The Economics of Civilian Victimization: Evidence from World War II Italy

What drives patterns of violence towards civilians in a war context and at the micro-level? We use evidence from the WWII Italian Campaign (July 1943 – May 1945), when Allied forces slowly pushed Axis troops North along the Italian peninsula. We exploit plausibly exogenous local variation in the activation of 32 front lines that affected both the potential gains from civilian victimization, due to increased insecurity, and its cost, resulting from decreased accountability. In a stacked Difference-in-Differences framework, we compare treated municipalities that happened to be in the combat zone (within 40km of the front line on the German side) with comparison municipalities that remained distant from the front line, on the day of activation of each front line. We find that indiscriminate violence (the likelihood of mass civilian killings and of vulnerable civilians ones) increased 10-fold. By contrast, the likelihood of more targeted, individual killings did not change. The effect is concentrated in areas away from division headquarters, tasked with policing soldiers, while locations more exposed to allied bombing and partisan resistance show relatively low levels of indiscriminate violence. We interpret the evidence as suggestive of an accountability mechanism that fostered indiscriminate violence through a reduction in the cost of misbehavior, rather than an increase in the gains linked to increased insecurity. Further, we explore treatment heterogeneity in terms of background of the perpetrators and local incentives to extraction in support of the war effort.