Estimation of Partially Linear Panel Data Models with Cross-Sectional Dependence

This paper studies the estimation of the partially linear panel data models, allowing for cross-sectional dependence through a common factors structure. This semiparametric additive partial linear framework, including both linear and nonlinear additive components, is more flexible compared to linear models, and is preferred to a fully nonparametric regression because of the ‘curse of dimensionality’. The consistency and asymptotic normality of the proposed estimators are established for the case where both cross-sectional dimension and temporal dimension go to infinity. The theoretical findings are further supported for small samples via a Monte Carlo study. The results suggest that the proposed method is robust to a wide variety of data generation processes.