On the Robust Measurement of Inequality
Résumé
In practice, a dataset used for calculating inequality index does not always present in a single statistic fashion. A robust inequality measure, in the face of multi-valued statistic dataset, is needed and should take into account both inequality and imprecision concerns in a proper way. However, we find that a commonly used approach is problematic in imprecision reduction. We therefore suggest a new approach of robust inequality measure which surmounts the difficulty. This approach naturally generalizes the Atkinson and Gini indices to measure multivalued problem. We finally axiomatize two social welfare functions which induce the robust Atkinson and the robust Gini indices.
Origine : Fichiers produits par l'(les) auteur(s)