A family of variability measures based on the cumulative residual entropy and distortion functions

Abstract

Variability measures are important tools in the construction of premium principles and risk aversions. In this paper, we propose a family of such measures based on a distorted weighted cumulative residual entropy, which follows by a sensitivity analysis of distortion risk measures. For this family, we obtain properties, connections with other measures, a covariance representation, and some useful interpretations. Furthermore, we explore an application on premium principles based on beta generated distributions, and we give an empirical estimation. We also provide bounds and numerical illustrations.