A question understudied in the literature on separatist conflict is why some governments permit secession to occur peacefully while others resist it militarily. To fill that gap, this dissertation presents a model that explains why dictatorships virtually always resort to military force to thwart self-determination movements while democracies usually permit secession. Dictators need a tough image to deter potential revolutionaries and palace coups and fear that ignoring separatist activity, even if it poses no direct threat to their rule, signals vulnerability that may encourage political rivals to challenge them. Peaceful secession from non-democracies only occurs when dictators initiate the separation for their own benefit, thereby avoiding the appearance of regime vulnerability.

The electorates of democracies, on the other hand, are reluctant to incur the costs of suppressing secessionist movements. An exception is when there exists in the territory claimed by the secessionists an enfranchised loyalist population that has reason to fear persecution should the secessionists win independence. When loyalists' livelihood depends on continued rule by the metropolis, the magnitude of the benefit to them of thwarting secession far exceeds the diffused cost to the metropolis's population of waging an anti-secessionist counterinsurgency. Consequently, if loyalists are enfranchised, they mostly become single-issue voters when secessionists threaten their privileged status and are able to out organize the peace camp.

This dissertation uses crisp set qualitative comparative analysis of the population of secessionist movements active between 1961 and 2007 and case studies to test its hypotheses. And, except for some anomalies surrounding India's counterinsurgencies in her Northeastern states and Kashmir, the model accounts for the entire population of cases.