An incidental parameters free inference approach for panels with common shocks

This paper develops a novel Method of Moments approach for panel data models with endogenous regressors and unobserved common factors. The proposed does not require estimating explicitly large number parameters in either time-series or cross-sectional dimension, T N respectively. Hence, it is free from the incidental parameter problem. In particular, suffer “Nickell bias” order O(T?1), nor bias terms that are O(N?1). Therefore, can operate under substantially weaker restrictions compared to existing procedures. Two alternative GMM estimators analyzed; one makes use fixed “averaged equations” à la Anderson Hsiao (1982), whereas other “stacked equations”, total which increases at rate O(T). It demonstrated both consistent asymptotically mixed-normal as N?? any value T. Low-level conditions ensure local global identification this setup examined using several examples.