A complete model for welfare analysis

Taking advantage of some of the lessons learned from income inequality comparisons over time and/or across space, we provide a complete framework of analysis to compare the social or aggregate welfare of independent cross-sections of household income and non-income household characteristics.This framework serves to clarify a number of traditional issues on i) the proper domain of the social evaluation problem; ii) the need to consider alternative mean invariant inequality notions; iii) the decomposition of changes in real welfare into changes of the mean at constant prices and changes in real inequality; iv) the nature of the interhousehold welfare comparability assumptions implicit in all empirical wok, and v) the strong implications of separability assumptions necessary for inequality and welfare decomposition by population subgroups. This review essay, written with an operational aim in mind, extends and updates the treatment found, for example, in Deaton and Muellbauer (1980). The main novelty is the analysis of the simplifying implications of the condition that income adjustment procedures for taking into account non-income needs are independent of household utility levelsan assumption originally introduced in the theoretical literature by Lewbel (1989) and Blackorby and Donaldson (1989), which is extended here to the absolute case