st: re: 2sls with multiple endogenous variables

[Kit Baum <baum@bc.edu>]

Ana replied:

I would like to ask for your advice regarding the
following model:

poverty=hhchar,farmchar,prod,mkt
prod=hhchar,farmchar,irrig,hhcomp
mkt=hhchar,farmchar,transport,road,ethnic

Would it be okay to estimate the model using SUR?
Could you please suggest a way to estimate the model?



NO. SUR is a systems estimator for systems of equations which each  
may be consistently estimated by OLS.

You have a recursive system here, in that both prod and mkt depend  
only on exogenous variables. You should do

ivreg2 poverty hhchar farmchar (prod mkt = irrig hhcomp transport  
road ethnic)

to estimate this model.

You could probably use -reg3- as well to impose the zero restrictions  
on instruments. I'm not sure that would make much difference, though,  
and Stata's -reg3- is much less capable than -ivreg2- (or official - 
ivregress-) in being able to deal with non-iid errors (e.g. it cannot  
even estimate using -robust-, let alone cluster() or HAC). I would  
worry about that in terms of the reliability of interval estimates  
and inference.


Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


*
*   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/