I think Nikolas is right, you cannot straightforwardly combine the
xtfrontier with IV commands.
However, if T is large in your data set, the endogeneity bias may not be a
big problem. In this case you can create the lagged dependent variable by
yourself (using gen ldv=L.y) and then estimate the model using xtfrontier
with the ldv included.
If T is small you cannot ignore the problem: then you may use xtabond2
(provided N is large enough) to estimate your production/cost function and
then extract the inefficiencies from the estimated fixed effects, as in
Schmidt, P. and Sickles, R. C. (1984) "Production Frontiers and Panel
Data", Journal of Business and Economic Statistics, 2:4, 367-374.
If N is small too, you may use xtlsdvc instead of xtabond2 and then extract
the inefficiencies in the same way.
Vasilis
On Aug 5 2005, Nicolas Couderc wrote:
Hi Fraser,
(1) The -xtfrontier- (at least in my version of Stata, 8.2. Maybe Stata
9...) do not include the B&C(95) model, only the B&C(92) and previous
models. But you can use the -frontier- command, with the -dist(truncated)
cm()- options, in order to run a one-stage regression. But it's not a
cross-sectional command. See the Stata manuals for details about this
command.
(2) To my knowledge, it's not possible. Why do you want to combine a
dynamic panel data estimator and a stochastic frontier estimator?
Hope this helps, kind regards,
Nicolas
--
If I could get my membership fee back, I'd resign from the human race.
Fred Allen (1894 - 1956)
*
* 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/
