09-23-2011 12:17 PM
I would like to carry out the Hansen's overidentifying restrictions testusing PROC MODEL. I'm told that this test can be done in the GMMframework in PROC MODEL, and is computed as N*objective in the GMM estimation.I entered gmm in the FIT statement, and observed N*objective in the output. However, there is no probability to accept/reject the null hypothesis. MaybeI'm doing this the wrong way. Does anyone have the correct SAS codes to do thistest?