B1804
Title: Efficiency bound under identifiability constraints in semiparametric models
Authors: Melanie Zetlaoui - Paris West University (France) [presenting]
Patrice Bertail - University of Paris-Ouest-Nanterre-La Defense (France)
Abstract: The purpose is to define an adequate efficiency bound in some models presenting some identification problems. It is shown how it is possible to define such bounds in some regular semi-parametric models (in the sense of Le Cam) when an identifying constraint is available, despite the degeneracy of the information matrix. A convolution theorem is proven in this framework. The method is illustrated by applying it to many models, ANOVA, single index, probit models, etc. It is also shown how a two-step procedure is still based on a preliminary estimator satisfying approximately the constraint, allowing the obtainment of an efficient estimator of the parameters.
