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

Re: st: cluster option of ivreg2


From   Giovanni Pica <[email protected]>
To   [email protected]
Subject   Re: st: cluster option of ivreg2
Date   Fri, 6 May 2005 10:45:19 +0100

Mark,

thanks for your quick reply. I updated ivreg2 yesterday and have the "ivreg2
1.0.27  02Nov2002" version. Is this the latest?

Let me just add that the error message "matrix not symmetric" appears not
only with the cluster option but also with the robust and gmm options.
Everything works fine with no options.

Thanks again,
giovanni







Quoting Mark Schaffer <[email protected]>:

> Giovanni,
>
> Quoting Giovanni Pica <[email protected]>:
>
> > Dear Stata Users,
> >
> > I have a problem with the cluster option of ivreg2.
> >
> > I am running the following regression:
> >
> > ivreg2 logw X P (T T*P = Z That*P), cluster(personid)
> >
> > this is using IV in a difference in difference regression where T is
> the
> > treatment group, P is post-reform dummy and T*P is the interaction
> > term of interest.
> > The use of That*P as an instrument, where That is obtained from a fist
> > stage regression of T on the exogenous regressors (X and P) and the
> > instruments Z, has been already suggested in Stata list in another
> context
> > (Mark Schaffer in reply to Steve Morris):
> > http://www.stata.com/statalist/archive/2003-11/msg00795.html
> >
> > The cluster(personid) is required because it's a panel of individual
> > wages.
> >
> > All this works when I use the ivreg command, when I use the ivreg2
> > command it says "matrix not symmetric".
>
> Have you updated your ivreg2 so that it's the latest available?
>
> There was a bug in earlier versions that resulted from Stata's matrix
> mulitiplication.  The matrix inversion command requires a symmetric
> matrix,
> and very occasionally (I think because of numerical issues) ivreg2 would
> multiply a series of matrices and end up with a matrix that should have
> been
> symmetric but wasn't.  I am pretty sure we fixed all the instances of
> where
> this could arise in ivreg2, but maybe we missed one.
>
> --Mark
>
> > Is there something wrong with the use of IV in the diff-in-diff
> > framework or the difference between ivreg and ivreg2 is due to the the
> way
> > the commands handles the cluster option?
> >
> > Best,
> > Giovanni
> >
> >
> >
> >
> > *
> > *   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 Schaffer
> Director, CERT
> Department of Economics
> School of Management & Languages
> Heriot-Watt University, Edinburgh EH14 4AS
> tel +44-131-451-3494 / fax +44-131-451-3294
> email: [email protected]
> web: http://www.sml.hw.ac.uk/ecomes
>
> -------------------------------------------------------------------
> DISCLAIMER
>
> This message is subject to http://www.hw.ac.uk/disclaim.htm
>
> -------------------------------------------------------------------
>
> *
> *   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