Panel.EstimandCharacterization
What popular panel regressions actually estimate: formal characterizations of TWFE and event-study estimands from the recent difference-in-differences literature.
EventStudyContamination 72 core · 20 supporting · 6 submodules Sun-Abraham: contamination of event-study coefficients under heterogeneous cohort effects and the interaction-weighted fix. FlexibleDIDMundlak 50 core · 26 supporting · 6 submodules Wooldridge's extended TWFE: the equivalence of pooled OLS with saturated interactions and imputation-style estimands. HeterogeneousTWFE 17 core · 18 supporting · 3 submodules de Chaisemartin-D'Haultfœuille: negative-weight diagnostics for TWFE under heterogeneous effects. ImputationEventStudy 34 core · 25 supporting · 5 submodules Borusyak-Jaravel-Spiess: the imputation estimator and the efficiency/characterization results for staggered adoption. OLSWeightDecomposition 27 core · 32 supporting · 4 submodules Sloczynski's characterization: the OLS coefficient as a convex combination of group-specific effects with reversed weights. StaggeredTWFEDecomposition 73 core · 66 supporting · 7 submodules This folder formalizes the Goodman-Bacon decomposition of a two-way fixed-effects (TWFE) coefficient estimated on a staggered-adoption panel.