Is there bias in computing household equivalence scales?

Household equivalence scales are often used to help perform welfare comparisons across households with different demographic composition. Abstracting from the problems of value judgments and ethical standards, the use of equivalence scales to perform welfare comparisons still faces several measurement problems, namely the identification problem and the endogeneity problem. This paper introduces and estimates an unconditional demand system that simultaneously addresses these two problems. By explicitly considering the demand for leisure, and the fact the household can choose some of its demographic characteristics, we deal with the endogeneity problem and obtain consistent estimates. We identify unconditional equivalence scales by estimating the demand for endogenous demographic variables along with the demand for leisure and consumer goods. More general equivalence scales allowing for better comparability are estimated and used for welfare comparisons.