ASSESSING THE ROBUSTNESS OF COMPOSITE INDICES RANKINGS

Abstract

Ranking objects in terms of different attributes is a crucial practice that is typically sensitive to the choice of attributes' weights. In this paper we present rigorous methods to assess the extent to which the weight‐based rankings are robust to the choice of alternative weights. Empirical illustrations are provided, showing the robustness of country rankings arising from the values of the UNDP Human Development Index, the Gender‐related Development Index, or the Human Poverty Index among others. The ideas and techniques presented in this paper can be used to assess the reliability of multiattribute rankings.