st: how to account for jointly determined decisions

["Khieu, Hinh" <hdkhieu@usi.edu>]

Dear Statalist members,

I have the following model (and panel data):
Dividend = investment + financial constraint dummy (1=constrained, 0=unconstrained) + investment x financial constraint dummy + other control variables
The problem I have is that dividend and investment decisions are jointly or concurrently determined. What is the appropriate method to handle this scenario?
1/ I thought about using SUREG, but I am stuck with the interaction term which I don't think SUREG syntax would account for.
2/ A second idea I have is to split the equation above into Constrained and Unconstrained and run SUREG without the interaction term?
That is:

sureg (dividend and control variables if Constrained==1)(investment and different set of control variables if Constrained==1)
sureg (dividend and control variables if Constrained==0)(investment and different set of control variables if Constrained==0)

I have a feeling that it is not correct. The question I am trying to answer is whether Investment is negatively or positively correlated with Dividend when firms change from constrained to unconstrained status or vice versa. I am not sure the split above would answer my research question since investment is no longer in the dividend equation and dividend is no longer in the investment equation.

Any suggestions?

Many thanks in advance.

Regards,
Hinh

*
*   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/