Dynamic Difference-in-Differences with Interference

Key Points

• This research aims to develop a framework for evaluating policy effects while accounting for interference among treated units.
• Developed a difference-in-differences framework using panel data with multiple posttreatment periods.
• Defined causal parameters under interference and established identification assumptions.
• Proposed doubly robust estimators for valid inference.
• Illustrated the framework through an empirical application showing its practical relevance and robustness.
• Provided evidence that accounting for interference enhances causal parameter estimation.

Abstract