Austin,
thank you for your help and forgive me for my late reply.
I am still comfused. What I took in the previous posting is an endogenous example arising from simultaneity, namely, to be cause and result by each other. I can find instruments for endogenous variables easily if the two equations are determined. With regards to endogeneity arising from omitted variables, how to look for instruments variable(s)?
For example,
if my model is
y1=cons+y2+z1+u1,
and I know y2 is an exdogenous independent variable due to omitted variables. So I need to look for instrument(s) for y2. Based on previous scholars' research, I get the following equation,namely, the determinants of y2.
y2=cons+z2+z3+z4+z5+...+z10+u2 (no y1 in the independent variables,which means not simultaneity)
I know as an instrument z, it should be satisfied with two conditions, cov(z,u1)=0 and cov(z,y2)!=0.
Based on the above information, does it mean the instrument(s) for y2 should be z2 or z3 or z4 or z5 or the combination of them? The first condition is unobservable and due to z2~z5 are determinants of y2, which means the second condition is satisfied.
In a word, how to find instruments for endogenous variable arising from omitted variables?
Thank you for any help!
Best regards,
Rose.
----- Original Message -----
From: Austin Nichols <[email protected]>
To: [email protected]
Subject: Re: st: endogeneity doubt about simultaneity and omitted variables
Date: 2009-9-13 20:52:11
Yes, e.g.
ssc inst ivreg2
ivreg2 y1 z1 (y2=z2 z3 z4 z5 z10), first liml
2009/9/12 <[email protected]>:
> y1=cons+y2+z1+u1
>
> y2=cons+y1+z2+z3+z4+z5+...+z10+u2
>
> For equation y1, if the endogeneity of y2 is not from simultaneity, but from omitted variables, how to solve it ?
>
> Can I still take z2 z3 z4 z5 ...z10 as iv instruments for y2?
*
* 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/
