re: st: 3 simultaneous equations

[Christopher Baum <kit.baum@bc.edu>]

<>
I am trying to estimate two-stage least squares simultaneous equations. I
have three endogenous variables Y1, Y2, and Y3 and one instrument for each
(X1,X2,X3)   

Y1 = a0 + a1*Y2 + a2*Y3+a3*X1 + a4*X4 + a5*X5 
Y2 = b0 + b1*Y1 + b2*Y3 + b3*X2 + b4*X4 + b4*X5 
Y3 = c0 + c1*Y1 + c2*Y2 + c3*X3 + c4*X4 + c4*X5

I tried to use xtivreg2 but I am not sure how it works with three equations.
I would really apprecite any hint.

Mark Schaffer's -xtivreg2- (SSC), like official -xtivreg-, is a single-equation estimator for panel data. That is not what you have here. You should use the -reg3- 
command if you are satisfied with the assumption of i.i.d. errors. As a recent posting indicated, you could also use the -gmm- command to estimate this sort
of system with robust or cluster-robust VCE. The -reg3- command cannot do that.

Each of the above equations is overidentified, so you can (and should) examine the overid test stats. Baum et al.'s -overid- on SSC will do that for -reg3-. 
The -gmm- command will produce a Hansen J statistic, which is a test of the overid restrictions.

Kit

Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                              An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.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/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/