Dear all,

I ran some DiD regressions to estimate the effect of a policy on the level of good supply per firm.

Now I would like to infer also implications for aggregate supply.
(which I do not want to infer from a simple pre-post comparison of actual aggregate supply as that would be biased by other inter-temporal changes).

Now to the extent to which effects are heterogeneous by firm size I cannot just say that if the average firm was triggered to grow its supply by 5% then so did aggregate supply. Instead I'd like to re-estimate the equation but assign more weight the larger the firm.

My first hunch was to estimate

xtset firm
xtreg supply post_treatment FE_firm* FE_time* [iweight = firmsize], cluster(firm) fe
or
reghdfe supply post_treatmentFE_time* [iweight = firmsize], absorb(firm) cluster(firm)

Stata told me that the -iweight- option is not allowed with -xtreg- or -reghdfe-, so then I did the same with [aweight = firmsize] which runs through.

Yet I'm not sure whether this procedure makes sense from what I've described and would be very grateful if you have any views on whether it does or whether you'd proceed differently?

Thanks so much!
PM