Skip to content Skip to navigation

A Computational Implementation of GMM

Aug 2015 to Jul 2018



National Science Foundation (NSF)

Intellectual Merit:  This project studies a method of implementing Quasi-Bayes estimators for GMM models allowing for nonlinearity and nonseparability. This method combines simulation with nonparametric regression in the estimation of nonlinear complex parametric models. We study both kernel and local polynomial methods, and allow for both exact and over identification. We also demonstrate the asymptotic validity of inference based on simulated posterior quantile regression. In ongoing work, we are studying the combination with sieve and bootstrap methods.

Broader Impact:  The results obtained from the project can provide very useful guidance to em-pirical researchers who make extensive use of computational intensive nonlinear models for which obtaining the estimator and conducting inference on the parameter of interest can both be nu-merically challenging. Beyond applications in economics, nonlinear models are also widely used in statistics and various disciplines in social sciences and natural sciences, where researchers often resort to simulation and method of moment based methods for estimation and inference. Our analysis provides a useful alternative computation scheme for empirical researchers applying these models to different types of data sets.