Semiparametric three step estimation methods in labor supply models

The aim of this paper is to provide an alternative way of specification and estimation of a labor supply model. The proposed estimation procedure can be included in the so called predicted wage methods and its main interest is twofold. First, under standard assumptions in studies of labor supply, the estimator based on predicted wages is shown to be consistent and asymptotically normal. Moreover, we propose also a consistent estimator of the asymptotic variance covariance matrix estimator. Secondly, we introduce a semiparametric estimator based on marginal integration techniques that allows for nonlinear relationships between working hours and other explanatory variables. We show the asymptotic properties of this estimator and we compare the results empirically against those obtained in standard three step estimators based on predicted wages.