[Limdep Nlogit List] Heckit model

[Randy Kesselring]

Marc,
 
You are going to have to do some investigating.  First, use the reject command to limit your observations to the ones you intend to use for the second stage regression and run that regression without the saved inverse Mills ratio and see what happens.  That should point you in the right direction.
 
Dr. Randall G. Kesselring
Professor of Economics
Arkansas State University
Phone: 870-680-8467

________________________________

From: limdep-bounces at limdep.itls.usyd.edu.au on behalf of M Goergen
Sent: Sat 12/1/2007 4:08 AM
To: Limdep and Nlogit Mailing List
Subject: RE: [Limdep Nlogit List] Heckit model



Dear Randall

Thank you for your reply.

I have 29 observations that equal one and 85 that are equal to zero. As the
selection equation contains a constant and just one variable, I am surprised
that I do not have enough degress of freedom. I have even recoded the dependent
variable in the probit, so that I have 85 observations which equal one, but
this still does not make a difference.

Marc


Quoting Randy Kesselring <RANDYK at astate.edu>:

> Marc,
>
> When you run the probit, it uses all of your observations and you have a
> sufficient number to accomplish that estimation.  When you use the
> select command it selects only the observations for the variable on the
> left-hand-side of the probit that equal one.  This significantly reduces
> the number of observations, and in your case, there are not enough
> observations to accomplish the estimation.
>
> Randall G. Kesselring
> Professor of Economics
> Arkansas State University
> Phone: 870-680-8467
> Randyk at astate.edu
> -----Original Message-----
> From: limdep-bounces at limdep.itls.usyd.edu.au
> [mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of Marc
> Goergen
> Sent: Friday, November 30, 2007 1:44 PM
> To: limdep at limdep.itls.usyd.edu.au
> Subject: [Limdep Nlogit List] Heckit model
>
> Dear all
>
> I am trying to run a Heckman selection model. The probit runs well, but
> when
> I try to run the linear regression using the 'select' command I get
> "error
> message 201: SELECT - selection leaves insufficient degrees of freedom".
> What is the cause of this?
>
> Best wishes,
>
> Marc
>
>
>
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au <http://limdep.itls.usyd.edu.au/> 
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au <http://limdep.itls.usyd.edu.au/> 
>


--
Prof. Marc Goergen
Sheffield University Management School (SUMS)
Sheffield University
9 Mappin Street
Sheffield S1 4DT
UK

Tel: +44 (0)114 222 3488 (direct line)
Tel: +44 (0)114 222 3379 (Ms Karen Hector - secretary)
Fax: +44 (0)114 222 3348

European Corporate Governance (ECGI) Research Associate
http://www.ecgi.org <http://www.ecgi.org/> 

SSRN author page:
http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=29035
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au <http://limdep.itls.usyd.edu.au/>