xtivreg2: endogeneity and overidentification

[Truong Quoc Phan]

Hello everyone,
I have an issue with interpret the result of xtivreg2:
- Overidentification is significant
- Endogeneity test is not significant

Please see the attached picture.
I don't know how I should interpret this result. The endogeneity test indicates that there is no endogeneity issue. This implies no need for the I.V. However, the overidentification test suggests the i.v. is a valid one?
Thank you

[Sebastian Kripfganz]

Your model is exactly identified - you have as many instruments as regressors. The overidentifying restrictions test is only applicable if your model (as the name suggests) is overidentified - you need more instruments as regressors for this test.

Aside from this: Even if all regressors are exogenous, your instruments could still be either valid or invalid (i.e. uncorrelated or correlated with the error term).