This paper develops the identification and estimation of nonlinear semi-parametric panel data models with mismeasured variables and their corresponding average partial effects using only three periods of data. The past observables are used as instruments to control the measurement error problem,...

We investigate a nonparametric panel model with heterogeneous regression functions. In a variety of applications, it is natural to impose a group structure on the regression curves. Specifically, we may suppose that the observed individuals can be grouped into a number of classes whose members...