CSDID: Stata module for the estimation of Difference-in-Difference models with multiple time periods
Fernando Ríos‐Avila, Pedro H. C. Sant’Anna, Brantly Callaway
Abstract
CSDID implements Callaway and Sant'Anna (2020) estimator for DID models with multiple time periods. The main idea of CSDID is that consistent estimations for ATT's can be obtained by ignoring 2x2 DID design that compare late treated units with earlier treated units. In the presence of heterogeneous and dynamic effects, this type of comparison can severely bias the estimation of Treatment effects. CSDID at its core uses DRDID for the estimation of all 2x2 DID designs to estimate all relevant ATTGT's (Average treatment effects of the treated for group G at time T). Post estimation commands can be used to obtain important aggregations. CSDID can be used with panel data and repeated cross-section.