Dear all,
I'm a phd student in economics and I'm dealing with a linear panel. I've to address potential heteroskedasticity problem.
(i) What kind of test I've to perform to detect heteroskedasticity? (hettest after a regress or lrtest on the estimation of xtgls (specififying "xtgls, panels(hetero)", saving the estimates and then running "xtgls, panels (iid)")
(ii) what estimator is consistent in presence of heteroskedasticity? ("xtgls, panels(hetero)" or "xtpcse, hetonly" that is only OLS with robust standard error (hence, the same results can be simply obtain running "regress, robust")
(iii) is there a test for choosing between OLS and GLS (ex. If hausman test tells me that random effect perform as well as fixed effect I've to delay on xtgls in presence of heteroskedasticity or I can perform my estimation using xtpcse?)
Thank you in advance for your kindly replay,
Best Regards,
Andrea Sisto
University of Tourin (Italy)
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
