A Comparative Analysis of Different Estimators for Dynamic Panel Data Models

It has become increasingly obvious that the estimation of dynamic panel data models has become one of the major issues in recent econometric research as evidenced by the plethora of papers on the subject. It is well known that the usual techniques for estimating panel data models are inconsistent in the dynamic setting. However, numerous consistent estimators have been proposed in the literature. In this paper, two new estimators are offered (one each for the fixed and random effects specifications), and their small sample performance compared with that of all of the existing estimators. It is hoped that the results of these experiments will provide invaluable guidance to applied researchers as to which is the preferred estimator(s). Finally, the divergences in point estimates of all of these estimators is illustrated with an application to a consumer demand schedule of laundry detergent in the metropolitan district of Melbourne, Australia.