st: question about two stage least square and ordered probit model

[Xuan Zhang <xuanzhan@buffalo.edu>]

Dear everyone,

I have a question about the 2SLS. My model is like below:

y=X'β1+γM+u1

and M is the IV that I will estimated from equation:

M=X'b+u2

However the M variable is a dummy variable which could be 1 2 3 4. So
in the estimation of M, I plan to use a ordered probit model. Then
substitutes the estimation into the first equation to get β1. I'm not
sure how could I deal with the problem in STATA.

I read some materials regarding this issue. I suppose I should use
smm. But I'm not sure whether I'm in the right direction.

My second question is: if I could solve the problem, could I apply it
into panel data. I'm now using a cross section data.

Thanks a lot.

Best,

Xuan Zhang

Department of Economics, SUNY Buffalo

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/