Uninformed Choices in Perishables

Abstract

We study consumer choices of perishable items and their effects on retail waste using a dataset tracking expiration-date-level consumer choices and on-shelf inventory. Our data include a field experiment with dynamic discounts for oldest-vintage items and retailer-driven shelf rotation. We begin with a theoretical framework under which we anchor a set of predictions related to consumer information, dynamic discounts, and shelf organization. Our empirical analyses lead to four main findings. First, we document strong evidence that consumers make uninformed choices: in almost half of all purchases, consumers choose an older item when a fresher one is available on the shelf at the same price, and simply rotating the items’ locations increases the oldest-vintage choice share by 24 percentage points. Second, these uninformed choices arise from two types of consumers: those with “mental gaps” who are indifferent between expiration dates, and those with “choice frictions” who prefer later dates but are hindered by search costs. Third, we find that we find that oldest-vintage discounts influence vintage choice by causing choice-frictions consumers to not search by compensating them to take the oldest vintage. Therefore, discounts have a direct effect on choice, as well a choice spillover of keeping the shelf organized for subsequent consumers, extending the benefits after a rotation. Fourth, we quantify the waste implications of this spillover using counterfactual simulations. Dynamic pricing and shelf rotation are both effective policies in isolation, reducing waste 8% and 13% respectively. But because discounts help keep the shelf organized, the combination of both policies reduces waste by about 50% more than their sum in isolation.

Keywords: expiration dates, consumer information, dynamic pricing, nudges, perishable inventory, supply-chain management, retailers, choice frictions, mental gaps

JEL Classification: D90, Q54, M30, H23, L66