I am trying to take into account unobserved heterogeneity in my Kenyan data that links four equations. In the first two equations (which are solved in reduced form), I use the Poisson distribution to determine the log odds of obtaining a birth and a death respectively. So I have for these two:
Eq 1:
xi: Poisson i.births i.agegroups i.Faith i.Husbeduc i.Womaneduc i.region etc. + unobserved heterogeneity
Eq 2:
xi: Poisson i.deaths i.agegroups i.Faith i.Husbeduc i.Womaneduc i.region etc. + unobserved heterogeneity
As I said, these two are in reduced form.
Then there are the other two - the fertility intentions and contraceptive use equations:
xi: mlogit i.fertintentions livingchildren i. landownership i.education opportunities etc. + unobserved heterogeneity
Equation 4.
xi: mlogit i.methodchoice fertintentions i.agegroups i.Faith etc + unobserved heterogeneity
These last two equations are endogenous (fertility intentions and contraceptive method choice are determined in the same system).
Now my question is: how do I solve (obtain consistent parameter estimates) for the four equations simultaneously?
Your suggestions are highly appreciated.
David Ojakaa,
D�partement de D�mographie,
Universit� de Montr�al,
2350 Blvd. �douard-Montpetit,
Montr�al (Qu�bec),
H3T 1J4 CANADA.
Tel. 514-343-6532 poste 44696
e-mail: [email protected] <mailto:[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/