Partition as a Solution to Ethnic Civil War: Statehood, Demography, and the Role of Post-War Balance of Power for Peace

Abstract

Partition has been proposed as a way to (i) end ethnic civil wars and to (ii) build a lasting peace after ethnic civil wars end. This dissertation builds on partition theory and the ethnic security dilemma in three ways, demonstrating empirical support for a novel theory of why violence recurs following the end of ethnic civil wars and how partition can be used to prevent such violence. The dissertation begins by introducing the puzzle of ethnic group concentration: the social sciences have demonstrated that concentrated ethnic groups produce both peace and violence. The first case study discredits the notion that ethnic group concentration produced during ethnic civil wars will produce an end to ethnic civil wars. I conducted detailed field research, producing a longitudinal study of ethnic migration and violence in the Georgia-Abkhaz civil war (1992-1993), which acts as a crucial case. I conclude that partitioning groups does not end ethnic war. This is the first accurate empirical test of the ethnic security dilemma. Next, the dissertation looks at partition's ability to build peace by concentrating ethnic groups in new homeland states, and I argue that post-partition violence is caused by weak states and the triadic political space endogenously created by partitions that do not separate ethnic groups completely. I call this the Third Generation Ethnic Security Dilemma, building on previous ethnic security dilemma research. I test this empirically by introducing an index measuring the degree to which partitions separate ethnic groups, and I compare all ethnic civil war terminations between 1945 and 2004, demonstrating that partitions which completely separate ethnic groups provide a better chance for peace. Third, I selected two cases (Moldova and Georgia) to examine the causal processes of post-war recurring violence. Georgia, which experienced post-partition violence, and Moldova, which did not, act as a structured case comparison. I conclude that mixed ethnic demography interacts with state-building to cause or avert renewed violence.