Bootstrap bias-adjusted GMM estimators

Abstract

The ability of four alternative bootstrap methods to reduce the bias of GMM parameter estimates is examined in an instrumental variable framework using Monte Carlo analysis. Promising results were found for the two bootstrap estimators suggested in the paper.