1. Introduction

Algorithmic competition refers to competition between businesses in markets where algorithms, artificial intelligence (AI), automated pricing systems, machine-learning models, and data-driven decision-making tools influence commercial behaviour.

Algorithms can help businesses determine prices, forecast demand, allocate resources, personalize offers, identify market opportunities, and respond rapidly to competitors. They can improve efficiency and intensify competition. However, algorithms can also facilitate price coordination, strengthen dominant market positions, discriminate between customers, and make anticompetitive conduct more difficult to detect.

Competition law must therefore examine not only agreements between businesses but also how automated systems affect market behaviour and competitive outcomes.

The central question is:

When does the use of algorithms represent legitimate independent competition, and when does it facilitate unlawful collusion, exclusionary conduct, or abuse of market power?

This question is particularly important in digital markets, where businesses may rely on real-time pricing, shared data, common software providers, recommendation systems, online marketplaces, and automated responses to competitors.

Theories of algorithmic competition provide analytical frameworks for understanding these developments. They help competition authorities, courts, businesses, and policymakers determine whether algorithmic conduct promotes rivalry or undermines the competitive process.

2. Meaning and Scope of Algorithmic Competition

Algorithmic competition concerns the interaction between automated decision-making systems and competitive market structures.

It includes several principal forms.

A. Algorithmic pricing

Businesses use algorithms to calculate or modify prices based on factors such as demand, inventory, competitor prices, customer behaviour, and market conditions.

For example, an online retailer may automatically reduce its price when a competing retailer lowers its price.

This may benefit consumers by intensifying price competition. However, if competing firms use algorithms to implement an unlawful pricing agreement, the same technology may facilitate collusion.

B. Algorithmic coordination

Algorithmic coordination occurs when automated systems produce parallel or mutually accommodating commercial behaviour.

Such coordination can arise through different mechanisms:

1. Explicit collusion: Competitors agree to coordinate prices or output and use algorithms to implement the agreement.
2. Algorithm-facilitated collusion: Businesses employ a common algorithm or pricing platform to implement or monitor a coordinated strategy.
3. Tacit coordination: Firms independently adjust their conduct in ways that sustain prices above competitive levels without necessarily communicating or entering into an unlawful agreement.
4. Autonomous algorithmic coordination: Self-learning systems may develop strategies that accommodate competitors, even where human operators have not explicitly programmed that outcome.

These categories must be distinguished carefully. Parallel pricing or the use of similar algorithms does not, by itself, establish an infringement of competition law.

C. Algorithmic exclusion

Algorithms may also be used to disadvantage competitors by manipulating rankings, restricting access to essential data, discriminating against rival sellers, or optimizing a dominant platform's systems in ways that exclude competing businesses.

The legality of such conduct depends on the applicable competition rules, the undertaking's market position, the effects on competition, and any legitimate business justification.

D. Algorithmic innovation and market entry

Algorithms can lower barriers to entry by allowing smaller businesses to automate activities previously requiring substantial resources.

However, advantages arising from exclusive datasets, computing infrastructure, network effects, switching costs, and economies of scale may reinforce incumbent market power.

Competition law must therefore assess both the procompetitive and anticompetitive possibilities of algorithmic technologies.

3. Major Theories of Algorithmic Competition

3.1. The Theory of Algorithmic Collusion

The theory of algorithmic collusion examines whether algorithms facilitate coordination between competing firms, potentially reducing competitive pressure without requiring conventional meetings or direct communications.

Traditional cartel arrangements often involve communications concerning prices, output, customers, or market allocation. Algorithms may make coordination easier by collecting information, monitoring competitors, and rapidly implementing agreed pricing strategies.

For example, several competing businesses might agree to maintain minimum prices and use automated systems to monitor compliance. The algorithms would make the agreement easier to implement, but the underlying infringement would be the unlawful agreement, not the mere use of software.

A more difficult question arises when firms independently adopt algorithms that learn to avoid price wars. Economic theory suggests that rapid monitoring and repeated interaction can sometimes make coordinated outcomes more sustainable.

Nevertheless, an economically coordinated outcome is not automatically a legally prohibited cartel. The applicable law may require proof of an agreement, concerted practice, or other legally specified conduct.

Competition-law significance: Authorities should investigate communications, contractual arrangements, common pricing instructions, information exchanges, algorithmic design, and market evidence rather than treating parallel algorithmic behaviour as conclusive proof of collusion.

3.2. The Theory of Tacit Algorithmic Coordination

Tacit coordination occurs when competitors independently recognize that maintaining higher prices may be more profitable than aggressively competing.

Algorithms may make this outcome easier to sustain because they can:

• Monitor competitors' prices continuously.
• Detect deviations from established pricing patterns.
• Adjust prices rapidly in response to market changes.
• Identify when a competitor begins discounting.
• Apply sophisticated strategies to repeated market interactions.

A market with a small number of competitors, transparent prices, frequent transactions, and stable demand may be particularly conducive to coordination.

However, the presence of these conditions does not establish an unlawful agreement. In some jurisdictions, unilateral conscious parallelism may not be prohibited unless additional legal requirements are satisfied.

The theory therefore reveals an important distinction between economic coordination and legally actionable collusion.

3.3. The Theory of the Digital Price Leader

Under this theory, one firm's algorithm may become a focal point for prices across a market.

A major platform or dominant retailer may adjust prices frequently in response to demand and competitive information. Other firms may independently follow those prices because doing so appears commercially rational.

A price leader can emerge without an explicit agreement between competitors.

Competition concerns arise if the pricing mechanism is used to facilitate an unlawful agreement, enforce a cartel, or support exclusionary conduct by a dominant firm.

The mere fact that competitors follow a market leader's prices is not sufficient to establish an infringement.

3.4. The Theory of Common Algorithmic Intermediaries

Several competitors may use the same third-party pricing software or optimization platform.

Such platforms can offer legitimate efficiency benefits, including lower operating costs, improved demand forecasting, and better inventory management.

Risks arise when a common intermediary:

• Collects commercially sensitive information from competing firms.
• Shares individualized future pricing intentions between competitors.
• Implements a coordinated pricing strategy.
• Enables firms to monitor compliance with a common arrangement.
• Uses contractual or technical mechanisms to restrict independent pricing decisions.

The intermediary's involvement may provide evidence of coordination, but its use does not automatically establish collusion.

Investigators must examine what information is collected, how it is processed, which participants can access it, and whether the arrangement restricts independent commercial decision-making.

3.5. The Theory of Self-Learning Algorithmic Competition

Machine-learning systems can adapt their strategies based on new data and observed outcomes.

Unlike traditional software that follows a fixed set of rules, some machine-learning systems alter their behaviour as they learn from market conditions.

This creates two competing possibilities.

Procompetitive effects:

• Faster responses to changing demand.
• Better price discovery.
• Lower operating costs.
• Improved allocation of resources.
• More effective competition from smaller businesses.

Anticompetitive risks:

• Automated price matching that weakens price rivalry.
• Strategies that discourage competitors from discounting.
• Discriminatory access to customers.
• Rapid exclusionary responses to new entrants.
• Reduced human oversight of market conduct.

The legal question is not simply whether an algorithm is autonomous. It is whether the relevant conduct satisfies the elements of an infringement under the applicable legal framework.

Businesses cannot necessarily avoid liability merely by claiming that an automated system generated the relevant decision.

3.6. The Theory of Algorithmic Market Transparency

Digital markets often make prices and commercial terms more visible than traditional markets.

Algorithms can process this information at a speed and scale unavailable to human decision-makers.

Greater transparency can help consumers compare prices and help businesses identify competitive opportunities. Yet, when competitors can observe and respond to one another's pricing almost instantaneously, transparency may also make coordination easier to sustain.

The effect depends on the market's structure, the nature of the information, and how firms use it.

Publicly available historical prices are not equivalent to confidential information about competitors' future pricing intentions. Competition authorities should distinguish between these categories when assessing potentially anticompetitive information exchange.

3.7. The Theory of Algorithmic Predation and Exclusion

Algorithms can support aggressive pricing strategies intended to weaken or eliminate competitors.

A dominant firm might use automated systems to identify vulnerable rivals, target particular customers, or selectively reduce prices in response to competitive threats.

Low prices are generally beneficial to consumers and should not automatically be treated as predatory.

A proper assessment may require examining:

• The firm's market power.
• The relationship between price and relevant costs.
• The duration and selectivity of below-cost pricing.
• The likelihood of excluding equally efficient competitors.
• The possibility of recouping losses, where relevant under the applicable legal test.
• The effects on consumers, innovation, and market entry.

Algorithmic exclusion can also occur without below-cost pricing, for example through discriminatory rankings, restrictions on interoperability, or strategic control over access to essential commercial data.

3.8. The Theory of Algorithmic Market Power

Market power in digital markets may derive from more than market share.

Important factors include network effects, access to data, switching costs, economies of scale, control over distribution, and the ability to influence the rules governing a digital platform.

Algorithms may reinforce these advantages by improving service quality as more data become available or by making a platform's products more attractive as its user base expands.

However, data possession and algorithmic sophistication do not automatically establish dominance or abuse.

A competition assessment must examine the relevant market, competitive constraints, entry conditions, customer alternatives, and the actual or likely effects of the conduct.

4. Important Case Laws on Algorithmic Competition

The following cases illustrate how competition law addresses algorithmic coordination, automated pricing, common intermediaries, information exchange, digital platforms, and market power.

Important qualification: Not every case discussed below involved artificial intelligence or a self-learning algorithm. Some concern traditional cartels, digital platforms, or pricing restrictions and are relevant because their legal principles can be applied to algorithm-driven markets. Decisions, judgments, regulatory findings, and settlements must also be distinguished according to their legal status.

Case 1: Eturas UAB and Others v Lietuvos Respublikos konkurencijos taryba (2016)

Court: Court of Justice of the European Union
Case number: C-74/14

Facts:

The case concerned a common online travel-booking system used by travel agencies. The system administrator sent a message indicating that discounts available through the system would be capped at 3%. Technical changes were subsequently made to implement the restriction.

The Lithuanian competition authority found that the travel agencies and the system administrator had participated in an unlawful agreement restricting discounts.

The dispute eventually reached the Court of Justice of the European Union.

Legal issues:

1. Could the use of a common computerized booking system facilitate a concerted practice?
2. Could travel agencies be presumed to have participated in coordination after receiving a message about the discount restriction?
3. What evidence was necessary to establish their awareness and participation?

Judgment and principles:

The Court held that Article 101 TFEU could apply where the circumstances established participation in a concerted practice implemented through a common computerized system.

However, the mere sending of a message did not establish that every recipient was aware of its contents. Where awareness and other relevant circumstances were established, participation could be presumed, subject to the possibility of rebutting that presumption.

Evidence capable of rebutting the presumption included publicly distancing oneself from the practice, reporting it to the authorities, or consistently applying discounts above the cap.

Relevance to algorithmic competition:

This is one of the most directly relevant European judgments for understanding algorithm-facilitated coordination.

It demonstrates that a common digital system can serve as a mechanism for implementing anticompetitive restrictions, even without conventional face-to-face cartel meetings.

It also establishes an important evidentiary safeguard: the use of a common platform does not automatically prove that every participant knowingly joined an unlawful arrangement.

The same principles are relevant when businesses receive automated pricing instructions or use shared software to implement a coordinated strategy.

Case 2: United States v. David Topkins (2015)

Court: United States District Court for the Northern District of California
Proceeding: Criminal antitrust prosecution

Facts:

The US Department of Justice investigated price fixing involving online sellers of posters and wall art.

The sellers used pricing algorithms to monitor and adjust prices on an online marketplace. The prosecution concerned an agreement between competing sellers to coordinate prices, with automated pricing tools used to implement the arrangement.

Topkins pleaded guilty to participating in the price-fixing conspiracy.

Legal issues:

The central question was whether competitors could be held responsible for conventional price fixing when they used automated pricing technology to implement their agreement.

Outcome and principles:

The prosecution established that the use of pricing software does not exempt a business from liability for participating in a price-fixing conspiracy.

The relevant prohibition was Section 1 of the Sherman Act, which addresses agreements that unreasonably restrain trade.

The case did not establish that independent algorithmic price matching is automatically unlawful. Its significance lies in the use of algorithms to implement an existing agreement.

Relevance to algorithmic competition:

The case demonstrates that competition authorities can apply established cartel rules to technologically sophisticated conduct.

For example, competing online retailers cannot lawfully agree to fix prices and then argue that their pricing systems, rather than their commercial decisions, produced the final prices.

The case also illustrates why investigators must examine communications, software configuration, commercial instructions, and evidence of coordination rather than relying exclusively on observed prices.

Key principle: Automation does not immunize an unlawful price-fixing agreement from antitrust liability.

Case 3: United States v. Apple Inc. (2013; appellate judgment in 2015)

Court: United States District Court for the Southern District of New York; United States Court of Appeals for the Second Circuit
Subject: E-book price fixing

Facts:

The US Department of Justice challenged arrangements involving Apple and major book publishers concerning the sale of electronic books.

The case arose during a period when publishers were concerned about the pricing practices of competing retailers, particularly Amazon.

The arrangements involved an agency-based pricing model under which publishers exercised greater control over retail prices, alongside contractual provisions affecting retailers' ability to discount.

Legal issues:

The central issue was whether the arrangements constituted an unlawful agreement to raise or stabilize e-book prices.

The dispute concerned coordinated commercial arrangements rather than a finding that a self-learning algorithm had independently created a cartel.

Judgment and principles:

The district court found that Apple had orchestrated and participated in a conspiracy with publishers to raise e-book prices. The Second Circuit affirmed the principal liability determination in 2015.

The case illustrates that a company may be liable for facilitating coordination among suppliers or competitors when the evidence establishes participation in an unlawful agreement.

Relevance to algorithmic competition:

The case is relevant to markets where pricing arrangements are implemented through digital platforms, automated contractual mechanisms, and centrally administered commercial systems.

A platform's role in organizing pricing relationships may require scrutiny if it facilitates an unlawful agreement among market participants.

However, a platform does not violate competition law merely because it introduces a new pricing model or provides technology that allows suppliers to control prices. Liability depends on the applicable legal test and the evidence of unlawful coordination.

Key principle: Digital intermediaries can face antitrust scrutiny when their conduct facilitates an unlawful pricing conspiracy.

Case 4: Meyer v. Kalanick (2016)

Court: United States District Court for the Southern District of New York
Subject: Algorithmic pricing and alleged horizontal coordination

Facts:

A consumer brought a proposed class action alleging that Uber's pricing model facilitated unlawful coordination among drivers.

The complaint challenged the use of Uber's pricing system, under which fares were calculated through a common platform rather than negotiated independently between each driver and passenger.

The plaintiff argued that drivers effectively participated in a price-fixing arrangement by accepting the platform's pricing mechanism.

Legal issues:

The dispute raised the question of whether competitors' use of a common pricing mechanism could support an antitrust claim when the arrangement allegedly limited independent price competition.

It also raised questions about whether a platform's pricing system could be treated as a mechanism through which competing service providers coordinated their commercial behaviour.

Procedural significance:

In 2016, the district court denied Uber's motion to dismiss the antitrust claim at that stage of the proceedings.

That decision allowed the claim to proceed; it was not a final determination that the alleged conspiracy had been proven.

Relevance to algorithmic competition:

The case illustrates the potential application of antitrust law to platform-mediated pricing.

It is particularly relevant where competing service providers rely on a common algorithm that determines prices or otherwise limits their independent pricing decisions.

The central issue is whether the arrangement represents legitimate platform organization or an unlawful agreement among economically competing participants.

The distinction matters because a common pricing system may improve efficiency, but it may also raise competition concerns depending on the nature of the agreement and the parties' economic relationships.

Key principle: A common algorithmic pricing mechanism can raise antitrust questions, but allegations must be distinguished from proven liability.

Case 5: VM Remonts and Others v Konkurences padome (2016)

Court: Court of Justice of the European Union
Case number: C-542/14

Facts:

The case concerned a public procurement procedure in Latvia and alleged coordination involving economic operators and an external service provider.

The dispute examined circumstances in which an undertaking's conduct may be attributable to it where coordination has occurred through a third party.

Legal issues:

The Court considered the conditions under which an undertaking could be held responsible for anticompetitive conduct facilitated by an independent service provider.

The case was not specifically about an AI pricing algorithm. Its relevance comes from its treatment of third-party involvement in competition infringements.

Judgment and principles:

The Court explained that liability for coordination through a third party depends on the applicable legal requirements and the undertaking's knowledge or circumstances surrounding the third party's conduct.

An undertaking cannot automatically be held liable merely because an independent service provider has engaged in unlawful conduct.

The evidence must establish the necessary connection between the undertaking and the infringement under the applicable rules.

Relevance to algorithmic competition:

The case is useful when examining businesses that outsource pricing or market optimization to external technology providers.

Suppose a retailer hires a software provider to manage its prices. The provider independently develops a system that exchanges competitively sensitive information with other clients.

Whether the retailer can be held liable would depend on the facts, including its knowledge, participation, and the applicable legal standards.

The same analysis may apply to shared AI providers, data intermediaries, and algorithmic consultants.

Key principle: The involvement of a third-party technology provider requires a fact-specific assessment of attribution and responsibility; third-party involvement alone does not establish liability.

Case 6: T-Mobile Netherlands BV and Others (2009)

Court: Court of Justice of the European Union
Case number: C-8/08

Facts:

The case concerned a meeting between mobile telecommunications operators at which competitively sensitive information relating to market conduct was exchanged.

The information concerned matters relevant to competition in the Dutch mobile telecommunications market.

Although this was not an algorithmic pricing case, the judgment is significant for understanding how information exchange can facilitate coordination.

Legal issues:

The Court considered whether an exchange of information between competitors could constitute a concerted practice under Article 101 TFEU.

Judgment and principles:

The Court emphasized that competitors must independently determine their market conduct.

An exchange of competitively sensitive information may reduce uncertainty about competitors' future behaviour and facilitate coordination.

In the circumstances addressed by the judgment, a single meeting could be sufficient to establish a concerted practice if the applicable legal requirements were met.

Relevance to algorithmic competition:

Algorithms can perform the information-processing functions that would otherwise require human communications.

For example, a shared pricing platform might collect competitors' intended price changes and use that information to recommend future prices.

If the arrangement facilitates an unlawful concerted practice, the fact that information is exchanged automatically rather than through a meeting does not necessarily remove the conduct from competition law.

However, the existence of a common data system or an exchange of information does not automatically establish an infringement. The content, sensitivity, purpose, and competitive context of the exchange must be examined.

Key principle: Information exchange that reduces strategic uncertainty between competitors can facilitate unlawful coordination, subject to the applicable legal requirements.

Case 7: Booking.com and Booking.com (Deutschland) GmbH, Case C-264/23 (2024)

Court: Court of Justice of the European Union
Subject: Online hotel-booking platforms and price-parity clauses

Facts:

Booking.com used contractual clauses restricting hotels from offering lower prices through certain alternative sales channels.

The dispute concerned both wide and narrow price-parity clauses. Wide clauses restricted lower prices on competing platforms and hotels' own channels, while narrow clauses principally restricted lower prices on hotels' own direct sales channels.

The case concerned contractual restrictions rather than a finding that algorithmic pricing itself was unlawful.

Legal issues:

The Court considered whether the parity clauses could be treated as ancillary restraints necessary for the operation of the booking platform.

Judgment and principles:

On 19 September 2024, the Court held that price-parity clauses of this kind could not, in principle, be regarded as ancillary restraints merely because they were intended to support the operation of the platform.

The Court emphasized that such restrictions could reduce competition between booking platforms and create risks for smaller platforms and new entrants.

Relevance to algorithmic competition:

Digital platforms frequently combine ranking algorithms, pricing tools, contractual restrictions, and automated monitoring.

This judgment demonstrates that competition authorities must assess the commercial restrictions surrounding a digital system, not merely the technical operation of its software.

For example, an algorithm that monitors compliance with price-parity obligations may raise different issues depending on whether the underlying contractual restriction is lawful.

The case also illustrates the importance of protecting competition between platforms, rather than focusing exclusively on price comparisons within a single platform.

Key principle: Digital-platform efficiencies do not automatically justify contractual restrictions that undermine competition between platforms.

5. Additional Relevant Competition-Law Authorities

The following authorities provide further principles that can help analyze algorithm-driven markets.

8. United States v. Apple Inc. — E-book pricing and coordinated market conduct

Beyond its significance as a digital-platform dispute, the case demonstrates the importance of distinguishing independent business decisions from coordinated conduct.

When algorithmic systems are used to administer contractual pricing arrangements, investigators should determine whether the arrangement facilitates an unlawful agreement or represents legitimate independent commercial activity.

9. Asus — European Commission resale-price-maintenance proceedings (2018)

The European Commission imposed fines on several electronics manufacturers, including Asus, for restricting the online resale prices of retailers.

The proceedings concerned restrictions on resale prices rather than a finding that the manufacturers had engaged in autonomous algorithmic collusion.

They are relevant because digital monitoring tools can help suppliers detect and respond to retailers' discounting.

The legal issue is whether such monitoring supports unlawful resale-price maintenance, not whether price monitoring itself is prohibited.

10. Competition and Markets Authority v. Compare The Market — UK price-comparison platforms

The UK Competition and Markets Authority investigated and imposed a penalty on ComparetheMarket concerning wide most-favoured-nation clauses in contracts with home insurers.

The restrictions limited insurers' ability to offer lower prices through competing channels.

This matter is relevant to algorithmic competition because digital comparison platforms rely on ranking systems, pricing data, and contractual arrangements that influence competition across distribution channels.

The principal concern was the restrictive contractual arrangement, rather than an independently established infringement involving a self-learning algorithm.

6. Comparative Analysis of the Major Theories

The principal theories can be compared as follows.

These theories are complementary rather than mutually exclusive. A single digital market may exhibit several of these characteristics simultaneously.

For instance, a dominant platform may use an algorithm to optimize prices, monitor competing businesses, rank sellers, and enforce contractual restrictions. A proper assessment should distinguish each form of conduct and apply the appropriate legal test.

7. Application of Algorithmic Competition Theory Under Indian Competition Law

India provides a useful example of how established competition-law principles can be applied to algorithm-driven markets.

The Competition Act, 2002, does not require an anticompetitive arrangement to take the form of a traditional face-to-face agreement.

Three provisions are particularly relevant.

A. Section 3: Anti-competitive agreements

Section 3 prohibits agreements that cause or are likely to cause an appreciable adverse effect on competition in India, subject to the statutory framework.

Section 3(3) addresses specified agreements between competitors, including price fixing, market allocation, and bid rigging.

An algorithm used to implement an unlawful agreement may form part of the evidence establishing the infringement.

For example, if competing sellers agree to maintain identical minimum prices and use a common algorithm to implement that agreement, the fact that prices are adjusted automatically does not necessarily prevent the arrangement from falling within Section 3.

However, independent adoption of similar software or parallel pricing is not, by itself, sufficient to establish an unlawful agreement.

B. Section 4: Abuse of dominant position

Section 4 prohibits the abuse of a dominant position.

Algorithmic conduct may be relevant where a dominant enterprise:

• Imposes unfair or discriminatory conditions.
• Engages in unfair or discriminatory pricing within the statutory framework.
• Restricts market access.
• Uses its position to exclude competitors.
• Leverages dominance from one relevant market into another in a manner prohibited by the Act.

A dominant digital platform may, for example, use its ranking algorithm to disadvantage competing businesses.

The relevant inquiry is whether the conduct constitutes abuse under Section 4, not merely whether the platform has a sophisticated algorithm.

C. Sections 19 and 26: Investigation and assessment

Section 19 sets out matters relevant to assessing anti-competitive agreements and abuse of dominance, including market conditions and the statutory factors relevant to the conduct under examination.

Section 26 provides the framework for the Competition Commission of India's consideration of information and initiation of investigative processes.

In algorithm-related investigations, relevant evidence may include:

• Pricing histories and transaction records.
• Software documentation and algorithmic instructions.
• Communications between competitors.
• Contracts with common software providers.
• Data-sharing arrangements.
• Changes in pricing following particular communications.
• Evidence of exclusionary effects or barriers to entry.

Investigators should also consider legitimate explanations for similar pricing, including common input costs, demand changes, and independently chosen commercial strategies.

Indian-law qualification: The statutory provisions can apply to algorithm-mediated conduct, but the existence of an algorithmic mechanism does not itself establish a contravention. The relevant statutory elements and evidence must be examined.

8. Economic Effects of Algorithmic Competition

Algorithmic competition can produce both positive and negative economic effects.

A. Consumer welfare

Algorithms may help consumers obtain lower prices, more relevant recommendations, faster service, and better product availability.

Conversely, automated pricing and personalization may create risks where businesses use customer data to charge different prices or exploit limited consumer alternatives.

Price differentiation is not necessarily unlawful. Its legality depends on the applicable legal framework, market circumstances, and the nature and effects of the conduct.

B. Innovation

Algorithmic competition can encourage businesses to develop better products, improve efficiency, and introduce new services.

However, entrenched algorithmic advantages may discourage entry if new competitors cannot access essential data, infrastructure, or distribution channels.

Competition authorities should therefore assess whether market advantages result from legitimate innovation or are reinforced by unlawful exclusionary practices.

C. Market entry

Automated pricing tools can help small businesses compete with larger firms by reducing operating costs.

At the same time, proprietary datasets, network effects, and economies of scale can make it difficult for entrants to match established platforms.

The competitive assessment should consider both the efficiencies created by algorithms and the barriers that may prevent rivals from challenging incumbent firms.

D. Market transparency

Real-time pricing information can make markets more responsive.

However, extensive transparency between competitors can sometimes facilitate coordination by making deviations from a common strategy easier to detect.

The central policy challenge is to preserve information that benefits consumers while preventing arrangements that unlawfully exchange sensitive commercial information or facilitate cartel behaviour.

9. Challenges in Enforcing Competition Law Against Algorithmic Conduct

9.1. Establishing an agreement

Algorithms may produce similar pricing outcomes without direct communication between competitors.

Authorities must distinguish independent parallel behaviour from legally prohibited coordination.

This requires careful examination of the evidence rather than an assumption that similar prices establish a cartel.

9.2. Identifying responsibility

Where several businesses use a common software provider, determining responsibility may be difficult.

The investigation may need to establish who designed the system, what instructions were provided, what information was exchanged, and whether the businesses knowingly participated in an unlawful arrangement.

Responsibility depends on the applicable law and the particular facts.

9.3. Understanding machine-learning systems

Some systems adapt their decisions over time, making their behaviour difficult to reconstruct.

Competition authorities may need access to technical documentation, model inputs, pricing records, system logs, and information about the objectives used to train or configure the system.

Technical complexity does not remove the need to prove the legal elements of an infringement.

9.4. Distinguishing aggressive competition from exclusion

Automated price reductions may benefit consumers even when they put substantial pressure on competitors.

Authorities must avoid treating every aggressive response as predatory conduct.

The analysis should focus on the applicable legal criteria, market power, pricing evidence, exclusionary effects, and relevant justifications.

9.5. Developing effective remedies

Where an infringement is established, remedies may include ending the unlawful agreement, removing restrictive contractual terms, changing information-sharing arrangements, or imposing penalties where authorized by law.

In suitable cases, compliance measures may include independent pricing controls, monitoring systems, and audits designed to prevent recurrence.

Remedies should address the identified infringement without unnecessarily preventing legitimate algorithmic innovation.

10. Principles for Effective Competition Governance

A sound framework for algorithmic competition should incorporate the following principles.

1. Technological neutrality: Apply competition law to the substance of commercial conduct, regardless of whether decisions are made by humans, conventional software, or AI systems.

2. Evidence-based enforcement: Establish infringements through reliable evidence of the conduct and the required legal elements.

3. Protection of independent decision-making: Prevent unlawful coordination while preserving the ability of businesses to make independent commercial decisions.

4. Proportionate transparency: Examine whether the collection and exchange of pricing information helps consumers or facilitates coordination between competitors.

5. Accountability for automated conduct: Businesses should maintain appropriate oversight of systems used to make commercially significant decisions.

6. Protection of market entry: Assess whether algorithmic advantages reflect legitimate efficiency or unlawful practices that prevent competitors from accessing customers and markets.

7. Procedural fairness: Ensure that businesses can challenge allegations, explain algorithmic outcomes, and contest the evidence relied upon by authorities.

8. International cooperation: Digital markets often operate across borders. Cooperation between competition authorities can help address conduct involving common platforms, international sellers, and shared technology providers.

11. Conclusion

Theories of algorithmic competition explain how automated pricing, machine learning, digital intermediaries, market transparency, and data-driven business models can transform competitive relationships.

The principal challenge is distinguishing legitimate technological innovation from conduct that unlawfully restricts competition.

The cases discussed establish several important propositions:

1. Eturas demonstrates that a common computerized system can facilitate a concerted practice, while emphasizing the need for appropriate evidence of participation.
2. United States v. Topkins shows that algorithms used to implement price fixing do not shield participants from antitrust liability.
3. United States v. Apple illustrates how digital commercial arrangements can facilitate unlawful price coordination.
4. Meyer v. Kalanick demonstrates the potential antitrust issues arising from platform-mediated pricing, while requiring care to distinguish procedural rulings from final findings of liability.
5. VM Remonts provides guidance on the role of third parties in competition infringements and the conditions for attributing responsibility.
6. T-Mobile Netherlands explains how exchanges of competitively sensitive information can facilitate coordination.
7. Booking.com illustrates the competition risks associated with contractual restrictions imposed by digital platforms.

Taken together, these authorities show that algorithmic competition is not a separate exemption from established competition-law principles. Rather, it is an evolving field in which established doctrines must be applied to new technological and economic circumstances.

An effective competition-law framework must protect independent commercial decision-making, prevent unlawful coordination, preserve opportunities for market entry, and encourage innovation without assuming that every algorithmic practice is either inherently beneficial or inherently anticompetitive.