A Supermarket-Level Analysis of Demand for Breakfast Cereals: A Random Coefficients Approach

This paper applies the BLP approach to the demand for ready-to-eat cereals (RTECs) at the supermarket-chain level in Boston using IRI monthly data. The Random Coefficient Model is used to estimate the demand for 37 brands of RTECs at the leading supermarkets in the Boston area. The empirical results provide a wealth of consumer behavior information, including own- and cross-price elasticites for 37 brands of RTECs at four leading supermarkets in Boston. The demand for RTECs is generally price elastic (ranging between -3 and -8). Consumers respond positively and strongly to promotion, negatively and strongly to price, calories and fiber, and weakly to sugar content. Income has a strong interactive effect with product characteristics and thus is a useful variable for market segmentation. In comparison, the results with the more commonly used Logit model indicate significantly lower price elasticities, provide a limited window on consumer behavior, and yield predicted brand and supermarket market shares that are quite divergent from observed values.