Competition Law and Governance of Scenario-Driven Markets

1. Introduction

Scenario-driven markets are markets in which competitive decisions are increasingly shaped by the creation, prediction, simulation, or evaluation of different future scenarios. Firms may use artificial intelligence, algorithms, predictive analytics, digital twins, automated pricing systems, market simulations, consumer forecasting, risk models, and real-time data to determine how they will compete.

Examples include:

• algorithmic pricing based on predicted competitor responses;
• AI systems forecasting consumer demand;
• digital platforms simulating alternative prices, rankings, or recommendations;
• financial and insurance markets using predictive models;
• energy markets using scenario modelling;
• logistics platforms predicting congestion and demand;
• autonomous systems deciding which suppliers or customers to prioritize;
• firms using common datasets or AI providers to generate competitive strategies.

Competition law becomes relevant where scenario-generation ceases to be an ordinary efficiency tool and becomes a mechanism for coordination, exclusion, discrimination, exploitation, information sharing, or strategic foreclosure.

The central legal question is therefore:

When does the use of predictive or scenario-based intelligence improve competition, and when does it reduce independent competitive decision-making?

2. Meaning of Scenario-Driven Markets

A scenario-driven market is one in which market participants make present competitive decisions by reference to predicted future market conditions.

A simplified model is:

Historical data → Predictive model → Alternative scenarios → Strategic decision → Market outcome

For example:

Competitor A's algorithm predicts that Competitor B will increase prices by 5%. A's system therefore increases its own price by 4%.

This may be legitimate independent competition.

However, the competition-law risk becomes greater if:

A and B use the same algorithm, receive commercially sensitive information from the same intermediary, or deliberately configure their systems to avoid aggressive competition.

Thus, scenario-driven markets create a distinction between:

Legitimate predictive competition

• forecasting demand;
• improving inventory;
• reducing waste;
• optimizing logistics;
• detecting fraud;
• improving service quality.

Potentially problematic predictive coordination

• coordinating prices;
• exchanging future pricing intentions;
• using common algorithms to implement parallel strategies;
• suppressing competitors through predictive exclusion;
• discriminating against particular users or suppliers;
• manipulating rankings based upon predicted responses.

3. Competition-Law Framework

Scenario-driven markets can engage several branches of competition law.

A. Anti-competitive agreements

The use of predictive technology does not eliminate the requirement for an agreement or concerted practice.

Competition authorities may examine whether competitors:

• exchanged future pricing information;
• agreed to use a common pricing system;
• relied on a common intermediary;
• coordinated through algorithms;
• shared strategic forecasts;
• deliberately aligned their scenario assumptions.

The technological mechanism is secondary to the underlying competitive relationship.

B. Abuse of Dominance

A dominant platform or infrastructure provider may use scenario-generation capabilities to:

• favour its own products;
• disadvantage rivals;
• deny access to predictive infrastructure;
• manipulate search or recommendation systems;
• impose discriminatory conditions;
• use data accumulated from downstream competitors to compete against them.

Potential theories include:

1. exclusionary abuse;
2. discriminatory access;
3. self-preferencing;
4. tying and bundling;
5. predatory or targeted pricing;
6. refusal to supply;
7. leveraging dominance from one market into another.

4. Algorithmic Coordination

One of the most important issues is algorithmic collusion.

Traditional cartel law generally looks for:

Human communication → agreement → coordinated conduct.

Scenario-driven markets can produce:

Data → algorithm → predicted competitor behaviour → automated response → parallel market outcome.

Parallel prices alone do not necessarily establish a cartel. Algorithms can independently produce similar prices because they respond to the same market conditions.

The stronger competition-law concern arises where firms intentionally use technology to reduce strategic uncertainty about competitors.

5. Common Predictive Infrastructure

Scenario-driven markets often depend upon common technological infrastructure.

For example, several competing firms might use the same:

• pricing algorithm;
• cloud-based decision system;
• market intelligence provider;
• data pool;
• demand forecast;
• AI foundation model;
• recommendation engine.

This creates a potential hub-and-spoke problem.

Structure

Competitor A
↓
Common algorithm/intermediary
↑
Competitor B

If the intermediary facilitates the exchange of competitively sensitive information or deliberately coordinates competitors' conduct, traditional competition-law principles may apply even though the communication is technologically mediated.

6. Data as a Competitive Input

Scenario-driven competition depends heavily on data.

Important categories include:

• historical prices;
• consumer behaviour;
• demand forecasts;
• inventory levels;
• capacity;
• future pricing intentions;
• supplier information;
• customer switching behaviour;
• competitor performance;
• location data;
• transaction histories.

The accumulation of large datasets can produce a competitive advantage because better data can produce better predictions.

Competition authorities may therefore investigate whether a dominant firm:

• unlawfully obtains competitor data;
• combines datasets in exclusionary ways;
• restricts interoperability;
• prevents rivals from accessing essential data;
• uses third-party data to disadvantage the businesses supplying it.

7. Scenario Manipulation

A distinctive problem is scenario manipulation.

A dominant firm may design its predictive system to generate assumptions that systematically favour its own business.

For example:

An online marketplace predicts that consumers are more likely to purchase the platform's private-label product and therefore gives that product greater visibility.

The competition-law issue is not simply that prediction is used. The issue is whether the prediction system is being used as a mechanism for exclusion or discriminatory treatment.

8. Six Major Case Laws

1. United States v. Apple Inc. — United States, 2024

The U.S. Department of Justice's antitrust action against Apple illustrates competition concerns involving digital ecosystems, control over platform architecture and restrictions affecting competitive access.

Relevance to scenario-driven markets

Large digital ecosystems possess enormous quantities of information concerning:

• consumer behaviour;
• application usage;
• payments;
• device usage;
• developer activity.

Such information can improve predictive capabilities.

The case demonstrates how competition law may examine whether control over a technological ecosystem allows a firm to shape the competitive environment in which other firms operate.

Principle

Digital infrastructure can itself become an important source of market power, particularly when control over the infrastructure affects competitors' ability to innovate or reach customers.

2. Eturas UAB v. Lietuvos Respublikos konkurencijos taryba — CJEU, 2016

This is particularly important for algorithmic and technology-mediated coordination.

An online travel-booking system transmitted a message to participating travel agencies concerning limitations on discounts.

The case concerned whether businesses participating in a common electronic system could be responsible for anti-competitive coordination.

Relevance

The significance lies in the fact that competition law does not require competitors to negotiate a traditional written cartel agreement.

Electronic communication and technological systems can facilitate a concerted practice.

Principle

Where competitors receive information through a common technological mechanism and knowingly participate in coordinated conduct, electronic infrastructure does not immunize the conduct from competition law.

3. AC-Treuhand AG v. European Commission — CJEU, 2015

AC-Treuhand concerned the liability of an intermediary that facilitated cartel arrangements.

The case is highly relevant to scenario-driven markets because modern competition can be coordinated through third-party infrastructure providers.

Relevance

An algorithm provider, data intermediary or market-intelligence provider may occupy a position analogous to a technological hub.

The important question becomes:

Did the intermediary merely provide a neutral service, or did it knowingly facilitate anti-competitive coordination?

Principle

Competition-law responsibility can extend beyond the firms fixing prices directly where an intermediary knowingly contributes to a cartel arrangement.

4. United States v. Topkins — U.S. Department of Justice, 2015

The Topkins matter involved online sellers using algorithms in connection with price coordination for posters and other products.

Relevance

It provides a direct illustration of the relationship between:

• online commerce;
• algorithmic pricing;
• communications between competitors;
• coordinated pricing.

The technological form of pricing does not alter the fundamental prohibition against price fixing.

Principle

A cartel does not become lawful merely because its implementation is automated.

5. United States v. Bazaarvoice, Inc. — U.S. District Court, Northern District of California, 2014

The Bazaarvoice litigation concerned the acquisition of PowerReviews by Bazaarvoice in the online ratings and reviews market.

Relevance

Online reputation systems are important inputs into scenario-driven markets because predictive systems frequently use:

• consumer reviews;
• ratings;
• seller reputation;
• product information;
• behavioural data.

The case demonstrates the importance of considering data-driven competitive dynamics when assessing digital-market transactions.

Principle

Market power analysis in digital markets may need to account for network effects, data advantages, switching costs and the competitive significance of information ecosystems.

6. FTC v. Amazon.com, Inc. — United States, 2023 onward

The U.S. Federal Trade Commission's action against Amazon raises issues concerning marketplace architecture, seller relationships, pricing practices and platform power.

Relevance

Large marketplaces can use sophisticated information systems to predict:

• consumer demand;
• seller behaviour;
• pricing;
• inventory;
• conversion rates;
• competitive responses.

The case illustrates how platform governance and algorithmic decision-making can become competition-law issues when platform rules potentially affect independent sellers and competing products.

Principle

The use of algorithms and predictive information by a dominant marketplace must be assessed alongside the platform's contractual and structural control over market participants.

7. Google Shopping — Google Search (Shopping), European Commission, 2017

The European Commission found Google had abused its dominant position by giving preferential treatment to its comparison-shopping service in search results.

Relevance

Search-ranking systems are essentially scenario-selection and prediction systems.

They determine which commercial options users are likely to see.

A platform controlling such an infrastructure can therefore influence competitive outcomes through:

• ranking;
• recommendation;
• visibility;
• traffic allocation;
• relevance scores.

Principle

A dominant platform's control over an algorithmic gateway can raise competition concerns where the system systematically disadvantages competing services.

8. Google Android — European Commission, 2018

The European Commission found several practices concerning Android devices to constitute abuses of dominance, including restrictions involving Google's search and browser services.

Relevance

The case demonstrates the importance of ecosystem governance.

An integrated technological environment can affect the competitive opportunities of complementary products and services.

In scenario-driven markets, control over an ecosystem can allow a dominant firm to influence:

Which services are installed → which services are visible → which services generate data → which services improve predictive models.

This can create a feedback loop.

Principle

Competition authorities may examine how contractual and technical restrictions reinforce dominance across interconnected digital markets.

9. Consolidated Case-Law Table

10. Scenario-Driven Markets and Merger Control

Merger analysis becomes more complicated where firms acquire predictive capabilities.

Suppose:

Company A owns a major consumer-data platform.

Company B owns an advanced predictive AI system.

A merger may combine:

Data + AI + distribution + customer access

This can create competitive advantages that are not captured adequately by traditional concentration measures.

Authorities may therefore examine:

• data accumulation;
• interoperability;
• network effects;
• economies of scope;
• vertical integration;
• access to AI infrastructure;
• foreclosure possibilities;
• innovation competition;
• future competitive constraints.

11. Scenario-Based Predatory Pricing

Predictive technology can make traditional predatory-pricing analysis more complicated.

A firm may use algorithms to identify:

• vulnerable competitors;
• customers likely to switch;
• geographic areas where rivals are weak;
• periods when competitors face financial pressure.

It could then selectively reduce prices.

The legal inquiry should distinguish:

Legitimate dynamic pricing

Prices change because demand, capacity or costs change.

Potential exclusionary pricing

Prices are strategically targeted to weaken or eliminate competitors.

Evidence may include:

• internal documents;
• algorithmic parameters;
• pricing histories;
• cost data;
• geographic targeting;
• competitor-specific targeting;
• subsequent changes after competitive exit.

12. Personalized Pricing

Scenario-driven markets facilitate increasingly sophisticated price discrimination.

Algorithms may predict a customer's:

• willingness to pay;
• likelihood of switching;
• urgency;
• purchasing probability.

This may produce individualized offers.

Competition law may become relevant where personalized pricing is combined with:

• dominance;
• exclusion;
• discriminatory access;
• tying;
• exploitation;
• exclusion of competing suppliers.

The mere existence of personalized pricing, however, does not automatically establish an antitrust violation.

13. Scenario-Driven Recommendation Systems

Recommendation engines may determine which products consumers see.

Examples include:

• e-commerce;
• app stores;
• search engines;
• streaming platforms;
• travel platforms;
• food-delivery platforms.

A dominant platform could potentially manipulate recommendations by:

1. favouring its own products;
2. disadvantaging competitors;
3. imposing discriminatory ranking criteria;
4. conditioning visibility on the purchase of another service;
5. exploiting seller data.

This connects scenario-driven competition with self-preferencing and platform neutrality.

14. AI and Predictive Competition

Generative AI and machine-learning systems introduce new competition concerns.

An AI system may:

• predict competitors' prices;
• forecast demand;
• generate strategic recommendations;
• automatically adjust prices;
• recommend suppliers;
• determine advertising allocation;
• predict customer churn.

Competition authorities therefore need to distinguish between:

Independent intelligence

Each firm independently uses its own data and makes its own decisions.

Coordinated intelligence

Several competitors use shared systems or information that materially reduces uncertainty about their competitive strategies.

The second situation creates greater competition-law risk.

15. Governance of Scenario-Driven Markets

An effective governance framework should contain several layers.

A. Data governance

Companies should establish controls over:

• data provenance;
• competitor information;
• commercially sensitive information;
• data-sharing arrangements;
• access rights;
• retention;
• automated data collection.

B. Algorithm governance

Algorithms should be reviewed for:

• discriminatory outcomes;
• exclusionary parameters;
• coordinated pricing;
• competitor-specific targeting;
• unexplained strategic constraints;
• automatic responses to competitor conduct.

C. Human oversight

Important competitive decisions should not always be completely automated.

Human review can be required for:

• major pricing changes;
• competitor-specific strategies;
• exclusionary restrictions;
• marketplace ranking changes;
• termination of suppliers;
• changes affecting access to infrastructure.

D. Auditability

Companies should maintain:

• algorithmic logs;
• version histories;
• decision records;
• training-data documentation;
• parameter changes;
• approval records;
• communications with third-party algorithm providers.

This is particularly important because algorithmic decisions can otherwise be difficult to reconstruct.

16. Competition Compliance Programme

A competition-compliance programme for scenario-driven markets should include:

1. Information classification

Identify:

• public information;
• commercially sensitive information;
• competitor information;
• aggregated information;
• confidential strategic forecasts.

2. Algorithm screening

Before deployment, ask:

Does the algorithm receive competitor-specific information?

Does it recommend coordinated conduct?

Does it automatically react to competitors?

Does it facilitate information exchange?

3. Third-party provider due diligence

Contracts with AI, data and pricing providers should be examined for:

• information-sharing provisions;
• common-model risks;
• access to competitor information;
• data pooling;
• algorithmic coordination.

4. Periodic competition audits

Algorithms should be periodically tested for:

• coordinated pricing;
• discriminatory outputs;
• exclusionary effects;
• self-preferencing;
• abnormal parallel conduct.

17. Regulatory Challenges

A. Attribution

Who is legally responsible for an anti-competitive algorithm?

Possibilities include:

• the company;
• its managers;
• the algorithm developer;
• the data provider;
• the intermediary.

Competition law must determine responsibility according to actual participation and control rather than simply attributing conduct to the technology.

B. Explainability

Competition authorities increasingly need to understand:

Why did the algorithm make this decision?

A black-box model can make enforcement difficult.

Consequently, explainability and audit trails may become increasingly important components of competition compliance.

C. Tacit coordination

The most difficult problem is automated parallel behaviour without explicit communication.

For example:

Algorithm A raises price when Algorithm B raises price.

Neither company may explicitly communicate with the other.

The legal question is whether this is simply rational independent adaptation or whether additional evidence demonstrates a coordinated practice or other prohibited conduct.

18. Economic Effects

Scenario-driven markets can produce both positive and negative competitive effects.

Potential pro-competitive effects

• lower costs;
• better demand forecasting;
• reduced waste;
• improved logistics;
• better consumer matching;
• increased innovation;
• faster market responses;
• improved resource allocation.

Potential anti-competitive effects

• automated collusion;
• increased entry barriers;
• data concentration;
• exclusionary personalization;
• discriminatory access;
• self-preferencing;
• strategic foreclosure;
• reduced transparency;
• reinforcement of dominant platforms.

Competition law must therefore avoid treating predictive technology itself as inherently anti-competitive.

19. Indian Competition-Law Perspective

In India, scenario-driven markets can be examined principally under the Competition Act, 2002, particularly:

Section 3

Prohibits agreements causing or likely to cause an appreciable adverse effect on competition.

Relevant issues include:

• algorithmic price coordination;
• information exchange;
• hub-and-spoke arrangements;
• technology-mediated cartel conduct.

Section 4

Addresses abuse of dominant position.

Potential concerns include:

• discriminatory conditions;
• denial of market access;
• leveraging;
• tying;
• exclusionary platform practices;
• unfair or discriminatory pricing.

Sections 5 and 6

Merger control becomes relevant when data, AI capabilities, platforms and predictive infrastructure are combined through combinations.

The Competition Commission of India may therefore need to consider not only current market shares but also:

• data advantages;
• network effects;
• ecosystem effects;
• access to AI infrastructure;
• innovation competition;
• switching costs;
• interoperability.

20. Essential-Facility Dimension

Some scenario-driven systems may become important competitive infrastructure.

Examples could include:

• dominant market-data systems;
• critical AI infrastructure;
• interoperability layers;
• major digital identity systems;
• indispensable transaction infrastructure.

If a dominant undertaking controls an infrastructure that competitors cannot reasonably replicate, refusal or discriminatory access may raise essential-facility/refusal-to-deal issues, subject to the applicable jurisdiction's legal test.

The important questions include:

1. Is the infrastructure genuinely indispensable?
2. Can competitors reasonably reproduce it?
3. Does denial eliminate effective competition?
4. Is there a legitimate business justification?
5. Can access be supplied on reasonable terms?

21. Future Competition

Scenario-driven markets require greater attention to future competition.

Traditional competition analysis often asks:

What is the competitive structure today?

Scenario-driven markets require an additional question:

How will control over predictive capabilities affect competition tomorrow?

For example, a firm with:

large dataset + superior AI + dominant distribution + customer feedback

may continuously improve its predictive system.

This produces a data-feedback loop:

More users → more data → better prediction → better service → more users → more data

Such feedback loops can strengthen network effects and increase barriers to entry.

22. Key Legal Principles

The principal competition-law principles emerging from scenario-driven markets are:

1. Technology does not alter the substance of cartel law.
2. Algorithmic coordination can create traditional competition concerns.
3. Independent parallel conduct is not automatically a cartel.
4. Common technological intermediaries can create hub-and-spoke risks.
5. Predictive data can constitute an important competitive asset.
6. Dominant platforms may face scrutiny over algorithmic ranking and recommendation.
7. Self-preferencing can become particularly significant where the platform controls the predictive gateway.
8. Merger control must consider data, AI and ecosystem effects.
9. Algorithmic transparency and auditability strengthen competition compliance.
10. AI should not be presumed anti-competitive merely because it is powerful or predictive.
11. Intent, structure, effects and market power remain central to legal analysis.
12. Competition law must distinguish efficiency-enhancing prediction from strategic foreclosure or coordination.

23. Conclusion

Scenario-driven markets represent a transition from conventional competition based on present information toward competition based increasingly on predictions of future behaviour.

Their central competition-law challenge is not the existence of prediction itself. Predictive technology can substantially improve efficiency, innovation and consumer welfare. The legal concern arises when predictive systems become instruments through which firms coordinate strategies, restrict access, exploit information asymmetries, discriminate against competitors, or reinforce existing market power.

The cases of Eturas, AC-Treuhand, Topkins, Google Shopping, Google Android, Bazaarvoice, Amazon and Apple demonstrate different aspects of the broader principle: digital technology and automated decision-making remain subject to ordinary competition-law principles, but their application requires attention to data, algorithms, intermediaries, network effects and technological ecosystems.

Accordingly, effective governance of scenario-driven markets should combine competition-law compliance, data governance, algorithmic auditing, human oversight, technological transparency and merger scrutiny. The objective is not to prevent firms from predicting market developments, but to ensure that predictive intelligence remains a tool of independent and innovation-oriented competition rather than a mechanism for coordination or exclusion.