IV estimators

[chou...@msu.edu]

 Hey guys
does anyone knows in what condition will IV estimators be unbiased?

thanks
Annie

[Doosoo Kim]

E[u|X,Z]=0

is one condition possible :D

Annie

--
You received this message because you are subscribed to the Google Groups "MSU econ 2012 study group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to msu-econ-2012-stud...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[riju joshi]

If Z is independent of u, is the IV/2SLS estimator unbiased?

[chou...@msu.edu]

Appartently I forgot to use group reply....
Below are quoted from dear Doosoo


For unbiasedness of IV, we need E[u|X,Z]=0.

none of them make IV unbiased in general since none of them implies E[u|X,Z]=0.

The condition i) doesn't give any information about joint distribution of u,X,Z. it only implies E[u|Z]=0.

The condition ii) implies E[u|X,Z]=E[u|X]. But we are suspecting E[u|X] is not zero since u and X are possibly correlated.

 

One possible condition is "E[u|X,Z]=0" as I mentioned earlier.

The other one is "given Z, u and X are independent & E[u|Z]=0

Then we have E[u|Z,X]=E[u|Z]=0

 

Best,

Doosoo

[riju joshi]

Thanks!