A0973
Title: Higher-order moments of GMM estimators
Authors: Paul Rilstone - York University (Canada) [presenting]
Abstract: Explicit representations of the second-order mean-squared error, skewness and kurtosis of standard GMM estimators are obtained. These are shown to be written as quadratic and cubic functions of the second through fourth moments of the influence functions of these estimators. Iterating on the optimal weighting matrix is shown to reduce the terms in each of these moments. However, there is no gain in the second-order moments obtained from additional iterations.
