eventstudyinteract is a Stata package that implements the interaction weighted estimator for an event study. Sun and Abraham (2020) proposes this estimator as an alternative to the canonical two-way fixed effects regressions with relative time indicators. Sun and Abraham (2020) proves that this estimator is consistent for the average dynamic effect at a given relative time even under heterogeneous treatment effects. As outlined in the paper, eventstudyinteract uses either never-treated units or last-treated units as the comparison group. A similar estimator is Callaway and Sant'Anna (2020), which uses all not-yet-treated units for comparison.
eventstudyinteract also constructs pointwise confidence intervals valid for the effect at a given relative time. The bootstrap-based inference by Callaway and Sant'Anna (2020) constructs simultaneous confidence intervals that are valid for the entire path of dynamic effects, i.e., effects across multiple relative times.
Callaway and Sant'Anna (2020) provides an R package did for their estimation and inference procedure.
eventstudyinteract can be installed easily via the github package, which is available at https://github.com/haghish/github. Specifically execute the following code in Stata:
net install github, from("https://haghish.github.io/github/")
To install the eventstudyinteract package , execute the following in Stata:
github install lsun20/eventstudyinteract
which should install the dependency packages. If not working, try manually install
ssc install avar
ssc install reghdfe
ssc install ftools
If you run into an error message of class FixedEffects undefined, this can usually be resolved by reghdfe, compile as suggested by the repository of reghdfe.
To update the eventstudyinteract package, execute the following in Stata:
github update eventstudyinteract
Documentation is included in the Stata help file that is installed along with the package. An empirical example is provided in the help file.
Liyang Sun
Preprint of Sun and Abraham (2020) is available on my personal website.