Hi everyone, I have a panel data (T=9 and N=24) and I have the three following problems:
- heteroscedasticity
-cross-sectional dependency
-time fixed effect
first of all, I run fixed via random model and the Hausman test reject the null hypothesis of Random effects. Then i run FGLS model with the following command (xtgls Y X i.year, panels(correlated)corr(psar1) rhotype(freg)) .
Do my estimation is right or false? if false, how can i resolve these problems?
Thanks
0 Response to xtgls command
Post a Comment