The problem is that "_se" is stored in the main tobit equation in
Stata versions prior version 9. This fools oaxaca. oaxaca assumes
"_se" be another covariate in the model and tries to compute its mean.
Note that oaxaca2 will not work even if you use Stata 9 because the
version is reset to 8.2 internally. I will change oaxaca2 so that it
does not reset version. In the meanwhile, estimate and store your
models under version 9 and then apply oaxaca to the stored estimates
(not oaxaca2). In your case:
version 9
tobit lasset_s sratio ... if sexo==0, ll
est sti sexo0
tobit lasset_s sratio ... if sexo==1, ll
est sti sexo1
oaxaca sexo0 sexo1
Hope this helps.
ben
On 11/21/06, Maria Cecilia Calderon <[email protected]> wrote:
Dear Ben and all,
I am trying to run an oaxaca decomposition using a Tobit
model for both groups. I obtain the model estimates, however, an error message
appears before the decomposition (variable _se not found). Below is my output.
Do you think you can help me?
Thanks a lot!
Cecilia
oaxaca2 lasset_s sratio cpiyear edumai edupai tierrapas_ss comun_d2 comun_d3
comun_d4, by(sexo) noisily cmd(tobit) cmdopts(ll)
[...]
variable _se not found
*
* 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/
