Inventory Information Frictions Explain Price Rigidity in Perishable Groceries

Abstract

Grocery retailers cannot engage in inventory-based pricing without physically auditing their shelves because their inventory records are inaccurate and incomplete. In particular, inventory data do not reflect actual, real-time quantity on the shelf, and standard UPC barcodes do not contain expiration dates. We argue that these inventory information frictions are much more powerful in explaining price rigidity in perishable groceries than physical menu costs are. We build our argument in two steps. First, we add inventory shrink, deteriorating product quality, and menu costs to Gallego and Van Ryzin's (1994) classic revenue-management model, and we derive the model's optimal pricing policy when the price setter can and cannot condition on real-time inventory information. Within the model, reducing a representative product's physical menu cost by 90% increases its price-updating frequency by at least 300% more when the pricing manager also has real-time inventory information than when they do not. Second, we provide proof of concept for our theory with data from two industry pilots: a U.K. grocery store that installed electronic shelf labels (ESLs)—a technology that reduces physical menu costs—and an E.U. grocery store that installed both ESLs and expanded barcodes—the latter being a technology that reduces inventory information frictions. Consistent with our theory, the technology upgrade increased price-updating frequency by 54% in the first pilot and by 853% in the second pilot. Our theory can explain the lack of wide-spread dynamic pricing of perishable groceries, which has significant consequences for profits, consumer surplus, and food waste.