Pricing Algorithms under EU Competition Law
Businesses increasingly rely on software tools to improve their pricing decisions. But whereas pricing algorithms have become commonplace in many industries, the antitrust framework for assessment is still the subject of debate. The use of pricing algorithms makes markets more efficient, for example by enabling sellers to understand and react swiftly to fluctuations in market demand and changes in supply conditions. However, algorithms could potentially also have anticompetitive effects. The European Commission’s (the Commission) review of its Guidelines on horizontal cooperation agreements has borne out that there is too little guidance on the use of algorithms. In the absence of specific guidance, some high level principles can be identified based on past investigations and existing Commission guidelines.
What is algorithmic pricing?
Pricing algorithms are software tools assisting businesses in their pricing decisions. Every rational company will determine prices for its products and services based on observed market conditions, and many will rely on a variety of tools to guide their decisions (e.g., market reports, customer surveys, price tracking data). A pricing algorithm is one such tool. It can perform complex calculations and data-processing functions that could be costly to execute for human beings. Pricing algorithms are used to adjust prices to changes in demand and supply (i.e., dynamic pricing) or to adjust prices across customers (i.e., personalised pricing). These algorithms may take into account a company’s own confidential information (e.g., inventory, cost base) as well as other observable information (e.g., competitors’ prices, demand fluctuations). There are also pricing algorithms that can be used by buyers, for example, price-tracking and price-forecasting websites and apps recommending when to buy and from whom.
Pricing algorithms are becoming commonplace. They have been in use in the airline industry for decades. The hospitality and financial industries have also been making use of pricing algorithms for a number of years. With the rise of e-commerce, pricing algorithms are becoming more common in retail markets as well. These software tools are easily accessible, even for small businesses that can use off-the-shelf solutions.
The rise of pricing algorithms has attracted the interest of antitrust commentators and regulators. The number of national and international reports and studies on the challenges of algorithms for competition policy and enforcement has been overwhelming, including reports by the Organisation for Economic Co-operation and Development (OECD), the European Union, the International Competition Network (ICN), the UK, Germany, France, the Netherlands, Portugal, Norway, Finland and Japan. Other studies are ongoing, including in the Netherlands and Greece. Some national competition authorities have created specialised digital economy units within their structure or have joined forces with other sector regulators, to develop in-depth expertise into algorithms as part of their remit, as is the case for example in France and the UK. Many of these reports ask for vigilance and call for increased monitoring of antitrust risks, but few of these reports offer empirical evidence of anticompetitive pricing, or insights as to what might guide the agencies’ assessment.
Pricing algorithms can have both pro- and anticompetitive effects. A pricing algorithm may intensify pricing competition and enhance consumer welfare. As noted by the OECD, ‘[d]ynamic pricing algorithms have been recognised to improve market efficiency, by allowing companies to react instantaneously to changes in supply conditions – such as stock availability, capacity constraints or competitors’ prices – as well as to fluctuations in market demand.’ The OECD has further indicated that ‘[p]ersonalised pricing, like any price discrimination, is typically pro-competitive and often enhances consumer welfare. As compared to more traditional forms of price discrimination, personalised pricing generally has more accentuated effects, having the potential to optimise static efficiency and incentives for innovation’. However, the use of pricing algorithms could also result in collusive pricing or price discrimination. The likelihood of anticompetitive pricing is dependent on very specific conditions.
Relevant EU framework
The ongoing Commission evaluation of the Guidelines on horizontal cooperation agreements has borne out that there is too little guidance in this area. However, pending the revised Guidelines, the existing Chapters on commercialisation agreements and information exchanges provide some relevant principles.
For example, a joint sales platform that sets uniform prices for all participating merchants would normally amount to a restriction under Article 101(1) but may well be justified under Article 101(3) TFEU. Importantly, the Guidelines on horizontal cooperation agreements indicate that restrictive effects are unlikely if the parties’ market shares do not exceed 15 per cent. Outside this safe harbour, no presumptions apply and restrictive effects need to be proven and assessed against efficiency gains (e.g., lower prices or better product quality or variety).
Moreover, the Guidelines recognise that information exchanges ‘may improve [companies’] internal efficiency through benchmarking against each other’s best practices’ and ‘may also help companies to save costs by reducing their inventories, enabling quicker delivery of perishable products to consumers, or dealing with unstable demand’. The Commission takes the view that ‘[i]n general, exchanges of genuinely public information are unlikely to constitute an infringement of Article 101’. Market coverage also matters.
Most of the decisions that have dealt with pricing algorithms to date concern traditional price fixing agreements (notably the UK Online Sales of Poster and Frames case) or resale price maintenance (notably the Commission Asus case). Outside the realm of hardcore restrictions, there is less case precedent but the existing cases support a more nuanced assessment of pricing algorithms.
A much discussed case is the E-TURAS online travel booking case. On 2 May 2016, after a reference to the Court of Justice of the European Union (CJEU) for a preliminary ruling, the Supreme Administrative Court of Lithuania (SAC) rendered judgment in an appeal against a decision by the Competition Council concerning Eturas UAB (the holder of exclusive rights to E-TURAS online travel booking system and the system’s administrator) and travel agencies using the E-TURAS system. This case concerned a commercial online travel booking platform for licensed travel agents. The platform administrator had sent a message to the travel agents, via the platform’s personal electronic mailbox, informing them that the discounts on tours sold through the system would be capped. The system underwent the technical modification necessary to implement that measure. The Competition Council had concluded that the travel agencies could – upon an assessment of the system’s operating principles and properties, the system message published by Eturas UAB, and the information about discounts published on the travel agencies’ websites – reasonably predict that all the travel agencies using the system would apply discounts of up to 3 per cent. The Competition Council held that the travel agencies indirectly expressed their common intention to behave in a certain way and qualified this as concerted actions. The SAC referred a question to the CJEU on the liability of the travel agents. The CJEU held that:
[I]f it cannot be established that a travel agency was aware of that message, its participation in a concertation cannot be inferred from the mere existence of a technical restriction implemented in the system at issue in the main proceedings, unless it is established on the basis of other objective and consistent indicia that it tacitly assented to an anticompetitive action.
Taking into account the CJEU ruling, the SAC grouped the travel agencies as follows: (1) travel agencies that knew about the imposed restriction and did not oppose it; (2) travel agencies that knew about and opposed the imposed restriction; and (3) travel agencies with respect to which there was no evidence whether they knew about the restriction imposed in the E-TURAS system. The SAC concluded that there was no basis to hold the group (2) and (3) travel agents liable. The SAC also concluded that the group (1) travel agents could be considered as participants to an anticompetitive practice. The SAC further found that Eturas UAB, who had arranged and introduced the restriction on the discount in E-TURAS system, violated competition law as well.
Equally relevant is the 7 June 2018 Luxembourg Competition Authority (LCA) decision exempting the algorithmic price-fixing arrangement of Webtaxi, a booking platform for taxi services in Luxembourg, from the prohibition of the national equivalent of Article 101 TFEU. Taxis belonging to several companies made use of the booking platform, which fixed the fares for the participating taxis with the help of a pricing algorithm. The LCA concluded that this arrangement qualified as a by-object restriction but went on to assess the claimed justifications. The LCA found that the fixed fares came with various benefits for the participating taxis, consumers and the environment. With respect to consumer benefits, the LCA assessed the algorithm and concluded that algorithm-based fares would always be equal to or lower than the meter price as the algorithm used a digressive price per kilometre. In addition, the LCA found that Webtaxi’s per-kilometre price was lower than that of its direct competitors. Given that Webtaxi’s estimated market share was only 26 per cent, the algorithm did not remove price competition in the market. Together with the benefits for the taxis and the environment, the LCA concluded that the restriction was justified.
In June 2020, Ageras A/S entered into a settlement with the Danish Competition Authority (DCA) and accepted to pay a fine for infringing the Danish Competition Act by using a price standardisation mechanism and setting minimum prices on the platform ageras.dk, which connects users (typically smaller companies) with professional service providers such as accountants, bookkeepers and lawyers (partners). If a partner offered a price below the market price estimated by Ageras, the partner would receive a pop-up message indicating that the offered price was below the estimated market price and would be given the opportunity to adjust its bid. In addition, the minimum partner-fee was calculated based on the estimated market price. The DCA found that by creating an algorithm and pop-up prompt informing individual partners of the ‘estimated market price’, Ageras invited the partners on the platform to enter into an illegal agreement with the intent of raising prices on the platform and that this amounted to a ‘by object’ restriction. The DCA found that the partners acquiesced to the arrangement by failing to publicly distance themselves from the practice or by explicitly consenting directly to Ageras. The DCA only targeted Ageras and not the partners as Ageras initiated the practices and implemented the algorithm and pop-up prompts.
Several investigations that bear relevance on the same subject are still pending. For example, the Spanish competition authority (CNMC) has opened an investigation into anticompetitive agreements in the real estate intermediation market. The CNMC is investigating whether this coordination was implemented by means of software and digital platforms and is exploring whether the conduct has been facilitated by IT firms offering real estate brokerage software and algorithms.
The MEO case has significance for the assessment of personalised pricing tools as it sets out the conditions for impermissible price differentiation. The case concerned a dispute between television broadcaster MEO and royalty collection organisation GDA. MEO had lodged a complaint with the Portuguese competition authority alleging, inter alia, that GDA had applied less favourable terms to MEO than to another television broadcast company. Importantly, the CJEU concluded that a finding of discrimination is not enough; there must also findings that it hinders the competitive position of some of its business partners. Relevant considerations are: the degree of market power, the degree of negotiating power, the conditions and arrangements for the charges, the duration of the price differences, the level of the price differences (notably the percentage of total costs), and the possible existence of a strategy aimed at hamstringing a particular business partner (or the absence of any interest to exclude a business partner).
The interest for pricing algorithms from antitrust commentators and regulators has to date not resulted in the formulation of practical guidance. Suggestions of increased monitoring or mandatory reporting of the details of pricing algorithms may increase the burden on companies without addressing any competition issues. Consumer laws will address some of the issues identified in recent reports. Moreover, the Commission’s review of its Guidelines on horizontal cooperation agreements may constitute an important opportunity to confirm the applicable framework and how it applies to the use of pricing algorithms.
1 Ingrid Vandenborre is a partner and Michael J Frese is an associate at Skadden, Arps, Slate, Meagher & Flom LLP. The authors wish to extend thanks to Caroline Janssens, senior professional support lawyer, non-practising solicitor, with Skadden, Arps, Slate, Meagher & Flom LLP for her invaluable assistance.
2 Commission Staff Working Document, Evaluation of the Horizontal Block Exemption Regulations, p. 44.
3 See also ‘Algorithms and Collusion: Competition Policy in the Digital Age’ (OECD, 2017), p. 9.
4 Algorithms and Collusion (Note from the European Union, 21–23 June 2017), p. 8. See also: Algorithms and Collusion: Competition Policy in the Digital Age (OECD, 2017), p. 18: ‘The development of algorithms has improved the ability to offer price comparison services either via search engines or comparison platforms. Price comparison websites (PCW) make it easier for consumers to compare the available offers and find the best alternative. Comparison platforms can also contribute to level the playing field and intensify competitive pressure.’
5 Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò and Sergio Pastorello, ‘Algorithmic Pricing: What Implications for Competition Policy?’, Review of Industrial Organization (2019) 55:155–171, 156.
6 Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò and Sergio Pastorello, ‘Algorithmic Pricing: What Implications for Competition Policy?’, Review of Industrial Organization (2019) 55:155–171, 156.
7 OECD, ‘Algorithms and Collusion: Competition Policy in the Digital Age’, 2017.
8 ‘Algorithms and Collusion’, note from the European Union, 21–23 June 2017. The EC’s proposal for a Digital Markets Act Impact Assessment does refer to algorithms, although only very briefly and primarily in reference to self-preferencing and lack of transparency and not with regard to pricing algorithms.
9 ICN, ‘The impact of digitalization in cartel enforcement’, 28 April 2020.
10 UK Competition and Markets Authority (CMA), ‘Pricing algorithms: Economic working paper on the use of algorithms to facilitate collusion and personalised pricing’ (CMA’s economic working paper on the use of algorithms), 8 October 2018; ‘Unlocking digital competition’ (the Furman report), a report of the Digital Competition Expert Panel appointed by the UK Chancellor of the Exchequer and chaired by Professor Jason Furman, former chief economist to former US President Obama, 13 March 2019; Ofcom, ‘Personalised Pricing for Communications’, 4 August 2020; CMA, ‘Algorithms: How they can reduce competition and harm consumers’ (CMA’s consultation paper on algorithms), 19 January 2021. See also, CMA, ‘Algorithms: How they can reduce competition and harm consumers. Summary of responses to the consultation’, May 2021, CMA141con; UK Department for Business, Energy and Industrial Strategy (BEIS) research paper on Personalised Pricing and Disclosure, 2021/008, 20 July 2021.
11 German Commission of Experts on Competition Law 4.0 report ‘A New Competition Framework for the Digital Economy’, 9 September 2019, the Commission concluded that ‘[i]f algorithms are trained with too little data or with data that is too uniform, this will have a negative impact on the algorithms’ abilities to deal with the problems they were supposed to solve.’; Joint study by the French Autorité de la Concurrence and the German Bundeskartellamt “Algorithms and Competition’ (Joint French-German study on algorithms), November 2019; German Monopolkommission, XXIII Biennial Report, 2020.
12 Joint French-German study on algorithms, op. cit..
13 Dutch Authority for Consumers & Markets (ACM), ‘Guidelines on the protection of the online consumer. Boundaries of online persuasion’, 11 February 2020; ACM, ‘Position Paper on Supervision of Algorithms’, 10 December 2020.
14 Portuguese Autoridade da Concorrencia, ‘Digital ecosystems, Big Data and Algorithms’, July 2019.
15 Norway Competition Authority, ‘Survey on the use of monitoring algorithms’, 3 February 2021.
16 Finland Competition Authority, ‘Collusion situations caused by algorithms’ and ‘Personalised pricing in light of consumer and competition policy’, 9 February 2021.
17 Japan Fair Trade Commission, ‘Report on algorithms and artificial intelligence’, 31 March 2021.
18 On 10 December 2020, the ACM launched a study into the functioning of algorithms in practice, i.e., a pilot investigation in which the ACM will map out the type of information it needs in order to study the use of algorithmic applications by companies in future investigations and supervision activities.
19 On 11 March 2020, the Hellenic Competition Commission launched an e-commerce sector inquiry to assess if artificial intelligence and algorithms are harming consumers. The final report is expected to be released in October 2021.
20 The French Competition Authority’s digital economy unit was launched on 9 January 2020. The French government also established a panel of leading practitioners in digital regulation to offer expertise about algorithms to the French competition authority and other government departments.
21 The CMA inaugurated its Digital Markets Unit (DMU) on 8 April 2021. In addition, the CMA has indicated that it intends to work closely with the Information Commissioner’s Office, the Communications regulator, and the Financial Conduct Authority through the newly set up Digital Regulation Cooperation Forum to share intelligence and take coordinated action regarding algorithms.
22 OECD, op. cit., p. 16. See also the CMA’s economic working paper on the use of algorithms, op. cit., Paras. 4.2-4.4.
23 OECD, Personalised Pricing in the Digital Era, 28 November 2018, p. 7.
24 See, e.g., ‘Pricing algorithms: Economic working paper on the use of algorithms to facilitate collusion and personalised pricing’ (CMA, 8 October 2018), 5.2.
25 See for more discussion on this subject Ingrid Vandenborre and Michael J Frese, ‘Algorithmic Pricing: Candidate for the New Competition Tool?’, E-Commerce Competition Enforcement Guide (third edition, 2020), p. 26.
26 Commission Staff Working Document, Evaluation of the Horizontal Block Exemption Regulations, p. 44.
27 Commission Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal cooperation agreements, OJ 14.1.2011, C 11/1, Para. 254.
28 id., Para. 250.
29 id., Paras. 57 and 96.
30 id., Para. 92. See, however, Case T-587/08, Del Monte Produce v. Commission, judgment of 14 March 2013, ECLI:EU:T:2013:129, Para. 369, where the General Court held that the fact the certain information could be obtained from other sources is not relevant as ‘the exchange system established enabled the undertakings concerned to become aware of that information more simply, rapidly and directly’.
31 Case 50223 Online sales of posters and frames, 12 August 2016. The CMA investigation followed similar investigations by the US DoJ in US v. Daniel William Aston and Trod Limited (2016) and US v. David Topkins (2015).
32 Case AT.40465 – Asus, Decision of 24 July 2018. The CMA has issued decisions in similar type cases: Online resale price maintenance in the synthesiser and hi-tech sector (29 June 2020); Online resale price maintenance in the electronic drum sector (22 July 2020).
33 See, e.g., Ingrid Vandenborre and Michael J Frese, ‘Algorithmic Pricing: Candidate for the New Competition Tool?’, E-Commerce Competition Enforcement Guide (third edition, 2020), pp. 27–28. Although outside the scope of this article, we would point out that there has also been enforcement activity against non-pricing algorithms, e.g., market-sharing arrangements that are implemented through algorithms, see the UK’s Office of Gas and Electricity Markets (Ofgem) 26 July 2019 decision regarding Economy Energy, EGEL and Dyball, but also for common algorithms used by insurance companies in settlement and underwriting, see the Italian Autorità Garante della Concorrenza e del Mercato (AGCM) 4 October 2021 decision accepting commitments concerning the anti-fraud project of the ANIA.
34 Case No. A-97-858/2016, SAC Decision of 2 May 2016.
35 See also Ingrid Vandenborre and Michael J Frese, ‘Algorithmic Pricing: Candidate for the New Competition Tool?’, E-Commerce Competition Enforcement Guide (third edition, 2020), pp. 27–28 for a discussion of these decisions.
36 Spanish Competition Authority, ‘The CNMC opens antitrust proceedings against seven firms for suspected price coordination in the real estate intermediation market’, press release, 19 February 2020.
37 Case C-525/16 MEO — Serviços de Comunicações e Multimédia v. Autoridade da Concorrência (MEO), ECLI:EU:C:2018:270.
38 id., Paras 31 and 34.