Thursday, May 9, 2013

Google's Plan B for Motorola's standard-essential patents: B as in 'baseball arbitration'

The abject failure of Google's strategy to leverage Motorola's standard-essential patents (SEPs) in order to address Android's proven patent infringement issues is clearer than ever after an already-famous rate-setting decision by a U.S. court last month and a preliminary antitrust ruling by the European Commission earlier this week. But after spending $12.5 billion on the Motorola deal and attaching a $5.5 billion book value to Motorola's IP, Google isn't going to give up... yet. Having learned the hard way that the outright pursuit of injunctive relief against allegedly-unwilling licensees doesn't work, it's already focusing on the second-biggest threat: excessive royalties (and damages for past "infringement").

In fact, this has already been Google's focus for a while, but it hasn't been widely noticed yet. Google didn't announce a sweeping withdrawal of injunction requests as Samsung did in Europe. It quietly began implementing Plan B while Plan A (injunctions) was falling apart:

  1. Last summer it accepted Apple's offer to take a license on court-determined terms to Motorola's cellular SEPs in Germany. It lost its opportunity to gain the maximum leverage it could have: an injunction. Instead, it's now going for a maximum amount of money.

  2. A few months later, after a rate-setting trial in Wisconsin had been canceled, the parties exchanged letters (which Apple attached to a court filing) about the possibility of resolving their FRAND terms dispute by other means. Google expressed a clear preference for having a worldwide FRAND rate set by a German court (presumably the Mannheim Regional Court), where a process for Motorola's German patents was already underway as a result of the aforementioned acceptance of Apple's offer to take a license. Google would ideally like to take advantage of the legal uncertainty caused by a scarcity of patent damages case law in Germany and an exceedingly patentee-friendly system and mindset, where some judges may actually think a royalty of a percent or two of someone's sales is acceptable even if U.S. courts would arrive (all other things being equal) at a fraction of those rates. However, the European Commission's preliminary assessment is that the license agreement underlying that rate-setting case came into being under anticompetitive circumstances, calling into question whether there even is an enforceable agreement in place.

  3. Alternatively to a process in Germany, Google "suggest[ed] the [arbitration] panel [in the United States] employ a 'baseball arbitration' procedure to resolve any areas in which the parties are unable to reach agreement" (emphasis mine). Apple and Google couldn't agree on key parameters of an arbitration proceeding, and there's no indication that they've agreed by now.

"Baseball arbitration" sounds like a whole lot of fun, but it isn't. In this context, it's just another form of extortion. I already criticized this proposal in November 2012, but I believe it's necessary to comment on it again, and in more detail, because Google and its friends apparently continue to push for a regime under which an implementer of a standard would have to commit to binding and generally non-appealable "baseball arbitration" -- otherwise the SEP holder should be free to pursue injunctions.

I don't believe any policymaker, regulator, standard-setting organization or businessperson who thinks the implications of "baseball arbitration" in this context through will ultimately support it. At first sight, however, it may be appealing to all those who want the problem of SEP abuse to be addressed but don't have the resources in place (and/or don't have a desire for political reasons) to set FRAND rates. In my opinion, access to courts of law -- as opposed to a de facto obligation to arbitrate -- must never be compromised. Judge Robart has shown that courts can set FRAND rates. Such processes are underway in other jurisdictions as well (for example, in the UK). If parties voluntarily -- without the threat of an injunction -- agree to submit to binding arbitration and prefer it over litigation, that's great. But they should not be obligated to do so.

Lemley-Shapiro paper promotes Google's preferred arbitration method and reflects unconcealed anti-Apple, anti-Microsoft bias

Two well-known California professors, Stanford's Mark Lemley (patent law) and Berkeley's Carl Shapiro (competition economics), have published a pro-Google policy paper on FRAND rate-setting styled as an academic working paper, entitled "A Simple Approach to Setting Reasonable Royalties for Standard-Essential Patents". The approach is "simple" indeed: it promotes SEP extortion. What's not so simple -- but not extremely hard either -- is to explain why "baseball arbitration", which at first sight means the same set of rules for both parties, has asymmetrical implications and favors SEP abusers.

These two professors were among the first ones to identify the problem of SEP abuse. Their write-ups are widely cited in this context. Their latest paper restates, without added value, the issues they identified years ago. So far, so good. But the practical "solution" they propose now would cause huge problems for honest implementers of FRAND standards -- while being the biggest opportunity, short of outright injunctions, for Google and a very few other companies seeking to leverage SEPs against non-SEPs in order to get away with infringement. BlackBerry is a company that openly lays out the same agenda, and I called its proposal a "wacko idea". The Lemley-Shapiro paper is far more sophisticated than BlackBerry's incredible bluntness -- but it promotes an idea that is almost as bad as BlackBerry's proposal to let SEP users create a scenario of mutually assured destruction in order to prevent non-SEP enforcement.

Before I go into details on the way in which "baseball arbitration" favors abuse in this context, I wish to clarify that I criticize the proposal for its undesirable implications and other shortcomings, and there are a couple of passages in the paper that expose the authors' bias, but the one thing I don't mean to suggest is that this was written on Google's behalf. It's pro-Google, yes. Written on Google's behalf? No. In a footnote on page 1, Professor Lemley discloses that he "has represented Google in matters related to the subject of this Article, but Google has provided no financial support for this project and the views offered here have neither been reviewed nor approved by Google". I have no problem with this because it's normal that the leading experts in this field do some work for major industry players at some point. They can still have independent views.

Proactive disclosure is the right thing. I, too, disclosed certain consulting relationships voluntarily, long before any judge ever asked questions about them.

Even if neither of the authors had ever received a cent from Google, there are two passages that are, in problematic ways, overtly partial. I want to show them before talking about "baseball arbitration" because they make it easier to understand the deficiencies of the proposal:

  1. Bottom of page 17: "If an implementer has a patent essential to a different standard that ends up in the same product, and that standard is not promulgated by an SSO that requires a FRAND commitment following our best practices, the patentee who has made a commitment will be at a disadvantage."

    This sentence points to footnote 65: "[W]e sympathize with the plight of Google-Motorola, which was investigated for antitrust violations for suing to enforce its patents only in response to being sued itself by Apple and Microsoft. [...] But the fact that Motorola made a FRAND commitment and Apple and Microsoft didn't means the situations are not the same."

    There you have it: they admit that they "sympathize" with Google, a company that the FTC in its proposed consent order and the European Commission in a preliminary antitrust ruling have found to have abused SEPs. They're Google apologists. They blame the mess Google created in no small part on companies that rely completely on non-SEPs in those Android-related disputes. They suggest that there's something wrong about Apple and Microsoft's enforcement of non-SEPs, though independent courts, literally around the globe, have identified numerous patent infringements by Android-based devices and almost the entire Android device industry, except for Google's Motorola, has recognized the need to take a license from Microsoft -- and these companies include major patent holders that collectively own more patents than Microsoft does, so they know a thing or two about patent licensing (and litigation).

    The authors wrote that footnote without even researching the chronology of events. Motorola sued Apple, then Apple countersued. It claimed that it had a "reasonable apprehension" of Apple suing it, but the statement that Motorola enforced its SEPs "only in response to being sued itself by Apple" is, as it stands, wrong. Wrong. To be clear, Professor Lemley is the U.S. patent law expert, in general. But he has a broad perspective on patent law that goes way beyond the smartphone patent disputes, which he doesn't always follow in detail (compare these tweets -- 1, 2 -- to the truth on the Apple-Samsung damages ruling).

    In the smartphone-related cases and investigations I watch, no court and no regulator has excused Google's (or Samsung's) conduct with the fact that other companies wanted to put an end to Android's proven, wide-ranging infringement. SEP and non-SEP issues were always kept separate.

    The authors wrongly suggest that the difference between Motorola's SEP assertions and Apple and Microsoft's assertions is just that one party made a FRAND commitment and the other didn't. The difference is not just the promise -- that one is key if a SEP case is based on contract law. From an antitrust perspective, the question is whether a patent can be worked around or not. The same patent, prior to inclusion in a standard, can be worked around, but not thereafter. And that's why there's a compulsory licensing regime for such patents that would even affect someone who never made a FRAND promise, as far as competition law is concerned.

    A patent has significant commercial value (i.e., not just "nuisance value" due to litigation costs) only if someone is willing to pay for a license because he will be at a disadvantage in the marketplace if an injunction is enforced. For a non-SEP that's hard to achieve: the broader it is, the more likely it is to be invalid; the narrower, the less likely valuable. For a SEP it's very easy: the fact that the specifications of a standard prescribe, for example, only one countdown method (that's what Motorola enforced against Apple in Germany), out of innumerable possible solutions, ensures that this one must be used regardless of whether it's innovative. That's why different sets of rules for SEPs and non-SEPs are appropriate from a competition and innovation policy perspective. It's a bad idea to conflate SEP and non-SEP issues.

    It's also misleading to suggest that a FRAND promise is an act of generosity, when it is, in connection with SEPs, merely dictated by necessity. If the leading players of a given industry define a standard, they use their collective market power (as a cartel) to exclude competing technologies. Regulators would already have to intervene at the standard-setting stage unless FRAND licensing is promised in order to ensure that the procompetitive aspects of standard-setting outweigh the anticompetitive ones.

  2. It's a very common approach by apologists to relativize a major wrongdoing (when it can't be denied anymore) by comparison with minor offenses (for example, suggesting that "concentration camps" also existed elsewhere before, during and after the years of the Nazi regime, which is of course a completely different problem than relativizing FRAND abuse, but an example everyone understands).

    The Lemley-Shapiro paper relativized Google's conduct in the passage I just quoted, suggesting that the "plight" (should we now all have tears in our eyes?) Motorola experienced due to Apple and Microsoft's non-SEP assertions justifies (or at least excuses in part) some SEP abuse. And they again relativize on page 14, saying that "[a]n implementer who agrees to participate only if it gets a result it likes is no different than a patentee who agrees to license on reasonable terms only if it gets to decide what is reasonable", adding that "[n]either party is acting in good faith" and pointing to the Apple v. Motorola FRAND case in Wisconsin dismissed (notably, without prejudice) after Apple said it would take a license on court-determined terms only if the rate doesn't exceed $1 per unit.

    No, it's not the same. There's a world of difference. The most Apple-critical position a reasonable, unbiased person could take on Apple's behavior in this case is that it was aggressive and, maybe, arrogant, and even if one describes this as a lack of good faith, it's not comparable at all to the anticompetitive pursuit of injunctive relief over SEPs. The maximum harm Apple's conduct could have caused is delay and a waste of litigation expenses. The harm Motorola was trying to cause was to wreak havoc to Apple's business in the two largest markets in the world unless it would do a deal on Google's/Motorola's terms -- with irreversible anticompetitive effects either way.

    The paper doesn't acknowledge that Apple was also willing to accept a higher rate provided that it would also apply to a license Motorola would have to take to Apple's (mostly acquired) SEPs. In the FRAND dispute with Microsoft in the Western District of Washington, it was actually Google's/Motorola's position that terms would have to be determined for a cross-license. This is, however, just detail. The real issue is that someone who relativizes SEP abuse is trying to apologize someone's misconduct and not making a constructive contribution to a debate on how to solve the problem.

"Baseball arbitration" doesn't effectively prevent strategic abuse -- and its efficiency gains are a myth

In the previous section I showed that the authors intend to give Google maximum leverage (under the circumstances) against non-SEP assertions necessitated by Android's wide-ranging infringement. They present their proposal as an efficient and balanced way to arrive at FRAND rates -- but closer scrutiny shows that the efficiency gains are, at best, highly doubtful, and there are scenarios, especially the ones that matter to Google (and to Samsung in the Apple dispute), in which the practical effects of the proposed approach would be detrimental to competition and innovation.

The authors describe baseball arbitration and its alleged benefits as follows (quoting from page 8 of the paper):

"We suggest that [standard-setting organizations] specify that disputes over what is FRAND be resolved through binding arbitration, 'baseball' style. In baseball arbitration, the parties produce evidence and argument before the arbitrator, and then they each propose a royalty number. The arbitrator must pick one of the two numbers offered and cannot come up with her own number. Using baseball arbitration logically drives the parties toward making reasonable proposals, because the party that asks for too much (or offers too little) risks losing the case altogether. FRAND disputes are well-suited to baseball-style arbitration, because the only thing at issue is which of two numbers in fact represents the more reasonable royalty. [...]

Baseball arbitration has a number of other advantages. The arbitrator does not need to decide whether any given patent is valid and infringed. Nor does she need to decide whether a particular patent is essential except in unusual circumstances. Both of those things may be contested, and the evidence on each question will likely influence the reasonableness of the competing royalty proposals. But unlike a court that might have to rule on any number of subsidiary fact reasons, the only thing the arbitrator needs to do is pick the better of two proposed royalty rates."

Even in baseball, it doesn't work this way for non-free agents

It's basically an insult for actual baseball arbitration to use this term for a proposal that is fundamentally less fair than salary arbitration in baseball. It works quite differently in baseball if the player whose salary must be determined is a non-free agent. In that case, he will have to stay with the same team, and the team is not allowed to make a cutthroat offer in order to protect this player against a scenario in which he would have to accept a very low annual salary (theoretically as low as 1 cent) only because it happened to be closer to the salary level the arbitrators deemed correct than the player's demand. I'll quote from

"No club may submit a salary figure that is less than 80% of the player's previous year's salary (or 70% of his salary two years' previous). [...} The maximum-cut rule does not apply for free agents in arbitration. [...] The last case where someone walked out of an arbitration hearing with less money than they made the previous year was Randy Milligan in 1994. Occasionally players [...] settle at a salary that is less than they made the year before, but even that is extremely rare."

Implementers of standards are also non-free agents. They must implement patents that are essential to the standard because the standard prescribes it. Just like a non-free agent can't just leave and sign a contract with a different team, implementers of standards can't just license other patents that solve the same technical problem if the standard dictates only one particular solution.

One could also argue that a patent holder shouldn't have to grant a license for a total of $1 for the sale of tens of millions of devices implementing hundreds of valid and infringed SEPs held by that company. From a competition point of view (considering that one party really needs a license), it would however be more acceptable to disincentivize overcharging than to allow situations in which massive overcompensation occurs.

Therefore, a correct application of baseball arbitration to FRAND standards would have to at least protect the implementer -- and possibly also the patentee -- against excessive or ridiculously "lowball" offers. But there's no comparable like the previous year's salary. In order to establish a range for the parties' offers, an arbitration panel would have to conduct almost all of the analysis needed to determine a FRAND range the way Judge Robart did in Seattle.

That's the dilemma. Baseball arbitration in connection with FRAND rates can be fair (like in baseball itself), but then it becomes inefficient. Or it can be seemingly efficient but definitely unfair in situations in which an abuser leverages SEPs against non-SEPs, as I'll explain now.

False assumptions about parties' behavior: someone like Google wouldn't behave the way the paper predicts

Let me repeat the passage that displays the primary flaw of the Lemley-Shapiro proposal:

"Using baseball arbitration logically drives the parties toward making reasonable proposals, because the party that asks for too much (or offers too little) risks losing the case altogether."

This sounds right but isn't always right. These are the underlying assumptions:

  1. Either actor has all of the relevant knowledge.

  2. Both actors behave rationally.

  3. Both actors look at each such arbitration from a broader, longer-term perspective, with the patentee being interested in a maximum financial return on investment.

  4. Maximizing the number of "baseball arbitrations" a patentee wins yields the best average royalty income across all arbitrated deals.

A good policy in connection with disputes has to work even if none of these assumptions is right. The first two assumptions are riskier than they appear. In reality, there will be incompetent actors and there will be emotional ones. That's not SEP-specific. The SEP-specific part is in items 3 and 4.

Let's assume there's a company that expects to go into (for example) roughly 100 arbitration proceedings over the next ten years over its SEPs. Let's furthermore assume, for the sake of the argument in this narrow context, that item 4 (the more arbitrations you win, the more money you make) applies (though it doesn't necessarily). Then it would indeed get the best average outcome by demanding (more or less) what it believes are FRAND rates in the opinion of the arbitrators.

But if a company like Google is looking for a strategy to get away with infringement of non-SEPs, then it isn't interested in ROI in a straightforward, simple sense. Google didn't pay billions for Motorola's patent portfolio because it thought it would generate licensing income exceeding the price. It wanted to reach a state of "mutually assured destruction" to stop non-SEP assertions against Android. So let's assume Google is this company expecting to go into 100 FRAND royalty arbitrations over its SEPs during the next ten years. In more than 90 of those arbitrations it may have no strategic interest related to non-SEPS and indeed act according to Lemley-Shapiro's underlying assumptions. But in its disputes with Apple and Microsoft it wouldn't. Here's why it wouldn't:

The only way that Motorola's SEPs (which aren't incredibly strong, as litigation has shown) can give Google meaningful leverage against a rival like Apple or Microsoft is, short of an outright injunction, if Google can collect a rate far in excess of a FRAND rate. Microsoft offered Motorola ten times the MPEG LA pool rate in Germany, and a $300 million (or higher) bond, but Motorola declined. If Judge Robart had awarded Google ten times as much as he has, it would get $18 million a year from Microsoft instead of $1.8 million. Even $18 million a year wouldn't enable Google to convince Microsoft that it should stop collecting royalties on Android devices. But with hundreds of millions of dollars per year, Google would have some leverage, and if the court had agreed with Motorola's original $4 billion demand, that probably would have been a game changer.

Therefore, it wouldn't have made sense for Google to maximize the likelihood of the arbitrator picking its number. It would then have been more interesting for Google to ask for the moon, knowing that it's a long shot, only to negotiate a settlement with Microsoft in a situation in which the latter can't rule out entirely that it may end up having to pay billions of dollars a year. That's not as good as an injunction, and not even as good as being likely to prevail. But it is a strategy for extortion, and under the circumstances it would still put Google in a better position than if it asked for "only" a FRAND rate, or even five or ten times a FRAND rate.

It's not just Google's (and Samsung's) strategic situation in which this approach is an attractive option for a patentee. I said before that even the fourth assumption -- the more frequently you win, the more money you make on average -- doesn't necessarily work. Let's now assume a non-strategic patent holder: a non-practicing entity (NPE). If that NPE feels that demanding an excessive, supra-FRAND rate will work in a small minority of its arbitrations, but it may occasionally hit the jackpot, then its calculus could mean that the expectancy value is actually greater if the demand is way in excess of FRAND. Think of the venture capital business where many investors assume at the outset that most of the companies they fund will go out of business, but if one of your investments becomes a Facebook, you'll do better than if you're too risk-averse and settle for mediocrity.

And that NPE will be managed by individuals. They may have short-term objectives in mind, such as the next quarterly report. Let's assume the NPE already has 50 license deals in place, and the average income doesn't please investors. Now the NPE is going into an arbitration with a substantial implementer. Management may decide to try a Hail Mary because it's under pressure to deliver results, and feels that it will be fired unless it scores a huge supra-FRAND win -- or a settlement with someone who will pay a premium only to eliminate the (low but existing) risk of a devastatingly high royalty rate.

There are additional scenarios in which the Lemley-Shapiro approach would fail to deter patent holders from taking extremely unreasonable positions. Strategic abuse of SEPs is what has given rise to most of the recent SEP-related antitrust investigations. If standard-setting organizations (and, above them, policymakers and regulators) now look for ways to prevent the reoccurrence of the phenomenon of $4 billion royalty demands, the analysis has to begin with one question: "What would Google (or Samsung) do to take advantage of this proposal?" And not with expressing sympathy for Google, and promoting its preferred arbitration method...

Another argument against the claimed efficiency: "just pick one of the two" is easier said than done

The Lemley-Shapiro paper claims (not only in a passage I quoted further above but also in several other contexts) that it's easier for an arbitration panel to decide only which one of the two offers is more reasonable than to set a rate. Just like with the "non-free agent" problem I described further above, it helps again to learn from what arbitrators actually do in baseball. Based on what I've been able to research (I must admit I know little about baseball, though I do know a thing or two about the professional sports business as I did some work on competition issues for the world's most famous club in the number one spectator's sport), arbitrators don't just pick one of the two proposed salaries but make their own determination, and then they choose the one of the two proposals that's closer to their finding.

This makes sense: how can you seriously determine that a point P1 is closer to a point F (as in "fair") than an alternative point P2 is to F without firstly deciding on what or where F is?

I encourage anyone who thinks this may be possible for FRAND to use Judge Robart's rate-setting opinion as an example. Read it from page 1 to page 207 and think about which part of the analysis would really be simplified by the fact that only one of two offers must be picked. Judge Robart looked at multiple comparables. He had to ask himself whether a certain deal Motorola did with another licensee was a reliable indicator of a FRAND rate or not. He had to analyze the extent to which Motorola's SEPs are actually used by Microsoft's products, and the value they add to the standard in general and Microsoft's products in particular. Only at the very end does he then state the result, based on all of the analysis performed. Could he have saved time in any area if he had been a "baseball arbitrator"? Not really. While Professors Lemley and Shapiro argue that there are only two outcomes to choose from, there are (as Judge Robart's landmark ruling shows) multiple considerations for FRAND. If the parties only disagreed on one such consideration, then the Lemley-Shapiro approach would work, but that's never going to happen. For example, if the only consideration on which the parties to an arbitration disagree was whether the royalty rate set forth in one particular license agreement signed by a third party with the patentee should also be acceptable to the implementer in the case before the arbitrator, then it may be possible to make a determination like this:

"Patentee argues that the same $10-per-unit royalty paid by Third Party is also a FRAND rate for Implementer. Implementer argues that its products to be licensed sell at a lower average price than Third Party's licensed products and proposes a $2-per-unit royalty. Parties agree on all other parameters except the way the different products use the standard: Patentee says that Implementer's products have an even greater benefit from the standard than Third Party's products. Even if Implementer was right, this royalty rate would at best be adjusted proportionally to the per-unit price, in which case it would be $7 according to Implementer's own representations, a FRAND rate to which Patentee's $10 demand is closer than Implementer's $2 offer, mooting the questions of (i) how the different products use the standard and (ii) whether Implementer (70% of Third Party's average price) or Patentee (80%) accurately calculated the average price of Implementer's products."

Again, this won't happen in practice. The parties to such arbitration proceedings will disagree on a multiplicity of factors and parameters, and then one has to ultimately go through the "Judge Robart exercise" of analyzing and weighting all of the relevant factors.

Participation in "baseball arbitration" to be coerced at threat of injunctive relief

The Lemley-Shapiro paper doesn't outline a dispute resolution mechanism that two parties may voluntarily consent to. Again, their intention clearly is to give leverage to someone like Google, so they want to let SEP holders use the threat of injunctive relief to force an implementer to submit to their unfair "baseball arbitration" process:

"[W]hile a court can and will compel a patentee to arbitration, there is no legal way to compel an implementer who wants to resolve the question in court to go to arbitration if they refuse. But what SSOs can and should do in that situation is release the patentee from the FRAND commitment not to seek an injunction. An implementer who wants to challenge the validity or infringement of the patents is free to do so in court, but if they are not willing to pay a FRAND royalty, they shouldn't benefit from the patentee's forbearance in seeking damages and injunctive relief." (quoted from page 21)

They want to preclude implementers from going to court. Access to courts is part of the rule of law. In the civilized world, arbitration is -- and should remain -- voluntary.

The Lemley-Shapiro approach is to let SEP holders leverage unfair, anticompetitive means (threat of injunctive relief with potentially devastating impact on someone's business) in order to coerce participation in a process designed to lead to unfair, anticompetitive results in disputes involving both SEPs and non-SEPs. Very disappointing.

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: