Hello everyone!

I have the following question:
I am working with instrumental variables. I perform the endogeneity test with the endog option
xtivreg2 TCAAPIBpc LNPib_i Var_Escol EV_i TGF_i Pobl_Rural (In_Ratio_20 = Gasto_PublS_i desemp_i), fe robust first endog(In_Ratio_20)

Endogeneity test of endogenous regressors: 4.117
Chi-sq(1) P-val = 0.0425
Therefore I could say that my variable is endogenous.. .

Then, I apply the rhausman test to compare the instrumental variables model with my model without IV (FE):

b1: obtained from xtivreg2 TCAAPIBpc LNPib_i Var_Escol EV_i TGF_i Pobl_Rural (In_Ratio_20 = Gasto_PublS_i desemp_i), fe robust first
b2: obtained from xtreg TCAAPIBpc LNPib_i In_Ratio_20 Var_Escol Pobl_Rural EV_i TGF_i, fe cluster()
Test: Ho: difference in coefficients not systematic
chi2(6) = (b1-b2)' * [V_bootstrapped(b1-b2)]^(-1) * (b1-b2)
= 3.77
Prob>chi2 = 0.7073

Since the probability is greater than 0.05 it should say that it's not necessary to use IV.
I'm right? Why do the tests give me different answers?

Or should I interpret the results as that the In_Ratio_20 variable is endogenous, but the rhausman test results suggest that the FE (b2) estimates are not significantly affect by endogeneity??

I would appreciate your guidance.
Sorry for my English. Thanks since now