Dear Paloys,
-biprobit- is designed to estimate two probit equations with correlated
disturbances, but not to estimate a simultaneous system. You can even use
-biprobit- if you have a recursive system. See Greene, William H.
"Econometric Analysis". Prentice Hall. Fifth Edition. Section 21.6.6, pages
715-19.
There is a user-written command -ivprob- that that estimates
probit models with endogenous regressors. Type in
findit ivprob
to locate, read about, and install it.
Gustavo
-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Alfredo Paloyo
Sent: Thursday, February 03, 2005 7:09 PM
To: [email protected]
Subject: st: biprobit model
Dear all,
Is it valid to run -biprobit- on a model described by the following:
y = f(x, a)
x = f(y, a, w)
In other words, the dependent variable of the first equation is among
the independent variables of the second equation, and vice-versa.
The results of the command gives exactly the same P>|z|'s for the common
arguments (i.e., a), with the coefficients having simply the opposite
signs (i.e., same absolute value). Variables x and y have exactly the
same coefficient and P>|z|. Standard errors are the same everywhere.
Much thanks.
--
Paloys
*
* 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/
*
* 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/