[Limdep Nlogit List] generalized residual of RE probit

[William Greene]

Satu:  It's not a stupid question, but nor is it an easy one.  In the random
effects model (as in other models), the generalized residual is the derivative
with respect to the constant term.  This has to be computed by numerical
integration, as Vella and Verbeek note in the appendix to their paper.  The 
problem, as you note, is computation of theta(i), which is done with Hermite
quadrature.  I can't vouch for the formula you give in this note, however. That
doesn't look right.  Either way, the computation you request is much more 
involved than a simple plug in. I'm not sure there is a simple way to do it
with LIMDEP.  (There may be; the preceding suggests the direction you have
to look).
/B. Greene

************************************************
Professor William Greene
Department of Economics
Stern School of Business
New York University
44 West 4th St., Rm. 7-78
New York, NY   10012
Ph. 212.998.0876
Fax. 212.995.4218
URL. http://www.stern.nyu.edu/~wgreene
Email. wgreene at stern.nyu.edu
************************************************

----- Original Message -----
From: Satu Nivalainen <snivalai at uci.edu>
Date: Thursday, March 23, 2006 12:55 pm
Subject: [Limdep Nlogit List] generalized residual of RE probit

> Dear All,
> 
> Is there any easy way to calculate the generalized residual of the 
> randomeffect probit model in Limdep?
> 
> That is, assumig we have a RE probit model:
> U(it)= gamma'w(it) + theta(i) + eps(it)
> 
> How can I obtain the generalized residual:
> gres = (2U(it)-1)*sigmaeps*normpdf[((gamma'w(it)+
> theta(i))/sigmaeps)(2U(it)-1)]/normcdf[((gamma'w(it)+
> theta(i))/sigmaeps)(2U(it)-1)]?
> 
> To be more specific, I haven't been able to figure out how to get the
> values of theta(i). Sorry if this is a stupid question, but for some
> reason my brains do not work at this point.
> 
> (I need the gen.resid. as an intermediate step in calculating the
> endogeneity correction terms a la Vella and Verbeek, 1998, Whose wages
> unions raise? J. Appl. Econ. So if anyone has any experience in 
> applyingthis method using Limdep, hints would be more than welcome)
> 
> Thanks,
> Satu
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au
>