Voting then, voting now: the long-term consequences of participation in South Africa's first democratic election

In a developing democracy, can elections change the behavior of participating citizens? I study the effect of voting in South Africa's first democratic election in 1994 on future voting and present evidence of the lasting behavioral effects of past participation. Eligibility to participate in 1994 affects future voting by 3 percentage points, with an average treatment effect of actually voting between 3.5 and 8.5 percentage points. Given low turnout rates, these effects account for 7%���20% of the size of the electorate. To explain this result, I propose a theoretical explanation that draws on psychological research���affective experience habituation. I argue that persistence (or habituation) in voting behavior is at least partly driven by the creation of associations between first-time voting and positive emotional states. This theory is tested within the context of the 1994 election by exploiting variation in electoral experiences.