In its ruling (here) on the European Commission's 500 page Intel decision, in a crisp 150 paragraphs, the EU Court of Justice (the Court) revisited forty years of jurisprudence on when a dominant company's rebate scheme may be abusive. Though no final decision for Intel, the case marks a potentially major departure from the arguably form-based per se (more accurately, "ordo-liberal") approach to rebates advocated by the EU's lower court. The ruling draws together the two lines of debate, looks to blend them together but, in doing so, comes down firmly on the "effects" based side - all an investigated party needs to do is raise the argument, with supporting evidence, that there is no actual foreclosure in order for the Commission to be required to look at the "as efficient competitor" arguments in detail.
- As efficient competitor (AEC) benchmark endorsed. The Court reverses Post Danmark II (§55-57) to bring rebates case law into line with "pure" pricing abuses (predatory pricing and margin squeeze). What matters is whether the rebate scheme would exclude competitors as efficient as the dominant company. If a scheme causes an inefficient competitor to exit, that is part of the competitive process. (§ 133-134).
- A full analysis must be undertaken if the defendant raises a no exclusion defence. Established case law holds that tying customers to a dominant company by exclusive dealing or pricing schemes with a similar effect will be abusive (§137). But with marked understatement, the Court finds that this case law "must be further clarified" if the defendant puts forward reasons why its scheme has no exclusionary effects. If that is the case, a full market analysis is required, and an assessment of a strategy aiming to exclude "as efficient" competitors. (§138-140)
- No per se illegal category of loyalty rebates. The lower court had not reviewed the adequacy of the Commission's foreclosure analysis, since it found there was no need to do so. The rebate schemes were exclusionary on their face. For this legal error, the Court reversed and remitted to the lower court for further review. The Court therefore reversed the lower court in its most contentious finding that certain types of rebate could be illegal per se - because they are conditioned on exclusivity or near exclusivity. Where a defendant raises a no-exclusion defence (likely always to be the case in practice), then a merits analysis must be undertaken. There is no per se (or 'by object') short cut.
- De minimis revived; Art 102 priorities guidance reprieved. A review of the scheme must involve an assessment of market power, market coverage of the scheme, the scheme mechanics and its potential to exclude as efficient competitors. By implication, therefore, Post Danmark II's denial of de minimis (lack of impact) defences in rebates cases is overruled. So too the Court implicitly endorses the AEC approach of the Article 102 enforcement priorities guidance, whose continued applicability was in doubt after the lower court's judgment.
The judgment is to be welcomed, for all that the Intel saga is far from over. Companies have greater scope for crafting compliant rebate schemes without the difficulty generated by the lower court's per se abusive rebate category. They can take comfort that properly devised schemes can be defended if they show no potential for exclusion.
Intel is an allegedly dominant supplier of central processing unit (CPU) chips for computers and servers. According to the Commission's 2009 decision, Intel agreed with its main desktop customers that it would pay them substantial rebates in return for exclusivity or near exclusivity, amounting to between 80% and 100% of their needs. The rebates in some cases totalled hundreds of millions of dollars. Intel also paid additional amounts to these customers and to a PC retailer to (i) not stock competitor-chip-based PCs, (ii) delay the introduction of rival chips, and (iii) confine the competitor chips to non-strategic products.
The Commission found this to be abuse of a dominant position and imposed fines of €1.06 billion. The decision was upheld by the EU's first tier appeal court, the General Court.
The lower court held that there was no economic defence to exclusivity-linked rebate schemes - for a dominant company, a rebate scheme linked to exclusivity, being the purchase of all or most of a customer's requirements, is presumed to be illegal irrespective of its actual effect.
The Court disagreed. Exclusive dealing or pricing schemes with a similar tying effect had been held to be an abuse of dominance in the past. But this case law required further clarification. It was not the case if the defendant put forward evidence showing that the scheme did not exclude competitors.
This exercise must involve consideration of (§139):
- "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 possible existence of a strategy aiming to exclude competitors that are at least as efficient as the dominant undertaking from the market."
The Court finds that demonstrating likelihood of "as efficient" competitor exclusion is also required by the concept of Article 102 objective justification. Without understanding whether as efficient competitors will be excluded, it is impossible to know whether a scheme shows countervailing benefits outweighing any restrictive effects. "That balancing of the favourable and unfavourable effects of the practice in question on competition can be carried out in the Commission’s decision only after an analysis of the intrinsic capacity of that practice to foreclose competitors which are at least as efficient as the dominant undertaking." (§140)
This conclusion reversed the lower court's finding that some types of rebate are abusive per se. It also, implicitly, overrules Post Danmark II (§74), to which the Court does not refer, that rules out de minimis defences. The Court in Intel suggests that the scope of the alleged illegal practice's market coverage is central to a foreclosure analysis. Of necessity therefore, de minimis scope must suggest lack of exclusion.
The Commission had put forward a detailed economic foreclosure model. It had analysed the "contestable share" of sales available to Intel rivals. It had then modeled the impact of Intel's rebates in preventing rivals competing for that share. It found that even if they had the same cost base as Intel (an "as efficient" competitor), they could not viably have competed on price. This analysis was skirted over by the lower court, despite Intel's challenges. It had found no need to review the AEC analysis because it held that other elements of the rebate schemes demonstrated exclusion. The Court reversed this finding, and remitted the case for further review by the lower court.
Intel argued wider jurisdictional points, claiming that the EU did not have jurisdiction over some of the relevant agreements and periods in so far as they related to conduct, parties, sales or purchases outside the EEA. The Court found sufficient nexus to the EEA in pricing practices aimed at Lenovo personal computers. There were sufficient indications of real, substantial and foreseeable impact on the EEA given the likelihood of potential exports from China to Europe.
In addition to the substantive elements, Intel appealed the General Court's finding that the Commission made no procedural errors in relation to a five-hour interview held with an employee of one of Intel's main desktop customers (Dell). Intel challenged the General Court's conclusion that it was sufficient for the Commission to provide Intel with a list of topics discussed rather than a record of the interview.[i] Intel also argued that the Court erred by concluding that it was up to the company to provide evidence that the Commission failed to record exculpatory evidence - i.e. evidence that would have helped Intel's case.
The Court concluded that the Commission was required to record these interviews. There were no such thing as informal interviews. However, as is often the case with procedural issues, it held the error had not invalidated the final result. It was incumbent upon Intel to show that the information provided would have been exculpatory in nature. Intel had not satisfied that burden - all parties agreed that the Commission had not relied upon the evidence in its decision. The Court sets a high bar for legal error here since it will generally be impossible to know the nature of unrecorded evidence. The Court's reasoning is also fanciful in part, saying that Intel could have sought to have Dell witnesses compelled to testify before the General Court as to the exculpatory nature of any such evidence, or sought to access the Dell witnesses by other means.
Volume-linked rebate schemes are a daily occurrence. They reduce prices and can increase output. They are an efficient means of solving price/volume negotiations where buyer and seller find it hard to predict future volumes. The new benchmark is easily invoked and offers greater flexibility for practitioners counselling on rebate schemes. There is no per se indefensible rebate category. Rather companies can defend well-crafted rebate schemes - small steps, low price increments, appropriate reference periods - based on the merits.
The case also signals a reprieve for the Commission's Article 102 Priorities Guidance, which adopts an AEC test in its analysis of rebate schemes. A hard line by the Court on rebates might have caused this to be revised.
The Intel saga is far from over. But this case brings a welcome (for many) re-injection of economic analysis into the debate.