Two-Way Fixed Effects Estimators with Heterogeneous Treatment Effects
Clément de Chaisemartin, Xavier D'Haultfœuille
Summary
The paper shows that linear regressions with group and period fixed effects estimate a weighted sum of treatment effects across cells where some weights can be negative. As a consequence the coefficient can be negative even when every cell-level effect is positive, undermining its interpretation as an average treatment effect. The authors propose an alternative estimator (DID_M) that is robust to heterogeneous effects and assess the empirical relevance of the negative-weighting problem.
Key findings
- Two-way fixed effects estimands are weighted averages of treatment effects with potentially negative weights.
- The estimate can have the opposite sign of every underlying treatment effect under heterogeneity.
- Proposes a heterogeneity-robust DID_M estimator and tools to gauge the severity of negative weighting in applications.
Subjects & keywords
Cite this paper
Clément de Chaisemartin, & Xavier D'Haultfœuille (2020). Two-Way Fixed Effects Estimators with Heterogeneous Treatment Effects. American Economic Review. https://doi.org/10.1257/aer.20181169
@article{chaisemartin2020twoway,
author = {Clément de Chaisemartin and Xavier D'Haultfœuille},
title = {Two-Way Fixed Effects Estimators with Heterogeneous Treatment Effects},
journal = {American Economic Review},
year = {2020},
doi = {10.1257/aer.20181169},
url = {https://doi.org/10.1257/aer.20181169}
}