NON-PARAMETRIC AND SEMIPARAMETRIC MAXIMUM-LIKELIHOOD ESTIMATORS AND THE VON-MISES METHOD .2.

Having shown in Part I of this paper that some well-known NPMLE's are compactly differentiable functions of the empirical data, and moreover solve a compactly differentiable generalized likelihood equation, we prove an efficiency theorem which shows that such an estimator asympto tically achieves the information bound in the generalized Hajek-Le Cam convolution and asymptotic local minimax theorems.