Inference of Structural Econometric Models: A Unified Approach

This paper considers the parametric inference of a wide range of structural econometric models. The class of models considered includes those with parameter-dependent support and those derived from game-theoretic models. Inference of those models has raised some important econometric issues. This paper addresses these issues within a unified framework using the indirect inference principle. Such an approach not only yields easy-to-implement consistent estimators with the standard root-n asymptotic normality, but also makes inference such as hypothesis testing and model specification and selection feasible. In particular, our approach leads to a method that resembles the conventional Cowles Commission structural econometrics spirit and can be viwed as an analog to the two-stage least squares (2SLS) estimator. Monte Carlo studies and an empirical analysis of the timber sale auctions held in Oregon illustrate the usefulness and feasibility of our approach.