Staggeringly Problematic: A Primer on Staggered DiD for Accounting Researchers

Abstract

This paper presents the staggered difference-in-differences (DiD) method in a manner that is accessible to a broad accounting research audience, offering insights from the perspective of an applied accounting researcher. It specifically surveys the challenges in DiD design when multiple units receive staggered treatment over time. Utilizing the Goodman-Bacon decomposition, I illustrate how heterogeneous treatment effects can introduce bias in the estimated treatment effect within a staggered DiD framework, especially when estimated using a two-way fixed effects regression. Additionally, the paper briefly reviews several recent proposals from the burgeoning field of econometrics, aimed at adjusting for this bias and effectively implementing staggered DiD designs. These proposals are exemplified through the case study of the staggered adoption of the 150-hour Rule.

Keywords: Difference in differences; staggered treatments; treatment effect heterogeneity; accounting research

JEL Classification: C1, C2, M4