On a simultaneous equations pre-test estimator

Abstract

This paper investigates the finite-sample properties of a pre-test estimator for the linear regression model when endogenous regressors render the ordinary least squares estimator inconsistent. A Wu-Hausman exogeneity test is applied to determine whether two-stage least squares is appropriate. Both analytic and nonparametric approximations to the pre-test risk and bias are discussed. © 1995.