A Complete Framework for Model-Free Difference-in-Differences Estimation

Viac o knihe

The book presents a comprehensive framework for conducting model-free difference-in-differences analysis with covariates, focusing on nonparametric estimation and testing. It details the process of selecting confounders and defining outcome scales, followed by estimating heterogeneous treatment effects and average effects for treated subjects. The authors discuss the behavior of estimators and tests, including bootstrap methods for standard errors and p-values, with automatic bandwidth selection. The framework's application is illustrated through a study on the Deferred Action for Childhood Arrivals program's impact on educational outcomes for non-citizen immigrants in the US.