Testing a Theory of Exact Aggregation

This article analyzes the importance of exact aggregation restrictions and the modeling of demographic effects in Jorgenson, Lau, and Stoker's (1982) model of aggregate consumer behavior. These issues are examined at the household level, using Canadian cross-sectional microdata. Exact aggregation restrictions and some implicit restrictions on household demographic effects are strongly rejected by our data. These results do not preclude pooling aggregate time series data with cross-sectional microdata to estimate a model of aggregate consumer behavior. They do suggest, however, an alternative basis for the aggregate model.