Erasmo is having problems posting to Statalist. While he gets this
sorted out, I am posting his email to the list plus some responses of my
own. Others should feel free to comment.
--Mark
> From: Giambona, E. [mailto:[email protected]]
> Sent: Friday, April 20, 2007 9:54 AM
> To: Schaffer, Mark E
> Subject: RE: hausman's test with xtivreg2
>
> Dear Statalist users,
>
> I am using the endog option with xtivreg2. I get a chi-2(6) = 6.9 with
a
> p-value of 0.3336 but I am not sure what the null hypothesis is.
> Does it mean that the variables are endogenous? Is there a reference
for
> this test?
The null is that you can treat the variables as exogenous.
Plenty of references for this standard test, starting with the xtivreg2
and ivreg2 help files and the citations therein.
> Also, is there a way to get the constant term when using xtivreg2?
Not directly, but you could use official xtivreg to do it.
> Finally, I have come across a reference claiming that the Hausman test
> is not appropriate when the regression includes firm-level fixed
effects.
> Does endog make any adjustment for such problem?
Not sure what you mean by "Hausman test" in this context. Do you mean
endogeneity test or fixed-vs-random effects?
In any case, I'm not aware of any particular problem with either.
Perhaps you could cite the full reference with an explanation of what
the "problem" is?
--Mark
>
> Any help and suggestions would be truly appreciated.
>
> Thanks,
>
> Erasmo Giamona
*
* 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/
