A Clash of Swords (The Intel Hearing, by Trevor Soames) (Part II)
Below is the second part of Trevor Soames‘ excellent, original and very detailed narration of the Intel Hearing held yesterday in Luxembourg. Chapter 1 deals with jurisdictional issues; Chapter 2 with rebates; Chapter 3 with procedural issues and Chapter 4 with fines. Enjoy!
Prior to the commencement of proceeding the Court asked the parties to address their oral pleadings to three questions, as follows “Pursuant to Article 61(2) of the Rules of Procedure, the parties are requested to focus their pleadings on the Commission’s jurisdiction over the agreements concluded between Intel and Lenovo for 2006 and 2007. The parties are also requested to state their views regarding the impact of the judgment in Post Danmark (C-23/14) on the characterisation as loyalty rebates applied to agreements concluded by Intel and on the procedural handling of the interview conducted by the Commission with Mr D.”
The following report seeks to reflect fairly and as precisely as possible what the different parties said, dealing with each of the issues in turn (although, for the sake of brevity, the closing speeches are not recorded here). As there is no Rapport d’Audience and as a result non-parties have no access to a summary of the pleadings, the oral argument provides an important insight into what each side has said to the Court in this very important case on a number of key issues. Also, and importantly, the questions posed primarily to the Commission and the extensive Q&A by AG Wahl in particular perhaps gives some insight into the direction of travel he may be following in the preparation of his Opinion. Will that Opinion provide a much hoped for clarification in this important area of competition law as well as a correction to the much-criticised judgment of the General Court in Intel? Will AG Wahl’s Opinion have the same importance and impact of a number of his other Opinions in the competition law arena, such as in Cartes Bancaires. We will have to wait and see. And, of course, even after the AG opines we will have to wait to see the extent to which the CJEU follows his advice.
The report therefore seeks to provide a logically structured yet still verbatim account of yesterday’s hearing, with some commentary and observation contained in the final part (which will come tomorrow).
Part the First: Jurisdiction
Daniel Beard QC for Intel referred to the 2006 and 2007 agreements between the US company Intel and the Chinese company Lenovo. Intel sold the CPU units in China for incorporation in China into devices. Jurisdiction over that conduct does not follow from Woodpulp. In that case, the CJEU held that the decisive factor in determining whether the Commission could exercise jurisdiction is the place of implementation of the agreement. Applying the principle of territoriality under international law. So, the key issue is where the place of implementation was, in this case that was where the CPUs were sold to purchasers –in China. Any implementation took place in China where those CPUs were sold. The General Court also sought to rely on the so called qualified effects test referred to in Gencor. However, the General Court in Gencor was not seeking to expand jurisdiction beyond what was contemplated in Woodpulp…. Qualified effects is not an additional test that expands jurisdiction beyond what Woodpulp states. Even if a qualified effects test could be applied where the implementation test is not met there is still no jurisdiction here. The direct and immediate effect was in China not the EEA. Intel chipsets were sold to a Chinese company in China, incorporated into products in China and only after those two critical events had occurred did Lenovo export its products to third countries which could include the EEA. Hardly substantial, immediate and foreseeable. Since pleadings closed in this case, we have had the proceedings in Innolux. In that case AG held that both the implementation test and the qualified effects test only allow jurisdiction over direct sales. CJEU held it was not necessary to address jurisdiction but at para 72 it held consistent with Woodpulp, showing support of the CJEU for the implementation test. As a result, the Commission did not have jurisdiction over Lenovo.
Nicholas Khan for the Commission stated that the General Court finding that Woodpulp implementation test and Gencor qualified effects test were both available and that both were satisfied on the facts of this case. Commission is not asserting some sort of universal jurisdiction as Intel claims.
There were two abusive arrangements with Lenovo both of which were intended to prevent Lenovo from purchasing any AMD CPUs for incorporation in a range of Lenovo devices to be sold worldwide, inevitably including the EU. Intel relies on a very narrow interpretation of Woodpulp and concept of implementation which would prevent EU jurisdiction over “stay at home” cartels in e.g. Japan such as in Seamless Tubes etc. Woodpulp is not an exhaustive statement on jurisdiction as is shown in the CJEU decision in the Air Transport Association of America case. Woodpulp was an Article 101 cartel case, fundamentally different from this. The pricing behaviour applied to an input which was sold to and incorporated by Lenovo into a final product outside the EU. But those practices were conceived with a view to shutting out of the EEA market products incorporating AMD CPUs.
On qualified effects, the Innolux case should be understood as moving the Gencor test beyond mergers (73, 74) it accepted that sales of the transferred product should be taken into account when setting the fine. Jurisdiction lay on the transformed products due to the application of the qualified effects test…. Intel’s abuse may have been manifested initially by the incorporation on Intel CPUs into Lenovo computers but that was not just the initial step, there was no supervening event that could have taken place thereafter that could have perfected the implications of that conduct on the EEA. It follows that the effects of Intel’s conduct regarding Lenovo were immediate, direct and foreseeable. Intel was well aware that competition at the input level of the market (chipsets) takes place not only at the level of the input market but also on the adjacent market for the finished products due to e.g. the intensive Intel Inside marketing program at the retail level, reimbursing OEMs for promoting the Intel brand at retail level.
Jean-Francois Bellis for intervener ACT noted that divergent positions have been taken. The General Court has held that there are two alternative jurisdictional bases, the implementation test and the qualified effects test. The General Court held that jurisdiction can be based on the qualified effects test rather than implementation. In Innolux however the General court held that there is only one jurisdiction test, namely the implementation test. The General Court held that this was there was jurisdiction where the components were sold outside the EEA if sales of finished products were made in the EEA by the infringing parties. In this case the General Court went even further claiming there was jurisdiction even where the EEA sales were made by unrelated third parties. A clarification by the CJEU of the jurisdictional base to be applied is needed to restore legal certainty.
The applicable test can only be the implementation test. Article 102 makes clear that it applies only to abuses of a dominant position within the common market. In Woodpulp the CJEU held that Article 101 applies only to agreements implemented to direct sales of the product concerned in the common market. No need to apply qualified effects test, Gencor case does not affect this conclusion as it was an ECMR case involving two South African companies….
In this case all elements of the infringing conduct re Lenovo occurred outside the EEA. How can this be an abuse of Intel’s dominant position within the common market, implemented within the common market? At para 311 of the judgment it suggests that the abuse was implemented by Lenovo selling exclusively notebooks incorporating exclusively Intel x86 CPUs. A most peculiar theory. How can Intel’s abuse be carried out by someone else. This is not about Lenovo’s conduct, it is Intel’s and this conduct was implemented in China, not in the EEA. There is no targeting of the EEA market by Intel.
Questions from Advocate General Nils Wahl
Advocate General Wahl: On jurisdiction, in Woodpulp the implementation was essentially direct sales into the common market so there was no need for the CJEU to do anything else. If I have understood AG Wathelet correctly he suggests that in Innolux he does not think that we should include indirect sales in assessing implementation. Is the concept of implementation different depending upon what type of anticompetitive conduct is being reviewed. Should the concept of implementation be adaptable to the type of behaviour, as in this case we are talking about sales to and incorporation in China. AG Wathelet seems think no but I assume that you would say, yes.
Nicholas Khan: Innolux is a cartel case and we submit that it is not right to treat Woodpulp as some form of statutory code as it is a cartel case. AG Wathelet was prepared to accept the effects doctrine in Innolux but it was a cartel.
Advocate General Wahl: AG Wathelet says that indirect sales are not covered by the term implementation. You seem to say that implementation can include almost anything.
Nicholas Khan: I am not saying that implementation can cover practically anything.
Advocate General Wahl: If you are saying that implementation happens on the territory, exclusion by its definition means that it does not happen on the territory, it happens somewhere else. Boycotts happen somewhere else, could they still be implemented. You said earlier that “stay at home” cartels were implemented, but surely they are not. Are you trying to stretch implementation even more here? Is that the correct approach?
Nicholas Khan: As regards “stay at home” cartels the Court has seen a succession many cartels of that type and the jurisdiction argument was not raised, so I must dissent from the proposition that a “stay at home” cartel is not a form of implementation recognised by the Court. As to the application to this particular case, an abuse is very different and an implementation has to be understood in the context of the conduct concerned and once Intel had sealed its deal with Lenovo, then the rest inexorably follows. Intel knew that Lenovo would export some of its production to the EEA, thereby securing the exclusion of AMD by an abusive method it thereby follows that the conduct is thereby implemented. Nothing further can happen to the Lenovo note books once assembled in China with their Intel chipsets inside to undo the exclusionary effect as regards AMD between their manufacture in China and their sale in the EEA. See para 17 Woodpulp re agents and subsidiaries, see also Commercial Solvents where the Commission found an abuse by Commercial Solvents as it had control of the distributor in Italy.
Advocate General Wahl: let us assume that we find jurisdiction on the basis of qualified effects and not as Intel claims, there was a limitation on that concept, you say that it should be a direct, immediate, substantial and foreseeable effect. What kind of effect is that? Is it the same type as the effects when it comes to a prohibition as such. The effects you should see in the EEA are we talking about the same effects as are prohibited under Article 102 and is it of the same magnitude. In other words, would you fulfil the jurisdictional criteria with the criteria used for the prohibition as such.
Nicholas Khan: In short the answer is yes, our concern is that these arrangements affect the structure of the market within the EEA.
Advocate General Wahl: But does that not mean that something that you call by nature anticompetitive, is by nature something over which Article 102 has jurisdiction over even though nit happened in China. If you apply the same reasoning.
Nicholas Khan: The infringement of competition by nature has an effect on the structure of competition.
Advocate General Wahl: I assume that the jurisdictional criteria that will be decided upon by the Court will be applicable across the board, not only for the enforcement by the Commission, the NCAs etc. Does that mean that private parties can use that jurisdiction to claim damages here in the EEA for conduct taking place in China. Could AMD sue Intel for damages for exclusion that happened in China? This is somewhere you should not go in my view from a PIL perspective.
Nicholas Khan: Regarding conduct occurring outside the EEA but with effects within the EEA, there would be an infringement decision which could provide a basis for follow on private damages claims. The Commission only asserts an infringement with effects within the EEA. It could therefore provide the basis for a follow on claim subject to the usual issues of causation and quantification of harm.
Part the Second: rebates
Daniel Beard QC for Intel argued that as regards rebates, the Commission and General Court applied an unduly formalistic test. Rebates may have an anticompetitive effect. It is generally accepted that depending on the circumstances loyalty rebates can have beneficial effects by eg lowering prices to customers. The Commission has acknowledged this and stated in its Article 102 Guidance at para 37 that all rebates including loyalty rebates may stimulate demand and benefit consumers. Supported also by a large body of academic writing. The issue for competition law is whether the rebate could have an exclusionary effect. Loyalty rebates can have negative effects on competition depending on the circumstances, other conditional rebates can also have also have negative effects and in fact can be significantly more harmful than loyalty rebates. In Post Danmark II the AG emphasised at paras 29 and 31 held that it is ultimately immaterial if the rebate scheme can be attributed to a traditional category of rebates, in particular quantity rebates or loyalty rebates. The key issue being whether the rebates are capable of producing on the relevant market an exclusionary effect which is not economically justified. Whether they have that capability must be examined in the light of all the relevant circumstances of the individual case. That is correct approach and consistent with Tomra which required an analysis of all the relevant factors and the EFTA court case of Posten Norge which assessed the CJEU case law re loyalty rebates and assessed the facts and circumstances.
So what are the facts and circumstances that must be considered? Inquiry needs to focus on whether there is an exclusionary effect. A key element of assessing whether a rebate is capable of having an exclusionary effect is market coverage. Rebates are unlikely to produce an exclusionary effect unless there is appreciable market coverage. Commission, General Court and CJEU all considered market coverage to be relevant in Tomra, and other cases such as Vandenbergh. In this case Intels’ market coverage was far below previous Article 82 cases, especially in the last two years of the alleged infringement and especially tiny if the Lenovo agreements are excluded (less than 1%).
When assessing the capability of a rebate scheme to produce an exclusionary effect it may be helpful to assess whether the rebate will prevent AEC from competing for particular sales. The CJEU in Post Danmark II recognised the point and although it was not required it held at para 58 that it was not excluded in principle. The Commission applied the AEC test in relation to Intel but after Intel said that the test was misapplied, the Commission went on to say that it was irrelevant. The Commission cannot ignore an inconvenient truth.
In Post Danmark II it is notable how the CJEU characterised loyalty rebates in the judgment. At para 28 the CJEU held that a loyalty rebate is conditioned on or with an obligation to purchase all or most of its requirements from a given supplier. The emphasis was on the purchaser being bound for if the purchaser is under no obligation to purchase most or all of its requirements from the dominant undertaking to obtain the rebate, the arrangements cannot be classified as a loyalty rebate. An incentive based on pricing alone is not enough. It is only where a rebate is coupled with an obligation or promise of exclusivity, something beyond the price incentive, it would only be able to be identified as a loyalty rebate where no assessment of the circumstances is required. In this case both the Commission and the General Court classified Intel’s rebates as loyalty rebates on the basis of a completely different criteria, namely whether they were de facto conditional on exclusivity without assessing whether they were conditional on an obligation or promise. Consider Dell. Completely different from all other cases where a rebate has been classified as a loyalty rebate. The General Court held that as a result of “impressions” given by Intel, Dell assumed that Intel might in future reduce its rebates if Dell switched away from Intel. The finding of what Dell assumed was based on the impression the General Court formed from an Article 18 response. Dell never gave a commitment to buy exclusively from Intel. A rebate that might incentivise Dell to purchase from Intel is not a loyalty rebate within the meaning of paragraph 28 of Post Danmark II. It was therefore necessary to consider all the relevant circumstances to determine whether the rebate at issue was exclusionary. Intel’s rebates to HP and Lenovo could not be classified as loyalty rebates as they affected in each case only a segment of each of their requirements, 30% in the case of HP.
Nicholas Khan for the Commission argued that Post Danmark II reinforces the correctness of the Commission’s position. Intel argues that its exclusivity rebates were just a low pricing practice and that the General Court failed to assess all the circumstances. We do not accept the low pricing practice argument, relying on pleadings. In assessing the rebate scheme in Post Danmark II the CJEU did not find it relevant to refer to Post Danmark I save as regards the AEC test. As to the second argument, the reliance on para 28 of Post Danmark II does not really help in view of the broader statement made by the CJEU at para 27. As to the arrangements in question, Dell was not left merely with an impression and refers to recital 323. As to the reliance of Intel on AG Kokott’s Opinion in Post Danmark II that does not contradict the Commission’s position. That case shows that even a loyalty inducing rebate scheme is not a form of competition on the merits, a fortiori a loyalty rebate scheme such as in the present case infringes Article 102 without any need for further assessment.
Intel relies on Tomra exhaustively interpreted in pleadings. Interpretation lies in the Commission’s favour as is clear from Post Danmark II. Hard to think of a better category of abuse that points to an infringement by object, using the lexecon of Article 101 in Cartes Bancaires. Para 54 provides no comfort to Intel.
Even if the CJEU was willing to overturn decades of settled case law and require the Commission to consider in an exclusivity rebates case all the circumstances beyond the obvious circumstance of exclusivity, Intel has not showed that the judgement erred as regards the fact that customers might take the prices offered by Intel into account. The evidence shows that the rebates had the effect of disincentivising customers from purchasing from AMD (eg AMD offered HP 1 million free CPUs in desperation, but HP could only use 160,000, see recitals 956 and 945 of decision).
On market coverage, our position is supported by Post Danmark II, which notes the obligation on dominant company not to weaken any further the structure of competition and rejected application of a de minimis threshold (at para. 73). As regards the AEC test, Intel’s arguments about having passed the AEC test are inadmissible on appeal as there is no finding to that effect. Intel relies on para 58 of Post Danmark II but more tellingly para 60 recognises that a Less AEC can still apply pricing pressure.
Jean-Francois Bellis for ACT emphasised that paragraph 28 of Post Danmark II is very important, it defines loyalty rebates as those which are coupled with an obligation or promise by purchasers to obtain all or a given proportion of their supplies. But this is not how the General Court identifies the rebates in this case when identifying the Category II rebates which are ipso facto an abuse. The General Court identified in Category II as a loyalty rebate any rebate conditioned on exclusivity or quasi exclusivity and held them to be per se abusive with no need to assess facts and circumstances. But what is the economic difference between such rebates and the rebates in Michelin [NB JFB was counsel in the case] which would fall within Category III. The Michelin II rebate scheme involved 47 turnover thresholds and is no different from rebates placed in Category II. The common characteristic of all those rebate schemes is that in each case the purchaser has the option not to buy all or most of its requirements from the dominant supplier. It has not entered into a commitment or given a promise of total or partial exclusivity thus the rebate will not necessarily lead to partial or total exclusivity. Therefore an assessment of circumstances is needed before concluding that a rebate is likely to have anticompetitive effects. The importance of this distinction between an exclusivity obligation on the one hand and an exclusivity option on the other hand is discussed very lucidly by Prof Nicolas Petit J at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2567628
[Note from the scribe: the citation of Professor Nicolas Petit (University of Liege) related to his insightful distinction between exclusivity obligations and exclusivity options, to which paragraph 28 of Post Danmark II seems to refer. This distinction had been entirely ignored by the General Court in the Intel judgment, as explained in his paper on the judgment published by the European Competition Journal (available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2567628). This is a problem because as Jean François Bellis rightly explained, this distinction has always inhabited the case-law of the Court since Hoffmann-La-Roche, Michelin I, Michelin II and British Airways. Pursuant to this line of judgments, exclusivity options – aka price incentives – cannot be deemed loyalty (or fidelity) rebates, and thus ought not to be reviewed under a per se prohibition rule. Instead, “all the circumstances” matter.]
Questions from Advocate General Nils Wahl
AG Wahl: the General Court takes a very clear position on the rebates in question, holding that they are prohibited by their nature (paras 84/85 of the judgment) and that there is never any need, according to that Court, to take into account effects, market coverage, duration or whether there is an obligation on the customer to purchase exclusively or semi exclusively. In the Commission’s pleadings, it claims that rebates “generally’ result in an abuse. Your position seems more nuanced. What does “generally” mean, almost always but not necessarily always or almost never?
Nicholas Khan: “Generally” has been used to seek to anticipate the possibility, not used by Intel in these proceedings, to provide some ability for the dominant undertaking to justify its conduct based on efficiencies.
Advocate General Wahl: Are you saying that even these types of rebates could be justified by way of efficiencies?
Nicholas Khan: We do not a priori exclude that possibility. It is difficult to say more as there was no evidence on that matter and no discussion before the General Court on the issue. It is recognised that in respect of any infringement of Article 82 that justification can be offered, see GSK and the Opinion of the AG there.
Advocate General Wahl: However, if one looks at what the General Court said at para. 89 of its judgment, the rebates are prohibited by their form, their form as such, their nature, I do not really see how you could evoke any type of efficiencies.. As a result, even if Intel had put forward an efficiencies defence it would not have amounted to anything and would have had no effect. Is that right? If so, your position is more generous than the position taken by the General Court.
Nicholas Khan: We are not intending to depart from the position taken in the judgment which we are defending. Para 89 to which the AG refers needs to be read in its context, together with para 94, which says that it is open to a company with an exclusivity rebate scheme to seek to justify it by raising an efficiency defence and refers to the classic jurisprudence of Hoffman La Roche and British Airways. So, paras 89 and 94 together are ad idem with the Commission’s response to this appeal.
Advocate General Wahl: I simply note that in para 89 the Court says that competition is always restricted. Leave that part. The point is that you say that you think it could be justified on certain occasions. Now, even so, it seems to me that the approach by the Commission and the General Court in this case is strict, in fact very strict. I cannot really see any other instance of an abuse of a dominant position where the Commission or the case law, as you interpret it, has such a strict approach. Is this really the worst form of Article 102 abuse, or is it the tip of the iceberg and you are prepared to find other stuff that would be prohibited simply by the form and nothing else, nothing else matters, not the duration, not the market coverage, not the individual requirements of the customers.
Nicholas Khan: We respectfully dissent the classification that this is an abuse by form, rather it is an abuse by nature of the arrangements. It is difficult to pronounce on the question because at para 94 that whilst admitting of the possibility that even if the arrangements are in principle restrictive there could be a justification advanced. Surely the position is no different from the position advanced under Article 101 that you can have an agreement which is restrictive but that is without prejudice to the possibility that the agreement may be justified under Article 101(3). This is indeed an Article 102 case but as previously indicated, by reference to Greek GSK case, this is not the first time that we see an acceptance in the jurisprudence about a justification being advanced in an Article 102 case. Since there was no debate about the justification, there was no evidence and no argument before the Court it is difficult to take that point much further.
Advocate General Wahl: moving onto the AEC test. If I understand the General Court correctly, and I refer to paras 150/151, it is of the opinion that an AEC test is never useful. First, it is not needed to implicate someone to find an abuse as such and secondly, even if the result of the test was positive (that is to say, that it showed that AECs could survive the ordeal, so to speak) it would not count for anything anyway as the market would still be foreclosed. The position of the Commission is the same. You do not think that an AEC test has any bearing on this. But, if I look at Post Danmark II at paras 58 and 61, it would seem as if the CJEU actually thinks that, on occasion, the AEC test might be of interest. Now, furthermore, if I look in your Priorities Paper, the Guidelines on the enforcement priorities of the Commission in Article 102 cases and refer to point 27 of that Paper, it actually says that a positive outcome of an AEC test could play a role for whether you would enforce or try to enforce the rules.
My question is this, assuming that it actually has a role to play and the Commission in its enforcement paper says that yes, positive indications are of importance, does that mean that you have different approaches when it comes to priorities for enforcement and then, later on, your position in Court? While one takes into account the positive results of the AEC test, but the other, not? If it went the other way, I would not have any problems with it. If you would say that we would concentrate our assessment of efficiencies (AEC) when we see that something is going on and have a different approach when you come to the Court that I can see. Is it just me, am I misunderstanding your position in this case, or what? Know you do not want to discuss the Priorities Paper but this is a more general question. Do you have a different approach, or not?
Nicholas Khan: as regards the Priorities Paper as is clear from the judgment it is not temporally relevant. Here, on appeal, Intel has continued to invoke the priorities paper but not even Intel claims that the Priorities Paper is of application here but Intel argues rather the substance of the Paper must be right and should be applied to the merits of the case. So the Priorities Paper, because of the findings of the General Court and the timings of this case, is not directly in issue here.
Advocate General Wahl: with respect, my question was not that. When you are looking at priorities in your work are you looking at that one [the Priorities Paper] a positive result [of the AEC test] would mean that you would not investigate it but every time you come to Court your position would be that the AEC test is never relevant. That is my question, not the time aspect.
Nicholas Khan: The time aspect is relevant in this case because the decision to investigate in the Intel case was taken before the Priorities Paper. The AEC test was mentioned simply to reinforce the Commission’s view that this was an abuse that merited investigation. The Priorities Paper is what it says on the tin, it is a priorities paper, it cannot decide what the law is. If the Court determines that the law in relation to what is or is not an abuse is in a certain direction which is not necessarily in line with how the Commission may identify priorities then indeed there could be a position where the Commission has a broad discretion over setting priorities over what it going to investigate by reference to a test which I not necessarily determinative when it comes to assessing the merits of the matter at the end of the investigation. An AEC test is not excluded in Post Danmark II but of course when we get to that part of the judgment we are not dealing with an exclusivity rebate as we are here (there it was a retroactive rebate. The Priorities Paper may play a role in deciding whether to reject a complaint, but in deciding what is or is not an abuse under Article 102, the Commission is guided by the jurisprudence in reaching its final determination.
Advocate General Wahl: I take the last sentence as confirmation that you do in fact have two different approaches, one for assessing priority and the second one that it does not play a role when coming to legal proceedings. Is that correct, or not?
Nicholas Khan: With respect to the matter before the Court today, that is correct as the Priorities Paper is not the guidance which the Commission decided that this was an infringement.
Advocate General Wahl: As regards market coverage, it is relevant to this appeal. You argue that market coverage is irrelevant because under all circumstances the exclusive rebate will take away the customers’ choice. To be honest, it seems to me, that you distinguish between exclusive rebates of the type that the General court was dealing with in this particular case and other types of exclusionary abuse primarily based on the impact on the customer of the dominant actor. Assuming that we are, in fact, looking at exclusionary abuse, namely that Intel is trying to exclude AMD, what is the importance of the customers’ free choice as regards AMD and the effects on AMD. Should one understand the Commission’s position as being that neither the market coverage, nor duration, the amounts of the requirements of each individual customer have any relevance for concluding on abuse once dominance is established and the exclusivity of the rebates in question.
Nicholas Khan: The exclusionary abuse in respect of AMD is made out in the Hoffman La Roche case law formula. Once an exclusionary abuse is made out, it deprives the customer of choice and thereby deprives AMD of the possibility to access that customer.
Advocate General Wahl: We are not talking about the individual customer here, we are talking about the exclusion of AMD. To be frank we don’t really care about one or two customers, whether they have a free choice or not, what we care about is the exclusion of AMD, as a competitor.
Nicholas Khan: I refer to Tomra dicta to the effect that the customer must not be restricted as to his choice of supplier.
Advocate General Wahl: I know that, that is why I am asking. In all other circumstances, when we talk about that particular thing, the importance for the customer we normally do so in combination with market coverage and the amount of exclusivity for that particular customer. That is why I am asking. The position that you take here gives me the impression that it is solely the customer’s free choice that is capable, but that one will be restricted. But what I am trying to figure out is whether we should take that into account rather than the exclusionary effects towards AMD which is what Tomra and all the other cases talk about.
Nicholas Khan: In combination with the finding of the General Court at para 116 as confirmed by Post Danmark II there is no de minimis rule. Then we would submit that there is no need to consider and calculate market coverage. We are not looking at the % of market that AMD was frozen out of, though we claim that it was sufficiently substantial.
Part the Third: procedural issues
Daniel Beard QC for Intel referred to the handling of Mr D’s interview, noting that the General Court recognised that it was an interview but the Commission does not. It was a 5 hour interview of a senior executive summoned to Brussels. Mr D should have been interviewed under Article 19 Reg 1/2003, which was the only proper legal basis. Even if there was an ability to have an informal interview, this was not one of them. The interview raised issues that went to the heart of the case, eg on the alleged conditionality of the rebates. There should have been a proper record. It is irrelevant that the interview was not used to incriminate Intel. Intel was denied the ability to use the interview to help it interpret Dell’s RFI response, to explain the case file documents and Dell’s “impressions”. It cannot be excluded that the testimony could have been useful and helpful to Intel’s defence as per Solvay. Failure to record the testimony must lead to an annulment at least as far as the Dell rebates are concerned. The General Court did something completely different, it suggested that the interview could be reconstituted. A remarkable approach. Reached conclusion on an aide memoire which says h gave very detailed explanations of key issue but then entirely fails to record them. All we have is bullet points, bullet points, not even a summary and even then the General court felt able to conclude that what Mr D said was more incriminating than exculpatory. The Court referred to getting an impression of what Mr D must have said from a brief sentence in the aide memoire and from the list of topics. It tells noting about the discussion. The interview could not be reconstituted. The Court engaged in speculation not whether the rights of defence had been respected.
Nicholas Khan for the Commission argued that the ground of appeal was inoperative as Intel admitted that it did not contest the General Court’s findings of fact. The Commission met Mr D as a Dell executive, very different from Solvay where the missing strand of evidence was from third parties. The value of Mr. D’s evidence depends entirely on whether (a) what he said at that meeting was the same as that which he said to the FTC and (b) that such evidence was exculpatory. Even if (a) was correct, (b) was not, it was not exculpatory
Jean-Francois Bellis for ACT argued that in relation to the procedural issues re handling of Mr D interview, every time that Intel argued that the interview could have been exculpatory the General Court repeatedly dismissed the claim on the basis that the evidence would not shed light on and was irrelevant to the determination whether the rebates were conditional on exclusivity or quasi exclusivity, Category II (see 545, 642, 652, 660 and 663). This was wrong, the Intel rebates should have been assessed on the basis of all the circumstances of the case and potential exculpatory evidence. Therefore the Dell rebates should have been assessed on the basis of all the facts and circumstances and one cannot exclude potential exculpatory evidence simply because there was a finding of conditionality of exclusivity.
Questions from Juge Rapporteur da Cruz Vilaca
Regarding interview with Mr D, what was the legal basis of the Commission acting in this way in terms of interviews and gathering information? What was the legal basis of the interview with Mr. D?
Nicholas Khan: the Commission has a general power to gather information and not every circumstance necessarily will be an Art 19 interview, there is no requirement that the Commission can only gather information under Article 18 and 19. In any investigation there are often eg telcons to ask if there was any relevant information with a view to formulating an RFI. If the Commission wants information it can issue an RFI and a response to an RFI carries weight due to the sanctions that can be imposed for incorrect or misleading information. As regards Mr D he was not convoked, it was agreed that he would come. No formal request was made under Article 19 and even if an Article 19 interview had been carried out it would not have had same weight as an Article 18 RFI due to the sanctions that can be imposed, which is not the case re Article 19 interviews.
Juge Rapporteur da Cruz Vilaca refers to para 617 of the General Court judgment which reads as follows, relevant words underlined:
“That interpretation of Article 19 of Regulation No 1/2003, read in conjunction with Article 3 of Regulation No 773/2004, does not mean that the Commission may decide arbitrarily, during an interview, which information it is to record. It is apparent from Article 3(1) of Regulation No 773/2004 that, where the Commission carries out an interview in accordance with Article 19 of Regulation No 1/2003, it must inform the person interviewed, at the beginning of the interview, of the legal basis and the purpose of the interview, and of its intention to make a record of it. It follows that the Commission must decide, at the beginning of each interview, whether it wishes to carry out a formal interview. If the Commission decides, with the consent of the person interviewed, to carry out such an interview, it cannot opt to omit certain aspects from the record. In that case, it is required, rather, to record the interview in full, without prejudice to the fact that the first sentence of Article 3(3) of Regulation No 773/2004 leaves the Commission free to decide on the type of record. However, in the present case, the Commission stated, without being contradicted on that point by the applicant, that the purpose of the meeting was not aimed at collecting evidence in the form of countersigned minutes or statements under Article 19 of Regulation No 1/2003, but simply to examine whether there were sufficient indications for antitrust concerns with regard to Intel’s business practices vis-à-vis Dell and to explore appropriate further investigative measures related to Dell. The meeting between the Commission’s services and Mr D1 thus did not constitute formal questioning for the purposes of Article 19 of Regulation No 1/2003”.
How for the purposes of applying Article 19 does the Commission distinguish between receiving exculpatory evidence and evidence regarding an indication for antitrust concerns..
Nicholas Khan: Mr D came to the Commission as an officer of Dell. The Commission met with D1 and both before and after the Commission obtained information before and after with RFIs. The meeting was a preparatory one to explore further avenues for investigation from Dell.
Juge Rapporteur da Cruz Vilaca interrupts and asks during an informal interview like this one, for 5 hours, when the interviewee provides information capable of constituting exculpatory evidence what does the Commission do, start recording or what.
Nicholas Khan: if the person is able to provide evidence, whether exculpatory or inculpatory the n if the person is there as the personification of the undertaking the proper course of action would be to seek that information from the undertaking through an Article 18 RFI. An Article 19 interview would normally inly be conducted if the information could not be obtained through the normal written means of communication with the undertaking, namely an Article 18 RFI
Juge Rapporteur da Cruz Vilaca asks what is the difference between “sufficient indications for antitrust concerns” and “collecting evidence” in para 617 of the General Court judgment?
Nicholas Khan: “Sufficient indications for antitrust concerns” is a wider concept relating to the general state of the Commission’s state of knowledge regarding Intel’s practices at that point, “evidence” is something more definite.
Juge Rapporteur da Cruz Vilaca asks Intel whether the failure to record an interview can be cured?
Daniel Beard QC: difficult to understand how one can cure the absence of a recording or proper note. At the time the interview is conducted then an account could have been noted shortly afterwards in sufficient detail, the question would be whether that was sufficient to be capable of conducting an evidential assessment. However, in this case the note made after the interview is an aide memoire which does not even begin to set out the material which was the content of Mr D’s statement. The first section talks about the Q&A focussing on the deposition and then bullet points relating to key issues regarding the infringement, that is all we have. There could be no ex post cure in this case and so the General court’s approach in trying to reconstitute by plucking little inferences and comments which are then misinterpreted does not work here.
Advocate General Wahl: Regarding the interview with Mr D, are you saying that you are at liberty, when something that is informal is turning out to be interesting you will either transfer to something more formal or you would stop it and have a formal application. The reason for this was that if it had been a formal interview it would have been recorded. It seems to me that particular aspect of your argument is circular. Is that your position, you decide whether it is a formal interview or not.
Nicholas Khan: In the first instance it can only be for the Commission to decide whether it is going to conduct an Article 19 interview with a natural or legal person. The usual way to obtain information from a legal person is by an Article 18 RFI. If unusually an individual is in possession of evidence that cannot be obtained otherwise than by speaking directly to that person then, yes, it would be appropriate to conduct an Article 19 interview. But not every contact the Commission has with third parties etc during an investigation needs to be recorded as an Article 19 interview.
Part the Fourth: Fines
Question from Judge Regan
The Commission decision is based on unexamined assumption that exclusivity rebates per se are contrary to Article 102 in effect irrespective of the impact on the EEA market and so that the arrangements at issue are a by object infringement. Given that it is the normal procedure of the Commission would be an economic approach to competition policy which is based on the effects on the market I would ask whether we can feel comfortable with the imposition of a fine of the magnitude that we have here on the basis of what is fundamentally an assumption without an economic analysis of all the relevant considerations which would normally be applied when undertaking an assessment of the effects on the market in question.
Nicholas Khan sought to justify the fine, emphasising that the conduct in issue here, the facts of which are no longer contested, demonstrates clearly abusive conduct, which of course justified the fine.