mined; Lagged

This paper investigates identification and root-n consistent estimation of a class of single index panel data models where: the index function is unspecified; the individual effects may be correlated with all the explanatory variables, and; all the explanatory variables may be prede-termined, including lagged dependent variables. The model is extended to allow for a general form of sample selection. For both estimators, we propose kernel based modified backfitting algorithms to estimate the finite and infinite dimensional parameters of interest. The algorithms fully implement all the identification restrictions of the models. We derive consistency and asymptotic normality results for the proposed estimators. Finally, Monte Carlo simulations are performed to study the finite sample properties of both estimators