did_multiplegt (Chaisemartin and D’Haultfœuille 2020, 2021)
Table of contents
Installation and options
ssc install did_multiplegt, replace
Take a look at the help file:
help did_multiplegt
Test the command
Please make sure that you generate the data using the script given here
Let’s try the basic did_multiplegt command:
did_multiplegt Y id t D, robust_dynamic dynamic(10) placebo(10) breps(20) cluster(id)
and we get this output:
DID estimators of the instantaneous treatment effect, of dynamic treatment effects if the dynamic
option is used, and of placebo tests of the parallel trends assumption if the placebo option
is used. The estimators are robust to heterogeneous effects, and to dynamic effects if the
robust_dynamic option is used.
| Estimate SE LB CI UB CI N Switchers
-------------+------------------------------------------------------------------
Effect_0 | -.0608394 .3699999 -.7860393 .6643604 78 23
Effect_1 | 8.49767 .3811963 7.750525 9.244815 78 23
Effect_2 | 17.64773 .4201394 16.82425 18.4712 78 23
Effect_3 | 25.9377 .5058977 24.94614 26.92925 78 23
Effect_4 | 34.62362 .8107131 33.03462 36.21262 75 23
Effect_5 | 42.85682 1.155268 40.5925 45.12115 64 19
Effect_6 | 51.93103 1.416187 49.1553 54.70676 64 19
Effect_7 | 60.13327 1.799572 56.60611 63.66043 64 19
Effect_8 | 68.82446 1.901396 65.09773 72.5512 64 19
Effect_9 | 77.30792 2.222771 72.95129 81.66455 64 19
Effect_10 | 85.78878 2.535131 80.81992 90.75764 55 19
Placebo_1 | -.1308918 .5886522 -1.28465 1.022866 78 23
Placebo_2 | .1944381 .4274514 -.6433666 1.032243 78 23
Placebo_3 | .0639963 .4441797 -.806596 .9345885 78 23
Placebo_4 | .2572878 .4284934 -.5825592 1.097135 78 23
Placebo_5 | .0679468 .3048067 -.5294744 .665368 78 23
Placebo_6 | -.082143 .2507972 -.5737055 .4094195 78 23
Placebo_7 | -.271289 .3715318 -.9994913 .4569133 78 23
Placebo_8 | .0338621 .2511709 -.4584328 .526157 78 23
Placebo_9 | -.1010115 .2640631 -.6185751 .4165522 78 23
Placebo_10 | .3842823 .3026827 -.2089757 .9775403 78 23
When dynamic effects and first-difference placebos are requested, the command does
not produce a graph, because placebos estimators are DIDs across consecutive time periods,
while dynamic effects estimators are long-difference DIDs, so they are not really comparable.
Even though we are warned in the output above that we should not compare the event study estimates, we can still plot these using the event_plot (ssc install event_plot, replace) command as follows:
event_plot e(estimates)#e(variances), default_look ///
graph_opt(xtitle("Periods since the event") ytitle("Average causal effect") ///
title("did_multiplegt") xlabel(-10(1)10)) stub_lag(Effect_#) stub_lead(Placebo_#) together
and we get this figure:
