Why would a leader ever commit mass murder? Most of the research on genocide/politicides essentially argues that broad political participation in a society is sufficient to ensure good public policy and benevolent political leaders. In a previous study we (Aydin & Gates, 2007) disaggregated political institutions into three dimensions; political participation, executive recruitment, and executive constraints. Our analysis demonstrated that institutional limitations on executive power are particularly effective in mitigating mass killing whereas simply mass political participation increases the likelihood that civilians are targeted by their own government. While these findings are important, they do not address why leaders target and kill their people in the first place.
This paper examines the question directly by formulating a formal theory of repression and the escalation of repression to mass murder. We argue that mass killings are not discrete outcomes where a leader either commits murder against her subjects or not. Instead, they are a product of an escalatory process where repressive government policies prior to the mass murder present an early warning system to the international community about its likelihood. As such, state leaders systematically repress a group or groups of citizens and deprive them of certain rights over a period of time which increase the risk of and may culminate in mass killings. Using a binary probit model, we empirically test our propositions regarding government repression and its escalation to the level of geno/politicides using Harff's (2003) dataset, 1955–2001.