Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Two stage least square estimation overidentified Sargan test


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Two stage least square estimation overidentified Sargan test
Date   Wed, 1 Jun 2011 11:44:21 -0400

Not linear transformations--nonlinear transformations.  Are you
referring perhaps to p 523 of
http://www.stata-journal.com/sjpdf.html?articlenum=st0136
? Note that if you assume Z and e are independent, rather than simply
that Z and e are uncorrelated, you can use all manner of
transformations of Z as additional excluded instruments.  You can also
interact Z with any X that you consider exogenous.  In each case, the
overid test is a diagnostic of how reasonable your assumption was.  Be
careful of multiple testing and pretesting bias though--if you have a
lot of poor instruments, some are bound to pass an overid test and
give you really poor estimates.

On Wed, Jun 1, 2011 at 6:11 AM, claudiac.4@libero.it
<claudiac.4@libero.it> wrote:
> Dear statalist,
>
> I am working with a just identified model thus I can not perform the validity
> test proposed by Stata 10 (Sargan test). I read in previous discussion that to
> avoid this problem it is possible to increment the number of instruments by
> adding linear transformations of these. Does someone know something more about
> this? What are the main theoretical references?Please let me know.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index