It's important to know also _why_ the problem
arises.
In essence, -by:- controls the fitting of
a sequence of models, while -predict- remembers the
last model fitted. The two don't mesh here in terms of what
you want.
Nick
[email protected]
Nick Cox
> As I understand it, you are fitting a separate
> model for each sector, and want to pool the predictions.
>
> Here is one way to do that:
>
> levels sector, local(S)
> gen te = .
> foreach s of local S {
> capture drop work
> qui xtfrontier
> logrgva time logwage logplnt if sector == `s', tvd
> predict work if e(sample)
> qui replace te = work if sector == `s'
> }
Vidya Mahambare
> > I am using 'xtfrontier' command for a C-D production
> > function for 13 sectors as
> > follows:
> > bysort sector: xtfrontier logrgva time logwage logplant, tvd
> >
> > When I try to generate 'predict te', it is generated only
> > for the last sector.
> >
> > Can I add any subcommand to create estimates of TE for each
> > of the sectors? or
> > do I have to run xtfrontier for each sector and generate te.
*
* 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/
