Algorithmic Holding Companies and Distributed Ownership Control

Introduction

Algorithmic holding companies and distributed ownership control describe corporate structures in which control over a group of companies is exercised through a combination of holding companies, subsidiaries, special-purpose vehicles, institutional investors, contractual rights, voting arrangements, and algorithmic systems rather than through straightforward majority ownership.

The central legal issue is that formal ownership and effective control may diverge. A parent company may own less than 50% of an undertaking yet exercise decisive influence through board appointments, veto rights, financing arrangements, shareholder agreements, common directors, technology infrastructure, data control, or centralized algorithms.

In competition law, this becomes particularly important where apparently independent firms are connected through a common ownership or control architecture. The question is not merely “Who owns the shares?”, but also:

Who can determine the undertaking's strategic commercial behaviour?

1. Meaning of Algorithmic Holding Companies

An algorithmic holding company can be understood as a holding-company structure in which important commercial decisions are coordinated or implemented through algorithmic systems.

These systems may determine:

• pricing;
• production;
• inventory;
• advertising;
• customer allocation;
• credit;
• investment;
• procurement;
• supply-chain decisions;
• platform access;
• risk management;
• resource allocation; and
• strategic coordination between subsidiaries.

A simplified structure may look like:

                    HOLDING COMPANY                           │             ┌─────────────┴─────────────┐             │                           │       Subsidiary A                Subsidiary B             │                           │       Algorithm A                  Algorithm B             │                           │       Customers                  Customers             │                           │             └──────────┬────────────────┘                        │                 Shared Data /              Common Infrastructure

The important point is that control can be distributed across several legal entities while technological infrastructure centralizes practical decision-making.

2. Distributed Ownership Does Not Necessarily Mean Distributed Control

A corporate group can have dispersed share ownership but concentrated control.

For example:

Thus, percentage ownership is only one indicator of control.

3. Economic Entity Doctrine

Competition law frequently distinguishes between separate legal personality and economic unity.

Two companies may be legally incorporated separately but constitute a single economic entity when one exercises decisive influence over the other's commercial policy.

This distinction is fundamental because competition rules concerning agreements between independent undertakings generally do not operate in the same manner where the entities form one economic unit.

Example

Suppose Holding Company H owns:

• 35% of Company A;
• 35% of Company B; and
• 30% of Company C.

If H also:

• appoints most directors;
• controls the common data infrastructure;
• determines algorithmic pricing parameters; and
• possesses veto rights over strategic decisions,

the legal analysis cannot stop at the share percentages.

The relevant question becomes whether H exercises decisive influence.

4. Decisive Influence

The concept of decisive influence is central to merger control and abuse-of-dominance analysis.

It can arise from:

A. Voting rights

A shareholder may possess enough voting power to determine corporate decisions.

B. Board appointment rights

A minority shareholder may control the appointment of directors.

C. Veto rights

Special rights concerning:

• budgets;
• business plans;
• senior management;
• major investments;
• acquisitions;
• technology;
• intellectual property

may create control.

D. Contractual control

Control can sometimes arise through long-term agreements or other contractual mechanisms.

E. Technological control

In an algorithmically integrated group, the entity controlling the core algorithm may exercise practical influence even without majority ownership.

5. Algorithmic Control as a New Form of Corporate Control

Traditional corporate law assumes that control is exercised through:

• shareholders;
• boards;
• officers;
• resolutions; and
• contracts.

Algorithmic organizations introduce another layer:

Share ownership       ↓ Corporate governance       ↓ Technology governance       ↓ Algorithmic decision-making       ↓ Commercial conduct       ↓ Market outcomes

The algorithm may effectively translate ownership power into market behaviour.

For example, a holding company may establish a central pricing engine used by several subsidiaries.

The subsidiaries technically determine their own prices, but the algorithm may:

• collect their pricing data;
• recommend prices;
• impose pricing boundaries;
• optimize prices collectively; and
• respond simultaneously to competitor movements.

The resulting commercial independence may therefore be considerably weaker than the corporate chart suggests.

6. Distributed Ownership and Common Ownership

A particularly important competition-law problem arises when an investment entity holds minority interests in several competing undertakings.

For example:

             Investment Holding Entity                 /       |       \                /        |        \              20%       18%       15%             Firm A    Firm B    Firm C

Each investment may individually appear harmless.

However, cumulative ownership can raise questions concerning:

• incentives to soften competition;
• information exchange;
• governance influence;
• common directors;
• strategic voting;
• reduced incentives to compete;
• portfolio-wide optimization; and
• algorithmic coordination.

This is sometimes described as the common-ownership problem.

7. Algorithmic Common Ownership

Algorithms can intensify common-ownership concerns.

An investment group could potentially operate a centralized analytical system that evaluates:

• the prices of portfolio companies;
• market shares;
• customer switching;
• production;
• margins;
• advertising expenditure;
• investment decisions; and
• competitive responses.

The resulting system might produce recommendations across the portfolio.

The competition-law question is whether this merely constitutes legitimate portfolio management or whether the arrangement facilitates coordination or exclusionary conduct.

8. Parent–Subsidiary Liability

A major issue is whether conduct implemented by a subsidiary can be attributed to its parent.

Where the parent exercises decisive influence, competition authorities and courts may treat the parent and subsidiary as part of the same economic entity.

This can have major consequences for:

• fines;
• merger notification;
• dominance;
• liability;
• compliance;
• information exchange; and
• calculation of corporate penalties.

9. Case Law

1. Europemballage Corporation and Continental Can Company Inc. v Commission — Continental Can

Court: Court of Justice of the European Union
Principle: Corporate structures cannot be analysed solely through formal legal separation.

The case concerned Continental Can's acquisition activity and the development of EU competition-law control over concentrations.

The decision is important because it illustrates the broader principle that competition law examines economic power and market structure, rather than relying exclusively on formal corporate classifications.

Relevance

For algorithmic holding companies, this supports examining:

• economic power;
• corporate relationships;
• control mechanisms;
• market structure; and
• the practical effects of corporate consolidation.

10. Viho Europe BV v Commission

Court: CJEU

This is one of the most important cases concerning the single economic entity doctrine.

The Court considered the relationship between a parent company and subsidiaries and recognized that entities belonging to the same corporate group may constitute a single economic unit where the subsidiary does not enjoy genuine autonomy in determining its market conduct.

Significance

The case demonstrates that:

Separate incorporation does not automatically establish independent economic actors.

Application to algorithmic holding companies

Where a holding company:

• owns multiple subsidiaries;
• controls their strategic policies;
• provides common technology;
• controls pricing algorithms; and
• determines commercial parameters,

the group may need to be analysed as an integrated economic structure.

11. Akzo Nobel NV v Commission

Case: Akzo Nobel NV and Others v Commission

Court: CJEU

This is a leading authority concerning parent-company liability.

The Court established the importance of the relationship between a parent company and its subsidiary when determining whether the subsidiary's conduct can be attributed to the parent.

The famous principle concerns the presumption associated with 100% ownership, subject to the possibility of rebuttal.

Importance

The case demonstrates that corporate liability may extend beyond the entity that physically committed the competition infringement.

Algorithmic relevance

Suppose a subsidiary operates an algorithm that produces exclusionary pricing.

The investigation would not necessarily end with the subsidiary.

Authorities may examine:

• who designed the algorithm;
• who approved it;
• who controlled its parameters;
• who monitored its operation; and
• whether the parent exercised decisive influence.

12. General Química v Commission

Case: General Química SA and Others v Commission

Court: CJEU

This case further developed the principles concerning attribution of liability within corporate groups.

The Court examined whether conduct of a subsidiary could be attributed to a parent based upon the relationship of control between the entities.

Significance

The case reinforces the importance of actual corporate influence, rather than simply identifying the entity whose employees directly participated in the conduct.

Algorithmic application

If a parent company establishes a centralized algorithmic compliance or pricing architecture used by subsidiaries, investigators may need to determine:

1. who designed the system;
2. who authorized deployment;
3. who controlled the parameters;
4. who received the resulting information; and
5. whether subsidiaries retained genuine commercial autonomy.

13. Dow Chemical v Commission

Case: Dow Chemical Ibérica and Others v Commission

Court: CJEU

The case is important for the broader concept of the undertaking as an economic unit and attribution of conduct within corporate structures.

Relevance

A competition investigation may therefore look beyond the immediate legal entity and examine the wider corporate organization.

For algorithmic holding companies, this is particularly important where several subsidiaries share:

• management;
• data;
• software;
• strategic planning;
• compliance systems; and
• pricing infrastructure.

14. Stora Kopparbergs Bergslags AB v Commission

Case: Stora Kopparbergs Bergslags AB v Commission

Court: CJEU

The Court examined the circumstances in which conduct of a subsidiary can be attributed to a parent company.

The case contributed to the jurisprudence concerning parental responsibility and control.

Algorithmic relevance

The case is useful for understanding why corporate groups cannot necessarily structure themselves to avoid competition-law responsibility merely by distributing operations among several legal entities.

An algorithmically coordinated corporate group could therefore face scrutiny concerning whether the apparent decentralization is genuine or merely organizational.

15. Sumal SL v Mercedes Benz Trucks España SL

Case: Sumal SL v Mercedes Benz Trucks España SL

Court: CJEU

This case significantly developed the economic-unit doctrine by addressing the circumstances in which liability connected with competition-law infringements can extend within a corporate group.

The Court emphasized the relationship between the parent and subsidiary as components of a single economic unit.

Importance for algorithmic holding structures

The decision is especially relevant to modern corporate architectures because legal responsibility may need to be examined at the level of the economic unit, rather than simply at the level of the individual corporation.

Where technology, governance and commercial strategy are centralized, the distinction between entities can become legally significant.

16. Mergers and Change of Control

Algorithmic holding companies are also important under merger-control law.

A transaction may involve:

Acquisition     ↓ Minority shareholding     ↓ Board rights     ↓ Algorithm/data access     ↓ Strategic influence     ↓ Possible change of control

A transaction does not necessarily need to involve a 51% acquisition to raise merger-control questions.

The relevant issue may be whether the transaction produces a lasting change in the quality of control.

17. Minority Shareholdings

Minority investments can be divided broadly into:

Passive minority investment

The investor receives:

• dividends;
• capital appreciation; and
• limited shareholder rights.

Strategic minority investment

The investor additionally obtains:

• board representation;
• veto rights;
• commercial agreements;
• access to sensitive information;
• technology rights; or
• strategic influence.

The second category creates substantially greater competition-law questions.

18. Algorithms and Interlocking Control

A particularly novel structure is:

                   Central Algorithm                         │        ┌────────────────┼────────────────┐        ↓                ↓                ↓     Company A        Company B        Company C        ↑                ↑                ↑        └────── Common Holding Entity ────┘

The holding entity may not issue direct pricing instructions.

Instead, the algorithm may:

• predict demand;
• optimize margins;
• recommend output;
• identify competitors;
• determine promotional strategy; and
• allocate capital.

The legal question then becomes whether algorithmic architecture itself constitutes a mechanism of control.

19. Information Control

Information may be more important than ownership.

A holding company with access to the following information across competitors could acquire substantial strategic influence:

• future prices;
• production forecasts;
• customer data;
• cost structures;
• tender strategies;
• capacity plans;
• investment plans.

Consequently, competition authorities may examine information flows, not merely shareholding percentages.

20. Algorithmic Firewalls

Corporate groups may attempt to preserve competitive independence through:

• data segregation;
• access controls;
• separate algorithms;
• independent boards;
• information barriers;
• separate management teams;
• compliance protocols.

These mechanisms can be legally significant because they may demonstrate genuine operational independence.

However, a firewall is meaningful only if it functions in practice.

A nominal separation that permits centralized access to commercially sensitive data may provide limited protection.

21. Risks of Algorithmic Holding Structures

A. Coordinated pricing

Common algorithms may produce parallel pricing.

B. Information exchange

A centralized system may expose one subsidiary's competitively sensitive information to another.

C. Market foreclosure

A holding company may prioritize its own portfolio companies over independent competitors.

D. Cross-subsidization

Profits from one market may finance exclusionary strategies in another.

E. Discriminatory access

Portfolio companies may receive preferential access to:

• data;
• infrastructure;
• APIs;
• customers;
• logistics;
• financing.

F. Strategic minority investments

Multiple minority stakes may create incentives to reduce competitive rivalry.

22. Algorithmic Holding Companies and Article 101-Type Concerns

Where entities are genuinely independent, arrangements involving them may potentially raise issues concerning:

• information exchange;
• coordination;
• market allocation;
• price coordination;
• output restrictions.

But where entities form a single economic unit, the legal analysis is different.

Therefore, the first question is often whether the firms are genuinely independent undertakings.

23. Article 102-Type Concerns

A holding company or technology-controlled corporate group can also face abuse-of-dominance issues.

Potential theories include:

• refusal of access;
• discriminatory access;
• tying;
• self-preferencing;
• exclusionary pricing;
• leveraging;
• margin squeeze;
• interoperability restrictions.

Algorithmic control can amplify these concerns because automated systems can apply exclusionary conditions at enormous scale.

24. Corporate Veil and Competition Law

Corporate law generally recognizes separate legal personality.

But competition law focuses on the economic reality of market conduct.

Consequently:

Separate incorporation        ≠ Automatically separate undertaking

Likewise:

Minority ownership        ≠ Automatically absence of control

And:

Algorithmic decentralization        ≠ Automatically independent decision-making

The actual legal analysis depends on the facts establishing control, influence and economic unity.

25. Regulatory Questions for Algorithmic Holding Companies

A competition authority may ask:

1. Who owns the undertaking?
2. Who controls voting rights?
3. Who appoints directors?
4. Who possesses veto rights?
5. Who controls the algorithm?
6. Who owns the relevant data?
7. Who determines algorithmic parameters?
8. Who can modify the model?
9. Who receives commercially sensitive information?
10. Are subsidiaries genuinely autonomous?
11. Does the holding company influence commercial policy?
12. Do portfolio companies compete with one another?
13. Does the structure create incentives for reduced rivalry?
14. Has a change of control occurred?
15. Can liability be attributed across the group?

26. Corporate Governance Implications

Boards increasingly need to understand that algorithmic governance is itself a form of corporate governance.

Traditional governance asks:

Who has voting power?

Algorithmic governance additionally asks:

Who controls the decision architecture through which corporate decisions are generated?

This creates new governance responsibilities involving:

• model approval;
• data governance;
• algorithmic audit;
• cybersecurity;
• access rights;
• model modification;
• human oversight;
• competition compliance.

27. Evidentiary Problems

Algorithmic holding structures create difficult evidence questions.

Important evidence may include:

• source code;
• model documentation;
• system logs;
• API records;
• board minutes;
• shareholder agreements;
• data-access permissions;
• model-training records;
• internal communications;
• pricing outputs;
• audit trails.

The absence of an express instruction does not necessarily resolve the issue.

Investigators may examine whether an algorithm's architecture itself facilitated coordinated conduct.

28. Compliance Framework

A multinational corporate group using centralized algorithms should consider:

Ownership mapping

Maintain an updated map of:

• shareholders;
• voting rights;
• beneficial interests;
• board rights;
• veto rights.

Algorithm mapping

Identify:

• centrally controlled algorithms;
• independently controlled algorithms;
• shared systems;
• common databases.

Information mapping

Determine who can access:

• competitor data;
• pricing information;
• customer information;
• strategic forecasts.

Governance separation

Where companies compete, establish appropriate:

• data barriers;
• governance separation;
• independent decision-making;
• access restrictions.

Auditability

Maintain records demonstrating:

• who approved algorithms;
• why parameters were selected;
• who modified them;
• how competition risks were assessed.

29. Six Core Legal Principles Emerging from the Case Law

30. Relationship Between Distributed Ownership and Control

The central conceptual distinction can be expressed as:

Ownership ≠ Control ≠ Management ≠ Algorithmic Decision-Making

They may overlap, but they do not necessarily do so.

For example:

Investor   │   │ owns 25%   ↓ Holding Company   │   │ appoints directors   ↓ Subsidiary   │   │ controls algorithm   ↓ Automated pricing   │   ↓ Market behaviour

The legal inquiry must follow the chain rather than stopping at the first ownership percentage.

Conclusion

Algorithmic holding companies and distributed ownership control represent a shift from traditional shareholder-centric control toward hybrid forms of corporate, contractual, informational and technological control.

The leading economic-unit and parent-liability jurisprudence—including Viho, Akzo Nobel, General Química, Dow Chemical, Stora Kopparbergs and Sumal—shows why competition law can examine relationships beyond formal corporate boundaries.

The emerging legal challenge is therefore not simply determining who owns a company, but determining who possesses decisive influence over its commercial decision-making architecture.

Where algorithms, shared data, common infrastructure and minority ownership operate together, regulators may need to examine the entire control architecture to determine whether apparently distributed enterprises are genuinely independent or function as components of a larger economic unit.

Key Case Laws

1. Continental Can v Commission
2. Viho Europe BV v Commission
3. Akzo Nobel NV v Commission
4. General Química v Commission
5. Dow Chemical Ibérica v Commission
6. Stora Kopparbergs Bergslags AB v Commission
7. Sumal SL v Mercedes Benz Trucks España SL

These cases provide the principal doctrinal foundation for analysing economic unity, decisive influence, parent-subsidiary relationships, attribution of liability and distributed corporate control in an increasingly algorithmic corporate environment.