Rain, temperature, and votes: Weather effects on electoral participation in Argentina

Compulsory voting was designed to make electoral participation immune to individual costs, yet this paper shows that rain on election day still keeps Argentine voters home. Exploiting quasi-random variation in precipitation across 526 departments and 14 national election days between 2011 and 2023, I estimate two-way fixed effects models with department and election fixed effects using ERA5 reanalysis data matched to department centroids. Each additional millimeter of voting-hours precipitation (8am–6pm) reduces turnout by 0.08 percentage points ( p < 0.01); on a day of average rainfall this implies about a 0.2 percentage-point reduction. The estimate survives an extensive identification battery: balance tests, a 9-day placebo window with horse races, grid-cell clustering, and a validation of the reanalysis data against independent rain-gauge stations. Using newly constructed vote shares by ideological bloc, I find that rainfall does not bias the partisan outcome—it leaves the incumbent vote share, the left/Peronist bloc, and the left–right balance unchanged, in contrast to the partisan effects documented under voluntary voting in Europe—and it has no effect on blank ballots, null votes, or fragmentation. Weather thus operates through the extensive margin (whether to vote) rather than the intensive margin (how to vote). If even a century-old compulsory voting institution cannot fully insulate turnout from weather shocks, institutional reforms such as postal voting, extended hours, or remote electronic voting deserve consideration even where voting is mandatory.