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

st: Re: Xttobit and completely determined panels.


From   "jyotsna puri" <pjyotsna@hotmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Re: Xttobit and completely determined panels.
Date   Wed, 27 Apr 2005 16:20:57 +0000

All,

I have an unbalanced panel dataset, and I am
running a tobit (random effects, panel i.e.
xttobit) for it.

I get, the line "two completely determined panels" at
the end of a run, with missing s.es etc. So I tabulated
the value of the dependent variable against the lone
independent variable (which is present along with a
time trend. xttobit is using (district) as the id variable.)
I could not identify anything as troubling.

Then I tabulated the means of the dependent variable,
against district. Sure enough, for 5 districts, the
standard deviation is zero for the dependent variable.

Do I have to take out observations for ALL five districts
from the xttobit? The specification runs fine, with no
warnings, if I take out only four of the districts (because
collinearity disappears?) One other question, if I take logs
of the dependent variable, with log(0) =0, this
warning never appears.

More importantly, why is this a cause for worry in tobit models
where one would expect many zeros.

Thanks!
-Jyotsna

Jyotsna (Jo) Puri
pjyotsna@hotmail.com


*
*   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