B1108
Title: A Stein-type shrinkage limited information maximum likelihood estimator
Authors: Muhammad Qasim - Jonkoping University (Sweden) [presenting]
Abstract: A Stein-type shrinkage estimator is proposed which is a weighted average of the ordinary least squares (OLS) and limited information maximum likelihood (LIML) estimators, with the weights inversely proportional to the Hausman test statistic. We derive the asymptotic distribution of the new estimator by means of local-to-exogenous asymptotic belief. In addition, the asymptotic risk of the Stein-type LIML estimator is calculated and shows that the risk is strictly smaller than the risk of the LIML under certain conditions. The Monte Carlo simulation and empirical application are considered to demonstrate the superiority of Stein-type LIML to the classical OLS and LIML estimators in the presence of many weak instruments and endogeneity.
