We study how seemingly benign changes to voting costs affect electoral turnout, exploiting plausibly exogenous variation in the assignment of polling places in Munich (Germany). Using an event study design, we find that polling place relocations cause a persistent shift from in-person to mail-in...