Household equivalence scales and welfare comparisons: A comment

Abstract

This note investigates recent claims by Lewbel regarding the use of equivalence scales in welfare analysis and finds them wanting. We go on to show that scaled income (income divided by the equivalence scale of the household) is an ordinal index of household well-being if and only if preferences are homothetic. Furthermore, we show that these scaled incomes can rank projects consistently with a Pareto-inclusive Bergson-Samuelson social-welfare function if and only if preferences satisfy full-homotheticity (an interhousehold homotheticity condition).