Review of economics and statistics vol:90 issue:2 pages:267-274
We compare the empirical performance of the unitary and the collective approach to modelling observed labour supply behaviour. Deviating from the mainstream literature, we conduct a nonparametric analysis, which avoids the distortive impact of an erroneously specified functional form for the preferences and/or the intrahousehold bargaining process. Our analysis specifically focuses on the goodness-of-fit of the two behavioural models. To guarantee a fair comparison, we complement this goodness-of-fit analysis with a power analysis. Our results strongly favour the collective approach to modelling the behaviour of multi-person households. More generally, they illustrate the usefulness of nonparametric testing tools for the empirical evaluation of theoretical behavioural models.