Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: IV conceptually


From   Christopher Baum <baum@bc.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: Re: IV conceptually
Date   Tue, 15 Apr 2003 07:11:13 -0400

On Tuesday, April 15, 2003, at 02:33 AM, Liliana wrote:

I wonder if there is a way in stata to perform instrumental
variables/2SLS
with three endogenous variables. I was told that the only way is
doing it
with two at a time. Is this true?

Doing a manual search I found the following command:

ivreg y1 x1 x2 (y2 y3= z1 z2 z3)

that this imply that in order to run this model (with 3 endogenous)
the
"z" variables have to be the same for the two endogenous y2 and y3
There is no restriction on the number of endogenous variables in ivreg. One might say

ivreg y1 (y2 y3 y4 y5 y6 y7 y8 y9 y10 = list of instruments) x1 x2 x3 ...

subject to the constraint of the order condition: that the "list of instruments" is at least as long as the number of R.H. endogenous variables. It will then be the case that all of those instruments (as well as x1 x2 x3...) will be used in the 'first stage' regressions -- and this is as it should be. IV from a theoretical standpoint does not allow one to choose which instruments appear in which first stage regressions -- a point often made on Statalist. This is not a Stata-imposed constraint; it is imposed by the underlying econometric theory. The textbook descriptions which suggest 'this is an instrument for that' are quite misleading in that regard.

Thanks
Kit

*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index