Is compulsory voting habit-forming? Regression discontinuity evidence from Brazil

Voting in one election increases one's propensity to vote in the future. It remains unclear, however, whether this pattern holds when voting is compulsory - as is the case in a quarter of all democracies. Is compulsory voting habit-forming? I address this question using a regression discontinuity design and administrative turnout data from Brazil, where voting is voluntary at age 16 and compulsory at age 18. I find no evidence that compulsory voting instils voting habits. Instead, the evidence points to a first-time compulsory voting boost, which gradually dissipates as voters grow older. I show that targeted mobilisation of first-time compulsory voters is a plausible mechanism behind the turnout boost. Alternative explanations find less support in the data. The results clarify the scope conditions of prior research on voting habits, and have important implications for the debate over the second-order effects of compulsory voting.