Markus,
The way you take care of the heterogeneity bias (i.e. correlation between
unobserved individual effects and observed variables) is by only looking at
the within variation. The only other option I know of is the correlated
random effects model (it is a Chamberlain idea but I don't know the cite
off-hand) where you specify the relationship between the unobserved
individual effects and observed variables (typically, a linear one). This
type of model can be estimated in stata by putting your data in wide form
and running sureg on it with the appropriate constraints. This is a bit
tricky and some versions of the model can't be estimated in stata because
they imply non-linear constraints. I have some sloppy do-file code which
estimates this type of model, which I am happy to share but don't want to
circulate. Contact me directly if you would like a copy of it.
Steve
> -----Original Message-----
> From: [email protected] [SMTP:[email protected]]
> Sent: Tuesday, February 25, 2003 2:39 AM
> To: [email protected]
> Subject: RE: st: Re: panel: within and between dimension/correlated
> effect s
>
> Scott,
>
> I cannot use a random effects estimator because of strong correlation
> between the individual effects and the regressors - heterogeneity bias
> according to Chamberlain. On the other hand, the fixed effects estimator
> throws out all the between variation. I know that there are ways of taking
> into account the heterogeneity bias, but do not know how to do it - that
> is
> why I wanted to know if anybody else does.
>
> Markus
>
>
> -------------
> > Greetings,
> >
> > I posted this question some time ago, without any reaction. Therefore, I
>
> > will try to make it clearer.
> >
> > I am estimating a macroeconomic panel data model, and I am looking for
> an
> > estimator that accomodates both the within and the between dimensions.
>
> Isn't the random effects estimator (xtreg, re) a weighted averaged of the
> between and within results?
>
> Have you taken look at xtgee - it allows for variety of correlation
> structures
> and semi-robust standard errors, though it is a population-averaged
> estimator.
>
> I hope this helps,
> Scott
>
> *
> * 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/
The information contained in this document is intended only for the
addressee and is not necessarily the views nor the official
communication of the Department of Labour. All final/official papers
which are sent from the Department will be sent by non-electronic
means, on appropriate letterhead, signed by authorised personnel.
*
* 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/
