Skip to main content
opinion
Open this photo in gallery:

Fuel is pumped into a vehicle at a gas station in Montebello, Calif. on May 15.FREDERIC J. BROWN/Getty Images

Vass Bednar is a contributing columnist for The Globe and Mail and host of the new podcast, Lately. She is the executive director of McMaster University’s master of public policy in digital society program.

Sometimes GasBuddy feels like a real friend. It’s a mobile app and website that tracks crowd-sourced prices of gas in Canada and the United States so you can quickly figure out where to find the cheapest gas in your neighbourhood. But imagine if gas stations are using GasBuddy to monitor each other in real time. Then it’s a collusion mechanism, not a consumer tool.

Which brings me to a concerning new phenomenon.

For long, we’ve had algorithmic pricing, which is a practice that automatically sets the price of an item for sale in a manner that maximizes profits. It is also sometimes referred to as “dynamic” pricing, and basically speaks to the near-constant microcalibration of prices that we see online, usually based on demand.

Now the use of algorithms is being taken up a notch. In contrast to what we’re used to seeing, algorithmic collusion refers to a situation where competitors use algorithms in deeper and more expansive ways, such as constantly monitoring competitors’ prices and then adjusting their own.

There is nothing inherently wrong with setting prices based on what a competitor charges, but these algorithms react quickly, are at work constantly and have access to a great volume of information. The result is that they mimic the effects of traditional price-fixing.

With such algorithms, companies might as well be in communication with each other with respect to pricing. And in fact, some effectively are when they use the same software. These algorithms help companies achieve anti-competitive outcomes without direct input from humans.

Price fixing by humans is currently illegal, with penalties of a fine, imprisonment for up to 14 years or some combination. Over the years, class-action lawsuits in Canada have been settled over the price-fixing of bread, chocolate, diamonds, drywall, car tires, LCD screens and more. But we haven’t decided whether we’re going to actually acknowledge how such collusion is now effectively being done by algorithms.

Canada is still far from explicitly condemning the practice as it pertains to computational activities. In the current consultation on AI and competition policy, the Competition Bureau has a section focusing on algorithmic conduct where it mentions algorithmic collusion as a “prominent theoretical challenge.” That characterization denies the fact that the practice is quickly becoming an established norm.

It’s hard to see how this practice could be considered “theoretical” when Amazon made headlines last fall after it was revealed that its secret algorithm (codenamed Project Nessie) tested how much it can get competitors to follow it in raising prices. A recent lawsuit in the US alleges that 90 per cent of Vegas Strip hotels used a revenue management program called Rainmaker to raise prices via algorithmic price collusion, matching their prices to each others’. An empirical analysis of the German gas market found that algorithmic pricing is rampant and raises prices overall. Prompted by reporting by Propublica, a series of class-action lawsuits were filed against RealPage alleging that their Yieldstar software used algorithmic collusion and elevated advertised residential rents in the U.S. by US$120 to US$790 per month.

Meanwhile, U.S. Federal Trade Commission chair Lina Khan pointed out last year that “the A.I tools that firms use to set prices for everything from laundry detergent to bowling lane reservations can facilitate collusive behaviour that unfairly inflates prices – as well as forms of precisely targeted price discrimination.” In March, the FTC and the Department of Justice filed a statement of interest explaining that hotels cannot collude on room pricing and cannot use an algorithm to engage in practices that would be illegal if done by a real person.

It is unrealistic to assume that hotel chains and other retailers in Canada haven’t eagerly adopted similar algorithmic strategies as those in the United States. It’s weird and disappointing that Canadian policy makers seem reluctant to acknowledge the consumer harms of algorithmic systems used on the down-low to maximize profits.

In early February, the U.S. introduced the Preventing Algorithmic Collusion Act, which presumes a price fixing “agreement” when direct competitors share sensitive information through a pricing algorithm to raise prices and seeks to ban the practice.

In Canada, we are more likely to require that these systems should be subject to oversight or audit at the direction of the AI and data commissioner that could eventually be realized through Bill C-27. But they should be similarly banned.

The third chapter of the most recent federal budget is dedicated to “Lowering Everyday Costs.” This pledge is not fully reflected in our legislative commitments as we have yet to take a clear position on the use of price-fixing algorithms in elevating the cost of everyday items and experiences.

Follow related authors and topics

Authors and topics you follow will be added to your personal news feed in Following.

Interact with The Globe