Dann's point is basically that when you do single-equation IV, you don't
need to make any assumptions about the appropriate functional form for the
equations in the system that you aren't estimating (in your case, the
treatment equation aka the probit). The endogenous dummy variable that you
are instrumenting can be treated just like any other endogenous variable,
and everything about IV goes through as normal, including the overid stat.
That said, what you gain in robustness by using single-equation IV, you
lose in efficiency by not estimating the system by treatreg (assuming that
the probit is the right functional form for the other equation in the
system). If treatreg will give you consistent estimates of the outcome
equation, then so will IV (because IV is consistent under a wider range of
assumptions than treatreg), but treatreg will be more efficient because it
uses information about the rest of the system that single-equation IV
ignores.
It would be nice to have an overid stat for treatreg ... and in fact I
thought I had programmed one and I was getting ready to offer it to Ron,
but I've just tested it again and I'm not sure it's right. If I manage to
get it working I'll get in touch.
--Mark
Quoting Ron Venezie <[email protected]>:
> Dann: Many thanks for the quick response. I'll take a look at this
>
> article as soon as I can get it. It turns out that our library at
> UNC
> doesn't carry this journal, but I'll explore some other option.
> Regards,
> Ron.
>
> --On Friday, January 03, 2003 2:47 PM -0600 DL Millimet
> <[email protected]> wrote:rrr
> > if you look at the special issue of Jrnl of Eco Perspectives that
> has
> > about 10 articles on various econometric models from a year or so
> ago,
> > there is one on IV methods by Angrist and Kreuger. In there, they
> state
> > that for cases of endog. dummy variable, one is better off
> estimating the
> > first stage as a linear probability model, assuming my memory
> recalls
> > this correctly, since the second-stage is consistent no matter
> what,
> > whereas if the first-stage is a probit/logit, the second-stage is
> only
> > consistent if the probit/logit model is the "correct" model. if i
> have
> > this right, this may mean you are fine using the ivreg2 command,
> and its
> > overid test.
> >
> > am i right on this?
> >
> > dann
> >
> >
> >
> >
> > -----Original Message-----
> > From: [email protected]
> > [mailto:[email protected]]On Behalf Of Ron
> Venezie
> > Sent: Friday, January 03, 2003 1:55 PM
> > To: [email protected]
> > Subject: st: diagnostics for the treatreg procedure
> >
> >
> > I am evaluating a preventive program using the treatreg procedure
> in Stata
> > 7.0. My outcome variable is continuous, and participation in
> the
> > preventive program is indicated by an endogenous dummy variable.
> Is
> > anyone aware of a way to test for overidentifying restrictions
> after using
> > the treatreg procedure? My search of the Stata manual and FAQ
> list turned
> > up nothing.
> >
> > Ron Venezie
> > Post-doctoral Fellow
> > Sheps Center for Health Services Research
> > UNC-Chapel Hill
> >
> > *
> > * 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/
> >
> > *
> > * 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/
> >
>
>
>
> *
> * 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/
>
Prof. Mark Schaffer
Director, CERT
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS
tel +44-131-451-3494 / fax +44-131-451-3008
email: [email protected]
web: http://www.sml.hw.ac.uk/ecomes
________________________________________________________________
DISCLAIMER:
This e-mail and any files transmitted with it are confidential
and intended solely for the use of the individual or entity to
whom it is addressed. If you are not the intended recipient
you are prohibited from using any of the information contained
in this e-mail. In such a case, please destroy all copies in
your possession and notify the sender by reply e-mail. Heriot
Watt University does not accept liability or responsibility
for changes made to this e-mail after it was sent, or for
viruses transmitted through this e-mail. Opinions, comments,
conclusions and other information in this e-mail that do not
relate to the official business of Heriot Watt University are
not endorsed by it.
________________________________________________________________
*
* 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/
