How does mandatory voting affect the risk of electoral violence? Most studies on the use of violence as a campaign strategy assume that parties use violence to decrease turnout among rival supporters. While there is not much systematic evidence yet on the effects of such violence, the most robust finding seems to be that it decreases turnout, rather than mobilizes voters to participate in an election. We argue that if this is the case, elites should have fewer incentives to engage in pre-electoral violence when voting is compulsory, as turnout cannot effectively be influenced. We thus expect the likelihood of electoral violence to be lower when voting is mandatory. If parties do engage in pre-electoral violence in such contexts, however, we expect parties to escalate violence, as high levels of violence will be necessary to deter voters from casting their compulsory vote. We test implications of our argument on a sample of all Latin American countries from 1990 to 2012, using data on violence from CREV and ECAV. Our research design exploits cross-sectional as well as temporal variation in the introduction and the enforcement of compulsory voting.
