In the "Tests on Parameters" section of the PROC MODEL documentation, the documentation erroneously states:

"For OLS and SUR, the likelihood ratio test statistic is computed as..."

followed by a formula for the LR test statistic. The above statement should read:

"For OLS and 2SLS, the likelihood ratio test statistic is computed as..."

In other words, the formula presented is for the OLS and 2SLS estimators, not the OLS and SUR estimators.

Note also that the log likelihood value, L, for the SUR and 3SLS estimators is computed as:

0.5*Objective*n

for the full and restricted models.