Unravelling the Veil: Unintentional Algorithmic Collusion in Oligopolistic Markets

An oligopolistic market structure is characterized by considerably fewer competitors, identical products, and inelastic demand. According to European and American legal jurisprudence, price parallelism and supra-competitive prices are only natural results of such market arrangements and are therefore not per se illegal. To determine whether such pricing parallelism is conscious price parallelism, the countries also consider additional variables such as communications and exchange of information.

Industry research in 2017 conducted by the European Union (“EU”) found that a substantial number of online retailers utilise pricing software in their businesses. Major retailers with a variety of several thousand unique items require an algorithm for automated pricing adjustments. The David Topkins and Trod cases in the United States (“US”) exposed instances of algorithmic collusion where price-fixing algorithms were used to maintain prices abnormally high and collaborate with a competing business. This is crucial when businesses use the same pricing software covertly, resulting in a situation where both algorithms set prices in the same general direction. As a result, the price of two enterprises may appear to be anticompetitive completely unintentionally.

Subverting the traditional notion of Anti-Competitive Agreements

In algorithmically adjusted pricing, the price of the competitor’s product will automatically be altered in response to the price increase. In algorithmic collusion, competitors may raise their prices as a signal for other competitors to follow. Prices are revised by algorithms based on a number of factors, with the pricing of competing goods receiving the greatest consideration. Price changes may therefore lead to artificially high pricing and price stabilization.

The Competition Commission of India’s (“CCI”) ruling in Samir Agrawal v. CCI (Samir Agrawal) is a missed opportunity to address algorithmic collusion. The regulator stated that collusive behaviour requires a consensus or meeting of the minds when establishing the direct evidence test. This strategy adopted by CCI was similar to that taken in Re: Domestic Air Lines, where the Director General (“DG”) came to the conclusion that airlines were acting, in the same manner, to raise rates following strikes by Air India employees. However, the CCI rejected the DG’s findings because there was insufficient material evidence establishing an agreement between the parties.

The matter of Re: Alleged Cartelization in the Airlines Industry (Airlines Industry) departs from the established legal precedent of CCI. A two-step test was established by CCI to address algorithmic collusion. CCI started by looking for inter-entity human communication. In the second step, algorithms’ function was looked at. Shikha Roy v. Jet Airways (Shikha Roy), which also indicated the beginning of a proactive investigation against the algorithmic collusion in India, marked the logical conclusion of the test. In contrast to the Airlines Industry case, when CCI examined alleged algorithmic collusion following DG’s report, CCI directed DG to examine the role of algorithms during the course of the investigation in Shikha Roy.

The two cases mentioned above point to the two-step test’s standardization, which the CCI may find customary in its use when handling cases of algorithmic collusion. Even though it is still growing, the CCI’s legal doctrine falls short on two counts. First, the test is more precise when assessing whether the rivals utilized a similar algorithm when determining algorithmic collusion. Competing businesses may utilize different platforms or software, which could prevent the regulator from finding them liable. Second, the test still reflects the old methodology because it calls for measuring “intention”. The level of “human intervention” is one of three crucial aspects in the test to determine guilt, as stated in the Airlines Industry case. The cartels can evolve in markets where computers organise and control the exchange of pricing information, even if the competitor’s initial goal was not to engage in illegal anti-competitive agreements. It is extremely challenging to translate and incorporate subjective concepts into the digital environment, such as intent.

Pricing algorithms have been in use for a long time, making the identification of suspicious collusive behaviour a major issue. As a result, it is still interesting to consider whether CCI is comparing anomalies with other anomalies that have become the “new” normal since they have replaced the normal market conditions ages ago. In the era of pricing algorithms, it is crucial for antitrust regulation to keep track of market changes. The German Monopolies Commission expressly suggests regularly monitoring marketplaces for price anomalies because the determination of intention to collude in pricing algorithms is still an unresolved issue.

A way ahead

The price stability of competing products may come under scrutiny by authorities when pricing decisions are not immediately related to market factors like the cost of raw materials, transportation, marketing, etc. However, the collusive acts may not be completely traceable because it is challenging to apply conventional ideas of “agreements” to such an issue. In this regard, even if the price decisions are not traceable, the US Policy Council of the Association for Computing Machinery favours the imposition of an obligation on each user of an algorithm.

When it comes to pricing algorithms, a transition from intent to consumer harm must be made, and the magnitude of customer harm must become a valid and workable threshold for discriminating between illegal collusion and acceptable oligopoly activity.