Intel and the role of less efficient competitors.

In this paper, we discuss the recent Intel case on rebates, which puts particular emphasis on the role of effects based analysis and the as-efficient competitor (AEC) principle and test. The Intel case is informative as in this case the European Court of Justice (ECJ) attempted to clarify some of the issues in the previous case law regarding abuse of dominance. The ECJ used of an effects-based approach, accepting the relevance of AEC test in relatively stronger terms. We consider that while adopting a more effects-based approach is a positive step, limiting the scope of the analysis to the protection of only as-efficient competitors could result in overlooking anti-competitive conduct in certain situations, and a more-nuanced approach may be desirable.

We begin by giving a short overview on the relevant case law on competition on the merits and the AEC principle and then analyse the Intel judgments with this background.

Since Continental Can^[1], the focus of the case law on abuse of dominance has predominantly been on exclusionary abuses.^[2] In Hoffmann-La Roche^[3], the ECJ distinguished anti-competitive behaviour from the competition on performance (or competition on the merits.) Similar to the earlier case-law, the Court used a mostly formalistic approach in prohibiting certain conduct without examining the likely or potential effects of it. Specifically, the Court iterated that pure volume-quantity rebates do not constitute an abuse of dominance whereas exclusivity rebates are by their very nature abusive.^[4] In United Brands, the ECJ stated that allegedly abusive conduct may be objectively justified.^[5] In Michelin I, a case on target rebates, the ECJ stated that one has to consider ‘all circumstances’ in assessing whether the target rebates has a tendency to restrict purchasers’ freedom and to exclude competitors.^[6] However, the ECJ did not consider demonstrable effects of the conduct.^[7]

The ECJ did however show its openness to an effects-based approach in the context of predatory pricing in Akzo (1991).^[8] To determine whether the conduct constitutes an abuse, the Court used a test by making use of two related approaches that originated in the USA, namely of Posner (1974) and Areeda and Turner (1975).^[9]

Posner (1974)^[10] stipulated, in the context of predatory pricing, only exclusionary conduct that has the purpose or likely effect of exclusion of an equally efficient competitor (AEC) should be prohibited. Posner distinguished between two situations. According to Posner, selling a product with a price below its short run marginal cost (SMC) can only have an exclusionary intent on an AEC, whereas selling a product with a price below the long-run marginal cost may possibly have such intent. The principle outlined above and the test based on it are called the AEC principle and the AEC test, respectively.

Areeda and Turner (1975)^[11] also considered that in a regular market any price below the SMC constitutes a profit sacrifice and cannot be explained other than for intention to exclude rivals and should be prohibited. Compared to Posner, Areeda and Turner were more permissive of prices above SMC, as it is less clear whether such prices could be anti-competitive. Furthermore, citing the difficulties in using the SMC directly, they advocated the use of average variable cost (AVC) as a proxy for SMC.

With this background, the ECJ introduced a test that compares the average total cost (ATC) and AVC to the price in Akzo. According to the test, prices below AVC were considered abusive, whereas prices between AVC and ATC were only considered abuse if they can be shown to be part of an exclusionary plan.^[12]

In 2009, the Commission published its Guidance Paper^[13] in which the Commission set forward an effects based approach for treating abuse of dominance cases. The Commission endorsed the consumer welfare standard and the importance of protecting effective competitive process rather than protecting competitors.^[14] Although the Commission accepted the possibility that the presence of less efficient competitors may be beneficial due to dynamic considerations, it embraced the AEC principle and (versions of) AEC tests for potentially exclusionary pricing practices.^[15]

In the context of predatory pricing, the Commission advocated a slight variation of the test used in Akzo. For margin squeeze and rebate cases, it embraced versions of the AEC test. In the case of conditional rebates, the Commission’s test investigated whether an AEC can profitably set a price to capture the contestable demand by offsetting the discount that a retailer would otherwise get from a rebate.^[16] Specifically, the Commission looked at the threshold price that offsets the discount and compared it to long-run average incremental cost (LRAIC) and average avoidable cost (AAC). If the threshold price is above LRAIC, this indicates that the practice allows an AEC to compete, whereas a price below AAC indicates clear foreclosure. Finally, a price between these rates would indicate that further investigation is necessary.

The Guidance Paper affected the case-law that came after, albeit not consistently. A case in which the effects could be seen was Post Danmark I, which was on selective pricing.^[17] In Post Danmark I the ECJ used a slight variation of the AEC test used in Akzo and it attempted to clarify the definition of competition on the merits and set out the underlying principle of AEC in the context of pricing practices:^[18]

“Thus, Article 82 EC prohibits a dominant undertaking from, among other things, adopting pricing practices that have an exclusionary effect on competitors considered to be as efficient as it is itself and strengthening its dominant position by using methods other than those that are part of competition on the merits.” ^[19]

It is important to note that the ECJ also stated that competition on the merits, “may, by definition” lead to departure of less efficient competitors, and that the Court did not mention the possible pro-competitive consequences that may rise from the presence of such competitors.^[20]

It may have been expected that the effect-based approach would also prevail also in the case of rebates. However, in Tomra^[21] the General Court (GC) did not find it necessary to review the relevant effects analysis of the Commission.^[22]

The first judgment of GC on Intel came just after Tomra. Intel sold x86 computer processors to various companies, including Dell, HP and Lenovo, and it used various rebate schemes that were conditional on whether the buyer buys all or significant portion of its processors from Intel. In 2009, the Commission found that these exclusivity rebates together with alleged naked restrictions constitute an abuse of dominance.^[23] To reach a decision, the Commission used formalistic arguments, but it also conducted an AEC test.

In 2014, the GC upheld the Commission’s decision^[24] stating that exclusivity practices, purchasing obligations or exclusivity rebates should be considered inherently anti-competitive^[25] and it ignored the AEC test-based analysis. In its ruling, the GC distinguished anticompetitive exclusivity rebates from quantity rebates that are typically not anti-competitive and a third category of rebates that may have fidelity-building effects which require detailed analysis. Intel together with supporting parties appealed the decision to the ECJ.

Just before the ECJ made its ruling on appeal, in Post Danmark II, a case on rebates, the ECJ stated that an AEC test is only “one tool among others” and its use is not obligatory for establishing the abusiveness of a rebate scheme.^[26]

However, on the Intel appeal the ECJ ruled that the judgement of the GC was inadequate and turned it back.^[27] According to the ECJ, the Commission cannot only rely on a purely formalistic approach and it should also consider the intrinsic capacity of the dominant undertaking to foreclose an as-efficient competitor.^[28] The ECJ further emphasized that the AEC test played a role in the Commission’s analysis and it was wrong for the Court (GC) to dismiss Intel’s arguments with respect to the Commission’s AEC test.

The ECJ clarified that effects-based analysis is distinct from the ability of the undertaking to present an objective justification. Furthermore, for the case in which the dominant undertaking presents evidence or arguments, the ECJ identified 5 factors^[29] that the Commission needs to consider:

• The extent of the undertakings’ dominant position.
• The share of the market that is covered by the alleged practice.
• The particular conditions and arrangements for granting the rebates.
• The duration and the amount of these particular conditions and arrangements.
• The possible existence of a strategy that aims to exclude at least as efficient competitors.

The last factor refers to the general AEC principle, although the ECJ did not state the necessity of an AEC test.

After taking the case back from the ECJ,^[30] the CG adhered to the principles defined in the ECJ’s judgment. Among other things, it stated that a system of rebates set up by dominant is not a per-se infringement, and that if the undertaking submits supporting evidence that indicates that its conduct was not capable of restricting competition or producing foreclosure effects, the Commissionneeds to consider the five criteria.^[31] Furthermore, in accordance with the ECJ judgment, the GC considered the AEC analysis of the Commission and the related arguments of Intel. The Court stated that the Commission’s AEC analysis contained errors; the Commission did not properly consider the share of the market covered by the alleged conduct and the duration of the conduct. As a result, the Commission was not in a position to determine the likely anti-competitive effects of the conduct nor its capability. Hence, the GC annulled the fine in relation to exclusivity rebates.^[32]

In its ruling, the Court cited the AEC principle of Post Danmark I. Furthermore, the ruling seems to indicate that the dominant undertaking can conduct their AEC tests and use it as supporting evidence to counter a claim of the Commission. Hence, effectively the AEC test may become a vital part of the abuse of dominance cases related to rebates.^[33]

It is however important to note that the Court did not mention the anti-competitive affects that may arise from the exclusion of less efficient competitors. This omission by the Court may be interpreted by undertakings to mean that such conduct is allowed, and that may have undesirable consequences for competition.

First of all, the economic theory tells us that less efficient competitors can put downward pressure on current prices. This will hold as long as the price in the scenario where the competitor is excluded is higher than the marginal costs of the excluded competitor. Furthermore, as others also pointed out, when products are differentiated, it makes little sense to compare the efficiency of competitors, as potentially less efficient competitors can provide valuable products for a subset of consumers.^[34]

Secondly, there is a real possibility that less efficient competitors become as-efficient in the longer run due to economies of scale. Nascent competitors and entrants may need sufficient time and investment to reach to the appropriate scale to be able to compete with dominant undertakings. Knowing this, dominant undertakings may attempt to exclude nascent competitors or foreclose the entry of entrants in order to guarantee that the current level of prices are also maintained in the future.

Hence, from an economics point of view, it may be desirable to pursue a more nuanced effects-based approach, that accepts the potential pro-competitive role of less efficient competitors. A nuanced approach would look at the present market structure including the market shares of at-least equally-efficient and less-efficient competitors, the amount of product differentiation, the barriers of entry, the economies of scale, uncertainties and bargaining power discrepancies between the players in the market.

For instance, in a market in which there is little product differentiation and several competitors with similar efficiency levels, one may be less wary of conduct that is potentially exclusive towards less-efficient competitors, as such conduct is not expected to increase the prices significantly. However, it would be more concerning if a significant portion of the competitors are relatively less-efficient, there are significant economies of scale and customers do not have sufficient bargaining power to counter-balance the potential consequences of the exclusive conduct.^[35]

Note that some level of flexibility for such considerations is potentially provided in the GC’s ruling- namely in the criterion on the share of the market covered by the alleged conduct. That criterion may be read to indicate that it is desirable to consider the role of less-efficient competitors for cases in which the rebate scheme affects a significant part of the market covered by such competitors. But this is speculative and it remains to be seen whether the Commission will incorporate such considerations into its analysis.

Lastly, the general applicability of the principles to abuses other than rebates, or to non-pricing exclusionary abuses, is not entirely clear. Post Danmark I had already established that the AEC principle is applicable to pricing practices. In some recent judgements such as Google Shopping, the ECJ did not embrace the use of the AEC test or AEC principle in non-pricing exclusionary abuses, whereas in the very recent judgment of Unilever, the ECJ stated that the relevance of AEC test cannot be ruled out, especially in the context of exclusivity clauses.^[36]

It may be argued that the case of Intel shows that the courts have begun to embrace a more effects-based reading of Article 102 in the case of rebates and in general exclusionary abuses. In that context, the ECJ introduced an analysis based on 5 criteria, which indicates that there will be a higher burden of proof on the Commission to establish an infringement. Furthermore, by requiring that the GC to analyse the AEC test of the Commission, the ECJ effectively gave an opportunity to dominant undertakings to conduct their own AEC test, possibly making the AEC test a standard tool in disputes related to the rebates.^[37] Unilever indicates that the same can be said for exclusivity clauses.

Shortly before the preparation of this paper, the Commission appealed the judgement of the GC and by extension the earlier judgement of the ECJ in a nearly unprecedented way.^[38] The details of the appeal are unknown, but it may be expected that the appeal would be against the arguments relating to the (in)adequateness of the Commission’s AEC test.

In any event, the broad AEC principle and the accompanying AEC test neglect potential anti-competitive effects that may arise from the exclusion of less-efficient competitors. Accordingly, it would be interesting to see whether the courts and the Commission will hold to the broad Post Danmark I standard, or will be more flexible in their approach accepting the relevance of the presence of less-efficient competitors.