Lowering Evidentiary Requirements for the Commission’s Decision Making: The Case for Purely Administrative Procedures

Leading up to Christmas 2024, the German Federal Government published its competition policy wishlist for the 2024–2029 EU Commission term. Anticipating the Commission’s revision of the regulation governing the enforcement procedure of Articles 101 and 102 since 2001, the wishlist includes a call ‘to simplify and speed up procedures […], such as allowing to enforce Article 102 TFEU in purely administrative procedures, without the threat of sanctions and fines’. What might at first seem like a major blow to enforcement, is actually meant to achieve the polar opposite: Lowering the standard of proof to make public enforcement more effective.

In the following, we want to show why the standard of proof in competition proceedings can be an obstacle to effective enforcement, and how the legislator could go on about lowering the standard of proof.

The ‘beyond any reasonable doubt’ standard of proof as an obstacle to effective enforcement

The problem stems from the strict ‘beyond any reasonable doubt’ standard of proof originating in criminal law. This standard governs the Commission’s proceedings in competition law matters. Such a high evidentiary threshold does not necessarily conflict with effective enforcement. As already showcased by the classic ‘Futurama’ episode ‘Fear of a Bot Planet’ - in which the robot attorney general is successful (although partially erroneously) before court in “demonstrat[ing] beyond 0.5 percent of a doubt that these humans are guilty of the crime of being humans” -, where the legal standard provides a clear set of facts that are relatively easy to prove, a strict standard of proof is generally compatible with effective enforcement.

Working with such an easy-to-prove legal test, however, is seldom the situation the Commission finds itself in enforcing modern day competition law. Regularly, the finding of an infringement hinges on a complex assessment of the relevant conduct’s effects on competition in the individual case or on a complicated assessment of an undertaking’s dominant position on a market yet to be defined. In such cases, the legal standard alone is challenging for the Commission, let alone proving the relevant facts beyond any reasonable doubt. The case law on abuse of dominance cases as well as cases of collusion moved away from relatively simple to prove conduct-based tests and rather emphasizes the importance of the analysis of the conduct’s effects ‘in the light of all the relevant factual circumstances’ and ‘on the basis of specific, tangible points of analysis and evidence’. Possibly most obvious, the Intel saga’s redefinition of the legal test for exclusive dealing serves as a textbook example: Under traditional EU competition law, each and every agreement including an exclusivity clause was abusive under Article 102 TFEU, irrespective of the peculiarities of the individual case. The Court’s judgment in Intel I modified this case law. Today, there is the possibility for the dominant undertaking to submit that, in the individual case, its behaviour was unlikely to produce exclusionary effects, adding considerable complexity to the legal test for abuse. This phenomenon is not limited at all to abuse of dominance, as the Court’s newer case law on resale price maintenance under Article 101 TFEU shows. 

Paired with the ‘beyond any reasonable doubt’ standard of proof, these challenges may evolve into actual obstacles to effective enforcement, as the Commission decisions in Intel and Qualcomm (Exclusivity payments) illustrate. Both decisions were annulled by the General Court largely on the grounds of an erroneous analysis of effects (see the judgments in Intel RENV and Qualcomm). In both judgments, the General Court referred to the presumption of innocence to justify its harsh judicial review.

Why ‘beyond any reasonable doubt’?

This leads us to the reason why the strict criminal standard of proof applies in competition proceedings. The courts do not expressly accept any single standard of proof in competition law. Still, the ‘beyond any reasonable doubt’ threshold must be the yardstick against which the case law on the standard of proof must be measured wherever the presumption of innocence applies, since the presumption requires any doubt to go to the benefit of the person charged. 

In competition proceedings before the Commission, the presumption of innocence broadly applies. As laid down in Article 48(1) of the EU Charter of Fundamental Rights, it requires that ‘[e]veryone who has been charged shall be presumed innocent until proved guilty according to law’. Even though the qualification of competition law enforcement as criminal or quasi-criminal is a touchy subject (as evidenced by Article 23(5) of Regulation 1/2003), the Union Courts since the 90s hold, that ‘given the nature of the infringements in question and the nature and degree of severity of the ensuing penalties, the principle of the presumption of innocence applies to the procedures relating to infringements of the competition rules’, as long as such procedures ‘may result in the imposition of fines or periodic penalty payments’. Since Regulation 1/2003 allows for fines against intentional or negligent infringements of Articles 101 and 102 TFEU, every single proceeding the Commission carries out under that regulation – irrespective of whether the Commission ultimately decides to impose a fine or not – is subject to the presumption of innocence.

How to lower the standard of proof

To not be subject to the presumption of innocence – and with it the ‘beyond any reasonable doubt’ standard of proof –, a proceeding would have to exclude the mere possibility for the Commission to impose a fine upon the undertaking. In this case, as clarified by the case law of the ECHR, the procedural safeguards do not apply in the same way, and leeway for a less strict standard of proof is opened.

This is where the German government’s proposal comes in, which aims at allowing the Commission to choose between two distinct tracks of enforcement. One, which includes fines as an option, and another, which excludes them from the outset. Only the former would be governed by the presumption of innocence, while the latter could make use of a lower standard of proof. This resembles the German system of public enforcement, in which the Bundeskartellamt has the possibility to choose between a type of procedure governed by criminal law standards and including the option to impose fines, and a purely administrative type of procedure without a possibility to fine the infringing undertakings. In Germany, this leads to the Bundeskartellamt predominantly imposing fines in cases of collusion, largely refraining from sanctioning undertakings in the realm of abuse of dominance. Against the background of the problem sketched above, this makes sense intuitively, since Article 102 TFEU, compared to Article 101 TFEU, comes with the inherently complex task of defining markets and assessing dominance, as well with situations in which the harm to competition is often less generalizable and subject to a more detailed assessment of the circumstances of the individual case.

Alternatively, and less intrusively, the revised procedural regulation could simply give the Commission the possibility to waive its power to impose fines at some early point in the proceeding, which would have to be defined more precisely in the regulation. While a more comprehensive option would allow to further differentiate the proceedings in accordance with the different levels of fundamental rights protection, for example, in how extensive the access to file provisions are worded, both proposals would provide the Commission with the option to conduct an infringement proceeding that excludes fines from the start. These would no longer be proceedings which ‘may result in fines’, and therefore, they would not be subject to the presumption of innocence and the criminal law standard of proof.

Which impact could it have?

The formulation of such a lowered standard is no simple task (even until now, the courts have avoided doing so for the enforcement of Articles 101 and 102 TFEU explicitly).

With care and attention to the substantive differences, notably a less clearly defined burden of proof, and a by-nature prospective analysis, merger control might provide a useful comparison, since the applied standard of proof resembles a ‘more likely than not’ or ‘balance of probabilities’ standard. Under such a standard, the court’s conviction that a fact is more likely true than not would be sufficient for that fact to be proven. In any case, absent the presumption of innocence applying, not ‘any reasonable doubt’ would prevent the courts from being convinced that the competition rules are, in fact, infringed.

This can be of particular value to the effectiveness of competition law remedies. If the Commission no longer must prove the relevant facts ‘beyond any reasonable doubt’, its assessment should be less intensive. The proceedings before the Commission will speed up, which enhances the effectiveness of remedies in different ways. In each individual case, expedited administrative proceedings ultimately make it easier for the final prohibition decision to remedy the respective distortion of competition because the anticompetitive conduct in question will have had less time to negatively affect the competitive structure. On a larger scale, individual cases will tie up resources in the Commission’s internal workflows for a shorter time, therefore enabling the Commission to ultimately investigate more cases and issue more prohibition decisions. This would be in line with a contemporary re-evaluation of antitrust’s error-cost framework under the impression that especially in highly dynamic environments, such as digital markets, the social costs of under-enforcement (or even enforcement that comes ‘too late’ to effectively counter the harm to competition already caused) are particularly high due to the tendency of such markets to consolidate beyond contestability.

The effects of such a change of procedure on deterring anticompetitive behaviour require a more differentiated analysis. In short, fines contribute to deterrence and, accordingly, omitting them decreases the deterrent effect of the rules being enforced. Yet, fines are not the only cost factor that undertakings account for in adapting their behaviour. Rather, a variety of possible costs, such as damages from private litigation, losses in equity due to reputational harm, or costs associated to a remedy-induced change in business model, must be taken into account. If the Commission would be enabled to pursue more cases overall, and even some kinds of cases it does not pursue under the current framework due to constraints in administrative capacity, costs from follow-on litigation and changes of the business model may increase. Additionally, the Commission would retain the option to conduct a fining proceeding and react to potential negative developments in this respect. After all this, a significant negative impact on deterrence seems unlikely.

Our analysis suggests that the option to introduce a type of procedure without the possibility of imposing fines should be seriously considered in the ongoing revision of Regulation 1/2003. A more comprehensive analysis can be found in our article ‘To fine, or not to fine’, that is available pre-print at SSRN, and forthcoming in the Journal of Antitrust Enforcement.