Explainability Requirements For Competition Compliance

 

Explainability Requirements as a Procedural Safeguard in GWB Enforcement

1. Introduction

In German competition law, explainability is increasingly important where the Bundeskartellamt uses complex economic analysis, algorithms, data-driven screening, or technically sophisticated evidence to investigate undertakings under the Gesetz gegen Wettbewerbsbeschränkungen (GWB).

Explainability means that an authority should be able to communicate, to an affected undertaking and ultimately to a reviewing court:

  • what evidence it relied upon;
  • what economic or legal methodology it applied;
  • how it interpreted relevant data;
  • why particular assumptions were made;
  • how alternative explanations were treated;
  • how the evidence supports the alleged infringement; and
  • why the selected remedy or intervention is proportionate.

It is therefore best understood not as a freestanding statutory right to demand disclosure of every internal analytical process, but as a component of broader procedural guarantees: the right to be heard, adequate reasoning, access to the administrative file, effective judicial review, equality of arms, and proportionality.

This becomes particularly important under GWB §§ 1, 19, 19a, 20 and 21, especially in digital-market investigations involving large datasets, algorithmic pricing, self-preferencing, data access, interoperability, and ecosystem effects.

2. Legal Basis for Explainability in GWB Enforcement

A. Duty to give reasons

German administrative decision-making is generally subject to requirements of adequate reasoning. For competition decisions, the authority must make sufficiently clear:

  1. the relevant facts;
  2. the legal provisions applied;
  3. the evidentiary basis;
  4. the economic assessment; and
  5. the reasoning connecting those facts to the legal conclusion.

A decision cannot ordinarily be justified merely by saying that an economic model, algorithm, internal investigation or expert assessment produced a particular result.

B. Right to be heard

The right to be heard is a central procedural safeguard. An undertaking must have a meaningful opportunity to respond to the case against it.

Where the authority's case depends on complex economic reasoning, meaningful participation may require disclosure of sufficient information concerning:

  • the relevant evidence;
  • the theory of harm;
  • important assumptions;
  • methodological choices; and
  • material objections raised by the undertaking.

Thus, explainability has an important participatory function.

C. Access to the administrative file

Access to the authority's file enables the undertaking to understand the evidentiary basis of the proceedings and contest adverse material.

The principle is particularly significant in data-heavy competition proceedings. If the authority relies upon millions of observations, econometric analysis or algorithmic classifications, procedural fairness may require enough information to allow the undertaking to test the reliability and relevance of the evidence.

D. Effective judicial review

Explainability is also necessary for judicial review.

A court cannot meaningfully review a competition decision if the underlying reasoning is effectively a black box.

The authority therefore needs to expose the chain of reasoning:

Evidence → factual finding → economic inference → legal qualification → competitive harm → remedy.

3. Explainability Under GWB § 19

GWB § 19 prohibits the abusive exploitation of a dominant position.

Traditional abuse analysis already involves substantial economic reasoning concerning:

  • market definition;
  • dominance;
  • foreclosure;
  • discrimination;
  • exclusionary effects;
  • consumer harm;
  • efficiencies; and
  • causation.

Explainability becomes particularly important where these concepts are established through sophisticated quantitative evidence.

For example, if the Bundeskartellamt concludes that an undertaking is dominant because of a combination of:

  • market share;
  • network effects;
  • access to data;
  • financial strength;
  • vertical integration;
  • switching costs; and
  • ecosystem advantages,

the authority should make clear how those factors collectively produce the legal conclusion of dominance.

4. Explainability Under GWB § 19a

GWB § 19a is particularly significant.

The provision permits the Bundeskartellamt to determine that an undertaking is of paramount significance for competition across markets and subsequently address specified anti-competitive practices.

The provision was designed partly for major digital ecosystems whose competitive importance cannot easily be captured by conventional single-market analysis.

Explainability is consequently important at two levels.

First: designation

The authority must explain why the undertaking possesses characteristics demonstrating paramount significance.

Relevant considerations can include:

  • position across multiple markets;
  • financial resources;
  • vertical integration;
  • access to data;
  • relevance for competition;
  • ecosystem effects; and
  • ability to leverage market power.

Second: individual conduct

After designation, the authority must explain why a particular practice falls within the relevant statutory prohibition and why it threatens competition.

The procedural challenge is therefore:

Broad economic discretion must not become unexplained economic discretion.

5. Explainability and Algorithmic Evidence

Digital competition enforcement increasingly involves evidence generated by computational systems.

Examples include:

  • algorithmic price comparisons;
  • ranking systems;
  • recommender systems;
  • automated discrimination detection;
  • network analysis;
  • transaction-screening algorithms;
  • econometric models;
  • data-mining systems; and
  • AI-assisted investigative tools.

The existence of algorithmic evidence does not automatically make the evidence legally unacceptable.

However, its use creates additional procedural questions.

The authority may need to explain, to an appropriate degree:

Input

What data was used?

Method

What analytical technique was employed?

Assumptions

What assumptions were incorporated?

Validation

Was the methodology tested for reliability?

Error

What are the relevant limitations or possible false positives?

Causation

How does the output demonstrate the alleged competitive effect?

Legal relevance

Why does the resulting economic finding satisfy the statutory test?

This does not necessarily require disclosure of source code or every internal investigative technique. Confidentiality, trade secrets and investigative interests remain relevant.

The requirement is better expressed as:

Sufficient explainability for effective defence and judicial review, subject to legitimate confidentiality restrictions.

6. Six Major Case Laws

1. WuW/E DE-R 3465 – Facebook/Meta, Bundeskartellamt

The Facebook proceedings are among the most important German precedents concerning digital markets and GWB § 19.

The Bundeskartellamt examined the relationship between Facebook's market power and its collection and combination of user data across different services.

The case demonstrated that conventional market analysis may be insufficient when assessing a digital ecosystem.

Explainability significance

The authority's reasoning needed to connect:

data collection → ecosystem position → market power → exploitative conduct → competitive relevance.

This illustrates why explainability is especially important under § 19 and § 19a.

A complex theory of harm cannot simply rely on the proposition that "data creates market power." The authority must explain which data, what competitive advantage it produces, and through what mechanism.

Principle

Digital-market intervention requires an intelligible economic chain connecting data practices with competition-law harm.

2. BGH, KVR 69/19 – Facebook

The German Federal Court of Justice's decision concerning Facebook confirmed the importance of examining the interaction between market power and data-related conduct.

The case is particularly significant because it illustrates how German competition law can address conduct that sits at the intersection of:

  • competition;
  • privacy;
  • data collection; and
  • platform power.

Explainability significance

The legal reasoning must distinguish between:

  • a privacy violation as such;
  • exploitative conduct by a dominant undertaking; and
  • competitive harm resulting from the conduct.

That distinction is critical.

An authority cannot simply substitute a general concern about privacy for the statutory competition-law analysis.

Procedural lesson

Where multiple regulatory objectives intersect, the authority should explain which statutory competition-law element is established by which evidence.

3. BGH, KVR 69/19 – Facebook/Meta and Effective Judicial Review

The Facebook litigation also illustrates a broader procedural point: highly complex digital-market cases require courts to understand the authority's underlying economic reasoning.

The judicial assessment cannot be reduced to asking whether the authority reached a plausible result.

It must be possible to examine:

  • the factual findings;
  • market-power assessment;
  • relationship between data and competition;
  • legal characterization; and
  • proportionality of intervention.

Explainability principle

The more innovative or economically complex the theory of harm, the greater the practical importance of a transparent reasoning chain.

This does not mean that the authority must disclose every internal deliberation. Rather, the operative reasoning supporting the decision must be sufficiently comprehensible.

4. ECJ, C-413/14 P – Intel

Although Intel is an EU competition-law case rather than a GWB case, it is highly relevant to German enforcement because German competition authorities operate within the broader European competition-law environment.

The Court of Justice emphasized the importance of examining the actual capability of allegedly exclusionary conduct to foreclose competitors where the undertaking provides evidence capable of calling the authority's theory into question.

Explainability significance

Economic evidence cannot be treated mechanically.

If an undertaking presents evidence concerning:

  • market coverage;
  • duration;
  • conditionality;
  • rebates;
  • foreclosure capability; or
  • competitor viability,

the authority must explain how that evidence affects the legal assessment.

Procedural principle

Relevant economic evidence must be confronted rather than merely acknowledged.

This is particularly important when sophisticated quantitative models are used.

5. General Court, T-286/09 – Intel

The General Court's earlier Intel judgment, later reviewed by the Court of Justice, provides an important illustration of why economic reasoning must be sufficiently articulated.

The dispute involved the assessment of conditional rebates and their potential exclusionary effects.

The subsequent development of the case demonstrates that an authority's legal conclusion cannot be insulated from economically relevant evidence simply by classifying conduct under a formal category.

Explainability lesson for GWB enforcement

Where the authority relies on an economic theory of harm, it should explain:

  1. the relevant competitive mechanism;
  2. the evidence supporting it;
  3. the relevance of counter-evidence; and
  4. why the legal test is satisfied.

This reasoning model is particularly relevant to sophisticated German abuse proceedings.

6. ECJ, C-377/20 P – Servizio Elettrico Nazionale

The Court of Justice addressed exclusionary abuse and emphasized the importance of considering the circumstances relevant to assessing whether conduct is capable of restricting competition.

The judgment is significant for modern competition-law reasoning because it reinforces the importance of identifying the actual mechanism of exclusion.

Explainability significance

A competition authority should not merely state:

"The conduct is capable of restricting competition."

It should explain:

How exactly does the conduct restrict competition?

That may involve:

  • raising rivals' costs;
  • reducing access;
  • exploiting informational asymmetry;
  • increasing switching costs;
  • weakening competitors;
  • preventing entry; or
  • reinforcing an existing ecosystem advantage.

For algorithmic conduct, this causal explanation becomes particularly important.

7. Commission Decision and EU Judicial Review: Microsoft

The Microsoft litigation before the EU courts is another important reference point for German digital-market enforcement.

The case involved complex issues concerning:

  • interoperability;
  • technological design;
  • network effects;
  • innovation;
  • foreclosure; and
  • remedies.

Explainability lesson

When competition authorities intervene in technologically complex markets, they must articulate why a technical or contractual feature has competitive significance.

A technical explanation alone is insufficient.

The authority must translate:

technical functionality → market effect → competitive harm → legal infringement.

This is closely analogous to modern GWB proceedings involving interoperability, platform access and ecosystem leverage.

8. T-201/04 – Microsoft Corp. v Commission

The General Court's Microsoft judgment is particularly relevant to procedural reasoning because the case involved extensive economic and technical evidence.

The court had to assess whether the Commission's conclusions concerning interoperability and tying were supported by the evidentiary record.

Explainability significance

The case demonstrates that complex technology does not eliminate the requirement of intelligible legal reasoning.

The authority must explain:

  • what technical restriction exists;
  • how competitors are affected;
  • why the restriction is competitively significant;
  • what counterarguments exist; and
  • why the remedy addresses the identified harm.

The same methodology can inform interpretation of GWB § 19 and § 19a.

9. European Commission / EU Courts and Airtours

Airtours v Commission is important for merger-control reasoning, although it does not directly concern GWB abuse enforcement.

The case demonstrated the importance of a coherent evidentiary and economic explanation when an authority relies on complex theories concerning coordinated effects.

Explainability significance

A complex economic theory must be supported by:

  • identifiable evidence;
  • logically connected reasoning;
  • consideration of alternative explanations; and
  • a sufficiently convincing evidentiary basis.

This is relevant to German competition proceedings because similar economic methodologies are increasingly used in merger and digital-market enforcement.

10. Explainability and the Burden of Proof

Explainability should not be confused with shifting the burden of proof entirely onto the authority at every procedural stage.

German competition law contains different evidentiary and procedural structures depending on the statutory provision.

Nevertheless, explainability affects the practical allocation of argumentative burdens.

If the authority produces an unexplained algorithmic conclusion, the undertaking may be unable to contest it effectively.

For example:

Authority: "Our algorithm identifies systematic exclusion."

The undertaking's response cannot be meaningful unless it knows:

  • what "systematic" means;
  • what data was examined;
  • what threshold was applied;
  • what comparator was used;
  • what error rate exists; and
  • what alternative explanations were considered.

Thus:

Opacity can create an artificial evidentiary advantage for the authority.

Procedural fairness seeks to prevent this.

11. Explainability and Confidentiality

A major limitation is that explainability does not automatically equal full disclosure.

Competition proceedings frequently involve:

  • trade secrets;
  • confidential business information;
  • personal data;
  • investigative methods;
  • third-party information; and
  • legally protected communications.

Consequently, an appropriate model is layered explainability.

Layer 1 — Public reasoning

The final decision should explain the essential legal and economic reasoning.

Layer 2 — Party access

The investigated undertaking receives sufficiently detailed information to defend itself, subject to legitimate confidentiality protections.

Layer 3 — Confidential access

Sensitive technical or economic information may be made available through appropriate confidentiality arrangements.

Layer 4 — Judicial scrutiny

The reviewing court can examine material that cannot be fully disclosed publicly.

This creates a balance between:

transparency + defence rights + confidentiality + effective enforcement.

12. Explainability and GWB § 19a Digital Ecosystems

Section 19a creates especially strong arguments for explainability because the authority may assess an undertaking's significance across markets.

The authority may need to explain:

Market dimension

Which markets are relevant?

Ecosystem dimension

How are the markets connected?

Structural dimension

What features give the undertaking its cross-market significance?

Data dimension

How does data reinforce the position?

Network dimension

How do network effects operate?

Behavioral dimension

What conduct exploits or reinforces the position?

Competitive dimension

How does the conduct affect actual or potential competitors?

Without such reasoning, § 19a enforcement risks appearing to rely on an essentially discretionary designation.

Explainability therefore acts as a constitutional and administrative-law constraint on regulatory discretion.

13. Explainability as an Anti-Arbitrariness Safeguard

One of the most important functions of explainability is preventing arbitrary enforcement.

Suppose two platforms engage in comparable conduct but the authority intervenes against only one.

The authority should be capable of explaining:

  • the relevant differences;
  • differences in market power;
  • differences in competitive effects;
  • differences in evidence; and
  • differences in legal circumstances.

Therefore:

Consistency requires reasons, and reasons require explainability.

This is particularly important where enforcement priorities involve highly discretionary economic assessments.

14. Explainability and Proportionality

Explainability also supports proportionality.

The authority should establish a rational relationship between:

identified harm → intervention → remedy.

For example, if the authority identifies a data-access problem, it should explain why:

  • disclosure;
  • interoperability;
  • data portability;
  • behavioural restrictions;
  • monitoring; or
  • structural intervention

is necessary.

A remedy that is more intrusive than necessary requires particularly strong justification.

15. Explainability of AI-Based Enforcement Tools

Future GWB enforcement may involve AI systems assisting authorities with:

  • cartel screening;
  • anomaly detection;
  • merger screening;
  • pricing analysis;
  • document classification;
  • market definition;
  • network analysis; and
  • identification of potentially abusive conduct.

This creates an important distinction:

AI-assisted investigation

The AI identifies potentially relevant evidence.

AI-supported reasoning

The AI contributes to economic or factual assessment.

AI-generated decision

The AI effectively determines the outcome.

The third category raises the greatest procedural concern.

German competition enforcement should preserve human legal responsibility for the final decision.

An authority should not be able to defend a decision merely by saying:

"The system classified the conduct as anti-competitive."

The responsible decision-maker must be able to articulate the legal and evidentiary reasoning independently.

16. Explainability and the Right to Contest Automated Evidence

Where automated systems are used, procedural safeguards should ideally permit an undertaking to contest:

  1. the underlying data;
  2. the classification methodology;
  3. relevant assumptions;
  4. false positives;
  5. model limitations;
  6. comparator selection;
  7. statistical significance;
  8. causation; and
  9. the translation of model output into legal conclusions.

The principle can be expressed as:

No decisive competition-law consequence should rest exclusively on an incomprehensible computational output.

17. Procedural Model for Explainable GWB Enforcement

A robust procedure can be represented as follows:

Investigation

↓

Data collection

↓

Economic/algorithmic analysis

↓

Disclosure of material theory of harm

↓

Opportunity for undertaking to respond

↓

Consideration of counter-evidence

↓

Reasoned final decision

↓

Judicial review

↓

Proportionate remedy

Explainability should operate throughout this chain rather than only at the end.

18. Six Core Procedural Requirements

An explainable GWB enforcement system should ideally satisfy six requirements.

1. Traceability

The authority should be able to identify the evidence supporting each material finding.

2. Methodological transparency

The authority should explain material analytical methodologies.

3. Contestability

The undertaking must have a realistic opportunity to challenge the analysis.

4. Reasoned adjudication

The authority must explain why competing arguments were accepted or rejected.

5. Reviewability

The reasoning must be sufficiently intelligible for judicial scrutiny.

6. Proportionality

The remedy must logically correspond to the established competitive harm.

19. Key Case-Law Principles at a Glance

CaseMain relevance to explainability
Facebook/BundeskartellamtData, dominance, platform ecosystems and GWB § 19
BGH Facebook, KVR 69/19Relationship between market power, data and abusive conduct
Intel, C-413/14 PEconomic analysis and treatment of evidence concerning exclusion
Intel, T-286/09Need for coherent economic reasoning in abuse analysis
Servizio Elettrico Nazionale, C-377/20 PIdentification and explanation of the actual exclusionary mechanism
Microsoft, T-201/04Technical evidence, interoperability and complex theories of harm
Airtours, T-342/99Evidentiary coherence in complex economic assessment

20. Critical Issues

A. Explainability versus investigative secrecy

Too much disclosure may undermine investigations; too little may undermine defence rights.

B. Explainability versus trade secrets

A platform may legitimately resist disclosure of proprietary algorithms or source code. The authority must nevertheless provide sufficient substantive reasoning to permit challenge.

C. Explainability versus complexity

Economic complexity is not itself an excuse for opaque reasoning.

D. Explainability versus administrative efficiency

Automated enforcement can make investigations faster, but efficiency cannot eliminate procedural safeguards.

E. Explainability versus judicial deference

Courts may give competition authorities substantial expertise-related latitude, but such deference cannot transform unexplained analytical outputs into unreviewable conclusions.

21. Legal Significance

Explainability performs four simultaneous functions in GWB enforcement:

Procedural function
→ permits meaningful participation.

Evidentiary function
→ identifies the basis of factual and economic findings.

Constitutional function
→ limits arbitrary administrative power.

Judicial function
→ permits effective review.

This is particularly significant after the expansion of German competition law into digital ecosystems through GWB § 19a.

22. Conclusion

Explainability requirements are best understood as a procedural safeguard embedded within existing German and European principles of due process rather than as an entirely independent competition-law doctrine.

In conventional GWB cases, the requirement means that the Bundeskartellamt should provide an intelligible connection between evidence, economic analysis and the statutory elements of infringement. In digital-market cases, particularly under § 19a, the requirement becomes even more important because market power may arise from complex combinations of data, network effects, ecosystems, vertical integration and technological advantages.

The emergence of AI-assisted enforcement makes this issue more significant. An algorithm may assist investigation, but its output cannot substitute for legally accountable reasoning.

The central procedural principle is therefore:

The more complex, data-driven or algorithmic the enforcement methodology, the more important it becomes that the authority's decisive reasoning remains traceable, contestable and reviewable.

Accordingly, explainability serves as a bridge between economic expertise and procedural justice in modern GWB enforcement.

LEAVE A COMMENT