The Retailer Winter 2018/19

PRICING ALGORITHMS

Alan Davis Partner, Head of Competition, EU & Trade, Pinsent Masons

Richard Snape Associate Pinsent Masons

1. Introduction A recent survey by the UK’s Competition and Markets Authority found that 72% of companies monitor their competitors’ pricing. It is equally common practice for retailers to monitor and react to competitor prices. Technological advances in artificial intelligence and ‘pricing algorithms’ have now started to revolutionise the way in which prices can be monitored and adjusted. Technology can gather and process large volumes of ‘real time’ data and automatically adjust a retailer’s prices. Competition law requires that any such monitoring and setting of prices and pricing strategies must be conducted ‘independently’ of competitors. Most retailers know that price fixing agreements and collusion between competitors are illegal and can result in significant fines of up to 10% of group worldwide turnover and even director disqualification and criminal prosecution. However, in certain circumstances, the use of pricing algorithms could risk infringing the requirement to act independently of competitors and lead to a variety of competition law concerns. It is important, therefore, for retailers to understand these risks and know how to mitigate against them. 2. Key risks Direct price fixing The immediate competition law concern with pricing algorithms is that, instead of being used to help consumers to find lower prices, such technologies are used to implement and monitor illegal direct ‘price-fixing’ agreements between competitors in order to keep prices high. In Trod (2016), a retailer who sold posters online via Amazon was found to have entered into an illegal price fixing arrangement with its competitor. The companies used a pricing algorithm to monitor each other’s prices and ensure that the illegal agreement was honoured. The software enabled the companies to spot any undercutting of each other’s prices in ‘real-time’. The UK competition authority fined the company and the director involved was disqualified from acting as a director for 5 years. Indirect price fixing Even in the absence of any agreement or communication between competitors, where competing retailers use the same third party algorithm supplier or data pool to determine prices, there is a risk that such an arrangement may create a ‘hub’ from which the ability or incentive to coordinate prices will arise.

Retailers may start to tacitly coordinate, knowing that each competitor is likely to act in the same way (and being able to monitor that this is the case), or otherwise delegate pricing decisions to the same intermediary leading to identical decisions being made. Even if retailers use different algorithms, they could be programmed to communicate and coordinate with each other or, over time, start doing so themselves through self-learning. Clearly, the likelihood of such circumstances arising will depend on a variety of market factors, but competition authorities are investing heavily in specialist skillsets in order to understand when such risks will arise and how to detect them when they do. Resale Price Maintenance An algorithm can also be used by a manufacturer or brand owner to monitor a retailer’s pricing and detect any deviations from ‘recommended’ retail prices (“RRPs”). While RRPs are generally permissible, a retailer must be free to reduce prices as it sees fit and the RRP must be a recommendation only. Therefore, the use of algorithms to monitor and subsequently enforce an RRP so that it becomes a fixed or minimum resale price constitutes a serious infringement of competition law, known as Resale Price Maintenance (“RPM”). The UK and EU competition authorities actively enforce the competition rules against RPM and significant fines can be imposed. In 2018, electronics manufacturers Asus, D&M, Philips and Pioneer were fined over €111m in total for engaging in RPM. All four manufacturers had used “sophisticated monitoring tools” to swiftly intervene if any member of their distribution networks reduced its retail pricing. Personalised pricing Concerns about the use of pricing algorithms are not limited to collusive pricing practices, but also that they may be used for ‘personalised pricing’ (i.e. using data to set prices based on what the retailer thinks is the maximum each customer or customer group is willing to pay). Clearly, such practices could have a detrimental result for certain customers, in particular for vulnerable consumers or in sectors which otherwise lack significant competitive pressures. However, from a competition law perspective, concerns are unlikely to arise unless the supplier of the products is dominant in the market.

28 | winter 2019 | the retailer