题目：Review of Treatment Effect Models and Some New Results

报告人：许育进

时间：2017年3月24日13:30-16:30

地点：暨南大学中惠楼106B会议室

主办方：暨南大学经济与社会研究院

报告人简介：

Yu-Chin Hsu (许育进) is currently an associate research fellow in the Institute of Economics, Academia Sinica. He got his Ph.D degree in May 2010 under Stephen G. Donald's supervision from the Department of Economics, University of Texas at Austin. His research focuses on econometrics.

讲题摘要：

Many empirical studies in economics and other social sciences are interested in the causal effects of programs or policies. The main problem studied in this literature is to evaluate the effect of the exposure of a set of units to a program, or treatment, on some outcome. In economic studies, usually the units are individuals, households, markets, firms, counties, states, or countries and the treatments can be job training program, educational programs, vouchers, laws and government policies. For example, we are interested in the effect of a job training program on individual’s income, hourly wage or employment status. For another example, we are also interested in the effect of attending colleges on individual’s future income.

In the last two decades, much research has been done on the econometric and statistical analysis of such causal effects. This course is served as a review on recent development with a focus on the Rubin Causal Model. We review various treatment effects under the assumption that the treatment assignment is unconfounded which is also known as selection-on-observables and conditional independence. Then we extend the results to cases where the treatment assignment is endogenous, but a valid binary instrument is available for the treatment assignment. We also discuss several tests for the unconfoundedness assumption. Last, we summarize some of my research in this area.

课程安排：

1.Estimation and Inference under unconfoundedness.

2.Estimation and Inference when the treatment assignment is endogenous.

3.Tests for unconfoundedness assumption.

4.Summary on Yu-Chin Hsu’s research.