B0846
Title: Efficient estimation of average treatment effect on the treated under endogenous treatment assignment
Authors: Trinetri Ghosh - University of Wisconsin-Madison (United States) [presenting]
Menggang Yu - University of Wisconsin - Madison (United States)
Jiwei Zhao - University of Wisconsin-Madison (United States)
Abstract: When evaluating a complex intervention, instead of the average treatment effect (ATE), researchers are more interested in the average treatment effect on the treated (ATT), a quantity that is more relevant and more interpretable to policy-makers. We consider the ATT estimation motivated by a case study, where the treatment assignment might depend on the potential untreated outcome and hence is endogenous. We focus on the efficient estimation of ATT by characterizing the geometric structure of the model, deriving the semiparametric efficiency bound for ATT estimation, and proposing an estimator that can achieve this bound. We rigorously establish the theoretical results of the proposed estimator. The finite-sample performance of the proposed estimator is studied through comprehensive simulation studies and an application to our motivating study.