AGGREGATION OVER CONSUMERS AND THE ESTIMATION OF A DEMAND SYSTEM FOR U.S. FOOD

This paper estimates a complete demand system for food for the United States using an extension of the Almost Ideal Demand System (AIDS) with household and aggregate data. The major purpose is to explore the implications of aggregation over consumers. Empirical evidence, based on data from the 1980-87 Continuing Consumer Expenditure Surveys, shows that the regression results and demand elasticities of the household and aggregate models and data can be very similar. Further results reveal factors which affect the similarity of the household and aggregate estimates.