Introduction
The traditional image of price-fixing — competitors meeting in a hotel room to carve up markets and fix prices — is increasingly anachronistic in the digital economy. Algorithmic pricing, where automated systems set prices in real time based on competitor pricing data, demand signals, inventory levels, and other inputs, is now standard across e-commerce, airline ticketing, ride-hailing, hotel booking, and financial markets. When multiple competitors use pricing algorithms that independently respond to each other’s prices — each raising prices when the other does and lowering them when the other does not — the result may be systematic price coordination that mimics the output of an explicit cartel agreement, without any human ever having communicated with a competitor.
This phenomenon, broadly termed “algorithmic collusion,” poses a direct challenge to the conceptual foundation of competition law’s treatment of horizontal agreements. India’s Competition Act 2002, like most competition law regimes, requires proof of an “agreement” between enterprises — a concurrence of wills — to establish an anti-competitive arrangement under Section 3. Where coordination emerges from parallel algorithmic responses to market information rather than from any human communication, the existence of an “agreement” in the legal sense is genuinely contested.
Legal Framework
Section 3(1) of the Competition Act 2002 prohibits agreements between enterprises that cause or are likely to cause an appreciable adverse effect on competition in India. Section 3(3) creates a presumption of appreciable adverse effect for horizontal agreements fixing prices, limiting supply, allocating markets, or rigging bids. The definition of “agreement” in Section 2(b) includes “any arrangement, understanding or action in concert, whether or not formal or in writing, or intended to be enforceable by legal proceedings.”
The breadth of the Section 2(b) definition — explicitly including “action in concert” — potentially accommodates algorithmic coordination that does not involve any explicit communication, if it can be characterised as “action in concert.” The Competition Commission of India’s decisional practice has historically interpreted “action in concert” broadly, finding concerted practices based on parallel conduct and market structure evidence without requiring proof of explicit communication. However, the CCI has not yet adjudicated a case where the coordination mechanism was purely algorithmic — where the parallel conduct was demonstrably the output of independently programmed pricing algorithms rather than any form of human agreement.
Judicial Developments
The European case law on algorithmic collusion, while not binding on Indian courts, provides the most developed analytical framework. The EU Court of Justice’s decision in Dyestuffs (1972) established the concept of “concerted practice” in EU competition law, covering coordination that, without rising to the level of agreement, “knowingly substitutes practical co-operation between undertakings for the risks of competition.” Under this standard, parallel pricing behaviour that can be explained only by the existence of prior coordination may constitute a concerted practice.
The Italian Competition Authority’s (AGCM’s) Eturas case, referred to the Court of Justice (Case C-74/14), addressed a situation where a common algorithm operated by a booking platform applied coordinated price caps to hotel booking commissions without the hotels having communicated with each other. The Court of Justice held that awareness of the algorithmic mechanism and participation in the platform could constitute tacit acquiescence sufficient for a concerted practice finding. This reasoning is the closest existing judicial guidance to the algorithmic collusion scenario, though it still involves human decision-making (deciding to participate in the platform) as a link.
In India, the CCI’s decisional practice in the airline fuel surcharge case and several other parallel pricing cases has applied a circumstantial evidence approach — inferring coordination from pricing parallelism combined with structural factors that make independent identical pricing unlikely. This approach is transferable to algorithmic pricing, in principle, though the evidentiary analysis becomes more technically complex.
Contemporary Issues and Analysis
The taxonomy of algorithmic collusion is essential to accurate legal analysis. Not all algorithmic price coordination raises competition concerns, and the competitive impact depends significantly on the mechanism:
“Messenger” algorithms that simply execute an explicit human agreement to coordinate pricing are traditional cartel facilitation through technology — the algorithm is a tool, and the agreement is among humans. This is the clearest case for competition law intervention and does not raise genuinely novel issues.
“Hub and spoke” algorithm cases — where competing firms use a common pricing algorithm provided by a third-party vendor, and that algorithm coordinates their prices — raise the question of whether the algorithm vendor and the using firms are parties to an agreement, or whether the using firms have entered an agreement with each other through the medium of the shared algorithm. The US Department of Justice’s 2024 indictment of RealPage, a property management software company whose algorithm was alleged to have facilitated landlord price coordination for rental properties, is the most significant real-world test of this theory.
Pure “parallel” algorithm cases — where each firm independently programmes its own algorithm to monitor and match competitor prices, without any common vendor or any form of communication — raise the hardest legal questions. If two hotels each independently programme their pricing algorithms to match the other’s price changes within 15 minutes, and prices consistently move in parallel as a result, is this an “agreement” or “concerted practice” under competition law? The case for intervention is strongest when the parallel response is to price increases (suggesting coordinated facilitation of supracompetitive pricing) and weakest when the parallel response is to price decreases (suggesting competitive matching that benefits consumers).
Comparative and International Perspective
The OECD’s 2017 report on Algorithms and Collusion is the foundational policy document, identifying the spectrum from algorithm-facilitated explicit collusion to tacit coordination through parallel algorithms and assessing the adequacy of existing competition law frameworks for each scenario. The OECD concluded that existing law is adequate for the messenger algorithm and hub-and-spoke scenarios but may be insufficient for pure parallel algorithm coordination, where the absence of communication prevents an “agreement” finding under traditional doctrine.
The UK Competition and Markets Authority (CMA) has conducted the most sustained work on algorithmic pricing, including a review of pricing algorithms across multiple sectors and guidance for businesses on compliance with competition law when using pricing algorithms. The CMA’s approach — focusing on the human decisions that programme and deploy algorithms as potential competition law violations — is more tractable than trying to characterise the algorithm’s output as the “agreement.”
Practical and Policy Implications
For e-commerce platforms and businesses that use pricing algorithms, the competition law compliance implications are significant. Algorithms that are programmed to monitor and respond to competitor prices — which describes most commercial pricing algorithms — potentially generate competition law risk if the competitor pricing being monitored is itself algorithmically set by a company that is monitoring the first firm’s prices. The circularity of this dynamic makes traditional compliance advice — “don’t communicate with competitors about prices” — insufficient.
Compliance programmes for algorithmic pricing should include: human review of algorithm design decisions for competition law implications, documentation of the independent business rationale for each algorithmic pricing parameter, and regular review of pricing outcomes for patterns that suggest coordination rather than independent competition.
Suggestions and Reforms
The Competition Commission of India should issue guidance on algorithmic pricing and competition law, specifying: (a) the factors relevant to determining whether parallel algorithmic pricing constitutes “action in concert” under Section 2(b); (b) the safe harbour for genuinely independent algorithmic pricing that responds to publicly available price signals; and (c) the circumstances in which the use of a common pricing algorithm vendor may constitute a hub-and-spoke arrangement under Section 3.
The CCI should also develop technical capacity for forensic analysis of pricing algorithm outputs — the ability to distinguish between coordination and independent competition through algorithmic pricing data analysis is a specialised econometric capability that the Commission currently lacks.
Conclusion
Algorithmic collusion is a genuine competition law problem that is becoming more pervasive as digital commerce expands. The legal framework for addressing it — built around the concept of “agreement” and “concerted practice” — is adaptable but requires thoughtful extension to accommodate the technological context. The primary challenge is not legislative but evidentiary and analytical: developing the forensic tools and economic methodologies to distinguish anti-competitive algorithmic coordination from pro-competitive algorithmic competition. India’s competition law regime has the conceptual resources to address this challenge; what it currently lacks is the decisional practice and technical capacity to apply those resources effectively.
