Xavier D'Haultfoeuille - Masterclass sur les limites de la méthode des différences de différences

Xavier D'Haultfoeuille - Masterclass sur les limites de la méthode des différences de différences

Title history
  • Xavier D'Haultfoeuille - Masterclass sur les limites de la méthode des différences de différences — observed on 20261110-0712 UTC (YouTube)

by Xavier D'Haultfoeuille · View the channel on PSI

Formal & Physical Sciences Economics & Finance KCEconomicsKCHEconometrics and economic statistics
🎙 Xavier D'Haultfoeuille 👥 1K 📅 November 6, 2026 ⏱ 51 min 👁 738 📄 literature review 🧭 2026-11-10
Available in: English (current) Français

Keywords

difference-in-differencestwo-way fixed effectsevent studyheterogeneous treatment effectsparallel trends

Summary

Xavier D’Haultfoeuille delivers a masterclass on the limitations of the difference-in-differences (DID) method, a widely used technique for evaluating public policies. He begins by tracing the method’s historical roots, citing Semmelweis’s hand-washing intervention and the classic Card and Krueger study on minimum wages. He then formalizes the DID framework, introducing key assumptions such as no anticipation and parallel trends. The talk highlights that traditional two-way fixed effects regressions, which work well in simple settings with two groups and two periods, can be biased when treatment timing is staggered or when effects are heterogeneous. He emphasizes recent methodological breakthroughs, including his own work with Clément de Chaisemartin, that propose new estimators and robust event-study designs. He also discusses problems of statistical power in standard pre-trend tests, referencing Jonathan Roth’s findings that many published results may be due to undetectable biases. Throughout, D’Haultfoeuille stresses the importance of correctly interpreting event-study graphs and selecting appropriate estimation strategies for credible causal conclusions.

158 words

Critical Evaluation

Value of the Information & Strength of the Argument

The information provided is highly valuable for practitioners and researchers applying DID methods. D’Haultfoeuille not only explains the limitations but also offers concrete solutions, which enhances the practical utility of the content. His argumentation is rigorous, grounded in published research and theoretical reasoning. He uses historical and contemporary examples to illustrate points, making the material accessible while maintaining academic depth. He is transparent about the assumptions required and warns against common pitfalls, such as misinterpreting pre-trend plots or relying solely on underpowered tests. The case presented is persuasive and well-supported by references to his own publications and other key works in the econometrics literature.

Scientific Rigor, Source Quality, Title Accuracy

The presentation demonstrates strong scientific rigor. D’Haultfoeuille explicitly cites relevant academic papers, including his own work in the American Economic Review and contributions by other scholars like Card, Krueger, and Roth. The methodological explanations are precise and align with current econometric theory. The title accurately describes the content, reflecting both the masterclass format and its focus on limitations. The sources mentioned are credible and appropriate for the subject matter. No misleading claims or overstatements were observed. The talk is well-structured, and the technical depth is suitable for an audience with background in statistics or econometrics.

214 words

Title / Content Match

The title accurately reflects the content as a masterclass focusing on the limits and recent solutions for difference-in-differences methods.

Quality & Reliability

9/10

The presentation is by a leading econometrician, draws on published research in top journals, and clearly explains methodological assumptions and limitations.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

  • ENSAI official website — Official website of the host institution, mentioned in the video description.
  • ENSAI LinkedIn profile — LinkedIn profile from the video description for following ENSAI.
  • De Chaisemartin and D'Haultfoeuille (2020) - Two-Way Fixed Effects Estimators with Heterogeneous Treatment Effects — Key paper cited in the masterclass on limitations of DID with heterogeneous effects.
  • Card and Krueger (1994) - Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania — Classic DID application discussed as an example in the talk.

Contribution & Novelties

The video provides an original synthesis of recent econometric developments regarding difference-in-differences, with a focus on heterogeneous treatment effects and event-study designs. D’Haultfoeuille, as a co-author of influential papers on the topic, offers a first-hand perspective that is both deep and accessible. The talk emphasizes practical pitfalls and proposes robust estimation strategies, making it a valuable resource for practitioners.

Pour aller plus loin :

  • Difference in differences (Wikipedia) — Overview of the method and its assumptions.
  • Event study (Wikipedia) — Introduction to event-study designs commonly used in DID analysis.
  • Heterogeneous treatment effects — Concept central to recent DID literature; see papers by de Chaisemartin & D’Haultfoeuille and Goodman-Bacon.

108 words

Radar Profile

The profile is well-balanced with high scores across all dimensions, reflecting a technically rigorous and thoroughly informative presentation. The lowest score is information quantity, but it remains high, indicating that the content is dense yet focused.

Reliability 9/10