Items ordered for curbside pickup sometimes sell out before their buyers arrive. This obliges the store to choose substitutes on the consumers’ behalf. Using novel data from a supermarket chain, I show that the store’s choice of substitute can influence consumers’ learning about brands. By selecting substitutes from hitherto-unfamiliar brands, the store increases the probability that consumers will purchase the brands’ products in the future. This finding motivates a structural model of curbside pickup under consumer learning, which I use to characterize the optimal substitution policy under a range of store objectives.
WiP
Preferences among Unpurchased Goods in Differentiated Products Demand Systems
Data on consumers’ preferences among unpurchased products can help identify demand elasticities. However, workhorse demand systems sometimes fail to replicate important substitution patterns in these data. I consider the following explanation for this shortcoming: both conditional and mixed logit impose a form of independence between consumers’ purchases and their preferences among unpurchased goods. Using data on curbside grocery pickup, I document substitution patterns that are inconsistent with both models’ independence properties. To quantify the influence of the independence property exhibited by mixed logit, I compare the model’s goodness of fit to that of mixed probit (which exhibits no such property).
WiP
Choice-Set-Dependent Preferences in Grocery Purchases
M. Conlin, K. Harris-Lagoudakis, and A. Zeyveld
(One-Sentence Summary) We utilize grocery scanner data to examine how the attributes of unchosen alternatives in a category (e.g., eggs) alter a person’s relative preference among products in that category.