This guest blog has been written by Prof. HAO Yuan of  Tsinghua University School of Law.

 China is facing a pressing need to build its innovation-driven economy. To facilitate key features of an innovative economy Chinese anti-monopoly authorities, along with their worldwide peers, face a daunting challenge of transition from a static regime to a more dynamic one.  Several recent judicial and administrative disputes, including Huawei v. IDC, Shenzhen Intermediate People’s Court, Shen Zhong Fa Zhi Min Chu Zi No. 857 (2011); Huawei v. IDC, Guangdong High People’s Court, Yue Gao Fa Min San Zhong Zi No. 306 (2013); and the NDRC’s administrative investigation against Qualcomm, Administrative Penalty Decision [2015] No. 1, point out the need to better understand the IP and antitrust intersection, particularly with regard to controversial issues such as  “unfairly high patent pricing (不合理专利高价)”.   This blog summarizes my recent paper (available on SSRN) which addresses this important issue (upcoming 2020 in GRUR International (Journal of European and International IP Law)).

Chinese Anti-Monopoly Law (“AML”), in contrast to US law but not facially dissimilar to EU competition law, pays substantial attention to a dominant market player’s unilateral “exploitative” conduct. Specifically, section 17(1) of the AML (2008) forbids a dominant undertaking from “abusing its market position” by selling at “unfairly high prices”. However, neither the Law nor later enacted Judicial Interpretations clearly define “unfairly high price” in the anti-monopoly sense. Correspondingly, courts and enforcement agencies have significant discretion in characterizing a market price as “unfairly high”, thereby potentially exposing an undertaking to harsh penalties.

Despite perhaps a legitimate institutional intention, this ex post legal risk of being found “unfairly high” could seriously curtail business entities’ ex ante incentive to innovate in China. Meanwhile, lacking adequate legal and economic guidance, this institutional discretion would likely result in significant error costs. Such costs are likely to be even more severe in the context of patent-intensive industries, particularly those in China’s burgeoning high-tech sector.

Section 55 of the AML arguably provides an IP “safety harbor”, providing that a proper exercise of IPR shall be immune from the AML scrutiny, while “an abuse of IPR, excluding or restricting on competition” shall not. As pointed out elsewhere on this website, this provision remains essentially unchanged in a recently proposed revisions to the AML.

Despite this statutory framework, recent cases indicate that the IPR immunity approach has been largely ignored in practice. In the 2011 Huawei v. IDC action, both courts found IDC’s patent pricing to be “unfairly high” primarily on three grounds: one, the licensing royalties IDC offered to Huawei were “apparently higher” than “comparable licenses”, i.e. those royalties IDC charged other companies previously (though whether these licenses were truly comparable with licenses made to Apple, Samsung and others was worthy of serious discussion); two, “IDC’s act of charging unfairly high licensing fee to Huawei, will force Huawei to either quit the competition in the relevant end product market, or accept the unfair pricing conditions, which will render Huawei to increased costs and decreased profits in the relevant end product market, directly restricting its capability to compete”; and three, IDC required Huawei to give global patent grant-backs on a royalty-free basis, an arguable violation of both the Antimonopoly Law and the then-existing Administration of Technology Import/Export Regulations^,. Similarly in the Decision against Qualcomm, the NDRC also disregarded Section 55 immunity, finding that Qualcomm’s licensing conditions were “unfairly high” due to these three factors: charging a flat fee for an ever-changing patent portfolio without proving the replenished patents are of equal value to expired ones; coercion for free grant-backs; and using the entire end product as royalty base. So far, Huawei v. IDC and Qualcomm have been the only two completed Chinese cases that entailed an extensive “unfairly high patent pricing” analysis.  Other currently ongoing cases include Iwncomm v. Apple (Beijing IP Court), Huawei v. Panoptics (Shenzhen Intermediate People’s Court), Xiaomi v. Sisvel (Beijing IP Court), and the anti-monopoly investigation by SAMR (China’s AML administrative enforcement authority) against Ericsson’s 3G/4G SEP licensing practices, etc..

It is true that both IDC and Qualcomm involved SEPs and  arguably an extra layer of FRAND commitment needs consideration, which may require additional law and policy analysis. For example, the courts might have looked to how to interpret FRAND in Chinese legal framework – is it a specific commitment to a SSO, or a general “principle （原则）” imposable by loosely grounded policy arguments? Will this FRAND interpretation affect an SEP holder’s anti-monopoly obligations in China?  And if so, how? Nevertheless, to a large extent these cases still reflected a practical departure from the statutory IPR immunity framework approach under Chinese law. Such an aggressive approach has also been reflected in Section 14 of the Anti-Monopoly Guidelines for IP Abuse (Draft for Comment 2017) as published by the Anti-Monopoly Commission of State Council. (Editors note: see the comments of the American Bar Association, as well as earlier drafts by enforcement agencies here). According to this rule-of-reason framework, “Business entities that enjoy dominant market positions may abuse their dominance by licensing intellectual property at unfairly high prices, excluding or restricting competition; while assessing whether such high pricing constitutes abuse of dominant market position, the following factors may be considered: (i) calculation method of license royalty, as well as the IP’s contribution to the value of goods; (ii) the business entity’s undertaking with regard to the IP licensing; (iii) the license history or comparable license standards; (iv) license conditions that may have led to unfairly high pricing, including territorial or product scope restrictions; and (v) whether a portfolio licensing includes expired or invalid IP”.

This laundry-list framework is not administrable. How these factors play out in a specific case, whether and how they would interact with each other, how much weight should be attached to each and every one of them, how to incorporate the pro-efficiency features into consideration … all these important practical questions remain unanswered. Facing such vague rules, patentees and other relevant market players would find it very hard to ascertain legal compliance. Vague rules combined with the high stakes often entailed in antitrust cases may lead to rent seeking and bad precedents, further upsetting a prior well-functioning market system, and harming market players’ confidence to  continuously invest in innovation.

One critical reason for this stark departure between the Law and enforcement may be the lack of an administrable test to differentiate “proper exercise” of patent right from “abuse”. Ambiguity lies in at least two aspects. First, due to the very mechanisms of patent regime in fostering innovation, even a proper exercise of right would necessarily restrict certain market competitions. Thus, it seems that an over-general standard of whether “to eliminate or restrict competition” cannot work as the ultimate test to differentiate abuse from legitimate exercise. Second, despite extensive use of “patent abuse” in the past decades worldwide, the exact contour of this concept remains elusive. Resorting solely to the various definitions and constructions in comparative law of sister jurisdictions, helpful as it is, would not solve these ambiguities adequately.

Mark Cohen, the editor of this blog, has also noted that one little-referenced basis for resolving the distinction between “proper exercise” and “abuse” is in fact found in Article 40 of the TRIPS Agreement, which creates a similarly vague dichotomy in permitting WTO members to provide remedies for abusive licensing practices.  Article 40 authorizes member states to address “licensing practices or conditions that may in particular cases constitute an abuse of intellectual property rights having an adverse effect on competition in the relevant market.”   Article 40 also permits member states to control mandatory grant backs of the type address in Huawei v. IDC.  The ambiguities in China’s AML and TIER may in no small part be due to the ambiguities within Article 40 itself, which may have been an important resource for the drafting of Article 55 of the AML.  Furthermore, as Prof. Cohen has separately noted, the legislative history of  Article 55 suggests that it was intended to provide greater assurance that enforcement of IP rights would not by themselves violate the AML, see Harris et al, Antimonopoly Law and Practice in China (2011), at 55. Prof. Cohen’s latter point was well proved in the little referenced, but authoritative, Statute Interpretation and Legislative Reasoning of the Anti-Monopoly Law (hereinafter “Statute Interpretation”) (published by the Legislative Work Commission of the NPC), which noted that, “To restrict competition by the exercise of IP right, is permitted by law after its balance of distinctive interests; in other words, certain restrictions must be imposed on competition to further technological innovation and improve competitive capability. Thus, monopoly status resulting from IP right as well as the restriction on competition because of the exercise of IP right, are legal acts based upon the legal authorization. Other jurisdictions generally treat the legal exercise of IP right as an exception to the application of antitrust laws” ; but on the other hand,  “[I]f an IP right holder abuses its right by exceeding the scope of exclusive right, in acquiring or strengthening a monopoly position, such act will not be protected, and if it excludes or restricts competition, the AML shall govern here.”

Though not a binding document in itself,  the Statute Interpretation can work as a useful guidance, since it embodied important legislative intents and reasoning at the time, particularly when no other legislative reports are accessible to the public. Therefore, to assess “unfairly high patent pricing”, a crucial question is whether the alleged excessive pricing act falls within the scope authorized by patent law. If the answer is positive, then even if this “excessive” pricing would restrict certain competition in the relevant market, such static efficiency loss is a necessary cost we deliberately pay for sustainable innovation, i.e. section 55 immunity applies. If and only if on the other hand, the alleged pricing has been proved to exceed the legally authorized scope of such patent, section 55 immunity will then be stripped and the high pricing act will be examined under section 17(1) of the AML. In other words, to be stripped of the IP clause immunity and subject to “unfairly high patent pricing” scrutiny under the AML lens, the alleged pricing must be not only excessive enough to exclude competition, but also in a way that departed from what patent law has originally contemplated.

Economic insights into the patent law jurisprudence of various jurisdictions, including China, US and EU, reveal that despite some rather subtle differences, all patent regimes promote innovation essentially through the instigation of dynamic competition. Specifically, the patent regime (1) bridges an innovator’s technological R&D to market demand: the better aligned a new technology is to consumer needs, the more valuable this “shell” of exclusive right would be in the marketplace, resulting in higher profit an innovator can harvest by excluding her imitating competitors (“competition by imitation” or “CBI”).  In addition, the patent regime enables innovators to signal potential “innocent” imitators utilizing mechanisms such as disclosure requirements, thereby reducing the amount of duplicative investment in innovation. See Edmund W. Kitch, The Nature and Function of the Patent System, Journal of Law and Economics, Vol. 20, No. 2 at 278; (2) with some caveats addressed in my full paper, by restricting CBI with reasonably tailored claims of the patent right coupled with disclosure requirements, the patent regime simultaneously induces social resources into “inventing around” activities, i.e. to provide better/cheaper substituting technologies, thereby encouraging competition by substitution (“CBS”).    In certain circumstances, the fruits of such CBS  may also be protectable with a new property (patent) right. For example, if a competitor’s “inventing around” technology is found sufficiently “inventive” or “non-obvious”, it would likely be entitled to a patent in itself.  See Robert P. Merges, Intellectual Property Rights and Bargaining Breakdown: The Case of Blocking Patents, 62 Tenn. L. Rev. 75 (1994); and (3) feeling the pressure of CBS, the earlier patentee/innovator would likely be incentivized to keep on her next run of “inventing around”, striving to stay a winner in the marketplace. In this way, a virtuous circle of dynamic competition could come into force.

On the other hand, a brief comb-through of the economic foundations and jurisprudential development of antitrust law suggests that this regime promotes innovation essentially through efforts to maintain an (optimally) heterogeneous competition ecosystem. Based on an understanding of both innovation-facilitating mechanisms, two insights are worthy of note here. First, a mutual ground shared by both patent and antitrust regimes in their disparate routes to innovation is the common facilitation of dynamic competition. And correspondingly, Chinese anti-monopoly law should respect the very patent mechanism, i.e. the instigation of dynamic competition, which pivots on the CBS precisely through efforts including a restriction upon CBI. Therefore, to put it in practical terms, if an alleged act at the intersection with anti-monopoly law, such as an accusation of “excessive” pricing, merely caused a restriction on CBI   as indicated in supra-competitive profits, AML must refrain from intervening. In other words, AML should not lightly disturb a well-functioning circle of dynamic competition, simply because under its lens CBI  or static efficiency seems restricted in a link, and local optimum seems not achieved – as a result, a supra-competitive profit enjoyed by a patentee should be found legal per se if static efficiency loss is the only proven harm. More importantly, the focus should be instead on whether the questioned act would not only restrict competition, but also in a way that departed from what the patent regime has contemplated, i.e. the restriction is to dynamic competition or CBS itself. As an example of the latter scenario, consider a joint agreement between two patentees owning substituting technologies or a patent pool consisting of competing technologies, with the patentees or pool members consenting to suppression of one technology in promotion of another (others).

Coming back to the context of patent pricing, is it possible for a high pricing to restrict dynamic competition, thereby constituting an abuse under the anti-monopoly lens? Maybe yes in theory. Arguably in exceptional circumstances, certain pricing may be so excessively structured to actually constitute a refusal to subsequent or follow-on innovators, or at least a significant “margin squeeze” for them, and such refusal or squeeze in theory may lead to foreclosure of the CBS itself or competition in downstream markets, thus frustrating the circle of dynamic competition.. A little more discussion on the interaction of patent and market competition may help here. In a nutshell, from the perspective of a patent law student, there often exist three tiers of market competition. Tier I is vanilla CBI, referring to those competing technologies with mere replication or insubstantial changes, the only advantage of which are their lower prices (mostly by saving duplicating R&D costs). Of course, CBI can be efficient and legal in those circumstances where patentee grants a license, but in most other cases, CBI is legitimately excludable under patent law and antitrust shall never interfere. Tier II, competition by improvement, is really a higher level of CBI, in that competing technologies here involve substantial improvements on the patented technology, though still falling into the scope of earlier claims. Note a caveat here – patent law is clever enough to give certain leeway to those competitors with own substantial improvement, for example by granting a new patent right if the improvement has been found to be “inventive” or “non-obvious”, so that these competitors would have stronger bargaining power to negotiate with the original patentee and harvest higher profits from the pair of sequential innovation. This power and resulting profit would further incentivize competitors to do follow-on improvements. This is actually the scenario of blocking patents. Generally speaking, the second tier only applies to those industries that are characterized with cumulative innovation. Despite the substantial improvements, Tier II is still legitimately excludable under patent law and antitrust shall not interfere. Tier III is CBS, referring to those technologies that are capable of truly substituting the earlier patented technology to fulfill consumer needs in the market. By their very name, in most circumstances Tier III is comprised of “inventing around” technologies that do not fall into the protected scope of earlier patent, thus there exists no serious risk of being excluded by the earlier patentee unilaterally. The theoretical worry can happen however, at the rare circumstance of tier II and III intersection: occasionally an infringing improvement could be so radical that it would truly substitute the commercialized earlier patent as a break-through technology.   A real-world example here is the long deadlock between the famous Marconi diode patent and the Lee De Forest triode patent during the early development years of radio technology. This kind of radical improvement, even though literally infringing on the pioneer patent, is precisely the type of “creative destruction” Schumpeter emphasized many years ago that fuels innovation in a most powerful way – an innovator’s descendants can actually become the instruments of his destruction, yet the society benefits in the long run.

Following this train of thought, if we temporarily leave the “unfairly high patent pricing” context and turn  to some of the classical NIE (New Institutional Economics) scholarly works in a more general sense, including the conventional hypothesis of patent “anti-commons” (Heller and Eisenberg 1998), worry of over-broad pioneer patents’ suppressing effect on follow-on innovation (Merges and Nelson 1990), and recent conjecture of patent hold-up and royalty stacking in the SEP context (Shapiro and Lemley 2007), arguably they all may be considered as embodying (more or less) such theoretical worries of patent foreclosure on dynamic competition.

Inspiring as these theories were, they, need to be tested in practice.  Curiously, decades of empirical research in various industries have showed little success in proving these theoretical worries happening in the real marketplace, at least on a macroscopic level. A specific concurrent example for this striking disintegration between theory and empirical results is the robust innovation and significant consumer benefits (as indicated in quality-adjusted prices) in the SEP intensive wireless communication industry, in the shadow of persistent “patent hold up and royalty stacking” predictions. See Alexander Galetovic, Stephen Haber & Ross Levine, An Empirical Examination of Patent Holdup, Journal of Competition Law & Economics 11(3) (Aug. 2015); Jonathan Barnett, Antitrust Overreach: Undoing Cooperative Standardization in the Digital Economy, 25 Mich. Telecomm. & Tech. L. Rev. 163 (2019). There might be different explanations to this empirical difference. I give a rough quantitative analysis in my paper why the seemingly plausible theoretical worry of dynamic competition foreclosure is unlikely to happen in practice.  In addition, I argue that perhaps precisely because CBS  has always been a critical link pivoted by patent regime in its instigation of dynamic competition, throughout ages patent law has developed an implicit yet critical awareness to safeguard CBS from being foreclosed per se. This implicit awareness has been built into a wide variety of rules and principles of patent jurisprudence prevalent in many jurisdictions, such as the eligibility doctrine, inventiveness/non-obviousness and disclosure/enablement requirements in the granting phase, all-element rule and reverse doctrine of equivalents in infringement assessment phase, as well as in deeper principles underlying and threading these specific rules together. A good example for the latter is the proportionality principle well illustrated by Prof. Merges in his Justifying Intellectual Property (Chap. 6, 2011).  Surely these built-in mechanisms are not ironclad, but they may have worked in a more successful way than we give faith to, and the AML  should not overlook it.

Based on the above, I further propose that a patentee’s unilateral pricing act should be generally found legal per se in China; or at the very least, presumed legal under section 55 of the AML unless an agency/plaintiff can prove otherwise – that the alleged pricing constitutes an “abuse” in that it would disrupt dynamic competition. Specifically, a patentee’s unilateral pricing act should be immune from the “unfairly high pricing” scrutiny under Section 17(a) of the AML, unless an anti-monopoly plaintiff or enforcement agency can overcome all three following hurdles with concrete evidence, in addition to a persuading economic analysis: (i) the patentee enjoys a real dominant market position; (ii) such pricing constitutes a de facto refusal to deal, and (iii) the refusal would likely foreclose the very type of competition patent law has aimed to promote, i.e. dynamic competition. More specifically, the harm to dynamic competition can be proven in either one of the following two dimensions: i. In the same market of patented technology, a monopolist patentee’s constructive refusal to license would render radical all follow-on substitutes impossible to be developed (restriction on competition by substitution); or ii. In the circumstances of a vertically integrated monopolist patentee, the constructive refusal to license would foreclose the competition or subsequent innovation in downstream markets in which the patentee also competes.

True this burden seems high, but it is justified by three cumulative resources, (i) the above economic insights into the intersection of patent and antitrust; (ii) the prevalent non-interventionist attitude toward “excessive patent pricing” in sister jurisdictions; and (iii) the inevitable limitations of antitrust law, manifested in the error costs due to lack of proper information and economic analysis methodologies on dynamic efficiency. See Frank H. Easterbrook, The Limits of Antitrust, 63 Tex. L. Rev. 1 (1984).

Despite existence of formal empowerments in some countries/regions, almost all jurisdictions with well-developed antitrust jurisprudence have exercised a very cautious attitude in condemning a market price as “unfairly high” in practice. Section 17(1) of the AML does not have a counterpart in conventional US antitrust jurisprudence, and despite the existence of a theoretical counterpart in EU (Art. 102(a) TFEU), it has been rarely invoked. When it comes to patent pricing, the EC has been taking a virtually non-interventionist approach. This non-interventionist approach may have originated from the Commission’s awareness that many general objections against exploitative excessive pricing actions, such as the danger of undermining investment incentives of new entrants as well as dominant firms, difficulty in assessing “excessiveness”, risk of improper price regulation, and undue space for political rent seeking etc., are particularly true in the context of patent-intensive innovative industries.

Antitrust is costly. As Judge Easterbrook pointed out many years ago, in reality judges and enforcement officials are always equipped with imperfect information about actual effects of the accused practice, and such costs of information and their corresponding actions are the limits of antitrust. Part of these costs comes from the judicial ignorance and inhospitality against business practices. Very often, if a poor defendant in an antitrust case cannot convince the judge that his practices promote competition, he is doomed. Unfortunately, “the gale of creative destruction produces victims before it produces economic theories and proof of what is beneficial.” Consequently, this unfortunate judicial inhospitality and ignorance would inevitably generate substantial positive costs in practice.

During the transition from  a planned regime to market economy, China needs to overcome all kinds of obstacles, ranging from formal restraints in laws and institutional infrastructure, to lingering outdated theories and prejudice that die hard and can still be quite powerful impeding the progress. On an institutional level for example, varied degrees of inertia may unavoidably exist on anti-monopoly agencies’ enforcement philosophy, especially when it comes to high pricing acts that seem to harm consumer interest (albeit short-term) on its face, and which had long been the subject of pervasive regulation through local pricing bureaus, even prior to the existence of the antimonopoly law.  In view of this potential institutional inertia, the likelihood of inhospitality a monopolist patentee, domestic or international, encounters in “unfairly high pricing” cases could be substantial.

Antitrust has two major analysis modes: per se rule and the rule of reason. If equipped with thorough information of market practices and perfect analysis methodologies, rule of reason is the route to precision and ultimate truth. Unfortunately, as discussed above, that is not the case in reality. One may contend that during the several decades after Judge Easterbrook’s seminal writing, rapid development of economic theories have provided more substantial guidance in many areas, but I am still reluctant to say that the improvement has been so significant to render his insight obsolete, particularly in context of dynamic efficiency and IP related issues. As such, an ambitious rule-of-reason framework as embodied in Section 14 of the Anti-monopoly Guidelines for IP Abuse, would inevitably generate significant error costs despite its good-willed intention.

Admittedly, our presumed legal per se framework is not cost-free. It may be conceivable that in exceptional circumstances, a monopolist patentee’s excessive pricing would disturb dynamic efficiency yet escape the law because the plaintiff simply cannot meet the high burden. On a systematic level however, I believe this possible negative error would be at least offset by the significant positive error costs avoided. A per se rule has always been used to condemn (or excuse) whole categories of practices, even though some of them are actually beneficial (or evil), and one cannot have the savings of decision by a particular rule without accepting its cost of errors.  When we choose which analysis mode to go, what really matters is the overall probability. In summary, we presume patent pricing to be legal per se, because both economic insights and comparative law already showed us that the happening chance of a real exclusionary excessive patent pricing would be extremely low (roughly estimated to be 1/100,000 – 1/100M in the paper), and partially confirmed by empirical studies so far. The exact reasons can be further explored, but the bottom line is clear – it seems that in a vast majority of circumstances, again the market mechanism coupled with strong patent protection has been functioning adequately well to facilitate innovation. Facing this extremely low probability of real foreclosure on dynamic competition, it would be unwise in every individual case to incur enormous administrative and error costs only to search a mere possibility. In a brief conclusion, if either way we are destined to make mistakes, we naturally choose the side with less cost.

In contrast to private law, anti-monopoly law is a much stronger form of interference with the market by government. Perhaps too many people today have omitted this (not only in China) – a blunt instrument as it is, antitrust law acquires legitimacy only in a minority of cases where market failure really happens, rather than a mere theoretical possibility. “The history of Chinese economic reform has clearly told us, whenever a market-oriented policy became dominant and market mechanisms were more frequently used to allocate resources, the quality and speed of Chinese economic development was better.”  Wu Jinglian, The Economic Development of China (The Great Encyclopedia Press 2018), at p. 3 (translation is my own). On the contrary, every time Chinese economic policy was influenced by the theories of planned economy, “both macroeconomic risks and microeconomic interests were affected deleteriously… … Therefore, to resolve the many problems we are confronted with during the economic reform process of China, the only answer is to insist and deepen our reform centering upon the confidence on market economy and rule of law, and further use the market mechanism to allocate resources; it should never be the pursuit of more state interferences.” Id. When stepping into the deep water of reform today, China should learn its historical lesson and be especially cautious with those legal instruments that potentially interfere with price mechanisms, the core feature of a market economy.

The author wishes to thank Hon. David Kappos, Prof. Robert Merges, Chief Judge Randall R. Rader (retired), and Prof. Mark Cohen for their editorial suggestions.  The opinions expressed herein are the author’s own.