Nonparametric Estimation in Large Panels with Cross Sectional Dependence

In this paper we consider nonparametric estimation in panel data under cross sectional dependence. Both the number of cross sectional units (N) and the time dimension of the panel (T ) are assumed to be large, and the cross sectional dependence has a multifactor structure. Local linear regression is used to …lter the unobserved cross sectional factors and to estimate the nonparametric conditional mean. A Monte Carlo study shows that the proposed estimator yields satisfactory …nite sample properties. Keywords: Cross sectional dependence; mixing process; large panels; local linear regression. JEL Classi…cation: C14; C23 The author thanks the editor and two referees for useful comments that signi…cantly improved the paper. The author also thanks Aman Ullah, Brett Katzman, Timothy Mathews and seminar participants at the 2006 Econometric Society Summer Meetings for helpful comments. All remaining errors are mine. Financial support from Kennesaw State University and U.C. Riverside is greatly acknowledged. Correspondence address: Department of Economics, Finance and Quantitative Analysis, M.D. 403, Kennesaw State University, Kennesaw, GA 30144, U.S.A. Phone: (770)423-6318. Fax: (770)499-3209. E-mail: [email protected]