Consumer Choice Over Shopping Baskets: A Linear Demand Approach

Popular demand estimation approaches impose unrealistic choice constraints and misstate market boundaries, biasing price elasticity estimates and affecting our understanding of market power and pass-throughs. I introduce a novel, scalable method that conditions the outcome of consumers' choice on the structure of their consideration set: a function of all the combinations of goods - shopping baskets - considered each purchase instance by consumers. I show that if consideration sets bind consumers' consumption bundles, joint purchases induce substitutes and complements, possibly making market concentration welfare-increasing. To allow for demand estimation across product categories while addressing dimensionality concerns, I develop a Slutsky matrix proxy from joint-purchase frequencies. I test the model's predictions and jointly determine price elasticities for 20 000 goods across 500 product categories for a Portuguese grocery store sample from 2020-23. The results match observed price volatility, profit margin surveys, as well as reports on shifting consumer tastes during the sample period. Mark-ups are found to have remained volatile around a stable mean, with peaks and troughs corresponding to COVID-related events.