> xtgls in STATA uses Feasible Generalized Least Square (FGLS) with
> cross-sectional time series linear models. My data is cross-sectional
> but only one-point in time. I wish to run a model which would do
>
> regress logconsumption <vars>
>
> and force FGLS estimation on that.
>
> Is that at all possible or meaningful!
To do anything like GLS, you need to have a clear idea of the covariance
structure of your errors. What -xtgls- does is it says, I know that my
errors are only correlated within the panels, and there is nothing going
on between panels. (Some further structures can be imposed, but this is
the primary simplifying assumption.)
As long as you know what your covariance structure of the errors should
be, you can do things by hand through -mat accum- (see help matrix), and
get the standard errors from the textbook matrix expressions. Those would
be wrong if you had to estimate the covariance parameters; the only way to
get the right standard errors would be to go to the maximum likelihood
procedures, see help ml.
--- Stas Kolenikov
-- Ph.D. student in Statistics at UNC-Chapel Hill
- http://www.komkon.org/~tacik/ -- [email protected]
* This e-mail and all attachments to it are not intended to provide any
* reasonable point of view and was transmitted to you in error. It
* should be immediately deleted by all recepients unless they really
* enjoy communicating with the author :). Other restrictions apply.
*
* 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/
