[Sergio Correia]

Oh I see your point; -test- only works if you are comparing the full vs the nested model (in which case, you only run the full model).
If you want to test *some* of the variables but not all, then you can't do it with a Wald test and you would need to use suest.

I found a way that would make -reghdfe- compatible with -suest-. It would require three things

• When you run reghdfe you need to save the fixed effects (within the absorb() option)
• Instead of using vce(unadjusted), the code in reghdfe would have to use vce(ols) everywhere
• The internal program -reghdfe_p- would have to call -_score_spec- in order to understand what suest is doing.

All in all, seems doable, but not entirely straighforward. If you still need this in a couple of days, send me an email ( [email protected] ) and I'll run the changes on reghdfe.

Cheers,
S