We estimate the causal effect of the type of electoral system on corruption by exploiting a feature of French municipal elections: Municipalities with fewer than 1,000 inhabitants use an individual majority system, while those above this threshold use a proportional closed-list system. Crossing the threshold therefore implies two opposite changes: reduced individual accountability, which may increase corruption, and stronger collective control, which may reduce it. Using a regression discontinuity design with both survey and actual corruption data, we find that the individual accountability effect dominates, as exceeding the threshold leads to higher levels of both perceived and actual corruption.