Why incumbents perpetrate election violence during civil war

Abstract

Civil conflict increases incumbents’ vulnerability, expands their coercive capacity, enervates public good provision, and stifles public opposition. Consequently, we expect that elections held during civil conflict will feature more incumbent-perpetrated election violence. We test our argument with disaggregated data on election violence, generating two principal findings. First, elections held during civil conflict are more likely to feature violent coercion by incumbents. Second, this effect does not depend on the conflict's intensity or political salience, but is endemic to conflict-affected societies as a class. This raises questions about the nature of elections in conflict-affected societies and the relationship between forms of political violence.