On Thursday Judge James L. Robart of the United States District Court for the Western District of Washington published his FRAND rate-setting decision in the Microsoft v. Motorola contract case (the court's findings of fact and conclusions of law, my blog post). It was easy to see that Microsoft won and Google (Motorola Mobility's parent company) lost. My second post on the decision lists 16 media reports that arrived at that conclusion, too. If someone initially demands $4 billion a year, still seeks hundreds of millions later in the game, and is awarded less than $2 million (i.e., less than a 20th of a percent of the initial demand), he obviously didn't win -- and he's also highly likely to lose the second trial over the breach-of-contract claim. In a round-up post I summarized Google's multiple major patent-related defeats during this month of April.
I've now taken a closer look at the rate-setting decision, which is the culmination of a truly impressive effort. The first thing I did was to create a table of contents because that's helpful when reading a 207-page document (that's the length of a book). I've shared my unofficial table of contents on Scribd because many of the professionals interested in FRAND patent issues read this blog and some of them may find the table of contents useful in their own analysis of Judge Robart's historic ruling.
This blog post has the following sections (click on any of the subheads to go directly to a particular section):
After an introductory part on the case (first eight pages), the parties (page 9) and standard-setting in general (pages 10-24), Judge Robart outlines his "economic guideposts for assessing [F]RAND terms" (pages 25-40). In that section he explains his decision to strike the FRAND balance (compensating fairly, but not overcompensating, SEP holders) within the Georgia-Pacific framework, the commonly-used set of criteria for reasonable royalties in a patent damages context corresponding to simulation of bilateral negotiation, applying several of its 15 factors with FRAND-specific considerations in mind. The decision says "[t]he Georgia-Pacific factors must be adjusted to account for the purpose of the [F]RAND commitment" and, accordingly, "modified" (emphasis mine in both quotes). Paragraph 92 explains that "the hypothetical negotiation under a [F]RAND obligation must be different than the typical Georgia-Pacific analysis historically conducted by courts in a patent infringement" and gives two reasons for this:
"First, the owner of an SEP is under the obligation to license its patents on [F]RAND terms, whereas the owner of a patent uncommitted to [F]RAND has monopoly power over its patent and may choose to withhold licensing. Second, the hypothetical negotiation almost certainly will not take place in a vacuum: the implementer of a standard will understand that it must take a license from many SEP owners, not just one, before it will be in compliance with its licensing obligations and and able to fully implement the standard."
The second sentence relates to the concern commonly called "royalty-stacking". The two points I just quoted amount to this: yes, this is a negotiation, but it's a structurally unique one because both parties can't walk away the way they usually can. It almost reminds me of the theory of the "specificity of sports" in competition law (rival sports teams need each other, and even need each other to be reasonably competitive, to create a salable product, while a monopoly can work, at least for the monopolist, in all other industries). Here, Judge Robart basically established the "specificity of SEPs" principle with a view to FRAND licensing negotiations. There's no room for a Hobson's Choice approach here: the SEP owner can't just exclude someone from the market, and the implementer can't just work around the patent (which is an option in non-SEP negotiations). Both sides must be reasonable and work it out. And no deal can be analyzed in isolation but one must think about the wider implications (royalty-stacking). This second part is not nearly as unique as the first one.
Going beyond these two fundamental points, the decision outlines several "basic principles" in paragraphs 70-74:
"A [F]RAND royalty should be set at a level consistent with the SSOs' [standard-setting organizations'] goal of promoting widespread adoption of their standards."
"In the context of a dispute concerning whether or not a given royalty is [F]RAND, a proper methodology used to determine a [F]RAND royalty should therefore recognize and seek to mitigate the risk of patent hold-up that [F]RAND commitments are intended to avoid."
"Likewise, a proper methodogology for determining a [F]RAND royalty should address the risk of royalty stacking by considering the aggregate royalties that would apply if other SEP holders made royalty demands of the implementer."
"At the same time, a [F]RAND royalty should be set with the understanding that SSOs include technology intended to create valuable standards. [...] To induce the creation of valuable standards, the [F]RAND commitment must guarantee that holders of valuable intellectual property will receive reasonable royalties on that property."
"From an economic perspective, a [F]RAND commitment should be interpreted to limit a patent holder to a reasonable royalty on the economic value of its patented technology itself, apart from the value associated with incorporation of the patented technology into the standard."
At the Daubert stage (motions to exclude expert opinions) the parties already defended, and attacked each other's, proposed frameworks, with Google's Motorola arguing for Georgia-Pacific, which Microsoft described as "open-ended" and "unweighted", promoting (against Motorola's resistance) as a superior alternative an ex ante (pre-standardization) multilateral negotiation of the kind that is needed to form a patent pool. In his October 2012 Daubert ruling Judge Robart allowed both parties to present their theories at trial. At that point the question was not who made the better proposal -- it was just about whether a given approach would result in inherently unreliable testimony, providing not even a useful indication for what the FRAND royalty should be. Even Motorola's unwillingness to apply the Entire Market Value Rule (EMVR) -- seeking contrary to the EMVR a percentage of sales of multifunctional products -- was only met with skepticism, but not penalized, as even an EMVR-ignorant approach can theoretically result in an indication for a FRAND rate. The hurdle was low.
In the rate-setting decision Judge Robart "generally agree[d] with Motorola's approach" and criticized several "flaws" in Microsoft's approach, but found in paragraph 76 that "ex ante examination of the incremental contribution of the patented technology to the standard can be helpful in determining a [F]RAND rate in the context of a dispute over a [F]RAND rate". What happened then is that the court used Google's (Motorola's) proposed framework, the 15 Georgia-Pacific factors, with the "adjustments" and parameters described above, but it would still be a mistake to believe that Google prevailed on methodology. The fact of the matter is that Judge Robart elected to thread the needle: he incorporated Microsoft's approach into Motorola's by using pool rates as a (key) indicator of FRAND rates that the parties to a hypothetical bilateral negotation would consider.
The result -- only $560,000 per year above Microsoft's proposal, but roughly 4 billion below Motorola's initial demand -- indicates that Motorola didn't get what it wanted, and that's not a coincidence. The court adopted Georgia-Pacific, but not Google's denomination of it. Google wanted to capture hold-up value. Microsoft's proposed framework wouldn't have allowed this to happen -- but Judge Robart's threading of the needle has the very same effect in this case. I'll talk about the defensibility of this approach in case of an appeal by Google (which will definitely happen unless the parties settle before) in the next section and about the implications for other FRAND rate-setting cases in the third section.
Comparison of approaches chosen by Judge Robart and Judge Posner
Judge Robart's rate-setting decision is the first of its kind, and the last high-profile decision in U.S. federal court that addressed the issue of FRAND royalties (and FRAND licensing negotiations) was Judge Posner's Apple v. Motorola ruling last June. Unless I missed something, I didn't see any reference to Judge Posner in Judge Robart's rate-setting decision. Both judges are equally aware of the relevant problems -- hold-up at the threat of injunctive relief, royalty demands capturing standardization value above and beyond the intrinsic value of a patent -- and committed to preventing abuse. But their styles are very different, and I mean this non-judgmentally: many roads lead to Rome.
Both judges denied Motorola injunctive relief. A difference between the two cases is that Judge Posner had to rule on a damages claim for alleged past infringement, while Microsoft asked for a rate determination. For patent damages there's plenty of precedential decisions -- not so for rate-setting. In the case before Judge Posner, Motorola was seeking hundreds of millions of dollars for a single cellular standard-essential patent; the initial royalty demand that gave rise to Judge Robart's case amounted to $4 billion annually and was later reduced to hundreds of millions. In Chicago, Motorola got nothing because its claim was thrown out with no chance for a do-over; in Seattle, it got 50% more than Microsoft deemed FRAND, but almost nothing compared to its original demand.
Judge Posner held that the FRAND value of an SEP must be determined ex ante, pre-standardization. The value of a "patent qua patent". He ridiculed the Georgia-Pacific framework, and Cisco, HP, Walmart and others support his skepticism of Georgia-Pacific in an amicus brief. Google's primary attack vector against the FRAND damages part of Judge Posner's ruling is that traditionally any patent damages based on a reasonable-royalties theory must be calculated as per the date when infringement began, not an earlier date. In that context I already expressed my view that the point in time when negotiations took (or would have taken) place doesn't matter as long as the intrinsic value of a patent is determined without post-standardization monopoly power. Each SEP has to be valued as though alternative technical solutions were still available (which they cease to be once a particular solution is prescribed by an industry standard). Judge Robart achieves this objective within his "modified"/"adjusted" set of Georgia-Pacific factors.
I believe it will be harder for Google to challenge Judge Robart's ruling on appeal. There's no single legal issue that Google can present -- apart from the jurisdictional question of whether the court can set a FRAND rate at all -- to get the ruling reversed or vacated. Applying Georgia-Pacific is a safe, conservative choice. If Judge Robart had adopted Microsoft's proposal, this would have been beneficial with a view to other FRAND rate-setting cases, but it would have required an appeals court to decide only one legal issue -- the reasonable-royalties framework -- differently in order for the whole ruling to be vacated, creating legal uncertainty for Microsoft as to what the outcome on remand would be. Here, even if Google prevailed on one or a couple of issues relating to the framework (which is unlikely since common sense dictates the application of Georgia-Pacific proposed by Judge Robart) and won a remand, the impact would most likely be rather limited. Maybe Google would get another few hundred grand. It wouldn't suddenly get hundreds of millions more.
If Judge Robart had determined that multilateral patent pool negotiations are the only or always the best way to arrive at FRAND rates, an appeals court might have disagreed and might have found that if a company elects not to participate in a pool, it should still be free to seek higher royalties through bilateral negotiations. But his decision uses pool rates only "under certain circumstances" and only as "indicators" of a FRAND rate. When major, sophisticated SEP holders agree to make their patents on a given standard available at certain rates, how can it not be an indicator of a FRAND rate?
Google can, and absent a settlement presumably will, challenge Judge Robart's application of the criteria he outlined -- for example, what weight he gave to certain findings. There's always something in a 207-page decision on a complex matter that someone can disagree with (in some cases even a reasonable person), but as a fact finder Judge Robart gets considerably more deference than as a developer or modifier of legal theories. There are numerous findings of fact in Motorola's favor -- far more than the dollar amount at which the court arrived suggests. Still, the result is far from what Google hoped to get out of this.
Another consideration with a view to an appeal is that Judge Robart's ruling is very business-like: focused on resolving a commercial dispute that only a court can put to rest. He's not trying to legislate or campaign from the bench. Undoubtedly he's aware of the economic dimension of this Microsoft-Google case in its own right and the wider implications of the issues it presents. He wouldn't have written a 207-page document and developed fairly complex theories (for example, footnote 23, starting on page 171, outlines the calculation performed by the court) for a small and unimportant case. But nothing in the decision suggests an agenda or any ambition other than someone doing his best to arrive at the right numbers, based on an appropriate but conservative methodology.
Implications for other FRAND determination cases
Judge Robart's focused, case-specific and conservative (evolutionary rather than revolutionary) approach worked well for Microsoft and for the reasons I outlined in the previous section looks pretty appeals-proof. One could argue that going forward this is even better for Microsoft than the approach it proposed. But there will be many more FRAND determinations going forward, and cases in which the reasonableness of a SEP royalty demand must be evaluated with a view to equitable defenses or (particularly in ITC investigations) public interest factors. Technically, Judge Robart's opinion sets a precedent only in the Western District of Washington, but being the very well-thought-through precedent that it is, it's going to be influential in other jurisdictions, throughout and beyond the United States. (The Georgia-Pacific framework is U.S.-specific, but the underlying concepts are applicable everywhere.)
In this particular Microsoft v. Motorola case, there's a robust body of evidence supporting the royalty determinations. This may not be easy to build in certain other cases, and then the application of Georgia-Pacific, even with Judge Robart's adjustments, could make it considerably harder for an implementer to obtain a license at a reasonable rate. Here are some of the key factors that helped Microsoft but for which there may not always be equivalents in other cases:
Microsoft managed to show that Motorola's H.264 and IEEE 802.11 SEPs aren't overly valuable in general, and particularly not for Microsoft's implementations of these standards. In October I reported on a Motorola expert's reference to a pirated Katy Perry video in a desperate attempt to talk up the commercial relevance of interlaced video, a technique invented in the 1940s. With respect to WiFi SEPs, Judge Robart's opinion mentions, in various footnotes, numerous patents other companies declared essential to IEEE 802.11 relating to the same techniques (for example, data modulation or encryption, concepts which the decision notes Motorola "did not invent") -- and even Google's Motorola had to concede that Microsoft uses, at best, 11 of Motorola's 24 WiFi SEPs at issue in this case. There may be other FRAND cases in which it's harder to counter a SEP holder's claims that its patents are far more valuable than the average patent declared essential to the same standard.
Judge Robart put a lot of effort into an analysis of Motorola's SEPs at issue in this case. He analyzed them and the fields of technology they relate to. He stopped short of full-fledged infringement (essentiality) and validity rulings but came pretty close to such findings in some cases. Other courts adjudicating other FRAND cases may be less willing to go into so much detail. Other license-seeking implementers than Microsoft may not be able to provide the court with the same quality and quantity of briefing and testimony in this regard. And there may be cases in which the number of SEPs at issue is far greater (for example, chip maker Marvell claims to own hundreds of WiFi SEPs; or think of a valuation of Qualcomm's cellular SEP portfolio), making it much harder or impossible for a court to provide such a granular analysis.
There are patent pools for both of the standards in question: MPEG LA's pool for H.264, and a Via Licensing pool (less successful than MPEG LA, but still useful as an "indicator") for IEEE 802.11. As I explained further above, the pool rates were merely used as an "indicator", but a fairly important one. In the MPEG LA case, what makes the pool rate even more useful as an indicator is, besides the fact that Motorola was considering joining and approved a couple of announcements of pool rates, that Google agreed to a grant-back clause. Judge Robart found that Microsoft is an intended third-party beneficiary of that commitment to reciprocity but didn't base his rate-setting decision on a strict application of the grant-back obligation. He did, however, note that "Google is a sophisticated, substantial technology firm", whose agreement to the grant-back clause "further corroborates that the MPEG LA H.264 pool arrangement is an appropriate benchmark for determining [F]RAND royalties in this case". It's only one of various reasons for which the court arrived at its conclusions, but it's certainly an interesting one -- and this indicator will be hard to replicate in other cases.
A unique and impactful circumstance relating to IEEE 802.11 is that Motorola had a consulting firm named InteCap perform a valuation analysis about ten years ago. Despite limited overlap between the WiFi SEPs it held then and the one it has now, and despite a need for the court to apply a more realistic share for Motorola of all WiFi SEPs, that valuation also provided some useful guidance that was fundamentally inconsistent with Motorola's royalty demand from Microsoft.
In other cases it may be necessary to assess in more detail the value of a standard relative to the other functionality of a product. Here, there were enough comparables that a particular percentage of, for example, the value of a video codec relative to a PC operating system didn't have to be determined.
That said, this FRAND determination provides enormously valuable guidance, and with case-specific and jurisdiction-related adjustments this approach can help solve many FRAND rate-setting problems going forward. This approach does not have a built-in guarantee that overcompensation of SEP holders will always be avoided. For example, I could imagine some problems if such a determination was left to a jury. The Georgia-Pacific factors, even with Judge Robart's FRAND-specific amendments, provide patentees with an opportunity to present even excessive demands to a jury. There was no jury trial in this rate-setting process, but juries may become involved with other processes of this kind.
There will be many more FRAND determinations soon. We're going to see innumerable citations to Judge Robart's rate-setting opinion. Rightly so.
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: