Does female leadership reduce corruption? We study this question using close mixed-gender elections in Brazilian municipalities over two decades and multiple measures of corruption: budget-based predicted corruption scores, audit irregularities, and legal sanctions. We find no evidence that electing a woman mayor affects corruption. This null result holds across time periods, mayor characteristics, and the electoral cycle. We only detect a negative effect in the small subsample of municipalities randomly audited in early terms, which coincides with a sharp imbalance in incumbency. Because incumbency directly impacts corruption, these previously documented effects likely reflect the impact of incumbency rather than gender.