Carlos,
I would just look these things up in an econometrics textbook (e.g.
Stock and Watson, Angrist and Pischke, Wooldridge, in increasing order
of difficulty), not ask the Stata mailing list about this. The latter
would tell you that all you need is a *correlation* between the
endogenous variable and the proposed instrument(s); the correlation
should not be too weak (otherwise you have a weak instruments problem)
and it should not violate the exclusion restriction (otherwise
instrumenting is pointless). Nothing about causality.
Davide
2009/3/16 Carlos Rodriguez <[email protected]>:
> Dear all
>
> Should the relationship between the proposed instrumental variable and
> the endogenous regressor for which it instruments be likely *causal*
> OR just a mere *correlation* for the IV to be a valid one?
>
> Second, if the relationship between the proposed instrument and the
> endogenous regressor has to be causal (or likely causal).
> would the IV only be valid if this relationship = "IV (likely) causes
> the endogenous regressor" or would the IV still be valid if the
> relationship is = "endogenous regressor likely causes IV"?
>
> How to ascertain this in Stata?
> thanks
> *
> * 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/
>
*
* 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/
