CALCULATING THE (LOCAL) SEMIPARAMETRIC EFFICIENCY BOUNDS FOR THE GENERATED REGRESSORS PROBLEM

The semiparametric efficiency bounds for a given model arc important for evaluating the asymptotic distribution of different estimators of the parameters of that model and may also be suggestive with respect to which technique should be chosen for estimation. Nevertheless. for many models these bounds may be difficult to derive and interpret. In this paper we demonstrate a straightforward method to derive and interpret the semiparametric efficiency bounds for the generated regressors model under the assumption that the disturbances of the model are normally distributed.