Fatma,
although I can just guess at your intended application (I presume you are
interested in things like the wage elasticity of the decision to become self-
employed vs. dependent employee), my bet is that you should be taking a
look at Stata's heckprob command instead.
Depending on your application, you could have a classical sample selection
problem where Y2 = self-employment vs . dep. employment is only observed
when Y1 = labour force participation == 1. This type of model is fit by
heckprob, which is Stata's discrete equivalent of the standard Heckman
sample selection model. Suprob or biprob are not really appropriate in the
above setting because you do not observe BOTH outcomes for all persons in
your sample (as an aside, I also think that the partial observability model is
not at all relevant because you are observing both steps of the decision tree).
So my guess is that you might be experiencing troubles with convergence
because Stata (correctly) drops observations that have missing values on
either Y2 or some variables in your X2 vector intending to predict self-
employment vs. dependent employment (e.g. actual wage measures are
observed only for Y1==1). So you might end up with a suprob/biprob
estimation sample where lfp==1 for (almost) everybody (and maybe there are
some coding errors to make the estimation work at all).
However, much of what I've said depends on the sort of question and data
you are actually working on - and I might simply not be guessing correctly
what you have done. Hope this helps nevertheless.
Markus
> Date: Mon, 12 Jul 2004 03:54:59 +0300
> From: fatma bircan <[email protected]>
> Subject: st: biprob/suprob HELP
>
> Fatma Bircan
> Middle East Technical University
> Department of Economics
> 210-2032
>
>
> Hi,
> Could anyone tell me the difference b/w suprob and biprob commands? I
> tried both commands but I dont get any results. It seems that
> loglikelihood does not become concave. why. I specified a labor force
> participation equation and a choice model b/w self-employment and
> wage-employment. I expect it to converge, to fit quite OK. I also use
> partial option since the positive outcome in my second equation is
> possible only when the first equation is positive, that is lfp=1, Why
> dont I get the results. Please help
>
> Thank you
>
>
>
> *
> * 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/
>
--
Dr Markus Gangl
Social Science Research Centre Berlin (WZB)
Reichpietschufer 50 Tel: +49.30.25491.141
D-10785 Berlin Fax: +49.30.25491.222
Germany Email [email protected]
*
* 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/
