Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Re: Xtfrontier

From   Eddy <>
Subject   Re: st: Re: Xtfrontier
Date   Mon, 22 Dec 2003 21:59:34 -0800 (PST)

I want to add my two cents to Scott's advice pertaining to
"environmental factors" in the inefficiency equation. Consider a
flexible/general specification of the inefficiency distribution
u~|N(mu, sigma2)|. If environmental factors are desired, one "should"
be able to add those variables through either mu, sigma2, or both.
However, Stata's current -frontier- model only allows for the
following configurations.

 (1) mu=0,       and sigma2=varlist;
 (2) mu=varlist, and sigma2=constant.

It does not support at least the following configurations:

 (3) mu=constant, and sigma2=varlist;
 (4) mu=varlist,  and sigma2=varlist.

Hung-Jen Wang (2002 Journal of Productivity Analysis, p.241-53) shows
why model (4) is particularly desirable. I think he provides his
Stata program, which should estimate all the models above, on his web
site (try search his name on google).

-- Eddy

Scott Merryman wrote:

> Mahmut,

> I am sending this to the entire list, as other people may find it
of use or I
> might say something dumb which others will catch.

> One possibility would be to use a fixed effects model, using
-frontier- with
> the -uhet- option to model the inefficiency effects.

> Below are some comments I made to a previous question that you
might find
> useful.


> The next question is: do you include these fixed effects (and/or
> factors) in production function or in the inefficiency equation
> (using -uhet- you can possibly solve two problems at once -
correcting for
> heteroskedasticity and incorporating environmental factors on
> see Kumbhakar and Lovell, 2000 "Stochastic Frontier Analysis").

Do you Yahoo!?
New Yahoo! Photos - easier uploading and sharing.
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index