LAUSR

This paper evaluates the effects of a program on Gini index where the selection to treatment depends on covariates. We propose a two-step nonparametric estimation procedure for the Gini inequality treatment effects. The proposed new estimator is shown to be consistent and has asymptotical normal distribution. We also show that the proposed estimator achieves semiparametric efficiency bound. Simulations confirm the theoretical results and show that the proposed estimator has good finite sample performance.