Yet another article on the 10-year anniversary of the iPhone

An extra-terrestrial alien visiting Earth in 2007 and returning, now, one decade later, might, at first glance, notice little difference in smartphones between times. For example, most-recent iPhone models superficially appear very similar to their predecessors including the first iPhone model in 2007. The external designs have remained fundamentally much the same including thin form factors, rounded corners and relatively large displays (with multi-touch operation) in comparison to featurephones.

                                                    Spot the difference

iPhone (2007)

iPhone 8 (as rumoured July 2017)

However, superficial appearances are very misleading: technological capabilities in mobile phones have improved massively with numerous valuable innovations from various contributors over the last decade, as did capabilities over the preceding couple of decades since the introduction of the first cellular “bricks” in the mid 1980s.

Another major milestone in cellular technology developments towards 5G

Recent new technology deployments with Gigabit LTE at Telstra in Australia, Sprint in the US and EE in the UK highlight how much cellular communications technologies have improved since the introduction of mobile data services with circuit-switched and then packet-switched offerings from around 20 years ago. Peak and average user data speeds on cellular networks have increased by a factor of 10,000 over 20 years. By way of comparison, microprocessor performance doubling every couple of years, as predicted by Moore’s Law, has increased only one thousand-fold over that period. Cellular performance improvements are therefore quite spectacular given the vagaries of connecting through the ether up to hundreds of metres, as well as processing those signals in the confines of around one square centimetre of baseband processor silicon!

2016 iPhone 7 is 1,000 times faster than the 2007 model

Whereas Apple has done an outstanding job in improving its iPhones in various ways and in motivating its customers to upgrade to later models, it is significantly dependent on other companies for many technical innovations that it includes in its devices.

While marketing departments and the press look for eye-catching new features on specific device models that might surge demand for the latter, it is relentless standards development work with innovations and performance improvements in cellular technologies to increase speeds, network capacity and reduce power consumption that provide the crucial underpinnings for these ––particularly for HD, 4K or even 8K video that sends or receives very large volumes of data over the mobile networks.

Inspiration and perspiration

Development work for this including 4G and 5G technologies is largely undertaken by a hard core of several major technology-developing firms. Research on attendance records of all the 3GPP working group meetings between 2005 and 2014 reveals that a few highly-active firms are largely responsible for the technical developments in that standards development organisation. Over this period, a total of 3,452,040 man hours were spent in 825 working group meetings, mostly in the development of 3G and 4G standards. Distribution of contributions to 3GPP is highly skewed, with a few firms submitting the vast majority. For example, the top two percent of firms (i.e. 9 of them) are responsible for submitting 60 percent of all contributions. Furthermore, approximately one-third of all participating firms (i.e. 161 of them) have not submitted a single contribution to 3GPP.

However, most of the activity in the public records of standards development organisation 3GPP is the mere tip of the iceberg in terms of the total amount of development work undertaken, with even more extensive other activities submerged from public view.

As I noted in a report on innovation and intellectual property protection, it is a popular misconception that innovation is random or serendipitous. In fact, it takes many ideas to find a few initiatives worth experimenting with, which may then enable some to be identified that are worth investing in significantly and might ultimately lead to a winner or two with sufficient development effort and investment. This work is largely undertaken outside of SDO meetings.

The numbers of patents and patent applications declared to the ETSI IPR database as possibly being essential to these cellular standards are also very skewed. A small number of mostly the same companies as above account for a large proportion of patent declarations. When I last checked, seven companies including Ericsson, Huawei, Nokia, Qualcomm and Samsung accounted for 70 percent of the many thousands of patents declared in the period 2008 to mid 2015.

Value for money in cellular patent licensing

Licensing fees paid in the smartphone industry are substantially for standard-essential patents and in some cases for non-SEPs. Total patent licensing costs for Apple and other smartphone OEMs at around only a few percent of revenues are good value given the development efforts and performance improvements delivered by technology developers.

Licensing fees pale in comparison to the profits generated by Apple. The original iPhone was introduced in June 2007 at a price of $599 in the US. This and subsequent iPhone models have generated very large profit margins, as illustrated by the difference between retail prices and manufacturing costs.

Substantial mark ups and profits to Apple on iPhone

	2007	2011	2016
Model	iPhone	iPhone 4s	iPhone 7
Standard/technology	2G EDGE	3G HSPA	4G LTE- Advanced
Version (storage)	8GB	16GB	32GB
Full retail price*	$599.00	$649.00	$649.00
BoM cost*	$222.55	$178.82	$246.91
Markup ($)	$376.45	$470.18	$402.09
Markup (%)	169%	263%	163%

* Source: TechInsights/Portelligent

According to Strategy Analytics, Apple sold 231 million iPhones with an operating profit (i.e. after some other operational costs) averaging $239 per phone in 2015. That represents 36 percent of its $669 average selling price that year.

Following the introduction of a new model every year at gradually increasing prices, “the 10th anniversary iPhone, the next model, expected to be massively redesigned and packed with state-of-the-art technology, could sell for as much as $1,200 to $1,400, according to some estimates”.

Analysts also estimate patent licensing fees paid to Qualcomm average about $10 to $20 per iPhone. Apple has stated that Qualcomm charges it "at least five times more in payments than all the other cellular patent licensors we have agreements with combined."

On that basis, Apple is paying a total of between $12.50 and $25.00 per iPhone in fees for licensing from all cellular patent licensors. That is equal to between two percent and four percent of iPhone prices. Licensing fees as a percentage of consumers’ total cellular expenditures over a smartphone's approximate two-year service life, including operator service fees averaging around $40 per connection per month in the US, for example, are considerably lower.

Happy anniversaries

It is also ten years since I published my abovementioned report, noting as well that innovation can occur in many ways, with a variety of different business models and that fully vertically-integrated companies had become a rarity in technology industries. I stated that explicit recognition of value through licensing was increasing innovation, competition and customer choice with third-party supply of IP, in addition to that for components and manufacturing. That conclusion still holds.

Adjusting the Balance in SEP Evaluations and Licensing

A European Commission DG Growth initiative described in its Roadmap on Standard Essential Patents for a European digitalised economy aims to increase information on SEPs so implementers can get a better idea about which of these they might be infringing. Additional disclosures on how patent claims might read on the standards could be beneficial. Requirements should reflect the dynamics and uncertainties in standards development and patent prosecution and must not be onerous to patent owners. These are issues for standards development organisations to consider.

A report DG Growth commissioned in support of its initiative entitled Transparency, Predictability, and Efficiency of SSO-based Standardization and SEP Licensing (the CRA report) proposes that SDOs or the European Patent Office could also help meet this objective by being appointinted the central assessor that would screen patent disclosures to determine and count which patents are truly essential. This would be undesirable intervention with various adverse consequences. As I wrote here for IP Finance in detail very recently, third-party determinations on large portfolios are inherently subjective, inconsistent and unreliable.

The creation of this EC-ordained “patent counting” database would also lend it to being used as an interventionist means of valuing SEP portfolios. In conjunction with the unwarranted imposition of maximum cumulative rates (i.e. royalty caps), this could facilitate the ill-conceived price regulation alluded to by the Competition Commissioner.

The CRA report also embraces defective patent hold-up and royalty stacking theories. General theories on hold-up and “Cournot complements” are misrepresented and do not apply to patents. There is a lack of supporting evidence on alleged patent hold-up, royalty stacking and much of it to the contrary including that for opposing effects from patent hold-out (i.e. patent trespass). 

SDO IPR policies are commonly misrepresented with the bogus notion that patent owners should be deprived a share of value from use of patents in standards. Neither the economics nor the law is settled here. Sharing in the “gains from trade” incentivises risky investments.

Private ordering has worked very well in 2G, 3G, 4G and it will continue to work well in IoT including 5G. Prospects are no more uncertain now than they were when these previous-generation standards were introduced and helped transform the communications markets. Decades of fruitful progress indicates it is not “too soon to tell” how things are playing out. By all measures these markets are extraordinarily competitive and successful, with large research and development investments, extensive resulting innovation, massive growth in subscribers and data consumption, reducing quality-adjusted prices, and dramatic shifts in market shares with new market entry, market exits, low and decreasing concentration in supply. I have been showing this with facts and figures here, here and here for many years, and as others have confirmed.

The CRA report is right to reject a mandatory switch to chip-based royalty rates and licensing, and to recognise the legitimacy of charging different royalties depending on “field-of-use” (e.g. an IoT lightbulb versus an augmented reality headset or a self-driving car). This well-established principle aligns costs with functionality used and value generated.

IoT is expected to be worth up to the teens of trillions of dollars to the global economy by 2025. That is 500 times more than the cost of licensing the communications technologies that are already providing the growth fulcrum for IoT developments. Undercutting royalties will diminish gains that could otherwise be obtained widely by leveraging reinvestments in intellectual property.

With it being much more difficult to obtain injunctions than it used to be, as the CRA report and Justice Birss in Unwired Planet v. Huawei also observes, the scales have already been tipped significantly in favour of implementers versus technology developers in terms of bargaining power. The balance here needs to be redressed here not swayed further. Royalties are flat or declining while opportunities and demands to invest in R&D for the good of all in IoT and 5G are increasing.

DG Growth should not interfere with SDO governance or try to pick winners among these or their IPR policies. Rather than speculating about how much aggregate licensing costs could be, costs should be measured by asking licensees what they are actually paying in cash royalties. Compare that with the value the resulting technologies deliver in the market.

Private ordering is preferable to public ordering and intervention is unnecessary. SDOs, patent pools, other licensing platforms and bilateral licensing under FRAND conditions can continue to serve the industry well and to the benefit of consumers.

However, if EC decides to intervene there should be impact assessments before intervening and empirical analysis of effects thereafter. DG Growth should also measure the results previous rulings— including those affecting the availability of injunctions— have already had on royalty rates and how long it takes to complete licensing agreements.

DG Growth’s analysis should be as open and transparent as possible, for scrutiny by all.

The above is a summary to my full article, here. This supplements my response to the DG Growth consultation on this topic in 2015.

FRAND licensing: A call for greater transparency

No topic in technology licensing is more vexing than standard-setting organizations (SSO’s), standard essential patent (SEP) owners and FRAND terms. Anders Møller, a recent graduate of Oxford engaged as an independent economist, describes some interesting research that shows how much transparent FRAND licensing negotiations still need to be in order to achieve their expressed goals.

With the IEEE having updated its patent policy in 2015, some standard essential patent (SEP) owners raised concerns about the economic effectiveness of its move. While it is still too early to tell what economic impact their updated patent policy will have, it is worth paying attention to an econometric analysis of the patent policy introduced by VITA (VMEbus International Trade Association ) in 2007, undertaken by the now Chief Economist of the EPO, Professor Yann Ménière, and François Lévêque, Professor at CERNA, MINES ParisTech.

SSOs generally require firms to license their patents for a given piece of technology to users on a Fair, Reasonable and Non-Discriminatory basis, or FRAND. The FRAND principle is adopted to avoid the moral hazard of license holders ex post (after negotiations) using their intellectual property to extract supra-normal rents from licensees, once technologies have become indispensable to a given manufacturing process. This is commonly referred to as patent holdup. As Lévêque and Ménière point out--

“the mainstream view is that the rationale for [F]RAND is to mitigate holdup and that the R in RAND stands for the level of royalty resulting from competition in advance of standard selection (pp. 5-6).”

In turn, this promotes the broader adoption of new standards and, hence, higher royalty earnings for licensors. However, there is no common definition of what constitutes a “reasonable” royalty, which creates vagueness and thus greater uncertainty in licensing negotiations.

Lévêque and Ménière argue that the FRAND principle has worked surprisingly well up until recently, given that both licensors and licensees have an incentive to abide by it (i.e. licensors gain an increase in uptake of their IP while licensees mitigate holdup risk). However, a few cases have been highly publicized involving antitrust litigations over IP holdup. In the words of Lévêque and Ménière, this has “instilled doubts for standard users about [F]RAND as an effective safeguard against holdup (p. 16).”

This creates a social welfare loss because both licensors and licensees are disincentivized to negotiate and license new technology standards. According to the authors--

“the threat of being held up results in allocative and productive inefficiencies that are detrimental to society, especially to consumers for whom standard-compliment products are produced in lower quantities, later and at a higher cost (p. 21).”

So, what can SSOs do about this? Lévêque and Ménière Leveque model two policy approaches. The first is based on the January 2007 updated patent policy of VITA, which states that members “must declare the maximum royalty rate for all patent claims... that may become essential to implement the Draft VSO Specification.” Furthermore, this declaration is “irrevocable,” and subsequent declarations covering the same "may only supersede the prior Declaration if the subsequent Declaration is less restrictive."

This is quite a departure from the practice of simply relying on FRAND in principle and negotiating without disclosures, which Lévêque and Ménière show to be a helpful tactic for SSOs to deal with the with the growing uncertainty associated with what ‘reasonable’ means in the FRAND regime.

Based on a mathematical model, the authors find that licensing commitments before manufacturers sink fixed costs into new product standards “induce lower prices for standard compliant products (p. 41).” This is because the expected costs are always lower when manufacturers are guaranteed a maximum royalty rate than with no assurance at all. Although the licensor accepts a certain amount of risk by committing to charging lower royalties ex ante before demand is realized, this is offset by the promotion of higher technology uptake (as lower risk encourages market entry of more manufacturers). There is therefore a convergence of interests between licensors and consumers as doing otherwise would reduce the consumption of tariff compliant products.

The second policy approach is to make declarations optional, whether of a specific royalty or a ceiling. Lévêque and Ménière’s model shows that – assuming rational action – licensors will always choose either a royalty ceiling or an exact royalty rate, as a high maximum rate is always better than having no assurance in place at all. Moreover, such a flexible policy approach allows licensors to choose the option which maximizes expected profits, thus leading to more innovation which is a significant positive externality.

The authors therefore conclude that “although the VITA IP policy is an attractive one, the policy of IEEE-SA is even preferable”, because it maximises both licensor and consumer benefits. The implications are simple yet profound for policy makers. With the European Commission as well as the Joint Research Council having issued several studies on the topic , and the European Union considering guidelines on SSO policies by early summer of this year, such insights are invaluable.

The study from Professor Lévêque and Professor Ménière clearly shows that there is a need for more transparency in FRAND licensing negotiations and that SSOs need to assume their responsibility so to further clarify what FRAND means. This is needed so to restore confidence in the FRAND regime. A VITA-like policy requiring a specific or maximum royalty rate declared up front is essential to maximize social welfare from standards and the respective patents that read on them.

The challenge for other SSOs, like ETSI or CEN/CENELEC, now lies not only in adopting policies that include flexible disclosure regimes, but also in how to institutionalize the practice by all negotiating parties in order to avoid any one licensor attempting to gain a first-mover advantage. The success of this approach will increase innovation, which is ultimately the goal of a prudent IP framework.