The New Empirical Industrial Organization (NEIO) literature allows to analyze competition, market power and welfare implications of public policies. This literature is based on the estimation of consumer substitution patterns, that is on the estimation of demand models. Standard discrete choice models are commonly used and assume that the consumer chooses only one unit of a single alternative. However, panelists are often shoppers making decisions for the entire household. A household is composed of several persons with various tastes. The current NEIO literature then often omits multiple brand choice and quantity choice of a product representing multiple tastes within the household. To tackle this point, we model multiple discrete/continuous choices of households following Bhat (2005a). Moreover, we deal with observed and unobserved household heterogeneity as well as omitted variable problem that occurs in this type of consumer choice model. Using French purchase data on the Soft Drink Market, we show that consumer substitution patterns could be biased if multiple brand and quantity choices are not taken into account. Indeed, our results suggest that own price elasticities could be significantly underestimated.
multiple purchases; discrete-continuous choices; soft drink market; price elasticities; household heterogeneity;
Céline Bonnet, and Olivier de Mouzon, “Analysis of Multiple Discrete-Continuous Choices: Empirical Evidence of Biased Price Elasticities under Standard Discrete Choice models on the Soft Drink Market”, TSE Working Paper, n. 14-520, August 2014.
TSE Working Paper, n. 14-520, August 2014