csdid

csdid estimates difference-in-differences effects when units begin treatment at different times. It supports balanced and unbalanced panels, repeated cross sections, covariates, and sampling weights.

Following Callaway and Sant’Anna (2021), it estimates an average treatment effect for each treatment cohort and period, then combines those effects into event studies, cohort averages, calendar-time averages, or one overall summary. Effects may differ across cohorts and over time. Inference uses a multiplier bootstrap with simultaneous confidence bands by default; analytical and clustered inference are also available.

How to install

Version 2.0.0 installs from GitHub. It replaces an earlier csdid under the same command name, so uninstall first and then install:

cap ado uninstall csdid
net install csdid, from("https://raw.githubusercontent.com/pedrohcgs/csdid-stata/main") replace

If your current copy came from SSC, ssc uninstall csdid is the first line instead. Uninstalling first is what keeps Stata tracking one csdid rather than two. Run csdid version afterwards to confirm what you are running.

SSC currently distributes csdid Version 1.82, the previous generation of the package. If that is what you want, it is one line:

ssc install csdid, replace

What it is built for

The package is for researchers who want to estimate and report staggered adoption designs in Stata. Choose doubly robust estimation (the default), outcome regression, or inverse probability weighting. Each aggregation has explicit weights, so its interpretation follows the question being asked.

Sample choices matter. The default bal(full) retains units observed in every period; bal(none) uses the available observations. If missingness changes which units remain, the population represented by an estimate may change. The unbalanced-panel guide explains these choices.

Other DiD commands can target different causal parameters. Our comparison guide shows when estimates coincide, when they differ, and how the comparison depends on the design.

Speed

Version 2.0.0 takes advantage of the shared structure across cohort-period comparisons. In the published comparisons with Version 1.82, the speedup ranges from 10x to 308x as the design’s size, periods, and cohorts vary. See Speed against Version 1.82 for the timings, machine, matched settings, and runnable code. The comparison with other commands reports both runtime and differences in what each command estimates.

Also in R and Python

The same estimators, from the same team, are available as did for R and csdid for Python (pip install csdid).

When comparing implementations, set the same sample, comparison group, base period, estimator, and inference options explicitly. This Stata package defaults to not-yet-treated controls and a universal base period; the R package defaults to never-treated controls and a varying base period.

If you find any discrepancy, please raise an issue, and we will address it. But also make sure you are using the same options!

How to use

csdid y x1 x2, ivar(id) time(year) gvar(gvar)   // every ATT(g,t), doubly robust
estat event                                     // the event study, uniform bands
estat group                                     // one effect per group/cohort

The syntax is as simple as described above. Covariates go right after the outcome, and the default estimation method is a doubly robust DiD estimator; you set method(reg) or method(ipw) options to use regression-adjusted or (normalized/Hajek-based) IPW DiD estimators. The default comparison group is not-yet-treated units (if you use the nevertreated option, the comparison group becomes the never-treated units). The default inference procedure is based on a multiplier bootstrap procedure paired with simultaneous bands, so we address head-on the issues of multiple hypothesis tests; if you want analytical standard errors, use the analytical option (add pointwise for pointwise intervals; otherwise the aggregation’s simultaneous band is still bootstrapped, with a note). For panel data, inference accounts for repeated observations of each unit. Use cluster(state), for example, when dependence extends across units within states. In repeated cross sections each observation is a separate unit unless you specify clustering.

Guides

We have prepared some user guides to help you use and understand our commands. Beyond actually reading Callaway and Sant’Anna (2021), you can find all our guides here. A few that stand out include:

Help Files

From inside Stata, you can get help for all our commands: help csdid, help csdid_postestimation, help csdid_estat, help csdid_stats, help csdid_plot.

Citing csdid

We ask that you cite both the method and the software when you use our csdid command: for the method, Callaway and Sant’Anna (2021); for the software, cite the version you ran, which csdid version reports for you.

@misc{CRS2026_csdidStata,
  author = {Callaway, Brantly and Rios-Avila, Fernando and Sant'Anna, Pedro H. C.},
  title  = {csdid: Difference-in-Differences with Multiple Time Periods in Stata},
  note   = {Stata module, version 2.0.0},
  year   = {2026},
  url    = {https://github.com/pedrohcgs/csdid-stata}
}