A problem about sureg

[Mindy Green]

Dear all:

My dependent variables are wives' and husbands' well being. When I regard wives' well being as dependent variable, husbands' well being will be regarded as one of the explanatory variables, should I use seemingly unrelated regression?

Best wish

[Carlo Lazzaro]

Zhuomin:
yes. See Example 1, -sureg- entry, Stata .pdf manual.

[Mindy Green]

Dear Carlo Lazzaro ,

Thanks so much for your answer. I just read the "Example 1, -sureg- entry, Stata .pdf manual.", it's about two equations with the same explanatory variables, but without the inclusion of each endogenous variable in the right side of the equation of the other. And I just read the paper Godwin, D. D. (1985). Simultaneous equations methods in family research. Journal of Marriage and the Family, 9-22. According to this paper, it seems that using 3SLS is better when I include each dependent variable in the right side of the equation of the other, and using sureg is better when there is the identity relationship between two dependent variables. So now I am hesitate which one to use. The equations are as follows:

Best wishes,
Zhuomin

[Carlo Lazzaro]

Zhuomin:
thanks for clarifying.
I was not aware of the endogeneity issue: 3SLS sounds better.

[Mindy Green]

Dear Carlo Lazzaro ,

Thanks for your help, it's really useful!

Best wishes,
Zhuomin