Estimation of quantile treatment effects with Stata

In this article, we discuss the implementation of various estimators proposed to estimate quantile treatment effects. We distinguish four cases involv- ing conditional and unconditional quantile treatment effects with either exogenous or endogenous treatment variables. The introduced ivqte command covers four different estimators: the classical quantile regression estimator of Koenker and Bassett (1978, Econometrica 46: 33–50) extended to heteroskedasticity consis- tent standard errors; the instrumental-variable quantile regression estimator of Abadie, Angrist, and Imbens (2002, Econometrica 70: 91–117); the estimator for unconditional quantile treatment effects proposed by Firpo (2007, Econometrica 75: 259–276); and the instrumental-variable estimator for unconditional quantile treatment effects proposed by Fr ̈olich and Melly (2008, IZA discussion paper 3288). The implemented instrumental-variable procedures estimate the causal effects for the subpopulation of compliers and are only well suited for binary instruments. ivqte also provides analytical standard errors and various options for nonpara- metric estimation. As a by-product, the locreg command implements local linear and local logit estimators for mixed data (continuous, ordered discrete, unordered discrete, and binary regressors).