Random Effects estimators - time-invariant variables effects benefit

[@rareclassicmovies6429]

what will happen if cov(race(i), x(i,t)) is not equal to 0? and the same for all time-independent variables? I mean is there any difference between alpha and the other time-independent variables? thanks

[@anshulgoel811]

That would be the case of collinearity, I suppose. If very high collinearity we will not able to get any estimate out of sample.