# Domestic Political Institutions, Leadership Survival and Conflicts in Dictatorships Open Access

## Koga, Jun (2013)

Permanent URL: https://etd.library.emory.edu/concern/etds/sx61dm83z?locale=en
Published

## Abstract

Under what conditions does an authoritarian leader eliminate rival elites from the regime and weaken the elites' capabilities to punish the leader through a coup? A dictator's strategy of eliminating strong rival elites in order to prevent a coup is a double-edged sword. The strategy is risky in the short term because it might prompt elites to immediately launch a coup. However, eliminating rival elites is beneficial in the long term because this diminishes the ability of elites to punish the dictator in the future. Using a formal model, I predict that a dictator is more likely to eliminate rival elites when the elites' coup-making capabilities are temporarily low. Essentially, I argue that a dictator takes advantage of the low risk of a coup in the present to address the future risk of a coup. Thus, somewhat paradoxically, my theory implies that a decrease in the probability of coup replacement will cause the dictator to reduce elites' coup-making capabilities, challenging the conventional argument that a dictator who faces a higher risk of coup is more likely to take coup-proofing efforts to reduce his coup risk. Moreover, focusing on coordination problems among individual elites and the role of coup events in updating elites' expectations about others' actions, I identify the conditions under which an elite audience's capability to punish a dictator becomes temporarily low. I maintain that coup events -- either in the form of a successful coup that puts a new dictator in power or a failed coup -- will temporarily decrease an elite audience's capability to coordinate to oust a dictator. Thus, they will provide a window of opportunity for a dictator to promote the process of consolidation of power. I test my hypotheses with an original data set on military purges for 438 dictators in 110 authoritarian countries from 1969 to 2003. Empirical results using the new data provide strong evidence for my theoretical arguments. Importantly, analyzing the process by which a dictator consolidates power at the expense of elites provides us with the root cause of why some dictators are more conflict prone or belligerent than others.

Contents

1 Introduction 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1.1 Research Question . . . . . . . . . . . . . . . . . . . . . . . . 5

1.2 The Existing Literature . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.2.1 Incentive Argument . . . . . . . . . . . . . . . . . . . . . . . . 8

1.2.2 Opportunity Argument . . . . . . . . . . . . . . . . . . . . . . 10

1.3 Theory of Consolidation of Power in Dictatorships . . . . . . . . . . . 12

1.3.1 Coordination Problem . . . . . . . . . . . . . . . . . . . . . . 13

1.4 Plan of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2 Theory 18

2.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.2 Formal Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.2.1 Comparative Statics Results . . . . . . . . . . . . . . . . . . . 25

2.3 Coordination Problem among Elites . . . . . . . . . . . . . . . . . . . 30

2.3.1 Coup Entry Argument . . . . . . . . . . . . . . . . . . . . . . 33

2.3.2 Failed Coup Argument . . . . . . . . . . . . . . . . . . . . . . 37

3 New Dataset on Military Purges 41

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.2 Coding Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.3 Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

4 Data and Model 54

4.1 Data and Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

4.1.1 Dependent Variables . . . . . . . . . . . . . . . . . . . . . . . 58

4.1.2 Independent Variables . . . . . . . . . . . . . . . . . . . . . . 60

4.1.3 Control Variables . . . . . . . . . . . . . . . . . . . . . . . . . 61

4.1.4 Empirical Model . . . . . . . . . . . . . . . . . . . . . . . . . 65

5 Empirical Results 66

5.1 Coup Replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.1.1 Coup Entry Argument (H2-1 & H2-2) . . . . . . . . . . . . . . 66

5.1.2 Failed Coup Argument (H4-1 & H4-2) . . . . . . . . . . . . . 75

5.2 Military Purge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

5.2.1 Coup Replacement Argument (H1) . . . . . . . . . . . . . . . 81

5.2.2 The Bayesian Analysis of Coup Risk . . . . . . . . . . . . . . 87

5.2.3 Coup-Entry Argument (H3-1 & H3-2) . . . . . . . . . . . . . . 104

5.2.4 Failed Coup Argument (H5-1 & H5-2) . . . . . . . . . . . . . 112

6 Conclusion 123

6.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

6.2 Implication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

6.2.1 Domestic Accountability and International Conict . . . . . . 125

6.2.2 Power-Sharing Institutions and Consolidation of Power in Autocracies . . . . . . 127

6.2.3 Temporary Shift in Relative Power and Commitment Problem 129

A Personalist Regimes for the period 1969-2003 132