Recent studies show a correlation between female representation in government and reduced corruption.

This paper examines French municipal elections using regression discontinuity and first difference designs. We track the impact of newly elected vs re-elected women mayors on corruption levels over time.

Our findings reveal:

• Women mayors initially reduce corruption risks through deliberate network exclusion
• This effect disappears in municipalities where women are reelected

These results align with marginalization theories but suggest adaptation to corrupt networks is key for longevity.