Identifying the effect of mobilization on voter turnout through a natural experiment

• We identify the effect of overall mobilization on voter turnout by leveraging a natural experiment in Japan.
• Almost concurrently held municipal elections boost voter turnout of prefectural elections by one to two percentage points.
• We argue that the estimated effect is attributable to mobilization, rather than cost sharing or psychological stimulus.

Although numerous get-out-the-vote field experiments have identified the effects of particular mobilization tactics (e.g., canvassing, phone calls, direct mails) on voter turnout, we do not yet have a full understanding of the causal effect of overall mobilization. We study this by leveraging a natural experiment in Japan, in which the timing of a municipal election is as-if randomly assigned. The results show that almost concurrently held municipal elections boost these municipalities’ voter turnout in prefectural elections by one to two percentage points. We argue that some unique settings in Japan allow us not only to mitigate omitted variable bias but also to attribute the estimated effect only to mobilization, rather than the effects of cost sharing and psychological stimulus.