This website requires JavaScript.

Doubly Robust Difference-in-Differences with General Treatment Patterns

Takahide Yanagi
Dec 2022
摘要
We develop a difference-in-differences method in a general setting in whichthe treatment variable of interest may be non-binary and its value may changein each time period. It is generally difficult to estimate treatment parametersdefined with the potential outcome given the entire path of treatment adoption,as each treatment path may be experienced by only a small number ofobservations. We propose an empirically tractable alternative using the conceptof effective treatment, which summarizes the treatment path into alow-dimensional variable. Under a parallel trends assumption conditional onobserved covariates, we show that doubly robust difference-in-differencesestimands can identify certain average treatment effects for movers, even whenthe chosen effective treatment is misspecified. We consider doubly robustestimation and multiplier bootstrap inference, which are asymptoticallyjustifiable if either an outcome regression function for stayers or ageneralized propensity score is correctly parametrically specified. Weillustrate the usefulness of our method by estimating the instantaneous anddynamic effects of union membership on wages.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答