m
MIME-Version: 1.0
Content-Type: text/plain; charset=3DISO-8859-7; format=3Dflowed
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from base64 to 8bit by hsphsun2.harvard.edu id l29Ekr=
J01440
Sender: [email protected]
Precedence: bulk
Reply-To: [email protected]
Errors-To: [email protected]
Hi,
I have panel data, and I am trying to estimate a model with
simultaneous equations similar to that in Altonji et al (2005) (but I
am using a linear framework while they use probit):
CH =3D X=E2 +u
Y =3D X=E3 + =E1CH + =E5
where the correlation between u and epsilon, p, is constrained to be:
p=3DCOV(X=E2, X=E3)/Var(X=E3)
I was considering trying to use reg3, but it looks like the
constraints for reg3 are constraints on the coefficients rather than
the error terms. Is there a way to impose this constraint on the
error terms? Also, is there a reg3 that is meant for estimation with
panel data?
I would really appreciate any ideas.
Thanks,
Molly
*
* 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/
