State-Sponsored Mass Murder

Abstract

The author argues that openings in the political opportunity structure, rather than the levels of concentration of power, best predict the onset of genocides or politicides and which states will engage in the most severe state-sponsored mass murder. These and other hypotheses are tested. Analysis of logit models reveals that civil war involvement is the most consistent predictor of the onset of genocides or politicides, and other political opportunity structure variables have some effects, especially when in combination with at least one of the other political opportunity structure variables. Analysis of negative binomial event-count models also reveals that political opportunity structure variables best account for the degree of severity of a given genocide or politicide. In sum, openings in the political opportunity structure are more important in understanding what affects the onset and degree of severity of genocides and politicides than other more static variables.