I’ve criticized the European Commission’s antitrust attack against Intel here and the resulting $1.44 billion fine. Now the EU is drawing fire for allegedly burying testimony, or at least failing to record it in a satisfactory manner, from Dell that it chose Intel’s chips not because of the coercive force of any of Intel’s rebates but because it preferred the performance of those chips over AMD’s product offerings. Part of the problem here is that the EU’s 542 page decision remains confidential and so it is impossible to tell what kind of impact this testimony would have on the issue of liability — and indeed the WSJ story sensibly suggests that there is little reason to believe that disclosure of this evidence would have changed the outcome. The WSJ story concludes by taking a swing at EU procedure:
The entire EU antitrust regime—in which the Commission can assess guilt without any semblance of a trial or the involvement of an independent judge or jury—is an affront to any reasonable notion of due process. Deliberately withholding exculpatory evidence is, and should be, grounds for retrial, and not merely a sympathetic report from a toothless Ombudsman.
It now falls to the EU’s Court of First Instance, to which Intel has appealed, to offer Intel redress. The Court may not have jurisdiction to give the Commission’s discredited antitrust theories the thrashing they deserve, but it can, at a minimum, stand up for the rights of the accused when they are trampled by Europe’s antitrust cops.
But to return to the liability issue for a moment, I think it is easy to understate the importance of this evidence from Dell in light of the 542 page Commission opinion by assuming it would have minimal weight. Indeed, that is likely true under EU law where Intel’s loyalty rebates with other OEMs would likely be sufficient to establish a violation. In that sense, the failing to disclose or record the exculpatory Dell testimony is likely to be “harmless error” only in the sense that it would not have impacted the liability ruling under Article 82 jurisprudence — but the error is obviously far from harmless in terms of calculating appropriate fines and from a procedural perspective.
But from an economic perspective, recall that the anticompetitive theory underlying the investigation is that Intel uses these loyalty rebates, conditioned on exclusivity, to foreclose its rival AMD from sufficient access to distribution to achieve minimum efficient scale. In other words, the rebates result in de facto exclusivity that deprives AMD of efficient scale, raises its costs, and allows Intel to maintain its monopoly power in the microprocessor market. A key part of that story is that AMD cannot compete on the merits for distribution because it is foreclosed. In the United States, a key analytical issue (in addition to whether any of the rebates resulted in below cost pricing, presumably under Brooke Group or some attribution test) is whether the distribution contracts foreclosed ENOUGH distribution to have an effect on competition. In this analysis, how the Dell business figures into the competitive analysis could be a very big deal. While Dell is just a single buyer, the OEM market is relatively concentrated with a handful of large buyers. Evidence that AMD was not foreclosed from Dell sales — or could have competed for those sales and earned them on the merits with an equivalent product — would be incredibly important (and tends toward a finding of no liability) under Section 2 analysis and conventional exclusive dealing/ exclusionary contract analysis generally.