The redacted version of Mr. Justice Marcus Smith’s decision in Optis v. Apple has been available now for a few weeks, and I have had some time to read and digest it. The bottom line, as I noted on June 20, is that the decision awards Optis $56.43 million ($5.13 million per year), plus interest (5% per annum, compounded) for Apple's past and future global access to Optis' portfolio of FRAND-committed SEPs; rejects the SSPPU rule and Apple's assertion that Optis abused a dominant position; but also rejects Optis' argument that Apple was an unwilling licensee, and finds Optis' proposed comparables unhelpful for calculating the royalties due. Many of the facts, particularly those relating to third-party licenses, remain for now redacted. Even so, the decision is as long as a standard book, and for present purposes I’m only going to discuss some of the highlights—beginning with a brief discussion of how the judge calculated the amount, before concluding with a summary of what I view as his two most interesting legal/economic conclusions relating to FRAND methodology—and will leave it to others to write up a more comprehensive account.
1. As others have noted, the relatively small amount awarded is clearly a win for Apple. In reaching the lump-sum figure of $5.13 million per year, the judge considered the expert testimony and the comparables put forward by each party. In regard to the experts, however, the judge writes that he “derived no benefit from the work of either valuation expert and . . . reject[s] their evidence as unhelpful in resolving the FRAND question” (para. 311). More to the point, he writes that each was too dependent on the guidance the parties provided them:
312. . . . [T]he nature of the comparables in this case has taken both experts far outside the zone of their proper expertise. Both experts, I am confident, would have approached a “typical” valuation exercise . . . with competence and skill. In that case, they would have applied their judgement in extrapolating from the objective data in the comparables, and derived a value for the Portfolio. Perhaps they would have disagreed, but only in relation to the exercise of judgement as regards a common and commonly understood set of comparators.
313. In this case, the Optis Comparables and the Apple Comparables were categorically different (as I have described). That is something that could have been handled by the experts, in that they could (although they did not) have sought to understand the differences underlying these different categories, and what these differences taught. This, then, was a surmountable difficulty, but it is one that neither party’s expert made any effort to surmount.
314. What was, in my judgement, an insurmountable difficulty was how what I have termed the three subjectivities (at [301]ff) were to be resolved. Inevitably, they would have to have been resolved by the exercise of judgement and had the valuation experts attempted to do so, no doubt they could have been accused of seeking to give the court the “final answer” to the FRAND Question, something experts are rightly told to avoid. Instead of doing so, the experts “unpacked” in accordance with the instructions of their clients . . . .
315. The fact is that both experts were tasked to articulate the ramifications of the cases that their respective clients were contending for. They provided no independent judgement – or, at least, insufficient independent judgement for me to be able to rely upon it – as to the merits or demerits of those cases. Neither expert approached the problem from a stance they had independently evolved. Their direction of travel was set down by the parties who instructed them, and all they did was (conscientiously and carefully) follow that direction of travel and “unpack” in accordance with their principal’s respective stances.
316. This sounds critical of both experts. It is not intended to be. . . .
These implicit criticisms (of the parties, if not the experts) notwithstanding, the judge concludes that neither party engaged in sanctionable holdup or holdout, and that Optis was not guilty of an abuse of dominant position (see paras. 374-89). He is nevertheless rather blunt in his criticism of Optis’ negotiating behavior. See, e.g., paras. 354-57 (characterizing Optis’ approach to negotiation as “inept,” lacking “any kind of rational underpinning for the prices they put forward,” “badly organized, verging on the incoherent,” and its case against Apple “overblown and wrong”).
2. Having found the expert testimony not helpful, the judge sets out to determine the FRAND royalty himself, giving consideration to the parties’ proposed comparables. His critique of the Optis comparables is withering. See, e.g., paras. 398(iii) (stating that “the Optis Comparables actually unpack in a manner that creates a price for the overall Stack that is indefensible,” with “ad valorem rates contended for by Optis that are overstated to a remarkable degree”; noting that “most of Optis’ counterparties in the Optis Comparables were small companies”; and stating that “it is difficult to avoid the conclusion that Optis was not dealing with these small counterparties for the revenue streams they would bring in, but because the licences they agreed would produce comparables that would assist Optis in this litigation”), 471 (“worse than useless”). By contrast, he finds that Apple’s comparables, though not without their own problems, are “more reliable” (para. 418).
3. Stating that his task is to price the value to Apple of the Optis portfolio, without “any assessment of the value of the individual patents” (para. 456), the judge considers what the ad valorem royalty would be if one nevertheless used Optis’ comparables to derive one. The ultimate number he comes up with is redacted, but he notes that “No Implementer would stay in business paying Optis’ rates” for 100% of the royalty stack (para. 467(iv)). As a result, Optis’ comparables “cannot be used to draw any wider conclusions as regards FRAND rates” (para. 470(ii)(c)). Moving on the Apple comparables, the judge assumes an overall royalty stack size of 22,000 (see paras. 459, 460(iv)(b)) and concludes that, in the present case, a lump-sum approach is preferable (see paras. 471, 480). The judge then derives for each relevant Apple license the lump-sum royalty that would be implied for Apple to acquire 100% of the stack. (For example, if I understand correctly, if Licensor A received $1 million for a 2-year license and owned 2% of the overall net stack, this would imply a 100% stack value/year of $25,000,000. The actual identities and figures are redacted.) He comes up with an adjusted average (para. 484, redacted), and then on the assumption that the amount should be divided 50/50 to account for past releases and forward licensing, divides this amount in two (para. 485), to give us an “annual rate for the entire Stack–whether as a backward looking release or a forward looking license” amounting to some redacted number (para. 486). He thereafter determines that Optis’ share of the stack is 0.38%, and that this results in an annual figure of $5.13 million (para. 494); so apparently the redacted estimated annual rate for the entire stack divided by two is $1.35 billion, if I did the math right. Interest will be at a 5% rate, compounded (para. 502). (I will note here that I have long favored the use of compound interest in patent cases, to reduce the risk that defendants will be rendered better off for having infringed.)
4. Although the decision is largely favorable to Apple, earlier in the opinion the judge rejected certain aspects of Apple’s general “framework” for negotiating royalties, as set forth in paras. 201-04. Most importantly, in my view, he rejects Apple’s advocacy of the “smallest saleable patent-practicing unit” (SSPPU) concept. As described in para. 201, which quotes Apple’s witness Heather Mewes, under Apple’s approach “The FRAND Framework is based on the following two starting premises, which seek to determine a reasonable aggregate royalty burden for the relevant cellular standard or standards: i) The smallest saleable patent practising unit (SSPPU) for cellular SEPs is the baseband chipset, where the functioning of the cellular standards is substantially practiced or embodied; and ii) The proportion of the overall cellular declared SEP royalty profits that is allocated to the cellular SEP holders is determined to be no more than the profits of the baseband chipset.” More specifically, as described by the judge:
215. Apple’s argument goes like this. The baseband chipset contains the cellular functionality of the Standards. I have concluded – although I have not set out my reasoning at any length – that the baseband chipset does contain the relevant technology.
216. Apple contend that this should frame the debate as to value or price. So, if the baseband chipset is valued at US$25 (a figure Optis did not seriously dispute), and the cost of production is US$20 (again, a figure Optis was prepared to accept, for the sake of argument), the manufacturer’s profit is US$5 (as a matter of simple arithmetic). The product is, however, unlicensed. The manufacturers of baseband chipsets do not (typically) seek or obtain licences to the SEPs comprising the Stack. . . .
217. Apple contended that the baseband chipset manufacturer should pay for the licence out of the US$5 profit, and that this therefore constituted the absolute limit that ought to be paid by anyone. Although I do not understand this to be a necessary part of the argument, only a supporting prop, Apple also contended that were the baseband chipset manufacturer to obtain a licence to the Stack, that licence would exhaust the intellectual property rights of the patent owner, such that manufacturers incorporating the baseband chipset into their product (say, a handset) would not themselves need a licence at all. Exhaustion of rights is a complex area of law, even in this jurisdiction; and its effects vary from jurisdiction to jurisdiction. For the sake of argument, I am prepared to assume that Apple are right, and that licensing of the baseband chipset manufacturer on a worldwide basis would exhaust the rights of SEP Owners in respect of any “downstream” products.
In the court’s view, however, this argument is “indefensible” (para. 213), “wrong in its essence” (para. 218), because “There is no reason why the baseband chipset manufacturer would have to fund the licence fee to the stack out of the US$5 profit that is being hypothesised. Indeed, it is quite absurd to presuppose that the manufacturer of a baseband chipset would forego any part of their profit unless absolutely compelled to do so. It is much more likely that baseband chipset manufacturers would increase the price of their product to reflect the added value to purchasers of that product of having a licence to the SEPs comprising the stack. Absent extremely clear market evidence, the assumption that the baseband chipset manufacturer would absorb the costs of the licence and not pass them on is almost certainly both unsafe and wrong. Certainly, it cannot be assumed. . .” (para. 218). For this reason, “the US$5 limit to the purchase price for the licence is arbitrary” (para. 219). (This is consistent with a critique that Norman Siebrasse and I have made concerning Judge Holderman’s use of the profit margin on chips as a royalty cap in the Innovatio case. See Norman V. Siebrasse & Thomas F. Cotter, The Value of the Standard, 101 Minn. L. Rev. 1159, 1226 n.198 (2017); Thomas F. Cotter, Patent Damages Heuristics, 25 Tex. Intell. Prop. L.J. 159, 207 (2018).)
The judge also addresses the argument, proposed by Apple’s expert witness Carl Shapiro, that the SSPPU is necessary to avoid an anchoring effect tied to the value of Apple’s products. See para. 221 (quoting Professor Shapiro’s argument that “When a patented technology is substantially embodied in one of many components of a complex product, using that end product as the royalty base may result in a royalty award above the FRAND rate. For example, when royalties for cellular SEPs are calculated in reference to handsets rather than baseband chipsets, the larger price and profit figures associated with the handsets are likely to ‘anchor’ royalties at a higher per-unit level even if the cellular SEPs are substantially embodied in only one of many handset components. This effect will be larger for high-end handsets because their higher prices will provide a higher anchor”). The court is not persuaded, however:
222. As an abstract proposition, I find Professor Shapiro’s identification of this issue important and helpful. However, it is necessary to make the following additional points:
i) The phenomenon is best demonstrated to exist where there is an objective answer. . . The problem here is that the FRAND Question does not have a known answer against which any biases that I might have due to anchoring can be measured. . . .
ii) The problem, in this case, is really that the price or value to be attributed to either the Portfolio or to the stack is not known. It is to be derived in this Judgment. . . .
b) . . . measuring royalties to an SEP Owner by reference to their ownership of an ascertainable percentage of the Stack that that portfolio comprises works very well, subject to issues of “quality” which I have resolved, but only where the value or price attributable to the Stack itself can be derived. Professor Shapiro can, of course, say that the total price of the Stack (derived, say, by adding the headline rates for the portfolios described above, and a bit more to take account of those not publishing headline rates) is “too high”. But that is as subjective as saying that it is “just right” or “not high enough”. Apple’s own method of valuing the stack – the notional profit of the manufacturer of the baseband chipset – is just as much susceptible of anchoring error as anything else. The fact that the price of the chipset is as low as US$25 skews our thinking, so that (unless we really force ourselves) we cannot imagine a chipset manufacturer charging US$50 or US$500 or US$5,000 for a chipset with a licence to all SEPs comprising the Stack. . . .
This too is consistent with something I’ve noted previously. See Patent Damages Heuristics, at pp. 202-03 (citing Nicolas Petit, The Smallest Salable Patent-Practicing Unit (“SSPPU”) Experiment, General Purpose Technologies and the Coase Theorem, SSRN at 3 (Feb. 18, 2016), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2734245) (noting that the SSPPU could increase or decrease accuracy, depending on whether you start from the premise that the patentee should receive some portion of “the increased value to the user resulting from complementarity”).
5. Another aspect of Apple’s proposed framework was the proposition that royalty should reflect the value of the technology, but not the value of the standard itself. Here is Mr. Justice Smith’s analysis of this issue, again in view of Professor Shapiro’s testimony:
224. Professor Shapiro’s opinion was that an SEP’s value (in terms of what an Implementer should pay to obtain a licence for it) should be calculated by reference to the usefulness of the SEP – its importance, as I have defined that term . . . – to the Implementer, disregarding the value of the Standard itself to the Implementer. . . .
226. Concentrating, for the moment, on the benefit of a wider market to Implementers, Professor Shapiro asserted that allowing SEP Holders to price taking into account the Standards of which the SEPs formed a part was both economically inefficient and wrong in principle. To quote from his report: . . .
“30 In the standard-setting context, the FRAND rate can be conceived of as the rate that would result from technology competition to be included in the standard. The rate resulting from such ex ante technology competition would be free of patent holdup. . . .
…
32 These economic principles establish that the FRAND royalty is the rate that would be negotiated between the SEP holder and potential licensees prior to the establishment of the standard. At that point in time, the industry is not yet locked into any particular technology. . . . Economics thus establishes an upper bound on the FRAND rate: the ex ante incremental value of the technology covered by the SEP compared with the next-best alternative that was available prior to standardisation.” . . .
228. I reject the existence of Professor Shapiro’s “upper bound” to the FRAND rate for the following reasons:
i) Practical impossibility. The upper limit to FRAND involves a patent-by-patent consideration of the importance of each SEP in issue. Whatever the technical merits of the approach, given the number of patent families in issue, an assessment of their merits (in addition to questions of validity and essentiality, that would no doubt also arise) renders the approach a non-starter, as Professor Shapiro accepted. . . .
ii) Unnecessary. The fact is that the decision in Unwired Planet (SC) resolves the problem of Hold Up. As I have described, the SEP Owner does not have a right to an injunction where (i) the owner has declared the patent to the Standard and (ii) where the implementer in question has expressed a willingness to abide by a court-determined FRAND licence. It follows that – whatever may have been the case in the past – the notion that rates will go up because of Hold Up or that rates will be computed to include a Hold Up “weighting” are not axiomatically right. Indeed, the whole point of the process is to answer the FRAND Question in a way that accords value neither to Hold Up nor to Hold Out.
iii) Wrong in principle. One of the consequences of Professor Shapiro’s approach is that the benefits of the Standard accrue away from the SEP Owner. The Apple Framework sought to exclude the value of the Standard itself from payments to SEP Owners. . . . Although a little equivocal, Ms Mewes was agreeing with the line taken by Professor Shapiro, namely that the SEP Owner should be rewarded for the technology inherent in the SEP and not for any contribution the SEP made to the Standard itself. I find this a proposition that is difficult to defend, even if it were practicable. The fact is that the SEPs are part of what makes the Standard. I agree with Ms Mewes that there are other contributions, but simply considering the position of the SEP Owner, the SEP Owner: (i) declares the SEP to the Standard, but (ii) as a direct consequence enters into the FRAND undertaking not to seek an injunction. This is a contribution going beyond the value of the SEP itself. It may be that the SEP Owner is rewarded through the higher volume sales achieved in having the SEP part of a standard, rather than a higher royalty payment for each sale. But that is a matter for negotiation of rates (whether volume justifies a lower per unit or ad valorem rate or not) between SEP Owner and Implementer and I can see no justification for excluding altogether from the SEP Owner’s reward the contribution made by SEPs to the Standard itself.
This is consistent with the conclusion of Mr. Justice Birss (as he then was) in Unwired Planet, and with the analysis presented in the Siebrasse & Cotter article The Value of the Standard cited above, though it is at odds with statements made by the Federal Circuit in Ericsson v. D-Link.
6. There is more that one could write about the details of the decision, and more that I am sure others will write (particularly if some of the redactions are lifted, or if the litigation continues on appeal). The above nevertheless should give you a flavor for how the judge reached his decision, and provide food for thought on some of the still highly disputed issues relating to SEP valuation.
