Autocratic elections are often marred with systematic intimidation and violence toward voters and candidates. When do authoritarian regimes resort to violent electoral strategies? I argue that electoral violence acts as a risk-management strategy in competitive authoritarian elections where: (a) the regime’s capacity for coopting competitors, local elites, and voters is low, and (b) the expected political cost of electoral violence is low. I test these propositions by explaining the subnational distribution of electoral violence during the most violent election in Mubarak’s Egypt (1981-2011): the 2005 Parliamentary Election. The results indicate that electoral violence is higher in districts where: the regime’s capacity for coopting local elites and competitors is low, clientelistic strategies are costlier and less effective, and citizens’ capacity for non-electoral mobilization is low. The conclusions provide lessons for efforts to contain electoral violence in less democratic contexts.