[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: The Sargan test in -xtabond(2)-
Date sent: Wed, 4 Aug 2004 11:11:57 +0100 (BST)
Subject: st: The Sargan test in -xtabond(2)-
From: "Clive Nicholas" <Clive.Nicholas@newcastle.ac.uk>
Send reply to: firstname.lastname@example.org
> Kostas Drakos wrote:
> > I am estimating a dynamic panel using the GMM Arellano-Bond routine.
> > It's proved impossible so far to find an instrument set that will
> > satisfy the Sargan test. How crucial is the significance of the Sargan
> > test?
> First of all, use a germane subject heading!
> Second of all, you didn't say whether you're using -xtabond- or
> -xtabond2-. I know this is a bit naughty to say, but if one doesn't give
> you the 'right' Sargan test statistic, why not try the other? (It is, of
> course, Hansen's J in -xtabond2-.)
They are, or should be, the same. Perhaps out of homage to Sargan,
Arellano and Bond use the term "Sargan statistic" throughout their
paper, but the heteroskedasticity-robust version in the paper should
probably be called a "Hansen's J statistic".
> I've had this trouble myself recently, but managed to be creative with the
> instruments I used to satisfy the J test. From what I've learned off
> others, this test is pretty crucial as it provides a strong indication
> that the instruments you're using are invalid.
Quite right. And, sadly, sometimes valid instruments are just
nowhere to be found. In any case, have you tried increasing the
minimum lag between the variables and the instruments? "Older"
instruments are more likely to be valid (but, unfortunately, less
likely to be relevant) than more recent ones.
Hope this helps.
> CLIVE NICHOLAS |t: 0(044)191 222 5969
> Politics |e: email@example.com
> Newcastle University |http://www.ncl.ac.uk/geps
> * 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
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS UK
44-131-451-3485 CERT administrator
* For searches and help try: