Hi all,

i'm running a panel regression using stata. i wanted to test whether there are indeed region fixed effects (FE) that reside in my data rather than random effects (RE). The Hausman test told me the differences in the FE model coefficients and the RE model coefficients are not systematic. However, the FE model told me the fraction of variance due to u_i (region effects) is 86%. In the RE model, this effect is zero. How do i reconcile these results? Do I need to do something about my data?

Thank you.