distribution factors : implications and identi cation .

This paper provides an exhaustive characterization of testability and identi…ability issues in the collective framework in the absence of price variation; it thus provides a theoretical underpinning for a number of empirical works that have been developped recently. We …rst provide a simple and general test of the Pareto e¢ ciency hypothesis, which is consistent with all possible assumptions on the private or public nature of goods, all possible consumption externalities between household members, and all types of interdependent individual preferences and domestic production technology; moreover, the test is proved to be necessary and su¢ cient. We then provide a complete analysis of the identi…cation problem; we show under which assumptions it is possible to identify, from the observation of the household consumption of private goods, the allocation of these goods within the household as well as the Engel curves of individual household members. JEL classi…cation: D13 Keywords: intrahousehold allocation, collective models, identi…cation, sharing. We thank G. Becker, A. Deaton, J. Heckman, Y. Weiss and participants at conferences and seminars in Paris, Cambridge, Chicago and Milan for useful comments and suggestions. Browning thanks the Danish National Research Foundation for support through its grant to CAM. Errors are ours. yWorld Bank zCAM, University of Copenhagen xDepartment of Economics, University of Chicago and Columbia University. Email: pc2167@columbia.edu