By Brian Kahin

Review of Patent Challenges for Standard-Setting in the Global Economy, Board on Science, Technology and Economic Policy, The National Academies, October 2013

The new National Academies report on patents and standards is a landmark effort to shed light on to the tensions between patents and standards in information and communications technology (ICT).  The website includes an impressive set of commissioned background papers that are a useful resource in themselves.  The principal discussion is narrowly focused on four issues:

• terms for licensing standard-essential patents (SEPs), typically, “fair, reasonable, nondiscriminatory (FRAND or RAND);
• requirements for disclosure of SEPs;
• transfer of licensing commitments; and
• injunctive relief for patents subject to licensing commitments.

Most recommendations are directed to standard-setting organizations (SSOs) rather than policymakers.  However, most SSOs are not eager to tackle divisive issues that may complicate priority work on standards.  Fortunately, the report will also be useful to competition agencies, who see SEP practice as something manageably within their purview, as opposed to the larger problems of patents and global patent warfare.

A chapter on China, India, and Brazil takes a broad look at the standards policies of these three emerging economies, but there is a conspicuous gulf between this chapter and focused discussion of the four legal issues.  Additional historical background would have helped bridge the gulf and enable the report to speak to nonspecialists.  For example, much of the competitive dynamic around standards results from a breakdown of the state-sanctioned monopolies and the homogeneity of traditional equipment, a consequence of deregulation and the rise of new technologies in mobile telephony and data communications.  This has been accompanied by specialization, diverging of patent interests, and a spectacular rise in portfolio patenting and the number of patent disclosures.  Smartphones have been destabilizing because of the convergence of computer, software, and Internet companies. These companies may lack essential patents or licenses for wireless technology – but hold patents for complementary technologies.  This imbalance, aggravated by the explosive growth of the smartphone market, has invited litigation and drawn attention from regulators and policymakers

The report is framed in terms of how SSOs can respond to the challenge of patents, rather than how patent laws and practice might respond to the unique importance of standards in ICT. It assumes that patents are an implacable force and has nothing to say about how the standards experience might inform patent policy.  To be sure, this was not part of its charge, which was determined by the funding agency, the US Patent and Trademark Office.  (The National Academies are not self-funded and must secure funding for research projects from Congress, agencies, or foundations.)

The China Chapter

The discussion of China commands attention, because of China’s aggressive engagement with both standards and patents as complementary elements of its indigenous innovation policy.  The report politely accuses China of “bias against the interests of patent holders” for apparently inconsistent reasons (compare p. 12 and p. 131).  Yet despite well-founded concerns about China’s enforcement of trademarks and copyright, China vigorously promotes patenting.  Its issuance and enforcement of large numbers of utility model patents has drawn international criticism, and it grants18 times as many design patents each year as the U.S.

China’s strategic approach to standards is not as anomalous as it might seem to a U.S. audience.  Europe, too, has had a distinctive standards strategy focused on market integration and has only recently come to terms with the global industry-driven consortium model developed in the U.S.  Previously, France and other countries viewed standards (such as the SECAM standard for analog television) as a vehicle for exporting technology and influence.  The U.S. has always been an outlier in its limited formal involvement in standards policy.  Nonetheless, federal support for the development of Internet standards outside of the international standards system has had a legendary long-term impact on standards development.  Core Internet standards remain freely available for anyone to use without seeking permission, not by law but by clear consensus.

Meanwhile China has been involved in a long legalistic dance over whether national mandated standards may bear royalties (background paper, pp. 11-13).

Chinese views have been shaped by its experience with high royalties for DVD players that left Chinese manufacturers with virtually no profit margin.  When Chinese manufacturers developed an alternative standard (AVD), DVD royalties charged to Chinese manufacturers dropped substantially.  The background paper reports a similar story for video compression (p. 19).

Those two cases involved patent pools, situations where the royalty stacking issue stands out.  The report views royalty stacking from the bottom up as a classic Cournot complements problem — where each essential patent owner seeks to maximize its own return and hold everybody else hostage.  But if viewed from the top down, royalty stacking raises difficult questions about the relative value of patent incentives, the social benefits of standardization, and the appropriate strategy for emerging economies of different sizes.  While it is reasonable to be critical of China for discriminating against foreign participation in developing standards, China’s efforts to limit the role of patents in mandated standards deserve a fuller, more nuanced discussion.  References to “international norms” just beg questions of what the norms are and what they should be.

Royalty-Free Licensing (FRAND-RF)

The report notes that six of the twelve of the SSOs require or prefer royalty-free licenses, but it gives a limited account of participant motivations for royalty-free licensing and when SSOs might favor RF.  In fact, there is considerable confusion about the desirability of royalty-free licensing. The first version of the European Interoperability Framework for government specified royalty-free; yet the initial DG COMP draft guidelines on horizontal cooperation were openly hostile to SSO policies favoring royalty-free.

The report acknowledges the incentive of broad and rapid uptake for royalty-free licensing, but it fails to note other incentives that might induce firms to allow royalty-free use of patented technology.  These incentives include: complementary assets (including patent claims that are not essential to practicing the standard), lead-time and learning-curve advantages, reputation, reciprocity, lower transaction costs, and diminished strategic behavior by participants (allowing for faster development of the standard).  Unfortunately, this omission reinforces the erroneous assumption that patents will not be licensed royalty-free because there is no benefit to the patent holder.  For example, as expressed by Judge Holderman in the recent Innovatio case:

[A] RAND rate must be set high enough to ensure that innovators in the future have an appropriate incentive to invest in future developments and to contribute their inventions to the standard-setting process. As Judge Robart put it, “[t]o induce the creation of valuable standards, the RAND commitment must guarantee that holders of valuable intellectual property will receive reasonable royalties on that property.”

This is plainly wrong as the preference for royalty-free in many contexts shows, although it is plainly less wrong in some contexts than in others.  ETSI with its thousands of SEP disclosures is at one end of a highly differentiated spectrum, W3C is at the other.  It is important to understand the difference – rather than burying them in the one-size-fits-all ideology of the patent system.

Comparing Patents and Standards

The report reads the growing number of disclosures over the past 25 years as a sign of growing importance of patents, but it can also be read as a sign of declining quality, as OECD analysis of indicators asserts.  There are now over 250,000 US patents issued annually.  According to the PTO Chief Economist’s working definition, since adopted by the Government Accountability Office, fully half are in “software-related” classes – which presumably also means ICT-related.

Patents may be more important in the aggregate because there are so many of them, but less important individually simply because the value of each is diluted.  If a proper apportionment framework is used to calculate royalties based on real markets in products or components, the value of individual patents will diminish if product prices stay constant and functionality increases.  On the other hand, to the extent that PAEs can use patents to extract value from sunk investments or nuisance settlements, patents will become more valuable in patent markets, but at the expense of fully functional products and consumers.

The report sees standards in terms of interoperability rather than the dynamics of market development.  It is missing a typology that might explain differences among SSOs (and pools) and their policies.  Yet it seems clear that in software, there is a preference for royalty-free, while the greater investments required in lower-level infrastructure can justify patents in terms of barriers to entry, higher investment requirements, revenues at stake, and the ability to manage the high transactions costs of patents.  Another dimension might be: thin, arbitrary standards for exchanging information vs thick standards for rich technology, like DVD.  There will be fewer patents for the former and more for the latter.  Another is the stage of development in the underlying technology:  Is the standard simply formalizing a mature product? – or an effort to legitimize and enable a new market?

Also undiscussed is what the market for standards should look like.  Should there be competition among standards and among standards organizations?  Firms and SSOs do not like to see duplication of effort, but in a free market, it is hard to object to a standard optimized for affordability competing with a standard optimized for performance?  China’s efforts to develop alternative standards appear to have reduced royalties even though the alternative standards have not been that successful.  There are many possible tradeoffs on technical and marketing factors beyond simple claims of performance or quality, and the global battle over the OOXML document format after the adoption of ODF provides ample precedent for competing standards.

The failure to address these larger issues head-on is a missed opportunity to examine the relationship between patents and standards more deeply.  While deference to the patent system may be implicit in a study funded by the PTO, the broader look at China raises larger issues of principle and political economy that cannot be answered effectively by handwaving about “international norms.”  The discussion of the four issues makes clear that there is plenty of inconsistency in SSO practice, as well as resistance to addressing fundamentals such as the meaning of FRAND.

At bottom, there is a radical difference between ICT standards and ICT patents in terms of how knowledge is managed, assessed, and exploited.  Standards are tested in the market:  Knowledge must be developed collectively, peer-reviewed, codified, clearly communicated, and must be successfully implemented.  Patents, by contrast, are applied for ex parte (in complete secrecy for the first 18 months), and must be granted unless the examiner can show good reason why not.  Quite unlike standards, which are the outcome of long, hard work by hundreds of experts, patents may be awarded by a single hurried government employee, carry an unjustifiably high presumption of validity (based on an average 18 hours of the examiner’s time), and do not require a working model.  Yet despite the resources (including the patented technology of multiple parties), negotiations, validation, and reliance interests that go into standards, standards are subordinated to patents – legally, institutionally, and politically.

The report does not confront this fundamental anomaly, but it is helpful for stressing the apportionment problem in digital products as discussed by the court in Microsoft v. Motorola.  I.e., royalties should reflect the relative value of the patented technology within the standard and the relative value of the standard and its SEPs within marketed products or components.  Within this framework, standards are an intermediate benchmark between patents and products, so it would be helpful to know not only the correct number of SEPs in each standard but also the number of standards in the marketed product.  For example, a study (not mentioned in the report) found 251 standards in a laptop, and one would imagine there would be many more in a smartphone.  There is additional value that should be taken into account at both levels: public domain technology, design, integration, testing, market research, marketing, etc., although these elements even may be more difficult to identify and value.

Disclosure, Transfers, and Injunctions

Any exercise on apportionment requires knowing what patents are truly essential to the standard.  Yet given the precision and predictability of digital technology, the interface with patents is strikingly inefficient and indeterminate. This is basically a patent problem, not a standards problem, so it does not get the attention it needs.  The report notes that firms resist full disclosure because searching their own portfolio is deemed too costly and indeterminate — and that, however motivated, there is substantial overdisclosure as well as underdisclosure.  In some SSOs, blanket disclosure is allowed – in one, no disclosure is required.  Perhaps there is good reason to waive explicit disclosure, since studies of two standards (WCDMA and GSM) indicate that less than 30% of the disclosed SEPs are found essential by neutral experts.

Licensing, too, is far less rigorous than one might assume.  The report notes the prevalence of bilateral portfolio cross-licensing among large firms, which obviates individual valuation of patents and diminishes the costs and risks of litigation.  In cross-licensing portfolios, distinctions are not normally made between SEPs and non-SEPs, again perhaps for good reason.  At same time, as the survey-based study commissioned by DG Enterprise shows, licenses are often not sought from small companies.  In short, business practices have sought to minimize transaction costs, while the report urges a more precise, lawyerly approach.

By contrast, there seems to be little disagreement that FRAND commitments and licenses should follow and bind the SEPs they reference or apply to, including in bankruptcy proceedings.

Injunctive relief stands out as the one area where the NAS committee could not come to agreement.  The dispute centers on whether FRAND-encumbered SEPs under a licensing commitment should be litigated in a forum that can only grant injunctive relief or its equivalent – i.e., a forum like the International Trade Commission (ITC) that is unable to assess monetary damages.  The majority say no, and further that damages and validity should be determined before an injunction is considered.  The minority would allow disputes to go forward when the only possible remedy is an injunction or the equivalent.

The committee split raises tough questions about the role of the ITC that are rarely aired, even though the role of the ITC in U.S. patent practice is in itself such a departure from “international norms.”  The split draws attention to the problem, but the recommendations are muted because they do not mention the ITC by name.  Yet the principal reason patent holders use the ITC is because they can get exclusion orders that bar producing companies from distributing imported products in the U.S., even when the patented technology is a miniscule contribution to the product.  Whereas, after the 2006 eBay decision, district courts must consider four factors before granting an injunction, including the sufficiency of monetary damages.

In a widely reported ITC case this year, Samsung won an exclusion order against Apple on what may have been standard-essential patents.  As the reports notes (p. 106), the ITC was not satisfied that the patents were standard-essential or that Samsung had failed to offer FRAND terms. Nevertheless, the US Trade Representative overruled the exclusion order on SEP-related policy grounds without specifically disagreeing with the ITC’s conclusions.  While the exclusion order is undesirable as a pure matter of policy, the USTR contradicts the ITC conclusions without explanation – in effect, abandoning the rule of law, for what might be read as political reasons.

This incident together with both majority and minority recommendations beg the question of whether the ITC’s exclusionary orders discriminate against imports – and whether this discrimination has outlived its purpose in a WTO world.  No other country imposes a similar barrier for imports – including China, which is subject to criticism for its particular flavor of mercantilism.  Quoting T.S. Elliot: “The last temptation is the greatest treason: To do the right deed for the wrong reason.”  Perhaps that puts it too strongly, but both the ITC ruling and the USTR letter of override undermine principles that the U.S. champions in other contexts.

Brian Kahin is a Fellow at the Center for Digital Business at the MIT Sloan School of Management and a Senior Fellow at the Computer & Communications Industry Association (CCIA).