st: r(3900) when doing xtabond2

[donaubaj@hsu-hh.de]

Dear statalist users,

I started using Stata only about one week ago, so I am not really an  
expert, but I try to be as specific about my problem as possible:
I have a panel with 21 countries and 8 years, including 7 variables  
plus time and cross section dummies. The panel is almost perfectly  
balanced, only few observations are missing. I am using a Stata  
version with 32bit on a laptop with 64 bit Windows and at a PC with 32  
bit Windows.
First, I tried to re-estimate the output that I produced with Eviews  
(mainly panel estimations with cross section and/or period fixed  
effects). This worked quite will, I managed to get exactly the same  
output as I did in Eviews. But when I try to estimate the same model  
with xtabond2 I am facing huge problems. Usually, I get the following  
error message:
. xtabond2 LHS-variable all RHS-variables year*, gmm(list of  
endogenous RHS variables) iv(year*)
Favoring space over speed. To switch, type or click on mata: mata set  
matafavor speed, perm.                      J():  3900  unable to  
allocate real <tmp>[268485,76710]
              _Explode():     -  function returned error
           _ParseInsts():     -  function returned error
         xtabond2_mata():     -  function returned error
                 <istmt>:     -  function returned error
r(3900);


Even xtabond does not work. Doing this, Stata returns "no  
observations" (r(2000)).
Unfortunately, I have no clue what could be wrong. I tried several  
options, including dropping several variables from the model. I saw  
that other users face similar problems. However, I could not find a  
viable solution.

Any help would be greatly appreciated.

Thanks!

Julian Donaubauer

Helmut Schmidt University
Hamburg


*
*   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/