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

st: Re:autocorrelation and panel data


From   Kit Baum <kitbaum@mac.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Re:autocorrelation and panel data
Date   Fri, 19 Mar 2004 21:01:01 -0500

On Mar 19, 2004, at 2:33 AM, Cordula wrote:

I am doing a fixed effects regression and I'm now trying to test for
autocorrelation. I tried the xttest2, but get the error message that the
correlation matrix is singluar. I found out from previous entries that
that's due to the fact that my N(=198) & my T(=8) are not equal. I've then
done the xtregar, fe regression, but how do I test whether the rho_ar are
significantly different from zero, as suggested by Kit Baum?
The test in xttest2 works like sureg: one must have N<T for the correlation matrix of the errors to have full rank.


Moreover, I've then done the same regression with "reg" using a dummy
variable estimator for the fixed effects. I know that I can then use
panelauto to test for autocorrelation, but I always get the error message
that the sample may not include multiple panels. Could anyone tell me what
the correct "if" command is to do those tests from panelauto?
panelauto is a workaround for those official Stata commands that will not work AT ALL if the data are tsset as a panel. The commands in panelauto will work on a single unit of a panel; thus there is no 'if' involving multiple units that will permit these tests to work.

Kit

*
* 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