Many thanks for your inputs.
I still have an unsolved issue which concerns the time dimension of my dataset.
Following the example you provided (http://www.stata.com/statalist/archive/2004-03/msg00101.html) I understand that to correct for
both autocorrelation and correlated residuals between equations I have to treat the consumption categories, which corresponds to the equations of the system, as panels and use xtgls with the option 'panels(correlated)' to replicate the sureg's treatment of the residuals.
The problem is now with the error message 'repeated time values within panel' as now for each consumption category, the panels, there are 15 time series data one for each country. Is there a way to overcome this problem or I simply accept a lower efficiency and go for an equation by equation estimation with xtgls?
Best,
Ignazio
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
