[Limdep Nlogit List] endogenizing a count estimation within a probit

William Greene wgreene at stern.nyu.edu
Sat Jun 14 18:24:35 EST 2008

(1) No, there is no routine for a probit model with an endogenous
count variable. You would have to work out how the endogeneity
got into the model. The Poisson model does not have a disturbance
in it.  So, what unobservables in the Poisson equation are 
correlated with the unobservables in the probit model to induce
the endogeneity?
(2) I don't understand the second question. (Irrespective of the
answer to (1).)  Is this suggesting a Poisson model with sample
selection - I have a couple working papers on this and you can 
find a very long discussion in my recent Foundations and Trends
paper - then somehow dropped into a third (probit) equation?  Or
something else.
/Bill Greene

----- Original Message -----
To: limdep at limdep.itls.usyd.edu.au
Sent: Thursday, June 12, 2008 5:15:49 PM (GMT-0500) America/Bogota
Subject: [Limdep Nlogit List] endogenizing a count estimation within a probit


In Limdep 9, there is a procedure that accommodates a probit endogenized within
a Poisson/negative binomial count model with selection.

I am in need of the converse.  I am constructing a model of "Days in Labor
Market Programs" for a European sample, with selection, followed by a probit
estimating the probability of subsequent migration.

Two questions:

1.  Is there a "hard wired" Limdep procedure for this type estimation?


2.  If the answer to #1 above is yes, does it allow for selectivity in the
first step of the estimation?

Thanks in advance.

Robert A. Nakosteen
Professor, School of Management
University of Massachusetts/Amherst
Amherst, MA 01003
Limdep site list
Limdep at limdep.itls.usyd.edu.au

Professor William Greene
Department of Economics
Stern School of Business
New York University
44 West 4th St., Rm. 7-78
New York, NY   10012

More information about the Limdep mailing list