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

st: Re: testing for cross-sectional independence


From   Christopher F Baum <baum@bc.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: Re: testing for cross-sectional independence
Date   Sat, 9 Oct 2004 09:33:30 -0400

On Oct 9, 2004, at 2:33, Levy wrote:


I have an unbalanced panel for several hundred firms. the maximium length
is 7 whereas the short is 1 for some of the firms. the mean length is 5.4.

I want to fit a fixed effect regression while one of my supervisor
prefers poolling regression. I don't know what kind of test can i use to
choose models between them

can u give me some suggestions?

....

i have tried xttest2, but it does not work, i will try xttest0 later, thanks
again


xttest2 is equivalent to the test that could be carried out after a -sureg- estimation. You cannot run -sureg- when N>T, and that is surely the case with your data. xttest2 tries to calculate the residual correlation matrix from your fixed effects model, and that matrix must be singular if there are more firms than observations per firm.

I do not see that a fixed effects regression is likely to be very successful in this context when you have so few observations (in some cases only one!) per firm. I would recommend using something like an industry fixed effect if you can argue on economic terms that firms in the same industry might share an 'industry effect', which you imagine will be more important than the within variation in an industry.

Kit Baum
Faculty Micro Resource Center [ fmrc@bc.edu ]
Boston College Academic Technology Services


*
* 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