Department of Economics Brown Bag with Richard Sparkes Establishing Empirical Facts about Price Rigidity

Abstract: This paper establishes new empirical facts about price rigidity in the United States using the full NielsenIQ retail scanner dataset. The analysis covers over 239 billion price observations from more than 2.26 million unique products, representing over $4.8 trillion in retail sales between 2006 and 2023. Price rigidity is measured using market reference prices, a novel concept defined as the modal price of identical products sold across all U.S. retail stores within the same week. I find that: (1) the median weekly market reference price-change frequency across products is 16.38%; (2) the median share of retail products experiencing at least one weekly market reference price-change within a month is 41.59%; (3) while product-level price-change frequency exhibits a strong upward trend over 2006–2023, market reference price-change frequency remains largely stable; (4) there is substantial heterogeneity in price-change frequencies across products and stores; (5) significant cross-state heterogeneity exists, driven by stronger price rigidity in rural areas; (6) in low-inflation environments, inflation is primarily driven by larger price increase sizes, whereas in high-inflation environments, inflation is increasingly driven by a rising frequency of price increases; (7) market reference prices are highly concentrated at digit-9-ending price points (65.20%) and display a strong tendency to transition between them (72.87%); and (8) the price selection effect is moderately weak. Overall, these findings highlight the need for more robust and empirically disciplined microeconomic foundations of price rigidity to underpin modern New Keynesian macroeconomic models.