st: Cnreg with fixed effect

[Milet Emmanuel <emmanuel.milet@cepii.fr>]

Dear Stata users,

I intend to run a Cnreg estimation with fixed effect. The routine on 
Stata does not include any fixed effect option, so I add them one by one.
I have a panel with many dimensions (year*country*firm*product) and my 
censoring variable is year*country*product specific.

I estimate exports (at the year*country*firm*product level) using GDP, 
Distance, Population, etc...

I replace all the zero flows by the smallest amount exported to a givern 
country for a given product a given year. So in the end, my export 
variable is both firm-specific and has some observations that are 
year*country*nef specific.

My problem is that the procedure takes too much time to run (when I 
include all the fixed effects, there are almost 8000).
I am looking for a way to transform my variables in order to get rid of 
the fixed effect (like substracting the mean if I were running an xtreg 
with 2 dimensions).
This cannot work here however, since my censoring variable is not 
firm-specific.


E.Milet
University Paris 1
Paris, France
*
*   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/