Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: atypical (?) 2sls question


From   "Gio Reidi" <gioreidi@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: atypical (?) 2sls question
Date   Tue, 29 May 2007 21:32:21 +0100

Thanks a lot Austin,
indeed I found other treads speaking about the forbidden regression I
couldnt find earlier.
Thanks for having the patience to direct me to this problem.

On 5/29/07, Austin Nichols <austinnichols@gmail.com> wrote:
Diego / Gio Reidi --
This is called the forbidden regression, and for good reason.  You
cannot plug in the fitted value. Instead, you have two endogenous
variables now, X and Ins*X, and you may need additional excluded
instruments.  You may want to consider interacting excluded
instruments... but be sure to check the diagnostic tests provided by
the newest copy of -ivreg2- (ssc install ivreg2, replace).

On 5/29/07, Gio Reidi <gioreidi@gmail.com> wrote:
> Dear Statalisters,
> sorry if this will reach you twice..
>
> I hope you could give me an insight about this problem, concerning
> econometrical theory and Stata commands for 2sls.
>
> I am trying to run a sort of two stages least squares regressing the
> impact of some institutions on how variable Y responds to variable X
> (endogenous):
>
> Y=a+B(X)+Z(Institution*X)+K
> (exogenous variables, including main effect of the institution)
>
> To do so I want to instrument X, for which I have my good
> instrument(s). I do not have an instrument for the interactions
> between
>  X and the institution, but I was thinking (and  maybe I am completely
> wrong) that I would not need one if I just plug the fitted values of X
> from the first stage regression and construct with them the
> interaction term.  At the end, this is  what I want to assess: the
> impact on the "purified" X of the institutions. The goal is to obtain:
>
> Y=a+B(fittedX)+Z(Institution
> *fittedX)+K(exogenous variables).
>
> I cannot do this using ivreg or ivreg2 because then they would
> consider the interaction as an exogenous variable and insert it in the
> first stage. So I was thinking to do two stage least squares by
> hand,including all the exogenous regressors in the first stage plus
> the additional instruments. and then plugging in the fitted X series
> and the fitted X sinteracted with the institutions.
*
*   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/



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