Econometric Methods with Randomly Censored Data: An Overview and Some Recent Results
2019/11/26
研討會日期 : 2019-11-26
時間 : 14:00
主講人 : Professor Pedro H. C. Sant'Anna
主持人 : Professor Yu-Chin Hsu
地點 : Conference Room B110
演講者簡介 : Professor Sant'Anna received his Ph.D. in Economics at Universidad Carlos III de Madrid, Spain in 2015. He is currently an Assistant Professor at Vanderbilt University. His research interests are Econometric Theory, Policy Evaluation, Semi and Nonparametric Methods, Specification Testing.
演講摘要 : In many economic applications, the outcome of interest, typically a duration variable is subject to right-censoring. A few examples include survival times, unemployment duration, time to recidivism, and time to corporate bankruptcy. In such cases, the use of standard econometric techniques that do not account for this data limitation usually lead to invalid inference procedures. The main goal of this presentation is to provide an overview of statistical and econometric procedures that are suitable to censored data, highlighting some results that are applicable to program evaluation and causal inference problems. We will cover conditional hazard, censored distribution regression, and censored quantile regression models. Furthermore, we will discuss some recent results on two-step Kaplan-Meier estimators that allow us to study a variety of treatment effect parameters under different identifying assumptions including unconfoundedness, local treatment effects, and nonlinear differences-in-differences.
