Power Sharing and Power Relations After Civil War 9781626377790

There are numerous studies on the role of power-sharing agreements in the maintenance of peace in postconflict states. L

187 116 3MB

English Pages 257 [268] Year 2022

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Power Sharing and Power Relations After Civil War
 9781626377790

Citation preview

Power Sharing and Power Relations After Civil War

Power Sharing and

Power Relations After Civil War edited by

Caroline A. Hartzell Andreas Mehler

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

Published in the United States of America in 2019 by Lynne Rienner Publishers, Inc. 1800 30th Street, Boulder, Colorado 80301 www.rienner.com and in the United Kingdom by Lynne Rienner Publishers, Inc. Gray’s Inn House, 127 Clerkenwell Road, London EC1 5DB

© 2019 by Lynne Rienner Publishers, Inc. All rights reserved

Library of Congress Cataloging-in-Publication Data A Cataloging-in-Publication record for this book is available from the Library of Congress. ISBN 978-1-62637-767-7 British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library. Printed and bound in the United States of America The paper used in this publication meets the requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992. 5

4

3

2

1

In memory of Chandra Lekha Sriram Dedicated scholar, generous friend and colleague, fierce advocate for justice

Contents

ix

List of Tables and Figures 1 Power Sharing and Power Relations in Postconflict States Andreas Mehler and Caroline A. Hartzell

1

Part 1 The Impact of Power Sharing on Power Relations 2 Government-Rebel Relations in the Wake of Power-Sharing Peace Agreements Martin Ottmann and Johannes Vüllers

19

3 The Transformation of Armed Organizations into Political Parties John Ishiyama

47

4 The Consequences of Power Sharing at the Local Level Andreas Mehler, Claudia Simons, Denis M. Tull, and Franzisca Zanker

67

Part 2 Power-Sharing Mechanisms at Work 5 Military Power Sharing: The Case of the Philippine Peace Agreement Rosalie Arcala Hall

vii

89

viii

Contents

6 Territorial Power Sharing: The Cohesion of Opposition Movements Kathleen Gallagher Cunningham

109

7 Economic Power Sharing: Potentially Potent . . . but Likely Limited Caroline A. Hartzell

125

Part 3 Power Sharing and the Quality of the Peace 8 Government Respect for the Physical Security of Postconflict Populations Matthew Hoddie

149

9 Shifting Public Attitudes? Power Sharing and Intergroup Tolerance Bernadette C. Hayes and John Nagle

169

10 Transitional Justice: Promoting or Hijacking Elite Accountability? Chandra Lekha Sriram

191

Part 4 Conclusion 11 The What, How, Where, and Who of Postconflict Power Sharing Caroline A. Hartzell and Andreas Mehler

215

Bibliography The Contributors Index About the Book

227 245 247 257

Tables and Figures

Tables 2.1 Descriptive Statistics for Practices of Power Sharing 2.2 Competing-Risks Coefficients for Transition to Nonviolent Power Relations 2.3 Competing-Risks Coefficients for Transition to Inclusive Power Relations 2.4 Competing-Risks Regression with Independent Variables Interacted with Time 2.5 Results of Endogeneity Test for Power-Sharing Practices 2.6 Comparison of Hypotheses and Empirical Findings 3.1 List of Fifty-Two Cases of Rebel Parties 3.2 Hazard Ratios for Rebel Party Splits After End of Civil War 3.3 Hazard Ratios for Rebel Party Splits After End of Civil War, Peace Agreement Cases 4.1 Effects of Power-Sharing Properties Down to the Local Level 6.1 Effects of Concession on Group Fragmentation 6.2 Fate of Inactive Factions Five Years After Concessions 8.1 Post–Civil War Power Sharing and Government Respect for Physical Integrity Rights, Aggregate Measure, 1984–2006 8.2 Post–Civil War Power Sharing and Government Respect for Physical Integrity Rights, Individual Measures, 1984–2006 9.1 Consociational Forms of Governance in Northern Ireland and Lebanon 9.2 Religious Affiliation in Northern Ireland and Lebanon

ix

22 35 37 41 43 44 57 61 62 85 120 122

162

164 174 177

x

Tables and Figures

9.3 Religious Affiliation, Behavior, and Conviction in Northern Ireland and Lebanon 9.4 Attitudes Toward Religious Intermarriage in Northern Ireland 9.5 Religious Affiliation and Obstacles to Interfaith Marriage in Lebanon 9.6 Religious Denomination and Attitudes Toward Intermarriage in Northern Ireland 9.7 Religious Affiliation and Perceived Obstacles to Marriage in Lebanon 10.1 Transitional Justice, Power Sharing, and Power Relations

178 180 180 183 183 210

Figures 2.1 Stacked Cumulative Incidence of Government-Rebel Relations for Different Power-Sharing Settings 3.1 Party Split Hazard over Time 3.2 Party Split Hazard over Time, Agreements Only 6.1 Fragmentation of Self-Determination Movements 7.1 Power Distribution by Socioeconomic Position and Social Group in Burundi 7.2 Power Distribution by Socioeconomic Position and Social Group in Liberia 7.3 Power Distribution by Socioeconomic Position and Social Group in the Republic of the Congo 7.4 Power Distribution by Socioeconomic Position and Social Group in Uganda 7.5 Power Distribution by Socioeconomic Position and Social Group in Sierra Leone 7.6 Power Distribution by Socioeconomic Position and Social Group in South Africa 9.1 Levels of Marital Homogamy in Northern Ireland, 1968–2012 9.2 Attitudes Toward Religious Intermarriage in Northern Ireland, 1989–2012

39 63 65 117 137 138 139 140 141 142 179 181

1 Power Sharing and Power Relations in Postconflict States Andreas Mehler and Caroline A. Hartzell

Power sharing in postwar countries has been the subject of growing attention by scholars and policymakers during the past two decades. Despite increased interest in power sharing as a means of ending intrastate conflicts, various dimensions of this complex set of institutional arrangements have yet to be well elucidated. Most scholarship focuses on the question of whether power sharing is able to help prevent war recurrence and achieve “peace” (Schneckener 2002; Hartzell and Hoddie 2003; Mukherjee 2006a; Joshi and Mason 2011). Although this is an important question, it is not the only one deserving of analysis. Furthermore, the responses to this question by different authors—who have employed varying concepts of power sharing and peace, indicators, selection of cases, and observation periods as well as differing statistical techniques—have been so diverse that it has become difficult to find any common ground (Binningsbø 2013). Notwithstanding these differences, what few scholars would contest is that power sharing has both positive and negative effects, many of which are likely to have been unintended by the architects of these measures. Power sharing might, for example, strengthen an elitist approach to politics, or it might create incentives for new actors to take up arms, or it might prove an impediment to the process of healing the wounds of violent conflict as the perpetrators of atrocities gain positions of power and influence over government affairs (Mehler 2009; Cheeseman 2011; Wolff 2009; Sriram 2008). Alternately, power sharing might lead to the creation of new identities, enhance the capacities of some groups of actors in unexpected ways, or provide actors with incentives that enable them to participate more effectively in postconflict politics. One thing is sure: the failure to investigate these types of effects thoroughly means that it is difficult to determine whether outcomes such as the durability of the peace are attributable to the core dispositions of power-sharing pacts themselves, as has been claimed, 1

2

Andreas Mehler and Caroline A. Hartzell

or are the product of heretofore unexplored changes in power relations engendered by power sharing. Accordingly, a central goal of this book is to extend the analysis of power sharing beyond the role that power-sharing institutions play in the duration of the peace to the potential that they have to impact the balance of power within, between, and among actors, groups, and institutions in the postconflict state. Power-sharing institutions may, as one of their by-products, generate incentives that encourage adversaries to keep the peace. However, such institutions are, first and foremost, a set of rules designed to apportion state power among a number of actors. As such, power-sharing institutions have the ability to alter preexisting power balances. Power-sharing arrangements can change the relative strength or influence wielded by key actors in the postconflict environment as well as affect the means by which they exercise power or influence. These types of transformations in the balance of power among actors are likely to have consequences—some short term and others more durable in nature, some intended and others not—for relations among actors in the postconflict state. Some of these consequences may, as noted above, have an impact on the duration of the peace. Others will certainly have an influence on the quality of the peace. Because power-sharing institutions are now one of the principal tools used in efforts to end civil wars, it is incumbent on those who work in the field of conflict management to know more about the consequences that follow from the use of these measures.1 This book seeks to advance understanding of this issue by developing a framework for the analysis of power relations among a variety of actors in the period that follows a power-sharing agreement. The next three sections of this chapter lay the groundwork for this framework. First, we highlight gaps in current knowledge regarding the types of effects power sharing has on postconflict power relations. Second, we focus on the influence that the modalities of power sharing—what, how, with whom, and where power is shared—may have on power relations. Third, we outline the central components of the analytical framework. Finally, we conclude with an overview of the contents of the book.

Bridging the Gaps: Power Sharing and Postconflict Relations As noted above, the extant literature on power sharing generally has neglected the potential that power sharing has to transform postconflict relations in various ways that could have an impact on war recurrence, the transition to democracy, or other important aspects of political life. In this section, we highlight major gaps in the research on power sharing and postconflict relations that this book seeks to bridge.

Power Sharing and Power Relations in Postconflict States

3

First, existing studies have focused on the design of only a few formal postagreement institutions such as electoral systems and government systems that are thought to help stabilize the country and facilitate a transition to democracy.2 However, researchers of power sharing have largely neglected other formal postconflict institutions influenced by power-sharing arrangements. These institutions include the security sector (generally the focus of a more technical literature), justice (the subject of works on transitional justice and rarely related to other institutions), and party regulations and systems (Reilly 2013).3 We know little, for example, about the impact that former rebel parties have on the development of the party system. Postconflict election victories by rebel parties in Burundi and Nepal, both countries in which power-sharing arrangements were designed as part of the process of ending civil wars, illustrate the relevance of this topic. Accordingly, some of the chapters in this book seek to highlight the transformation of institutions in the wake of a postconflict power-sharing arrangement. Second, there currently is virtually no information regarding how powersharing agreements transform power resources or the means by which actors wield power and influence within a country. Two issues about which relatively little is known, for example, are the consequences that follow from military and economic forms of power sharing. Existing studies on military power sharing have highlighted the importance of military integration as a credible commitment or argued that economic opportunities offered by disarmament, demobilization, and reintegration (DDR) programs explain the willingness of rebel groups to consent to military integration (Hoddie and Hartzell 2003; Glassmyer and Sambanis 2008). They have not, however, considered the role of mid-level commanders who could spoil a DDR process or the consequences of failed implementation of power-sharing arrangements on the rank-and-file level.4 Research on economic power sharing, which is surprisingly limited in nature, has concentrated on the group level and the distribution of wealth with a special focus on resources (Binningsbø and Rustad 2012; Lujala and Rustad 2012). The possibility that actors who are given access to economic power as part of a power-sharing agreement may use it for patronage or for personal gain has hardly been discussed in the literature, nor has much attention been given to the implications such arrangements may pose for shifts in economic power from the political center to the regions of the country. As the chapters on military and economic power sharing in this book seek to make clear, understanding what benefits different actors believe they will derive from different forms of power sharing, as well as the extent to which such arrangements actually deliver on those expectations or not, is important for understanding the consequences that stem from the use of power sharing. Third, the link between issues of representation and inclusion, both affected by power-sharing deals, has not yet been sufficiently explored.

4

Andreas Mehler and Caroline A. Hartzell

Whether or not settlements make explicit reference to the matter, elites are included in power-sharing institutions in the name of representation of a group with grievances. This is particularly true where prominent members of a rebel movement are concerned. However, it is far from clear whether these individuals perform better than political parties and elected members of parliament in bringing the concerns of their constituents to the table. This raises an important question regarding the extent to which power sharing fosters group representation, a topic that has yet to be empirically explored. Another issue that merits analysis is power sharing’s impacts on the inclusion of social groups in the political system. This mostly pertains to identity groups, some of them of rather recent origin, but not in all cases: sometimes social stratification and exclusion have ossified to a point where the inclusion of lower strata in the political system has become a main goal of rebel movements (e.g., Maoist rebels in Nepal) and, therefore, is high on the agenda of power-sharing negotiations. The process of selecting the movements that will sit at a negotiation table is one in which power is attributed to some groups, which most likely will be part of the power-sharing arrangement, and not others (Nilsson 2008). Inclusion of one group therefore may equate with the exclusion of another (Tull and Mehler 2005; Jarstad 2008). Power relations within rebel organizations or between rebel organizations and civilian political parties are affected by the scope of inclusiveness, which may indirectly affect the broader representation of group interests (Mehler 2011). Fourth, we know little about the ways in which actors themselves can be transformed by power-sharing agreements. To the extent that existing studies have dealt with this issue at all, it has been to focus on the impact that spoilers have on peace processes or the duration of the peace (Nilsson and Söderberg Kovas 2011). One of the topics that may well have an impact on postconflict relations, for example, includes the degree of cohesion demonstrated by the conflict parties in the postwar period. That the fragmentation of rebel groups is of relevance during the course of conflicts is not an issue that is disputed (K. G. Cunningham 2011). However, the fragmentation of groups in postconflict periods and the influence that power-sharing arrangements may have on group cohesion following the end of civil wars are issues that have been subject to little theorizing or empirical analysis. Our knowledge regarding how and why these changes occur is minimal. Information regarding the potential for power-sharing arrangements to produce these types of transformation of actors is particularly important since such changes could produce new conflicts within the various parties as has been highlighted in other research (Kalyvas 2003; De Juan 2013). Finally, more information is needed regarding the ways in which power-sharing agreements transform the wider society. The literature on postwar societies is broad ranging in nature and has not, to our knowledge,

Power Sharing and Power Relations in Postconflict States

5

theorized the impact that power-sharing arrangements have on changes in postconflict societies (Jarstad and Sisk 2008; Roeder and Rothchild 2005). Topics worth studying include, for example, the possible creation of new identities (ethnic, regional) and of subgroups as a result of power-sharing arrangements. A thorough analysis of rebel-to-party transitions could also help to enhance our understanding of the development of new societal identities and social cleavages in postconflict countries. Yet another example of potential societal transformations that may be induced by power sharing is long-run conceptions of citizenship and empowerment.

The Modalities of Power Sharing and Changes in Power Relations If power sharing really means the sharing of power, it is important to know exactly what is being shared, how it is shared, with whom, and where the locus of power sharing takes place.5 Each of these modalities of power sharing has the potential to produce a different type of effect on the balance of power among actors. Because conflict-ending settlements often differ in the manner in which they address each of these issues, it can be difficult to predict exactly what types of changes in power relations a settlement may produce. Nevertheless, an awareness of the types of changes in power relations that may stem from each of these aspects of a power-sharing agreement can help actors to better think through their consequences for power relations in the short, medium, and long term. What

The question of “what” power consists of or what type of power is being shared is far from trivial. In numerous cases of contemporary political settlements, one may suspect that power sharing is principally about the sharing of spoils (and, hence, of material rewards), while power is commonly seen as something different—but what, exactly? Power may be a less-thanclear concept if we examine its Latin epistemology as well as social scientists’ use of the term. In our attempt to define the concept in what follows, particularly in relation to our contributors’ use of the term power sharing, we concentrate on those perspectives that speak to the ambition of this book. The Latin word potestas stands for the legal and symbolic aspects of power (i.e., the formal entitlement to give out orders); the Latin word potentia refers to the disposal of means or instruments of power. While both understandings may resonate with what is meant by power sharing, it is probably the second aspect that is more important, particularly in a postconflict setting. Responding to the “security dilemma” with a power-sharing

6

Andreas Mehler and Caroline A. Hartzell

deal should give all conflict parties enough instruments of power to be able to avoid risks to the survival of a group or of an organization.6 In early modern times, Niccolò Machiavelli developed a different connotation of power, abstracting from legal underpinnings and normative justifications and moving power to the center of the ambitions of all true politicians. This understanding may play a role in the zero-sum game thinking that is deeply inscribed in most power-sharing deals. Sharing power in a Machiavellian sense would have to be understood—from the perspective of The Prince—as losing power. Such an interpretation, however, abstracts from contextual conditions and habitual behavior. It might be perfectly rational and far from undesirable to share power as long as this can be interpreted as a deliberate act of delegation. In patrimonial systems, this sort of asymmetrical relationship between the top of the system and second-rank “barons” is even constitutive for the power position at the top. This understanding of power and its relation to power sharing may be of particular relevance given the fact many of the states engaging in power-sharing deals nowadays have a strongly patrimonial political culture (Bayart 1989; Chabal 1992). German sociologist Max Weber made the influential distinction between power (macht) and authority (Herrschaft), with the latter referring to a form of power that consists of having legitimacy (in distinct ways), while power itself is defined as “the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability rests.”7 This definition stresses the relational aspect of power, a conceptualization of power that is obviously central to a book focusing on changes in power relations stemming from the use of power sharing. Implicitly, under Weber’s definition of power, those giving in may not be completely devoid of power resources even if the balance of power does not favor them in a given situation. This suggests that the ability of some actor to wield absolute power is highly unlikely, particularly as any form of resistance on the part of other actors implies that they are able to employ some element of power. Power relations change when an actor becomes involved in the comanagement of state affairs; they also change when resistance grows or diminishes. Although frequently juxtaposed, these Weberian elements of power are, it should be noted, quite compatible with Michel Foucault’s concept of power.8 How

How is power shared within political settlements? Most efforts to answer this question have focused on negotiated elite pacts, which are frequently facilitated by international mediators and become part of a written agreement. These pacts are formal in nature with signed texts committing all signatories to fulfill promises. Power-sharing measures that are part of such

Power Sharing and Power Relations in Postconflict States

7

agreements may even become enshrined in interim or permanent constitutions, thus further formalizing power sharing. However, the promises enshrined in these agreements are not always fulfilled. How much power is actually shared in the aftermath of a peace deal and following the enactment of a new constitution are empirical questions that cannot be answered solely by relying on the letter of the agreement or the constitution. A second reflection regarding the manner in which power is shared pertains to the formality of power sharing: recurrent practices form informal institutions. This may link up best with the more radical understandings of power as omnipresent and not limited to state actions as stressed by Foucault (1978, 1991). According to Foucault, power is not confined to the field of politics and is exercised throughout the social body; many other sociological power theorists would be in conformity with this view. However, such a wide understanding of power may not help in detecting when informal power sharing aimed at achieving or consolidating peace takes place. Some working ethnic or religious quota arrangements may not have been spelled out, for example, but are respected and clearly constitute an institution. Additionally, it is also possible that informal power-sharing arrangements are at work at the subnational or local levels where fighting has taken place during wartime and where opposing camps have to live together again, even though a document has never been signed by the groups in question. With Whom

Who shares power with whom is also an important question. A rebel movement’s recognition as a party to a negotiation process confers some form of power on the group not only as veto players at the negotiation table, but also with respect to its relations with the constituency that it supposedly represents. However, all conflict parties do not always become part of a peace deal. Settlements that are not inclusive have consequences for the duration of the peace, but also for power relations between those actors involved in the pact as well as between those in and out of the pact. More intriguingly, power sharing may have impacts within organizations, altering power relations between those who sign a pact and profit from it, and those excluded from the deal. For instance, while the political wing of a movement is likely to gain power in peacetime, the military wing of the same organization may get sidelined during the implementation of a power-sharing deal. This has occurred quite often within rebel movements, prompting scholars to produce a distinct branch of research focusing on splits within such groups (K. G. Cunningham 2011; Bakke, Cunningham, and Seymour 2012). The identity of the actors involved in power-sharing agreements has not always been fully elucidated in much of the literature. For example,

8

Andreas Mehler and Caroline A. Hartzell

ethnic or religious groups frequently have been equated with parties or rebel movements even though the latter may only have usurped a certain identity marker. This raises an interesting question: Just what is or should be the relevant unit of analysis when focusing on power sharing? Is it individuals? Organizations (rebel movements vs. governments)? Groups? Much of the consociational literature has focused on the group level while most of the (data-driven) work on postconflict power sharing has employed an organizational level. Although each of these approaches may be justified, it should be made clear that these are different units of analysis and that each has different implications for postconflict power relations. What are some of the ways in which the type of actor that is the focus of a settlement matters for power relations? In some peace negotiations, the elite nature of power sharing reigns supreme, with little or no attention given to group grievances. In these instances, the distribution of top positions becomes key, with the agreement specifying which individual gets what. In other cases, power-sharing constitutions (of consociational/corporatist or centripetalist/liberal forms) focus on group representation and minority rights, making the accommodation of group grievances and interests the primary goal of the arrangement.9 Finally, power sharing also has played a role in pacts that are supposed to help terminate an authoritarian system, with the balance of power between state and society strongly altered in favor of the latter. Where

The question of where power is shared can be understood spatially or according to the levels of a government system. Spatially speaking, some agreements are meant to fix the composition of a governing elite in the capital of a given country, with minimal if any attention given to the periphery of the country.10 The texts of other peace agreements, however, go to some length to specify a quota for local government institutions, thereby aiming to have an impact on (local) peace as well as on local power relations (Simons et al. 2013). More abstractly, power sharing also has been conceptualized as rules that allocate power on a sectoral basis. Four sectors or bases of state power have been identified as realms in which power can be distributed among actors. Political power sharing is concerned with proportionality in the distribution of central government authority, with collectivities guaranteed a degree of representation within state institutions as a function of their membership in a group. The strategies that can be used toward this goal are electoral proportional representation, administrative proportional representation, and proportional representation in the central government’s executive branch. Military power sharing distributes authority within the state’s coercive apparatus. This can be accomplished by bringing together adversaries’

Power Sharing and Power Relations in Postconflict States

9

armed forces within a unified state security force; by appointing members of the subordinate group or groups to leadership positions in the state’s military; or, in rare cases, by letting opposing sides keep their weapons to maintain their own security forces. Territorial power sharing divides authority among levels of government by creating forms of decentralization based on territory.11 Finally, economic power sharing provides groups in divided societies access to or control of state resources by distributing wealth, income, natural resources, or production facilities on the basis of group identity (Hartzell and Hoddie 2007). Each of the foregoing types of power sharing has the potential to shape power relations in different ways. Power sharing that allocates positions at the political center to the representatives of groups that formerly have been excluded from power could succeed in producing a sense of inclusion based on a rough balance of political power between majority and minority groups. Political power sharing of this nature would not, however, necessarily be expected to alter power relations between elite representatives and their followers. Military power sharing that calls for integrating former rebel troops into the state security forces could serve as a check on the state’s power to use coercive force against the population, thereby potentially altering power relations between the state and society. Economic power sharing could facilitate rent seeking by individuals placed in positions of control over state economic resources, thereby economically empowering certain elites. Alternately, some forms of economic power sharing, particularly if used in conjunction with territorial power sharing, could alter the balance of power between the central government and certain regions of the country.

Central Components of an Analytical Framework This book seeks to move beyond the current debate on power sharing that focuses on whether or not power-sharing arrangements contribute to the peace. This question is so large in scope that even a definitive yes-or-no answer is likely to be of little help to those seeking advice on how to construct an effective and just power-sharing agreement. We believe that by asking, and answering, more specific questions regarding the nature of the power that is shared, how power is shared, who shares power with whom, and in what areas or sectors it is shared, we can learn more about the mechanics of how power sharing works. Once we have a better understanding of how power sharing alters power relations, we should be better able to comprehend the potential that power-sharing measures have to affect a variety of outcomes, including the duration and quality of the peace. A key component of the analytical framework used in this book is the identification of the types of power sharing employed as part of the conflict

10

Andreas Mehler and Caroline A. Hartzell

settlement (i.e., political, military, territorial, or economic). The authors of the chapters each focus on one or more of these forms of power sharing with the goal of understanding the effects that they have on changes in power relations. The second central component of the analytical framework that we employ in this project is the use of levels of analysis. Power relations can be conceptualized in terms of the balance of power that exists within, between, or among actors. The particular levels of analysis used by each of the chapter authors vary depending on whether the author or authors seek to determine how power sharing affects: • the balance of power within a unit—that is, a group (e.g., a religious community or an ethnic group), an organization (e.g., a rebel organization or political party), the government apparatus (e.g., the executive, legislative, and judicial branches of government; the levels of government from national to local), or a territorial unit; or • the balance of power between/among groups (e.g., between religious confessions), organizations (e.g., the national government and rebel groups), state and societal actors, and the central government and subnational units of government. While some of the contributors to this volume employed qualitative methods and others used quantitative tools, a common analytical framework as well as a set of central questions informs the research process in each chapter and helps to ensure that the works cohere around the issue of changes in power relations. Each chapter thus describes the actor or set of actors within or among which a relationship of power exists, noting why this type of power relation is important; identifies the mechanisms via which one or more forms of power sharing shape power relations within or between/among the actors in question; and discusses the short-term versus the long-term impact that power sharing has on the nature of and shifts in power relations within or among the actors.

Organization of the Book We structured the book along the following lines. In Part 1, three chapters focus on the effects power sharing has on power relations between, among, and within groups and levels of government in countries that have experienced internal armed conflict. Part 2 consists of three chapters, each of which explores means by which one form of power sharing—military in the first chapter, territorial in the second, and economic in the third—has the potential to shape various aspects of power relations in postconflict coun-

Power Sharing and Power Relations in Postconflict States

11

tries.12 In Part 3, three chapters bring together the book’s foci on types of power sharing and their effects on power relations to explore how these factors influence rights, representation, and inclusion, each of which has the potential to affect the duration and the quality of the peace in states emerging from violent intrastate conflict. Part 4 concludes with a chapter in which we assess central findings, specify the contributions that the book makes to the study of power sharing’s effects on power relations in postconflict states, and consider potential policy implications. Part 1: The Impact of Power Sharing on Power Relations

Part 1 of the book focuses on the potential that power-sharing institutions have for transforming power relations between and among a variety of groups in post–civil war states. In Chapter 2, Martin Ottmann and Johannes Vüllers examine the effects that the implementation of political and military forms of power sharing have on the balance of power between the government and rebel groups following the end of a civil war. Employing a crossnational dataset, they found that the implementation of some variants of power sharing, but not others, fosters more politically inclusive power relations by providing rebel groups with space in the political system. Although Ottmann and Vüllers conclude that power-sharing institutions can help to alter power relations between the government and rebels from ones characterized by a zero-sum logic to one that is positive-sum in nature, they also emphasize that not all power-sharing measures may generate the effects that they are intended to produce. In Chapter 3, John Ishiyama analyzes the manner in which power sharing affects the balance of power within political parties established by former rebel groups. Different types of power-sharing measures, Ishiyama argues, generate different types of pressures for rebel groups transforming into political parties. One of the unanticipated effects that some types of power-sharing measures may have, he posits, is to aggravate potential fault lines that exist within rebel political parties in such a manner as to increase the likelihood that the parties will split. Ishiyama tested this proposition using data from fifty-three armed rebel organizations that transformed into political parties to contest elections following a civil conflict and found that territorial power-sharing arrangements in particular promote organizational splits in rebel parties in postconflict politics. In Chapter 4, Andreas Mehler, Claudia Simons, Denis M. Tull, and Franzisca Zanker focus on the potential that power sharing has to reorder local power relations. Observing that power sharing is usually introduced in the form of a pact signed by national elites who are often distant from where conflicts originated or were fought, they argue that many features of elite-centered peace agreements will have unexpected impacts on local

12

Andreas Mehler and Caroline A. Hartzell

power relations. Elites included in power-sharing agreements are frequently well anchored in the local arena and, thus, play a two-level power game: they gain influence locally from their position at the national level while drawing critical support from their area of origin. Nevertheless, note the authors, some local elites may stand to lose critical local influence if they accept positions in the national government. Based on their analysis of the cases of Burundi, the Democratic Republic of Congo, Kenya, and Liberia, Mehler, Simons, Tull, and Zanker conclude that while some stakeholders gain locally from measures that call for sharing political power at the national level and territorial power sharing at the local level, others stand to lose influence. Part 2: Power-Sharing Mechanisms at Work

The second part of the book investigates the means by which power-sharing measures shape power relations. Rosalie Arcala Hall’s contribution, Chapter 5, employs a case study, the 1996 Final Peace Agreement between the Philippine government and the Moro National Liberation Front (MNLF), to map the shifts in state-society relations that took place as a result of the agreement’s provision for military power sharing. Military power sharing called for the absorption of 7,500 ex-rebels and their proxies into mixed units of the Philippine army and police, which were then deployed for internal security missions in Mindanao. Hall found that the merger enhanced the state’s influence over Muslim communities by improving the army’s performance in nation-building projects and symbolic inclusiveness as an institution. In turn, the MNLF’s leadership initially was strengthened by its selective distribution of integration slots to loyal commanders, although that gain in power was later diluted by factional rivalries. In Chapter 6, Kathleen Gallagher Cunningham examines the effects that territorial power sharing has on opposition movements active within self-determination disputes. She argues that although conventional wisdom suggests that territorial concessions by governments to opposition movements are expected to produce splintering within the groups, territorial power sharing is actually likely to lead to a short-term increase, but longerterm decrease, in movement fragmentation. Using large-n statistical analysis, Cunningham found that territorial power sharing decreases fragmentation, an outcome she attributes to the fact that many groups, satisfied with concessions they have received, cease to press self-determination claims. In Chapter 7, Caroline A. Hartzell explores the potential that economic power sharing has to alter power relations in states emerging from civil war. She posits that economic power sharing has the potential to induce three different types of shifts in power relations: it could change the balance of power between regions and the central government, it could alter

Power Sharing and Power Relations in Postconflict States

13

power relations horizontally or among identity groups, and it could affect the power of individual political elites. Hartzell observes that state actors have demonstrated a reluctance to use economic power sharing as a means of ending civil wars and, in those instances in which they have agreed to do so, often have been lackadaisical in its implementation. Additionally, global economic forces and the actions of international actors have frequently served to limit the economic power sharing measures’ potential impact on power relations. Accordingly, Hartzell concludes that the type of change in power relations that economic power sharing is most likely to produce is to add to the power of individual elites. Part 3: Power Sharing and the Quality of the Peace

Part 3 of the book examines the impact that power-sharing measures have on rights, representation, attitudes, and inclusion in postconflict states, factors that have the potential to shape the quality of the peace. In Chapter 8, Matthew Hoddie considers the relationship between the adoption of power sharing following civil war and a government’s respect for the physical security of the population. In particular, he investigates the common expectation that power sharing diminishes a government’s respect for the principle that individuals should be free from state aggression in the forms of extrajudicial killing, torture, political imprisonment, and disappearances. Based on cross-national statistical analysis, Hoddie found that there is only limited and inconsistent support for the view that the adoption of powersharing institutions diminishes a government’s respect for physical integrity rights. Hoddie concludes that power sharing is only one factor among many that determine a postwar government’s commitment to respect the rights of citizens to be free from these different forms of harm. Bernadette C. Hayes and John Nagle examine the impact that power sharing has on the balance of power within and outside religious confessions in Chapter 9. Hayes and Nagle note that while some scholars believe consociational power-sharing arrangements can be effective in helping to manage identity conflicts, others contend that they entrench and perpetuate divisions as well as marginalizing or facilitating the targeting of minority groups outside the dominant cleavage. In an effort to engage with this debate, Hayes and Nagle concentrate on public attitudes toward intermarriage, long considered the most salient indicator of communal division, in Northern Ireland and Lebanon. Focusing on the implementation of two different forms of consociational power-sharing agreements—the “liberal” 1998 Belfast Agreement in Northern Ireland and the “corporate” 1988 Taif Agreement in Lebanon—they found that the type of power-sharing arrangement has a differential impact on attitudes toward intermarriage. While public tolerance toward interfaith marriages is notably greater in Northern

14

Andreas Mehler and Caroline A. Hartzell

Ireland than it is in Lebanon, it is also considerably higher than support for intermarriage across other identity groups. In Chapter 10, Chandra Lekha Sriram considers the effects of transitional justice processes on power relations and, potentially, on power-sharing arrangements themselves in states emerging from violent conflict. In particular, Sriram seeks to assess the widespread assumption that transitional justice processes may alter power relations within a state by removing, delegitimizing, or otherwise changing the incentives faced by relevant political or military actors. Focusing on the use of international or internationalized criminal tribunals in African countries, Sriram found that while there is limited evidence that transitional justice can reshape power relations by removing key perpetrators, it may also fail to reshape power relations or may have unexpected consequences. Among the latter, transitional justice may create incentives for the accused to embed themselves further in political power, may be hijacked by political or military actors seeking to gain the upper hand, or may cause the latter actors to choose to engage selectively with it. Part 4: Conclusion

In Chapter 11, the concluding chapter of the book, Caroline A. Hartzell and Andreas Mehler assess the central findings that emerge from the analyses, identifying how they contribute to the study of power sharing’s effects on power relations in postconflict states. Hartzell and Mehler consider the utility and the limitations of the analytical framework employed in the book for the study of those effects on power relations and identify factors that appear to influence the manner in which power sharing shapes postconflict power relations. The chapter concludes with thoughts regarding policy implications stemming from the findings in the book.

Notes 1. Potential consequences include resistance by actors to the use of power sharing as a tool of peacebuilding. See, for example, Mehler (2016). 2. In the case of electoral systems, attention has centered primarily on shifts to more proportional electoral representation. See, for example, Bogaards (2013) and Horowitz (2008). 3. Efforts are currently being made to widen the approach to the study of postconflict institutions. The Institutions for Sustainable Peace (ISP), a network in which both editors of this volume are active, was created to pursue an integrated approach to institutional challenges in divided and postwar societies. For more information, see isp.giga-hamburg.de. 4. Recent exceptions include Themnér (2011) and Utas (2012). 5. Our focus on these questions echoes the approach popularized by Lasswell (1950).

Power Sharing and Power Relations in Postconflict States

15

6. A security dilemma exists in situations in which, lacking any effective central authority to enforce rules or contracts, groups seek to acquire more power and capabilities to gain an extra margin of safety, a process that can lead to a “vicious circle of security and power accumulation” (Herz 1950, 157). 7. Weber (1922) as translated in Weber (1978, 53). All translations of Weber’s text have trouble capturing the full meaning of the German original. Weber uses the term chance and not Wahrscheinlichkeit, which is literally the retranslation of probability. Chance contains a strong element of opportunity, which would have been a better translation. Dahl’s definition of power is similar to Weber’s: “A has power over B to the extent that he can get B to do something that B would not otherwise do” (Dahl 1957, 202–203). 8. It is not easy to pin down Foucault’s understanding of power. In his The History of Sexuality, there are numerous explanations of what power is not, but not a clear definition of what it is. The most important notions of power in Foucault’s work, however, can be distilled from his writings, and may be summarized as follows: power is relational, omnipresent, and productive; builds on consent; and is also created from below. See Foucault (1978, 92–96). One may add that, according to Foucault, (disciplinary) power is invisible (1991, 194). 9. While the consociational subtype of power sharing, emphasizing elements of group autonomy, has arguably received more attention academically and in practice, there is a distinct school of thought that proposes centripetalism as an opposite form, emphasizing systematic incentives for cooperation across identity groups. 10. This general neglect of the periphery by many power-sharing agreements is notable since case studies suggest that power-sharing agreements have ramifications at the local level of a polity. See, for example, Heitz (2009). 11. Decentralization, federalism, and regional autonomy, it should be noted, have quite distinct mechanisms in terms of attributing power positions. For a discussion on the relative rarity of territorial autonomy in Africa, see Hartmann (2013). 12. Given the fact that political power sharing has been the central focus of the literature on power sharing, in this book we sought to include chapters on the three forms of power sharing—military, territorial, and economic—about which, comparatively speaking, less has been written.

PART 1 The Impact of Power Sharing on Power Relations

2 Government-Rebel Relations in the Wake of Power-Sharing Peace Agreements Martin Ottmann and Johannes Vüllers

Since 1989, civil conflicts have increasingly been resolved by peace agreements between governments and rebel groups (Kreutz 2010). Even when these peace agreements succeed in ending armed combat, the interaction between the government of a state and the former rebel group remains the defining relationship in postconflict countries. Past research on powersharing institutions has focused primarily on the risk of civil conflict recurrence in postconflict situations (e.g., Hartzell and Hoddie 2007; Mason et al. 2001; Walter 2004). While this research undoubtedly has contributed greatly to our understanding of postconflict situations, it has addressed only part of what is going on in the relationship between the two conflict parties. Avoiding renewed violence is important, but under what circumstances do government and rebels in postconflict countries actually engage in nonviolent and inclusive relations with each other? In this chapter we provide a new perspective on government-rebel relations in the aftermath of a civil conflict. Following the conceptualization of power relations in Chapter 1 of this volume, we understand power relations as the relative strength between the government and a rebel group after the signature of a peace agreement. In the aftermath of a civil conflict, the power relations between these actors have to be transformed to enhance the quality of peace. This especially requires changes in the fundamental behavior of the two actors, from violence to nonviolence, and changes from an exclusive to an inclusive political system. These important changes in the power relations between the two sets of actors, however, are hard to achieve in the direct aftermath of a civil conflict. Power relations between the government and the rebels follow a zerosum logic during the civil conflict, according to which changes in their relations will be to the advantage of only one actor. Power-sharing institutions are able to change this into a positive-sum logic. 19

20

Martin Ottmann and Johannes Vüllers

In line with past research (e.g., Hartzell and Hoddie 2007; Jarstad and Sundberg 2007), we consider power-sharing institutions as the central element of negotiated peace agreements and, therefore, as the main determinant of government-rebel relations in postconflict situations. Power-sharing encompasses any arrangement between the government of a state and a rebel group that promises to establish joint control of power on the national level of government (Ottmann and Vüllers 2015). Nowadays, most peace agreements incorporate some sort of power-sharing institutions between the government and rebels. Power-sharing institutions include, for example, the inclusion of both conflict parties in the government at the cabinet level, parliamentary quotas for the rebels, and the integration of former rebel fighters into the national army. Against this background, our research question is: What impact do power-sharing institutions have on the transformation of power relations between the government and the rebels in a postconflict period? The classical and the postconflict power-sharing schools do not offer adequate answers to this particular question. Classical power-sharing theories have focused on improving the quality of democracy, although scholars working in this vein have not referred directly to a postconflict context (Lijphart 1977; McGarry and O’Leary 2004; Nordlinger 1972). The research on postconflict power sharing, in contrast, has focused directly on the impact of power-sharing institutions on civil conflict recurrence. This research has a slightly different concept of power-sharing institutions, differentiating between political, economic, military, and territorial variants of rules for the sharing of power (Hartzell and Hoddie 2007). Power-sharing institutions function as credible commitment mechanisms between the former conflict parties in the aftermath of a civil conflict and, thus, decrease the likelihood of conflict recurrence by integrating the conflict parties into the state system (Walter 2002). Neither of the schools mentioned above has directly addressed changes in power relations between the government and the rebels. One shortcoming associated with both of the schools is that they have employed rather abstract means of conceptualizing conflict actors. The relevant actors are ethnic groups—often ambiguously defined—or the social groups in whose name the rebel group fights (Hoddie and Hartzell 2005; Mattes and Savun 2009). We address this weakness by focusing instead on the power relations between the government and the rebels (e.g., K. G. Cunningham et al. 2012; Findley and Rudloff 2012). Another deficiency associated with the classical and postconflict power-sharing schools is that each has had a distinct rationale and a rather dichotomous understanding of the impact of power-sharing institutions—they contribute to democracy (or not), or they contribute to peace (or not) (e.g., Hartzell and Hoddie 2007; Lijphart 1977). The research, thus, has not systematically taken into account that power-sharing institutions address the more general questions of the willingness of both conflict parties to transform their

Government-Rebel Relations

21

power relations in regard to the form of action (from violence to nonviolence) and regarding the political system (from exclusive to inclusive). We seek to overcome this dichotomy through the use of the concept of power relations between the government and the rebels. First, we present our theoretical framework linking power-sharing institutions to the change of power relations between the conflict parties by changing the logic of the game. Then, we statistically test our hypotheses using a global sample of civil conflicts resolved by peace agreements between 1989 and 2006 and a new dataset on power-sharing events (Ottmann and Vüllers 2015). We found no evidence that cabinet power sharing affects government-rebel power relations, but did find that parliamentary power sharing makes nonviolent and inclusive power relations more likely. We also found that military integration increases the chances for inclusive power relations. Finally, in the concluding section, we discuss the implications of our findings regarding the functioning of the power-sharing institutions, their partially unintended consequences, and the effect of the transformation of power relations on the quality of peace.

Changes in Government-Rebel Power Relations in the Postconflict Period Power relations between the government and the rebels, as the most influential actors in the immediate postconflict period, continue to follow a zero-sum logic in the direct aftermath of a civil conflict. Power-sharing institutions, especially cabinet power sharing, parliamentary power sharing, and the military integration of former rebel fighters, help to overcome this zero-sum logic of the old power relations (e.g., Hartzell and Hoddie 2007; Jarstad and Sundberg 2007). The power-sharing institutions of a peace agreement provide incentives for the conflict parties to transform their power relations and, therefore, have a positive impact on the quality of the peace. Governmentrebel power relations thus can change from violence to nonviolence and from politically exclusive to inclusive. Changes in these two dimensions of power relations, however, come with costs and benefits for both sides. In effect, each conflict party will, first and foremost, see the costs associated with changes in the power structure due to the remaining security dilemma (Snyder and Jervis 1999). The zerosum logic of the civil conflict, according to which every loss of power is always a disadvantage, is still present in its direct aftermath. For example, the government or the rebel group can lose a fight on the battlefield during a civil conflict. Depending on the importance and the strategic consequences of the battle, the winner will increase its power while the other side will lose power. This zero-sum logic can be transformed into a positive-sum logic by

22

Martin Ottmann and Johannes Vüllers

implementing power-sharing institutions in the postconflict period. Powersharing institutions provide incentives (e.g., the integration of rebels into the cabinet or the acceptance of the state’s monopoly on the use of force by the rebels) that can help induce both sides to accept a loss of power such as the disarmament of the rebel fighters or the acceptance of the rebels as a legitimate political group. Thus, the logic of power relations becomes a positivesum game with the help of power-sharing institutions. In this chapter, we concentrate on government-rebel power relations in the military and the political dimensions. Cabinet and parliamentary power-sharing institutions and military integration have the greatest impact on the transformation of power relations. The political dimension is important because both conflict parties make claims about political goals during a civil conflict. A change in the political field is thus a sign of the acceptance of common rules by both sides. The military dimension is of equal interest concerning the quality of peace in a postconflict period because changes in this field signal that both conflict parties accept a nonviolent means of managing future conflicts. The subsections that follow outline the impact of the three power-sharing institutions in the political and security dimensions that most likely will have an effect on the changes in power relations between the former conflict parties. To illustrate our argument, we make use of our Power-Sharing Event Dataset (PSED), which records the practices of power sharing during the first five years of each postconflict period that began between 1989 and 2006 (Ottmann and Vüllers 2015). A summary of the descriptive statistics for the various forms of power sharing in the PSED is provided in Table 2.1.

Table 2.1 Descriptive Statistics for Practices of Power Sharing Power-Sharing Event Number of total postconflict months in the PSED Number of postconflict months in which rebels occupy cabinet position rebels occupy three or more cabinet positions rebels occupy senior cabinet position rebels occupy three or more senior cabinet positions rebels occupy nonsenior cabinet position rebels occupy three or more nonsenior cabinet positions rebels take over guaranteed seats in the national parliament rebel-military integration takes place

Frequency

Percentage

3,123

100

979 501 619 190 916 467 216 1,305

31 17 20 6 29 15 7 42

Note: The following cabinet positions and ministries were coded as being senior: president, vice president, prime minister, deputy prime minister, foreign affairs, defense, interior, justice, finance, economy, and resources. PSED, Power-Sharing Event Dataset.

Government-Rebel Relations

23

From Violent to Nonviolent Power Relations

A state must have the monopoly on the use of force accepted by all its citizens to stabilize and to guarantee the internal order while the rebels have a “private army” that still opposes the state monopoly after a civil conflict. The rebels therefore have to give up their private armies by demobilizing fighters and accepting the state monopoly on the use of force to transform power relations from violent to nonviolent (Walter 2002). According to the zero-sum logic of power relations, this change will disadvantage rebels in the short term due to the demobilization of their fighters and the consequent loss of (military) power capabilities. The government, in contrast, increases its power in the military field because a pivotal counterpart, the rebels, accepts the state monopoly on the use of force. Some power-sharing institutions are able to change the logic of this game to one that is positive-sum in nature, thus increasing the likelihood that power relations will indeed change from violent to nonviolent while others will not (Hartzell and Hoddie 2007; Walter 2002). The first powersharing institution that may help accomplish this is the integration of rebel representatives into the national cabinet. From a government perspective, the integration of rebel representatives into the cabinet is not a costly signal. While the government publicly accepts the rebels as a political actor with their integration into the cabinet, the rebels often take over cabinet portfolios with rather limited power that cannot directly impact the policies of the state. The data reported in Table 2.1, for example, show that rebels occupied only three or more senior cabinet positions in just 6 percent of all postconflict months recorded in the PSED. The costs of integrating rebels into the national cabinet are reasonable for the government because in return it receives the acceptance of its monopoly on the use of force from the rebels. From a rebel perspective, the integration into the cabinet is beneficial because all cabinet positions are a public signal that the government accepts the rebels as a political counterpart. Unsurprisingly, our data show a respectable 29 percent of postconflict months in which rebels occupied socalled nonsenior cabinet positions (see Table 2.1). These positions often encompassed the health, tourism, environment, culture, or education portfolios. Rebels receive political power by sitting at the political center when governing decisions are made. In that way, rebels are also involved in the discussion of decisions regarding the national army and the military sector. The inclusion of rebel representatives in the national cabinet should dampen the costs associated with the demobilization of rebel fighters. In effect, the integration of rebels into the national cabinet helps to transform government-rebel power relations from ones of violence to nonviolence.

24

Martin Ottmann and Johannes Vüllers

Hypothesis 1a: The integration of rebel representatives in the national cabinet makes it more likely that power relations become nonviolent. Another important power-sharing institution is guaranteed seats for a rebel group in the national parliament. This can be seen as a credible signal by the government that it will change power relations fundamentally by integrating the rebels into the political system. The costs for the government, however, are rather limited due to the often low number of rebels in the parliament and the limited political power of the national parliament in the aftermath of a civil conflict. From a rebel perspective, the integration of its members into the national parliament will not help to counterbalance the costs associated with the demobilization of its fighters. Members of the parliament have no direct influence on the security field that will be governed by the state leadership in the national cabinet. In effect, the benefits for the rebels should be minimal because the parliamentary seats do not provide the same amount of influence that cabinet positions do. Table 2.1 shows that parliamentary power sharing occurred in just 7 percent of all postconflict months in the PSED. In sum, the integration of the rebels into the parliament has a rather limited ability to change the logic of the game into one that is positive-sum in nature and to fundamentally change power relations from ones of violence to nonviolence. Hypothesis 1b: The integration of rebel representatives in the national parliament has no effect on the likelihood that power relations become nonviolent. Finally, the military integration of former rebel fighters into the national army makes the transformation of power relations more difficult. From the perspective of the government, the benefit of the integration is getting control over the rebel fighters. The government can closely monitor the behavior of the former rebels in the army. For the rebels, the integration of their members has benefits, but it hampers a transformation of power relations to nonviolence. Rebel leaders often hold control over their fighters while the fighters are integrated into the national army. In many cases, the rebels are integrated in groups into the national army so that the social network and hierarchical structure of the rebel group continues undiminished (Baaz and Verweijen 2013). As a result, the national army finances and educates the integrated rebel fighters, but the rebel network remains unchanged. Table 2.1 shows that a rebel-military integration process took place in 42 percent of all postconflict months of the PSED. In sum, the integration of rebel fighters into the national army can be beneficial for both sides, but the integra-

Government-Rebel Relations

25

tion provides the rebel leadership with the possibility to easily mobilize armed combatants for their own purposes. Hypothesis 1c: The integration of rebel fighters into the national army makes it less likely that power relations become nonviolent. From Exclusive to Inclusive Power Relations

The second important dimension of power relations for the conflict parties is the political field. Civil conflicts occur because the political system is exclusive or the rebels perceive it in such a way. In the postconflict period, however, both conflict parties have to agree on the fundamental rules of the political game. In line with our focus on power relations between the government and the rebels, we define the political system as inclusive if the rebel group is fully integrated into the system. Full integration is achieved if the former rebel group as a group is able to organize itself to participate in politics and is accepted as a political actor by the government. Following almost all civil conflicts that end with a negotiated settlement, the state government holds the political power in the capital while the rebels are excluded from institutionalized political power. The rebels thus have no chance to politically influence national policies as a means of reaching their goals through activities in the political system. To improve the quality of peace and to stabilize the political system, government-rebel power relations in the political field have to become inclusive. In general, this will come with costs for the government and benefits for the rebels. In a long-term perspective, however, it can also be to the benefit of the government which, by being seen as the peacemaker, can help to stabilize its own political power position. The first power-sharing institution that helps to change power relations in the political dimension is the integration of rebel representatives into the national cabinet. As argued above, the government can publicly show its willingness to integrate the rebels into national politics by providing them with seats in the cabinet. The government, however, gives only limited political power to the rebels due to the rather low number of ministerial portfolios offered to them. Looking at the first five years of each postconflict period included in the PSED, for example, there has been cabinet power sharing in only 30 percent of the observed postconflict months (Ottmann and Vüllers 2015). Moreover, rebels received more than three cabinet positions in just 17 percent of all postconflict months. From a rebel perspective, the integration of members into the cabinet provides them with political experience and with access to the administrative apparatus that can be used in upcoming election campaigns. Ministers also have better access

26

Martin Ottmann and Johannes Vüllers

to the press than normal members of the parliament. The ministerial positions, moreover, provide status symbols, such as a chauffeur and security officials, that increase the public’s awareness of the rebels as legitimate and pivotal political actors. The rebels should thus see cabinet positions as providing them with advantages for their political campaigns. It is the case, however, that the ministerial portfolios assigned to rebel representatives are limited in duration. The rebel group will not be guaranteed representation at the political center after the expiry of the powersharing institution. That is, the actual political system has not explicitly changed because of the time-limited implementation of a power-sharing institution. We argue that rebels profit from the integration of rebel representatives into the political center for a certain period of time, but this power-sharing institution does not ensure that inclusive power relations have been created. Nevertheless, the associated benefits that this form of power sharing provides the rebels can help to facilitate their acceptance of the political system. Hypothesis 2a: Rebel representatives in the national cabinet make it more likely that power relations become inclusive. The second power-sharing institution that impacts the transformation of a rebel group into a political party, and thus political power relations, is the integration of the rebels into the national parliament. For the government, the costs of integrating the rebels into the parliament are low. In most cases, the parliament has little decisionmaking power in the direct aftermath of a conflict (Samuels 2006). Moreover, governments generally do not view rebel parties as a strong political competitor due to their rather restricted core constituency and because they assume that any atrocities they may have committed during the civil conflict will make an electoral victory unlikely. From a rebel perspective, seats in the parliament are essential to being viewed as a political party. The rebels can act as members of the parliament in public and they gain access to political resources such as financial means or invitations to public events. The integration of the rebels into the parliament, moreover, provides them with the opportunity to present themselves as a group rather than only as single members of the parliament. It is important to distinguish between the power-sharing institution “rebel parliamentary seats,” in which the rebels take over seats in the parliament without election, and parliamentary seats that rebel representatives gain later as a result of elections. Only the first is a power-sharing institution because the latter is a normal result of a democratic election. The time dimension is once again important: while the power-sharing institution rebel parliamentary seats allows the participation of rebel representatives in the national assembly for a certain period of time, the possibility of losing their seats is a real risk

Government-Rebel Relations

27

for the rebels. Depending on various factors, such as the assumed and real public support for the rebels as a political actor, rebels will decide during the time that this particular power-sharing institution is in place whether they are willing to change the power relations into inclusive ones. Hypothesis 2b: The integration of rebel representatives into the national parliament makes it more likely that power relations will become inclusive. Finally, military integration should help both conflict parties to accept a change in power relations in the political dimension with power relations becoming more inclusive in nature. The government views the integration of former rebel soldiers into the national army as a credible commitment on the part of the rebels to transform from an armed actor into a political actor. Additionally, the government gains control over the rebel soldiers through the army command structure so that the threat of a future outbreak of a civil conflict by these soldiers will be reduced. The government, thus, is willing to allow the rebels to become a recognized political party if the rebels give a credible signal that they generally accept the rules of the political game. The integration of former rebel fighters into the military constitutes such a signal. The importance of rebel-military integration is reflected in the already reported high frequency of postconflict months in which such a process was active. Almost half of all postconflict months included in the PSED experienced some form of ongoing or already concluded rebel-military integration process. For the rebels, military integration provides a means of contending with the numerous rebel fighters who cannot play a prominent role in the rebel party. The parties of former rebel groups often lack financial and organizational resources, meaning that only a small number of former fighters can be paid through this channel. Military integration is thus a welcome means for the rebels to provide their fighters with some material benefits after the fight (Glassmyer and Sambanis 2008). The rebels, as an organization, thus have the potential to prevent future institutional disruption by dissatisfied members due to a lack of finances. Moreover, the rebel group can show that it cares about the fighters and is able to provide them with prospects for the future. The transformation of the rebel group into a normal political actor thus should be easier if the military integrates some of the rebel soldiers. As a result, the integration of rebels into the national army can be expected to change the logic of the game into a positive-sum one and increase the likelihood that power relations become inclusive. Hypothesis 2c: The integration of rebel fighters into the national army makes it more likely that power relations become inclusive.

28

Martin Ottmann and Johannes Vüllers

Data To test our hypotheses, we statistically examined the effects that powersharing institutions have on power relations between the government and rebels during the postconflict period. Our unit of analysis was the governmentrebel dyad during the postconflict period following a peace agreement. With this focus on the government-rebel dyad, we explicitly investigated the inclusion of a particular rebel group in the power-sharing institutions of a postconflict country. If there was more than one rebel signatory to a peace agreement, we treated each government-rebel dyad of this peace agreement separately. Following a convention established by past research (e.g., Hartzell and Hoddie 2007; Walter 1999), we observed each government-rebel dyad for a five-year postconflict period after the signature of the peace agreement. If the power relations between the government and rebels changed within these five years, the observation period ended on the day this change occurred. If no change in the power relations occurred in the first five years of a postconflict period, the observation period of this government-rebel dyad was censored. Finally, a postconflict period was also terminated if the civil conflict recurred, an event we operationalized as government-rebel violence causing more than twenty-five battle-related deaths. We used the Power-Sharing Event Dataset (Ottmann and Vüllers 2015) to arrive at our universe of cases. The PSED includes all peace agreements signed between the government and rebels in civil conflicts from 1989 to 2006 as identified by the Conflict Encyclopedia of the Uppsala Conflict Data Program (UCDP) (Uppsala Conflict Data Program 2013). There were 110 government-rebel dyads active in 46 civil conflicts in 41 countries in our sample. Dependent Variable

The dependent variable in our study was the status of the power relations between the government and rebels on any given day of the dyad’s postconflict period. The power relations between the government of a state and a rebel group were “frozen” directly after both conflict actors had signed a peace agreement. The government-rebel dyad remained within this state of frozen power relations until it experienced a transition to another state of power relations. Two alternative states of power relations in a postconflict government-rebel dyad were possible: the power relations became nonviolent or they became inclusive. We operationalized the transition of government-rebel power relations to a nonviolent state by identifying the date when the rebel group ceased existing as an armed actor. A rebel group was understood to have undergone a change in this basic group characteristic when it was publicly announced

Government-Rebel Relations

29

that it had concluded a disarmament, demobilization, and reintegration (DDR) process. In effect, we considered such an announcement to mean that the rebel group acknowledged the legitimacy of the government’s monopoly of violence. Similarly, we operationalized the transition to inclusive power relations by identifying the date on which a rebel group transformed itself into a political party.1 This rebel-to-party transformation marked the acknowledgment by the government that the rebel group could continue to exist as a political organization within the political system of the state. A rebel group became a political party when it was officially and publicly recognized as such by the government of the state. To code these outcomes of our dependent variable, we relied on raw data collected for the PSED (Ottmann and Vüllers 2015). A transition to either one of these states of power relations could occur only once; the occurrence of such a transition terminated the observation period for this outcome. Even if a transition to one of these power relations conditions occurred, the dyad was still at risk of experiencing a transition to the other status. These two different states—that is, nonviolence and inclusiveness—were assumed to be unordered and clearly distinct from each other. Thus, government-rebel power relations did not have to become nonviolent to be inclusive and vice versa. The Sudan People’s Liberation Army (SPLA) in Sudan, for example, did not begin its disarmament process until more than two years after the January 2005 peace agreement, but was already acknowledged as a political party in February 2005. In contrast, the Macedonian Ushitra Çlirimtare e Kosovës (UCK) concluded its disarmament in September 2001 and only then became a political party in June 2002. It is also possible that the government-rebel power relations only experience a transition to either nonviolence or inclusion, but not both. The fate of the Uganda National Rescue Front II (UNRF II) serves as an example of a rebel group that disarmed, but was never included in the country’s political system as a political party. The rebel groups in Democratic Republic of Congo, on the other hand, became political parties, but retained their arms. Finally, it also often happens that neither of these two events occurs and government-rebel power relations remain unchanged from their status following the signing of a peace agreement. Examples of this include the socalled frozen conflicts in Georgia, Moldova, and the Comoros. Overall, 56 of the total of 111 government-rebel dyads in our sample experienced a change in their power relations. There were 20 cases of government-rebel power relations becoming nonviolent in our sample of 111 government-rebel dyads. Interestingly, rebel groups even became nonviolent when the government and rebels did not agree on either a rebelmilitary integration or DDR process. The mean duration of a postconflict period until this transition occurred was a little over two years (783 days). In contrast, we had almost twice as many transitions (thirty-nine) to inclusive

30

Martin Ottmann and Johannes Vüllers

government-rebel power relations. In this instance, the mean duration of a postconflict period until such a transition took place was substantially shorter. On average, it took less than one and a half years (546 days) for former rebel groups to become a political party. That is, government-rebel power relations were more likely to turn inclusive than nonviolent. Explanatory Variables

To assess the impact of power-sharing institutions on the power relations between the government and rebels, we relied on the power-sharing measures provided by the PSED (Ottmann and Vüllers 2015). The PSED has an actorcentric approach toward power sharing and, therefore, focuses exclusively on those power-sharing institutions that apply to the conflict actors. The PSED disaggregates power-sharing institutions along dimensions of power. For this study, we exclusively employed those PSED indicators that measure the practices of power sharing in the national executive, the national parliament, and the national army. Power-sharing practices were coded as existing if powersharing institutions were implemented, regardless of whether or not the conflict parties agreed to them as part of the peace agreement.2 Power sharing in the national executive (or cabinet power sharing) occurs when the rebel signatory to a peace agreement takes over ministries in the national government. Parliamentary power sharing occurs when the rebel signatory is given guaranteed seats in the national parliament. Power sharing in the national army is operationalized as the integration of former rebel fighters into the national army. We used a binary variable for each power-sharing institution. If a government-rebel dyad engaged in one of these power-sharing practices, the respective variable assumed the value of 1 on the particular day this practice started. It remained 1 as long as this power-sharing practice was active and changed back to 0 when the power sharing was terminated.3 Looking at the distribution of power-sharing arrangements across the government-rebel dyads in our sample, one can see that power sharing was a relatively infrequent feature of the postconflict period. For example, only 29 percent of all dyads in our sample experienced cabinet power sharing before either nonviolent or inclusive power relations emerged. Parliamentary power sharing was even more infrequent, with only 6 percent of all recorded dyads experiencing this form of power sharing. Finally, military integration of rebel fighters in the national army took place in 17 percent of all recorded postconflict dyads. Control Variables

Our analysis also included control variables to account for alternative explanatory factors affecting the power relations within a government-rebel

Government-Rebel Relations

31

dyad: the characteristics of the conflict and the postconflict situation, additional power-sharing arrangements, promises of power sharing, country development, and rebel group characteristics. As we explain in the next section, these control variables were included in the regression models in a stepwise manner. First, we controlled for the characteristics of the armed conflict between the government and rebels (Hartzell and Hoddie 2007; Walter 2002). We assumed that conflicts between the government and rebels that are fought over the political status of parts of the country are more likely to increase the likelihood that the postconflict power relations become nonviolent and inclusive. Such territorial conflicts leave more political space for accommodation between the conflict actors than conflicts over national power. We included a binary variable taking the value of 1 when the conflict had been fought over a particular territory within a country and 0 otherwise. We also assumed that longer-running conflicts are less likely to result in nonviolent and inclusive power relations. The divisions between government and rebels are more entrenched after long-lasting conflicts. We used the number of months between the first time armed combat between the government and rebels had resulted in more than twenty-five battle-related deaths in a calendar year and the day the peace agreement had been signed. Moreover, civil conflicts with high levels of battle violence between the government and rebels could reduce the probability of nonviolent and inclusive power relations. Similar to the hypothesized effect of conflict duration, intense conflicts are likely to deepen the divide between the conflict actors. We used a binary variable taking the value of 1 when the government-rebel dyad had caused more than 1,000 battle-related deaths in at least one calendar year of the overall conflict. The data for all three variables were taken from the UCDP Dyadic Dataset (Harbom, Melander, and Wallensteen 2008). Second, we included two postconflict control variables (Collier, Hoeffler, and Söderbom 2008). We assumed that although the presence of a UN peacekeeping operation might be able to prevent civil conflict recurrence, it might also freeze the existing divisions between the government and rebels and thus prevent the development of cooperative power relations. We used a dummy variable taking the value of 1 when a UN peacekeeping mission was present in the country under analysis (Hegre, Hultman, and Nygård 2011).4 We then controlled for the impact of multiple rebel signatories to the peace agreement on civil conflict recurrence. The assumption was that the political and military competition between multiple rebel groups decreases the probability that nonviolent and inclusive power relations will occur. We included a binary variable taking the value of 1 when the peace agreement had been signed by two or more rebel groups and 0 otherwise. The data for this variable were taken from the PSED (Ottmann and Vüllers 2015).

32

Martin Ottmann and Johannes Vüllers

Third, we controlled for power-sharing arrangements aimed at the ethnic group level or the country as a whole. While such power sharing does not directly address government-rebel power relations, we nevertheless assumed that it had an indirect effect, increasing the likelihood for nonviolent and inclusive power relations to emerge by satisfying the rebels’ constituency directly. We included a binary variable taking the value of 1 when a territorial power-sharing arrangement granting autonomy or enacting regional devolution had been passed by the national parliament. In such a case, we expected that rebel activity would be focused on the respective region of the country. The likelihood of inclusive power relations at the national level should thus decrease. We also included a binary variable taking the value of 1 when an electoral law had been passed promising proportional representation. Proportional representation should increase the likelihood for nonviolent and inclusive power relations by addressing the initial grievances resulting in the rebellion and also by reducing the opportunity for rebellion (Cammett and Malesky 2012). Both variables kept the value of 1 as long as the laws were in effect. Fourth, we tested whether our results held after we controlled for the mere promises of cabinet, parliamentary, and military power sharing as agreed on in the peace agreement. It is possible that the effects of the practices of power sharing would disappear once we considered what had originally been promised to the rebel groups in the peace agreement. If that was to occur, it would imply that the promises of power sharing made at the outset of a postconflict period would have a greater impact on postconflict power relations than any practices of power sharing taking place during the postconflict period. For each of the three possible power-sharing institutions introduced above, we included a binary variable taking the value of 1 when a measure had been agreed to as part of a negotiated settlement— irrespective of whether or not it was implemented in the postconflict period. The data for all of the power-sharing variables presented in this paragraph were taken from the PSED (Ottmann and Vüllers 2015). The fifth control variable was the average life expectancy in the country under analysis. Life expectancy functions as a proxy for the economic development of a country. We argue that nonviolent and inclusive power relations are more likely in a more developed country as the rebels see more economic benefits from integrating themselves into the country’s political system. The data on life expectancy were taken from the World Bank Development Indicators (World Bank 2012). Finally, we assumed that the strength of the rebel group relative to government forces affects the nature of the power relations between the government and rebels (Gent 2011). We expected a rebel group whose military strength was equal to or even surpassed the military strength of the government to be less likely to engage in either nonviolent or inclusive power

Government-Rebel Relations

33

relations. In such instances, rebel groups’ relative strength induces them to hold out and hope for a better deal in the future. While one would expect that stronger rebel groups would not even agree to a negotiated settlement, it nevertheless occurs. Six of the 111 government-rebel dyads under analysis featured a relatively stronger rebel group. We used two binary variables to measure relative rebel strength. The first took the value of 1 when government and rebel forces were equally strong during the civil conflict and 0 otherwise. The second took the value of 1 when rebel forces were stronger than government forces during the civil conflict and 0 otherwise. The data for these variables were taken from the Non-State Actor Dataset (D. E. Cunningham, Gleditsch, and Salehyan 2013).

Empirical Analysis We assumed that a government-rebel dyad can experience two distinct changes in power relations at any time during the postconflict period. Government-rebel power relations can become nonviolent first, or they can become inclusive first. We further hypothesized that power-sharing arrangements within a government-rebel dyad do not necessarily have the same effect on the first occurrence of each of these two outcomes. Therefore, the question we seek to answer is: What impact do power-sharing institutions have on the transformation toward either inclusive or nonviolent power relations between government and rebels in a postconflict period? A competing-risks event history model is able to provide us with an answer to this question (Cleves et al. 2010). Commonly, the Cox proportional hazard model is used to estimate separate models to investigate when the transition to a particular event occurs. The alternative competing risks are treated as randomly censored. However, such censoring disregards that it is the prior occurrence of these competing risks that prevents the risk under analysis from occurring first. That is, the failure times for each of the risks under analysis are likely to be correlated with each other. As a consequence, the results from a Cox proportional hazard model can be difficult to interpret as the failure time of each risk under analysis depends on the failure times of the competing risks. Jason P. Fine and Robert J. Gray’s (1999) competing-risks regression model offers an alternative option. This model estimates the subdistribution hazards. It explores the “instantaneous probability of failure from cause i at time t given either no failure before t or failure from another cause before t” (Cleves et al. 2010, 382). It accomplishes this by keeping units in the overall risk pool that have experienced a failure while also assuming that these failed units cannot experience another failure due to a competing risk anymore. For example, once a government-rebel dyad transitioned to nonviolent

34

Martin Ottmann and Johannes Vüllers

power relations, our analysis excluded the possibility that this dyad could experience a transition to any other state of power relations. Therefore, the sample used for the statistical analysis included only twelve out of twenty total transitions to nonviolent power relations and thirty-four out of thirtynine total transitions to inclusive power relations.5 We also employed robust standard errors clustered on government-rebel dyads to account for heteroscedasticity and serial correlation.6 We conclude this section with two robustness checks discussing the proportional subhazards assumption of the competing-risks regression. We also address possible endogeneity concerns. Regression Results

Using Fine and Gray’s (1999) competing-risks regression model, we estimated a series of models with different sets of control variables for each of the possible power relations outcomes. The objective of this procedure was to ascertain whether the estimated effects remained stable in terms of statistical significance and substantive effect when controlling for different context factors. The first model—Model 1—is the baseline model; it includes only the measures of cabinet and parliamentary power sharing as well as military integration. In Model 2, we added the control variables measuring the characteristics of the civil conflict and the postconflict situation to this baseline model. Past studies considered these variables as the central control variables; we therefore included them in all of the following models. In Model 3, we also controlled for additional power-sharing institutions; that is, territorial power sharing and the introduction of a proportional representation voting system. Model 4 evaluated whether the effect of our main explanatory variables on the dependent variable changed once we controlled for the promises of political and military power sharing. In Model 5, we included life expectancy as a control for the economic development of the country under analysis. In Model 6, we controlled for the military strength of the rebel group relative to the government. Table 2.2 reports the coefficients and their statistical significance for the competing-risks regressions with the emergence of nonviolent power relations as the event of interest. Contrary to what we had hypothesized, the presence of cabinet power sharing in a government-rebel dyad appeared to decrease the likelihood that nonviolent power relations would emerge. However, it should be noted that the coefficient was not statistically significant in all estimated models. In Model 1, the coefficient measuring the impact of parliamentary power sharing was positive, but not statistically significant. With the inclusion of alternative explanatory factors in Models 2a to 6a, however, the coefficient became very large and was statistically significant. Once additional control variables accounted for some of the variation in the dependent variable, it appeared that there was an almost

35 Table 2.2 Competing-Risks Coefficients for Transition to Nonviolent Power Relations (inclusive power relations)

Cabinet power sharing Parliamentary power sharing Military integration

Model 1a

Model 2a

Model 3a

Model 4a

Model 5a

Model 6a

–0.810 (0.596)

–0.572 (0.600)

–0.672 (0.591)

–0.323 (0.702)

–0.588 (0.627)

–1.152 (0.701)

1.072 (0.556) –0.872 (1.069)

16.405*** (0.669) –0.801 (1.113) –0.637 (0.618) –1.692 (1.127) 0.000 (0.000) 0.307 (0.583) –15.866*** (0.486)

18.071*** (0.751) –0.778 (1.072) –0.911 (0.702) –1.538 (1.185) 0.000 (0.000) 0.439 (0.612) –17.405*** (0.495) 0.979 (1.101) –16.986*** (0.747)

34.191*** (1.130) –0.342 (1.162) –0.882 (0.678) –2.332 (1.443) 0.000 (0.000) 0.305 (0.573) –17.268*** (0.709)

Conflict incompatibility Conflict intensity Conflict duration UN peacekeeping Multiple rebel signatories Territorial power sharing Proportional representation

–18.212*** (0.885) 2.584*** (0.738) –0.619 (0.735)

Promise of cabinet power sharing Promise of parliamentary power sharing Promise of military integration Life expectancy (ln) Rebels as strong as government Rebels stronger than government Government-rebel dyads Transitions to nonviolent power relations Transitions to competing event Log pseudo-likelihood Akaike information criteria (AIC) Bayesian information criteria (BIC)

16.699*** 17.873*** (0.622) (0.603) –0.319 –0.807 (1.193) (1.131) –1.035 –0.673 (0.637) (0.605) –1.650 –1.912 (1.059) (1.014) 0.000 0.000 (0.000) (0.000) 0.248 0.221 (0.530) (0.555) –15.636*** –16.151*** (0.596) (0.487)

4.022 (2.217)

111 12

111 12

111 12

111 12

111 12

34

34

34

34

33

0.443 (0.715) –16.876*** (0.824) 111 12 34

–49.242 104.485

–44.151 104.303

–42.732 105.463

–39.309 100.618

–42.214 102.429

–41.892 103.784

115.827

134.549

143.271

142.206

136.428

141.592

Note: Robust standard errors clustered on government-rebel dyad are in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

36

Martin Ottmann and Johannes Vüllers

perfect relationship between the presence of parliamentary power sharing and nonviolent power relations. Thus, our theoretical consideration underestimated the impact of parliamentary power sharing: when rebels occupied guaranteed seats in a national parliament, government-rebel relations were likely to become nonviolent. Moving on to military power sharing, we found a negative coefficient indicating that military integration decreased the likelihood of nonviolent power relations, which supports our theoretical assumption. However, the coefficient was not statistically significant. Table 2.2 also reports the coefficients of the control variables. The most important findings pertain to the presence of multiple rebel signatories, the introduction of proportional representation, promises of power sharing, and relative rebel strength. We found that nonviolent power relations were less likely when the peace agreement involving the government-rebel dyad included additional rebel signatories. In Model 3a, we also found that the adoption of a law introducing proportional representation reduced the likelihood of nonviolent power relations. Interestingly, Model 4a showed that the promise of cabinet power sharing made at the outset of the postconflict period was a statistically significant predictor, reducing the probability of nonviolent power relations. Taken together with the negative but statistically not significant coefficient for the practice of cabinet power sharing, this casts even more doubt on Hypothesis 1a. In contrast, the promise of parliamentary power sharing increased this probability. Finally, Model 6a showed that nonviolent power relations were less likely to emerge when rebels were militarily stronger than government forces. Next, we examined the competing-risks regressions analyzing the emergence of inclusive power relations (see Table 2.3). To begin with, we once again found that the presence of cabinet power sharing had a negative impact, but was not statistically significant. We also found the presence of parliamentary power sharing to exert a statistically significant positive impact on the outcome variable. In line with Hypothesis 2b, government-rebel power relations were more likely to become inclusive when rebels occupied guaranteed seats in a national parliament. Regarding our final power-sharing measure, the findings showed that military integration increased the chances for inclusive power relations. The coefficient was statistically significant across all models and therefore supports Hypothesis 2c. Turning to the control variables, our estimates showed conflict duration, territorial power sharing, and proportional representation, as well as relative rebel strength, to be of relevance. Contrary to our expectation, long-lasting civil conflicts made inclusive power relations more likely. However, the substantive effect of this variable was negligible. Once territorial power sharing—that is, devolution or even autonomy—was established, the probability for government-rebel power relations to become inclusive was reduced. In contrast, the introduction of proportional representation

37 Table 2.3 Competing-Risks Coefficients for Transition to Inclusive Power Relations (nonviolent power relations)

Cabinet power sharing Parliamentary power sharing Military integration

Model 1b

Model 2b

Model 3b

Model 4b

Model 5b

Model 6b

–0.492 (0.361) 1.183** (0.418) 0.901** (0.341)

–0.340 (0.400) 1.475** (0.499) 0.952* (0.429) –0.445 (0.518) –0.540 (0.577) 0.000*** (0.000) –0.111 (0.399) 0.231 (0.511)

–0.078 (0.406) 1.209* (0.544) 0.964* (0.418) –0.051 (0.551) –0.440 (0.571) 0.000** (0.000) –0.208 (0.408) 0.227 (0.498) –17.467*** (0.662) 0.816* (0.343)

–0.413 (0.520) 1.380* (0.606) 0.890* (0.440) –0.277 (0.578) –0.620 (0.624) 0.000*** (0.000) –0.048 (0.423) 0.329 (0.596)

–0.325 (0.391) 1.553** (0.511) 1.215* (0.477) –0.844 (0.614) –0.842 (0.526) 0.000*** (0.000) –0.083 (0.401) 0.370 (0.534)

–0.107 (0.443) 1.217* (0.600) 1.045* (0.417) –0.400 (0.518) –0.313 (0.584) 0.000*** (0.000) –0.176 (0.408) 0.166 (0.520)

Conflict incompatibility Conflict intensity Conflict duration UN peacekeeping Multiple rebel signatories Territorial power sharing Proportional representation Promise of cabinet power sharing Promise of parliamentary power sharing Promise of military integration Life expectancy (ln) Rebels as strong as government Rebels stronger than government Government-rebel dyads 111 Transitions to inclusive 34 power relations Transitions to 12 competing event Log pseudo-likelihood –134.213 Akaike information 274.427 criteria (AIC) Bayesian information 285.769 criteria (BIC)

0.120 (0.458) –0.082 (0.748) 0.377 (0.484) 1.001 (1.421)

111 34

111 34

111 34

111 33

12

12

12

12

–0.992* (0.398) 0.392 (0.646) 111 34 12

–126.780 269.560

–123.937 267.873

–126.360 274.719

–119.851 257.702

–125.240 270.480

299.806

305.681

316.307

291.701

308.287

Note: Robust standard errors clustered on government-rebel dyad are in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

38

Martin Ottmann and Johannes Vüllers

increased this probability. Both group-level power-sharing variables showed the hypothesized effect. Finally, inclusive power relations were less likely if government and rebel forces were roughly equal in their military capabilities. Substantive Effects

To illustrate the substantive impacts of our statistically significant explanatory variables on our outcome variable, we modeled their effects on cumulative incidence functions. That is, we estimated the probability that nonviolent or inclusive power relations would occur on a particular postconflict day given the absence or presence of parliamentary or military power sharing. Figure 2.1 displays these cumulative incidences as stacked plots (Cleves et al. 2010). These plots show the probabilities for all competing risks—including the possibility that no failure event occurs—when neither parliamentary nor military power sharing occurs (Figure 2.1a), when parliamentary power sharing occurs (Figure 2.1b), and when military power sharing occurs (Figure 2.1c). The probabilities for each competing risk and the possibility of no failure event ranged between 0 and 1 and—when summed up—equaled 1. In each plot, the cumulative incidence for each possible failure event is given by the distances between the curves. Figure 2.1a displays the probabilities of the competing risks of power relations turning inclusive or nonviolent when neither parliamentary nor military power sharing occurs. One can see that government-rebel power relations have indeed an increasing probability to turn either inclusive or nonviolent over time. However, these probabilities are rather low. The most likely outcome is that no change at all occurs. In the following, we compare this baseline plot to instances when either parliamentary or military power sharing occurs. Figure 2.1b shows that the presence of parliamentary power sharing substantially increases the probability that government-rebel power relations turn inclusive. While the probability of power relations becoming inclusive is on average around 5 percent in the first year without parliamentary power sharing (see Figure 2.1a), the introduction of this powersharing institution more than quadruples this probability to around 22 percent. This effect then sharply increases over time. Without parliamentary power sharing, the baseline plot shows that the probability of inclusive power relations only reaches a maximum of 27 percent in the last year of a postconflict period. In contrast, the presence of parliamentary power sharing pushes this probability to a maximum of 77 percent in the last year of a postconflict period. The effect of parliamentary power sharing on nonviolent power relations is even more drastic. In the baseline plot, the area between the curves

Figure 2.1

Stacked Cumulative Incidence of Government-Rebel Relations for Different Power-Sharing Settings Figure 2.1a Baseline CIF

Figure 2.1b Parliamentary Power-Sharing CIF

Figure 2.1c Military Power-Sharing CIF

Note: CIF, cumulative incidence function.

39

40

Martin Ottmann and Johannes Vüllers

marking the inclusive power relations probability and the nonviolent power relations indicates an average probability of 18 percent that power relations become nonviolent in the first year and 33 percent in the fifth year of a postconflict period. If parliamentary power sharing is introduced, however, the probability of this competing risk is 1 for every possible point in time. Since there cannot be a cumulative incidence higher than 1, the dashed line in Figure 2.1c is therefore just an illustration of the substantive effect of parliamentary power sharing on nonviolent power relations. This also means that the introduction of parliamentary power sharing definitely results in a change of government-rebel power relations over the course of a postconflict period. As shown in Figure 2.1c, the introduction of military power sharing substantially increases the probability that government-rebel power relations turn inclusive when compared to the baseline probability (see Figure 2.1a). In the first year of a postconflict period, the average probability is 14 percent. This increases to an average probability of 56 percent for the fifth postconflict year. In contrast, the reduced area between the curves marking the inclusive power relations probability and nonviolent power relations probability illustrates to what extent military power sharing reduces the emergence of nonviolent government-rebel power relations. With military power sharing in place, the average probability for nonviolent power relations for the first 365 postconflict days is reduced from 18 percent to 5 percent. This gap widens over time as the average probability for the last postconflict year under observation decreases from 33 percent to 16 percent. Robustness Checks

In a final step, we checked the robustness of our empirical findings. First, we explored whether our models met one of the core assumptions of the Fine and Gray (1999) competing-risks regression model: the proportional subhazards assumption. This assumption implies that the relative subhazard of each covariate is fixed over time. It might be possible, however, that the effects of some covariates will change over time. This might especially be so because the Fine and Gray competing-risks regression model uses the last available covariate values of those units that have experienced failure due to a competing risk, but have been kept in the risk pool. This is potentially problematic as the power-sharing institutions under analysis were only temporary arrangements, but the information regarding a possible change in the power-sharing institutions might be lost when the unit was censored. To test the proportional subhazards assumption, we reran the competing-risks regressions, but interacted each of the independent variables with time. If one of the interactive terms was statistically significant, it indicated that the subhazard varied with time. For the sake of clarity, we

Government-Rebel Relations

41

reported only the findings of Models 2a and 2b, which include our key explanatory variables and control variables (see Table 2.4). As our results make clear, we did not find any sign that the effect of our explanatory or control variables varied over time. Second, we explored the possibility that our key explanatory variables capturing power-sharing practices were endogenous to context factors also

Table 2.4 Competing-Risks Regression with Independent Variables Interacted with Time (Models 2a and 2b) Model 2a Constant effects Cabinet power sharing Parliamentary power sharing Military integration Conflict incompatibility Conflict intensity Conflict duration UN peacekeeping Multiple rebel signatories Time-varying effects Cabinet power sharing Parliamentary power sharing Military integration Conflict incompatibility Conflict intensity Conflict duration UN peacekeeping Multiple rebel signatories Log pseudo-likelihood Akaike information criteria (AIC) Bayesian information criteria (BIC)

–1.390* (0.575) 15.377*** (0.945) –0.178 (0.924) –2.266 (1.202) –2.577* (1.141) 0.000 (0.000) –0.031 (0.982) –16.352*** (0.549) 0.002 (0.001) 0.002 (0.003) –0.003 (0.002) 0.006 (0.004) 0.002 (0.002) 0.000 (0.000) 0.002 (0.003) 0.004 (0.003) –41.132 114.263 174.755

Model 2b

–0.265 (0.424) 2.077** (0.716) 0.614 (0.516) 0.218 (0.588) –0.636 (0.735) 0.000** (0.000) 0.305 (0.547) –0.528 (0.712) 0.001 (0.001) –0.002 (0.002) 0.001 (0.002) –0.003 (0.002) 0.001 (0.002) 0.000 (0.000) –0.001 (0.002) 0.002 (0.001) –121.208 274.416 334.908

Note: Robust standard errors clustered on government-rebel dyad are in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

42

Martin Ottmann and Johannes Vüllers

determining postconflict power relations. Particular instances of powersharing practices are not necessarily adopted randomly, but might be shaped by certain characteristics of the civil conflict or the postconflict situation. If this is the case, these potential confounding variables might affect the impact of power-sharing practices on the emergence of nonviolent and more inclusive power relations between the government and rebels. Unfortunately, we do not know of any statistical package implementing sophisticated instrumental variable estimation or treatment effect models for the Fine and Gray (1999) competing-risks regression models employed in our analysis. Following previous studies (Hartzell and Hoddie 2007; Mattes and Savun 2009), we therefore employed an alternative approach and estimated statistical models in which the presence of each of our central explanatory variables was predicted by characteristics of the civil conflict and the postconflict country. If we found any statistical association between the characteristics of the civil conflict or the postconflict situation and our power-sharing measures, we might have found a confounding variable potentially driving our findings. For this robustness check, we employed logistic regression on a sample of 111 postconflict periods. We collapsed each postconflict period in our sample into one distinct observation with the binary variables measuring cabinet power sharing, parliamentary power sharing, and military integration taking the value of 1 when the respective power-sharing practice occurred at any point during the postconflict period. We included the following factors as independent variables: conflict incompatibility, intensity and duration, the presence of UN peacekeeping, relative rebel strength, and the postconflict country’s life expectancy. Each of these variables took the value it assumed at the day the peace agreement was signed. The results of the logit regressions of these independent variables on our power-sharing variables are presented in Table 2.5. First, we can see that none of the included independent variables had a statistically significant effect on the presence of cabinet power sharing. The most common form of postconflict power sharing between the government and rebels was apparently not determined by any factor under analysis. Turning to parliamentary power sharing, conflict incompatibility perfectly predicted the dependent variable. This is due to the fact that there was not one instance in our sample in which rebel groups fighting in a territorial conflict received parliamentary seats. We also found that a country’s life expectancy and whether rebels were stronger than government forces were statistically significant predictors of parliamentary power sharing. The last model in Table 2.5 reveals that the presence of multiple rebel signatories made military integration more likely. Moreover, we found that the presence of stronger rebel groups also perfectly predicted this outcome. Overall, the measures for conflict incompatibility, the presence of multiple rebel signatories, rebels being stronger than government forces, and a

Government-Rebel Relations

43

Table 2.5 Results of Endogeneity Test for Power-Sharing Practices

Conflict incompatibility Conflict intensity Conflict duration UN peacekeeping Multiple rebel signatories Rebels as strong as government Rebels stronger than government Life expectancy (ln) Constant Log pseudo-likelihood Akaike information criteria (AIC) Bayesian information criteria (BIC) Observations

Cabinet Power Sharing

Parliamentary Power Sharing

Military Integration

–0.511 (0.717) –0.047 (0.513) 0.000 (0.000) –0.933 (0.587) 0.410 (0.444) 0.428 (0.496) 0.492 (0.945) 0.954 (1.587) –4.172 (6.269) –71.047 160.095 184.399 110

Omitted

–0.349 (0.647) 1.234 (0.653) –0.000 (0.000) –0.900 (0.622) –2.247*** (0.617) –0.760 (0.577) Omitted

1.528 (0.889) –0.000 (0.000) –0.884 (1.454) 1.518 (0.865) –0.655 (1.075) 3.122** (1.096) 8.728** (3.026) –37.688** (12.318) –25.68 67.368 87.186 88

–0.231 (1.655) 1.260 (6.591) –56.245 128.491 149.646 104

Notes: Robust standard errors clustered on government-rebel dyad are in parentheses. Omissions are due to perfect prediction. *p < 0.05, **p < 0.01, ***p < 0.001.

country’s life expectancy were potentially confounding variables affecting the impact of either parliamentary power sharing or military integration. As we showed in Table 2.2, however, the inclusion of conflict incompatibility, rebel strength, and life expectancy did not fundamentally alter our findings regarding the impact of parliamentary power sharing on nonviolent power relations. The same was true for the effect of multiple rebel signatories and rebel strength on the relationship between military integration and the emergence of inclusive power relations (see Table 2.3). We, therefore, conclude that our findings are reasonably robust.

Discussion In this chapter, we argued that political and military power-sharing institutions can help to transform government-rebel power relations from violent to nonviolent and from exclusive to inclusive. The statistical analyses support our theoretical argument that power-sharing institutions help to change

44

Martin Ottmann and Johannes Vüllers

a zero-sum logic into one that is positive-sum in nature (see Table 2.6). The findings widen our view on the possible impacts of postconflict power-sharing institutions on government-rebel relations and, thus, contribute new aspects for analysis by both of the power-sharing schools mentioned earlier. We found, however, that some power-sharing institutions have unintended consequences for power relations. While military power sharing showed the hypothesized causal direction, cabinet and parliamentary power sharing did not act in the expected ways. The most surprising effect is that cabinet power sharing decreased the likelihood of nonviolent and inclusive power relations. Although the findings should not be overestimated due to the lack of statistical significance, they are in line with a growing number of studies that see postconflict power sharing in a critical light (Tull and Mehler 2005). An explanation for this effect could be that members of the rebel leadership take over the positions in the government and, thus, are able to exploit resources through corruption linked to these power positions. Therefore, the rebel leadership has great interest in freezing the power relations to remain in a position to spoil the peace process if their demands are not met by the government. As a consequence, the rebels will have an incentive to take action to disturb the transformation process of power relations. Our finding that parliamentary power sharing had a positive effect on the transformation of government-rebel power relations in the military and political dimensions constitutes good news for the international policy community. While cabinet power sharing focuses on the individual level, parliamentary power sharing focuses on the group level. We think that these distinct foci of the power-sharing institutions explain their different impacts. Parliamentary power sharing allows the rebels to transform their organization into a political party and to remain a group.

Table 2.6 Comparison of Hypotheses and Empirical Findings Nonviolent Power Relations

Inclusive Power Relations

Hypothesis

Finding

Hypothesis

Finding

Cabinet power sharing

Increasing effect

Increasing effect

Parliamentary power sharing

No effect

Military integration

Decreasing effect

Decreasing effect, but not significant Increasing effect Decreasing effect, but not significant

Decreasing effect, but not significant Increasing effect Increasing effect

Increasing effect Increasing effect

Government-Rebel Relations

45

The discussion of unintended consequences of power sharing in a postconflict situation highlights the need to go beyond the main foci on democracy and civil conflict recurrence of the two dominant power-sharing schools. As we show in this chapter, power-sharing institutions designed to stop the violence and to prevent conflict recurrence affect government-rebel power relations in different ways. Some power-sharing institutions, foremost among them parliamentary power sharing, are able to improve the quality of peace by transforming power relations. Cabinet and partly military power sharing, however, also work in the opposite direction and hamper the transformation process. Thus, power-sharing institutions have different effects on the quality of peace in the postconflict period that are also different from their effects on peace duration.

Notes 1. Not every rebel group will ultimately transform into a political party. Aside from becoming a fighting actor once again, rebel groups can simply cease to exist. Examples include the Ugandan UNRF II and the Forces Armées pour la République Fédérale (FARF) in Chad. 2. In our empirical analysis, we also controlled for the mere promises of power sharing made in the peace agreement resolving the civil conflict. 3. Parliamentary power sharing is always terminated once a new parliament constitutes itself based purely on national elections. In such a case, no seats can be guaranteed to the rebel group anymore. 4. We thank Håvard Hegre, Lisa Hultman, and Håvard Mokleiv Nygård for granting us access to their UN peacekeeping data. We extended their data collection up to 2011. 5. The decrease in the number of failure events for each state of power relations had no implications for the statistical analysis. When it came to sample size considerations, it was the overall number of failure events that mattered, not the number of events for each type of failure under analysis. 6. We also estimated models with robust standard errors clustered on conflicts and peace agreements. By and large, the results remained the same. Regression results of these additional models are available on request.

3 The Transformation of Armed Organizations into Political Parties John Ishiyama

In this chapter, I examine how different types of power sharing that appear in postconflict peace settlements affect power relations within rebel parties. Thus, consistent with the central question addressed by this book, I focus on the question of how different power-sharing arrangements impact the balance of power within groups; in this case, the former rebel political parties. I contend that all rebel groups contain potential fault lines along which organizational splits may occur, and that different types of power sharing affect whether parties split based on these lines of internal cleavage. I hypothesize that different power-sharing arrangements create different pressures facing rebel groups transforming into political parties (or what I refer to as “rebel parties”),1 which can then impact on the probability of such organizations splitting. Indeed, as some scholars have noted, the transformation of actors could result in new conflicts within the various conflict parties (Kalyvas 2003). Whether they split or not can also affect the likelihood that “spoilers” emerge that may jeopardize the duration of the peace (Asal, Brown, and Dalton 2012; for a discussion of the role of spoilers, see Nilsson and Söderberg-Kovacs 2011). In this chapter, I offer a theory as to why certain types of power sharing (political, military, economic, and territorial) have differential effects on whether a rebel party experiences organizational splits in the postconflict period. I propose a set of hypotheses regarding whether or not a rebel party will experience an organizational split, and test these hypotheses using data from fifty-two armed organizations that transformed into political parties that sought to compete in the first postconflict elections. I employ both Weibull’s and Cox’s hazard estimation techniques to examine the relationship between power-sharing arrangements and the probability that a rebel party will split following the end of civil war. 47

48

John Ishiyama

The Effects of Types of Power Sharing A considerable amount of work over the years has focused on the positive impact that types of power sharing have on promoting peace after civil wars (Roeder 2005). Caroline A. Hartzell and Matthew Hoddie (2003, 2007) argue that postconflict settlements that include key power-sharing provisions between the government and rebels help to alleviate the commitment problem and provide some guarantees that neither side will renege on peace agreements (see also Mattes and Savun 2009). Madhav Joshi and T. David Mason (2011) argue that power sharing helps expand the “winning coalition” and, thus, provides for checks against reneging on peace agreements and the restart of civil wars. From the rebel’s perspective, power-sharing provisions “ensure domestic groups that they will not become victims of discrimination and violence in the new state” (Mattes and Savun 2009, 140). However, there has not always been consensus on how to define the various forms that power-sharing provisions take. Barbara Walter (2002) argues that power sharing takes political, territorial, and economic forms, whereas Hartzell and Hoddie (2003) add military power sharing as a further category (see also Mattes and Savun 2009; Jarstad and Nilsson 2008). Although these approaches tend to identify political power sharing as involving the provision of political offices to the rebel leadership as part of the peace settlement, Bumba Mukherjee (2006a; 2006b) includes as part of his definition of political power-sharing settlements in which rebels are merely allowed to transform into political parties and participate in elections. Generally, the literature on power sharing has focused on the impact of power-sharing provisions on peace duration and the probability of conflict resumption. However, much of the empirical evidence has been rather mixed (Binningsbø 2013). For instance, although Walter (2002) argues that political and territorial power sharing improve the prospects for maintaining peace (albeit fortified by third-party guarantees), and Michaela Mattes and Burcu Savun (2009) found that political power sharing prolongs peace duration, Philip G. Roeder (2005) contends that political “power-sharing dyads” tend to be more conflict ridden than dyads that do not involve power sharing. Further, Mukherjee (2006b) found that political power sharing in the wake of a settlement resulting from a stalemate (which is most often the case) shortens the length of postconflict peace duration. Anna Jarstad and Desirée Nilsson’s findings (2008) suggest that military and territorial power-sharing arrangements encourage postconflict peace duration, whereas political power-sharing arrangements do not have any effect on peace duration and prevention of conflict restart. Finally, Hartzell and Hoddie (2003, 2007) suggest that the impact of power sharing is cumulative; that is, if a settlement includes more power-sharing components, the likelihood of a sustained peace is greater.

Transformation of Armed Organizations into Political Parties

49

What generally has been missing from the literature is the impact that power sharing has on other aspects of the postconflict political environment. As Chapter 1 of this volume points out, although power sharing in postwar countries has received a fair share of scholarly attention over the past decade, most of this has focused on postconflict peace duration. What have been less explored are the effects that may have “been unintended by the architects of these measures.” Indeed, a virtually unexplored consequence of power sharing is the impact it has on the transformation of actors after a civil war. Although there has been some work on spoilers (e.g., Nilsson and Söderberg Kovacs 2011) and also some work on the fragmentation of rebel groups and the effect of this fragmentation on the course of a conflict and how it ends (K. G. Cunningham 2011), there has been no work to my knowledge that has examined the impact of power-sharing arrangements on the fragmentation of rebel parties during the postconflict period.

Toward a Theory of Power Sharing and Rebel Party Splits A considerable amount of literature has recently appeared that has examined the transformation of rebel groups into political parties. However, the focus of such literature has been on whether rebel groups “successfully” transition into a political party, rather than on the organizational transformation of these parties once they begin to engage in political competition (Manning 2004, 2007; Söderberg-Kovacs and Hatz 2016).2 This is somewhat surprising in that, as some scholars point out, the organizational transformation is the most important transformation for successful adaptation to new political circumstances. According to Jeroun de Zeeuw (2007), a key challenge facing a rebel group’s transformation into a political party is the necessary transformation of internal power configurations (e.g., the internal devolution of power) and the accompanying organizational restructuring. The organizational demands for a rebel group-turned-party are very different when the purpose is military operations as compared to campaigning for elections. However, perhaps the organizational transformation of rebel groups as they transform into political parties is the most crucial. An important exception in the literature is de Zeeuw’s (2007) work on rebel group transformations into political parties. He highlights the internal struggle within the rebel group as it transforms into a political party. De Zeeuw differentiates between (1) successful transformations (where the reformist perspective has won); (2) partial transformations (where an internal stalemate exists); and (3) facade transformations (where effectively the conservatives have won the internal struggle). The successful transformation from rebel group to political party is indicated by the complete disarmament and demobilization of its fighters, the renunciation of violence, a demonstrated

50

John Ishiyama

commitment to implement the peace accords, and an acceptance of elections as the only legitimate means to political power. Successful transformations from this perspective include the National Resistance Movement/Army (NRM/A) in Uganda, the Ethiopian People’s Revolutionary Democratic Front (EPRDF) in Ethiopia, the Frente Sandinista de Liberación Nacio-nal (FSLN) in Nicaragua, the Mozambican National Resistance (RENAMO), and the Frente Farabundo Martí para la Liberación Nacional (FMLN) in El Salvador. A partial transformation occurs when rebel leaders recognize that their movement has to change to participate in political life, but they do not completely cease the armed struggle, resulting in a hybrid military political organization. The Irish Republican Army (IRA) is an example of this model, as are Hamas (in Palestine), Hezbollah (in Lebanon), Fatah in Palestine, the Conseil National Pour la Défense de la Démocratie–Forces pour la Défense de la Démocratie (CNDD-FDD) in Burundi, and the Sudanese People’s Liberation Movement/Army (SPLM/A) in Sudan. In a facade transformation, the group speaks the rhetoric of rebel-to-party transformation as a pretext for continuing to receive money and support, erecting a facade political front to mask their military activities. Examples include Colombia’s Ejército de Liberación Nacional (ELN) and Fuerzas Armadas Revolucionarias de Colombia (FARC), Afghanistan’s Hizb i Islami, and the Liberation Tigers of Tamil Eelam (LTTE) in Sri Lanka. This work highlights the importance of internal struggle and power relations within rebel parties, and how they evolve after the end of a conflict. However, there are two shortcomings to this approach. First, the De Zeeuw approach (2007) assumes that once a rebel group becomes a party, these internal contradictions end (which the most recent experience of RENAMO in Mozambique would question). It is likely that these tensions will continue for some time, even after the decision to compete in elections. Indeed, it may be the case that some in the rebel party see the necessity of transforming into a political party to win elections, whereas others in the same party may see participation in elections as a way to buy time until the “people’s struggle” can begin again, even among those parties that, in their view, have made a successful transition. Second, De Zeeuw’s approach does not examine how characteristics of a peace settlement, particularly how types of power-sharing arrangements, impact on these internal contradictions within rebel party organizations. Thus, to address these issues, in this chapter I seek to formulate a theory that focuses on how types of power sharing impact internal “fault lines” within rebel parties. To do so I begin with a discussion of the literature on political parties, particularly those that experience organizational transformation. In general, parties are the organizational conduit that links political entrepreneurs with the market of voters and political elites. Parties provide organizational resources to participants, which include two different kinds

Transformation of Armed Organizations into Political Parties

51

of capital—ideational and administrative (Hale 2005). Ideational capital is the identity and values that attract voters to particular party symbols such as sympathy for the party’s platform and proposed policies or psychological benefits. Administrative capital is more material—it includes the offices and power that can be doled out to ambitious politicians. As Angelo Panebianco (1988) notes, as parties develop over time they tend to move from systems of solidarity to a system of interest. A system of solidarity is based on the concept of a “community” of equals in which the participants’ ideological ends coincide (and where ideational capital dominates). A system of interest, on the other hand, is where the party becomes a collection of individuals with different ambitions, goals, and aspirations, and the party becomes an organizational means to manage these differences. It stands to reason, then, that a similar transformation should also occur within a rebel group transformed into a political party. Indeed, as the rebel group enters into electoral competition, and especially if it wins seats in the legislature, this is likely to activate the office-seeking impulse within the party, both by transforming existing leaders into ambitious politicians and by attracting outside individuals to the party who see it as a potential vehicle for their own political ambitions. However, the differences between the ideationally motivated members of the organization and the office seekers are also likely more acute than in other parties that have not fought a civil war. Since the risks of participating in a conflict are so high for potential participants, and the immediate benefits are so low, it is likely that the largest proportion of rebel group members are those who are attracted for reasons other than administrative capital. Offering the social incentive of fighting for a good cause instead of loot is likely to generate highly committed and ideologically oriented members (Weinstein 2007). These individuals are unlikely to view favorably the compromise of ideals and values for the sake of electoral gain. Thus, there are potential fault lines within rebel political parties. What are the potential lines along which rebel parties can fracture? In general, there are two sets of potential lines of cleavage within a rebel party that can lead to internal divisions within the group. These can be divided into horizontal and vertical cleavages. Horizontal cleavages involve potential disputes between factions led by contending leaders within the rebel groups. These usually involve interpersonal rivalries between contending leaders. Although these can be based on ideological differences, they are often driven by personal rivalries and animosities. A second fault line is the vertical cleavages, or those between the leadership of the rebel organization and the rank-and-file supporters, and armed fighters, in the group. Generally, as many scholars have argued, all political groups, and particularly parties that seek political power via elections, are made up of different kinds of participants. These participants are

52

John Ishiyama

motivated by different goals and aspirations, some by the pursuit of ideational benefits (e.g., attraction to a party’s ideology, or deriving some psychological benefit from belonging to a cause) and some by the pursuit of administrative or career benefits (Panebianco 1988; Schlesinger 1985). It is important to note that these are not mutually exclusive aspirations. Indeed, being a “true believer” may translate into upward career mobility. However, there are individuals who are more motivated by ideational concerns, and others who emphasize career advancement relatively more than attraction to the organization’s ideal purpose. Thus, as the rebel groups transform from an organization designed to fight a civil war, to one designed to win an election or to govern, there is tension that emerges within a rebel organization between those who believed in the cause and those who participated for reasons of personal ambition. For the former, even the act of participation in a power-sharing agreement represents a “selling out” of the movement’s principles. How power sharing is designed in a peace settlement will thus impact how these fault lines develop. Types of Power Sharing and Their Impacts on Party Splits

Hartzell and Hoddie (2003, 2007) identify four types of power-sharing arrangements that are often part of a negotiated settlement: (1) political; (2) military; (3) territorial; (4) economic. As mentioned above, the previous literature has suggested that each of these types have differential effects on peace duration (Joshi and Mason 2011). Political power-sharing arrangements involve providing the rebel party with access to offices in the negotiated settlement. This may involve the allocation of ministerial or cabinet portfolios to rebel leaders, which provides rebels access to power and resources.3 On the one hand, having access to such resources allows leaders to reward followers. This should have a powerful dampening impact on the likelihood of party splits, inasmuch as there is now much to lose by splitting, even for those within the rebel party who suspect the leadership has sold out the revolution’s ideals. Hypothesis 1a: Political power sharing will reduce the likelihood of rebel party splits. On the other hand, because political power sharing often involves only a few political offices to distribute (usually just to top rebel leaders), one might expect the opposite to occur. The availability of only a few top political offices to distribute may activate potential vertical cleavages within a rebel party, given that there may not be many to distribute to the rank and file. Thus, we also might expect:

Transformation of Armed Organizations into Political Parties

53

Hypothesis 1b: Political power sharing will raise the likelihood of rebel party splits. Military power sharing involves some degree of integration of rebel fighters into the national army. Joshi and Mason (2011) argue that as a result, the inclusion of rebel forces makes it difficult for either government or rebels to launch a surprise attack against one another or monopolize the military as an instrument of policy. Thus, military power sharing should provide incentives on the part of both rebels and government to pursue their goals through peaceful means. However, military power sharing also provides opportunities for career advancement for the rank-and-file fighters within the rebel organization. Indeed, by integrating rebel fighters into the national military, this should help assuage the concerns of those fighters that their leaders are selling out the principles of the revolution. Indeed, unlike political power sharing, military power sharing that calls for military integration would mitigate vertical cleavages within the rebel party because there are many more positions to be distributed in the military to rank-andfile members of rebel groups. Thus, military power sharing should have a dampening effect on party splits after the peace settlement. Hypothesis 2: Military power sharing will reduce the likelihood of rebel party splits. Territorial power sharing promotes the decentralization of authority to self-governing regional units. Such arrangements institute self-governance and provide a framework for regional parties to compete for office at the local level. Consolidating that base would also allow rebel parties leverage to extract concessions from the national government. Further, access to local power also allows for the building of a springboard to launch a bid for national power if they seek to do so (Brancati 2006). However, although territorial power sharing is attractive to the leaders of rebel groups in that it gives access to power and resources, albeit at the regional level, it also means that the goals of independence in the case of separatist movements or the overthrow of the existing government in the case of a revolutionary group would remain essentially unfulfilled. This would provide elements within the rebel group with an opportunity to claim that the leadership betrayed the ideals of the group and sold out the revolution. This could create resentments that align along the fault lines identified above. Thus, we would expect that territorial power-sharing arrangements would be most likely to result in rebel party splits. Hypothesis 3: Territorial power sharing will raise the likelihood of rebel party splits.

54

John Ishiyama

Economic power sharing involves some redistribution of economic resources as part of the peace settlement. Thus, by granting access to resources, particularly to economically marginalized groups that may have provided the rebel group with its rank and file, this should provide incentives for the members of the rebel group not to defect from a rebel party. On the other hand, economic resources may provide an incentive for rival factions within the rebel party to contest for control of these resources. Hypothesis 4a: Economic power sharing will reduce the likelihood of rebel party splits. Hypothesis 4b: Economic power sharing will increase the likelihood of rebel party splits. Control Variables

In addition to the above propositions, the literature has suggested a number of control variables that would need to be included in the analysis. First, there are the characteristics of the rebel organization itself, which impact on whether that organization splits after the conclusion of a peace settlement. One of these is the age of the organization. In the literature on changing party identities (Janda et al. 2005; Ishiyama 2005), for instance, frequency of name changes is taken to signify a lack of institutionalization of the party. Applying this logic to the rebel organization, the older the organization, the more likely that the identity of the organization has become institutionalized and thus the less likely that party splits will occur after the end of the civil conflict. A second organizational characteristic is whether the rebel group originated as an alliance of independent parties and groups (which was the case, e.g., with the FMLN in El Salvador) or as a single organizational unit. Splits presumably would be more likely among rebel parties that began as alliances of disparate groups, united only by a common enemy. Third, whether or not groups had a history of organizational splits prior to a peace settlement should impact the likelihood of a postconflict split. On the one hand, a history of organizational splits may make future splits more likely. On the other hand, the departure of dissident groups within a rebel organization prior to the peace settlement may reduce the likelihood of future organizational splits. Another factor that may impact on whether a former rebel organization splits after a peace settlement is whether or not the group represents an ethnic or religious minority. A considerable amount of literature has discussed the political impact that civil wars fought on the basis of ethnicity have on postwar politics. For instance, it is widely argued that when ethnic divisions

Transformation of Armed Organizations into Political Parties

55

have either caused civil war or been hardened by ethnically based violence (Kaufmann 1996), they will make democratization particularly problematic. Similarly, ethnically or religiously based rebel organizations that represent such groups may have greater reasons not to split organizationally, given the hardening effects of ethnicity or religion on group cohesion. Further, historical contextual factors are also identified in the rebel groups-to-political parties literature that can affect the organizational development of rebel groups-turned-political parties. Importantly, as de Zeeuw (2007) notes, one of the legacies of intense military conflict is the tendency to lead to a greater centralization of power within an organization. Thus, we would expect that the legacy of the previous conflict, in terms of intensity and violence, will likely increase the tendency in the successor organization to maintain this tradition of centralization. Further, the more intense the conflict, the less likely the former rebel party will make the successful transformation to a political party. On the one hand, this might suggest that such organizations will be more cohesive and less likely to split. On the other hand, the greater the intensity of the conflict, the more likely it is that at least a portion of a rebel group will hold a great deal of resentment and suspicion about selling out the revolution (de Zeeuw 2007). This may result in greater pressures for the rebel party to split. In either case, the nature of the organization should impact on whether it remains cohesive or splits (nature meaning whether the organization is ethnic or ideological). In addition to these controls, there should also be some consideration of the environment in which organizational development occurs, particularly the economic and political environments. Indeed, one might expect that in wealthier economic environments, there may be an incentive for dissidents to strike out on their own (i.e., the probability of an organizational split rises) because of greater access to resources than in poorer environments. Further, one might also expect that in more open political systems (i.e., relatively more democratic ones), there would be a greater incentive to more openly express organizational dissidence; hence, raising the probability of organizational splits.

Data and Variables To test the foregoing hypotheses, I collected data on fifty-two rebel organizations that chose to participate in the postconflict political process after the end of a civil war (the list is shown in Table 3.1). This list includes both cases in which the rebel organization won the civil war and then transformed into the governing party (seventeen cases) and cases where there was a peace agreement (thirty-six cases). These included not only the development of organizations after the conclusion of a comprehensive peace settlement

56

John Ishiyama

(as in Mozambique, Nepal, and Tajikistan), but also those that were successful in a war of separation (as in Timor-Leste, Nagorno-Karabakh, and Transnistria) and victorious rebel organizations (as in Uganda and Rwanda). The dependent variable for this study was coded as 1 for the year in which a major organizational split occurred within the rebel party involving the creation of a separate political organization apart from the original rebel party. A leader within the rebel party defecting to another party did not count as an organizational split—only if the split involved the establishment of a rival organization was it considered a major organizational split suffered by the rebel party. The unit of analysis that I employed in this study was party organization/year, from the time that the end of the civil war occurred until either the moment that the first major party organizational split occurred or the end year of the study, which was 2014 (meaning the data were right censored).4 The total number of cases in the base dataset numbered 743, with 52 rebel parties and 22 organizations experiencing major splits. To test the impact of power-sharing arrangements on whether a rebel party splits after the end of a civil war, I first identified whether military, political, economic, and territorial power-sharing agreements existed between the government and the rebels. These data were originally from Hartzell and Hoddie (2003), which were later updated by Joshi and Mason (2011) who used additional information for each case from Third-Party Intermediaries and Negotiated Settlements Dataset (Frazier and Dixon 2006) and Keesing’s Record of World Events. I further augmented and updated the data until 2014, using similar sources. As with the previous studies, the power-sharing variables were dichotomously coded. However, unlike Hartzell and Hoddie but like Joshi and Mason, these variables were coded only if the power-sharing provisions were implemented. To measure the characteristics of the rebel group itself, I calculated the number of years from the reported founding of the rebel organization until the year of the observation after the end of the civil war. This meant that, with each passing year, the organization was one year older. The older the organization, the more institutionalized the organization will be, and hence the less likely the rebel party will split. Second, whether or not the rebel party originated as a federation or alliance of groups and parties was coded dichotomously. A rebel party was coded as a federation or alliance of parties if, when the rebel organization was established, it was made up of a merger of different organizations such as the Communist Party of Nepal (Maoist) or the FMLN in El Salvador. Finally, whether rebel groups experienced organizational splits prior to the end of the civil war was also coded dichotomously. These data were derived, where possible, from the official rebel party websites, or from sources such as Keesing’s, the Political Handbook of the World and Political Parties of the World (Banks, Muller, and Overstreet, 2017; Sagar 2009).

57 Table 3.1 List of Fifty-Two Cases of Rebel Parties Name of Armed Group

Name After War Ended

Islamic Society Unity Party

Jamayat-e-Islami Hizb-e Wahdat-e Islami Afghanistan

National Islamic Movement of Afghanistan Armenian Revolutionary Federation (Nagorno-Karabakh) National Union for Total Independence of Angola (UNITA) Popular Movement for the Liberation of Angola

Junbish-i-Milli Islami Afghanistan

Jamayat-e-Islami Party split Hizb-e Wahdat-e Islami Afghanistan Junbish-i-Milli Islami Afghanistan

Dashnaktsutyun

União Nacional para a Independência Total de Angola

Country

English Name

Afghanistan Afghanistan

Afghanistan

Azerbaijan

Angola

Angola

Angola Bosnia Bosnia

Burundi

Burundi Burundi

Central African Republic Chad Democratic Republic of Congo (DRC) Democratic Republic of Congo

Movimento Popular de Libertação de Angola–Partido do Trabalho National Front Frente Nacional de of Angola Libertação de Angola Serb Democratic Party Srpska Demokratska Stranka Croatian Democratic Hrvatska demokratska Union zajednica Bosne i Hercegovine National Council for Conseil National Pour the Defense of la Défense de la Democracy (CNDD) Démocratie–Forces 1996–2012 pour la Défense de la Démocratie (CNDD-FDD) 2012 National Front de Liberation Front Libération Nationale Party for the Parti pour la libération Liberation of the du peuple hutu Hutu Peoples Forces nationales de libération Union for Democratic Union des Forces Forces for Unity Démocratiques pour le Rassemblement Patriotic Salvation Mouvement Movement Patriotique du Salut Rally for Congolese Rassemblement Democracy (Goma) Congolais pour la Démocratie (Goma) Forces for Renewal

Rassemblement Congolais pour la Démocratie-Kisangani

Founded

Year War Ended

1968 1989

2003 2002

1992

2002

Dashnaktsutyun

1990

1994

União Nacional para a Independência Total de Angola

1966

2002

Movimento Popular de Libertação de Angola–Partido do Trabalho Frente Nacional de Libertação de Angola Srpska Demokratska Stranka Hrvatska demokratska zajednica Bosne i Hercegovine Conseil National Pour la Défense de la Démocratie–Forces pour la Défense de la Démocratie

1956

1992

1954

1992

1992

1992

1990

1992

1994

2002

Front de Libération Nationale Forces nationales de liberation

1985

2002

1992

2006

Union des Forces Démocratiques pour le Rassemblement Mouvement Patriotique du Salut Rassemblement Congolais pour la Démocratie (Goma)

2004

2007

1990

1993

1999

2003

Forces for Renewal

1999

2003

continues

58 Table 3.1 continued Name of Armed Group

Country

English Name

Democratic Republic of Congo Djibouti

Movement of Mouvement de Liberation of Congo Libération du Congo Front for the Restoration of Democracy Front for the Restoration of Democracy Farabundo Marti National Liberation Front Eritrean People’s Liberation Front Ethiopian People’s Revolutionary Democratic Front Guatemalan National Revolutionary Unity Aceh Independence Movement Al-Madhi Army

Djibouti

El Salvador

Eritrea Ethiopia

Guatemala Indonesia Iraq Iraq

Front pour la Restoration de l’Unité et de la Démocratie Front for the Restoration of Democracy–Radicals Frente Farabundo Martí para la Liberación Nacional Eritrean People’s Liberation Front Ethiopian People’s Revolutionary Democratic Front Unidad Revolucionaria Nacional Guatemalteca Gerakan Aceh Merdeka National Independent Cadres and Elites Partîya Demokrata Kurdistan Yeketî Nihtîmanî Kurdistan Palestinian National Liberation Movement (now Fatah) Islamic Resistance Movement (Hamas) Mouvement populaire ivoirien du Grand Ouest Mouvement patriotique de Côte d’Ivoire Ushtria Çlirimtare e Kosovës Hizb Al-Kata’eb AlLubnaniyah Harakat al-Mahrumin

Kurdish Democratic Party Iraq Patriotic Union of Kurdistan Israel and Palestinian National Palestine Liberation Movement–Fatah Israel and Islamic Resistance Palestine Movement (Hamas) Côte d’Ivoire Popular Movement of Côte d’Ivoire Côte d’Ivoire Patriotic Movement of Côte d’Ivoire Kosovo Kosovo Liberation Army Lebanon Lebanese Phalanges Party Lebanon Movement of the Dispossessed, the Lebanese Resistance Detachments Lebanon Hezbollah Hezbollah Liberia National Patriotic National Patriotic Front Front of Liberia of Liberia Liberia Liberians United Liberians United for Reconciliation for Reconciliation and Democracy and Democracy

Name After War Ended

Founded

Year War Ended

Mouvement de Libération du Congo

2003

2003

Front pour la Restoration de l’Unité et de la Démocratie Front for the Restoration of Democracy–Radicals Frente Farabundo Martí para la Liberación Nacional People’s Front for Democracy and Justice Ethiopian People’s Revolutionary Democratic Front Unidad Revolucionaria Nacional Guatemalteca Partai Aceh

1991

1994

1991

2000

1980

1992

1970

1994

1989

1994

1982

1996

1976

2005

2003

2004

1946

1992

1975

1992

1965

1993

1987

1993

2003

2007

2005

2007

Partia Demokratike 1996 e Kosovës Hizb Al-Kata’eb Al1936 Lubnaniyah Amal Hope Movement 1974

1999

Hezbollah 1985 National Patriotic Party 1989

1990 1996

Merged with Unity Party

2003

Merged with United Iraqi Alliance Partîya Demokrata Kurdistan Yeketî Nihtîmanî Kurdistan Palestinian National Liberation Movement (now Fatah) Islamic Resistance Movement (Hamas) Forces Nouvelle Forces Nouvelle

1999

1990 1990

continues

59 Table 3.1 continued

Country

English Name

Moldova

United Work Collective Council Mozambique Liberation Front (Frelimo) Mozambican National Resistance (RENAMO) Communist Party of Nepal (Maoist) Sandinista

Mozambique

Mozambique

Nepal Nicaragua Papua New Guinea The Philippines Rwanda Sierra Leone Sudan

Tajikistan Timor-Leste

Uganda Venezuela

Zimbabwe

Name of Armed Group

Name After War Ended

United Work Collective Council Frente de Libertação de Moçambique

Respublikanskaya 1989 Partiya Pridnestroviya Frente de Libertação 1962 de Moçambique

Resistência Nacional Moçambicana

Resistência Nacional Moçambicana

1975

1992

Unified Communist Party of Nepal Frente Sandinista de Liberación Nacional Bougainville Peoples Congress Moro National Liberation Front Front Patriotique Rwandais Revolutionary United Front Party Sudan People’s Liberation Movement (SPLM) Islamic Renaissance Party Frente Revolucionária de Timor-Leste Independente National Resistance Movement Movimiento V [Quinta] República

1994

2006

1961

1979

1988

2001

1969

1996

1987

1994

1991

1999

1983

2005

1990

1997

1974

2002

1981

1986

1982

1994

1963

1979

Communist Party of Nepal (Maoist) Frente Sandinista de Liberación Nacional Bougainville Bougainville Revolutionary Army Revolutionary Army Moro National Moro National Liberation Front Liberation Front Rwanda Patriotic Front Patriotique Front Rwandais Revolutionary Revolutionary United Front United Front Sudan People’s Sudan People’s Liberation Liberation Movement Movement (SPLM) (SPLM) Islamic Renaissance Islamic Renaissance Party Party Revolutionary Front Frente Revolucionária for an Independent de Timor-Leste East Timor Independente National Resistance National Resistance Movement Movement The Revolutionary Movimiento Bolivariano Bolivarian Revolucionario–200 Movement–200 Zimbabwe African Zimbabwe African National Union National Union– Popular Front

Zimbabwe African National Union– Popular Front

Founded

Year War Ended 1992 1975

The second variable was also coded as a dummy variable indicating whether or not the rebel organization was based on support of an ethnic or religious group. An organization was coded as ethnic if its basis of support was judged to lie with a single ethnic group. A religious organization was also coded as 1 if that religious group was a political minority in a country. Thus, for example, an Islamic organization in Afghanistan would not be coded as 1 but the Shia Hazara organization would be, as would the ethnic Uzbek organization, Jumbish-i-Milli. Finally, from Joshi and Mason (2011), I measured the intensity of the previous conflict by using the log of battlerelated deaths of the previous civil war.

60

John Ishiyama

Finally, to control for some aspects of the economic and political environment, annualized gross domestic product (GDP) per capita (purchasing power parity [PPP] indexed by 2011) data and Polity 2 scores were employed to measure relative wealth annually and the level of democracy. As noted above, it is possible that wealthier contexts would provide some greater incentive to split, because it may be possible for internal organizational dissidents to gain access to resources by striking out on their own. Further, the higher the level of democracy, the more openings exist for entrance into the political market, thus providing a potential incentive to split. Whatever the case, these two variables were incorporated as general controls.

Analysis Tables 3.2 and 3.3 show the results of tests of the hypotheses using a hazard analysis. I used Weibull’s and Cox’s techniques because both have their advocates and there is some debate as to which is the superior technique. In health sciences, economics, and international relations, the Weibull technique has been widely used (see Carroll 2003; Werner 1999). On the other hand, the Cox proportional hazards regression technique has been popular in most of the rest of political science, stemming largely from extensive experience in its application and the fact that it is distribution free—no assumption needs to be made about the underlying distribution of survival times to make inferences about relative failure rates. The Cox model is a statistical method that produces a hazard rate that reflects “the risk an object incurs at any given moment in time, given an event has not occurred” (Box-Steffensmeier and Jones 1997, 1419; see also Box-Steffensmeier, Reiter, and Zorn 2003). By estimating this model, I obtained a hazard rate of an organizational split occurring for a given type of power-sharing agreement (political, military, territorial, economic). An increase in the hazard rate suggests that there is an increase in the likelihood of an organizational split, and a decrease in the hazard rate suggests there is a decrease in the likelihood of an outbreak of civil war. To check for the proportionality assumption, the Schoenfeld residualbased test is recommended to determine the presence of nonproportional hazards (Box-Steffensmeier, Reiter, and Zorn 2003, 35). I conducted the test for nonproportionality for all models to determine which variables violated the proportionality assumption. The p-values for each covariate and global test indicated that the assumption that the hazard rates were proportional was not violated by any of the models (Box-Steffensmeier and Jones 2004). In addition, to check for the potential for multicollinearity in all models, I conducted a variance inflation factor (VIF) test. All VIF scores were less than 2 for any of the explanatory variables, thus indicating that multicollinearity was not a problem for the analyses.

Transformation of Armed Organizations into Political Parties

61

Table 3.2 Hazard Ratios for Rebel Party Splits After End of Civil War

Variable Political power sharing (0 = no , 1 = yes) Economic power sharing (0,1) Military power sharing (0,1) Regional power sharing (0,1) Age of armed organization (in years) Was armed group an alliance? (0,1) Did armed group have history of splits prior to settlement? (0,1) Ethnic party? (0,1) Religious party? (0,1) Number of battle deaths during civil war (LN) Annual GDP per capita (PPP 2011) Annual Polity 2 score Wald chi-square Probability of chi-square Observations Number of subjects at risk Failures (splits)

Model 1 (Weibull) Hazard Ratio .77 (.69) 2.58 (2.75) 6.51* (7.16) 19.82**** (20.89) .96 (.06) 6.35** (4.90) 1.37 (1.41) 2.31 (2.27) .73 (.49) .78 (.17) .99 (.00) 1.19 (.13) 78.63 .000 697 52 22

Model 2 (Cox) Hazard Ratio .82 (.69) 2.32 (2.40) 5.79* (6.15) 15.57*** (15.08) .96 (.05) 5.76** (4.26) 1.31 (1.42) 2.12 (1.89) .74 (.49) .78 (.17) .99 (.00) 1.17 (.12) 78.64 .000 697 52 22

Notes: Robust clustered standard errors are in parentheses. GDP, gross domestic product; PPP, purchasing power parity. Two-tailed tests: *p < 0.10, **p < 0.05, ***p < 0.01, ****p < 0.001.

Table 3.2 shows the results of a Weibull estimation technique (Model 1) and a Cox technique (Model 2) for all the organizations listed in the dataset. This includes organizations that may have attained power via means other than a negotiated settlement (e.g., a rebel victory). The results reported for each model are remarkably similar. For both Models 1 and 2, although the reported hazard ratios indicated that political power sharing reduces the hazard of party organizational splits (which is suggested by Hypothesis 1a), the relationship was not statistically significant. This means that the null hypothesis of no relationship cannot be rejected. This was also true for economic power sharing (which counters Hypothesis 4a). Interestingly, military power sharing increased the hazard of an organizational split (by a factor of 6 in each model) and the relationship

62

John Ishiyama

Table 3.3 Hazard Ratios for Rebel Party Splits After End of Civil War, Peace Agreement Cases

Variable Political power sharing (0 = no, 1 = yes) Economic power sharing (0,1) Military power sharing (0,1) Regional power sharing (0,1) Age of armed organization (in years) Was armed group an alliance? (0,1) Did armed group have history of splits prior to settlement? (0,1) Ethnic party? (0,1) Religious party? (0,1) Number of battle deaths during civil war (LN) Annual GDP per capita (PPP 2011) Annual Polity 2 score Wald chi-square Probability of chi-square Observations Number of subjects at risk Failures

Model 3 (Weibull) Hazard Ratio 1.60 (1.45) 1.09 (.69) 45.37* (88.23) 22.67*** (26.09) .92 (.08) .98 (1.69) .57 (.69) .30 (.39) .26 (.49) .68 (.18) .99 (.01) 1.27 (.14) 49.98 .000 494 36 16

Model 4 (Cox) Hazard Ratio 1.43 (1.28) 1.14 (.72) 23.44* (31.88) 18.27** (22.13) .94 (.07) 1.18 (1.82) .43 (.50) .31 (.40) .27 (.49) .66 (.21) .99 (01) 1.24 (.15) 49.91 .000 494 36 16

Notes: Robust clustered standard errors are in parentheses. GDP, gross domestic product; PPP, purchasing power parity. Two-tailed tests: *p < 0.10, **p < 0.05, ***p < 0.01, ****p < 0.001.

was statistically significant, but opposite of the direction suggested by Hypothesis 2. Although this will require further analysis, a potential reason for this is that integration of rebel forces into the national army may exacerbate the tensions between various wings of the rebel party (particularly if there is some debate over which units are integrated and which are not). Importantly, Hypothesis 3 was supported. Regional power sharing significantly increased the hazard of an organizational split in the rebel party by a factor of between 15 and 19. This result suggests that regional power sharing may activate tensions within rebel parties inasmuch as territorial autonomy may create resentments within the rebel party that the principles

Transformation of Armed Organizations into Political Parties

63

of the revolution had not been realized (e.g., independence or victory) and that the leadership of the rebel party had betrayed the people’s struggle. However, it should be noted that although regional power sharing may increase the hazard of a rebel organizational split relative to a nonregional power-sharing arrangement, the likelihood of a split was not overall particularly great (there is not much substantive difference between a hazard ratio of .005 and .015). A graphical representation of the relationship between regional power sharing and organizational splits over time is presented in Figure 3.1. As indicated in Figure 3.1, as time progressed (in years), the likelihood of a party split grew at an increasing rate when there had been regional power sharing compared to when there had been no regional power sharing. Further, the likelihood of a split was much more pronounced even in the first few years after a regional power-sharing agreement had been reached Finally, among the control variables, rebel groups that were alliances of parties or those that were ethnic organizations increased the likelihood that a rebel party would split after the end of a civil war. In addition, splits were slightly less likely to occur in relatively wealthier countries, but the marginal effect as suggested by the hazard ratios was small. It might be argued that countries that did not experience an actual peace agreement should be excluded from the analysis. In other words, if

Figure 3.1 Party Split Hazard over Time

64

John Ishiyama

there was no peace agreement (i.e., the rebels were victorious), there would be no power-sharing arrangements. Thus, to check the results reported in Table 3.2, I replicated the analyses, but using only the thirty-six cases where there were actual peace agreements that involved some measure of power sharing (or not). As indicated in Table 3.3, the only relationship that remained statistically significant (and in the posited direction) was the relationship between regional power sharing and whether a rebel party organizationally split (replicating the relationship found in Table 3.2). However, all other relationships were now statistically insignificant (at the standard p < .05 level). Figure 3.2 again, however, illustrates that although regional powersharing arrangements after a civil war may increase the hazard rate when compared to peace agreements that do not include regional power-sharing arrangements, overall, the likelihood of an organizational split of a rebel party was generally low even when projecting across time (noting again that there is not much substantive difference between a hazard ratio of .005 and .015).

Conclusion In this chapter, I presented some preliminary results regarding whether a rebel party splits as the result of power-sharing arrangements found in a peace settlement after a civil war. I formulated a set of hypotheses that focused on four different types of power-sharing arrangements—political, military, territorial, and economic. Although the likelihood of party organizational splits was not particularly great over time, clearly a good number of rebel parties experienced splits (twenty-two of fifty-two cases). Generally, I found that regional power sharing raises the likelihood of a rebel party split substantially when compared to settlements that do not have regional power-sharing arrangements. Now, of course, this observation should be tempered with the possibility that there could be something particular about the rebel parties that were involved in arrangements that included regional power sharing. In other words, it may be the case that rebel parties that seek independence, such as separatist parties, are more likely to split in the first place, quite apart from the characteristics of the power-sharing agreement (i.e., the relationship between regional powersharing arrangements and rebel party organizational splits was spurious). However, although not a perfect proxy for a separatist party, as indicated in the results in Table 3.2, ethnic and religious parties were not more or less likely to split than other parties. Thus, this might support the idea that power-sharing arrangements exert an independent effect on party splits, quite apart from characteristics of the rebel organization.

Transformation of Armed Organizations into Political Parties

65

Figure 3.2 Party Split Hazard over Time, Peace Agreements Only

A considerable amount of additional work needs to be done on this topic. Although I demonstrated in this chapter an empirical relationship between regional power-sharing and rebel party splits, only through careful qualitative analysis of individual cases can we really tell if the internal dynamics within parties were a result of the structure of incentives generated by types of power sharing. However, that will take much more space and time to develop, far beyond the range of this chapter. It should be noted that this study did not examine the nature of splits. Although in previous work (Ishiyama and Batta 2011; Ishiyama 2005), I examined whether rebel organizational splits were peaceful or violent and the effect this had on the durability of postconflict peace, such work was not done regarding rebel party splits. As we noted in 2011, violent splits among rebel organizations were highly correlated with the resumption of a civil war. This finding is likely also true for violent splits within rebel parties. However, this will require further analysis beyond the scope of this chapter. As conflict scholars acknowledge the importance of transforming rebel organizations into political parties or turning “bullets into ballots” as the best guarantee for the development of political stability and potentially for democracy after the end of the conflict, the question of how rebel organizations transform will become increasingly important. In particular, whether rebel parties split, and whether such splits lead to conflict resumption, will become an important research question.

66

John Ishiyama

Notes 1. I refer to the parties that emerge from armed groups as rebel parties. A rebel party is the organization that is the direct successor to the armed group, either because that group registers as a political party itself or that group creates a political party to run in its name. A rebel party is not one that individual leaders defect to or join. 2. An exception is a recent piece by Ishiyama and Batta (2011) that examines the internal organizational transformation of rebel parties. 3. It is important to note here that Mukherjee’s (2006b) inclusion of the situation where rebel parties are allowed to organize is not counted as an instance of political power sharing. This is because allowing an armed group to organize as a party does not guarantee access to government and resources, and as such represents potential (but not actual) power sharing. 4. I corrected for the right-censored nature of the data by using the stset command in Stata 13.

4 The Consequences of Power Sharing at the Local Level Andreas Mehler, Claudia Simons, Denis M. Tull, and Franzisca Zanker

A quick glance through the literature on postconflict power sharing reveals that the local level has been grossly neglected. With few exceptions, the focus of most studies has been on national elites or international brokers (exceptions include Simons et al. 2013; Heitz 2009). This is surprising insofar as most contemporary violent conflicts escalate—and are fought—in precise locations. This happens because of particularly antagonistic interests on the part of competing groups, specific mobilization capacities of locally based rebel organizations, or locally confined transboundary dynamics. While it is not uncommon to interpret war and peace in terms of winners and losers within a given national elite, what power sharing means for the distribution of power at the local level remains obfuscated at best or ignored in most observations. Power-sharing agreements, despite some consistent features, show important variations that can be expected to impact local power constellations. Based on this observation, we ask four key questions: 1. Power-sharing agreements are national elite pacts, but do they play out at all at the local level? Are other matters more important for local power constellations, such as political culture or the immediate effects of war? 2. Power-sharing agreements can be more or less inclusive in terms of who takes part in the negotiations and who gets what. Does inclusiveness matter at the local level? 3. The characteristics of the parties to an agreement will always vary, especially in relation to rebel fragmentation. As a result, how does the relative cohesiveness of the parties impact local power constellations?

67

68

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

4. Finally, there is variation in terms of temporality: some agreements create interim arrangements, others are meant to be permanent, and some power-sharing content is inscribed within postconflict reforms of core institutions with a potentially stronger impact on the local power constellations than the elite pact. When and for how long does power sharing, and related postconflict reform, impact on local power constellations? The academic literature has been mostly silent on the issue of the locallevel effects produced by power-sharing arrangements. A short review helps to identify some useful approaches for the analysis of key processes at stake, although no single approach touches directly on the topic as discussed here. One dimension within the four-level definition advocated by the influential work of Caroline A. Hartzell and Matthew Hoddie is clearly relevant to our argument (Hoddie and Hartzell 2005; Hartzell and Hoddie 2007). The territorial dimension of power sharing explicitly touches the subnational level, mostly translating into some form of decentralization or, less frequently, federalism (see also Hartmann 2013).1 A territorial element can have considerable appeal to former armed groups because it allows them to control politics and resources within a certain territory. This not only bolsters these groups’ security, but also gives them an extraordinary chance to consolidate their power base (and to dominate local competitors) and potentially even resume the war. Subnational interests and concerns are assumed to get a better chance to play out at the national level as well if territorial power sharing is implemented, yet antagonistic interests within a local arena rarely are taken into account. The literature on rebel governance has not paid much attention to the specific postconflict rationale of legitimizing the maintenance of power positions as well as the interplay with state actors at the local level; that is, the production of order (e.g., Weinstein 2007; Humphreys and Weinstein 2007; Persson 2012). Research by Till Förster (2010) on northern Côte d’Ivoire (Korhogo in particular), by Kathrin Heitz (2009) on the rebel-held city of Man in the west of the same country, and by Kristof Titeca (2011) on northeastern Congo are notable exceptions. Conversely, the existing literature on local peace initiatives (Chopra 2009; Frank 2002; Paffenholz and Lundqvist 2003) has tended to ignore the national context, the power games of new (rebel), modern (state, party), and old (traditional) local elites as well as the patron-client ties that may link them to national politics. The interaction between international actors and local actors in liberal peacebuilding models has received some attention, although this has not explicitly touched on power-sharing situations and has either merged the national level with the local level or focused on one exclusively (e.g., Autesserre 2010; Manning 2003). What a transition from rebel organization to a political party involves

The Consequences of Power Sharing at the Local Level

69

in general terms has been aptly analyzed by Carrie Manning (e.g., 2007) and others (see also references cited above). However, the process by which a rebel elite actor “reinvents” himself or herself by taking on a civilian role and at times competing locally with established elites is underresearched. A growing body of literature within the peacebuilding field is focusing on the local level and problematizing the term local itself as opposed to the liberal (international) peacebuilder, particularly within critical security studies (e.g., Mac Ginty and Richmond 2013; Donais 2012; Richmond and Mitchell 2011). This is partly enlightening for our purpose, but there has been no particular attention to local power sharing within this school of thought (for an exception, see Hirblinger and Simons 2014). In summary, there have been only a handful of contributions on the effect of postwar power sharing at the local level despite the hypothesized importance of these measures and especially of territorial power sharing. The effect of such elite pacts on power constellations, in particular at the local level, has not yet received the attention it merits. In this chapter, we identify some of the relevant mechanisms at play. We observe many variations in both conditions and outcomes, offering a basis for more systematic reflection. In terms of case selection, we chose to focus on four African postconflict countries, each with a national power-sharing accord in place during a similar time period (2002–2008). These cases vary on the basis of the historical center-periphery relations in place as well as the levels of violence each experienced during their respective periods of armed conflict. Although Burundi and Liberia are small countries, only in Burundi did the government continuously try to control the entire territory. Kenya is middle-sized and has a fairly strong state, enabling it to control both territory and its borders, while the Democratic Republic of Congo (DRC) is one of the biggest territorial states in Africa with notoriously weak control over its territory. The DRC, Burundi, and Liberia all fought extensive civil wars; Kenya witnessed a rather short period of intense postelectoral violence. We conducted research in two local arenas—former hot spots during periods of widespread violence—in each country. We defined these arenas territorially, considering them to be administrations and populations of subnational units outside the capital and below national state institutions. In our case studies, this refers to more or less urban provincial or county capitals (Goma and Bukavu in the DRC, Ganta and Gbarnga in Liberia, Nakuru and Eldoret in Kenya) as well as more rural areas (Kalehe and Sake in the DRC, and the rural area around Gitega and Bubanza in Burundi). Our understanding of local elites includes: (1) locally deployed administrative elites; (2) elected political elites; (3) traditional elites; (4) economic elites; and (5) military commanders, both with an army and a rebel background. Obviously, each of those groups may have distinct properties. Administrative elites may just have been redeployed or stayed on during

70

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

the conflict; they may have resources, equipment, and legitimacy, or not. Political elites may have more or fewer followers, resources, and legitimacy. Traditional elites are in some cases elected, in others appointed, or they inherit their position—with potential consequences again for followership, resources, and legitimacy.2 Economic elites may or may not have close connections to the political sphere. Demobilization, disarmament, and reintegration programs also will impact the power base of local military commanders (e.g., through access to weapons) as well as potentially affecting their legitimacy and the number of followers each has. Finally, we understand local power sharing in a twofold manner (see Simons et al. 2013). First, powerful local actors might be included directly in the peace negotiations and subsequent agreement as well as central government institutions. Alternately, national elites negotiating peace and sharing power at the center might address specific local concerns in a top-down fashion; consider this a means of pacifying the margins. We define this as the local content of national power-sharing agreements. Second, power sharing can be extended to the subnational level. We define arrangements as constituting an instance of local power sharing if they take the form of power-sharing formulas between belligerents at the local level that are identical or similar to the power-sharing formula at the national level (power balance) or if strategic local arenas are divided among belligerent groups (e.g., the government exercises control over a certain province while leaving or giving other provinces to one or more rebel groups), an arrangement we refer to as power dividing or what we define as a juxtaposition of local power monopolies. We also take into account how the types of power sharing (political, military, economic, and territorial) play out at the local level. We pay special attention to elements of territorial power sharing (including informal types of territorial power sharing; see Zanker, Simons, and Mehler 2015) as we expect that this form of power sharing should have the strongest effects at the local level. We now turn to our empirical analysis, drawing on insights we gained during fieldwork to respond to the four questions listed above.

Kenya Elites frequently have power at the local and the national levels of politics, which is particularly evident in Kenya. Here, elites can play a two-level game to mutually entrench their power capacities at both levels. A national powersharing pact potentially offers additional opportunities for elites to strengthen their position locally, thereby altering preexisting power constellations. Such an impact proceeds from the formation of a (power-sharing) postconflict government that implicitly or explicitly, informally or formally, follows rules of

The Consequences of Power Sharing at the Local Level

71

geographic proportionality. In addition, an agreement can also contain some privileges for a hitherto neglected or rebellious part of a country. This will resound in particular areas, especially when “a son of the soil” is co-opted into an important position, which might potentially end the exclusion of a locally based identity group. In the Kenyan case, Kalenjins in the Rift Valley or Luos in Kisumu were satisfied to some degree when their sons were included in the interim government of Kenya after the disputed presidential elections in 2007. Violence had broken out after disputed election results were decided in favor of the Kikuyu Mwa Kibaki at the cost of the Luo Raila Odinga. Over a thousand people were killed and many thousands displaced in the course of the violence (e.g., Cheeseman 2008). At the time, the Kalenjin group was largely in a coalition with the Luos under the auspices of the Orange Democratic Movement (ODM), though this coalition later split (see below). Many Luos were convinced that their candidate, Raila Odinga, had won the election. When he finally became prime minister as a result of a power-sharing deal, this was at least a strong consolation. One of Kenya’s best-known ethnopolitical entrepreneurs, William Ruto (now vice president of the country), acceded to his postviolence position to represent the Kalenjin ethnic group at the national level, an event that also was meaningful in his hometown of Eldoret. Taking part as a negotiator in the power-sharing negotiations, he became a minister in the power-sharing cabinet. This was of high symbolic importance, especially as Eldoret had been one of the major hot spots during the election violence. Ruto was thus able to confirm his political grip on his North Eldoret constituency in Rift Valley Province, where he had been the incumbent member of parliament (MP) since 1997. In contested areas like Eldoret and its surroundings, such appointments could mean taking sides with only one conflict party. Explicit representation of the local ethnic majority may in fact accentuate majority-minority confrontations in specific areas, as local minorities feel marginalized when the strongman of the opposing dominant group is promoted to a national position and handed a patronage position that permits him or her to lavishly support their own constituency. Scant representation of Kikuyu opposition leaders in, for example, the town hall (as town councilors) showed that this group was limited in its political scope by the clear personal powerhold that existed in the town. This was a strong indication that Kikuyus had lost the power game in Eldoret. One major reason for examining these patterns is to identify whether a given episode of violence was the expression of an elite game regarding national power positions that was being played out at the local level, or whether it was essentially a local conflict about the distribution of resources and power that escalated to reach national proportions. While undoubtedly national political actors played an important part in inciting violence after the

72

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

2007 elections, local roots and historical grievances, which were particularly prominent in the Rift Valley, also played a significant role in the violence that unfolded. Locally relevant actors are arguably of importance in Kenya on a national level because their identity groups are strongly linked to different regions. Nic Cheeseman argues that “the Kenyan crisis needs to be placed in the local understandings of citizenship, belonging and exclusion . . . in order to fully understand the dynamics of conflict, and the state’s inability to respond, we must connect these local debates and frustrations to the way in which the Kenyan state structures both political competition and the means of coercion” (2008, 170, emphasis added; see also Anderson 2010). As a result of their strong local support, two out of the four mediators put forward by the ODM side were MPs from the Rift Valley.3 Each of those individuals went on to receive a ministerial post in the power-sharing cabinet. In 2011, Ruto was relieved of his ministerial duties after a series of corruption allegations.4 By that time, the International Criminal Court (ICC) had also charged him with crimes against humanity due to his alleged role in the postelection violence. Both events led to renewed tensions in Eldoret. Rumors, threats, and discrimination were rife, partially related to the intimidation of ICC witnesses (or their families) and human rights activists (see Human Rights Watch 2013). The ICC cases introduced a rather surprising dimension to power politics in Kenya, moving the focus much more to the national level. Arguing the case of a neocolonial conspiracy with the potential to infringe on Kenya’s sovereignty, Ruto joined the Kikuyu Uhuru Kenyatta (also indicted by the ICC) in the 2013 election campaign under the Jubilee Coalition. This marriage of convenience enabled a narrow election victory (which was further consolidated in 2017 with the next elections), and in theory ensured both Kalenjin and Kikuyu representation at the national level. Locally, this gain in national power—a complete change of affairs after being fired from the cabinet and the ICC allegations—further entrenched Ruto’s position in his constituency. Similarly, over time, as power sharing becomes entrenched or is changed, so will local power constellations. A widely lauded reform process that started in Kenya included a new constitution, promulgated in 2010. As a result, administrative and political devolution were introduced with the first presidential and parliamentary elections under the new constitutional framework held in 2013. While, strictly speaking, the implementation of these new measures was not planned for the interim period of power sharing (2008– 2013), even the mere discussion and eventual inclusion of devolution in the new constitution had effects—partly unexpected—on local power relations. The official reasons given for devolution in the new constitution focus on the fairer distribution of resources and access to government in all parts of Kenya as well as some efforts to resolve long-term conflicts (see Article 174, Kenyan Constitution of 2010). The rationale provided for devolution is inter-

The Consequences of Power Sharing at the Local Level

73

preted differently among various stakeholders in different provinces. At worst, devolution reminds some of ethnoregional decentralization related to the notorious idea of “majimboism” (or regionalization in Swahili). Prominent in the 1990s, majimboism was seen as something akin to ethnic cleansing. It was perceived as especially dangerous in the Rift Valley for Kikuyus who, until 2013, formed a majority in the administration and who feared that “for the ‘indigenous’ Rift Valley . . . devolution . . . would provide an opportunity to reclaim lost lands” (Nyanjom 2011, 26; see also Anderson 2010). Hence, the emphasis in official public discourse was put on equal access to services, a priority reflected in much of the policy literature on the matter (e.g., World Bank 2011). Most importantly, this emphasis can be noted in the semantic preference for devolution rather than decentralization. Even prior to the reform implementation, rolled out in 2013 with the elections of governors, deputy governors, and senators, fears related to the upcoming devolution led to a fragile balancing of local power relations. By 2011, for example, two years before the elections and eventual implementation, relations in Nakuru, one of the hot spots of violence during the postelection violence in 2008, were so fragile that local authorities had to intervene. A local Council of Elders started brokering power-sharing agreements to decide who should run (or be allowed to run) for which seat. In 2012, the National Cohesion and Integration Commission (NCIC) took up this issue and went on to broker about twenty different local peace agreements related to the upcoming devolution. In Nakuru, for example, elders from the Kalenjin and Kikuyu communities signed a peace agreement in August 2012, promising to refrain from violence until the elections. In Uasin Gishu County where Eldoret is located, 74 percent of the electorate voted for the Kenyatta-Ruto Jubilee Alliance in the presidential election of 2013. Notably, it was someone from Ruto’s United Republican Party (URP), Jackson Kiplagat, who won the governor’s seat with 73 percent of the vote. Devolution with a governor aligned with Kenyatta would never have worked in this county. More importantly, the office had to be assumed by someone with proven loyalty to the strongman in Eldoret, Ruto. The elected governor had only one opposition candidate, Margaret Kamar, who had been an MP and minister of education (as part of ODM), during the power-sharing period between 2008 and 2013. When Ruto broke off his alliance with Odinga’s ODM, Kamar did not follow suit, preferring to show her loyalty to Odinga. As a result, observers, administrators, and other local town councilors moved to ostracize her, at least locally, swearing she would never again be elected in Eldoret and Uasin Gishu County.5 Ruto—who by now was firmly playing national politics—made sure to entrench his local power position through his loyalists. In contrast, in Nakuru, one of the only places in the Rift Valley with a significant Kikuyu population, Kinuthia Mbugua, who belongs to President

74

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

Kenyatta’s political party, The National Alliance (TNA), was elected as governor in March 2013. The Kenyatta vote of 80 percent in Nakuru County can be attributed not only to the support for Ruto as elsewhere in the Rift Valley, but also to popular support for a Kikuyu president (and not only the Kalenjin coalition partner Ruto). Considering the Kikuyu population in Nakuru, it was essential that the group be represented under the new devolution arrangement to address majimbo-related fears. This outcome is also likely to be related to conflict-induced population movements. Many Kikuyus from all over the Rift Valley fled to Nakuru from late December 2008 onward, seeking a safe haven. In the aftermath of the postelection violence, this once heterogeneous city now has much more homogenous neighborhoods (or electoral wards) that will indubitably affect election results and local power constellations. Interestingly, Agnes Cornell and Michelle D’Arcy (2014, 181) argue that the reason Mbugua won (while having a Kalenjin running mate) was based on his close connections to the core national elite. This once again reinforces the idea of the two-level power game.

Democratic Republic of Congo The case of the DRC, and more specifically eastern Congo’s North Kivu Province, shows that the appeal of national power sharing is sometimes insufficient to overcome the agendas of rebels with deep-seated roots in local fiefdoms. In North Kivu, insurgents were unwilling to trade their local power base for national positions, partly because they embraced a longterm horizon that went beyond a transitional period of power sharing. As a consequence, the local configuration of political power and, thus, power relations contravened a buy-out strategy centered on national political and economic spoils. At the same time, the resulting decoupling of local and national processes was apparent only to the extent that local groups opposed to continued rebel rule in the province stepped up their contestation. Aligning themselves with national actors, they saw an opportunity to turn the tables on the Rassemblement Congolais pour la Démocratie (RCD) rebellion to reclaim local dominance. In the DRC, nearly continuous warfare since 1998 was formally ended in late 2002 when a comprehensive peace agreement was signed by the war’s main protagonists; that is, the government under President Joseph Kabila and the two major insurgent movements, the RCD and the Mouvement de Libération du Congo (MLC) (see Prunier 2011; Reyntjens 2009). The accord contained a wide-ranging and detailed transitional power-sharing formula, which would lead to democratic elections within four years at the latest (Pretoria Agreement of 2002). Power sharing reached from the top executive level down to provincial administration (political power sharing),

The Consequences of Power Sharing at the Local Level

75

but also embraced the security apparatus (military power sharing) as well as public enterprises (economic power sharing). At the apex, Kabila retained the presidency, but was flanked by no less than four vice presidents, including the two main rebel leaders, Azarias Ruberwa (RCD) and Jean-Pierre Bemba (MLC). The ministerial cabinet, the National Assembly, the newly created Senate, and various public transitional institutions were also shared among the belligerents and, to a largely cosmetic degree, representatives of political parties and civil society. It is no exaggeration to say that the designers of the Pretoria Agreement equated peace with power, deploying a buy-in approach that sought to incentivize the insurgents to abandon their struggle (Tull and Mehler 2005). By mid-2003, the main insurgent leaders from the RCD and MLC had arrived in Kinshasa to take up positions as vice presidents of the republic while registering their armed groups as political parties. They were accompanied by insurgent officials who were to be appointed cabinet ministers, army generals, senators, and so forth. In principle, these developments signaled the territorial unification and pacification of the divided country. Power sharing in the DRC called for power within all institutions (provincial administration, parastatals, etc.) to be distributed among the former belligerents; for example, a governor would come from the Kabila camp, a first deputy from the RCD, and the second deputy from the MLC. In other words, power sharing was supposed to be practiced inside institutions, permitting power balancing and mutual control. Among the formerly controlled rebel territories, the strategically most significant areas were the provinces of North Kivu and South Kivu in eastern Congo, both of which had been fiefdoms of the RCD since 1998. Bringing the latter into the government was also meant to bring the Kivus back into the fold of the state. A key question therefore was whether the RCD would fully comply with the peace accord, meaning it would have to abandon or at least weaken its foothold in the Kivus. Participation in the unity government in Kinshasa had unintended consequences within the RCD insurgent movement, which in turn had an impact on the peace process. The power-sharing accord provoked a de facto, if undeclared, split of the movement. While a compliant group headed by RCD leader Ruberwa had left Goma to participate in the unity government, a second group decided to stay in their fiefdom of North Kivu (Tull 2004). The latter group was led by Laurent Nkunda, a senior military commander in the RCD, and Eugene Serufuli, an RCD founding member and, since 2001, North Kivu’s governor. While most parts of the hitherto RCD-ruled territory fell under the control of the unity government, the dissident group entrenched itself in South Kivu and North Kivu. Although it lost the former after heavy fighting one year into the unity government, it managed to retain control of the most important parts of North Kivu. A first

76

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

consequence for local power relations thus was that the Ruberwa-led group, now based in Kinshasa, largely lost its influence in North Kivu and eastern Congo at large. Paradoxically, the group’s entry into national politics doomed it to political demise. The group was in no position to play a twolevel power game. With the conclusion of the transition in 2006, the party ceased de facto to exist. The RCD’s internal split was rooted in diverging strategic assessments. The dissidents perceived the power-sharing transition as a losing option, certainly in the long term. Quite correctly, the dissidents had a bleak view of the RCD’s political future, especially in formal politics. They were acutely aware that the outcome of the postconflict elections as the concluding point of the power-sharing period was preordained to their disadvantage. Given the group’s unpopularity on its home turf in the Kivus, not to mention the national level, democratic elections amounted to a rope with which to hang itself. Enjoying the spoils of short-term power sharing (2003–2006) was an unappealing reward in view of the expected electoral outcome. Facing the prospect of losing the local political and economic supremacy that it had gained during the war, the RCD’s dissident faction saw little merit in complying with the power-sharing peace process (Stearns 2008). More specifically, the control over local power and resources was a major driver in the strategic decision to stay out of the unity government. These not only concerned the control over the area’s mineral resources, but also control over land and local markets. Influential local businesspeople were said to play a leading role in supporting the dissidents. These interests also tied in with the deep local social roots of the dissidents whose leaders were almost exclusively ethnic Tutsi or Hutu from the districts of Masisi and Rutshuru. This approach, it should be emphasized, was supported by neighboring Rwanda, the key ally of the RCD dissidents, which since the first insurgency in 1996 had relied on Congolese armed groups to protect its interests in eastern Congo. The protection of local and partly regional interests was diametrically opposed to the consequences of joining the national government, due to the design of the peace process, which called for a replication of power sharing at the local level. The Goma-based dissidents actively resisted the local implementation of the power-sharing agreement to protect their local power base in core parts of the province. For example, officials from the MLC and the Kabila camp who arrived in Goma to take up their positions alongside RCD cadres were effectively prevented from working. They were sidelined, intimidated, and sometimes killed.6 Many of them left the region in frustration. Consequently, the preponderance of primarily local interests directed the dissidents to maintain the status quo in terms of power relations, which worked as a disincentive to become co-opted into the central government. There was a second specific aspect whereby the local consequences of power sharing threatened the interests of the dissidents. As mentioned

The Consequences of Power Sharing at the Local Level

77

before, war and rebellion since 1998 had permitted the RCD’s Kivu-based Rwandophone elites (Hutu and Tutsi) to acquire significant control over local power and resources. This process had been accompanied by a concomitant marginalization of other ethnic, so-called autochthonous groups7 in Kivu, notably the demographically and economically most important group, the Nande. The resultant power shift away from the Nande and other groups was even more noteworthy in view of the fact that the Rwandophone communities, and the Tutsi in particular, had long been a politically marginalized group with a tenuous status at best. For example, local political competitors had time and again questioned their Congolese citizenship. Rwandophone dominance in the guise of the Rwanda-backed RCD rebellion was therefore fiercely contested by emerging ethnic autochthonous militias since 1998. In principle, and given the hatred of Rwandophone rule in North Kivu, the peace process, the unification of the country, and with it the expected weakening of the RCD authority over North Kivu were perceived by the Nande and other autochthonous groups as a chance to regain supremacy. Given North Kivu’s demography, the prospect of postconflict elections was a golden opportunity for autochthonous groups for the same reason that it was a threat to the Rwandophone rebel elite. Seeing the writing on the wall, the latter therefore all but refused to bring the power-sharing process to its logical conclusion, resulting in renewed violence, insecurity, and human rights violations, especially since late 2006 when the end of the transition was in sight. In summary, the case of North Kivu during (and after) the transitional power-sharing process in the DRC exemplifies an instance where wellanchored local elites found it more advantageous to protect their local fiefdoms than to join the national unity government. Under specific circumstances, the perception of a trade-off between local and national power is dominant. In the case of the dissidents of North Kivu, these circumstances included a lack of local and national political support, access to local resources, the structural weakness of the central state, and considerable backing from a powerful neighboring country. A long-term perspective of the rebels, which erased the potential merits of short-term power sharing, informed these considerations.

Liberia In contrast to Kenya and the DRC, Liberia’s top politicians did not overtly emphasize their local bases of support in the competition for national power. While claiming to be a descendant from the hinterland has become something of a ritual on the part of Liberian politicians seeking to distance themselves from the despised Americo-Liberian elite that governed the country

78

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

until 1980, the scene was not set for a two-level game. Some locally relevant actors were present at the peace negotiations ending the second civil war (1999–2003), among them many members of the Charles Taylor government. However, their presence at the negotiating table was not due to their local importance. In Ganta, a border town near Guinea in northern Liberia, the elected representative Nohn Kidau (2005–2011) was present at the negotiations as part of the civil society delegation, but never claimed to represent local grievances. Though Kidau came from Ganta and later used this as a beneficial claim of authenticity in her election campaign, she came to the peace talks as a representative of the diaspora, having been in the United States since her late teens when she fled the earlier conflict in Liberia. In places where Taylor and his political party, the National Patriotic Party (NPP), continued to have loyal followers, including the centrally located town of Gbarnga (which during the early 1990s had been Taylor’s alternative capital to Monrovia), there was a much stronger feeling of being represented at the peace talks. The presence of many pro-Taylor loyalists at the talks was enough for some of Taylor’s followers to feel that their issues had been included (at least when talking to local elites). George Mulbah (2005–2017), a representative from Gbarnga who was reelected in 2011 with record-high percentages, was not present at the talks, though he was a member of the interim government. This does not seem to matter, however. The national level had either completely merged with the local one (Gbarnga), or there was no expectation of local inclusion in the first place (Ganta). Such an inclination toward national-level politics for local power gains does not mean, however, that some more unusual arrangements were not taking place. It had been agreed in the negotiations that local government positions would not be reallocated according to a power-sharing formula, but left as they were (i.e., with personnel appointed by Taylor remaining in place) until elections, following which the new executive could (re)appoint them. The continued influence of former pro-Taylor militia commanders in Ganta was visible throughout this period. In 2003, Ganta’s first postwar mayor was appointed by the infamous rebel commander, Adolphus Dolo (better known under his nom de guerre General Peanut Butter), a Taylor loyalist. The mayor’s successor, appointed in 2008, was a former National Patriotic Front of Liberia (NPFL) fighter (one example of conversion from combatant to local power broker). The local setup did not provide for a counterweight from the Mandingo or the Liberians United for the Restoration of Democracy (LURD) faction, resulting in continued tensions on issues of landownership in the region. LURD was the political project of the marginalized Mandingo group that had long been integrated into Liberian society, at least to a degree. Nevertheless, at least from the 1980s onward, “a formal, ritualized relationship of reciprocity and mutuality has been replaced by an

The Consequences of Power Sharing at the Local Level

79

antagonistic relationship expressed in terms of exclusive, ethnic and religious identities” (Højbjerg 2010, 285). In contrast stands Gbarnga where informal efforts have been made to ensure Mandingo representation, including giving the position of assistant superintendent (of Bong County, where Gbarnga is situated) to a Mandingo in 2011. While there have also been incidents in Gbarnga between Mandingo and the locally dominant Kpelle group, this has not occurred to the same degree as in Ganta. LURD faction leaders took on roles in the interim government (and, where they could, beyond), leaving behind and ignoring not only all their mid-level commanders and foot soldiers but also worries and fears, whether justified or not, on the part of the larger community now associated with their rebellion. The faction leaders’ aim had been to gain national power; when this was achieved, the local level no longer mattered. In a locally hostile situation such as Ganta this can have specific repercussions, as antagonistic relations had increased throughout the war. Was this due to carelessness or rationality? Kenya’s political system, in comparison, has always provided for rather strong center-periphery links, making it inconceivable that such local power and conflict dynamics would be ignored. This has never been the case in Liberia, making such a result unsurprising. Even as the centralization of politics has begun to change in Liberia, this has not necessarily improved the importance of local minority elites. Generally speaking, Liberia’s decentralization reform has the potential to affect the distribution of power between the capital and the rest of the country. The disparity between the haves (a small elite based largely in Monrovia) and the have-nots (especially those based in the so-called hinterlands) and a long-standing emphasis on deep overcentralization has long been recognized as a source of conflict in Liberia (see Sawyer 2005, 115). On the surface, decentralization reform is widely proposed and is considered to be a necessary, at least rhetorically, reform, often with reference to the phrase “Monrovia is not Liberia.” Implementation plans have been extremely slow, however, suggesting a lack of political will at the center. Undoubtedly, decentralization would have significant impacts on local power constellations. As it stands, at the time of this writing the local governance structure (which is a local administration rather than a local government) is just beginning to function again, with fifteen local county governments (consisting of appointed district commissioners, superintendents, and development superintendents; and, as of 2013, also finance superintendents) and seventeen city corporations with appointed mayors. Both county and city governance remain highly dependent on the president, in whose office political decisions concerning local governance are still largely made. It is the executive who has the final say over the appointment of superintendents, mayors, district councilors, and the new finance superintendents. No local elections have been conducted since 1985 due to financial constraints in the postwar

80

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

period. Therefore, one of the major fears related to decentralization, expressed not only by Mandingos but also majority representatives, is that elections will favor majority groups and result in further sidelining minorities, which are currently able to secure at least some representation through a system of presidential appointment.8 One of the major changes envisaged is the election of superintendents and district commissioners. Despite risks of not being elected, current superintendents tend to see this as a positive development, as they hope it will improve their political legitimacy, which they seek to maintain in power struggles with the much stronger county legislators, namely the elected representatives and senators. Elected superintendents may have a much stronger role to play in the disbursement of development funds. And an elected county parliament will threaten the local importance of national legislators. Representatives and senators largely play the role of unopposed local patrons to this day as the nearly exclusive providers of government resources. Especially in the more rural counties, they make rare appearances, providing occasional personal financial contributions and public goods in a sporadic fashion, which in turn buys them loyalty at election time. It is perhaps no surprise that two important MPs in Gbarnga and Ganta were local businessmen who had put into place water pipes and electricity lines, respectively, prior to being elected in 2011. These legislators have no interest in changing a process that helped them to win elections. Decentralization will seriously damage their control of government resources or state institutions. Other locally significant actors, however, are throwing themselves behind the planned decentralization, including the chiefs who fear being increasingly sidelined by the national government in a process of the redrawing of chiefdom boundaries. Locally based actors such as superintendents and chiefs can only win from planned decentralization, while elected legislators sitting in Monrovia fear losing out in a zero-sum game in which their grip on power at the local level can only be lost. To date, the power game is far from being lost by “central” actors: the Local Government Bill is waiting to be discussed in the legislature.9 The Governance Commission finalized the draft bill in 2013 and has started to lobby individual senators and representatives to guarantee support for the bill once it reaches the House of Representatives and the Senate. But many government employees are not overly enthusiastic about the impending decentralization for fear of being posted outside of Monrovia. Many uncertainties and lingering issues remain (see Zanker 2014). Monrovia may not be Liberia, but for the time being it does not seem to be willing to give up its prime position. In Burundi, in contrast, the omnipresence of a transformed rebel movement benefiting from power-sharing arrangements has been entrenched locally. We provide details regarding this shift in power in the next section.

The Consequences of Power Sharing at the Local Level

81

Burundi Burundi differs from the three other cases in the geographic and socioeconomic distribution of identity groups. In this instance, national elites can be relatively cohesive and representative of large parts of the population, effectively equalizing national power with local power. During the civil war, the Hutu rebel group Conseil National Pour la Défense de la Démocratie–Forces pour la Défense de la Démocratie (CNDD-FDD) increasingly won popular support. It almost monopolized Hutu representation through two consecutive moves: first, it backed out of the peace negotiations leading to the Arusha Peace Agreement of 2000, exposing Front pour la Démocratie au Burundi (FRODEBU), the traditional Hutu party and origin of the CNDD-FDD, as too willing to compromise with the privileged Tutsi minority. Second, it signed a cease-fire (effectively ending the civil war) and security protocol just before the first postwar elections of 2005, thereby granting Hutu widespread representation and important positions in the army—which, during decades of Tutsi rule, was seen as the main obstacle to peace. It was partly disillusionment with the Arusha peace process that played into the hands of the CNDD-FDD (Curtis 2013, 86). Quickly transforming itself into a political party, the CNDDFDD established itself as the true representative of Hutu interests and the group responsible for the end of the civil war, which guaranteed its victory in the 2005 elections (Nindorera 2012, 27). This stands in contrast to the actions of its two main competitors, FRODEBU (as mentioned above) and the long-standing Hutu rebellion of the Forces Nationales de Libération (FNL), which failed to join the institutions before losing more and more of its legitimacy and strength. When the FNL finally laid down their arms, the group did so on the basis of terms established almost entirely by the new government headed by the CNDD-FDD. Since the CNDD-FDD fighters were part of the Hutu majority, they benefited from the quota system agreed on in the Arusha Peace Agreement, which stipulated that 50 percent of the security institutions and 60 percent of the political posts in government institutions had to be allocated to Hutu. In fact, for the first time in history, Hutu were massively integrated in state structures down to the local level. Because the CNDD-FDD established itself as the major representative of Hutu interests soon after they transformed into a political party in 2004, they gained a lot of ground in many parts of the country compared to the period before and during the war. Furthermore, just before the elections of 2005 when the CNDD-FDD’s victory was foreseeable, quite a few MPs of other parties, including FRODEBU and Union pour le Progrès National (UPRONA), changed sides and became CNDD-FDD members.10 This dynamic continued and the following years saw the increasing dominance of the ruling party: from the national to the

82

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

most localized level, more and more people joined the CNDD-FDD for opportunistic reasons or simply to “secure their job and life.”11 As Hutu and Tutsi populations are not geographically confined, the increasing dominance of the CNDD-FDD unfolded nationwide. This does not mean that historically there are no differences between provinces. The southern province of Bururi should be mentioned as historically Tutsi dominated, just as a comparatively large number of rebels of the CNDD-FDD and the FNL came from the western provinces of Bujumbura Rural, Bubanza, and Citiboke. To this day, the west is more affected by continued political violence than other parts of the country. The CNDD-FDD leadership is, of course, sensitive to the regional background of those members in high positions. However, the main fault lines during the war, as well as afterward, were and are ethnic and about political affiliation, not about regionalism.12 Decentralization therefore has a different connotation in Burundi than, say, in Kenya. Decentralization had been discussed since the 1990s, but was implemented only after the transition period (see Senate of Burundi 2010). The most important change was the democratization of local government; that is, the introduction of elected councils at the level of the commune and the hill (subcommunal level). In both postconflict elections of 2005 and 2010, the CNDD-FDD won the vast majority of seats in communal councils in most parts of the country. Additionally, almost all opposition parties denounced the 2010 communal elections (in which the CNDD-FDD won over 60 percent of the seats) as fraudulent and, afterward, told their elected council members to abstain from meetings. The CNDD-FDD thus gained control even over communes in which other parties were initially strong. As has been argued elsewhere, the change in power relations between members of the CNDD-FDD and the old regime became particularly visible in struggles for local authority at the level of the smallest administrative entity (the hill) and between a completely new political institution, an elected local government (conseils collinaires), and the traditional institution of Bashingantahe (Hirblinger and Simons 2014). The latter two institutions are tasked with local development and conflict resolution. Due to the historical links of the Bashingantahe to the UPRONA regime, the CNDD-FDD considered the traditional institution a potential threat to its own power in rural communities and, therefore, discredited the Bashingantahe while simultaneously promoting the conseils collinaires. The sidelining of the Bashingantahe by the CNDD-FDD is manifest in legal changes (their special status was erased from the communal law in 2010) and in the everyday conduct of local politicians.13 Thus, the new ruling party also benefited from the decentralization of government. In conclusion, as opposed to the other cases laid out in this chapter, power sharing had a power-consolidating and power-enforcing effect on the CNDD-FDD, not least felt on a local level where the rebel organization

The Consequences of Power Sharing at the Local Level

83

quickly transformed itself into a dominant party with single party habits, tightening its grip on power. Local power was not considered more important than national power, as in the case of the DRC, nor was local power of little importance to national elites, as in the case of Liberia. In fact, in the small state of Burundi, which has historically strong ties between the central government and the periphery and where conflicting groups are not geographically confined but rather evenly spread across the national territory, national power largely equals local power.

Conclusion Drawing on four African cases, we showed that power sharing has strong effects on the dispensation of power at the local level, though in distinct ways. The effect is dependent on properties of the arrangement, but also on the local composition of groups and local political cultures. The first question we asked at the outset of the chapter—whether national elite pacts have an impact on power relations at the local level—can be answered affirmatively. However, it is less clear whether this effect necessarily outweighs the influence of established ties between center and periphery. We consider the latter factor a core context condition that influences the scope of the impact that power-sharing arrangements have locally. The intense exchange between provincial capitals and Nairobi in the case of Kenya, the neglect of the hinterland in Liberia, the omnipresence of the state in Burundi down to the local level, and the remoteness and relative disconnect of the Kivus from Kinshasa in the DRC were tangible facts before and after power-sharing measures were implemented. These features obviously factored into the power calculations of both national and local elites during the implementation of the accords. Surely, properties of the wars (and of political violence in the case of Kenya) played important roles as well. The fact that LURD in Liberia was not able to control conquered territory over time (in contrast to the RCD in the DRC) may explain the relative failure of the former’s ethnic constituency (Mandingo) to win in terms of local power after the war and the latter’s interest in upholding a military dominance in their zone of control, regardless of the letter of power-sharing agreements. The spatial segregation of ethnic groups as a result of electoral violence seems to play into the hands of local-national ethnic entrepreneurs in the case of Kenya, just as much as the dominant interpretation of political power sharing as a game where players need a firm regional base. The inclusiveness of national power-sharing governments is not a guarantee for altering power relations at the local level. We found that inclusion at the national level represented a danger for the local political survival of RCD leaders while it boosted the already strong position of local strongmen in Kenya. Contrary to the Kenyan experience, Liberian negotiators standing

84

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

for their home region did not gain power locally from being included in the power-sharing agreement. Again, context seems to matter strongly. We also found that power sharing impacts the cohesiveness of signatory parties. The prospect of profiting from power sharing further solidified Burundi’s CNDD-FDD at the local level, fostering its attractiveness to local strongmen and motivating politicians from other parties to change to the CNDD-FDD. The prospect of losing power on a local level due to such agreements, however, enhanced fragmentation in the case of the RCD in the DRC. Power sharing may have been an accelerator of such processes. Table 4.1 summarizes the diverging effects of power sharing in our cases. Political and military power sharing have immediate effects on the local level, but the longer-term effects of territorial power sharing, including decentralization reforms, may be stronger. In all four countries, some sort of decentralization was foreseen and rather strongly affected the strategies of local and national actors. Fears that decentralization could exacerbate tensions led to the use of conflict prevention strategies in a place like Nakuru, Kenya, whereas hopes that decentralization would legitimize county superintendents in Liberia met with dilatory tactics from national legislators and the Liberian government. Detailed regulations and the specification of an ethnic quota down to a local level profited the new dominant party in Burundi, leading to a further decline of traditional leaders who were perceived as being associated with the old regime. As the four cases make clear, power-sharing arrangements and their proximate effects create winners and losers on both the national and local levels. Wartime chiefs versus civilian politicians, appointed versus elected governors, traditional versus modern authorities—all constitute examples of the juxtaposition one gets with any sort of political transition. However, we found that two moments matter strongly: the moment at which a national deal is signed, and the date of the implementation of more proximate reforms pertaining to the local level. The former can create incentives for cooperation across previously inimical camps (as was the case in Kenya), thus resulting in a positive-sum game, whereas the latter inspires defensive or offensive strategies to maintain or gain power in a perceived zero-sum game. Stepping out of our changing power relations framework, we note that the effect of power sharing on the quality of (local) peace is not linear, but depends on context and on preestablished power relations and changes over time. Some emancipatory effects of territorial power sharing—namely, electing local administrators and legislators—may prove beneficial to peace in the long run, but could also put shaky power balances under severe stress in the short term. In this chapter, we sought to explore some of the undetected local mechanisms at play when national elites sign power-sharing pacts, focusing in particular on the level where violent conflicts once escalated. This perspective, we argue, has real implications for peace.

Table 4.1 Effects of Power-Sharing Properties Down to the Local Level

De Facto Local Power Sharing

Inclusiveness (of Power Sharing at the National Level; Local Effects)

Cohesiveness (of Conflict Parties as a Consequence of Power Sharing)

Temporality (Time Period of Power Sharing and Other Postconflict Reforms)

Devolution as result of related constitutional reform

Replication of national power-sharing deal in Nakurua

High, accentuating majority-minority confrontations locally

High, profitable to local-national power brokers

Interim (2008–2013), but ongoing effects due to devolution implementation

Liberia

None planned

Variable/weak

High, with little effects locally

Uneven

Interim (2003–2005), but potential effects with decentralization policy

Democratic Republic of Congo (DRC)

Detailed prescriptions on filling positions in provinces

Implementation subverted, instead local monopolies of power

High, accentuating majority-minority confrontations locally

Low, power sharing leading to organizational splits of main rebel organization

Interim period (2003–2006), plans for decentralization reversed

Burundi

Detailed prescriptions on filling positions in the entire politicoadministrative system

Implementation favoring one conflict party

High, reducing majority-minority confrontations locally

High, power sharing leading to organizational consolidation

Interim period (2000–2005), permanent with new constitution (since 2005)

Power Sharing Playing Out at the Local Level (Local Content) Kenya

Note: a. See Zanker, Simons, and Mehler (2015).

85

86

A. Mehler, C. Simons, D. M. Tull, and F. Zanker

Notes 1. This is linked to Lijphart’s (1977) category of “group autonomy” as part of his work on consociational democracy. 2. In our cases, local religious elites played a subordinate role; they may, however, show similarities to traditional leaders. 3. The ODM Negotiation Team included Musalia Mudavadi (MP in Western Province), William Ruto (MP in Rift Valley Province), Sally Kosgei (MP in Rift Valley Province), and James Orengo (MP in Nyanza Province). 4. In 2010, Ruto was suspended, reinstated in a lower position, and then suspended again, and in 2011, he was formally fired. 5. Margaret Kamar, interviewed by Zanker, Eldoret, November 20, 2011. 6. Interviews by Simon of a researcher and a journalist, Goma, August 25, 2011. 7. Note that the term autochthonous is highly contested within the area. It is often used as a means to discredit Rwandophones, questioning their rights as citizens and discursively constructing them as outsiders even though parts of the Rwandophone population(s) have been living in the area for more than a century. 8. Mandingo Chairman, interviewed by Zanker, Ganta, August 22, 2013; city mayor, interviewed by Zanker, Ganta, January 27, 2014. However, evidence of such appointed representation is lacking (see also Neumann and Winckler 2013, 672). 9. The bill was further put on hold because of the Ebola epidemic that greatly affected Liberia and its neighboring countries in 2014–2015, enforcing a state of emergency in Liberia between August and November 2014. 10. FRODEBU MP, interviewed by Simons, Bujumbura, November 10, 2011. 11. Former CNDD-FDD administrator, interviewed by Simons, Bubanza, December 5, 2011. 12. Note, however, that during almost thirty years of Tutsi (military) governance in Burundi, a small faction of Tutsi from Bururi dominated politics and the military. Nevertheless, as in most so-called ethnic conflicts a small elite on each side of the divide was able to mobilize on general ethnic grounds, effectively obfuscating stratifications within each ethnic group. 13. Bashingantahe, interviewed by Simons, Gitega, January 22, 2014.

PART 2 Power-Sharing Mechanisms at Work

5 Military Power Sharing: The Case of the Philippine Peace Agreement Rosalie Arcala Hall

In 1996, the Philippine government and the Moro National Liberation Front (MNLF), a secessionist group in Mindanao, signed the landmark Final Peace Agreement (FPA) ending a two-decades-long conflict that had seen widespread internal displacement and economic dislocation in an area populated by Muslims, Christians, and indigenous peoples. The agreement was widely seen as a fulfillment of the bargains made under the 1976 Tripoli Agreement, which the previous Ferdinand Marcos dictatorship had not carried out. The FPA provided for political power sharing, with transition arrangements under the Southern Philippines Zone of Peace and Development (SPZPD) to be administered by the Southern Philippines Council for Peace and Development whose membership was shared fifty-fifty by the government and the MNLF. The FPA included more robust autonomy and fiscal resource injections into the territorially expanded Autonomous Region for Muslim Mindanao (ARMM), which, it was expected, would be electorally dominated by the MNLF. The agreement also contained a proviso for military power sharing through the individual-based absorption of 7,500 MNLF ex-combatants and their proxies into mixed army and police units largely deployed in Mindanao. The Philippine FPA was one of the twenty-two cases (out of forty-one) of negotiated conflict termination agreements identified by Caroline A. Hartzell and Matthew Hoddie (2003) for the period 1945–1998 that contains a military power-sharing element. It was also one of the few cases of military integration that was largely endogenous and involved little or no external support by the UN or by a third party (Licklider 2014). Unlike other cases where the absorption of ex-rebels into the security forces was bundled with disarmament and demobilization processes, the FPA curiously lacked this element. The FPA provided for a social development program covering the more numerous MNLF ex-combatants not integrated into the 89

90

Rosalie Arcala Hall

security forces, but this process was carried out separately and largely through external donor funding. By contrast, the military power-sharing project was handled exclusively by the Philippine military and the MNLF leadership. Largely touted as a success by virtue of having met the quantitative targets and time lines stated in the FPA, the military power-sharing arrangement benefited the Philippine army’s internal security campaign in Mindanao and its credentials as a culturally inclusive institution (Hall 2014, 114). As an employment vehicle for MNLF ex-combatants and proxies, the integration also shored up the MNLF’s legitimacy vis-à-vis its own members and supporters. In this chapter, I examine how military power-sharing arrangements contained in the 1996 FPA transformed the balance of power between the military, as agent of the state, and the MNLF, as well as between the Philippine state and the Muslim Mindanao community. The military power-sharing agreement entailed the absorption of MNLF combatants and proxies into the army, based on processes (selection, recruitment, training, placement, and deployment) jointly defined by the MNLF and military leadership. Because the arrangement had direct consequences for each group’s membership or personnel and capabilities, I explored the manner in which the integration of ex-MNLF rebels into the army affected the groups’ respective interests. The Bangsamoro area (which comprises the provinces of Maguindanao, Lanao del Norte, Lanao del Sur, Basilan, Sulu, and Tawi-Tawi) has a complex security landscape made up of other formal and informal actors. The Philippine military, whose numerous ground forces and assets have long been invested in the Mindanao war, is a ubiquitous stand-in agent of the state in erstwhile hostile Muslim areas. The police, whose force is locally recruited but nationally administered, relatively speaking is a lower-key state security force hampered by the strong pull of clan loyalties and control by mayors. The largely ethnic Tausug-dominated MNLF competes for representation of the Bangsamoro armed struggle with the splinter group Moro Islamic Liberation Front (MILF), which has consolidated more field commands in the Maguindanao and Maranao areas of central Mindanao since 1977. Powerful local chief executives—bosses or political warlords with their own coterie of heavily armed private security providers—wield control over their domains using their own brand of patronage and intimidation. Allied with the national government (in the sense that they deliver votes to national candidates in exchange for access to resources), these politician-attached private armed groups (PAGs) are tolerated as an organic expression of the Muslim masculine penchant for guns. Within this landscape are also loose groupings of heavily armed men—village militias, kidnap-for-ransom (KFR) gangs— brought into pragmatic networks of convenience with the army, police, and mayors for political and economic gain.

Military Power Sharing

91

The military integration project also has perceptible consequences for the balance of power between the local army units and other armed groups. It is my contention in this chapter that military integration furthered the army’s reach into Muslim society, thereby gaining greater acceptance for it as a proxy government institution in areas where formal civilian authority was weak or lacking. Through its civil-military operations activities, conducted by culturally savvy integree-recruits, the army was able to penetrate and build social capital among the Muslim population. By producing better and more stable employment for Muslim men, the military provided an alternative patronage network to illicit groups and private armies. Local elites were not threatened by the military integration because the army’s increased capability was not directed at dismantling their own private security apparatus; rather, the parallel reorganization and personnel infusion from the police integration expanded their local control. Even without accompanying provisions for MNLF disarmament or demobilization, the power-sharing arrangements improved the army’s internal security performance, enhanced its relationship with Muslim communities in Mindanao, and improved its public image as an inclusive institution to Muslim minorities. State control over the security forces located in the autonomous area was maintained simply by renaming the regional police force as the Special Regional Security Force (SRSF) provided in the agreement. However, there were marginal returns to local police performance from the integration project. The state made gains through the security sector, but the agreement failed to produce parallel improvements in civilian governance.

Power-Sharing Arrangements and Outcomes A power-sharing arrangement is one of the pathways to civil war termination (apart from partition and military victory). It pertains to negotiated institutional designs that “share, divide or balance power among competing groups” and can be political, economic, territorial, or military in nature (Hoddie and Hartzell 2005, 29). David Lake and Donald Rothchild (1996, 58) see power-sharing arrangements in ethnic conflict–ridden societies as those with “representative ruling coalitions in the cabinet, civil service, military and high party” that cover both formal and informal arrangements. Power-sharing arrangements, whether at the negotiation, agreement, or implementation stage, are argued to contribute to peace by allowing parties to convey credible commitment through their willingness to endure costs (Hoddie and Hartzell 2005). Lake and Rothchild (1996, 59) also see powersharing arrangements, when bundled with other institutional features (e.g.,

92

Rosalie Arcala Hall

regional autonomy and federalism), as leading to improved management of ethnic conflict by providing security guarantees to minority groups. Military power sharing refers to the absorption, merger, or integration of rival armed forces (state and rebel) based on an agreed formula of representation. This type of power sharing comes in two forms: (1) amalgamation of previously opposed military forces into a new state security force; and (2) merger of members of nonstate armed groups into an existing state security force. In both cases, representation can be measured in terms of proportional personnel strength, in the number of leadership positions, and in whether antagonists are in separate or mixed units within the force. Sven Gunnar Simonsen (2007, 585) sees military power sharing as reducing the armed group’s numerical strength as well as its ability to mount renewed threats against the government. In their examination of sixteen peace agreements with military power sharing from 1980 to 1986, Matthew Hoddie and Caroline A. Hartzell (2003) found that implementation increases the prospects of peace. Peace was sustained in seven out of the eight completed cases and even half of the failed/partially completed pool. Katherine Glassmyer and Nicholas Sambanis (2008) doubt that military power sharing has a unilateral effect on peace building. In their study of thirty-four cases (as part of a negotiated conflict settlement, both strict and lenient) of military power sharing since 1945, they note that military integration was driven by economic incentives. They conclude that military integration is better at promoting peace when it is pursued in conjunction with political powersharing arrangements and with international or third-party support to build economic capacities. Military Power Sharing Under the 1996 Final Peace Agreement: Gains and Deficits

Integration was part of the agenda of the four formal rounds of MNLFgovernment negotiations. Men of strong military credentials (former military chiefs Feliciano Gacis and Eduardo Ermita) who understood the security forces’ institutional viewpoint were part of the government negotiating panel. The final number of individuals to be integrated into the state security forces (5,750 for the army and 1,750 for the police) was arrived at with a great deal of difficulty and involved serious haggling between the MNLF’s high demand of 15,000 and the military’s proposed number, which was based on the proportion of Muslims in relation to the entire population (Rodil 2000, 13; Feliciano Gacis, personal communication, March 18, 2010). At that time, the military estimate of MNLF strength was between 7,500 and 10,000 individuals, while the Philippine armed forces’ strength stood at 110,000. Disarmament and demobilization were removed as topics of negotiation as these were unacceptable to MNLF leader Nur Misuari (Rodil 2000).

Military Power Sharing

93

The military power-sharing components of the 1996 Final Peace Agreement are contained in Sections 19 and 20a to 20f, which call for: (1) the integration of 1,750 MNLF elements into the Philippine National Police (PNP) and 5,750 MNLF members into the Armed Forces of the Philippines (AFP); (2) utmost government effort to create the necessary conditions for the eventual integration of the maximum number of the remaining MNLF forces into the SRSF and other government agencies and instrumentalities; and (3) a special government socioeconomic cultural program for MNLF forces not absorbed into the state security forces. The second and third elements were stated in promissory terms and without a set quantitative target. Other mechanisms provided for in the agreement include: (1) a waiver of requirements and qualifications for entry of MNLF forces into the AFP; (2) ranks and grades of MNLF forces joining the AFP to be decided on by the president as commander in chief; (3) an unspecified transition period during which the MNLF integrees are grouped into distinct units deployed in the autonomous area; and (4) the appointment of a deputy commander for the Southern Command from among the MNLF forces tasked with assisting in the command, administration, and control of the separate units during the transition period from 1996 to 1999.1 Sections 73 to 93 of the 1996 Final Peace Agreement enumerated in great detail the establishment and mechanisms for the SRSF, including provisions for the head of the autonomous government to be ex-officio chairman of the Regional Police Commission (RPC) with powers to exercise operational control, supervision, and disciplinary powers; and to employ, deploy, assign, and reassign officers and personnel through an appointed regional police director. The agreement also provided for the Regional Legislative Assembly to enact laws governing the regional police command. However, the ensuing 2000 Organic Act for the Autonomous Region of Muslim Mindanao passed by the Philippine Congress diluted much of this purported local civilian control mechanism over what the MNLF expected to be a separate and more numerous SRSF. The law treated the Philippine National Police units stationed in the autonomous area (the ARMM Police) with the previously absorbed MNLF integrees as the SRSF that it was obligated to create under the FPA. Since the regional police force is comprised of local police units stationed in various municipalities and cities covered by the ARMM, control over their deployment was de facto by the local chief executives. Operationally speaking, the FPA thus did not alter the civilian control infrastructure for the police. The integration processes in the police and the military followed separate streams, but both allowed the MNLF full control over identification of potential recruits. The MNLF distributed police and integration quotas and slots to its various commands. Both the military and the police accepted MNLF combatants and proxies. Unlike the military, which opened a special

94

Rosalie Arcala Hall

recruitment window for MNLF integration, the police had theirs folded into the regular recruitment cycle for units in the ARMM. The military required prospective integree-recruits to surrender a firearm, but the police did not. The police integrees were deployed locally where they were recruited while army integrees were individually absorbed by infantry units based in Mindanao, save for a six-month interim period during which they were organized as separate rifle and engineering companies. By all accounts, the military and policy integration efforts experienced minimal attrition rates. Soliman Santos Jr. (2009) mentions approximately fifty recruits dropping out during the police integration process, although the Office of the Presidential Assistant on the Peace Process (OPAPP 2007) reported an attrition rate of less than 10 percent for the army integrees. Military integration proceeded without an enabling law and was carried out under Department of National Defense Order 295, which set a threeyear (1996–1999) time frame. A Joint Force Integration Board (JFIB), composed of equal numbers of AFP and MNLF members, decided on details such as: (1) a waiver of normal military entry requirements; (2) a list of candidates plus alternates, combatants, and their proxies, for initial screening by the AFP; (3) surrender of a weapon as a requirement for processing; and (4) the interim formation of separate rifle and engineering companies composed exclusively of integrees. The JFIB agreed on another recruitment for attrited cases in 2009. Because the integrees were individually absorbed into mixed units of the regular army infantry, the Philippine military’s organizational logic understood the ARMM Unified Command (created in 2003 under Executive Order 212) as falling under the direct supervision and control of the AFP Southern Command.2 Consequently, troop disposition and deployments (which are by unit), as well as personnel officers and assignments, follow the same principle of command and control. Many of the mixed integree units were under the operational control of the First, Fourth, Sixth, and Tenth Infantry Divisions whose area of operations comprise the whole of Mindanao. Ultimately, however, it is expected that these units can and will be redeployed to nonARMM areas in Mindanao and other parts of the Philippines as needs arise. As such, there was no change in the civilian control complements in place prior to the 1996 Final Peace Agreement. The national government through the Department of National Defense neither shared nor relinquished part of its control over the armed forces located inside the autonomous region. Ante-Agreement Status: Informal Power Constellations Inside the Bangsamoro Conflict Theater

To understand how the balance of power was altered by the military powersharing terms of the 1996 Final Peace Agreement, it is important to map the

Military Power Sharing

95

political ecology of the Bangsamoro area prior to the agreement. Populated by native indigenous tribes, Muslims, and Christian settlers, Mindanao was under the firm control of local elites. Religion (Islam) and ethnicity (e.g., Tausug, Maranao, Maguindanao, or Yakan) are the traditional leaders’ political assets, supplemented by patronage, public works funds, and control over the police. These traditional politicians, descended from families of the former sultan or who have emerged as datus, or chiefs, are tied to the state in a clientilistic fashion—brokering deals with their local population and delivering block votes for national candidates (Abinales 2004, 8). During the Marcos dictatorship, traditional leaders were superseded by local strongmen types or bosses whose focus on extracting the natural resources of their areas parallel those of the predatory state (Sidel 2000). These warlords (e.g., Ali Dimaporo of Lanao and Andal Ampatuan Sr. of Maguindanao) rule under the proverbial guns, goons, and gold formula; their control over public resources (including the local police) are matched by the strength of their private armies. Along with the warlords, the military was Marcos’s institutional ally, systematically suppressing insurgents and dissidents alike and propping up the dictatorship for over two decades. The armed secessionist movement under the MNLF was antithetical to the Muslim traditional leaders and political strongmen whose fortunes were tied with the central government. The MNLF capitalized on Muslim marginalization and oppression by the Christian-dominated central government apparatus to press for a separate Bangsamoro state. At the height of intensive fighting between Philippine government troops and the MNLF in the early 1970s, MNLF strength peaked at 15,000. The campaign also produced parallel communal conflicts in many mixed-population Mindanao communities, spawning civilian militias. In 1976, the MNLF negotiated and signed the Tripoli Agreement with the Marcos government, which the dictatorship failed to implement. Tensions within the MNLF leadership, particularly between the Maguindanao and Maranao field commanders versus the mainly Tausug front liners, led to the splitting off of the Moro Islamic Liberation Front in 1977. Other commanders (e.g., Magic 5) were also persuaded to leave the MNLF by the Marcos government’s lucrative offers of positions and perks (Abinales 2004, 10). As the MNLF wound down, the MILF Bangsamoro Islamic Armed Forces (BIAF) grew in strength to 15,000 by the end of the 1990s. It has consolidated control over communities in Maguindanao and Lanao Provinces, even forming self-contained shadow governments in areas it controls. The links between the Moro National Liberation Front and the local elites were not necessarily antagonistic. For instance, G. Carter Bentley (1994, 254) argues that the Alonto and Lucman families in Lanao Province, rival to Marcos’s ally Dimaporo, had strong links with the MNLF, with one family member having been one of the key founders of the MNLF. He,

96

Rosalie Arcala Hall

therefore, read the MNLF’s opposition to Dimaporo as resident warlord-state proxy as stemming from being mapped into the competition between the rival clans for dominance in Lanao politics. In the years running up to the 1986 snap elections between President Marcos and Corazon Aquino, MNLF leader Misuari instructed Lanao commanders to support her (Bentley 1991, 245). Although formal negotiations between the Aquino government and the MNLF collapsed by May 1987, President Aquino’s peace overtures toward the group (as opposed to rival Moro Islamic Liberation Front) solidified the MNLF’s claim of being the legitimate representative of the Bangsamoro people. The MILF effort to mobilize (à la People’s Power) in Cotabato failed to gain traction vis-à-vis the new set of national politicians, who were more familiar with MNLF leader Misuari (Bentley 1991, 245). This recognition from the central government was an important symbolic gain for the MNLF, which had lost much support on the ground to the MILF. The Philippine police historically has been under the control of local chief executives. Under martial law in the 1970s, Marcos tried to undercut mayors and governors by placing all police under one roof—the Integrated National Police (INP), which was merged with the Philippine constabulary and transferred under the effective control of the armed forces (Hernandez 1988). Reforms carried out in 1990 yielded the eventual dissolution of the Philippine constabulary and the remaking of the Philippine National Police into a civilian outfit divorced from the armed forces (operationally under the control of the Interior and Local Government Department) and national in scope. The latter proviso elevates decisions on appointment and troop deployment to the National Police Commission and is intended as a safeguard against potential abuse by local chief executives. The PNP shared the task of internal security operations with the military and, briefly from 1991 to 1998, was assigned the lead role. Correspondingly, it maintained Special Forces (later renamed Public Safety Battalions) at the regional and provincial levels for internal security operations (ISO). Under Republic Act 8551 (1995), the police were relieved of this lead role in internal security operations in favor of the military. This unusual tasking yielded a police force that is complacent and deferential to the military when it comes to the job of internal security (Santiago-Oreta and Tolosa 2012). By contrast, the military has an overwhelming presence in the Bangsamoro landscape. The Philippine military is organizationally configured for internal security operations with ground forces (army and marines and, until 1991, the constabulary) comprising the bulk of its regular troops plus a paramilitary arm, called the Civilian Home Defense Forces (CHDF) under the Marcos dictatorship but from 1988 onward the Citizen Armed Geographic Unit (CAFGU) Active Auxiliary (CAA). These paramilitaries serve as territorial defense units and as force augmentation in the military’s localized operations.3 Since the 1970s, the army’s campaign strategy against the Islamic

Military Power Sharing

97

separatists has featured heavy reliance on conventional warfare with ground troops supported by air strikes and naval bombardments to neutralize rebel strongholds from one town to the next. In these contested spaces, the military maintained numerous checkpoints, detachments, and garrisons. Some civil-military operations (CMO) were included, such as the dropping of leaflets in rebel strongholds to warn the population of an approaching artillery bombardment and to induce the rebels to relinquish hold over the town, but the overall campaign was marred by poor conduct by ground troops, human rights violations, and large-scale displacements (Russell 2013, 111). Even with the start of negotiations leading to the Tripoli Agreement between the Philippine government and the MNLF in 1976, the military was locked into using overwhelming firepower in its campaign. Throughout the martial law era (1972–1985), it was not inaccurate to describe the military as enjoying unprecedented political power given its sheer presence and the relative dysfunctionality of the civilian government apparatus throughout the conflict zone (Abinales 2000; Casper 1995). This physical presence and tacit political influence did not wane with the formation of the Autonomous Region for Muslim Mindanao in 1986. The relationship between the local bosses, the police, and the military was not framed by any measure of civilian control but by pragmatic cooperation. Clan and ethnic ties underscore dynamics between various local armed actors in these local settings. As local police forces require nomination by the local chief executive, operationally the police comprise troops who are relatives of the mayors and function more as their personal security force. In the contested Bangsamoro areas, the local police are deemed unreliable public security providers amidst the impunity of shadow economies (drug dealing, cattle rustling, smuggling, and kidnapping) (see McCoy 2009; Lara and Schoofs 2013). The dysfunctional nature of the local police is linked to the broader issue of the absence or weakness of formal government structures at the local level (apart from warlord personalist-style politics) for most parts of the Muslim-dominant areas (Hall 2013). The collocated military units, staffed by officers and enlisted personnel who are nonlocals, tolerated the local bosses’ private armies while the local bosses considered the military’s presence in their area as good insurance against MNLF attacks. A good example of this pragmatic engagement is Lanao under strongman Dimaporo. Dimaporo was backed by the military, which allowed him to mobilize and deploy his private army at will; at the same time, Dimaporo cleared the military to carry out its operations against the MNLF however the latter saw fit (Bentley 1994, 254). Where local bosses were absent and in MNLF strongholds, the army and constabulary commanders functioned as the de facto government (Casper 1995). Many private justice and security providers operate alongside the local bosses’ private armies, the military, and the police. These grassroots

98

Rosalie Arcala Hall

security-self-provisioning groups are characteristic of Mindanao: variants include village-based militia, religious cults, MNLF “lost commands,” death squads, and KFR gangs. Many of these groups have informal ties with the military and have been intermittently used for the latter’s counterinsurgency operations. After 1986, several of these groups were accused of human rights violations, prompting then president Corazon Aquino to issue an executive order that placed many of them under the ambit of army control (and, consequently, distinguished them from “private armed groups” that had been declared illegal under the 1987 Constitution). Renamed the Citizen Armed Geographic Unit Active Auxiliary, these armed local volunteers, which can be used in internal security operations, are supposedly subject to human rights screening and supervision by the Philippine army. The government also allowed special CAAs—private security groups, organized by agricultural and mining companies, with army training and supervision. Kidnap-for-ransom gangs, whose formation is facilitated by arms proliferation and an abundance of young unemployed men, constitute part of Mindanao’s “dark networks”—groups with no fixed membership and no chain of command that coalesce or disband depending on opportunity. Referring to the KFR gangs operating in Cotabato and Lanao during the 1980s, Eric Gutierrez (2013, 141) considers them “corporate ventures between politicians, policemen, military” that pocket the proceeds from ransom payments. MNLF and military “lost commands”—local commanders and their loyal men that the MNLF or armed forces leadership could neither control nor discipline—also comprise some of these KFR networks as subcontractors for which hostages are either sold or bought at a markup. Gutierrez (2013) alleges that local politicians and MNLF leaders are not entirely disconnected from these shadow groups. Post-1996, the links between these KFR gangs and the MILF as well as with the Abu Sayyaf Group (ASG) were given as a rationale for military offensives against these groups’ known lairs (Yegar 2002, 344–345). Military Power Sharing and the Remaking of the Security Landscape in the Autonomous Region

At the time of the Final Peace Agreement in 1996, the state security forces and the MNLF were of varying strength and importance given then emerging threats. Military deployment in the Bangsamoro zone increased further with sustained campaigns against the MILF and the appearance of a new threat from the Abu Sayyaf Group.4 The police, which had assumed primary responsibility over internal security operations from 1991 onward, had half of its strength (20,000 out of 40,000) devoted to the ASG threat and kidnapfor-ransom activities, which spiked during this period (Philippine Congress,

Military Power Sharing

99

Senate Finance Committee 1996, 34). The military was also assisting the police in responding to the ASG and other law enforcement challenges (de Villa 1996, 98). The MNLF’s symbolic gain from having entered into negotiations with the Corazon Aquino and Fidel Ramos administrations shored up its political capital vis-à-vis Muslims in Mindanao. In terms of armed strength and popular support, it was now confined to the Tausug population in the island provinces (Sulu, Basilan, and Tawi-Tawi), having lost much ground to the MILF that seized effective control over the Maguindanao and Lanao areas of central Mindanao. In the new democratic political order after 1986, the local elites with their private armies remained in place, despite calls by the central government to disband such illegal armed groups. Various civilian militia were reconstituted into paramilitaries under the army’s control, bringing at least a portion of the loose firearms through the Bangsamoro area within the ambit of military supervision. The integration project increased the army’s size by 4 percent and constituted the single largest intake of Muslims into the predominantly Christian army. The majority of the integrees were ethnic Tausug, paralleling the ethnic composition of the MNLF’s leadership and membership profile at that time. More importantly, a substantial number of the integrees were young and educated proxies. Although the MNLF independently determined who got on the list, its internal recruitment process, whereby political leaders and commanders themselves determined the distribution of slots, resulted in the inclusion of the younger male kin of combatants. In retrospect, this process worked in favor of the military which, despite having agreed to waive education, age, and height requirements as part of the integration project, nevertheless netted positive gains in its personnel pool from these young proxies.5 For the army, the integration project was a triple boost. Institutionally, it was able to acquire additional personnel with the right credentials (young and Muslim) to carry out effective civil-military operations. Previous assessments have lauded integrees’ effectiveness in civil-military operations in Muslim-populated communities (Depayso 2004, as cited in Santos 2009; Lidasan 2006; Jacildo 2003). Mindanao units with integree elements were found to be especially good at Muslim community dialogues, with integrees acting as cultural bridges and emissaries. Russell (2013, 167) argues that the Philippine military was able to fine-tune its civil-military operations in the Muslim theater in 2000 through the Special Advocacy on Literacy/ Livelihood and Advancement for Muslims (SALAAM) units in large part because it had a sufficient number of Muslim soldiers. While the SALAAM program did not come into full force until 2004, its combined elements of community immersion, problem identification, and socioeconomic projects produced more positive attitudes toward soldiers on the part of many Muslim communities. The reliability of integree-soldiers has been tested in the

100

Rosalie Arcala Hall

subsequent military campaigns against the MILF (e.g., 2000 all-out war, 2001 Buliok offensive). Although there were small numbers of reported desertions (troops who were absent without leave or who walked out of camp because of a disagreement with commanders) and attrition (soldiers who were killed while off duty as a result of an altercation or clan feuding), the bulk of the integrees remained with the army. By all indications, the army was also transformed for the better by the integration project. To absorb culturally distinct Muslims into the predominantly Christian-anchored military culture, the army had to institute changes. An assimilation program was designed, which took integree units on study tours to the General Headquarters, Philippine Military Academy, Corregidor, and the Iloilo province-based battalions to expose them to various aspects of military culture.6 Culturally sensitive practices (e.g., halal food preparation and Friday prayers) were adopted during training, which filtered down the line into parallel practices by Christian officer-commanders stationed in the autonomous area. When queried, many of the integrees I interviewed said their commanders by-and-large accommodated their special religious needs at camp or headquarters. As one noncommissioned officer observed, “Sa holiday at Ramadan, may privilege din kami 15 days. [Pero] Nasa individual na yon, pwede man magpaalam sa CO nila.” [We have a 15-day privilege during Ramadan. But that depends on the individual; each can seek permission from the commander.] Another integree-enlistee said, “Dito kami sa loob ng Kampo, pwede natin ma-apply ang pagka-Islam dito. Sa mga lugar walang masjid, hindi tayo makasambahayang.” [We can apply being a Muslim when we are inside the camp. However, in places with no mosque, we can’t do group prayer.] A Christian personnel officer who I interviewed in Maguindanao also said that he routinely allowed integree-enlistees to take on lighter duties during Ramadan, particularly if he knew they were fasting. Some senior officers have devised informal cultural briefings by a Muslim officer to enlistees as a way to overcome prejudice against Muslims. Then colonel Raymundo Ferrer, who commanded troops in Basilan, pioneered the institutionalization of cultural sensitivity by involving nongovernmental organizations in the design and training of peacebuilding and conflict management techniques for paramilitary, officers, and enlisted men under his watch (Philippine Army 2011, 86). While the military power-sharing agreement precluded disarmament, the army implemented a modified guns-for-cash program in which integreerecruits are required to turn in their weapons to be registered, given compensation for each firearm surrendered, and reissued arms for training purposes (Feliciano Gacis, personal communication, March 18, 2010).7 According to Santos (2009, 9), during the period from 1996 to 1999, integree-candidates turned in 4,874 firearms, mostly old M1 Garands and carbines. The program allowed for retention of firearms provided they were registered, but high-

Military Power Sharing

101

caliber and crew service weapons were exempt from these arrangements (meaning they had to be turned in without compensation). The MNLF leaders agreed to this scheme, but with instructions to their integree-candidates to privately procure arms to turn over to the army. Thus, while the guns-for-cash component of the military power-sharing scheme had little impact on the number of MNLF arms (Ferrer 1999; Makinano and Lubang 2000), the army nevertheless was able to collect a substantial amount of the loose firearms in the Bangsamoro zone, at least those held by individuals. For the army, this was a small gain in its efforts to reign in some of the lawless activities and clan fighting in the conflict zone, which are fed by gun proliferation. By contrast, the effect of military power sharing on the police was marginal. Although additional police personnel were provided through the integration project, this carried little symbolic import as the local police in Muslim-dominated areas already had considerable numbers of Muslim officers in their ranks. The absorption also carried the least disruption as it was piggybacked on the PNP’s normal recruitment cycle and did not involve waivers of age and educational requirements. The extra infusion of 1,750 personnel did little to address the operational challenges besetting the local police forces previously mentioned. The local police continued to be as they were before—tied up with the local chief executive and performance deficient. The additional personnel from the integration mostly went into the regular police force, not the Special Forces units that were going after the Abu Sayyaf Group and kidnap-for-ransom gangs. Because of the local residency–local posting norm, the integration project also did little to ameliorate cases of horizontal conflict. In Jowel Canuday’s (2014) study of clan feuding (rido) in Maguindanao and North Cotabato Muslim communities, family members in the security forces (including paramilitary and the nonstate armed groups) were considered important clan assets that could be mobilized. The local police, for the most part, are rendered helpless in these clan wars or drawn into the fray (Hall 2013). Unlike the military, the police did not require integree-candidates to turn in a firearm. Thus, integration into the police provided a better employment option for young Muslim men, particularly when combined with the perks of local posting. Military power sharing served the corporate interests of the MNLF leadership fairly well. By giving field commanders and political leaders discretion over the distribution and the filling up of integration slots, they were able to use the integration project as a means to shore up support within their own ranks and within their community of supporters (given the proviso for proxy). The integration slots became an important patronage currency, enabling respective MNLF commanders and leaders to strengthen loyalty claims among their troops. With secure government employment through the integration program, it also possibly dissuaded some troops from joining rival armed groups and, therefore, stemmed further MNLF

102

Rosalie Arcala Hall

membership loss. As previously mentioned, the guns-for-cash scheme did not reduce the MNLF’s firepower given internal instructions for integreecandidates to privately procure weapons to turn in. Despite the military power-sharing agreement, the MNLF, particularly its units in Sulu Province, retained its ability to mobilize forcibly against the state as it did on two occasions: during the 2001 Jolo siege in support of embattled outgoing ARMM governor and MNLF chair Misuari, and again in 2013 during the Zamboanga siege (Ferrer 1999). Vis-à-vis the state, the MNLF suffered dual losses from the integration scheme: (1) the subsequent dissolution of integree-only units (companies) and the individual integration of their members into various mixed units scattered across Mindanao meant former MNLF commander-turned-army officers lost potential control over presumably loyal integree-enlistees; and (2) the government’s unilateral interpretation that the ARMM police (with the 1,750 police integrees) was the Special Regional Security Force contained in the FPA, not a separate force as expected by the MNLF, meant loss of control in favor of local strongmen (mayors and governors) who now have more of these local police officers under their thumb. The police integrees who were fielded in municipalities and provinces became add-ons to the power of local politicians. Military power sharing did little to alter the balance of power between the state security forces and local elites. The integration program tackled only the quantitative increase (and ultimately ethnoreligious composition) of the army and police, but did not touch the broader issue of how to address the security problems of the autonomous region. The dismantling of private armed groups, many of which directly supported local strongmen and the military’s counterinsurgency efforts, was not part of the military powersharing bargain. Rather, both parties (army and local elites) succeeded in recycling these private armed entities into legitimate paramilitary outfits or as special Civilian Volunteer Organizations (CVOs), which presumably operate with army oversight. The local army units remain as they were, unwilling and unable to contest the independent power of local politicians. This is partly explained by the fact that the army is alien: an important but nevertheless external actor to the Bangsamoro political ecology that is strongly anchored in ethnic and clan relations. Army commanders and units come and go, but the local police, the paramilitary, and the civilian militia all stay and are much more wedded to the complex security arrangements of the area. The army units understand enough that if they forcibly go against the local warlords, they run the risk of being numerically overwhelmed because of the cultural phenomenon of pintakasi (spontaneous mob attack by an entire community of armed men against an outsider). How did the Muslim communities view the military power-sharing agreement? On the whole, MNLF communities regarded the integration scheme as positive because it provided employment opportunities for mem-

Military Power Sharing

103

bers of their family. However, among many MNLF elites and Muslim scholars, the MNLF got the shorter end of the bargain in the military power-sharing agreement. Council of 15 leaders Muslimen Sema (personal communication, March 4, 2009) and Uttuh Salim (personal communication, April 6, 2009), as well as former ARMM governor Parouk Hussin (2005), claim that the government reneged on previous assurances that the MNLF units would form separate units within the AFP and that they would be stationed within the area of autonomy. They also accuse the government of reneging on the promise of a separate Special Regional Security Force with an independent funding source. Mashur Bin-Ghalib Jundam (personal communication, February 21, 2009) and Amina Rasul (2005) also regard the integration process as not having been fully implemented. MNLF supporters (Bai Albaya Wampa, personal communication, March 3, 2009; Monawara Esmael Sanayatin, personal communication, March 3, 2009; Abdullawi Hadji Ebrahim, personal communication, March 3, 2009) who I interviewed several years ago echo the complaints of the MNLF leaders regarding the lack of a separate MNLF unit inside the army, the integrees’ deployment outside the autonomous area and their use in combat operations against the MILF, and the failure to create an SRSF.

Conclusion The decades-old Bangsamoro conflict in Mindanao festers in part because of the weakness of the Philippine state in developing a national agenda by way of its civilian government apparatus or through its security forces (Cook and Collier 2006). In large part, the Philippine state’s inability to contain the conflict is due to a deficient security force structure and capability, which in the years following the democratic transition of 1986 has not seen significant investments in modernization. The creation of the Autonomous Region for Muslim Mindanao in 1986 did not dampen this identity-based conflict. At a historic moment in 1996, the state was presented an opportunity to remake the Mindanao political landscape, with implications for its security force structure. In negotiated conflict terminations with a military power-sharing dimension, some kind of bargaining outcome that reduces state power is intuitive. Yet in this case study of the 1996 Final Peace Agreement between the Philippine government and the MNLF, it is evident that state power through its security forces expanded in the Bangsamoro conflict zone. The inclusion of MNLF integrees in the army ranks strengthened the state’s footprint on contested Bangsamoro ground, compensating for the continued deficits in local civilian governance. First, the type of military power sharing under the 1996 FPA, which involved individual absorption of MNLF

104

Rosalie Arcala Hall

forces into mixed regular army units, rather than as separate units, allowed the military better control over the use of such mixed units. This arrangement further diluted the possibility of the MNLF’s informal influence over former commanders and personnel, inasmuch as they are scattered across four infantry divisions within and outside the autonomous area. The inclusion of proxies into the integree list also enabled the army to balance the potential damage that could have resulted from having waived the educational and age requirements for the MNLF ex-combatants. The proxy-integrees were younger, more professional-mobility oriented, and less attached to the symbolic aspects of the identity-based conflict in Mindanao. Because the primary motivation for joining the integration program was economic (i.e., gainful and steady employment), the army was able to bank on the added personnel to carry out its internal security operations in the Bangsamoro area. The mixed integree units were widely used in combat operations against the Moro Islamic Liberation Front in the 2000 campaign, with integree discipline holding even in subsequent campaigns against breakaway Misuari loyalist factions from Sulu. The integration project provided the army with the personnel assets necessary for its civil-military operations, presaging the SALAAM program and other institutionalized conflict resolution and peacebuilding initiatives that allowed the army to build better relationships with Muslim communities. The Muslim integrees became effective point forward linkages that opened otherwise hostile areas for penetration of nation-building projects carried out by the military. The power-sharing arrangement did not compromise the state’s control over the army and police inside the ARMM. Through the equal membership Joint Force Integration Board, the state was able to accommodate MNLF demands (i.e., generating their own integration list and filling up attrition cases beyond the transition time frame), but retained substantial leverage over the process (i.e., its insistence on validation and firearm surrender as prerequisites to integration). In the case of police integration, the legal maneuver to treat the ARMM Special Police Force as the promised Special Regional Police Force under the 1996 FPA quashed all hopeful calculations by the MNLF for control over a larger police force. Because military power sharing in this instance precluded a shift in the security forces’ orientation, the state was in no danger of losing its foothold in Mindanao. Both the army and police were and remain enmeshed in internal security operations, albeit no longer with the MNLF but with other armed groups. While the integration project was a boon to military capability, it yielded marginal returns on the capability of the local police forces, which remained deficient in their task against illegal economies. Part of that was because the FPA was not accompanied by internal reforms within the police, unlike the military that began experimenting with its internal security operations strategy in the Bangsamoro theater. Against the backdrop of

Military Power Sharing

105

local bosses with their personal armed coteries and the persistence of clan feuds, adding ex-MNLF elements neither strengthened nor induced a change in the weak police force. Neither does integration of former MNLF members have symbolic value as the local police already have a number of Muslims in their ranks. An important dimension in a negotiated agreement is the parties: With whom does the state share this power? As observed by Andreas Mehler (2009) in African cases, this type of elite bargain favors armed groups and precludes civilian parties, particularly civil society organizations and communities whose concerns regarding security are paramount. The tendency of armed groups to factionalize, as the MNLF did with the subsequent split between the Council of 15 and the Misuari wings, translated little to overall security outcomes. The military power-sharing agreement ignored the parties who matter most in terms of local security: the powerful local bosses, the kidnap-forransom gangs, and the Abu Sayyaf Group. Except for the Philippine army, which became a better fighting force against the MILF as well as a more culturally sensitive force, military power sharing did little to improve security outcomes in places where the police remain handicapped. The Philippine case study of the 1996 FPA provides a unique lens for viewing how an endogenously crafted military power-sharing arrangement, with no concomitant disarmament, demobilization, and reintegration (DDR) or security sector reform, pans out for the state. Even in the absence of an external force intervention factor, the Philippine state nevertheless managed to remake power dynamics in Mindanao in its favor. The civilian control complements exerted by the national government remained unchanged, and a gradual exit or withdrawal of state forces typical of a normalization process was absent. Instead, the articulation of state power through the state security forces became more pronounced. The military deployed in the ARMM increased in number and became better at providing government services, largely through the aid of Muslim integrees acting as bridges to otherwise impenetrable Muslim communities. While relatively marginal compared to more organic local nonstate armed groups, the military remains the embodiment of state will in these contested spaces. The Mindanao landscape has seen profound changes since 2009. These include the emergence of new terror groups and the preferential treatment given to the Moro Islamic Liberation Front by the administration of President Benigno Aquino III (2010–2016). The government reopened peace talks with the MILF in 2010, yielding the 2014 Comprehensive Agreement on the Bangsamoro (CAB) which, unlike the FPA with the MNLF, does not contain any provision on military power sharing. The CAB provides for a normalization process—the phased disarmament and demobilization of MILF combatants and other private armed groups in the Bangsamoro core territory—and a Bangsamoro Police Force under the supervision of a new

106

Rosalie Arcala Hall

Bangsamoro Political Entity (BPE). The BPE, as outlined in the Bangsamoro Basic Law (BBL) bill (at the time of this writing, pending in the Philippine Congress), is expected to supplant the ARMM government structure, with more robust economic and political power sharing vis-à-vis the national government. The MNLF Misuari wing’s perceived exclusion from this process resulted in the three-week siege in Zamboanga City in 2013. The conflict episode was contained in Zamboanga City, without parallel eruptions in central Mindanao where Misuari-loyalist armed groups still operate, largely because co-located military units were able to persuade local MNLF commanders to not participate in the siege. This level of accommodation stems from years of positive engagement between the military and the MNLF, on foundations laid by the rebel integration process. Current president Rodrigo Duterte reopened talks with the MNLF and reconstituted the Bangsamoro Transition Commission (tasked with revising the BBL bill) to include representatives from the MNLF Misuari wing. This gesture is expected to appease the MNLF’s concerns regarding political outcomes, particularly where the powers of the regional government are concerned. It is not expected that new demands for military power sharing, or departures from the planned Bangsamoro Police Force, will be made under this new dispensation. Meanwhile, increased military deployments for antiterror tasks in Sulu Island against the Abu Sayyaf Group in 2016, and in Marawi against the Islamic State in Iraq and Syria (ISIS)–connected group of Isnilon Hapilon and Maute brothers in 2017, have brought state power to new heights. With martial law declared in Mindanao (May 2017 to December 2018), the military and the police have assumed more important roles in local governance. Local government units, while not abrogating their authority, are aware of the heightened security regime in their localities. The military victory in Marawi and the stellar public approval rating of the armed forces that followed speak to the institutional gains from these deployments. The military is expanding its reach, even into erstwhile isolated enclaves of Marawi City and Lanao del Sur Province, using civil-military operations strategies it honed with its Muslim integrees from two decades ago. Despite these gains, local security outcomes have been tarnished by President Duterte’s war on drugs. The government’s antidrug operations have placed the police in a lead role with the military lending support. With a death toll of thousands and mounting, this new front points to a drastic shift in the use of state power. As an early case of military power sharing, the Philippines did not have the DDR complement or a security sector reform framework that would later anchor many negotiated conflict termination agreements supported or brokered by the UN. Its relevance lies in the fact that even on a modest scale, military power sharing can shore up the national government’s reach in contested areas by improving relations with communities of ethnic Oth-

Military Power Sharing

107

ers through the enhanced civil-military relations work of the integrees and by improving the military’s overall performance against other security threats. The scheme can also be an easy sell to armed groups by providing them with patronage resources (e.g., employment and prestige), without the state opening up its local military command structure to rebel influence through its integrees. Because the power sharing was directed at only one conflict party, it did not disturb local power configurations since the military did not use its integree assets to dismantle private armed groups or challenge local politicians’ control of police forces, which potentially could lead to security deterioration. The Philippine case also points to an insight that power sharing by way of reconstitution and local control of police forces, rather than through a national military that is deployed for internal security, may be a more viable form of military power sharing. When adopted in conjunction with territorial self-governance, this form of power sharing can incentivize both state and armed actors toward greater collaboration in addressing localized security threats.

Notes 1. The original Autonomous Region for Muslim Mindanao (ARMM), under Republic Act 6734 of 1986, covers four provinces (Sulu, Tawi-Tawi, Lanao del Sur, and Maguindanao). This act was superseded by Republic Act 9054 of 2000, which expanded the area of autonomy to Basilan Province and Marawi City. The autonomous area covers only a part, not all, of Mindanao. 2. Executive Order 212 Section 4 also provides that the ARMM governor can request the president to call on the armed forces to suppress violence in the region where the regional police are not able to do so. 3. The Civilian Home Defense Forces (CHDF) were mainly organized by Christian communities against Muslim opponents. They were related to the Ilaga, armed militia created by Ilonggo Christian settlers against the Muslim Black Shirts in the early stage of the Bangsamoro armed conflict. To date, these civilian militias remain in existence. Renamed Civilian Volunteer Organizations (CVOs), they are the communities’ first line of defense against harassment by Muslim armed groups operating in the Liguasan marsh bordering Maguindanao and North Cotabato Provinces. 4. The radical Abu Sayyaf Group at this point had been responsible for a series of kidnappings for ransom, abductions, and killings of locals and foreigners alike in the Sulu, Tawi-Tawi, and Basilan areas. The group’s members were treated primarily as criminal elements during this period. After the September 11 attacks and the subsequent opening of the US Operation Enduring Freedom with Mindanao as a front, the ASG was rebranded a terrorist organization with links to the Jemaah Islamiyah, an al-Qaeda affiliate operating in Southeast Asia. 5. The downside of the integration project for the army was the remedial measures necessary for the older and less educated ex-combatant integrees. Through the Paaral (education) Program, integree officers were given scholarships to obtain college or university degrees within seven years after integration. Under the Army Literacy Patrol Program (ALPS), integree-enlistees went through two to three hours of biweekly remedial literacy sessions to help them obtain formal equivalency. Even so, it was

108

Rosalie Arcala Hall

difficult for these ex-combatant integrees to catch up. In effect, the integration program created a two-track system of true-blue MNLF ex-combatants just putting in time on a government payroll en route to early retirement, and the relatively young college-educated proxy integrees with better prospects of career advancement paralleling that of their nonintegree peers. Then Sixth Infantry Division commanding general Alfredo Cayton, who I interviewed in 2008, admitted that the promotion process for the integrees was much slower compared to that of their regular counterparts. On the whole, however, being an integree has little bearing on individuals’ identities within the force inasmuch as neither any overt physical marker nor even a special serial number is assigned to them (Hall 2014). 6. According to Colonel Dickson Hermoso (personal communication, May 22, 2014), former head of the Force Integration Branch, the inclusion of Corregidor and Iloilo in the study tour itinerary was designed to counter whatever remaining ill feelings the integrees may have had related to their identity. Corregidor was the site of the Jabidah massacre in the 1970s, where government forces summarily executed Muslim trainees preparing for a secret invasion of Malaysia. Christian Ilonggo settlers in Mindanao organized as armed Ilaga were the Muslims’ key antagonists in the spiral of violence in the 1970s. 7. This is a subset of the larger Balik-Baril (Return-Gun) program that targeted communist rebel returnees and was also administered by the army.

6 Territorial Power Sharing: The Cohesion of Opposition Movements Kathleen Gallagher Cunningham

The literature on power sharing has focused on the stability and quality of postconflict peace and has increasingly assessed the implications of different types of power sharing.1 Among the varieties of power-sharing arrangements, territorial power sharing specifically has been linked to successful transitions away from civil war (Hoddie and Hartzell 2005; Lake and Rothchild 2005). Recent work has also shown that the nature of nonstate actors, in particular their degree of internal cohesion or fragmentation, matters for a number of conflict processes, including the chance of war starting (K. G. Cunningham 2013), incentives to spoil peace agreements (Pearlman 2008–2009), the scope of concessions within settlements and whether they forestall a return to war (K. G. Cunningham 2011), infighting among armed groups (K. G. Cunningham, Bakke, and Seymour 2012; Fjelde and Nilsson 2012), violence against civilians (Weinstein 2007; Wood 2010), collaboration with the state (Staniland 2012), political and military effectiveness of movements (Krause 2013), and human rights violations (Carey, Mitchell, and Lowe 2013).2 The primary focus of this book is on the effect of different powersharing arrangements on power relations. In Chapter 1 of this volume, Andreas Mehler and Caroline A. Hartzell use the term power relations to refer to the relative strength or influence wielded by actors and the means by which the actors use this power or influence, typically in the postconflict environment. In this chapter, I link this concept of power relations to a broader conceptualization of opposition fragmentation. Examining relationships of power and influence among actors, particularly within the opposition side of a dispute with the state, suggests that we already have a clear idea of who the key actors are and assume that they will retain a position of influence. Yet this determination is often a product of a number of

109

110

Kathleen Gallagher Cunningham

factors such as negotiations with the state, support from outside actors, charismatic individuals, or grassroots support, to name but a few. Rather than attempt to trace specific shifts in power among opposition actors, in this chapter I explore the effects of territorial power sharing on the underlying fragmentation of the opposition. The degree of opposition movement fragmentation or cohesion determines the extent to which power can be diffused or centralized. It provides the scaffolding on which power dynamics change in the opposition. Highly fragmented oppositions have a large number of actors that can both compete for influence and potentially step into positions of power created by territorial power sharing. In this sense, fragmented oppositions provide many opportunities for political engagement and progress. Yet fragmented movements are also associated with greater conflict potential (with respect to the outbreak of civil war and the length of such wars). What is missing thus far is an understanding of how power-sharing arrangements affect actor fragmentation. This is critical because of the effects that actor fragmentation has on many processes related to conflict and conflict resolution. Anecdotally, there is some reason to think that the process that leads to power-sharing arrangements might also lead to more fragmentation. Wendy Pearlman’s (2011) work highlights this with the Palestinians, as she demonstrates that the incentives created by different settlement possibilities can lead some actors to splinter or try to subvert settlement. We can also see splintering between political and military wings of opposition movements during negotiations that end civil war. For example, in Burundi, the political wings of the National Council for the Defense of Democracy (CNDD) and Palipehutu split with their militant counterparts to join the Arusha peace process in 1998 to 2000.3 These anecdotes suggest that the process of trying to settle a dispute and the contents of concessions made to nonstate actors in intrastate disputes contribute to the underlying degree of fragmentation within actors. This, in turn, influences the power dynamics among opposition actors. However, the link between settlement (process or outcome) and the cohesion or fragmentation of the opposition side has not been explored systematically. I argue in this chapter that there are reasons that concessions emphasizing territorial power sharing (whether they constitute full settlement of a dispute or not) should lead to greater cohesion of opposition movements. I examine the link between territorial power sharing and opposition fragmentation in the context of self-determination disputes. These disputes are ongoing challenges over the status of ethnonationalist minority groups. For example, the Scots in the United Kingdom and the South Ossetians in Georgia each have long-standing challenges over self-determination with their host states. There are several advantages to exploring the role of power sharing in the context of self-determination disputes. First, self-determination disputes

Territorial Power Sharing

111

include both violent and nonviolent challenges to the state. About half of self-determination disputes included in this study have seen the outbreak of civil war at some point in time while others have seen a limited use of militant politics techniques or are totally peaceful. By investigating the effects of power sharing on opposition fragmentation among all of these disputes, I examine a much larger range of cases than studies that have looked only at war-ending power-sharing agreements. The advantage of a broader look at power sharing is that many instances of power sharing occur outside the context of civil war settlements (see also Chapter 4 by Andreas Mehler in this volume). Many concessions made by states in and out of civil war transfer meaningful powers to the local level. My study’s findings speak to the role of power-sharing arrangements more generally and as it relates to the power dynamics in movements that challenge the state in a number of different ways. Second, the set of self-determination disputes over time is large, including 142 separate disputes that have been active for nearly 4,000 dispute years collectively from 1960 to 2005. Among these disputes, I found 165 instances of transfers of power from the state to the group that can be considered territorial power sharing. This means that the set of territorial power sharing arrangements that I examined was much larger than in studies that focused exclusively on civil war settlements. Third, I was able to examine over-time trends in the fragmentation that underlay the extent to which power can be dispersed or centralized in such movements. There is a tendency, particularly in the quantitative conflict literature, to see groups such as those seeking self-determination as essentially unitary. Yet nearly all movements for self-determination are made up of multiple actors that act on their behalf, making demands on the state. Among the most fragmented movements is the well-known Palestinian case, where many (often competing) organizations press claims on Israel. By identifying all actors involved in making demands over self-determination, I captured a broader picture of movement fragmentation than by examining only rebel groups or rebel group splintering. Thus, looking beyond civil war settlement to self-determination disputes more generally provides a fertile ground for evaluating the effects of territorial power sharing on opposition movements. This approach to understanding the effect of power sharing on fragmentation differs from much of the existing literature in two key ways. First, I examined territorial power sharing inside and outside the context of civil war. Many existing studies have emphasized the role that power sharing has in creating a successful transition from civil war to democracy. These studies necessarily limit our understanding of the role played by institutional arrangements to situations in which there has been a significant breakdown in order. In contrast, I examined a wide range of contexts from almost totally conventional politics (e.g., the Scottish in the United

112

Kathleen Gallagher Cunningham

Kingdom) to persistent civil war (the southerners in Sudan before the separation of South Sudan). The second difference between this study and other research is how I defined territorial power sharing. Because existing studies have focused on war-ending settlements, power sharing has typically been defined as a specific type of institutional arrangement, with territorial power sharing including elements of decentralization or devolution (see Chapter 1 of this volume). Here, I examined a set of instances where power has transferred to the substate group regardless of whether it occurred during conflict or was institutionalized. Such transfers can occur through formal change in institutions, such as the creation of the modern Scottish parliament in the United Kingdom, but do not necessarily do so. Mehler and Hartzell defined territorial power sharing as an institution that “divides authority among levels of government by creating forms of decentralization based on territory.”4 Studies that focus on postwar settlement deals have examined territorial power sharing as a set of rules. However, a great deal of the actual sharing of power occurs through more incremental changes in the status of groups in society, rather than comprehensive institutional reforms. For example, the Bodos in India, the South Tyrolians in Italy, and the Buganda in Uganda all have had territorial power sharing that occurred outside the context of war-ending settlements. While I did include all instances of authority or power transfer to the more local level, I excluded concessions made to groups that were primarily cultural in nature (e.g., language recognition) in the subsequent analysis.

Explaining Opposition Fragmentation A large body of literature has addressed why some actors are more fragmented than others. These explanations have emphasized a range of factors that affect group fragmentation that can be examined in the context of selfdetermination movements or rebellion. A long-standing literature in sociology has emphasized the role that repression can play on group cohesion. Competing arguments have suggested that repression increases cohesion and that it increases fragmentation. Georg Simmel (1955) argues that confrontation or conflict with a common enemy will lead to greater cohesion. Arthur A. Stein (1976) suggests that this response will be the case only when internal leadership is strong. Theodore McLaughlin and Wendy Pearlman (2012) make a similar argument about repression, suggesting that it can generate cohesion within the opposition, but only when group members are satisfied with the status quo in terms of the group’s institutional configuration. Others have argued that conflict will lead to fragmentation for several reasons. Charles Tilly (1978) links repression to fragmentation of the opposition, arguing that repression can raise the costs of collective action, and

Territorial Power Sharing

113

thus unity, as costs rise for individuals. James DeNardo (1985) argues that even if individuals remain committed to a cause in the face of repression or conflict, rising costs can generate disagreement about how to respond to these costs and this could promote the fracturing of the opposition. Beyond the social movements literature, a number of works have examined multiple actors in conflict with the state and the role that connections between nonstate actors play in conflict processes for civil war.5 Some studies have emphasized the degree to which rebel organizations cooperate (Bapat and Bond 2012), or what determines splintering. Hanne Fjelde and Desirée Nilsson (2012) argue that fragmentation (particularly the entry of new rebel groups) in civil war will occur when costs to entry are low (e.g., in shorter disputes, or those over the government rather than territory). Michael Findley and Peter Rudloff (2012) suggest that the chances of fragmentation are higher for relatively weak groups. Fotini Christia (2012) also emphasizes power, suggesting that the distribution of power among coalitions of actors in a conflict can lead to fragmentation, as each actor will ally with or turn on others in an effort to be part of a minimum winning coalition in the dispute.6 In sum, the existing literature has tended to connect fragmentation and cohesion of oppositions to repression or conflict against a common foe (with competing predictions about the effect this has), the costs of joining the dispute, and the power relationship between actors during civil war.

Territorial Power Sharing and Opposition Fragmentation Beyond some expectations about the possible emergence of organizations that try to spoil a peace process, there is little work that has spoken directly to whether and what type of power-sharing arrangement could contribute to fragmentation of oppositions.7 Stephen John Stedman (1997) argues that “peace processes create spoilers” (p. 7), and that individuals that occupy spoiler roles help determine what settlements are possible. Kelly M. Greenhill and Solomon Major (2006–2007) as well as Pearlman (2008–2009), in contrast, suggest that the nature of a settlement plays a large role in determining whether incentives for spoiling will occur. These works have tended to emphasize the behavior of actors (whether they try to subvert a settlement) rather than the degree to which the opposition becomes more or less fragmented in response to power sharing. Here, I focus more explicitly on the outcome and argue that territorial power sharing is likely to have divergent effects over time on the internal structure of opposition movements for self-determination. I define Territorial power sharing as accommodation of the self-determination group that includes some transfer of power or control to the group. Examples include

114

Kathleen Gallagher Cunningham

the creation of the Saami Parliament in Finland in the mid-1990s and the decentralization of greater economic powers to Corsica in France in 2004. This is in keeping with Mehler and Hartzell’s definition in Chapter 1 in that self-determination groups are all geographically concentrated, and federalism and decentralization that transfer power to the group fit in this definition. This definition, however, is potentially more inclusive of powersharing instances than the general definition offered by Mehler and Hartzell. Increases in political powers that are transferred down to groups can occur via a change to a more decentralized system, or within an existing system of territorial power sharing as powers are expanded. Instances of increasing territorial power sharing should have varying effects on the fragmentation or cohesion of the opposition over time. Territorial power sharing is designed to create a sense of security in postconflict situations (see Lake and Rothchild 2005), but also to resolve underlying discontent on the part of the self-determination group. In the short term, territorial power sharing can lead to the emergence of new organizations in several ways. First, organizations may form with the express goal of rejecting a possible change in territorial power sharing. This can happen because people within the group oppose greater power sharing, or because they believe the increase in power is deficient and want to press for more. Second, territorial power sharing can also spur splintering of existing organizations where there was internal disagreement over the change. Yet after settlement, there are several mechanisms through which territorial power sharing can lead to a reduction in fragmentation over time, including government co-optation, satisfaction of specific factions’ demands, and the preemptive satisfaction of unmobilized individuals. First, territorial power sharing can be designed to integrate opposition organizations, either through the creation of new local institutions that are part of the state or by bringing them into the larger governance structure. Either path could lead organizations to decrease demands against the state and potentially abandon them altogether. For example, in the dispute between the government of Comoros and the Anjouanese, the opposition Mouvement Populaire Anjouanais (MPA) became the local government of Anjouan with the consent of Comoros in 1997. Second, outside of explicit integration into the state, some organizations will stop making demands after territorial power sharing is enacted because they are satisfied, thereby decreasing the overall mobilization and fragmentation of an opposition movement. For example, the Mizo Union party in India ceased calling for greater self-determination after the Mizoram state was created in 1986. Third, territorial power sharing can preempt further mobilization that might generate a more highly fragmented opposition by reducing grievances when concessions are successfully implemented.

Territorial Power Sharing

115

Thus, there are multiple reasons to believe that territorial power sharing that satisfies some demands of the group will reduce opposition fragmentation over time despite the potential for a short-term increase in the fragmentation of a movement when territorial power sharing initially occurs. This discussion leads to two hypotheses: Hypothesis 1: Territorial power sharing leads to immediate fragmentation of self-determination movements. Hypothesis 2: Territorial power sharing leads to longer-term cohesion of self-determination movements. I tested these predictions by examining a set of disputes over selfdetermination. An advantage of looking at self-determination disputes is that there are common grievances across all groups in terms of how they define themselves vis-à-vis the state. All self-determination groups make claims to a distinct national identity and demand greater rights based on this. Moreover, I could observe the fragmentation of the movement by examining all actors that made claims surrounding the issue of self-determination over time. This allowed me to look at a wide range of actors (from pressure groups to political parties to rebel groups). Because self-determination identity is relatively stable, I was also able to examine fragmentation of these movements over time, not only in periods of intense political change (as would likely be the case for regime change disputes).

Empirical Exploration Does territorial power sharing decrease or increase opposition fragmentation? In this section, I present a set of statistical analyses on changes in fragmentation in opposition movements seeking greater self-determination from the state. I found that territorial power sharing leads to decreased fragmentation, but not immediately. Measuring Fragmentation

To examine changes in fragmentation, I measured the number of organizations making claims over self-determination in any given year. The dataset includes 142 disputes over self-determination from 1960 to 2005, taking place in seventy-seven countries. The disputes were identified by the Center for International Development and Conflict Management’s Peace and Conflict Report (Marshall and Gurr 2003). This list identifies a set of groups

116

Kathleen Gallagher Cunningham

that have grievances related to autonomy or their status in the state and have an ethnonationalist basis for mobilization. Using this list, I created a measure of internal fragmentation of each movement every year by counting the number of organizations that made demands over the group’s status in any specific year.8 Organizations are selfidentifying through the demands they make related to self-determination status for the group. Demands related to self-determination included maintaining the status quo, autonomy, independence, union or reunion with another state, and the creation of a pan-ethnic state. All organizations must represent the same group of people. Thus, to be considered part of the movement and included in the fragmentation measure, an organization had to claim to represent the interests of the self-determination group and to make demands on the state about the status of the group. This measure captures fragmentation in that more organizations reflected a greater degree of movement fragmentation. However, it did not necessarily capture the entire scope of fragmentation, which could include aspects of institutionalization and the distribution of power across organizations (see Bakke, Cunningham, and Seymour 2012; Krause 2013). Measuring power distributions among organizations in the same movement would require several dimensions of types of power and would need to reflect the venues in which power is likely to be exercised. For example, different bases of power would matter differently in the event of an outbreak of war or the outbreak of a mass nonviolent campaign. Yet other bases of power would matter for conventional political contestation. Focusing on the number of organizations as one critical dimension of movement fragmentation, I found that there was a wide range of degrees of fragmentation in self-determination both across cases and over time within cases. The number of internal factions in self-determination groups ranged from one organization in a given year (e.g., the Berbers in the 1960s in Algeria) to thirty-nine organizations (the incredibly fragmented Kashmiri Muslims in 2002). Among all self-determination movement years from 1960 to 2005, I found that in about 68 percent of the observations the movement was comprised of two or more organizations. Figure 6.1 shows the distribution of observations by the number of organizations in a movement year (up to fifteen organizations). In a number of years, self-determination movements had only one organization; however, there was a great deal of variation among movement years when I examined internally divided movements. Not only were many movements fragmented, but the process also was dynamic. There was a great deal of change in fragmentation within movements for self-determination over time. Nearly 90 percent of movements experienced some fragmentation (i.e., an increase in the number of organizations pressing demands related to self-determination). Just over 70 percent of movements increased cohesion from one year to the next at some point in time.

Territorial Power Sharing

117

Figure 6.1 Fragmentation of Self-Determination (SD) Movements

Measuring Territorial Power Sharing

States made a variety of concessions to opposition movements seeking selfdetermination, both within and outside the context of open conflict. From 1960 to 2005, I found over 200 concessions made over self-determination in these 142 disputes. Some of these were war-ending settlements (e.g., the Dayton Peace Accords that ended the conflict in Bosnia), others entailed further transfers of power to groups already enjoying autonomy (e.g., the series of concessions to the Flemish and Walloons in Belgium), and still others addressed basic cultural rights (which would not be considered territorial power sharing). To assess the effect of territorial power sharing on movement fragmentation, I examined all concessions that entailed a transfer of power to the group represented by the opposition movement (e.g., the creation of a new local parliament, increased powers over local economic policymaking, or inclusion in the national government as a representative of the group).9 I excluded concessions that were purely cultural in nature such as the creation of language protection rights. A key reason to use such an expansive conceptualization of territorial power sharing was that all of these instances of conferring greater power to the subnational group had the potential to

118

Kathleen Gallagher Cunningham

shift the balance of power within it. The dividends to individuals or factions within the movement for receiving an increase in influence at the center, economic, or local political power did not necessarily rely on the institutionalization of such changes. Many instances of territorial power sharing are multifaceted, including provisions that we would consider territorial power sharing as well as additional concessions that are cultural in nature. For example, the creation of the 1989 semiautonomous districts in the Chittagong Hill territory in Bangladesh included political, economic, and cultural elements. The 1985 agreement between the Assamese and the Indian government involved political rights as well as cultural protection. Of all concessions made to self-determination movements from 1960 to 2005, about 79 percent entailed some dimensions of territorial power sharing. Nearly 25 percent of these instances of territorial power-sharing concessions were made during a period of conflict with the state (measured as active civil war).10 Methodology and Control Variables

To assess the effect of territorial power sharing on movement fragmentation, I used an ordinary least squares regression on the change in the number of organizations at one-, three- and five-year intervals. For example, the Chittagong Hill peoples received territorial power sharing in 1989 and 1997. There was no immediate change in the number of organizations representing the group following either instance of territorial power sharing. However, changes in fragmentation occurred sometime after each of the extensions of territorial power sharing. Five years after the 1989 territorial power-sharing change, the movement had two more active organizations than in 1989. Five years after the 1997 accommodation, the movement had one fewer active organization. Coding change over these intervals allowed me to examine the effect of concessions in a given year on the change in the movement’s fragmentation immediately after concessions were made, but also several years later when we would expect the fruits of successful territorial power sharing to be born. The choice of one-, three-, and fiveyear increments reflects the potential for immediate changes, but also two points (three and five years) at which we could reasonably expect that either the territorial power sharing was influencing opposition fragmentation or not. Studies of recurrent conflict have employed a five-year window to determine success or failure (see Licklider 1995; Walter 1997; Hartzell 1999). While that designation may be somewhat arbitrary, it is also a reasonable time frame for determining the effects on mobilization patterns. I clustered standard errors on the dispute level. I included a series of control variables that could influence changes in fragmentation or cohesion of an opposition movement. First, I controlled

Territorial Power Sharing

119

for whether there was active civil war that year as well as repression by the state. The civil war measure, which was based on the Uppsala Conflict Data Program (UCDP/PRIO) Armed Conflict Database, used a threshold of twenty-five battle deaths for each year.11 The repression measure was from the UCDP One-Sided Violence data.12 Both repression and civil war can impose costs on an opposition movement, which others have suggested can engender fragmentation (c.f. Stein 1976; McLauchlin and Pearlman 2012). I also controlled for two characteristics of the group—the number of organizations in the movements in that year and the size of the group’s population (logged).13 More fragmented movements may be more likely to further fragment or become more cohesive just by nature of having many organizations relative to more cohesive movements. Larger populations may also be more likely to be represented by more organizations. I also controlled for several characteristics of the state, including whether it was federal,14 whether it was democratic, and the size of the state’s economy, military (measured as personnel), and overall population.15 Both federal states, with local venues for competition, and democratic states may engender more organizations by offering more opportunities for them. The measures of state economy and military related to the strength of the state. Stronger states may dissuade mobilization that leads to more organizations. States with larger populations tend to see more conflict generally (Hegre 2008), which may influence the dynamics of movement growth. The analyses in Table 6.1 show that territorial power sharing had a negative effect on fragmentation. Models 1 through 3 returned a negative coefficient on territorial power sharing. This was statistically significant at the 0.01 level in Models 2 and 3, the three years after and five years after models. This provides support for Hypothesis 2, but not for Hypothesis 1. Opposition movements seeking greater self-determination were likely to be more cohesive three years after concessions and even more so five years later. The effect of territorial power sharing across these time intervals increased. Territorial power sharing was associated with a decrease of 0.27 organizations three years after concessions, and 0.37 organizations five years after concessions were made. The other control variables showed some significant effects. These included a positive effect (i.e., increasing movement fragmentation) for states with greater state military personnel and larger populations. Thus, these findings provide some limited evidence that fragmentation of the opposition movement is more likely in stronger and larger states. The finding on more populous states may just reflect the larger, and likely more diverse, pool of participants that mobilize in a dispute. The positive finding on military personnel suggests that stronger states have more divided opposition movements. This may reflect the ability of such states to prevent the emergence of a unified opposition, which weaker states would be less able to do.

120

Kathleen Gallagher Cunningham

Table 6.1 Effects of Concession on Group Fragmentation

Variables Territorial power sharing Civil war Repression Number of organizations Group size Federal state Democracy GDP per capita (log) Military personnel State population (log) Constant Observations R-squared

(1) 1 year after

(2) 3 years after

(3) 5 years after

–0.11 (0.11) 0.03 (0.04) 0.07 (0.07) 0.00 (0.02) 0.00 (0.01) 0.05 (0.03) –0.00 (0.04) 0.00 (0.02) 5.09* (2.77) 0.02** (0.01) –0.19 (0.14) 3,012 0.01

–0.27** (0.11) –0.00 (0.09) –0.04 (0.12) –0.00 (0.06) 0.02 (0.02) 0.13 (0.10) –0.02 (0.11) –0.02 (0.05) 17.48** (8.49) 0.06** (0.03) –0.42 (0.36) 2,966 0.02

–0.37** (0.17) 0.00 (0.15) –0.05 (0.17) –0.02 (0.09) 0.03 (0.04) 0.21 (0.16) 0.05 (0.18) –0.06 (0.08) 31.07** (14.17) 0.11** (0.05) –0.53 (0.58) 2,821 0.02

Note: Ordinary least squares regression coefficients with standard errors clustered on dyad are in parentheses. GDP, gross domestic product. Two-tailed tests: ***p < 0.01, **p < 0.05, *p < 0.1.

The findings on the effect of territorial power sharing decreasing fragmentation over time were robust to a number of alternative specifications of the models. The findings held when I included controls for regional and temporal effects using regional dummies and decade dummies as well as a dummy for the Cold War period. The coefficients on territorial power sharing remained positive and statistically significant at the 0.05 level or better for Models 2 and 3 when I used a measure of civil war onset instead of civil war incidence (or excluded civil war altogether), when I controlled for whether the movement had ethnic kin in an adjoining state, and when I controlled for the number of movements for self-determination in the state.16

The Process of Cohesion The preceding statistical analyses demonstrate that territorial power sharing is associated with a decrease in movement fragmentation. That is, territorial power sharing appears to make these movements more cohesive over time,

Territorial Power Sharing

121

though this effect is not immediately observed when territorial power sharing is agreed to. This provides support for my argument that territorial power sharing can work to decrease fragmentation, but the statistical relationship does not speak to how this happens. From the standpoint of states facing such opposition movements, it is important to understand how territorial power sharing decreases fragmentation. If the decline in fragmentation is the product of satisfaction and the co-optation of moderate organizations into the state, then territorial power sharing could be seen to be playing a positive role in terms of dispute management and resolution. Another scenario could be that territorial power sharing leads organizations to merge together and present the state with a more unified, stronger challenge. This is less likely to be seen as constituting successful management of a dispute. Utilizing data on organizations making claims over self-determination, I examined the fate of those organizations that become inactive after instances of territorial power sharing. To explore the fate of these organizations, I traced the reasons for inactivity of seventy-one organizations that ceased making demands related to self-determination within five years after an instance of territorial power sharing occurred. Recall that instances of territorial power sharing can refer to the creation of new structures of governments (e.g., regional parliaments) and also to the transfer of greater power to the group that does not entail the creation of new institutions.17 Table 6.2 shows the fate of these organizations. I found several paths to inactivity for organizations seeking selfdetermination. Some were disbanded (and in some cases disarmed) in response to the territorial power sharing agreement. For example, the Shanti Bahini who fought for the autonomy of the Chittagong Hill peoples disarmed after a territorial power-sharing treaty with the Bangladeshi state. Other organizations, such as the Corsican Peasant Liberation Front, ceased activity after being banned by the state. About 17 percent of the organizations in Table 6.2 disbanded, either directly as a result of territorial power sharing or for some other reason. The majority of organizations (52 percent) did not have a definitive end that I could find; rather, I considered them to be inactive because I could no longer find evidence of their making demands related to self-determination. This could be the result of the organizations having died off, or could have occurred because they shifted away from making claims related to the group’s status. The second most common outcome for organizations was merging into other organizations (either new organizations or joining an existing organization). Finally, 10 percent of the organizations were coded as having been co-opted or integrated into the state in some way. For example, a faction of the Front for Restoration of Unity and Democracy (FRUD) representing Afars in Djibouti joined with the dominant national party after territorial power-sharing concessions were made.

122

Kathleen Gallagher Cunningham

Table 6.2 Fate of Inactive Factions Five Years After Concessions Disbanded via Settlement 7% (5)

Disbanded Other

Inactivity

Merged

Co-opted/ Integrated

10% (7)

52% (37)

21% (15)

10% (7)

What does this pattern suggest? First, fears that territorial power sharing creates stronger, more cohesive challengers to the state were not supported.18 Less than a quarter of the organizations that ceased activity after territorial power sharing could be linked to a merger with other organizations seeking the same goal. Second, just over a quarter of the organizations integrated or disbanded and just over half ceased activity in terms of making open demands related to self-determination. This provides at least some anecdotal evidence to suggest that the satisfaction and co-optation paths to greater cohesion are exercising an effect on the fragmentation of opposition movements.

Territorial Power Sharing, Fragmentation, and Stable Peace In this chapter, I explored the relationship between territorial power sharing and the fragmentation of opposition movements for national self-determination. Examining the role of territorial power sharing on opposition fragmentation is critical for understanding how power sharing will translate to changes in the extent to which different elements in society gain and retain power. The underlying fragmentation of the movement sets the stage for understanding the centralization or decentralization of power among the opposition actors. Opposition fragmentation, in turn, has been demonstrated to have substantial effects on processes that are likely to make territorial power sharing successful in managing the dispute or not, including spoiling behavior, conflict onset, and civilian abuse. Territorial power sharing decreases opposition fragmentation, seemingly in ways that increase the potential for persistent peace. I did not find that the reduction in the number of organizations pressing claims for selfdetermination in the wake of territorial power sharing was the product of significant merging. That is, reduced opposition fragmentation appeared to be the product of integration with the state or cessation of demands much of the time. If this reflects greater integration of group demands into the state structure and satisfaction, then territorial power sharing promotes a stable peace for such groups. While my analyses provide evidence that territorial power sharing decreases opposition fragmentation and give us a clearer picture of the process of decreased fragmentation, some questions remain. Although it is clear from Table 6.2 that the fate of most organizations after an instance of

Territorial Power Sharing

123

territorial power sharing is not to band together to form a metamovement, more work is needed to explore the dynamics of how and why specific organizations cease to press claims. In the absence of detailed information on the role that state repression plays in the fate of these organizations, we cannot attribute all observed cases of inactivity to satisfaction on the part of organizations. Clearly, some organizations were banned and subsequently disbanded, but other tactics may have been used by the state to limit the activity of organizations seeking to press self-determination claims. Differences in the ways that states co-opt and integrate organizations also may be worth examining in greater detail as these could influence the prospects for stable peace between the group and the state. Some organizations cease making demands related to self-determination and others become part of a local administrative structure under the oversight of the state. The debate over whether territorial power sharing, and in particular regional devolution, exacerbates tensions seems to assume that a unique identity will necessarily be maintained as a basis for political action. Empirically, however, it appears that there is some variation in the degree to which becoming or joining the local political apparatus means integration with the state. In conclusion, territorial power sharing leads to decreased fragmentation of opposition movements seeking greater self-determination. As such, this process is potentially another route through which power sharing might make conflict less likely. Fragmented movements for self-determination are more likely to see the onset of civil war, and civil wars with more parties last longer (D. Cunningham 2006). My analysis suggests that concerns that territorial power sharing leads to fragmentation are unwarranted and, in fact, territorial power sharing may actually lead to longer periods of peace by reducing opposition fragmentation.

Notes 1. See Chapters 2–10 in this book, Hartzell and Hoddie (2007), as well as Roeder and Rothchild’s (2005) volume on power sharing after civil war. 2. See also Stedman (1997), Varshney (2002), Beissinger (2002), Kydd and Walter (2002), E. Wood (2003), King (2004), Kalyvas (2006, 2008), Greenhill and Major (2006–/2007), and Lawrence (2010) on aspects of fragmentation. Fragmentation has also been linked to the shape of authoritarian regimes (e.g., Smith 2005; Brownlee 2007), and the trajectory of social movements (e.g., McCarthy and Zald 1987) to labor movements (e.g., Olson 1982). 3. This example illustrates splintering in a peace process, though not necessarily one focused on territorial power sharing. 4. They note the differences that can exist between different forms. 5. These include Atlas and Licklider (1999) and Nilsson (2008). 6. Others, such as Staniland (2012) and Driscoll (2012), emphasize the role of the state in “flipping” or manipulating actors into defecting.

124

Kathleen Gallagher Cunningham

7. There is a debate over whether accommodation related to territorial power sharing (in particular, devolution) leads to fragmentation of states. See Brancati (2009) for a summary of this debate. 8. This measure was coded using both aggregate sources of information, such as the Minorities at Risk Project and the Uppsala Conflict Database as well as news reports from Keesing’s Record of World Events and LexisNexis Academic news sources (all English language news). See K. G. Cunningham (2014) for a detailed description of coding procedures. 9. It is possible that such inclusion could be a token concession, not actually leading to increased power. Chapter 4 in this volume also addresses how center elites can address what they call “local” issues. 10. Civil war is measured as twenty-five or more battle deaths in a year based on the Uppsala Conflict Data Program/Peace Research Institute Oslo (UCDP/PRIO) Armed Conflict Database (N. P. Gleditsch et al. 2002). 11. I matched the actors and dispute issues in the UCDP data to the lists of selfdetermination movements to determine which civil wars occurred in these disputes. 12. One-sided violence includes intentional attacks on civilians by governments or other organized armed groups resulting in at least twenty-five deaths in a given year. The measure here is dichotomous, indicating whether there was any one-sided violence by the state in a year. See Eck and Hultman (2007) and Sundberg (2009). 13. The number of organizations is based on my coding. The group size estimates are used from Ethnic Power Relations Dataset (Wimmer, Cederman, and Min 2009), the Minorities at Risk Project (Bennett and Davenport 2003), and then supplemented by Central Intelligence Agency (CIA) values where possible (CIA World Factbook). 14. I also reproduced the models in Table 6.2 with a one-year lag of the federalism dummy as well as a dummy indicating whether the territorial power sharing was a change to federalism. The results for territorial power sharing are similar to Table 6.1 in all models. The federalism indicators were not statistically significant. 15. Democratic states are those scoring 6 or greater on the Polity score (Marshall and Jaggers (2000). Military personnel data is from Correlates of War Project National Military Capabilities Data, Version 3.2 (Singer, Bremer, and Stuckey 1972). Both gross domestic per capita and state population data are from K. S. Gleditsch (2002). 16. The analyses in Table 6.1 confirm a relationship between instances of territorial power sharing and a subsequent decrease in opposition fragmentation. Territorial power sharing was coded at the time the agreement or decision to enact territorial power sharing occurred. In some cases, there was a time delay between implementation of the territorial power sharing and agreement time. I recoded the territorial power-sharing variable to reflect the year of implementation and reproduced the analyses in Table 6.1. Delay in implementation was typically one to two years and the findings are similar to the initial analyses. Territorial power sharing was associated with decreased movement fragmentation three and five years after. The coefficient on the territorial power-sharing variable was negative and statistically significant at the 0.10 level in two-tailed tests. 17. The findings in Table 6.1 were similar if I examined only territorial power sharing that created new institutions of local government separate from those that did not. 18. This is a relatively broad concept of “challenging the state,” which could be either violent or nonviolent movements that make demands on the state. A notable nonviolent case that illustrates this is the, to many, surprisingly strong independence campaign in Scotland.

7 Economic Power Sharing: Potentially Potent . . . but Likely Limited Caroline A. Hartzell

On the face of it, economic power sharing is a measure that seemingly has the potential to shape power relations in countries emerging from civil war. Depending on their nature, changes in rules governing who gets what insofar as the economic realm of state power is concerned could fundamentally alter the balance of power in a society. Given the types of economic power-sharing measures that have been agreed to as part of civil war settlements, three different types of changes in power relations seem feasible. One is a shift in power between the political center and a region. Another is the alteration of power between or among groups. A third involves increases in or diminutions of the power of individual political elites. Despite the potential that economic power sharing has to alter power relations in states emerging from civil war, there are a number of reasons for believing that its effects on structures and relations of power may be limited in scope and form. First, state actors have demonstrated a reluctance to use economic power sharing as a means of ending civil wars, a stance that has been reflected in generally limited forms of economic power sharing in those cases in which such measures have been agreed to. Second, in those instances in which economic power sharing has been included as part of a civil war settlement, its implementation has often been limited or incomplete. Third, economic power sharing may also be more limited in its impact on power relations than other types of power sharing due to the effects of global economic forces or the actions of international actors. Determining whether economic power sharing is potent or limited in its effects on postconflict power relations proves to be quite challenging. As I discuss in more detail below, the lack of appropriate data, potential selection effects, and the difficulty of disentangling the effects of economic power sharing from other forms of power sharing make it difficult to empirically assess the impact of economic power sharing on the distribution of 125

126

Caroline A. Hartzell

power within states emerging from civil war. As a result of these constraints, I limited myself to conducting a plausibility probe of the effects that economic power sharing has on horizontal power relations or power relations between or among groups.

Economic Power Sharing: A Brief Overview Power-sharing institutions have been defined as rules that allocate rights to and limit the exercise of power by groups that have been engaged in intrastate conflict (Hartzell and Hoddie 2007). Economic power sharing, one of the four types of power-sharing mechanisms most commonly agreed to as part of civil war settlements, seeks to manage societal conflict as well as regulate relations between the state and society through the design of institutions that define who gets what—and how they are to get it—where the economic dimension of state power is concerned. The element of state power bears emphasizing insofar as the definition of economic power sharing that I employ here is concerned. Economic power sharing involves the allocation of resources controlled by the state, per some set of defined rules, to one or more units, groups, or actors. Although natural resources (e.g., oil, diamonds, and timber) tend to receive the bulk of attention in studies that focus on the resource base of civil war states, the types of economic resources available for distribution by the state can extend to state-owned enterprises, public service employment, and other forms of preferential economic treatment. State policies that call for the seizure or shift of privately owned resources, such as land, from one individual or group to another are not encompassed in this definition of economic power sharing. Although such actions, which involve the exercise of state power, obviously can have an effect on the wealth of different individuals or groups (and thus have the potential to affect power relations), they do not involve the (re)distribution of economic bases of state power. Rules for sharing economic power can take a variety of forms. In some cases, ownership of a natural resource may be allocated to a region within the postconflict state; this often occurs in conjunction with an autonomy arrangement, as was the case in Aceh. In other instances, the state and a rival group may enter into an agreement that designates how the revenue associated with a resource is to be apportioned (e.g., Sudan’s Comprehensive Peace Agreement). Groups also can gain access to economic assets or rents through the appointment of one of their members to a position in government that gives that individual responsibility for state-controlled resources; for example, the 1994 Lusaka Protocol’s allocation of the post of minister of geology and mines to a member of the National Union for Total Independence of Angola (UNITA). Finally, states also can design preferen-

Economic Power Sharing

127

tial policies or rules that draw on a variety of measures—regulations, court orders, administrative rules, and other public interventions—to distribute state-owned economic resources among rival groups (Weiner 1983). The logic that is most often advanced for the design of rules specifying how some measure of state economic power is to be controlled and used in the postwar state is to help secure, and extend, the peace. Economic powersharing measures are thought to help achieve this end through three mechanisms. One is by limiting the ability of any single group within the postconflict state to control state economic power and allocate resources in a manner that threatens the security of rival groups (Hartzell and Hoddie 2007). A second mechanism is by helping to ensure that a group that feels aggrieved by the share of state-controlled economic resources it receives, particularly in comparison to other groups, will receive a larger share of resources (Binningsbø and Rustad 2012). Finally, and most crassly, resources controlled by the state also can be used to buy off or induce potential spoilers to abide by the terms of a peace agreement (Stedman 2002). Somewhat surprisingly, given the frequency with which economic factors, and particularly access to state-controlled natural resources, have been identified as motivations for actors or groups to engage in armed conflict against the state (cf. Collier and Hoeffler 2004; Rustad and Binningsbø 2012), relatively few civil war settlements have called for economic power-sharing measures as a means of ending intrastate conflicts.1 Of the 127 civil wars ended during the period 1945–2006, 68 called for one or more forms of power sharing (political, military, economic, and territorial). Of the sixty-eight civil war settlements calling for some form of power sharing, only twenty-nine included a provision for economic power sharing. This stands in contrast to the fiftythree settlements that called for political power sharing and the forty-seven that contained a commitment to share military power, but is on par with the twenty-nine settlements that called for territorial power sharing.2

Types of Economic Power Sharing and Their Potential Impacts on Postconflict Power Relations Establishing the types of effects that economic power sharing has on the access of regions, groups, and individuals to the economic resources controlled by the state is an important endeavor because changes in the rules governing access to those resources have the potential to alter the balance of economic power in a society. The balance of economic power matters because of the potential that it has to shape the exercise of political power in a society. Individuals who suffer economic exclusion to a point of being unable to meet their most basic needs, for example, will most likely be unable to participate in any meaningful manner in the political realm (Sen 1999).3

128

Caroline A. Hartzell

In an effort to assess the potential that economic power sharing has to alter power relations in countries emerging from civil wars, I turn now to the different types of economic power sharing measures that have been employed as part of civil war settlements. First, I focus on those measures that have the potential to alter the balance of power between regions and the central government. Next, I turn to measures with the potential to alter power relations horizontally or among identity groups. Then, I discuss the potential that economic power sharing has to affect the power of individual political elites. In the case of each of the measures, I note the logic behind its use—that is, how it is supposed to work—as well as factors that may limit its potential to alter power relations. Shifts in Power Between the Political Center and Regions of a Country

Separatist conflicts or civil wars in which territorially defined groups have sought to exercise greater control over their own interests have at times been ended with the creation of settlements intended to alter the balance of power between the central government and one or more regions of the country. One way in which such a change in the balance of power potentially can be effected is through the use of territorial power sharing; for example, by shifting decisionmaking power to subunits of the state or ensuring that territorially based groups are included in the institutions of the federal government (Hartzell and Hoddie 2007). Arguably, however, such alterations in power relations may amount to little real change in the balance of power if subunits and the groups that live in them do not obtain the resources necessary to implement the types of policies they seek to promote. Autonomy measures are also likely to do little to address the grievances of territorially based groups that stem from their economic exclusion unless accompanied by economic power-sharing arrangements that allocate some measure of control over resources to the groups in question. In light of the above, a number of substate units and territorially based groups seeking to alter the balance of power with the central government have called for the inclusion of one of the following economic power-sharing measures in their respective civil war settlements: (1) the allocation of natural resource ownership to a region within the postconflict state; (2) an agreement between the state and a rival group designating the apportionment of revenues associated with a resource; (3) the design of preferential policies or rules that distribute state-controlled economic resources among groups; or (4) the appointment of a member of a group that has engaged the state in armed conflict to a position in the government that gives that individual responsibility for state-controlled resources.

Economic Power Sharing

129

The first type of economic power-sharing measure that groups have called for in some civil war settlements is: the allocation of natural resource ownership to a region within the postconflict state. As Nicholas Haysom and Sean Kane (2009, 8) observe, “The property-ownership regime for natural resources is a highly emotive issue that is often intertwined with identity-based conflicts.” In cases in which grievances or security concerns emanate from perceptions that resource revenues have been distributed in an inequitable or unjust manner, marginalized groups may insist on acquiring ownership of the resources as a condition of ending an armed challenge to the state (Binningsbø and Rustad 2012). Although arrangements of this variety have most often been arrived at as part of civil war settlements in federal states, unitary states in which groups in certain regions of the country claim autonomy also have had ownership of local resources ceded to them as part of a peace agreement (Haysom and Kane 2009). The transfer of resource ownership from the central government to a region is fraught with symbolism. Given that central governments are loath to give up any of their power, some actors are likely to perceive an economic power-sharing agreement of this type as a sign of state weakness, a concession forced on the government by virtue of the nonstate actors’ military success or their potential to reinitiate armed conflict. A transfer of resource ownership is not just symbolic, however. If the central government is highly reliant on the resource in question for revenues, a shift in ownership of the resource could negatively impact the state’s economic power— and, presumably, add to the power of the region or group in question. A number of factors may, however, limit the potential that this form of economic power sharing has to alter power relations between regions and the political center of the country. Ownership, as Haysom and Kane (2009, 6) note, does not necessarily equate with “the power to control, regulate and manage natural resources,” powers that are “potentially more significant than ownership rights in themselves.” If national authorities retain the power to develop and exploit natural resources, there may be little real shift in economic power from the state to subnational actors or regions following the transfer of resource ownership. Additionally, if nonstate actors or regions lack the capacity to manage and develop the natural resource over which they have acquired ownership, they will have attained little meaningful economic power. Finally, the impact that control of a resource has on the balance of economic power between the state and other actors can change over the course of a conflict as well as in the postconflict period. This appears to have happened in the case of Aceh, for example, where the Indonesian government, once loath to surrender control of oil and gas deposits in the region, was persuaded to do so once it became clear that reserves there were becoming depleted (Wennmann 2011).

130

Caroline A. Hartzell

A second type of economic power-sharing measure that groups have called for in some civil war settlements is: an agreement between the state and a rival group designating the apportionment of revenues associated with a resource. Rather than ceding ownership of a resource to a group or region, the state can agree to a formula for sharing the revenues associated with a natural resource. Sudan’s Comprehensive Peace Agreement stands as a case in point. The formula agreed to in that instance called for 50 percent of revenues to be allocated to the government of Sudan and 50 percent to the government of Southern Sudan, with producing regions each first receiving a 2 percent share. In those instances in which the share of revenues designated to a group or region is transferred in a timely fashion and in a manner free from political interference, economic power-sharing arrangements of this type can end up shifting significant resources from the central government to other actors and thus may serve to alter the economic balance of power between the center and region in some meaningful fashion. However, as Haysom and Kane (2009) make clear, based on past experiences, groups often have reason to doubt that revenues will be distributed in such a transparent manner. Indeed, this lack of trust led the government of Southern Sudan to insist that it directly collect its share of revenue from oil, taxes, and fees rather than relying on a centralized tax system, a provision that arguably entailed a significant shift in economic power to the region as well as providing the potential future state with its own revenue base (Wennmann 2011). Changes in Horizontal Power Relations

Grievances of an economic nature can extend beyond questions concerning the control and distribution of natural resources or the revenues produced by those resources.4 Nonstate actors may seek to alter other types of state policies they perceive to be responsible for producing, or sustaining, economic disparities or horizontal inequalities among groups. Economic power-sharing measures designed to address this type of grievance have often taken the shape of redistributive policies designed to provide public goods “such as admission into schools and colleges, jobs, promotions, business loans, and rights to buy and sell land on the basis of membership in a particular . . . group” (Weiner 1983, 35; see also Stewart 2002). The 1996 Philippine Peace Accord constitutes a relevant example, where the state was called on to provide the Special Zone of Peace and Development with resources to foster development in a marginalized area of the southern Philippines. Policies specified in the accord included basic services such as water and socialized housing, entrepreneurial development support in the form of livelihood assistance, and credit facilities (Hartzell and Hoddie 2007).

Economic Power Sharing

131

A third type of economic power-sharing measure that groups have called for in some civil war settlements, in particular, appears to have the greatest potential to produce a change in horizontal power relations: the design of preferential policies or rules that distribute state-controlled economic resources among groups. Economic power sharing can be used to reduce horizontal inequalities by increasing the amount of wealth, income, and opportunities available to marginalized or poorer populations. If increased access to resources has the effect of empowering marginalized populations and enhances the ability of poorer groups to participate more meaningfully in the political realm, shifts in power relations among groups may be expected to follow. Generally speaking, any effects that this type of economic power-sharing measure has on power relations can be expected to take some time to manifest themselves. Affirmative action policies in the area of higher education, for example, may take years to translate into a rise in the proportion of professionals from a previously marginalized group. Failures fully to implement preferential policies may also serve to limit the degree to which this form of economic power sharing transforms power relations. According to the Kroc Institute for International Peace Studies’ Peace Accord Matrix, for example, it took seven years to move from minimal to intermediate implementation of the economic and social development measures called for in the 1996 Philippines Peace Accord (Joshi and Darby 2013). Change in the Power of Individual Elites

A fourth type of economic power-sharing measure that groups have called for in some civil war settlements is: the appointment of a member of a group that has engaged the state in armed conflict to a position in the government that gives that individual responsibility for state-controlled resources. By having a group representative appointed to a position in the central government, nonstate actors seek to ensure that the individual will be able to monitor, if not influence or control, the manner in which certain state-controlled resources are employed. This form of economic power sharing was employed in Sierra Leone with the appointment of Foday Sankoh, leader of the Revolutionary United Front, to chair of the Commission for the Management of Strategic Resources, National Reconstruction and Development as part of the 1999 peace agreement. This form of economic power sharing potentially could produce a change in horizontal power relations if the group representative were to use the proceeds of the resources under his or her control in a manner that reduced intergroup economic disparities or otherwise empowered the group’s followers. However, there are reasons to doubt that this particular type of power sharing would produce such an effect. First, appointment as

132

Caroline A. Hartzell

a minister within the postwar government does not necessarily translate into wholesale control of some set of state-controlled resources. Even if the minister was so inclined, it is doubtful that individual would have the ability to use the position to direct sufficient resources, for a sufficiently long enough period of time, to the members of the group so as to effect a shift in horizontal power relations. Second, once the elite representative of a group is a member of the central government, a desire to maintain that position of power may make the individual less inclined to challenge power relations, at least in any manner that may be perceived as threatening by other members of the government. Although the type of economic power sharing discussed in this section is unlikely to produce a shift in horizontal power relations, it clearly has the potential to affect the power of individual elites. Individuals who are newly appointed to positions in the central government gain access to resources that can make them even more influential power brokers vis-à-vis the groups or regions they represent. The obverse effect is likely to hold in those instances in which individuals have had to cede a position in the central government to the representative of a rebel group.

Assessing the Effects of Economic Power Sharing A number of challenges exist where efforts to assess the effects that economic power-sharing measures have on postconflict power relations are concerned. Some of these are conceptual in nature, while others involve data and modeling issues. I briefly discuss these challenges below in an effort to provide some context for the analysis that follows. Conceptual Challenges

In the overview provided above, I focused on four different types of economic power-sharing measures, positing that two have the potential to alter power relations between the political center and regions of a country and that the other two measures could alter horizontal power relations in one instance and the power of individual elites in another. Although I believe these claims rest on a solid theoretical basis, it is important to acknowledge that the effects produced by each of the different forms of economic power sharing may not be as discrete in character as discussed above. Indeed, in the discussion of the fourth form of power sharing I acknowledged as much, noting that although the intention behind the inclusion of such a measure in a settlement is ostensibly to provide benefits to marginalized groups (an action which could serve to alter horizontal power relations), the effect is most likely to be to enhance the power of individual elites.

Economic Power Sharing

133

The following are some other reasons for believing that there may well be a blurring of the lines between the different forms of economic power sharing and their effects on postconflict power relations: • Theoretically speaking, all four types of economic power sharing discussed above could have an impact on horizontal power relations. In the case of the first and second types of economic power sharing (the allocation of natural resource ownership to a region within the postconflict state, and an agreement between the state and a rival group designating the apportionment of revenues associated with a resource), it is generally understood that such measures should redound to the benefit of some group(s) located in the region in question, resulting in a reduction in horizontal inequalities between those group(s) and those traditionally associated with control of the political center. This could occur if the territorially based group gains control of sufficiently valuable resources or a large enough apportionment of the revenues associated with a resource. Whether or not there is a concomitant alteration of horizontal power relations in a manner that benefits the group(s) in question, however, depends on a host of factors including: the degree of follow-through or implementation of the power-sharing measure and the actions of regional elites with regard to the distribution of the rents or revenues associated with the natural resources. • The use of either of the first two forms of economic power sharing could also have the effect of producing an alteration in the power of individual elites. The case of Aceh is illustrative. Following the Helsinki peace agreement, interelite competition emerged in the region among different elites vying for control over “lucrative government contracts and business opportunities” (Barron, Rahmant, and Nugroho n.d.). As noted by Mohammad Hasan Ansouri (2012), actors in the upper echelons of the Free Aceh Movement (GAM) insurgency secured access to significant state-controlled economic resources and, concomitantly, political power. For example, Nur Djuli, senior GAM negotiator at the Helsinki peace talks, became chair of Badan Reintegrasi Aceh (the Aceh Re-integration Agency), receiving “a high income and other special privileges provided by the Agency” (Ansori 2012, 35). • Finally, the potential for a blurring of the lines between the different types of economic power sharing and the effects that each produces is also likely to be compounded by the fact that more than one type of economic power sharing has the potential to be included in a settlement. In such instances, it can be difficult to hypothesize what effect the economic power-sharing measures will have on postconflict power relations. While some types of economic power-sharing measures might have complementary effects, other combinations could work at cross-purposes, effectively limiting or canceling their effect on power relations.

134

Caroline A. Hartzell

A different type of conceptual challenge stems from the difficulty of operationalizing some of the dependent variables—that is, different types of power relations—discussed above. Although it makes theoretical sense to talk about a relationship of power between the political center and regions of the country as well as the notion that these could change over time, what exactly are the components of that power relationship? How does one operationalize such a relationship to track whether changes in it have taken place? The same types of questions can also be asked about the power of individual elites. What factors contribute to a political elite’s base of power? Are they the same for elites at the political center and regionally based elites? How can changes in the power of individual elites be measured? And does an increase in the power of one elite necessarily equate to the decline in power of a comparably placed elite? As these questions suggest, operationalizing some of the types of changes in power relations that I have posited may be associated with economic power-sharing measures can be rather challenging. Data and Modeling Challenges

A number of data-related and modeling issues also make it difficult to empirically test the effects that economic power sharing has on postconflict power relations. Perhaps the most basic of these is the lack of data appropriate for testing some of the relationships specified above. This is particularly true of the first type of changes in power relations discussed above. Even if it were possible to operationalize a measure of the balance of power between the political center and regions of the country, regional-level data would be necessary to calculate whether a change in that balance had taken place. At present, cross-national time series datasets containing appropriate regional-level variables do not exist. And although datasets containing information on the leaders of individual states do exist (e.g., Archigos, created by Goemans, Gleditsch, and Chiozza 2009), these do not cover political elites at the group or regional levels of analysis. Potential selection effects are another challenge that must be confronted by analysts of economic power sharing. Scholars have suggested that power-sharing measures do not appear in civil war settlements randomly; instead, they may be used in the cases of civil wars that have proven the most difficult to end (Hartzell and Hoddie 2015). Although econometric models can and have been employed in an effort to contend with selection effects where power sharing is concerned, the fact that economic powersharing measures are always used in conjunction with other forms of power sharing raises further concerns regarding efforts to analyze their effects. Do actors agree to include economic power-sharing measures in a civil war settlement only when other types of power sharing have been included? If so,

Economic Power Sharing

135

why is that the case? Do actors representing the government expect that other types of power sharing will serve to attenuate the effects of the economic power-sharing measures that are included? If actors choose to include economic power sharing in a settlement for these types of reasons, merely controlling for the other forms of power sharing included in the settlement will not serve to address the potential selection effects that such strategic decisionmaking represents.

Analysis In light of the challenges discussed above, my effort to analyze the relationship between economic power sharing and postconflict changes in power relations was limited to a focus on horizontal power relations. My decision to focus on only one type of potential change in power relations was driven by the lack of data appropriate for examining potential changes in the balance of power between regions and the central government and in the power of individual elites. My analysis of the potential effects that economic power sharing has on horizontal power relations in the wake of civil war took the form of a plausibility probe. I focused on six countries in sub-Saharan Africa that have experienced civil war; in some cases, some of the countries have experienced more than one episode of civil war. Two of the countries I examined, Burundi and Liberia, initially did not include economic power sharing as part of the settlement ending a civil war, but added the measure to later settlement(s) of their respective conflicts. The Republic of the Congo and Uganda never included an economic power-sharing measure in any of their respective civil war settlements.5 Sierra Leone included economic power sharing as part of all three of its civil war settlements as did South Africa in its single civil war settlement. The following logic guided my selection of cases: by focusing on Burundi and Liberia, I hoped to be able to see whether the introduction of economic power sharing in a later civil war settlement, following one or more settlements that initially had not called for economic power sharing, had any effect on power relations. I examined the Republic of the Congo and Uganda to see whether changes in horizontal power relations take place in the absence of an agreement to share economic power. Finally, by looking at Sierra Leone and South Africa, I hoped to gain some insight into the effects that committing to economic power sharing as part of an initial civil war settlement (South Africa) or as part of serial settlements (Sierra Leone) may have on horizontal power relations. I relied on two indicators of power distribution from the V-Dem: Varieties of Democracy dataset. The first measure centers on the distribution of power

136

Caroline A. Hartzell

by social groups. The V-Dem codebook notes that a “social group is differentiated within a country by caste, ethnicity, language, race, region, religion, or some combination thereof” (Coppedge et al. 2016, 250). In the case of this indicator, country experts, typically citizens or residents of the country being coded, were asked to respond to the question “Is political power distributed according to social groups?” by selecting one of the following responses: 0: Political power is monopolized by one social group comprising a minority of the population. This monopoly is institutionalized, i.e., not subject to frequent change. 1: Political power is monopolized by several social groups comprising a minority of the population. This monopoly is institutionalized, i.e., not subject to frequent change. 2: Political power is monopolized by several social groups comprising a majority of the population. This monopoly is institutionalized, i.e., not subject to frequent change. 3: Either all social groups possess some political power, with some groups having more power than others; or different social groups alternate in power, with one group controlling much of the political power for a period of time, followed by another—but all significant groups have a turn at the seat of power. 4: All social groups have roughly equal political power or there are no strong ethnic, caste, linguistic, racial, religious, or regional differences to speak of. Social group characteristics are not relevant to politics.6

The second measure of power distribution focuses on the distribution of power by socioeconomic position. Coders were asked to respond to the question “Is political power distributed according to socioeconomic position?” by selecting one of the following responses: 0: Wealthy people enjoy a virtual monopoly on political power. Average and poorer people have almost no influence. 1: Wealthy people enjoy a dominant hold on political power. People of average income have little say. Poorer people have essentially no influence. 2: Wealthy people have a very strong hold on political power. People of average or poorer income have some degree of influence but only on issues that matter less for wealthy people. 3: Wealthy people have more political power than others. But people of average income have almost as much influence and poor people also have a significant degree of political power. 4: Wealthy people have no more political power than those whose economic status is average or poor. Political power is more or less equally distributed across economic groups.7

In the following, I used the V-Dem data to graph trends in each of the two types of power distribution—that is, the distribution of power by social groups and the distribution of power by socioeconomic position—for each

Economic Power Sharing

137

of the six countries over a period of years.8 In the following subsections, I briefly describe each case, noting when a civil war came to an end, how it was ended, and whether or not a settlement called for economic power sharing (or any of the other three forms of power sharing).9 Burundi

Burundi’s first civil war broke out in 1965, following the first set of postindependence elections. The conflict ended some four years later via a military victory; no power-sharing measures were agreed on. Civil war next broke out again in 1972. The conflict, which lasted three months, saw mass killings of Hutu by the Tutsi-led army following an initial attack on Tutsis by Hutu members of the gendarmerie. There was no agreement to share power of any kind following this episode of genocidal conflict. Political power sharing was called for, however, following a third short-lived outbreak of war in 1988. Economic power-sharing measures also were included in settlements of the conflict that stretched from 1993 to 2000, the conflict that subsequently followed from 2001 to 2003, and the 2004–2006 conflict. Each of the latter three settlements also called for political and military power sharing measures. As shown in Figure 7.1, there appears to be little correlation between economic power-sharing measures and the distribution of power by social group and socioeconomic group in the case of Burundi. The distribution of

Figure 7.1 Power Distribution by Socioeconomic Position and Social Group in Burundi

138

Caroline A. Hartzell

both forms of power remained unchanged following the settlement of Burundi’s first civil war while the distribution of power by social group became slightly less equitable following the 1972 conflict. A small improvement (i.e., a move toward greater equality) in the distribution of power by social group took place in 1988. Interestingly enough, in that year political power sharing was included as part of the civil war settlement. A shift toward a more equitable distribution of power by social group and by socioeconomic group took place in 1993–1994, years that corresponded with the outbreak of a new conflict in Burundi. The settlement of that conflict in the year 2000, although calling for various forms of power sharing (including economic power sharing), produced no shift in the distribution of power. Two years after the 2003 settlement, which resembled the 2000 civil war settlement in calling for political, military, and economic power sharing, Burundi saw movement toward less equitable distributions of power by social group and socioeconomic group. The country experienced no change in the distribution of either form of power following the 2006 settlement, which also called for the same three forms of power sharing. Liberia

An examination of three settlements of the civil war in Liberia also yielded little evidence of a correlation between economic power sharing and shifts in the distribution of horizontal power in that country (see Figure 7.2).

Figure 7.2 Power Distribution by Socioeconomic Position and Social Group in Liberia

Economic Power Sharing

139

Following the outbreak of civil war in 1989, a settlement of the conflict was reached in July of 1993. That settlement, which called for political power sharing, was followed by a slight improvement in the distribution of power by social group; no change took place in the distribution of power by socioeconomic group. Following a renewed outbreak of the conflict in July 1994, a settlement reached in 1996 called for political, military, and economic forms of power sharing. That power-sharing agreement subsequently saw the distribution of power by social group become less equitable in character. Conflict reerupted in 1999, ending in 2003 with a settlement that again called for political, military, and economic forms of power sharing. That settlement was followed by a marked movement toward greater equality in the distribution of power by social group and a smaller shift of the same nature in the distribution of power by socioeconomic group. The Republic of the Congo

Shifts in the distribution of horizontal power in the Republic of the Congo appear to bear little association with power sharing. As can be observed in Figure 7.3, the distribution of power by social group and by socioeconomic group exhibited the biggest shift in a more equitable direction in the years 1990–1991, before war broke out in the country. Following the 1993–1994 conflict, no power-sharing measures were included as part of the truce ending

Figure 7.3 Power Distribution by Socioeconomic Position and Social Group in Republic of the Congo

140

Caroline A. Hartzell

the conflict. The distribution of power by social group and socioeconomic group became less equitable between 1996 and 1998. During that time, armed conflict broke out again, lasting for four months in 1997. That conflict episode was ended without any agreement to share power. An agreement to share military power was arrived at following the outbreak of violence in 1998–1999. There were no apparent shifts in the distribution of power following that agreement nor the subsequent settlement reached in 2003 following another year of armed conflict. The 2003 settlement also called for military power sharing. Uganda

Uganda has experienced four episodes of civil war during the period following its independence. The first civil war, a short-lived conflict that took place in 1966, and the second conflict, which lasted for six months during 1978–1979, each ended in military victory. In neither case was any form of power sharing agreed on. Political and military forms of power sharing were offered as part of the conclusion of the conflict that lasted from 1980 to 1986, although that war also ended in a military victory. War once again broke out in 1986. Its conclusion, in 2006, saw no offers to share power. Again, as Figure 7.4 makes clear, there is little apparent relationship between power sharing and the distribution of political power in the case of

Figure 7.4 Power Distribution by Socioeconomic Position and Social Group in Uganda

Economic Power Sharing

141

Uganda. Shifts in power by socioeconomic group were, interestingly enough, more marked than changes in power by social group in the case of this country. Following the end of the conflict in 1966, there was a slight worsening in the distribution of power by social group. The distribution of power by socioeconomic group became markedly more equitable following the end of the civil war in 1979. In neither of these two instances was there any offer of power sharing. After 1986, a slight deterioration in the distribution of power by socioeconomic group took place. Sierra Leone

Three attempts were made to end the civil war in Sierra Leone. The first, which came in 1996 following five and one-half years of fighting, called for political, military, and economic forms of power sharing. Two subsequent settlements, in 1999 and 2001, also called for the same forms of power sharing. Neither of the first two settlements saw any subsequent change in the distribution of power by social group or socioeconomic group. The absence of any change might well be due to the fact that renewed outbreaks of conflict soon followed each of those settlements. After the third settlement, which did produce a durable peace, there was (as can be seen in Figure 7.5) some movement toward a more equitable distribution of power by social group and by socioeconomic group.

Figure 7.5 Power Distribution by Socioeconomic Position and Social Group in Sierra Leone

142

Caroline A. Hartzell

South Africa

South Africa’s civil war settlement was characterized by the inclusion of all four kinds of power sharing. Concomitant with the adoption of the settlement, there was a marked movement toward a more equitable distribution of power by social group and a notable, although smaller, movement toward greater equality in the distribution of power by socioeconomic group, each of which can be traced in Figure 7.6. Following that single shift in power, there was no further movement in the distribution of power for the remainder of the period under observation.

Discussion What does this brief examination of six country cases, which together account for twenty-one civil war episodes and their associated settlements, tell us about the relationship, if any, between economic power sharing and horizontal power relations in post–civil war states? Based on a review of the evidence, it appears that economic power sharing has a negligible effect on horizontal power relations. In most of the post–civil war episodes that I reviewed, there was no shift in the distribution of power among social or socioeconomic groups following the adoption of economic power sharing.

Figure 7.6 Power Distribution by Socioeconomic Position and Social Group in South Africa

Economic Power Sharing

143

In fact, in those instances in which a change in the distribution of power occurred, it sometimes occurred in the midst of an armed conflict and at other times took place following the end of a conflict but in the absence of any type of power-sharing arrangement. Finally, there were also instances in which the inclusion of economic power sharing in a settlement was followed by a shift in a less equitable direction in the distribution of power between social and socioeconomic groups, an outcome that is the opposite of the effect typically thought to follow from the use of this measure. It is possible that the nature of the plausibility probe that I conducted may have led me to miss or overlook some influential effect of economic power sharing on horizontal power relations. One potential limitation of the analysis, for example, is the limited time period observed following a number of the relevant settlements of civil wars. As I noted earlier in the chapter, some types of economic power-sharing measures can be expected to take considerable time to have an effect on the balance of power between or among groups. It may be the case that shifts of this nature occurred at some time after the period on which I focused in the analysis. Another possible shortcoming is that the V-Dem measures for the distribution of power that I employed are national in nature. It is conceivable that economic powersharing measures extended to relatively small groups or to delimited areas of the country might have succeeded in altering horizontal power relations locally or in some other manner that failed to register at the national level.10 The results of the plausibility probe cannot, of course, be considered conclusive. First of all, the probe was based on a limited number of cases, all of which were drawn from one region of the world. Second, in none of the cases was it possible to isolate the effects of economic power sharing. Although this reflects the fact that economic power-sharing measures have always been adopted in conjunction with at least one other form of power sharing, the absence of any cases that adopted only economic power sharing makes it difficult to know whether it or some other form of power sharing was doing the work of altering horizontal power relations (in cases where such a change occurred). Additionally, although I attempted to include cases that differed on the basis of how they were terminated (negotiation vs. military victory), the number of conflict episodes they experienced, and whether or not they called for economic power sharing, the analysis failed to control for variables such as the severity of the conflict, the role of outside actors, and, critically, the extent to which the terms of a settlement were implemented, all of which have the potential to influence the dependent variable in question. Although the results of the plausibility probe are only suggestive in nature, there are reasons for believing that they reflect the limited potential that economic power sharing has to alter horizontal power relations. I address each of these in turn below.

144

Caroline A. Hartzell

First, although the number of economic power-sharing measures that have been adopted as part of agreements to end civil wars is not negligible, it is significantly lower than the number of political and military power-sharing measures that have been included as part of settlements. The restrained use of this measure could be the product of supply-side factors, demand-side factors, or both. On the supply side, it seems clear that groups that control, and derive benefits from, the economic levers of state power will be reluctant to agree to any measures that reallocate economic resources of the state to other groups. Insofar as the demand side is concerned, the apparent lack of interest in or willingness on the part of groups with less power to press for economic power sharing is difficult to account for. One potential explanation is that such groups put more effort into securing political forms of power sharing in the belief that those have a greater likelihood of effecting a change in horizontal power relations. Whatever the explanation, the relative lack of emphasis on economic power sharing suggests that civil war actors generally have little interest in using it to produce more equitable power relations between or among groups. Second, the potential that economic power sharing has to impact horizontal power relations is obviously subject to whether or not the relevant measures are implemented, as well as the extent to which and manner in which they are implemented. Groups that are the beneficiaries of unequal power relations can be expected to try to secure less than full implementation of measures that may have the effect of reducing their political power. And even in those instances in which the state may be inclined to implement economic power-sharing measures fully, factors beyond the state’s control may make it difficult for the central government to follow through on its commitments. A case in point includes pressures by some international actors for countries emerging from civil war to adopt neoliberal economic policies that may make it challenging for the state to engage in redistributive activities (Paris 2004). Although my analysis focuses only on the relationship between economic power sharing and horizontal power relations, some of the factors that I identified as limiting the potential for economic power sharing to alter those relations are also likely to apply to the use of economic power sharing to effect a change in the balance of power between region and center. In particular, one can expect that actors at the political center will have little interest in ceding control of resources, revenues, and rents once dominated by the center to regional actors. Accordingly, even if such commitments are made as part of a settlement, there may well be less than full implementation of such measures. The one type of potential change in power relations that seems least likely to be subject to the types of limitations that I discussed is a change in the power of individual elites. The commitment to appoint a group repre-

Economic Power Sharing

145

sentative to a position in the government that gives that individual control of state-controlled resources is one that it is nearly impossible to only partially implement. Although members of the government may be reluctant to follow through on such a commitment, the highly visible nature of such an appointment makes it more likely that they will follow through on it. One can thus expect that the inclusion of this type of economic power-sharing measure in a civil war settlement will add to the power of the elites in question. Economic power-sharing measures intended to effect a shift in power between the political center and regions of the country may also have the effect of adding to the power of the regional elites responsible for managing the resource or allocating the rents from them. Exactly what these changes in the power of individual elites look like, how long they endure, and what they imply for the rest of society are all critical questions—and ones beyond the scope of this analysis.

Conclusion At the outset of this chapter I noted that although economic power sharing theoretically has the potential to alter power relations in states emerging from civil war, in all likelihood the effects that this type of power sharing has on power relations are limited. Although I only attempted to analyze this proposition via a focus on horizontal power relations, the results of the plausibility probe that I conducted appear to provide support for this outlook. Additionally, I enumerated a number of reasons for believing that economic power sharing will have limited effects on a shift of power from the political center to the region of a country. The type of change in power relations that economic power sharing is most likely to produce is to add to the power of some individual elites (and perhaps lessen the power of others). The limited potential of economic power sharing to transform certain kinds of power relations should not be taken to suggest that it is not an important mechanism for ending civil wars. It may well be the case that civil war actors see economic power sharing as a short-term tool to be used to enhance their security, rather than as an implement for rebalancing power relations. Groups may instead opt to rely on political power sharing as a means of altering power relations and redressing grievances. The fact that economic power sharing is always coupled with at least one other form of power sharing suggests that civil war actors may in fact see some forms of power sharing as complementary or as buttressing one another.11 The potential for some type of interactive effects between or among different forms of power sharing, thus, is one issue to which scholars should devote attention in the future. There clearly is still a good deal to be learned about the effects that economic power sharing has on postconflict power relations. My study

146

Caroline A. Hartzell

admittedly has generated more questions regarding these processes than it has provided definitive answers. It is my hope that by thinking theoretically about some of the effects that economic power sharing might have on power relations, as well as calling attention to some of the factors that currently make it difficult empirically to test those relationships, this work will serve to convince others of the importance of devoting further attention to this topic.

Notes 1. Aning and Atuobi (2011) made the same point referring to the relative lack of economic power-sharing measures included as part of negotiated peace settlements in West Africa. 2. This is based on data drawn from Hartzell and Hoddie (2015). Although twenty-nine civil war settlements called for economic power sharing and twentynine called for territorial power sharing, only fourteen settlements called for both of these types of power sharing. 3. As research by Cederman, Gleditsch, and Buhaug (2013) demonstrates, the economic exclusion of groups exerts a strong effect on the risk of civil war. 4. At the heart of horizontal power relations is a focus on the distribution of power between or among groups. These groups are typically distinguished on the basis of identity-based categories. 5. My logic in selecting cases in which there was no agreement to share economic power was to attempt to observe whether factors (which I did not identify in this analysis) other than economic power sharing might regularly be associated with shifts in the distribution of power among groups. 6. Responses were converted from an ordinal scale to interval by the measurement model employed by V-Dem. 7. Responses were converted from an ordinal scale to interval by the measurement model employed by V-Dem. 8. I used V-Dem’s online graphing tool to generate the graphs. 9. I focused on all four types of economic power sharing in the analysis that follows since, as discussed previously, each has the potential to shape postconflict power relations. 10. I thank Matthew Hoddie and Andreas Mehler for making these points. 11. One piece of evidence in support of this proposition is Ottmann and Vüllers’ (2015) finding that, among power-sharing measures, the implementation of promises to allocate political and economic posts to rebel groups had the effect of reducing the risk of civil conflict recurrence.

PART 3 Power Sharing and the Quality of the Peace

8 Government Respect for the Physical Security of Postconflict Populations Matthew Hoddie

In this book we consider how power sharing in the aftermath of civil war has the potential to transform postconflict states and societies. In particular, we seek to identify and evaluate the different ways that power relations are shaped by the new set of institutional incentives that emerge with the adoption of power sharing. The question of how these institutions perform in practice, as the introduction (Chapter 1) by the editors serves to highlight, has largely been neglected as much of the research on power sharing has instead focused on the debate over whether these institutions inhibit or encourage a return to war.1 In this chapter, I consider one means by which power sharing may influence the practices of the postwar state: its respect for the principle that individuals should be protected from government threats to their physical integrity in the forms of extrajudicial killing, torture, political imprisonment, and disappearances. My consideration of this topic provides an opportunity to explore the degree to which post–civil war power-sharing conditions one aspect of the balance of power between state and society. I do so by evaluating whether post–civil war countries that adopt power sharing have a stronger tendency to threaten the physical security of their own populations in comparison to postconflict states that eschew such mechanisms. The conventional wisdom that has been articulated in prior research on this topic is that power-sharing structures tend to foster an environment that is antithetical to the protection of physical security or physical integrity rights. In this sense, existing work has suggested that power sharing cultivates an environment in which the balance of power favors government actors and provides them with the capacity to threaten the security of citizens. These studies, however, have tended to rely largely on anecdotes from individual cases rather than systematic evidence. My chapter’s contribution to the study of the relationship between power sharing and physical 149

150

Matthew Hoddie

integrity rights is that I consider this topic through the analysis of a series of quantitative tests. Based on an analysis of a dataset for post–civil war states between the years 1984 and 2006, the central finding of my study is that there is little support for the view that the adoption of power-sharing structures diminishes a country’s ability to protect physical integrity rights. This proves to be the case in analyses considering the aggregate number of power-sharing structures that were put into place following conflict, as well as tests focusing on individual dimensions of power sharing. I find instead that other aspects of the postwar environment prove more influential in determining the degree to which a state respects physical integrity rights. These are the conflict intensity of the concluded civil war, the size of the country’s population, and the number of international nongovernmental organizations (INGOs) that had a presence within the state. These results suggest the need for caution when making claims about the influence that power sharing has on the politics and societies of post– civil war states. Given the virtual omnipresence of power sharing in negotiated settlements to end civil wars, there is a temptation to attribute almost all subsequent developments in postconflict states to the influence of these institutional structures. Yet claims of power sharing’s effects should always be subjected to empirical evaluation rather than assumed. Such an approach allows for a distinguishing between those phenomena for which power sharing has a genuine and meaningful influence and those instances in which other factors prove more important. This chapter is organized as follows. I begin by defining the two concepts that are central to this study: power sharing and physical integrity rights. Then, I summarize the arguments of recent studies, each suggesting that a negative relationship should exist between the creation of power-sharing institutions following civil war and a state’s respect for physical integrity rights. Next, I present the research design, including details about how I coded the variables included in the statistical analyses. In the next section, I summarize the findings, highlighting the lack of consistent evidence to indicate that power sharing diminishes a state’s commitment to the protection of physical integrity rights. I also speculate about the reasons why power sharing may not prove particularly important in shaping a given post–civil war state’s commitment to the protection of these rights. Finally, I discuss the implications for research on power sharing that might be drawn from the study.

Key Terms: Power Sharing and Physical Security Rights Since the goal of this study is to explore the possibility that a relationship exists between power sharing and physical security or physical integrity

Government Respect for the Physical Security

151

rights in the context of states emerging from civil war, I begin with a discussion of how I define these key terms. Power Sharing

The definition of power sharing that I adopt for this study characterizes these mechanisms as “rules that, in addition to defining how decisions will be made by groups within the polity, allocate decision-making rights, including access to state resources, among collectivities competing for power” (Hartzell and Hoddie 2003, 320). This approach seeks to account for the diverse means through which civil war rivals may share power by taking into account four different dimensions: political, military, territorial, and economic power sharing. Political power sharing is concerned with proportionality in the distribution of central government authority, with collectivities guaranteed a degree of representation within state institutions as a function of their membership in a group. The strategies that can be used toward this goal are electoral proportional representation, administrative proportional representation, and proportional representation in the central government’s executive branch. Military power sharing distributes authority within the state’s coercive apparatus. This can be accomplished by bringing together adversaries’ armed forces within a unified state security force; by appointing members of the subordinate group(s) to leadership positions in the state’s military; or, in rare cases, letting opposing sides keep their weapons or maintain their own security forces. Territorial power sharing divides authority among levels of government by creating forms of decentralization based on territory. Finally, economic power sharing provides groups in divided societies access to or control of state resources by distributing wealth, income, or control of natural resources or production facilities on the basis of group identity (Hartzell and Hoddie 2003, 2007). Physical Security or Physical Integrity Rights

The outcome of interest for my analyses is the protection of physical security or physical integrity rights, defined as the right of people to “be free from arbitrary physical harm and coercion by their government” (Cingranelli and Richards 2010, 410). I am thus interested in the degree to which the power-sharing arrangements that are established at the end of civil war shape people’s sense of physical security vis-à-vis the postwar state. There clearly are other forms of human rights that also merit consideration in terms of the degree to which post–civil war governments are committed to their protection and promotion. These include rights with the effect of providing individuals with greater autonomy from their governments.

152

Matthew Hoddie

Such empowerment rights are comprised of items such as freedom of movement (both international and domestic), freedom of speech, and freedom of religion (Cingranelli and Richards 2010, 418). I believe that an initial emphasis on only physical integrity rights is particularly appropriate given the common finding that violations of these types of rights increase precipitously in the context of war—whether it is civil or international (Davenport 1995; Mason 2004). This suggests that, in many cases, states emerging from civil war will find it challenging to be trusted as protectors of the physical integrity rights of their citizens. Maggie Beirne and Fionnuala Ní Aoláin (2009, 227) highlight this point by suggesting that peacebuilding efforts are highly complex as they are “processes of political transition in which the state is both responsible for or complicit in human rights violations and responsible for creating the means to hold and achieve accountability for such violations.” It is also the case that a focus on the contrasting degrees to which governments respect the protection of physical integrity rights is one of the most straightforward means of identifying differences in the relative balance of power between state and society following civil war. A balance of power that favors postwar governments is apparent if the state is largely unconstrained from attacks on physical integrity in forms such as extrajudicial killing and torture. Alternatively, a balance of power that favors society would be in evidence if the government proves to largely respect the principle that it must protect the physical integrity of its citizens. I follow the Cingranelli-Richards (CIRI) Human Rights Database in terms of identifying the types of harm that would constitute a violation of physical integrity rights. I thus take into account the degree to which post– civil war governments respect the principle that individuals have the right to be free of the following government actions: extrajudicial killings, torture, political imprisonment, and disappearances.

Post–Civil War Power Sharing: A Threat to Physical Security Rights? The adoption of power-sharing institutions at the end of civil war is a common strategy for managing conflict among former rivals. Andrew Finlay (2011, 1) notes that scholars have come to understand power sharing as the “‘dominant’ or ‘default’ response of the international community when it comes to conflict resolution.” This status as the preferred means of ending civil wars is further underscored by the fact that the United Nations now consistently includes a power-sharing expert among the members of the Department of Political Affairs’ Mediation Support Unit Standby Team (McCrudden and O’Leary 2013, 4).

Government Respect for the Physical Security

153

Critics of the adoption of power sharing have noted a range of concerns about employing these mechanisms following civil war. Studies have called into question whether these tactics actually serve to promote stability (Roeder and Rothchild 2005; Tull and Mehler 2005; Jarstad and Sisk 2008), have emphasized that aspects of power sharing are out of step with the principles of open and free democratic competition (Roeder and Rothchild 2005; Jarstad 2008), and have suggested that there is likely an association between the adoption of power-sharing institutions and the abuse of human rights (Jarstad and Sisk 2008; Finlay 2011). The research in this study is intended to address only the last of these claims: the suggestion of a causal relationship between the establishment of power-sharing structures and a failure to protect physical integrity rights in the aftermath of civil war. The basis for anticipating that such a relationship exists rests on four key arguments, which I consider in detail below. First, critics of the influence of power sharing on physical integrity rights have emphasized that power-sharing arrangements use group identities as the basis for distributing both state authority and access to resources. This results in power-sharing governments tending toward the unequal treatment of individuals, with rights and privileges contingent on whether a person holds membership in a favored group or not. The collectivities that tend to fare the worst in these arrangements are those that lacked a significant military presence during the war and are thus often unrepresented at the peace negotiations to establish power sharing (Jarstad 2008; Mehler 2009; Sriram and Zahar 2010). In recognition of the fact that power sharing allows for a degree of inequality, Paul R. Brass (1991, 334) asserts that such an arrangement “inevitably violates the rights of some groups and the rights of some individuals.”2 An example of this tension between power sharing’s emphasis on group identities and the principle of individual rights is the court case of Dervo Sejdic and Jakob Finci versus Bosnia and Herzegovina. These individuals were denied the opportunity to stand as candidates for elections, given that the 1995 power-sharing accords that ended Bosnia’s civil war mandated that state power be distributed exclusively among three groups: Bosnians, Serbs, and Croats. Sejdic’s Roma identity and Finci’s Jewish identity had the effect of barring them from elected office. A 2009 ruling by the European Court of Human Rights found that this aspect of the peace accords was discriminatory, and required that Bosnia bring an end to the practice of determining eligibility for office on the basis of ethnicity.3 What is the connection between the state’s unequal treatment of groups under power-sharing arrangements and violations of physical integrity rights? Given that access to state power and benefits is contingent on wartime identities, power-sharing structures tend to discourage individuals from building common ground or nurturing shared identities with their rivals. The result, critics of power sharing have asserted, is an environment

154

Matthew Hoddie

defined by the continuation of tensions between groups (Finlay 2011). At times, these tensions may be expressed through practices more consistent with a state at war rather than a country embracing an incipient peace, including violations of physical integrity rights through actions such as extrajudicial killings and torture. A second and related issue concerns the question of how efforts to monitor and protect the physical integrity rights of individuals will be carried out in the context of a power-sharing arrangement. During civil wars, rivals often highlight the human rights abuses of the opposing side(s) as a strategy intended to delegitimize the opposition and to motivate their own supporters (Beirne and Aoláin 2009, 212; Hannum 2009, 245). Some scholars have suggested that post–civil war power-sharing arrangements have the potential to perpetuate this practice as each group’s representatives in government and civil society organizations seek partisan advantage by pointing out the violations of physical integrity rights committed by rivals (Bell 2000, 299). Finlay (2011, 75–87) describes evidence that he interprets as confirming the partisan manipulation of human rights concerns in the cases of the postconflict environments of Northern Ireland and Bosnia and Herzegovina. In both instances, he reported that human rights organizations were linked to a particular group and were perceived as protecting the interests of only their own ethnic community.4 Finlay concludes by suggesting that whatever else they have achieved, the human rights processes provided for in the Good Friday Agreement and the Dayton Accords have not been successful. Rather than acting as a unifying force, human rights processes have been sources of on-going tension and pseudo-crisis. The old “liberal” critique of consociation which complains about the institutionalization of ethnic divisions might lead us to expect this: that human rights would become a new terrain upon which old communal conflicts are fought out. (86)

A further hazard associated with the partisan manipulation of human rights abuses is that the public and governmental response to these claimed attacks on physical integrity rights may prove more muted than if they were not tied to a particular group’s political agenda. The potential thus exists that abuses of physical integrity rights will be less thoroughly constrained in a context in which there exists public skepticism about the motivations that underlie accusations of human rights abuses. A third connection between power-sharing institutions and the potential for violations of physical integrity rights focuses on the individuals that claim authority through these arrangements. Scholars and activists concerned with human rights have noted that power sharing can at times place into government positions those individuals that are known to have engaged in atrocities during the civil war. This raises the possibility that these actors will continue to violate human rights once the war is over, now taking

Government Respect for the Physical Security

155

advantage of the authority they have at their disposal as leaders of the post– civil war state. An individual associated with wartime atrocities who was later included in a planned post–civil war power-sharing arrangement is Foday Sankoh, leader of Sierra Leone’s Revolutionary United Front (RUF). RUF fighters had a notorious reputation during the country’s civil war for terrorizing the civilian population through mass rapes and mutilations. Despite this history of human rights violations, the 1999 Lomé Peace Accord included provisions that granted the RUF significant powers. Sankoh himself was to be both vice president and chairman of a new Commission for the Management of Strategic Resources, National Reconstruction and Development that would oversee the country’s extensive gold and diamond resources (Melrose 2009, 136). With the peace accord signed, Sankoh obstructed its implementation and the peace process largely stalled until after his arrest in 2000 by UN forces. Human rights advocates believe that the adoption of these types of power-sharing arrangements at the end of the conflict have the potential to contribute to a sense of impunity among those who engage in wartime atrocities. Based on the same logic, they also oppose peace agreements that include promises of amnesty from prosecution for human rights violations taking place during the civil war; they instead favor trials and imprisonment for those guilty of abuses. The hope and expectation is that denying these individuals positions of power, and even threatening them with the possibility of prison, should serve to deter others from engaging in these types of actions (Sikkink 2011). In short, the claim is that the protection of human rights requires that individuals guilty of wartime atrocities should have seats in prison rather than seats in government. A fourth and final concern regarding the influence of power-sharing institutions on human rights centers on the potential silencing of public debate over divisive issues (Daalder 1974; Rothchild and Roeder 2005, 36– 37). As elite-centered pacts, the intention behind power-sharing arrangements is to construct political structures that address the central cleavage within a society. The leaders that craft these arrangements, as beneficiaries of the mechanisms of accommodation, have a vested interested in protecting these structures and insulating them from criticism. Efforts to protect a power-sharing pact from public denigration have the potential to take the form of antidemocratic behavior by elites. Hans Daalder (1974, 607) expresses this idea of elite protection of power-sharing arrangements as: “Strongly divided societies can be stabilized by a conscious effort on the part of political elites, provided they deliberately seek to counteract the immobilizing and destabilizing effects of cultural fragmentation. Patterns of inter-elite accommodation therefore form an independent variable that may impede and reverse the centrifugal forces at the level of the masses.” The

156

Matthew Hoddie

literature has been largely silent on the question of the strategies that elites might employ to impede and reverse actions that might challenge a powersharing arrangement. Yet it does not require a great leap of logic to anticipate that some of these actions would be contrary to the principles of protecting human rights. Possibilities include censorship and vocal critics being imprisoned or disappearing for statements and activities hostile to a power-sharing bargain. Taken together, these four arguments suggest that there should be a negative relationship between the adoption of power-sharing institutions following civil war and a state’s protection of physical integrity rights. To this point, however, there has not been a study that directly tests whether there is systematic evidence that supports this expectation. In what follows, I provide the details of tests intended to determine the nature of the relationship between power sharing and a government’s respect for physical integrity rights.

Research Design I employ a panel time series dataset to consider the effect that power sharing has on state respect for physical integrity rights. The analysis focuses on those countries that had fought and ended a civil war between the years 1984 and 2006. The unit of analysis is the post–civil war country year. I focus exclusively on those years following the end of a conflict during which the country remains at peace. If a country experienced more than one civil war, I analyze the years following the conclusion of each conflict episode. Key Explanatory Variables

The central explanatory variables for this study are the power-sharing mechanisms to which governments commit themselves in the aftermath of civil war. As I noted earlier in this chapter, I define post–civil war power sharing as comprised of possible four dimensions. These are the political, military, territorial, and economic bases of state authority. When making the determination concerning whether or not a particular form of power sharing is present or absent within a post–civil war state, the focus is on the initial public promise by those in authority to establish these types of institutions. In the context of a negotiated settlement, this commitment is most often apparent in the signing of a peace agreement with formal rivals. Following a military victory, power sharing may be indicated through the winning side’s promise to allow defeated group(s) to play a role in determining the trajectory of the postwar state.

Government Respect for the Physical Security

157

I operationalize the power-sharing indicator using different approaches, each providing an alternative means of taking into account its four potential dimensions. In some tests, I employ an aggregate measure of power sharing in the aftermath of civil war with values that range from 0 to 4. A score of 0 reflects the absence of power-sharing dimensions adopted following civil war; a score of 4 identifies a case of war termination in which all potential dimensions of power sharing are in evidence at the end of the conflict. Such an approach is in keeping with earlier research, which has reported that the aggregate number of power-sharing dimensions specified at a war’s end has a positive influence on the durability of the peace (Hartzell and Hoddie 2003, 2007). In other tests, I disaggregate the power-sharing measure into its different political, military, territorial, and economic dimensions in an effort to assess the degree to which a particular form of sharing power might influence physical integrity rights. These tests are thus concerned with determining if some forms of power sharing are more important than others in shaping a post–civil war state’s commitment to respect the physical integrity rights of individuals. Dependent Variables

To consider a state’s respect for physical security or physical integrity rights, I draw on data made available from the CIRI project. The data provide annual measures reflecting the degree to which a state respects the principle that its population should be free from government threats to physical integrity in the form of extrajudicial killing, torture, imprisonment for political reasons, and disappearances For each of these types of threat to physical integrity rights, an ordinal value is assigned varying between 0 and 2. A score of 0 indicates an absence of respect by a government for the principle that individuals have the right to be free from a particular act threatening to physical integrity. A value of 1 identifies a government with an inconsistent commitment to ensuring that the state is restrained from engaging in actions threatening to a particular form of physical integrity. Finally, a score of 2 indicates full respect by the government for the principle that individuals should be free of state behaviors that threaten the relevant form of physical integrity (Cingranelli and Richards 2010). Employing a separate dependent variable for each form of physical integrity rights provides a means of addressing the question of whether power sharing has an influence (either positive or negative) on particular types of human rights abuses, but not others. It thus accommodates the possibility that power sharing may have an influence, for example, over extrajudicial killing, but not disappearances.5

158

Matthew Hoddie

Control Variables

Power-sharing institutions are not the only factor with the potential to shape the capacity of post–civil war states to protect physical integrity rights. In recognition of this fact, I include a series of control variables in the analyses. I categorize these variables in terms of whether they are most strongly associated with one of the following categories: the concluded civil war, the domestic environment, or the global environment. Post–civil war conditions. The first category of control variables is a set

of indicators that represent the conditions linked to the war and its aftermath. For each factor, there are reasons to anticipate that it should be associated with a country’s protection of physical integrity rights. 1. Civil war intensity. I include a variable representing the intensity of the concluded civil war based on the expectation that people’s memories of how the war was conducted will serve to inform their behavior following conflict termination. I anticipate that civil wars that are tied to greater numbers of deaths will tend to generate higher levels of post-war animosity and distrust between wartime rivals. This has the potential to spill over into the postconflict environment in the form of violations of the physical integrity rights of individuals. I operationalize this indicator as the number of battle deaths that occurred in the context of the concluded conflict. I employ Hartzell and Hoddie’s data (2007) for this measure, which includes all terminated civil wars to the year 1999; I update the scores to the year 2006. 2. Peacekeeping. I expect that the presence of a peacekeeping operation in the aftermath of a conflict should have a positive effect on respect for these rights among post–civil war states. The presence of peacekeepers enhances the potential to monitor and protect against violations of physical integrity rights. It is also the case that peacekeepers often engage in efforts to reform state institutions, including the military and police forces so that their performance is consistent with the practices of established liberal democracies. Research by Amanda Murdie and David R. Davis (2010, 64–67) lends some support for this expectation, with their finding that peacekeeping operations with a humanitarian purpose tend to improve a state’s respect for physical integrity rights. I score this variable as a 1 for those countries to which peacekeeping missions were deployed at the end of a civil war and 0 otherwise. The data for this measure originates with the work of Virginia Page Fortna (2004), Michael Doyle and Nicholas Sambanis (2006), and the Uppsala Conflict Data Program/Peace Research Institute Oslo (UCDP/PRIO) Armed Conflict Database. 3. Time since last war termination (TSLWT). The final variable I include in the model with relevance to the concluded civil war is an indicator that rep-

Government Respect for the Physical Security

159

resents the passage of time since peace was achieved. It is operationalized as the number of years of peace up to 2006. This measure is included with the expectation that an ended civil war’s influence on physical integrity rights, whether it is positive or negative, has the potential to be shaped by the passage of time. This makes intuitive sense, as the significance of wartime bargains should become less important as peace proves more durable. Domestic environment. The second category of variables looks beyond

factors tied to the civil war itself and considers other domestic conditions that also have the potential to influence a country’s respect for physical integrity rights. 1. Democracy. There is a consensus among academics that regime type matters in shaping a given country’s level of respect for human rights. Democracies are understood to have lower rates of human rights abuses in comparison to nondemocratic states (Davenport 1999; Poe, Tate, and Keith 1999; Hafner-Burton and Tsutsui 2005; Cole and Ramirez 2013). To control for the influence of democracy achievement on physical integrity rights, I include an indicator for democracy. I draw this measure from the Polity IV dataset (Marshall and Jaggers 2011). The values for this variable range from –10 (identifying the most autocratic states) to +10 (for the most democratic countries). 2. Gross domestic product (GDP) per capita. Like democratic achievement, state wealth is strongly associated with a country’s record of human rights protection. Wealthy countries are considered to be more likely to respect the rights of their populations (N. J. Mitchell and McCormick 1988). To capture the role of wealth in shaping a state’s respect for physical integrity rights, I include in the models an indicator for GDP per capita, employing a log transformation. These data originated with the Penn World Tables (Heston, Summers, and Aten 2009). 3. Population. The last control variable with a focus on domestic conditions that I incorporate into the study reflects the size of a country’s population. I include this measure based on the suggestion that larger populations heighten the competition for a state’s scarce resources and are thus associated with a stronger potential for instability. This instability invites government repression in response (Keith 1999; Hafner-Burton and Tsutsui 2005; Carey, Gibney, and Poe 2010; Cole and Ramirez 2013). As with the indicator for country wealth, I collected data for this indicator from the Penn World Tables and employed a log transformation. Global environment. The final category of control variables I include in

the analysis are intended to take into account factors related to globalization that academics suggest have the potential to play a role in shaping a state’s posture concerning the physical integrity rights of individuals.

160

Matthew Hoddie

1. Economic globalization. I anticipate that the degree to which a state is integrated into the global economy will play a role in its level of respect for physical integrity rights. This is based on the expectation that countries that are more integrated into the global economy will be reluctant to engage in repression out of concern that it will signal instability and encourage the withdrawal of foreign interests (Richards, Gelleny, and Sacko 2001, 235). The measure of economic globalization I employed was drawn from the 2010 KOF Index of Globalization (Dreher 2006). The KOF Index scores economic globalization on the basis of both actual economic flows and existing restrictions on economic flows. The values for the indicator range between 0 and 100. 2. International nongovernmental organizations. Scholars have characterized INGOs as tending to have a positive influence on a state’s respect for human rights. INGOs have this influence by virtue of their ability to monitor a country’s human rights performance, drawing attention to evidence of violations, and mobilizing a domestic and international response to abuses (Keck and Sikkink 1998; Simmons 2009). The indicator I employ in the analyses for INGOs is the “number of INGOs that count at least one citizen or domestic organization as a member” (Cole and Ramirez 2013, 712); this measure is transformed by its natural log. The data for the indicator originates with the reference handbooks of the Union of International Associations.6 3. Convention Against Torture (CAT; CAT individual and CAT interstate). Research by Wade M. Cole found that states that are signatories to international human rights agreements tend to have stronger human rights records. Of particular relevance to this study, he notes that the signing of the Convention Against Torture has a positive influence on a given country’s protection of physical integrity rights (Cole 2012).7 I thus include two dichotomous indicators relevant to the CAT. A first variable reflects whether or not a country allowed complaints of CAT violations to originate with individuals; a second variable indicates whether or not a country allows these complaints to originate with other states (Cole and Ramirez 2013, 712). Each is scored with a 1 if the form of complaint was accepted and a 0 otherwise. 4. Trends in physical integrity rights (TPIR). The final variable consistent with the category of the global environment that I include in the tests concerns the global trend in respect among states for the physical integrity rights of individuals. This indicator is intended to account for the fact that changes in the protection of physical integrity rights may be a function of overall global trends rather than influences originating within a particular country. I measure the variable based on the annual lagged average CIRI score for physical integrity rights for all countries for which CIRI data are available.

Government Respect for the Physical Security

161

Method

With my focus on only the years following a civil war in which a country remained at peace, the panels in this study prove unbalanced and relatively small. These dataset characteristics present challenges for the use of fixedeffects analysis with nonlinear models (Cohen 2013). I therefore opt to use an ordered probit model with robust standard errors. I account for autocorrelation by including one-year lagged dichotomous variables for each of the categories of dependent variables, excluding the most repressive category of 0 from the analyses.8 I favor this approach, rather than the more common practice of adding a one-year lag of the dependent variable, given the nonlinear nature of CIRI’s measures of human rights (Hafner-Burton 2005; Peksen 2011). Following Dursun Peksen (2011), I also employ a one-year lag of the time-variant independent variables in the models to address the potential complication of simultaneity bias.

Results I present the initial tests of the relationship between post–civil war power sharing and physical integrity rights in Table 8.1. For these analyses, I test for an association between the aggregate measure of power sharing (varying between 0 and 4) and its influence on a government’s respect for the principle that its population should be free of threats in the form of extrajudicial killing, torture, political imprisonment, and disappearances. The results of these tests do not provide support for the view that power sharing serves to diminish a government’s respect for physical integrity rights. While there are negative coefficients for the aggregate measure of power sharing among three of the four tests, in each of these analyses this key explanatory variable fails to achieve statistical significance. This suggests that one cannot, with any degree of confidence, identify a causal relationship between the adoption of power-sharing mechanisms and a decline in a state’s protection of the forms of physical integrity rights that are the focus of this study. What factors prove more influential in shaping a government’s posture vis-à-vis physical integrity rights? In the context of these tests, the variables that appear to have an influence most consistently are the intensity of the concluded civil war, a state’s population size, and the number of INGOs present within the post–civil war state. The indicator representing the intensity of the civil war is negative in all four tests, and statistically significant for those results associated with the dependent variables of torture and disappearances. This suggests that relatively high levels of wartime violence may play a role in weakening the state’s commitment to the protection of physical integrity rights once the conflict has concluded.

162

Matthew Hoddie

Table 8.1 Post–Civil War Power Sharing and Government Respect for Physical Integrity Rights, Aggregate Measure, 1984–2006

Number of powersharing dimensions Civil war intensity Peacekeeping TSLWT Democracy GDP per capita Population Economic globalization INGOs CAT individual CAT interstate TPIR Log pseudo-likelihood Pseudo R-squared N

Extrajudicial Killing

Torture

Political Imprisonment

Disappearances

–0.018 (0.094) –0.083 (0.062) 0.011 (0.237) –0.025 (0.022) –0.018 (0.018) –0.248 (0.152) –0.505** (0.109) 0.006 (0.008) 0.623** (0.167) 0.084 (0.229) –0.223 (0.268) –0.57 (1.153) –221 0.328 310

0.125 (0.103) –0.185** (0.062) 0.028 (0.215) –0.026 (0.02) –0.017 (0.016) –0.095 (0.162) –0.333** (0.125) –0.014 (0.008) 0.457* (0.224) 0.414 (0.593) –0.001 (0.654) 0.126 (1.305) –182 0.207 310

–0.061 (0.11) –0.023 (0.055) 0.682* (0.3) –0.012 (0.02) 0.07** (0.019) –0.255 (0.201) –0.374** (0.097) 0.001 (0.007) 0.402 (0.205) –0.826** (0.33) 0.835** (0.408) –0.612 (0.749) –179 0.472 310

–0.006 (0.074) –0.113* (0.053) 0.054 (0.236) 0.056 (0.023) 0.01 (0.019) –0.083 (0.138) –0.471** (0.094) 0.001 (0.008) 0.454* (0.179) –0.076 (0.62) 0.011 (0.588) –0.006 (0.831) –178 0.342 307

Notes: Ordered probit model. Robust standard errors are in parentheses. All models include binary variables for each level of the dependent variable, lagged, excluding most repressive category (not shown due to space constraints). TSLWT, time since last war termination; GDP, gross domestic product; INGOs, international nongovernmental organizations; CAT, Convention Against Torture; TPIR, trends in physical integrity rights. **p < .01, *p < .05.

The measure of population size is negatively signed and significant in all four tests. In keeping with expectations, this indicates that countries with large populations tend to be associated with higher rates of both instability and government repression (Keith 1999; Hafner-Burton and Tsutsui 2005; Carey, Gibney, and Poe 2010; Cole and Ramirez 2013). Finally, the influence of INGOs within a post–civil war state appears to pull in the opposite direction from civil war violence and population size, with the variable positively signed and significant in each of the four tests. This result offers support for the claim that INGOs tend to improve the protection of physical integrity rights through their activities of monitoring and

Government Respect for the Physical Security

163

denouncing the human rights abuses that take place within states (Keck and Sikkink 1998; Simmons 2009). The second cluster of tests I present are intended to take into consideration the possibility that different dimensions of power sharing may have distinctive influences on a state’s respect for physical integrity rights. I thus include in the model separate independent variables for the sharing of power in the post–civil war state across its political, military, territorial, and economic dimensions. I present the results of these analyses in Table 8.2. These analyses again do not offer support for the expectation that there is a negative relationship between power sharing in the aftermath of civil war and a government’s level of commitment to the protection of physical integrity rights. In every instance, the variables representing the individual dimensions of power sharing across the four tests do not achieve statistical significance. Thus, it appears that power sharing is not an important factor in shaping a state’s respect for physical integrity rights, whether its dimensions are included in the model in aggregate form or individually. In keeping with the test results I present in Table 8.1, the same three control variables appear influential in these analyses. The measures of conflict intensity and population size are statistically significant and negatively signed in a number of tests, indicating that these factors tend to diminish protection of physical integrity rights. The indicator representing the number of INGOs within the post–civil war state is positively signed and appears as significant in three of the four tests. The sole exception is the test for a relationship between power-sharing mechanisms and the frequency of politically motivated disappearances. Taken together, the results shown in Tables 8.1 and 8.2 suggest the absence of any compelling evidence supportive of the claim that the adoption of power-sharing institutions in the aftermath of conflict has deleterious consequences for a state’s respect for the physical integrity rights of individuals. Rather than power sharing, a given post–civil war state’s human rights performance appears to be driven by factors largely unassociated with the specifics of efforts to resolve the conflict: conflict intensity, population size, and the presence of INGOs within the country.9 Summary and Some Speculation

The central finding of these analyses is that there is no consistent evidence of a relationship between the commitment to establish power-sharing institutions following civil war and a country’s subsequent commitment to the protection of physical integrity rights. What might explain this unanticipated result? One possible answer, although necessarily speculative given the absence of data that directly capture this dynamic, is that power sharing has the potential to introduce a degree of paralysis into the policymaking

164

Matthew Hoddie

Table 8.2 Post–Civil War Power Sharing and Government Respect for Physical Integrity Rights, Individual Measures,1984–2006

Political PS Military PS Territorial PS Economic PS Civil war Intensity Peacekeeping TSLWT Democracy GDP per capita Population Economic globalization INGOs CAT individual CAT interstate TPIR Log pseudo-likelihood Pseudo R-squared N

Extrajudicial Killing

Political Torture

Imprisonment

Disappearances

0.202 (0.311) –0.38 (0.289) 0.079 (0.212) 0.114 (0.293) –0.064 (0.056) –0.112 (0.213) –0.024 (0.022) –0.027 (0.02) –0.247 (0.163) –0.499** (0.1) 0.008 (0.01) 0.585** (0.188) 0.115 (0.289) –0.297 (0.423) –0.597 (1.184) –220 0.333 310

0.355 (0.266) –0.245 (0.21) –0.027 (0.301) 0.5 (0.389) –0.18** (0.055) –0.079 (0.271) –0.032 (0.023) –0.036 (0.024) –0.081 (0.163) –0.31** (0.116) –0.008 (0.008) 0.438* (0.178) 0.531 (0.574) –0.221 (0.653) –0.183 (1.345) –180 0.217 310

–0.23 (0.269) 0.417 (0.26) –0.431 (0.222) –0.028 (0.189) –0.086 (0.064) 0.933 (0.36) –0.016 (0.022) 0.078** (0.023) –0.198 (0.206) –0.364** (0.097) 0.003 (0.008) 0.491** (0.17) –0.648 (0.374) 0.569 (0.51) –0.942 (0.802) –176 0.481 310

–0.577 (0.312) 0.292 (0.268) 0.09 (0.216) 0.169 (0.282) –0.12* (0.051) 0.064 (0.234) 0.06* (0.026) 0.015 (0.023) –0.12 (0.15) –0.503** (0.112) –0.001 (0.008) 0.367 (0.207) –0.445 (0.771) 0.642 (0.843) 0.385 (1.305) –176 0.349 307

Notes: Ordered probit model. Robust standard errors are in parentheses. All models include binary variables for each level of the dependent variable, lagged, excluding most repressive category (not shown due to space constraints). PS, power sharing; TSLWT, time since last war termination; GDP, gross domestic product; INGOs, international nongovernmental organizations; CAT, Convention Against Torture; TPIR, trends in physical integrity rights. **p < .01, *p < .05.

process that makes it impossible for the postwar government to have an influence over either the strengthening or weakening of protections for physical integrity rights. The claim that political systems based on power sharing are particularly prone to deadlock among the government’s competing interests is not new. Research on power sharing has frequently noted that these institu-

Government Respect for the Physical Security

165

tional structures invite the potential for governmental inefficiency.10 The most commonly cited case of policymaking paralysis in the aftermath of civil war is Bosnia and Herzegovina’s power-sharing arrangement. The requirement that representatives of each of the country’s major ethnic groups support new policy initiatives has resulted in little in the way of legislative accomplishments. As a result, there has been growing evidence of dissatisfaction and unrest among the country’s population (Bilefsky 2014). The suggestion that political paralysis explains why there is little evidence of a relationship between power sharing and respect for physical integrity rights following civil war is not entirely satisfying. Not all of the arguments I previously identified for anticipating a negative relationship between power sharing and the protection of physical integrity rights focused on the identities of government actors or their behaviors. Notably, emphasizing a role for government paralysis does not explain away the potential for threats to physical integrity that are a function of reinforced divisions between groups or skepticism about claims of human rights violations. Noting a role for government deadlock does nevertheless offer at least a partial explanation for the apparent absence of a relationship between power sharing and a state’s performance vis-à-vis physical integrity rights.

Limitations Associated with the Reported Findings Admittedly, it is surprising that the results that I report in this chapter do not offer support for the prevailing view that the adoption of power-sharing institutions following civil war serves to diminish a government’s respect for physical integrity rights. As suggested earlier, this may be a function of the fact that power sharing introduces a form of paralysis into the policymaking process that makes it impossible for government elites to influence, either negatively or positively, the state’s level of commitment to the protection of physical integrity rights. It may also be the case that other factors are simply more influential in shaping a postwar government’s posture concerning this important issue. The common emphasis placed on power sharing’s role in postwar governance may have had the unfortunate effect of drawing scholarly attention away from the role played by such influential items as population size, conflict intensity, and the number of INGOs present within a country. At the same time, it should be acknowledged that there are some limitations associated with the quantitative analyses that suggest that the results should not be interpreted as definitive. First, given constraints associated with data availability, the period under consideration in my research is limited to civil wars that ended between 1984 and 2006. A further restriction on the analyses is that I include only those years in which these post–civil war states had been at peace. The result is that the number of cases available for

166

Matthew Hoddie

analysis stands at just over 300. The possibility thus exists that if a more robust number of cases had been available for study, it might have produced a more definitive finding concerning the relationship between power sharing and a government’s protection of different forms of physical integrity rights. A second, similarly problematic, aspect of the quantitative tests concerns the issue of how the variables for power sharing are operationalized. The power-sharing data that I employ reflect only the public commitments of elites to share power with their rivals rather than the implementation of these arrangements. This seems an appropriate initial approach for understanding the relationship between power sharing and physical integrity rights given the use of these same data to analyze power sharing’s capacity to promote postwar stability and democracy (Hartzell and Hoddie 2007, 2015). However, there is a growing body of research highlighting instances in which the power-sharing commitments of elites have not been consistently realized in practice.11 An important next step in this research agenda would thus be to determine if a relationship exists between power sharing and a state’s commitment to protect physical integrity rights when focusing the statistical analyses exclusively on those cases of implemented powersharing arrangements.12

Conclusion In a study of the relationship between electoral systems and the durability of peace among post–civil war states, Shaheen Mozaffar (2010) developed an argument in favor of what he describes as a context-sensitive approach. A defining feature of this approach is a view that electoral systems must be understood as embedded within a broader set of postwar circumstances that may equal or exceed voting rules in terms of their influence over the political future of a given state emerging from civil war. Among the other factors Mozaffar identifies as influential are the legacies of the civil war, the role of the international community, and the other institutional arrangements (e.g., the adoption of a parliamentary or presidential system) associated with the state. While the concerns of this chapter are in some ways different from those of Mozaffar’s study, the results I have presented serve as a further demonstration of the value of accounting for context in the study of post– civil war states. The findings of the statistical analyses indicate that once other aspects of the post–civil war environment are taken into account, the claim of a relationship between power sharing and declining state respect for physical integrity rights appears to lack merit. Instead, other conditions within post–civil war states proved to have a more consistent influence on

Government Respect for the Physical Security

167

the level of commitment to protect physical integrity rights. These include the intensity of the civil war, the size of the country’s population, and the number of INGOs with a presence within the state. Given that they are high-profile aspects of the effort to end civil wars, efforts at power sharing and peacekeeping often receive much of the credit or blame for how countries fare once they emerge from conflict. The results I presented in this chapter, however, suggest that it is also important to recognize that the influence of these factors over some aspects of the postwar state may actually prove quite limited. It thus appears that a key remaining challenge for studies of post–civil war states is to attempt to disentangle and identify where the influence of power-sharing institutions matters for shaping future power relations and where it does not.

Notes 1. See, for example, Hartzell and Hoddie (2003, 2007) and Roeder and Rothchild (2005). 2. Brass (1991), as quoted in McCrudden and O’Leary (2013, 11). To be precise, Brass’s quote criticizes the concept of consociational democracy. 3. The description of this case is drawn from McCrudden and O’Leary (2013), which makes the court’s ruling a centerpiece of their study. 4. Among the studies, Finlay (2011) cites in support of this perspective are Marcon et al. (2008) and Whitaker (2010). 5. I do not employ a dependent variable that aggregates these four separate measures of physical integrity rights into a single indicator. This does not prove practical for the models I employ as it introduces a degrees of freedom problem— more parameters specified than country clusters represented in the model. 6. While the data originated with these handbooks, the data was provided to me based on the work of Cole and Ramirez (2013). I thank Wade Cole for generously providing access to the INGO data. 7. This positive relationship may largely depend on the level of democratic achievement of the state signing the agreement. Hafner-Burton and Tsutsui (2005) found that signing human rights treaties had little positive influence on the human rights practices of the world’s most repressive states. 8. Autocorrelation is a concern when the error terms are not independent of one another. Given that I am analyzing time series data that includes multiple observations per country, the presence of autocorrelation needs to be addressed. Autocorrelation is an issue as it violates one of the key assumptions associated with most regression models. 9. To test the robustness of the analyses, I performed these same tests when including variables representing the means by which a civil war came to an end: negotiated settlement, negotiated truce, or military victory. For those tests that employ an aggregate measure of power sharing, the power-sharing variable consistently fails to attain statistical significance. For analyses that disaggregate powersharing into its individual components, there is some isolated, but far from systematic, support for the claim that power sharing may decrease a state’s commitment to the protection of physical integrity rights. In the context of these tests, there are two

168

Matthew Hoddie

significant findings linking a particular form of power sharing to a decline in human rights performance. The results identify an association between military power sharing and decreased protections against torture; they also suggest that political power sharing reduces a state’s respect for the principle that individuals should not be subject to disappearances. 10. See, for example, Rothchild and Roeder (2005, 39) and Hartzell and Hoddie (2007, 151). 11. See, for example, Hoddie and Hartzell (2003), Jarstad and Nilsson (2008), and Ottmann and Vüellers (2015). 12. A final objection that may be raised concerning the results of the quantitative analyses is the potential for endogeneity bias to influence the findings. The concern is that violations of physical integrity rights that occurred during the war simultaneously shape the adoption of power-sharing arrangements as well as a country’s postwar performance in protecting these rights. Such a scenario is not accounted for in the statistical tests. I am skeptical that this is a problem for my reported findings as there is no reason to anticipate that elites would respond to high rates of wartime human rights violations through an emphasis on power sharing. Given a consensus that power sharing is antithetical to the protection of human rights, it seems implausible that the creation of these institutions would be the preferred means of ensuring the protection of physical integrity rights.

9 Shifting Public Attitudes? Power Sharing and Intergroup Tolerance Bernadette C. Hayes and John Nagle

Consociational power sharing has evolved in the post–Cold War era into an instrument for resolving civil war and intrastate conflict as evident in notable arrangements in Lebanon, Macedonia, South Africa, Burundi, Bosnia, Northern Ireland, Kenya, Afghanistan, and Iraq. In distinction to earlier consociational pacts (e.g., Switzerland, Belgium, and Austria), the new wave of postwar agreements is no longer informal socalled gentleman’s pacts, but represents formal constitutionally bound frameworks that typically include the signatures of a number of states and the support of international institutions. As one leading theorist puts it, “Consociationalism has become the prescribed method of conflict regulation of the international community” (O’Leary 2005, 4). Despite its apparent ubiquity, power sharing invites vociferous debate and no little acrimony between proponents who claim that “consociational institutions offer a viable strategy to build peace, states, and democracy” (Wolff 2011, 1796) and opponents who maintain that it aggravates “the malady it is allegedly designed to treat” (Shapiro 1996, 102). Thus, while proponents hope that the process of elite-level bargaining and compromise will eventually soften antagonistic intergroup boundaries, opponents fear a reverse dynamic in which power sharing expedites a dangerous hardening of ethnic cleavages. Yet debates about these politically elite-driven approaches to conflict regulation cannot be divorced from existing public opinion. As the collapse of the earlier but ill-fated power-sharing experiment in Northern Ireland and the suggested attempt to include a Civic Forum in both Northern Ireland and Lebanon in the postsettlement period attest,1 the willingness of citizens to accept such power-sharing arrangements during and after these negotiations remains paramount. In fact, so great was the unease of the political elite in Northern Ireland concerning the views of their electorate during the Northern Ireland peace process that political polling was used throughout the process, 169

170

Bernadette C. Hayes and John Nagle

in which party negotiators were directly involved. As Stephen Tierney (2012) notes, this not only allowed politicians on both sides of the conflict to identify areas of compromise and common ground, but it also addressed one of the “big problems of elite negotiation in divided societies, namely the need to carry along the communities which political elites represent” (251). It is with these two competing expectations and accompanying caveat in mind that we focus in this chapter on public attitudes toward intermarriage between the main religious groupings in Northern Ireland and Lebanon. A key aim of this investigation was to examine the degree to which changing public attitudes may influence the dynamic of power relations within these two societies. More specifically, we wished to explore the relationship between public attitudes toward intermarriage in Northern Ireland and Lebanon and elite-level power relations. As noted earlier, a major principle underlying the application of consociationalism is that the successful regulation of ethnic cleavages via elite-level cooperation will eventually have “a beneficial impact upon societal ethnic rigidities, allowing differences to be managed peacefully and contributing to their eventual erosion” (Tonge 2009, 53). Yet, if public attitudes do show some evidence of amelioration under consociational frameworks, does this also reveal a correlative waning of the power of ethnic elites? If the control of ethnic elites largely rests on their capacity to instrumentalize their groups via the politics of fear and communal defense, then it could be argued that greater levels of intercommunal tolerance can be perceived as a threat to their grip on power. A comparative study of Northern Ireland and Lebanon in relation to this issue may be considered particularly appropriate for the following reasons. First, Lebanon and Northern Ireland represent two deeply divided societies based on ethnonational or religious cleavages that introduced consociational arrangements to stabilize the balance of power between the main protagonists to the conflict. Second, in both cases, the nature of these power-sharing arrangements was further complicated by the role of powerful external state benefactors who were deeply implicated in maintaining and, in the case of the Muslim community in Lebanon, further fracturing these communal divisions (Lustick 1993; Salamey 2014). Third, although the political arrangements in both societies are explicitly based on consociationalism, they also represent different forms of power sharing; namely, the liberal consociationalist approach in Northern Ireland versus the corporate form of consociationalism in Lebanon. Fourth, while there is much debate regarding the differing impact that each of these forms of consociationalism has had on communal divisions within these two societies, to date, empirical research in relation to this issue has been extremely limited and has tended to focus on elite-level political institutions rather than the views of ordinary citizens. The chapter proceeds in three stages. First, we outline the differing nature of the consociational settlements—liberal versus corporate—in both

Shifting Public Attitudes?

171

of these societies and discuss their suggested consequences for intercommunal relations. Second, using a range of nationally representative data, we examine the nature and extent of religious differences within these two societies.2 Finally, building on this analysis, we investigate religious differences in public attitudes toward what has long been considered the most salient indicator of communal division, interfaith marriages. A key purpose of this investigation was to examine the degree to which the differing political arrangements within the two societies have given rise to a divergence in attitudes in relation to this issue. The use of attitudes toward interfaith marriage as a key indicator of communal division may be considered justified for the following reasons. First, given the primacy of marriage as a key social institution in maintaining communal division, public attitudes on intermarriage can be viewed as a major bellwether of the impact of consociationalism (see Lijphart 1975). For example, social scientists have long considered the development of more tolerant attitudes toward intermarriage as strongly indicative that groups, once divided by violent antagonism, are now more willing to peacefully coexist (Smits 2010). Second, as consociationalism in many divided societies operates through the process of allocating citizens into clear ethnoreligious categories for the purposes of parity in political representation, the control of marriage as a mechanism to reproduce communal division is often perceived to be at the very foundation of the consociational system. This is particularly the case in Lebanon, where the religious control of marriage—all marriages in Lebanon are performed by a religious authority and, until very recently, civil marriages have been impossible—is viewed as the key factor in perpetuating and maintaining the sectarian balance of power within the political system (Larkin 2012; Cammett 2014).3 This is not to discount, however, the degree of variation in which these ethnic or religious identities may be embedded within power-sharing consociational arrangements (McGarry and O’Leary 2007; McCulloch 2014). Neither does it omit their varying consequences for the political acceptance of marital exogamy (marrying outside of one’s group) along religious and ethnic lines. For example, in Northern Ireland, where there is an official recognition of intermarriage, albeit with some public reservation particularly among the Protestant community, elites rarely, if ever, castigate intermarriage or threaten sanctions against those members who engage in the practice. In Lebanon, however, given that the corporate model is deployed to maintain communal boundaries at all costs, ethnoreligious leaders often use fiery rhetoric to censor those who support intermarriage.4 Finally, because consociational arrangements are, by their nature, simultaneously inclusionary and exclusionary—while the major groups in conflict are accommodated and their interests are privileged through power sharing, less powerful groups can find themselves marginalized—such processes of

172

Bernadette C. Hayes and John Nagle

inclusion and exclusion can have a differential impact on attitudes toward communal division within deeply divided societies. For example, while proponents of consociationalism point to the eventual development of a greater tolerance between the major groups accommodated within the power-sharing arrangement, critics of this approach suggest that this higher degree of acceptance between the major groups is accompanied by increased intolerance toward out-group members such as sexual minorities or minority groups that are not officially included within the remit of power sharing (see Nagel 2003; Taylor 2009). In summary, by exploring how different forms of power-sharing influence public attitudes toward intermarriage, we are given a crucial insight into how consociationalism can affect power relations in a divided society. On the one hand, consociationalism may provide a framework for stabilizing power relations and, thus, encourage intergroup tolerance to slowly emerge between communities previously engaged in sectarian violence. On the other hand, consociationalism can simultaneously act to disempower minority groups that are excluded from the official remit of power sharing and, thus, lead to their increasing marginalization and denigration within society.

Liberal and Corporate Consociationalism Consociational power sharing is based on the idea that conflict resolution in divided societies is achieved through the accommodation of the political elites representing the salient segments of society and institutionally anchored by inclusive coalitions and proportionality in public appointments (Andeweg 2000). By including equally the main groups in the divided society, consociationalism aims to transform a zero-sum game for political power into one that encourages consensus and sharing of power. By sharing power, the object is to equalize power relations in a context where conflict was formerly generated by an asymmetrical balance of power. In so doing, power sharing may be applied—as this book highlights—to the political, economic, territorial, and military domains. As Caroline A. Hartzell and Matthew Hoddie (2007) argue, power sharing is more likely to sustain peace in the duration if it is embedded across these four dimensions. While the military (including police) dimensions of power sharing are, in different ways, present in Lebanon and Northern Ireland,5 our focus in this chapter concerns the social consequences of political power sharing in these two societies as evidenced in public views toward interfaith marriage. As noted earlier, consociational power sharing was introduced in both Northern Ireland and Lebanon to redress and then stabilize the balance of power between the main groups in conflict. In Northern Ireland, the introduction of power sharing under the auspices of the 1998 Good Friday

Shifting Public Attitudes?

173

Agreement sought to give equal rights and recognition to nationalist and unionist identities within the polity. Irish nationalists, who were historically marginalized from political power in the region, were now given equal political representation with the hitherto dominant unionist majority. In Lebanon, the 1989 Taif Accord redistributed political power in a way so as to establish equality between Christians and Muslims. As such, the two accords—Good Friday and Taif—were based on the formula that the war had created neither winners nor losers and that peace was to be fostered through leveling power relations between the groups. Political power sharing rarely operates as a one-size-fits-all system (Lijphart 2004). To date, theorists (McGarry and O’Leary 2007; McCulloch 2014) have made a distinction between two types of consociationalism— corporate and liberal—based on their perceptions of the salience of ethnic or religious group identities—fixed versus malleable—as well as their consequent approach to governance—constitutionally determined versus dependent on electoral outcomes. As noted earlier, Northern Ireland and Lebanon are categorized as, respectively, representatives of corporate and liberal consociationalism. Table 9.1 illustrates the main differences between these two approaches in terms of the four key elements of their consociational arrangements, as specified in their respective agreements. Focusing initially on the element of grand coalition, a key distinction between the two societies is to be found in the manner in which they allocate positions within government. Under Lebanon’s corporate system, positions at the executive level are predetermined and reserved for the three main religious denominations—the presidency (Christian Maronite), the premiership (Sunni), and the speaker of the house (Shia)—via the so-called troika method of allocation. By contrast, Northern Ireland’s liberal consociational power-sharing structure seeks to ensure that the representation of the two main ethnonational groups—Protestant/Unionist and Catholic/Nationalist—emerges as a self-determined process. Hence, the premiership diarchy (first minister and deputy first minister), which heads the power-sharing executive, is not predetermined but rather appointed separately by the two largest parties elected to the Northern Ireland Assembly (NIA). Important differences also emerge when the element of proportionality— methods to guarantee the representation of the main groups in the legislature and public employment—is considered. Under Lebanon’s corporate system, a quota system is applied to guarantee the representation of the main religious groups6 in the Chamber of Deputies (the Lebanese parliament), with the Chamber’s 128 deputies divided equally between Christians and Muslims and proportionately among the denominations of each sect (Salamey 2014). The state’s electoral law operationalizes the quota system for the Chamber of Deputies. Up until the 2018 election, Lebanon had been divided into twentysix electoral qada (districts), wherein a quota of seats is reserved in each

174

Bernadette C. Hayes and John Nagle

Table 9.1 Consociational Forms of Governance in Northern Ireland and Lebanon

Type Foundation Form of governance Grand coalition

Northern Ireland

Lebanon

Liberal Belfast Agreement (1998)

Corporate Taif Agreement (1989)

Premiership diarchy voted in by Premiership troika predetermined the two largest parties in the by ethnoreligious affiliation Northern Ireland Assembly (NIA) Proportionality Assembly seats achieved via Parliamentary and ministerial seats electoral outcomes and ministerial distributed on a fifty-fifty quota positions allocated by the d’Hondt between Christians and Muslims. algorithm. The NIA’s 108 seats are All major public positions also elected from18 constituencies via subject to a fifty-fifty split. The single transferable vote (proportional 128 parliamentary seats are representation) elected from 26 districts via the majoritarian bloc vote, first-past-the-post system Mutual veto Vetoes restricted to key legislation Vetoes at executive (cabinet) and legislative level (parliament) Segmental autonomy De facto in selective areas such as Parallel civil institutions and separate schooling autonomy of control

district for specific sects based on what the Ministry of Interior estimates to be the sectarian demography. To ensure the representation of the main groups, the postwar electoral system is majoritarian bloc vote, first past the post, in which voters select as many candidates as there are seats allocated to their district. This system has had the effect of “hardening, rather than ameliorating, sectarian cleavages” (Salloukh 2006, 637–638). Appointments in the public sector are similarly based on the sectarian quota system. In Northern Ireland, by contrast, no quota system exists in the legislature and is applied extremely selectively to the public sector only in regard to policing. Seats for the Northern Ireland Assembly are allocated on the basis of party performance in elections. For the purpose of elections to the NIA, until the 2017 election, Northern Ireland had been divided into eighteen constituencies that each returned six assembly members elected via the single transferable vote (STV) system, a type of proportional representation (McEvoy 2008). Thus, while the Lebanese quota mechanism aims to maintain the “balance between the confessional elites in order to avoid conflict” (Salamey and Payne 2008, 463), Northern Ireland’s proportional system provides scope for transformation.7 Finally, there are some notable differences when the two remaining elements—veto powers and segmental autonomy—of consociationalism are considered. In Lebanon, the mutual veto system is deeply embedded in the executive and legislative branches to protect group interests, thus producing

Shifting Public Attitudes?

175

a system that is prone to policy deadlock. A two-thirds majority is required for important issues decided by the cabinet (e.g., international treaties and the budget), and for the parliament to make constitutional amendments (Tonge 2014). In Northern Ireland, by contrast, key legislation can only be passed either by weighted majority or by “parallel consent.” And while segmental autonomy is a key feature of Lebanese society, in that salient ethnoreligious groups are able to exercise legal authority over a range of personal matters—such as marriage, divorce, inheritance, and the custody or guardianship of children—via their own civil courts, it is largely lacking in Northern Ireland. In summary, then, although the political arrangements in both Northern Ireland and Lebanon are explicitly based on a consociational form of power sharing, they differ considerably in terms of their perceptions of the salience of ethnic or religious group identities as well as their consequent approach to governance. More so than any other factor, it is this difference in approach to accommodating group identities that has led some commentators to suggest that in comparison to their liberal counterparts, corporate forms of consociationalism are notably more likely to institutionalize and even exacerbate communal division (McGarry and O’Leary 2007; McCulloch 2014). In fact, the nomenclature of “liberal” to describe power sharing in Northern Ireland in distinction to the description of “corporate” for Lebanon leads to the argument that Northern Ireland’s structures have greater potential to disavow the role of ethnic identity in society. Proponents have argued that liberal consociationalism “is more likely to transform identities in the long run” (McGarry 2001, 124) and “provide a hospitable environment for the erosion of difference” (Coakley 2009, 145) than the corporate variant of consociationalism.

The Nature and Extent of Ethnoreligious Division Northern Ireland and Lebanon have been characterized as two fragmented and deeply divided societies along ethnonational or religious lines. Although the conflict in Northern Ireland is perceived as a bipolar religious conflict between a majority Protestant and minority Catholic population, it is generally agreed that the primary political cleavage is one of competing national identities, underpinned by religious affiliation (McGarry and O’Leary 1995). Thus, the conflict in Northern Ireland is best understood as a struggle between those who wish to remain part of the United Kingdom (Protestant/Unionist) and those who wish to see the unification of Ireland (Catholic/Nationalist). And while most commentators have agreed that religion is not the source of the conflict, it remains a key source of identity and acts as a crucial boundary marker in perpetuating these communal divisions

176

Bernadette C. Hayes and John Nagle

(C. Mitchell 2006), despite recent shifts to parity in the demographic composition of the two main religious communities in recent years.8 By contrast, the main lines of division in Lebanon are ethnoreligious or confessional, as it is commonly known in Lebanon. Consociational systems in Lebanon have sought to accommodate the main ethnoreligious sects in the state, which currently amount to eighteen distinct groups that are formally recognized. Although it is too simplistic to state that the cause of the 1975–1990 Lebanese civil war was purely religious, the main cleavage was Christian/Muslim, despite the presence of important political and ideological schisms within each of the two blocs. Even more so than in Northern Ireland, religious identity has been of overriding importance in defining the population and demarking communal division in Lebanon.9 As a number of commentators have noted, internal conflict resolution, renewal, and reform in Lebanon have all been determined by religious cleavages (Tonge 2014). And although recent surveys of the population have suggested that the religious composition of the two communities has now shifted to a Muslim majority,10 religion remains a key source of identity and the major cleavage within society. Thus, while Taif set out to regulate the traditional Muslim/Christian cleavage, since 2005 Lebanese politics aligns along a Sunni/Shia division and Christians are discordantly split between the Muslim blocs. Despite the reordering of historical cleavages, there has not been a transformation of the saliency of traditional ethnic identities. Indeed, “the sectarian players have . . . redefined their roles but the sectarian structure they are embedded in is still intact” (Knio 2005, 226).11 Recent survey estimates reflect the high rates of religious belonging as well as its changing composition in both of these societies (see Table 9.2). For the purposes of this analysis and due to the absence of suitable distinguishing data in Lebanon, only the two main religious groupings in both societies—Protestant/Catholic and Christian/Muslim—were included in the analysis. Focusing initially on Northern Ireland, the results in Table 9.2 are clear. Despite the subjective and voluntary nature of religious identification within this society,12 Northern Ireland continues to maintain a high rate of religious belonging. Not only did the vast majority of the adult population—or 85 percent in this instance—claim a Christian religious affiliation, but they are now almost equally divided between those who self-identified with either the Catholic or Protestant faith. This is not to deny, however, the significant proportion of the population who disavowed any religious identification, currently at around 14 percent. A similar bipolar division, albeit along Christian-Muslim lines, emerged when the religious identification of the Lebanese population was considered. However, at 56 percent of the adult population, it is now Muslims, and not Christians, who stand out as the majority group. Thus, extremely high levels of religious belonging characterize both societies, and religious identity remains pervasive in each.

Shifting Public Attitudes?

177

Table 9.2 Religious Affiliation in Northern Ireland and Lebanon (in percentage) Affiliation Christian Protestant Catholic Muslim Other No religion Total (N)

Northern Ireland 84.7 41.3 43.4 0.1 0.8 14.4 100.0 (1,180)

Lebanon 43.8 — — 56.2 — — 100.0 (1,386)

Sources: Northern Ireland Life and Times Survey, 2012; Arab Barometer, 2010–2011. Note: Questions: “Do you regard yourself as belonging to any religion?” (NI); “Religion” (LB). NI, Northern Ireland; LB, Lebanon.

It is important to note, however, that these high levels of religious affiliation are not merely nominal, but also translate into high levels of religious observance. As the data in Table 9.3 demonstrate, not only did a significant proportion of the population in both societies engage in formal religious practices such as daily prayer and weekly church attendance, but this was overwhelmingly the case in Lebanon. For example, whereas a staggering eight out of every ten adults in Lebanon claimed that they engage in prayer on a daily basis, the equivalent proportion in Northern Ireland was 49 percentage points lower at just 31 percent. Similar, though somewhat less marked, divisions occurred when church attendance rates and subjective religious views were considered. For example, while nearly two-thirds of respondents in Lebanon attended church on a weekly basis, the equivalent proportion in Northern Ireland was around half this amount, or just under a third. Finally, at least as far as church attendance patterns are concerned, there is evidence to suggest that there were also some marked differences between the two main religious groupings in both societies in relation to this issue. While members of the Catholic community stood out as the group with the greatest level of such formal religious practices in Northern Ireland, or 41 percent in this instance, it was the Christian community that emerged as the most religiously observant in Lebanon. In summary, then, both Lebanon and Northern Ireland remain deeply religious societies. In both societies, not only did the overwhelming majority of individuals claim a religious identity, but this religious identity also was accompanied by high levels of religious attachment. To what extent does this strong sense of religious belonging influence attitudes regarding interreligious marriages? Moreover, are there any notable differences between the two main religious communities—Protestant/Catholic and Christian/Muslim—in each society in relation to this matter? It is to an investigation of this issue that we now turn.

178

Bernadette C. Hayes and John Nagle

Table 9.3 Religious Affiliation, Behavior, and Conviction in Northern Ireland and Lebanon (percentages responding every day, yes, or mostly)

Pray (daily)

Quran/Bible (daily)

Church Attendance (weekly)

Religious Person

Northern Ireland Protestant Catholic All

30.6 31.9 31.1

— — —

26.4 40.5 32.4

17.1 12.0 14.9

Lebanon Christian Muslim All

78.9 80.8 80.0

64.1 67.6 66.1

68.4 58.7 63.0

46.5 42.4 44.2

Affiliation

Sources: Northern Ireland Life and Times Survey, 2008; Arab Barometer, 2010–2011. Note: Questions: “How often do you pray?” (NI); “Do you pray daily?” (LB); “Listen to or read the Quran/Bible?” (LB); “Apart from special occasions such as weddings, funerals, baptisms and so on, how often nowadays do you attend services or meetings connected with your religion?” (NI); “Attend Friday prayers/Sunday Services?” (LB); “Would you describe yourself as religious (extremely/very/somewhat) or non-religious (somewhat/very/extremely)?” (NI); “Generally speaking would you describe yourself as religious, somewhat religious, not religious?” (LB). NI, Northern Ireland; LB, Lebanon.

Attitudes Toward Interreligious Marriage Religious segregation is a long and established feature of life in Northern Ireland and Lebanon. For the most part, the various religious communities lead largely separate lives in a situation of “benign apartheid.” This is particularly the case in Lebanon where segregation occurs not only along religious lines, but is also further divided along confessional or denominational lines, otherwise known as sects (Khalaf 2012; Cammett 2014). In both societies the two communities are educated separately, segregated residentially and, to a somewhat lesser extent, in their place of work, and display extremely low levels of religious mobility (Nagle and Clancy 2010; Larkin 2012; Khalaf 2012; Hayes and McAllister 2013). However, perhaps nowhere is the segregation more prevalent than in kinship patterns or rates of marital homogamy. While the reasons for this finding differ between the two societies,13 it is commonly agreed that interfaith marriage remains extremely rare and to a large extent highly unacceptable in both of these societies, although there is some evidence to suggest that the traditional opposition to such unions, particularly among the Protestant population, has declined in recent years (see C. Mitchell 2006, 61). While no official figures are currently available because of the highly sensitive nature of religious identification in Lebanon,14 recent estimates from Northern Ireland suggest that only about one in ten of all marriages are now religiously mixed.

Shifting Public Attitudes?

179

The results in Figure 9.1 lend further empirical support to the findings regarding Northern Ireland. Although the number of marriages between people of the same religion has declined by 6 percent since 1968, single-faith marriages remain, by far, the dominant feature of marital life in Northern Ireland. A staggering 97 percent of individuals reported that their marital partner was of the same religion in 1968, or just before the most recent outbreak of the most recent phase of the conflict; by 1995, this had fallen to 91 percent. By the following decade, the rate of intra-community marriage declined by a further 6 percentage points to reach its lowest level of 85 percent in 2005 or exactly seven years after the ratification of the Belfast Agreement. Since then, however, rates of marital homogamy have increased steadily so that, by 2012, around nine out of ten of all adults in Northern Ireland has a husband, wife, or partner of the same religion as their own.15 To what extent do these high rates of marital homogamy coincide with negative attitudes toward religious intermarriage? Moreover, are there any differences between Northern Ireland and Lebanon in relation to this issue? Tables 9.4 and 9.5 address this question by focusing on religious differences in attitudes toward religious intermarriage in both of these societies. In Northern Ireland, attitudes toward religious intermarriage were assessed in terms of the degree to which respondents would “mind” or “not mind”

Figure 9.1 Levels of Marital Homogamy in Northern Ireland, 1968–2012

'%%#

!"# !"#$"%&' $"%&'

&"# &%# $"# $%# !"# '&($# '&$&# '&&)# '&&"# '&&$# *%%)# *%%"# *%%!# *%'%# *%'*# (")#'*+',-#."/' (")#'*+',-#."/' Sources: Northern Ireland Loyalty Survey, 1968; Northern Ireland Social Attitudes Survey, 1989, 1993, 1995; Northern Ireland Life and Times Survey, 1998, 2003, 2005, 2007, 2010, 2012.

180

Bernadette C. Hayes and John Nagle

Table 9.4 Attitudes Toward Religious Intermarriage in Northern Ireland (percentage)

“Would mind a lot” “Would mind a little” “Would not mind” (N)

Protestant

Catholic

9.9 18.9 71.2 (475)

4.0 7.8 88.2 (498)

Total 6.9 13.3 79.9 (973)

Source: Northern Ireland Life and Times Survey, 2012. Note: Question: “Would you personally mind or not mind if one of their close relatives were to marry someone of a different religion?”

if one of their close relatives were to marry someone of a different religion. In Lebanon, by contrast, the slightly different but still comparable question was the degree to which respondents viewed a person of a different religion or denomination as an obstacle to their acceptance of a marriage involving an immediate family member. Focusing initially on Northern Ireland (see Table 9.4), the results are clear. The vast majority of the Northern Ireland adult population, or around eight out of every ten individuals, claimed not to mind if a close relative were to marry someone of a different religion. By contrast, only 20 percent stated their opposition to this position. And while just 7 percent who adopted this negative stance said they would mind a lot, about double this amount, or 13 percent, endorsed the somewhat less negative view, claiming that they would mind just a little. This is not to deny, however, some differences between the two main religious groupings in relation to this issue, with Protestants being notably more negative in their response than Catholics. For

Table 9.5 Religious Affiliation and Obstacles to Interfaith Marriage Among Family Members in Lebanon (percentage)

Different religion/denomination: “Obstacle to a great extent” “Obstacle to a medium extent” “Obstacle to a limited extent” “No obstacle whatsoever” (N)

Christian

Muslim

All

58.3 11.2 6.4 24.1 (607)

60.4 12.7 9.0 17.9 (777)

59.5 12.1 7.9 20.6 (1,384)

Source: Arab Barometer, 2010–2011. Note: Question: “To what extent do you consider the following factors [different religion/denomination] obstacles to accepting your son/daughter/sister/brother’s marriage?”

Shifting Public Attitudes?

181

example, whereas nearly nine out of every ten Catholic adults, or 88 percent, claimed that they would not mind if a close relative was to marry someone of a different religion, the equivalent proportion among the Protestant community was markedly lower at 71 percent. A similar, albeit converse, pattern emerged when opposition toward intermarriage was considered; only 12 percent of Catholics as compared to 29 percent of Protestants endorsed this position. And while only 10 percent of Protestants said that they would mind a lot, just under double this amount—19 percent—claimed that they would mind a little, or agreed with the least negative view. Thus, at least as far as the Northern Ireland adult population is concerned, only a small minority of individuals were opposed to interfaith marriage, the vast majority of which were disproportionately drawn from the Protestant faith. The results in Figure 9.2 lend further empirical support to these findings. Although the number of individuals who were opposed—would mind a little or a lot—to interfaith marriage has declined markedly since 1989, Protestants remained notably less supportive of such unions than Catholics. While just over a third of all adults were opposed to such unions in 1989, by 2002 this had fallen to 25 percent. By the following decade, the rate of intracommunity marriage declined by a further 5 percentage points to reach its lowest level of 20 percent in 2012. While opposition to intermarriage

Figure 9.2 Attitudes Toward Religious Intermarriage in Northern Ireland, 1989–2012

Sources: Northern Ireland Social Attitudes Survey, 1989; Northern Ireland Life and Times Survey, 1998, 2002, 2005, 2008, 2010, 2012.

182

Bernadette C. Hayes and John Nagle

has declined across the two religious groupings, the decline has been greatest among Protestants. For example, in 1989, 47 percent of Protestants were opposed to interreligious marriage; by the mid-1990s, that proportion had declined to 39 percent. Since then, the percentage of individuals who oppose interfaith marriage has declined by almost an equivalent margin, so that currently only 29 percent of Protestants remained opposed to interfaith marriage. By contrast, opposition to interfaith marriage among Catholics has remained consistently lower and relatively stable over the same period, declining from just 17 percent in 1989 to 12 percent in 2010. Turning now to the question of attitudes toward interfaith marriage in Lebanon, we begin by assessing the degree to which members of the two main religious communities—Christian and Muslim—viewed individuals from a different religious denomination or affiliation as an obstacle to their acceptance of the marriage of an immediate family member. The results in Table 9.5 are conclusive. A staggering seven out of every ten adults, or 72 percent in this instance, viewed such unions as an obstacle to their acceptance to a greater (60 percent) or medium (12 percent) extent. The vast majority of the Lebanese population thus was opposed to religious intermarriages, and markedly more so than was the case in Northern Ireland. By contrast, just over a fifth of the adult population was willing to accept such interfaith unions unconditionally, while a further 8 percent would be willing to do so to a limited extent. Moreover, there were no major differences between the two main religious communities in terms of the strength of their opposition; 58 percent of Christians viewed such interfaith marriages as a “great” barrier to acceptance as compared to 60 percent of Muslims. Christians, however, were somewhat more likely than Muslims—24 percent as compared to 18 percent—to accept such unions unconditionally. To what extent are these patterns repeated when other forms of marital unions are considered? For example, is the increasing acceptance of interfaith marriages in Northern Ireland also reflected in a similarly positive view of other forms of unions—such as marriages between members of other religious or ethnic minority communities—within this society? Or, as discussed earlier, have such traditional antagonisms between the two main religious communities simply been redirected to other increasingly marginalized out-groups such as members of the gay community or other ethnic groups outside the body politic of the nation?16 Tables 9.6 and 9.7 address this question by focusing on religious differences in attitudes toward intermarriage across a range of other ethnic and social groups in Northern Ireland and Lebanon. In Northern Ireland, attitudes toward intermarriage regarding other social groups were assessed in terms of the degree to which respondents would be willing to accept by way of marriage a relative from a member of another ethnic minority group—or, in this case, individuals from within the Chinese or South Asian community—

183

Shifting Public Attitudes?

a member of the Muslim faith, or a member of the Irish Traveller community.17 In Lebanon, by contrast, respondents were asked whether they viewed a person of a different class or nationality as well as a person who does not pray as an obstacle to marriage with an immediate family member. Focusing initially on Northern Ireland, the results are conclusive (see Table 9.6). Irrespective of whether the Protestant or Catholic population was considered, there was considerable public opposition to intermarriage involving individuals from other ethnic or religious groups. Of these various groups, however, it was marriages with members of the Muslim or Irish Traveller community that stood out as the least publicly acceptable. For example, whereas just over two-fifths of respondents would not accept a marriage by a relative with a member of the Chinese or Asian community, the equivalent proportions in relation to Muslims or the Traveller community were significantly higher and almost identical at 56 percent and 57 percent, respectively.

Table 9.6 Religious Denomination and Attitudes Toward Intermarriage Among Other Ethnic and Religious Groups in Northern Ireland (percentage “not willing to accept”)

Other ethnic group Muslims Travellers

Protestant

Catholic

All

49.6 66.3 68.0

35.4 45.8 47.1

42.3 55.8 57.3

Source: Northern Ireland Life and Times Survey, 2012. Note: Questions: “Could you please indicate whether you agree with the following statements about [Irish Travellers; other ethnic groups such as Chinese or Asian; people belonging to the Islamic faith, that is, Muslim]: I would be willing to accept them as a relative by way of marriage to a close member of my family.”

Table 9.7 Religious Affiliation and Perceived Obstacles to Marriage Among Family Members from a Different Religion and from Other Social Groups in Lebanon (percentage who perceive obstacle to a “great extent”)

Different class Not praying Different nationality Different religion

Christian

Muslim

All

10.7 16.1 31.8 58.3

14.0 15.1 33.9 60.4

12.5 15.5 33.0 59.5

Source: Arab Barometer, 2010–2011. Note: Questions: “To what extent do you consider the following factors [different class/not praying/different nationality/different religion] obstacles to accepting your son/daughter/sister/ brother’s marriage?”

184

Bernadette C. Hayes and John Nagle

Marked differences also emerged between the two religious groupings in relation to this issue, with Protestants again notably more opposed to such marriages than Catholics. For example, whereas 68 percent of Protestants stated that they would not accept a marriage by a relative involving a member of the Traveller community, only 47 percent of Catholics endorsed this view. Thus, for Northern Ireland at least, the results suggest that the traditional enmity toward intermarriage between the Catholic and Protestant population may have given way to a similar and firmly entrenched antipathy toward intermarriage between members of other ethnic or religious groups outside the traditional Protestant-Catholic cleavage. A similar, albeit converse, pattern emerged when attitudes toward intermarriage among other social groups in Lebanon were considered (see Table 9.7). Here, intermarriage along religious or denominational lines remained, by far, the least acceptable form of union. While six out of every ten adults viewed such unions as an obstacle to a great extent, only one in three shared this view in relation to differences in nationality, and around one in seven in relation to any class difference or difference in religious observance. Moreover, and irrespective of the social group considered, there were no major differences between the two main religious communities in terms of their level of opposition. For example, whereas 32 percent of Christians viewed differences in nationality as a primary obstacle to marriage, the equivalent proportion within the Muslim community was almost identical at 34 percent. Thus, at least as far as the Lebanese community was concerned, it was differences in religious denomination that remained the key impediment to intermarriage. In summary, religious denomination remained the primary barrier to marriage in Lebanon and there were no differences between the two religious communities in relation to this issue. Irrespective of whether the Christian or Muslim population was considered, the vast majority of respondents remained opposed to these interfaith unions. This was not the case in Northern Ireland, where opposition to such marriages was not only significantly less than that toward other forms of union, such as those with a member of the Muslim or the Irish Traveller community, but also a clear minority view. This is not to discount, however, the significant differences between the two religious communities in relation to this issue, with Protestants being notably less supportive than Catholics in their views.18

Conclusion In this chapter, we addressed the question of how consociational institutions can alter the balance of power by assessing the relationship of two contrasting power-sharing forms—corporate and liberal—on public attitudes

Shifting Public Attitudes?

185

toward interfaith marriage in Lebanon and Northern Ireland. Ordinary citizens’ views on intermarriage are a strong indicator of levels of tolerance in divided societies, and these attitudes profoundly expose existing power relations between groups. Using differences in attitudes toward interfaith marriages as the basis for our analysis, we evaluated the claim that Lebanon’s corporate system would give rise to less favorable public attitudes in relation to this issue than those in Northern Ireland. The results lend much support to this view. Our arguments in support of this position are twofold. First, irrespective of whether the Christian or Muslim communities were considered, the vast majority of the adult population in Lebanon was opposed to such interfaith marriages, viewing them as an obstacle—either to a great or a medium extent—to their acceptance of a family member. Furthermore, there were no major differences between the two main religious communities in terms of the strength of their rejection, and this finding remained even when a range of control variables were included in the investigation. This was not the case, however, in Northern Ireland, where not only was opposition to such marriages a clear minority view, but there also were marked differences between the two main religious communities in relation to this issue, with Protestants being notably less accepting of such marriages than Catholics. Second, there is also evidence to suggest that not only was intolerance toward interfaith marriages notably greater in Lebanon than in Northern Ireland, but it was considerably higher than intermarriage across other social groups. Thus, at least as far as Lebanon is concerned, it was intermarriage along religious or denominational lines that remained, by far, the least acceptable form of marital union. A similar, though converse, pattern emerged when attitudes toward such interfaith unions in Northern Ireland were considered. Irrespective of whether the Protestant or Catholic population was considered, it was support for interfaith marriages between the two main religious communities, and not unions across other ethnic or religious groups, that stood out as the most endorsed view. What may explain these cross-national differences in our findings? In other words, why are Lebanese adults notably more likely to disapprove of such interfaith marriages than their Northern Ireland counterparts? Part of the explanation, we suggest, rests with the differing perceptions concerning the salience of ethnic or religious group identities in these societies and the subsequent nature of their representation. As noted earlier, although the political arrangements in Northern Ireland and Lebanon address the identitybased nature of the conflict via a consociational form of power sharing, they differ considerably in terms of their perceptions of the rigidity of such identities as well as their legal accommodation in political and social life. While both societies recognize officially that these identities form the basis for

186

Bernadette C. Hayes and John Nagle

electoral office, only in Lebanon are such identities viewed as immutable and thus predetermined at the executive, legislative, and public sector levels. In fact, some commentators have gone so far as to suggest that not only has the predetermined confessional basis of politics further exacerbated these divisions, but it also has perpetuated a small and intergenerational political sectarian elite that lacks any national accountability and simply acts as a representative of their particular religious community, or sect, rather than of the nation as a whole (Salamey 2014). Moreover, unlike Northern Ireland, this confessional rigidity in political representation is also accompanied by a high degree of segmentation, such as legal autonomy over the educational system as well as a range of personal matters—including marriage, child custody, inheritance and divorce—along confessional lines. As noted earlier, legal autonomy is more far-reaching in Lebanon than in Northern Ireland, and all marriages in Lebanon are performed exclusively by a religious authority. It is this factor— the rigidity in identities as well as their widespread legal accommodation in the social as well as the political life of the country—that we suggest sets Lebanon apart and explains its citizens’ greater antipathy toward intermarriage along religious lines. By influencing levels of intergroup tolerance, corporate and liberal consociations also have a direct impact on power relations between groups in Northern Ireland and Lebanon. Consociationalism, as noted earlier, is purposely designed to transform politics in divided societies from a zerosum conflict over power into a dynamic characterized by consensus and equal sharing of power. In so doing, the hope is that the culture of elitelevel political cooperation will eventually translate into greater levels of tolerance required for peaceful coexistence. Yet, as we showed in this chapter, liberal and corporate consociations can reorient power relations in decidedly contrasting and complex ways. In fossilizing ethnic identities, Lebanon’s corporate model has a correlative impact on power relations. The fact that levels of tolerance for intermarriage remain low in Lebanon can be read as symptomatic of an institutional system that steadfastly aims to preserve the balance of power between groups at all costs, for fear that any alteration will violently destabilize society. Northern Ireland’s liberal system also seeks to maintain the balance of power between the main protagonists, but its greater flexibility has allowed space for increasing tolerance of interfaith marriage between Protestants and Catholics in this society. Yet, while liberal consociations can be beneficial for those groups that are officially included and empowered, a different picture is evident for those groups that lie outside the boundaries of institutional recognition. In Lebanon, tolerance for intermarriage between members of the salient and nonsalient groups, while low, is higher than that recorded between members

Shifting Public Attitudes?

187

of the salient cleavages. It can be argued, therefore, that the low level of trust between the main groups accommodated in Lebanon’s corporate system ensures that antagonism between the main sects sharing power remains high. In Northern Ireland, alternatively, we note that tolerance for intermarriage significantly declines when it involves the attitudes of members of the salient groups included within power sharing toward some nonsalient groups such as migrants and Irish Travellers. Liberal consociations, therefore, can be seen as expediting two interrelated dynamics that simultaneously privilege and marginalize different groups: those that formally share power and those that are omitted. It could nevertheless be argued that liberal and corporate forms do not directly lead to differing outcomes regarding intergroup tolerance. Rather, and as noted in our earlier caveat, a reciprocal or even reverse causality could be hypothesized in light of the influence of public opinion on intergroup relations. In other words, liberal and corporate institutions may merely reflect the existing levels of intergroup tolerance in these societies prior to the formation of consociational structures. However, our argument against this proposition in terms of our present study is twofold. First, there has been a significant deterioration of intercommunal relations in Lebanon since the reintroduction of corporate consociationalism in 1989. For example, scholars have drawn attention to the “impoverishment of public life” as well as the “erosion of civility” after the Taif Agreement. As Samir Khalaf (2012, 4) puts it, “The modicum of pluralism the country once enjoyed is now generating large residues of paranoia, hostility and differential bonding.” Central to this outcome is Lebanon’s power-sharing structure, which has been accused of institutionalizing and exacerbating communal divisions (see Salamey 2014). In Northern Ireland, by contrast, although the liberal agreement appeared at first to reward the so-called political hardliners engaging in “ethnic outbidding,” research has convincingly shown that these leaders have moderated their policies once in power (P. Mitchell, Evans, and O’Leary, 2009). This is not to suggest, however, that Northern Ireland’s power sharing is an unmitigated success story unblemished by interethnic tensions. For example, at the communal level, Northern Ireland continues to be plagued by segregation and ethnonational disputes over symbols as group rights become war by other means. It is with these observations in mind that future research should not only investigate the influence of different consociational forms of governance on public attitudes, but also their impact on groups that remain on the periphery of power sharing. A key focus of this investigation should be to examine the degree to which changing public attitudes toward intergroup relations are either derivative of these power-sharing arrangements or are dependent on and constrained by public opinion prior to, as well as during, the implementation of the power-sharing arrangements. Given the absence

188

Bernadette C. Hayes and John Nagle

of previous research in this area as well as the assumed orthodoxy of the top-down or elite-driven institutional determining model, it is to an investigation of this issue—the possible interaction between the institutional influence of power sharing on popular opinion versus the influence of the prior views of citizens—that future research should be directed. Only via such an investigation can the nature of the interaction between these two factors—the top-down influence of institutional designs initiated by political elites versus the bottom-up role of public opinion—be understood in terms of determining either the success or failure in resolving ethnic divisions in transitional societies emerging from conflict. Notes 1. While both peace accords—the 1998 Belfast Agreement in Northern Ireland and the 1989 Taif Accord in Lebanon—advocated the establishment of a Civic Forum to promote deliberative democracy, only in Northern Ireland was the institution established. Designed as a consultative body detached from the political party system and composed of representatives from organizations in civil society, the forum was launched in 2000, albeit with considerable hostility, most notably from Democratic Unionists. With the dissolution of the Northern Ireland Assembly in 2002, the Civic Forum was suspended but, unlike the assembly, was not recalled with the restoration of devolved powers in 2007. 2. In Northern Ireland, the two datasets used primarily in the analysis were the Northern Ireland Life and Times Survey of 2008 and 2012. The dataset for Lebanon was the Arab Barometer of 2010–2012. All three datasets are representative of their respective adult populations and are publicly available at the websites: www.ark .ac.uk (Northern Ireland data) and www.arabbarometer.org/ (Arab Barometer data). 3. Until recently, all marriages in Lebanon were performed exclusively by a religious authority, or by a member of one of Lebanon’s eighteen officially recognized sects, and civil marriages, including intermarriage, were legally impossible. As Haddad (2003, 99) notes, “Individuals who wished to marry outside their religious faith had either to convert or become a religious ‘dissident.’” Lebanese couples wanting a civil marriage were forced to marry abroad, although such marriages were subsequently officially recognized by the Lebanese state. However, legislative changes in 2009 that allowed Lebanese citizens to remove religious identity from all state records as well as their identity cards paved the way for the introduction of civil marriage in Lebanon. The first civil marriage in Lebanon—between a Sunni and a Shiite—took place on October 10, 2012. 4. For example, the religious leader of Lebanon’s Sunnis issued a fatwa against any Sunni politician “trying to sow the bacteria of civil marriage in Lebanon” (NOW 2013). 5. In Lebanon, power sharing is applied to the military and the police service. The position of commander of the Lebanese armed forces is reserved for a Maronite Christian. In fact, two commanders of the army have become president (Haddad 2009, 413). The position of head of the Internal Security Forces (ISF)—the police service—is reserved for Sunnis. In Northern Ireland, power sharing was applied to the Police Service Northern Ireland (PSNI), which operated a fifty-fifty Catholic/ Protestant recruitment policy between 2001 and 2011.

Shifting Public Attitudes?

189

6. Although the Lebanese constitution officially recognizes eighteen religious groups, only six of them are of major demographic importance, three of which are Muslim (Sunnis, Shiites, and Druze) and three of which are Christian (Maronites, Greek Orthodox, and Greek Catholics). 7. It should be noted that the power-sharing arrangements for both Lebanon and Northern Ireland have undergone some key reforms. The number of seats for the Northern Ireland Assembly has been reduced to ninety, while Lebanon’s 2018 election used proportional representation to elect deputies from fifteen districts. 8. According to the most recent 2011 census, the Northern Ireland population is now almost evenly divided by religious identity, with 40.7 percent describing themselves as Catholic and 41.5 percent claiming to belong to either a Protestant denomination (35 percent) or another Christian faith (5.8 percent). 9. In contrast to Northern Ireland, not only is religious identity in Lebanon considered an ascribed status, sanctioned by the state and assigned at birth, but the use of religion as a communal marker of identity was so pervasive that until recently religious affiliation was also encoded on all national identity cards and official registry documents. In 2009, however, following a directive from the minister of interior, citizens are no longer required to have their religious affiliation encoded on their identity cards or official documents and can have it either removed or changed if they so wish. 10. Because of the extremely sensitive political nature of religious identification in Lebanon, there are no definite figures concerning its distribution within the population. The last census, which was conducted in 1932, reported the Christian population at 52 percent and Muslims at 48 percent. However, more recent estimates, based on a variety of demographic and political sources, suggest that of the three major groupings, which make up at least 68 percent of the total population of citizens, Christian Maronites comprise 22.5 percent, while each of the two largest Muslim sects—Shiites and Sunnis—are believed to be evenly balanced at 28.1 percent each (see Faour 2007). 11. This was a difference that found political expression with the establishment of the March 8 (pro-Syrian) and March 14 (anti-Syrian) coalitions subsequent to the Cedar Revolution, triggered by the assassination of former prime minister Rafik Hariri on February 14, 2005. 12. Although the Christian churches in Northern Ireland have an entry rite of baptism, they have no formal mechanisms of membership renewal or registration. Moreover, in direct contrast to Lebanon as well as in some continental European countries, the state plays no role in identifying or recording the church membership of citizens. Thus, whether or not one belongs to a particular religious denomination in Northern Ireland is as much a matter of subjective perception and identification as of externally observable behavior (Fahey, Hayes, and Sinnott 2006). 13. Examples include the legally devolved nature of personal status matters such as marriage to religious authorities in Lebanon versus the traditional demand by the Catholic Church in Northern Ireland and elsewhere that all children from these unions be brought up in the Catholic faith. 14. Previous estimates from a survey conducted in 1970, based exclusively on the female population, suggest that such marriages are negligible in Lebanon, with only about 1 percent of couples marrying outside their faith, and the vast majority marrying family members, or “cousins,” within their own sect (Chamie 1980, 178). More recent research, however, has suggested a notable decline in such consanguineous marriages, although the vast majority of Lebanese still marry within their own faith (Barbour and Salameh 2009).

190

Bernadette C. Hayes and John Nagle

15. This is not an unexpected finding given public perceptions of the deteriorating nature of relations between the two main religious communities—Catholics and Protestants—since the immediate postagreement period. This phenomenon, it should be noted, was most marked among the young (Hayes and McAllister 2013). 16. In fact, so great was the upsurge in racial and homophobic incidents in the immediate postagreement period that Northern Ireland became known as the “race hate capital of Europe” (see Jarman and Monaghan 2003). 17. According to the most recent 2011 census, 1.8 percent of the resident population of Northern Ireland belongs to a minority ethnic group. The main groups are Chinese, Indian, mixed, and other Asian, each accounting for around 0.3 percent of the resident population. Irish Travellers is one of the smallest and most-discriminatedagainst ethnic groups, accounting for 1,300 individuals, or only 0.1 percent of the population. 18. Multivariate analyses lend further support to this view. Even when a range of sociodemographic background variables as well as religious conviction and attitudes toward marriage among other out-groups were included in the analysis, Protestants were significantly less supportive of religious intermarriage than Catholics in Northern Ireland; there was no difference between the two religious communities in Lebanon.

10 Transitional Justice: Promoting or Hijacking Elite Accountability? Chandra Lekha Sriram

Traditional power-sharing literature has focused on arrangements in divided, but democratic, societies and, in the past two decades, has expanded to encompass peace agreements in countries that have experienced severe violence and often have little recent experience of democracy. The more recent academic literature has, alongside gray literature and the concerns of practitioners, focused on whether power-sharing peace agreements help to promote stable peace and, ideally, stable and durable states. Recent literature critical of power-sharing arrangements has identified a range of weaknesses, including their failure to promote durable peace, their failure to attend to the needs of the populations affected, challenges of their implementation, and their potential to promote impunity and undermine transitional justice (Mehler 2009; Sriram 2008; Vandeginste and Sriram 2011). In line with the aims of this volume, which are to attend to the relationship between power sharing and the transformation of power relations within a state, in this chapter I ask a question not previously examined in the power-sharing or transitional justice literatures: How if at all do transitional justice measures impact on power relations in a transitional or postviolence state, and can any specific impact on, or interactive effect with, power-sharing measures be discerned? This is a complex set of possible relationships, not easy to simplify or disaggregate, given that both power sharing and transitional justice seek to, and are expected to, alter power relations in a transitional or post-transition state. In this chapter, I examine common presumptions regarding the effects of transitional justice on political power relations and in particular the expectation that the prosecution of top state or rebel officials alters power relations by removing them from power directly, or altering or eliminating their political leverage. Following the approach taken in this book, I treat power relations as the relative strength or influence, or changes in relative strength and influence, 191

192

Chandra Lekha Sriram

of key political and military players, whether formerly state or nonstate, in new or postsettlement institutions of state. This involves not only the relative power of state and rebel or formerly rebel forces or leaders, but where relevant also involves relative power of government and opposition political parties or leaders, or a multiplicity of political parties or leaders. The focus is thus explicitly on elite leaders, not power relations across the state and society. For reasons discussed in the Methodology and Case Selection section, in this chapter I examine only countries in which international or internationalized criminal tribunals have been operational contemporaneously with power-sharing arrangements. I investigate a series of hypothetical causal claims through the review of experiences of states in Africa in which prosecutions of top actors have been pursued or have occurred and power-sharing agreements have been reached.

Power Sharing as Peace Promotion My focus in this chapter is not on the vigorous debates in the literature and among practitioners regarding the durability of peace agreements and the utility of different types of provisions, including power sharing (Hartzell 1999; Stedman, Rothchild, and Cousens 2002; Hoddie and Hartzell 2003; Roeder and Rothchild 2005; Hartzell and Hoddie 2007; Sriram 2008; Brown and Zahar 2008; Mehler 2009; ACCORD 2009). Such provisions, which may include political, economic, territorial, and military modalities, have been implemented in the hope that they provide strong incentives for former combatants to remain committed to a peace agreement and prevent a resurgence of conflict. They have also been used in settlements of violence short of civil war, as with the national accord in Kenya following postelection violence in 2007–2008. In particular, policymakers have expected power sharing to reassure weaker parties or manage ethnic tensions and rivalries at the end of civil wars (Rothchild 2005). Numerous peace agreements have utilized at least one type of power sharing and many involve more than one; the Comprehensive Peace Agreement for Sudan in 2005 included all four dimensions of power sharing. While advocates argue that such measures can be critical for durable peace agreements, critics argue that they are not sustainable or involve problematic compromises and may reward criminals.

Power Sharing as Undermining Accountability or Transitional Justice In this chapter, I also do not revisit in detail the arguments about the ways in which power-sharing agreements may undermine transitional justice or

Transitional Justice

193

limit the possibilities for accountability more generally. As a coauthor and I have argued elsewhere, transitional justice measures may sometimes be hampered, not by explicit amnesties, but by power-sharing arrangements that provide positions in governance for former fighting factions that may enable them to shield themselves from accountability measures. This, we suggested, might in some cases enable implicit, rather than explicit, amnesty, which may not be permanent but can endure for some time (Vandeginste and Sriram 2011). Rather than considering the effects of power sharing on transitional justice per se, I consider the reverse, that is to say, the possible effects of transitional justice measures on power-sharing agreements and in reshaping power relations in states emerging from violence and civil war. It should be noted that because power sharing and transitional justice measures likely have reciprocal and interactive effects, this makes a simple unidirectional causal analysis problematic.

Power Sharing as Reshaping Power Relations Power-sharing arrangements are not simply expected to provide inducements to warring parties to agree to peace and to implement agreements. They also are designed to reshape institutions of state, and the relationships among those institutions and elite leaders within them. As is discussed in the introduction (Chapter 1) of this volume, this is about more than the four dimensions of power sharing: economic, political, military, and territorial. It is also about inclusion and recognition, particularly of rebel groups that not only may have been excluded from power but also may be considered illegal. Their inclusion, whether in political governance structures, control over the economy, restructured security forces, or over territorial units, shifts different types of power to elites within those groups and away from some traditional elites. It can also trigger shifts in authority within traditional elites or within newly incorporated rebel groups as Andreas Mehler and Caroline A. Hartzell point out in Chapter 1. Since the introduction and other chapters in this book examine these shifts in greater detail, I focus here on the effects of transitional justice in shaping power relations in postconflict states with and without power-sharing arrangements.

Transitional Justice as Reshaping Power Relations Transitional justice is a vast field of practice with a burgeoning scholarly literature analyzing it, which cannot be dealt with in any detail here. However, transitional justice generally involves a range of judicial and nonjudicial measures to address past abuses in countries emerging from violent

194

Chandra Lekha Sriram

conflict or engaged in transitions to democracy (United Nations 2010; Teitel 2000; Mani 2002). These can include trials (whether national or international), commissions of inquiry, vetting or lustration, memorials or reparations, and amnesties, and increasingly they are expected to be victimfocused and may include measures of “traditional” justice (Sriram 2004; Sriram et al. 2013). Given the range of measures, it should not be surprising that a variety of effects on a post-transition state may be ascribed to them. I examine some of these claims briefly. Many have been theoretical, although a more critical recent literature has interrogated them in largely single-country studies. Transitional justice measures not only seek to address past harms inflicted on victims, but also frequently to promote reconciliation, institutional reform, and address the underlying causes or grievances that may have underpinned past conflict and abuses. As such, these not only are legal practices, but also eminently political ones, which often seek directly to alter the state and political practices, or which might be expected to do so indirectly. The claims made for the transformative capacities of transitional justice are sometimes breathtakingly broad, and until recently have been examined largely through single- or small-n comparative empirical research (Thoms, Ron, and Paris 2008). More recently, quantitative research has also sought to examine the correlation between transitional justice measures and democracy or human rights records of states (Wiebelhaus-Brahm 2010; Olsen, Payne, and Reiter 2010; Sikkink 2011). However, while a growing literature has been examining whether transitional justice measures cause, or are at least correlated with, improved human rights or democracy records, the specific pathways of causal change have been less specified. Less frequently still have scholars made claims about, or sought to explain, how transitional justice measures can alter power relations in a state. Nonetheless, there have been many implicit claims and some critical arguments, which I survey before turning to my examination. A leading argument for transitional justice generally, and international criminal accountability specifically, has historically been deterrence. This argument was made most forcefully in a seminal article by Payam Akhavan, former legal adviser to the Office of the Prosecutor of the International Criminal Tribunal for the Former Yugoslavia (Akhavan 2001). He argues that contrary to claims that the conflicts in the former Yugoslavia and elsewhere were driven by primordial ethnic hatreds and thus impossible to prevent, they were the result of decisions by politicians and warlords. Akhavan contends that, given that the violence was based on elite calculations, international criminal justice that targeted leaders could send an important message about the cost of using “hatred and ethnic violence as an instrument of power” (8). Akhavan’s argument also incorporates a separate claim, one of deterrence or prevention of atrocities through the removal of the leaders foment-

Transitional Justice

195

ing them. The latter claim depends on the prosecution of top leaders, not mid-level or rank-and-file accused. Human Rights Watch significantly elaborated on this argument in a later report (Human Rights Watch 2009). The international nongovernmental organization (INGO) argues that rather than impeding peace negotiations, the indictment of certain leaders can help promote agreements and enable the marginalization and eventual removal of such leaders, suggesting that this was the case with former Yugoslav president Slobodan Milosevic (Human Rights Watch 2009). This argument finds support in the work of Rachel Kerr, who argues that the indictments of Radovan Karadzic and Ratko Mladic helped to enable the Dayton peace negotiations insofar as they excluded these individuals from the talks and encouraged Muslim representatives to attend (Kerr 2004). Richard Holbrooke argues that the “tribunal served as a valuable instrument of policy that allowed us to bar Karadzic and other indicted war criminals from public office” (Holbrooke 1998, 190). Human Rights Watch similarly followed the INGO International Center for Transitional Justice’s claim that in Liberia, President Charles Taylor’s agreement to step down from office and not seek a role in a transitional government as part of Liberian peace negotiations was a direct result of the unsealing of his indictment by the Special Court for Sierra Leone (SCSL) (Human Rights Watch 2009). The organization also suggests that the arrest warrants for leaders of the Lord’s Resistance Army (LRA) in northern Uganda may have helped to isolate the organization and weaken support from international sponsors, such as the government of Sudan, as well as helping to push the organization to the negotiating table in 2006 (Human Rights Watch 2009), although it appears that Sudan has subsequently renewed its support for the LRA. Advocates of transitional justice have often argued that it helps to reshape political structures and democratize them. They frequently have contended that trials, among other measures, help to promote rule of law through demonstration effects as well as removal of abusive leaders and help to promote faith in, and the legitimacy of, new democratic structures and leaders while delegitimating and disempowering old violent leaders and excombatants. Transitional justice advocates also have made the case that trials in particular mark a break with the past and serve purposes of wider political and societal transformation (Teitel 2000). They sometimes have argued that such processes help to empower victims, also broadly reshaping the political spectrum in a transitional state. However, my collaborative research on rule of law, transitional justice, and peacebuilding suggests that these effects have often been overstated (Sriram, Martin-Ortega, and Herman 2011; Sriram et al. 2013). I argue, instead, that the relationship between a range of political actors and accountability is rather more complicated, and that a consideration of those actors’ interests and levels of vulnerability is important (Sriram 2013). Somewhat more optimistic, however, is the work

196

Chandra Lekha Sriram

of Lara Nettelfield, who argues that the International Criminal Tribunal for the Former Yugoslavia helped to promote nascent democracy in Bosnia and Herzegovina while nonetheless recognizing that local social context very much shapes the degree to which such a tribunal can affect political culture (Nettelfield 2010). Transitional justice measures, such as trials, should not be assumed to help promote accountability or limit the impunity and political power of abusive leaders. State or opposition or rebel elites can also deploy transitional justice measures in an attempt to reshape political or military power relations. Jelena Subotic argues, for example, in the context of the International Criminal Tribunal for the Former Yugoslavia, that local elites were able to hijack justice, responding to that tribunal and international pressure to create local transitional justice mechanisms to serve their own political interests (Subotic 2009). It has been suggested that President Yoweri Museveni sought to refer only the situation in northern Uganda, and only as it related to crimes allegedly committed by the Lord’s Resistance Army, to the International Criminal Court (ICC) in an attempt to gain the upper hand in a long-standing military conflict and at the same time shield the national army from accountability. President Ahmed Tejan Kabbah of Sierra Leone similarly sought the creation of an international tribunal to address only the alleged crimes of the Revolutionary United Front (RUF) and was able to secure the creation of a hybrid tribunal that did not address crimes allegedly committed by those under his own direct command (Sriram 2005). Elites may use or seek to use transitional justice measures to remove enemies and protect themselves, but also to attempt to increase their own domestic legitimacy and claims to power by promoting a narrative of their enemies as criminals and themselves as upholders of the state or as protectors of their specific (political, regional, ethnic, religious) group. They may even, as suggested by Valerie Arnould, use transitional justice measures as a tool to alter the balance of power imposed domestically by power-sharing arrangements (Arnould, forthcoming). Finally, it is possible that transitional justice measures do not reshape power relations in affected states, nor do they simply get co-opted by government or opposition elites to promote their own interests. Rather, elites might engage in “strategic compliance,” as Christopher Lamont argues in the context of Croatian government compliance with the International Criminal Tribunal for the Former Yugoslavia (ICTY), by performing acts of cooperation with an international tribunal to glean benefits or avoid punishment from international actors such as the European Union (Lamont 2010). This model allows for both cooperation and defiance, and accordingly for the same political actors to permit prosecution of some leading political actors and resist them in others, in complex relation to their own political interests.

Transitional Justice

197

Study Questions The core question of this study, then, is: How, if at all, do transitional justice measures impact power relations in a transitional or postviolence state on their own, or in tandem with power-sharing measures? Subsidiary questions are: 1. Does transitional justice signal to elites the risk of using certain types of violence in their pursuit of power and change their behavior? 2. Does transitional justice help to remove certain elites from the political equation or change their calculations and behavior: by isolating or delegitimating them, or by removing them from peace negotiations, altering their behavior in peace negotiations, or promoting moderates or other forces in negotiations? 3. Does transitional justice help to reshape political structures, for example, by promoting rule of law and democratization, or altering political balances and behaviors? 4. Can transitional justice measures be hijacked by elites to promote their own interests? 5. Can transitional justice measures be engaged with selectively by elites pursuing their own interests? 6. Are the effects on power relations of transitional justice operating alongside power-sharing arrangements different than those of transitional justice operating alone? It is entirely possible that the answer to each of these questions, in a given situation at a particular point in time, might be yes, and it is beyond the scope of this chapter for me to answer these questions comprehensively. Further, assessing each of these questions is complicated if one attempts to do so in tandem with assessing the impact of transitional justice measures on powersharing agreements as such. At the moment, because this is a relatively novel set of inquiries, I simply attempt to trace apparent effects in a select group of countries, as explained in the methodology section below. It is possible to observe the apparent effects of transitional justice measures on key political actors, and conversely the choices that such actors make with regard to transitional justice, but it is difficult to simultaneously and autonomously assess the interactive effects of a transitional justice measure with power-sharing arrangements. This is the case because power-sharing arrangements, like transitional justice measures, seek to alter power relations, so the best that one can hope to do is assess the overall change in a situation in which both are present and try to trace the effects on the former. Therefore, as I discuss below, I selected only situations with both power-sharing arrangements and transitional justice measures. But I

198

Chandra Lekha Sriram

treated effects of transitional justice measures and power-sharing arrangements on power relations as necessarily overlapping while recognizing that this may have failed to properly assess the autonomous effects of the power-sharing arrangements on power relations themselves. The latter analysis I must leave to the other chapters in this book. While I focused on countries that have had both power-sharing and transitional justice processes, they have not all had them introduced simultaneously, so the effects of sequencing may or may not be discernible.

Methodology and Case Selection Many of the theoretical claims and previous empirical examinations have been drawn from experiences in the Balkans, particularly with reference to the International Criminal Tribunal for the Former Yugoslavia and particularly in the context of power sharing and democratization in parts of the former Yugoslavia. In this chapter I chose to focus only on countries in Africa that have been subject to the jurisdiction of an international or internationalized criminal tribunal and that have had, contemporaneously, one or more power-sharing agreements in place. I further focused only on those countries in which one or more top state leaders or top rebel/militia leaders were both the subject of accountability and a participant in the powersharing arrangement directly or through their party or organization.1 My selection of African cases was made to keep the size of the study manageable; in principle, this inquiry could be repeated in other regions of the world. Africa provides a significant number of possible cases, as it has been the site of numerous power-sharing agreements; a database of these is available through the University of Antwerp (2014). Africa is also the site of a significant number of international criminal justice interventions, by the International Criminal Court as well as temporary courts such as the Special Court for Sierra Leone and the International Criminal Tribunal for Rwanda. 2 It is possible that African countries, their conflicts, peace agreements, and prosecutions may have atypical features, which a larger comparative study would correct. Many of the conflicts have been fairly intractable and have been the subject of multiple peace agreements featuring power-sharing arrangements, thus complicating analysis and also potentially indicating acute instability in power relations and institutions more generally. At the same time, many of the countries that I examined have been the subject of significant international peacekeeping operations, which is also likely to alter power relations and state institutions, and may or may not increase the likelihood of arrests of accused or the implementation of power-sharing agreements. More detailed comparative analysis of these countries (intraregion) as well as of these countries with

Transitional Justice

199

those in other regions, such as the Balkans or Latin America (cross-region), might well reveal important patterns and trends. In this chapter I examined situations in which the transitional justice mechanism was an international or internationalized criminal tribunal to enable the strongest possible causal test. Domestic trials in the wake of conflict seldom have the capacity to prosecute top leaders or former leaders, and such prosecutions will be even less likely where these leaders are members of a power-sharing government and in a position to stall or block prosecutions. Trials of former officials who have been excluded by way of a power-sharing arrangement do not offer a reasonable test, as these are more likely an indicator that power relations have already been shifted. While international and internationalized tribunals do not have extremely strong enforcement capacity, they had greater capacity and will than the state in the situations of concern. I did not consider other transitional justice measures such as vetting/lustration because they seldom reach top-level leaders and are as likely an indication of a prior shift in power relations; neither did I examine the reports of commissions of inquiry in any detail. However, I mention both in my country studies as appropriate to elaborate on the wider political context. Finally, I examined only situations where top state leaders or top rebel/militia leaders were the subjects of prosecutions. The reason for this is simple. This is in line with a focus on elite power relations, one of the levels of analyses employed in this book. Prosecutions of lower-level perpetrators are far less likely to affect power relations within the state, except perhaps where they take place in large numbers. Low-level prosecutions are more likely to be evidence either that material evidence needed for highlevel prosecutions is lacking, or that the state is successfully resisting prosecutions of more senior actors. Because the impact to be scrutinized was of transitional justice alone, and with power sharing, on power relations within the state, I examined only situations in which the accused were not only subjects of prosecutions but also the beneficiaries, directly or through their parties or organizations, of power-sharing arrangements. These criteria should present relatively easy cases to test the effects of one specific measure of transitional justice—internationalized prosecutions— on power relations. This is the case because such prosecutions are most likely to target current political leaders/opposition or rebel leaders, and most likely to actually secure custody of them. The goal in setting an easy test is to determine whether the causal claim appeared to work in the most auspicious of settings; future research can and should set more difficult tests. At the same time, this provided an opportunity to examine the effects of transitional justice both alone and in tandem with power-sharing arrangements. While all selected countries have experienced both measures, they have not all done so simultaneously.

200

Chandra Lekha Sriram

For the sake of completeness, I examined ten African countries here, largely based on desk study. My discussions do draw on my primary fieldwork in three of these countries: Sudan, Sierra Leone, and Kenya.

Country Experiences The ten African countries in my study have both experienced the intervention of an international or internationalized criminal tribunal and one or more agreements involving one or more dimensions of power sharing at the same time or in proximate temporal sequence. These are Rwanda, Liberia, the Democratic Republic of Congo (DRC), Central African Republic (CAR), Côte d’Ivoire, Kenya, Sierra Leone, Mali, Uganda, and Sudan. I considered situations in which power sharing preceded the introduction of international criminal justice but the power-sharing arrangement had failed or been superseded by the time a tribunal was introduced, and in which power sharing preceded accountability but the two operated alongside one another. Of these ten countries, all involving conflict resolution in the 1990s until recently, none experienced a tribunal preceding an operational power-sharing mechanism (except, arguably, Mali). Power Sharing Preceding Accountability Rwanda. Rwanda had a power-sharing agreement that preceded accountability measures, and which had formally ceased to function before any such measures. The 1993 power-sharing peace agreement between the government of Rwanda and the Rwandese Patriotic Front (RPF) preceded the death of President Juvenal Habyarimana (University of Antwerp 2014; Rothchild 2005) and the genocide and conflict that followed. The arrangement did not persist in postgenocide Rwanda and did not operate contemporaneously with transitional justice measures. Following the conflict and invasion of the RPF, which helped to end the genocide, the new Rwandan state insisted on the rejection of ethnic distinctions, meaning that no power sharing on such a basis could continue, although the dominance of the RPF across the state meant that an unspoken dominance of Tutsi elites was established (Brown 2011; Straus and Waldorf 2011). While the International Criminal Tribunal for Rwanda was set up in the wake of the genocide in 1994, it was only able to pursue Hutu accused, and the government of Rwanda vigorously resisted attempts to pursue members of the RPF for alleged war crimes committed during the conflict and immediately after. The impact of transitional justice on power relations, however, does not appear to have been that prosecutions removed powerful individuals from office or otherwise affected their rule. While some relatively

Transitional Justice

201

high-level former officials of Rwanda were prosecuted, they were members of the government that had been removed, and those in power were able effectively to limit the scope of prosecutions to their own benefit. Côte d’Ivoire. Côte d’Ivoire has had a history of power-sharing arrangements, many of which predate the involvement of the International Criminal Court. The most recent power-sharing peace agreement prior to ICC involvement came in 2007. It was negotiated by Laurent Gbagbo’s government and rebel groups, and resulted in sharing between the president and Guillaume Soro, leader of the northern-based rebel movement who became prime minister, with government ministers allocated among Gbagbo’s party, from rebel forces, and from the political parties of former president Henry Konan Bedie and former prime minister Alassane Ouattara (University of Antwerp 2014; Mehler 2009; Rothchild 2005). However, in the wake of election-related violence and a rebel takeover in 2011, the country accepted the ICC’s jurisdiction through a procedure open to nonstate parties. The acceptance of the ICC’s jurisdiction appears to have been part of an attempt to utilize the court to target political enemies rather than an attempt to refer all crimes committed during the violence. Those charged by the ICC are former president Gbagbo (in custody), his wife Simone (arrest warrant still pending), and alleged co-perpetrator Charles Blé Goudé (in custody) (ICC 2014a). However, the effects of power sharing are less clear because Gbagbo was forcibly removed from power as the 2007 power-sharing arrangement began to fail. Following his removal, there has been no power-sharing arrangement between Ouattara and proGbagbo forces. Nonetheless, there is arguably a more general connection between transitional justice and power relations insofar as the acceptance of the court’s jurisdiction and self-referral and surrender of Gbagbo and others appear to have been designed to secure their exclusion from power and future influence. Central African Republic. Central African Republic is one of the countries

into which the ICC has opened an investigation. The country also has a history of political and military power-sharing arrangements in multiple peace agreements among multiple parties between 1997 and 2008 (University of Antwerp 2014). These power-sharing arrangements both preceded and followed the investigations and initiation of cases by the ICC. The situation in CAR was referred by a new government following the ouster of the previous one after serious political violence in 2002–2003, in which fighters of the Mouvement de Liberation du Congo (MLC) led by Jean-Pierre Bemba allied with then president Ange-Felix Patasse to battle a rebel movement led by Francois Bozize, former chief of staff of the Central African Armed Forces. Bemba was charged with crimes against humanity,

202

Chandra Lekha Sriram

including rape, murder, and pillaging, committed by forces under his command. The referral was arguably an attempt to ensure Bemba was prosecuted as the government of CAR was unable to pursue him. However, in the first instance, ICC prosecutions have targeted only actors from the Democratic Republic of Congo and not domestic CAR actors, such as Bemba for violence in CAR (ICC 2014c). Therefore, the power relations here may be rather different, in relation to a foreign actor, with a state apparently using the ICC to accomplish what the state itself could not. Arrest warrants have subsequently been issued against one CAR national, Narcisse Carido, for allegedly engaging in witness tampering in the case relating to Bemba. While in 2014 the ICC prosecutor announced the opening of a second investigation into crimes against humanity allegedly committed since 2012 in the context of sectarian violence and the 2013 ouster of Bozizé and his replacement with Michel Djotodia, no arrest warrants have been issued. Further, peace negotiations in 2015 have sought to include Bozizé and Djotidia, parallel to but without engaging with the ICC processes. The implications for power relations of these ICC interventions are thus somewhat unclear. Sierra Leone. Sierra Leone was the site of the first hybrid criminal tribunal

in Africa, the Special Court for Sierra Leone, established in 2002 to prosecute those who bore the greatest responsibility for crimes committed during the decade-long civil war in that country. The creation of the court followed the end of the country’s conflict, brought to a halt by significant international military intervention after the collapse of two major power-sharing peace deals. The final agreement, the Lomé Peace Accord of 1999, drew particular international condemnation for offering an “absolute and free pardon,” or amnesty, to all participants in the conflict, effectively foreclosing domestic prosecutions of serious international crimes (Sriram 2005). It also included political, economic, and military power-sharing provisions. The agreement provided for a restructuring of the armed forces, which would include elements of the national army, the rebel Revolutionary United Front, and the Civil Defense Forces (CDF). It provided for a body to oversee the management of national resources, to be chaired by RUF leader Foday Sankoh and to include representatives of the government, the RUF, civil society, and political parties; and it provided for cabinet posts to be allocated to the RUF and the post of vice president to Sankoh (University of Antwerp 2014; ACCORD 2009). However, not long after the agreement was concluded, the RUF returned to fighting. While the agreement was never formally renounced, its implementation certainly failed and, when peace was finally restored, power-sharing arrangements were not put into place. However, while the SCSL was compelled to consider whether the amnesty in the agreement remained valid (and, if so, for whom), there was no serious discussion regarding the power-sharing elements of the agree-

Transitional Justice

203

ment. While the RUF did transform into a political party, it did not benefit from those elements of the agreement and its electoral shares were vanishingly small. It merged with the All People’s Congress in 2007. The SCSL did not operate, therefore, contemporaneously alongside operational power-sharing provisions, but in the immediate wake of failed ones. It might have been designed to, or succeeded in, altering power relations within the country. Certainly, when he requested that the United Nations create an international criminal tribunal for Sierra Leone in 2000, President Ahmed Tejan Kabbah had a particular vision. He sought an international tribunal, not the hybrid tribunal that was ultimately created and, more notably, he sought a tribunal to try only the crimes of the Revolutionary United Front. He sought, arguably, to use a transitional justice tool to maintain his politically advantageous position. Further, while the remit of the tribunal was ultimately broader, encompassing other rebel actors and the Civil Defense Forces that at least indirectly supported his government and his military strategies during the conflict, neither Kabbah nor any members of the national army were the subjects of prosecutions, although the statute did not explicitly exclude it (Special Court for Sierra Leone 2002). The SCSL prosecuted, among others, Sankoh as the head of the RUF and Chief Sam Hinga Norman of the CDF, which may have had the effect not only of removing them from the political equation but of delegitimating their groups. This was potentially true of the RUF, but this is difficult to assess in light of the fact that the RUF’s notoriously brutal tactics may well have done this. In prison Norman remained a potent political symbol, and the CDF remained a strong presence in the country but was not active in national-level politics (Sriram 2005). The death of each of the accused in custody means that it is impossible to assess the effects of their convictions, had they taken place. Thus, it is evident that, although President Kabbah sought to use the creation of the SCSL to his own political ends, it is less clear whether or how the court’s operation affected national power relations. Uganda. Recent conflicts in Uganda have not been the subject of any

power-sharing peace agreements (University of Antwerp 2014). However, in 1985 a power-sharing arrangement was reached with leaders of the Acholi in the north, but later collapsed, which was one of the causes of the conflict that then ensued. Uganda is included in this discussion because the final round of peace negotiations to resolve the conflict in the north involved discussions of power-sharing elements, including integration of members of the rebel Lord’s Resistance Army into the national army. In 2004, the ICC issued arrest warrants for the head of the LRA, Joseph Kony, and four of his deputies (ICC 2014f). Peace talks were revived in Juba, South Sudan, in 2006, under the shadow of the arrest warrants. The Juba peace talks ultimately broke down amidst allegations that they were

204

Chandra Lekha Sriram

undermined by the arrest warrants, although there is significant policy and scholarly debate on this point. When he referred the situation to the ICC, President Yoweri Museveni specifically sought to refer the situation in northern Uganda, and crimes committed by the LRA only. While the prosecutor publicly embraced this self-referral and the two men stood together in announcing it, the former subsequently clarified that the referral could not be limited in the manner apparently intended, meaning that investigations and prosecutions could in principle pursue members of the national army as well. This clearly was not the intent of President Museveni, who appeared to have hoped to use the ICC as another tool to combat the LRA. While the ICC issued arrest warrants for five LRA members, none have been arrested and two are presumed dead, and there are no prosecutions under way for this situation. Following debates regarding the effects of the ICC on peace negotiations, calls for amnesty and traditional justice processes by certain Acholi leaders, and what appear to be altered preferences on the part of Museveni, the government of Uganda began to take a hostile stance toward the ICC, setting up an International Crimes Division domestically in what it said was unilateral implementation of one element of the incomplete Juba agreement. The Ugandan military is excluded from the jurisdiction of this court (Republic of Uganda, the Judiciary 2014). There has been only one trial to date, of an LRA fighter, which was fraught with legal and procedural challenges. Power Sharing Alongside Accountability The Democratic Republic of Congo. The Democratic Republic of Congo was the subject of a complex power-sharing arrangement over a decade ago, reached alongside the opening of an investigation by the ICC into the situation in that country. The Sun City Agreement of 2003 provided for the existing president Joseph Kabila to retain his post as president and provided for four vice presidents, who were former rebel leaders and political opposition members (University of Antwerp 2014; Rothchild 2005). This agreement was put in place in the same year as the government referred the situation in the country to the ICC, with investigations opening shortly thereafter. The first completed cases at the ICC (one conviction and one acquittal) pertain to the DRC. However, none of the rebel leaders prosecuted by the ICC or targeted with arrest warrants were part of the 2003 power-sharing arrangement, and each was considered relatively low-level in the context of the country’s politics, although significant in local conflicts. Thus far, the only leader involved in the 2003 agreement who has been accused is Jean-Pierre Bemba, who is being pursued for alleged crimes committed outside the DRC. Only Bosco Ntaganda is a specific subject of both power sharing and accountability, as

Transitional Justice

205

he was part of a subsequent agreement that integrated him and members of his group into the army of the DRC. He defected, however, returning to fighting in the eastern DRC and then surrendering to the ICC in the face of an arrest warrant (ICC 2014b). Liberia. Liberia is in one sense an atypical case because it was neither the

site nor the subject of an international criminal tribunal. However, as is well known, its former head of state, Charles Taylor, was prosecuted and convicted by an internationalized tribunal, the Special Court for Sierra Leone, for his role in the conflict in that country. Although the SCSL was not evidently designed or expected to shape power relations in Liberia, it is worth examining what effect it may have had in practice. Liberia’s conflict prompted numerous peace negotiations, resulting in several power-sharing agreements over a decade between 1993 and 2003, each of which had some or all of three types of power-sharing provisions: political, economic, and military. The country experienced repeated relapses into conflict. However, in 2003, during peace negotiations that were being held in Ghana, the SCSL unsealed an indictment for Taylor, a sitting president, and Ghana was asked to surrender him. The Ghanaian government, apparently seeking to avoid a diplomatic crisis and any effect on the peace negotiations, did not arrest him. Instead, a peace deal was reached and, notoriously, Taylor was allowed to go into exile in Nigeria (Sriram 2005). The 2003 peace deal included economic, military, and political power-sharing elements. The agreement provided for the restructuring of the armed forces to include former state, former rebel, and civilian actors; for a process to negotiate distribution of leadership of various public corporations and commissions among parties to the agreement and civil society; and an inclusive transitional government with seats allocated to the government, rebel groups, and civil society (University of Antwerp 2014; Mehler 2009). While the agreement has been termed a “model” by some analysts, who have suggested that the indictment of Taylor helped to delegitimate him, they have also speculated that it may have helped embolden rebels and delay talks (ACCORD 2009, 12–14). It is unclear whether the unsealing of the indictment for Taylor helped convince him to agree to the peace deal and to go into exile, and it is impossible to prove the counterfactual. It is clear that the power-sharing transitional government was able to operate, and that neither the prosecution of the former head of state nor the recommendation of the national truth commission that the president, Ellen Johnson Sirleaf, be excluded from public office for thirty years altered the composition of the postagreement government (Raddatz 2013). It is also difficult to assess whether the trial of Taylor helped to alter the power of his former supporters, or conversely to increase the relative power and influence of Johnson Sirleaf; this would require

206

Chandra Lekha Sriram

more targeted population surveys and election analysis than are available. It is clear that the prosecution of Taylor did not undermine implementation of the power-sharing agreement. Liberia presents a very different picture than its neighbor Sierra Leone, despite each having been the subject of power-sharing peace agreements, as in the former these elements were implemented while in the latter they were not. Kenya. The postelection violence sparked by disputed election results in 2007 in Kenya resulted in both a power-sharing arrangement and the eventual involvement of the International Criminal Court (Sriram and Brown 2012; University of Antwerp 2014; ICC 2014d). While the power-sharing arrangement put in place by a negotiated national accord phased out as planned, a rather unusual form of power sharing now appears to be in place in the form of the current government. In a coalition of adversaries who are both the subjects of prosecutions by the ICC, Uhuru Kenyatta and William Ruto ran and were elected as, respectively, president and deputy president in 2013. While these men were not candidates in the original contested election, they were leading figures in the parties of the leading candidates in that election. The power-sharing arrangement negotiated in 2008 ensured that both the incumbent and his challenger were placed in positions of leadership: Mwai Kibaki retained the title of president and Raila Odinga was given the new position of prime minister, and a coalition government with balanced portfolios was agreed on. The parties also agreed to a constitutional review process, which resulted in the 2010 constitution, passed by referendum, which entailed significant reform for the judiciary, the police, and other governmental institutions in principle. At the same time, the national dialogue resulted in the creation of a Commission of Inquiry into Post-Election Violence, which proposed a hybrid tribunal based in Kenya, but with the proviso that if a tribunal was not approved nationally by a specified deadline (later extended), a list of the potential accused and related documentation were to be passed by Kofi Annan, the mediator for the negotiation process, to the International Criminal Court. Following several delays and machinations by politicians, the local tribunal was not created, and once documentation was given to the ICC, the Office of the Prosecutor requested and was granted permission to open an investigation into the postelection violence in 2010, attended by strong expectations that it would help promote a new era for rule of law, accountability, and democracy in Kenya (Sriram and Brown 2012; Höhn 2014). At that time there was strong public support for accountability, and for the ICC in Kenya, which may have translated into the strong public support for the new constitution, with tighter rule-of-law provisions. However, while this might have affected political actors, and indeed there were mixed domestic reactions initially to the naming of Kenyatta and Ruto in 2011 and

Transitional Justice

207

the confirmation of charges against them in 2012, this did not prevent either man from remaining politically active. Indeed, the two accused formed a coalition, the Jubilee Coalition, and ran for office together. Arguably, here the ICC did alter political power relations, insofar as former adversaries were brought together for strategic reasons to pursue top offices in the country, calculating that ethnic allegiances that supported each would strengthen a coalition party. Speculation is rife in Kenya that, as evidentiary difficulties with the case against Kenyatta make it likely to be dismissed, the coalition of convenience will collapse (Sriram 2014). Mali. Mali has experienced a number of armed conflicts, particularly in the

north of the country, based on demands by nomadic groups and others for greater autonomy, by groups demanding a separate Tuareg state, and by Islamist groups. Power-sharing arrangements to resolve violence in the north date back to at least 1991, at which time an agreement between the government, the Mouvement de l’Azaouad, and le Front Islamique was reached which included military power-sharing provisions and mixed territorial and economic power-sharing elements (University of Antwerp 2014). Subsequent agreements with the Mouvement de l’Azaouad expanded to include political power sharing, through inclusion of persons from movements in the north into the National Assembly and into public administration. In June of 2013, two Tuareg rebel groups, including the National Movement for the Liberation of the Azawad (NMLA), agreed to a power-sharing peace deal involving military, political, and territorial elements. Provisional security forces in Kiddal, the affected area in the north, were to be comprised of state security forces and those who had deserted; a provisional regional council was set up for more devolved local governance; and agreement was reached for greater involvement of qualified persons from the north in governance with an implementation committee comprised of members from all parties (University of Antwerp 2014). In January 2013, the prosecutor of the ICC formally opened an investigation into crimes allegedly committed in the north of Mali dating back to January 2012, following a referral of the situation by the government of the country (ICC 2014e). The investigation was into alleged war crimes and crimes against humanity committed during fighting between the government and several organized armed groups—Mouvement National de Libération de l’Azawad (National Movement for the Liberation of Azawad, MNLA), al-Qaeda in the Islamic Maghreb (AQIM), Ansar Dine, and the Mouvement pour l’unicité et le jihad en Afrique de l’Ouest (Movement for Oneness and Jihad in West Africa, MUJAO)—and “Arab militias,” and among those armed groups themselves (ICC, OtP 2013). The MNLA has its roots in earlier Tuareg rebellions in the north dating to the early 1960s, but it is not strictly related to the organization of a similar name involved in the

208

Chandra Lekha Sriram

1991 power-sharing agreement; the MNLA is, however, party to the most recent power-sharing arrangement. To date, the situation is under investigation but no accused have been named and no trials are under way, so limited impact may be expected. Whether the government had any political or military agenda in referring the situation to the ICC in advance of elections in August 2013 or while the most recent agreement was being negotiated is difficult to gauge based on available information. Sudan. The conflict in Sudan between the Khartoum-based government in

the north and the Sudanese People’s Liberation Movement/Army (SPLM/A) in the south formally ended with the completion of the Comprehensive Peace Agreement in 2005 (University of Antwerp 2014; Sriram 2008; ACCORD 2009). The signing of the agreement came alongside the escalation of conflict in the Darfur region of the country, and a referral of the situation by the UN Security Council to the ICC (ICC 2014g). The situation in Sudan presents an interesting test case as the peace agreement involves a distinct conflict from that presented by Darfur, the subject of ICC investigations and prosecutions; nevertheless, it merits examination because the same head of state is involved. The Sudanese Comprehensive Peace Agreement involves power-sharing arrangements on all four dimensions: military, economic, political, and territorial (University of Antwerp 2014; Sriram 2008; Mehler 2009). The agreement provided for an interim period prior to a referendum (held in 2011), which would determine the status of southern Sudan, during which time the government and SPLM/A forces would remain autonomous, with the plan of integrating into a new national army, save for a small number of joint integrated units to be created earlier. It also provided for wealth sharing to manage oil resources as well as creating a national petroleum commission with participation by both sides. Political power-sharing provisions placed then head of the SPLM/A as first vice president of Sudan and president of the interim government in the south and allocated representation in the National Assembly and Council of States to southerners as well as stipulating the representation of southerners versus the national ruling party in southern Sudan. Finally, the agreement provided for autonomy in the south and for the possibility of independence, which was granted following the 2011 referendum. During negotiations for the Darfur Peace Accord, the rebel groups Justice and Equality Movement (JEM) and the Sudan Liberation Army (SLA) sought provisions similar to these. Clearly the agreement did alter power relations at one level, bringing in new players at the national level for a time and ultimately resulting in the separation of South Sudan; however, I argue that at a deeper level, the existing structures of power and bureaucracy limited the influence and operational capacity of many SPLM/A members who joined (Sriram 2008).

Transitional Justice

209

While implementation of the Comprehensive Peace Agreement has encountered significant difficulties as new conflicts have emerged and old ones continued, the ICC entered the situation, notably officially issuing an arrest warrant for President Omar al-Bashir in March 2009, following earlier arrest warrants issued for the former national minister of humanitarian affairs, the alleged head of the militia janjaweed, and key leaders of rebel groups in Darfur. The first warrant of arrest for a sitting head of state by the ICC might have been expected to have some effect on power relations within the state, or potentially on the power-sharing arrangement itself. However, the warrant seemed only to have made President al-Bashir more resistant to international pressures over Darfur and certainly did not appear to weaken his political power. Neither did the warrants for al-Bashir and others appear to affect the power-sharing arrangement in place: while the SPLM/A had previously vowed solidarity with other oppressed regions, its representatives in government did not break ranks over the warrant.

Analysis A review of the countries that I surveyed suggests that rather than having one set of effects, transitional justice measures, specifically internationalized trials, may have several effects on power relations in a country emerging from violence and conflict and that these may vary depending on the presence or not of a currently operational power-sharing arrangement. The effects that are expected, and to a degree partly borne out, by experience are mixed (see Table 10.1 for an overview of these effects): • Transitional justice helps to reshape power relations by removing some significant perpetrators, who may be leaders of one or more key political/military factions; • Transitional justice helps shape power relations by embedding past perpetrators, elites, and their political/military factions in power; • Transitional justice is invoked by leading political/military elements in an attempt to maintain power or undermine political challengers (hijacking); • Transitional justice is engaged with selectively by leading political/ military elements; • Transitional justice fails to affect power relations. A number of observations follow from the countries that I examined and from Table 10.1. First, there is no strong evidence, based on these ten examples of post–Cold War African countries that experienced both powersharing agreements and the operation of an international criminal tribunal,

210

Chandra Lekha Sriram

Table 10.1 Transitional Justice, Power Sharing, and Power Relations Power Sharing Precedes Transitional Justice and Terminates

Power Sharing Precedes Transitional Justice and They Operate in Part Simultaneously

International(ized) tribunal helps to reshape power relations/removal of elites

No cases

No cases

International(ized) tribunal helps shape power relations/embedding elites

Rwanda, Sierra Leone

Kenya

International(ized) tribunal invoked by elites in attempt to shape power relations

Rwanda, Uganda, Côte d’Ivoire, Sierra Leone

Democratic Republic of Congo, Central African Republic

Selective engagement

Rwanda, Uganda

International(ized) tribunal fails to have an effect on power relations

Sudan, Mali, Liberia, Central African Republic

that having the latter promotes any significant change in power relations. Second, despite the first finding, international trials may still have an effect on power relations, which is to embed those elites already in power, whether or not there is a currently operational power-sharing arrangement. Nonetheless, in a significant number of countries, there is a lack of evidence that the operation of an international tribunal has any effect on power relations. Finally, in a majority of situations, elites in power with or without the operation of a power-sharing agreement seek to use the presence of an international tribunal to their own political advantage. On the basis of this evidence, it is clear that while elites in many states are able to control any impact of international trials on their overall political power, many elites also seek to use such tribunals to reinforce their power, and sometimes succeed. This is marginally more common in states in which a previously operational power-sharing arrangement is not in effect, but the evidence at this stage is limited. Further research, including wider and deeper qualitative studies, and quantitative studies, would be necessary to check the generalizability and accuracy of these preliminary findings and to enable the development of a nuanced explanation. The trends that I identified are not easy to explain, and further research is needed to elaborate on them. However, some speculative observations follow. While trials may have a weak effect in terms of removal and delegitimation, they do have effects on power relations that previously have not

Transitional Justice

211

been well articulated. Specifically, they may be deployed by elites to promote their own political goals and have the effect of embedding elites or, as in Kenya, creating unusual alliances. This suggests that they have the capacity to affect power relations, but that this is not based on their substantive aims or purported messages. Instead, they may be susceptible to redeployment by such elites. This may be because trials, targeting a small number of specific leaders, are inherently unable to effect sweeping change, but also because even international prosecutions are vulnerable to political manipulation in content or message. If this is indeed the case, then advocates of transitional justice may wish to exercise caution; similarly, those who develop power-sharing arrangements may want to plan for this effect.

Notes 1. This does not necessarily refer to a head of state or head of a rebel organization; important deputy officials may also be relevant. 2. And indeed, there are relatively few internationalized criminal tribunals operational outside of Africa, compared to the number created for Africa or the caseload of the International Criminal Court—those external to the continent are hybrid processes in Timor-Leste and Kosovo (now complete, with a new Kosovar process recently initiated) and the continuing tribunal in Cambodia run with the support of the United Nations; hybrid tribunals in Bosnia and Herzegovina, Croatia and Serbia, and the UN-mandated Special Tribunal for Lebanon.

PART 4 Conclusion

11 The What, How, Where, and Who of Postconflict Power Sharing Caroline A. Hartzell and Andreas Mehler

Power-sharing arrangements of the kind designed to help end intrastate conflicts are posited to work by altering the balance of power between warring groups in such a manner as to secure the peace.1 As we emphasized in the introduction (Chapter 1) of this volume, although a number of studies have examined the impact of power sharing on conflict recurrence and the duration of the peace, the effects that these measures have on the distribution of power within, between, and among actors, groups, and institutions in the postconflict state by and large have been neglected. This neglect strikes us as problematic for two reasons. First, if as has been hypothesized, power sharing “works” by effecting shifts in the balance of power, then it seems important to examine whether, and to what extent, such changes take place following agreements to share power. Second, the possibility exists that power sharing may have unintended effects, including on power relations. Although some of those effects might be positive or peace prolonging in nature, others might serve to deepen grievances, exacerbate inequalities, or in some other fashion undermine the peace. If we want to understand whether, and how, power sharing works, tracing its effects on the distribution of power in postwar states strikes us as a good place to start. The contributors to this book were asked to examine the effects that power sharing has on power relations in states emerging from intrastate conflict. The authors of the chapters sought to do so by employing different levels of analysis—that is, by focusing on what occurs to the distribution of power within, between, or among different types of actors or units—and examining the effects that different forms of power sharing—that is, political, military, territorial, economic, or some combination thereof—have on the balance of power. Whereas some of the chapters took the form of in-depth case studies of individual countries or a small number of countries that the authors analyzed using qualitative methods, others relied on 215

216

Caroline A. Hartzell and Andreas Mehler

cross-national samples and employed statistical analyses in an effort to arrive at some generalizations regarding power sharing. As editors, our goal in this chapter is to assess the central findings that emerge from the analyses in the book as well as to specify the contributions that this volume makes to the study of power sharing’s effects on power relations in postconflict states. We begin by summarizing the various chapters’ findings regarding the types of effects that power sharing has on the balance of power within, between, and among actors, groups, and institutions following intrastate conflict, complementing our synopsis with some observations regarding the potential significance of the authors’ results. Following that, we consider the utility and the limitations of the analytical framework we employed in the book for the study of power sharing’s effects on power relations. Next, we identify factors that appear to influence the manner in which power sharing shapes postconflict power relations. We conclude with a few thoughts regarding policy implications that stem from this work.

Effects of Power Sharing on Postconflict Power Relations: Findings Not surprisingly, given the different types of power relations and forms of power sharing analyzed in the volume, as well as the methodologies that are employed, our contributors found that power sharing has a number of different effects on power relations. In some instances these effects are, as Martin Ottmann and Johannes Vüllers put it in Chapter 2, positive-sum in nature. Focusing on power relations defined in terms of the relative strength, militarily and politically speaking, between the government and rebel groups, Ottmann and Vüllers found evidence that parliamentary power sharing makes nonviolent and inclusive power relations between these two sets of actors more likely, while military integration increases the chances for inclusive power relations. They also found, however, that power sharing at the cabinet level exercises a negative (although not statistically significant) effect on government-rebel power relations. This outcome, which Ottmann and Vüllers attribute to the desire of the rebel leaders who are appointed to cabinet positions to freeze power relations and thus remain in positions of power, stands as one example of the potential that power-sharing institutions have to produce unintended effects on power relations. Rosalie Arcala Hall’s contribution (Chapter 5) on military power sharing in the Philippines offers rich insights into the mechanisms that played a role in altering power relations between the state and society as well as at the local level in Mindanao. Generally speaking, the effects that military power sharing appears to have had on power relations in this case seem to be positive-sum in nature. Integration of rebel forces into the state’s security

The What, How, Where, and Who

217

apparatus appears to have helped strengthen the state as well as generating some degree of legitimacy for those forces. Military power sharing also produced some material benefits for those Muslim communities that took part in the integration of forces. The many types of ways in which military power sharing, a measure typically included in peace settlements to provide weaker actors with a sense of security, shaped the balance of power between state and society in Mindanao serve to draw attention to the need to investigate heretofore unexplored mechanisms through which power sharing may shape power relations. Juxtaposing two consociationalist cases, the liberal 1998 Belfast (Good Friday) Agreement in Northern Ireland and the corporate 1989 Taif Accord in Lebanon, Bernadette C. Hayes and John Nagle provide evidence in Chapter 9 that the design of power-sharing institutions can have an effect on whether postconflict power relations prove to be positive- or zero-sum in nature. Focusing on issues of interfaith marriage, Hayes and Nagle examined the ways in which the two sets of power-sharing arrangements have shaped social relations. The authors found that the groups sharing power in Lebanon’s corporate system have not reduced their mutual distrust because the antagonism between them is strongly institutionalized, authorities with powers regarding marriage are segmented, and the balance of power relies on a reproduction of their antagonism. As a result, intermarriage is not approved of in Lebanon while, at least where attitudes toward intermarriage between the two main religious camps are concerned, individuals in Northern Ireland are decisively more tolerant. As the analysis in this chapter suggests, power relations within groups can be shaped by the extent to which institutions foster (or fail to foster) tolerance. A number of other chapters sound a cautionary note regarding unanticipated and potentially negative effects that power sharing may have on postconflict relations. Based on an analysis of four African countries, Andreas Mehler, Claudia Simons, Denis M. Tull, and Franzisca Zanker conclude in Chapter 4 that, in some instances, local elites who accept power-sharing positions in the national government may actually lose critical local influence. Concomitantly, some political actors may also gain financially from their positions at the local and national levels by playing a two-level game that has the effect of reinforcing already strong elements of patronage. The authors’ elucidation of the two-level game in this chapter makes clear the ways in which institutions designed to shape the balance of power between or among certain actors, organizations, or units may unexpectedly end up altering other sets of power relations. John Ishiyama provides further food for thought in Chapter 3, in which he studied the effects that political, military, territorial, and economic power sharing have on the fragmentation of rebel movements. Ishiyama’s finding that territorial power-sharing arrangements promote organizational splits in

218

Caroline A. Hartzell and Andreas Mehler

rebel parties is one that peace settlement architects will want to take into account, given that such splits may give rise to potential spoilers to peace agreements. Kathleen Gallagher Cunningham’s examination in Chapter 6 of the effects that territorial power sharing have on the cohesion of selfdetermination movements makes for an interesting counterpart to Ishiyama’s contribution to this book. Cunningham found that, by fulfilling the core demands of self-determination groups, power sharing can lead to their disbanding, their disarmament, and the cessation of violent conflict as former opponents are integrated into a newly defined, more decentralized, or federal state apparatus. Although Ishiyama and Cunningham focus on different sets of cases (civil wars in the former and self-determination movements in the latter), the two works have in common the assumption that specific forms of power sharing create distinct incentives shaping actors’ behavior. Finally, it is worth noting that three chapters (Chapters 7, 8, and 10) yielded a number of null findings where the effects of power sharing on postconflict power relations are concerned. In Chapter 8, Matthew Hoddie investigated the implications that the adoption of each of the four forms of power sharing focused on in this book have for the respect by postconflict governments of individuals’ physical integrity rights. Hoddie found that, contrary to conventional wisdom on this topic, the adoption of power-sharing measures as part of a peace settlement has virtually no effect on this particular aspect of the balance of power between state and society. Instead, he found that a number of other factors related to the postconflict environment proved more influential in shaping states’ respect for physical integrity rights. Hoddie’s nonfindings thus serve as a corrective to the temptation to overattribute developments in postconflict states to the contents of peace settlements, including power-sharing institutions. Focusing on economic power sharing, Caroline A. Hartzell concluded in Chapter 7 that this particular form of power sharing has had a limited impact in shaping power relations in postconflict states. Hartzell adduced a number of reasons for this, including reluctance on the part of some actors fully to implement agreements to share economic power and an apparent perception that other forms of power sharing may better serve to alter power relations. Employing a plausibility probe, she found little evidence that economic power sharing successfully has been used to alter horizontal power relations and argues that there is little reason to believe it has served significantly to change the balance of power between regions and the central government. The apparent inability Hartzell identified on the part of economic power sharing to alter some types of power relations suggests that, as is the case with other types of institutions, power-sharing institutions can be either effective or ineffective in nature. Chandra Lekha Sriram’s examination in Chapter 10 of the effects transitional justice processes have on power relations and, through them, on power-

The What, How, Where, and Who

219

sharing arrangements, serves as an important reminder that other factors that shape power relations may have the potential to shape power-sharing institutions themselves. Although Sriram found little evidence that transitional justice mechanisms have produced a fundamental reshaping of power relations, her study of a number of African cases led her to conclude that the interaction of transitional justice processes with power-sharing institutions can have unexpected consequences, including the ability for some key actors to manipulate these devices to advance their own interests.

Analytical Framework of the Book: Its Utility and Limitations The central puzzle motivating this work was the desire to learn whether and, if so, how, power-sharing arrangements have an effect on the balance of power within, between, and among diverse sets of actors and units in the postconflict context. Although each set of chapter authors was afforded some latitude regarding the manner in which they explored this issue, to lend coherence to this endeavor and to ensure that the parts would sum up to a theoretical whole, we asked them to make use of a common analytical framework. The analytical framework consists of a series of specific questions meant to help improve our understanding of how power sharing alters power relations. In what follows, we briefly discuss the analytical leverage gained through the use of each of these questions before moving on to discuss the utility and limitations of the framework as a whole. The authors of each chapter were asked to address the question What? in two ways. First, authors were instructed to focus their analyses on some aspect of power in relation to the state. Given our focus on power-sharing arrangements as rules designed to apportion state power among some subset of actors, we sought to determine what effect such institutions had on the balance of power in relation to the state. Accordingly, as we noted in the introduction to this volume, authors were asked to adopt a definition of power that stresses its relational aspect. Second, authors were asked to describe the nature of some aspect of power relations in the postconflict state. In this sense, the what of interest constituted the outcome or dependent variable in each chapter. A number of different types of power relations were examined in the chapters. These range from a focus on state-society relations to relations between government and former rebel groups, relations between the political center and local governments, relations within rebel groups, and more. Although not exhaustive, the variety of power relations explored in this volume is instructive, highlighting as it does the multiplicity of ways in which power sharing can have (or may not have, in the case of the null effects that authors noted) an effect on different balances of power in the postconflict state.

220

Caroline A. Hartzell and Andreas Mehler

The question How? that authors were asked to consider was also multidimensional in nature. One dimension consisted of analyzing whether power was meant to be distributed among actors on the basis of a formal political settlement negotiated among elites and often with the involvement of the international community, or whether power sharing was informal in nature and perhaps based on informal practices rather than on clearly delineated agreements or as part of constitutions. The second dimension that the chapters were meant to address was just how much power actually was shared in the aftermath of a conflict-ending settlement. The extent to which the authors were able to fully address these issues depended, in no small measure, on the empirical approaches they adopted. Chapters that focused on a single power-sharing deal or a limited number of cases were in a better position to unearth informal power-sharing practices than were chapters that employed cross-national analyses based on coding the contents of settlements. Determining just how much power was shared in the postconflict environment proved a challenging task for everyone given that there exists no metric for measuring changes in power relations. Generally speaking, however, chapters that wrestled with the question of the extent to which power-sharing measures actually were implemented may have provided more traction on this issue. Authors were also asked to delve into the question of Where? power was shared in the cases that they examined. In many instances, power-sharing agreements focused on the distribution of power among elites at the political center. In others, power-sharing settlements encompassed center-periphery relations while, in a few cases, some measure of power was specified for local governments. Power sharing and power relations thus can be understood to have a spatial dimension to them, something that a number of the chapters touch on. Power sharing also has a sectoral dimension to it in that state power can be distributed on the basis of one or more of four sectors— political, military, territorial, and economic—of state power. This dimension comes through clearly in all of the chapters in the book, with authors analyzing the effects that one or more forms of power sharing have on postconflict power relations. In keeping with new scholarship on this issue, some chapters even developed a more detailed understanding of the content of one or more of the forms of power sharing than that established in the original scholarly specifications of power-sharing institutions as part of war-ending agreements. Finally, authors were asked to identify Who? were the actors among which power was to be distributed as part of a power-sharing agreement. Perhaps the most innovative element that our analytical framework introduces is its use of the levels of analysis as a means of answering this question. Authors were asked to think about the potential that power-sharing arrangements could have for altering the balance of power within actors,

The What, How, Where, and Who

221

between actors, and among actors. Working within this framework, the contributors to this book identified a number of different actors and dimensions along which power relations might potentially be altered by power sharing. Although, in some instances, power sharing appeared to have little to no effect on some of these sets of power relations, the simple act of identifying a variety of types of power relations on which power sharing could conceivably have an effect should be considered a significant contribution to the literature on power sharing. We believe that the utility of the analytical framework employed in this book is twofold in nature. First, and perhaps most importantly, the framework focuses attention on the balance of power between the state and some set of actors (or unit, in the case of substate territories or regions), the relationship at the heart of a power-sharing arrangement. The chapter authors were asked to consider whether the adoption (and implementation) of some set of power-sharing measures has the effect of altering that balance of power and, if so, how it does so. Are the effects those that were expected? Or might power sharing have unintended effects, perhaps altering the balance of power within one of the parties to the settlement? Put another way, the questions at the core of the framework attempted to get at the mechanics of how power sharing works. If students of power sharing can gain a clear understanding of these mechanics, we arguably will be in a better position to understand how power-sharing arrangements shape other outcomes of interest, including the longevity and quality of the peace. A second way in which this book’s analytical framework may prove of use to those interested in learning more about power sharing is that it provides a means for tracing, and comparing, pathways from the adoption of power-sharing arrangements on through to their effects on power relations. This could be done by setting up a table, for example, with the different forms of power sharing arrayed along one axis and the actors with the potential to be involved in or affected by the power-sharing measures listed along the other axis. Each of the cells in the table could then be populated by noting whether, and if so to what extent, power relations between the state and a particular actor were affected, with particular attention given to how they were affected—that is, as a result of a shift in the balance of power between or among actors or within an actor or unit. Although this type of analysis would not allow one to control for all of the factors that might potentially have an impact on postconflict power relations, it still could serve as a means of identifying potential causal factors, processes, or even outcomes that previously may have been overlooked in the study of power sharing. As we alluded to in the previous paragraph, one limitation of the analytical framework developed for this book is that it does not encompass all of the explanatory factors that have the potential to shape power relations. Lacking from the list of questions we asked the authors to consider, for

222

Caroline A. Hartzell and Andreas Mehler

example, is Why? We did not ask the contributors to consider why actors had agreed to share power, perhaps because the reason—to end intrastate conflict—seemed so obvious. However, elites may adopt power-sharing mechanisms for other reasons such as ensuring that the national balance of power continues to favor them. We also did not ask authors to analyze why in some cases some forms of power sharing, but not others, are adopted. This strikes us as a question well worth asking given the possibility, raised by Sriram and Hartzell, that actors may think strategically about what types of power-sharing institutions to adopt given judgments about what will and will not work and what effects the institutions will have. These are questions that anyone who considers making use of this analytical framework in the future may well want to take up.

Factors Influencing How Power Sharing Shapes Postconflict Power Relations In what follows, we draw attention to some factors that have the potential to shape the manner in which power sharing affects—or fails to affect— postconflict power relations. Context

As a number of the chapters in this book reveal, the contexts in which powersharing arrangements are adopted and carried out have an influence on postconflict power relations. Mehler, Simons, Tull, and Zanker, for example, attribute quite strongly diverging effects of similar types of power-sharing measures in the four African cases they examined to contextual factors such as the history of center-periphery relations. Hall notes that a history of rido, or clan feuding, in some Mindanao communities vitiated the security effects of integrating Muslims into the local police forces given that, within that context, family members in the security forces are considered clan assets available for mobilization. Hoddie found that conflict intensity is far more important in predicting whether or not governments will show respect for physical integrity rights, particularly the propensity to use torture, than any form of power sharing. All of these points, and many others, serve as an important reminder regarding the need to be careful not to overgeneralize the degree of influence that power sharing has on postconflict power relations. Interestingly enough, our contributors seemed to find little evidence that global structural factors played much of a role in shaping power sharing and postconflict relations in the cases that they analyzed. This should not be taken to indicate that global structural factors do not play a role. Rather, this nonfinding may stem from a choice on the part of most of the authors not to

The What, How, Where, and Who

223

focus on this type of contextual factor. In her chapter on economic power sharing, Hartzell does speculate that one reason this form of power sharing has only limited effects on power relations is that global economic forces or international actors such as the international financial institutions curtail its effects. Although she offered no evidence in support of this claim, scholars may want to consider the role that factors such as relations among the great powers, the global economy, and global technology (e.g., access to the Internet and the spread of social media) may play in shaping power-sharing arrangements and power relations. While global structural factors received relatively little attention in the chapters in this book, there are other contextual factors operating outside the borders of a state or region that students of power sharing may want to consider as having the potential to shape power sharing and power relations. One that merits consideration is the influence of the area in which a country considering adopting a power-sharing arrangement finds itself. Lars-Erik Cederman, Kristian Skrede Gleditsch, and Julian Wucherpfennig (2018) found, for example, that the more inclusive a world region is, the more likely a shift to power sharing becomes, although they did not identify the impact this factor has on the nature of power relations. Also of interest to explore would be whether ideologies or movements that are prevalent at the regional or hemispheric level have any effect on power-sharing arrangements and power relations. Design

One important way in which power-sharing arrangements differ is in their design. Here, we address two fundamental variations in design: the types of power-sharing measures included as part of a settlement and the inclusiveness of those measures. Power-sharing agreements negotiated as a means of ending an intrastate conflict may contain as few as one of the four forms of power sharing focused on by the authors in this book or as many as four. In addition, a power-sharing arrangement can contain a number of power-sharing institutions of one form or another—for example, the leader of a rebel group may be appointed as head of a ministry and a system of electoral proportional representation may be installed, each of which is an instance of political power sharing.2 Ideally, the types of power-sharing measures included in a settlement should be selected on the basis of their ability to halt the fighting by structuring power relations in a manner that mitigates security concerns and addresses the primary cleavages at the center of the conflict. In practice, as demonstrated in this volume, elites may insist on the inclusion of powersharing measures meant to favor their political or material interests, measures designed to address a cleavage may generate unintended effects that

224

Caroline A. Hartzell and Andreas Mehler

weaken some actor’s ability to play by peaceful political rules, and the effectiveness of other measures may be undermined by entities ranging from local actors to global forces. The chapters in this volume also make clear that power-sharing arrangements can vary in terms of their inclusiveness. A number of the chapters indicate that, frequently, only belligerents were included as part of a power-sharing deal, with civilian actors such as political parties or civil society organizations relegated to the sidelines. Additionally, work by the contributors suggests that power-sharing arrangements were often inclusive in the sense of attempting to address central conflict cleavages—for example, ethnoterritorial, religious, or ideological—but that little effort generally was made to include actors representative of secondary cleavages in the peace agreements. As we noted in the introduction, in some instances power-sharing arrangements can have an impact on power relations even in those instances in which an actor is not included as part of the arrangement. Thus, it is difficult, on the face of it, to conclude that less inclusive settlements are necessarily less effective than more inclusive settlements. It seems unlikely, however, that changes in power relations that are unintended in nature, perhaps as a result of the exclusion of some actors from a settlement, will bode positively for the duration or quality of the peace. Temporality

Another manner in which power-sharing arrangements differ is in the length of time that they are meant to be in place. Some power-sharing arrangements can be short term in nature, and may even be informal, while others, codified in constitutions, are meant to be permanent. It seems reasonable to expect that the distinction between interim measures and permanent institutions can have an effect on actors’ behavior and, thus, on the nature of postconflict power relations. Interim power-sharing measures, for example, could either prompt those who gain a position in government to prepare to win elections by trying to gain popular legitimacy, creating functioning support structures, building alliances, or influencing the rules of the game to stay in power for a longer period of time—or induce actors in positions of power to enrich themselves as fast as they can if they believe they face the prospect of being excluded from power following elections. Some changes in actors’ behavior may materialize only after a long period of time. Hayes and Nagle nicely illustrate this process in their chapter on forms of consociational power sharing in Ireland and Lebanon. Political power sharing is also strongly associated with the transformation of rebel movements into political parties. Previous research has shown that, when this process, which takes some time to play out, is successful, it tends to generate both winners and losers within former rebel movements (de

The What, How, Where, and Who

225

Zeeuw 2007). There also is liable to be another facet to temporality, however, with the likelihood being lower that wartime events will shape public debate or the individual perceptions of wartime leaders involved in postwar power sharing with a greater of passage of time.

Policy Suggestions Drawing on the contributions to this book, we end this chapter by offering the following suggestions to the policy community regarding the use of power sharing as a means of ending civil wars. Think through and, where possible, trace out, the various effects that power sharing could potentially have on power relations. Power-sharing arrangements are designed to work—to help end the fighting and preserve the peace—by altering power relations. As we attempted to demonstrate in this book, power-sharing measures can alter power relations between or among actors in ways that are expected. They can also do so in unexpected ways, however, including, at times, within actors or even among actors that were not directly included as part of a power-sharing deal. Attempting to anticipate the effects that various types of power-sharing measures may have on power relations before they are included as part of a peace settlement can give settlement architects, policymakers, and stakeholders an opportunity to alter the settlement design where possible. In those instances in which significant alteration of power-sharing mechanisms proves difficult, actors can at least prepare to attempt to mitigate any of the negative effects on power relations that certain power-sharing arrangements might be expected to produce. Do not oversell the potential that power sharing has for altering power relations in ways that will help to create a lasting and just peace. As some of the contributors to the volume made clear, in some cases it is factors other than power-sharing arrangements that exercise a significant influence on power relations. In other cases, power-sharing measures do not appear to have the types of effects attributed to them. It may even be the case that actors fail to adopt some power-sharing measures because they expect that they will not be effective. Whatever the reason, power-sharing measures clearly are not omnipotent. Overselling their utility could be risky if it leads actors to overlook other possible solutions to a conflict or, once having oversold them, fail to consider power-sharing measures in those instances in which they might well be viable. Buttress functional power-sharing arrangements with supportive structures and aid. In those instances in which power-sharing measures appear to alter power relations in ways that help yield a durable and high-quality peace, the policy community should take action to help support the arrangement. Such support might involve the provision of peacekeeping forces to provide a

226

Caroline A. Hartzell and Andreas Mehler

sufficient sense of security for actors to fully implement the relevant powersharing measures, financial support (e.g., of rebel groups as they form political parties or funding for a decentralized unit that has been granted autonomy during a transition period), or a reprieve from the implementation of neoliberal economic policies that could prove corrosive to power relations. In conclusion, neither the suggestions made above nor the analyses provided by the contributors to this volume should be considered the final word on the effect that power sharing has on power relations in postconflict states. Our hope is that this book will serve as a starting point for a conversation among scholars, policymakers, and stakeholders regarding the possibilities and limitations that power sharing holds as a means of altering the postconflict landscape in a manner commensurate with building a stable and just peace.

Notes 1. Power-sharing arrangements are generally thought to work by helping to strengthen the weaker party to a peace settlement by giving it access to some element(s) of state power. 2. For a detailed examination of the influence that the choice of institutions and rules made as part of a political power-sharing settlement can exercise on postconflict stability, governance, and “justice, fairness, and democracy” (see McGarry 2017, 268).

Bibliography

Abinales, Patricio. 2000. Making Mindanao: Cotabato and Davao in the Formation of the Philippine State. Quezon City: Ateneo de Manila University Press. ———. 2004. “American Military Presence in Southern Philippines: A Comparative Historical Overview.” Working Paper No. 7. Honolulu: East West Center. ACCORD (African Centre for the Constructive Resolution of Disputes). 2009. “Peace Agreements and Durable Peace in Africa.” Durban: ACCORD. Akhavan, Payam. 2001. “Beyond Impunity: Can International Criminal Justice Prevent Future Atrocities?” American Journal of International Law 95: 7–31. Anderson, David. 2010. “Majimboism: The Troubled History of an Idea.” In Our Turn to Eat: Politics in Kenya Since 1950. Daniel Branch, Nic Cheeseman, and Leigh Gardner, eds. Berlin: Lit Verlag. Andeweg, Rudy. 2000. “Consociational Democracy.” Annual Review of Political Science 3: 509–536. Aning, Kwesi, and Samuel Atuobi. 2011. “The Neglected Economic Dimensions of ECOWAS’s Negotiated Peace Accords in West Africa.” Africa Spectrum 46, 3: 27–44. Ansori, Mohammad Hasan. 2012. “From Insurgency to Bureaucracy: Free Aceh Movement, Aceh Party and the New Face of Conflict.” Stability: International Journal of Security and Development 1, 1: 31–44. Arnould, Valerie. Forthcoming. The Politics of Transitional Justice and Peacebuilding: The Case of the Democratic Republic of Congo. London: Routledge. Asal, Victor, Mitchell Brown, and Andrew Dalton. 2012. “Why Split? Organizational Splits Among Ethnopolitical Organizations in the Middle East.” Journal of Conflict Resolution 56, 1: 94–117. Atlas, Pierre M., and Roy Licklider. 1999. “Conflict Among Former Allies After Civil War Settlement: Sudan, Zimbabwe, Chad, and Lebanon.” Journal of Peace Research 36, 1: 35–54. Autesserre, Séverine. 2010. The Trouble with the Congo: Local Violence and the Failure of International Peacebuilding. Cambridge: Cambridge University Press. Baaz, Maria Eriksson, and Judith Verweijen. 2013. “The Volatility of a Half-Cooked Bouillabaisse: Rebel-Military Integration and Conflict Dynamics in the Eastern DRC.” African Affairs 112, 449: 563–582.

227

228

Bibliography

Banks, Arthur S., Thomas. C. Muller, and William. R. Overstreet, eds. 2017. Political Handbook of the World. Washington: CQ Press. Bakke, Kristin M., Kathleen Gallagher Cunningham, and Lee J.M. Seymour. 2012. “A Plague of Initials: Fragmentation, Cohesion, and Infighting in Civil Wars.” Perspectives on Politics 10, 2: 265–283. Bapat, Navin, and Kanisha Bond. 2012. “Alliances Between Militant Groups.” British Journal of Political Science 42, 4: 793–824. Barbour, Bernadette, and Pascale Salameh. 2009. “Consanguinity in Lebanon: Prevalence, Distribution and Determinants.” Journal of Biosocial Sciences 41, 4: 505–517. Barron, Patrick, Erman Rahmant, and Kharisma Nugroho. n.d. The Contested Corners of Asia: Subnational Conflict and International Development Assistance—The Case of Aceh, Indonesia. Executive Summary. Asia Foundation. https://asiafoundation.org/resources/pdfs/ContestedCornersOfAsia.pdf. Bayart, Jean-François. 1989. L’Etat en Afrique: la politique du ventre. Paris: Fayard. Beirne, Maggie, and Fionnuala Ní Aoláin. 2009. “Rights After the Revolution: Progress or Backslide After the Good Friday Agreement?” In Human Rights and Conflict Resolution in Context: Colombia, Sierra Leone, and Northern Ireland. Eileen F. Babbitt and Ellen L. Lutz, eds. Syracuse: Syracuse University Press. Beissinger, Mark R. 2002. Nationalist Mobilization and the Collapse of the Soviet State. New York: Cambridge University Press. Bell, Christine. 2000. Peace Agreements and Human Rights. Oxford: Oxford University Press. Bennett, D. Scott, and Christian Davenport. 2003. “MARGene: Minorities at Risk Data Generation and Management Program.” www.cidcm.umd.edu/mar/data.asp. Bentley, G. Carter. 1991. “People Power and After in the Islamic City of Marawi.” In From Marcos to Aquino: Local Perspectives on Political Transition in the Philippines. Benedict Kirkvliet and Resil Mojares, eds. Quezon City: Ateneo de Manila University Press. ———. 1994. “Mohamad Ali Dimaporo: A Modern Maranao Datu.” In An Anarchy of Families: State and Family in the Philippines. Alfred McCoy, ed. Quezon City: Ateneo de Manila University Press. Bilefsky, Dan. 2014. “Protests over Government and Economy Roil Bosnia.” New York Times, February 8. Binningsbø, Helga Malmin. 2013. “Power Sharing, Peace and Democracy: Any Obvious Relationships?” International Area Studies Review 16, 1: 89–112. Binningsbø, Helga Malmin, and Siri Aas Rustad. 2012. “Sharing the Wealth: A Pathway to Peace or a Trail to Nowhere?” Conflict Management and Peace Science 29, 5: 547–566. Bogaards, Matthijs. 2013. “The Choice for Proportional Representation: Electoral System Design in Peace Agreements.” Civil Wars 15, S1: 71–87. Box-Steffensmeier, Janet M., and Bradford S. Jones 1997. “Time Is of the Essence: Event History Models in Political Science.” American Journal of Political Science 41, 4: 1414–1461. ———. 2004. Event History Modeling—A Guide for Social Scientists. Cambridge: Cambridge University Press. Box-Steffensmeier, Janet M., Dan Reiter, and Christopher Zorn. 2003. “Nonproportional Hazards and Event History Analysis in International Relations.” Journal of Conflict Resolution 47, 1: 33–53. Brancati, Dawn. 2006. “Decentralization: Fueling the Fire or Dampening the Flames of Ethnic Conflict and Secessionism.” International Organization 60, 3: 651–685.

Bibliography

229

———. 2009. Peace by Design: Managing Intrastate Conflict Through Decentralization. Oxford: Oxford University Press. Brass, Paul R. 1991. Ethnicity and Nationalism: Theory and Comparison. Thousand Oaks, CA: Sage Press. Brown, Stephen. 2011. “The Rule of Law and the Hidden Politics of Transitional Justice in Rwanda.” In Peacebuilding and Rule of Law in Africa: Just Peace? Chandra Lekha Sriram, Olga Martin-Ortega, and Johanna Herman, eds. London: Routledge. Brown, Stephen, and Chandra Lekha Sriram. 2014. “‘The Big Fish Won’t Fry Themselves’: Criminal Accountability for Post-Election Violence in Kenya.” African Affairs 111: 244–260. Brown, Stephen, and Marie-Jöelle Zahar. 2008. “Committing to Peace: Soft Guarantees and Alternative Approaches to Power Sharing in Angola and Mozambique.” Journal of Peacebuilding and Development 4: 75–88. Brownlee, Jason. 2007. Authoritarianism in an Age of Democratization. New York: Cambridge University Press. Cammett, Melani. 2014. Compassionate Communalism: Welfare and Sectarianism in Lebanon. New York: Cornell University Press. Cammett, Melani, and Edmund Malesky. 2012. “Power Sharing in Postconflict Societies: Implications for Peace and Governance.” Journal of Conflict Resolution 56, 6: 982–1016. Canuday, Jowel. 2014. “Big Wars, Small Wars: The Interplay of Large-Scale and Community Armed Conflicts in Five Central Mindanao Communities.” In Rido: Clan Feuding and Conflict Management in Mindanao. Wilfredo Torres III, ed. Quezon City: Ateneo de Manila University Press. Carey, Sabine C., Mark Gibney, and Steven C. Poe. 2010. The Politics of Human Rights: The Quest for Dignity. Cambridge: Cambridge University Press. Carey, Sabine C., Neil J. Mitchell, and Will Lowe. 2013. “States, the Security Sector, and the Monopoly of Violence: A New Database on Pro-Government Militias.” Journal of Peace Research 50, 2: 249–258. Carroll, Kevin J. 2003. “On the Use and Utility of the Weibull Model in the Analysis of Survival Data.” Control Clinical Trials 24, 6: 682–701. Casper, Gretchen. 1995. Fragile Democracies: The Legacies of Authoritarian Rule. Pittsburgh: University of Pittsburgh Press. Cederman, Lars-Erik, Kristian Skrede Gleditsch, and Halvard Buhaug. 2013. Inequality, Grievances, and Civil War. New York: Cambridge University Press. Cederman, Lars-Erik, Kristian Skrede Gleditsch, and Julian Wucherpfennig. 2018. “The Diffusion of Inclusion: An Open-Polity Model of Ethnic Power Sharing.” Comparative Political Studies 51, 10: 1279–1313. Chabal, Patrick. 1992. Power in Africa: An Essay in Political Interpretation. London: Macmillan. Chamie, Joseph. 1980. “Religious Groups in Lebanon.” International Journal of Middle East Studies 11, 2: 175–187. Cheeseman, Nic. 2008. “The Kenyan Elections of 2007: An Introduction.” Journal of Eastern African Studies 2, 2: 166–184. ———. 2011. “The Internal Dynamics of Power-Sharing in Africa.” Democratization 18, 2: 336–365. Chopra, Tanja. 2009. “When Peacebuilding Contradicts Statebuilding: Notes from the Arid Lands of Kenya.” International Peacekeeping 16, 4: 531–545. Christia, Fotini. 2012. Alliance Formation in Civil War. Cambridge: Cambridge University Press. Cingranelli, David L., and David L. Richards. 2010. “The Cingranelli and Richards (CIRI) Human Rights Data Project.” Human Rights Quarterly 32, 2: 401–424.

230

Bibliography

Cleves, Mario, Roberto G. Gutierrez, William Gould, and Yulia V. Marchenko. 2010. An Introduction to Survival Analysis Using Stata, 3rd ed. College Station, TX: StataCorp LP. Coakley, John. 2009. “Implementing Consociation in Northern Ireland.” In Consociational Theory: McGarry and O’Leary and the Northern Ireland Conflict. Rupert Taylor, ed. London: Routledge. Cole, Wade M. 2012. “Human Rights as Myth and Ceremony? Reevaluating the Effectiveness of Human Rights Treaties, 1981–2007.” American Journal of Sociology 117, 4: 1131–1171. Cole, Wade M., and Francisco O. Ramirez. 2013. “Conditional Decoupling: Assessing the Impact of National Human Rights Institutions, 1981 to 2004.” American Sociological Review 78, 4: 702–725. Collier, Paul, and Anke Hoeffler. 2004. “Greed and Grievance in Civil War.” Oxford Economic Papers 56: 563–595. Collier, Paul, Anke Hoeffler, and Måns Söderbom. 2008. “Post-Conflict Risks.” Journal of Peace Research 45, 4: 461–478. Cook, Malcolm, and Kit Collier. 2006. “Mindanao: A Gamble Worth Taking.” Paper No. 17. Double Bay, New South Wales: Lowy Institute for International Policy. Coppedge, Michael, John Gerring, Staffan I. Lindberg, Svend-Erik Skaaning, Jan Teorell, David Altman, Michael Bernhard, et al. 2016. “V-Dem Codebook v6.” Varieties of Democracy (V-Dem) Project. Cornell, Agnes, and Michelle D’Arcy. 2014. “Plus Ça Change? County-Level Politics in Kenya After Devolution.” Journal of Eastern African Studies 8, 1: 173–191. Cunningham, David. 2006. “Veto Players and Civil War Duration” American Journal of Political Science 50, 4: 875–892. Cunningham, David E., Kristian Skrede Gleditsch, and Idean Salehyan. 2013. “Non-State Actors in Civil Wars: A New Dataset.” Conflict Management and Peace Science 30, 5: 516–531. Cunningham, Kathleen Gallagher. 2011. “Divide and Conquer or Divide and Concede: How Do States Respond to Internally Divided Separatists?” American Political Science Review 105, 2: 275–297. ———. 2013. “Actor Fragmentation and Civil War Bargaining: How Internal Divisions Generate Civil Conflict.” American Journal of Political Science 57, 3: 659–672. ———. 2014. Inside the Politics of Self-Determination. Oxford: Oxford University Press. Cunningham, Kathleen Gallagher, Kristin M. Bakke, and Lee Seymour. 2012. “Shirts Today, Skins Tomorrow: Dual Contests and the Effects of Fragmentation in Self-Determination Disputes.” Journal of Conflict Resolution 56, 1: 57–93. Curtis, Devon. 2013. “The International Peacebuilding Paradox: Power-Sharing and Post-Conflict Governance in Burundi.” African Affairs 112, 446: 72–91. Daalder, Hans. 1974. “The Consociational Democracy Theme.” World Politics 26, 4: 604–621. Dahl, Robert. 1957. “The Concept of Power.” Systems Research and Behavioral Science 2, 3: 201–215. Davenport, Christian. 1995. “Multi-Dimensional Threat Perception and State Repression: An Inquiry into Why States Apply Negative Sanctions.” American Journal of Political Science 39, 3: 683–713. ———. 1999. “Human Rights and the Democratic Proposition.” Journal of Conflict Resolution 43, 1: 92–116. De Juan, Alexander. 2013. “Devolving Ethnic Conflicts: The Role of Subgroup Identities for Institutional Intergroup Settlements.” Civil Wars 15, 1: 78–99.

Bibliography

231

de Villa, Renato. 1996. “Securing Economic Growth.” In Fookien Times Philippines Yearbook 1995–1996. Manila: Fookien Times. de Zeeuw, Jeroun. 2007. From Soldiers to Politicians: Transforming Rebel Movements After Civil War. Boulder: Lynne Rienner. DeNardo, James. 1985. Power in Numbers. Princeton: Princeton University Press. Donais, Timothy. 2012. Peacebuilding and Local Ownership: Post-Conflict Consensus-Building. New York: Routledge. Doyle, Michael, and Nicholas Sambanis. 2006. Making War and Building Peace: United Nations Peace Operations. Princeton: Princeton University Press. Dreher, Axel. 2006. “Does Globalization Affect Growth? Evidence from a New Index of Globalization.” Applied Economics 38, 10: 1091–1110. Driscoll, Jesse. 2012. “Commitment Problems or Bidding Wars? Rebel Fragmentation as Peace Building.” Journal of Conflict Resolution 56, 1: 118–149. Eck, Kristine, and Lisa Hultman. 2007. “Violence Against Civilians in War.” Journal of Peace Research 44, 2: 617–631. Fahey, Tony, Bernadette C. Hayes, and Richard Sinnott. 2006. Conflict and Consensus: A Study of Values and Attitudes in the Republic of Ireland and Northern Ireland. Leiden: Brill. Faour, Muhammad A. 2007. “Religion, Demography, and Politics in Lebanon.” Middle Eastern Studies 43, 6: 909–921. Ferrer, Mirriam C. 1999. Integration of the MNLF Forces into the PNP and AFP: Integration Without Demobilization and Disarmament. UP Project on Assessment of the Implementation of the GRP-MNLF Peace Agreement, Phase I. Quezon City: UP Center for Integrative and Development Studies. Findley, Michael, and Peter Rudloff. 2012. “Combatant Fragmentation and the Dynamics of Civil War.” British Journal of Political Science 42, 4: 879–901. Fine, Jason P., and Robert J. Gray. 1999. “A Proportional Hazards Model for the Subdistribution of a Competing Risk.” Journal of the American Statistical Association 94, 446: 496–509. Finlay, Andrew. 2011. Governing Ethnic Conflict: Consociation, Identity, and the Price of Peace. New York: Routledge. Fjelde, Hanne, and Desirée Nilsson. 2012. “Rebels Against Rebels: Explaining Violence Between Rebel Groups.” Journal of Conflict Resolution 56, 4: 604–628. Förster, Till. 2010. “Maintenant, on Sait Qui Est Qui: Statehood and Political Reconfiguration in Northern Côte d’Ivoire.” Development and Change 41, 4: 699–722. Fortna, Virginia Page. 2004. “Does Peacekeeping Keep Peace? International Intervention and the Duration of Peace After Civil Wars.” International Studies Quarterly 48, 4: 269–292. Foucault, Michel. 1978. The History of Sexuality, vol. 1: An Introduction. New York: Pantheon Books. ———. 1991. Discipline and Punish: The Birth of a Prison. London: Penguin. Frank, Emily. 2002. “A Participatory Approach for Local Peace Initiatives: The Lodwar Border Harmonization Meeting.” Africa Today 49, 4: 69–87. Frazier, Derrick V., and William J. Dixon. 2006. “Third Party Intermediaries and Negotiated Settlements, 1946–2000.” International Interactions 32: 385–408. Gent, Stephen E. 2011. “Relative Rebel Strength and Power Sharing in Intrastate Conflicts.” International Interactions 37, 2: 215–228. Glassmyer, Katherine, and Nicholas Sambanis. 2008. “Rebel-Military Integration and Civil War Termination.” Journal of Peace Research 45, 3: 365–384. Gleditsch, Kristian S. 2002. “Expanded Trade and GDP Data.” Journal of Conflict Resolution 46, 5: 712–724.

232

Bibliography

Gleditsch, Nils Petter, Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg, and Håvard Strand. 2002. “Armed Conflict 1946–2001: A New Dataset.” Journal of Peace Research 39, 5: 615–637. Goemans, Henk E., Kristian Skrede Gleditsch, and Giacomo Chiozza. 2009. “Introducing Archigos: A Dataset of Political Leaders.” Journal of Peace Research 46, 2: 269–283. Greenhill, Kelly M., and Solomon Major. 2006–2007. “The Perils of Profiling: Civil War Spoilers and the Collapse of Intrastate Peace Accords.” International Security 31: 7–40. Gutierrez, Eric. 2013. “Bandits, Kidnappers and Bosses: Kidnappers of the Southern Philippines.” In Out of the Shadows: Violent Conflict and the Real Economy of Mindanao. Francisco Lara Jr. and Steven Schoofs, eds. Manila: International Alert. Haddad, Simon. 2003. The Palestinian Impasse in Lebanon: The Politics of Refugee Integration. Portland: Sussex Academic Press. ———. 2009. “Lebanon: From Consociationalism to Conciliation.” Nationalism and Ethnic Politics 15, 3–4: 398–416. Hafner-Burton, Emilie M., and Kiyoteru Tsutsui. 2005. “Human Rights in a Globalizing World: The Paradox of Empty Promises.” American Journal of Sociology 110, 5: 1373–1411. Hale, Henry. 2005. “Why Not Parties? Electoral Markets, Party Substitutes, and Stalled Democratization in Russia.” Comparative Politics 37, 2: 147–166. Hall, Rosalie. 2013. “What Works: Documenting Local Conflict Dynamics and Internal Security Operations Strategies in Maguindanao Province, Philippines.” Paper presented at the Inter-University Seminar on Armed Forces and Society, Chicago, October 25–27. ———. 2014. “From Rebels to Soldiers: An Analysis of the Philippine Policy of Integrating Former Moro National Liberation Front Combatants into the Armed Forces.” In New Armies from Old: Merging Competing Military Forces After Civil Wars. Roy Licklider, ed. Washington, DC: Georgetown University Press. Hannum, Hurst. 2009. “The Impact of European Human Rights Jurisprudence on the Conflict in Northern Ireland.” In Human Rights and Conflict Resolution in Context: Colombia, Sierra Leone, and Northern Ireland. Eileen F. Babbitt and Ellen L. Lutz, eds. Syracuse: Syracuse University Press. Harbom, Lotta, Erik Melander, and Peter Wallensteen. 2008. “Dyadic Dimensions of Armed Conflict, 1946–2007.” Journal of Peace Research 45, 5: 697–710. Hartmann, Christof. 2013. “Territorial Power Sharing and the Regulation of Conflict in Africa.” Civil Wars 15, S1: 123–143. Hartzell, Caroline. 1999. “Explaining the Stability of Negotiated Settlements to Intrastate Wars.” Journal of Conflict Resolution 43, 1: 3–22. Hartzell, Caroline A., and Matthew Hoddie. 2003. “Institutionalizing Peace: Power Sharing and Post–Civil War Conflict Management.” American Journal of Political Science 47, 2: 318–332. ———. 2007. Crafting Peace: Power-Sharing Institutions and the Negotiated Settlement of Civil Wars. University Park: Penn State University Press. ———. 2015. “The Art of the Possible: Power Sharing and Post–Civil War Democracy.” World Politics 67, 1: 37–71. Hayes, Bernadette C., and Ian McAllister. 2013. Conflict to Peace: Politics and Society in Northern Ireland over Half a Century. Manchester: Manchester University Press.

Bibliography

233

Haysom, Nicholas, and Sean Kane. 2009. “Negotiating Natural Resources for Peace: Ownership, Control and Wealth-Sharing.” Briefing Paper. Geneva: Henry Dunant Centre for Humanitarian Dialogue. Hegre, Håvard. 2008. “Gravitating Toward War: Preponderance May Pacify, but Power Kills.” Journal of Conflict Resolution 52, 4: 566–589. Hegre, Håvard, Lisa Hultman, and Håvard Mokleiv Nygård. 2011. “Simulating the Effect of Peacekeeping Operations 2010–2035.” Social Computing, BehavioralCultural Modeling and Prediction: Lecture Notes in Computer Science 6589, 1: 325–332. Heitz, Kathrin. 2009. “Power-Sharing in the Local Arena: Man—A Rebel-Held Town in Western Côte d’Ivoire.” Africa Spectrum 44, 3: 109–131. Hernandez, Carolina. 1988. “Reforming the Military and Police Agencies in the Philippines.” Philippine Journal of Public Administration 32, 1–2: 152–162. Herz, John H. 1950. “Idealist Internationalism and the Security Dilemma.” World Politics 2, 2: 157–180. Heston, Alan, Robert Summers, and Bettina Aten. 2009. “Penn World Table Version 6.3.” Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania. Hirblinger, Andreas, and Claudia Simons. 2014. “The Good, the Bad and the Powerful: The Coaxing Up of the Local in Peacebuilding.” Paper presented at the annual meeting of the International Studies Association, Toronto, March 26–29. Hoddie, Matthew, and Caroline A. Hartzell. 2003. “Civil War Settlements and the Implementation of Military Power-Sharing Arrangements.” Journal of Peace Research 40, 3: 303–320. ———. 2005. “Power Sharing in Peace Settlements: Initiating the Transition from Civil War.” In Sustainable Peace: Power and Democracy After Civil Wars. Philip G. Roeder and Donald S. Rothchild, eds. Ithaca: Cornell University Press. Höhn, Sabine. 2014. “New Start or False Start? The ICC and Electoral Violence in Kenya.” Development and Change 45, 3: 565–588. Højbjerg, Christian K. 2010. “Victims and Heroes: Manding Historical Imagination in a Conflict-Ridden Border Region (Liberia-Guinea).” In The Powerful Presence of the Past. Jacqueline Knörr and Wilson Trajano Filho, eds. Leiden: Brill. Holbrooke, Richard. 1998. To End a War. New York: Modern Library. Horowitz, Donald. 2008. “Conciliatory Institutions and Constitutional Processes in Post-Conflict States.” William and Mary Law Review 49, 4: 1213–1248. Human Rights Watch. 2009. “Selling Justice Short: Why Accountability Matters to Peace.” www.hrw.prg/print/reports/2009/07/07/selling-justice-short. ———. 2013. “Kenya: Rights Defenders Under Attack.” www.hrw.org/news /2013/10/04/kenya-rights-defenders-under-attack. Humphreys, Macartan, and Jeremy M. Weinstein. 2007. “Demobilization and Reintegration.” Journal of Conflict Resolution 51, 4: 531–567. Hussin, Parouk. 2005. “Excerpts from the Presentation as ARMM Governor: Meeting with the Committee of Eight.” Indonesian Embassy, Makati, Philippines, August 15. ICC (International Criminal Court). 2014a. “Situation in Côte d’Ivoire.” www.icccpi .int/EN_Menus/ICC/Situations%20and%20Cases/Situations/ICC0211/Pages /situation%20index.aspx. ———. 2014b. “Situation in the Democratic Republic of Congo.” www.icc-cpi.int /en_menus/icc/situations%20and%20cases/situations/situation%20icc%200104 /Pages/situation%20index.aspx.

234

Bibliography

———. 2014c. “Situation in the Central African Republic.” www.icc-cpi.int /EN_Menus/ICC/Situations%20and%20Cases/Situations/Situation%20ICC% 200105/Pages/situation%20icc-0105.aspx. ———. 2014d. “Situation in the Republic of Kenya.” www.icc-cpi.int/en_menus /icc/situations%20and%20cases/situations/situation%20icc%200109/Pages /situation%20index.aspx. ———. 2014e. “Situation in the Republic of Mali.” www.icc-cpi.int/en_menus/icc /situations%20and%20cases/situations/icc0112/Pages/situation%20index.aspx. ———. 2014f. “Situation in Uganda.” www.icc-cpi.int/uganda. ———. 2014g. “Situation in Darfur, Sudan.” www.icc-cpi.int/en_menus/icc /situations%20and%20cases/situations/situation%20icc%200205/Pages /situation%20icc-0205.aspx. ICC, OtP (International Criminal Court, Office of the Prosecutor). 2013. “Situation in Mali: Article 53 (1) Report.” www.icc-cpi.int/en_menus/icc/situations%20 and%20cases/situations/icc0112/Documents/SASMaliArticle53_1Public ReportENG16Jan2013.pdf. Ishiyama, John. 2005. “The Former Marxist Leninist Parties in Africa After the End of the Cold War.” Acta Politica 40, 4: 459–479. Ishiyama, John, and Anna Batta. 2011. “Swords into Plowshares: Rebel Legacies, Electoral Competition and the Transformation of the Communist Party of Nepal.” Communist and Post-Communist Studies 44, 4: 369–379. Jacildo, Nerlyne C. 2003. “Experiences of MNLF Integrees in Basilan and Zamboanga: Issues and Problems.” Master’s thesis, University of the Philippines– Diliman. Janda, Kenneth, Robert Harmel, Christine Edens, and Patricia Goff. 1995. “Changes in Party Identity: Evidence from Party Manifestos.” Party Politics 1: 171–196. Jarman, Neil, and Rachel Monaghan. 2003. Racist Harassment in Northern Ireland. Belfast: Institute for Conflict Research. Jarstad, Anna. 2008. “Power Sharing: Former Enemies in Joint Government.” In From War to Democracy: Dilemmas of Peacebuilding. Anna Jarstad and Timothy Sisk, eds. Cambridge: Cambridge University Press. Jarstad, Anna, and Desirée Nilsson. 2008. “From Words to Deeds: The Implementation of Powersharing Pacts in Peace Accords.” Conflict Management and Peace Science 25, 3: 206–223. Jarstad, Anna, and Timothy Sisk, eds. 2008. From War to Democracy: Dilemmas of Peacebuilding. Cambridge: Cambridge University Press. Jarstad, Anna K., and Ralph Sundberg. 2007. “Peace by Pact: Data on the Implementation of Peace Agreements.” In Globalization and Challenges to Building Peace. Ashok Swain, Ramses Amer, and Joakim Öjendal, eds. London: Anthem Press. Joshi, Madhav, and John Darby. 2013. “Introducing the Peace Accords Matrix (PAM): A Database of Comprehensive Peace Agreements and Their Implementation, 1989–2007.” Peacebuilding 1, 2: 256–274. Joshi, Madhav, and T. David Mason. 2011. “Civil War Settlements, Size of Governing Coalition and Durability of Peace in the Post–Civil War States.” International Interactions 37, 4: 388–413. Kalyvas, Stathis. 2003. “The Ontology of ‘Political Violence’: Action and Identity in Civil Wars.” Perspectives on Politics 1, 3: 475–494. ———. 2006. The Logic of Violence in Civil War. New York: Cambridge University Press. ———. 2008. “Ethnic Defection in Civil War.” Comparative Political Studies 41: 1043–1068.

Bibliography

235

Kaufmann, Chaim. 1996. “Possible and Impossible Solutions to Ethnic Civil Wars.” International Security 20, 4: 136–175 Keck, Margaret E., and Kathryn Sikkink. 1998. Activists Beyond Borders: Advocacy Networks in International Politics. Cambridge: Cambridge University Press. Keesing’s Record of World Events Online. 2012. https://www.eui.eu /Research /Library/ElectronicResources/E-Resources/2012/01-14-KeesingsRecordof WorldEvents. Keith, Linda Camp. 1999. “The United Nations International Covenant on Civil and Political Rights: Does It Make a Difference in Human Rights Behavior?” Journal of Peace Research 36, 1: 95–118. Kerr, Rachel. 2004. The International Criminal Tribunal for the Former Yugoslavia: An Exercise in Law, Politics and Diplomacy. Oxford: Oxford University Press. Khalaf, Samir. 2012. Lebanon Adrift: From Battleground to Playground. London: Saqi Books. King, Charles. 2004. “The Micropolitics of Social Violence.” World Politics 56: 431–455. Knio, Karim. 2005. “Lebanon: Cedar Revolution or Neo-Sectarian Partition?” Mediterranean Politics 10, 2: 225–231. Krause, Peter. 2013. “The Political Effectiveness of Non-State Violence: A TwoLevel Framework to Transform a Deceptive Debate.” Security Studies 22, 2: 259–294. Kreutz, Joakim. 2010. “How and When Armed Conflicts End: Introducing the UCDP Conflict Termination Dataset.” Journal of Peace Research 47, 2: 243–250. Kydd, Andrew F., and Barbara Walter. 2002. “Sabotaging the Peace: The Politics of Extremist Violence.” International Organization 56: 263–296. Lake, David, and Donald Rothchild. 1996. “Containing Fear: The Origins and Managements of Ethnic Conflict.” International Security 21, 2: 41–75. ———. 2005. “Territorial Decentralization and Civil War Settlement.” In Sustainable Peace: Power and Democracy After Civil Wars. Philip G. Roeder and Donald S. Rothchild, eds. Ithaca: Cornell University Press. Lamont, Christopher. 2010. “Defiance or Strategic Compliance? The Post-Tuđman Croatian Democratic Union and the International Criminal Tribunal for the Former Yugoslavia.” Europe-Asia Studies 62: 1683–1705. Lara, Francisco, Jr., and Steven Schoofs. 2013. Out of the Shadows: Violent Conflict and the Real Economy of Mindanao. Manila: International Alert. Larkin, Craig. 2012. Memory and Conflict in Lebanon: Remembering and Forgetting the Past. London: Routledge. Lasswell, Harold D. 1950. Politics: Who Gets What, When, How. New York: Peter Smith. Lawrence, Adria. 2010. “Triggering Nationalist Violence: Competition and Conflict in Uprisings Against Colonial Rule.” International Security 35, 2: 88–122. Licklider, Roy. 1995. “The Consequences of Negotiated Settlements in Civil Wars, 1945–1993.” American Political Science Review 89, 3: 681–690. ———, ed. 2014. New Armies from Old: Merging Competing Military Forces After Civil Wars. Washington, DC: Georgetown University Press. Lidasan, Nasser P. 2006. “The Integration of MNLF and MILF Combatants into the Philippine Armed Forces: Implications to Conflict Resolution.” Master’s thesis, University of Bradford. Lijphart, Arend. 1975. “The Northern Ireland Problem: Cases, Theories, and Solutions.” British Journal of Political Science 5, 1: 83–106. ———. 1977. Democracy in Plural Societies. New Haven: Yale University Press.

236

Bibliography

———. 2004. “Consociational Design for Divided Societies.” Journal of Democracy 15, 2: 96–101. Lujala, Pävi, and Siri Aas Rustad, eds. 2012. High-Value Natural Resources and Post-Conflict Peacebuilding. London: Earthscan. Lustick, Ian. 1993. Unsettled States, Disputed Lands: Britain and Ireland, France and Algeria, Israel and the West Bank-Gaza. Ithaca: Cornell University Press. Mac Ginty, Roger, and Oliver P. Richmond. 2013. “The Local Turn in Peace Building: A Critical Agenda for Peace.” Third World Quarterly 34, 5: 763–783. Makinano, Merliza M., and Alfredo Lubang. 2000. “Disarmament, Demobilisation and Reintegration: The Mindanao Experience.” In South Asia at Gun Point: Small Arms and Light Weapons Proliferation. Dipankar Banerjee, ed. Colombo, Sri Lanka: Regional Centre for Strategic Studies. Mani, Rama 2002. Beyond Retribution: Seeking Justice in the Shadows of War. Cambridge, UK: Polity Press. Manning, Carrie. 2003. “Local Level Challenges to Post-Conflict Peacebuilding.” International Peacekeeping 10, 3: 25–43. ———. 2004. “Armed Opposition Groups into Political Parties: Comparing Bosnia, Kosovo and Mozambique.” Studies in Comparative International Development 39: 54–76. ———. 2007. “Party-Building on the Heels of War: El Salvador, Bosnia, Kosovo and Mozambique.” Democratization 14: 1–20. Marcon, Giulio, Sergio Andreis, Thorsten Bonacker, Christian Braun, Francesca Nicora, Valentina Pellizzer, and Inger Skjelsbaer. 2008. “Conflict Society and the Transformation of the Bosnia-Herzegovina Question.” Case Study ReportWorking Paper No. 5. SHUR: Human Rights in Conflicts: The Role of Civil Society. International Peace Research Institute, Oslo (PRIO); University of Marburg’s Center of Conflict Studies; and Lunaria, Italy. Marshall, Monty G., and Ted Robert Gurr. 2003. Peace and Conflict 2003: A Global Survey of Armed Conflicts, Self-Determination Movements, and Democracy. Center for International Development and Conflict Management. College Park: University of Maryland. Marshall, Monty G., and Keith Jaggers. 2000. “Polity IV Project: Political Regime Characteristics and Transitions, 1800–1999.” Unpublished manuscript, University of Maryland, Center for International Development and Conflict Management. ———. 2011. “Polity IV Project.” www.systemicpeace.org/polity/polity4.htm. Mason, T. David. 2004. Caught in the Crossfire: Revolution, Repression, and the Rational Peasant. Lanham, MD: Rowman and Littlefield. Mason, T. David, Mehmet Gurses, Patrick T. Brandt, and Jason Michael Quinn. 2001. “When Civil Wars Recur: Conditions for Durable Peace After Civil Wars.” International Studies Perspectives 12, 2: 171–189. Mattes, Michaela, and Burcu Savun. 2009. “Fostering Peace After Civil War: Commitment Problems and Agreement Design.” International Studies Quarterly 53, 3: 737–759. McCarthy, John D., and Mayer N. Zald. 1987. “Resource Mobilization and Social Movements: A Partial Theory.” Social Movements in an Organizational Society 82, 6: 15–42. McCoy, Alfred. 2009. Policing America’s Empire: The United States, the Philippines and the Rise of the Surveillance State. Madison: University of Wisconsin– Madison Press. McCrudden, Christopher, and Brendan O’Leary. 2013. Courts and Consociations: Human Rights Versus Power-Sharing. Oxford: Oxford University Press.

Bibliography

237

McCulloch, Alison. 2014. Power-Sharing and Political Stability in Deeply Divided Societies. Abingdon: Routledge. McEvoy, Joanne. 2008. The Politics of Northern Ireland. Edinburgh: Edinburgh University Press. McGarry, John. 2001. “Northern Ireland, Civic Nationalism, and the Good Friday Agreement.” In Northern Ireland and the Divided World: Post-Agreement Northern Ireland in Comparative Perspective. John McGarry, ed. Oxford: Oxford University Press. ———. 2017. “Conclusion: What Explains the Performance of Power-Sharing Settlements?” In Power-Sharing: Empirical and Normative Challenges. Allison McCulloch and John McGarry, eds. London: Routledge. McGarry, John, and Brendan O’Leary. 1995. Explaining Northern Ireland: Broken Images. London: Wiley-Blackwell. ———, eds. 2004. The Northern Ireland Conflict: Consociational Engagements. Oxford: Oxford University Press. ———, eds. 2007. “Iraq’s Constitution of 2005: Liberal Consociation as Political Prescription.” International Journal of International Law 5, 4: 670–698. McLauchlin, Theodore, and Wendy Pearlman. 2012. “Out-Group Conflict, In-Group Unity? Exploring the Effect of Repression on Intramovement Cooperation.” Journal of Conflict Resolution 56, 1: 41–66. Mehler, Andreas. 2009. “Peace and Power Sharing in Africa: A Not So Obvious Relationship.” African Affairs 108, 432: 453–473. ———. 2011. “Rebels and Parties: The Impact of Armed Insurgency on Representation in the Central African Republic.” Journal of Modern African Studies 49, 1: 115–139. ———. 2016. “Adapted Instead of Imported: Peacebuilding by Power-Sharing.” In Peacebuilding in Crisis. Rethinking Paradigms and Practices of Transnational Cooperation. Tobial Debiel, Thomas Held, and Ulrich Schneckener, eds. Abingdon: Routledge. Melrose, Joseph H. 2009. “The Sierra Leone Peace Process.” In Human Rights and Conflict Resolution in Context: Colombia, Sierra Leone, and Northern Ireland. Eileen F. Babbitt and Ellen L. Lutz, eds. Syracuse: Syracuse University Press. Mitchell, Claire. 2006. Religion, Identity and Politics in Northern Ireland. Aldershot: Ashgate. Mitchell, Neil J., and James M. McCormick. 1988. “Economic and Political Explanations of Human Rights Violations.” World Politics 40, 4: 476–498. Mitchell, Paul, Geoffrey Evans, and Brendan O’Leary. 2009. “Extremist Outbidding in Ethnic Party Systems Is Not Inevitable: Tribune Parties in Northern Ireland.” Political Studies 57, 2: 397–421. Mozaffar, Shaheen. 2010. “Electoral Rules and Post–Civil War Conflict Management: The Limitations of Institutional Design.” In Strengthening Peace in Post–Civil War States: Transforming Spoilers into Stakeholders. Matthew Hoddie and Caroline A. Hartzell, eds. Chicago: University of Chicago Press. Mukherjee, Bumba. 2006a. “Does Third-Party Enforcement or Domestic Institutions Promote Enduring Peace After Civil Wars? Policy Lessons from an Empirical Test.” Foreign Policy Analysis 2, 4: 405–430. ———. 2006b. “Why Political Power-Sharing Agreements Lead to Enduring Peaceful Resolution of Some Civil Wars, but Not Others?” International Studies Quarterly 50, 2: 479–504. Murdie, Amanda, and David R. Davis. 2010. “Problematic Potential: The Human Rights Consequences of Peacekeeping Interventions in Civil Wars.” Human Rights Quarterly 32, 1: 49–72.

238

Bibliography

Nagel, Joane. 2003. Race, Ethnicity and Sexuality: Intimate Intersections, Forbidden Frontiers. Oxford: Oxford University Press. Nagle, John, and Mary-Alice C. Clancy. 2010. Shared Society or Benign Apartheid? Understanding Peace-Building in Divided Societies. New York: Palgrave Macmillan. Nettelfield, Lara. 2010. Courting Democracy in Bosnia and Herzegovina. Cambridge: Cambridge University Press. Neumann, Hannah, and Joel Gwyn Winckler. 2013. “When Critique Is Framed as Resistance: How the International Intervention in Liberia Fails to Integrate Alternative Concepts and Constructive Criticism.” International Peacekeeping 20, 5: 618–635. Nilsson, Desirée. 2008. “Partial Peace: Rebel Groups Inside and Outside of Civil War Settlements.” Journal of Peace Research 45, 4: 479–495. Nilsson, Desirée, and Mimmi Söderberg Kovacs. 2011. “Revisiting an Elusive Concept: A Review of the Debate on Spoilers in Peace Processes.” International Studies Review 13, 4: 606–626. Nindorera, Willy. 2012. “The CNDD-FDD in Burundi: The Path from Armed to Political Struggle.” Berghof Transitions Series 10. Berlin: Berghof Foundation. Nordlinger, Eric A. 1972. Conflict Regulation in Divided Societies. Cambridge: Center for International Affairs, Harvard University. NOW. 2013. “Lebanon Grand Mufti Moves Against Civil Marriage.” January 29. https://now.mmedia.me/lb/en/lebanonnews/lebanon_grand_mufti_moves_against _civil_marriage. Nyanjom, Othieno. 2011. Devolution in Kenya’s New Constitution. Nairobi: Society for International Development. O’Leary, Brendan. 2005. “Debating Consociational Politics: Normative and Explanatory Arguments.” In From Power-Sharing to Democracy: Post-Conflict Institutions in Ethnically Divided Societies. Sid Noel, ed. Montreal: McGillQueen’s Press. Olsen, Tricia D., Leigh A. Payne, and Andrew G. Reiter. 2010. Transitional Justice in Balance: Comparing Processes, Weighing Efficacy. Washington, DC: United States Institute of Peace. Olson, Mancur. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation and Social Rigidities. New Haven: Yale University Press. OPAPP (Office of the Presidential Assistant on the Peace Process). 2007. Update on the MNLF Integration Program (as of September 26, 2006). Annual Report. Pasig, Metro Manila: OPAPP. Ottmann, Martin, and Johannes Vüllers. 2015. “The Power-Sharing Event Dataset (PSED): A New Dataset on the Promises and Practices of Power-Sharing in PostConflict Countries.” Conflict Management and Peace Science 32, 3: 326–350. Paffenholz, Thania, and Kristina Lundqvist. 2003. Community-Based Bottom-Up Peacebuilding: The Development of the Life and Peace Institute’s Approach to Peacebuilding and Lessons Learned from the Somalia Experience (1990– 2000). Uppsala: Life and Peace Institute. Panebianco, Angelo. 1988. Political Parties: Organization and Power. Cambridge: Cambridge University Press. Paris, Roland. 2004. At War’s End: Building Peace After Civil Conflicts. New York: Cambridge University Press. Pearlman, Wendy. 2008–2009. “Spoiling Inside and Out: Internal Political Contestation and the Middle East Peace Process.” International Security 33, 3: 79–109. ———. 2011. Violence, Nonviolence, and the Palestinian National Movement. Cambridge: Cambridge University Press.

Bibliography

239

Peksen, Dursun. 2011. “Foreign Military Intervention and Women’s Rights.” Journal of Peace Research 48, 4: 455–468. Persson, Mariam. 2012. “Demobilized or Remobilized? Lingering Rebel Structures in Post-War Liberia.” In African Conflicts and Informal Power: Big Men and Networks. Mats Utas, ed. London: Zed Books. Philippine Army. 2011. Military Leadership in Challenging Times. Fort Bonifacio, Metro Manila: Philippine Army. Philippine Congress, Senate Finance Committee. 1996. Discussion on the Proposed 1997 Department of National Defense Budget. 10th Cong., 2nd sess., September 20. Poe, Steven C., C. Neal Tate, and Linda Camp Keith. 1999. “Repression of the Human Right to Personal Integrity Revisited: A Global Cross-National Study Covering the Years, 1976–1993.” International Studies Quarterly 43, 2: 291–313. Prunier, Gérard. 2011. Africa’s World War: Congo, the Rwandan Genocide and the Making of a Catastrophe. Oxford: Oxford University Press. Raddatz, Rosalind. 2013. “Tempering Great Expectations: Peacebuilding and Transitional Justice in Liberia.” In Transitional Justice and Peacebuilding on the Ground: Victims and Excombatants. Chandra Lekha Sriram, Jemima GarcíaGodos, Johanna Herman, and Olga Martin-Ortega, eds. London: Routledge. Rasul, Amina. 2005. Broken Peace? Assessing the 1996 GRP-MNLF Final Peace Agreement. Manila: Philippine Council for Islam and Democracy. Reilly, Benjamin. 2013. “Political Parties and Post-Conflict Peacebuilding.” Civil Wars 15, S1: 88–104. Republic of Uganda, the Judiciary. 2014. “International Crimes Division.” www.judicature.go.ug/data/smenu/18/International_Crimes_Division.html. Reyntjens, Filip. 2009. The Great African War: Congo and Regional Geopolitics, 1996–2006. Cambridge: Cambridge University Press. Richards, David L., Ronald D. Gelleny, and David H. Sacko. 2001. “Money with a Mean Streak? Foreign Economic Penetration and Government Respect for Human Rights in Developing Countries.” International Studies Quarterly 45, 2: 219–239. Richmond, Oliver P., and Audra Mitchell. 2011. “Introduction—Towards a PostLiberal Peace: Exploring Hybridity via Everyday Forms of Resistance, Agency and Autonomy.” In Hybrid Forms of Peace: From Everyday Agency to PostLiberalism. Oliver P. Richmond and Audra Mitchell, eds. Basingstoke: Palgrave Macmillan. Rodil, B. Rudy. 2000. Kalinaw Mindanao: The Story of GRP-MNLF Peace Process 1975–1996. Davao City: Alternate Forum for Research in Mindanao. Roeder, Philip G. 2005. “Power Dividing as an Alternative to Ethnic Power Sharing.” In Sustainable Peace: Power and Democracy After Civil Wars. Philip Roeder and Donald Rothchild, eds. Ithaca: Cornell University Press. Roeder, Philip, and Donald Rothchild, eds. 2005. Sustainable Peace: Power and Democracy After Civil Wars. Ithaca: Cornell University Press. Rothchild, Donald. 2005. “Reassuring Weaker Parties After Civil Wars: The Benefits and Costs of Executive Power-Sharing Systems in Africa.” Ethnopolitics 4: 247–267. Rothchild, Donald, and Philip G. Roeder. 2005. “Power Sharing as an Impediment to Peace and Democracy.” In Sustainable Peace: Power and Democracy After Civil Wars. Philip Roeder and Donald Rothchild, eds. Ithaca: Cornell University Press. Russell, Delilah Ruth. 2013. “Civil Military Operations (CMO) in the Philippines: Examining Battle Space Management in the Past and the Present.” Doctoral dissertation, Graduate School of Asia-Pacific Studies, Waseda University.

240

Bibliography

Rustad, Siri Aas, and Helga Malmin Binningsbø. 2012. “A Price Worth Fighting For? Natural Resources and Conflict Occurrence.” Journal of Peace Research 49, 4: 531–546. Sagar, D. J. 2009. Political Parties of the World, 7th ed. London: John Harper Publishing. Salamey, Imad. 2014. The Government and Politics of Lebanon. London: Routledge. Salamey, Imad, and Rhys Payne. 2008. “Parliamentary Consociationalism in Lebanon: Equal Citizenry vs. Quotated Confessionalism.” Journal of Legislative Studies 14, 4: 451–473. Salloukh, Bassel. 2006. “The Limits of Electoral Engineering in Divided Societies: Elections in Postwar Lebanon.” Canadian Journal of Political Science 39, 3: 635–655. Samuels, Kirsti. 2006. “Post-Conflict Peace Building and Constitution-Making.” Chicago Journal of International Law 6, 2: 663–682. Santiago-Oreta, Jennifer, and Kathline Ann Tolosa. 2012. The AFP’s Bayanihan Plan: Assessing the Readiness of the Military Institution and the Commitment of Security Stakeholders to the AFP’s Internal Peace and Security Plan (IPSP). Working Group on Security Sector Reform, Department of Political Science, Ateneo de Manila University, Quezon City. Santos, Soliman, Jr. 2009. “MNLF Integration into the AFP and PNP: Successful Co-Optation or Failed Transformation? (Case Study).” In Primed and Purposeful: Armed Groups and Human Security Efforts in the Philippines. Geneva: Small Arms Survey; Manila: South-South Network on Non-State Armed Group Engagement. Sawyer, Amos. 2005. Beyond Plunder: Toward Democratic Governance in Liberia. Boulder: Lynne Rienner. Schlesinger, Joseph. 1985. “The New American Political Party.” American Political Science Review 79: 1152–1169. Schneckener, Ulrich. 2002. “Making Power-Sharing Work: Lessons from Successes and Failures in Ethnic Conflict Regulation.” Journal of Peace Research 39, 2: 203–228. Sen, Amartya. 1999. Development as Freedom. New York: Oxford University Press. Senate of Burundi. 2010. Le Sénat et La Politique de Décentralisation Au Burundi. Bujumbura. Shapiro, Ian. 1996. Democracy’s Place. Ithaca: Cornell University Press. Sidel, John. 2000. Capital, Coercion and Crime: Bossism in the Philippines. Stanford: Stanford University Press. Sikkink, Kathryn. 2011. The Justice Cascade: How Human Rights Prosecutions Are Changing World Politics. New York: Norton. Simmel, Georg. 1955. Conflict and the Web of Group-Affiliations. New York: Free Press. Simmons, Beth A. 2009. Mobilizing for Human Rights: International Law in Domestic Politics. Cambridge: Cambridge University Press. Simons, Claudia, Franzisca Zanker, Andreas Mehler, and Denis M. Tull. 2013. “Power-Sharing in Africa’s War Zones: How Important Is the Local Level?” Journal of Modern African Studies 51, 4: 681–706. Simonsen, Sven Gunnar. 2007. “Building ‘National’ Armies—Building Nations? Determinants of Success for Post-Intervention Integration Efforts.” Armed Forces and Society 33, 4: 571–590. Singer, J. David, Stuart Bremer, and John Stuckey. 1972. “Capability Distribution, Uncertainty, and Major Power War, 1820–1965.” In Peace, War, and Numbers. Bruce Russett, ed. Beverly Hills: Sage.

Bibliography

241

Smith, Benjamin. 2005. “Life of the Party: The Origins of Regime Breakdown and Persistence Under Single-Party Rule.” World Politics 57, 3: 421–451. Smits, Jeroen. 2010. “Ethnic Intermarriage and Social Cohesion: What Can We Learn from Yugoslavia?” Social Indicators Research 96, 3: 417–432. Snyder, Jack L., and Robert Jervis. 1999. “Civil War and the Security Dilemma.” In Civil Wars, Insecurity and Intervention. Barbara F. Walter and Jack L. Snyder, eds. New York: Columbia University Press Söderberg-Kovacs, Mimmi. 2007. From Rebellion to Politics. Stockholm: Uppsala University. Söderberg-Kovacs, Mimmi, and Sophia Hatz. 2016. “Rebel-to-Party Transformations in Civil War Peace Processes, 1975–2011.” Democratization 23, 6: 990–1008. Special Court for Sierra Leone. 2002. Statute of the Special Court for Sierra Leone. www.sc-sl.org/LinkClick.aspx?fileticket=uClnd1MJeEw%3d&tabid=176. Sriram, Chandra Lekha. 2004. Confronting Past Human Rights Violations: Justice vs. Peace in Times of Transition. London: Routledge. ———. 2005. “Wrong-Sizing Transitional Justice? The Hybrid Tribunal in Sierra Leone.” Fordham International Law Journal 29: 472–506. ———. 2008. Peace as Governance: Power-Sharing Armed Groups and Contemporary Peace Negotiations. Basingstoke: Palgrave Macmillan. ———. 2013. “Spoilers of Justice.” Nordic Journal of Human Rights 2: 248–262. ———. 2014. “Field Interviews in Nairobi.” (Currently not for attribution.) Sriram, Chandra Lekha, and Stephen Brown. 2012. “Kenya in the Shadow of the ICC: Complementarity, Gravity, and Impact.” International Criminal Law Review 12, 2: 219–244. Sriram, Chandra Lekha, Jemima García-Godos, Johanna Herman, and Olga MartinOrtega, eds. 2013. Transitional Justice and Peacebuilding on the Ground: Victims and Excombatants. London: Routledge. Sriram, Chandra Lekha, Olga Martin-Ortega, and Johanna Herman, eds. 2011. Peacebuilding and Rule of Law in Africa: Just Peace? London: Routledge. Sriram, Chandra Lekha, and Marie-Joelle Zahar. 2010. “The Perils of PowerSharing: Africa and Beyond.” Africa Spectrum 44, 3: 11–39. Staniland, Paul. 2012. “Between a Rock and a Hard Place: Insurgent Fratricide, Ethnic Defection, and the Rise of Pro-State Paramilitaries.” Journal of Conflict Resolution 56, 1: 16–40. Stearns, Jason. 2008. “Laurent Nkunda and the National Congress for the Defence of the People (CNDP).” In L’Afrique des Grands Lacs: Annuaire 2007–2008. Stefaan Marysse, Filip Reyntjens and Stef Vandeginste, eds. Paris: L’Harmattan. Stedman, Stephen John. 1997. “Spoiler Problems in Peace Processes.” International Security 22: 5–53. ———. 2002. “Introduction.” In Ending Civil Wars: The Implementation of Peace Agreements. Stephen John Stedman, Donald Rothchild, and Elizabeth M. Cousens, eds. Boulder: Lynne Rienner. Stedman, Stephen John, Donald Rothchild, and Elizabeth M. Cousens, eds. 2002. Ending Civil Wars: The Implementation of Peace Agreements. Boulder: Lynne Rienner. Stein, Arthur A. 1976. “Conflict and Cohesion: A Review of the Literature.” Journal of Conflict Resolution 20, 1: 143–172. Stewart, Frances. 2002. “Horizontal Inequalities: A Neglected Dimension of Development.” Queen Elizabeth House Working Paper, No. 81. Oxford: Oxford University Press. Straus, Scott, and Lars Waldorf, eds. 2011. Remaking Rwanda: State Building and Human Rights After Mass Violence. Madison: University of Wisconsin Press.

242

Bibliography

Subotic, Jelena. 2009. Hijacked Justice: Dealing with the Past in the Balkans. Ithaca: Cornell University Press. Sundberg, Ralph. 2009. “Revisiting One-Sided Violence: A Global and Regional Analysis.” Department of Peace and Conflict Research, Uppsala University. Taylor, Rupert. 2009. “The Injustice of a Consociational Solution to the Northern Ireland Problem.” In Consociational Theory: McGarry and O’Leary and the Northern Ireland Conflict. Rupert Taylor, ed. London: Routledge. Teitel, Ruti G. 2000. Transitional Justice. Cambridge: Cambridge University Press. Themnér, Anders. 2011. Violence in Post-Conflict Societies: Remarginalization, Remobilizers and Relationships. London: Routledge. Thoms, Oskar N.T., James Ron, and Roland Paris. 2008. “The Effects of Transitional Justice Mechanisms.” Ottawa, Centre for International Policy Studies Working Paper. Tierney, Stephen. 2012. Constitutional Referendums: The Theory and Practice of Republican Deliberation. Oxford: Oxford University Press. Tilly, Charles. 1978. From Mobilization to Revolution. Reading, MA: Addison-Wesley. Titeca, Kristof. 2011. “Access to Resources and Predictability in Armed Rebellion: The FAPC’s Short-Lived ‘Monaco’ in Eastern Congo.” Africa Spectrum 46, 2: 43–70. Tonge, Jonathan. 2008. “From Conflict to Communal Politics: The Politics of Peace.” In Northern Ireland After the Troubles: A Society in Transition. Colin Coulter and Michael Murray, eds. Manchester: Manchester University Press. ———. 2014. Comparative Peace Processes. Cambridge: Polity Press. Tull, Denis M. 2004. Congo Facing a Third War? Possible Repercussions of the Gatumba Massacre. Berlin: Stiftung Wissenschaft und Politik. Tull, Denis M., and Andreas Mehler. 2005. “The Hidden Costs of Power-Sharing: Reproducing Insurgent Violence in Africa.” African Affairs 104, 416: 375–398. United Nations. 2010. “Guidance Note of the Secretary-General: United Nations Approach to Transitional Justice.” www.unrol.org/files/TJ_Guidance_Note _March_2010FINAL.pdf. University of Antwerp. 2014. “African Peace Agreement Database.” www.uantwerpen .be/en/rg/law-and-development/research-topics/power-sharing-and-human -rights/african-peace-agreement-database/. Uppsala Conflict Data Program. 2013. “UCDP Conflict Encyclopedia.” Uppsala: Uppsala University. Utas, Mats, ed. 2012. African Conflicts and Informal Power: Big Men and Networks. Uppsala: Nordic Africa Institute; London: Zed Books. Vandeginste, Stef, and Chandra Lekha Sriram. 2011. “Power Sharing and Transitional Justice: A Clash of Paradigms.” Global Governance 17: 489–505. Varshney, Ashutosh. 2002. Ethnic Conflict and Civic Life. New Haven: Yale University Press. Walter, Barbara. 1997. “The Critical Barrier to Civil War Settlement.” International Organization 51, 3: 335–364. ———. 1999. “Designing Transitions from Civil War: Demobilization, Democratization, and Commitments to Peace.” International Security 24, 1: 127–155. ———. 2002. Committing to Peace: The Successful Settlement of Civil Wars. Princeton: Princeton University Press. ———. 2004. “Does Conflict Beget Conflict? Explaining Recurring Civil War.” Journal of Peace Research 41, 3: 371–388. Weber, Max. 1922. Wirtschaft und Gesellschaft. Tübingen, Germany: JCB Mohr (Paul Siebeck).

Bibliography

243

———. 1978. Economy and Society. Berkeley: University of California Press. Weiner, Myron. 1983. “The Political Consequences of Preferential Politics: A Comparative Perspective.” Comparative Politics 16, 1: 35–52. Weinstein, Jeremy. 2007. Inside Rebellion: The Politics of Insurgent Violence. New York: Cambridge University Press. Wennmann, Achim. 2011. The Political Economy of Peacemaking. London: Routledge. Werner, Suzanne. 1999. “The Precarious Nature of Peace: Resolving the Issues, Enforcing the Settlement, and Renegotiating the Terms.” American Journal of Political Science 43, 2: 912–934. Whitaker, Robin. 2010. “Debating Rights in the New Northern Ireland.” Irish Political Studies 25, 1: 23–45. Wiebelhaus-Brahm, Eric. 2010. Truth Commissions and Transitional Societies: The Impact on Human Rights and Democracy. London: Routledge. Wimmer, Andreas, Lars-Erik Cederman, and Brian Min. 2009. “Ethnic Politics and Armed Conflict: A Configurational Analysis of a New Global Dataset.” American Sociological Review 74, 2: 316–337. Wolff, Stefan. 2009. “Complex Power-Sharing and the Centrality of Territorial SelfGovernance in Contemporary Conflict Settlements.” Ethnopolitics 8, 1: 27–45. ———. 2011. “Post-Conflict State Building: The Debate on Institutional Choice.” Third World Quarterly 32, 10: 1777–1802. Wood, Elizabeth. 2003. Insurgent Collective Action and Civil War in El Salvador. Cambridge: Cambridge University Press. World Bank. 2011. Navigating the Storm, Delivering the Promise with a Special Focus on Kenya’s Momentous Devolution. Washington, DC: World Bank. ———. 2012. World Development Indicators. Washington, DC: World Bank. Yegar, Moshe. 2002. Between Integration and Secession: The Muslim Communities of the Southern Philippines, Southern Thailand and Western Burma/Myanmar. Lanham, MD: Lexington Books. Zanker, Franzisca. 2014. “‘Monrovia Is Not Liberia’: A Rocky Path Towards Decentralisation.” Conflict Trends 3: 10–16. Zanker, Franzisca, Claudia Simons, and Andreas Mehler. 2015. “Power, Peace, and Space in Africa: Revisiting Territorial Power Sharing.” African Affairs 114, 454: 72–91.

The Contributors

Kathleen Gallagher Cunningham is associate professor in the Department of Government and Politics at the University of Maryland and is also affiliated with the Center for International Development and Conflict Management. Rosalie Arcala Hall is professor and University of the Philippines scientist I at the University of the Philippines Visayas, Miagao, Iloilo. Caroline A. Hartzell is professor of political science at Gettysburg College and editor of the journal Conflict Management and Peace Science. Bernadette C. Hayes is professor of sociology and director of the Institute for Conflict, Transition, and Peace Research (ICTPR) in the School of Social Sciences at the University of Aberdeen. Matthew Hoddie is professor of political science at Towson University. John Ishiyama is University Distinguished Research Professor of Political Science at the University of North Texas and former editor in chief of the American Political Science Review. Andreas Mehler is director of the Arnold Bergstraesser Institute and professor of political science at Freiburg University and coeditor of the Africa Yearbook. John Nagle is reader in sociology at the University of Aberdeen.

245

246

The Contributors

Martin Ottmann is lecturer and Birmingham Fellow in International Security at the University of Birmingham. Claudia Simons is senior program officer and deputy head of the Africa Division at the Heinrich-Boell-Foundation and freelances as a consultant in the field of foreign and security policies. The late Chandra Lekha Sriram was professor of international law and international relations and director of the Centre on Human Rights in Conflict at the University of East London. Denis M. Tull is currently researcher at the Institute for Strategic Research (IRSEM) in Paris, on leave from the German Institute for International and Security Affairs (SWP). Johannes Vüllers is senior research fellow in the Department of Politics and Public Administration at the University of Konstanz. Franzisca Zanker is research fellow at the Arnold Bergstraesser Institute at the University of Freiburg, where she heads a research team on “Patterns of (Forced) Migration.”

Index

Abu Sayyaf Group, 98–99, 101, 105– 106, 107n4 Accountability, 192–193, 200–209 Administrative capital, 50–51 Africa. See specific countries Akhavan, Payam, 194–195 Angola, 126–127 Annan, Kofi, 206–207 Ante-agreement, 94–98 Aquino, Corazon, 96, 98–99 Armed Forces of the Philippines. See Philippines Armed organizations: analysis of, 60–64, 61tab–62tab, 63fig; control variables for, 54–55; data for, 55–56, 57tab– 59tab, 59–60, 66n4; diplomacy for, 48–49; party splits for, 52–54; politics for, 47, 51–52, 64–65, 65fig, 66n3; theory for, 49–50 Asia, 183–184 Authority: analysis of, 2, 9–10; democracy and, 119; descriptive statistics for, 22tab; diplomacy and, 1; for elites, 13, 132–133; government as, 10–12, 103–104; for groups, 116; ICC, 72; institutions and, 2, 224; in integration, 99; law and, 5–6; military as, 3, 10, 12–13, 27; peace and, 10, 13–14; politics of, 5–9, 15n7, 127– 128, 136; psychology of, 15n8; research on, 5; security and, 15n6; socioeconomics and, 141; for stakeholders, 12, 226; theory of, 125– 126, 136–137

Autocorrelation, 167n8 Autonomy: for groups, 15n9, 86n1; human rights and, 151–152; for institutions, 91–92; peace and, 98– 103, 121; sovereignty and, 72; of territory, 15n11

Balkans, 194–196, 198–199, 211n2 Bamba, Jean-Pierre, 201–202 al-Bashir, Omar, 208–209 Bemba, Jean-Pierre, 75–77, 204–205 Blé Goudé, Charles, 201 Bosnia, 153–154 Bozize, Francois, 201–202 Burundi, 3, 12, 84, 85tab; economics of, 135–138, 137fig; local politics in, 81– 83, 86n12; rebels in, 80; research on, 69–70

Cambodia, 211n2 Carido, Narcisse, 202 Case studies: analysis of, 142–145; on conflict, 118; economics in, 146n5; methodology for, 215–216; on military, 119–120, 120tab; for negotiations, 89–90; on peace, 12, 105–106; on rebels, 111, 146n11; on security, 166–167; on social movements, 137–142; theory and, 149–150; for transitional justice, 197– 198 CAT. See Convention Against Torture Catholicism. See Christians Cayton, Alfredo, 107n5

247

248

Index

Censorship, 155–156 Central African Republic, 201–202 Centralization: decentralization, 15n11, 68, 79–80, 82–83; of government, 95; politics of, 114; theory of, 55; transformation and, 129 Chad, 45n1 Christians, 100, 107n3, 173, 176, 185– 188. See also Religion Cingranelli-Richards (CIRI) Human Rights Database, 152, 157, 160–161 Civilians: consociationalism for, 172– 175, 174tab; elections for, 14n2; ethnoreligious divisions for, 175–178, 177tab–178tab; government and, 149–150, 161–167, 162tab, 164tab; grassroots security for, 97–98, 101– 102, 107n3; identity for, 5, 69, 175; marriage for, 178–184, 179fig, 180tab, 181fig, 183tab; military for, 96–97; in politics, 224; postconflict relations for, 49; psychology of, 14n1, 15n10, 169–172, 184–188, 190n15; religion and, 12; settlements for, 4 Cleavages, 51–52, 176, 223–224 CNDD-FDD. See Conseil National Pour la Défense de la Démocratie-Forces pour la Défense de la Démocratie Cohesion, 115, 118–119; analysis of, 122–123, 122tab; territory and, 120– 122, 122tab; transformation and, 116, 117fig Cohesiveness, 67–70, 85tab Cold War, 120, 209–210 Community: for Christians, 176; diplomacy in, 222; elites in, 77–78; ethnicity and, 106–107; global environment as, 225–226; for Muslims, 99–100, 176; solidarity as, 51; violence in, 153 Conceptual challenges, 132–134 Conflict: case studies on, 118; economics of, 23–24, 92; from elections, 73; horizontal power relations in, 130– 131, 135–136; leadership in, 112–113; military in, 22, 140–141; for Muslims, 95–96; peace and, 45n6; in politics, 25, 75–76, 111–112, 125, 223–224; psychology of, 31, 43–45, 44tab, 72; from religion, 89, 96–97, 216–217; research on, 187–188; from

resources, 127; risk in, 40–42, 41tab; separatists in, 128; from settlements, 138–139; from territory, 68; UCDP, 28–33; UN in, 42–43, 43tab; variables in, 55–56, 57tab–59tab, 59– 60; violence in, 84, 106, 193–194, 198–199, 209–211, 210tab. See also Postconflict relations Conflict resolution. See Transitional justice Congo. See Democratic Republic of Congo Conseil National Pour la Défense de la Démocratie-Forces pour la Défense de la Démocratie (CNDD-FDD), 81– 84, 85tab Consociationalism, 169, 171–175, 174tab, 184–188 Constitutions, 8, 72–73, 189n6 Context, in policy, 222–223 Control variables, 30–33, 118–119, 158– 160 Convention Against Torture (CAT), 160, 167n9 Corporate consociationalism. See Consociationalism Côte d’Ivoire, 201 Culture: of Asia, 183–184; economics and, 118; ethnicity and, 97; of Muslims, 89–90, 100; of politics, 196

Data: analysis of, 22tab, 23; for armed organizations, 55–56, 57tab–59tab, 59–60, 66n4; autocorrelation, 167n8; on democracy, 124n15, 135–136; for economics, 134–135; on federalism, 124n14; for opposition fragmentation, 115–120, 117fig, 120tab; on peace, 156–158; PSED, 22–27, 22tab, 28– 33; on religion, 177tab–178tab, 179fig, 180tab, 181fig, 183tab; on risk, 150; on self-determination, 118– 122, 120tab, 122tab, 124n11; UCDP, 28–33; on violence, 119. See also Power-Sharing Event Dataset; Variables DDR. See disarmament, demobilization, and reintegration Decentralization, 15n11, 68, 79–80, 82– 83 Demobilization, 23–24

Index Democracy, 161–165, 162tab, 164tab; authority and, 119; data on, 124n15, 135–136; in domestic environment, 159; human rights in, 167n7, 194; politics of, 2–3, 20, 191 Democratic Republic of Congo (DRC), 12, 83–84, 85tab, 135–137, 204–205; economics in, 139–140, 139fig; local politics in, 74–77; MLC, 74–77; RCD in, 74–77; research on, 69–70 Dependent variables, 28–30, 157–158, 167n5 Descriptive statistics, 22tab Design, in policy, 223–224 Dictatorships. See Philippines Diplomacy, 50; for armed organizations, 48–49; authority and, 1; in community, 222; economics of, 125– 127, 142–146, 142fig; empirical analysis of, 33–34; estimations in, 47; for groups, 144; inclusion and, 83–84; institutions in, 126–127; in Philippines, 91–92; politics and, 26– 27, 69–70, 216; psychology of, 21–22; security and, 8–9; segregation as, 178; for territory, 109–112; theory of, 20, 119–120, 120tab; violence and, 125n18 Disarmament, demobilization, and reintegration (DDR), 3, 28–30 Djotodia, Michel, 202 Domestic environment, 159 DRC. See Democratic Republic of Congo Duterte, Rodrigo, 106

Economics, 15n12, 217–218; analysis of, 135–137; of Burundi, 137–138, 137fig; in case studies, 146n5; of conflict, 23–24, 92; culture and, 118; data for, 134–135; of diplomacy, 125– 127, 142–146, 142fig; in DRC, 139–140, 139fig; in elections, 79–80; GDP, 60; in global environment, 160; government and, 12–13; of Liberia, 138–139, 138fig; politics and, 55; of postconflict relations, 127–132; PPP, 60; for rebels, 225–226; of security, 27; of Sierra Leone, 141, 141fig; socioeconomics, 136–137, 139–141; of South Africa, 142, 142fig; territory

249

and, 9; theory of, 52–54, 132–134, 151; in Uganda, 140–141, 140fig. See also Local politics Education, 131 El Salvador, 50, 54, 56 Elections: for civilians, 14n2; conflict from, 73; economics in, 79–80; ethnicity in, 83; for government, 45n3; ideology of, 51–52; institutions and, 3, 24; nationalism in, 78–80; policy for, 174; politics and, 95–96, 102–103; security and, 25–26; violence in, 71–72, 206–207 Elites: authority for, 13, 132–133; in community, 77–78; government and, 8; in groups, 132; identity for, 81; in Liberia, 83, 195; national elite pacts, 67–70; nationalism for, 11–12, 67, 70, 84; negotiations for, 70–71, 81; policy for, 134; in politics, 134–135, 169– 170, 220; psychology of, 221–222; religion for, 86n2; resources and, 144–145; in settlements, 4; violence and, 197; wealth and, 136–137. See also Local politics Empirical analysis, 45n2; of Burundi, 81–83; of diplomacy, 33–34; of DRC, 74–77; of Kenya, 70–74; of Liberia, 77–80; of regression results, 34, 35tab, 36, 37tab, 38; of robustness checks, 40–43, 41tab, 43tab; of substantive effects, 38, 39fig, 40; of territory, 115–120, 117fig, 120tab Environment, 158–160 Estimations, 47 Ethiopia, 50 Ethnicity: community and, 106–107; culture and, 97; in DRC, 76–77; in elections, 83; ethnoreligious divisions, 175–178, 177tab–178tab; identity and, 54–55, 86n7, 116, 186; in Northern Ireland, 190n17; religion and, 59, 78–79; violence and, 72–73, 76–77, 86n12 Exclusivity, 25–27 Explanatory variables, 30, 156–157

Federalism, 124n14 Ferrer, Raymundo, 100 Final Peace Agreement (FPA), 89–91 Foucault, Michel, 6–7, 15n8

250

Index

FPA. See Final Peace Agreement Fragmentation. See Opposition fragmentation Front pour la Démocratie au Burundi (FRODEBU), 81–83

Gbagbo, Laurent, 201 Gbagbo, Simone, 201 GDP. See Gross Domestic Product Ghana, 205 Global environment, 159–160, 222–223, 225–226 Government: analysis of, 43–44, 44tab; as authority, 10–12, 103–104; centralization of, 95; civilians and, 149–150, 161–167, 162tab, 164tab; data for, 28–33; decentralization of, 15n11; economics and, 12–13; elections for, 45n3; elites and, 8; empirical analysis of, 33–34, 35tab, 36, 37tab, 38, 39fig, 40–43, 41tab, 43tab; incentives for, 129; integration for, 114; in Northern Ireland, 172– 175, 174tab; in Philippines, 107n5; postconflict relations for, 21–27, 22tab; rebels and, 19–21, 48, 131– 132; research on, 156–161, 165–166; resources for, 126; security and, 150– 156; violence and, 93 Grassroots security, 97–98, 101–102, 107n3 Gross Domestic Product (GDP), 60, 159, 161–165, 162tab, 164tab Groups: Abu Sayyaf Group, 98–99, 101, 105–106, 107n4; authority for, 116; autonomy for, 15n9, 86n1; cohesion for, 119; diplomacy for, 144; elites in, 132; identity for, 128, 153–154; integration of, 123; PAGs, 90; in policy, 128–129; rebels as, 50; settlements and, 10; in society, 20–21; socioeconomics for, 139–140

Habyarimana, Juvenal, 200–201 Hairiri, Rafik, 189n11 Hazards. See Risk Helsinki peace talks, 133 Hermoso, Dickson, 108n6 History, 6, 114, 169–172 The History of Sexuality (Foucault), 15n8 Homophobia, 190n16

Horizontal cleavages, 51–52 Horizontal power relations, 130–131, 135–136, 143, 146n4 Human rights: autonomy and, 151–152; CAT, 160; CIRI, 152, 157, 160–161; in democracy, 167n7, 194; for media, 155–156; negotiations for, 195; politics of, 109, 154–155; in postconflict relations, 168n12; theory of, 77, 155 Hussin, Parouk, 103 Hutus, 76–77, 81–83, 137–138

ICC. See International Criminal Court Ideational capital, 50–51 Identity: for Christians, 173; for civilians, 5, 69, 175; for elites, 81; ethnicity and, 54–55, 86n7, 116, 186; for groups, 128, 153–154; for Muslims, 12, 91; religion and, 13–14, 189n10; research on, 7–8; in selfdetermination, 115; violence and, 48 Ideology, 51–52 Incentives, 113–114, 129 Inclusion, 25–27, 67–70, 85tab; diplomacy and, 83–84; psychology of, 86n7, 124n9; representation and, 3–4 Indonesia, 129 INGOs. See International nongovernmental organizations Institutions: authority and, 2, 224; autonomy for, 91–92; in diplomacy, 126–127; elections and, 3, 24; peace and, 2, 7; in policy, 25–26, 226n2; psychology of, 19, 23 Institutions for Sustainable Peace (ISP), 14n3 Integration: authority in, 99; for government, 114; of groups, 123; for military, 99–100, 104–105, 107n5; policy for, 93–94, 104; politics of, 101–102; religion in, 102–103; violence from, 108n6 International brokers, 67 International Criminal Court (ICC), 72, 196, 198, 200–209. See also Transitional justice International nongovernmental organizations (INGOs), 150, 160– 167, 162tab, 164tab, 195

Index Interreligious marriage. See Marriage Iraq, 106 ISIS. See Islamic State in Iraq and Syria Islam. See Muslims Islamic State in Iraq and Syria (ISIS), 106 ISP. See Institutions for Sustainable Peace

Kabbah, Ahmed Tejan, 203 Kabila, Joseph, 204–205 Karadzic, Radovan, 195 Kenya, 12, 83–84, 85tab, 206–207, 211; local politics in, 70–74; politics in, 79; research on, 69–70; violence in, 192 Kenyatta, Uhuru, 73–74, 206–207 KFR. See Kidnap-for-ransom Kibaki, Mwai, 206–207 Kidau, Nohn, 77–80 Kidnap-for-ransom (KFR), 90, 98–99, 101, 107n3 Kiplagat, Jackson, 73 Kony, Joseph, 203–204 Kosgei, Sally, 86n3

Law: authority and, 5–6; ICC, 72; martial law, 96, 106; in Philippines, 107n2; politics of, 197–198; theory of, 194 Leadership: in conflict, 112–113; in local politics, 86n2; military as, 89–91; for MNLF, 12; for Muslims, 188n4; for rebels, 51, 66n1, 199, 223–224 Lebanon, 13–14; constitutions, 189n6; history of, 169–172; marriage in, 188n3; military in, 188n5; Northern Ireland and, 175–184, 177tab–178tab, 179fig, 180tab, 181fig, 183tab, 189n7, 189n12, 217; politics in, 172– 175, 174tab; religion in, 184–188, 189n9 Legislation, 94 Liberal consociationalism. See Consociationalism Liberalism, 69 Liberia, 12, 84, 85tab, 86n9, 135–137; economics of, 138–139, 138fig; elites in, 83, 195; local politics in, 77–80; NPP in, 78–80; research on, 69–70, 205–206 Life expectancy, 32

251

Local politics: in Burundi, 81–83, 86n12; in DRC, 74–77; in Kenya, 70–74; leadership in, 86n2; in Liberia, 77– 80; theory of, 67–70, 83–84, 85tab

Macedonian Ushitra Çlirimtare e Kosovës (UCK), 29 Machiavelli, Niccolò, 6 Mali, 207–208 Marcos, Ferdinand, 89, 95–97 Marriage: analysis of, 190n18; in Lebanon, 188n3; politics of, 186; religion and, 170–171, 178–184, 179fig, 180tab, 181fig, 183tab; research on, 189n14 Martial law, 96, 106 Mbugua, Kinuthia, 73–74 Media, 155–156 Methodology, 118–120, 120tab, 198– 200, 215–216 MILF. See Moro Islamic Liberation Front Military, 15n12, 217–218; as authority, 3, 10, 12–13, 27; case studies on, 119– 120, 120tab; for civilians, 96–97; in conflict, 22, 140–141; demobilization of, 23–24; integration for, 99–100, 104–105, 107n5; as leadership, 89– 91; in Lebanon, 188n5; in Philippines, 92–94, 98–103; policy for, 38, 39fig, 40, 103–104; politics and, 8–9, 14, 24–25, 110, 144, 191– 192; rebels and, 9; theory of, 52–54, 151; violence and, 167n9 Milosevic, Slobodan, 195 Misuari, Nur, 93, 96, 102 Mladic, Ratko, 195 MLC. See Mouvement de Libératíon du Congo MNLF. See Moro National Liberation Front Moro Islamic Liberation Front (MILF), 90, 95–96, 98–100, 103, 105–106 Moro National Liberation Front (MNLF), 12, 89–91 Mouvement de Libératíon du Congo (MLC), 74–77 Mozambique, 50 Mudavadi, Musalia, 86n3 Mulbagh, George, 78 Museveni, Yoweri, 204

252

Index

Muslims, 173; Christians and, 107n3, 185–188; community for, 99–100, 176; conflict for, 95–96; culture of, 89–90, 100; identity for, 12, 91; leadership for, 188n4; MILF, 90, 95– 96, 98–100, 103, 105–106; negotiations for, 92–93, 107n1; in politics, 104; as rebels, 207–208. See also Religion

National elite pacts, 67–70 National Patriotic Party (NPP), 78–80 Nationalism: in elections, 78–80; for elites, 11–12, 67, 70, 84; horizontal power relations and, 143; selfdetermination and, 122–123 Negotiations: case studies for, 89–90; for elites, 70–71, 81; for human rights, 195; for Muslims, 92–93, 107n1; for peace, 73, 109–110; in politics, 83– 84, 105; with rebels, 205; resources in, 130, 208; for settlements, 56 Nepal, 3, 56 Nicaragua, 50 Nigeria, 205 Nkunda, Laurent, 75–77 Norman, Sam Hinga, 203 Northern Ireland, 13–14, 154, 188n1; analysis of, 188n2; ethnicity in, 190n17; government in, 172–175, 174tab; history of, 169–172; Lebanon and, 175–184, 177tab–178tab, 179fig, 180tab, 181fig, 183tab, 189n7, 189n12, 217; politics in, 184–188; religion in, 172–173, 189n8 NPP. See National Patriotic Party Ntaganda, Bosco, 204–205

Odinga, Raila, 70, 73, 206–207 One-sided violence, 119, 124n12 Opposition fragmentation, 109–115; data for, 115–120, 117fig, 120tab; policy for, 120–122, 122tab; territory and, 112–115, 122–123 Orengo, James, 86n3 Organizations. See Armed organizations Ouattara, Alassane, 201

PAGs. See Private armed groups Palestine, 110–111 Parallel consent, 175

Party splits, 52–54, 56, 62–65, 63fig, 65fig Patasse, Ange-Felix, 201–202 Peace, 188n1; ante-agreement, 94–98; authority and, 10, 13–14; autonomy and, 98–103, 121; case studies on, 12, 105–106; conflict and, 45n6; data on, 156–158; FPA, 89–91; Helsinki peace talks, 133; for INGOs, 195; institutions and, 2, 7; legislation for, 94; negotiations for, 73, 109–110; in Philippines, 89–91, 103–107; policy for, 11, 126–127, 158, 192; politics of, 9, 93; for rebels, 55–56; resources and, 133; security and, 218; settlements for, 50, 81–82, 97, 117– 118, 140, 141–142, 221, 226n1; theory for, 21, 48; transitional justice for, 191–192; UN for, 31, 155. See also Institutions for Sustainable Peace Philippine National Police, 94–98, 101 Philippines, 216–217; ante-agreement in, 94–98; diplomacy in, 91–92; government in, 107n5; horizontal power relations in, 130–131; law in, 107n2; military in, 92–94, 98–103; peace in, 89–91, 103–107; politics in, 12; religion in, 107n1 Plausibility probes, 143–145, 218 Policy: context in, 222–223; design in, 223–224; for education, 131; for elections, 174; for elites, 134; groups in, 128–129; institutions in, 25–26, 226n2; for integration, 93–94, 104; for military, 38, 39fig, 40, 103–104; for opposition fragmentation, 120– 122, 122tab; for peace, 11, 126–127, 158, 192; for postconflict relations, 225–226; for security, 106–107; temporality in, 224–225; transitional justice as, 193–196; for violence, 100–101, 107n2 Politics: analysis of, 4; for armed organizations, 47, 51–52, 64–65, 65fig, 66n3; of authority, 5–9, 15n7, 127–128, 136; of censorship, 155– 156; of centralization, 114; civilians in, 224; of cohesion, 118; of Cold War, 209–210; conflict in, 25, 75–76, 111–112, 125, 223–224; culture of, 196; of DDR, 3; of democracy, 2–3,

Index 20, 191; diplomacy and, 26–27, 69– 70, 216; economics and, 55; elections and, 95–96, 102–103; elites in, 134– 135, 169–170, 220; of human rights, 109, 154–155; of integration, 101– 102; in Kenya, 79; of law, 197–198; in Lebanon, 172–175, 174tab; of marriage, 186; military and, 8–9, 14, 24–25, 110, 144, 191–192; Muslims in, 104; negotiations in, 83–84, 105; in Northern Ireland, 184–188; party splits, 52–54; of peace, 9, 93; in Philippines, 12; Political Handbook of the World (reference book), 56, 57tab–59tab; Political Parties of the World (reference book), 56, 57tab– 59tab; psychology of, 1–2, 172–173; rebels in, 11, 49–55, 66n2, 224–225; regulation of, 3; religion in, 94–95; risk in, 80; of security, 5–6; of selfdetermination, 217–218; of settlements, 6–7; theory of, 52–54, 151; transformation in, 68–69, 152; of transitional justice, 218–219. See also Local Politics Population, 159 Postconflict relations: analysis of, 14, 43–44, 44tab, 145–146, 215–225; for civilians, 49; constitutions in, 8; economics of, 127–132; exclusivity in, 25–27; for government, 21–27, 22tab; human rights in, 168n12; inclusion in, 25–27; policy for, 225– 226; psychology of, 64–65, 65fig; for rebels, 4, 7, 21–27, 22tab, 47; religion in, 54–55; research on, 2–5; risk in, 152–156, 166–167; security in, 149– 150, 156–165, 162tab, 164tab; transformation in, 49–52; violence in, 23–25. See also Conflict Power-Sharing Event Dataset (PSED), 22–27, 22tab, 28–33 PPP. See Purchasing power parity The Prince (Machiavelli), 6 Private armed groups (PAGs), 90 Property, 129 Protestants. See Christians PSED. See Power-Sharing Event Dataset Psychology: of authority, 15n8; of civilians, 14n1, 15n10, 169–172, 184–188, 190n15; of conflict, 31, 43–

253

45, 44tab, 72; of diplomacy, 21–22; of elites, 221–222; history of, 6; of inclusion, 86n7, 124n9; of institutions, 19, 23; of politics, 1–2, 172–173; of postconflict relations, 64–65, 65fig; of territory, 12; of violence, 49–50 Public attitudes. See Civilians Purchasing power parity (PPP), 60

Racism, 190n16 Ramos, Fidel, 99 Rassemblement Congolais pour la Démocratie (RCD), 74–77, 83–84, 85tab Rebels: analysis of, 43–44, 44tab; in Burundi, 80; case studies on, 111, 146n11; CNDD-FDD, 81–83; cohesiveness for, 67–70; data for, 28– 33; economics for, 225–226; empirical analysis of, 33–34, 35tab, 36, 37tab, 38, 39fig, 40–43, 41tab, 43tab; FRODEBU, 81–83; government and, 19–21, 48, 131–132; as groups, 50; Hutus as, 81–83; leadership for, 51, 66n1, 199, 223– 224; military and, 9; MLC, 74–77; Muslims as, 207–208; negotiations with, 205; peace for, 55–56; in politics, 11, 49–55, 66n2, 224–225; postconflict relations for, 4, 7, 21–27, 22tab, 47; RCD, 74–77; research on, 55–56, 57tab–59tab, 59–64, 61tab– 62tab, 63fig. See also Armed organizations Reform, 72–73 Regional power, 62–63 Regression results, 34, 35tab, 36, 37tab, 38 Regulation, 3 Religion: civilians and, 12; conflict from, 89, 96–97, 216–217; data on, 177tab– 178tab, 179fig, 180tab, 181fig, 183tab; for elites, 86n2; ethnicity and, 59, 78–79; ethnoreligious divisions, 175–178, 177tab–178tab; identity and, 13–14, 189n10; in integration, 102–103; in Lebanon, 184–188, 189n9; marriage and, 170–171, 178– 184, 179fig, 180tab, 181fig, 183tab; in Northern Ireland, 172–173, 189n8;

254

Index

in Philippines, 107n1; in politics, 94– 95; in postconflict relations, 54–55 Representation, 3–4 Republic of Congo. See Democratic Republic of Congo Research: on authority, 5; on Burundi, 69–70; on conflict, 187–188; on DRC, 69–70; on government, 156– 161, 165–166; on identity, 7–8; on international brokers, 67; ISP for, 14n3; on Kenya, 69–70; on Liberia, 69–70, 205–206; on marriage, 189n14; on postconflict relations, 2– 5; PSED for, 28–33; on rebels, 55–56, 57tab–59tab, 59–64, 61tab–62tab, 63fig; on social movements, 113; on society, 4–5; theory and, 60, 224–225; on violence, 13, 157–158; World Bank Development Indicators for, 32 Resources, 126–130, 133, 144–145, 208 Risk: analysis of, 33–34, 38, 39fig, 40, 45n5; in conflict, 40–42, 41tab; data on, 150; for party splits, 62–65, 63fig, 65fig; in politics, 80; in postconflict relations, 152–156, 166–167; Schoenfeld residual-based test, 60; utility and, 225; VIF in, 60–62, 61tab–62tab; violence and, 34, 35tab, 36, 37tab, 38 Robustness checks, 40–43, 41tab, 43tab Ruberwa, Azarias, 75–77 Ruto, William, 71–74, 86nn3–4, 206–207 Rwanda, 76–77, 86n7, 200–201

Salim, Uttuh, 103 Sankoh, Foday, 131 Schoenfeld residual-based test, 60 Security: authority and, 15n6; case studies on, 166–167; diplomacy and, 8–9; economics of, 27; elections and, 25–26; government and, 150–156; grassroots security, 97–98, 101–102, 107n3; peace and, 218; policy for, 106–107; politics of, 5–6; in postconflict relations, 149–150, 156– 165, 162tab, 164tab Segregation, 178 Self-determination: analysis of, 115–116, 117fig; data on, 118–122, 120tab, 122tab, 124n11; history of, 114; identity in, 115; nationalism and,

122–123; politics of, 217–218; theory of, 110–111 Sema, Muslimen, 103 Separatist conflict, 128 Serufuli, Eugene, 75–77 Settlements: analysis of, 5, 105–106, 112, 225; for civilians, 4; conflict from, 138–139; in El Salvador, 54; elites in, 4; groups and, 10; incentives in, 113–114; negotiations for, 56; for peace, 50, 81–82, 97, 117–118, 140, 141–142, 221, 226n1; politics of, 6–7; stalemates in, 48; temporality in, 68– 70; for territory, 11 Sierra Leone, 131, 135–137, 195, 202– 203; economics of, 141, 141fig; Violence in, 155 Sirleaf, Ellen Johnson, 205–206 Social movements, 113, 137–142 Society, 4–5, 20–21 Socioeconomics, 136–137, 139–141 Solidarity, 51 South Africa, 142, 142fig Sovereignty, 72 Stakeholders, 12, 226 Stalemates, 48 Statistics. See Data; Research Substantive effects, 38, 39fig, 40 Sudan, 192, 195, 208–209 Syria, 106, 189n11

Taylor, Charles, 77–80, 195, 205–206 Temporality, 68–70, 85tab, 124n16, 224– 225 Territory, 15n12, 217–218; analysis of, 124n16; autonomy of, 15n11; cohesion and, 120–122, 122tab; conflict from, 68; diplomacy for, 109–112; economics and, 9; empirical analysis of, 115–120, 117fig, 120tab; opposition fragmentation and, 112– 115, 122–123; psychology of, 12; settlements for, 11; theory of, 52–54, 151 Terrorism, 106, 107n4 Time since last war termination (TSLWT), 158–165, 162tab, 164tab Torture, 160, 167n9 TPIR. See Trends in physical integrity rights

Index Transformation: centralization and, 129; cohesion and, 116, 117fig; in politics, 68–69, 152; in postconflict relations, 49–52; theory for, 50–51 Transitional justice: accountability and, 192–193, 200–209; analysis of, 209– 211, 210tab; case studies for, 197–198; methodology for, 198–200; for peace, 191–192; as policy, 193– 196; politics of, 218–219 Trends in physical integrity rights (TPIR), 160–165, 162tab, 164tab TSLWT. See Time since last war termination Tutsis, 76–77, 81–83, 86n12, 137–138

UCDP. See Uppsala Conflict Data Program UCK. See Macedonian Ushitra Çlirimtare e Kosovës Uganda, 135–137, 140–141, 140fig, 195, 203–204 Uganda National Rescue Front II (UNRF II), 29, 45n1 United Nations (UN), 31, 42–43, 43tab, 155, 211n2 UNRF II. See Uganda National Rescue Front II Uppsala Conflict Data Program (UCDP), 28–33 Utility, 14, 219–222, 225

Variables: in analysis, 167n9; autocorrelation, 167n8; in conflict, 55–56, 57tab–59tab, 59–60; control, 30–33, 54–55, 118–119, 158–160;

255

dependent, 28–30, 157–158, 167n5; explanatory, 30, 156–157; in methodology, 118–120, 120tab; temporality, 124n16; TSLWT, 158– 159 Variance inflation factors (VIF), 60–62, 61tab–62tab Vertical cleavages, 51–52 VIF. See Variance inflation factors Violence: Abu Sayyaf Group, 98–99, 101, 105–106, 107n4; analysis of, 19; CAT, 160, 167n9; in community, 153; in conflict, 84, 106, 193–194, 198– 199, 209–211, 210tab; data on, 119; diplomacy and, 125n18; in elections, 71–72, 206–207; elites and, 197; ethnicity and, 72–73, 76–77, 86n12; government and, 93; identity and, 48; for INGOs, 161–165, 162tab, 164tab; from integration, 108n6; ISIS, 106; in Kenya, 192; KFR, 90, 98–99, 101, 107n3; military and, 167n9; onesided, 119, 124n12; PAGs, 90; policy for, 100–101, 107n2; in postconflict relations, 23–25; psychology of, 49– 50; research on, 13, 157–158; risk and, 34, 35tab, 36, 37tab, 38; in Sierra Leone, 155; terrorism, 106, 107n4; in theory, 44–45, 44tab

Wealth, 136–137 Weber, Max, 6, 15n7 World Bank Development Indicators, 32

Yugoslavia, 194–196, 198–199, 211n2

About the Book

There are numerous studies on the role of power-sharing agreements in the maintenance of peace in postconflict states. Less explored, however, is the impact of power sharing on the quality of the peace. Do powersharing institutions in fact transform the balance of power among actors in the aftermath of civil wars? And if so, how? As they address these issues, seeking to establish a new research agenda, the authors provide a rich new analytical approach to understanding how power sharing actually works. Caroline A. Hartzell is professor of political science at Gettysburg College. Andreas Mehler is director of the Arnold-Bergstraesser-Institut and professor of development policy and the theory of development at the University of Freiburg.

257