Weapon Technology and Civilian Targeting: Evidence from Ukraine

• In Event: Empirical Perspectives on the War in Ukraine

Abstract

The killing of civilians is a persistent feature of modern warfare. Consistently over the last twenty years, about 15 to 35 percent of all casualties were civilians (Khorram-Manesh et al., 2021; Högbladh, 2023). However, the exact reasons why civilians are targeted with such regularity are yet only partially understood. Extant theory generally interprets civilian targeting as motivated by an urge to dissuade collaboration with enemy forces. Civilians, this strand of the literature argues, can provide important resources to combatants such as shelter, food, and, most crucially, information (e.g., on troop movements or bases) on their adversary. Armed actors, therefore, strategically use violence against the civilian population to extract these resources and to prevent civilians from passing on intelligence to their enemy. However, this theory was developed in the context of civil wars where there are overlapping spheres of influence, and does not easily translate into settings of conventional war.

Expanding the theory, we argue, that motives of information control become relevant in interstate wars once the influence of modern weapon and information technology is accounted for. In the presence of long-range weapons, civilian collaboration poses a critical threat to combatants, not least because modern means of communication make it possible for civilians to transmit precise targeting information in real-time. An awareness of this threat, we argue, can lead armed actors to employ higher levels of violence against civilians when within the range of fire. We demonstrate our theory with geospatial data from the ongoing Russian war against Ukraine, a conflict that has seen widespread abuse of Ukrainian civilians at the hands of Russian soldiers. Our data comprises information at the 10x10km grid level for the first year of the conflict, including on the location of the front line on a weekly basis.

We show that areas occupied by Russia that were in range of Ukrainian cannon artillery systems saw 7 times higher levels of civilian targeting. This effect holds up to a variety of measurement strategies, including a treat-control design (comparing areas narrowly within and narrowly beyond artillery range) and panel analyses using fixed effects and difference-in-difference estimation (exploiting the movement of the front line over time). The relationship is especially strong before the introduction of long-range rocket artillery and in areas with comprehensive mobile network coverage. Our findings are of both theoretical and practical importance, shedding light on potential unintended side effects of the supply of long-range weapons, and enabling us to make better predictions about where and why one might expect civilians to be particularly vulnerable.

Authors

• Julius Valentin Koch, Uni Hamburg

• Max Schaub, University of Hamburg