This article provides an applied introduction to Bartik instruments. The instruments attempt to reduce familiar endogeneity concerns in differential exposure designs (e.g., panel regressions with unit and time fixed effects). They isolate treatment variation due to the differential impact of common shocks on units with distinct pre-determined exposures. As a result, the instruments purge the treatment variation of possibly confounding factors varying across units over time. Given their broad applicability, Bartik instruments promise to provide researchers with a versatile new tool in their empirical toolbox to investigate relevant accounting questions.

JEL Classifications: C51; M40.

You do not currently have access to this content.