Polling Place Location and the Costs of Voting

We study how distance to one’s polling place affects the likelihood of voting using a geographic regression discontinuity design with data from Pennsylvania and Georgia. A one-mile increase in distance to the polling place reduces the likelihood of voting in person by 1 to 3 percentage points. Effects are two to three times higher among those closest to the polling place. When available, voters substitute to mail-in voting as distance to the polling place increases. In counterfactual exercises, we identify turnout-maximizing polling places. Some precincts have large potential gains in turnout, even when choosing from existing buildings. (JEL D72, K16, R23)