Competition Law and Institutional Memory Concentration Risks

1. Introduction

Institutional memory refers to the knowledge, records, practices, historical information, technical experience and accumulated organisational understanding that enable an institution to make decisions over time.

In modern digital economies, institutional memory is increasingly stored and mediated through:

enterprise databases;

cloud systems;

knowledge-management platforms;

customer relationship systems;

proprietary datasets;

AI models;

decision-support systems;

document repositories;

industry databases;

workflow software;

transaction histories; and

organisational analytics.

This creates a new competition-law question:

What happens when the historical knowledge required to compete effectively becomes concentrated in the hands of one undertaking or a small group of undertakings?

Institutional memory can become a source of market power, entry barriers, informational advantages, switching costs and ecosystem dependence.

Competition law therefore has an interest in preventing the strategic use of concentrated institutional memory where it materially restricts competition.

2. Meaning of Institutional Memory

Institutional memory is broader than ordinary data.

It may include:

Historical information

previous transactions;

customer histories;

market conditions;

pricing patterns;

failures and successes.

Organisational knowledge

technical procedures;

engineering knowledge;

business processes;

operational experience.

Relationship knowledge

supplier histories;

customer preferences;

contractual relationships;

distribution networks.

Technological knowledge

software architecture;

system configurations;

technical specifications;

accumulated research.

Predictive knowledge

patterns derived from historical information;

demand forecasts;

customer behaviour;

risk assessments.

Thus:

Institutional memory is the accumulated informational and organisational capability that allows an undertaking to understand and operate within a market.

3. Institutional Memory as a Competitive Asset

A firm with twenty years of accumulated customer and market knowledge may have a substantial advantage over a new entrant.

Suppose:

Firm A: 20 years of historical transaction information.

Firm B: newly established competitor.

Firm A may know:

which customers buy particular products;

when demand changes;

which suppliers are reliable;

which product features succeed;

how customers respond to price changes;

how competitors typically react.

This accumulated knowledge can reduce uncertainty.

The advantage may be entirely legitimate.

The competition problem arises when the advantage is reinforced through exclusionary control over information infrastructure.

4. Institutional Memory and Data Concentration

Institutional memory increasingly exists in digital databases.

The competitive structure may therefore become:

Data accumulation → institutional memory → better decisions → better products → more customers → more data.

This creates a feedback loop.

Data feedback loop

More customers

↓

More data

↓

More institutional knowledge

↓

Better prediction

↓

Better service

↓

More customers

This can generate strong competitive advantages for established firms.

5. Competition Law Does Not Require Equal Knowledge

An important qualification is that competition law does not ordinarily require competitors to possess identical information.

Competition itself can reward firms that:

invest in research;

maintain better records;

develop superior databases;

employ skilled personnel;

develop better analytics.

Therefore:

A superior institutional-memory capability is not by itself an antitrust violation.

The competition concern arises where a dominant undertaking uses control over indispensable or strategically important information to foreclose rivals or prevent effective competition.

6. Institutional Memory and Entry Barriers

A new entrant may have sufficient capital and technology but still lack years of accumulated knowledge.

For example:

Incumbent: 15 years of customer behaviour.

Entrant: no historical customer dataset.

The entrant may therefore struggle with:

product design;

demand prediction;

credit assessment;

fraud detection;

customer segmentation.

If the information is replicable through ordinary competition, the disadvantage may simply be part of competitive rivalry.

If the information is uniquely controlled and rivals are deliberately prevented from accessing it, competition-law concerns become stronger.

7. Institutional Memory as a Switching Barrier

Institutional memory can also create lock-in.

A business may accumulate years of:

transaction records;

customer profiles;

workflows;

preferences;

software configurations;

historical analytics.

If these cannot easily be transferred to another provider, switching costs increase.

The relationship becomes:

Historical data → institutional memory → switching costs → customer lock-in.

This can make entry difficult even when alternative suppliers technically exist.

8. Data Portability

Data portability can therefore promote competition.

If customers can transfer their historical information from:

Provider A → Provider B

then Provider B may be able to compete more effectively.

Without portability:

Provider A → accumulated memory → lock-in

The competition-policy importance of portability is therefore particularly strong in:

banking;

cloud computing;

healthcare;

enterprise software;

digital platforms;

accounting;

e-commerce.

9. Institutional Memory and Cloud Computing

Cloud platforms can store enormous quantities of organisational knowledge.

A business may store:

customer information;

internal documents;

operational histories;

analytics;

software;

AI models.

If the business becomes dependent on one cloud ecosystem, switching to another provider may be costly.

Potential competition issues include:

data portability;

interoperability;

migration barriers;

proprietary formats;

contractual restrictions;

technical lock-in.

The competition question is whether these barriers arise naturally from technical complexity or are deliberately reinforced to exclude competitors.

10. Institutional Memory and AI

AI makes institutional memory particularly significant.

Modern AI systems can transform historical information into a continuously updated organisational knowledge base.

For example:

Historical documents + transactions + customer interactions + internal decisions → AI knowledge system.

The system can then answer:

What worked previously?

Which customers are likely to leave?

Which products succeeded?

Which technologies failed?

Which suppliers are reliable?

The AI system effectively becomes an institutional memory engine.

This can create substantial competitive advantages.

11. AI Memory and Competition

Suppose a dominant platform's AI system has access to ten years of:

consumer interactions;

product searches;

transaction histories;

customer complaints;

supplier relationships.

A new entrant cannot easily recreate this knowledge.

The resulting advantage can become a barrier to entry.

Competition authorities may therefore need to distinguish:

Legitimate learning advantage

The firm earned its data through competition.

Exclusionary advantage

The firm prevents competitors from accessing information or interoperability necessary to compete effectively.

12. Institutional Memory and Labour Mobility

Institutional memory does not exist only in databases.

Employees themselves carry organisational knowledge.

When employees move between firms, knowledge can spread.

Competition law must therefore consider:

non-compete clauses;

employee mobility;

no-poach agreements;

information restrictions.

Excessive restrictions on employee mobility may prevent knowledge diffusion and reduce innovation competition.

13. Institutional Memory and No-Poach Agreements

If competing employers agree:

“We will not hire each other's employees.”

they may preserve institutional knowledge within existing firms.

But they can also:

restrict employee mobility;

reduce wage competition;

prevent knowledge transfer;

strengthen incumbent advantages.

Competition authorities have therefore treated certain no-poach arrangements as potentially serious labour-market restraints.

This is especially important for technology markets where skilled workers carry substantial institutional and technical knowledge.

14. Institutional Memory and Merger Control

Mergers can combine two large institutional-memory systems.

For example:

Firm A database + Firm B database

may create a uniquely powerful informational asset.

The combined entity could obtain:

better customer intelligence;

better forecasting;

broader market knowledge;

more accurate risk models;

greater ability to identify emerging competitors.

Merger review can therefore consider data and knowledge concentration.

15. Institutional Memory and Killer Acquisitions

An incumbent may acquire a startup partly because of its:

technical team;

research database;

historical user data;

proprietary knowledge;

AI models.

Even if the startup's current revenues are small, its institutional knowledge may have significant future competitive value.

The acquisition could eliminate an independent source of innovation.

This connects institutional-memory concentration with potential competition.

16. Important Case Law

Case 1: United States v. Microsoft Corp.

The Microsoft litigation is a foundational technology competition case.

Microsoft's control over the Windows operating-system ecosystem provided substantial advantages in related software markets.

The case concerned exclusionary conduct designed to protect Microsoft's position against emerging competitive threats.

Relevance to institutional memory

A dominant technological ecosystem accumulates extensive knowledge about:

developers;

users;

software compatibility;

applications;

technological behaviour.

Control over such an ecosystem can reinforce incumbent power.

The case therefore illustrates how platform control and accumulated technological knowledge can support exclusionary strategies.

17. Case 2: Google Search (Shopping)

The European Commission's Google Shopping decision concerned Google's dominant search engine and preferential treatment of its own comparison-shopping service.

The case demonstrated the competitive significance of control over a major information infrastructure.

Relevance

A search engine accumulates enormous amounts of institutional knowledge about:

user behaviour;

search patterns;

commercial queries;

websites;

consumer preferences.

Control over that information environment can provide advantages in adjacent markets.

The case therefore helps explain why information infrastructure can become an important source of market power.

18. Case 3: Google Android

The European Commission's Android case involved Google's position in mobile operating systems and related restrictions affecting manufacturers and competing services.

Relevance

Operating systems create extensive institutional memory about:

users;

applications;

developers;

devices;

software ecosystems.

When the same undertaking controls multiple layers of the ecosystem, it can potentially leverage knowledge and infrastructure from one market into another.

The case illustrates the importance of examining ecosystem effects rather than treating every market as completely isolated.

19. Case 4: Microsoft Corp. v. Commission

General Court, Case T-201/04

The Microsoft case concerning interoperability and access to information required for work-group server interoperability is particularly relevant.

Microsoft controlled information necessary for competitors to achieve effective interoperability with Windows systems.

The Commission and General Court treated the refusal to provide interoperability information as an important component of the abuse analysis.

Relevance to institutional memory

This provides a direct conceptual connection:

Proprietary technical information → interoperability capability → rival access → competitive conditions.

Institutional memory can similarly become competitively important where accumulated technical information is necessary for rivals to participate effectively.

20. Case 5: IMS Health v NDC Health

CJEU, Case C-418/01

IMS Health concerned copyright-protected pharmaceutical sales data and the possibility of compelling access to a protected information structure.

The case established important principles concerning compulsory access to intellectual-property-protected information.

Relevance

The case is particularly useful for understanding the boundary between:

legitimate control over information; and

exceptional circumstances in which refusal of access can raise Article 102 concerns.

Institutional memory systems may similarly involve valuable proprietary databases.

The case shows that not every information asset must be shared with competitors.

21. Case 6: Bronner v Mediaprint

CJEU, Case C-7/97

The Bronner case concerned access to a newspaper home-delivery system.

The Court adopted a demanding standard for treating infrastructure as an essential facility.

Relevance

The principle is important when considering whether institutional-memory infrastructure must be made available to competitors.

A firm controlling an important database or knowledge platform does not automatically have a duty to provide access.

Questions such as:

indispensability;

elimination of competition;

lack of alternatives;

feasibility of duplication

are important.

22. Case 7: Commercial Solvents

CJEU, Joined Cases 6/73 and 7/73

Commercial Solvents involved a dominant undertaking's refusal to supply an essential input to a downstream competitor.

The Court recognised that a dominant undertaking can abuse its position where it cuts off supplies to a downstream market in circumstances capable of eliminating effective competition.

Relevance

The principle can apply conceptually to institutional memory where a dominant undertaking controls an essential informational input and uses that control to foreclose downstream competitors.

23. Case 8: Facebook/Meta Data Case

Bundeskartellamt — Facebook

The German competition authority examined Facebook's combination of user data from different sources.

The case connected data collection and combination with competition concerns involving a dominant digital platform.

Relevance

The case is particularly important for institutional-memory concentration because combining information from different services can produce a more comprehensive behavioural profile.

The broader competitive mechanism is:

Data combination → deeper knowledge → stronger service capability → stronger market position.

This illustrates why accumulation and combination of data can matter to competition analysis.

24. Institutional Memory and Refusal to Supply

Suppose a dominant undertaking controls a historical dataset that rivals cannot realistically reproduce.

A refusal to provide access might potentially raise abuse-of-dominance concerns.

But competition law does not automatically impose a general duty to share information.

Authorities must assess:

whether the undertaking is dominant;

whether the information is indispensable;

whether alternatives exist;

whether refusal eliminates effective competition;

whether access is technically feasible;

whether legitimate business justification exists.

25. Institutional Memory and Essential Facilities

The essential facilities doctrine provides a conceptual framework for examining access to infrastructure.

For institutional memory, the relevant question could be:

Is the accumulated knowledge or database so indispensable that competitors cannot effectively compete without access?

This threshold is generally demanding.

Ordinary competitive disadvantages are not sufficient.

26. Institutional Memory and Foreclosure

Foreclosure may occur when an incumbent uses its informational advantage to make rival entry or expansion substantially more difficult.

Examples include:

refusing interoperability;

preventing data portability;

restricting API access;

blocking integration;

withholding technical information;

imposing discriminatory access conditions.

The key issue is whether the conduct materially weakens competitive constraints.

27. Institutional Memory and Self-Preferencing

A platform with extensive historical information may use that information to favour its own products.

For example:

Marketplace data → identify successful product → launch competing private-label product → preferential placement.

This creates a potential conflict between:

the platform's role as infrastructure; and

its role as competitor.

Competition authorities may therefore examine whether the platform uses accumulated market knowledge to disadvantage businesses dependent upon it.

28. Institutional Memory and Platform Governance

Platform governance itself can become a competition issue.

A platform may possess years of knowledge regarding:

seller performance;

consumer behaviour;

algorithmic ranking;

transaction patterns;

demand.

If the platform changes its rules in a way that exploits dependent businesses' lack of equivalent information, competitive concerns may arise.

The relevant issue is not simply possession of knowledge but how market power is exercised through that knowledge.

29. Institutional Memory and Algorithmic Advantage

Historical data can improve algorithmic performance.

Suppose:

Algorithm A: trained on 10 years of market data.

Algorithm B: trained on one year.

Algorithm A may have a significant performance advantage.

If this advantage is earned through competition, it may benefit consumers.

But if the incumbent controls the only practical source of training data and deliberately denies rivals access, entry barriers can increase.

30. Institutional Memory and Data Portability

Data portability can reduce these effects.

Suppose customers can transfer:

transaction history;

preferences;

business records;

customer relationships;

historical analytics.

A new provider can begin with meaningful institutional knowledge.

This can reduce switching costs and improve contestability.

However, portability must also respect:

privacy;

cybersecurity;

confidentiality;

intellectual property;

third-party rights.

31. Institutional Memory and Interoperability

Interoperability can allow competing systems to preserve institutional knowledge.

For example:

Old CRM → new CRM

or:

Cloud A → Cloud B

or:

Banking platform A → Banking platform B.

Without interoperability, users may lose accumulated knowledge or face high migration costs.

This can create artificial lock-in.

32. Institutional Memory and Cloud Lock-In

Cloud providers increasingly store an organisation's entire operational history.

Migration may involve:

enormous datasets;

proprietary formats;

application dependencies;

security requirements;

technical integration.

The longer a firm remains with one cloud provider, the more institutional memory becomes embedded in that ecosystem.

Competition policy therefore has an interest in switching and portability conditions.

33. Institutional Memory and AI Foundation Models

Foundation models may become repositories of organisational knowledge.

Businesses may train models on:

internal documents;

customer interactions;

technical manuals;

legal records;

operational histories.

If the AI provider controls the model and its data environment, switching to another provider can become difficult.

Potential competition issues include:

model portability;

data portability;

interoperability;

ecosystem lock-in;

proprietary interfaces.

34. Institutional Memory and Labour Markets

Institutional knowledge can also create employer market power.

Suppose a small number of firms employ most specialised workers in an industry.

No-poach agreements may prevent workers from moving between those firms.

This can:

reduce wage competition;

restrict knowledge diffusion;

reduce innovation;

strengthen incumbent organisational memory.

Therefore, labour-market competition is relevant to institutional-memory concentration.

35. Institutional Memory and Innovation

Innovation frequently depends upon accumulated knowledge.

A firm with extensive historical knowledge can innovate more efficiently because it knows:

which experiments failed;

which technologies succeeded;

which customers need particular features;

which designs are inefficient.

Concentration of institutional memory can therefore affect dynamic competition.

If competitors cannot reproduce the knowledge base, innovation barriers may rise.

36. Institutional Memory and Merger Efficiencies

Mergers can also generate legitimate knowledge efficiencies.

For example:

Firm A's engineering knowledge + Firm B's manufacturing knowledge

may create a new technology faster than either could independently develop it.

Competition law must therefore distinguish:

Efficient knowledge integration

from

Anti-competitive knowledge concentration.

The merger may create legitimate innovation benefits where complementary knowledge is combined.

37. Institutional Memory and Mergers: Key Questions

Authorities may ask:

What databases are being combined?

Are they substitutable?

Do the parties compete using those datasets?

Does the merger eliminate an important independent source of knowledge?

Does the combination create a unique informational advantage?

Can rivals replicate the combined dataset?

Will users become more locked in?

Does the merged entity gain incentives to deny access?

Are there privacy or regulatory constraints?

Can behavioural or structural remedies preserve competition?

38. Institutional Memory and Digital Ecosystems

Institutional memory can become concentrated across an entire ecosystem.

Consider:

Operating system

↓

App store

↓

Search

↓

Payments

↓

Cloud

↓

AI

Each layer generates information.

If one company controls several layers, it may accumulate a uniquely comprehensive picture of users and businesses.

This can create cross-market informational leverage.

39. Competition Risks From Institutional Memory Concentration

The major risks can be summarised as follows:

40. Potential Remedies

Where competition concerns are established, possible remedies may include:

Data portability

Allowing customers to transfer their historical information.

Interoperability

Allowing competing systems to communicate.

API access

Providing technical access on non-discriminatory terms.

Non-discrimination

Preventing an infrastructure provider from favouring affiliated businesses.

Restrictions on data combination

Separating data collected in distinct markets where combination produces competitive harm.

Structural remedies

In exceptional cases, separation or divestiture may be considered.

Labour-market remedies

Enforcement against unlawful no-poach or wage-fixing arrangements.

41. Limits of Competition Law

Not every concentration of institutional memory requires intervention.

A company may legitimately possess superior knowledge because it:

invested in R&D;

served customers for many years;

developed better databases;

retained skilled employees;

created superior software.

Competition law should generally avoid converting every informational advantage into a mandatory sharing obligation.

The objective is to protect competitive opportunities, not to equalise every firm's knowledge.

42. Key Legal Principles

The case law supports several important principles.

1. Information can constitute a competitive asset.

2. Proprietary information is not automatically subject to compulsory sharing.

3. Dominant firms may face greater scrutiny when control over information is used to exclude competitors.

4. Interoperability can become important where technological information is necessary for effective competition.

5. Data combination can reinforce market power in digital ecosystems.

6. Network effects can turn accumulated information into a significant entry barrier.

7. Merger review may need to consider whether a transaction excessively concentrates strategic information.

8. Labour-market restraints can prevent the diffusion of valuable organisational knowledge.

43. Conclusion

Institutional memory is becoming an important form of competitive capital.

Historically, institutional memory resided primarily in:

employees;

paper records;

business relationships;

organisational culture.

Today, it increasingly exists in:

databases;

cloud infrastructure;

AI systems;

predictive models;

enterprise software;

digital platforms.

This transformation creates new competition-law questions.

The principal risks are:

informational entry barriers;

data concentration;

platform lock-in;

interoperability restrictions;

refusal to provide essential information;

self-preferencing;

cross-market leverage;

knowledge consolidation through mergers;

AI-driven informational advantages; and

restrictions on employee mobility.

The cases of Microsoft, Google Shopping, Google Android, Microsoft v Commission, IMS Health, Bronner, Commercial Solvents and Facebook demonstrate different legal principles relevant to these concerns.

The fundamental distinction is between knowledge acquired through legitimate competition and strategic control of institutional memory used to restrict competitive opportunities. Competition law is primarily concerned with the latter where dominance, exclusionary conduct, coordination or substantial foreclosure can be established.