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

Re: st: Two stage estimation: Can the first stage be a count model


From   "Mark Schaffer" <M.E.Schaffer@hw.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Two stage estimation: Can the first stage be a count model
Date   Thu, 03 Feb 2005 18:43:12 -0000

Tinna,

Related questions come up from time to time on the list.  See below.

Date sent:      	Thu, 3 Feb 2005 10:47:04 -0500
From:           	TinnaLaufey Asgeirsdottir <statalist@gmail.com>
To:             	statalist@hsphsun2.harvard.edu
Subject:        	st: Two stage estimation: Can the first stage be a count model
Send reply to:  	statalist@hsphsun2.harvard.edu

> Hello fellow stata users,
> 
> I am currently using the ivreg command for a two stage least squares
> estimation but am wondering how fitting it is in my case.  The
> variable I am instrumenting is a count measure 0, 1, 2, 3 (number of
> specific alcohol problems individuals have) and my instrument is a 0 -
> 1 indicator variable (daily smoker or not).  I have two questions for
> you:
> 
> A)  Could it cause a problem in my estimation that the first stage
> consists of an indicator explaining a 0 - 1 - 2 - 3 variable?

Mostly no.  The consistency of the coefficient estimates for the main 
equation doesn't rely on specifying the correct functional form for 
the first stage.  However, you can get more efficient estimates if 
you estimate as a system, e.g., the Poisson regression becomes a 
structural equation to be estimated along with the other eqn.  See 
e.g.

http://www.stata.com/statalist/archive/2004-07/msg00710.html

> B)  Is there a different command that does two stage estimations with
> the first stage being some type of a count model (e.g. Poisson)?

I don't think one is around, though -findit- might turn up something. 
 A related alternative is Wooldridge's suggestion to do a first-stage 
Poisson by hand, get the predicted counts, and then use this variable 
as the excluded instrument in an IV regression.  See e.g.

http://www.stata.com/statalist/archive/2004-09/msg00337.html

Hope this helps.

Cheers,
Mark

> 
> Any insights are greatly appreciated
> Tinna
> *
> *   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/

Prof. Mark E. Schaffer
Director
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS  UK
44-131-451-3494 direct
44-131-451-3296 fax
44-131-451-3485 CERT administrator
http://www.sml.hw.ac.uk/cert

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index