On 24 February the Competition and Markets Authority (“CMA”), the UK’s competition law enforcement body, launched an investigation into suspected breaches of the Competition Act 1998 by a number of hotel chains and a provider of hotel data services. The investigation is the latest to raise competition issues relating to the use of “algorithmic pricing tools”.
The CMA’s concerns relate to the way in which the hotel analytics tool STR owned by CoStar collects data from hotel chains. Three chains – Hilton, IHG and Marriott – are under investigation for the use of STR.
The prohibition on information exchange
Chapter I of the Competition Act 1998 prohibits any arrangement or concerted practice between undertakings that has the object or effect of distorting competition. This includes prohibiting actual or potential competitors from sharing commercially sensitive information with each other. Put simply, where a business (“A”) receives information from one of its competitors or potential competitors (“B”) about B’s current or future commercial strategy, that allows A to take account of that information when deciding A’s own strategy. The CMA’s horizontal guidance puts it this way:
“A key principle of competition is that each undertaking should determine independently its economic conduct on the relevant market. This principle does not prevent undertakings from adapting [to changing market conditions]. However, it does preclude any direct or indirect contact between undertakings of such a kind as either to influence the conduct on the market of an actual or potential competitor or to reveal to such a competitor the conduct which an undertaking has decided to follow itself or contemplates adopting on the market, where the object or effect of those contacts is to [distort competition].”
For example, if ABC Limited knows that XYZ Limited is about to raise its prices, or reduce production, then ABC will be able to take decisions to raise its own prices or otherwise alter its behaviour on the market in the knowledge that XYZ will be a less attractive alternative and therefore that ABC’s customers are less likely to move their custom to XYZ.
ABC is of course free to observe what XYZ is doing on the market when XYZ does it, and to respond to that. For example in a retail context ABC may send its employees into XYZ’s stores to observe the prices that XYZ is charging, and ABC may then reduce (or increase) its own prices in response. That ability to observe and respond is a fundamental feature of a competitive market.
However ABC cannot explicitly or implicitly agree with XYZ that it will change its own prices when XYZ does. Where XYZ sends information to ABC about its business strategy, competition law assumes that XYZ is doing so with the intention that ABC will coordinate its own behaviour accordingly. Such coordination between businesses is unlawful.
Businesses are therefore prohibited from sharing commercially sensitive information directly, and also from doing so through agents or consultants acting on their behalf (known in competition law as a 'hub and spoke' arrangement).
Pricing algorithms
A recent feature of online markets in particular has been the development of dynamic pricing. The price of a good or service (such as a flight, a hotel room, or a concert ticket) is no longer set periodically for every customer, and occasionally adjusted manually in response to changes in supply and demand, but is continuously adjusted in real time, sometimes on a customer-by-customer basis.
A dynamic pricing algorithm is a software tool that constantly monitors data and responds to changes in that data by making automatic changes to pricing. ABC can for example observe how many visitors to its website are searching for certain goods or services. When demand for a product appears to be increasing based on that traffic, the algorithm can be set to automatically increase the price of that product, for example a hotel room on a given date. The algorithm can also monitor broader data sets that may indicate higher or lower demand, for example the announcement of local concert dates or school holidays.
The potential issue begins to arise where this “data scraping” takes account of the prices XYZ is charging. Again, ABC is entitled simply to observe XYZ’s prices and to change its own in response. ABC can do this by instructing a third party to monitor market conditions. However, if ABC and XYZ both use the same third party (“TP”), using the same algorithm, then they will both know how that algorithm will respond to changes in their prices. ABC knows that if it puts its price up by 10% then any of its competitors using TP will also increase their prices automatically. If enough of ABC’s competitors use TP, then ABC may be able increase its prices without the risk that it will be undercut by other providers. Conversely, ABC knows that if it reduces its own prices to boost demand, all of its competitors will automatically do the same, and so it will gain no competitive advantage. If reducing its revenues will not therefore give it any advantage in the market, it is disincentivised from doing so.
The approach to algorithmic pricing in competition law
The CMA published a research paper in 2021 on the potential adverse effects of algorithmic pricing on competition, and continues to keep the use and impact of algorithms under review. The CMA recognises that algorithms are not inherently anti-competitive. The CMA’s Horizontal Agreements Guidance, published in August 2023, notes that:
“Undertakings can independently use algorithms to monitor the prices of competitors and to inform their own price setting. This may lead to more intense competition, reduced costs and/or reduced barriers to entry. However, it is possible that the use of such algorithms could result in a concerted practice.”
The concern that the CMA has with algorithms is not therefore simply the automatic effects that they have – since these can push prices down as well as up. Rather, the horizontal guidance suggests two principal means by which algorithms may be used in an anti-competitive way. Firstly, the algorithm may be a means of exchanging information in order to coordinate parameters of competition. That is a particular risk where competitors use the same third party algorithmic tool in the knowledge (and with the intention) that they can use this to signal pricing to each other. If they use different tools then each will have less insight into the way the other will respond to price signals, and this will be less likely to give rise to a competition law breach.
Secondly, algorithmic tools can be used to implement and monitor an agreement to coordinate prices. If such an agreement exists independently of the algorithm (for example if two companies have implicitly or explicitly agreed that they will both use the same algorithm, so that each knows how the other’s prices will move) then that agreement, not the presence of the algorithm, is the underlying breach.
For example, an algorithm that stopped reducing prices at a set “floor” was found to breach EU competition law because the setting of that floor constituted a concerted practice among the algorithm’s users. Similarly an algorithm that had been tailored by two businesses to track the pricing of other businesses but not each other was found to breach UK competition law because it was a manifestation of an agreement between those businesses not to undercut each other.
Penalties for a breach of competition law
A business found to have breached competition law is liable to a fine of up to 10% of its global group turnover, as well as being exposed to damages claims by customers or competitors who suffered a loss as a result of the breach. It will also be excluded from most public procurement opportunities. Company directors can also be liable to disqualification for up to 15 years. Fines and director disqualifications can usually be avoided by being the first participant in a competition law breach to report it to the CMA, under its 'leniency' program. Any business that thinks it may have breached competition law should therefore seek legal advice ASAP.
Our team has significant experience of advising clients on information exchange issues, as well as with implementing compliance programs and training to mitigate the risk of competition issues arising. If your business needs assistance with these or any other competition law matters, please get in touch with Jamie Dunne, Charles Livingstone, or your usual Brodies contact.
Contributors
Legal Director
Partner
