A characterization of inequality measures based on bilateral inequality (Q1592818)

A function \(I^n: \mathbb{R}_{+}^n\mapsto \mathbb{R}\) is called an inequality measure if it fulfils common axioms like the principle of progressive transfers, normalization and symmetry. The authors include the strict monotonicity axiom and compare it with the principle of progressive transfers when \(I^n\) is homogeneous of degree \(0\). The main result concerns inequality measures \(I^n\) of the form \(I^n(x_1,\dots,x_n)=\varphi_n(\sum_{i=1}^n\sum_{j=1}^nI^2(x_i,x_j))\), \(n=2,3,\dots\), where all \(\varphi_n\) are strictly monotonically increasing and \(\varphi_n(0)\). Under suitable conditions it is shown that \(I^2\) must be convex in each of its variable, and then \(I^2(x,y)=c|x-y|\) for some positive constant \(c\), so that \(I^n\) becomes close to the Gini measure.