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

RE: st: biprobit

From   "Conway, Karen" <[email protected]>
To   <[email protected]>
Subject   RE: st: biprobit
Date   Wed, 13 Oct 2004 09:11:25 -0400

Just a follow-up question to Marc's clear discussion of bivariate probit... Suppose that one finds rho is not significantly different from zero in the case of a simultaneous probit model.  Does that mean that you can treat the endogenous dichotomous regressor as exogenous when you estimate the two probits separately?  In other words, is rho a de facto exogeneity test as well?  My understanding is that you can include the observed dichotomous regressor on the rhs, even though its endogenous, BECAUSE you are estimating with a bivariate probit (Greene 2003, p. 716).  

Karen Conway

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Marc F. Bellemare
Sent: Wednesday, October 13, 2004 7:22 AM
To: [email protected]
Subject: Re: st: biprobit

Hi Victoria:

A bivariate probit will estimate a pair of probits. For example, you might 
want to estimate simultaneously whether or not a college student works 
during a given semester (y_{1} = 1 if she works, 0 otherwise) and whether 
or not she drops out of school (y_{2} = 1 if she drops out of school, 0 
otherwise). If you need to see the likelihood function of the bivariate 
probit, turn to Maddala (1983) (maybe Gouri�roux, 2001, also has it).

It can be used whenever you have a simultaneous or sequential pair of 
probits. The coefficients are indeed consistent, assuming the hypotheses of 
maximum likelihood estimation are satisfied and the regularity conditions 
hold true.

The main question is: should you estimate a bivariate probit or two 
separate probits? The biprobit command will return a correlation 
coefficient ("rho") which is the correlation coefficient between the 
residuals of each of the two probits. If rho is statistically significantly 
different from zero, then you should estimate the two probits 
simultaneously. If rho is not statistically significantly different from 
zero, then you can stick to estimating two separate probits.


At 09:13 AM 10/13/2004 -0300, you wrote:
>I need your help.
>Does anybody knows exactly, theoretically speaking, what the biprobit 
>comando does? When exactly can be used? Are the coefficients 
>Thank you all for your help.
>Victoria ROdriguez
>*   For searches and help try:

Marc F. Bellemare
Ph.D. Candidate
Department of Applied Economics and Management
Cornell University
Ithaca, NY 14853-7801

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index