Weighting with individuals , equivalent individuals or not weighting at all . Does it matter empirically ?

Therefore, practitioners have adapted standard tools in a simple way: first, convert the heterogeneous distribution of income and non-income characteristics into an equivalent distribution of living standards and second, apply the preferred standard tool to the distribution of living standards, either unweighted, or weighted by the number of individuals. The essence of the living standard function is thus to compress both income and non-income characteristics into a single comparable measure of well-being. This compression is often performed by dividing income by an equivalence scale factor to obtain equivalent incomes. Glewwe (1991) suggests that transfers from poor to rich could decrease inequality. Both Ebert (1997) and Shorrocks (1995) analyze this problem formally and show that, under some standard assumptions, the problem essentially boils down to an incompatibility between Pareto indifference1 — which requires indifference between two distributions with equal living standards for all individuals — and a between type PigouDalton (BTPD) transfer principle — which socially prefers mean-preserving income