Modified generalized instrumental variables estimation of panel data models with strictly exogenous instrumental variables

Abstract

In models for panel data, strict exogeneity assumptions generate many more moment conditions than are used by standard estimators. An efficient estimator based on a much smaller set of moment conditions can often be obtained by the generalized instrumental variables (GIV) principle. However, for certain classes of models the GIV estimator is not consistent. Such models include Hausman and Taylor-type models with unrestricted autocorrelation, and models with time-varying individual effects. We propose a modified GIV estimator and show how it leads to efficient estimation of these models. Acknowledgment The first author gratefully acknowledges the financial support of the College of Business and Dean’s Council of 100 at Arizona State University, the Economic Club of Phoenix, and the alumni of the College of Business. The second author gratefully acknowledges the financial support of the National Science Foundation.