It seems to be well-documented (here: https://www.statalist.org/forums/for...port-fvvarlist or here: https://www.stata.com/statalist/arch.../msg00707.html) that xtoverid does not work when factor variable are included in a regression using xtivreg.

I am using factor variables in an xtivreg regression, and I would like to know the first stage F-stat for my excluded variables. Is there any way to do this w/out using xtoverid?

If there is no post-estimation command that works to do this, I can of course separately run what I think is the 1st stage, and test my excluded variables myself. From page 20 of the manual (https://www.stata.com/manuals/xtxtivreg.pdf) it looks like I would first (a) remove all fixed effects using xtreg, then (b) run a 2SLS regression of my 1st stage using ivreg or ivreg2. Does anyone know if this is indeed the best manual approximation of the first stage of xtivreg?