Competition Law and Simulation-Based Competitive Advantages

1. Introduction

Simulation-based competitive advantage refers to the ability of an undertaking to use computer simulations, digital twins, predictive models, artificial intelligence, machine learning, scenario modelling, or algorithmic experimentation to make better competitive decisions than rivals.

Examples include:

• simulating consumer demand before changing prices;
• modelling competitors' likely responses to a price reduction;
• using digital twins to optimise production capacity;
• simulating supply-chain disruptions;
• predicting customer switching;
• testing thousands of pricing strategies;
• modelling merger effects;
• forecasting congestion or network capacity;
• simulating advertising-auction outcomes; and
• using competitor or marketplace data to improve one's own products.

Competition law does not generally prohibit obtaining a competitive advantage through superior simulation technology. The legal issue arises when the simulation system is used to facilitate exclusionary conduct, coordination, discrimination, exploitation, information misuse, or an otherwise unlawful restriction of competition.

Modern competition authorities increasingly examine algorithms and data-driven systems because technological sophistication does not change the underlying competition-law assessment. The FTC and DOJ, for example, have expressly stated that using an algorithm does not make conduct lawful if the same conduct would constitute unlawful price fixing when performed by humans.

2. Meaning of Simulation-Based Competitive Advantage

A firm can create competitive advantage through simulation in several ways.

A. Pricing simulation

A company may simulate:

“If we reduce price by 5%, how will customers respond, and how will competitors react?”

This can produce legitimate efficiencies.

However, competition concerns arise if competing firms use a common algorithm or common pricing intermediary that effectively coordinates prices.

B. Demand simulation

A platform may use historical transactions to predict:

• consumer demand;
• peak purchasing periods;
• geographic demand;
• price elasticity;
• customer churn;
• competitor entry.

This may constitute normal competition based on superior analytics.

C. Capacity simulation

Manufacturers, airlines, electricity suppliers and logistics companies can simulate:

• production capacity;
• transportation bottlenecks;
• inventory;
• network utilisation;
• infrastructure constraints.

Such systems may produce efficiencies while simultaneously creating opportunities for exclusionary conduct.

D. Competitor-response simulation

A dominant undertaking might model:

“What happens if a rival cuts prices?”

The model itself is ordinarily not unlawful. But the resulting conduct could raise concerns if it involves predatory pricing, exclusionary rebates, discriminatory access, retaliation, or coordinated conduct.

E. Digital-twin competition

A digital twin reproduces a physical or economic environment digitally so that alternative strategies can be tested before implementation.

For example:

Physical market → Digital model → Scenario testing → Prediction → Commercial decision

A digital twin could therefore become an important competitive asset where rivals lack equivalent data or computational capabilities.

3. Central Competition-Law Principle

The important distinction is:

Lawful competitive advantage

“We have better technology and therefore compete more efficiently.”

versus

Potentially unlawful competitive advantage

“We use technology to exclude rivals, coordinate behaviour, exploit non-public information, or distort competitive conditions.”

Competition law normally protects competition on the merits, not competitors from technological superiority.

Thus, a company normally should not be penalised merely because:

• its simulations are more accurate;
• its algorithms are faster;
• its models use more computing power;
• it has invested more in R&D;
• it has better forecasting;
• it has superior logistics optimisation.

The legal question is how that advantage is obtained and exercised.

4. Major Competition-Law Issues

I. Algorithmic Coordination

Simulation systems may predict competitor behaviour with increasing accuracy.

If several competitors independently use sophisticated algorithms, they might reach similar prices without traditional communications.

This creates a difficult distinction between:

parallel competitive behaviour

and

concerted coordination.

The existence of similar prices alone normally does not establish an agreement. Authorities may examine communications, algorithm design, common intermediaries, data inputs, contractual arrangements and evidence of coordination.

The FTC and DOJ have specifically emphasised that algorithmic mechanisms cannot be used to circumvent conventional prohibitions on price fixing.

5. Information Advantage

Simulation depends heavily upon data.

A firm possessing:

• competitors' prices;
• customer-level information;
• transaction histories;
• inventory information;
• seller data;
• bidding information;
• demand forecasts

may obtain an advantage unavailable to competitors.

The competition-law problem becomes particularly serious where a dominant platform obtains non-public competitor information through its intermediary position and then uses that information to compete against the businesses supplying it.

This issue was central to the European Commission's Amazon Marketplace investigation. The Commission's preliminary concerns involved Amazon's use of non-public data obtained from third-party sellers to benefit its own retail business.

6. Simulation and Predatory Pricing

Simulation can make exclusionary pricing more sophisticated.

A dominant company may simulate:

1. rival's cost structure;
2. likely response to price cuts;
3. expected customer migration;
4. duration of rival's financial pressure;
5. probability of market exit;
6. post-exit price increases.

A model could therefore be used to identify the most effective strategy for weakening a competitor.

The simulation itself is not necessarily unlawful. The subsequent pricing strategy may nevertheless attract scrutiny under rules concerning predatory or exclusionary pricing.

The FTC has specifically discussed the possibility that pricing algorithms can make predatory pricing more feasible and effective.

7. Simulation and Self-Preferencing

A vertically integrated digital platform may simulate:

• ranking outcomes;
• consumer click-through rates;
• conversion probabilities;
• competitor performance;
• placement of its own products.

It may then alter its ranking system.

The issue becomes whether the platform uses its control over an important intermediary to systematically favour its own downstream services.

The Google Shopping litigation is particularly relevant because the Court of Justice examined Google's preferential treatment of its own specialised search results and the resulting competitive effects.

8. Simulation and Network Effects

Simulation becomes particularly powerful in markets with:

• network effects;
• economies of scale;
• data advantages;
• switching costs;
• interoperability barriers.

A dominant platform can simulate the effects of:

• changing APIs;
• modifying rankings;
• increasing switching costs;
• altering interoperability;
• changing commission rates;
• introducing loyalty incentives.

If the model enables the firm to reinforce an existing network advantage and foreclose competitors, authorities may examine the conduct under abuse-of-dominance rules.

9. Simulation and Merger Analysis

Simulation is increasingly relevant to merger control.

Economic models can estimate:

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

For example:

Firm A + Firm B → simulate post-merger market → estimate prices/output/innovation → compare with counterfactual.

Simulation therefore operates on both sides of merger review.

Authorities may use simulation to test:

• whether the merger substantially lessens competition;
• whether products are close substitutes;
• whether efficiencies offset competitive harm;
• whether coordinated effects are plausible.

Merging parties may use it to demonstrate:

• efficiencies;
• increased capacity;
• lower marginal costs;
• improved innovation;
• stronger competition against larger rivals.

10. Simulation and Essential Facilities

A dominant infrastructure provider may simulate network capacity and determine how much access competitors receive.

Examples include:

• telecommunications;
• electricity grids;
• payment systems;
• cloud infrastructure;
• transport networks;
• digital platforms.

A concern arises where simulation is used to create apparently technical justifications for denying or degrading access.

The authority may therefore examine whether:

“technical optimisation”

is genuinely based on capacity and efficiency or is being used as a mechanism of exclusion.

11. Simulation and Discriminatory Treatment

A sophisticated platform can simulate each user's or seller's likely behaviour.

It could theoretically apply different:

• prices;
• rankings;
• commissions;
• advertising requirements;
• access conditions;
• recommendations.

Competition law may become relevant where discriminatory treatment is capable of harming competition, particularly where a dominant undertaking discriminates between similarly situated trading partners to disadvantage rivals.

12. Simulation and Consumer Data

Simulation models often depend upon large quantities of personal or behavioural data.

Data protection and competition law can therefore intersect.

A firm that combines:

massive data + superior computing + predictive simulation

may obtain an advantage that competitors cannot easily reproduce.

Competition authorities may examine whether the data advantage:

• creates barriers to entry;
• prevents effective competition;
• facilitates exclusion;
• strengthens market power;
• supports discriminatory conduct.

However, having more data is not automatically an antitrust violation.

13. Simulation-Based Competitive Advantage and the Counterfactual

One of the most important concepts is the counterfactual.

Authorities may ask:

What would competition have looked like without the challenged conduct?

Simulation can actually help answer that question.

For example:

Actual world

Dominant firm changes ranking algorithm.

↓

Rival's traffic falls 40%.

Counterfactual simulation

Ranking algorithm treats rival neutrally.

↓

Rival's traffic remains substantially higher.

↓

Potential evidence of competitive foreclosure.

The CJEU's Google Shopping judgment illustrates the importance of analysing the competitive counterfactual and causal relationship between the conduct and its effects.

14. Six Important Case Laws

1. United States v. Apple Inc. — E-books

Court: U.S. District Court, Southern District of New York / Second Circuit

Apple concerned coordination in the e-books market.

Relevance to simulation

The case illustrates the fundamental distinction between:

• independent strategic decision-making; and
• coordinated conduct facilitated through interactions among market participants.

For simulation-based systems, sophisticated modelling cannot be treated as an independent justification if the underlying commercial strategy is the product of an unlawful agreement.

Principle

Technology does not immunise coordinated conduct from antitrust scrutiny.

15. Eturas UAB v Lietuvos Respublikos Konkurencijos Taryba

Court: Court of Justice of the European Union

This is particularly important for algorithmic competition.

An online travel-booking platform distributed a message through its system concerning limits on discounts available to travel agencies.

The CJEU considered when businesses using the platform could be regarded as participating in concerted practices.

Relevance

The case demonstrates that:

• electronic systems can facilitate concerted practices;
• algorithmic or technological communication can have competition-law significance;
• knowledge and participation remain important;
• merely using the same platform does not automatically establish liability.

Simulation relevance

If several competitors use a common simulation or pricing system, investigators would examine whether the system merely provides independent analytical tools or actually facilitates coordinated conduct.

16. Google Shopping

Case: Google and Alphabet v European Commission

The case concerned Google's treatment of its own specialised comparison-shopping service within general search results.

The CJEU's 2024 judgment addressed whether Google's conduct constituted an abuse of dominant position and examined issues including competitive effects and causation.

Relevance to simulation

A dominant platform can use:

• ranking models;
• prediction;
• consumer behaviour data;
• optimisation algorithms;
• simulated click behaviour

to improve its own service.

The competition-law issue is not the existence of an algorithm but whether the resulting conduct gives the dominant undertaking an unlawful advantage through exclusionary self-preferencing.

17. Amazon Marketplace

Case: European Commission, Amazon Marketplace

The Commission's investigation concerned Amazon's use of non-public information obtained from third-party sellers operating on its marketplace.

The concern was that such information could be used to benefit Amazon's own retail activities.

Relevance to simulation

This is highly relevant to simulation-based competitive advantages.

Suppose a platform receives:

• seller sales data;
• inventory information;
• customer demand;
• conversion rates;
• pricing information.

It could feed those inputs into predictive models and simulate which products it should launch, stock, price or promote.

Competition issue

The critical question becomes whether the platform's informational advantage results from legitimate competition or from exploiting its intermediary position to obtain and use competitors' confidential commercial information.

18. FTC v Amazon

The U.S. FTC and state authorities sued Amazon alleging a series of practices designed to maintain monopoly power, including alleged anti-discounting mechanisms and manipulation of marketplace conditions. These remain allegations in litigation, not final findings on every claim.

Relevance to simulation

Large platforms can continuously experiment with:

• search rankings;
• seller visibility;
• pricing;
• advertising;
• fulfilment requirements;
• customer behaviour.

Simulation and experimentation can therefore become tools for optimising the platform's competitive strategy.

The competition-law question is whether such optimisation represents legitimate competition or systematically excludes rivals and constrains sellers.

19. United States v RealPage

Subject: Algorithmic rental pricing

The U.S. authorities have addressed the use of algorithmic pricing systems in residential rental markets.

The FTC and DOJ have expressly argued that using an algorithm does not make price fixing lawful and that firms cannot evade antitrust rules by outsourcing pricing decisions to software.

Relevance to simulation

A pricing algorithm may effectively simulate:

• demand;
• occupancy;
• competitors' prices;
• expected responses;
• optimal rent levels.

If competing landlords independently use an algorithm merely as an analytical tool, the legal analysis differs from a situation in which competitors share competitively sensitive information through a common pricing mechanism.

Principle

Algorithmic optimisation remains subject to ordinary antitrust rules.

20. Intel v European Commission

Case: Intel Corp. v European Commission

The case concerned Intel's conduct involving rebates and alleged exclusionary effects.

The modern legal significance of Intel lies partly in the assessment of whether conduct by a dominant undertaking is capable of producing anticompetitive foreclosure and the role of economic analysis.

Simulation relevance

Simulation can be used to estimate:

• rival foreclosure;
• customer switching;
• effective prices;
• contestable demand;
• competitive constraints.

Therefore, sophisticated economic modelling can become important evidence both for the authority and the defendant.

21. Additional Case: T-Mobile Netherlands

Case: T-Mobile Netherlands BV v NMa

The CJEU addressed concerted practices and the exchange of strategically sensitive information.

Simulation relevance

Simulation systems become especially problematic where competitors feed them with:

• future pricing intentions;
• capacity;
• planned discounts;
• strategic forecasts.

The distinction between independent market intelligence and concerted information exchange is therefore fundamental.

22. Competition-Law Framework

Simulation-based competitive advantages can generally be analysed through five major legal categories.

23. Key Legal Test

A useful analytical framework is:

Step 1 — Identify the simulation

What exactly is being simulated?

• prices?
• demand?
• capacity?
• competitors?
• consumer behaviour?
• network effects?

Step 2 — Identify the inputs

What information does the model use?

• public data;
• proprietary data;
• customer data;
• competitor data;
• confidential seller information;
• shared industry information?

Step 3 — Identify the market position

Is the undertaking:

• a small competitor;
• an emerging platform;
• a vertically integrated firm;
• a dominant undertaking;
• an essential infrastructure provider?

Step 4 — Identify the output

What does the simulation produce?

• lower costs;
• better products;
• lower prices;
• exclusionary prices;
• discriminatory ranking;
• coordinated prices;
• restricted access?

Step 5 — Examine competitive effects

Ask whether the conduct:

• increases efficiency;
• improves innovation;
• lowers prices;
• increases output;

or instead:

• forecloses rivals;
• raises barriers to entry;
• facilitates coordination;
• reduces consumer choice;
• exploits dependent businesses.

24. Simulation as an Efficiency Defence

Simulation technology can produce substantial efficiencies.

For example:

Better forecasting → less waste → lower costs → lower prices

or:

Digital twin → fewer production failures → greater output → lower marginal costs

or:

Demand simulation → improved inventory → fewer shortages → better consumer welfare

These benefits can be important in competition analysis.

A competition authority should therefore distinguish between:

innovation that makes a firm a more effective competitor

and

technology deliberately designed to undermine the competitive process.

25. Evidence in Simulation-Based Competition Cases

Authorities may examine:

Technical evidence

• source code;
• model architecture;
• algorithmic rules;
• training data;
• system logs;
• API records;
• version histories.

Commercial evidence

• pricing decisions;
• internal strategy documents;
• communications;
• contracts;
• discount policies.

Economic evidence

• market shares;
• diversion ratios;
• elasticity;
• price-cost relationships;
• foreclosure rates;
• entry barriers;
• counterfactual simulations.

Behavioural evidence

• changes before and after algorithm deployment;
• competitor responses;
• customer switching;
• ranking changes;
• price movements.

The Amazon and Google matters illustrate the growing importance of analysing platform data, ranking systems and economic effects rather than looking only at conventional written agreements.

26. Important Distinction: Advantage vs Abuse

This distinction is crucial for examination purposes.

Simulation-based advantage is ordinarily legitimate when:

• it results from independent innovation;
• the inputs are lawfully obtained;
• the firm does not coordinate with competitors;
• consumers benefit from efficiencies;
• rivals remain able to compete;
• the firm does not misuse confidential competitor information.

It may create competition concerns when:

• competitors jointly use it to coordinate;
• a dominant platform uses competitor data obtained through its intermediary role;
• simulations support exclusionary pricing;
• algorithms systematically disadvantage rivals;
• access decisions are manipulated to foreclose competitors;
• the system facilitates discriminatory treatment;
• the technology is used to strengthen an existing monopoly through exclusionary conduct.

27. Emerging Issue: AI-Based Simulation

Generative AI and reinforcement learning create an additional problem.

An AI system may repeatedly simulate:

competitor action → firm response → competitor response → firm response

and optimise the firm's strategy.

The system could theoretically discover pricing patterns without an explicit human instruction to coordinate.

This creates an important competition-law question:

Can autonomous algorithmic learning create coordinated outcomes even without traditional human communications?

Current competition law generally continues to focus on established concepts such as agreement, concerted practice, dominance, exclusionary effects and consumer harm. The fact that an AI system produced the strategy does not automatically create a separate category of lawful conduct.

28. China-Specific Perspective

If the topic is applied to China, simulation-based competitive advantage can be examined principally through the Anti-Monopoly Law of the People's Republic of China, together with relevant rules concerning:

• monopoly agreements;
• abuse of dominant market position;
• economic concentration;
• platform-economy conduct;
• data and algorithms;
• discriminatory treatment;
• refusal to deal;
• tying;
• unreasonable trading conditions.

Particular attention would be appropriate for digital platforms because simulation capabilities may be combined with enormous datasets, algorithmic ranking and network effects.

For example:

Platform data → simulation → predicted seller behaviour → algorithmic ranking → market outcome

The legal question would be whether this represents legitimate technological competition or is being used to reinforce or exploit market power.

29. Practical Compliance Framework

Businesses using simulation systems should adopt:

1. Data controls

Record where every important dataset originates.

2. Competitor-information controls

Prevent employees from uploading confidential competitor information into shared systems.

3. Algorithm governance

Document important algorithmic objectives and constraints.

4. Competition-law testing

Review algorithms for potential:

• price coordination;
• discriminatory treatment;
• exclusionary ranking;
• predatory strategies.

5. Audit trails

Maintain records showing:

• model versions;
• changes;
• decision-makers;
• input data;
• outputs.

6. Human oversight

Important competition-sensitive decisions should be subject to compliance review.

7. Independent operation

Where competitors use similar analytical systems, ensure that each firm independently determines its competitive strategy rather than using a mechanism that coordinates market behaviour.

30. Conclusion

Simulation-based competitive advantage is not inherently anticompetitive. Competition law generally permits firms to gain advantages through better technology, forecasting, modelling and innovation.

The legal boundary is crossed when the simulation becomes an instrument for coordination, exclusion, exploitation, discriminatory treatment, misuse of competitor information or reinforcement of market power through anticompetitive conduct.

The central principle can therefore be stated as:

Competition law protects competition on the merits, but technological sophistication does not provide immunity from antitrust rules.

The most important authorities for studying this area include Eturas, G