Competition Law and Regulation of Algorithmically Coordinated Markets

1. Introduction

Algorithmically coordinated markets are markets in which firms use pricing, bidding, recommendation, allocation, or monitoring algorithms that interact with market data and competitors’ conduct. Algorithms can increase efficiency by adjusting prices rapidly, forecasting demand, reducing transaction costs, and improving inventory management. However, the same technology can facilitate coordination, price alignment, information exchange, exclusionary conduct, or tacitly sustained supra-competitive prices.

The central competition-law question is not simply whether an algorithm produced similar prices. The critical question is whether firms have used technology as a mechanism for an agreement, concerted practice, exchange of competitively sensitive information, exclusionary strategy, or other conduct prohibited by competition law.

The OECD has identified algorithmic collusion as an important competition issue while also emphasizing that algorithms can have substantial pro-competitive effects.

2. Meaning of Algorithmic Coordination

Algorithmic coordination occurs when algorithms used by competing firms contribute to a reduction in independent competitive decision-making.

It may arise through several models.

A. Messenger model

Human competitors first reach an unlawful agreement and then use algorithms to implement it.

Example: Two sellers agree not to undercut each other and configure repricing software accordingly.

This is the easiest case legally because the underlying human agreement supplies the conventional evidence of collusion.

B. Hub-and-spoke algorithmic coordination

Several competitors use the same platform or software provider, which becomes the intermediary through which competitively sensitive information or pricing instructions circulate.

The central question is whether the hub merely provides neutral software or facilitates coordination among competitors.

C. Common-algorithm model

Several competitors use the same algorithm that receives information from competing firms and generates pricing recommendations.

The legal risk increases where the algorithm incorporates non-public competitor information and causes competitors to abandon independent pricing decisions.

D. Autonomous or tacit algorithmic coordination

Independent algorithms observe market conditions and react to each other's behaviour without an express human agreement.

This is the most difficult category because conventional competition law generally requires evidence of an agreement, concerted practice, or unilateral abusive conduct rather than merely parallel outcomes.

The OECD has noted that algorithmic pricing can increase market transparency and interaction frequency, potentially making coordination easier to sustain.

3. Why Algorithms Create Competition Concerns

3.1 Increased price transparency

Algorithms can continuously monitor competitors' prices.

Traditional human monitoring may occur once a day or once a week. Automated systems can monitor competitors almost instantaneously.

This may make it easier to:

• detect deviations;
• punish discounting;
• match price reductions;
• maintain common price levels;
• stabilize coordination.

The European Commission's e-commerce inquiry found extensive use of automated price-monitoring and price-adjustment systems and identified circumstances in which real-time information could facilitate automated coordination.

3.2 Increased frequency of interaction

Algorithms can make thousands of pricing decisions much faster than humans.

Repeated interaction can make deviations from a coordinated outcome easier to detect and punish.

For example:

Firm A lowers price → Firm B's algorithm detects it → B immediately lowers price → A's algorithm responds → both algorithms converge again.

Even where such behaviour is not itself unlawful, competition authorities may examine whether the underlying design or communications created an unlawful coordination mechanism.

3.3 Use of competitively sensitive information

Algorithmic coordination becomes particularly problematic where competitors provide a common software provider with:

• current prices;
• future pricing plans;
• discounts;
• inventory information;
• customer information;
• occupancy rates;
• production capacity;
• margins;
• strategic business information.

The RealPage litigation illustrates this concern: the U.S. Department of Justice alleged that competing landlords supplied non-public information that was incorporated into algorithmic rental-price recommendations. A proposed 2025 settlement required RealPage to stop using competitors' non-public competitively sensitive information in runtime pricing and redesign specified features.

4. Applicable Competition-Law Principles

A. Prohibition of agreements between competitors

Under systems such as:

• Article 101 TFEU in the European Union;
• Section 1 Sherman Act in the United States;
• Chapter I Competition Act 1998 in the United Kingdom;
• Section 3 Competition Act 2002 in India;

agreements or concerted practices restricting competition may be prohibited.

The fact that the agreement is executed through software does not automatically make it lawful.

B. Hub-and-spoke arrangements

Where competitors use a common platform or intermediary, authorities may examine whether:

1. competitors knew of the coordination mechanism;
2. sensitive information was exchanged;
3. the intermediary facilitated the exchange;
4. participants accepted or implemented the common strategy;
5. the arrangement reduced independent decision-making.

C. Abuse of dominance

Algorithms can also create unilateral competition concerns.

A dominant firm could potentially use algorithms to:

• self-preference its products;
• discriminate against competitors;
• deny interoperability;
• degrade rivals' access;
• manipulate rankings;
• impose exclusionary rebates;
• engage in predatory pricing;
• exploit data advantages;
• foreclose competing platforms.

Thus, algorithmic competition law is broader than algorithmic price fixing.

5. Important Case Laws

1. United States v. Topkins, No. CR 15-00201 (N.D. Cal. 2015)

Facts

David Topkins and other online sellers participated in an arrangement concerning prices of posters sold on Amazon Marketplace.

The participants used pricing algorithms to implement the agreed pricing strategy.

Legal principle

The case demonstrates the messenger model of algorithmic collusion.

The unlawful conduct was not transformed into lawful conduct merely because software performed the pricing adjustments.

Significance

The case established an important practical proposition:

An algorithm can be the instrument through which a conventional cartel operates.

Competition authorities therefore examine communications, configuration instructions, code, pricing rules, and internal documents—not merely the final prices.

6. Trod Ltd. and GB eye Ltd. — UK CMA (2016)

This is one of the leading UK algorithmic-pricing cases.

Two online sellers of posters and frames agreed not to undercut one another on Amazon Marketplace.

They used automated repricing software to implement the arrangement. The CMA found that the sellers had infringed competition law; Trod was fined £163,371, while GB eye received immunity after reporting the cartel and cooperating.

Legal principle

The use of automated software did not remove liability for the underlying price-fixing agreement.

Importance

The case demonstrates:

Human agreement → algorithmic implementation → coordinated prices → competition-law liability.

It remains an important illustration of the distinction between:

• lawful independent algorithmic pricing; and
• algorithmically implemented cartel conduct.

7. Eturas UAB v. Lietuvos Respublikos konkurencijos taryba, Case C-74/14 (CJEU, 2016)

Facts

Travel agencies used the Eturas online booking platform.

The platform administrator sent a system message indicating a restriction on online discounts and technically implemented a maximum discount.

The European Court of Justice considered whether the participating travel agencies could be regarded as engaging in a concerted practice.

Legal principle

The mere receipt of a platform message is not automatically sufficient to establish participation in an unlawful concerted practice.

However, knowledge of the restriction combined with circumstances demonstrating participation or failure to distance oneself may be relevant to establishing concerted conduct.

Importance

Eturas is particularly significant for platform-mediated algorithmic coordination.

It demonstrates that authorities must examine:

• what competitors knew;
• what the platform communicated;
• whether firms responded;
• whether firms accepted the restriction;
• whether independent competitive conduct continued.

8. Samir Agrawal v. Competition Commission of India, (2021) 3 SCC 136

Facts

The case concerned allegations relating to algorithmic pricing used by ride-hailing platforms, particularly Ola and Uber.

The allegation was that algorithmically determined fares reduced the ability of individual drivers to independently compete on price.

Legal principle

The Supreme Court's analysis is important because algorithmic price determination does not by itself establish a cartel.

There must be evidence satisfying the statutory requirements for an agreement or concerted practice.

Importance

The case illustrates an essential distinction:

Common pricing mechanism ≠ automatically illegal price fixing.

Competition authorities must identify the legally relevant relationship and evidence of coordination.

This is particularly important in digital platform markets where independent service providers may operate through a common technological infrastructure.

9. United States and State Plaintiffs v. RealPage, Inc.

Facts

The U.S. Department of Justice filed an antitrust action against RealPage concerning algorithmic pricing in rental housing.

The complaint alleged that competing landlords provided RealPage with non-public, competitively sensitive rental information and that this information was used to generate pricing recommendations.

The DOJ alleged violations of Sections 1 and 2 of the Sherman Act.

Subsequent development

In November 2025, the DOJ announced a proposed settlement requiring RealPage to stop using competitors' non-public competitively sensitive information for runtime rental-price determination and to redesign features that allegedly limited price decreases or aligned pricing.

Further settlements involving major landlords followed in 2025–2026. For example, proposed settlements involving Greystar, LivCor, Willow Bridge and Pinnacle included restrictions on algorithmic coordination and sharing competitively sensitive information.

Importance

RealPage is particularly significant because it moves beyond the simple “human cartel + algorithm” model.

The controversy concerns whether a common algorithm and shared sensitive data can function as a mechanism for coordinating competitors even where the traditional cartel structure is less visible.

10. Apple Inc. v. Pepper, 587 U.S. 273 (2019) — Digital Platform Competition Context

Although not an algorithmic-collusion case, Apple v. Pepper is relevant to understanding competition problems in algorithmically mediated markets.

The case concerned Apple's App Store and the relationship between the platform, app developers and consumers.

Competition-law significance

Digital platforms can simultaneously act as:

• intermediary;
• marketplace;
• rule-maker;
• data collector;
• ranking mechanism;
• transaction processor.

This creates potential competition concerns where the platform controls the technological infrastructure through which rivals compete.

Relevance to algorithmic markets

An algorithmically coordinated market should therefore be examined not only for price coordination but also for platform governance and control over competitive parameters.

11. United States v. Apple Inc. — Algorithmic and Platform Competition Context

The U.S. government's 2024 antitrust case against Apple provides another useful framework for algorithmically mediated competition.

The case concerns alleged exclusionary practices involving Apple's control over the iPhone ecosystem.

Although it is not principally a case of algorithmic price collusion, it demonstrates the broader principle that technological architecture can constitute an important mechanism of competitive exclusion.

The relevant lesson is that competition law may scrutinize:

• interoperability restrictions;
• access conditions;
• technological restrictions;
• platform rules;
• control over ecosystem functionality.

Thus, algorithmic competition regulation extends beyond price algorithms.

12. Excel Crop Care Ltd. v. Competition Commission of India, (2017) 8 SCC 47

This Indian Supreme Court decision concerned cartelisation and the assessment of competitive harm.

It is not an algorithmic-coordination case, but it provides useful principles for analysing coordinated conduct.

Relevance

The decision emphasizes the need to examine:

• market circumstances;
• actual competitive effects;
• relevant participants;
• economic evidence;
• proportionality of penalties.

In algorithmic cases, such analysis can be particularly important because similar prices may arise from legitimate market responses rather than unlawful coordination.

13. Rajasthan Cylinders and Containers Ltd. v. Union of India, (2018) 13 SCC 472

Again, this is not specifically an algorithmic case, but it is useful to the evidentiary analysis.

Principle

Parallel pricing or similar commercial behaviour does not automatically establish cartelisation.

Authorities should examine whether there are additional circumstances demonstrating coordination.

Importance for algorithms

This principle becomes particularly important when algorithms produce:

• identical prices;
• synchronized price changes;
• similar discounts;
• similar output decisions.

Algorithmic parallelism should not automatically be equated with an unlawful agreement.

14. Algorithmic Coordination: Main Legal Tests

A competition authority generally needs to investigate several questions.

Step 1 — Are the firms competitors?

If the firms do not compete in the relevant market, ordinary horizontal cartel analysis may not apply in the same way.

Step 2 — What does the algorithm actually do?

The authority should understand:

• inputs;
• objectives;
• constraints;
• pricing rules;
• training data;
• reinforcement mechanisms;
• feedback loops;
• output controls.

Step 3 — What information is exchanged?

Particular attention should be given to:

• current prices;
• future prices;
• costs;
• inventory;
• capacity;
• discounts;
• margins;
• customer data.

Step 4 — Is there human involvement?

Investigators should examine:

• emails;
• WhatsApp/Slack/Teams communications;
• contracts;
• software specifications;
• developer instructions;
• board documents;
• implementation manuals.

Step 5 — Is there an agreement or concerted practice?

This remains the critical legal issue in many jurisdictions.

Step 6 — Did the algorithm merely respond independently?

Independent algorithmic reaction to publicly observable prices may be materially different from an arrangement in which competitors deliberately provide confidential information to a common algorithm.

15. Types of Algorithmic Competition Concerns

16. Tacit Collusion and the Difficult Legal Problem

The most difficult question is:

What happens when algorithms independently learn to coordinate without an explicit human agreement?

Suppose:

• Firm A uses Algorithm A;
• Firm B uses Algorithm B;
• neither communicates with the other;
• both algorithms observe market prices;
• both learn that aggressive price reductions reduce profits;
• both eventually maintain higher prices.

There may be economic coordination, but proving a traditional legal agreement may be difficult.

This produces a distinction between:

Economic coordination

Algorithms reach a stable coordinated outcome.

Legal collusion

There is sufficient evidence of conduct satisfying the relevant jurisdiction's legal test for an agreement, concerted practice, or other prohibited conduct.

These concepts should not automatically be treated as identical.

The OECD has specifically recognized that the legal status of autonomous algorithmic behaviour remains an important unresolved issue in some jurisdictions.

17. Evidence in Algorithmic Competition Investigations

Competition authorities increasingly need technical evidence.

A. Source-code evidence

Authorities may investigate:

• pricing rules;
• optimization functions;
• constraints;
• decision trees;
• model architecture.

B. Training-data evidence

Particular attention may be given to whether competitors' confidential data was used.

C. Audit logs

Logs can reveal:

• price changes;
• algorithmic decisions;
• input variables;
• communications;
• override decisions.

D. Internal communications

Human evidence remains highly important.

An email stating that competitors should configure their pricing systems in a particular way may be much more probative than thousands of apparently parallel prices.

E. Econometric evidence

Authorities can examine:

• price convergence;
• frequency of price changes;
• margins;
• deviation responses;
• structural breaks;
• effects of algorithm deployment.

F. Counterfactual analysis

Authorities may compare:

algorithmic market outcome

with

competitive market outcome without the relevant coordination mechanism.

18. Regulation and Remedies

Competition authorities can employ several remedies.

18.1 Prohibition of information sharing

Competitors may be prohibited from supplying current competitively sensitive information to a common pricing system.

18.2 Algorithmic auditing

Independent audits can examine whether algorithms:

• incorporate competitor data;
• implement exclusionary rules;
• systematically prevent discounting;
• favour particular firms.

18.3 Data firewalls

Competitively sensitive information can be segregated so that one competitor cannot obtain or influence another competitor's information.

18.4 Human oversight

Firms may be required to maintain human review over important pricing or market-allocation decisions.

18.5 Algorithm certification

In high-risk markets, third-party algorithms could potentially be subject to certification or compliance verification.

Recent U.S. RealPage settlements illustrate this direction: proposed orders have included restrictions on competitively sensitive data and monitoring requirements for certain third-party pricing algorithms.

18.6 Structural remedies

Where algorithmic coordination is combined with market power and exclusionary conduct, authorities may consider:

• divestiture;
• interoperability;
• access obligations;
• data portability;
• separation of business functions.

19. Compliance Framework for Businesses

Companies using algorithmic pricing should establish an Algorithmic Competition Compliance Programme.

Before deployment

1. Identify competitors.
2. Identify all data sources.
3. Classify competitively sensitive information.
4. Determine whether competitor information enters the model.
5. Review pricing objectives.
6. Document independent decision-making.

During operation

1. Maintain algorithm logs.
2. Monitor unusual price convergence.
3. Restrict employee communications concerning competitor pricing.
4. Prevent unauthorized data sharing.
5. Conduct periodic legal and technical audits.

After deployment

1. Preserve audit trails.
2. Review model modifications.
3. Investigate unexplained coordinated behaviour.
4. Maintain records showing legitimate business justification.

20. Distinction Between Lawful and Potentially Unlawful Algorithmic Pricing

The crucial distinction is independent competitive decision-making versus coordinated decision-making.

21. Key Case-Law Principles — Consolidated

The first five are particularly useful for an answer specifically focused on algorithmic coordination; the remaining cases provide broader doctrinal support.

22. Emerging Regulatory Approach

The current regulatory direction can be summarized as:

Algorithms are not inherently anti-competitive.

They can improve:

• efficiency;
• demand forecasting;
• inventory management;
• consumer search;
• price discovery;
• allocation of resources.

But regulators increasingly focus on whether algorithms:

facilitate an agreement → exchange sensitive information → reduce independent pricing → stabilize coordination → exclude competitors → or exploit market power.

The OECD's recent work similarly identifies both efficiency-enhancing and anti-competitive possibilities and emphasizes investigation of the specific circumstances rather than treating algorithmic pricing itself as unlawful.

23. Conclusion

Competition law does not generally prohibit firms from using algorithms to compete. The central principle remains independent competitive decision-making.

The major legal challenge arises when algorithms transform traditional coordination into a faster, less visible and potentially more persistent mechanism. The cases involving Topkins, Trod/GB eye, Eturas, Samir Agrawal and RealPage demonstrate different points along this spectrum—from an ordinary cartel implemented by software to sophisticated common-algorithm and information-sharing theories.