Number of placebo periods

[JorgeGarciaHombrados85]

I am writing because I am using the updated version of the stata command did_multiplegt_dyn for the estimation of a difference-in-difference with staggered treatment adoption and continuous treatment indicator. However, the event study is only showing one placebo period when I belive it should be showing two:

-Panel of counties -4 periods

• Treatment variable in the first two periods is 0 in all observation -Then some observations receive the treatment with different intensities in period 3 -Some new observations receive the treatment with different intensities in period 4 -Many observations remain without treatment for all the periods

This is the specification: did_multiplegt_dyn fgc_religion LGA year_interview sh_population_3g_LGA , effects(2) placebo(2) cluster(LGA) continuous(1) bootstrap(50,1)

I am unable to figure out why the command is estimating only one placebo period rather than 2. I also estimate two post-period. I am missing something about how the placebo effects are calculated?

Thank you very much,

Jorge

[DiegoCiccia]

Hello Jorge, This is Diego from Clement de Chaisemartin's RA Team. Thanks for your interest in did_multiplegt_dyn! I am sorry for the late reply, but your inquiry sparked a lot of interest and we ended up building a full extension of the program aimed at retrieving the placebos that are normally not computed by the vanilla version of did_multiplegt_dyn. You can find here a vignette where we present a simple method to retrieve such placebos using a spin-off function of did_multiplegt_dyn that can be downloaded from GitHub. Please let me know if you have any issues/questions. Lastly, given your design, you should drop the continuous option, since your baseline treatment is 0 for all groups and, as a result, you do not have any issue in finding groups that have the same status quo treatment. To this end, your command line should look like this:

did_multiplegt_dyn fgc_religion LGA year_interview sh_population_3g_LGA , effects(2) placebo(2) cluster(LGA)

I hope this helps! Best, Diego