Hi,

I am currently doing an econometrics project (due in 6 days' time). When I run fixed effects regression in Stata, the time invariant dummy variable in my model (whether a country is a Latin American country or not) of course gets omitted, and when I run random effects it doesn't. However when I run a Hausman test it says fixed effects is the better model. The time invariant dummy (Latin) is one of my main variables of interest, so how should I go about all this? Do I use random effects instead? Or some other regression for panel data?

I know similar questions have been asked but I can't seem to find an answer about what exactly to do about it!

Many thanks,
Riaz