Page 1 of 1
IV diagnostics regressor endogeneity
Posted: Tue Aug 06, 2013 6:50 pm
by k100z
After running an instrumental variable estimation for cross sections, one can check for the endogeneity of the regressors: View / IV Diagnostics & Tests / Regressor Endogeneity Test
It works fine when running a GMM estimation, but there are no results for the TSLS. According to the handbook it should be available for GMM and TSLS. It is executing the test, but test statistics are simply not displayed. the are no numbers for
Value df Probability
___________________
Re: IV diagnostics regressor endogeneity
Posted: Tue Aug 06, 2013 7:38 pm
by EViews Gareth
Could you provide your workfile, along with EViews version and build date (available from Help->About EViews).
Re: IV diagnostics regressor endogeneity
Posted: Wed Aug 07, 2013 12:12 pm
by k100z
Could you provide your workfile, along with EViews version and build date (available from Help->About EViews).
Hi,
Eviews 8, Aug 1 2013build, workfile attched.
Re: IV diagnostics regressor endogeneity
Posted: Wed Aug 07, 2013 12:56 pm
by EViews Gareth
Thanks.
Re: IV diagnostics regressor endogeneity
Posted: Tue Sep 03, 2013 12:44 pm
by k100z
Hi Gareth,
did you find any solution/explanation to my problem?
Re: IV diagnostics regressor endogeneity
Posted: Tue Sep 03, 2013 12:49 pm
by EViews Gareth
I believe it is fixed in the latest patch.
Re: IV diagnostics regressor endogeneity
Posted: Tue Sep 03, 2013 12:53 pm
by EViews Gareth
Hmmm, I just checked the latest patch, and actually it isn't fixed.
We will investigate.
Of course, in the mean time, you can always calculate the statistic manually - the secondary estimation output is shown, you just need to find the difference between the original J statistic and the secondary estimation's.