The traditional approach to household behavior is to treat the household as a single individual. Large empirical evidence, however, rejects the unitary approach in favor of a more general model, the collective, where each households member has his or her own preference and choices are Pareto optimal. Testable restrictions of the two approaches are based on properties of expenditure data (symmetry of the Slutsky matrix or collective forms of GARP). When households' choices do not involve prices (e.g. transfers, charitable giving etc.) we are not able to test. This project aims at developing a price-free test of household behavior. We will operationalize the "inferential approach", where preferences of the household's members are inferred from observables (revealed preferences).
