I feel like jumping in to the conversation although the original
question is not posted by me.
My question is: If one wants to report the first stage estimations of
a models as the one described before. Should all instruments be used
in the regression predicting x1 and again in x2? That is, should there
be two first stages, with different dependent variable (x1 and x2) but
the same regressors?
Tinna
On 12/1/05, Nick Cox <[email protected]> wrote:
> What I know of econometrics is very limited, but this statement of Kit
> Baum appears relevant:
>
> Instrumental variables involves a matrix of regressors X
> and a matrix of instruments Z. All variables considered
> exogenous (or predetermined) are in Z. Some may also be
> in X.
>
> There is no one-to-one correspondence between regressors
> and instruments. All elements of Z are being used to
> generate the 'X-hat' matrix; that is, each column of X
> is regressed on the elements of Z. This is not a whim of
> Stata's design; it is inherent in the notion of an IV estimator.
>
> Nick
> [email protected]
>
> Andrei Simonov
>
> I wonder if you can help me. I need to estimate quite simple 2SLS reg, say
> Y=a+b1*x1 +b2*x2 +exog. Vars +e.
>
> The problem is that I want to specify different lists of instruments for x1 and x2. Is there any way to deal with 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/
>
*
* 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/
