In a paper of 1986 published in Econometric Theory, Baltagi suggests that in order to choose an error component (RE or FE) model or the Parks-Kmenta model, one should apply the Bartlett's test for homoskedasticity to the OLS residuals.
I've been checking the command for the Bartlett's test (WNTESTB), however Stata reports an error specifying that such command works only with time-series data.
I assume then, that the likelihood ratio test that checks the panel-level heteroskedasticity can be used as an alternative. Can anybody confirm please?
Thank you for your help.
Ilaria
*
* 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/
