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

st: RE: RE: Testing for endogeneity with ivreg2


From   "Shaw, Jim (NIH/NCI)" <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   st: RE: RE: Testing for endogeneity with ivreg2
Date   Fri, 25 Jul 2003 10:21:21 -0400

Thanks for your help!  One other question regarding ivreg2 and GMM.  When
using ivreg2, I noticed on one occasion that the addition of variables to an
IV-GMM regression led to a reduction in both the centered and uncentered
R^2.  This appears to have been related to the use of GMM since when I used
2SLS to estimate the same regression (both with and without the additional
variables), adding the variables yielded an increase in R^2 (which would be
expected).  Any ideas as to what might have happened?

--
Jim

 

-----Original Message-----
From: Steven Stillman (LMPG) [mailto:[email protected]]
Sent: Thursday, July 24, 2003 8:33 PM
To: '[email protected]'
Subject: st: RE: Testing for endogeneity with ivreg2


Jim,
You are 90% of the way to getting your endogeneity test.  As you state, HOLS
is the appropriate alternative for GMM IV.  To test for endogeneity in this
context, run the HOLS model with your suspect endogenous variable in the
orthog option.  So in your example the syntax would be:

ivreg2 y1 ( = x4 x5) x1 x2 x3 , gmm robust cluster(subjid) orthog(y2)

Steve






> -----Original Message-----
> From:	Shaw, Jim (NIH/NCI) [SMTP:[email protected]]
> Sent:	Thursday, July 24, 2003 11:10 PM
> To:	Statalist (E-mail)
> Subject:	st: Testing for endogeneity with ivreg2
> 
> Dear Statalist:
> 
> Does anyone know how the orthog option is used to test for endogeneity
> with
> ivreg2?  Suppose I have 2 models (with clustering) where
> 
> y1 = y2 + x1 + x2 + x3
> 
> and 
> 
> y2 = x1 + x4 + x5
> 
> Using ivreg2 to estimate the first model via GMM, the command would look
> something like this:
> 
> ivreg2 y1 (y2 = x4 x5) x1 x2 x3 , gmm robust cluster(subjid)
> 
> I would like to test whether y2 is orthogonal to the errors.  My
> understanding is that, if that were the case, OLS would be consistent and
> more efficient than the IV estimator.  Actually, I suspect that the OLS
> alternative to the GMM IV estimator would be heteroscedastic OLS (HOLS).
> Is
> this correct?
> 
> --
> James Shaw
> Research Associate
> College of Pharmacy
> The University of Arizona
> *
> *   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/

The information contained in this document is intended only for the
 addressee and is not necessarily the views nor the official 
communication of the Department of Labour.  All final/official papers 
which are sent from the Department will be sent by non-electronic
means, on appropriate letterhead, signed by authorised personnel.
*
*   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/
*
*   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 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index