Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: "Forbidden Regressions"


From   Erasmo Giambona <e.giambona@gmail.com>
To   statalist <statalist@hsphsun2.harvard.edu>
Subject   st: "Forbidden Regressions"
Date   Thu, 2 Apr 2009 14:31:44 +0200

Dear Statalist,

I have the following model: y = a +bX+cZ+d(X*Z) + fK + e. Both X and Z
are endogenous making X and Z also endogenous. K is exogenous. I am
fitting the model using IV methods. I instrument X and Z respectively
with X1 and Z1, and I use X1*Z1 to instrument X*Z.

I understand from a STATA thread
(http://www.stata.com/statalist/archive/2003-11/msg00795.html) that a
natural instrument for X*Z would also be the the product of the
predicted values of Xp and Zp from the first stage regressions (i.e.,
Xp*Zp).

Would the following also be plausible instruments for X*Z: X1*Xp,
Xp^2, (Xp*Zp)^2, etc..?

I hope to hear some comments on this issue.

Regards,

Erasmo
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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