15 Some aspects of generalized method of moments estimation

Abstract

Publisher Summary This chapter provides an introduction to generalized method of moments (GMM) which is designed to acquaint the applied researcher with the basic ideas behind GMM and its statistical properties. The emphasis is placed on intuition and not mathematical rigor. (GMM) estimation have a considerable impact on econometrics. It provides a unifying framework for the analysis of many familiar estimators and includes least squares, instrumental variables and maximum likelihood as special cases. It also offers a convenient method of estimation in certain models which were computationally very burdensome to estimate by more traditional methods. GMM has been applied to a wide variety of models including the estimation of probit models with panel data and nonlinear rational expectations models. Because of the generality of the GMM estimation principle, many theoretical treatments of the estimator are at an advanced level. While this generality is very desirable for practical purposes, it can make it more difficult for the less technical reader to understand the intuition behind the estimation procedure. The chapter discusses instrumental variables (IV) estimation in the classical linear regression model, Hansen's (1982) GMM estimator for nonlinear dynamic models, the application of this technique by Hansen and Singleton (1982) to the estimation of Euler equation models, and recent research on GMM based estimation and inference.