Sensitivity Analysis for Unmeasured Confounding in Difference-in-Differences
Editor assessment Waiting for Editors 5d left- Editor
- Unassigned
- Submitted
- Nov 3, 2026
- Due
- Nov 10, 2026
- Reviews
- 0 of 0
- Status
- Waiting for Editors
Assign an editor to this article
Choose a handling editor to move this submission into the active editorial queue.
Abstract
We derive sharp bounds for the average treatment effect in difference-in-differences designs when the parallel trends assumption is violated by unmeasured confounders. Our approach extends Manski-type partial identification to panel settings and provides closed-form sensitivity parameters with natural interpretations.