From the standard of living as a latent variable to the estimation of equivalence scales and other indices

A recent approach to the estimation of equivalence scales (S) suggests that a three-step procedure be followed: one must first form clusters of households with the same apparent standard of living (a latent dimension, to be inferred from selected, “well-behaved” indicators), then estimate within-cluster equivalence scales and finally, with a weighted average, obtain the general equivalence scale. This paper further elaborates on these ideas and illustrates how the same logic can also lead to the estimation of inflation and PPP (purchasing power parity). Thanks to its flexibility, the method can be applied not only to “standard” databases (expenditure surveys) but also to income surveys (e.g. the Bank of Italy SHIW—Survey on Household Income and Wealth), and to any other database including an indicator of resources (e.g. income or total expenditure) and a few “well-behaved” indicators of economic well-being. Empirical results for Italy (2004–2010) are presented and discussed.