Estimation of a binary dependent variable sample selection model with endogeneity under a nonparametric selection mechanism

Abstract

In this paper, we consider an estimation of a sample selection model with a binary dependent variable and possible endogenous regressors. Under the semiparametric framework, we impose neither the parametric specification of the error distribution nor the functional form of the selection equation, which largely reduces the risk of model misspecification. Throughout this paper, we present the identification 条件 and propose a two-step semiparametric maximum likelihood estimator with the first step being a nonparametric regression for the selection variable. A control function approach is used to control for the possible endogeneity. The proposed estimator is shown to be consistent and asymptotically normal. In the simulation studies, we compare the finite sample properties of our estimator with those of existing alternatives, demonstrating the significant advantages of our approach especially in robustness to the model misspecification. Finally, an economic application on labor economics is carried out to illustrate the usefulness of our estimator.