What's Trending in Difference-in-Differences? A Synthesis of the Recent Econometrics Literature

Abstract

This paper synthesizes recent advances in the econometrics of difference-in-differences (DiD) and provides concrete recommendations for practitioners. We begin by articulating a simple set of "canonical" assumptions under which the econometrics of DiD are well-understood. We then argue that recent advances in DiD methods can be broadly classified as relaxing some components of the canonical DiD setup, with a focus on piq multiple periods and variation in treatment timing, piiq potential violations of parallel trends, or piiiq alternative frameworks for inference. Our discussion highlights the different ways that the DiD literature has advanced beyond the canonical model, and helps to clarify when each of the papers will be relevant for empirical work. We conclude by discussing some promising areas for future research.

Conducting Research in Marketing with Quasi-Experiments

“…Therefore, the treatment here is the review a book receives, and the After Treatment period occurs at different times for different books. Athey and Imbens (2022), Borusyak, Hull, and Jaravel (2022), Callaway and Sant'Anna (2020), De Chaisemartin and d'Haultfoeuille (2020), Goodman-Bacon (2021), and Roth et al (2022) explore the effects of variation in treatment timing. The issue is that because a fixed-effects DID estimator is a weighted sum of the treatment effect in each group and at each period, even though the weights sum to one, negative weights may arise when there is a substantial amount of heterogeneity in the treatment effects over time.…”

Government Demand and Domestic Firms Growth: Evidence from Uganda

“…The event study estimates could be biased even in the absence of pre-trends. A recent strand of econometric literature has highlighted that event study estimates could be contaminated by the treatment effects from other periods (for an overview of the issues and proposed solutions see Roth et al (2022) and De Chaisemartin and D'Haultfoeuille (2022)). Specifically, when the treatment is -as it is in our case -heterogeneous over time across different units, the effects we find may not provide the correct weighted average of treatment effects across units (and time).…”

Selection and parallel trends

“…Our second contribution is to the more recent literature on DiD methods. See, e.g., de Chaisemartin and D'Haultfoeuille (2021) and Roth, Sant'Anna, Bilinski, and Poe (2022) for recent surveys. Within this strand of the literature, our paper is most closely related to Roth and Sant'Anna (2021), Imbens (2022), andArkhangelsky, Imbens, Lei, andLuo (2021), though our focus greatly differs from theirs.…”