Ran two -step GMM estimation . With the two-step, I need to employ the Hansen test rather than the AR test I guess.
However the P- Values of my Hansen-test are very high eg Hansen:(P-value) for my models : 0.996, .999; 0.986; 0.990.
What does this means for my estimation.
Can it be said that my estimation is over identified?
I am sorry that you did not get an answer so far. Anyways, who would see your post and is looking for an answer like you; after performing two-step GMM, you need to check the following :-
(1) AR(1) p-value 0.05 = insignificant ( no second-order serial correlation).
(3)Hansen test >0.05 = insignificant of overidentification restrictions of the excluded instruments, which means the adequacy and validity of the included instruments. But, if you applied one-step, you shall check Sargen test instead. two-step is used when there is heteroscedasticity problem
(4) no. instruments is less or equal to the number of sections = no proliferation problem of instruments that weakens Hansen/Sargen tests.
- Badges
- Science topic
- Similar topics
- Mathematics