Patent privateering - a comment on the FOSS Patents campaign

In my previous post, I questioned the stance taken by FOSS Patents blogger Florian Mueller on what he calls "patent privateering". It seems like a veritable FOSS Patents campaign has been launched against this phenomenon, with crowd-sourced name-and-shaming and all. 

I feel that this campaign is - at best - unhelpful, since its basis is rather lacking in analysis and it carries an unwarranted presumption of sinister causes on behalf of SEP holders. 
In short, FOSS Patents goes after an effect rather than a cause.

More questioning of the campaign has recently come from Richard Lloyd of IPR newspaper Intellectual Asset Management (IAM): "the notion behind the FOSS Patents initiative - that privateering is bad and should be stamped out – is just plain wrong".

However, in a more recent FOSS Patents post, Mr. Mueller does ask the relevant question "...why companies with such vast resources and enormous sophistication (in-house and externally, in legal and in technical respects) need help from little guys with a controversial business model to do license deals with the very same licensees with which they've already done deals before and do deals with all the time."

But the question was rhetorically put, i.e. still presuming sinister motives. It shouldn't be rhetorical though. It's an excellent question. At least if you really want to try to understand the problem. My answer, as I indicated last year and more recently last week, is that patent holdout is an important driving factor behind SEP sales. So to the extent we need a campaign, patent holdout should be the target. To cite IAM's Richard Lloyd again: "In short, it doesn’t look like the privateer who is being unreasonable and abusive, it is the companies who will not sit down and talk turkey. Maybe this is what we really want to be shining a light on. Anyone want to help IAM put together a list?"

In what looks like an attempt to isolate a particular group of "good guys" from "evil-doers", Mr. Mueller goes on to ask the question "...why a number of major right holders generally don't sell patents to PAEs. For example, I'm not aware of Qualcomm doing this... Or IBM. ... Or even Microsoft".

Again rhetorical, and again it needn't be. This one is particularly easy to explain. To begin with, in a cellular SEP context we can forget about IBM and Microsoft since neither is a significant SEP holder. So let's look at Qualcomm. It's the only significant SEP holder that sells chipsets and not end-user products. And not just any old chipsets, but the most popular ones on the market. So for Qualcomm, SEP-licensing is more straightforward than for other SEP-holders since there's already an established and keen business interest on the part of its potential licensees. Indeed, Qualcomm hardly ever initiates SEP infringement lawsuits, and yet it has the industry's highest SEP license royalty revenue. Clearly, Qualcomm has so far been subject to far less patent holdout than other significant SEP-holders, and therefore hasn't felt the need to engage in SEP selling for "patent privateering".

So, while perhaps contrary to Mr. Mueller's intention, his Qualcomm example is the perfect corollary to my own conclusion: SEP selling is largely a rational response to patent holdout. 

Patent privateering - cause and effect

In a recent FOSS Patents blog post, there was a discussion on "patent privateering" in relation to SEPs. The definition used in the post was "the act of large companies feeding trolls with patents in order to maximize their patent monetization income and/or drive up their competitors' total cost of defense". Whether this definition is appropriate or not could be debated, but the blog post is nevertheless interesting as one reaction to what seems to be a relatively new form of behaviour in the industry.

However, there's a key statement in the post: "Many privateering deals involve SEPs and are part of a scheme to circumvent FRAND licensing obligations", that is worth quite some scrutiny.

Regarding FRAND licensing obligations, my experience is that:

1) A growing number of SDOs and competition jurisprudence ensure that FRAND licensing obligations transfer with the SEPs to the new owners. Therefore, SEP-holders typically contractually bind the buyer to abide by FRAND terms for the purchased SEPs. Sellers may also impose royalty caps versus third parties on the buyer. 

2) The post inherently assumes that when an SEP holder sells off SEPs, the seller's FRAND rate is unchanged. It doesn't explore whether there may be legitimate reasons for maintaining the same rate or whether sellers reduce their FRAND rates by a "corresponding" amount after sale. Sellers can contractually agree with the buyer to do so, as well as permit buyers to disclose that information to potential licensees. Then the transfer won't affect the cumulative FRAND rate for OEMs.

But the most troubling aspect of the FOSS post is that it carries a presumption of sinister motives being present; "part of a scheme to circumvent...obligations". No attempt is made to explore what might be legitimate motives for selling SEPs.

My experience from real-world licensing is overwhelmingly that companies do have legitimate motives for their business decisions, so to simply and sweepingly assume the opposite strikes me as a rather extreme position to take.

In fact, as I've already mentioned in an earlier post, there is indeed a major legitimate cause of SEP selling. It's called patent hold-out.

Several of today's phone OEMs with large product sales are relatively new to the business and as regards to SEP licensing tend to behave differently from their equivalents in the past. Instead of trying to negotiate necessary SEP-portfolio licenses in good faith, there's a tendency to proclaim virtually any license offer as being "non-FRAND". Then simply wait for the patent holder to try to obtain FRAND value on its global SEP-portfolio through expensive and inefficient patent-by-patent, country-by-country litigation - a process that may never lead to FRAND value recoupment. This behaviour is known as "patent hold-out" and is becoming widely recognized by courts and agencies worldwide as being a major problem. For a discussion on the mechanism, reasons and effects of patent hold-out, I refer interested readers to my earlier post on the subject.

As a result of all this, several holders of significant global SEP-portfolios have in recent times increasingly become unable to obtain FRAND value from their portfolios in traditional ways, i.e. through bi-lateral patent licensing. These SEP-portfolio owners don't have the corporate culture or relevant experience to extensively enforce their patents through courts and so have turned to NPEs, who are much more experienced with patent litigation, in order to recoup value from their patents.

If we really want to mitigate "privateering" going forward, we should voice our support for appropriate changes to mitigate its actual causes. We should find ways to facilitate fair and efficient adjudication of FRAND license terms of global SEP-portfolios, and adopt clear rules that allow injunctive relief against those who refuse such an adjudication. And of course, regulators, courts and agencies should continue to acknowledge, highlight and enforce against the growing patent hold-out problem.

With such measures effectively implemented, I'm convinced that SEP transfers seen as "privateering" will decrease and SEP licensing efficiency will increase.

IPR & Make in India

Under new prime minister Modi, India has launched the ambitious "Make in India" campaign, with the clear aim to attract global companies to locate their manufacturing resources in India. The timing of the campaign seems excellent, coinciding with global companies looking around for alternative manufacturing hubs to China. It could be just what India needs.

Although India hasn't really kept up with China's level of technology investment for some time, the overall IPR legal framework in India is comprehensive and there's a strictly independent judiciary that Indians are rightly proud of. In recent years, a rapidly growing number of patent cases in various fields have been handled by Indian courts, continually further developing their specific competence on patents - arguably the most complex form of IPR - in the process. In fact, today India even serves to demonstrate that major international patent enforcement is no longer restricted to the US and Europe. India was recently selected by Ericsson for the filing of cellular SEP infringement lawsuits against some alleged infringers, including the Chinese OEM Xiaomi and the Indian OEM Intex.

In relation to the Make in India campaign, the Indian Cellular Association (ICA), a grouping including many Indian mobile phone OEMs, recently submitted a report to the Indian government, apparently pertaining to IPR issues related to the campaign. Although the report is currently not public, statements from it have been quoted in a recent article in the Indian newspaper Economic Times. Sadly, those statements are at best confusing, and at worst misleading.

According to the article, ICA in its report seems to highlight the above-mentioned case of Ericsson enforcing its cellular SEP portfolio against alleged infringement in India, and then makes generalized, negative and far-reaching conclusions about India's future as a manufacturing hub based on that. 

What seems to be a key message is:

"There can never be any manufacturing set up in India since the global giants have cartelized and self declared their patents and they would be charging these incredibly high rates only from the Indians and not from the international companies."

To begin with, this statement presumes a clear link between manufacturing and patents. But with respect to cellular SEP licensing, the licensing point is at the OEM level, which is downstream from the manufacturing point. This makes the manufacturing territory virtually irrelevant. As far as cellular SEP licensing is concerned, for the typical Indian OEM selling products largely in India, it doesn't matter if the products are manufactured in India or not. So actually, the supposedly "showcase example" of SEP licensing highlighted in the report has nothing to do with Make in India in the first place.
There may well be other technology areas, for example pharmaceuticals, where IPR may be more coupled to manufacturing, but highlighting cellular SEP licensing in a manufacturing context is simply not relevant.

The allegation in the same quote that "global giants" would for some reason discriminate against Indians in particular with "incredibly high rates" is truly odd to say the least. Assuming still that the reference is cellular SEP licensing, it can have no basis in reality nor in common sense. 
Firstly, it's highly questionable what rationale a licensor would have to even attempt such discrimination. 
Secondly, there are FRAND commitments in place to prevent unfairness, unreasonableness and discrimination. Licensors' existing license databases are regularly scrutinized by courts and agencies worldwide to ultimately enforce FRAND terms, including mitigating discrimination. 

Thirdly, Indian OEMs are so far generally unlicensed under cellular SEPs, i.e. they pay zero actual royalties to SEP holders. So if anyone would be discriminated against, it's the already licensed international companies operating in India. And to add to the irony, the potential discriminators are Indian OEMs themselves, including members of ICA (!). Indeed, the High Court of Delhi in its recent infringement judgement in Ericsson vs. Intex found that Intex - an ICA member - had "acted in bad faith" and became an "unwilling licensee as per its overall conduct". It further decided on interim payments by Intex in the region of 0.8-1.3%, obviously a far cry from the alleged "incredibly high rates".

With regards to "global giants have cartelized and self-declared their patents", I'm afraid it's a bit of a challenge making any sense of the statement at all. The only feasible conclusion I can draw from it is that ICA has not fully understood the simultaneously collaborative and competitive process of cellular standards creation and the FRAND licensing regime for technology sharing. It would be unfortunate if ICA failed to understand how these processes actually enable the standards to be deployed worldwide and allow billions of dollars of R&D and network investments to be made and the entire related eco-system - which ICA partly represents - to prosper.

Another key statement in the article reads:

"This in complete contrast to regimes such as China, where 0.019 per cent royalty can be charged and USA, where the Courts have directed 0.5-2 per cent royalties to be charged on the value of smallest saleable practicing unit which is royalty on chipset value and not on the phone value"

In fact, the quoted examples of allegedly lower rates are not from particularly relevant rulings relating to cellular SEP licensing. The quoted China ruling is most likely from Huawei vs InterDigital in the Guangdong High Court of China. An early use of the Chinese Anti-Monopoly-Law (AML), it is widely seen as a purely political ruling against so-called Non-Practicing Entities (NPEs). If ICA really wanted to use China as a yardstick, a far more important measure would be the conclusion of the recent AML investigation against Qualcomm - importantly being a non-NPE and major cellular SEP licensor - where China allowed Qualcomm to charge a 3.25% royalty on the phone value. 

Regarding the reference to the US and the "royalties to be charged on ...chipset value" argument, it is well-known to be nonsensical. In the US, chipset value has been squarely dismissed as a mandatory royalty base in e.g. CSIRO vs Cisco and by the US Federal Circuit in Ericsson vs. D-Link. Also in China it has been dismissed through the AML ruling on Qualcomm where phone value was maintained as royalty base. Based on e.g. such powerful international references, the High Court of Delhi also dismissed the argument in its recent judgement in the Ericsson vs. Intex case.

To the extent that ICA should be making recommendations concerning IPR, it should obviously strive to provide relevant, true and objective facts about IPR in the mobile industry, and their relevance - if any - to "Make in India". But in its report, ICA unfortunately does not do that. Instead it provides incorrect and irrelevant information about cellular SEP FRAND licensing in particular. I'm afraid I can't see how conclusions drawn from such information can be helpful to the government of India.

ICA is now, according to the same article, part of a Fast Track Task Force set up by the Indian government to rejuvenate mobile manufacturing in the country. To be clear, ICA may well be a competent advisor on many aspects of mobile phone manufacturing. But as far as cellular SEP licensing goes, as put recently by reputed IPR newspaper Intellectual Asset Management, "Perhaps FRAND matters are best left in the hands of the judges".

Apple and Ericsson - a first look

On January 12, 2015, Apple filed a complaint in the US District Court for the Northern District of California asking for a declaratory judgment that i) seven Ericsson US patents are not essential to 4G nor infringed by Apple or alternatively, if found essential and infringed, ii) the court sets reasonable royalties using a royalty base of "at most, the component that substantially embodies the alleged invention". It also asked the court to prevent Ericsson from seeking injunctive relief or exclusion orders based on the patents-in-suit.

Shortly after, Ericsson filed a complaint in the US District Court for the Eastern District of Texas asking for a declaratory judgement that i) Ericsson's global license offers have "complied with its FRAND commitment" and ii) Ericsson has "complied with its contractual obligations under its FRAND commitment".

There's been more filings in this dispute since, and of course there could be more to come. But in this post I'll specifically examine certain aspects of this first filing by Apple. There are some parts that I find interesting and related to what I have touched upon in previous posts. Also in relation to both of these first filings, I'll revisit the issue of SEP portfolio license valuation.

The first point in Apple's complaint concerns the "royalty base". Basically, a "royalty base" is the value of "something" that a royalty rate percentage is multiplied by, to arrive at a payable royalty for a licensed product. In its complaint, Apple argues that "the law requires" that the royalty base be selected as "(at most) the smallest salable unit", and explains that such a unit would correspond to "(at most) the baseband processor chip" inside its products. Apple further implies that the "current technological and legal environment" is one where this particular Apple opinion prevails, and that Ericsson has refused to "adapt" to it.

Let's first look at the law. The fundamental basis for US patent damages can be found in 35 USC 284, "...the court shall award the claimant damages adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer..." (emphasis added). Clearly, there are no limitations of the type advanced by Apple to be found there.

So how about the "smallest salable unit"? Well, the phrase is actually part of the Entire Market Value Rule (EMVR) concept in US patent damages law. But Apple has embraced it in a completely out-of-context manner. The concept, as clarified only recently by the US Federal Circuit in Ericsson vs. D-Link in 2013, is only an "evidentiary principle" to be used to specifically assist a US layman jury to arrive at a reasonable royalty when there is a risk of it being misled (biased) by a higher end-product value. Moreover, in the very same breath the Federal Circuit made it crystal clear that i) licenses are "generally negotiated without consideration of the EMVR" and ii) the "substantive legal rule" is that the “ultimate reasonable royalty”– e.g. the result of multiplying a royalty rate and a royalty base–”must be based on the value that the patented invention adds to the end product” (emphasis added).

And if that's not enough, also the market has rejected Apple's position. It's an indisputable fact that the end-product selling price has been used as a royalty base in cellular SEP licensing for more than 20 years. That is, since more than a decade before Apple even got into the mobile phone business. During all those years, several hundreds, if not thousands, of cellular SEP license agreements were signed using precisely that royalty base.

If those hundreds of licenses would have been so clearly wrong, how could the world's legal systems - including the US legal system - have let such practices flourish for decades? Not to mention the tremendous global growth in the mobile phone business during those 20 years, from virtually zero to 7 billion mobile subscriptions and 2.5 billion devices sold annually today? All that, based on something that's basically against the law? I'm afraid I find that quite hard to believe.

Clearly, Apple's opinions about what "the law requires" and the "current legal environment" on the subject of royalty base do seem quite odd to say the least.

Moving now to the question about injunctive relief. Here Apple asked the court for "a ruling that Ericsson cannot seek injunctive relief or exclusion orders against Apple" based on any of the patents-in-suit found to be infringed and essential.

As I've mentioned before, there's no legal support whatsoever for banning the seeking of injunctive relief for infringement of SEPs. In fact, the US Federal Circuit in Apple vs. Motorola in 2014 could not have been clearer on this point: "To the extent that the district court applied a per se rule that injunctions are unavailable for SEPs, it erred." Indeed, "an injunction may be justified where an infringer unilaterally refuses a FRAND royalty or unreasonably delays negotiations to the same effect." Clearly, an SEP holder does have the right to seek an injunction and it's up to the court to decide on a case by case basis whether an injunction is warranted, based on e.g. whether the infringer is deemed to have engaged in patent hold-out. An actual hold-out situation can be very damaging to the SEP holder and especially to other willing and existing licensees, as I discussed in a specific post on that subject. This is presumably why the US Federal Circuit believes that an injunction may be appropriate in such cases.

So, like the royalty base issue, Apple's request about injunctions also seems out of touch with US law.

The last aspect I'll examine here is that of SEP portfolio license valuation, looking at both Apple's and Ericsson's filings.

Apple seems to view the entire issue from a patent-by-patent aspect, and asked the court for declaratory judgements on seven particular Ericsson patents in terms of essentiality and infringement, and assuming those are fulfilled, value.

Ericsson on the other hand appears to treat the issue as a SEP portfolio licensing effort, and asked the court to determine whether its terms for its entire worldwide alleged SEP portfolio as a whole - i.e. not just seven US patents - are to be considered FRAND.

So essentially we have two approaches pitted against each other; i) patent-by-patent license value adjudication and ii) global patent portfolio license value adjudication.

The patent-by-patent approach requested by Apple may be suitable when the matter at hand is a license to a limited explicit set of patents granted in the US. But that's really not the case here at all. The majority of Apple's sales are outside the US, and it sells tremendous volumes on every continent. And an SEP-holder like Ericsson surely has granted SEPs on every continent too. In addition, major standards-contributors tend to obtain new SEPs over time, adding to their portfolios. For at least these reasons, the license scope of the Apple-Ericsson negotiation preceding these court filings must have concerned a global SEP portfolio license. Anything else would have been highly inefficient if not irrational for companies such as these.

As I elaborated in an earlier post, when it comes to a license to a large global SEP portfolio, a  patent-by-patent, country-by-country approach to adjudication can never be a complete solution, and when pursued by an infringer it may in practice amount to a patent hold-out situation. Patent hold-out can in turn lead to the unreasonable result that the SEP-holder cannot efficiently enforce its SEP portfolio simply because it's so large.

Looking now at the "one-stop shopping" approach requested by Ericsson. It does have an obvious attraction point; it directly focuses on the very topic of the actual negotiation between the parties - the value of a license to a global SEP portfolio. And in cases where significant databases of existing license agreements to the very same global SEP portfolio are available as references, this approach should have the potential to be both efficient and fair. In particular if the vast majority of those existing license agreements have been negotiated in good-faith without litigation, something we know is generally true for SEP portfolio licensing. So if Ericsson's public statements about having "more than 100 patent-licensing agreements in place" are to be believed, this case should surely qualify for this approach.

In recent times the "one-stop-shopping" approach has indeed gained support from courts and agencies worldwide. In Microsoft vs Motorola in the US District Court for the Western District of Washington, although some dubious calculation methodology was used as I've mentioned earlier, the court did in fact determine a global FRAND rate to Motorola's SEP portfolio. And the “Request for a FRAND Determination” process endorsed by the US Federal Trade Commission in the consent degree of Motorola vs Google also expressly concerned a global SEP portfolio license. Even the Chinese National Development and Reform Commission (NDRC) in 2014 imposed a similar process on InterDigital for a global SEP portfolio license.

Concluding this review of the initial filings of Apple and Ericsson, I find the argumentation coming from Apple surprisingly unconvincing so far. Its arguments about royalty base and injunctions seem to be at odds not only with US law - including recent US Federal Circuit opinion - but also with decades of regular SEP-portfolio licensing. So for Apple's sake, one would hope that it comes up with some better argumentation as the case continues. Furthermore, the global license value adjudication sought by Ericsson seems to be rather more appropriate for the case at hand - a global SEP portfolio license - as compared to the patent-by-patent approach initiated by Apple.

China and Qualcomm - a new reference

On Feb 9, US chip maker and wireless SEP-holder Qualcomm reached a concluding settlement with the National Development Reform Commission (NDRC) of China, under the Anti-Monopoly Law investigation it has been subject to.

In practice, the settlement results in Qualcomm's cellular SEP portfolio royalty rate being 3.25% of the net selling price of 3G- or 3G/4G-compliant devices sold in China. Clearly the Chinese authorities deemed Qualcomm's original rate to be unreasonably high, since the imposed discount is almost 35%.

But besides that specific message to Qualcomm, a broader message can also be detected here. China tells the world that it does not embrace market disruptiveness with respect to the cellular SEP FRAND licensing model per se. It confirms the applicability of long-standing basic SEP-portfolio licensing principles, with the end-product price as the royalty-base and royalty rates in the order of lower single-digit percentages for strong SEP-portfolios.

This message from China is actually a powerful endorsement of the importance of basic wireless R&D to the ecosystem. Hopefully regulators and policy makers worldwide will consider this input when confronted with various SEP-devaluing proposals popping up in recent times.

New IEEE patent policy approved

On February 9, new IEEE bylaws with radical changes to the patent policy were adopted by the IEEE board. I commented on the alarming content of that policy in an earlier post.

If the policy is allowed to stay this way, I fear that IEEE will over time become irrelevant as a telecoms standards setting organization due to its sudden anti-innovation stance.

A key clause of the policy describes factors for determining a reasonable rate, and the final wording "should include, but need not be limited to, consideration of" was used. But if both "new" and "traditional" factors are to be considered, I think there can be confusion, if not contradiction. For example, the new policy essentially says that the value contributed to the chipset, or perhaps even only a part of it, should be considered. So how can you also consider the value contributed to the end-product, which is something entirely different. And for that matter, are you supposed to compare or combine these valuations, and if so, how? 

In its recently issued Business Review Letter, the US Department of Justice (DoJ) appeared to base its decision to clear the content of the policy on a broad and theoretical reading of that very clause; "Significantly, the Update makes clear that the determination of Reasonable Rates 'need not be limited to' these factors. The Update, then, does not mandate any specific royalty calculation methodology or specific royalty rates.";

I for one don't agree with the US DoJ. I rather predict that "should include, but need not be limited to" followed by an exhaustive list of factors, will - in practice - be at the exclusion of other factors that cause confusion or contradiction, as mentioned above. Which is why I reviewed the clause from that viewpoint.

Unfortunately the DoJ clearance may ultimately be to the detriment of US consumers, as the innovation content of WiFi and other standards is likely to be diluted. In fact, it may already have begun, considering e.g. Qualcomm's recent statement to not abide by the new policy.

For IEEE to stay relevant, it needs to be inclusive with regards to the world's top innovators and innovations. It should be focusing on making the world's most technically advanced and future-proof standards. It can't be wasting its time designing around technology that would ultimately have been beneficial for consumers.

Follow-up on cellular FRAND royalty levels

Recently, a Fierce Wireless article by Keith Mallinson discussed the issue of cellular SEP FRAND royalty rates and in my opinion made some very good points on the cumulative level, based on a practical rather than theoretical argumentation.

In the article, there's an estimation of cumulative royalty rates, where the actual combined licensing revenues of the world's major cellular SEP licensors are basically divided by the actual number of phones sold. Based on such a calculation, it was estimated that "a total aggregate SEP royalty across all handsets worldwide is most likely to be no more than a mid-single-digit percentage". That estimation is probably quite accurate, and I find the "reality-check" approach very refreshing. Not least considering the amount of dubious theoretical argumentation one can find on this subject. 

But since I myself mentioned the figure of "around 10%" in an earlier post on cellular FRAND royalty levels, I felt that a clarification might be needed to remove any potential confusion. Mr Mallinson's calculation method gives a high-level view; a global average estimation of the cumulative royalty rate. On the other hand, my post focused on the  question: "What cumulative SEP FRAND royalty does an OEM without own basic R&D or SEP holdings have to pay?". These are two different things. Let me explain.

Essentially, there are three categories of mobile phones sold on this planet:

1) Phones sold by OEMs with own basic wireless R&D and SEP portfolios.

Many (other) asserting SEP portfolio holders are themselves implementers, and thus need a "grantback" license to these OEMs' SEP portfolios. Therefore, this subset of phones is subject to lower SEP royalty rates due to "royalty netting". As I described in my post and is also clear from Mr Mallinson's article, this is not discriminatory, due to fair value transfer. The cumulative royalty for this category of phones is somewhere in the 0-10% interval.

2) Phones sold by OEMs without own basic wireless R&D and SEP portfolios.

Instead of doing heavy investments in cellular wireless R&D, these OEMs have chosen to focus their business efforts on e.g. design, brand management, localization, feature development, marketing or logistics. The basic cellular technology is procured through the SEP FRAND licensing regime. This is the category of phones I addressed in my post, and in my experience, its cumulative royalty is around 10%.

3) Phones sold by OEMs that are "unwilling licensees".

Due to factors such as patent hold-out and the emergence of major localized OEMs in jurisdictions with less effective IPR enforcement regimes, a number of phones sold globally today are unlicensed. That is, they constitute infringing devices and no royalty at all is payed for them.

So a new OEM has two legitimate choices: a) get into basic wireless R&D to contribute to the standards and in the longer term pay somewhere between 0% and 10% or b) let others do all that and pay around 10%. As is worth repeating, there are no free lunches.

Considering that the phones in the three categories are subject to cumulative royalties of something like 0-10%, 10% and 0% respectively, it's perfectly possible that the global average might end up somewhere around 5% as indicated by Mr. Mallinson's calculation. Having said that though, it's important - for new OEMs but also for courts and agencies - to appreciate that the global average is not necessarily a relevant reference per se for a given OEM. 

Finally, it's of course quite alarming that a subset - and probably a sizable minority - of phones sold globally are currently not covered by SEP portfolio FRAND licenses and corresponding royalty payments. Or to put it more bluntly, are sold by OEMs who "willfully infringe" other's SEP portfolios. Partly a result of the "patent hold-out" challenges discussed in a previous post, this is a grossly under-reported problem that I hope to return to in another post.