Historical Institutionalism, Prospect Theory and an Alternative Theory Of
Total Page:16
File Type:pdf, Size:1020Kb
HISTORICAL INSTITUTIONALISM, PROSPECT THEORY AND AN ALTERNATIVE THEORY OF COLLECTIVE VIOLENCE: THE CASES OF RWANDA AND BURUNDI by STACEY MARIE GIBSON MITCHELL (Under the Direction of Christopher S. Allen) ABSTRACT Rwanda and Burundi are both poor, densely populated, multiethnic countries that formally began the process of democratization in the early 1990s. Shortly thereafter, resistance to reform led to coups d’état and the assassination of both countries’ presidents. It is here that the parallel paths of the two similar countries diverge. In Rwanda, a total genocide of the country’s Tutsi population was perpetrated by extremist Hutu factions in 1994. In Burundi, however, genocide did not occur. The country instead collapsed into civil war. This variation in outcome is inadequately explained by the micro-, meso-, and macro-level theories frequently applied to one or both cases. This theoretical deficiency is attributed to problems of structure and agency, and an over reliance on the assumption that actors behave as rational utility maximizers. To better explicate the relationship between democratization and collective violence in Rwanda and Burundi, this dissertation utilizes an integrative approach that combines two distinct, but compatible theories: historical institutionalism and prospect theory. This dissertation argues that institutional legacies influenced the success of democratization and the degree of collective violence that occurred in both countries. Consistent with the core assumptions of both theories, the results of a most similar systems analysis demonstrate that an institutional legacy characterized by an asymmetrical distribution of power between groups induces incumbent leaders to perceive joint-rule from the domain of losses. Conversely, when a country has had greater experience with formal and informal democratic traditions and practices, the likelihood that multiparty democracy will be perceived by incumbent elites from the domain of gains is increased. The results of the content analysis performed by this study further demonstrate that actors who perceive multiparty democracy as a loss are more likely to favor risk-seeking strategies in response to political liberalization. Additionally, the analysis reveals that actors who perceive multiparty democracy as a gain are more likely to favor a risk-averse policy course. INDEX WORDS: Burundi, Collective violence, Democratization, Ethnic Conflict, Genocide, Hybrid Democracy, Institutionalism, Prospect Theory, Rwanda. HISTORICAL INSTITUTIONALISM, PROSPECT THEORY AND AN ALTERNATIVE THEORY OF COLLECTIVE VIOLENCE: THE CASES OF RWANDA AND BURUNDI by STACEY MARIE GIBSON MITCHELL B.A. Georgia State University, 1993 M.A. Georgia State University, 2000 A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY ATHENS, GEORGIA 2007 © 2007 Stacey Marie Gibson Mitchell All Rights Reserved HISTORICAL INSTITUTIONALISM, PROSPECT THEORY AND AN ALTERNATIVE THEORY OF COLLECTIVE VIOLENCE: THE CASES OF RWANDA AND BURUNDI by STACEY MARIE GIBSON MITCHELL Major Professor: Christopher S. Allen Committee: Jeffrey Berejkian Loch K. Johnson Abdulahi Osman Jaroslav Tir Electronic Version Approved: Maureen Grasso Dean of the Graduate School The University of Georgia May 2007 ACKNOWLEDGEMENTS I want to first acknowledge the scores of ethnic conflict and genocide scholars whose studies form the backbone of this dissertation. I am especially indebted to René Lemarchand, whose work on the history of conflict in Rwanda and Burundi served as the inspiration for much of the argument developed by this dissertation. I want to further acknowledge all of the members of my graduate committee to whom I owe a debt of gratitude for their assistance in the development of this dissertation. I am especially indebted to Christopher Allen whose guidance and instruction have been invaluable to me. I would also like to thank the other members of my committee, Jeffrey Berejikian, Loch Johnson, Abdulahi Osman and Jaroslav Tir for their assistance in the completion of this dissertation and the development of my academic career. I owe a special debt of gratitude to my husband. Without his love, advice, and patience this dissertation would not have been completed. iv TABLE OF CONTENTS Page ACKNOWLEDGEMENTS…………………………………………………………………………….iv LIST OF TABLES……………………………………………………………………………………...ix LIST OF FIGURES……………………………………………………………………………………xii LIST OF MAPS……………………………………………………………………………………….xiii CHAPTER 1. INTRODUCTION………………………………………………………………………….1 Summary of Existing Explanations………………………………………………….....2 The Argument……………………………………………………………………….....7 Methodology………………………………………………………………………......12 Plan of Dissertation…………………………………………………............................15 2. EXPLANATIONS OF COLLECTIVE VIOLENCE IN RWANDA AND BURUNDI………………………………………………………………………......20 Macro-level Explanations of Genocide and Collective Violence……………………………………………………………………………….21 Meso-level Explanations: Democratization and Collective Violence…………………………………………………………………………….....27 Micro-level Explanations of Collective Violence…………………………………….39 Conclusions………………………………………………………................................44 3. HISTORICAL INSTITUTIONALISM, PROSPECT THEORY AND THE EXPLANATION OF COLLECTIVE VIOLENCE…………………………...47 The Need for an Integrative Theory………………………………...………………....49 Historical Institutionalism: Core Assumptions and Problems of v Structure and Agency…………………………………………………………….......52 Prospect Theory: Core Assumptions and the Problem of Predicting Reference Points………………………………………………………………….......63 Historical Institutionalism and Prospect Theory: An Integrative Theory Of Collective Violence………………………………………………..………….......71 Conclusion…………………………………………………………............................79 4. METHODOLOGY AND RESEARCH DESIGN……………………………………......82 Variation between Cases: The Use of a Two-Case Study………………………........85 Dependent Variable: Total Genocide…………………………………………….......91 Explanatory Variables: Social Structures and Traditional Political Institutions………………………………………………………………….92 Control Variables…………………………………………………………………….97 Variation within Cases: The Use of a Within-Case Analysis……………………….106 Intervening Variable: Multiparty Democracy……………………………………….110 Conclusion…………………………………………………………………………..122 5. STRUCTURAL EXPLANATIONS OF GENOCIDE………………………………......124 Rwanda and Burundi: A Study of Two Similar Cases……………………………....126 Pluralized Society…………………………………………………………………....128 Domestic Conflict…………………………………………………………………....130 Demographic Stress…………………………………………………………….........144 Economic Health………………………………………………….............................148 Conclusion…………………………………………………………………………...154 6. HISTORICAL INSTITUTIONALISM AND COLLECTIVE VIOLENCE IN RWANDA AND BURUNDI, PART: 1: SOCIAL STRUCTURES…………………………………………………………………………..192 Cultural Differentials Index…………………………………………………….........194 vi Economic Differentials Index………………………………………………….........195 Civil Society…………………………………………………………………………196 Kinship Groups………………………………………………………………………215 Regional Identities…………………………………………………...........................221 Patron-Client Systems…………………………………………………………..........222 Informal Institutions of Conflict Resolution…………………………………............229 Conclusion…………………………………………………………………………....231 7. HISTORICAL INSTITUTIONALISM AND COLLECTIVE VIOLENCE IN RWANDA AND BURUNDI, PART 2: TRADITIONAL POLITICAL INSTITUTIONS……………………………………………………………………….....241 Political Differentials Index……………………………………………………….....243 POLITY2 Scores……………………………………………………………………..244 Political Centralization…………………………………………………………….....245 Executive Competition and Recruitment……………………………………………..248 Executive Constraints………………………………………………………………...251 Political Participation and Recruitment……………………………............................252 Political Paradigms…………………………………………………………………...259 The Process of Democratization: An Initial Assessment………………………….....262 Conclusion…………………………………………………………………………....275 8. REFERENCE POINTS, DECISION FRAMES AND COLLECTIVE VIOLENCE IN RWANDA AND BURUNDI……………………………………..........286 Rational Choice Explanations of Collective Violence in Rwanda And Burundi…………………………………………………………........................287 Content Analysis: Results and Discussion…………………………………………..288 Discussion and Qualitative Evidence of Risk Assessment………………………….300 Conclusion…………………………………………………………….......................312 vii 9. CONCLUSIONS…………………………………………………………........................314 The Influence of Social Structures and Traditional Political Institutions on Institutional Development and Change……………………………...316 Institutional Change and Actors’ Assessment of Options…………………………...321 Implications of This Study and Areas for Further Research………………………...325 REFERENCES………………………………………………………………………………..............334 APPENDIX A………………………………………………………………………………………...354 viii LIST OF TABLES Page Table 4.1: Multiparty Democracy …………………………………………………………………….111 Table 4.2: Use of Reference Point and Policy Preference…………………………………………….121 Table 4.3: Type of Reference Point Used and Policy Preference……………………………………..121 Table 5.1: Inter- and Intra-Communal Disputes, Burundi…………………………………………….157 Table 5.2: Inter- and Intra-Communal Disputes, Rwanda…………………………………………….160 Table 5.3: Intra-Communal Conflicts, Burundi……………………………………………………….170 Table 5.4: Intra-Communal Conflicts, Rwanda……………………………………………………….171 Table 5.5: Inter-Communal Conflict (COMCO index), Burundi……………………………………..171 Table 5.6: Inter-Communal Conflict (COMCO Index),