st: Re: multiple eqn GMM

[Kit Baum <baum@bc.edu>]

You cannot assign instruments to regressors in any form of instrumental 
variables estimation. If you were to estimate these equations 
separately, X3 and X4 would also be instruments for the first equation, 
and X1 and X2 would also be instruments for the second equation. 
Presumably there are structural equations for each of the X's somewhere 
in your system which render them endogenous.

 If a systems estimator is to be used, you list all of the instruments 
(Z1-Z5 plus X1-X4) and specify the two equations as you have done. You 
can use reg3 with constraints to estimate this system -- it will not be 
full GMM, but it will allow you to do the systems estimation and apply 
the cross--equation constraints. At present I am not aware of a full 
GMM systems estimator for Stata.

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html

On Jan 11, 2005, at 2:33 AM, Gordon wrote:


> I wonder how to use STATA to estimate a multiple equation GMM like the
> following:
>
>
> Equation 1: Y1 = a1*X1+a2*X2. I have Z1,  Z2 and Z5 as instruments for 
> X1 and X2
>
> Equation 2: Y2 = a1*X3+a2*X4. I have Z3 and Z4 as instruments for X3 
> and X4.
*
*   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/