Staggeringly Problematic: A Primer on Staggered DiD for Accounting Researchers

Abstract

This paper presents the staggered difference-in-differences (DiD) method in an accessible language to a broad accounting research audience from an applied researcher's perspective. Specifically, the paper examines DiD design problems when multiple units are treated and when the treatment is conducted at staggered periods in time. Using the Goodman-Bacon decomposition, I show how heterogeneous treatment effects can bias the estimated treatment effect in a staggered DiD when estimated using a two-way fixed effect regression. Finally, using the staggered adoption of the 150-hour Rule as an example, I briefly review several proposals to adjust for the bias and correctly implement staggered DiD designs that a bourgeoning literature has put forward in econometrics.