Description:
This study investigates the relationship between women's political empowerment and quality of government in European regions. Given that the association is expected to be endogenous, we use exogenous variation in pre-industrial societal traits, legal origins, geographic characteristics, and historical socio-demographic features to build instrumental variables for the quality of government and female empowerment. As a novelty, we use Random Forest forecasting, which is a machine learning technique that helps build strong instruments with greater predictive accuracy than other approaches. Our findings show that women's political empowerment raises the quality of government, and that the quality of government also boosts female empowerment. This result is robust to the effect of influential observations and outliers, the measurement of women’s empowerment, the treatment of the endogeneity of other covariates, or alternative estimation strategies allowing for institutional spillovers across regions