When Do Agreements Restrict Competition in EU Competition Law? A Four-Step Roadmap Under Article 101

Oct 1

10 min read

Updated: Oct 7

This post is based on the paper “When do agreements restrict competition in EU competition law?” by Associate Professor Christian Bergqvist, Ph.D., University of Copenhagen (SSRN abstract 3650755, updated until December 2021). All credit for the analysis, case discussion and framework summarised here goes to Professor Bergqvist. This post is a simplified summary for learners and is not a substitute for reading the paper or for legal advice.

Article 101 of the Treaty on the Functioning of the European Union bans anti-competitive agreements, but the Treaty gives very little guidance on how to tell whether a particular agreement crosses the line. Case law has produced patterns, with some practices condemned by their very nature and others assessed in detail, but no workable definition of the boundary. Professor Bergqvist’s paper offers a roadmap, and this post walks through it in plain language.

The Four-Step Roadmap

The paper explains that an enforcer should start by building a theory of harm, which explains how an agreement could damage competition, and then test it. In practice this means asking four questions:

Who are the parties? Are they actual or potential competitors, making the agreement horizontal, or do they sit at different levels of a supply chain, making it vertical?
Does the agreement restrict competition by object or by effect?
Is the damage appreciable? Or is it too small, ancillary or insignificant to matter?
Can the agreement be exempted under Article 101(3)? Either individually, or under a block exemption.

The paper stresses that the form of the agreement, its official purpose, the parties’ intent and even ignorance of Article 101 do not matter by themselves. What counts is the actual or plausible consequence for competition. The European Commission’s guidelines add that the protection is meant for the benefit of consumers, and that merely restricting the freedom of action of a party is not enough.

Step 1: Horizontal or Vertical?

Nothing in Article 101 itself separates horizontal from vertical agreements, but enforcement practice treats them very differently. Horizontal agreements between competitors form the core of enforcement. Vertical agreements, such as distribution arrangements, are assessed more leniently, with a presumption that they often produce efficiencies. Even so, in the Consten and Grundig case, the Court of Justice rejected the argument that Article 101 covers only horizontal agreements.

The labels can also blur. According to the paper, an agreement that looks vertical can acquire a horizontal element, for example where sensitive price information passes between a wholesaler and its retailer and on to others, as in the Hasselblad case, or where a distributor also competes with the supplier downstream.

Potential competitors. The parties do not have to be actual competitors. It is enough that they are potential competitors, and the paper explains the test: would entry be feasible without the agreement in response to a small but significant and lasting price increase, often called the SSNIP test. Entry within two to three years is a common benchmark, but the legal standard from the VISA case requires entry with sufficient speed to exert a competitive constraint. The cases show practical limits:

In Toshiba and the Franco-Japanese ball-bearings agreement, parties that agreed to stay out of each other’s home markets were treated as potential competitors, and the agreement itself was even used as evidence.
In E.ON Ruhrgas, the General Court overturned a decision because the German market was closed to entry by law, so there was no competition to restrict.
In pharmaceutical pay-for-delay cases such as Lundbeck and Generics, the existence of patents gives a rebuttable presumption that must be tested against real barriers to entry and evidence of preparatory steps.

Cartel facilitators. The paper also highlights that a business that is not a competitor can still be pulled into a cartel. In AC-Treuhand, an intermediary that coordinated a cartel was held liable, and in the Yen Interest Rate Derivatives case the concept was accepted, though liability failed on the facts for lack of knowledge of the whole infringement. The lesson is that knowing assistance that is instrumental to a cartel can be enough.

Inter-brand and intra-brand competition. Inter-brand competition is between suppliers of competing brands, and intra-brand competition is between distributors of the same brand. Article 101 covers both, and gains in one cannot simply offset restrictions in the other, as the Metropole case held. But restrictions limited to intra-brand competition are seen as much less harmful, which is why franchising, exclusive distribution and selective distribution are treated leniently.

Step 2: Restriction by Object or by Effect

The text of Article 101(1) speaks of agreements with the object or effect of restricting competition, and the Court of Justice made clear in Société Technique Minière that these are alternatives, not cumulative requirements. If an agreement restricts competition by object, there is no need to prove actual effects.

What counts as a restriction by object? Classic examples are price fixing, output limits and market or customer sharing between competitors, and for vertical arrangements, fixing minimum resale prices and some territorial restrictions. The justification offered is that such practices are by their nature harmful, and the paper compares this with risk offences in criminal law such as drink driving, where punishment does not wait for an accident. The paper identifies several features of the case law:

The assessment looks at the content of the agreement, its objectives and its legal and economic context, and the concept must be used restrictively.
Subjective intent, lack of a commercial interest in restricting competition, an acceptable purpose, and acting openly do not stop an agreement from being a restriction by object.
In the Budapest Bank and Generics rulings of 2020, the Court of Justice indicated that an agreement is restrictive by object when it has no plausible purpose other than restricting competition, and the enforcer must establish this.
Advocate General Bobek proposed reserving the category for obvious infringements, comparing it to recognising a fish at first sight, with a closer look needed only when something looks odd.
In Groupement des cartes bancaires, the Court of Justice reversed a finding of restriction by object on interbank fees and required more attention to context and market conditions.
In the Krka case, the General Court accepted that a patent settlement is not automatically hidden market sharing, since the documents and royalty level pointed to a commercial explanation.

The paper also notes two competing readings of the doctrine, one with a content-and-context analysis and another separating obvious from less obvious object restrictions, and says there is still no operative definition that can be applied directly. The only firm elements are the need for restrictive application and for the enforcer to explain why there is no alternative explanation.

What consequences follow? A restriction by object is caught by Article 101(1) without proof of effects. Any pro-competitive benefits are handled later under Article 101(3). It is also irrelevant whether the agreement was implemented or whether it is small in scale.

Restriction by effect. Where the object is not clearly anti-competitive, the effect must be shown to be real. The paper explains this as a counterfactual analysis, comparing competition with the agreement against competition without it. A tangible negative impact on parameters such as price, quantity and quality must be plausible, and a theory of harm must be linked to evidence. Examples include:

In European Night Services, the Commission’s decision was overturned because it did not substantiate the foreclosure claim and the parties’ market shares were small.
In Van den Bergh Foods, freezer exclusivity was unlawful because of the cumulative effect of many agreements tying up a large part of the market.
In UK Tractor, an information exchange system in a concentrated market was held anti-competitive because it created high market transparency, even without sharing sensitive internal data.
In Servier, pay-for-delay settlements were held restrictive by object and the Commission also examined market power and altered incentives of potential entrants.

The paper also notes a relevant US contrast. US antitrust law divides conduct between per se rules and a rule of reason that balances pro and anti-competitive effects. The European Union has clearly refused to import that balancing into Article 101(1), and its category of by object infringements is more dynamic and decided case by case, rather than a fixed catalogue.

Step 3: Is the Effect Appreciable?

Under case law, only an appreciable restriction is caught. The paper groups the situations where an agreement escapes in three ways:

By content: The restriction by its nature cannot restrict competition, for example certain quality-based requirements.
By limited impact (de minimis): Under the Commission’s 2014 Notice, an agreement is of minor importance if the parties’ combined market share is no more than 10 per cent where they are competitors, or no more than 15 per cent for each party where they are not. In mixed or unclear cases, 10 per cent applies. Where parallel networks of agreements cover 30 per cent of the market, the thresholds fall to 5 per cent. Exceeding the thresholds does not itself prove an appreciable effect, and the safe harbour never applies to restrictions by object.
By context (ancillary restraints): A restriction that is objectively necessary for, and proportionate to, a legitimate main operation is not assessed in isolation.

Commercial ancillary restraints. In Gøttrup-Klim and Metropole, purchase and non-compete clauses were accepted as necessary to make a partnership possible. The Court of Justice requires that the main operation would be impossible without the restriction, and being merely more difficult or less profitable is not enough. The paper lists further factors: the restriction should be directly linked to the main operation, adopted with it, confined to the parties, and judged objectively, and the main operation should be pro-competitive or neutral. The doctrine also covers vertical arrangements, such as selective distribution in Metro and franchising in Pronuptia, where restrictions protect know-how or the uniformity of the concept.

Regulatory ancillary restraints. The Wouters case, involving a Dutch bar rule banning partnerships between lawyers and accountants, and Meca-Medina, involving anti-doping rules, suggest that some rules adopted by self-regulatory bodies can fall outside Article 101(1) if they pursue legitimate objectives, are inherent in pursuing them and are proportionate. The paper warns that this doctrine is open-ended. In the International Skating Union decision, rules that served only the union’s financial interests and could shut out rival organisers did not qualify.

Step 4: Exemption Under Article 101(3)

Even an agreement that restricts competition by object or effect may be exempted if it meets four cumulative conditions:

It improves production or distribution of goods, or promotes technical or economic progress, through efficiencies that are plausible and attainable.
Consumers receive a fair share of the benefits.
The restrictions are reasonably necessary to achieve those efficiencies, with stricter scrutiny for more severe restraints.
The agreement does not allow the parties to eliminate competition for a substantial part of the market.

The paper notes that the parties bear the task of showing the nature of the efficiencies, the link to the agreement, their likelihood and magnitude, and how and when they will reach consumers. Notification to the Commission is no longer required, since Article 101(3) can be invoked directly before European and national enforcers and courts.

Block exemptions. The Commission has issued block exemptions covering areas such as vertical agreements, research and development and intellectual property licensing. These create safe harbours, usually with market share thresholds of 20 to 30 per cent and a list of hardcore restrictions that cannot be included. Exceeding the market share threshold does not create a presumption of infringement, but including a hardcore restriction is treated as a restriction by object, which could only be saved by an individual exemption.

Key Takeaways for Practitioners

Start with the relationship between the parties, because horizontal arrangements, including those between potential competitors, are treated far more strictly than vertical ones.
Check whether a clause could be a restriction by object, such as price fixing, market sharing or minimum resale prices, because no effect needs to be shown and the de minimis safe harbour does not apply.
Where object is not clear, expect the enforcer to build a theory of harm and use a counterfactual analysis based on market power and market structure.
Use the ancillary restraints doctrine and the de minimis notice as defences on context and size, and plan for exemption under Article 101(3) or a block exemption.
Document legitimate commercial reasons and efficiencies at the time of drafting, because the absence of a plausible alternative purpose is central to the by object analysis.

Build Practical Skills in Commercial Contract Drafting

Clauses on exclusivity, non-compete, territorial limits and resale prices are where competition law risk enters commercial contracts. To learn how to draft and review such agreements with practical templates, explore the Advanced Legal Drafting 6.0 course by Into Legal World.

For daily legal updates, join the Into Legal World WhatsApp Channel.

Frequently Asked Questions

1. What is the difference between a restriction by object and by effect under Article 101?

A restriction by object is one that by its nature is harmful to competition, such as price fixing or market sharing, so no proof of actual effects is needed. A restriction by effect requires proof that competition is in fact, or plausibly, harmed to an appreciable extent, usually through a counterfactual analysis.

2. Can a vertical agreement infringe EU competition law?

Yes. The Court of Justice made clear in Consten and Grundig that Article 101 covers vertical agreements too, although they are generally less likely to be harmful than horizontal ones and are assessed more leniently.

3. What is the de minimis rule in EU competition law?

Under the Commission’s 2014 Notice, an agreement of minor importance is generally outside Article 101(1) if the combined market share of competitors is no more than 10 per cent, or each party’s share is no more than 15 per cent for non-competitors. It never protects restrictions by object.

4. What is an ancillary restraint?

It is a restriction that is objectively necessary for, and proportionate to, a legitimate main operation, such as a joint purchasing arrangement or a franchise. Such a restriction is assessed together with the main operation instead of in isolation.

5. When can an anti-competitive agreement be exempted?

Under Article 101(3), if it improves production or distribution or promotes progress, passes a fair share of benefits to consumers, is indispensable to those benefits, and does not eliminate competition for a substantial part of the market. Block exemptions provide safe harbours for categories of agreements.

Keywords

When do agreements restrict competition in EU competition law, Article 101 TFEU, restriction by object, restriction by effect, anti-competitive agreements, de minimis notice, ancillary restraints, block exemptions, Article 101(3) exemption, horizontal and vertical agreements, potential competitors, cartel facilitation, Christian Bergqvist competition law

Credits and Source

This post is an original summary prepared for educational purposes. All credit for the ideas and analysis goes to:

Author: Associate Professor Christian Bergqvist, Ph.D., University of Copenhagen
Title: When do agreements restrict competition in EU competition law?
Publication: SSRN, abstract 3650755 (paper updated until December 2021)

Readers are encouraged to read the full paper for the complete case analysis and footnotes.