Military capacity and state-perpetrated killings during internal conflicts

Existing quantitative studies of conflict rarely move beyond treating military capacity as a control variable to explicitly examine its relationship with civilian victimization. Addressing this gap, this paper argues that military capacity is a double-edged sword. On the one hand, lack of military capacity leads to agency loss, makes soldiers more “desperate”, and inhibits selective applications of violence. However, should physically harming civilians serve a strategic or ideological purpose for the government, military capacity would only facilitate the government’s implementation of this policy. Which side of the sword prevails, this paper argues, depends on the political costs of civilian victimization: military capacity aggravates one-sided violence when the government faces low costs. The paper evaluates this theoretical argument using dyadic data on one-sided violence from 1990 to 2011. The dataset includes all intrastate conflicts during this period that resulted in 25 or more battle deaths, encompassing 60 governments and 195 rebel groups. Empirically, military capacity increases one-sided violence in contexts where the government experiences limited political costs from victimization: in ethnic outgroups and autocracies. These associations are robust to alternative measures of the variables and different model specifications.