st: Re: Re: Limdep and Stata

["Kenneth G. Liu" <gao.liu@gmail.com>]

I thought I could also. But the default option of -xtregar, fe- just gave me 
quite different estimations. Anyone knows the reason?

Thanks

Gao





----- Original Message ----- 
From: "Rodrigo A. Alfaro" <raalfaroa@gmail.com>
To: <statalist@hsphsun2.harvard.edu>
Sent: 2007?6?29? 5:27 PM
Subject: st: Re: Limdep and Stata



> I think that you can replicate LIMDEP results using -xtregar, fe-
> Rodrigo.
>
> ----- Original Message ----- 
> From: "Kenneth G. Liu" <gao.liu@gmail.com>
> To: <statalist@hsphsun2.harvard.edu>
> Sent: Friday, June 29, 2007 5:07 PM
> Subject: st: Limdep and Stata
>
>
>
> > Hi,
> >
> > I tried to analyze a panel dataset containing observations with a maximum 
> > of 21 years in 12 countries (unbalanced panel). Fixed effect is used and 
> > the disturbance term is modelled as an AR1 process.
> >
> > I used both Stata and Limdep to estimate the model ("xtregar,fe" for 
> > Stata, and "; fe; AR1;" for Limdep). However, the results (pho, 
> > coefficients  and t values) turned out to significantly different. The 
> > estimated pho in  Limdep is much smaller.
> >
> > I know that Limdep uses a 2sls estimation with the Cochrane-Orcutt 
> > transformation, but don't know how stata estimates the model. Why did 
> > they lead to significantly different results? Which one is more 
> > appropriate?
> >
> > Thanks
> >
> >
> > Gao
> > *
> > *   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/
> *
> *   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/ 
*
*   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/