Competition Law and Simulation Governance Competition Issues

1. Introduction

Simulation governance refers to the use of simulated environments, digital twins, computational models, artificial intelligence, algorithms, virtual marketplaces, or controlled “sandbox” environments to design, test, predict, or manage commercial conduct before it is implemented in the real market.

In competition law, simulation governance creates an important question:

When firms use simulations to model competitors, prices, consumer behaviour, market entry, allocation, or platform rules, can the simulation itself become a vehicle for anticompetitive coordination or exclusion?

The issue is particularly significant where competing firms use a common algorithm, platform, data pool, digital twin, market simulator, or third-party optimisation system. Competition authorities increasingly examine algorithmic decision-making because algorithms can facilitate coordination even where traditional human-to-human communications are difficult to identify. The OECD has specifically recognised that algorithms create challenges for traditional concepts of agreement and tacit collusion.

China's amended Anti-Monopoly Law is also particularly relevant because Article 9 expressly addresses the use of advantages in data, algorithms, technology, capital and platform rules for anticompetitive practices.

2. Meaning of Simulation Governance

Simulation governance can operate at several levels:

A. Market simulation

A company constructs a virtual market to predict:

• competitor reactions;
• pricing;
• demand;
• market entry;
• consumer switching;
• capacity;
• supply shortages;
• effects of mergers.

B. Algorithmic simulation

AI or machine-learning systems simulate competing strategies and recommend conduct likely to maximise profits.

C. Digital-twin governance

A digital replica of a market, infrastructure system, supply chain or platform is used to test commercial decisions.

D. Regulatory sandbox governance

A regulator permits businesses to test innovative technologies under controlled conditions.

E. Platform simulation

A digital platform simulates:

• ranking;
• recommendations;
• search results;
• seller allocation;
• pricing;
• advertising;
• consumer targeting.

F. Competitor-response simulation

A firm repeatedly models how identified competitors would respond to its pricing or strategic decisions.

The competition-law problem arises when the simulation stops being merely an internal forecasting tool and becomes a mechanism for coordinating market conduct.

3. Principal Competition-Law Concerns

I. Algorithmic Collusion

The most obvious risk is that simulations allow competitors to reach similar pricing or output decisions without conventional communication.

For example:

Competitors A, B and C provide competitively sensitive information to the same simulation provider. The provider's model predicts the price that maximises collective profitability. All three companies subsequently follow the recommended price.

Even if no executive says “let us fix prices,” the system may effectively facilitate coordinated behaviour.

The modern issue therefore becomes:

Does competition law require an express human agreement where an algorithm performs the coordinating function?

The answer depends upon the jurisdiction and evidence. The European Court of Justice has already demonstrated that automated systems can form part of the evidentiary chain establishing a concerted practice.

4. Case Laws

1. Eturas UAB and Others v Lietuvos Respublikos konkurencijos taryba

Case C-74/14, Court of Justice of the European Union, 2016

This is one of the most important cases for simulation and algorithmic governance.

Several travel agencies used a common computerised booking system. The system administrator sent a communication concerning a restriction on discounts, and the system automatically implemented a maximum discount.

The CJEU considered whether the automatic technical restriction could form part of a concerted practice.

The Court recognised that participation in a common electronic system, combined with knowledge of an anticompetitive mechanism and continued participation, could be relevant evidence of concerted conduct. However, the authorities still had to establish the necessary knowledge and participation of the individual undertakings.

Principle

A technological system does not necessarily become legally irrelevant merely because the coordination is implemented automatically.

Relevance to simulation governance

A common market simulator could similarly become evidence of coordination where:

• competitors participate in the same system;
• they know the system produces competitively sensitive recommendations;
• they receive or act upon those recommendations; and
• the system facilitates parallel market conduct.

5. RealPage Algorithmic Pricing Litigation — United States

In United States v. RealPage, Inc., filed in 2024, the U.S. Department of Justice alleged that RealPage's revenue-management software facilitated coordination among competing landlords.

The DOJ alleged that competing landlords supplied non-public information about rental rates and lease terms to RealPage, whose algorithm generated pricing recommendations using that information. The lawsuit alleged violations of Sections 1 and 2 of the Sherman Act.

Competition significance

The important issue is not simply that an algorithm recommends prices.

The concern is the combination of:

competitor data + common algorithm + pricing recommendations + competing users.

The DOJ alleged that landlords would otherwise independently compete over prices, discounts and lease terms.

Simulation-governance relevance

Suppose competing businesses use a common simulation platform to model:

• optimal market price;
• competitor response;
• supply allocation;
• discount levels.

The same legal concern can arise where the simulation effectively substitutes a common coordinating mechanism for independent decision-making.

6. Cornish-Adebiyi v. Caesars Entertainment — Algorithmic Hotel Pricing

In Cornish-Adebiyi v. Caesars Entertainment, U.S. antitrust authorities addressed allegations involving algorithmic hotel pricing.

The FTC and DOJ stated that competitors cannot use an algorithm to accomplish conduct that would be unlawful if performed directly by humans. They also explained that an unlawful arrangement does not necessarily require competitors to communicate directly with each other where an intermediary or algorithm facilitates the coordination.

Principle

Technology does not immunise otherwise unlawful coordination.

Simulation relevance

A simulation provider cannot necessarily be treated as a neutral technological intermediary if its system is being used to coordinate competing businesses.

7. Alibaba Group / SAMR — China

The Alibaba antitrust decision is highly relevant to simulation governance because it demonstrates the importance of platform architecture, algorithms, data and internal monitoring mechanisms.

In 2021, China's State Administration for Market Regulation found Alibaba to have abused its dominant position through the “choose one from two” strategy, restricting merchants from dealing with competing platforms. SAMR imposed a RMB 18.228 billion penalty.

The conduct involved platform mechanisms capable of monitoring and influencing merchants' behaviour.

Principle

Competition analysis can examine not only contractual terms but also:

• platform design;
• monitoring systems;
• ranking;
• incentives;
• penalties;
• technological enforcement.

Simulation-governance relevance

A simulated platform environment could be problematic where the platform tests different algorithmic mechanisms to determine the most effective way to prevent merchants from switching to competitors.

The simulation itself may therefore become relevant evidence of exclusionary strategy.

8. Meituan — China

In the Meituan antitrust investigation, SAMR examined the online food-delivery platform's alleged “choose one from two” exclusivity arrangements.

The investigation considered Meituan's ecosystem, market control, financial resources, technical conditions and barriers to entry, as well as mechanisms used to promote and monitor exclusivity.

Competition significance

The case illustrates that competition analysis in digital markets may extend beyond a single contractual restriction.

Authorities can consider:

ecosystem + data + technology + monitoring + platform dependence.

Simulation relevance

If a platform uses simulated environments to determine:

• which sellers should receive visibility;
• how competitors should be treated;
• which incentives maximise exclusivity;
• which algorithmic penalties reduce multi-homing,

those systems can potentially become relevant to an abuse-of-dominance analysis.

9. Amazon Marketplace / Buy Box — European Union and United Kingdom

The European Commission raised concerns concerning Amazon's use of third-party seller data and its Buy Box mechanism.

The Commission's preliminary assessment concerned the possibility that Amazon's systems favoured its own retail offers and sellers using Amazon's fulfilment services when selecting the prominently displayed Buy Box offer.

Separately, the UK's Competition and Markets Authority investigated Amazon's use of third-party seller data, Buy Box selection and Prime-related practices before accepting commitments.

Competition principle

An algorithmically governed marketplace can raise competition concerns where the operator controls the infrastructure while simultaneously competing with participants using that infrastructure.

Simulation-governance relevance

If a platform simulates alternative ranking or Buy Box rules and selects the rule that systematically disadvantages competing sellers, competition authorities may investigate the purpose, effects and implementation of that governance mechanism.

10. Apple – App Tracking Transparency (ATT) — France

In 2025, the French Competition Authority fined Apple €150 million concerning implementation of its App Tracking Transparency framework.

The Authority found that although data protection could constitute a legitimate objective, the particular implementation was neither necessary nor proportionate to that objective and amounted to an abuse of Apple's dominant position.

Earlier, the Authority had specifically considered whether Apple's rules might result in discriminatory treatment or self-preferencing.

Simulation-governance relevance

This case demonstrates that competition authorities may scrutinise system architecture and governance rules, not merely traditional contracts.

A simulation system that produces different competitive conditions for:

• the platform itself;
• affiliated businesses; and
• independent competitors

can therefore create self-preferencing or discrimination concerns.

11. Google Shopping — European Union

The Google Shopping litigation provides another important analogy.

The European Commission found that Google systematically favoured its comparison-shopping service in general search results and demoted competing comparison-shopping services.

The legal significance is that the competitive effect resulted substantially from the design and operation of an algorithmically governed platform rather than from a traditional contractual restriction.

Simulation-governance relevance

Where a dominant platform uses simulations to determine:

• ranking rules;
• visibility;
• recommendation algorithms;
• traffic allocation;
• advertising placement,

competition authorities may examine whether the resulting architecture systematically disadvantages rivals.

12. Simulation Governance as a Form of Digital Coordination

The principal risk can be represented as follows:

Competitors

↓

Common data environment

↓

Common simulation model

↓

Competitor-response prediction

↓

Recommended strategy

↓

Parallel implementation

↓

Reduced independent decision-making

This creates a spectrum:

Low-risk situation

Company independently simulates its own market.

↓

Intermediate-risk situation

Company uses publicly available market information.

↓

Higher-risk situation

Several competitors use the same commercially sensitive dataset.

↓

Very significant risk

Competitors use a common system that generates coordinated pricing or output recommendations.

The final category creates the strongest potential antitrust concerns.

13. Simulation Governance and Relevant Market Definition

Simulation technologies can also affect market definition.

Traditional market-definition techniques may become difficult where:

• products are customised;
• prices are personalised;
• consumers multi-home;
• markets are two-sided;
• AI predicts substitution;
• products change rapidly;
• virtual and physical markets interact.

Simulation models may be used to estimate:

• demand elasticity;
• diversion ratios;
• consumer switching;
• critical loss;
• hypothetical monopolist tests;
• network effects.

However, authorities must examine the assumptions underlying the model rather than automatically treating simulation outputs as factual market evidence.

14. Simulation Governance and Abuse of Dominance

A dominant platform could potentially use simulation to identify the most effective exclusionary strategy.

For example:

A dominant platform creates 10,000 simulated market scenarios and discovers that reducing a rival's visibility by 20% causes the greatest decline in rival traffic.

If it then implements that strategy, the competition issue is not necessarily the simulation itself.

The central questions become:

1. Does the undertaking possess dominance?
2. What conduct was implemented?
3. Did the conduct exclude rivals?
4. Was the conduct objectively justified?
5. Were efficiencies generated?
6. Could less restrictive alternatives achieve the same objective?

15. Simulation Governance and Merger Control

Simulation is increasingly relevant to merger analysis.

Merging parties may use simulations to predict:

• unilateral effects;
• coordinated effects;
• price increases;
• capacity reduction;
• innovation effects;
• customer switching;
• efficiencies.

Competition authorities can scrutinise whether the model:

• uses reliable data;
• properly represents competitive constraints;
• incorporates entry;
• accounts for multi-market interactions;
• understates competitive harm.

A merger simulation should therefore not be treated as conclusive evidence merely because it is mathematically sophisticated.

16. Simulation Governance and Information Exchange

One of the most important legal problems concerns inputs.

Consider:

The competition risk therefore often lies less in the software itself and more in what information enters the simulation.

17. Third-Party Simulation Providers

A particularly important emerging issue is the role of independent technology providers.

Imagine:

50 competing firms subscribe to the same AI market simulator.

The provider receives:

• pricing information;
• production data;
• demand forecasts;
• strategic plans.

The provider's algorithm then recommends market strategies.

The legal question becomes whether the provider is merely supplying software or is functioning as a facilitator of coordination.

The RealPage litigation illustrates why this distinction matters. The DOJ's allegations specifically focused on the combination of competitor information and common pricing software.

18. Simulation Governance and Self-Preferencing

A vertically integrated platform may use simulation to test how different governance rules affect its own products.

For example:

Platform owns marketplace + competing product

→ simulation tests ranking options

→ platform product receives higher visibility

→ independent competitors receive less traffic.

This can create concerns involving:

• self-preferencing;
• discriminatory access;
• leveraging;
• foreclosure;
• exclusionary product design.

The Amazon Buy Box investigation demonstrates the relevance of automated selection systems to this category of competition concerns.

19. Simulation Governance and Essential Facilities

A simulation environment may itself become strategically important where competitors depend upon it.

Potential examples include:

• common industry digital twins;
• shared logistics simulations;
• common charging-network models;
• financial-market simulators;
• cloud-based AI environments;
• industry-wide data platforms.

If a dominant provider controls a critical simulation infrastructure and denies access selectively, competition concerns may involve:

• refusal to deal;
• discriminatory access;
• interoperability;
• tying;
• leveraging;
• essential-facility principles.

20. Governance Remedies

Competition authorities may consider several remedies.

A. Data separation

Competitors should not receive competitors' confidential data.

B. Data aggregation

Sensitive data may be aggregated sufficiently to reduce coordination risks.

C. Independent audits

Algorithms can be periodically examined for discriminatory or collusive outputs.

D. Access controls

The system should prevent competitors from viewing competitively sensitive information.

E. Explainability

Users and regulators should understand important elements of algorithmic decision-making.

F. Human independent decision-making

Businesses should retain genuine independent pricing and strategic decisions rather than mechanically following common algorithmic recommendations.

G. Firewalls

Separate business units can be prevented from accessing competitively sensitive information.

H. Interoperability

Dominant simulation platforms may be required to permit reasonable access or interoperability where appropriate.

China's platform-governance literature similarly identifies algorithmic transparency, explainable AI and audit mechanisms as potential methods for managing competition risks while protecting legitimate trade secrets.

21. Compliance Framework for Simulation Governance

A competition-compliant simulation governance framework can be structured as:

Step 1 — Identify participants

Who operates or accesses the simulation?

↓

Step 2 — Classify data

Is the information public, historical, aggregated or competitively sensitive?

↓

Step 3 — Examine algorithm design

Does the model facilitate coordination or exclusion?

↓

Step 4 — Examine outputs

Does it recommend independent conduct or coordinated conduct?

↓

Step 5 — Examine implementation

Did firms actually act on the recommendations?

↓

Step 6 — Assess market power

Is any participant dominant?

↓

Step 7 — Assess competitive effects

Price, output, innovation, quality, entry and consumer choice.

↓

Step 8 — Test justification

Are efficiencies legitimate, verifiable and proportionate?

↓

Step 9 — Audit continuously

Monitor model updates, training data and outputs.

22. Six Core Legal Principles

The above cases establish several important principles:

Principle 1 — Automation does not eliminate antitrust responsibility

Eturas demonstrates the importance of examining automated systems in concerted-practice analysis.

Principle 2 — Algorithms can facilitate coordination

The RealPage litigation illustrates the competition risks associated with common algorithmic pricing systems using competitor information.

Principle 3 — A common intermediary can be legally significant

The hotel algorithm litigation illustrates that competitors cannot necessarily avoid antitrust scrutiny merely because coordination occurs through an algorithmic intermediary.

Principle 4 — Platform architecture can constitute a competition issue

Alibaba and Meituan demonstrate the importance of platform governance, monitoring and technological mechanisms in analysing digital-platform conduct.

Principle 5 — Algorithmic selection can create exclusionary concerns

Amazon's Buy Box investigation demonstrates how automated marketplace-selection mechanisms can attract competition scrutiny.

Principle 6 — Legitimate objectives do not automatically justify discriminatory implementation

The Apple ATT decision demonstrates that even a legitimate objective such as privacy protection can be examined for necessity, proportionality and discriminatory effects under competition law.

23. Emerging Legal Issues

Simulation governance will raise increasingly difficult questions concerning:

1. AI-to-AI coordination without human communication.
2. Digital twins used by competing firms.
3. Synthetic market data and whether it can facilitate coordination.
4. Reinforcement-learning agents learning to avoid competitive price reductions.
5. Autonomous pricing systems.
6. Common AI infrastructure used by competing undertakings.
7. Simulation of merger effects.
8. Regulatory sandboxes and competition neutrality.
9. Algorithmic self-preferencing.
10. Simulation-based exclusion of market entrants.
11. Competition risks from foundation-model ecosystems.
12. Auditing of black-box commercial algorithms.

Recent comparative scholarship notes that China, the EU and the United States are converging in recognising that algorithms, data and digital infrastructure create distinctive competition-law challenges, although their regulatory approaches differ.

24. Conclusion

Simulation governance is not inherently anticompetitive. Simulation is an important tool for innovation, forecasting, risk management and regulatory experimentation.

The competition-law problem arises when a simulation becomes a mechanism for:

• coordinating competitors;
• exchanging sensitive information;
• fixing or stabilising prices;
• allocating markets;
• restricting output;
• excluding rivals;
• discriminating against dependent businesses;
• facilitating self-preferencing; or
• entrenching the power of a dominant platform.

The central legal principle can therefore be stated as:

Competition law should examine the economic function of a simulation system rather than merely its technological form.

A system called a “simulation,” “digital twin,” “AI optimiser,” “sandbox,” or “decision-support tool” does not automatically fall outside antitrust law. The decisive questions are who controls it, what information it receives, what recommendations it produces, whether competitors participate, whether the recommendations influence actual conduct, and what effects the resulting governance structure has on competition.

Key Cases Covered

1. Eturas UAB v Lithuanian Competition Council, C-74/14 — automated systems and concerted practices.
2. United States v RealPage, Inc. — algorithmic pricing and competitor information.
3. Cornish-Adebiyi v Caesars Entertainment — algorithmic hotel-price coordination.
4. Alibaba / SAMR — platform governance and exclusionary “choose one from two.”
5. Meituan / SAMR — platform ecosystem, monitoring and exclusivity.
6. Amazon Marketplace / Buy Box — algorithmic selection and self-preferencing concerns.
7. Google Shopping — algorithmic ranking and exclusionary self-preferencing.
8. Apple ATT / French Competition Authority — digital governance, discrimination and proportionality.