Semiparametric Estimation of Long-Term Treatment Effects

By Jiafeng Chen et al

Published on Aug. 21, 2023

Read the original document by opening this link in a new tab.

Table of Contents

1. Introduction
2. Problem Formulation
3. Estimands
4. Identifying Assumptions
5. Latent Unconfounded Treatment
6. Statistical Surrogacy
7. Notation

Summary

This paper develops semiparametric methods for estimating long-term treatment effects by combining short-term experimental and long-term observational data sets. The authors propose two models and derive efficiency bounds for estimating long-term average treatment effects. They analyze the performance of estimators and compare different approaches. The paper contributes to the literature on missing data models and semiparametric efficiency in missing data models.