The author of this paper, available here, was for a long time the President of the Court of First Instance (now the ECJ’s General Court). More importantly for our purposes here, he was also the CJEU judge responsible for drafting the Intel judgment.
The paper is structured as follows:
A first section reviews how EU courts approach judicial review in complex matters, and how this approach has evolved over time.
For a number of years, the Court of Justice (ECJ) has taken a careful approach to the scope and intensity of review of Commission decisions as regards complex economic matters. From the outset, the Court conceived its role in competition matters as being limited to reviewing legality, and not as involving unlimited jurisdiction or full merits review (except as regards the imposition of fines). Since Consten & Grunding in 1966, the ECJ has acknowledged that the Commission must engage in complex evaluations of economic matters. The judicial review of these evaluations recognises the Commission’s expertise and confines itself to examining the relevance of the facts in evidence and of the legal consequences flowing from them.
This approach has been widely criticized as ‘formalistic’. Critics have advocate for “a more economic approach”, an “economics based approach”, or even an “effects based approach”. In response to these criticisms, the EU courts have progressively moved towards a more intense scrutiny of competition cases. Still, when the Commission pursues complex economic appraisals, the EU courts must limit themselves to a complete review of legality – in the sense that such review extends to all the defects normally examined by those courts in the context of an action for annulment. This is in line with how the ECJ reviews decision by the Commission or any other EU authority whenever they are called upon, in the performance of their duties, to make complex assessments of an economic, scientific or technical nature.
The following sections (II to IV) look at the Intel judgment.
Intel was seen as an occasion for the ECJ to abandon its quasi-per se rule of illegality approach to exclusive dealing and loyalty rebates, and replace it with a structured rule of reason analysis. However, the CJEU is not bound by precedent; instead, the Court takes each case on its own merits and does not purport to lay down a rule to be followed inexorably in the future. Furthermore, all judges must agree on the content of the judgment. The way the Court necessarily operates helps to explain why one should not expect to find in the Court’s case law a monolithic body of jurisprudence obeying a single pattern of decision-making or a unique conception of EU law. At the same time, the Court is able to rely on a wealth of prior case law that serves as a guide and beacon: any departure from a well-established line of case law must be transparent, duly motivated, well-reasoned, and respectful of legitimate expectations.
In short, the case law evolves incrementally. This was what occurred in Intel, where the law was clarified and the prevailing interpretation of Hoffman LaRoche (reflected in the General Court’s approach in Intel) rejected. According to this interpretation, some types of conduct by a dominant undertaking, such as so-called exclusivity rebates, are inherently anticompetitive or anticompetitive per se. The reason for clarifying the law on rebates was: ‘the fact that, during the administrative procedure, the undertaking concerned had submitted, on the basis of supporting evidence, that its conduct was not capable of restricting competition and, in particular, of producing the alleged foreclosure effects. The Court notably found that if, in a decision finding a rebate scheme abusive, the Commission after all carries out an analysis of the capacity of the practice to foreclose competitors as efficient as the dominant undertaking, the General Court must then examine all of the applicant’s arguments seeking to call into question the validity of the Commission’s findings’ (emphasis in the original).
Importantly, the Court of Justice did not take a position on whether the application of an as-effective-competitor (AEC) test was, in itself, necessary or not in the case at stake. The author notes that this is in line with the decision in Post Danmark II, which held that the AEC test is merely a possible tool among others for identifying abusive conduct. Yet, since the AEC test had played an important role in the Commission’s assessment, the Court of Justice held that the General Court was required to examine all of Intel’s arguments concerning the AEC test.
A third part of the paper seeks to describe what inferences can be legitimately drawn from Intel.
The author believes that this judgment – which he wrote – constitutes an important, albeit limited, step towards strengthening judicial control as regards the application of Article 102 TFEU. In short, he thinks it is fair to hold that it was the fundamental premise, adopted by the Commission and upheld by the General Court, that an examination of all the circumstances was not necessary to conclude that a system of exclusive rebates was capable of restricting competition in breach of Article 102 TFEU. This assumption led the Commission, and subsequently the General Court, wrongly to fail to take into consideration Intel’s argument seeking to expose alleged errors committed by the Commission in the application of the AEC test.
At the same time, the author does not think that Intel brings about any revolution in the way the Court of Justice looks at the conduct of a dominant undertaking in order to determine whether such conduct is abusive. Even in Hoffman La Roche the: ‘Court reached its conclusion that a system of exclusivity or quasi exclusivity rebates, such as the one in question in that case, is abusive, save in exceptional circumstances, only after carrying out a thorough analysis of the legal, contractual and economic context (‘all the circumstances’) of the practice in order to conclude that it had the ability to foreclose competitors’ (emphasis mine).
It is true that the meaning of the expression ‘all the circumstances’ was not further clarified in Intel. It is suggested that one should rely on the Post Danmark II list of factors, in which the Court stressed the importance of taking into account not only the criteria and rules governing the grant of the rebate, but also the extent of the dominant position and the particular conditions of competition prevailing in the relevant market. In any event, such an expression does not require a detailed effects analysis, because so-called exclusivity or quasi-exclusivity rebates offered by a dominant undertaking are, by their very nature, capable of leading to a finding of an infringement in a large number of cases, as the likelihood of having foreclosing effects are generally high.
At the same time, it is dubious that the Court decision can be read as setting out a presumption of anticompetitive behaviour when such a rebate system is at issue on its own – ‘That was not, in any case, the path explicitly taken by the Court.’ In any event, any such presumption would have to be grounded in economics and the objectives underpinning competition rules. More importantly, such a presumption would be rebuttable. In this case, the Court stopped short of admitting that the absence of anticompetitive effects could contribute to justifying the practice / rebutting the presumption of unlawfulness – instead, it held that it is not necessary to examine whether the practice in question did actually produce its intended exclusionary effects on the market.
In short, he thinks that Intel does nothing more than correctly to apportion the burden of proof between the Commission and the sanctioned company. The specificity of the assessment of a system of rebates that includes an exclusivity clause is that, once the Commission has discharged its duty to establish the likelihood of anticompetitive conduct, the hurdles will be particularly high for the sanctioned company when trying to rebut the Commission’s findings.
A last section points out that the Intel decision occurred in a specific procedural context.
In an appeal against a decision of the General Court, the Court of Justice is bound to accept the limits of the dispute as they have been set by the applicants. The Court cannot ‘invent’ new grounds of appeal that have not been raised by the applicants. It simply has to give an answer to the pleas raised and the arguments advanced in support thereof. That is what the Court did here.
Furthermore, the Court faces the challenge of turning economic theories into solid legal criteria, capable of securing the clarity of the concepts and their adaptability to a complex reality. The Court must also enhance legal certainty and predictability in the application of the law. In this task, courts have an arsenal of legal weapons available to them; the most obvious are the rules on the burden of proof.
This is an interesting analysis of the Intel decision – and, without wanting to sound like an originalist, the most authoritative one. Without false modesty, I think it fits with my reading of the judgment as being a limited development of the case law that focuses on the allocation of burdens of proof in the specific context of exclusive dealing / loyalty rebates, which I elaborated on here and here.
However, this reading came to me as much from the reasoning of the Court in Intel as from years of studying EU law decisions on other fields. Experience has taught me to be very careful about deducing any grant theories or sweeping changes from any individual decision adopted by the European courts. Which brings me to a concern: what does it say about a decision – and its reasoning style – when it needs to be clarified three times, as this judgment did?