], the first projects during my second term will take place in Touba.” Intermediaries’ Influence While it is easy to document the widespread use of intermediaries by politicians, the reader will wonder whether such intermediaries actually influence the voting behavior of their followers. I showed earlier that intermediaries play an important role in the lives of their dependents, affecting their economic well-being and enjoying a certain level of moral authority and deferential treatment on the basis of their higher social status. There is indeed direct evidence that intermediaries are able to convert their social clout into political influence and that they affect electoral choices of their followers. Numerous scholars who studied the electoral process in Senegal quote their respondents stating that they were affected in their electoral choices by their local leaders. Schaffer’s excellent study of the 1988 election provides particularly clear examples. One of his interviewees simply explained his following of the voting order from the Khalifa Général of the Mouride brotherhood by saying: “If Serign Touba orders us to do something, we do it. Even if it pains us, we have to do it”

70 71 72

73

Author’s interview with Mamadou Sombre, Dakar, February 14, 2007. Author interview, Dakar, November 29, 2006. Author’s interviews with Ibrahima Faye, Modou Diop and Tidiane Diop, Mbour, February 20, 2007. Le Quotidien, March 3–4, 2007.

104

Mobilization Strategies and Electoral Outcomes

(1998: 111). Another voter said that he would have preferred in fact to vote for the opposition candidate Abdoulaye Wade, but voted for his marabout’s choice, the incumbent, President Diouf (1998: 110). This evidence is not only anecdotal. Two different surveys show that people are influenced in their electoral choices by intermediaries. A study conducted in 1999 in the departments of Thiès and Diourbel by Gercop, a group of researchers from the Senegalese University of Saint-Louis, found that 38 percent of their respondents admitted following voting orders (consignes de vote) from religious leaders.74 The 2005 Afrobarometer survey in Senegal also asked respondents “who (if anyone) can influence their political choices.”75 Of those interviewed, 23 percent of Wolof respondents reported that they are influenced “a lot” by their religious leader, whereas 13 percent are influenced a lot by relatives.76 Not surprisingly, the number of respondents influenced by family members is higher among women, standing at 18 percent. The self-reported influence of intermediaries is thus considerable and since admitting to being told how to vote might be somewhat shameful, the real number is likely to be higher.77 Incremental Variation in the Role of Intermediaries: The Mourides versus Tijanis There is evidence not only that people are influenced by their religious leaders, but that the tighter the tie, the more they follow such leaders’ suggestions. A comparison of the two largest Senegalese Sufi brotherhoods, the Mouridiyya and Tijaniyya, illustrates this variation. The relationship 74

75 76

77

“Etude sur le comportement électoral dans les régions de Thiès et Diourbel,” Mai 1999. The sample size was 4,877 (2,900 in the department of Thiès and 1,977 in Diourbel). Question 88a_Sen, Round 3, 2005. The difference between the two surveys might be due to the fact that the two departments surveyed by Gercop are in the heartland of the Senegalese Muslim brotherhoods. In contrast, the Afrobarometer surveyed all regions in Senegal, including zones where the dominant type of intermediary comes from the traditional and not religious elites, and where we thus wouldn’t expect voters to turn to religious leaders for help. I limited the sample to Wolof respondents to make it more comparable to the Gercop study, since the Wolof are most likely to be influenced by this particular type of intermediary. The sample still contains Wolof outside the Groundnut Basin where religious intermediaries are weaker. Thus the larger role of religious leaders’ influence in the Gercop study is consistent with our expectations. More recently, Gottlieb (2014) found using behavioral games in Senegal that a high share of voters are willing to vote for leader optimal (but voter suboptimal) options, indicating their deference to local leaders.

Electoral Mobilization in Senegal

105

between Mouride disciples and their leaders is uniformly regarded by scholars and politicians alike as tighter than the one among the Tijanis, because of the Mouride leaders’ greater control of material resources and the brotherhood’s more stringent norms of obedience. Lucy Behrman notes in her comparative study of the Senegalese brotherhoods that there is a “significant difference between the Mourides and other groups in their degree of marabout control over disciples” (1970: 68). First, the Mourides are more dependent economically on their marabouts than the Tijanis. The Mouride brotherhood extends to greater extent into the economic realm, in particular through their control of vast areas of land, mainly in the Groundnut Basin (Villalón 1995: 120), as well as their infiltration of the sectors of urban trade and transportation. Second, the various brotherhoods have different philosophy and they promote distinct values, which affect the nature of ties between leaders and followers. The key contrasting feature, which influences the strength of hierarchical ties within the two largest brotherhoods, the Tijaniyya and the Mouridiyya, is their approach to obedience to religious authority. The Mourides place a much higher emphasis on personal obedience than the Tijani marabouts. The Mouride ideology espouses the virtues of obedience and submission to a much greater degree than in other brotherhoods. Cruise O’Brien (1971) argues that Mourides are more submissive and more devoted to their marabouts than members of other brotherhoods. Already the ritual act of submission, which each person performs to become a disciple of a marabout, reflects the different role of obedience. Even the words in the act of submission are much more deferential among the Mourides (Coulon 1981: 104–105). Whereas for members of other brotherhoods the act of submission is “part of a process of Muslim religious socialization,” for the Mourides “it is simply an engagement of total personal obedience, unaccompanied by religious instruction” (Cruise O’Brien 1971: 86). In comparison to the Mourides’ deferential act of submission, the same ritual is much more relaxed among the Tijanis, and it reflects more “vague forms of dependency” (1971). This high level of submission and obedience is an integral part of Mouride ideology. Villalón explains that Mouride ideology “enshrined the act of submission to a marabout as a central defining feature of the order” and “this submissive aspect of Mouridism has figured prominently in all analyses of the order” (1996: 119). For example, Behrman (1970) notes that the Mourides obey their marabouts to a greater extent, a behavior dictated by the brotherhood’s norms and values. She also highlights that Mourides define one’s

106

Mobilization Strategies and Electoral Outcomes

“demonstration of devotion to Islam through work for and obedience to the marabouts” (Behrman 1970: 62). At the same time, the two brotherhoods approach social changes differently, such as the spread of education and literacy, which could reduce their disciples’ dependence on their leaders. While the Tijani clerics encourage education and independent thinking among their followers, the Mouride marabouts try to suppress it, fearing that it will undermine their control. Behrman compares the teachings of Amadou Bamba, the founder of the Mouride brotherhood to his Tijani contemporary, a legendary marabout, Malik Sy, and concludes that their different attitudes to education were clearly apparent (1970: 63). In his authoritative book The Mourides of Senegal, Donal Cruise O’Brien (1971) argues that in comparison to other orders, the Mouride marabouts actually discourage their followers from pursuing education, because they fear that education would decrease their submission. The Mourides sensed that “the educated disciple tends to become alienated from their authority” (Cruise O’Brien 1975: 74). Many marabouts expressed overt hostility to government primary schools. In response to the expansion of secular education, marabouts often encouraged parents to keep their children at home. As a result, there were “very few schools in the central Mouride zone (due to saintly [marabout] influence on the state authorities), and those which exist[ed] [were] poorly attended (due to saintly [marabout] influence on disciple parents)” (Cruise O’Brien 1975: 74). Educational achievement data confirm that compared to other brotherhoods, about 50 percent fewer Mouride children can speak, read or write in French (1975). The higher deference to Mouride leaders is also reflected in the size of monetary contributions made by disciples to their marabouts; Mourides are much more regular and lavish in their payments78 to their marabouts (Cruise O’Brien 1971: 91–96). Given the different nature of marabout–disciple relations, we would thus expect Mourides to wield more electoral influence over their followers, including in the realm of politics. The more clout religious leaders have, the better they should perform in the role of electoral intermediaries. Indeed, survey data indicate that Mouride marabouts indeed wield more electoral influence over their disciples. The 2005 Afrobarometer survey asked the Senegalese respondents who might influence their vote, and the answers to this question revealed a substantial difference between the

78

An example of such payment is hadiyya, a personal gift to the leader.

Electoral Dynamics in Senegal

107

Tijanis and the Mourides. While only 11 percent of Tijanis say that religious leaders affect them a lot in their electoral choices, the figure is more than double among the Mourides (23 percent).79 These findings are consistent with earlier survey research. A Gercop study of 2900 voters in the regions of Thiès and Diourbel, almost all of them belonging to one of the two brotherhoods in question, found that Mourides were more likely than Tijanis to follow voting orders from their religious leaders.80 These differences are certainly not lost on Senegalese politicians. When asked about the possibility of mobilizing voters from various brotherhoods, most politicians across the political spectrum perceive the Mouride brotherhood as having superior potential to deliver votes. Moustapha Diakhaté from the ruling PDS argues that Mourides have always followed voting orders from the religious establishment to a much larger extent.81 Opposition politicians, Mamadou Ly (AFP) and Ibrahima Sène (PIT), add that Tijani religious leaders are less likely to influence the voting behavior of their followers.82 In the words of Baidy Sall (AJ/Pads), Tijanis follow such advice to a much lesser extent.83 Sall argues that Tijanis feel much freer than their Mouride counterparts. Magatte Diop (PDS) also claims that Mourides follow voting orders “to the letter” whereas in other brotherhoods, people are much freer.84 Madior Diouf (RND) concurs with his fellow politicians that Tijanis and members of other brotherhoods are more emancipated from the influence of religious leaders than the Mourides.85 Similarly, Abbas Bâ from the Parti de la vérité pour le développement (PVD) believes in Mourides’ mobilizational superiority, based on the strength of ties between Mouride marabouts and their disciples.86

electoral dynamics in senegal As the preceding paragraphs outlined, politicians in Senegal and Benin followed qualitatively different mobilization strategies. While in Senegal 79 80

81 82 83 84 85 86

Question 88a_Sen, Round 3, 2005. 40 percent of Mourides, but only 34.9 percent of Tijanis said that they accept voting orders from religious leaders. A previous Gercop study, conducted in Saint-Louis, also found that Tijanis are less influenced by religious leaders. Author interview, Dakar, January 31, 2007. Author interview, Dakar, February 13, 2007. Author interview, Dakar, February 14, 2007. Author interview, Dakar, February 20, 2007. Author interview, Dakar, March 3, 2007. Author interview, Dakar, February 26, 2007.

108

Mobilization Strategies and Electoral Outcomes

politicians used electoral intermediaries extensively, their Beninese counterparts did not, relying instead on bonds of shared ethnicity. Their chosen modes of mobilization, in turn, had implications for electoral outcomes, namely, the presence or absence of identity-based voting blocs. In contrast to direct mobilization of voters along ethnic lines, which simply translates identity into vote choice, resulting, not surprisingly, in identity-based voting blocs, hierarchical mobilization mediates the relationship between voters and politicians through intermediaries, loosening the constraints of shared identity and allowing politicians to win votes across ethnic lines. Electoral intermediaries deliver voters to politicians across ethnic lines, helping parties and candidates build diverse electorates. Before describing the actual electoral outcomes in Senegal, this section elaborates on the mechanism of mobilization through intermediaries. In brief, it will elucidate why this type of mobilization does not inevitably lead to the emergence of identity voting blocs. As I argued in Chapter 2, we should not expect intermediaries to limit themselves to working with coethnic politicians. First, intermediaries have fewer constraints than ordinary voters. In their interactions with politicians, intermediaries, unlike individual voters, are less dependent on shared identity to enforce politicians’ promises. Intermediaries with a substantial base have strong bargaining chips; they could defect to a different candidate and they cannot be easily “cheated” by national politicians. More importantly, it is in the intermediaries’ material interest to keep their options open. Limiting the range of patrons only to their coethnics would narrow their prospects of material rewards. The implication of low constraints and the primacy of material rewards is that intermediaries want to cooperate with a politician who offers them the best deal rather than with a coethnic, unless the two overlap. At the same time, politicians want to hire effective intermediaries regardless of their identity. Consequently, while in any individual case or point in time, an intermediary may work for a coethnic, in the aggregate and over time, we should not see fixed alliances. The rest of this section shows that empirically Senegalese intermediaries have not limited themselves to working with politicians from the same ethnic or religious background. It provides concrete evidence that intermediaries stand to gain access to material resources through their involvement in electoral politics and that these material rewards influence intermediaries’ selection of patrons.

Electoral Dynamics in Senegal

109

In intermediaries’ dealings with politicians, material resources play an important role.87 Politicians’ ability to provide resources is crucial because intermediaries try to facilitate acquisition of public goods for a whole village or a community. Village chiefs, for example, can bargain with politicians and demand communal goods in return for the votes of their village. Playing such a coordinating role is quite common. Several politicians reported being asked by village chiefs during the 2007 campaign for various public goods such as funds for the construction or renovation of a village mosque, or repairs of a broken water pump.88 Similarly, many scholars have documented how the marabouts have been very successful in securing various infrastructure projects in return for the support of their followers. The renovation of the Great Mosque in Touba, alongside electrification and road construction are some of the most salient examples (e.g. Schaffer 1998: 112–113). At a local level, roads to villages, electricity, wells and bore-holes are some of the most frequently demanded goods (Cruise O’Brien 1971: 276). Cruise O’Brien gives an example of “a marabout whose village bordered on Senegal’s central desert, [who] secured the installation of 4 diesel-pumped wells” (1971: 177). In another case, “a marabout procured an 11-mile extension of a tarred road, terminating in his own village and at his front door” (1971). At a higher level, during the 1960 General Election, the UPS contributed over 100,000 pounds to the mosque construction fund in Touba, alongside promises of loans, creation of trading centers, attribution of subsidies and higher prices for the main cash crop, the groundnut (Cruise O’Brien 1971: 267).89

87

88 89

Material rewards might not be the only benefit of being an intermediary. Apart from all the material benefits offered to intermediaries, there is a psychological dimension to their relationships with politicians in the form of prestige and feelings of self-importance. Politicians want to make sure that intermediaries feel appreciated. The language used to address intermediaries is full of flattery and it stresses their important position in their community. Politicians admit that they try to appeal to people’s sense of their own importance. One MP, Babacar Gaye, argues that by being approached by politicians, intermediaries feel valued (ils se sentent valorisés). Gaye stresses the fact that these psychic benefits have roots in Senegalese culture. Signs and feelings of regard, respect or esteem are very important. Unfortunately, it is impossible to quantify the utility gained from feeling flattered or important. However, I mention the possible psychic benefits to indicate that material rewards might not be the only motivating force and that nonmaterial factors should not be regarded as irrelevant. Author’s interview with Madieyna Diouf, Dakar, April 2007. The marabouts show prowess not only in what they can secure but also in what they can avert. Cruise O’Brien shows that the Mouride marabouts were able to block a campaign effort to train peasants in modern agricultural techniques (1971: 275). Such a scheme

110

Mobilization Strategies and Electoral Outcomes

Cruise O’Brien generalizes that the allocation of government expenditure in rural areas was directed toward places with powerful rural notables with many disciples. The government distributed in this manner roads, deep wells, schools, dispensaries, agricultural credit, as well as staple food – millet and rice (1975: 175–176). As he sums up, the leading powerbrokers were in a position to demand many favors for their communities and the subsequent distribution of money and resources favored the communities which were the most effectively organized (1975: 180). From this perspective, peasants who were “leaving politics to their marabouts” were making a rational decision, given how resources were allocated (1975: 177). Schaffer shows that not only were many marabouts successful in securing public goods for their communities, but also that the marabouts’ followers attributed the existence of these goods directly to their leaders. One of Schaffer’s interviewees, talking about the Khalifa Général of the Mouride brotherhood, describes his prowess in extracting communal benefits: “Water, electricity, paved roads – he [Abdou Ahad] got everything he wanted” (1998: 112). He then adds that based on the acquisition of these goods, it was right for the Khalifa to support President Diouf and urge his followers to do so. Abdoulaye Wade has continued the tradition of promising public investment and infrastructure to the religious capitals. Even though the religious establishment in Touba did not endorse candidate Wade in 2000, immediately following his election, President Wade sought the support of the Khalifa Général of the Mouride brotherhood, offering his city important infrastructure, including an airport (Beck 2002: 541). During the 2007 election, President Wade offered religious leaders in Touba 100 billion francs CFA in public funding for the city.90 Yet, importantly, intermediaries also stand to gain personal goods. Apart from extracting funds for various infrastructure projects or other public goods, these prominent local leaders often receive personal benefits, both money and gifts in kind, such as sponsored pilgrimages to Mecca. Cruise O’Brien argues that the flow of money from political sources made it possible for certain marabouts, the Mourides in particular, to “acquire

90

would potentially make peasants more independent vis-à-vis their religious leaders. This example also shows the limits of the maraboutic social safety net system. The marabouts are willing to secure goods for their community only if those goods do not make their followers less dependent on their leaders. Le Quotidien, March 9, 2007.

Electoral Dynamics in Senegal

111

wealth on a scale hitherto unimaginable” (1971: 263). In fact, he goes as far as to say that politics became the greatest single source of revenue for the Mouride elite. Cash is not the only form of benefit offered to intermediaries. President Diouf, for example, gave marabouts diplomatic visas ahead of the 2000 election (Beck 2002: 541). The tradition of rewarding intermediaries with personal goods continued uninterrupted, despite an alternation of power, which took place in 2000. During the 2007 election, there were many reports of cash payments, varying in size according to an intermediary’s importance. Two different MPs reported to me that marabouts who were heads of a lineage were receiving around 50 million francs CFA (about 100,000 USD) and a 4 by 4 car.91 One MP from the ruling party, Babacar Gaye, admitted that he personally bought a new Mercedes for a marabout in his home department of Kaffrine.92 Strikingly, Gaye was in no way secretive about his gift. Indeed, as he explained, he wanted people in Kaffrine to know that he was “taking good care of their marabout.” At a more local level, marabouts were sometimes rewarded with a pilgrimage to Mecca, whereas the lowest level clerics, such as village imams, were given beautifully bound editions of the Koran.93 Village chiefs also often get money. For example, one of the politicians interviewed reported giving 10,000 USD to various village chiefs.94 During the 2012 election, intermediaries continued to be rewarded with personal gifts. Even the state daily, Le Soleil, described the president’s “extreme generosity toward religious dignitaries.”95 Other newspapers reported many “presents” given by the president to local leaders. Various sums of money, all-terrain vehicles and diplomatic passports were among the most oft-cited gifts.96 In one instance of such “electoral charm offensive,” on the eve of the 2012 election, President Wade gave cars and “allowances” of between 50,000 and 70,000 francs CFA to 30,000 village chiefs across Senegal. In a more shocking example, Wade’s entourage sent a sum of 2 billion francs CFA to the religious establishment in Touba, the capital of the Mouride brotherhood, via an ambulance to avoid detection.97 91

92 93 94 95 97

Author’s interviews with Madior Diouf, Dakar, March 1, 2007, and Madieyna Diouf, Dakar, April 2007. Author interview, Dakar, December 1, 2006. Author’s interview with Semou Pathé Guèye, Dakar, March 1, 2007. Author’s interview with Samir Abourizk, Rufisque, October 16, 2006. 96 Le Soleil, February 4–5, 2012. Le Quotidien, March 2, 2012. Le Quotidien, March 2, 2012.

112

Mobilization Strategies and Electoral Outcomes The Logic of Intermediaries’ Support

Knowing what is offered to intermediaries helps us understand what motivates their political allegiance and why it is not in their interest to limit themselves to supporting their coethnics. An anecdote recounted to me by Madieyna Diouf, one of the top leaders of Alliance des Forces de Progrès (AFP), illustrates well the effect of material benefits on intermediaries’ decision making. Diouf remembers that when his boss, the leader of AFP, Moustapha Niasse, campaigned in his home region of Kaolack, in one village a local marabout asked for money for the construction of a mosque. Niasse gave him around 5 million francs CFA (10,000 USD).98 When Niasse visited the village again four days before the election, the marabout asked him for more money. Later that day Niasse sent another 10 million francs CFA via a messenger. But the news of the transaction got out and the ruling party, PDS, became aware of it. PDS then contacted the same marabout and promised him 100 million francs CFA. 50 million was paid upfront and the remaining 50 million was to be paid after the election if the marabout delivered the villagers’ votes. Diouf admitted that these tactics undoubtedly worked since President Wade won overwhelmingly in that village. This story of one village illustrates a broader phenomenon. It shows that the marabout did not have a set preference for a party or a candidate and cared only about his and his dependents’ material reward. He also seemed ready to work with multiple benefactors. First, he was willing to sell his services to AFP but ultimately supported PDS, which proved more lucrative. This type of behavior is characteristic of many intermediaries, including the most famous in Senegalese political history, namely Muslim religious notables, the marabouts. Over the course of political competition in Senegal, the marabouts had the tendency to support candidates on the basis of prospective material rewards, regardless of their religious affiliation, rather than candidates from their own brotherhood. Even before Senegal’s independence, Muslim marabouts were courted by candidates across the political spectrum and faced choices in their political allegiance. Already at the time, promises of rewards, rather than any affinity with candidates, seemed to be the main determinant of support for various candidates. The marabouts were ready to support a candidate that seemed to offer them the best deal. This dynamic was clear during the 1951 legislative election to the French National Assembly. In 1951, the election 98

Author interview, Dakar, April 2007.

Electoral Dynamics in Senegal

113

pitted Léopold Sédar Senghor, a Catholic, against a Muslim, Lamine Guèye. To access the substantial rural electorate both candidates were seeking support of powerful Muslim dignitaries. Arguably, the most important among them was Serigne Falilou Mbacké, the Khalifa Général of the Mouride brotherhood. An account of Serigne Falilou’s interactions with both candidates sheds light on the decision-making process in choosing to support a political candidate. As a Senegalese historian, El Hadj Ibrahima Ndao (2003) recounts, on the eve of the election, both candidates went to the Mouride capital, Touba, to seek the Khalifa’s support. The Khalifa told his confidants that he would support the candidate who came to him and offered resources to continue the construction of the Great Mosque of Touba. Ndao chronicles the Khalifa’s meetings with both candidates: Lamine Guèye came to Touba during the course of his campaign. He promised to help the countryside (le monde paysan) achieve development, in the event of his victory. But he didn’t make any reference to the construction of the mosque . . . . When Senghor arrived for the same reasons, he repeated his objectives and ambitions for Senegal. He reaffirmed his pledge (sermon) of never betraying Islam and above all stressed his decision to help the Khalifa complete the construction of the Great Mosque of Touba. . . . It is from that day that Serigne Falilou Mbacké took up Senghor’s cause and launched his first voting order (ndigal) to the Mouride voters. (2003: 132)

A similar openness to the possibility of supporting either candidate in the 1951 election is also evident in the case of the Khalifa Général of the Tijan brotherhood, Ababacar Sy. The Khalifa was initially more predisposed to his fellow Muslim Lamine Guèye, but after being courted much more by Senghor than Guèye, he decided to support the former (Ndao 2003: 132). With the support of the two most prominent marabouts in Senegal, Senghor beat Lamine Guèye by more than a two-to-one margin (Ndao 2003: 138). In his interactions with the Muslim marabouts, it would appear that the odds of gaining their support were against him. If the marabouts were driven by affective ties they would surely support Lamine Guèye.99 But in the contest for marabouts’ support, it was Senghor who seemed more in tune with what the marabouts wanted and was able to make them more appealing promises.

99

The 1920 election to the French Assembly provides another example of a contest pitting a Muslim against a Catholic candidate, albeit at a time of limited suffrage. The marabouts supported a Catholic, Blaise Diagne, against a Muslim, Galandou Diouf (Beck 2008: 72–73).

114

Mobilization Strategies and Electoral Outcomes

Since then during most, if not all, elections the marabouts tended to support those already in power, including the French administration, earning this practice of legitimizing incumbents the term vote légitimiste. Cruise O’Brien (1971) asserts that in the 1958 referendum, the three great local brotherhoods who had patronage ties to the French administration, sided with the French and helped postpone Senegalese independence, by convincing their followers to vote against it. After independence, the marabouts overwhelmingly supported President Senghor and his party, UPS, which subsequently became a single party. This support continued after a limited liberalization under Senghor and then after restoration of multiparty politics under Abdou Diouf. With a return to competitive politics, the marabouts could have supported different candidates, but they have thrown their weight mostly behind President Diouf. Both in 1983 and 1988 elections, the religious leaders issued a ndigal to their followers to vote for Diouf (Schaffer 1998: 107). In 1993 and 2000, the main marabouts, namely heads of brotherhoods, abstained from issuing a ndigal, but many others did. As electoral contests became more competitive, there has been a certain diversification of voting orders, with some marabouts supporting the main opposition candidate at the time, Abdoulaye Wade. Yet, still the majority of marabouts involved sided with the incumbent (Coulibaly 2006: 235–236). The fact that some marabouts chose not to support the incumbent and abstain was most likely caused by diminished patronage flowing from the Diouf administration. Galvan argues that the shrinking of the Senegalese state, due to the structural adjustment reforms, reduced financial resources that could be channeled to the marabouts, thus souring Diouf’s relationship with the marabouts (2001: 53). At the time of the 2000 election, some commentators thought that the Senegalese society has matured beyond being susceptible to voting orders from religious leaders. Yet, in the light of the increased involvement of religious leaders in the 2007 and to a lesser extent in the 2012 election, such analysis appears to be premature.100 The economic factors seem much more convincing in explaining their less fervent involvement in the two previous elections (Audrain 2004). Intermediaries also showed a preference for the incumbent under the Wade presidency. Whereas Wade struggled to acquire intermediaries as a challenger, he found it much easier as an incumbent. When he was in

100

See Dahou and Foucher (2004) for a similar verdict on the 2000 analyses.

Electoral Dynamics in Senegal

115

opposition Wade did not have the backing of any high-ranking religious leaders, even among his fellow Mourides. As an incumbent, during the 2007 election, Wade enjoyed the support of some of the country’s most prominent religious leaders, including the Khalifa Général of the Mouride brotherhood. However, by 2012, Wade’s ability to acquire intermediaries was much diminished in comparison to the previous election, despite the fact that his campaign courted intermediaries as vigorously as ever before. Wade had the backing of many secondary marabouts,101 but the upper echelons of the religious hierarchies in Touba and Tivaouane remained silent, despite multiple visits by the president and frequent “requests for prayers.” In this respect, the differential behavior of marabouts between 2007 and 2012 mirrors their fluctuating support over time for the Diouf regime. While intermediaries tend to side with those already in power, their support for the incumbent is contingent on his likelihood of winning. When the incumbent is in a strong position, intermediaries are keen to issue voting orders. When a president running for reelection faces strong challenges, intermediaries are much less likely to embrace him. This is because intermediaries’ actions are not only based on concrete goods offered at the time of election, but also on the prospects that the incumbent will be reelected. As in the example of the Khalifa Général’s backing of President Wade in 2007, apart from gifts issued during the campaign, part of the reward was to be provided after the president’s successful reelection. When the incumbent’s reelection appears uncertain, the value of promises of post-electoral transfers is greatly diminished. Moreover, intermediaries tend to be risk averse, not wanting to damage their reputation among their followers by supporting a losing candidate. Indeed, President Wade’s political camp understands that marabouts and other potential intermediaries are less inclined to support a struggling incumbent. Babacar Gaye, who was closely involved in courting marabouts on behalf of the Wade campaign in 2007, explained that in 2012 “marabouts preferred to abstain, given the situation.”102 In other work (Koter 2013: 672–674), I document the growing opposition to Wade’s regime and his 101

102

Among others, Wade gained open support from the Khalifa of Darou Mousty, Serigne Cheikh Khady Mbacké (Le Soleil, February 15, 2012), Cheikh Béthio Thioune, Cheikh Ndigel Fall (Sud Quotidien, February 21, 2012), Seigne Fallou Mbacké (Le Soleil, March 1, 2012), the marabout Pathé Kébé (Le Soleil, March 6, 2012), Serigne Bakhé Mbacké (Le Soleil, March 9, 2012), several religious dignitaries and imams in the department of Kolda (Le Soleil, March 13, 2012). Author’s interview with Babarcar Gaye, Dakar, May 29, 2013.

116

Mobilization Strategies and Electoral Outcomes

deteriorating electoral prospects, which stemmed from Wade’s failed attempt to change the constitution, his controversial bid for a third term and the grooming of his widely disliked son, Karim, as the heir apparent. The importance of the incumbent’s likelihood of winning further indicates that intermediaries are driven by access to resources, state coffers in particular, rather than affective ties. This motive was in fact explicitly articulated by some of my interviewees. A Mouride religious leader and head of PVD, Modou Kara Mbacke, illustrates well this dynamic. In the fall of 2006, Kara, as he is commonly called, was nominated by his party to be PVD’s presidential candidate.103 Yet, not long afterward he joined President Wade’s campaign. During our interview, PVD’s second in command, Abbas Bâ, provided useful details which help us understand why Kara flocked to the president’s side. As he explained, Kara wanted to act as an intermediary for the president rather than for other challengers because “it is Wade who is in power; if we want to be in power, we have to go towards power.”104 Kara’s new allegiance was promptly formalized when on December 25, 2006 Bâ signed on behalf of PVD a “strategic political alliance” with the president’s party, which he proudly showed me.105 Aliou Dia, a MP from the Forces Paysannes, who comes from a maraboutic family in Louga, provides another example of supporting a candidate based on material calculations rather than any other factors, such as ideology or ethnic affinity. In our first interview more than two months before the 2007 election, Dia was very critical of President Wade. However, as elections neared, Dia drifted into the incumbent camp. In our subsequent interview in February 2007, the month of the election, we broached the topic of Dia’s change of position. As he explained, on January 4, 2007 an offer came from President Wade. The president solicited his “help” in his reelection bid.106 Although Dia did not want to disclose the exact terms of their deal, he acknowledged that he signed a “partnership accord” (accord de partneriat). He made no effort to convince me that there was any other rationale behind his support of Wade, apart from the fact that he got the best deal. He said explicitly that he was open to working for other candidates, but supporting President Wade was more advantageous.

103 105

106

104 Walfadjri, November 20, 2006. Author interview, Dakar, February 26, 2007. The pact was signed by the then prime minister Macky Sall, who was also doubling as Wade’s campaign director. Author interview, Dakar, February 9, 2007.

Electoral Dynamics in Senegal

117

In the run-up to the 2007 presidential campaign, there were several highly publicized instances in which intermediaries switched their allegiance to support the incumbent apparently on material grounds. In some cases, the newly professed loyalty to the president was quickly rewarded with a government post. For example, Serigne Mamoune Niasse, a marabout leader of a small party Rassemblement du Peuple (RP), used to support his fellow Tijan, Idrissa Seck, a former Prime Minister who later came second in the 2007 presidential election. Yet, by November 2006, Niasse abandoned Idrissa Seck for President Wade. As the press reported, Wade and his entourage offered to compensate the marabout generously for his change of allegiance. The reward came that same week. After a ministerial reshuffle, which saw an expansion of the number of ministers to a hitherto unseen 41, Serigne Niasse was named “Minister of state beside the president” (Ministre d’Etat auprès du président). As the media concluded, “[T]his was the price paid by [President] Wade to acquire the services of Mamoune Niasse.”107 Importantly, one of President Wade’s closest allies, the third vice president of the National Assembly, Babacar Gaye, admitted that they did not name Niasse a minister because of his “expertise,” but because they “needed links with a marabout.”108 Politicians who “steal” intermediaries from other parties, view their own actions as an effective electoral strategy, much to the lament of other politicians. When Idrissa Seck, a presidential contender who finished second in the 2007 poll, was confronted by journalists at a press conference about luring intermediaries from other political formations, he said that although he is not an advocate of such tactics, he “is

107 108

Walfadjri, November 24, 2006. Author interview, Dakar, December 1, 2006. That same cabinet reshuffle was behind the change of allegiance of Abdourahmane Agne. A former MP from the then ruling PS, Agne founded a small party, Le parti de la réforme (PR) and was involved with an opposition coalition, Coalition pour alternative (CPA), campaigning against President Wade. But as the media reported, he was rewarded for his split with the CPA, by getting a portfolio of a Minister of Micro-finance and International Decentralized Co-operation. Two other ministers in Wade’s cabinet, Ousmane Ngom and Mamadou Diop, joined him after switching from the former ruling party. Madior Diouf, a MP and member of the CPA, gave additional examples of politicians stolen by the PDS: in the department of Louga, Abdourahman Sow, formerly aligned with the PS, switched sides and became first a Vice President of the National Assembly, then an ambassador (Author interview, Dakar, March 1, 2007). In the region of Kaolack, the PDS lured Abdoulaye Diack. Ousmane Tanor Dieng, the presidential candidate of the PS, adds even more names to this list. For example, the mayor of Kaffrine abandoned the PS in favor of President Wade two days before the election (Author interview, Dakar, April 14, 2007).

118

Mobilization Strategies and Electoral Outcomes

for the recruitment of the best.”109 Similarly, Babacar Gaye admitted in our interview that PDS steals intermediaries, but they do it in the spirit of competition, because they want the best ones.110 Intermediaries’ lack of loyalty to specific politicians or parties was further exemplified during the 2012 election. Numerous intermediaries switched allegiance in between the first and second round of the election, abandoning the incumbent, President Wade, during the course of a few days after his poor showing in the first round. Even among Macky Sall’s coethnics in the Fouta region, many intermediaries switched their support to Sall only in the second round, after having backed President Wade in the first round. In the town of Matam several prominent figures, including the town’s mayor, “left the sinking ship,” and disavowed Wade and the PDS in favor of Sall.111 At the same time, the PDS lost three MPs in the Sine region and several party members in Saint-Louis.112 Not surprisingly, given the intermediaries’ instrumental approach to politics, Mamoune Niasse, the religious leader who flocked to President Wade’s camp in 2007, did not remain loyal to him for very long. By the time of the 2012 election, he decided to support Wade’s bitter rival, Idrissa Seck, commonly called Idy. When asked by journalists why he abandoned Wade, Niasse replied “I have never associated myself with a loser.”113 The argument that intermediaries’ selection of patrons is motivated by material rewards rather than affective ties finds general support beyond the specific illustrations provided thus far. It is a view held by most influential actors: politicians, scholars and journalists. Politicians have a consistent view of intermediaries’ motives. Babacar Gaye, who talked so openly about stealing intermediaries, conceded that the game is stacked in the ruling party’s favor, precisely because intermediaries expect rewards. As he explained, “[I]ntermediaries do not have much consideration for everyone; you have to have the means. Relays prefer people with money, and thus those in power.”114 This is especially true of the marabouts who, in the aggregate, “are always on the side of those in power.”115 Among others, opposition politicians, Mamadou Ly and Ibrahima Sène, also voiced their opinion that intermediaries are usually 109 111 113 115

110 Press conference on July 14, 2005. Author interview, Dakar, December 1, 2006. 112 Sud Quotidien, March 12, 2012. Walfadjri, March 20, 2012. 114 Jeune Afrique, September 2011. Author interview, Dakar, December 1, 2006. In fact, the marabouts have a convenient way of justifying their pro-incumbent support. As Gaye argues, the marabouts often explain their choice to their followers by saying that God already chose the country’s leaders, now they just have to support them. The use of “God’s will” in justifying intermediaries’ support for the incumbent has

Electoral Dynamics in Senegal

119

on the side of incumbents. As did the representative of LD/MPT in Mbour, who explicitly added that religious leaders are always in favor of those in power because of money. The same view is advanced by more dispassionate political observers. Professor Thierno Diop, a sociologist at the Cheikh Anta Diop University in Dakar, said that intermediaries choose politicians who give them the most, and consequently, they almost always end up on the side of those in power.116 Similarly, Professor Adolphe Dansou argues that in Senegal one changes loyalties “to go toward people with resources.”117 A similar interpretation of intermediaries’ behavior can be gleaned from the Senegalese media. A famous investigative journalist, Abdou Latif Coulibaly, describes religious intermediaries as “rentiers, keen on consumer goods” (2006: 239). In a similar vein, in its analysis of President Wade’s political entourage, a popular Dakar daily, Walfadjri, argues that: those who are around him [Wade] are nothing but courtesans who practice their job; they would do it for anyone else; they don’t know how to do anything else; The proof: many of them were already at the service of other masters . . .; they are mercenaries who want to weigh themselves for future haggling.118

The analysis of the behavior of intermediaries shows that it is indeed very different from the paradigms of ethnic politics. As the empirical evidence indicates, intermediaries are driven by material benefits, rather than ethnic considerations. They prefer to choose the highest bidder, very often the incumbent, rather than a coethnic. The actors referred to in this section provide concrete illustrations of how they diverge from the predictions of the primacy of ethnicity or other identities. For example, during the power struggle between a Catholic, Léopold Senghor, and a Muslim, Lamine Guèye, the Muslim dignitaries decided to support the Catholic candidate. Shared religious identity could not have been a factor and most likely, according to historical sources cited earlier, the decision was based on Senghor’s promise to deliver funds for the construction of the Great Mosque in Touba. In general, the behavior of the marabouts, especially the more politically involved and influential Mourides, shows great inconsistency with the

116 117

been widely noted by other scholars and observers and it is not limited to President Wade’s term in office, but extends back to his predecessors. Author interview, Dakar, December 2006. 118 Author interview, Saint-Louis, November 7, 2006. Walfadjri, February 8, 2007.

120

Mobilization Strategies and Electoral Outcomes

instrumental importance of identity.119 As documented, the Mourides worked as intermediaries for the French, then for a Catholic president, followed by a Tijan and only most recently, a Mouride. Rather than showing consistency in supporting a member of their group, they showed consistency in having a preference for the incumbent. Many maraboutic families changed their political allegiance following an alternation in power. For example, in 1988 Serigne Abdoul Ahad Mbacké issued a ndigal for the then president, Abdou Diouf, but in 2001 his son gave a ndigal for President Wade.120 Other actors mentioned here illustrate the same pattern. The village marabout, from Madieyna Diouf’s anecdote, decided to support President Wade over Moustapha Niasse, who is a native of that very region. Similarly, Serigne Mamoune Niasse, a Tijan, shifted his support from a fellow Tijan, Idrissa Seck, to President Wade, a Mouride, and took a position in Wade’s government, before switching back to Seck in the subsequent election. Even intermediaries who share identity with their patron, as for example Modou Kara Mbacké with President Wade, made their decisions based on material incentives. In Kara’s case, the fact that he supported President Diouf, a Tijan, against Wade in the 2000 election before offering his support to President Wade in 2007, shows that shared identity is not driving his behavior.121 Rather than always supporting a coreligionist, he consistently supported the incumbent. The repeated switching of political patrons by various political intermediaries discussed here is inconsistent with the effect of ethnicity on allegiance. Another behavior that is inconsistent with ethnic affinity is the high frequency of split support for different candidates within maraboutic families. The phenomenon of a father and son, or a set of brothers supporting different candidates is relatively common among Senegalese intermediaries. For example, toward the end of PS’s reign, Serigne Sidi Mbacké supported the then opposition leader Wade, but the rest of the family of Serigne Falou Mbacké was won over by the PS.122 Similar splits were noted in 2012 between the marabout Béthio Thioune and his relatives, or between Aïda and Pape Mbodji, a well-known political family in Bambey.123 119

120 121 122 123

For additional examples of prominent intermediaries supporting candidates from other groups, see Diouf (1994) and Beck (2008: 72–73). Walfadjri, April 19, 2001. Kara’s support for Diouf is discussed in Audrain (2004). Sud Quotidien, May 22, 1998. Walfadjri, March 5, 2012; Sud Quotidien, March 22, 2012.

Electoral Outcomes in Senegal

121

One explanation for this behavior is that intermediary families want to “hedge their bets.” If members of the family support different candidates, they are increasing the chances that one of them will be on the winning side. In addition, marabouts sometimes align themselves with different politicians to play out their internal family rivalries and potential succession battles (Behrman 1970: 73, 77, 88). Beck notes that supporting opposition candidates against other members of the family who back the incumbent is often a result of jockeying for status within the brotherhoods (2008: 73). Beck dates this practice all the way to the first mass elections in the 1950s. As she highlights, political campaigns in Senegal River Valley during the 1950s were fueled by an intense rivalry among the political dynasties of toorobe families who varyingly allied with parties led by Lamine Guèye and Léopold Senghor (Beck, 2008: 127). When junior members of a religious family are overlooked by the incumbent, sometimes they offer to support opposition politicians to draw attention to their mobilization capacity and to establish their worth. By working against their family members, intermediaries can also try to renegotiate their internal family position and force their senior relatives to take them more seriously.

electoral outcomes in senegal Consistent with an intermediary-type politics, we do not observe ethnic candidates or ethnic parties in Senegal. As Makhtar Diouf sums up: “[T]he Senegalese landscape has never been traversed by ethnic or religious cleavages, not in the past, not in the present” (1994: 44–45). None of the political parties has (or had) an ethnic or confessional base. Lambert notes that none of Senegal’s political parties “have been credibly accused of favoring the interests of specific ethnic groups” (1998: 598). This is the perception that both Senegalese voters and academics have of the political landscape, but more importantly it is clearly supported by the electoral data. These nonethnic patterns are not a new development in Senegalese politics, but rather a continuation of electoral outcomes since the onset of mass politics. Already in the 1950s, political parties were not confined to ethnic blocs. Even in the absence of individual-level data, aggregate data from early elections highlight a few features incompatible with ethnic politics. During the 1957 elections to the territorial assembly, Léopold Sédar Senghor, a Catholic Serer, gained support for his BPS throughout the country, winning seats in every constituency, where there are no native Serer or Catholics. After independence, President Senghor actually enjoyed the highest levels of support in regions dominated by non-Serer

122

Mobilization Strategies and Electoral Outcomes

Muslims. Similarly, President Diouf won with the highest electoral margins in peripheral regions dominated by non-coethnics (Beck 2008: 2). Various indices measuring association between ethnicity and vote choice using individual-level data from the Afrobarometer, facilitate the comparison of the Senegalese results from recent years with other African countries. Using their index, CVELI, Down and Driessen estimate that only 11 percent of vote choice in Senegal can be predicted by ethnicity, compared with Benin’s 41 percent. Working with the same data, Cheeseman and Ford (2007) calculated an index of ethnic polarization, namely the extent to which support for a given party varies between a country’s ethnic groups, and ethnic diversity, i.e. the range of ethnic groups represented within different parties. Based on this metric, Senegal has one of the lowest polarization levels among African countries (0.13 on a scale between 0 and 1) and high ethnic diversity of parties (0.7) and the authors conclude that ethnicity is not a significant factor in influencing party affiliation (see Table C.4 in Appendix C). Further analysis of individual-level data from several Afrobarometer surveys provides the best evidence that candidates’ and parties’ support is not determined by their identity, be it ethnicity, religion or brotherhood. Respondents’ electoral choices from the 2000 election, recorded in Round 2 of the Afrobarometer survey (2002), show lack of importance of identity factors as predictors of voting. The ethnic composition of Abdoulaye Wade’s winning electorate is very similar to the electorate of his main rivals, Abdou Diouf and Moustapha Niasse. Similarly, his scores among each ethnic group are highly proportional to each group’s size, with no single group dominating his electorate (see Table 4.3). Similarly, candidates’ vote shares among Muslims and Catholics, as well as among the main brotherhoods, appear evenly distributed and proportional to the group size (see Table 4.4). When Wade, a Mouride, ran against Diouf, a Tijan, their levels of support among the two groups were almost indistinguishable. Diouf, a Tijan, got 39 percent of his votes from Tijanis, and 41 percent from Mourides. Wade, a Mouride, got 45 percent of his votes from Mourides and 39 percent from Tijanis. The results of the 2001 legislative elections similarly highlight the lack of discernible ethnic or other identity voting patterns. The three main parties – PS, PDS, and AFP – enjoyed comparable level of support among the country’s Muslims and Catholics, in proportion to their overall scores in that election (see Table C.5 in Appendix C). The same can be said about the breakdown of their support among Mourides and Tijanis or among the main ethnic groups. Most importantly, neither party relied heavily on

Electoral Outcomes in Senegal

123

table 4.3 Ethnic Composition of the Electorates of the Main Presidential Candidates in Senegal

Ethnic group

Candidates’ share of their total vote received from a given group

Groups’ share of the population

A. Wade

A. Diouf

M. Niasse

42% 28% 13% 5% 5%

42% 30% 11% 5% 4%

43% 26% 20% 3% 4%

42% 28% 13% 5% 5%

Wolof Pulaar Serer Mandika Diola

Note: These candidates gathered together 90% of the vote. Respondents were asked, “[W]ho did you vote for in the 2000 presidential election?” Source: Afrobarometer, Round 2 (2003). N = 1147.

table 4.4 Religious and Brotherhood Composition of the Electorates of the Main Presidential Candidates in Senegal

Religious group Muslims Catholics Muslim brotherhood Mouride Tijan

Groups’ share of the population

Candidates’ share of their total vote received from a given group A. Wade

A. Diouf

M. Niasse

95% 4%

93% 5%

91% 9%

93% 5%

40% 39%

41% 39%

45% 39%

37% 49%

Source: Afrobarometer, Round 2 (2003). N = 1147.

support from any given group or coalition of groups (see Table C.6 in Appendix C). Given these numbers the major parties certainly do not fit the definition of an ethnic party. The three subsequent Afrobarometer surveys, conducted in Senegal in 2005, 2008 and 2013, reveal very consistent patterns over time, with candidates’ electorates broadly similar to the general electorate (see Tables C.7, C.8 and C.9 in Appendix C). President Wade managed to increase the size of his electorate since his days in opposition but the ethnic and religious breakdown of his supporters remained remarkably stable, as highlighted by the three surveys. As the next chapter will show, Wade’s levels of support in urban and

124

Mobilization Strategies and Electoral Outcomes

rural areas changed dramatically since he became the incumbent, highlighting that the urban–rural cleavage is much more pronounced in Senegal than ethnic or religious cleavages. Electoral data from the 2012 election, coupled with ethnic composition at the department level, confirm Wade’s cross-ethnic appeal. Wade, a Wolof, certainly did not benefit from any ethnic favoritism from his coethnics. Indeed, the correlation between the share of departments’ population that is Wolof and Wade’s vote share is slightly negative (−0.27).124 In the 2012 election, some observers looked for ethnic favoritism toward Wade’s main adversary, Macky Sall, among the Peul, due to Sall’s mixed Peul (Halpulaar)125 and Serer background. During the campaign, President Wade accused Macky, as he is commonly called, of ethnic bating (faire d’ethnicisme).126 However, the case against Macky Sall as an ethnic candidate is not very convincing for several reasons. First, Sall won by a landslide in the second round of the 2012 election, scoring very high across Senegal. While Sall’s scores were positively correlated with a department’s Peul population (0.39), so were Wade’s scores (0.25), albeit to a lesser degree. Several academics, including Mousta Sall and Amadou Ly, contest the charges of ethnic vote in favor of Sall. Ly points out that Wade won in several predominantly Peul areas, such as Kolda, Medina Yoro Fula, Ranerou, Vélingara and he concludes that “people who talk about ethnic voting are sore losers.”127 Even PDS members in Matam deny that there was ethnic voting for their opponent. Abdoulaye Dramé, PDS’s secretary general in the region, points out that Ranerou, which is the most Halpulaar of all communes in Matam, voted for Wade. Furthermore, he added that many in the area see Sall as Serer and not Halpulaar.128 Indeed, what is remarkable in Senegal is that the four presidents since independence – two Serers, a Wolof and a Peul/Serer have constructed very similar winning electorates, representative of the national population. Importantly, Senegalese inclusive politics cannot be seen as a result of Wolof hegemony. It is simply not the case that smaller ethnic groups yield to the Wolof. A Wolof president has ruled Senegal for only 12 of the 55 years since the country’s independence. What is more striking is that 124

125 126 128

Data on the ethnic composition of departments comes from the Senegalese Census (Recencement de Population du Sénégal, 1988). The most recent census did not provide departments’ ethnic composition. The terms Peul and Halpulaar are frequently used interchangeably in Senegal. 127 Le Soleil, February 14, 2012. Walfadjri, 5989, February 2012. Walfadjri, March 16, 2012.

Electoral Outcomes in Senegal

125

rather than dominating electoral politics, the Wolof are consistently open to backing non-coethnic politicians. Sub-nationally, the only exception to the generally successful crossethnic political integration is Casamance, where, as previously noted, hierarchical ties and local leadership have always been weak. Since 1982, the region, home to the Diola and Balant groups, who constitute about 5 percent of the Senegalese population, has been the site of an ethnic rebellion and a separatist movement, Mouvement des Forces Démocratiques de la Casamance (MFDC).129 Not coincidentally, social structure frequently features in explanations of the insurgency. First, the insurgency is commonly explained as a result of the central government’s marginalization and neglect of Lower Casamance, inhabited largely by the Diola group. Social structure in Lower Casamance had a significant effect on the region’s incorporation into the Senegalese state and subsequent distribution of resources and political power. By some accounts, the area’s marginalization stems from the weakness of its local hierarchies. For example, Boone notes that “state building in Lower Casamance proceeded along very different trajectory” as “the Dakar regime found no rural leaders with whom to broker a stable and secure political alliance” (2003: 94). She further elaborates that the “Diola social structure did not offer Dakar secure footholds, or possibilities for alliances with local elites that could have magnified the authority and influence of the state” (Boone 2003: 94). Consequently, “Dakar never built here a deep-reaching party-state apparatus” (Boone 2003: 97). The main parties struggled to secure electoral bases, including the ruling party, which had “a very weak hold in the region” (Boone 2003: 111, 115). Similarly, Beck observes that Lower Casamance’s “egalitarian society offered no infrastructure for a [clientelist] mode of national integration” (1997: 196). While politicians in Dakar failed to find effective intermediaries in Casamance, the Diola peasants struggled to claim their fair share of resources and power from the central state, having “very feeble possibilities for access to the apparatuses of the state” (Darbon 1988 quoted in

129

For more detail on the ethnic rebellion, see de Jong (2005), Faye (1994) and Lambert (1998). The view of the rebellion as ethnic is widespread. There is some disagreement on whether MFDC is a Diola rebellion or whether it is based on a more encompassing Casamançais identity, of which Diola is the largest component (see Lambert 1998). The vast majority of MFDC members are Diola. Lambert (1998) also adds that the Senegalese state is keen to portray the rebellion as ethnic, making it less acceptable to the Senegalese public, given the general absence of ethnic cleavages in Senegalese politics.

126

Mobilization Strategies and Electoral Outcomes

Boone 2003: 121). The marginalization of Casamance has been further compounded by the region’s geographic isolation, due to its separation from the bulk of the Senegalese territory by the Gambian state. Another narrative, which also focuses on the effect of social structure, postulates that the lack of hierarchy makes the Diola naturally hostile to overarching authority and thus the central state. Many explain Casamançais separatism “in terms of the long history of Diola resistance to hierarchical political systems.”130 Whether or not the Diola are hostile to central authority, their lack of robust local leaders has placed them at a disadvantage vis-à-vis the rest of Senegalese society. The weakness of their social hierarchies made them less appealing electorally to national politicians and contributed to their economic and political marginalization. In this light, the insurgency can be seen as a response to the unsuccessful integration of Lower Casamance within Senegalese political structures. Because of the Diola’s small size, ethnic electoral mobilization would not help them accede to power, or gain more resources.

conclusion The existence of strong local leaders who can act as credible intermediaries between politicians and voters has profound implications for electoral politics. Whenever there are such leaders, they are a valuable asset to political actors, who are eager to use them. In Senegal, where there was a plethora of strong local leaders, politicians relied on them to a great extent to mobilize voters. In Benin, where local leaders were considerably weaker than their Senegalese counterparts, politicians could not use them and they relied on shared ethnic identity to win voters’ support. Intermediaries are not merely an additional layer of electoral mobilization but a truly distinct strategy. As the evidence discussed in this chapter showed, intermediaries have very strong material motives in their political behavior. They stand to gain significant personal rewards, be it money, gifts in kind or political positions, while at the same time they try to acquire goods for their community. Both these goals make them predisposed toward candidates who can deliver benefits. It is more advantageous to support a candidate with resources rather than a coethnic politician who might not be able to provide comparable benefits. Such incentives, which are driving intermediaries’ behavior, imply that intermediaries will frequently try to cooperate with politicians from different 130

See Lambert (1998: 593) for a review.

Conclusion

127

ethnic and religious backgrounds, allowing them to win voters outside their ethnic group. In Senegal, where intermediaries were used extensively, no ethnic voting patterns emerged. Intermediaries frequently delivered blocs of voters to politicians from different ethnic backgrounds, creating cross-ethnic electorates. In contrast, in Benin, where hardly any intermediaries were used and politicians relied on shared ethnic identity to forge links with voters, we find enduring ethnic voting blocs.

5 Intermediaries in Urban and Rural Settings

Do electoral intermediaries work equally well in all settings? Chapter 3 showed how the existence of important local leaders varies across countries, as well as different regions in a single state, whereas Chapter 4 analyzed how this variation presented politicians in Senegal and Benin with different mobilization options. This chapter addresses another type of variation in the strength of hierarchical ties, namely the difference between urban and rural milieus. The role of local leaders varies systematically between the urban and rural setting. Urban areas present challenges to the power of authority figures; ties between leaders and followers are weaker. Because these ties are the very basis of notables’ ability to act as intermediaries, their strength affects intermediaries’ performance. Leaders with weaker bonds with their followers are less viable as intermediaries as they are less likely to influence voters. This chapter provides evidence for this argument. First, it shows that on average local leaders play a greater role in rural than urban areas. Both survey research and qualitative studies of specific authority figures, be they religious or traditional, suggest that urbanites are less dependent on, and deferential to, their leaders. This systematic difference makes rural intermediaries more effective. This chapter presents two types of evidence for this varying effectiveness. Most directly, in the Afrobarometer and Gercop surveys rural voters report being more likely to be influenced by religious leaders in their voting behavior. Second, both scholars and politicians assess urban intermediaries as more influential. Politicians not only perceive differences between the two milieus, but they also act on it. They are less likely to employ intermediaries in urban areas. 128

The Role of Local Leaders in Urban and Rural Areas

129

The chapter then addresses an additional observable implication of the different role of leaders in the urban and rural setting. It combines the finding that intermediaries are more effective in rural areas with the incumbency advantage of an average African president. Based on these two tendencies, we should expect that all else equal the incumbent should perform better in rural rather than urban areas. This chapter provides robust evidence for this, combining electoral data from seven elections in Senegal, as well as data from other African countries. The case of Senegal is particularly suited for studying the variation in the strength of hierarchical ties between urban and rural areas since most regions of Senegal, with the notable exception of Casamance, have important religious and/or traditional elites. The variation in the repertoire of authority figures throughout Senegal helps us isolate the effects of urban and rural milieus, which play a role across different sets of hierarchical ties. This chapter proceeds as follows: The first section discusses the different role of local leaders in urban and rural areas. The following section shows that intermediaries are more influential in the countryside. The chapter then examines the electoral implications of this tendency. The remaining sections discuss the external validity of the findings and address alternative explanations.

the role of local leaders in urban and rural areas The nature of relationships between local leaders and their dependents is markedly different in urban and rural settings. Ties between leaders and their followers tend to be weaker in urban settings in comparison to rural areas. Whereas most social relationships in the countryside are strong, stable and durable, urban ties are more superficial and temporal. Rural patron–client ties, in contrast to urban ones, tend to be based to a larger extent on deferential relationships, which enhance the capacity of local brokers to mobilize political support (Beck 2008: 12). Urbanization, in turn, has the tendency to weaken this traditional type of ties because it reduces the dependence of followers on their leaders. While local leaders can control both economic and social activities of a village, they cannot manage to the same extent the complicated urban existence of their followers. The relationship between the followers’ dependence and their leaders’ influence means that the weaker the leaders’ control, the less likely they are to serve as effective intermediaries.

130

Intermediaries in Urban and Rural Settings

Urban settings undermine the role of authority figures, both religious and traditional. Several prominent scholars of Senegalese society note that urbanization weakens the relationship between Muslim religious leaders, the marabouts, and their followers. Cruise O’Brien (1971) noted already a few decades ago that urbanization would result in declining marabout power over urban disciples. He observed that “the cumulative effect of urban contact has been to weaken the talibes’ [followers’] allegiance” to their leaders (1971: 238). Similarly, Villalón noted the “weakening in the practice of submission that follows migration to an urban area” (1995: 119). There are structural explanations of this tendency, as religious organization differs between urban areas and the countryside. Cruise O’Brien elucidated this difference with his case study of the Mouride order. First, most shaykhs, religious leaders with followers, actually reside in rural areas (Cruise O’Brien 1971: 238). Thus, migration to the urban area entails a physical separation of the disciple from his marabout, affecting the frequency of their interactions. Second, the most cohesive units of religious organizations, the collective maraboutic farms known as daaras, are obviously possible only in the rural setting. Religious organization in the city developed its own form, the daaira, an association of followers of the same marabout but, as Cruise O’Brien argues, urban daairas “cannot compare favorably as an instrument of socialization by the religious leaders” and they are often more “active on paper than in reality” (1971: 259). He concludes that urban residence is less conducive to the preservation of the marabout-disciple ties than life in isolated villages, where the marabouts can dictate the terms of social existence (1971: 99). An easily quantifiable reflection of the different deference to religious leaders in the two settings is the size of monetary contributions made by the followers to their marabouts. Urban taalibes reputedly give a much lower proportion of their income than do rural ones (Cruise O’Brien 1971: 248). The same structural factors affect the role of traditional leaders. To an even greater extent than religious leaders, traditional leaders reside in rural areas. Separated by physical distance, they cannot exert the same influence over their (former) dependents, especially secondgeneration urbanites. For example, Beck (2008) notes that the Tukulor from the Fouta region often stop paying their customary tithes to the traditional toorobe elite. She concludes that the influence of the toorobe is tightly linked to physical presence on the land under their control. As people move to urban areas, their deference to traditional authority

The Role of Local Leaders in Urban and Rural Areas

131

wanes. Although this discussion focuses on the Senegalese case, the same tendency has been noted about traditional authority elsewhere in Africa. In Botswana, for instance, the paramount chief of BakgatlaBa-Kgfela, Kgosi Linchwe II, argues that the power and prestige of the chieftaincy is being eroded by migration to urban areas (Herbst 2000: 179). The differing role of local leaders in urban and rural areas is reflected in survey research. Question 32 from the 2005 Afrobarometer survey, which asks respondents how often they turn for help to different leaders, provides a good proxy for the level of dependence on local leaders in the two respective settings. The data show that, compared to urban residents, a much larger proportion of rural respondents turn to local leaders for help. While 27 percent of Senegalese respondents in the countryside report relying on religious leaders, only 18 percent of urbanites do so. This difference is even more marked in the case of traditional leaders. Only 5 percent of urban dwellers are dependent on traditional leaders in this way, but this number is over four times higher in the countryside (23 percent). Such differences in the strength of traditional and religious leaders between urban and rural areas are by no means limited to Senegal. Religious and traditional leaders play a bigger role in the lives of rural than urban respondents in other countries with strong local leaders. In Mali, around twice as many rural respondents are reported relying on a traditional or religious leader frequently for help than their urban counterparts.1 The relationship between urbanization and weaker ties with local leaders also seems to be present even in places such as Benin where hierarchical ties are generally weak.2 Other Afrobarometer questions gauging the role of local leaders reinforce these patterns. Rural residents are much more likely to exhibit high levels of trust in local leaders than their urban counterparts. Among 13 countries that participated in Round 4 of the Afrobarometer survey, the average number of rural residents who trust local leaders a lot was 12 percentage points higher than among urbanites (45 percent and

1

2

The reported numbers are 19 percent and 9 percent, respectively, for religious leaders and 16 percent and 9 percent, respectively, for traditional leaders. Question 32, Afrobarometer, Round 3. The reported numbers are 6 percent and 3 percent, respectively, for religious leaders and 7 percent and 4 percent, respectively, for traditional leaders. Question 32, Afrobarometer, Round 3.

132

Intermediaries in Urban and Rural Settings

33 percent, respectively).3 A larger share of rural respondents in those 13 countries reports that traditional leaders have a great deal of influence in their community (27 percent and 21 percent, respectively).4

local leaders as intermediaries This reduced social role has a clear political implication: local leaders acting as intermediaries should be less successful in the urban setting. Intermediaries leverage their social power to influence how people vote, thus the more social power they have, the more political influence they can wield. Successful intermediaries transform voting from an individual act in which voters make decisions independently to collective voting, or bloc voting, when voters follow the advice of their leader, such as a village chief or a marabout.5 The ability to influence voters and produce collective voting rests on intermediaries’ clout. Survey evidence shows that rural voters report being more likely to be influenced by traditional and religious leaders in their voting behavior. Politicians whom I interviewed and scholars of Senegalese politics consistently believe that intermediaries are less capable of influencing urban than rural voters. Several observers of Senegalese politics note that local leaders in the city play a lesser role as intermediaries (e.g. Samson 2000). People’s dependence on these authority figures is less extensive, and consequently, the brokers’ ability to affect voters is weakened. Alioune Badara Diop argues that the process of secularization observed in Dakar reverses the traditional dynamic of voting orders issued by intermediaries (2002: 40). Similarly, Mame Less Camara observed that whereas marabouts are social mediators in the countryside, in the city their role is reduced as they cannot solve the most important problems of their followers.6 He points out that urban dwellers and secondgeneration urbanites, in particular, act as individuals when voting and they have lesser tendency to subject themselves to their marabout or traditional chief.

3

4 5

6

Based on Question 49i (“How much do you trust traditional leaders?”) from 13 countries included in the Round 4 Afrobarometer survey (2008). Each country has a sample size of approximately 1,200 respondents. Question 65, Round 4 Afrobarometer survey (2008). Gottlieb (2014) finds empirical evidence of bloc voting in rural Senegal. In all the villages that she sampled, the majority of inhabitants of a given village converged on a single candidate. Author interview, Dakar, October 17, 2006.

Local Leaders as Intermediaries

133

This alleged link between differences in social dependence in urban and rural areas and the success of intermediaries finds support in survey data. The Gercop survey provides concrete evidence that urban voters are indeed influenced to a lesser extent by their local leaders. Respondents between the age of 18 and 34 in the departments of Thiès and Diourbel were asked whether they were likely to follow voting orders from their religious leaders.7 The difference between urban and rural respondents was fairly substantial. In comparison to only 29 percent of urban dwellers, 40 percent of rural inhabitants admitted being influenced by religious leaders. The Afrobarometer survey found the same pattern. In comparison to only 9 percent of urban interviewees, 18 percent of rural respondents in Senegal say that they are influenced a lot in their electoral choices by religious leaders.8 A systematically different assessment of the strength of ties in urban and rural areas also emerges from the politicians whom I interviewed. Politicians from all sides of the political spectrum notice the difference in social structure and the effectiveness of local intermediaries in the two milieus. They all agree that intermediaries are used extensively for the purposes of electoral mobilization, but they all believe that intermediaries are weaker, and thus less viable as an electoral strategy, in the urban setting. Moustapha Diakhaté from the ruling PDS confirms that his party is relying heavily on electoral intermediaries (porteurs de voix) but he acknowledges that this strategy is much harder in the capital, Dakar.9 As he explains, in Dakar, it is not about “big personalities”; politics is individualized. From a politician’s point of view, you cannot submit the inhabitants of Dakar “to be part of your fief.” He elaborates that in a classic fief relations are based on local solidarity, but in the city life is individualized. As urban dwellers often explain, in the city, “we live as individuals, everyone for himself.” Babacar Gaye, a MP from the ruling party and the third vice president of the National Assembly, also paints a clear difference between urban and rural voters. For him, urbanites are

7

8

9

“Etude sur le comportement électoral dans les régions de Thiès et Diourbel,” May 1999. The original French wording was “sensible à la consigne de chefs religieux.” Unfortunately, this question was not asked in any other country. In the Senegalese case the question did not ask about the possibility of being influenced by traditional leaders. It is also worth pointing out that the question most likely estimates the lower bound of maraboutic influence on voters. Given civil society groups’ vigorous challenge to marabouts’ vote peddling, voting orders have become a shameful act. Thus many respondents might be reluctant to admit to following them. Author interview, Dakar, January 31, 2007.

134

Intermediaries in Urban and Rural Settings

more individualistic.10 In contrast, in rural areas there is more collective voting. Not only do entire families vote together, but often whole villages. As Gaye explains, in the rural setting, “vote carriers” play a bigger role and voters are deeply affected by them. Since people vote as a bloc, it is enough for a politician to convince a village chief, a prominent religious dignitary, or any other person who plays a key role in the community, to gain the votes of the whole village. As he sums up, in the rural world, people follow their leaders, and leaders follow their interests. Opa Diallo, an opposition MP from LD/MPT, also argues that it is easier to acquire whole blocs of voters in the countryside as people have higher tendency to vote together. As he argues, from the point of view of brokers, rural voters are more “stable” and they rarely “betray.” This point is important in an electoral setting with a secret ballot. Although there are attempts during elections to monitor how people voted, they are far from perfect.11 Most politicians admit that such efforts are not reliable enough and that they have to rely instead on social control. This type of control is also found to be effective in other elections in which clientelism plays an important role. For example, Stokes (2005) observed that Peronist brokers in Argentina relied on their personal interactions with people to assess how they voted. Signs such as avoiding eye contact were deemed as possible evidence that people did not vote as expected. Ousmane Tanor Dieng, presidential candidate of the former ruling party, PS, also alludes to the role of social ties in affecting electoral patterns. According to Dieng, Muslim religious leaders, one of the most prominent types of electoral intermediaries in the Senegalese context, play a much bigger role in the countryside, whereas the limitations of their influence in the city are apparent.12 Similarly, Baidy Sall, from the opposition party AJ/PADS, argues that in the city people are freer and more emancipated, whereas in the countryside people are much more “structured.”13 An implication of this structure is that during a political 10 11

12 13

Author interview, Dakar, December 12, 2006. During presidential elections voters are given separate ballots with the name and picture of each candidate. They are supposed to cast the ballot with their preferred candidate and throw away all the rest in the bin in the voting booth. Brokers sometimes ask voters to bring back all the unused ballots, so they can verify that the voter voted as promised. Yet, other voters’ unused ballots could be taken from the bin in the voting booth and voters could thus show brokers other people’s unused ballots. The best evidence that these ballots were not well guarded is that for several days after the 2007 election one could find unused ballots among trash in the streets. Author interview, Dakar, April 14, 2007. Author interview, Dakar, February 14, 2007.

Electoral Implications

135

campaign it is sufficient to talk to a rural leader and convince him, rather than search individual votes. As another politician, Samir Abourizk, put it, “you spend two hours on the floor with a chief, give him 500,000 francs CFA (1000 USD) and he will sell his soul to the devil.”14 In sum, both political analysts and politicians believe that intermediaries play a much greater role in the rural setting, making influencing voters’ behavior easier in the countryside. Intermediaries’ operations in rural areas are also facilitated by smaller voting bureaus than in urban centers.15 These small polling stations, typically one per village, provide relatively fine-grained data on voting behavior at the local level, which are useful for post-electoral payoffs or sanctioning.16

electoral implications The different role of intermediaries in urban and rural areas has in turn electoral implications. Social structure in rural areas, with its tighter ties between leaders and followers, makes rural voters more easily mobilizable through intermediaries than their urban counterparts. Intermediaries, however, are not equally likely to support all candidates. Incumbents, who typically have unrivaled resources at their disposal, are better placed to secure the support of intermediaries. There are noticeable discrepancies in funds available to Senegalese incumbents and their opponents (e.g. Patterson 2002). For example, during the 2012 election, journalists noted that while President Wade was giving out millions, most of his adversaries “had to scramble.”17 They concluded that Wade’s campaign is “without doubts the richest” and they noticed a “visible lack of resources” among some of the main contenders, including Moustapha Niasse and Ousmane Tanor Dieng.18

14 15

16

17

Author interview, Rufisque, October 16, 2006. Author’s interview with Babacar Kanté, Vice President of the Constitutional Council. Dakar, Senegal, February 2, 2007. Although individual vote buying is difficult to enforce, politicians can observe the collective behavior of voters in a given polling station because results are publicly displayed for each station. Indeed, politicians do take advantage of their ability to monitor the behavior of an entire village; they frequently offer rewards conditional on the voting behavior of the village (Author’s interview with Madeyna Diouf, Dakar, April 2007). One indicator of the difficulty in bribing individual voters to vote a certain way was the fact that President Wade’s camp during both the 2007 and 2012 elections frequently bought and destroyed voting cards from urban voters, thus preventing them from casting a ballot for the opposition, rather than their actual votes. 18 Sud Quotidien, February 22, 2012. Sud Quotidien, February 22, 2012.

136

Intermediaries in Urban and Rural Settings

Because of this asymmetry of resources, clientelism overwhelmingly favors the incumbent, who tends to have the most money at his disposal. Indeed, as Chapter 4 showed, Senegalese intermediaries have always had a preference for incumbent presidents. Over the course of Senegalese electoral politics, the most-valued intermediaries, the marabouts, have supported overwhelmingly the incumbent. Other intermediaries, such as members of the toorodo aristocratic lineages in the Fouta, also tend to side with the incumbent. A local intermediary in the department of Vélingara, Abdoulaye Bâ, illustrates this point. When asked how he determines who his dependents should support, he simply answers “we always voted for the party in power.”19 If intermediaries are more likely to drum up support for the incumbent, and if mobilization through intermediaries is much more widespread in rural than urban areas, then we should expect to see higher levels of support for the incumbent in the countryside than in urban areas, all else equal. The data show that there is in fact a robust and timeconsistent pattern of higher support for the incumbent in rural areas, over several decades and different incumbents. The difference between urban and rural voters in their willingness to vote for the incumbent is evident in Senegalese politics. Throughout the rule of Parti Socialiste, the ruling party between 1960 and 2000, major urban areas, and Dakar in particular, always voted least eagerly for the incumbent. During all presidential elections since the reintroduction of multiparty politics in 1978, the incumbent presidents, Léopold Sédar Senghor and then Abdou Diouf, always underperformed in Dakar, receiving on average 14 percentage points fewer than in the rest of the country. PS’s incumbents’ score in the capital was 5–20 points lower in every single election (see Table 5.1). Conversely, the main opposition candidate, Abdoulaye Wade, always performed better in Dakar than in the rest of the country. For example, in the first round of the 2000 election, Wade scored 49 percent in the capital and an average of only 25 percent in the remaining departments.20 The incumbent party also underperformed in Dakar in legislative elections. In the 1993 legislative election, the PS’s score in greater Dakar was roughly half of its average result in other parts of the country (33 percent and 61 percent, respectively). Likewise, in the 1998 legislative election, the PS scored around 29 percent of the vote in the two

19

Le Quotidien, March 8, 2007.

20

Cour Constitutionnelle.

Electoral Implications

137

table 5.1 Incumbent Vote Share in Presidential Elections in Senegal Presidential election year 1978 1983 1988 1993 2000 (1st round) 2000 (2nd round) 2007 2012 (1st round) 2012 (2nd round)

Vote share in Dakar

National vote share

70.4% 78.2% 58.2% 41.2% 21.0% 23.4% 52.0% 22.0% 23.5%

82.3% 83.6% 73.2% 58.4% 41.3% 41.5% 55.0% 34.5% 34.2%

Sources: Cour Constitutionnelle, Dakar, Senegal; Agence de Presse Sénégalaise; and Beck (2008: 235). 2000 data includes Greater Dakar.

departments that cover the capital, whereas its mean score in the rest of the country was 53 percent. This is not merely a difference between Dakar and the rest of the country, but between urban and rural areas in general. This trend holds when we use a finer, continuous measure, a district’s urbanization level, namely, the percentage of a district’s population living in urban areas. These levels range from 100 percent in Dakar to 5 percent in the district of Nioro, with a lot of variation in between.21 In every election, urbanization levels were negatively correlated with incumbent score (−0.56 for the 1998 election, −0.75 in 2000, −0.25 in 2007 and −0.52 in 2012), demonstrating that the more rural the district, the higher the incumbent score on average. This relationship also holds at the individual level. Based on data from the Afrobarometer survey conducted in Senegal in 2003, which recorded how people voted in 2000 along with their place of residence,22 it is evident that rural voters were more likely to support the incumbent. Twenty-four percent of rural respondents stated that they voted for the incumbent, President Diouf, compared to only 14 percent of urban respondents. Place of residence is a significant predictor of support for the incumbent even when we control for respondents’ various social

21 22

Projection de population du Sénégal, Recensement de 2002. Afrobarometer coded respondents’ place of residence as a dichotomous variable: “urban” or “rural.”

138

Intermediaries in Urban and Rural Settings

characteristics, such as age, gender, education and brotherhood affiliation (see Table D.1 in Appendix D). It is important to note that these trends say something specific about the urban and rural environment, but not about individual parties or candidates. After 40 years of one-party hegemony, Senegal experienced an alternation in power when Abdoulaye Wade won the 2000 presidential election, beating the incumbent president, Abdou Diouf, in the second round. As indicated earlier, Wade rode to power on an urban vote, garnering twice as many votes in the capital as in the rest of the country. Yet, during his first term in office Wade was able to increase substantially his hitherto limited support in rural areas. By 2007, when running for reelection, he was actually scoring higher outside the capital (57 percent to 54 percent). In fact, some of the least urbanized departments in Senegal, with urbanization levels under 20 percent, made the most significant switch from voting against challenger Wade in 2000 to strongly supporting President Wade in 2007. For example, in the 2000 election the largely rural departments of Linguère, Matam and Podor cast over 70 percent of the vote for the incumbent, Abdou Diouf, and gave less than 10 percent of the vote to challenger Wade. Yet, in the 2007 election, when Wade was running as an incumbent, he received over 60 percent of the vote in these departments, which amounted to a 50 percentage point gain since he became the incumbent. Indeed, the ranks of Wade’s 2007 supporters were full of former Diouf voters. Together with two research assistants we questioned close to 400 voters in four areas of Senegal about their electoral preferences shortly after the 2007 election. Among those who were old enough to vote in both 2000 and 2007, 33 declared having voted for Diouf in 2000. Yet, only 13 people from that group voted for Diouf’s anointed successor from Parti Socialiste, Ousmane Tanor Dieng; 16 others switched support from the former incumbent to President Wade. In other words, more people left for the new incumbent than remained loyal to the party that they previously supported. The longer President Wade was in office, the more noticeable was the difference between his levels of support in urban and rural areas. During the first round of the 2012 presidential election, Wade’s vote share in Dakar was 25.3 percent, compared with his national score of 34.5 percent. Likewise, in the second round, Wade polled 11 percentage points lower in Dakar than nationally (23.5 percent and 34.2 percent, respectively). Wade eventually lost the election to Macky Sall in the second round by a wide margin nationally, but the only parts of the country in which the majority of voters supported Wade over his challenger were some of the most rural

Electoral Implications

139

areas, such as the departments of Sédhiou, Kédougou, Ranerou and Bignona. Out of 30 largely rural departments, with urbanization levels below 50 percent, 24 voted for Wade at a higher margin than his national score and only 6 voted below average.23 Between 2000 and 2012, Wade’s electoral base underwent a dramatic change. It is striking that while Wade was a challenger, his score was positively correlated with urbanization level, but once he was an incumbent, this relationship became negative. As a challenger, Wade relied disproportionately on the support of the capital; the longer he stayed in office, the more rural became his electorate. In 2000, almost a third (32.3 percent) of Wade’s vote came from the capital,24 which encompasses approximately 23 percent of registered voters.25 By 2007, only a quarter (24.3 percent) of Wade’s support was found in the capital, and by 2012 Dakar contributed merely 17.5 percent of the president’s total vote.26 Thus although Wade in 2012 had a lower score in rural areas than in 2007, the disparity between his support in urban and rural areas grew larger and the share of the votes coming from rural voters, as a percentage of his electorate, increased. These different levels of support for the incumbent and the opposition in urban and rural areas, evidenced by the electoral data, were palpable during the campaign season. Whereas in 2000 challenger Wade staged in Dakar regular mass rallies (les marches bleus), full of young people, in 2007 he was having a hard time turning out urban supporters. Instead, Wade’s campaign events were full of people bussed in from the countryside, a practice commonly called by the Senegalese “la politique de ndiaga ndiaye,” named after the colorful busses used to bring outside supporters. Wade also largely avoided visiting the campus of Cheikh Anta Diop University, the country’s leading institution, where he was a regular visitor as an opposition candidate. These urban–rural patterns are in fact deeply engrained in political actors’ analysis of their electoral prospects. Politicians view them as a recurring tendency: Regardless of who is in power, incumbents can win over rural voters more easily but they struggle with urbanites. During the 2007 election, politicians in the incumbent camp voiced their misgivings about their electoral chances in the urban milieu. Mbaye-

23

24 26

Agence Nationale de la Statistique et de la Démographie (ANSD), Agence de Presse Sénégalaise (APS). Those six departments were Fatick, Nioro, Mbour, Matam, Louga and Tivaouane, with the last three very close to the national score. This statistic includes the suburb of Pikine. 25 Cour Constitutionnelle. Cour Constitutionnelle; APS.

140

Intermediaries in Urban and Rural Settings

Jacques Diop, the president of the Conseil de la République, admits that campaigning in Dakar is much more difficult for the incumbent.27 Diop, who has been in politics since the pre-independence era, clarifies that this is a historical tendency. For any incumbent since independence, winning in big cities, especially Dakar, has always been difficult. As a close associate of President Wade, and the fourth person in command in Senegal, Diop admits that throughout the presidential campaign Wade was worried about winning in Dakar and its suburbs. Similarly, on the eve of the 2007 election, Moustapha Diakhaté had obvious doubts about the president’s prospects in the city.28 It is worth stressing that Diakhaté always talked about Dakar as difficult for “an incumbent,” implying that this tendency is largely apolitical. It appears that any incumbent would run into problems, confronted with the demanding Dakar electorate. Babacar Gaye, who was heavily involved in Wade’s re-election campaign, also expressed fears about Dakar, several months before the poll.29 As he put it, the incumbent’s camp is afraid that the city will “sanction them.” At the same time, he viewed Dakar as an easier target for the opposition. In fact, from the opposition’s side, there emerges a much more sanguine view of electoral prospects among urban voters. Baidy Sall (AJ/Pads) places great hopes in urban districts.30 Similarly, Ousmane Tanor Dieng (PS) argues that cities are easier for the opposition and that opposition always starts in the city.31

beyond senegal The differences between urban and rural areas in the levels of support for the incumbent are by no means confined to Senegal. In Mali, we find the same pattern. The previous incumbent, Amadou Toumani Touré, enjoyed much larger electoral margins in rural areas. During his 2007 reelection bid, Touré, or ATT, as he is commonly called, scored 53 percent in the district of Bamako, the capital, compared to his national result of 71 percent.32 The result was the exact opposite for his closest rival, Ibrahim Boubacar Keïta, known as IBK, who received 38 percent in the 27 28 29 30 31 32

Author interview, Rufisque, March 1, 2007. Author interview, Dakar, January 31, 2007. Author interview, Dakar, December 12, 2006. Author interview, Dakar, February 14, 2007. Author interview, Dakar, April 14, 2007. Rapport Général, Cour Constitutionnelle, Bamako, Mali (2007). I thank Jessica Gottlieb for generously sharing these data with me.

Beyond Senegal

141

capital, twice as high as his nation-wide score of 19 percent. The same pattern can be found in neighboring Burkina Faso. During the 2005 presidential election the then incumbent, Blaise Campaoré, scored 71.6 percent in the capital, compared with his national average of 80.4 percent.33 The opposite trend was true for the main opposition candidate, Bénéwendé Sankara, whose result in the capital was almost three times as high as his national average, 14.5 percent and 4.9 percent, respectively. Scholars noted similar patterns in elections across the continent. Jeffries (1998) shows that since the return to democracy in Ghana, both in the 1992 and 1996 elections the incumbent, Jerry Rawlings, performed better in rural areas. In Botswana, urban voters supported the ruling Botswana Democratic Party (BDP) at much lower levels than their rural counterparts, contributing only 13 percent of BDP’s total vote in the 1989 election, despite constituting over 25 percent of the population (Parson 1992: 68–71). At the same time, opposition garnered around two-thirds of their votes in urban areas (Molutsi 1998: 370). Similarly, observers of Zimbabwean politics highlight President Mugabe’s strength in rural areas and the opposition’s higher scores in the city (Maphosa 2005). Afrobarometer survey data indicate that 35.4 percent of rural Zimbabwean respondents describe themselves as close to the incumbent party, but the rate is only 21.3 percent among city dwellers.34 Conroy-Krutz (2007) finds a negative relationship between urban areas and incumbents’ scores with cross-sectional data for 22 African countries.35 Some of these studies also provide anecdotal evidence specifically supporting the notion that variation in clientelist linkages is the mechanism behind urban and rural differences in the level of support for the incumbent. Writing about Botswana Parson notes that “[I]n rural areas, patron-client relations continued to be important underpinnings for a favorable election outcome for the [ruling] BDP” (1992: 84). Yet, the BDP could not use the same strategy in urban areas, as patron–client economic relations and traditional political ties could not solve all of voters’ problems to the same extent as in rural areas (Parson 1992: 76).

33

34

Data obtained from africanelections.tripod.com. The publicly available results of the 2010 presidential election are not disaggregated by region. 35 Afrobarometer, Round 2, 2003. See also Harding (2010) on this topic.

142

Intermediaries in Urban and Rural Settings

alternative explanations While Chapter 1 already discussed the most plausible alternative explanations of African electoral outcomes, let us consider a few other possible predictors of voting patterns, particularly pertinent to the urban–rural cleavage. As I showed in Chapter 1, the Senegalese electoral patterns, with the striking difference between urban and rural areas, cannot be ascribed to factors such as ethnicity or religion. But could these patterns be simply a reflection of different levels of voters’ education in the two settings? Urban voters are undoubtedly better educated on average than people in the countryside and there are reasons to suspect that education plays a role in how people vote. In fact, politicians often describe urban voters as “brighter,” “more awakened” or more “cultivated.”36 Yet, voters’ place of residence plays a role independent of their education. Rural voters have a higher tendency to vote for the incumbent, even if we take their level of education into consideration. Data from the Afrobarometer survey indicate that among voters with the same education level, urbanites vote for the incumbent at a lower rate. Thus, the higher propensity in the countryside to vote for the incumbent is not an artifact of different education levels of urban and rural voters. Nor is the incumbent’s greater success in rural areas a result of differential turnout rates, as there is very little variation in the level of voter participation between the metropolitan area and the rest of the country.37 The higher support for the incumbent in rural areas cannot be ascribed to government policies either. In fact, the existing electoral outcomes seem to be in many ways counterintuitive, given the well-documented urban bias of most African governments. As Lipton (1977) first pointed out, urban bias exists unambiguously whenever an outcome, such as an endowment of schools or the setting of prices, persistently favors urban rather than rural people, and thereby harms income distribution. In his seminal book, Markets and States in Tropical Africa, Robert Bates (1981) elucidated the policies through which successive African governments disadvantaged their rural constituents to bring disproportionate benefits to the cities. Among other tactics, state marketing boards for cash crops and foodstuffs paid producers well below world market prices. For example, in Senegal the 36

37

Author’s interview with Ousmane Tanor Dieng, Dakar, April 14, 2007, and Moustapha Diakhaté, Dakar, January 31, 2007. The turnout rates were 71.6 percent and 70.3 percent, respectively, in 2007; 65.7 percent and 62.5 percent in the first round in 2000; 61.6 percent and 61.9 percent in the second round in 2000 (Cour Constitutionnelle).

Alternative Explanations

143

government controlled the marketing of its main export crop, the groundnut. By the late 1960s, the Senegalese government reduced the official price of groundnuts by 15 percent and with a system of delayed payments to farmers, managed to decrease the actual price even further (Bates 1981: 85). The government then diverted groundnut earnings to “feed the bureaucracy, while the producers went hungry” (Fatton 1987). Because African farmers were legally obliged to sell their products to the marketing boards, short of resorting to the black market, they were effectively price takers at the mercy of their government. The key effect of this policy was a major transfer of resources from rural producers to urban consumers. Although structural adjustment reforms alleviated some of these harmful government policies, they have not fully reversed the fate of the countryside. Bezemer and Headey (2008) argue that systematic bias against the rural economy still persists in most parts of Africa. Moreover, many rural areas in Africa receive only a small share of public expenditures, and are poorly endowed with schools and roads, further reducing the standard of living well below that found in an urban environment (Lipton 1993). Sahn and Stifel (2003) pointed out a continued large urban–rural gap in the delivery of public services and they saw no evidence that this disparity is decreasing. In the case of Senegal, many villagers complain that too much money is spent on infrastructure in the capital. President Wade seemed preoccupied with planning flagship urban infrastructure projects, including a new airport outside Dakar, a four-star hotel and a conference center, the embellishment of the Corniche, Dakar’s main coastal road, and future construction of a new capital, dubbed “Dubai on the Atlantic.” At the same time, only a small part of the country’s scarce resources gets channeled into rural areas: Although farmers and their dependents constitute 60 percent of the Senegalese population, this sector receives only 10 percent of public investments.38 Many indicators show that the standard of living and the quality of available infrastructure are considerably lower in the countryside than in the city. For example, although 68 percent of Senegalese urban households have running water inside the house, only 11 percent of rural households do.39 Only 38 percent of people in the countryside have access to safe drinking water, compared to 79 percent of urban dwellers.40 The gap in the availability of electricity between urban and rural households is equally glaring: 76 percent and 10 percent, respectively.41 Rural areas in Senegal are also disadvantaged in terms of 38

ANSD.

39

ANSD.

40

Social Indicators of Development.

41

ANSD.

Intermediaries in Urban and Rural Settings

144

access to essential medical services and they suffer from higher poverty rates (Harsch 2003). Admittedly, urban existence in Senegal is by no means easy. Urban life creates a distinct set of problems, such as higher cost of living and greater social inequality. Urban voters’ access to important services is imperfect. Although urban areas have better roads, they also have persistent traffic jams. Many urban areas in Senegal experience frequent and prolonged power outages, which cause much frustration among urbanites. Yet, despite the significant inconveniences of urban life, urban dwellers have consistently received more state funding than their rural counterparts. Policies harmful to rural areas are a common thread in Senegalese postcolonial history; they do not tend to vary too much, regardless of who is in power. Under Senegal’s first president, Léopold Sédar Senghor, agriculture was in crisis, which led to poverty and food insecurity throughout the rural areas (Harsch 2003). The state of the rural sector further suffered under Senghor’s successor, Abdou Diouf, when significant agricultural services, such as the provision of fertilizer, were reduced (Harsch 2003). After being ruled continuously by the PS for the first 40 years of independence, many Senegalese had high hopes for their new president, Abdoulaye Wade, and his PDS. Yet, following his arrival in office in 2000, the pace of change in the countryside remained disappointing. Instead, Senegal witnessed a continuous rural exodus and a staggering growth in illegal migration to Europe. President Wade should be credited with the construction of some new roads and health clinics in rural areas, but the majority of the infrastructure projects were nonetheless concentrated in the capital. Rural areas received a smaller share of public spending than their share of population would merit.42 Crucially, Wade failed to make any significant improvement to agriculture, which is the most important pillar of economic activity in the countryside. As one commentator put it, Wade’s plan for the development of agriculture “is not a serious response to the serious problems which Senegalese farmers experience.”43 That the state of the rural sector is dire became obvious when tens of thousands of farmers from across the country marched on the capital for the first time in Senegalese history to demand policies that would end their status as “second-class citizens” (Harsch 2003: 14). Given that rural areas are more disadvantaged by government policies than urban areas, rural voters do not have a good policy reason to support

42

ANSD.

43

Jeune Afrique, February 4–10, 2007.

Conclusion

145

the incumbent more eagerly than their urban counterparts. In contrast, looking at the social organization of the two respective milieus, and their implications for mobilization strategies using electoral intermediaries, allows us to understand why incumbents perform better electorally in the countryside.

conclusion The different social structure of urban and rural areas affects the role of local leaders. The urban environment loosens ties between leaders and their followers. All else equal, urbanites tend to be less dependent on both traditional and religious leaders. Dependency is the foundation of intermediaries’ electoral power. The consequence of the systematically different levels of social power is that the role played by intermediaries is bound to differ between the rural and urban milieus. Rural intermediaries are much more likely to be able to influence voters, and since it is a wellknown tendency, politicians are less likely to rely on intermediaries in the urban setting. This chapter provided support for this argument. It showed that scholars have characterized rural ties between dependents and their leaders as much tighter than their urban equivalents. This analysis is further supported by survey data which show that inhabitants of rural areas report much higher levels of dependence on their leaders. This difference is then reflected in the role of intermediaries in the two settings. Both qualitative and quantitative evidence suggests that intermediaries are less likely to be able to affect voters’ behavior in the urban environment. Politicians across the political spectrum view urban voters as less likely to be influenced by intermediaries and they report being less inclined to pursue mobilization based on intermediaries in the urban setting. The perception of difference is thus backed up by concrete actions. While this chapter established that intermediaries are more effective in rural areas, the previous chapter showed that intermediaries are more likely to support the incumbent rather than opposition candidates. These two patterns combined would suggest that incumbents should perform better in rural areas. Electoral data confirm this expectation. They show that incumbents consistently score lower in urban areas both across time and space.

6 Social Structure and Ethnic Politics in Africa and Beyond

The varying landscape of local authority had important implications for electoral mobilization in Senegal and Benin, determining when politicians use ethnic appeals. The in-depth focus on these cases allowed us to trace the mechanism of electoral mobilization and carefully document the variation in social structure across time and politicians’ use of intermediaries or ethnic appeals. But is the elucidated dynamic generalizable? Does social structure affect electoral politics beyond the cases discussed so far? This chapter addresses these questions by providing additional analyses to test the main theory of this book. I selected two additional controlled comparisons, one from Francophone and one from Anglophone Africa. The comparison of Guinea and Mali, and Kenya and Botswana, reveals the dramatically different role of local leaders in these countries by the time of the founding elections. Whereas Mali and Botswana had robust hierarchical ties that were used to connect politicians with voters, preventing ethnic cleavages from arising, Kenya and Guinea had weak local leadership, leaving ethnic appeals as the easiest default option for politicians since independence. The chapter sequentially discusses the four countries’ social structures, electoral mobilization and electoral patterns. The final section then probes the generalizability of the theory in Africa and beyond.

mali and guinea The neighboring West African countries of Mali and Guinea have much in common, including a similar level of socioeconomic development and even some of the same ethnic groups, especially the Peul and Malinké, residing 146

Mali and Guinea

147

along the two countries’ border. Mali and Guinea have almost identical levels of ELF, 0.78 and 0.75, respectively, and they both have the crosscutting social institution of cousinage (joking kinship). Yet, ethnicity has played a dramatically different role in electoral politics in the two countries. Whereas Guinea’s politics are marred by ethnic cleavages and ethnic violence, ethnicity is a very poor predictor of vote choice in Mali. The PREG index also highlights this disparity. Despite the two countries’ similar levels of ethnic diversity, the PREG score is only 0.13 in Mali, but 0.48 in Guinea, indicating a much higher role of ethnicity in Guinean politics. The following sections will show that the role of local leaders was strikingly different in the two countries at the time when mass elections began. This gap has subsequently widened since independence, following a campaign against traditional authority waged by Guinea’s first president, Ahmed Sékou Touré. This varied role of local leaders impacted politicians’ modes of electoral mobilization. As highlighted in the following paragraphs, ethnic differences play an important role in political campaigns in Guinea, but Malian politicians have avoided ethnic appeals or building ethnically dominated parties, using influential social leaders to build cross-ethnic electorates instead. Social Structure in Mali As in Senegal, Malian society has strong hierarchical ties with influential local leaders, although Malian social structures tend to be less centralized than in Senegal.1 There are similar prominent forms of social stratification as in Senegal, such as Sudanic castes and Sufi brotherhoods, with their powerful marabouts (Hodgkin and Morgenthau 1964). All major groups have a caste system and the vast majority of people are members of one of the Sufi orders, mainly the Tijaniyya, Quadriyya or the Hamalliyya (Imperato 1989: 82, 86). In addition to Sufi orders and castes, there are other forms of hierarchical linkages, including ranked patrilineages (Hopkins 1972: 25). Throughout most of rural Mali, social 1

There are indications that the level of centralization of social structures varies subnationally in Mali. Anecdotally, northern Mali is believed to have more centralized social structure than southern Mali. Northern Mali also has higher turnout rates and, arguably, northern communities do a better job mobilizing voters and supporting candidates backed by traditional authority (personal communication with Jaimie Bleck, January 2010). This variation would be consistent with the patterns documented in Senegal (in Chapter 4), wherein the tighter the hierarchical ties, the more useful they are for electoral politics.

148

Social Structure and Ethnic Politics in Africa and Beyond

structure is based on “village-based gerontocracies, which through a complex system of age grades, initiation societies and land tenure control capital accumulation” (Imperato 1989: 81). The system of gerontocracy establishes a clear division of functions in a village, allocating “government over people and administration of goods to the elders, and the function of work to the young” (Diop 1971: 64). Importantly, just as in Senegal, the traditional forms of hierarchies survived until the onset of mass politics. Unlike in southern Benin, the French did not alter the existing structures of authority too much because they did not encounter resistance from the main groups, such as the Bambara. French colonial officials named men from traditional aristocracy as chefs de cantons rather than replacing them (Morgenthau 1964: 257), as they did throughout most of southern Benin. The continued strength of Mali’s local leaders is evidenced by data from the Afrobarometer survey. Sixty-five percent of respondents report that they trust traditional leaders, en par with levels of trust in Senegal, but twice as high as in Benin.2 Thirty-four percent of respondents state that traditional leaders influence “a great deal” the governance of their local community.3 Sixteen percent of men reported that they often turned for help to religious leaders during the course of the last year, whereas 14 percent reported that they relied on traditional leaders.4 These numbers are slightly below those of Senegal but about three times higher than in Benin. Social Structure in Guinea Although Guinea shares many similarities with neighboring Mali and Senegal, including some of the same ethnic groups, the Malinké and Peul, the standing of local leaders on the eve of mass politics was much weaker, and it further diverged under the leadership of Guinea’s independence leader, Ahmed Sékou Touré. Around the time of first mass politics in the 1950s, Guinean chiefs were very unpopular (Suret-Canale 1966: 479). They were widely seen as colonial agents and they overexploited the population, especially in their capacity as tax collectors for the French. Sékou Touré and his Rassemblement Démocratique Africain (RDA) party

2

3

Round 4, 2008. As in Senegal, trust in traditional leaders is much higher than in political actors. Only 38 percent of Malians trust the president. Trust in the National Assembly stands at 32 percent and in the ruling party at 26 percent. 4 Round 4, 2008. Round 3, 2005.

Mali and Guinea

149

capitalized on the chiefs’ lack of popularity and mobilized voters against the chefferie (Suret-Canale 1966: 484). After winning the founding election, Sékou Touré embarked on a project of destroying traditional elites. On December 31, 1957 chefferie was abolished in Guinea. This was no mere symbolic piece of legislation. Touré was committed to effecting the chiefs’ demise not only on paper but also in reality. Following Touré’s virulent attack, chiefs’ goods were confiscated and many chiefs had to go into exile. Touré similarly set out to undermine Muslim religious leaders. The political secretary of Guinea’s single party, Parti Démocratique de Guinée (PDG) “exhorted the PDG sections to wage a vigilant war against Muslim associations and leaders because of their ‘political nonconformism’” (Kaba 2000: 192). “Demaraboutization” was the official policy during the early years of Touré’s regime and “Guineans were asked to war against the hustlers and swindlers licensed as ‘marabouts’” (Kaba 2000: 193). Touré’s decision to destroy chiefs and marabouts had strong ideological underpinnings. By some accounts, Touré was the most “radical” politician in West Africa (Alexandre 1970a, Cooper 2002: 70), with fierce dislike of colonialism and traditional hierarchies, despite his own illustrious Malinké lineage. The French used the term “radical” to describe those leaders who were seeking autonomy and a new social order, instead of accommodation with the French, as did Senghor in neighboring Senegal. The clearest testament of Touré’s alleged radicalism came at the time of the 1958 referendum when voters throughout French West Africa had to decide whether to seek outright independence or continued membership in a French-led union. Touré was the only West African politician to campaign for independence and Guinea was the only colony where the majority of voters supported independence. As a result, Guinea became the first French West African country to gain independence in 1958, two years before France unilaterally ceded control of the rest of West Africa. The “yes” vote in Guinea was viewed as an indication of Touré’s radicalism because France put significant pressure on all West African countries to remain in its orbit. The French president, Charles de Gaulle, issued veiled threats to Guinea not to cut its ties to France and when Touré went ahead with independence, de Gaulle immediately froze all aid to Guinea, while the departing French personnel removed all valuable possessions, infamously ripping out telephone cords from the walls. Touré’s willingness to incur high costs for his country by standing up to de Gaulle illustrates that his commitment to certain causes was much stronger

150

Social Structure and Ethnic Politics in Africa and Beyond

than any sense of pragmatism, which characterized most other African leaders in Francophone Africa. Touré’s ideology, with his anticolonialism and his dislike of the feudal order that the chiefs embodied, played an important role in his decision to abolish the chefferie. Yet, the fact that the chiefs were widely unpopular at the time most likely also played a role. The chiefs’ weakness allowed Touré to defeat them; his task would have been much harder had the chiefs enjoyed widespread support among the population. Electoral Mobilization and Electoral Patterns in Mali In Mali, as in Senegal, the existing hierarchical ties created local leaders who were strong enough to serve as electoral intermediaries. The studies of early political competition indicate that politicians targeted voters through intermediaries instead of making appeals to ascriptive identities. Morgenthau (1964) documents how the first successful parties relied on intermediaries for their support. Parti Soudanais Progressiste (PSP) built its electoral support across ethnic lines through chiefs, who came from the traditional aristocracy, and the marabouts, who were particularly strong in the rural areas (Morgenthau 1964: 256–257, Hodgkin and Morgenthau 1964: 223). The party thrived on the support of rural peasants who followed their chiefs (Snyder 1965: 59). As Snyder observes, the PSP relied on loyalty and ties to traditional leaders, and its main competitor, Union Soudanaise (US), tried to combat it by appealing to the urban milieu, where such ties were weaker. The two main parties, PSP and US, were both national parties, covering the whole country, while parties with ethnic appeals never acquired much significance (Hodgkin and Morgenthau 1964: 221). Tribal identities had little use in explaining political competition. The main cleavage between PSP and US was urban–rural. The PSP was a traditional rural party, whereas the US built its base in the urban areas, largely in opposition to the PSP’s intermediary model (Hodgkin and Morgenthau 1964: 221–222). Mali followed a typical political trajectory in sub-Saharan Africa: After initial mass political contests in the 1950s and early 1960s, the victorious US party consolidated power and became a single party, later replaced by a military regime. After a three-decade-long suspension of competitive politics, multiparty elections resumed in 1992, during the Third Wave of democratization. The post-democratization elections revealed the nonsectarian nature of Malian politics, which continues to this day. Dunning and Harrison point out that ethnic identity is a poor predictor of vote choice

Mali and Guinea

151

and parties do not form along ethnic lines (2010: 21). Similarly, Wing notes that “despite the country’s ethnic diversity, Malian politics seem never to have been shaped mainly by ethnic interests” (2008: 9). In their study of the 2007 presidential elections, Baudais and Sborgi (2008) highlight that the incumbent, Amadou Toumani Touré, won throughout Mali and across ethnic groups, beating his rivals even in their supposed strongholds. Basedau and Stroh (2011) measure the extent of ethnicization of the party system and note that for Mali the figure is 0 percent, in contrast to highly ethnicized party systems, such as in Benin, where the figure is 44 percent. They also add that Mali’s main parties have weak levels of regionalization. As they sum up, in Mali “no indications can be found that there is either a substantial ethnic or regional support base of parties” (Basedau and Stroh 2011: 18). The individual-level data from the Afrobarometer confirm that parties in Mali have cross-ethnic bases of support, representative of the broader ethnic makeup of the Malian society (see Table 6.1).5 Even the last presidential election in the summer of 2013, which followed a coup in 2012 and a French military intervention, was not contested along ethnic cleavages. Tensions between the Tuareg minority, under 10 percent of the population, and the “black” majority remain significant, but there were no tensions or overt competition between the different “African” ethnic groups.6 A systematic search of campaign coverage by major media outlets that reported on the 2013 campaign, including Al Jazeera, the BBC, Le Figaro and Jeune Afrique, found no mention of ethnic mobilization by presidential candidates. Unlike in neighboring Guinea, press articles do not even make references to candidates’ ethnicities. Nor do they talk about ethnic groups, with the 5

6

The ethnic composition of these parties is almost identical when using an alternative question, namely “Which party do you feel close to?”; however, the sample size for that question is smaller. The same questions were also asked in Round 3 and yield similar results. There are several characteristics that set the Tuaregs apart from the rest of Malian society. First, they are considered Caucasoid, rather than Negroid. Despite their dark skin, they typically view themselves as distinct from “Black” Africans. The Tuaregs are also nomadic. They are spread across several Sahelian states, including Mali, Niger and Algeria. They do not respect official state boundaries, and since the colonial time, the Tuaregs have sought disengagement from the state, first from the French and subsequently from the independent state of Mali. Tuaregs have first fought against the French colonial army, and since Mali’s independence, there have been three separate periods of rebellion against the state, starting in 1962. The rebellion that broke out in January 2012 was the latest wave of insurgency, which aims to accomplish independence, or at least significant autonomy from the Malian state. For details on the Tuareg insurgency, see Keïta (1998) and Wing (2013). Importantly, the Tuaregs want to leave the state, but they do not compete along ethnic lines in elections.

152

Social Structure and Ethnic Politics in Africa and Beyond

table 6.1 Ethnic Composition of the Electorates of the Main Parties in Mali Parties Ethnic group Bambara Peul Senufo Soninke Sonrhai Dogon Malinké

Group’s share of the population

Adema

ATT/Mouvement Citoyen

RPM

31% 15% 10% 8% 7% 7% 7%

28% 15% 11% 6% 5% 14% 6%

33% 16% 5% 12% 13% 5% 4%

27% 16% 4% 3% 4% 7% 18%

Source: Afrobarometer, Round 4 (2008).

exception of the separatist Tuaregs. The eventual winner, Ibrahim Boubacar Keïta, better known as IBK, won convincingly throughout Mali. The parliamentary elections that followed a few months later, in November 2013, also saw no ethnic mobilization. At the same time, there was evidence of involvement of intermediaries in the electoral process. The eventual winner, IBK, enjoyed the support of influential religious leaders, who called on people to vote for him.7 Electoral Mobilization and Electoral Patterns in Guinea While Sékou Touré campaigned against the chiefs and French colonialism in the 1950s (Cooper 2002: 71), upon coming to power he created an ethnically dominated single party. The new single party, the PDG, was run mainly by Touré’s Malinké group and, as Horowitz pointed out, it was a clear example of ethnic hegemony (1985: 434). For example, among the seven top members of the Guinean politbureau, there were five Malinké, even though the Malinké share of the population was only 34 percent (Adamolekum 1976). In the 1970s, Touré embarked on a repression of the country’s largest group, the Peul, accusing them of plotting against him.8 While the Peuls were persecuted, the Forestier group felt marginalized and neglected by Touré’s regime.9

7 8

BBC, July 29, 2013, August 11, 2013, August 13, 2013, Jeune Afrique, August 6, 2013. 9 Jeune Afrique, June 23, 2010. Jeune Afrique, June 25, 2010.

Mali and Guinea

153

Authoritarian rule lasted in Guinea much longer than in neighboring countries. Whereas one-party rule was the norm throughout Africa until the end of 1980s, most countries experienced political liberalization and the return to competitive multiparty elections in the 1990s. Guinea is thus a continental laggard, with the first free multiparty elections taking place only in 2010, after the death of Guinea’s second authoritarian president, Lansana Conté, and a brief rule by a military junta. The 2010 presidential election, which was the country’s first competitive poll since independence, was tinged by ethnic politics from the start. The two candidates who won the most votes in the first round and thus advanced to the second round run-off, Alpha Condé and Cellou Dalein Diallo, were seen as representing the country’s two main ethnic groups, the Malinké and the Peul, respectively. Indeed, it is hard to find any articles describing the election that do not explicitly make references to leaders’ ethnic background and to voters’ communal identities. The two main parties, Condé’s Rassemblement du Peuple de Guinée (RPG) and Diallo’s Union des Forces Démocratiques de Guinée (UFDG), are seen as representing mainly their leaders’ ethnic groups. In the midst of the presidential run-off, there was a prevalent perception that the Malinké would not accept the victory of the Peul candidate, Cellou Dalein Diallo, whereas the Peul would oppose the Malinké, Alpha Condé.10 The Peul, in particular, felt that it was their “turn” after being governed by a Malinké (Sékou Touré), a Sousou (Lansana Conté) and a Forestier (the junta leader, Moussa Dadis Camara). Apart from ethnic mobilization, there was substantial fear of ethnic violence throughout the campaign. For example, the president of the National Transitional Council, known by its French acronym CNT, Rabiatou Serah Diallo, publicly expressed concerns about the possibility of clashes between ethnic communities.11 Toward the end of the campaign, ethnic skirmishes erupted in the capital, Conakry, and in cities in the eastern part of the country. Attacks against the Peul took place in towns in Upper Guinea, leading some Peul to flee the area.12 In the northern part of the country, hundreds of Peuls fled their homes in majority Malinké villages to escape ethnic violence.13 Under international pressure, the two candidates in the run-off issued a joint appeal for calm and “fraternity” after having previously accused each other of

10 12

Jeune Afrique, June 23, 2010. Jeune Afrique, July 11, 2010.

11 13

Jeune Afrique, June 11, 2010. New York Times, July 11, 2010.

154

Social Structure and Ethnic Politics in Africa and Beyond

fomenting ethnic tension that led to clashes.14 New ethnic riots erupted, nonetheless, when the initial results of the polls were announced.15 Ethnic tensions did not end once the presidential campaign was over. A UN news agency, IRIN, reported a year after the presidential election that politics remained ethnically divisive and public discourse has been peppered with ethnic rhetoric.16 Vincent Foucher, a researcher at the International Crisis Group, argued that “Ethnic tensions are getting worse, not better. Everyone is playing the ethnic card.”17 Corinne Dufka, head of the Human Rights Watch in West Africa, notes that the current administration has fomented ethnic tension and engaged in ethnic favoritism, discriminating against the Peul.18 Unfortunately, detailed census and electoral data for Guinea are hard to find. Any discussion of electoral results thus relies on broad patterns, taking into consideration the known homelands of different ethnic groups in Guinea. While we do not have an exact regional ethnic breakdown, the country’s Northwest is known as the Peul heartland, whereas the West is dominated by the Malinké. The two candidates in the run-off, a Peul and a Malinké, scored very high in their groups’ regions. In the two administrative regions in the Malinké heartland, Kankan and Nzerekoré, the Malinké candidate, Alpha Condé, scored 93 percent and 84 percent, respectively. In the Peul heartland, in the administrative regions of Labé and Mamou, Condé garnered only 4 percent and 7 percent, respectively.19 Given how lopsided these scores were, even without more disaggregated data, it is evident that ethnicity played an important role in determining vote choice. These broad patterns, taken together with overt ethnic references and ethnic clashes during electoral campaigns, highlight the extent of ethnic politics in Guinea.

kenya and botswana Kenya and Botswana have radically different levels of ethnic politics. While ethnic identity accounts for 36 percent of vote choice in Kenya, in Botswana it explains only 10 percent (Dowd and Driessen 2008). Indeed, 14 15

16 19

Jeune Afrique, July 11, 2010 Al Jazeera, November 17, 2010. Fears of new rounds of ethnic clashes were also reported by Agence France Presse on November 10, 2010. 17 18 IRIN, June 12, 2011. IRIN, June 12, 2011. IRIN, June 12, 2011. Electoral data comes from the Guinean Electoral Commission (CENI), published in Jeune Afrique November 15, 2010. Spatial distribution of Guinea’s ethnic groups comes from a 1973 map, available on www.vidiani.com.

Kenya and Botswana

155

Kenya is often cited as the quintessential example of ethnic politics, whereas Botswana is touted as an African success story with an absence of ethnic tensions. It might be tempting to ascribe Botswana’s lack of ethnic politics to its relative homogeneity as the majority of the population belongs to the Tswana group. However, the common Tswana language masks underlying diversity. Pitcher, Moran and Johnston (2009) note that ethnic homogeneity should not be overstated in Botswana as eight main tribes make up the Tswana group. This tribal diversity could have been easily exploited by political entrepreneurs for the purpose of political mobilization. Thus the absence of identity politics in Botswana should be explained rather than taken for granted. As the following paragraphs will show, many accounts stress the continuous importance of chiefs in electoral politics in Botswana. Arguably, this reliance on chiefs allowed politicians to build expansive electorates across different groups and preclude the need for mobilization of tribal identities. In the following paragraphs, I will highlight the striking difference between Kenya and Botswana in terms of the strength of local leadership by the time of first mass politics. While Kenya never had much social stratification or influential local leaders because its main ethnic groups were traditionally chiefless, Botswana historically had very robust chiefs who managed to weather the colonial period, benefiting from Britain’s benign neglect of the country. Botswana’s almost irrelevance to the British was a blessing for the country’s traditional authority. Subsequent sections will highlight how central these traditional figures have been to electoral politics in Botswana. Social Structure in Botswana At the onset of mass politics in the early 1960s, Botswana had powerful traditional leaders. The precolonial traditional system was very hierarchical, with clear authority figures. Due to very limited British involvement in Botswana, traditional authority was not affected much by the colonial experience. Tribal chiefdoms emerged in the territory of present-day Botswana during the eighteenth and early nineteenth century (Holm 1993: 179). Each group had an elaborate political system under the leadership of a chief, kgosi (Somolekae and Lekorwe 1993: 187). These chiefs were powerful rulers: they allocated land, controlled movement of goods, and owned and leased cattle, the main source of wealth in the area (Holm 1993: 181–182). As Holm describes, the chiefs “allocated [cattle] in chains of clientage to

156

Social Structure and Ethnic Politics in Africa and Beyond

subordinates” (1988: 179). They dominated all important forms of collective action in the community, making the indigenous polities of Botswana one of the most unequal and stratified in all of southern Africa. The precolonial position of chiefs was much less disrupted by the British rule than elsewhere in Africa. Botswana was the backwater of the British Empire (Parson 1992: 65). It was of “marginal importance to Britain” (Molutsi 1998: 365) and its colonial experience can be described as “benign neglect” at the hands of the British authorities, who administered the Bechuanaland Protectorate, the colonial name of present-day Botswana, from across the South African border (Dunning 2008: 261). The British exercised a very mild degree of sovereignty, with little interest in effecting any direct control (Holm 1988: 184). As a result, the British rule had minimal impact on traditional authority, compared to the rest of Africa, and the communal structures and traditional authority were almost left intact (Holm 1988: 183–184, Molutsi 1998: 365, Onoma 2006: 109). Consequently, the Tswana elites never lost their right to rule (Englebert 2000: 112). In the 1960s, at the start of mass politics and the eve of independence, traditional authority still played an important role in the lives of the various Tswana tribes. The political structure entailed a strongly institutionalized central authority, which focused on the chief (Pitcher et al. 2009). Molutsi notes that at the time Botswana could be characterized as “a weak state with strong traditional leadership”(1998: 365). Similarly, Onoma describes Botswana in the 1960s as “a collection of small communities ruled by chiefs” (2006: 61). These chiefs “kept a strong political and social grip over the lives of their subjects” (Molutsi 1998: 365); their legitimacy “ran deep” (Onoma 2006: 109); and they controlled crucial resources, including land, most of it under customary land tenure (Holm 1988: 186, Onoma 2006: 64). Since independence, the new government of Seretse Khama took several measures to reduce the role of chiefs, including passing legislation, such as the Tribal Land Act of 1968, which introduced land boards to take away chiefs’ control of land (Somolekae and Lekorwe 1998: 194). Yet, in practice, these reforms did not eliminate the power of chiefs who still shape land tenure and are still very influential in their communities (Onoma 2006: 64). As Somolekae and Lekorwe note, “Chieftaincy remains significant to a great many Batswana, especially in rural areas. The present government is quite aware of this fact, hence it has not abolished the institution. Such action would have been tantamount to political suicide” (1998: 196).

Kenya and Botswana

157

Social Structure in Kenya In contrast to Botswana, colonial reports and ethnographic studies of Kenyan groups highlight the weakness, or outright absence, of local leaders. Most segments of Kenyan society, including the leading Kikuyu and Luo groups, have long manifested weak social hierarchies.20 The Ethnographic Atlas ranks the Kikuyu and Luo, the two largest groups in the country as having little hierarchy, with a score of 0 and 1, respectively, on a 0–4 scale, in stark contrast to the Senegalese groups, which score 2 or 3 (Murdoch 1967, 1981). Tignor notes that much of Kenya was traditionally chiefless. He describes the Kikuyu, Kamba and Masai on the advent of colonialism as “decentralized conciliar polities” where most adult males had the right to voice their views (1971: 342). As he notes, “[B]y 1914, they [the British] had collected enough information to realize that the Kikuyu, before colonization, had had no chiefs” (Tignor 1971: 342). Yet, the British had trouble understanding and governing decentralized societies, so when faced with chiefless groups, they preferred to create chiefs. The British creation, however, was a resounding failure. The new chiefs never became popular or trusted by their communities. Tignor argues that most chiefs were simply powerless, whereas those who tried to accumulate power were highly unpopular and mistrusted. The Kikuyu chiefs, who tried to establish authority, fell in the latter category. Tignor highlights widespread opposition to Kiambu Kikuyu chiefs and argues that these “antagonisms were generated directly by the issue of collaboration, by the willingness of chiefs to do the bidding of a colonial government which was forcing unpopular programs on the people” (1971: 345). There were frequent reports that these chiefs were very corrupt and abusive, needing military bodies to impose their rule. Moreover, the chiefs were viewed as an outside imposition. As Tignor explains, “All knew that they were designated by the colonial authorities and represented the outsider’s interests. Their authority rested not on reciprocal relationships, such as increased political security and accelerated economic development, but on force” (1971: 352). He concludes that “very few chiefs were genuinely admired” and most Kikuyu interviewees spoke instead of chiefs’ “use of force to terrorize local communities and their illegal attainment of wealth” (Tignor 1971: 352). Whereas Kikuyu chiefs tried to establish authority by force, Masai and Kamba chiefs were harmless but marginal. As Tignor describes, “Masai 20

Murdoch (1967).

158

Social Structure and Ethnic Politics in Africa and Beyond

chiefs were simply insulated from the rest of society, the activities of which were conducted without regard to them. They were nearly powerless . . . Their influence over the rest of the tribe was minimal” (1971: 357). Similarly, Kamba chiefs’ rule was a failure due to Kamba’s individualism and lack of preexisting social divisions. Electoral Mobilization and Electoral Patterns in Botswana In the context of robust social structure in Botswana, successive governments since independence have relied on traditional leadership to harness electoral support, as well as gain backing for various government reforms and programs. After independence, “the chiefly class became the core leadership for the state” (Pitcher et al. 2009). As Pitcher et al. note, “the BDP aggressively drew chiefs into its base” and “drew strength from traditional structures.” The independence leader, Seretse Khama, enjoyed considerable support of traditional authority (Holm 1988: 186). He succeeded in uniting many of the traditional chiefs and leaders of the eight Tswana tribes behind his state-building project, but also used the traditional authority of Tswana leaders to limit challenges to the ruling elite (Dunning 2008: 261, 263). Somolekae and Lekorwe note that to this day MPs and government officials go through chiefs in most of their dealings with people (1998: 194). Chiefs play an intermediary role in mobilizing rural people for government policies, and in this respect chieftainship has a key function in the politics of Botswana, and will most likely do so for the years to come (Somolekae and Lekorwe 1998: 194). Chiefs and their headmen still survive as a potent force in politics, and since the government is aware of their popular status, it uses them to legitimate its new structures and to encourage political participation (Holm 1988: 199, Holm 1993: 107). In addition, Holm highlights the government’s use of the kgotla, the traditional village forum for communal decision making, as a means to communicate with the people (Holm 1993: 99–100). Among this broad-based incorporation of traditional structures for the purposes of state-building and development, the use of traditional authority for electoral gains is unmistakable. Parson notes the reliance on “traditional political status and legitimate authority” to forge links between the BDP, the only ruling party since independence, and the peasants, who constitute the majority of the electorate (1992: 76). The BDP has been even characterized by some as “a collection of local notables” (Holm 1988: 190). As Holm explains, “[T]he BDP succeeds by gaining the support of critical local notables and offers

Kenya and Botswana

159

them patronage to distribute to their supporters” (Holm 1988: 191–192). Similarly, Onoma argues that “the BDP translated chiefs’ support into consistent electoral success” (2006: 102). Appreciation of the influence of traditional leaders is not limited to the BDP. As Holm notes, despite his socialist sympathies, the leader of the opposition Botswana National Front (BNF) also “recognized that he would need the support of traditional authority to challenge the BDP successfully” (1988: 189). Unfortunately, for the BNF, he had very limited success in acquiring chiefs’ help. Apart from the fact that political parties search for chiefs’ backing, there is evidence that chiefs’ support for a given party does have an actual effect on people’s vote choice. Molutsi describes “the continuous influence of the chiefs on the voting pattern of their people” (1998: 370). The chiefs themselves are eager to promote such a view of their electoral prowess. Chief Linchwe argues that in Botswana “people still rally more behind the chief than behind the politician” (cited in Logan 2008). A couple of concrete examples can illustrate chiefs’ electoral pull. The second most powerful chief, Bathoen II Gaseitsiwe of Bangwaketse, joined the opposition BNF after he had a falling out with the ruling BDP. After throwing his weight behind the BNF, he helped them win three seats in his area (Molutsi 1998: 366, Somolekae and Lekorwe 1998: 196). Yet, after Bathoen II died, the BNF lost all the Bangwaketse seats (Molutsi 1998: 371). The case of the Kgatleng district shows a similar pattern. When the main chief in the area was against the BDP and sympathetic toward Botswana People’s Party (BPP), the district’s Bakgatla people voted for BPP in large numbers. However, when the chief’s attitude toward the BDP became favorable, there was a shift to the BDP in the 1970s and early 1980s (Molutsi 1998: 371). Electoral Mobilization and Electoral Patterns in Kenya Lacking viable alternatives, Kenyan politicians have consistently used ethnic mobilization as their means of garnering support. In fact, many analysts cite Kenya as one of the quintessential examples of ethnic politics in Africa (Ndegwa 1997, Ajulu 2002, Elischer 2013). The individual-level data from the Afrobarometer indicate that Kenya is one of the countries with the strongest association of voters’ ethnic identity and vote choice, with around one-third of vote choice correctly predicted by ethnic identity alone. In the Afrobarometer’s sample of 18 countries, only in Benin ethnicity was there a better predictor of vote choice.

160

Social Structure and Ethnic Politics in Africa and Beyond

The use of ethnic appeals dates back to Kenya’s independence leader, Jomo Kenyatta. As Branch (2011: 98–100) documents, the Kenyatta years were filled with tribalism and the dominant and later single party, Kenya African National Union (KANU), “seemed to be for Kikuyu” only. Once victorious, the president was determined to retain power by consolidating his Kikuyu base through land resettlement schemes, which privileged his coethnics, or by purging the military, among other means. As Branch notes, in his effort to bolster his support among the Kikuyu, Kenyatta “abandoned all but the most perfunctory pretense that his government was for all Kenyans” (2011: 102). Kenyatta’s successor, Daniel arap Moi, continued relying on his coethnics as his core supporters. President Moi shifted the balance of power from the Kikuyu to his fellow Kalenjin. When Moi faced domestic and international pressures to restore multiparty politics in the early 1990s, he defended the single party as necessary to prevent violence, claiming that democracy in Kenya would mean overt competition between groups and ethnic bloodshed (Branch 2011: 197). Unfortunately, Moi’s predictions came to fruition in 1991, with Luo families forced out of their homes by Kalenjin youths. Following the first outbursts of violence, Moi was quick to point out that “since multipartyism came, you can see ethnic clashes” (Branch 2011: 199). Further ethnic violence associated with electoral contests took place throughout the 1990s (Boone 2011). When Moi was barred from seeking further reelection in 2002 due to term limits, the first democratic alteration in power brought Mwai Kibaki, a Kikuyu, to the presidency. Despite democratic gains in the country, Kibaki followed his predecessors’ strategy of ethnic mobilization. Kibaki’s tactic was to galvanize support among fellow Kikuyu voters and those of other closely linked groups, such as Embu and Meru (Branch 2011: 267). Once in office, Kibaki’s government was accused of deliberately neglecting the Kalenjin to punish them for Moi’s earlier marginalization of the Kikuyu (2011). Ethnic mobilization remained the dominant strategy during Kibaki’s reelection bid. As Horowitz (2011) notes, “[T]he 2007 election was filled with divisive ethnic appeals that played on and exacerbated resentments and hostilities between ethnic communities.” The election, perceived as a duel between the Kikuyu, represented by Kibaki, and Raila Odinga’s Luo group, resulted in dramatic violence, when Kibaki was proclaimed the winner by a narrow margin, amidst accusations of fraud. Although much effort has been placed to prevent another outburst of ethnic violence, such as the one after the 2007 election that claimed approximately

Local Leaders and Ethnic Politics Across and Beyond Africa

161

3,000 lives, politics continues to be dominated by ethnic cleavages. Most prominent politicians are associated with specific groups. After Kibaki’s retirement, Uhuru Kenyatta, the winner of the 2012 election and the son of Kenya’s first president, became the representative of the Kikuyu. William Ruto, the current vice president, is described as the lynchpin of Kalenjin politics, succeeding former president Daniel arap Moi in this role (Branch 2011: 266). Raila Odinga, a serial presidential election loser, has been unable to expand sufficiently beyond his Luo base. Unfortunately, the fact that both the current president and vice president have been indicted by the International Criminal Court in the Hague for their role in stoking ethnic violence in 2007 does not bode well for the prospects of transcending ethnic politics in the near future.

local leaders and ethnic politics across and beyond africa Even though the strength of local leaders is important for electoral politics, unfortunately, it is not easily measured and there are no “offthe-shelf” indicators for it that would allow us to test its effect on a large number of cases. The six case studies that I analyzed in this book so far included ethnographies and other secondary sources to document the strength of hierarchical ties in a given polity. This strategy cannot be easily pursued for multiple cases in a single work. With this limitation in mind, and to supplement the evidence garnered from my in-depth cases, I use two proxies to probe further the generalizability of my theory. I then briefly discuss other empirical regularities in Africa and elsewhere that relate to the theory’s portability. First, I use the percentage of interviewees in a given country surveyed by the Afrobarometer stating that they trust traditional leaders “a lot” as a proxy for that country’s strength of hierarchical ties.21 The dependent variable measures the association between ethnic identity and vote choice, CVELI, as discussed in Chapter 1. The following graph gives a visual representation of the relationship between the two variables (see Figure 6.1). It shows that the lower the trust in traditional leaders, the higher the level of ethnic voting. The correlation between these two variables is −0.43. Although this cannot be treated as a definitive test, the relationship

21

Question 49i from Round 4, 2008.

Social Structure and Ethnic Politics in Africa and Beyond

162 0.45 0.4 Ethnic Politics (CVELI)

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0

10

20

30 40 Trust in Leaders

50

60

70

figure 6.1 Trust in Local Leaders and the Level of Ethnic Politics

that we find provides some suggestive evidence, as it is in the direction predicted by the theory. The same relationship holds if we use another proxy for the strength of hierarchical structures, namely the degree of influence that traditional leaders have in governing their community.22 I use the percentage of interviewees that say “a great deal” as a proxy. The correlation between the two variables is −0.4. Just as was the case with the previous question, the relationship between the variables is in the direction predicted by the theory. The greater the role played by traditional leaders, the lesser the extent of ethnic politics. Beyond cross-national variation, there is also subnational variation consistent with the theory. For example, in Ghana the group with the highest strength of hierarchical ties seems to engage less in ethnic politics than other groups in the country. The northern Dagomba show much higher levels of trust in traditional leaders than the two main ethnic groups, the Akan and Ewe, located in the South: 70 percent of Dagomba say they trust traditional leaders a lot, compared with 37 percent and 41 percent, respectively, for the other two groups.23 Whereas Dagomba chiefs openly back politicians, Ashanti chiefs play a rather marginal and 22

Question 65, Round 4, 2008.

23

Question 49i, Afrobarometer, Round 4, 2008.

Local Leaders and Ethnic Politics Across and Beyond Africa

163

ceremonial role in electoral politics. A member of the New Patriotic Party (NPP) in the Ashanti region explains, “People [here] don’t like it when chiefs get involved in politics. They should keep it to themselves. Here chiefs don’t do it. They wouldn’t have influence; people are independent.”24 At the same time, while the Akan and Ewe are each associated with a political party, the NPP and the National Democratic Congress (NDC), respectively, the Dagomba seem much more fluid in their electoral support and there is no “Dagomba” party or presidential candidate.25 There is also evidence that in many settings chiefs’ allegiance is driven by material interests rather than identity. As Van Kessel and Oomen argue in the case of South Africa, “[C]hiefs often align themselves, whether wholeheartedly or for tactical reasons, with the powers that seem to offer the best chances of safeguarding their positions” (1997: 562). The same can be said about the Ghanaian intermediaries. During the early post-independence years, the Ya-Na of Dagbon, the most prized intermediary among the Dagomba, wanted to support the most likely winner. Although he initially sided with the NPP, as Nkrumah’s CPP grew in strength, the Ya-Na switched his allegiance to the CPP (Staniland 1975: 146). Elsewhere in Ghana, the Brong chiefs also decided to back Nkrumah for strategic reasons, because he promised them to take the Brong area away from Asante and turn it into a separate region. After the chiefs supported Nkrumah, he fulfilled his promise to them and created a new region, the Brong Ahafo (Rathbone 2000: 78). Another aspect of some Ghanaian chiefs’ behavior also indicates that their allegiance is often strategic, rather than based on ethnic solidarity. In many chiefly families, just as in Senegalese maraboutic families, there were splits, with some members of the family supporting one candidate and others throwing their weight behind another one (Rathbone 2000: 79). Even within northern chiefly families from the same ethnic group, different chiefly “gates,” the Abudu and Andani, associate themselves with the two rival political parties.26 The two main parties, in turn, vie for chiefly support by backing different candidates for succession. The chiefs’ political support is thus based on their internal family rivalries and competition for power rather than ascriptive ties with politicians.

24 25 26

Author’s interview with Kwesi Kyei, Kumasi, Ghana, June 19, 2014. Michelitch (2015) provides a good description of Ghanaian ethnic electoral allegiances. Author’s interview with Professor Aminu Dramani, Kumasi, Ghana, June 19, 2014.

164

Social Structure and Ethnic Politics in Africa and Beyond Generalizing Beyond Africa

While I focused thus far on Africa, different cases from outside of Africa also provide some evidence of generalizability of this theory, as they show patterns of electoral politics that are consistent with the argument. First, Southeast Asia exhibits several examples of the usefulness of hierarchical ties to politicians, to the exclusion of ethnic appeals. Scott shows that patron–client linkages, rather than ethnic bonds between members of the same groups, served as building blocks of electoral politics in many settings throughout Southeast Asia (1972: 103). As he explains, instead of electoral competition between separate groups, more often in Southeast Asia “a situation prevails in which a number of patrons with separate followings within the same communal group compete for the most advantageous links to the outside” (1972: 104, emphasis in the original). In addition to highlighting competition between patrons, rather than groups, Scott makes explicit references to the variable strength of social structures. As he notes, vertical links in some cases are strong, whereas sometimes they are “decentralized” (1972: 103). He highlights Philippines and Thailand as the cases with strong vertical ties and with lack of ethnic electoral competition (1972: 103, 105). While the cases of Philippines and Thailand are consistent with my theory, they are not as useful for determining causality. Scott notes a general weakness of ethnic ties in these countries. Without more detailed process tracing, we cannot determine whether these ties were always weak, or whether they lost strength as a result of their irrelevance to the functioning of society in terms of distribution of goods or providing avenues of advancement. Undoubtedly, it would be more desirable for the validity of the theory to show a scenario where ethnic ties were robust, but when presented with strong hierarchical ties politicians chose the latter. Even though such a scenario would most likely have to be based on a counterfactual, Scott does not seem to think that more robust ethnic ties would have eliminated the prevailing patron–client hierarchical ties. Rather, he voices his skepticism about the usefulness of nominal ethnic ties to provide adequate access to goods and services (1972: 103). Scott’s cases also support another element of the theory, namely politicians’ adaptation of mobilization strategies to local structure. As he argues, “throughout Southeast Asia all parties had to adapt themselves to these differences in social structure in different regions of the country: The PNI of Indonesia operated differently in Central Java than in the Outer Islands; U Nu’s AFPFL faction could not win the support of the hill tribes in the same way they won the vote of the lowland Burmese, and the

Conclusion

165

Nationalists in the Philippines campaigned differently in central Luzon than in Mindanao” (1972: 111). There are also indications that when presented with important material benefits, voters beyond Africa are willing to support candidates from different ethnic and religious backgrounds. Thachil (2011) showed in the case of India that lower caste voters whose communities benefited from social services provided by organizations affiliated with the Bharatiya Janata Party (BJP) – a Hindu nationalist party – were surprisingly willing to support the BJP, despite its upper caste image. Even more remarkably, there are instances in India of Muslim brokers working for the Hindu BJP.27 Intermediaries that link voters with politicians can potentially create nonethnic alliances.

conclusion This chapter showed that the argument about the effect of local leaders on electoral politics presented in this book extends to other African countries, beyond Senegal and Benin. The comparison of Mali and Guinea, two Francophone countries with similar ethnic groups but local leaders of different stature, revealed that whereas local leaders worked as electoral intermediaries in Mali, producing ethnically diverse voting patterns, traditional leaders in Guinea have been all but eliminated and could play no role in helping the country overcome its pronounced ethnic political cleavages. Similarly, the contrast between Kenya and Botswana highlighted the weakness of traditional leadership in Kenya and the strong and sustained position of traditional leaders in Botswana. Whereas the former were in no position to play a meaningful role in electoral politics and ethnic mobilization became the norm in Kenya, local leaders in Botswana have worked with a range of politicians over the past 50 years, producing inclusive voting patterns. The chapter also showed a general tendency across Africa of lower levels of ethnic politics in polities with stronger local leaders.

27

Personal communication with Adam Auerbach, May 2014.

7 Conclusion

This book was motivated by a desire to explain why in seemingly similar environments electoral politics cleave along ethnic lines, whereas in other comparable settings politicians create inclusive electorates, representative of the population at large. Finding the answer to this question is important, to the extent that inclusive parties and candidates are both much closer to the democratic ideal and are consequential in practice for democratic accountability. Until recently, the assumption that all electoral politics in Africa is ethnic was widespread and the only question of interest was which dimension of ethnicity will become salient. Since then, scholars have produced a much more nuanced empirical picture of African elections, showing that ethnicity is by no means the only predictor of vote choice and that ethnic census elections are not ubiquitous. While we know that ethnicity plays a markedly different role in multiparty elections across the continent, we have made much less progress understanding this variation. As I have discussed in Chapter 1, many existing theories fail to account for empirical variation. Despite the importance of the question, the number of theories grappling with this question is rather limited. The theory that I develop in this book challenges two prevalent assumptions, which, I would argue, prevented us from understanding the apparent puzzle of nonethnic politics in ethnically diverse societies. First, I show that ethnic ties are not the only useful and viable ties that can be used for non-programmatic electoral mobilization. I highlight that politicians can use a different kind of clientelist or personalistic linkages, namely ties between voters and local authority figures, such 166

Conclusion

167

as chiefs or religious dignitaries. While programmatic appeals are also an alternative to ethnic mobilization, they have not eliminated nonprogrammatic, clientelist electioneering in most African countries, especially in rural areas. The second assumption that I challenge is the notion that a country’s demographics, such as the sizes of different ethnic groups, is the only relevant factor in determining politicians’ electoral strategies. Societies vary widely in the landscape of social authority and politicians respond to this variation in crafting their electoral strategies. This variation in social structure is crucial. In this book I argued that where there are robust local leaders, politicians have a viable alternative to appeals to ethnic ties. In the absence of such leaders, ethnic ties remain the “go-to” strategy for politicians. I showed that the two principal cases examined in this book, Senegal and Benin, had remarkably different local hierarchies – with powerful local leaders in Senegal and weak ones in Benin – and that politicians adapted their electoral strategies accordingly, enlisting the help of intermediaries in Senegal, but relying on ethnic appeals in Benin. I further examined two other sets of countries – Kenya and Botswana, and Mali and Guinea – highlighting the same pattern of qualitatively different electoral mobilization. I also made the case that electoral intermediaries are not merely another layer of electoral mobilization; they are not simply a different means that lead to the same outcome. Instead, intermediaries are a pivotal actor because they change ensuing electoral patterns. As the Senegalese case illustrated in detail, intermediaries have both the incentives and the means to work with politicians across the ethnic spectrum, delivering votes of their dependents across ethnic lines. Furthermore, this book’s focus on social structure, with a different strength of local leaders in urban and rural areas, allows us to understand the different voting behavior and electoral patterns in urban and rural Africa. As Chapter 5 showed, the urban–rural cleavage is one of the most robust electoral patterns across Africa. Yet, theories emphasizing ethnic composition of the electorate cannot account for it. Instead, I show that urban and rural areas vote differently because of their varied levels of dependence and vulnerability to clientelism. In the remainder of this chapter, I consider the boundary conditions of my theory and discuss possible sources of change to social structure and electoral outcomes. I then reflect on broader implications and contributions of this book’s theory.

168

Conclusion

generalizability and boundary conditions The cases reviewed in Chapter 6 showed the generalizability of the theory developed in this book. Beyond Africa, the theory should apply to other developing countries with multiparty politics, low institutionalization and largely non-programmatic politics, and with a significant amount of clientelism. Under what conditions would the theory not apply? One of the key elements of my argument is that mobilization through intermediaries changes electoral outcomes because intermediaries are able to work with politicians across the ethnic spectrum. Intermediaries are expected to be very fluid in their electoral support, easily changing allegiance in between elections and following the highest bidder. The case study of Senegal highlighted the intermediaries’ openness to different alliances and it showed that they do not remain loyal to specific candidates when the candidates’ electoral prospects, and funds at their disposal, change. The behavior of intermediaries in many cases resembles a free market. But, would there be circumstances when intermediaries are not as free in their selection of patrons? The high flexibility of intermediaries and their lack of loyalty are possible because of the absence of salient ideological differences among the electorate and because of the organizational weakness of political parties. These factors currently apply to most African countries, but in the context of higher salience of ideology and stronger parties, I would expect intermediaries to be more limited in their behavior. Intermediaries can change allegiance to different candidates and parties because these parties or candidates do not represent palpably different programs. They promise general development and access to resources. To the extent that voters do not see a marked difference between candidates and they are simply trying to ensure that they will not be shut out of the distribution of resources, they do not object to intermediaries’ changing alliances. They are willing to follow intermediaries to maximize their material well-being. In the absence of substantial ideological differences between political actors, voters may view intermediaries’ maneuvers as pragmatic rather than unscrupulous. This could change with the rising salience of position issues. If political candidates took opposing stances on important issues, it would be harder to justify switching one’s loyalty between them. If parties acquire salient ideological labels, it will be more difficult to convince voters to shift their support. An increase in ideological differences producing higher levels of partisan attachments would limit intermediaries in their choices. For this

Persistence and Change

169

reason, political brokers in Latin America, for example, are more constrained in their behavior and they thus typically work with the same party over time. Intermediaries also have little reason to remain loyal to parties as long as parties remain weak and poorly institutionalized. Because of African parties’ high levels of personalism, limited funding and weak brands, intermediaries have little to gain by staying loyal to parties, especially those in opposition. They are better off acting as free agents, especially when politicians switch parties freely and frequently as well. If parties become stronger, more institutionalized and more prosperous, intermediaries might remain more loyal. At present, supporting parties in opposition is not lucrative, as these parties have few resources. If parties develop more substantial bases of funding, not dependent on being in power, they will have better chances of retaining intermediaries.

persistence and change As the previous section highlighted, there are several conditions that would limit the applicability of the theory and certain changes in key parameters that would alter the predictions of the theory. Important changes to the landscape of authority would produce different electoral outcomes. Because social structure is not immutable, the theory is dynamic in nature. As I explained in Chapter 1, electoral competition is path dependent. The relative stickiness of electoral patterns seen since the first mass elections in the 1950s is a testament to the effect of the founding elections that set the tone for subsequent electoral contests. Politicians generally keep winning strategies, thus reproducing existing patterns. However, the endurance of social structure also helps reinforce the existing patterns. Changes in social structure can alter electoral dynamics, albeit slowly, as politicians update their priors. For example, if local leaders experience a loss of authority and trust, they will become less effective as intermediaries. Chapter 5 already showed that changes such as increasing urbanization transform social structure because local leaders play a lesser role in urban areas than in rural areas. Growing urbanization across Africa will thus undermine to some extent the role of local leaders. Other developments that reduce voters’ dependence on local leaders, such as increasing wealth or education, could also undermine the effect of electoral intermediaries. At the same time, these developments might create space for an infusion of programmatic politics.

170

Conclusion

However, local leaders are unlikely to lose authority and become dispensable in the near future. The continued weakness of most African states and the political apparatus will continue to create demand for the services of intermediaries. Indeed, some politicians might see the need to strengthen local leaders rather than weaken them, given their electoral usefulness. Baldwin (2014) found widespread instances of politicians’ efforts to further empower local authority by increasing their control over land. There is no sign of any reforms that would undermine the role of local leaders in Senegal, whereas in Benin there are efforts to resurrect the largely ceremonial royalty. President Yayi has increased resources available to both religious and traditional leaders. Shortly after coming into office, he created two institutions – Cadre de concertation des autorités communales and Fonds d’appui aux communautés religieuses – to bolster state funding to local authority figures. The president also introduced salaries for traditional leaders and royals. It is unlikely that the Beninese authority figures will rival in strength their Senegalese counterparts in the foreseeable future, but President Yayi’s policies are a clear attempt to increase local leaders’ clout. Anecdotally, many Beninese saw these policies through the prism of electoral politics. It is also possible to see the entrance of new intermediaries. For example, in the last couple of decades evangelical churches have gained ground across Africa, providing many essential benefits to their members. McCauley (2012) found in a study of Ghana that Pentecostal leaders formed new clientelist networks. If these networks become the dominant sources of help in certain communities, Pentecostal leaders might become appealing intermediaries. In Benin, Pentecostal leaders are some of the biggest beneficiaries of the new sources of funding provided by the government. Such changes are most likely to be gradual, but over time it is possible to see a different landscape of social authority.

broader implications The main argument of the book was that social structure impacts the type of electoral mobilization and electoral patterns. But do these patterns have any broader implications beyond the presence or absence of ethnic politics? These modes of electoral mobilization are important because they could impact voters’ sense of inclusion and their access to resources. In Benin, where ethnic mobilization is widespread, voters without a coethnic in power feel discriminated vis-à-vis the president’s coethnics. When asked by the Afrobarometer survey how often their group is treated

Broader Implications

171

unfairly, President Yayi’s coethnics, the Bariba, report such sentiment much less frequently than other ethnic groups. Whereas only 18 percent of Bariba say that they are sometimes treated unfairly by the government, and only 3 percent say “often”, these numbers are much higher among the Fon, 34 percent and 11 percent, and the Adja, 31 percent and 19 percent, respectively.1 The same question yielded different results in Senegal. First, a much larger percentage of the Senegalese than Beninese respondents report that their group is never treated unfairly by the government – 84 percent and 54 percent, respectively. There was also no evidence that President Macky Sall’s coethnics, the Serer and Peul, felt treated more fairly than the Wolof.2 Indeed, the number of Peul and Serer respondents stating that they are sometimes or often treated unfairly, 8 percent and 7 percent for Peul and 11 percent and 7 percent for Serer, was higher than among the Wolof (3 percent and 1 percent). Not surprisingly, given the Diola’s historical marginalization largely due to the weakness of local elites, the group reports highest perceived levels of unfair treatment, with 14 percent of respondents saying that their group is sometimes treated unfairly and further 16 percent saying “often.” These responses lend credence to the suggestion that electoral patterns translate into more tangible aspects of voters’ well-being. Admittedly, a sense of unfair treatment is not equal to actual discrimination. Yet, voters’ perceptions are important, whether they are empirically grounded or not, as they shape their confidence in the political system. There are some indications that the strength of local leadership has impact on voters’ access to resources. Bratton, Coulibaly and Machado (2001) found that despite much more dire economic conditions in Mali, twice as many Malians as Nigerians report being able to turn to a community group during a time of food shortage. Malians also report two to three times higher levels of trust in traditional authority than do Nigerians (Logan 2008). In this case, the strength of local leadership corresponds with tangible benefits. While reliable data on social expenditures and the provision of services is very difficult to obtain, there is anecdotal evidence that well-organized communities enjoy higher levels of material support. In the Senegalese case, Casamance,

1

2

Question 85A, Round 5, 2011. The recently released data from the Afrobarometer Round 6 survey confirm these results, with the gap between the president’s coethnics and noncoethnics even widening. The survey was carried out almost a year after Sall’s inauguration.

172

Conclusion

with its weakness of local leadership, has long been marginalized in the distribution of resources vis-à-vis the more hierarchical groups. Even among groups with strong leadership, the most tightly organized ones seem to be noticeably better placed to extract resources. Touba, the center of the Mouride brotherhood, with powerful local leaders par excellence, seems visibly better endowed in comparison to other Senegalese cities. All Senegalese presidents since independence have lavished the seat of the brotherhood with money and even policy concessions. While much more research is needed into redistributive effects of different modes of political mobilization, there are reasons to believe that the type of electoral mobilization is consequential. The fluidity of political alliances allows, at least in principle, more consistent access to resources. It should also affect voters’ sense of inclusion and the integration of the national political system.

contributions The theory and the findings of this book have implications for several different sets of literature. First, this book provides an account of the variation in social structure, shedding light on the wide differences in the degree of power of local authorities throughout Africa. While the main purpose of understanding the origin of the variation in social structure was to establish its exogeneity to mass electoral politics, the historical examination of the fate of local leaders uncovered an important pattern, namely the impact of colonial rule on local leadership. This adds to the growing body of work on the political impact of colonialism on African societies. Just as Posner (2003) found that the colonial rule shaped ethnic identities, this book documents how colonial rule impacted the landscape of traditional and religious leadership, beyond the precolonial variation in the standing of those elites. While readers might not be surprised that colonial rule reshaped traditional leadership, I highlight a less appreciated pattern, namely that colonial officials were much more effective at destroying existing social structures than at altering them or building new ones. Wherever traditional and religious leaders posed a threat to the colonial enterprise, colonial powers weakened or eliminated local hierarchies. However, there are few examples of successful creation of respected local leaders. This pattern highlights very important limitations of the colonial enterprise. As much as it was desirable to create local leaders molded by French or British interests, such leaders were rarely popular, trusted or respected

Contributions

173

by the population, which they were supposed to control. Despite instances of creation of new chiefs, or replacement of traditional chiefs with those backed by colonial officials, this strategy did not translate into strong hierarchical ties between local leaders and their followers, much to the dismay of the colonial officials. The cases reviewed in this book showed uniformly that local authorities molded by the colonizer were viewed as usurpers and were highly unpopular. Local leaders remained popular and influential among their base only in cases where colonial officials left them some autonomy. Second, and most obviously, the book makes a contribution by providing an explanation of the puzzling variation in the extent of ethnic politics across Africa. In doing so, it adds to the burgeoning literature on political competition in Africa. Furthermore, the focus on social structure breaks new ground as it is a clear departure from the dominant paradigms of electoral politics in Africa, which typically have highlighted ethnic composition of the electorate. Yet, as I show, demographics are not the only relevant variable determining politicians’ strategies. Social structure matters because it reflects the different type and strength of ties that are most salient in voters’ lives. Some earlier works on politics in developing countries (e.g. Scott 1972, 1976) highlighted the importance of social structure in conditioning politicians’ behavior. In this sense, this book is bringing social structure “back in.” However, it goes beyond the earlier scholarship by showing that social structure has an impact not only on electoral strategies but also on electoral outcomes. The argument about the effect of local leaders on the presence or absence of ethnic cleavages is novel. The theory presented here also helps disentangle the relationship between ethnicity and clientelism. Many works view clientelism in Africa as intertwined with ethnicity (Bates 1974, Posner 2005). This book, in contrast, shows that ethnic clientelism is just one possible form, however widespread, of clientelist networks. In environments with strong local leaders, the ties between local leaders and their followers provide alternative clientelist channels. Rather than assuming that ethnic ties provide the default networks for the distribution of resources, I contend that politicians use the dominant preexisting selfhelp networks to channel resources to voters. As a result, ethnic groups and clientelist networks do not always go hand-in-hand. This book also speaks to the growing literature on traditional authority in Africa. In doing so, it takes up Sklar’s (1993) challenge for Africanists to study more seriously and rigorously the neglected topic of dual authority,

174

Conclusion

or mixed government, namely the coexistence of bureaucratic and traditional forms of power. Indeed, Sklar argued that the phenomenon of dual authority is an area of study where Africanists can make the biggest impact on Political Science as it is of general interest to the wider discipline (1993: 86–87). After several decades when local leaders were ignored by academics, or expected to cease to exist as countries modernized, they are now seen again as important social and political actors. Indeed, instead of withering away, the influence of traditional leaders in Africa has grown during the period of multiparty politics in the last two decades (Logan 2008, 2011, Baldwin 2014). The general weakness of African states and the difficulty of state-building and forging linkages with rural populations provide a backdrop to understanding the continued usefulness of local leaders. In contrast to dictatorial one-party rule, competitive politics increased politicians’ need to form ties with voters and to gain legitimacy. Recent scholarship showed that local leaders were historically useful to rulers, both colonial and postcolonial, for the purposes of state-building (Herbst 2000, Boone 2003) and development (Cruise O’Brien 1971, 1975, Baldwin 2013). Englebert also made the case that traditional leadership had important implications for a given state’s economic prospects and its level of legitimacy (2000). Beyond Africa, Slater (2009) found that the strength of what he eloquently called communal elites had important implications for democratic mobilization in Southeast Asia. This book adds to this burgeoning field by stressing the importance of the variable strength of local leadership for ethnic politics, and electoral mobilization more broadly. As Sklar (1993) noted, even when traditional authorities exist, they are not always “the power behind the throne,” raising the crucial question as to why their role in elections differs so widely. This book provides an explanation of the variable role of traditional elites in electoral politics and the transformative effects of their involvement.

appendix a Socially Salient Identities in Senegal and Benin

table a.1 Identity Repertoire in Senegal General category

Classification

Group size

Ethnic group Wolof (descent-based) Halpulaar (Peul and Tukulor) Serer Diola Manding

Religion

Muslim Christian (mainly Catholic)

Brotherhood (Sufi Order/ tariqa)

Tijaniyya Mouridiyya Quadriyya

Caste (Sudanic type)

Freemen Non-artisan/ Non-casted “géér” Casted/Artisan castes/ “neeno” (castes of blacksmiths, griots (praise-singers), leatherworkers, etc.)

175

Meaning/indicators

Different family names; stereotypes especially based on traditional 14.8% occupation of each 5.5% group – for example, 4.6% traditionally the Wolof are merchants, the Peul pastoralists, the Tukulor sedentary agriculturalists, the Serer fishermen, etc. 93.8% Different first names (French for 4.3% Catholics; Africanized Arabic for Muslims); dress 41.6% Variation in religious practice and 28% celebration (e.g. 10% pilgrimages, holy places); allegiance to different leaders Not available Based on traditional occupation; a frequent barrier to marriage (endogamy) 43.7% 23.2%

176

Appendix A table a.2 Identity Repertoire in Benin

General category

Classification

Ethnic group (descent-based)

Group size

Fon (South) Adja (South) Yoruba (South) Bariba (North) Peul (North) Region (geography- South based) North

39% 15% 12% 9% 7% 2/3 1/3

Religion

36%

Christians [27% Catholic; 5% Protestant] Muslims Vodoun

24% 17%

Meaning/indicators In some cases associated with territorial homeland/precolonial kingdom; stereotypes Stereotypes – e.g. Northerners perceived as backward (even savage); Southerners as “foreigners,” cooperating with Europeans Different first names (French for Catholics; Africanized Arabic for Muslims); dress; different religious practice and customs

appendix b The Role of Social and Political Leaders in Senegal and Benin

table b.1 Reliance on Different Types of Leaders in Senegal and Benin

Local government councilor MP Official of a government ministry Political party official Religious leader Traditional leader

Senegal

Benin

14.7% 2.3% 2.4% 6.3% 22.9% 15.5%

5.5% 0.8% 0.3% 1.7% 5.0% 5.3%

Source: Question 32, Afrobarometer, Round 3 (2005).

177

appendix c Additional Electoral Data for Senegal and Benin

178

table c.1 1970 Presidential Election in Benin Region

179

Ouémé Atlantique Zou Mono Borgou Atakora Total

No. of registered voters 234,664 229,309 144,197 221,137 167,919 190,000

No. of voters 112,729 109,278 73,613 137,024 132,703 n/a 565,347

Turnout (percent voting) Ahomadegbe 47.7 47.7 51 62 78.9 n/a

8,679 (8%) 56,843 (52%) 31,961 (43%) 99,592 (73%) 3,017 (2%) n/a

Apithy

Maga

92,295 (82%) 6,129 (5%) 38,009 (35%) 6,100 (6%) 19,725 (27%) 5,705 (8%) 26,511 (19%) 6,211 (5%) 288 (0.2%) 128,406 (97%) n/a n/a

Note: The election was suspended by the military before voting took place in Atakora, the last department scheduled to vote. Source: Afrique Contemporaine, May–June 1970: 7 (reprinted in Decalo 1970: 455).

Zinsou 4,104 (4%) 5,724 (5%) 4,074 (6%) 3,494 (3%) 257 (0.2%) n/a

Appendix C

180

table c.2 Ethnic Composition of the Electorate of the Main Presidential Candidates in Benin (2008) Candidates Ethnic group Fon Adja Bariba Yoruba Goun Otomari

Group’s share of the population

Yayi (FCBE)

Soglo (RB)

30% 18% 12% 13% 7% 7%

20% 17% 20% 13% 4% 8%

75% 17% 0% 4% 2% 0%

Houngbedji Amoussou (PRD) (PSD) 40% 6% 0% 20% 33% 1%

0% 100% 0% 0% 0% 0%

Source: Question 97, Afrobarometer, Round 4, N = 1200.

table c.3 Ethnic Composition of the Electorate of the Main Presidential Candidates in Benin (2011) Candidates Ethnic group Fon Adja Bariba Yoruba

Group’s share of the population 42% 16% 9.2% 14%

Yayi (FBCE) Houngbedji (UN) 32% 15% 14% 16%

63% 16% 0% 20%

Bio Tchané (ABT) 18% 13% 8% 13%

Notes: Based on Question 99 asking respondents if the presidential elections were held tomorrow, who would you vote for (a candidate from which party)? Only 65% of respondents indicated a specific candidate that they would support. Almost 20% said they didn’t know who they would vote for and almost 8% refused to answer. I include here all candidates that received support from at least 5% of respondents. All other candidates mentioned received below 1% of declared support. Data on respondents’ ethnicity are based on Question 84. Some of the available answers changed from Round 4. For example, the Goun self-identification was no longer provided as an option to respondents. These changes concern only smaller ethnic groups, and the four largest ethnic categories remained unchanged. Group’s share of the population size is based on the group’s share of the sample (rather than the population as a whole). Since it is an equal probability sample, it should very closely approximate group’s share of the population. Source: Afrobarometer, Round 5. N = 1200.

Appendix C

181

table c.4 Extent of Ethnic Politics in Senegal and Benin Measure CVELI (Dowd and Driessen 2008) Party Nationalization Score (Elischer 2013)

Meaning Percentage of vote choice predicted by voter’s ethnicity The spread of party’s support 0–1, with higher values indicating more national support

Senegal 11%

41%

All above African average PS: 0.84–0.92 PDS: 0.78–0.9

All but 1 below African average PSD: 0.44–0.49 PRD: 0.32–0.67 MADEP: 0.32 FARD-Alafia 0.31 ADD 0.56 UBF 0.65 FCBE 0.72

a. The lower the a. 0.03 (2003), a. Ethnic 0.13 (2006) Polarization value (0-1), the less ethnically b. 0.7 (2003) b. Ethnic 0.6 (2006) Diversity concentrated (Cheeseman the vote and Ford 2007) b. The higher the value (0-1), the more ethnically diverse parties’ electorate is Ethnicization of The extent to which parties’ Parties Basedau and Stroh electorate is (2011) ethnically concentrated

N/A

Number of top ethnic candidates

0 for all elections

Over 75% of vote from a single ethnic group

Benin

N/A

2 strongly ethnicized parties (PSD and RB); 1 medium ethnicized (PRD) and 1 low ethnicized (FCBE); Benin overall: 44% ethnicized; All parties have strong or medium level of regionalization 3 out of 4 (2005); 2 out of 4 (2008), 0 (2011)

Appendix C

182

table c.5 Ethnic Composition of the Electorate of the Main Parties in Legislative Elections in Senegal Parties Ethnic group

Group’s share of the population

PDS

PS

AFP

42% 28% 13% 5% 5%

39% 31% 13% 5% 4%

43% 28% 20% 5% 5%

45% 25% 11% 6% 4%

Wolof Pulaar Serer Mandinka Diola Source: Afrobarometer, Round 2.

table c.6 Religious and Brotherhood Composition of the Electorate of Main Parties in Legislative Elections in Senegal Parties Religious group Group’s share of the population

PDS

PS

AFP

Muslims Catholics Muslim brotherhood Mouride Tijan

95% 4%

95% 4%

90% 4%

91% 9%

40% 39%

43% 38%

41% 42%

32% 51%

Source: Afrobarometer, Round 2.

table c.7 Ethnic Composition of the Main Candidates/Parties in Senegal (2005)

Ethnic group Wolof Pulaar Serer Mandinka Diola

Candidates

Group’s share of the population

Wade (PDS)

Dieng (PS)

Niasse (AFP)

48% 25% 12% 3% 5%

46% 25% 11% 5% 5%

51% 17% 22% 3% 5%

57% 14% 17% 0% 5%

Source: Afrobarometer, Round 3.

Appendix C

183

table c.8 Ethnic Composition of the Main Candidates/Parties in Senegal (2008) Candidates Ethnic Group Wolof Haalpulaar Serer Diola Mandinka

Group’s share of the population

Wade (PDS) Dieng (PS)

46% 25% 13% 6% 7%

45% 27% 10% 8% 7%

40% 21% 21% 2% 10%

Niasse (AFP)

Sagna (TDS)

55% 25% 9% 7% 2%

48% 23% 15% 7% 3%

Source: Afrobarometer, Round 4.

table c.9 Ethnic Composition of the Main Candidates/Parties in Senegal (2013)

Ethnic group Wolof Peul Serer Diola

Candidates

Group’s share of the population

Sall (APR)

43% 25% 14% 5%

37% 32% 19% 2%

Source: Afrobarometer, Round 5.

Niasse (AFP) Wade (PDS) 48% 15% 11% 6%

41% 23% 6% 14%

appendix d Incumbent Support in Urban and Rural Areas in Senegal

184

table d.1 Determinants of Incumbent Support in Senegal

185

Incumbent vote

|

Coefficient

Standard error

z

P>|z|

Milieu Education Age Brotherhood Gender Constant

| | | | | |

0.4800215 −0.1127794 0.0003659 0.0003681 −0.1706428 −1.787198

0.2219195 0.0609973 0.0005638 0.0003858 0.1874053 0.5009928

2.16 −1.85 0.65 0.95 −0.91 −3.57

0.031 0.064 0.516 0.340 0.363 0.000

[95% Conf. interval] 0.0450672 −0.2323318 −0.0007392 −0.0003881 −0.5379505 −2.769126

0.9149758 0.0067731 0.0014709 0.0011243 0.1966648 −0.8052703

Note: Logistic regression of vote for the incumbent on place of residence (urban versus rural milieu), level of education, age, brotherhood affiliation and gender. N = 741. Source: Senegal, Afrobarometer, Round 2.

References

Adamolekum, ‘Ladipo. 1976. Sékou Touré’s Guinea: An Experiment in Nation Building. London: Methuen & Co Ltd. Afrobarometer Briefing Paper, 17. 2005. “Démocratie au Bénin: le point de vue de la population béninoise à partir de l’Enquête Afrobaromètre 2005.” Ajulu, Rok. 2002. “Politicised Ethnicity, Competitive Politics and Conflict in Kenya: A Historical Perspective.” African Studies, 61, 2: 251–268. Aldrich, John. 1995. Why Parties? The Origin and Transformation of Political Parties in America. Chicago: The University of Chicago Press. Alexandre, Pierre. 1970a. “Chiefs, Commandants and Clerks: Their Relationship from Conquest to Decolonization in French West Africa.” In Michael Crowder and Obaro Ikime, eds., West African Chiefs: Their Changing Status under Colonial Rule and Independence. Ile-Ife, Nigeria: University of Ife Press. Pp. 2–13. Alexandre, Pierre. 1970b. “The Problems of Chieftaincies in French Speaking Africa.” In Michael Crowder and Obaro Ikime, eds., West African Chiefs: Their Changing Status under Colonial Rule and Independence. Ile-Ife, Nigeria: University of Ife Press. Pp. 24–78. Arriola, Leonardo. 2013. Multiethnic Coalitions in Africa: Business Financing of Opposition Election Campaigns. New York: Cambridge University Press. Audrain, Xavier. 2004. “Du « ndigel avorté » au parti de la vérité.” Politique Africaine, 96: 99–118. Auyero, Javier. 2001. Poor People’s Politics: Peronist Survival Networks and the Legacy of Evita. Durham, NC: Duke University Press. Bako-Arifari, Nassirou and Pierre-Yves LeMeur. 2003. “La chefferie au Bénin: une resurgence ambiquë” In Claude-Hélène Perrot and François-Xavier FauvelleAymar, eds., Le retour des rois. Paris: Editions Karthala. Pp. 125–144. Baldwin, Kate. 2013. “Why Vote with the Chief? Political Connections and Public Goods Provision in Zambia.” American Journal of Political Science, 57, 4: 794–809. Baldwin, Kate. 2014. “When Political Leaders Cede Control of Resources: Chiefs, Land, and Coalition Building in Africa.” Comparative Politics, 46: 253–271. 186

References

187

Banégas, Richard. 2003. La démocratie à pas de caméléon: transition et imaginaires politiques au Bénin. Paris: Karthala. Banégas, Richard. 2014. “Briefing Benin: Challenges for Democracy.” African Affairs, 113, 452: 449–459. Bartolini, Stefano. 2000. Political Mobilization of the European Left, 1860–1980: The Class Cleavage. Cambridge: Cambridge University Press. Basedau, Matthias ad Alexander Stroh. 2011. “How Ethnic are African Parties Really? Evidence from Four Francophone Countries.” International Political Science Review, 33, 1: 5–24. Bates, Robert. 1974. “Ethnic Competition and Modernization in Contemporary Africa.” Comparative Political Studies, 6: 457–484. Bates, Robert. 1981. Markets and States in Tropical Africa: The Political Basis of Agricultural Policies. Berkeley: University of California Press. Battle, Martin and Jennifer Seely. 2007. “It’s All Relative: Competing Models of Vote Choice in Benin.” Afrobarometer Working Paper, 78. Baudais, Virginie and Enrico Sborgi. 2008. “The Presidential and Parliamentary Elections in Mali, April and July 2007.” Electoral Studies, 27: 769–773. Bayart, Jean-François. 1987. Etat en Afrique: La Politique du Ventre. Paris: Fayard. BBC. 2012. “Country Profiles: Guinea.” At: www.bbc.co.uk/news/world-africa13442053, accessed March 15, 2012. Beck, Linda. 1997. “Senegal’s ‘Patrimonial Democrats’: Incremental Reform and the Obstacles to the Consolidation of Democracy.” Canadian Journal of African Studies, 31, 1: 1–31. Beck, Linda. 2001. “Reining in the Marabouts? Democratization and Local Governance in Senegal.” African Affairs, 100: 601–621. Beck, Linda. 2002. “Le clientelism au Sénégal: un adieu sans regrets.” In MomarCoumba Diop, ed., Sénégal Contemporain. Paris: Karthala. Pp. 529–549. Beck, Linda. 2008. Brokering Democracy in Africa: The Rise of Clientelist Democracy in Senegal. New York: Palgrave Macmillan. Behrman, Lucy. 1970. Muslim Brotherhoods and Politics in Senegal. Cambridge: Harvard University Press. Bezemer, Dirk and Derek Headey. 2008. “Agriculture, Development, and Urban Bias.” World Development, 36, 8: 1342–1364. Bierschenk, Thomas. 2006. “The Local Appropriation of Democracy: An Analysis of the Municipal Elections in Parakou, Republic of Benin, 2002–3.” Journal of Modern African Studies, 44, 4: 543–571. Bleck, Jaimie and Nicolas van de Walle. 2011. “Parties and Issues in Francophone West Africa: Towards a Theory of Non-Mobilization.” Democratization, 18, 5: 1125–1145. Bleck, Jaimie and Nicolas van de Walle. 2013. “Valence Issues in African Elections: Navigating Uncertainty and the Weight of the Past.” Comparative Political Studies, 46, 11: 1394–1421. Bokoma, Lioutao. 2011. “Entreprises et marchés politiques lors des elections communales de 2008 dans la commune de Kalalé.” Master’s Thesis. University of Parakou. Boone, Catherine. 2003. Political Topographies of the African State: Territorial Authority and Institutional Choice. Cambridge: Cambridge University Press.

188

References

Boone, Catherine. 2011. “Politically Allocated Land Rights and the Geography of Electoral Violence: The Case of Kenya in the 1990s.” Comparative Political Studies, 44, 10: 1311–1342. Brambor, Thomas, William Roberts Clark and Matt Golder. 2007. “Are African Party Systems Different?” Electoral Studies, 26: 315–323. Branch, Daniel. 2011. Kenya: Between Hope and Despair, 1963–2011. New Haven: Yale University Press. Bratton, Michael and Nicholas van de Walle. 1997. Democratic Experiments in Africa. Cambridge: Cambridge University Press. Bratton, Michael, Massa Coulibaly and Fabiana Machado. 2001. “Popular Perceptions of Good Governance in Mali.” Afrobarometer Working Paper, 9. Bratton, Michael and Mwangi Kimenyi. 2008. “Voting in Kenya: Putting Ethnicity in Perspective.” Economics Working Papers. Bratton, Michael, Ravi Bhavnani and Tse-Hsin Chen. 2011. “Voting Intensions in Africa: Ethnic, Economic or Partisan?” Afrobarometer Working Paper, 127. Brusco, Valeria, Marcelo Nazareno and Susan Stokes. 2007. “Poverty, Risk and Clientelism.” Paper presented at the Annual Meeting of the American Political Science Association, Chicago, August 30–September 2. Chandra, Kanchan and Daniel Metz. 2002. “A New Cross-National Database on Ethnic Parties.” Paper presented at Midwestern Political Science Association. Chandra, Kanchan. 2004. Why Ethnic Parties Succeed: Patronage and Ethnic Head Counts in India. Cambridge: Cambridge University Press. Chandra, Kanchan. 2005. “Ethnic Parties and Democratic Stability.” Perspectives on Politics, 3, 2: 235–252. Chandra, Kanchan. 2006. “What Is Ethnicity and Does It Matter?” Annual Review of Political Science, 9: 397–424. Chazan, Naomi. 1982. “Ethnicity and Politics in Ghana.” Political Science Quarterly, 97, 3: 461–485. Cheeseman, Nicholas and Robert Ford. 2007. “Ethnicity as a Political Cleavage.” Afrobarometer Working Paper, 83. Cliffe, Lionel. 1967. One Party Democracy. Nairobi: East African Publishing House. Conroy- Krutz, Jeffery. 2007. “African Cities and Incumbent Hostility: Explaining Opposition Success in Urban Areas.” Unpublished Manuscript. Conroy- Krutz, Jeffery. 2012. “Information and Ethnic Politics in Africa.” British Journal of Political Science, 43, 2: 345–373. Cooper, Frederick. 2002. Africa Since 1940: The Past of the Present. New York: Cambridge University Press. Coulibaly, Abdou Latif. 2006. Une Démocraties Prise en Otage par Ses Elites: Essai Politique sur la Pratique de la Démocratie au Sénégal. Dakar: Les Editions Sentinelles. Coulon, Christian. 1981. Le Marabout et le Prince: Islam et Pouvoir au Sénégal. Paris: Editions A. Pedone. Coulon, Christian. 1988. “Senegal: The Development and Fragility of Semidemocracy.” In Larry Diamond, Juan Linz and Seymor Martin Lipset, eds., Democracy in Developing Countries: Africa. Boulder, CO: Lynne Rienner Publishers. Pp. 141–178.

References

189

Creevey, Lucy. 2006. “Senegal: Contending with Religious Constraints.” In Gretchen Bauer and Hannah Britton, eds., Women in African Parliaments. Boulder, CO: Lynne Rienner Publishers. Crowder, Michael and Obaro Ikime. 1970. “‘Introduction’ in West African Chiefs: Their Changing Status under Colonial Rule and Independence.” In Michael Crowder and Obaro Ikime, eds., West African Chiefs: Their Changing Status under Colonial Rule and Independence. Ile-Ife, Nigeria: University of Ife Press. Cruise O’Brien, Donal. 1971. The Mourides of Senegal. Oxford: Claredon Press. Cruise O’Brien, Donal. 1975. Saints and Politicians: Essays in the Organization of a Senegalese Peasant Society. Cambridge University Press. Cruise O’Brien, Donal. 1992. “Le contrat social sénégalais à l’épreuve.” Politique Africaine, 45: 9–20. Cruise O’Brien, Donal. 2002a. “Chefs, Saints et Bureaucrates: La Construction de l’Etat Colonial.” In Donal Cruise O’Brien, Momar-Coumba Diop and Mamadou Diouf, eds., La Construction de L’Etat au Sénégal. Paris: Karthala. Cruise O’Brien, Donal. 2002b. “Les negotiations du contrat social sénégalais.” In Donal Cruise O’Brien, Momar-Coumba Diop and Mamadou Diouf, eds., La Construction de L’Etat au Sénégal. Paris: Karthala. Dahl, Robert. 1956. A Preface to Democratic Theory. Chicago: The University of Chicago Press. Dahou, Tarik and Vincent Foucher. 2004. “Le Sénégal: entre changement politique et révolution passive.” Politique Africaine, 96: 5–21. Decalo, Samuel. 1970. “Full Circle in Dahomey.” African Studies Review, 13, 3: 445–457. Decalo, Samuel. 1973. “Regionalism, Politics, and the Military in Dahomey.” The Journal of Developing Areas, 7, 3: 449–478. de Jong, Ferdinand. 2005. “A Joking Nation: Conflict Resolution in Senegal.” Canadian Journal of African Studies, 39, 2: 389–413. Diamond, Larry. 1988. Class, Ethnicity, and Democracy in Nigeria: The Failure of the First Republic. Syracuse, NY: Syracuse University Press. Diop, Abdoulaye-Bara. 1981. La société Wolof: Les systèmes d’inégalité et de domination. Paris: Karthala. Diop, Alioune Badara. 2002. “Logiques sociales et démocratie électorale au Sénégal. Essai de reconstruction et d’interprétation d’une trajectoire de crise: l’exemple de Fouta Toro (1983–2001).” Unpublished PhD Thesis. Insitut d’Etudes Politiques de Bordeaux, Université Montesquieu – Bordeaux IV. Diop, Majhemout. 1971. Histoire de classes sociales dans l’Afrique de l’ouest (I): le Mali. Paris: François Maspero. Diop, Momar-Coumba, Mamadou Diouf and A. Diaw. 2000. “Le Baobab a été déracine. L’alternance au Sénégal.” Politique Africaine, 78: 157–179. Diop, Momar-Coumba and Mamadou Diouf. 2002a. “L’administration, les confréries religieuses et les paysanneries.” In Donal Cruise O’Brien, MomarCoumba Diop and Mamadou Diouf, eds., La Construction de L’Etat au Sénégal. Paris: Karthala. Diop, Momar-Coumba and Mamadou Diouf. 2002b. Le Sénégal sous Abdou Diouf. Paris: Karthala.

190

References

Diouf, Mamadou. 1992. “Le clinetelisme, la « technicratie » et après?” In Momar Coumba Diop, ed., Sénégal: Trajectoire d’un Etat. Diouf, Makhtar. 1994. Sénégal: Les ethnies et la nation. Paris: L’harmattan. Dowd, Robert and Michael Driessen. 2008. “Ethnically Dominated Party Systems and the Quality of Democracy: Evidence from Sub-Saharan Africa.” Afrobarometer Working Paper, 92. Dresang, Dennis. 1974. “Ethnic Politics, Representative Bureaucracy and Development Administration: The Zambian Case.” American Political Science Review, 68, 4: 1605–1617. Dunning, Thad. 2008. Crude Democracy: Natural Resource Wealth and Political Regimes. Cambridge University Press. Dunning, Thad and Lauren Harrison. 2010. “Cross-Cutting Cleavages and Ethnic Voting: An Experimental Study of Cousinage in Mali.” American Political Science Review, 104, 1: 21–39. Elischer, Sebastian. 2013. Political Parties in Africa: Ethnicity and Party Formation. New York: Cambridge University Press. Englebert, Pierre. 2000. State Legitimacy and Development in Africa. Boulder, CO: Lynn Rienner Publishers. Englebert, Pierre. 2002. “Born-again Buganda or the Limits of Traditional Resurgence in Africa.” Journal of Modern African Studies, 40, 3: 345–368. Fatton, Robert. 1986. “Clientelism and Patronage in Senegal.” African Studies Review, 29, 4: 61–78. Fatton, Robert. 1987. Making of a Liberal Democracy: Senegal’s Passive Revolution, 1975–1985. Boulder, CO: Lynne Rienner Publishers. Faye, Ousseynou. 1994. “L’instrumentalisation de l’histoire et de l’ethnicité dans le discours séparatiste en Basse Casamance (Sénégal).” Afrika Spektrum, 29, 1: 65–77. Ferree, Karen. 2004. “The Microfoundations of Ethnic Voting: Evidence from South Africa.” Afrobarometer Working Paper, 40. Ferree, Karen. 2010. Framing the Race in South Africa: The Political Origins of Racial-Census Elections. New York: Cambridge University Press. Foltz, William. 1969. “Social Structure and Political Behavior of Senegalese Elites.” Behavior Science Notes, 4, 2: 145–163. Franck, Raphaël and Ilia Rainer. 2012. “Does the Leader’s Ethnicity Matter? Ethnic Favoritism, Education, and Health in Sub-Saharan Africa.” American Political Science Review, 106, 2: 294–325. Galvan, Dennis. 2001a. “Market Liberalization as a Catalyst for Ethnic Conflict in Senegal and Central Java, Indonesia.” Paper presented at the 97th Annual Meeting of American Political Science Association. Galvan, Dennis. 2001b. “Political Turnover and Social Change in Senegal.” Journal of Democracy, 12, 3: 51–62. Galvan, Dennis. 2004. The State Must Be Our Master of Fire: How Peasants Craft Culturally Sustainable Development in Senegal. Berkeley and Los Angeles: University of California Press. Galvan, Dennis. 2006. “Joking Kinship as a Syncretic Institution.” Cahiers d’études africaines, 4, 184: 809–834.

References

191

GERCOP. 1999. “Etude sur le comportement électoral dans les régions de Thiès et Diourbel.” Saint-Louis, Sénégal. Glélé, Maurice. 1969. Naissance d’un état noir: L’évolution politique et constitutionnelle du Dahomey, de la colonisation a nos jours. Paris: Librairie Générale de Droit et de Jurisprudence. Gottlieb, Jessica. 2014. “The Determinants of Local Leader Influence in Elections: A Lab-in-the-field Experiment in Senegal.” Paper presented at the Annual Meeting of the American Political Science Association, Washington, DC. Harding, Robin. 2010. “Urban-Rural Differences in Incumbent Support Across Africa.” Afrobarometer Working Paper, 120. Harsch, Ernest. 2003. “Senegal: to Fight hunger, Modernize Farming.” Africa Recovery, 17: 1. Herbst, Jeffrey. 2000. States and Power in Africa: Comparative Lessons in Authority and Control. Princeton, NJ: Princeton University Press. Hodgkin, Thomas and Ruth Schachter Morgenthau. 1964. “Mali.” In James Coleman and Carl Rosberg, eds., Political Parties and National Integration in Tropical Africa. Berkeley and Los Angeles: University of California Press. Hoffman, Barak and James Long. 2013. “Parties, Ethnicity, and Voting in African Elections.” Comparative Politics, 45, 2: 127–146. Holm, John. 1988. “Botswana: A Paternalistic Democracy.” In Larry Diamond, Juan Linz and Seymour Lipset, eds., Democracy in Developing Countries: Africa. Boulder, CO: Lynne Rienner Publishers. Holm, John. 1993. “Political Culture and Democracy: A Study of Mass Participation in Botswana.” In Stephen Stedman, ed., Botswana: the Political Economy of Democratic Development. Boulder, CO: Lynne Rienner Publishers. Hopkins, Nicholas. 1972. Popular Government in an African Town. Chicago and London: University of Chicago Press. Horowitz, Donald. 1985. Ethnic Groups in Conflict, Berkeley: University of California Press. Horowitz, Jeremy. 2011. Campaigns, Ethnic Polarization, and Violence in Kenya’s Multiethnic Democracy. Ph.D. diss., University of California, San Diego. Ichino, Nahomi and Noah Nathan. 2013. “Crossing the Line: Local Ethnic Geography and Voting in Ghana.” American Political Science Review, 107, 2: 344–361. Imperato, Pascal. 1989. Mali: A Search for Direction. Boulder, CO: Westview Press. Iroko, Félix. 2003. “Rois et chefs en République du Bénin (1960–1999)” In Claude-Hélène Perrot and François-Xavier Fauvelle-Aymar, eds., Le retour des rois. Paris: Editions Karthala. Pp. 111–124. Ishiyama, John. 2011. “Ethnic Parties: Their Emergence and Political Impact.” Party Politics, 17, 2: 147–149. Jeffries, Richard and Clare Thomas. 1993. “The Ghanaian Elections of 1992.” African Affairs, 92, 368: 331–366. Jeffries, Richard. 1998. “The Ghanaian Elections of 1996: Towards the Consolidation of Democracy?” African Affairs, 97, 387: 189–208.

192

References

Joseph, Richard. 1987. Democracy and Prebendal Politics in Nigeria: The Rise and Fall of the Second Republic. New York: Cambridge University Press. Kaba, Lansiné. 2000. “Islam in West Africa: Radicalism and the New Ethnic of Disagreement, 1960–1990.” In Nehemia Levtzion and Randall Pouwels, eds., The History of Islam in Africa. Athens, OH: Ohio University Press. Kalyvas, Stathis. 1996. Rise of Christian Democracy in Europe. Ithaca: Cornell University Press. Kasara, Kimuli. 2007. “Tax Me if You Can: Ethnic Geography, Democracy, and the Taxation of Agriculture in Africa.” American Political Science Review, 101, 1: 159–172. Kasfir, Nelson. 1979. “Explaining Ethnic Political Participation.” World Politics, 31, 3: 365–388. Kaspin, Deborah. 1995. “The Politics of Ethnicity in Malawi’s Democratic Transition.” Journal of Modern African Studies, 33, 4: 595–620. Keïta, Kalifa. 1998. “Conflict and Conflict Resolution in the Sahel: The Tuareg Insurgency in Mali.” Small Wars and Insurgencies, 9, 3: 102–128. Kimenyi, Mwangi. 2006. “Ethnicity, Governance and the Provision of Public Goods.” Journal of African Economies, 15, AERC Supplement 1: 62–99. Kitschelt, Herbert and Steven Wilkinson. 2007. “Citizen-politician Linkages: An Introduction.” In Herbert Kitschelt and Steven Wilkinson, eds., Patrons, Clients and Policies. Cambridge: Cambridge University Press. Pp. 1–49. Klein, Martin. 1968. “The Evolution of the Chefferie in Senegal.” In Arnold Rivkin, ed., Nations by Design: Institution Building in Africa. Garden City, NY: Anchor Books. Koter, Dominika. 2013. “Urban and Rural Voting Patterns in Senegal: The Spatial Aspects of Incumbency, c.1978–2012.” The Journal of Modern African Studies, 51, 4: 653–679. Kramon, Eric and Daniel Posner. 2012. “Ethnic Favoritism in Primary Education in Kenya.” Unpublished manuscript. Kramon, Eric and Daniel Posner. 2013. “Who Benefits from Distributive Politics? How the Outcome One Studies Affects the Answer One Gets.” Perspectives on Politics, 11, 2: 461–474. Krishna, Anirudh. 2011. “Gaining Access to Public Services and the Democratic State in India: Institutions in the Middle.” Studies in Comparative Political Development, 46, 4: 98–117. Laitin, David. 1986. Hegemony and Culture: Politics and Religious Change among the Yoruba. Chicago: Chicago University Press. Lambert, Michael. 1998. “Violence and the War of Words: Ethnicity V. Nationalism in Casamance.” Africa, 68, 4: 585–602. Lake, David and Donald Rothchild. 1998. International Spread of Ethnic Conflict: Fear, Diffusion and Escalation. Princeton, NJ: Princeton University Press. LeBas, Adrienne. 2011. From Protest to Parties: Party-Building and Democratization in Africa. Oxford: Oxford University Press. Lehoucq, Fabrice. 2007. “When Does a Market for Votes Emerge?” In Frederic Schaffer, ed., Elections for Sale: The Causes and Consequences of Vote Buying. Boulder, CO: Lynne Rienner Publishers.

References

193

Lemarchand, René. 1972. “Political Clientelism and Ethnicity in Tropical Africa: Competing Solidarities in Nation-Building.” American Political Science Review, 66, 1: 68–90. Lemarchand, René and Keith Legg. 1972. “Political Clientelism and Development: A Preliminary Analysis.” Comparative Politics, 4, 2: 149–178. Lemarchand, René. 1988. “The State, the Parallel Economy, and the Changing Structure of Patronage Systems.” In Donald Rothchild and Naomi Chazan, eds., The Precarious Balance: State and Society in Africa. Boulder, CO: Westview Press. Pp. 149–170. Levtzion, Nehemia and Randall Pouwels. 2000. “Introduction.” In Nehemia Levtzion and Randall Pouwels, eds., The History of Islam in Africa. Athens, OH: Ohio University Press. Pp. 1–18. Lieberman, Evan. 2007. “Ethnic Politics, Risk, and Policy-Making: A Cross-National Statistical Analysis of Government Responses to HIV/AIDS.” Comparative Political Studies, 40: 1407–1432. Lieberman, Evan. 2009. Boundaries of Contagion: How Ethnic Politics Have Shaped Government Responses to AIDS. Princeton, NJ: Princeton University Press. Lieberman, Evan and Prerna Singh. 2012. “The Institutional Origins of Ethnic Violence.” Comparative Politics, 45, 1: 1–24. Lieberman, Evan and Gwyneth McClendon. 2013. “The Ethnicity-Policy Preference Link in Sub-Saharan Africa.” Comparative Political Studies, 46, 5: 574–602. Lindberg, Staffan. 2003. “It’s Our Time to ‘Chop’: Do Elections in Africa Feed Neo-patrimonialism Rather Than Counteract it?” Democratization, 10, 2: 121–140. Lindberg, Staffan and Minion Morrison. 2008. “Are African Voters Really Ethnic or Clientelistic? Survey Evidence from Ghana.” Political Science Quarterly, 123, 1: 95–122. Lipset, Seymour and Stein Rokkan. 1967. Party Systems and Voter Alignments: Cross-National Perspectives. New York: Free Press. Lipton, Martin. 1977. Why Poor People Stay Poor: Urban Bias in World Development. Cambridge: Harvard University Press. Lipton, Martin. 1993. “Urban Bias: Of Consequences, Classes and Causality.” Journal of Development Studies, 29, 4: 229–258. Logan, Carolyn. 2008. “Traditional Leaders in Modern Africa: Can Democracy and the Chief Co-Exist?” Afrobarometer Working Paper, 93. Logan, Carolyn. 2011. “The Roots of Resilience: Exploring Popular Support for African Traditional Authorities.” Afrobarometer Working Paper, 128. Lombard, Jacques. 1965. Structures de type “féodal” en Afrique Noire: Etudes des dynamismes internes et des relations sociales chez les Bariba du Dahomey. Paris: Mouton & Co. Lombard, Jacques. 1967. Autorités Traditionnelles et Pouvoirs Européens en Afrique Noire: Le Déclin d’une Aristocratie sous le Régime Colonial. Paris: Librairie Armand Colin. MacLean, Lauren. 2010. Informal Institutions and Citizenship in Rural Africa: Risk and Reciprocity in Ghana and Côte d’Ivoire. New York: Cambridge University Press.

194

References

Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and its Demise in Mexico. New York: Cambridge University Press. Mahoney, James and Dietrich Rueschemeyer. 2003. “Comparative Historical Analysis: Achievements and Agendas.” In James Mahoney and Dietrich Rueschemeyer, eds., Comparative Historical Analysis in the Social Sciences. Cambridge: Cambridge University Press. Mama, Machoud. 2008. Entrepreneur politique et marché politique au Bénin: le Président Boni Yayi et les elections municipals de 2008 à Parakou. Masters Thesis. University of Parakou. Mayrargue, Cédric. 2006. “Yayi Boni, un president inattendu?” Politique Africaine, 102: 155–172. McCauley, John. 2012. “Africa’s New Big Man Rule? Pentecostalism and Patronage in Ghana.” African Affairs, 126, 446: 1–21. McLaughlin, Fiona. 1995. “Haalpulaar Identity as a Response to Wolofization.” African Languages and Cultures, 8, 2: 153–168. Melson, Robert and Howard Wolpe. 1970. “Modernization and the Politics of Communalism: A Theoretical Perspective.” American Political Science Review, 64: 1112–1130. Michelitch, Kristin. 2015. “Does Electoral Competition Exacerbate Patterns of Interethnic or Interpartisan Economic Discrimination? A Field Experiment on Price Bargaining in Africa: Evidence From a Field Experiment I Market Price Bargaining.” American Political Science Review, 109, 1: 43–61. Miguel, Edward. 2004. “Tribe or Nation? Nation Building and Public Goods in Kenya Versus Tanzania.” World Politics, 56, 3: 327–362. Miller, Norman. 1968. “The Political Survival of Traditional Leadership.” The Journal of Modern African Studies, 6, 2: 183–201. Molutsi, Patrick. 1998. “Elections and Electoral Experience in Botswana.” In Wayne Edge and Mogopodi Lekorwe, eds., Botswana: Politics and Society. Pretoria: J. L. van Schaik Publishers. Morgenthau, Ruth. 1964. Political Parties in French-Speaking West Africa. Oxford: Clarendon Press. Mozaffar, Shaheen and James Scarritt. 2005. “The Puzzle of African Party Systems.” Party Politics, 11, 4: 399–421. Murdoch, George. 1967. Ethnographic Atlas. Pittsburgh, PA: University of Pittsburgh Press. Murdoch, George. 1981. Atlas of World Cultures. Pittsburgh, PA: University of Pittsburgh Press. Ndao, El Hadj Ibrahima. 2003. Sénégal, histoire des conquêtes démocratiques, Les Nouvelles Editions Africaines du Sénégal. Ndegwa, Stephen. 1997. “Citizenship and Ethnicity: An Examination of Two Transition Moments in Kenyan Politics.” American Political Science Review, 91, 3: 599–616. Nyerere, Julius. 1967. Freedom and Unity. Oxford: Oxford University Press. Onoma, Ato. 2006. Securing Property Rights: Politics on the Land Frontier in Postcolonial Africa. Ph.D. diss., Evanston, IL: Northwestern University. Onoma, Ato. 2009. The Politics of Property Rights Institutions in Africa. New York: Cambridge University Press.

References

195

Padró i Miquel, Gerard. 2007. “The Control of Politicians in Divided Societies: The Politics of Fear.” Review of Economic Studies, 74: 1259–1274. Parson, Jack. 1992. “Liberal Democracy, the Liberal State and the 1989 General Elections in Botswana.” In Stephen Stedman, ed., Botswana: the Political Economy of Democratic Development. Boulder, CO: Lynne Rienner Publishers. Patterson, Amy. 2002. “The Impact of Senegal’s Decentralization on Women in Local Governance.” Canadian Journal of African Studies, 36, 3: 490–529. Pélissier, Paul. 1966. Les Paysans du Sénégal: Les Civilisations Agraires du Cayor à la Casamance. Saint-Yrieix: Imprimerie Fabrègue. Pitcher, Anne, Mary Moran and Michael Johnston. 2009. “Rethinking Patrimonialism and Neopatrimonialism in Africa.” African Studies Review, 52, 1: 125–156. Posner, Daniel and David Simon. 2002. “Economic Conditions and Incumbent Support in Africa’s New Democracies: Evidence from Zambia.” Comparative Political Studies, 35, 3: 313–336. Posner, Daniel. 2003. “The Colonial Origins of Ethnic Cleavages: The Case of Linguistic Divisions in Zambia.” Comparative Politics, 35, 2: 127–146. Posner, Daniel. 2004a. “Measuring Ethnic Fractionalization in Africa.” American Journal of Political Science, 48, 4: 849–863. Posner, Daniel. 2004b. “The Political Salience of Cultural Difference: Why Chewas and Tumbukas Are Allies in Zambia and Adversaries in Malawi.” American Political Science Review, 98: 4. Posner, Daniel. 2005. Institutions and Ethnic Politics in Africa. Cambridge University Press. Powell, John. 1970. “Peasant Society and Clientelist Politics.” American Political Science Review, 64, 2: 411–425. Rathbone, Richard. 2000. Nkrumah and the Chiefs: The Politics of Chieftaincy in Ghana 1951–1960. Athens, OH: Ohio University Press. Resnick, Danielle. 2011. “In the Shadow of the City: Africa’s Urban Poor in Opposition Strongholds.” Journal of Modern African Studies, 49, 1: 141–166. Resnick, Danielle. 2014. Urban Poverty and Party Populism in African Democracies. New York: Cambridge University Press. Riedl, Rachel. 2014. Authoritarian Origins of Democratic Party Systems in Africa. New York: Cambridge University Press. Robinson, David. 1999. “The Murids: Surveillance and Collaboration.” The Journal of African History, 40, 2: 193–213. Robinson, Kenneth. 1960. “Senegal: The Elections to the Territorial Assembly, March 1957.” In W. J. M. Mackenzie and Kenneth Robinson, eds., Five Elections in Africa. Oxford: Clarendon Press. Ronen, Dov. 1974. “Colonial Elite in Dahomey.” African Studies Review, 17, 1: 55–76. Ronen, Dov. 1975. Dahomey: Between Tradition and Modernity. Ithaca and London: Cornell University Press. Rothchild, Donald. 1985. “State-Ethnic Relations in Middle Africa.” African Independence: The First Twenty Five Years. Bloomington: Indiana University Press. Pp. 71–96.

196

References

Sahn, David and David Stifel. 2003. “Urban-rural Inequality in Living Standards in Africa.” Journal of African Economies, 12, 4: 564–597. Samson, Fabienne. 2000. “La Place du Religieux dans L’Election Présidentielle Sénégalaise.” Afrique Contemporaine, 194: 5–11. Schaffer, Frederic. 1998. Democracy in Translation. Ithaca: Cornell University Press. Schaffer, Frederic and Andreas Schedler. 2007. “What is Vote Buying?” In Frederic Schaffer, ed., Elections for Sale: the Causes and Consequences of Vote Buying. Boulder, CO: Lynne Rienner Publishers. Schumacher, Edward. 1975. Politics, Bureaucracy, and Rural Development in Senegal. Berkeley: University of California Press. Scott, James. 1972. “Patron-Client Politics and Political Change in Southeast Asia.” American Political Science Review, 66, 1: 91–113. Scott, James. 1976. The Moral Economy of the Peasant. New Haven: Yale University Press. Seely, Jennifer. 2007. “The Presidential Election in Benin, March 2006.” Electoral Studies, 26: 196–231. Skinner, Elliott. 1975. “Competition within Ethnic Systems in Africa.” In Leo Despress, ed., Ethnicity and Resource Competition in Plural Societies. The Hague: Mouton. Pp. 131–157. Sklar, Richard. 1993. “The African Frontier for Political Science.” In Robert Bates, V. Mudimbe and Jean O’Barr. eds., Africa and the Disciplines: The Contributions of Research in Africa to the Social Science and Humanities. Chicago: The University of Chicago Press. Slater, Dan. 2009. “Revolutions, Crackdowns, and Quiescence: Communal Elites and Democratic Mobilization in Southeast Asia.” American Journal of Sociology, 115, 1: 203–254. Slater, Dan and Daniel Ziblatt. 2013. “The Enduring Indispensability of the Controlled Comparison.” Comparative Political Studies, 46, 10: 1301–1327. Smith, Etienne. 2006. “La nation « par le côté »: Le récit des cousinages au Sénégal.” Cahiers d’études africaines, 4, 184: 907–965. Snyder, Frank. 1965. One-Party Government in Mali: Transition toward Control. New Haven and London: Yale University Press. Somolekae, Gloria and Mogopodi Lekorwe. 1998. “The Chieftaincy System and Politics in Botswana, 1966–95.” In Wayne Edge and Mogopodi Lekorwe, eds., Botswana: Politics and Society. Pretoria: J.L. van Schaik Publishers. Staniland, Martin. 1973. “The Three-Party System in Dahomey: I, 1946–56.” The Journal of African History, 14, 2: 291–312. Staniland, Martin. 1975. The Lions of Dagbon: Political Change in Northern Ghana. Cambridge: Cambridge University Press. Stokes, Susan. 2005. “Perverse Accountability: A Formal Model of Machine Politics with Evidence from Argentina.” American Political Science Review, 99, 3: 315–325. Stokes, Susan. 2007a. “Pork, by Any Other Name: Building a Conceptual Scheme of Distributive Politics.” Manuscript, Yale University. Stokes, Susan. 2007b. “Political Clientelism.” In Carles Boix and Susan Stokes, eds., Handbook of Comparative Politics. Oxford: Oxford University Press.

References

197

Suret-Canale, Jean. 1966. “La fin de la chefferie en Guinee.” Journal of African History, 7, 3: 459–493. Tamari, Tal. 1991. “The Development of Caste Systems in West Africa.” Journal of African History, 32: 221–250. Thachil, Tariq. 2011. “Embedded Mobilization: Nonstate Service Provision as Electoral Strategy in India.” World Politics, 63, 3: 434–469. Thompson, Virginia. 1963. “Dahomey.” In Gwendolen Carter, ed., Five African States: Responses to Diversity. Ithaca: Cornell University Press. Pp. 161–262. Tignor, Robert. 1971. “Colonial Chiefs in Chiefless Societies.” The Journal of Modern African Studies, 9, 3: 339–359. van de Walle, Nicholas. 2001. African Economies and the Politics of Permanent Crisis. Cambridge University Press. van de Walle, Nicholas. 2003. “Presidentialism and Clientelism in Africa’s Emerging Party Systems.” Journal of Modern African Studies, 41, 2: 297–321. van de Walle, Nicholas. 2007. “Meet the New Boss, Same as the Old Boss? The Evolution of Political Clientelism in Africa.” In Herbert Kitschelt and Steven Wilkinson, eds., Patrons, Clients and Policies. Cambridge: Cambridge University Press. Pp. 50–67. Van Kessel, Ineke and Barbara Oomen. 1997. “One Chief, One Vote: The Revival of Traditional Authorities in Post-Apartheid South Africa.” African Affairs, 96: 561–585. van Rouveroy van Nieuwaal, Emile. 1987. “Chiefs and African States: Some Introductory Notes and an Extensive Bibliography on African Chieftaincy.” Journal of Legal Pluralism, 25–26: 1–46. Varshney, Ashutosh. 2002. Ethnic Conflict and Civic Life: Hindus and Muslims in India. New Haven, CT: Yale University Press. Vicente, Pedro and Leonard Wantchekon. 2009. “Clientelism and Vote Buying: Lessons from Field Experiments in African Elections.” Oxford Review of Economic Policy, 25, 2: 292–305. Villalón, Leonardo. 1994. “Democratizing a (Quasi) Democracy: The Senegalese Elections of 1993.” African Affairs, 93: 163–193. Villalón, Leonardo. 1995. Islamic Society and State Power in Senegal. Cambridge University Press. Wallerstein, Immanuel. 1965. “Elites in French-Speaking West Africa: The Social Basis of Ideas.” The Journal of Modern African Studies, 3, 1: 1–33. Wantchekon, Leonard. 2003. “Clientelism and Voting Behavior: Evidence from a Field Experiment in Benin.” World Politics, 55, 3: 399–422. Weghorst, Keith and Staffan Lindberg. 2013. “What Drives the Swing Voter in Africa?” American Journal of Political Science, 57, 3: 717–734. Wilkinson, Steven. 2004. Votes and Violence: Electoral Competition and Ethnic Riots in India. Cambridge, UK; New York: Cambridge University Press. Wimmer, Andreas, Lars-Erik Cederman and Brian Min. 2009. “Ethnic Politics and Armed Conflict: A Configurational Analysis of a New Global Data Set.” American Sociological Review, 74, 2: 316–337. Wing, Susanna. 2008. Constructing Democracy in Transitioning Societies of Africa: Constitutionalism and Deliberation in Mali. New York: Palgrave Macmillan.

198

References

Wing, Susanna. 2013. Mali’s Precarious Democracy and the Causes of Conflict. Washington, DC: US Institute of Peace. Wittenberg, Jason. 2006. Crucibles of Political Loyalty: Church, Institutions and Electoral Continuity in Hungary. Cambridge University Press. Wrong, Michela. 2009. It’s Our Turn to Eat: The Story of a Kenyan Whistleblower. New York: HarperCollins Publishers. Young, Crawford. 1976. Politics of Cultural Pluralism. Madison: University of Wisconsin Press. Young, Crawford. 1982. “Patterns of social conflict: state, class, and ethnicity.” Daedalus: 71–98. Young, Crawford and Babacar Kante. 1992. “Governance, Democracy, and the 1988 Senegalese Elections.” In Goran Hyden and Michael Bratton, eds., Governance and Politics in Africa. Boulder, CO: Lynne Rienner Publishers. Pp. 57–74. Zolberg, Aristide. 1968. “The Structure of Political Conflict in the New States of Tropical Africa.” American Political Science Review, 62, 1: 70–87. Zuccarelli, François. 1988. La vie politique sénégalaise (1940–1988). Paris: CHEAM. Zucco, Cesar. 2008. “The President’s ‘New’ Constituency: Lula and the Pragmatic Vote in Brazil’s 2006 Presidential Elections.” Journal of Latin American Studies, 40, 1: 29–49. Zucco, Cesar. 2013 “When Pay Outs Pay Off: Conditional Cash-Transfers and Voting Behavior in Brazil 2002–2010.” American Journal of Political Science, 47, 3: 810–822.

Index

Ahomadegbe, Justin, 82–83, 91–92 Alliance des Forces de Progrès (AFP), 107, 112 Alliance pour un Bénin Triomphant (ABT), 178 Amoussou, Bruno, 42, 82, 89 And-Jëf/Parti Africain pour la Démocratie et le Socialisme (AJ/PADS), 107, 134 Apithy, Sourou, 82–83, 90, 91–92 authoritarian regimes, 14 Balant ethnic group, Senegal, 60 Bates, Robert, 142 Benin Bariba, 74–75, 95 double-ethnic strategy, politicians views, 85 elections context, 42 elections, media descriptions, 85 electoral data, 178 electoral outcomes, mobilization strategies, 81, 91–95 electoral rules and, 11 eligible voters, right to vote, 59 ethnic composition of electorate, main presidential candidates, 93–94 ethnic politics in, 8–9 ethno-regionalist parties in, 84–86 French colonialism and, 71–74 intermediaries in, 89–90 Kérékou dictatorship, 75–76 local leaders, intermediaries in, 75, 77–79, 90–91 Local Revolutionary Committees, 76

map, 72 minimum-winning coalition theory and, 10–11 mobilization strategies, electoral outcomes, 81, 83–90 political figures in, 82–83 reliance on types of leaders, 177 socially salient identifies, 176 Somba, 74 sub-national differences, 90–91 traditional, religious leaders contemporary strength, 77–79 Beninese National Conference, 82 Bharatiya Janata Party (JBP), 165 Bloc Démocratique Sénégalais (BDS), 98–99 Bolsa Familia, 50 Botswana, 141 colonial period, 155, 156 electoral mobilization, electoral patterns, 158–159 social structure, 155–156 tribal chiefs, traditional authority, 155–156, 158–159 Tribal Land Act of 1968, 156 Botswana Democratic Party (BDP), 141, 158–159 Botswana National Front (BNF), 159 Botswana People’s Party (BPP), 159 Burkina Faso, 141 Campaoré, Blaise, 141 colonialism, 43–45, 71–74, 155, 156 Condé, Alpha, 153 Conombo, Joseph, 36

199

200

Index

Conté, Lansana, 153 Convention People’s Party (CPP), 36 Cramer’s V Ethno-linguistic Voting Index (CVELI), 53–54 Cruise O’Brien, Donald, 67, 109–110 Democratie Citoyenne, 99 Diallo, Cellou Dalein, 153 Dieng, Tanor, 48, 97 Diola ethnic group, Senegal, 60 Diop, Mbaye-Jacques, 101, 102 Diouf, Abdou, 7, 96, 99 electoral mobilization. See mobilization strategies electoral politics (African), normative foundations of analysis, 25–28 electorate diversity, multiethnic coalitions vs., 14–15 ethnic politics material interests vs., 163 trust in local leaders and, 161–163 ethnic politics, 1, 6–7, 166–167 cross-cutting ties and, 12 electoral rules and, 11 ethnic bonds, strategic choice, 1 existing theories of variation in, 8–15 indices, 53–54 joking kinship and, 12–13 multi-party politics and, 26–27 nation-building after independence and, 13–14 over-lapping cleavages and, 11–12 politician’s accountability and, 26 politicization of ethnicity and, 25–26 summary of measures of, 53–56 violence and, 26–27 ethnic stereotypes, 9–10 Ethnolinguistic Fractionalization Index (ELF), 147 Forces Cauris pour un Bénin Emergent (FCBE), 85, 87 French Socialist Party (SFIO), Senegal, 98–99 Ghana, 8, 141 Guinea electoral mobilization, electoral patterns, 152–154

ethnic tensions, 153–154 ethno-linguistic fractionalization (ELF), 147 Politically Relevant Ethnic Group Index (PREG), 147 social structure in, 148–150 hierarchical ties, 35 material control and, 52 moral authority and, 52–53 Senegal, sub-national level, 65–66 Horowitz, Donald, 53 Houngbedji, Adrien, 82, 88 incumbency advantage, 47–51, 114–115, 151 India, 165 Indonesia, 36–37, 164 intermediaries, local leaders. See also specific countries electoral implications, urban vs. rural areas, 34–37, 135–140, 145 importance of, 33–34 party loyalty and, 168–169 politicians use of, 37–39, 167 urban vs. rural areas, 129–135, 167 voter financial dependence, trust, 37–39 Keïta, Ibrahim Boubacar, 140–141 Kenya, 159 electoral mobilzation, electoral patterns, 159–161 ethnic identity, vote choice in, 15 nation-building policies in, 13–14 social structure, 157–158 Kenya African National Union (KANU), 160 Kenyatta, Jomo, 13, 160 Kenyatta, Uhuru, 161 Kérékou, Mathieu, 75–76, 82, 91, 93 Khama, Seretse, 156, 158 Kibaki, Mwai, 15, 160–161 Léopold Sédar Senghor, 96, 98–99 Ligue Démocratique / Mouvement pour le Parti du Travail (LD/MPT), 99–100 Local Revolutionary Committees, Benin, 76 Loi cadre (The Enabling Act), 59 Maga, Hubert, 36, 82–83, 84, 90, 101 Mali, 140–141

Index electoral mobilization, electoral patterns, 150–152 ethno-linguistic fractionalization (ELF), 147 Politically Relevant Ethnic Group Index (PREG), 147 social structure in, 147–148 Sufi orders, castes, 147–148 village-based gerontocracies, 147–148 mass (electoral) politics, 20–21, 45 material control, hierarchical ties, 52 Mbacké, Modou Kara, 120 Mbacké, Serigne Falilou, 112–113, 120 Miller, Norman, 34 minimum winning coalition strategy, 41–42 mobilization strategies. See also intermediaries, local leaders demographics and, 167 electoral outcomes, Benin vs. Senegal, 81, 107–108 empirical evidence, 20–25 ethnic bonds and, 1 ethnic vs. non-ethnic, 45 hierarchical ties and, 17 local leader authority and, 22–23 mobilization modes, electoral patterns and, 49–50 national electoral outcomes and, 18–19 non-programmatic electoral mobilization, 17 predicted dominant modes, 39–40 pre-existing conditions and, 2 social structure and, 1–2, 39–43 Moi, Daniel arap, 160 moral authority, hierarchical ties, 52–53 Mouvement Africain pour la Développement et le Progrès (MADEP), 178 Mouvement de Forces Démocratiques de la Casamance (MFDC), 125–126 Mugabe, Robert, 141 National Domain Law of 1974, Senegal, 77 National Electoral Commission (CENA), Benin, 86 National Transitional Council (CNT), 153 New Patriotic Party (NPP), 163 Niasse, Mamoune, 117, 118, 120 Niasse, Moustapha, 97, 112

201

Nkrumah, Kwame, 40 Nyerere, Julius, 13, 27 Odinga, Raila, 15, 160–161 Parti de la Vérité pour le Développement (PVD), 107 Parti de l’Independence et du Travail (PIT), 100, 107 Parti Démocratique de Guinée (PDG), 149, 152 Parti Démocratique Sénégalais (PDS), 96–97, 107 Parti des Nationalistes Dahoméens (PND), 91–93 Parti du Renouveau Démocratique (PRD), 91–92 Parti Social Démocrate (PSD), 93 Parti Socialiste (PS), 96, 136–137 Parti Socialiste du Bénin (PSB), 85 Parti Soudanais Progressiste (PSP), 150 Party Nationalization Scores (PNS), 53–54 Philippines, 164–165 Politically Relevant Ethnic Group Index (PREG), 6, 147 Rassemblement du Peuple de Guinée (RPG), 153 Rassemblement Démocratique Dahoméen (RDD), Benin, 91–93 Rassemblement du Peuple (RP), 117 Rassemblement National Démocratique (RND), 107 Rassemblent Démocratique Africain (RDA), 148 Rawlings, Jerry, 36, 141 Renaissance du Bénin (RB), 85 Sall, Macky, 97, 118, 123–124 Sankara, Bénéwendé, 141 Seck, Idrissa, 97, 118, 120 Senegal Balant, 60, 69–71 brotherhood affiliation in, 9–10 democratic practice, 97 Diola, 60, 69–71, 125–126 elections context, 96–97 electoral data, 129 electoral dynamics, 107–111 electoral rules, 11 electorate diversity in, 6

202

Index

Senegal (cont.) eligible votes, right to vote, 58–59 ethnic composition of electorate, main presidential candidates, 122 ethnic politics in, 8–9 Grand Magal (2007) hierarchical ties, sub-national level, 65–66 incumbent vote share, 136–140 intermediaries’ influence, 98–104, 126–127 intermediaries’ material rewards, government infrastructure projects, 108–111 intermediaries’ support, logic thereof, 112–121 intermediary families, 120–121 local leaders in, 69–71, 77–79 map, 61 marabouts, 61–63, 66, 67, 69 mobilization strategies, electoral outcomes, 81, 98–103, 121–126 Mourides vs. Tijanis, incremental variation in role of intermediaries, 104–107 National Domain Law of 1974, 77 political figures, 96–97 reliance on traditional, religious leaders, 77–79 reliance on types of leaders, 177 religious/brotherhood composition, main presidential candidates, 93–94 religious elites, Sufi hierarchy, 60–63 ruling elites, post-independence period, 76–77 rural areas, systemic bias against, 143–145 Senegalese social contract, 77 Serer aristocrats, social hierarchy in, 64–65, 69–71 socially salient identifies, 175 social organization, 59–60 Toorodo revolution, 69 traditional elites, caste-based aristocracy, 63–65 traditional elites primacy, Northern Senegal, 68–69 traditional, religious leaders contemporary strength, 77–79

Tukulor elites, toorobe, 63, 64, 69 voters’ education, 142 voting orders, from intermediaries, 103–104 Wolof elites, 66, 69–71 Senghor, Léopold Sédar, 119, 121–122, 136–137 social structure accounting for variation in, 43–45 electoral patterns, outcomes and, 45–47 implications of variation in, 39–43 onset of mass politics and, 45 origins of, 57–58 persistence, change in, 169–170 voter treatment, access to resources, 171 Soglo, Nicéphore, 82 Somba (Benin), 74 state strength, 51 Sufi hierarchy, Senegal, 62 Tanzania, nation-building policies, 13–14 Tchané, Abdoulaye Bio, 86 Thailand, 164 Touré, Ahmed Sékou, 147, 148–150, 152 Touré, Amadou Toumani, 140–141, 151 Uganda, 8 Union Démocratique Dahoméenne (UDD), Benin, 91–93 Union fait la Nation (UN), Benin, 86 Union des Forces Démocratiques de Guinée (UFDG), 153 Union Progressiste Sénégalaise (UPS). See Parti Socialiste (PS) Union Soudanaise (US), 150 urbanization, Africa, 32 Wade, Abdoulaye, 7, 48, 96–97, 100–103, 110, 123–124, 136, 138–140 West Africa, founding elections in, 58–59 Yayi Boni, 83, 86, 94–95, 170 Zimbabwe, 141 Zinsou, Emile, 84