Household intertemporal behaviour: A collective characterization and a test of commitment

In this paper, a formal test of intra-household commitment is derived and performed. To that end, two models of household intertemporal behaviour are developed. In both models, household members are characterized by individual preferences. In the first formulation, household decisions are always on the ex ante Pareto frontier. In the second model, the assumption of intra-household commitment required by ex ante efficiency is relaxed. It is shown that the full-efficiency household Euler equations are nested in the no-commitment Euler equations. Using this result, the hypothesis that household members can commit to future allocations of resources is tested using the Consumer Expenditure Survey. I strongly reject this hypothesis. It is also shown that the standard unitary framework is a special case of the full-efficiency model. However, if household members are not able to commit, household intertemporal behaviour cannot be characterized using the standard life-cycle model. These findings have two main implications. First, policy makers can change household behaviour by modifying the decision power of individual household members. Second, to evaluate programmes designed to improve the welfare of household members, it would be beneficial to replace the standard unitary model with a characterization of household behaviour that allows for lack of commitment.