Dear all
I'm encountering a statistical problem about tests for the endogeneity in panel data. Here I found two methods: for the first is xtivreg2, it states the sargan test results which implies the instruments are useful if it rejects the null hypothesis; for the second is dmexogxt, it tests for the exogeneity and the rejection of null hypothesis implies the instruments are useless.
The question is that I got inconsistent results in these two tests, more specifically, the sargan test and exogeneity test are both non-reject, how can I judge the endogeneity? Thanks anyway.