by Fatma Ceren Morbel
Recent technological advancements have enabled companies to collect and utilize data in order to improve their profitability. Globally, algorithmic pricing has become an integral part of market operations. Although pricing algorithms have several advantages, they are also associated with concerns regarding their possible role in facilitating collusion within the digital marketplace.
When purchasing a product through an online website, a consumer may be faced with a pricing algorithm, in which the price can be adjusted according to the consumer's preferences. Therefore, you may already be familiar with pricing algorithms if you have encountered price changes when you added a product to your basket. But pricing algorithms can be described as “a computer program that autonomously adjusts prices based on current and past data related to demand, cost, or rivals’ prices.”[1]
Since data is regarded as the new oil, companies cannot avoid using pricing algorithms. As stated in the Final Report on the E-commerce Sector Inquiry, a survey conducted by the European Commission in 2017 showed that 78% of the retailers who admit to using price tracking software adjust their own prices to match those of their competitors.[2] Furthermore, pricing algorithms enable companies to develop pricing strategies based on consumer behaviour, preferences, likes and dislikes, such as personalized pricing.
In addition to monitoring prices more efficiently than a human, pricing algorithms would also enable quicker response to changes in the market. The use of pricing algorithms is an essential tool for the development of innovative business models for the purpose of improving product quality and customizing products. For instance, pricing algorithms may enhance price transparency and reduce search costs by giving consumers the ability to compare prices, but they also present a substantial risk of causing market distortions due to tacit collusion.
The ways of using algorithms to reach anti-competitive collusion can be differentiated on two aspects. First, an already existing price-fixing agreement can be facilitated by the algorithms. Second, tacit collusion can be achieved through algorithms that are designed to achieve that outcome. In the first situation, if there is an already price-fixing agreement, the algorithm should be evaluated in conjunction with the main agreement that is enabled by it. However, the latter would be very challenging for competition enforcement. Advanced algorithms may learn on their own how to collude among themselves without human intervention. Some commentators even claim that the concept of machines self-learn to collude seems like science-fiction, Professor Nicholas Petit expressed it by stating “Antitrust and Artificial Intelligence literature is the closest ever our field came to science fiction.”[3]
Pricing algorithms can be in different forms such as discriminating algorithms, monitoring algorithms, parallel algorithms, signalling algorithms, and self-learning algorithms.
A form of price discrimination occurs when firms charge consumers different prices based on their behaviour. The use of algorithmic pricing may facilitate price discrimination by enabling firms to identify customers' willingness to pay at a given price. By providing lower prices to consumers who might not otherwise participate in the market, price discrimination may improve efficiency.[4]
Monitoring algorithms are used to implement and monitor the cartel by competitors. A pre-existing contract can be enforced through the implementation of a monitoring algorithm. Use of monitoring algorithm would make cartels stronger and more long-lasting.[5] For example, in the UK, a fine has been imposed by the Competition and Markets Authority against two competing online poster and frame vendors for agreeing to restrict price competition between themselves by utilizing an “automated repricing system” that would automatically determine product prices.[6]
Parallel algorithms also referred as “hub and spoke conspiracy” are type of algorithms which competitors use it to fix prices and respond to market conditions through a collusion. In this situation, tacit collusion can occur when rival companies (spokes) choose not to develop their own price algorithms, but to use those developed by third parties. As a result, if rivals purchase pricing algorithms from the same supplier (hub), the supplier or with other words the hub may be able to raise prices above the level of competition. Thus, the hub can facilitate a horizontal cartel between the spokes through the purchased same pricing algorithm. In 2016, a customer of Uber filed a complaint to the US District Court against Uber by alleging that Uber's vertical agreements with individual drivers lead to horizontal coordination due to the parallel use of the same pricing algorithm. As argued in this case, Uber acted as a hub and its drivers as spokes, colluding with one another through third-party Uber.[7]
Signalling algorithms enable companies to monitor and modify prices according to the market conditions without explicit agreement among competitors. It is claimed that signalling algorithms can improve market transparency and reduce strategic uncertainty. Companies that use this type of algorithm have a will to achieve supra-competitive prices, but there is no explicit agreement among them. Nevertheless, it is possible for artificial intelligence (AI) to collude with other artificial intelligences (AIs) if the AI has developed a self-learning mechanism.[8]
From the above-mentioned types of algorithms, problems related to monitoring and parallel algorithms can be resolved under the framework of the existing competition law. In contrast, it would be difficult to address the problems associated with other types of algorithms. Thus, a reformulation of the competition law framework is necessary to deal with problems of advanced technology.
[1] Zach Brown & Alexander MacKay, Are online prices higher because of pricing algorithms?, Brookings, 2022.
[2] Report from the Commission to the Council and the European Parliament, Final report on the E-commerce Sector Inquiry, 2017.
https://ec.europa.eu/competition/antitrust/sector_inquiry_final_report_en.pdf
[3] Ashwin Ittoo & Nicolas Petit, Algorithmic Pricing Agents and Tacit Collusion: A Technological Perspective, October 2, 2017. Chapter in L'intelligence artificielle et le droit, Hervé JACQUEMIN and Alexandre DE STREEL (eds), Bruxelles: Larcier, 2017, pp. 241-256, Available at SSRN: https://ssrn.com/abstract=3046405 or http://dx.doi.org/10.2139/ssrn.3046405
[4] Karan Sangani, Examining the Antitrust Implications of Pricing Algorithms in the United States, European Union, and India, Robotics, Artificial Intelligence & Law / March-April 2020, Vol. 3, No. 2, 2020. p. 122.
[5] Sumit Singh Bhadauria, Lokesh Vyas, Algorithmic Pricing & Collusion; The Limits of Antitrust Enforcement, Nirma University Law Journal: Volume-8, Issue-2, July 2019.
[6] UK CMA, Decision of the Competition and Market Authority: Online sales of posters and frames Case 50223, https://assets.publishing.service. gov.uk/media/57ee7c274ofob6o6dcooool8/case-50223-final-non-confidential-infringementdecision. pdf.
[7] Bas Braeken & Jade Versteeg, Algorithmic collusion in digital markets and AI: science fiction or reality?, April 2022.
https://www.bureaubrandeis.com/algorithmic-collusion-in-digital-markets-and-ai-science-fiction-or-reality/?lang=en
[8] Karan Sangani, Examining the Antitrust Implications of Pricing Algorithms in the United States, European Union, and India, Robotics, Artificial Intelligence & Law / March-April 2020, Vol. 3, No. 2, 2020. p. 125.
