# Re: st: Heteroscedasticity and endogeneity in panel

 From "R.E. De Hoyos" To Subject Re: st: Heteroscedasticity and endogeneity in panel Date Sun, 8 Feb 2004 19:14:12 -0000

```Vincenzo,

What about the multicolinearity between Xs and Zs?

You are running a model where Y = f [ X(Z), Z] hence the Cov(X,Z) is not
zero.

Andrea,

To account for heteroscedasticity across panels you can use the -xtgls-
command (use the -force- option if your panel is an unbalanced one)

The command -xtivreg- will help you with the endogeneity problem by doing a
similar 2-stage least squares test as the one outlined by Vincenzo.

Hope this helps,

Rafa

----- Original Message -----
From: "Vincenzo Verardi" <vincenzo_verardi@hotmail.com>
To: <statalist@hsphsun2.harvard.edu>
Sent: Saturday, February 07, 2004 12:52 PM
Subject: RE: st: Heteroscedasticity and endogeneity in panel

> I do not know if it is 100% orthodox, but I would probably act like this:
>
> If you have a model, lets say:
>
> yit=a+bxit+czit+e where xit is endogenous and you would like to instrument
> it by wit (zit is a set of exogenous variables).
>
> I would first estimate
>
> xit=a2+c2zit+gwit+e2
>
> (where a2 and c2 are just the coefficients associated respectively to the
> constant and the z variable in this equation and e2 is an error term)
>
> and fit the xit (let's call it xxit).
>
> then I would estimate
>
> yit=a+bxxit+czit+e
>
> with the heteroskeastic structure you want.
>
> Normally this should solve your problem (or at least I hope)
>
> VV
>
> P.S. For the tests, this is fairly standard. Check xttest for
> heteroskedasticity and maybe a Durbin-Wu-Hausman for endogeneity.
>
>
>
>
>
> >From: "Andrea Molinari" <A.Molinari@sussex.ac.uk>
> >To: <statalist@hsphsun2.harvard.edu>
> >Subject: st: Heteroscedasticity and endogeneity in panel
> >Date: Thu, 5 Feb 2004 13:44:46 -0000
> >
> >Dear Statalisters,
> >
> >I am trying to estimate a model in panel data, but which has both
> >heteroscedastic variances and an endogeneity problem.  I was wondering
> >whether there would be a suitable model to do this at once.  I also need
to
> >test for both of these.
> >
> >My dataset is unbalanced, and it has a big N (420 individuals) and a
> >relatively smaller T (22 years).
> >
> >I would really appreciate any suggestions!
> >
> >Thanks,
> >Andrea Molinari
> >
> >
> >*
> >*   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/
>
> _________________________________________________________________
> Charla con tus amigos en línea mediante MSN Messenger:
> http://messenger.latam.msn.com/
>
> *
> *   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/
```