Data/Methods: IKEA is a multinational retailer of home furnishings. IKEA’s iconic catalog is the company’s primary advertising device, representing 70% of IKEA’s annual advertising budget. The catalog contains photos, prices, and written descriptions of items, with prices denominated in the customer’s currency. IKEA prices are famous for remaining unchanged for the life of the catalog year. The IKEA data are free of two major problems that arise in other micro-pricing studies: product substitution and temporary sales. This paper analyzes IKEA catalog price data for Germany, France, Italy, Sweden, the UK, US and Canada, accounting for 2/3 of IKEA’s annual revenue worldwide. All information on each good in the catalog was professionally double-keyed. Record linkage was achieved using information from the item’s name, dimensions, color, and description.
Results/Expected Results: During the Great Recession, IKEA raised prices on a larger fraction of goods, and by a higher percentage amount, than in the years before or since. This is surprising since the prices of raw materials—wood and labor—declined through this period. We find that IKEA price changes respond to lagged inflation with an estimated pass-through rate of about 100%, but we find zero pass-through of lagged exchange rates. IKEA price changes show low coordination across countries, even among countries in the Euro Zone. For example, prices for identical goods are no more likely to move together between Germany and France than between Sweden and Canada. Prices for very similar products, e.g., the “Billy” bookcases, made of the same materials, with variation only in size, do not move together, and often move in opposite directions. These facts imply that IKEA’s pricing decisions are inconsistent with standard pricing models, including time-dependent models, state-dependent models, reference price models, and models with increasing returns to scale in menu cost. By contrast, we find substantial evidence that IKEA catalog prices are heavily influenced by factors associated with behavioral/psychological determinants of price setting. Specifically, we find substantial evidence of pricing that makes sense only if customers are believed to have first-digit bias. We conclude that IKEA catalog pricing is inconsistent with many predictions of standard pricing models, but are consistent with behavioral/psychological pricing practices.