On Mon, 19 Jan 2004, Si Li wrote:
> Basically, the model is as follows:
>
> Y1_star=X*b1+e1
> Y2_star=X*b2+e2
>
> Y1=1 if Y1_star>=0 and Y2=1 if Y2_star>=0. Let Zi=Y1i*Y2i. We have data
on
> Zi=0, 1. But no info about Y1* or Y2*. Prob(Zi=1)=Prob(Y1=1 and
> Y2=1)=F(X*b1,X*b2,rho). log-likelihood
> fn=sum[Zi*prob(Zi=1)+(1-Zi)*prob(Zi=0)].
>
> This model is similar to a bivariate probit in spirit. The difference is
that
> it puts all the observations of (0 1), (1 0), and (0 0) into Zi=0.
>
> Do you know how I can use Stata to estimate the above model? Your
response
> will be highly appreciated.
>
>
Dear Si,
The -biprobit- command has an option called "partial" to estimate
Poirier's partially observable model.
--Brian
[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/
