Odd Prices at Retail Gasoline Stations: Focal Point Pricing and Tacit Collusion

• Published date: 01 September 2015
• Date: 01 September 2015
• DOI: http://doi.org/10.1111/jems.12103
• Author: Matthew S. Lewis

Odd Prices at Retail Gasoline Stations: Focal Point

Pricing and Tacit Collusion

MATTHEW S. LEWIS

John E. Walker Department of Economics

Clemson University

Clemson, SC 29634-1309

mslewis@clemson.edu

This study empirically investigates the theory that odd-numbered pricing points can be used

as focal points to facilitate tacit collusion. Like other retailers, gasoline stations in the United

States disproportionately sell at prices ending in odd digits. I show that station prices are higher

and change less frequently in locations using more odd prices (particularly those ending in 5

or 9), even after controlling for other market characteristics. The evidence suggests that the use

of pricing points can be an effective mechanism for tacitly coordinating prices, providing an

alternative explanation for the widespread use of odd prices in retail markets.

1. Introduction

Economists and marketing scholars have long discussed the widespread use of odd-

numbered “pricing points” and the preponderance of prices just below round numbers

in many retail markets (Bader and Weinland,1932; Schindler and Wiman, 1989). Despite

its long history,the cause for and effect of these “customary prices” is still widely debated.

Although many chalk the convention up to tradition, others suggest that these prices

have a psychological advantage leading consumers to purchase distinctly more than

they would at the slightly higher round-number price.1A number of studies, including

recent work by Kashyap (1995) and Levy et al. (2011), show that more frequent use of

such pricing points by firms leads to increased price rigidity. However, there has been

essentially no investigation of how the use of pricing points relates to the degree of

competition or the resulting price level in the market.

There are a number of reasons why prices might be higher in markets that rely

more heavily on pricing points. One possibility is that, for whatever reason, consumers

are attracted to buy at certain pricing points, but firms are only able to keep prices at

these customary levels when they have a sufficient level of market power.The incentives

to undercut a rival may simply be too strong in highly competitive markets (even if one

is undercutting an attractive pricing point). In other words, market power may lead to

an increased use of pricing points.

An alternative possibility is that the use of pricing points can actually help firms

keep prices higher by providing focal points for competitors to coordinate pricing deci-

sions. Scherer (1967) provides a nice illustration of the potential for coordination:

I thank Jason Blevins, David Blau, Severin Borenstein, P.J.Healy, Ryan Kellogg, Dan Levin, Howard Marvel,

Steve Puller,and Bruce Weinberg as well as the editor,coeditor, and anonymous referees for helpful comments

and suggestions.

1. See Gedenk and Sattler (1999) for a survey of empirical studies attempting to identify the demand

effects of 9-ending prices.

C2015 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume24, Number 3, Fall 2015, 664–685

Odd Prices at Retail Gasoline Stations 665

By setting its product price at some such focal point, a firm tacitly encour-

ages its rivals to follow suit without undercutting. Conversely, if one firm

announces a price which has no such compulsion, a rival is tempted to set its

own price just a cent or two below. This leads to a further small retaliatory

cut, precipitating a downward spiral which, in the absence of focal points,

has no clear cut stopping point. In setting one’s price at the focal point, one

in effect asks rhetorically, “If not here, where?” implicitly warning rivals of

the danger of downward spiraling.2

In this case, the presence of customary pricing points serves as a type of facilitating

device that firms with market power may use to more effectively keep prices high.

This study empirically examines the extent to which the prevalence of customary

prices is associated with higher price levels using data on retail gasoline prices from

gas stations throughout the United States during 2009 and 2010. Retail gas stations

overwhelmingly charge prices for gasoline that end in odd numbers, particularly 5 and 9.

After truncating off the ubiquitous nine tenths of a cent per gallon that is added to the

posted gasoline prices at all stations, 49% end in either a five or a nine and over 76%

of prices end in an odd number. Although the use of these pricing points may have

evolved from tradition or may be a response to consumer perception, they may also

have become an integral part of how firms exercise their market power.

There are certainly many other product markets where pricing points, such as

round number prices or 9-ending prices, are used much more extensively than in retail

gasoline. However, it becomes difficult to uncover the competitive effects of pricing

points if they are always in use. Local retail gasoline markets, on the other hand, exhibit

significant variation in the use of odd prices—across cities, across neighborhoods within

a city,and even across stations within a neighborhood. This makes it possible to compare

stations’ pricing behavior within an area while controlling for unobserved factors, such

as costs or consumer tastes, that affect all stations.

My empirical results consistently show a relationship between high price levels,

price rigidity, and the use of pricing points. I find that areas or individual stations that

exhibit a higher share of odd prices and prices ending in 5 or 9 change their prices

significantly less often than other nearby stations. For example, when comparing zip

codes within a city, zip codes that are one standard deviation below the mean with

respect to the share of odd prices and the share of odd prices that end in 5 or 9 change

prices about 50% more often than zip codes that are one standard deviation above the

mean. More importantly, these areas or stations also have significantly higher prices. Zip

codes in which the share of odd prices that end in 5 or 9 is one standard deviation above

the mean have prices that are 2.5 cents per gallon higher on average than zip codes one

standard deviation below the mean. This represents a significant price difference for an

industry in which net retail margins are only a few cents per gallon on average.3Ifind

very similar differences in price level and the degree of price rigidity when comparing

stations within the same zip code that differ in their use of odd pricing points. Additional

analysis confirms that these relative price differences acrosszip codes remain even when

controlling for zip code characteristics such as income, population, and station density

that have been shown to impact local retail gasoline prices. Finally, odd pricing appears

2. Scherer (1967, pp. 497–498).

3. According to the National Association of Convenience Stores, gross retail gasoline margins average

about 14 cents per gallon, but expenses such as credit card fees, labor,rent, and utilities make up 12 to 13 cents

per gallon.