Does Time-Use Data Support the Unitary Model? A Test using Swedish Time-Use Data from 1984 and 1993

Mainly because of data limitations, the unitary model (dating back to Becker, among others) has been the common-used theoretical framework in microanalysis of the household labor supply. Because of its simplicity, the household members are assumed to allocate time and consumption in consensus, having the implication that it is the households, not the individuals themselves, that are the interesting economic agents. In this paper, I study the household decision to supply labor and test if the unitary model holds. What is subject to a test in this paper is the resulting symmetry of the Slutsky matrix, i.e., that the compensated cross-wage effects are equal. The test uses Swedish time-use data from 1984 and 1993. As empirical specification, I use the trans-log, thereby avoiding some of the restrictive assumptions about functional forms that can be found in other studies. Contrary to most other studies, the conclusion is that symmetry cannot be rejected. This suggests that observed household labor supply could be an outcome of the single-criterion model.