Thursday, January 25, 2018

A Puzzle: Why Do Retail Chains Charge Uniform Prices Across Stores?

Imagine yourself as the profit-seeking owner of a chain of retail stores. Would you charge the same (or nearly the same) price across all the stores? Or would you vary prices according to average income level of consumers who use that store, or according to whether the local economy was  robust or shaky, or according to whether the store had geographically nearby competitors?

In their working paper on "Uniform Pricing in US Retail Chains," Stefano DellaVigna and Matthew Gentzkow argue that most retail chains do in fact charge the same (or nearly the same) prices across stores, but that profits would be higher if they did varied prices instead (Stanford Institute for Economic Policy Research Working Paper No. 17-042, November 14, 2017). Obviously, this finding poses a puzzle. They describe the data on 73 retail chains in this way: 
In this paper, we show that most large US food, drugstore, and mass merchandise chains in fact set uniform or nearly-uniform prices across their stores. ...  Our analysis is based on store-level scanner data for 9,415 food stores, 9,977 drugstores, and 3,288 mass merchandise stores from the Nielsen-Kilts retail panel. ... Our first set of results documents the extent of uniform pricing. While we observe no cases in which the measured prices are the same for all products across stores, the variation in prices within chains is small in absolute terms and far smaller than the variation between stores in different chains. This is true despite the fact that consumer demographics and levels of competition vary signifi cantly within chains: consumer income per capita ranges from $22,700 at the average 10th percentile store to $40,900 at the average 90th-percentile store, and the number of competing stores within 10 kilometers varies from 0.6 at the 10th-percentile store to 8.3 at the 90th-percentile store. Prices are highly similar within chains even if we focus on store pairs that face very different income levels, or that are in geographically separated markets. 
In one calculation, if the stores raised prices in stores where the average buyer had higher incomes, they could increase profits by 7%. Why don't stores do this? Indeed, they point to other evidence on European retailers, and on the sale of certain-brand-name products in US markets, which suggests that this pattern of strangely uniform prices is widespread. In the style of any good detective story, there are a list of suspects and clues.

For example, advertising might create a situation where a certain price is publicized to a wide area.  However, the chains in this study advertise mostly on a city-by-city basis, and it would not be difficult for them to vary their advertising across cities.

Another possible explanation is that chain stores don't want to get into a price war with their competitors. With uniform prices, they are signalling to their competitors that they won't be locally flexible in their price choices. However, the evidence shows the same uniformity both when chains are facing lots of competitors, and when they are not, which suggests that worries about avoiding a price war aren't the main issue.

Perhaps varying prices across stores would appear unfair to consumers, and thus damage the brand name of the store. But not many consumers are going to comparison-shop across stores in widely separate areas. And if prices are  higher for stores in high-income areas than in low-income areas, it's hard to imagine that lots of consumers would view this as violation of some ethical rule.

The explanation these authors find most likely involves managerial decision-making costs: that is, figuring out how to set varying prices across stores, and how to adjust those prices over time, is a substantial task. Unless the payoff in terms of higher profits is large, the inertia of uniform pricing becomes attractive. The authors find some suggestive evidence that store-level price flexibility does seem higher, and seems to be increasing, in settings where the profit potential is larger, which they ascribe in part to the ability of improved information technology to keep track of varying prices across stores.

The authors forthrightly note that "none of this evidence is de finitive," which means that the phenomenon of what seem to be overly uniform prices is a good talker for courses in microeconomics and business schools, and an interesting research topic.

In addition the pattern of overly uniform prices is more than just an intellectual puzzle, as the authors point out. For example, less-uniform prices might mean lower prices for areas with low-income consumers, while "redistributing" the higher prices to higher-income consumers, and in this way reduce inequality. Less-uniform prices would mean bigger price cuts when a local economy has a shaky time, which in turn could help that local economy recover. On the other side, less-uniform prices would presumably mean higher prices for those in remote areas, with less geographic competition.