search

grf-labs / grf

Generalized Random Forests
https://grf-labs.github.io/grf/
GNU General Public License v3.0
952 stars 248 forks source link

Causal forest in DID setting #1064

Open 992bin opened 2 years ago

992bin commented 2 years ago

Hello:

I am new here. Does anybody know how to set up a causal tree in a diff-in-diff setting with observational data? Is anyone aware of any papers where this has been done?

Any help would be much appreciated!

swager commented 2 years ago

Do you have 2 time periods or many? In the case of 2 time periods, the easiest is to just run causal forests as-is on first differences Y{i,post} - Y{i,pre}. This paper discusses the statistical setting:

Abadie, Alberto. "Semiparametric difference-in-differences estimators." The Review of Economic Studies 72.1 (2005): 1-19.

For an application using this kind of approach, see:

Wang, Guihua. "The Effect of Medicaid Expansion on Wait Time in the Emergency Department." Management Science, forthcoming.