Identifiability of Household Equivalence Scales through Exact Aggregation: Some Empirical Results.

General and commodity-specific equivalence scales are estimated with Canadian microdata. The former are identifiable when exact aggregation is imposed. However, these identifiability restrictions can be tested, as can homogeneity and symmetry in the model employed here. Exact aggregation is not rejected but homogeneity and symmetry are. The estimated equivalence scales exhibit increasing returns to scale in family size. They also rise with age, exhibit strong regional effects, and rural households' scales are generally lower than those of urban households. Since scales are sensitive to imposing homogeneity and symmetry restrictions, caution should be exercised in using such equivalence scales for policy purposes.