Small Sample Bias of Alternative Estimation Methods for Moment Condition Models: Monte Carlo Evidence for Covariance Structures and Instrumental Variables

Abstract

It is now widely recognized that the most commonly used efficient two-step GMM estimator may have large bias in small samples. This problem has motivated the search for alternative estimators with better finite sample properties. Two classes of alternatives are considered in this paper. The first includes estimators which are asymptotically first-order equivalent to the GMM estimator, namely the continuous-updating, exponential tilting, and empirical likelihood estimators. Analytical and bootstrap bias-adjusted GMM estimators form the second class of alternatives. Two extensive Monte Carlo simulation studies are conducted in this paper for covariance structure and instrumental variable models. We conclude that all alternative estimators offer much reduced bias as compared to the GMM estimator, particularly the empirical likelihood and some of the bias-corrected GMM estimators analyzed.