endogeneity tests

[uctum]

Hello. I am an Eviews 7 user. I was trying to perform a regressor endogeneity test following instructions in the Help menu: "View/IV Diagnostics and Tests/Regressor Endogeneity Test". I cannot find IV Diagnostics and Tests under View. I checked View under both Equation window and Workfile window. I would appreciate any help with this.

[EViews Gareth]

That view will only appear under Equations that have been estimated using instrumental variable methods (TSLS, GMM and LIML).

[startz]

This may be a foolish question, but did you use the TSLS command to do the estimation?

[lilc]

In connection with this topic, i would like to know how to interprete the results of the endogeneity test? I read the notes from the eviews help guide but i couldn't really understand. The test explains that if the selected exogenous variables are suspected to be endogenous it will be explained by the set of instruments. However, what the endogeneity test does is it adds the "endogenous" variables into the GMM model and see if the difference in the J-stat is significant or not. My understanding is that if the difference in J-stat is not significant, there is endogeneity as the additional variables do not help to explain the model better. However, H0 states the candidate variables are exogenous. How do I read the results? Compare with the Chi-square table of 6 df @ 5% is 12.59. So does it mean that the difference is insignificant and we fail to reject H0?

Value df Probability
Difference in J-stats 1.494689 6 0.9598