Generalized RIF Regressions

This paper suggests a generalization of covariates shifts to study distributional impacts on inequality and distributional measures. It builds on the recentered influence function (RIF) regression method designed for location shifts on covariates, but it extends to general policy interventions such as location-scale or asymmetric interventions. Numerical simulations for Gini, Theil and Atkinson indexes show it has a very good performance on a myriad of cases and distributional measures. An empirical application to studying changes in Mincerian equations illustrates the method.