Testing the Life Cycle Model of Consumption Using Information on Households without Children

Ekaterina V. Stepanova, University of Washington

This paper tests the life-cycle hypothesis of consumption using household data from the U.S. Consumer Expenditure Survey (1980-1998). The main puzzle posed by the data is that a household’s consumption follows an inverted U-shape profile during the family’s working life. This would seem to be inconsistent with optimal consumption smoothing behavior. We focus on demographic explanations (number and spacing of children over time) in order to reconcile the theory with the data. Specifically, we work with a sample of families who do not have children currently present in the household – an approach suggested by Browning et al. (2002) - and “extract” the “never have children” households to question whether the consumption of these families has a non-linearity relationship with age. Demographically adjusted, consumption loses its age-dependence non-linearities supporting a demographic explanation within the life-cycle framework. This means that a household’s consumption behavior can be viewed as been intertemporally consistent.