When Can We Use Two-Way Fixed Effects (TWFE)? A Comparison of TWFE and Novel Dynamic Difference-in-Differences Estimators

The conventional two-way fixed effects (TWFE) estimator has come under increasing scrutiny in settings with staggered treatment adoption. Recent research shows that TWFE can be biased when treatment effects vary across time or groups, prompting the development of several alternative dynamic difference-in-differences (DiD) estimators. Yet for applied researchers, it often remains unclear when TWFE is biased, what the newer estimators solve, and what limitations they all have. The authors provide an accessible overview of this literature and compare conventional TWFE with five dynamic DiD estimators using Monte Carlo simulations. The simulations consider realistic scenarios with time- and group-varying treatment effects, anticipation effects, and violations of parallel trends. The results show that much of the bias in TWFE reflects functional-form misspecification: when treatment effects are modeled in event time, TWFE performs well under time heterogeneity, with only modest bias in later posttreatment periods when strong group-specific heterogeneity is present. None of the new estimators is universally superior, and all involve trade-offs, with some being more robust to anticipation effects and others less sensitive to deviations from parallel trends. The authors illustrate these findings in an application on the motherhood earnings penalty using UK panel data, in which the dynamic treatment profiles are broadly similar across estimators despite some differences in long-run effects. The authors conclude that applied researchers should carefully consider potential effect heterogeneity and model it appropriately, whatever estimator they use. The authors also challenge the presumption that TWFE should invariably be abandoned in favor of newer DiD estimators.