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

RE: st: Hausman and robustification?


From   "Salvati, Jean" <JSalvati@imf.org>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: Hausman and robustification?
Date   Thu, 20 Oct 2005 16:59:11 -0400

I wrote an implementation of the regression-based test described on
pages 290-291 of Wooldridge (2002).

Basically, the program estimates equation (10.79) on page 290, and then
tests the hypothesis that the coefficients of the de-meaned variables
(or a subset of the de-meaned variables) are null. This is the procedure
suggested by Wooldridge on page 291, paragraph 2.

I haven't read the post by Vince Wiggins yest, so I don't know we
implemented the same test.

If you would like to try the ado, send me an email at jsalvati@imf.org.
However, I must point out that I cannot guarantee that I will respond to
support requests.
 
Jean Salvati

Reference:

Jeffrey M. Wooldridge (2002), "Econometric Analysis of Cross Section and
Panel Data", MIT Press.


> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Joana Quina
> Sent: Thursday, October 20, 2005 4:25 PM
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: Hausman and robustification?
> 
> Aurora,
> 
> I had a similar problem a while ago.  Vince Wiggins provided 
> a very useful suggestion - an asymptotically equivalent test 
> for random vs.
> fixed effects using an augmented regression (Hausman and 
> Taylor, 1981).
> You can find the code in this post:
> http://www.stata.com/statalist/archive/2005-08/msg00853.html
> 
> I suggest you read the thread of this post and any references therein.
> 
> I hope this helps,
> 
> Best regards,
> Joana
> 
> On 20/10/05, Aurora Mordonu <Aurora.Mordonu@ugent.be> wrote:
> > Hi,
> >
> >
> >
> > Using a Hausman test, I would like to decide whether I 
> should take a 
> > fixed effect or a random effect model. Correcting for 
> > hereroskedasticity (I take fe robust and re robust) implies 
> that the 
> > Hausman test fails to meet the asymptotic conditions. So I 
> get the following message:
> >
> > model fitted on these
> >
> >                                        data fails to meet the 
> > asymptotic
> >
> >                                        assumptions of the Hausman 
> > test;
> >
> >                                        see suest for a generalized 
> > test
> >
> >
> >
> >
> >
> >
> >
> > Consequently, I try to use suest, but I get the following message:
> >
> > unable to generate scores for model fixed
> >
> > suest requires that predict allow the score option
> >
> > r(322);
> >
> > Could anybody tell me how to decide on a FE or a RE model in this 
> > case? This happens only after I correct for heteroskedasticity.
> >
> > Any hint would be very appreciated!!
> >
> >
> >
> > Thank you,
> >
> > Aurora
> >
> > --
> >
> > *
> > *   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/
> 

*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index