sdid (Arkhangelsky et. al. 2021)
Table of contents
Introduction
The sdid command is written by Damian Clarke and Daniel Pailañir based on the Arkhangelsky et. al. 2021 paper Synthetic Difference-in-Differences. A detailed description is provided on GitHub.
Installation and options
ssc install sdid, replace
Take a look at the help file:
help sdid
Test the command
Please make sure that you generate the data using the script given here
sdid Y id t D, vce(bootstrap) seed(1000)
Since we are using bootstrapped standard errors, we fix the seed for replicability. We get this output:
Bootstrap replications (50). This may take some time.
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.................................................. 50
Synthetic Difference-in-Differences Estimator
-----------------------------------------------------------------------------
Y | ATT Std. Err. t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------
D | 131.07490 8.42889 15.55 0.000 114.55458 147.59522
-----------------------------------------------------------------------------
95% CIs and p-values are based on Large-Sample approximations.
Refer to Arkhangelsky et al., (2020) for theoretical derivations.
The command also has a built in graph option:
sdid Y id year D, vce(bootstrap) seed(1000) graph
The command saves a set of graphs with prefix g2_ that represent synthetic DiD graphs split by the year of the first treatment:
The weights used to average pre-treatment periods are shown as area fills at the bottom of the figures.