Catch and release: Expansion, contraction, and the downsizing of states

This dissertation asks two questions about secession. First, what explains the proliferation of states and secessionist movements in the post-1945 period? Second, when do governments fight to prevent secessionist movements from forming an independent state? None of the existing research addresses these puzzles in their entirety. One of the chief gaps in the literature is an account of how governments balance external security concerns with domestic demands for independence.;My central argument is that the international environment has altered the costs and benefits of holding territory, but it is the internal structure of states that governs how they downsize. Just as competitive pressures in the 19th and early 20th centuries compelled states to expand and secure territory and economic resources, changing security and economic conditions in the post-1945 period have led to an unwinding of that process as numerous states have chosen to contract and permit the secession of peripheral regions. But while the international system has rendered secession more likely, it is the internal characteristics of states that govern how they downsize. First, states have to beware lest the release of one unit sets a precedent for others and risks a cascade of further secessionism. I argue that the internal organization of states matters and that governments rely on notions of administrative rank when determining which groups can secede without fear of setting a precedent, and who they must fight in order to maintain a credible reputation. Second, democracies are more likely to permit secession provided the attempt is made through non-violent and institutional means.;To evaluate this argument, I collected original data on internal administrative units (proto-states) and secessionist movements over the period 1816-2005. I then tested my hypotheses in a mixed-methods approach using quantitative analysis and case studies on three countries: England, Russia, and India.