Research

Book

With Clément de Chaisemartin · Princeton University Press, forthcoming

Previous version (SSRN)

Show abstractHide abstract

This textbook offers an up-to-date survey on difference-in-differences. It covers classical designs but also staggered adoption designs, so-called heterogeneous adoption designs and still more general designs. Inference, robustness with respect to heterogeneous treatment effects and relaxations of parallel trends are also discussed extensively.

Working papers 9

With Laurent Davezies, Louise Laage · under revision at Review of Economic Studies

Show abstractHide abstract

We show that the sharp bounds on the average marginal and treatment effects (AME and ATE) can be obtained very simply in fixed effect logit models, without any optimization. We consider two related estimation methods. The second is very simple but yields asymptotically larger confidence intervals, though it performs very well in finite samples.

With Clément de Chaisemartin · conditionally accepted at American Economic Review

Show abstractHide abstract

The common practice of conducting inference on a parameter of interest conditional on not rejecting a specification test is valid (though generally conservative) if the tested conditions hold, under general conditions. In particular, this requires no assumption on the asymptotic dependence between the estimator of the parameter of interest and the specification test. Conditional inference may also have better properties than unconditional inference under local alternatives.

With Philippe Février, Ao Wang, Lionel Wilner · revised and resubmitted to Quantitative Economics

Show abstractHide abstract

We quantify gains and losses from revenue management relative to uniform and flexible pricing using French railway data. The analysis combines temporal variation in relative prices with restrictions on consumer behavior and firms’ pricing decisions to identify demand and evaluate alternative strategies.

With Laurent Davezies, Yannick Guyonvarch · under revision at American Economic Review

Show abstractHide abstract

The usual analytic variance estimator for two-way clustering may not be positive and is invalid for some DGPs. We propose a simple fix that solves both issues. When the estimator of the parameter of interest is asymptotically Gaussian, the corresponding tests are asymptotically exact and equivalent to usual ones. Otherwise, the new tests are asymptotically conservative.

With Clément de Chaisemartin, Félix Pasquier, Gonzalo Vazquez-Bare, Doulo Sow · under revision at Quantitative Economics

Show abstractHide abstract

We propose new DID estimators for continuous treatments. We also assume that from two periods, the treatment of some units, the movers, changes, while the treatment of other units, the stayers, does not change. Then, our estimators compare the outcome evolution of movers and stayers with the same initial value of the treatment. Our estimator is robust to heterogeneous treatment effects. We also extend this framework to an IV setup.

With Clément de Chaisemartin, Diego Cicca, Felix Knau

Show abstractHide abstract

We consider treatment-effect estimation in designs in which no unit is treated initially, and then units simultaneously receive heterogeneous and strictly positive treatment doses. Under a parallel-trends assumption and quasi-untreated units, a causal effect is identified by an estimand mixing DID and RDD. We also consider designs without quasi-untreated units and propose a test of the homogeneous-effect assumption underlying TWFE regressions.

With Christophe Gaillac, Arnaud Maurel

Show abstractHide abstract

We consider linear regressions when the outcome is not observed in the same dataset as some of the covariates. Without further restriction, we show that the coefficients are partially identified and the bounds take a very simple form. We also show their asymptotic normality.

With Isis Durrmeyer, Jean-François Fournel, Alessandro Iaria

Show abstractHide abstract

We investigate the consequences of introducing an online distribution channel in the French car industry, where currently most of the sales take place in person through car dealers relying on third-degree price discrimination. To this end, we develop a structural model of demand with unobserved third-degree price discrimination and transportation costs related to visiting car dealers.

With Clément de Chaisemartin

Show abstractHide abstract

We consider the estimation of a scalar parameter, when two estimators are available. The first is always consistent. The second is inconsistent in general, but has a smaller asymptotic variance than the first, and may be consistent if an assumption is satisfied. We propose to use the weighted sum of the two estimators with the lowest estimated mean-squared error (MSE). This third estimator dominates the other two from a minimax-regret perspective.

Journal articles and chapters 40

For published papers, titles link to the journal or publisher. “Working paper version” links to arXiv where available, otherwise to an earlier manuscript.

With Clément de Chaisemartin · forthcoming in Econometric Society Monograph (13th World Congress)

Show abstractHide abstract

The paper studies identification of dynamic treatment effects under parallel trends, with possibly non-binary and non-absorbing treatment. The first part clarifies which event-study effects are identified without further restrictions. Those effects can be helpful for ex post analysis, but cannot be used for ex ante analysis. Then, we consider another restriction, namely a random coefficients distributed-lag linear model. We illustrate our findings by revisiting Gentzkow et al. (2011).

With Clément de Chaisemartin · Review of Economics and Statistics, 108 (2026)

Working paper version

Show abstractHide abstract

We consider treatment effect estimators with a panel of groups under parallel trends. We propose event-study estimators of the effect of being exposed to a weakly higher treatment dose for ℓ periods. We also propose normalized estimators, that estimate a weighted average of the effects of the current treatment and its lags. Contrary to usual estimators, they are valid even if the treatment effect is heterogeneous and beyond staggered adoption designs.

With Christophe Gaillac, Arnaud Maurel · Review of Economic Studies, 92 (2025)

Working paper version

Show abstractHide abstract

We consider partially linear model, when the outcome of interest and some of the covariates are observed in two different datasets that cannot be linked. We obtain a simple characterization of the identified set and develop tractable inference. We apply our results to intergenerational income mobility in the US, without relying on exclusion restrictions behind the two-sample TSLS, unlike previous approaches.

With Clément de Chaisemartin, Gonzalo Vazquez-Bare · AEA Papers & Proceedings, 114 (2024)

Working paper version

Show abstractHide abstract

Applied researchers usually rely on two-way fixed effect regressions to estimate the effect of a continuous treatment. However, such estimators are not robust to heterogeneous treatment effects and rely on the linearity of treatment effects. We propose estimators that do not impose those restrictions, and that can be used even if the treatment of all units changes from one period to the next.

With Pujee Tuvaandorj · Quantitative Economics, 15 (2024)

Working paper version

Show abstractHide abstract

We develop a new permutation test for inference on a subvector of coefficients in linear models. The test is exact when the regressors W and the error terms u are independent, but also consistent under, mainly, two conditions: (i) a slight reinforcement of cov(W, u)=0; (ii) the number of strata, defined by values of the regressors not involved in the subvector test, is small compared to the sample size.

With Stefan Hoderlein, Yuya Sasaki · Journal of Econometrics, 240 (2024)

Show abstractHide abstract

In a nonparametric triangular structure typical of the control function literature, and provided the instrument satisfies a ‘local irrelevance’ condition, we show that we can test the exclusion restriction. Second, if the ‘instrument’ also directly affects the outcome variable, we show that the identification of average causal effects can be achieved in linear random coefficients models and single index models.

With Clément de Chaisemartin · Econometrics Journal, 26 (2023)

Working paper version

Show abstractHide abstract

Linear regressions with period and group fixed effects are widely used to estimate policies’ effects. Yet, those regressions may produce misleading estimates, if the policy’s effect is heterogeneous between groups or over time. This survey reviews a fast-growing literature documenting this issue and proposing alternative estimators robust to heterogeneous effects. We use those alternative estimators to revisit Wolfers (2006).

With Laurent Davezies, Martin Mugnier · Quantitative Economics, 14 (2023)

Working paper version

Show abstractHide abstract

We show that the impossibility result of Chamberlain for fixed effects binary choice models (with iid errors, the slope parameter is point identified only with logistic errors) only holds with T=2. With three or more periods, we exhibit a family of distributions for which point identification can be achieved. Identification leads to a GMM estimator, which reaches the semiparametric efficiency bound when T=3. We apply our method to revisit Brender and Drazen (2008).

With Marianne Bléhaut, Jérémy L’Hour, Alexandre Tsybakov · in Belomestny, D. et al. (eds) Foundations of Modern Statistics. Springer (2023)

Working paper version

Show abstractHide abstract

The synthetic control program does not admit a unique solution when n, the number of control units, is larger than p, the dimension of the variables used to construct the synthetic control. We propose a method that solves this issue and enjoys a double robustness property. We then extend it to the case where p is potentially larger than n, and show root-n consistency and asymptotic normality of the corresponding ATT estimator.

With Stefan Hoderlein, Yuya Sasaki · Journal of Econometrics, 234 (2023)

Working paper version

Show abstractHide abstract

We extend the change-in-change model of Athey and Imbens (2006) to continuous treatments. We require an exogenous change over time of the distribution of the treatment inducing a crossing in their cdfs. This crossing plays the role of a ‘control group’, and allows us to average and quantile treatment effects in a nonparametric way.

With Laurent Davezies, Yannick Guyonvarch · Statistics and Probability Letters, 188 (2022)

Working paper version

Show abstractHide abstract

We show a Marcinkiewicz-Zygmund law of large numbers (both in L^r and a.s.) for jointly and dissociated exchangeable arrays. As a result, we obtain a law of iterated logarithm for such arrays under a weaker moment condition than the existing one.

With Christophe Gaillac, Arnaud Maurel · Quantitative Economics, 12 (2021)

Working paper version

Show abstractHide abstract

We construct the best possible test of rational expectations (RE) when we observe expectations on future variables on a certain sample, and realizations of that variable on another sample of different units. We apply our methodology to test for RE about future earnings.

With Laurent Davezies, Yannick Guyonvarch · Annals of Statistics, 49 (2021)

Working paper version

Show abstractHide abstract

We show the weak convergence of empirical processes and some bootstrap processes with multiadic (e.g., dyadic) data or multiway clustering. These results imply asymptotic normality and the validity of the bootstrap for a large class of nonlinear estimators. We illustrate our results with trade data.

With Lucas Girard, Roland Rathelot · Stata Journal, 21 (2021)

Show abstractHide abstract

The Stata command Segregsmall estimates usual segregation indices (e.g. Duncan or Theil) when units (=geographical areas, firms, classrooms…) are small. In such cases, naive estimators are biased upwards. The command computes in particular the estimators described in R. Rathelot’s JBES paper and in our joint QE paper.

With Clément de Chaisemartin · American Economic Review, 110 (2020)

Working paper version

Show abstractHide abstract

We show that if treatment effects are not constant, regressions with groups and time fixed effects identify weighted averages of treatment effects across groups and time periods, with potentially (many) negative weights. We suggest sensitivity checks and better estimands under testable restrictions on the design.

With Arnaud Maurel, Xiaoyun Qiu, Yichong Zhang · Stata Journal, 20 (2020)

Working paper version

Show abstractHide abstract

This paper presents the Stata command eqregsel, which computes the extremal quantile regression estimator for sample selection developed in our paper ‘Extremal Quantile Regressions for Selection Models and the Black-White Wage Gap’.

With Philippe Février · Quantitative Economics, 11 (2020)

Show abstractHide abstract

To what extent do people react to incentives? Are observed contracts (nearly) optimal? We answer these questions using a nonparametric principal agent model and an exogenous variation in contracts between the French national institute of statistics and its interviewers.

With Isis Durrmeyer, Philippe Février · Review of Economic Studies, 86 (2019)

Working paper version

Show abstractHide abstract

We consider inference on a demand and supply model for differentiated products with price discrimination that is unobserved by the econometrician. We show how to extend BLP’s GMM estimation to this setting, using restrictions on marginal costs. We apply our framework to the French automobile market.

With Clément de Chaisemartin, Yannick Guyonvarch · Stata Journal, 19 (2019). The Stata command fuzzydid is available on the SSC repository

Show abstractHide abstract

This paper presents the Stata command fuzzydid, which computes various estimators of the LATE and LQTE for fuzzy DID designs, following our paper ‘Fuzzy DID’. It can handle non-binary treatments, multiple periods and groups, covariates and partial identification.

With David Haziza, Eric Lesage · Journal of the American Statistical Association, 114 (2019)

Show abstractHide abstract

We show that the calibration method based on instruments, proposed by Deville (2002), leads to a large variance when the instrumental variables are poorly related to the calibrating variables. If the exclusion restriction is violated, the bias is also large under the same condition.

With Clément de Chaisemartin · Review of Economic Studies, 85 (2018)

Working paper version

Show abstractHide abstract

In many applications of the DID method, the treatment rate only increases more in the treatment group. In such fuzzy designs, we show that the popular ‘Wald-DID’ (the DID of the outcome divided by the DID of the treatment) identifies a LATE only if two homogeneous treatment effect assumptions hold. We then propose two alternative estimands that do not rely on such assumptions.

With Arnaud Maurel, Yichong Zhang · Journal of Econometrics, 203 (2018)

Working paper version

Show abstractHide abstract

We consider models with endogenous selection and no instrument nor large support regressors. Identification relies on the independence between the covariates and selection, when the outcome tends to infinity. We propose a simple estimator based on extremal quantile regressions and apply it to the evolution of the black-white wage gap in the US.

With Roland Rathelot · Quantitative Economics, 8 (2017)

Working paper version

Show abstractHide abstract

Suppose that an individual in a small unit j (a classroom, a small firm…) belongs to a minority with a probability pj. To measure segregation of this minority, one would ideally use an inequality index on the pj, but they are unobserved. Using the observed proportion instead leads to an overestimation. The segregation indices are actually partially identified. We provide tractable bounds and develop inference.

With Isis Durrmeyer, Philippe Février · International Journal of Industrial Organization, 47 (2016)

Show abstractHide abstract

We study the factors of the decrease in average CO2 emissions of new cars in France between 2003 and 2008. We show that the evolution of consumers’ preferences account for 43% of this decrease, and that these changes follow two environmental policies put in place during this period.

With Philippe Février · Journal of Econometrics, 189 (2015)

Working paper version

Show abstractHide abstract

Suppose that observed variables (X1,…,XK) are independent conditional on a continuous and unobserved variable X*. We show that the distributions of Xi conditional on X* are identified if the bounds of the conditional support of Xi are strictly increasing with X*. We also develop a test of this condition.

With Pauline Givord · Economie et Statistique, 471 (2014)

Show abstractHide abstract

This article (in French) is an introduction to quantile regression, with an emphasis on its interpretation.

With Arnaud Maurel · Journal of Econometrics, 174 (2013)

Working paper version

Show abstractHide abstract

Consider an extended Roy model where a binary decision depends on expected gains and an unobserved cost. The model is identified without instruments if, basically, the unobserved cost only depends on covariates. Applying our results to French data, we show that nonpecuniary components are a key factor for going to college.

Journal of Econometrics, 154 (2010)

Working paper version

Show abstractHide abstract

Consider the sample selection model under the nonstandard IV restriction that D is independent of Z conditional on Y. Nonparametric identification is achieved under a completeness condition between Y and Z. Partial identification can also be obtained if one replaces independence by monotonicity restrictions.

With Laurent Davezies, Denis Fougère · Econometrics Journal, 12 (2009)

Working paper version

Show abstractHide abstract

A linear-in-means model close to the one of Manski (1993) is identified provided that we observe groups with three distinct sizes. This applies even if one does not observe all members of the group, and can also be extended to binary outcomes.