PATHINTDID: Stata module performing Path-Integrated Difference-in-Differences (PI-DiD): Identification, Estimation, and Inference

In empirical policy evaluation, interventions such as worker training programs, cash transfers, localized health programs, and digital financial services often produce dynamic, transitory effects where individual outcome paths temporarily diverge from counterfactuals before converging again due to skill depreciation or control-group catch-up. Standard Difference-in-Differences models suffer from endpoint-subtraction bias because they evaluate policy efficacy using static endline snapshot contrasts, which can erroneously indicate zero impact when trajectories rejoin despite substantial historical utility generated while the program was active. The pathintdid package overcomes this limitation by implementing a path-integrated framework that measures the total net stock of accumulated welfare generated over the post-treatment window by integrating the area trapped between realized and counterfactual outcome trajectories. This cumulative impact is also scaled into an annualized per-unit treatment flow, providing an accurate metric of policy success under dynamic rejoining conditions while accounting for arbitrary within-unit temporal correlation through cluster-robust standard errors.