Centre for Development Economics
Department of Economics
Delhi School of Economics
ANNOUNCE A SEMINAR
Staggered Treatment and Heterogeneous Effects: Efficient Estimation using Forbidden Comparisons
by
(Ashoka University)
(Thursday, January 15, 2026, at TBA )
Venue: TBA
Abstract:-
The two way fixed effect regression (TWFE) has been a workhorse model to estimate the average treatment effect on treated (ATT) under difference in difference (DiD) setting where the treatment and control group satisfy the parallel trend assumption. However, recent literature discusses how including an already treated group as control (forbidden comparison) in the estimation recovers a biased ATT when adoption is staggered and the cohort specific treatment effect is heterogeneous over time. In response, several authors (Callaway and Sant’Anna (2021), De Chaisemartin and d’Haultfoeuille (2020), Sun and Abraham
(2021), Gardner (2022), Borusyak et al. (2024) etc.) proposed unbiased estimators of ATT by excluding these forbidden comparisons. In this paper, we argue that dropping forbidden comparisons, even though it ensures unbiasedness, costs loss of information and increases the variance of the estimator. We propose a framework which incorporates the forbidden comparisons while accounting for the bias under heterogeneous effects and recovers an unbiased, efficient, and asymptotically normal estimator. The proposed estimator is efficient not only under heterogeneous effects, but also under homogeneous effects. We confirm these results in a simulation study in which we simulated cohorts with differing treatment timings under homogeneous and heterogeneous effects. In the application, we replicate Beck et al. (2010) using ours and alternative estimators and found that the estimated ATTs are sensitive to the approach used.