Re: st: Testing joint significance of fixed effects in presence of heteroskedasticity and auto-correlation

[Christopher Baum <kit.baum@bc.edu>]

<>
On Jun 9, 2010, at 2:33 AM, Christian wrote:

> using  - xtreg y x, fe - we obtain the p-value of the joint
> significance of firm-specific fixed effects from the common output (F
> test that all u_i=0).
> 
> In order to correct for heteroskedasticity and auto-correlation, we
> would like to use -xtreg y x, fe vce(cluster clusterid) - to receive
> robust standard errors. The F-test mentioned above is not calculated
> anymore when this option is applied. Stata help files indicate that
> "The F test of u_i = 0 is suppressed because it is too difficult to
> compute the robust form of the statistic when there are more than a
> few panels." (xt p. 452)
> 
> Kit Baum proposed a test for the joint significance of fixed effects
> based on estimates robust to heteroscedasticity, but not
> autocorrelation (
> http://www.stata.com/statalist/archive/2005-05/msg00373.html ).
> 
> Is there any way to test the joint significance of fixed effects in
> the presence of both heteroscedasticity AND autocorrelation?

The inclusion of -robust- in the quoted example was a red herring. You will get the same results without
-robust-, and you would get the same results with a HAC estimator in the first part (which you could compute
with Mark Schaffer's -xtivreg2-). If only residuals are involved in the formula, consistency of the point estimates
makes this computable without any reference to the VCE.

Kit



Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                              An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.html
   An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/