Abstract
Elections define representative democracies but also produce spikes in physical mobility if voters need to travel to polling places. In this paper, we examine whether large-scale, in-person elections propagate the spread of COVID-19. We exploit a natural experiment from the Czech Republic, which biannually renews mandates in one-third of Senate constituencies that rotate according to the 1995 election law. We show that in the second and third weeks after the 2020 elections (held on October 9-10), new COVID-19 infections grew significantly faster in voting compared to non-voting constituencies. A temporarily related peak in hospital admissions and essentially no changes in test positivity rates suggest that the acceleration was not merely due to increased testing. The acceleration did not occur in the population above 65, consistently with strategic risk-avoidance by older voters. Our results have implications for postal voting reforms or postponing of large-scale, in-person (electoral) events during viral outbreaks.