The DOI matches its Crossref record; 39 of the 42 references it cites were found in a registry or an index, and 3 are not checked yet; 10 of the 30 sources its claims cite were checked for support.
Two-Way Fixed Effects, the Two-Way Mundlak Regression, and Difference-in-Differences Estimators
Open accessAbstract
I establish the equivalence between the two-way fixed effects (TWFE) estimator and an estimator obtained from a pooled ordinary least squares regression that includes unit-specific time averages and time-period specific cross-sectional averages, which I call the two-way Mundlak (TWM) regression. This equivalence furthers our understanding of the anatomy of TWFE, and has several applications. The equivalence between TWFE and TWM implies that various estimators used for intervention analysis – with a common entry time into treatment or staggered entry, with or without covariates – can be computed using TWFE or pooled OLS regressions that control for time-constant treatment intensities, covariates, and interactions between them. The approach allows considerable heterogeneity in treatment effects across treatment intensity, calendar time, and covariates. The equivalence implies that standard strategies for heterogeneous trends are available to relax the common trends assumption. Further, the two-way Mundlak regression is easily adapted to nonlinear models such as exponential models and logit and probit models.
Verified 2026-10-10 · 3 references not checked yet
Overview
CoChat Summary
TL;DR
The paper shows that the two-way fixed effects (TWFE) estimator is numerically equivalent to a pooled OLS regression that includes unit-specific time averages and time-period-specific cross-sectional averages—the two-way Mundlak (TWM) regression. It uses this equivalence to show that flexible pooled OLS/extended TWFE, cohort imputation, random effects, and event-study estimators produce identical average treatment effects on the treated under standard staggered difference-in-differences assumptions.
Key findings
- TWFE and TWM are algebraically equivalent: adding both unit-specific time averages and period-specific cross-sectional averages to pooled OLS reproduces the TWFE coefficients on time-varying regressors.
- Under no anticipation, conditional parallel trends, linearity, and random sampling, cohort imputation, pooled OLS on cohort dummies, random effects, TWFE, and Borusyak et al. imputation yield the same ATTs; the flexible leads-and-lags estimator is equivalent to Sun and Abraham (2021) and the regression-based Callaway and Sant’Anna (2021) estimator.
- In the Walmart application using 1,288 U.S. counties over 1977–1999, the lags-only and leads-and-lags weighted average effects are 9.4% and 7.3%, respectively; allowing cohort-specific linear trends reduces the average effect to about 2.6% (t ≈ 2.71), with effects in the first six years between 2% and 3% and a peak just under 5% nine years out.
- The paper provides aggregation formulas for overall and exposure-time ATTs and shows how to allow heterogeneous cohort trends within the same regression framework.
Methods
- Theoretical econometric paper deriving algebraic equivalences between TWFE, TWM, pooled OLS, random effects, imputation, and event-study estimators, with proofs in an appendix.
- Identification is developed for staggered interventions under SUTVA, no bad controls, no anticipation, conditional parallel trends, linearity, and random sampling across units.
- Empirical application uses Brown and Butts (2025) county data: 1,288 counties, 1977–1999, log retail employment, three time-constant 1980 controls, 14 treated cohorts, county-clustered standard errors, and Stata margins for exposure-time effects; Fig. 2 uses 1,000 bootstrap replications.
Limitations
- The paper states that the equivalences assume a balanced panel and time-constant covariates; with time-varying covariates, POLS, TWFE, and RE are no longer identical unless the Mundlak equation is expanded, and with unbalanced panels the equivalences no longer hold.
- The paper states that inference can be problematic with few treated units per treated cohort, and that the Walmart application’s parallel trends assumption is clearly violated, so the lags-only and leads-and-lags estimates cannot be relied on as good causal estimates; cohort-specific linear trends reduce but do not perfectly eliminate pre-trends.
- The paper notes that linearity of conditional means in covariates is assumed, and that the leads-and-lags estimator can exacerbate bias if parallel trends violations carry into treated periods.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
6 Valid · 7 Not valid · 17 Not checked · 184 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“Moreover, this estimator turns out to be identical to the 2 × 2 regression-based estimators in Callaway and Sant'Anna (2021)[CS (2021)].”
What the sources say
- Supportedcites Callaway and Sant'Anna (2021) source
The Introduction describes outcome-regression DiD estimators, and Appendix B says its outcome-regression estimands generalize the two-group, two-period setup. The citation identifies the estimators being compared; the assertion that the citing authors’ estimator is identical to them is their own comparison, not a result attributed to Callaway and Sant’Anna. Page numbers are unavailable in our copy. (LLM, conf=0.83)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Without covariates x i , this cohort imputation method is the same as that proposed by Gardner (2022)-which he calls 'two-stage difference-in-differences.'”
What the sources say
- Supportedcites Gardner (2022) source
The Introduction describes Gardner’s proposed two-stage approach: estimate group and period effects using untreated observations, subtract them from outcomes, then regress the adjusted outcomes on treatment status. Section 3 and the estimation steps specify this procedure without covariates, and Section 3.3 calls it the “two-stage difference-in-differences estimator.” (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Borusyak et al. (2024)[BJS (2024)] propose an imputation estimator based on TWFE estimation in the imputation step.”
What the sources say
- Supportedcites Borusyak et al. (2024) source
The Introduction describes an imputation estimator that fits unit and period fixed effects using untreated observations, then uses them to impute untreated outcomes for treated observations. Section 4.1 explicitly says the imputation procedure requires estimating a simple TWFE model. (LLM, conf=0.94)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“As shown by these earlier authors, the resulting estimate, ˆ τ T W FE , depends on 2 × 2 difference-in-differences estimates, several of which are based on 'forbidden comparisons.'”
What the sources say
- Supportedcites De Chaisemartin and d'Haultfoeuille (2020) source
Section 3.1, “A decomposition result” (page not available in our copy), writes the TWFE coefficient in its example as the average of two DID estimates. It explains that the second DID uses a group treated in both periods as its control—the kind of comparison the claim calls “forbidden.” (LLM, conf=0.88)
- Partly supportedcites Callaway and Sant'Anna (2021) sourceExperimental — accuracy not yet measured
Section 3 (page unavailable) says the TWFE coefficient is a weighted average of underlying treatment-effect parameters and that some weights can be negative. Section 2.4 identifies never-treated and not-yet-treated units as comparison groups for the paper’s approach. Our copy does not state that the TWFE estimate depends on 2×2 DiD estimates or that several use ‘forbidden comparisons.’ (LLM, conf=0.83)
- Supportedcites Goodman-Bacon (2021) source
Theorem 1 states that the two-way fixed-effects estimate is a weighted average of 2×2 difference-in-differences estimates. Section I explains that some of those comparisons use already-treated units as controls for later-treated units—the comparisons the citing sentence calls “forbidden.” (LLM, conf=0.92)
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“the methods the imputation and pooled OLS methods can be extended to the case of repeated cross sections, provided we impose essentially the same assumptions in Sect. 4 along with a stable population over time.”
What the sources say
- Supportedcites Deb et al. (2024) source
Section 2.2 develops imputation and OLS estimation for repeated cross-sections, and Section 2.1 assumes a stable population sampled over time. The reference to assumptions in Sect. 4 appears to refer to the citing document’s own assumptions, not a claim about this source. Page numbers are unavailable in our copy. (LLM, conf=0.91)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
- Supported
109 more claims
Create a free account to see all 214 claims and every source verdict.
- Probably not a substantial claim. 52
- It may state two separate claims as one. 61
- The quote only partly states this claim. 3
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Note that Wooldridge (2021) has also proposed an estimation strategy connected, and in some cases numerically equivalent, to that of Borusyak et al. (2021).”
From Two-Way Fixed Effects and Differences-in-Differences with Heterogeneous Treatment Effects: A Survey (2022) · cites this paper as Wooldridge (2021)
The cited source says what the claim attributes to it.
In the citing paper
“Indeed, BJS prove that when Assumption 4 holds, their estimator is efficient under homoskedasticity and serially uncorrelated errors; see also Wooldridge (2021).”
From What's trending in difference-in-differences? A synthesis of the recent econometrics literature (2023) · cites this paper as Wooldridge (2021)
The cited source says what the claim attributes to it.
In the citing paper
“Wooldridge (2021) proposes an alternative regression adjustment procedure that is suitable for staggered setups.”
From What's trending in difference-in-differences? A synthesis of the recent econometrics literature (2023) · cites this paper as Wooldridge (2021)
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“On the other hand, if these additional assumptions are violated, Wooldridge (2021)'s estimator may be more biased than Callaway and Sant'Anna (2021)'s.”
From What's trending in difference-in-differences? A synthesis of the recent econometrics literature (2023) · cites this paper as Wooldridge (2021)
Experimental — accuracy not yet measuredIn the citing paper
“His proposed estimator differs from the Callaway and Sant'Anna (2021) regression adjustment estimator as he exploits additional information from pre-treatment periods, which, in turn, can lead to improvements in precision.”
From What's trending in difference-in-differences? A synthesis of the recent econometrics literature (2023) · cites this paper as Wooldridge (2021)
Citations
Cites
(2001) · RePEc: Research Papers in Economics · Cited by 27,577
Self-citationSpecification Tests in Econometrics
Hausman, J. A. (1978) · Econometrica · doi:10.2307/1913827 · Cited by 18,820 · cited 3×
“For one, it leads to a robust, variable addition version of the Hausman (1978) test”
InfluentialEconometric Analysis of Panel Data
Baltagi, B. H. (2021) · Springer Texts in Business and Economics · doi:10.1007/978-3-030-53953-5 · Cited by 15,104 · cited 2×
“In the large-T panel literature, where one is interested in obtaining valid inference on β as T → ∞ (usually along with N → ∞ ), a “double-demeaning” characterization is used for ˆ β FE ; see, for example, Baltagi (2021).”
InfluentialDifference-in-Differences with multiple time periods
Callaway, B., & Sant’Anna, P. H. C. (2021) · Journal of Econometrics · doi:10.1016/j.jeconom.2020.12.001 · Cited by 9,126
Difference-in-differences with variation in treatment timing
Goodman-Bacon, A. (2021) · Journal of Econometrics · doi:10.1016/j.jeconom.2021.03.014 · Cited by 8,046 · cited 1×
“The work by De Chaisemartin and d’Haultfoeuille (2020), Callaway and Sant’Anna (2021), Goodman-Bacon (2021), and others use the phrase “two-way fixed effects” to refer to a particular, restrictive model, namely, (5.8) For simplicity, I have dropped the control variables.”
InfluentialConsistent Covariance Matrix Estimation with Spatially Dependent Panel Data
Driscoll, J. C., & Kraay, A. C. (1998) · Review of Economics and Statistics · doi:10.1162/003465398557825 · Cited by 6,865
Estimating dynamic treatment effects in event studies with heterogeneous treatment effects
Sun, L., & Abraham, S. (2021) · Journal of Econometrics · doi:10.1016/j.jeconom.2020.09.006 · Cited by 5,518
Matching as an econometric evaluation estimator: Evidence from evaluating a job training programme
Heckman, J. J., Ichimura, H., & Todd, P. E. (1997) · The Review of Economic Studies · doi:10.2307/2971733 · Cited by 5,486 · cited 2×
“Another important algebraic equivalence involving the FE estimator, usually invoked in microeconometric settings, is the equivalence between the FE estimator that removes unit-specific effects – the one-way FE estimator (OWFE) – and the Mundlak (1978) device, which includes unit-specific time averages of time-varying…”
On the Pooling of Time Series and Cross Section Data
Mundlak, Y. (1978) · Econometrica · doi:10.2307/1913646 · Cited by 5,412 · cited 2×
“3 The two-way Mundlak regression Mundlak (1978) showed that, with x it including explanatory variables that vary across ( i , t ) or t only (including time-period dummies), the one-way FE estimator can be obtained as a particular GLS estimator by adding the time averages, ¯ x i · , as additional…”
Two-Way Fixed Effects Estimators with Heterogeneous Treatment Effects
de Chaisemartin, C., & D’Haultfœuille, X. (2020) · American Economic Review · doi:10.1257/aer.20181169 · Cited by 5,031 · cited 1×
“Further, using N, T → asymptotics with independent sampling in the cross-sectional dimension and weak dependence in the time-series dimension, Hansen (2007) shows that the TWFE estimator is consistent and asymptotically normal without strict exogeneity.”
Cited by
Revisiting Event-Study Designs: Robust and Efficient Estimation
Borusyak, K., Jaravel, X., & Spiess, J. (2024) · Review of Economic Studies · doi:10.1093/restud/rdae007 · Cited by 2,280
What's trending in difference-in-differences? A synthesis of the recent econometrics literature
Roth, J., Sant’Anna, P. H. C., Bilinski, A., & Poe, J. (2023) · Journal of Econometrics · doi:10.1016/j.jeconom.2023.03.008 · Cited by 2,027
Revisiting event study designs: robust and efficient estimation
(2022) · arXiv (preprint) · doi:10.47004/wp.cem.2022.1122 · Cited by 607
Two-way fixed effects and differences-in-differences with heterogeneous treatment effects: a survey
de Chaisemartin, C., & D’Haultfœuille, X. (2023) · The Econometrics Journal · doi:10.1093/ectj/utac017 · Cited by 522
Climate policies that achieved major emission reductions: Global evidence from two decades
Stechemesser, A., Koch, N., Mark, E., Dilger, E., Klösel, P., Menicacci, L., Nachtigall, D., Pretis, F., Ritter, N., Schwarz, M., et al. (2024) · Science · doi:10.1126/science.adl6547 · Cited by 322
Metadata
- Authors
- Wooldridge, J. M.
- Affiliations
- Michigan State University (US)
- Corresponding author
- Wooldridge, Jeffrey M.
- Venue
- SSRN / The Econometrics Journal
- Publisher
- Elsevier BV
- Published
- 2021
- Volume
- 69
- Issue
- 5
- Pages
- 2545-2587
- ISSN
- 0377-7332 (ISSN-L)
- OpenAlex
- W4413754779
- Cited by
- 714
- References
- 39
- Citations by year
- 2026: 218 · 2025: 199 · 2024: 155 · 2023: 78 · 2022: 45 · 2021: 11
- FWCI
- 225.81
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Empirical Economics (publishedVersion, cc-by)
- SSRN Electronic Journal (submittedVersion)
- Funders
- Agricultural and Applied Economics Association
- University College London
- College of Engineering, Michigan State University
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Spatial and Panel Data Analysis
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Keywords
- two-way fixed effects, difference-in-differences, estimator equivalence, treatment effect heterogeneity, average treatment effect on the treated, parallel trends assumption, intervention analysis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.