Semiparametric estimation of long-term treatment effects*

被引：5

作者：

Chen, Jiafeng ^{[1
]}

Ritzwoller, David M. ^{[2
]}

机构：

[1] Harvard Sch Business, Boston, MA USA

[2] Stanford Grad Sch Business, Stanford, CA 94305 USA

来源：

JOURNAL OF ECONOMETRICS | 2023年 / 237卷 / 02期

基金：

美国国家科学基金会;

关键词：

Long-Term Treatment Effects; Semiparametric Efficiency; Experimentation; PROPENSITY SCORE; MISSING DATA; INFERENCE; MODELS; ECONOMETRICS; EFFICIENCY; TRIALS; LASSO; GMM;

D O I：

10.1016/j.jeconom.2023.105545

中图分类号：

F [经济];

学科分类号：

02 ;

摘要：

Long-term outcomes of experimental evaluations are necessarily observed after long delays. We develop semiparametric methods for combining the short-term outcomes of experiments with observational measurements of short-term and long-term outcomes, in order to estimate longterm treatment effects. We characterize semiparametric efficiency bounds for various instances of this problem. These calculations facilitate the construction of several estimators. We analyze the finite-sample performance of these estimators with a simulation calibrated to data from an evaluation of the long-term effects of a poverty alleviation program.

引用

页数：18

共 78 条

[1]

Athey S, 2020, Arxiv, DOI arXiv:2006.09676

[2] Using Wasserstein Generative Adversarial Networks for the design of Monte Carlo simulations☆ [J].