Leniency Programmes Vs Algorithmic Collusion - An Equal Fight

Author: LawSphereX

Category: General Law

Published: 2026-08-23T20:32:40.807253+00:00

Updated: 2026-08-26T17:10:37.123+00:00

LENIENCY PROGRAMMES VS ALGORITHMIC

COLLUSION: AN EQUAL FIGHT?

Katyayani Shekhar, Student, Amity Law School, Amity University Noida Sonakshi Varshney, Assistant Professor, Amity Law School, Amity University Noida

ABSTRACT

The arrival of digital technologies has turned upside down the traditional competition law enforcement, challenging the effectiveness of the leniency programs that has since ages have been monitoring cartel behaviour. This paper aims to examine the growth of algorithm collusion where autonomous systems coordinate anti-competitive behaviour without explicit human involvement, and the profound implications it holds for antitrust regulation. Through careful analyses of incidents like Amazon Marketplace and European grocery markets, the learning reveals that traditional leniency frameworks, dependent on human whistleblowers, struggle to address the opacity, speed, and autonomy of algorithm-driven collusion.

It analyses global responses, including the European Union’s Digital Markets Act and proposed reforms under India’s Competition Amendment Bill, 2022, highlighting the regulatory gaps and enforcement difficulties posed by algorithmic collusion. A comparative review of EU, US, and Indian competition law models further emphasizes the need for innovation-centric, dynamic regulatory approaches. The study concludes that without proactive reforms—such as algorithm auditing, enhanced transparency requirements, and cross-border cooperation—regulators risk falling behind rapidly evolving digital markets, ultimately undermining consumer welfare and fair competition.

LENIENCY PROGRAMMES VS ALGORITHMIC COLLUSION: AN EQUAL

FIGHT?

Evolution of Leniency Programs Leniency programs have been instrumental in the battle against cartelization in conventional industries, acting as an efficient whistle-blower tool that disrupts illegal pacts between competing businesses. But the age of digitalisation, with their advanced algorithms and data driven decision making, have presented challenges to the relevance of these programs. Creating a new category of cartels, algorithmic collusion where algorithms facilitate anti-competitive behaviour without human involvement has reshaped the composition of cartels and highlighted deficiencies in current regulatory systems. The advent of leniency programs in jurisdictions such as the United States, European Union, and India led to a paradigm shift in antitrust enforcement.

The premise behind these programs is simple: Offer cartel operatives immunity or lower sentences in exchange for information, even if it means cooperating with authorities. These programs have taken down some of the most infamous cartels in the world. 1 Case Studies of Success: Lysine Cartel (1996): Whistleblowing led the U. S. Department of Justice (DOJ) to uncover

Cooperation in full: Companies need to demonstrate strong proof and cooperate with investigations according to investigations Worldwide Approach: More than 80 jurisdictions have adopted leniency programs (e.g., the Indian regime implemented under Section 46 of the Competition Act, 2002).

2. Algorithmic Collusion: The Myth becomes Reality

On traditional markets, cartels depended on written agreements between human agents. But the digital era has birthed algorithmic collusion, where automated systems analyse pricing strategies, track competitors, and alter prices in real-time without explicit human involvement.4 Definitional features of algorithmic collusion: Tacit Coordination Without Communication: Algorithms use public data (e.g., competitor prices) to reach collusive outcomes without explicit agreements. Speed and Precision: Algorithms react in a split-second to changes in the market making collusion more effective and durable. The Effects Are Obscure & Complex: Proprietary coding & AI are major obstacles to regulatory scrutiny. Algorithmic Collusion: Some Examples The Amazon Marketplace Case (2015): The DOJ looked into e-commerce companies that

Regulatory Breaches stemming from Algorithmic Collusion Traditional leniency programs are not well-placed to take proper account of the nuances of algorithmic collusion for a number of reasons:5 No Whistleblowers: Algorithms do not possess human agency, so they do not defect and testify for evidence. Legal Ambiguity: Existing legislation, including the Sherman Act in the United States and the Competition Act in India, does not specifically prohibit tacit collusion that is enabled via algorithms. Technical Barriers: Investigators may not have the means or expertise to decipher complex AI models. Algorithmic collusion has led to some recent reforms, notably proposed under the European Commission's Digital Markets Act. This includes accident audits of algorithms and increased fines for non-compliance.

The FTC has called for amending the Sherman Act to clarify that it applies to algorithmic collusion. In line, the CCI has initiated studies on algorithmic pricing and is also considering amending the Competition Act, 2002, to fill the gaps in the regulation of digital markets. 6 The results from leniency programs show their power pace in conventional industries and weakness in deterring algorithmic collusion, indicating the necessity for the evolution of enforcement strategy in antitrust. (Collective action: Some solutions, like banning certain kinds of algorithms, require money and political action. ) In this digital battleground, the antitrust regulators must adapt as quickly as the technology they are trying to regulate.

The Next Tech Revolution AI and Advances in Algorithms have Disrupted Global Markets, Innovating and Making Them More Efficient But these developments have also created new and unexpected challenges for competition enforcement. Algorithmic 5 Gal, Michal S. , Algorithmic Challenges to Autonomous Competition, Journal of Antitrust Enforcement, 201. 8, 2020, pp. 295–322. 6 Competition Commission of India, “Market Study on E-commerce in India”, CCI Report (2020), available at https://www. cci. gov. in/ (last visited Jan 5,2025)

collusion that allows such anticompetitive practices as price-fixing and market allocation to occur through autonomous or semi-autonomous algorithmically-enabled vehicles has emerged as a matter of urgent concern. Whereas traditional cartels rely on explicit agreements among humans, algorithmic collusion typically happens without direct human involvement, by virtue of the speed, complexity, and opacity of modern digital systems. This transition poses an enormous challenge for governments around the world. 7 Algorithmic collusion occurs when companies use artificial intelligence and machinelearning algorithms to coordinate their actions so as to achieve outcomes similar to what an explicit collusion would achieve, including higher prices or less output, but without any explicit deal among the participants.

It is dependent on the algorithm's computational capacity to forecast, accommodate, and adjust to conditions in the marketplace, which includes rival activity. In such a scenario, this phenomenon commonly occurs in infrastructural industries featuring digital platforms, dynamic pricing equations, and abundant real-time data, such as e-commerce and ride-hailing services. Fundamentals of Algorithmic Collusion: Autonomy: Algorithms are independent, they don't require human oversight to decide between conflicting goals without human action. Quick and Accurate: Algorithms analyse vast amounts of data in real-time, allowing for

surge pricing, which unintentionally leads to identical pricing across competitors during peak hours. Competition algorithms 8keep track of competitors’ prices in real time, adjusting their own prices to them, thus forming a stable pricing balance. Take, for instance, e-commerce platforms that provide pricing algorithms to customers to protect against being undercut by their competitors, while still allowing profitability. 9 Example. , In a case involving Amazon’s marketplace, for example, sellers had used pricing algorithms that matched competitors’ prices, thus eliminating price wars in the process and artificially inflating consumer costs. Hub-and-Spoke Models: Such arrangements are called hub-and-spoke conspiracies: one centralized source (the hub) interacts with a number of other players (the spokes), allowing them to act in concert (for instance, charging the same prices).

This is evident in platforms like Uber and Ola that are ride-hailing services, which operate under central algorithms to determine the fare and often result in even identical pricing between competing providers. Machine learning algorithms, especially reinforcement learning, can automatically find strategies that maximize profit even collusive behaviour. Such systems adjust their price points based on market feedback to achieve stable, high-price equilibriums. A study in science showed that reinforcement-learning algorithms empowered to maximize

profit will, on their own, develop collusive pricing strategies, even without overt instructions to do so. Challenges Brought About by Algorithmic Collusion Algorithmic collusion poses multiple challenges to traditional models of competition law enforcement.10 8 Gal, Michal S., “Algorithmic Challenges to Competition Law,” Journal of Antitrust Enforcement, 2019, vol. 6, 2020, pp. 109–135. 9 Ezrachi, Ariel and Stucke, Maurice E., “Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy”, Harvard University Press, 2016.

10 Id at 67.

Lack of Formal Agreements: Whereas traditional cartels commonly involve explicit agreements between parties, a clear smoking gun of collusion, algorithmic collusion often does not. The absence of human intent adds a layer of complexity to applying existing antitrust laws. Determining Liability: A basic question behind algorithmic collusion is accountability: who is responsible, the

deploying company, the developers, or the algorithms themselves? Legal structures like U.S. legislation like the Sherman Act or India's Competition Act, 2002 do not contemplate action by way of AI. Opacity of Algorithms: Many algorithms are black boxes: their proprietary coding and machine learning processes are not easily interpretable. Regulators have a steep challenge ahead to understand how algorithms work in collusion, and whether their behaviour constitutes collusion. Speed and Scale: Quicker than traditional enforcement mechanisms, algorithms can track and react to changes in the market in real time. Their magnitude enables them to affect worldwide markets quickly, magnifying their influence. The complexity of Algorithmic Collusion: Case Studies Amazon Marketplace (2015):

were no explicit agreements discovered; the algorithms were clearly colluded tacitly by keeping prices identical. 11 Indian E-Commerce Platforms: According to reports from the Competition Commission of India (CCI), the pricing algorithm offered on websites such as Amazon and Flipkart. Instead, these algorithms automatically tracked and matched their competitors' prices, and this collusion gave the illusion of competition and kept prices elevated. Responses to Algorithmic Collusion: Antitrust Regulators have been increasingly pressing for mandated disclosures about how algorithms function. This includes requiring companies to disclose information about their algorithmic decision-making processes and their potential impacts on the market. Further, auditing algorithms by third parties before they go operational (sandboxing) can identify and address collusive tendencies.

However, because digital markets are global, the international cooperation is needed. International organizations such as OECD and ICN are developing harmonized frameworks for algorithmic collusion. Countries are rethinking antitrust laws for an AI-driven market. However, the European Union’s Digital Markets Act (DMA) aims to combat anticompetitive behaviour in digital markets, including algorithmic collusion. 12 The Competition Amendment Bill, 2022, India, seeks to set up a framework to regulate digital markets, most of which are dominated by AI, data-driven businesses. Indeed, algorithmic collusion is a major and ongoing challenge of the digital age. As AI and algorithms facilitate efficiency and innovation, risks of anticompetitive behaviour remain.

It is very much a tightrope that regulators have to walk in encouraging innovation whilst ensuring a level playing field, which may require proactive action from regulators such as transparency requirements, increased auditing, and new legal frameworks. How, 11 European Commission, “Digital Markets Act: Addressing Collusive Behaviour in the Digital Age,” EC Policy Paper (2021). 12 OECD, “Algorithms and Collusion: Competition Policy in the Digital Age,” OECD Report, 2017, available at https://www. oecd. org/ (last visited Jan 5, 2025)

as the market algorithmic intersects consumers, we regulate will set the course for future competition in the market and consumer welfare. Comparative Analysis Globally, competition law frameworks work to promote a free market by outlawing anticompetitive practices. Such standards and tests differ across jurisdictions in light of economic and political conditions on the ground. The competition law regimes in the EU, US and India have developed in relation to their own unique priorities and methodology, particularly in relation to the evaluation of the effect on the competitive landscape of Mergers & Acquisitions (M&A).

The European Union (EU) competition law system, largely based on the Treaty on the Functioning of the European Union (TFEU) and administered by the European Commission (EC), is centred around accepting the SIEC (Significant Impediments to Effective Competition) test enshrined in the Merger Regulation. 13 Under the SIEC framework, the EC reviewed this $2. 1 billion deal and expressed concern that, by acquiring Fitbit, Google will be able to use the health data of Fitbit to enhance its monopoly power in digital advertising. The merger was only approved after Google accepted strict behavioural commitments to ensure data interoperability and protect consumer privacy. 14 Criticisms: More specifically, the SIEC test can be unduly restrictive by unnecessary hindering procompetitiveness mergers. Behavioural remedies typically need to be monitored consistently, making them difficult to enforce.

In the United States, antitrust enforcement is grounded in the Sherman Act (1890), the Clayton Act (1914) and the Federal Trade Commission Act (1914). The foundation of US 13 Neven, Damien, “How Important is the SIEC Test? A European Perspective,” Journal of Competition Law & Economics, Vol. 6, No. 3, 2010, pp. 549–570.

14 Ibid

antitrust jurisprudence is the Consumer Welfare Standard, which assesses whether business practices tend to diminish consumer prices or quality, and to curtail innovation. Some Key Features of the Consumer Welfare Standard: Price-Centric Analysis: A merger is reviewed to see that it will not result in higher prices or less output. This is

especially true in industries where pricing is a key competitive parameter. Efficiency Gains: The standard looks to whether the merger produces efficiencies that can, ultimately, benefit consumers, in the form of lower production costs or better products. Focus on Consumer Impact: Whereas the EU focuses on structural changes, the US approach prioritizes immediate consumer impacts over abstract competition fears. Thumbing up this 3. 7 billion merger, the Federal Trade Commission (FTC) argued that the combination would help consumers via better logistics and lower costs. Critics, however, said the merger increased Amazon’s market clout in retail and groceries and set off long-term competition worries. 15 The Competition Act, 2002, administered by the Competition Commission of India (CCI), governs India’s competition law framework.

Although the Indian policy takes inspiration from global practices, it also lays strong emphasis on competition in innovation markets and concerns about digital ecosystems. 16 Main Features of India’s Model: Innovation Market Analysis:

The CCI is responsible for ensuring that mergers do not affect innovation markets, especially in industries such as pharmaceutical, technology, and e-commerce. This can include assessing the R&D capabilities and potential innovation pipelines, and the consumer access to innovative products. Key Dynamic Market Considerations: India’s approach acknowledges the dynamic features of competition in digital markets, which include entry barriers, network effects and data monopolization. Behavioural and Structural Remedies: The CCI utilizes a combination of behavioural and structural remedies to mitigate competition concerns, like requiring data-sharing agreements or divestments. The $206 million purchase removed Uber Eats from the competition, with Zomato and Swiggy emerging as the two dominant players in India’s food delivery market.

The union was approved by the CCI after evaluating its impact on consumer choice, pricing and barriers to entry in the market. 17 The EU, US and Indian approaches showcase their domestic priorities and market realities. The EU prioritizes market structure and prolonged competition, while the US values immediate consumer benefits. India’s stance is one of moderation, tailored to the unique challenges of these markets for digital and innovation. Now, the international cooperation needed to harmonise these diverse frameworks will be key to ensure that global markets are effectively regulated. Future of Merger Control The challenge for merger control going forward will be to adapt to the rapidly changing landscape of digitized and innovation-driven markets.

Antitrust paradigms that have been successful in the 20th century do not apply well to these unconventional markets because of their unique features, including network effects, the crucial role of data, and fast pace of innovation. Reforms proposed in that report focus on fostering market dynamics and

17 Ibid

the role of data-driven valuation metrics in analysing mergers. These reforms are designed to facilitate the ability of competition authorities to regulate mergers to promote innovation and preserve competition. In determining if a merger is likely to substantially lessen competition, we evaluate competitive effects in terms of how the merger would affect rapid product innovation, economic growth related to innovation, and long-term consumer welfare. In contrast to static analyses that focus on having an instantaneous effect on price, dynamic market analysis has a more expansive view and takes into account a wider range of factors, including barriers to market entry, innovation pipelines and future competitive dynamics.

18 Dynamic Analysis of Markets — Major Elements: Focus on Innovation Markets: Today, mergers in areas like technology, pharmaceuticals and artificial intelligence are often evaluated based on their R&D capabilities and innovation impact. One notable instance relates to the merger between the pharmaceutical giants Pfizer and Wyeth, which was criticized for its possible impacts on drug discovery pipelines. 19 Long-Term Market Impact: Dynamic analysis takes into account the effects of mergers on the time-is-dynamic nature of market structures, e. g. , monopoly creation or strengthening. Predictive Models:

in assessing the impact of the merger. Authorities in multiple jurisdictions pointed to the low likelihood of continued competition for cloud gaming services. At Open Data Institute we see data becoming more and more central to the valuation of companies in digital markets. The purchase price of many tech companies captures not only their existing assets, but also their data stores, which can provide considerable competitive edge. Data as Key Factor in Merge Assessments Monetary Value of Data: It has become a central asset for many companies, shaping valuations and the dynamics of mergers. For instance, Fitbit’s vast health data was a key factor behind Google’s acquisition of the company in 2021. Competitive Advantage: Data helps companies to enhance products, micro-target customers, and control markets.

Mergers that concentrate data in the hands of a few firms can also lead to monopolistic outcomes. Data Portability and Interoperability:

Changes to Merger Control Proposed Embracing Dynamic Market Models: This would point to the need for competition authorities to incorporate models of dynamic competition in terms of innovation and market entry barriers, or technological disruption. For example, this preventive approach is present in the EU’s Digital Markets Act, which explicitly highlights the need to avoid market foreclosure within digital ecosystems. Guidelines for Data-Driven Mergers: The updated guidelines should consider the data role in valuations and competitive dynamics. These guidelines can contain:21 Measures for the valuation of data stockpiles. Technical and institutional standards for data accessibility and portability. Rules to prevent data concentration from undermining competition.

Cross-Border Cooperation: With the global character of these digital markets, cooperation between the competition authorities is of vital importance. Common frameworks can help to address issues such as forum shopping and jurisdictional overlaps. Enhanced Expertise and Tools:

Stronger behavioural Remedies: To the extent that structural remedies are fall short, behavioural remedies (e. g. , commitments to ensure data interoperability) should be added. For example, Google’s commitments in the Fitbit acquisition set a precedent for using behavioural remedies to alleviate concerns related to data. Dynamic, evidence-based merger control approaches are integral to the next era of merger control in digital markets. With an emphasis on long-term market dynamics and the role of data at the centre of this new economy of “power” competition authorities are in a position to strike that balance of promoting innovation while guarding against the monopolistic outcomes the market can create.

The need for these reforms is not just necessary, they are urgent: The pace of technological change and the growing dominance of digital giants demand foreseeability and principles that allow us to think through future challenges.