-xtpcse- should do exactly what you asked -xtgls- to do.
Only a few commands correct for correlations across cross-sectional units, i.e. that type of correlation tested by -xtcsd-: -xtscc- let you do this for fixed effects only. I am not aware of any command allowing to do the same in a random effects environment (unless you think that -xtgls- are random effects), but you may ask to the authors of -xtcsd- (from ssc) if they have some suggestions.
Nicola
At 02.33 05/11/2007 -0500, "dora sepour" wrote:
>Thank you very much Nicola for your help. xtpcse seems a good alternative,
>but can i use it to correct for hetero and correlation with fixed or random
>effects models?
>Dora
>
>Move to:
>xtpcse y x1 x2, corr(ar1)
>Alternatively, you may try -xtgee-, or -xtscc- (from ssc)
>Nicola
>
>At 02.33 31/10/2007 -0400, "dora sepour" wrote:
>>Dear Stata users,
>>
>>I have an unbalanced panel data. I detected heteroskedasticity,
>>autocorrelation (using xtserial) and cross sectional correlation
> (using
>>pesaran test).
>>
>>Using xtgls y x1 x2, panels(corr) corr(ar1) to correct for hetero and
>>dependence, a message tells me that the panels must be balanced. Is
> someone
>>have an idea to correct for hetero and dependence with unbalanced
> panels
>>data?
*
* 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/
