Is your use of pricing algorithms putting you at risk of competition investigation?

The CMA recently posted a blog warning about the risks of pricing algorithms. Whilst the CMA acknowledges that the widespread use of such algorithms can bring substantial gains, it is concerned to ensure that they are not "misused" in ways that reduce competition and harm consumers. The blog emphasises that not knowing what an algorithm is doing is "no excuse" and that firms must ensure that they understand how their algorithms work including how prices are set.

This CMA blog forms part of a wider focus by competition authorities on the use of AI - see our earlier posts here.

Pricing algorithms

The CMA's blog focuses on pricing algorithms which it defines as a system that sets prices or recommends prices to be set, usually based on current and past data about market conditions.

The key competition concern is that pricing algorithms may be used to implement or monitor price-fixing. For example, in 2016 the CMA found that two online sellers of posters, featuring popular artists such as Justin Bieber and One Direction, and frames had participated in an illegal price-fixing cartel by agreeing that they would not undercut each other’s prices for products sold on Amazon’s UK website (see here). The sellers used automated re-pricing software to give effect to the cartel. This case resulted in one of the firms facing a fine of over £160,000, whilst the other avoided financial penalties by reporting its conduct to the CMA under the CMA's leniency policy.

The CMA emphasises that firms may break competition law even if there is no explicit agreement to fix prices. This is because the use of pricing algorithms may involve the exchange of confidential information (such as pricing plans, future market strategies, stock levels or spare capacity) between rival businesses, for example if they use the same algorithmic system to set prices. It is also possible that as pricing algorithms become more sophisticated, they could allow competing firms to coordinate their behaviour.

The risks of breaching competition law can be avoided by firms seeking to understand the software they use to ensure that:

• Confidential information is not shared with competitors, whether directly or indirectly (for example, through a pricing consultant or pricing software).
• Any pricing advice from 'pricing solutions' is not influenced by confidential information from competitors (even if the underlying confidential information is not shared directly). 
• A cautious approach is taken when the same algorithm is used by competing firms.  

As noted above, a firm may be found in breach of competition law even if it has no direct knowledge of how the algorithm gives rise to the issue. For example, it may be sufficient that it is reasonably foreseeable that a pricing recommendation could be drawing on confidential competitor information.

Price monitoring software 

Price monitoring software can also give rise to concerns under competition law as a means of facilitating resale price maintenance (RPM). RPM occurs where a supplier of goods or services seeks to set a fixed or minimum price at which its resellers must sell the goods or services.  

For example, the CMA has imposed fines totaling more than £13 million on musical instruments suppliers for breaking competition law by restricting online discounting. This included a finding that three of the suppliers (Casio, Roland and Korg) used price monitoring software which made it easier to monitor online prices in real time and ensure widespread compliance with their pricing policies. It also meant that individual retailers had less incentive to discount for fear of being caught and potentially sanctioned.  

In response to this investigation, the CMA launched its own price monitoring tool to detect RPM activity, improving the CMA’s chance of detecting breaches of competition law. 

Avoiding competition law risk 

If you use pricing algorithms in your business, it is crucial to ensure that you have sufficient understanding of how the algorithms operate. Ask your supplier or your in-house team: 

• What information does your algorithm draw on before making pricing recommendations? 
• Is any of that information drawn from information of other users of the algorithm?  
• Is any of that information made available to other users of the algorithm?  
• How are prices which are recommended or set ultimately determined?  
• What measures are built in to avoid competition law risk?