【12月18日讲座】美国加利福尼亚大学圣迭戈分校徐轶青：Causal Inference with Time-Series Cross-Sectional Data

讲座题目：Causal Inference with Time-Series Cross-Sectional Data

主讲人：徐轶青，美国加利福尼亚大学圣迭戈分校

主持人：谢岳，BETVLCTOR伟德唯一官网长聘教授

讲座时间：2018年12月18日（周二）14：00-15：30

讲座地点：上海交大徐汇校区 新建楼239会议室

讲座摘要：

Difference-in-differences and two-way fixed effects models are commonly used for causal inference with time-series cross-sectional (TSCS) data. They require the “parallel trends” assumption, which states that the average outcomes of treated and control units would have followed parallel paths in the absence of the treatment. In practice, this assumption is often violated due to the presence of time-varying confounders. To address this problem, I introduce two novel methods: the generalized synthetic control method (Xu 2017) and trajectory balancing (Hazlett and Xu 2018). The former adopts a model-based approach and imputes treated counterfactuals using a latent factor model; the latter employs a reweighting approach and seeks balance in pre-treatment outcome trajectories and covariates between the treatment and control groups. I illustrate these two methods using several empirical examples from political science.