Thomas Bayes Goes to the Virtual Supermarket: Combining Prior Food Price Elasticities and Experimental Data in a Large Demand System of Dietary Choices

Abstract

Food price elasticities (PE) are essential for evaluating impacts of food pricing interventions. Existing econometric estimates of food PEs are often poor, being based on single observational data sets without much variation in prices and failing to utilise prior information. In order to provide better PE estimates for policy analysis, this paper innovates the use of experimental purchasing data from a recent virtual supermarket experiment to estimate the PE matrix for a large set of foods via a Linear Almost Ideal Demand System (AIDS) and proposes an approach to incorporate PE results from observational data studies in the empirical results within a Bayesian estimation framework. We combine a multi-stage Bayesian approach to estimate a set of demand systems using the Edgerton approach to aggregate elasticities to obtain Marshallian and Hicksian PE matrices for a total of 23 food groups.