|
Competition at work in gas stations Results of a Canadian study shed light on how competing gas stations within a single town adjust their prices at the pumps Alan Bailey Petroleum News
Why is it that gasoline prices always seem to rise faster than they fall? And do variations in gasoline prices from one area to another indicate a lack of competition in the gasoline market?
To try to answer these questions and enable a better understanding of how gas stations set their prices a Canadian research team recorded gas prices at 27 gas stations in Guelph, Ontario, every two hours between 8 a.m. and 10 p.m. from Aug. 14 to Nov. 24, 2005. At the time of the research Guelph had a population of about 106,000, and there were a total of 28 gas stations in the town.
Benjamin Atkinson, Andrew Eckert and Douglas West published the research results in the June 2008 issue of Economic Inquiry.
Competitive pricing? An informal theory for gasoline pricing, promoted by the oil industry and government, says that consumers shop around for the lowest possible gas prices in a region, with large price placards at gas stations helping with the consumers’ search, the Economic Inquiry article says.
Gasoline retailers should set prices non-cooperatively. As a consequence, prices should tend to be fairly uniform, with price changes moving rapidly through the market. The majority of gas stations should respond to price levels set by a small number of key competitors, with price reductions radiating outwards from an initial source, like a tumbling pack of dominoes.
But is this theory correct? There are a number of factors that could complicate a simple picture of price competition. For example, a gas station may be located conveniently next to a car wash or a store.
And the business structure of the gas station could impact pricing policy. Some gas stations form part of oil company-owned chains; others are lease-operated within a brand-name chain while some are independently owned. To what extent do oil companies set prices within owned or leased gas station chains? And do independent operators tend to focus more on a primary business such as automotive repair rather than jumping onto every regional change in retail gas prices?
Four categories In fact the Canadian researchers distinguished four general categories of gas station in Guelph: brand-name stations where the prices were known or thought to be controlled by the brand; brand-name stations where the prices were thought to be controlled by the local operator; large independent stations; and small independent stations.
The researchers found little difference between the different types of gas station when it came to the frequency of gas price rises — on average during the 103-day study stations of each type raised their prices between 17 and 23 times.
But price reductions occurred much more often. And the frequency of reductions varied substantially between different types of station. Stations with prices controlled by the brand lowered prices 162 times, brand-name stations with independent pricing lowered prices 84 times, large independents lowered prices 120 times and small independents lowered prices 46 times.
However, the fact that many of the stations with brand control of pricing are located in a similar part of the town caused some uncertainty about the extent to which location played a role in price reduction frequency.
Pricing cycle When the researchers plotted the most frequent gas price in each two-hour data collection period, they found that the variation of price over time cycled up and down in a saw-tooth pattern; relatively slow price declines followed rapid price rises. The price changes did not appear to correlate with any variation in the cost of the crude oil used to refine the gasoline (during the period of the study the oil price did not vary greatly).
The researchers concluded that the pattern of observed gas price fluctuations corresponded to what economists term an “Edgeworth cycle.” In this type of cycle, different retailers undercut each other’s prices to cause an overall price decline during a period of one week or more. However, once the price reaches a floor equal to the wholesale cost of the product, one company will raise its prices back to a more profitable level. Once the price rise is initiated, all other companies will rapidly follow suit. Competition then starts to push prices down again.
But although this finding lends credence to the retail gas market being competitive, the data from Guelph show significant variations in gas prices between different gas stations at any one time, rather than the uniform pricing predicted by gasoline pricing theory.
“Excluding the first two days of each (pricing) cycle, 47 percent of prices are at least one-half of a cent above or below the current mode price,” the Economic Enquiry article says.
The researchers also found that varying policies regarding which digit to use at the end of a price created artificial pricing discrepancies between some stations. For example, some stations always ended their prices with the digit six, while others ended their prices with the digit nine.
The data suggested that each station tends to match the prices at a group of other stations, rather than at all of the gas stations in town, the article said. Each station’s price had a high degree of correlation with at least one other station in town. And eight of the 17 major brand stations most frequently matched the price of another station in the same brand chain.
Proximity The extent to which the proximity of gas stations determines the correlation between their gas prices was not entirely clear from the data — in some case rival stations close to each other showed fairly well-correlated pricing, while on average the price correlation for closely located stations was quite low.
An analysis of price changes between successive two-hour observation periods also showed an uncertain impact of gas station proximity. In some cases the analysis showed that one station responded to price changes in a neighboring station while in other cases factors other than proximity seemed to come into play.
And it appeared that brand-name stations tended to be price leaders in driving price reductions.
A further examination of the timing of price reductions revealed that many stations responded to price reductions by other stations within two hours. The data indicated that independent gas stations tend to respond to price changes more slowly than the brand-name stations.
Domino effect An examination of the speed with which price cuts propagated from one of the price leader gas stations revealed broad consistency with the concept of a domino effect, with more distant stations lowering their prices later than the closer stations. However, the contract type of the gas station (brand owned vs. independent, for example) appeared to have more impact than distance between stations in determining the speed of response. And the propagation of price reductions took several days to occur, a slower response to the initial price cut than pricing theory would suggest.
When it came to price increases, the researchers found that all of the gas stations raised their prices within 24 hours of a price increase by a price leader, although it took some stations several hours to respond to the increases posted by competitors.
As a general result of their study, the researchers concluded that the informal theory for retail gasoline pricing is essentially correct, in that gas stations tend to set prices to match small numbers of rival stations; price leaders tend to initiate price changes that then ripple though all the gas stations in town.
However, the theory considerably oversimplifies reality, the researchers found. Stations within the same brand chain tend to follow a consistent pricing pattern; there are significant variations in the speed with which different gas stations respond to price changes; and gas stations do not necessarily react to the prices set by their nearest neighbors.
| 