If the handful of companies controlling a market — say, gasoline — could meet in a ‘smoke-filled room’ and agree on price strategy, the result would likely be higher prices at the pump, higher profits for all the companies, and the consumer footing the bill.

That’s roughly the Economics 101 definition of an oligopoly — a market with a relatively small number of dominant companies — engaging in collusive behavior, which prevents rewarding the most efficient firms, who set lower prices and still make profits.

But research from Professor Leslie M. Marx, the Robert A. Bandeen Professor of Economics at Duke University’s Fuqua School of Business, shows that oligopolists might no longer need smoke-filled rooms or even written communication to negotiate their deals.

In the working paper, “Negotiating Price Coordination: A Study in Retail Gasoline,” Professor Marx studied the retail gasoline market in Melbourne, Australia, and detailed a system of coordination facilitated by online platforms posting real-time price movements at gas stations.

Marx and co-researchers — David P. Byrne and Andy Wu of University of Melbourne, Nicolas de Roos of University of Liverpool, and Matthew S. Lewis of Clemson University — found that Melbourne area’s gasoline retailers adopted a system where some gas stations signaled price increases on a subscription-only online platform, triggering the other retailers to also raise prices and consolidate to the new equilibrium.

“They were basically having a conversation through these price changes and the rapid ability to observe each other's prices,” Marx said.

An Australian government’s antitrust lawsuit against five “petrol” retailers and the online platform found that “the information sharing arrangements (...) had the effect or likely effect of substantially lessening competition in markets for the sale of petrol in Melbourne.”

Professor Marx — who was an expert witness for the Australian government — said that as a result of the settlement of the case, the online platform that was facilitating the collusive system is now required to make its data available for purchase “on reasonable terms.”

This allowed the researchers to study how the collusive system worked, shedding light on how similar information-sharing platforms might facilitate oligopolistic coordination around the world.

Digital signals that drive prices skyward

Marx and colleagues studied the price cycles in the gasoline market of Melbourne, Australia, between 2007 and 2014.

The market was dominated by five large retailers — BP, Caltex, Coles, Woolworths, 7-Eleven — that owned about two-thirds of the gas stations in that market.

The researchers noted that Melbourne prices go through cyclical phases, including coordinated increases, cheating (some firms try to undercut rivals’ prices, lowering prices to attract more consumers), punishment (other companies react by lowering prices even further), negotiation and “restoration”, when everyone again raises prices together.

Marx and colleagues mapped out how, in the first few years of available data, the five dominant players in Melbourne worked out a system based on a day of the week as the focal point when gas stations were supposed to raise prices.

However, after 2011 the gasoline retailers transitioned to a “signal-and-consolidate” pattern, facilitated by an information-sharing platform owned by Informed Sources.

The new system involved a small subset of the stations owned by one retailer — usually about 20% — taking the lead and raising their prices, followed by a larger subset the following day. The new prices, shared almost in real-time on the Informed Sources platform, signaled to the other retailers the intent to start a new price cycle, and were followed by a period of “consolidation,” during which all the other retailers increased their prices accordingly

“It seems the previous arrangement wasn't working satisfactorily for all of the retailers in the Australian market, and they managed to reformulate their mechanism so that instead of having a focal day, they would wait for price signals and respond to those price signals,” Marx said.

The researchers found that in 2013, it took between 14 and 21 days on average for all the gas stations to complete the price-raising process.

Price signals as non-verbal, collusive language of communication 

Historically, antitrust authorities around the world have searched for explicit communication and agreements as evidence of collusive behavior.

However, the antitrust case brought by the Australian Competition and Consumer Commission in August 2014 argued that the rich communication medium provided by the platform Informed Sources — and the pattern of coordination between price movements across the five retailers — was evidence that the gasoline retailers were deliberately coordinating with one another. 

“We can see one gas station raise the price to $1 at 10% of its stations, and then the other one raises its price to $1 at 10% of theirs, and then the other one and the other one, so that in the next hours, they've all agreed to go up to $1,” Marx explained. “You can make a judgment about whether that was coordinated communication among the stations or not.”

The platform provided station-level price data at 15- or 30-minute frequencies to its subscribers, the researchers write, and was unavailable to non-subscribers (like smaller independent stations) and consumers.

Are digital platforms the new smoke-filled rooms for market collusion? the researchers ask.

Antitrust concerns around the world

Data analytics platforms providing market intelligence services, including real-time prices, are used by industries around the world.

Some of these platforms use “pricing algorithms” that might facilitate collusion and harm consumers, the researchers write.

In the U.S., the Department of Justice filed antitrust lawsuits against RealPage (2020) and AgriStats (2023), claiming these digital platforms may have enabled collusion in the rental and meat processing markets, respectively.

As these lawsuits make their way through the system, researchers will be working to understand better the competitive effects of online price-sharing and price-recommending platforms, Marx said.

“The key concern is that they are coordinating in a way that enables them to raise the price higher than they otherwise would, or more frequently than they otherwise would, or spend more time with the high prices than the low prices,” Marx said.

At the same time, collusion continues to be an issue in markets without electronic sharing, she said. 

“There's a retail gasoline case where the retailers would call each other and say, ‘It's time to take a drive,’ meaning it's time for you to go drive around and look at the price boards at the other gas stations,” Marx said. “In order to see whether their rivals had raised the price, they had to go drive around.”

How policy makers and consumers can confront anticompetitive, coordinated behavior

The ups and downs in price cycles exist in many markets, and they don’t always favor the consumer, Marx said. “Sometimes the price cycles are referred to as ‘rockets and feathers,’ because the price goes up like a rocket and then comes down slowly, like a feather,” she said.

When dominant companies change prices on focal days of the week, consumers can learn from the pattern and work around it, Marx said.

It is harder for consumers with the new platform-facilitated signaling system, she said, because they would need to follow the actual prices in real time.

Sometimes, however, an expanded use of publicly available platforms may empower consumers to counter companies’ information advantage, Marx said. In research studying Chilean gas stations, she noted, the researchers compared cities where, along with price-sharing platforms that helped retailers, consumer apps were also largely used, allowing consumers to shop aggressively for lower prices. The research found that in cities where consumers used the data, the prices actually went down. But in cities where the gas stations were mostly using the platforms, the prices went up.

“There's a little bit of a race here,” Marx said. “Who's making a better use of the data? The firms, trying to use it to raise prices? Or the consumers, trying to use it to ‘comparison-shop’ effectively?”

Platforms usually argue that their information-sharing service helps companies to price more accurately, because they argue that it gives them a better sense of total demand and supply in the market, she said.

However, the study of Melbourne’s gasoline market indicates that platforms are also used to enhance price coordination, Marx said, and it may not be enough for policymakers to simply hunt for verbal communications between dominant companies.

“You don't just say, ‘it's fine because there's no verbal communication,’” she said. “These price-sharing platforms offer a rich enough set of communication opportunities that they could facilitate anti-competitive conduct in the same way as verbal communication.”