Hi,
I am trying to perform the Hausman test on an unbalanced panel but gets the
result that V_b-V_B is not positive definite. As a respone to this I try to
run the proposed a generalized Hausman test using the Suest command. See
below for the commands,.
regress var L.var yr*
estimates store OLS, title(The OLS estimator)
xtreg var L.var yr*, fe
estimates store WG, title(The Within Group estimator)
suest WG OLS
And as a response I get:
. suest WG OLS
WG was estimated with a non-standard vce (conventional)
r(322);
I am lost concerning this. Since I do not use the VCE option in xtreg it
implies that xtreg uses vce(conventional) and as far as I understand from
the reference book on the Suest command I am supposed to use conventional
vce in my estimation.
So why do I get this error and how can solve this problem? Has it to do with
the fact that I am running the analysis on an unbalanced panel?
Joachim Landstrom
*
* 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/
