[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: predicting efficiency with xtfrontier
On 11/19/2005 2:54 AM, joserra coco wrote:
> Hi everybody,
> I was wondering if anybody could offer some guidance on two problems I
> have dealing with xtfrontier.
> Using panel data, I am trying to estimate technical efficiency for a
> group of firms, over 10 years. The main problem I encounter is that,
> when I predict the technical efficiency scores, I get a single value for
> each firm, instead of one value per firm per year. My understanding is
> that if I use the time varying efficiency option (tvd), these values
> should be different within firms from year to year. This is an example
> of the code I use.
> xtfrontier lny lnk lnl lnm , tvd iterate(26)
> predict efficiency, u
> Alternatively, I have also tried to include time dummies to account for
> technical progress. Same situation.
> I would appreciate any guidance on this matter. Also, once I reach the
> technical scores, I would like to make them dependent on a set of firm
> attributes (not the ones I have used in the stochastic production
> frontier). Is this supposed to be a separate analysis that cannot be
> estimated through xtfrontier?.
> Thanks a lot for any feedback.
you should be using "predict var, te" instead of
"predict var, u" for getting technical efficiency
And it works fine:
xtfrontier lnmachines lnworkers lnwidgets, tvd
predict te, te
br id te
and technical efficiency is really time-varying!
* For searches and help try: