RE: st: Weak identification with a large number of endogenous variables

["Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>]

Marjan,

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Roger Harbord
> Sent: 21 February 2010 18:28
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: Weak identification with a large number of 
> endogenous variables
> 
> On 21 February 2010 00:21, PETRESKI Marjan 
> <M.Petreski@staffs.ac.uk> wrote:
> > Dear all,
> >
> > Can somebody advise me on a solution of the missing 
> Stock-Yogo critical values for a large number of endogenous 
> variables under xtivreg2? I do have 14 endogenous variables, 
> and get Stock-Yogo F-stat of about 2.5, which seems low.

You may be able to reach a conclusion by considering the test statistics
reported for underidentification.

If something is underidentified, it is also weakly identified, by
definition.

I suspect you will fail to reject the null of underidentification.  If
the test stats fail to reject the null of underidentification, you have
a serious identification problem.  You can call it weak identification
if you want, but it's actually worse than that.

HTH,
Mark

> However, increasing the instrument set leads to only marginal 
> improvement of the F-stat.
> >
> > Is the IV bias reduced with a large number of endogenous 
> variables and if yes, can you please lend me some reference?
> >
> > Best regards,
> >
> > Mr. Marjan Petreski
> > Staffordshire University
> 
> I think the best solution might be to reduce the number of 
> endogenous variables:
> <http://www.mostlyharmlesseconometrics.com/2010/02/multiple-en
> dogenous-variables-what-now/>
> 
> Roger.
> --
> Roger Harbord
> http://www.epi.bris.ac.uk/staff/rharbord.htm
> 
> *
> *   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/
> 


-- 
Heriot-Watt University is a Scottish charity
registered under charity number SC000278.


*
*   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/