Some elements on nonparametric random coefficients models

Abstract

In this talk I will give a review of nonparametric random coefficients models. We will start with the linear random coefficients model and discuss new topics such as the identification of such models when we want to allow for nonlinearities and regressors with limited variation. We will then present binary and multinomial discrete choice models. Finally we will discuss estimation of linear models with continuous or binary endogenous regressors using a triangular system with one or multiple sources of unobserved heterogeneity.