st: Simultaneous Heckman Method

[Sakib Mahmud <smahmud@uwyo.edu>]

Dear All, 

I am trying to estimate a private disaster risk mitigation model where a household has two risk-reduction strategies: averting (A) and mitigating (M). Based on my household survey data, I have questions regarding whether a household participates in either A or M. If 'yes,' then the follow up questions are how much each household spend on either A or M. 

However, it is likely that there is an endogeneity problem since the alternative risk management actions taken by the household are unlikely to be independent. That is, the choice of A is unlikely to be independent of the choice of M and vice-versa. 

Hence, I have two simultaneous bivariate probit model for participation in A and M followed by two simultaneous expenditure equations on A and M. That is, 
ProbA (probability to participate in A) = a0 + a1M + a2 X1 +a3X2 + a4X4+u1 
ProbM (probability to participate in M) =b0+ b1A + b2X1 + b3X5 + b4X6+u2
ExpendA (expenditures on A) = c0 + c1ExpendM + c2X1 + c3X2+c4X4+u3
ExpendM (expenditures on M)= d0 + d1ExpendA + d2X1 + d3X5 + d4X6 + u4 
 
My concern is whether a simultaneous-heckman method can be performed under stata. 

I know that there is a -heckman- command in stata to perform the two-step heckman or heckit method. But I am not sure how the -heckman- command can be performed when you are dealing with simultaneous equations. Since I have a simultaneous equations problem with four equations, I am bit concerned how to approach my econometric problem with stata. 

Any assistance will be appreciated. 

Thanks,
Sakib 
PhD Student, University of Wyoming 


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