Difference in difference regression with multiple event

[Par Broek]

Dear statalist community,

I'm fairly new to stata and have a question on something I'm trying to do for my bachelor thesis.
I'm trying to find out if natural disasters (storms, floods) affect firm performance (employment growth, capital growth, productivity, etc) in Germany, and want to do this with a difference in difference regression.
I know how to do a diff in diff regression for a single disaster, but ideally, I would like to look at multiple disasters over a longer time period (2013-2020) and see something like a cumulative effect. Is this possible, or do I need to do the regression for each disaster?

Kind regards,

Par

[Jared Greathouse]

I go over this here

[Par Broek]

Thank you Jared for your comment!

I go over this here

Are you suggesting I use allsynth instead of DiD? Like I said I am fairly new to stata, so I dont really know how to do synthetic control method. Do you maybe have a video/article where this is explained for "beginners"?

Thanks in advance,

Par

[Jared Greathouse]

Allsynth isn't really for beginners, what I was really referring to was my first paper that I linked to, where the authors explain all the advances in Difference-in-differences which is likely the better design you'll want to go with.

Considering that your time period of interest is quite short, I wouldn't recommend SCM in this instance. Perhaps, read the paper i linked, and then look into the didmultiple_gt, csdid and flexpaneldid commands. Simply Googling them should guide you to the resources you seek.

But, I really do advise you to read the first paper I mentioned in my link. Staggered addoption DD is sort of a different monster than normal DD, so you'll need to read the details to ensure that the estimator you use is doing that which you need it to do

[Par Broek]

Thank you again Jared, really appreciate your help.

[Helen Chang]

Thank you again Jared, really appreciate your help.

Hi! Did you figure this out? I am also trying to figure out how to conduct such staggered DiD? Thanks.