The European Court of Justice (CoJ) just came back to business with a bang. On September 6, 2017, it finally unveiled its long-awaited judgment in the Intel case. The CoJ refers the case back to the General Court (GC), which must now re-do its homework and take into account economic arguments put forward by Intel in its defense against the European Commission’s (Commission) findings that its exclusivity rebates were anticompetitive.
Run-up to the Judgment
On May 13, 2009, the Commission imposed a record fine of €1.06 billion on Intel Corporation (Intel) for engaging in abusive practices on the market for central processing units of the x86 architecture (‘x86 CPUs’, i.e. a key computer component) from 2002 to 2007. The Commission found that Intel engaged in two specific forms of illegal practices, namely: (1) exclusivity rebates, i.e. rebates granted to manufacturers on the condition that they would buy all or almost all of the x86 CPUs from Intel; and (2) so-called ‘naked restrictions,’ i.e. direct payment to a retailer on condition that it stocks only computers with Intel x86 CPUs (hereafter, together, ‘the Intel rebates’).
The Commission took the view that the Intel rebates were, in and of themselves, sufficient to find an infringement. Owing to their very form, the Commission could assume that they were capable of restricting competition. Therefore, it was not necessary to show that the Intel rebates had anticompetitive effects. For the sake of completeness, the Commission did nevertheless examine whether such anticompetitive effects could arise, on the basis of the ‘as efficient competitor’ test (AEC test).
Intel sought the annulment of the Commission decision before the GC. In particular, Intel argued that the Commission was under the obligation to look at all the relevant circumstances to ascertain whether the rebates at hand were in fact capable of restricting competition, and had failed to do so correctly. Intel also argued that the AEC test conducted by the Commission was flawed.
In a judgment dated June 12, 2014 (T-286/09), the GC dismissed Intel’s action in its entirety. With regards to the Intel rebates, the GC’s analysis revolved around two issues:
- The Commission was not required to assess Intel’s rebates in light of all the relevant circumstances. In support for this affirmation, the GC distinguished three categories of rebates:
- Category 1: volume-based rebates, which are presumed lawful;
- Category 2: exclusivity rebates, such as Intel’s, which are presumed unlawful. As a result, according to the GC, the Commission was not required to examine whether Intel’s practices were, in fact, capable of restricting competition;
- Category 3: rebates based on a mechanism that may have a fidelity-inducing effect, for which an analysis of all the circumstances is necessary in order to determine whether such rebates are capable of restricting competition.
- In any event, the GC took the view that the analysis of the rebates in question taken in their relevant context did show that Intel’s rebates were capable of restricting competition. The GC did not examine Intel’s claim that the Commission had committed serious errors in its AEC test, since it considered that the AEC test was irrelevant to the analysis.
On August 26, 2014, Intel appealed the GC’s judgment before the CoJ. In a very critical opinion issued on October 20, 2016, Advocate General (AG) Wahl recommended that the CoJ set aside the GC judgment (see our briefings here and here).
Exclusive Rebates Are Not Anticompetitive ‘By Object’
Contrary to the lower court, and in line with the recommendation of the AG, the CoJ makes no distinction between exclusivity and loyalty rebates, i.e. category 2 and category 3 rebates in the GC’s classification.
Instead, the CoJ considers that both categories of rebates are subject to the same legal regime under Article 102 TFEU, specifically: (1) exclusivity and loyalty rebates are presumptively anticompetitive, unless (2) the investigated dominant undertaking provides evidence that a rebate, either exclusive or not, ‘is not capable of restricting competition and, in particular, of producing the alleged foreclosure effect.’ Arguably, the CoJ sets a presumption of illegality. However, in practice, dominant companies will invariably fight this presumption by producing economic evidence showing that the impugned rebate scheme is not anti-competitive. In sum, the CoJ instructs the Commission to take a close look at the economic effects of all rebates before concluding on their illegality under Article 102 TFEU.
In examining such effects, the Commission must undertake a two-step analysis:
- First, it must assess whether the rebate is capable of producing a foreclosure effect, by examining all of the relevant circumstances, in particular:
- The extent of the undertaking’s dominant position on the relevant market;
- The share of the market covered by the challenged practice;
- The conditions and arrangements for granting the rebates in question, their duration and their amount;
- The existence of a strategy aiming to exclude as efficient competitors.
- Second, once it is established that the rebate is capable of foreclosing as efficient competitors, the Commission must examine whether those effects are outweighed by efficiencies which may benefit consumers.
Critically, the CoJ makes repeated references to the fact that anticompetitive foreclosure under Article 102 TFEU only arises where the practice under investigation excludes an as efficient competitor from the market. The exclusion of less efficient competitors, however, will not be deemed anticompetitive. This is a welcome clarification, albeit entirely consistent with the (non-binding) Guidance paper of the Commission on the application of Article 102 TFEU.
The CoJ noted that, in examining Intel’s rebates, the Commission did undertake the above two-step analysis – although on a ‘for the sake of completeness’ basis. More specifically, the Commission carried out an AEC test, which led it to conclude that the Intel rebates were capable of generating anticompetitive foreclosure effects.
In its judgment, the CoJ takes issue with the GC’s refusal to examine Intel’s arguments that the Commission’s AEC analysis was flawed.
Therefore, the CoJ annuls the GC judgment and refers the case back to the GC for a review of the factual and economic arguments put forward by Intel (NB: the CoJ could have decided to give a final judgment on the matter, but considered that the state of the proceedings did not permit it – presumably because the review of the AEC test requires a complex assessment of economic and factual evidence).
Beyond Rebates – Towards the Extension of the Extraterritorial Reach of EU Competition Law
The Intel judgment is also significant insofar as it addresses another critical issue in EU competition law enforcement, namely its extraterritorial reach.
The CoJ rejects Intel’s argument that the Commission did not have jurisdiction over two agreements entered into by Intel and Lenovo in Asia (‘the Lenovo agreements’). In doing so, the CoJ makes two very interesting points:
- The CoJ recognizes for the first time that the qualified effects doctrine, which was recognized in the context of merger control proceedings in Gencor, applies to antitrust investigations as well. Under this doctrine, the Commission’s jurisdiction is deemed established provided that the conduct has substantial, immediate and foreseeable effects in the European Economic Area (‘EEA’). In practical terms, this doctrine extends the Commission’s jurisdiction to cases where a given practice does not generate sales in the EEA (e.g. market sharing agreement or refusal to deal).
- Contrary to the recommendation of its AG, the CoJ finds that the Lenovo agreements had qualified effects in the EEA, insofar as they formed part of an overall strategy aimed at foreclosing Intel’s competitor from the market. According to the CoJ, ‘to do otherwise would lead to an artificial fragmentation of comprehensive anticompetitive conduct.’
Conclusion - A Whole New World Indeed
Assessing the legality of rebates under Article 102 TFEU is a headache for companies in search of practical solutions that provide a high level of legal certainty. Over the years, the administrative practice and the jurisprudence have considered different approaches to that end. Under one approach, exclusive rebates should be regarded as anticompetitive, regardless of their concrete effects on the market. This rather formalistic approach stems from established case-law, which includes Hoffmann-La Roche as well as the GC judgement in Intel.
The CoJ judgment in Intel marks a major departure with this line of case-law. From now on, all rebates, whether exclusivity or loyalty-enhancing, must be examined in light of all of the relevant circumstances, including their economic effects, and against an as efficient competitor standard. Arguably, this new approach is a – probably small – opening for dominant businesses willing to offer exclusivity rebates: under certain circumstances, such rebates might well withstand close antitrust scrutiny.
What the CoJ judgment does not tell us is: what are the market circumstances under which an exclusivity rebate may not be regarded as anticompetitive? The judgment lays down the parameters (e.g. the share of the market covered by the practice, the duration and amount of the rebate) to be taken into account in the economic assessment of rebates, but provides no guidance as to the thresholds beyond which an exclusivity rebate would be illegal under Article 102 TFEU (e.g. what market coverage, duration and amount would make the rebate exclusionary?). The CoJ is also critically silent on how the AEC test should work in practice.
Hopefully, these questions will now receive adequate attention by the GC when it reviews again Intel’s rebates.