Tuesday, February 26, 2013

For disparate reasons, Apple and Qualcomm raise concerns about arbitration of FRAND terms

This is my third post in a row on the two dozen submissions the competition enforcers at the Federal Trade Commission received in connection with the proposed Google and Motorola Mobility decision and consent order. I previously reported on Apple and Microsoft's criticism of Google's conduct and Qualcomm's thinly-veiled threat to duke it out in court with the FTC should its own pursuit of injuntions based on standard-essential patents (SEPs) come under antitrust scrutiny. What I don't want to spend time on is a bunch of filings that just reiterate the usual pro-injunction talking points previously filed with the ITC (for example, BlackBerry's submission consist almost entirely of copies of two letters to the ITC). I'm primarily interested in submissions that make new arguments or raise new issues.

Under the envisioned FTC-Google consent decree one of the things an implementer of a FRAND standard can do to fend off injunction requests is to trigger binding arbitration of the royalty rate and other relevant terms of a SEP license agreement. Even though the FTC held Google (Motorola Mobility) to have acted anticompetitively, the proposed consent decree imposes requirements on the conduct of innocent implementers unless they opt out of the potential benefits of the antitrust deal. They must do something, and initiating arbitration is one option, the other being a FRAND determination action in federal court.

Many people associate with the word "arbitration" something far more peaceful and helpful than it actually is. Arbitration isn't superior over litigation in all respects -- it has its specific advantages and drawbacks. Even if parties "agree" to arbitrate, it's merely a procedural agreement, irrespectively of which they could still be light years apart on the substance of a dispute. A few months ago, Apple and Google discussed the possibility of having their FRAND dispute resolved through arbitration, and this led some commentators to believe that a settlement was near. There are probably many thousands of lawsuits pending right now in federal and state courts in which no one is even contemplating arbitration, yet many of the parties to those disputes are considerably closer to a settlement than Apple and Google were last year when they explored this possibility. If judges (or regulators) encourage arbitration, they often do so simply because it disposes of cases they otherwise have to rule on. This doesn't make arbitration a panacea.

It's interesting to see that Apple and Qualcomm, while standing on opposite sides of the FRAND debate, are both somewhat uncomfortable and fear that arbitration could result in non-FRAND terms. They both own SEPs and need SEP licenses from others, but Apple is clearly focused on building products that add tremendous value on top of industry standards while Qualcomm derives a large part of its revenue from SEP licensing and depends on SEPs in its product business. Therefore, Apple is more concerned about arbitration resulting in royalties and other terms that are too onerous on licensees, while Qualcomm is worried about ways in which willing licensees might game the system and benefit exceedingly from the arbitration option the FTC envisions. It's no secret that I'm much closer to Apple's positions on FRAND than Qualcomm's, but both companies express concerns about arbitration that are worth thinking about.

The path to a FRAND determination is very important -- more than important enough to talk about it, and starting with the next paragraph I will summarize the concerns raised by Apple and Qualcomm. But in the greater scheme of things the parameters of arbitration are not the most important issue relating to the FTC-Google deal. The circumstances under which injunctive relief is available are the most important part of this. An injunction can cause more consumer harm than unfairly high or low royalties -- and an injunction, or sometimes just a credible threat of one, provides a SEP holder with the leverage to become "the dictator of the royalties", which is more than he'll ever be in any reasonably-conducted arbitration. In terms of priorities, I continue to believe that the "defensive use" exception, a recipe for mayhem, should be the first and foremost issue for the FTC to reconsider, especially in light of what's happening between Ericsson and Samsung on the one hand and between HTC and Nokia on the other hand. Microsoft's submission focuses on this concern, while other filings place the emphasis on corporate interests that are not closely related to consumer interests. That said, if arbitration results in unfair terms, that's also a bad thing for consumers, though not as directly as a sales or import ban.

Qualcomm would be much less worried about the FTC-Google settlement if the Commission did not suggest that the process outlined in the consent order should serve as a generally applicable model. Qualcomm's concern is that others are going to follow that process in their dealings with it. It concedes that implementers of standards can raise FRAND defenses, but it wants to be able to pursue an injunction from the start rather than having to go through a FRAND determination process first. It wants leverage. It wants to be the dictator of the royalties rather than have others set royalty rates and other contract terms. And it doesn't want defendants to delay the conclusion of a license agreement.

Section IV of Qualcomm's submission (starting on page 14) addresses "arbitration vs. litigation". Once again, Qualcomm opposes the "general applicability" of the FTC-Google settlement, and it expresses fears that "foreign regulators" might also look at this as a template or guideline. The regulatory agency Qualcomm is presumably thinking of in the first place is the European Commission's Directorate-General for Competition (DG COMP).

Qualcomm notes that "a FRAND licensing commitment to ETSI (or to other SSOs relevant to the cellular industry) does not contain any agreement to arbitrate future disputes, or any waiver of the fundamental and statutory right of recourse to courts of law". In Qualcomm's opinion, "an arbitration requirement should not be made a routine component of [antitrust] remedies". And these are Qualcomm's specific concerns about arbitration:

  1. Qualcomm says that arbitration "tends to result in 'split the baby' outcomes in an effort to find the middle ground, rather than make sharper decisions clearly favoring the position of one side over the other, even if that may be the right result". I agree. I believe Apple is also concerned that, for example, an unjustifiable 2.25% royalty demand by Google countered with a FRAND rate that is only a fraction of that amount could lead an arbitration tribunal to arrive at a royalty in the 1% range. A middle ground it would be, but not a FRAND rate. Qualcomm would probably go into arbitration with a demand that is far more realistic, relative to the strength of its portfolio, than what Google's Motorola has asked for -- and Qualcomm would not want to see a reasonable demand cut in half by the arbitrators. Qualcomm warns that "if arbitration looms over a negotiation, then both parties have a perverse incentive to exaggerate their demands, to increase the gap, in the hopes that by asking for a high-ball (or low-ball) result, the final 'middle ground' ruling from the arbitrator will be at an attractive point". In principle, I agree with Qualcomm, except that I believe that in most cases this will benefit SEP holders. Qualcomm is a special case because it really owns a huge SEP portfolio. It simply can't exaggerate the way that companies with far weaker SEP portfolios, such as Google or Samsung, have done in some ongoing litigations. If Qualcomm wanted to exaggerate by the same factors as Samsung and Motorola, it would have to ask for double-digit, if not triple-digit percentages. For example, a 20% demand by Qualcomm would be much more easily identified as unrealistic than Samsung's 2.4% or Motorola's 2.25% positions, even though a 20% royalty for Qualcomm would be less of a relative exaggeration than Motorola's 2.25%.

    While Qualcomm is right that arbitration has a proclivity for a "middle ground", this is also a problem, even if not to the same extent on average, in litigation. I recently heard Judge Voss of the Mannheim Regional Court mention that in his experience court-appointed damages experts in Germany almost always arrive at figures in the 2% to 5% range -- which is also way too high in most cases given the huge number of patents that need to be licensed to build a wireless device.

  2. Qualcomm furthermore dislikes the notion of cherrypicking of FRAND terms by implementers. The proposed FTC consent decree is designed to limit the scope of arbitration or litigation. Therefore, willing licensees can request arbitration or court determination with respect to only those terms the parties can't agree on. Qualcomm recognizes that this is "a well meant attempt to narrow issues in dispute", but believes it's better to assess the reasonableness of a license agreement as a whole than to create an incentive for implementers to play the following kind of game:

    "Picture, for example, a licensing negotiation in which the prospective licensee has indicated that it would be willing to bear unusual risk in the form of a large, one-time, up-front payment if by doing so it could enjoy a lower running royalty rate. Responding in good faith, the SEP-holder makes a good faith offer including the requested low running royalty rate and a large, one-time, up-front fee. Under the process laid out in the [proposed FTC-Google decision and consent order], upon receipt of a licensing offer from the SEP-holder, the licensee could seek arbitration of the up-front fee alone, thereby locking in the running royalty rate, as well as any or all of the other terms of the agreement--the very terms through which compromises are often reached in a private negotiation. The result is that arbitration would become a tool that infringers could use to seek better terms, without any risk of having to compromise on other terms."

    I consider the above a valid point, though I disagree with what Qualcomm subsequently goes on to say: Qualcomm raises these concerns to argue that the best framework is designed to "get to yes" as quickly as possible, and Qualcomm wants this to happen with the threat of an injunction hanging above an implementer's head. And once again I think that Qualcomm's concern cuts both ways. Pre-arbitration (or pre-litigation) posturing by SEP holders is at least as much of a possibility as pre-arbitration posturing by willing licensees, even though licensees have the unilateral option to initiate arbitration.

  3. Qualcomm's third point in this context is one I don't consider valid in the slightest. Qualcomm basically says that implementers of standards could get the best possible result for their purposes from arbitration and still improve the deal terms by subsequently asserting non-SEPs (which are typically not subject to such rules) against the SEP holder (unless the SEP holder is a non-practicing entity, of course), and then get terms for a "whole-portfolio license" (relating to the parties' SEPs as well as non-SEPs) that result in a reduction of the royalties paid for the SEPs involved.

    While the outcome that Qualcomm describes can of course correspond to what happens in certain cases, I don't see why this would be relevant from an antitrust point of view as long as there is no abuse of non-SEPs. The objective here, in the FTC-Google and wider SEP context, is to ensure that SEPs aren't leveraged anticompetitively to force companies to give away their non-SEPs or license them at rates that are too low. That's because SEPs always raise antitrust issues and come with a FRAND licensing obligation, while non-SEPs don't. Non-SEPs can be worked around, while SEPs cannot. If a SEP holder believes that a cross-licensing counterpart is demanding too much for its non-SEPs (such as a massive reduction of SEP royalties), the solution is to walk out on the deal and work around the non-SEPs involved. Qualcomm implies that SEP holders are generally entitled to whole-portfolio cross-licenses. But they are not. Patents are a property right, but SEPs come with particular obligations, for good reasons.

    From an innovation policy point of view it's absolutely correct and desirable to distinguish between encumbered patents (SEPs and a few others) and unencumbered ones. Anybody can get leverage from SEPs by simply sitting at the standard-setting table and doing a deal with the other participants. In this context I'd like to point to a very interesting empirical study conducted by three European researchers (Rudi Bekkers, René Bongard and Alessandro Nuvolari) on "the determinants of essential patent claims in compatibility standards". Their key finding is that "the involvement in the standardization process is a stronger determinant than the technical value ('merit') of the patent". I've watched many SEP assertions and I've been consistently underwhelmed by the technological merits of the SEPs-in-suit.

  4. I agree with Qualcomm on the next item:

    "Fourth, it is worth noting that arbitration is not an inherently superior dispute-resolution mechanism compared to litigation. Experience teaches that it is not always faster, and not always cheaper. Often, the primary benefit that drives parties to choose arbitration is confidentiality--a benefit that does not appear to be pertinent to the Commission's concerns. The {proposed consent decree] structure, of course, does not mandate arbitration; it merely requires Google to offer binding arbitration. But where arbitration is likely to be the more efficient route and both parties are acting in good faith, SEP owners and implementers will always be free to elect arbitration, so the efficiencies of arbitration, if any, are always within reach."

  5. Qualcomm's fifth and final point is that the proposed deal structure, if it becomes a generally applicable model, results in a shift to a "rate-regulation regime". While Qualcomm is right that the parties understand their own businesses better than arbitrators (or judges and court-appointed experts, for that matter), the problem in a growing number of FRAND negotiations is not lack of knowledge and insight but simply that certain SEP holders make prohibitive demands. Arbitration or court determination of FRAND terms are preferable over a might-makes-right regime in which the threat of lethal SEP-based injunctions determines the outcome.

Apple's arbitration-related concerns are perfectly complementary to those raised by Qualcomm:

  1. Apple's primary concern is that "Google may try to frustrate the arbitration process, cause it to fail in whole or in part, and then seek injunctions upon the claim that it complied with its duties under the Decision and Order". Apple's submission provides specific examples of how Google could derail arbitration in order to later claim that it was willing to arbitrate, but Apple (or any other party) was not a willing licensee because it rejected some of Google's preconditions for arbitration:

    "Arbitration could fail at the outset because the parties are unable to agree on the terms or scope of arbitration. For example, Google might seek to condition its Offer to Arbitrate in unacceptable ways, e.g., through 'creative' interpretations of 'field of use' in Section I.D.4.a or by limiting the scope (and not just individual terms) of the Relevant Licensing Agreement in Section III. Alternatively, in line with the discussion above, Google may decline to participate in an arbitration where it has the usual burden of proof on the question of infringement or another issue."

    Apple's proposed solution is to "include a process to resolve an impasse and require the arbitration to go forward", and to ensure that Google can't benefit from undermining the arbitration process: it should never be allowed to seek an injunction even if arbitration fails.

    This is Apple's priority in connection with arbitration. It additionally makes some other constructive suggestions, which I'll discuss in the remainder of this post.

  2. Apple proposes a "representative set" approach. I previously reported that Apple highlights Samsung and Google's dismal litigation track record with their multijurisdictional SEP assertions. Apple stresses that a SEP holder must prove actual infringement of a valid patent just like any other patent holder. In order to avoid that an arbitration tribunal or court sets a portfolio rate based on an unjustified presumption of essentiality, Apple proposes that a "representative set" of patents from the portfolio in question be chosen and "tested for actual essentiality/infringement, validity, performance benefits of the technology for the accused product, and other issues (e.g., patent exhaustion)". Given that it's not feasible for courts or arbitration tribunals to analyze dozens or hundreds of SEPs in one proceeding, I think Apple's suggestion is a very pragmatic approach. Most declared-essential patents aren't both valid and truly essential, and even truly essential ones aren't infringed if a feature of a standard is optional and not actually implemented. All of these issues must be assessed in order to determine the fair market value of a SEP portfolio.

  3. Apple wants licensees to be entitled "to meaningful discovery into the terms of the patent holder's other licenses, among other evidence critical to determining FRAND license terms in both arbitrations and declaratoryjudgment actions under the proposed framework". In the U.S. discovery goes quite far in federal court. Apple presumably wants to ensure that it will be equally productive in arbitration and in foreign jurisdictions.

  4. By demanding a "transparent, reasoned decision by arbitrators" (though "transparent" doesn't rule out redactions of confidential business information), Apple proposes an approach that "will create a body of decisions", which is a good thing per se, but which runs counter to a key reason for which parties frequently elect arbitration: privacy. In a passage I quoted further above, Qualcomm says that "[o]ften, the primary benefit that drives parties to choose arbitration is confidentiality".

  5. In light of the commercial significance of SEP licensing terms it's absolutely key in Apple's view to have access to a "substantive appellate review". Apple wants to ensure that arbitration rulings will be "subject to judicial review for errors of fact or law". From a rule-of-law point of view this makes sense, but it also makes arbitration less binding. Apple's submission doesn't specify a particular standard of review. What it says sounds like de novo.

  6. Apple proposes, for the method of appointment of arbitrators, "that each party picks one arbitrator and the two appointed arbitrators pick a third arbitrator". This is the most common approach, and Apple wants it to be applied consistently.

  7. Apple wants there to be competition between arbitration organizations and encourages the FTC to add more organizations to its list of qualified arbitration organizations. But there's always a problem with having courts, or arbitration organizations in this case, compete with each other: it may lead them to favor the interests of those who pick the court. In federal court, it's usually the patent holder who sues and picks a venue. In the FTC-Google consent decree context, implementers of standards would potentially benefit from more competition. The other suggestions that Apple makes are also in the public interest, but this one here, while not completely unreasonable, appears too self-serving.

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents and Google+.

Share with other professionals via LinkedIn: