Re: st: RE: instruments in ivreg2

[Bidisha Mandal <mandal.7@osu.edu>]

Steve,

Many thanks! I went over the literature, and have realized my mistake in 
interpreting the workings of 2SLS. Now I have another question. Is it 
alright to use time invariant variables as valid excluded instruments 
for time varying endogenous variables? For example, will it be ok to use 
parents' marital status as excluded instruments for the /change/ in 
individuals' marital status? My equations look somewhat as follows (d = 
delta):

d(mental health) = a1 + d(marital status) + d(income) + e1
d(marital status) = a2 + d(mental health) + parents' marital status + e2

Thanks,

Bidisha Mandal

Dept. of AED Economics
The Ohio State University  


Steven Stillman wrote:

> Bidisha,
>
> I think you are confusing the tests available in ivreg2.  The 'endog' option
> test whether one or more of your included RHS variables are exogenous given
> that you have a valid excluded instrument.  So, you seem to be indicating
> that you are using state employment rates an excluded instrument with labour
> market status as your suspect endogenous variable and that you are then
> finding that labour market status is, in fact, endogenous.
>
> It is not possible to test whether state employment rates are a valid
> instrument unless you have more than one potentially valid instruments, in
> which case you can then run an overidentification test of the joint validity
> of the excluded instruments.
>
> This is all explained in great detail in Christopher F Baum & Mark E.
> Schaffer & Steven Stillman, 2003. "Instrumental variables and GMM:
> Estimation and testing," Stata Journal, StataCorp LP, vol. 3(1), pages 1-31,
> March which is also available as Boston College Working Papers in Economics
> 545.
>
> Cheers,
> Steve
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of Bidisha Mandal
> Sent: Wednesday, October 18, 2006 6:13 AM
> To: statalist@hsphsun2.harvard.edu
> Subject: st: instruments in ivreg2
>
>
> Hi,
>
> I am working with panel data, and using ivreg2 for estimation purposes.
> My dependent variable is a measure of individual's mental health status.
> One of the independent variables is individual's labor market status,
> which could be endogenous. Using the 'endog' option I tested that it is
> in fact endogenous. I used state unemployment rate as a proxy, due to
> high correlation between this and individual's labor market status. But
> the J-statistic turns out to be significant. Seems unrealistic, how can
> my mental health affect state unemployment rate? Is this result because
> the test JOINTLY tests instruments are uncorrelated with the error AND
> correctly excluded from the estimated equation? Is there anyway to JUST
> test whether the instrument is appropriate?
>
> Thanks,
>
> --
> Bidisha Mandal
>
> Dept. of AED Economics
> The Ohio State University
>
> *
> *   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/
>
>   
*
*   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/