l have a state-by-year panel data where my RHS treatment variable is turning on and off. So I'm trying to estimate dCDH estimator. However, I am running into issues where I am finding different estimates and "# of switchers" btwn the two dCDH command did_multiplegt vs. did_multiplegt_dyn. The code below is a very simple code that creates fake data and runs did_multiplegt and _dyn command.
Is there a reason why the results are different. Which command should I use??
HTML Code:
set seed 12345 clear set obs 100 gen time = ceil(_n/10) gen id = _n-10*(time-1) gen x = runiformint(0,1) gen y = rnormal() did_multiplegt_dyn y id time x did_multiplegt y id time x
