EU antitrust remedies study finds less than half fully effective despite high implementation rates

The European Commission's antitrust enforcement toolkit has a effectiveness problem. A comprehensive study commissioned by DG COMP and published this week finds that while the overwhelming majority of remedies imposed under Regulation 1/2003 are implemented by the companies targeted, fewer than half actually achieve their intended competitive effect on the market.

The research, conducted by a consortium led by law firm Grimaldi Alliance and economic consultancy NERA, with monitoring trustee Thomas Hoehn, represents the first systematic retrospective evaluation of EU antitrust remedies since the modern enforcement regime began in 2004. Merger remedies have been studied repeatedly since a landmark 2005 Commission review; antitrust remedies, by contrast, have escaped comparable scrutiny until now.

A novel dataset spanning two decades

Drawing on the Commission's COMP Case Search database, the authors built a dataset covering every non-cartel antitrust decision adopted between the entry into force of Regulation 1/2003 on 24 January 2003 and 31 December 2022. The total: 108 decisions, comprising 57 prohibition decisions under Article 7 and 51 commitment decisions under Article 9. A handful pre-date the Regulation's formal application on 1 May 2004, including the seminal Microsoft I case.

Each decision was coded for the nature of the competition concern, horizontal agreements, vertical agreements, single-firm exclusionary conduct, single-firm exploitative conduct, or single market concerns, and for the remedy type: purely behavioural, structural, or behavioural with structural elements such as the transfer of airport slots.

The dominance of cease-and-desist orders

The dataset reveals a striking pattern. Approximately 80% of Article 7 decisions are simple cease-and-desist orders. Over time, basic prohibitions have given way to "like object or effect" formulations, which additionally require undertakings to refrain from any conduct having the same or equivalent object or effect, a formulation first seen in the Slovak Telekom decision. Only one Article 7 case in the entire period imposed a structural remedy: the ARA foreclosure decision, which proceeded under the cooperation procedure with national competition authorities.

Article 9 commitment decisions show more variety. Purely behavioural remedies remain the most common, but structural remedies were accepted in six cases, all in energy markets, and behavioural remedies with structural elements in seven more. The energy sector's prominence reflects the Commission's long-running focus on market integration and entry barriers in gas and electricity.

The monitoring trustee gap

Flanking measures such as reporting obligations and monitoring trustees are intended to secure compliance. Yet no monitoring trustee has been appointed in an Article 7 case since Microsoft I, where the General Court ruled that delegating investigative powers to a trustee and requiring the addressee to pay for its services was unlawful. That judgment has cast a long shadow: in Article 9 cases, by contrast, trustees have been appointed in roughly half of purely behavioural cases, in the vast majority of hybrid cases, and in every structural case.

The asymmetry is not accidental. Article 9 commitments are negotiated; the Commission can condition acceptance on the appointment of a trustee. Article 7 remedies are imposed unilaterally, and the Microsoft I precedent limits the Commission's ability to attach meaningful oversight powers without risking annulment.

Selecting the cases that matter

From the universe of 108 decisions, the researchers selected twelve for deep retrospective evaluation: five Article 7 and seven Article 9 cases. Simple cease-and-desist orders were excluded, as were decisions under judicial review at the time of selection, Google Search (Shopping), Google Android, International Skating Union, and Google Search (AdSense), and decisions fully or partly annulled, such as CISAC Agreement.

The five Article 7 cases were straightforward to identify: they were the only remaining ones with substantive remedies. For Article 9, the authors constructed a quantitative index combining decision length and download counts from the COMP Case Search website, then selected the highest-ranking cases while ensuring coverage across time, legal basis, competition concern, and remedy type. Broadcom, the sole Article 8 interim measures decision to date, was included but yielded no firm conclusion because its obligations are still ongoing.

Implementation versus effectiveness

The core finding is arresting. Three quarters of the twelve evaluated remedies were fully implemented by the addressees. Yet less than half were fully effective in achieving their stated objectives, whether stopping the infringement, preventing repetition, or removing anticompetitive effects. The gap points squarely to remedy design: measures that look adequate on paper fail to restore competition in practice.

Article 7 remedies exhibited more numerous and severe implementation problems than Article 9 commitments. Purely behavioural remedies underperformed relative to structural remedies and hybrid behavioural-structural packages. Encouragingly, the Commission's practice appears to have improved over time; the most troubled cases cluster in the earlier years of the regime.

Why Article 7 struggles

The study identifies several structural weaknesses in Article 7. The instrument is hybrid: it simultaneously establishes the infringement and imposes the remedy, leaving little room for a dedicated remedy design phase. There is no statutory requirement for market testing of proposed remedies, unlike Article 9, where Article 27(4) mandates it. The Regulation subordinates structural remedies to behavioural ones, stipulating that the Commission may impose structural remedies only where behavioural remedies are insufficient. And the Microsoft I judgment constrains the use of monitoring trustees, depriving the Commission of a tool that has proven valuable in commitment cases.

These are not new criticisms. The ECN+ Directive, adopted in 2019, already signals a shift: its Article 10 provides that remedy choice should be left to the application of effectiveness and proportionality in the individual case, without a priori preference for behavioural measures. The Draghi Report's proposal 8 on revamping competition enforcement likewise calls for systematic ex post evaluation, precisely what this study delivers.

Reform proposals: separation, testing, advisors

The study makes three principal recommendations. First, in particularly complex cases, the infringement decision and the remedy decision should be separated, allowing a dedicated phase for remedy design and greater transparency. Second, Article 7 should incorporate the market testing requirement that already applies to Article 9 commitments, giving market participants a structured opportunity to comment on proposed remedies before they are finalised. Third, the Commission should appoint independent advisors to assist in remedy design, particularly where structural measures are under consideration.

These proposals would align antitrust remedy practice more closely with the merger regime, where the 2005 remedies study led to detailed guidance and a notice that has shaped practice for nearly two decades. No equivalent guidance exists for antitrust remedies, a gap the authors argue is no longer defensible given the regime's maturity.