Semiparametric Estimation and Inference in Partially Identified Econometric Models

Trustees Of Boston University, Boston

Investigators

Abstract

This award funds research in statistical methods for the analysis of economic data. The goal is to expand the scope of the existing statistical theory that explains the properties of estimation and inference methods for a partially identified model with set identified parameters. The research clarifies the main challenges to expanding the scope of the existing theory of semiparametric inference, develops new approaches to overcome these challenges, and develops practical suggestions for constructing asymptotically efficient estimators within semiparametric models. The research develops new methods that are robust against lack of identification, robust against misspecification, and asymptotically efficient within semiparametric models. Broader impacts will come from the use of these methods to evaluate the results of a variety of government and business decisions. The PI will make software code available to other researchers who want to use these new methods.