Feasible bias-corrected OLS, within-groups, and first-differences estimators for typical micro and macro AR(1) panel data models

Abstract

In this paper we suggest several alternative ways of constructing feasible bias-corrected (FBC) pooled least squares, within-groups, and first-differences estimators for AR(1) panel data models. In a Monte Carlo simulation study involving data with the qualities normally encountered by both microeconomists and macroeconomists we found that the estimators proposed seem to possess better finite sample properties than the GMM estimators usually employed in this setting: most FBC estimators are unbiased, even when the time series is highly persistent, display less variability, and are not affected by the relative magnitude of the variances for the individual effect and the idiosyncratic error.