無歧視:通信標準必要專利許可(上)

《無歧視：通信標準必要專利許可》目錄

 

一、   標準必要專利許可歧視現象和問題

(一)  愛立信向TCL要約的專利許可費率是歧視性的

(二)  愛立信對中國的專利許可費率不公平，歧視的問題更加嚴重

(三)  法院對「無歧視」的理解仍然充滿爭議

(四)  印度競爭委員會的初步決定認為愛立信的許可費率是歧視性的

(五)  專利許可中的歧視行為極為嚴重和普遍

(六)  開風氣之先，愛立信擬按每臺設備收取固定金額專利許可費

二、   反壟斷法/競爭法所定義的歧視行為

(一)  我國專利法和反壟斷法均明確禁止實施歧視性許可，各國普遍認為智慧財產權應當受反壟斷法/競爭法對歧視行為同樣的規制，對標準必要專利歧視性許可的規制應當更加嚴格

A. 反壟斷法和專利法均明確禁止歧視行為，行使智慧財產權過程中的歧視行為沒有、也不應該按例外情況對待

B. 普遍認為對智慧財產權行使行為進行反壟斷規制應採用與其他財產性權利相同的規制標準

C. 專利法的宗旨和目的禁止歧視性許可，該行為應當受反壟斷法/競爭法對歧視同樣的規制

D. 對標準必要專利歧視性許可的規制應當更加嚴格

(二)  中國、美國和歐盟反壟斷法/競爭法規制歧視行為的條文

A. 無「歧視」的法律規則比「公平」、「合理」的要具體明確

B. 中國《禁止濫用市場支配地位行為暫行規定》第十九條

C. 歐盟Section 102(c) of the TFEU

D. 美國反壟斷法對專利許可歧視行為的規制歷史悠久、案例豐富、學術探討充分，引領世界潮流

E. 美國section 2(a) of the Robinson-Patman Act對歧視行為規制的原則和方法對於專利許可具有重要指導作用，它已經通過§ 5 of the FTC Act適用於執法和司法實踐

F.  歧視行為的法律定義：條件相同的交易，差別待遇

G. 對條件相同的交易實行差別待遇即構成歧視行為。不必考慮對競爭的影響或者是否存在正當理由，更不必考慮交易相對人自身的特點

H. 歐盟與美國反壟斷法/競爭法對歧視行為規制的側重點不同

(三)  條件相同的交易

A. 我國「條件相同的交易相對人」與美國goods of like grade and quality、歐盟equivalent transactions with other trading parties的含義應協調一致，應按「條件相同的」「交易」的「相對人」而不是「條件相同的」「交易相對人」來理解

B. 在標準必要專利許可違反「無歧視」承諾的合同之訴中，「similarly situated firms」含義與上述反壟斷法/競爭法的「條件相同的交易相對人」、goods of like grade and quality、equivalent transactions with other trading parties也相同

三、   標準必要專利許可中的若干歧視行為

(一)  按整機售價百分比計算許可費

A. Apportionment

B. Smallest Salable Patent-Practicing Component

C. Entire Market Value Rule

D. Hold-Up The Profit On That Complementary Good

(二)  Level Discrimination

(三)  Volume or Quantity Discounts

A. Specific Cost Savings Resulting From The Larger Purchases

B. Equally And Practically Available To All Purchasers

(四)  Loyalty Rebates

(五)  Cross-Licenses

四、   標準必要專利許可歧視行為對競爭具有不利影響

(一)  專利許可中的歧視行為屬於Secondary-line Injury 和Input Market Discrimination，其對競爭具有嚴重不利影響

A. A Price Difference Not Justified By A Difference In Cost May Distort Competitive Relationships And Impair Efficiency At The Buyer Level.

B. Price Discrimination Is A Condition Fostering And A Symptom Of Monopoly Or Cartel Pricing At The Seller Level, Price Discrimination Is Seen As Aggravating Rather Than Alleviating The Monopoly Problem.

C. 價格歧視對利潤和財富分配的惡劣影響

D. 中小企業應當受到特別保護

E. Divide-and-Conquer Strategies

(二)  Secondary-Line Injury推定具有反競爭效果

A. Secondary-Line Injury

B. Morton Salt Inference: Substantial, Sustained Price Differences Between Competing Customers May Justify The Inference Of The Requisite Injury To Competition

C. Direct evidence

D. 我國通訊設備製造業利潤微薄，競爭慘烈，加強對標準必要專利許可中歧視行為的規制，關乎電子信息領域中國製造的興衰存亡、關乎創新驅動戰略能否實現，刻不容緩

E. 原告證明對個別競爭者造成傷害即可，無需再證明對整體競爭的傷害

(三)  中國反壟斷法對歧視行為採用本身違法原則

A. 採用本身違法原則，無需單獨證明排除、限制競爭效果

B. 相反，有觀點認為應當根據合理性原則判斷，並且反競爭的標準應提高至「顯著」程度

(四)  Secondary-line Injury完全有可能是強勢標準實施者推動的，以謀求相對於弱小標準實施者的競爭優勢；標準必要專利權人是否有激勵進行價格歧視、是否有市場支配地位不是反競爭效果的必要條件

A. Pivotal Licensee Capture

(五)  Input Market Price Discrimination

A. Input Market Price Discrimination對競爭具有顯著不利影響

B. 縱向一體化的壟斷者具有實施Input Market Price Discrimination扭曲下遊產品市場競爭的激勵，中美歐均嚴厲規制

(六)  Burden-Shifting

(七)  法律經濟學分析

A. 經濟學分析應用於具體法律案件容易引起嚴重問題

B. 經濟學對歧視行為反壟斷法/競爭法規製造成了嚴重幹擾

五、   標準制定行為對競爭具有不利影響，該不利影響集中體現在標準必要專利許可中的歧視行為。為限制該不利影響，必須嚴格執行「無歧視」承諾。違反「無歧視」承諾，必然對競爭具有不利影響，無需再另外證明

(一)  標準制定行為對競爭具有不利影響，該不利影響主要體現在標準必要專利許可時索要過高的和/或歧視性的許可費

A. 標準制定行為排除、限制競爭，專利權人因此獲得了或者顯著增強了市場支配地位；而標準實施者被鎖定於標準必要專利所覆蓋的技術，面臨著來自上遊標準必要專利權人和同行標準實施者兩個方面的更為不利的競爭處境

B. 專利劫持

C. 因標準制定行為而獲得或增強的那部分市場支配地位，不是專利法激勵創新所必須的，並且危害他人創新，違背專利法和反壟斷法目的

D. 專利劫持主要是傷害最終消費者

E. SSOs Cannot In General Be Counted On To Adopt Effective FRAND Policies

F.  許可費堆疊

(二)  為緩解標準制定行為對競爭的不利影響，標準化組織和標準必要專利權人必須對標準必要專利許可設定必要、合理的約束機制。FRAND承諾作為僅有的約束機制，對它的違反必然對競爭造成不利影響

A. 標準制定與標準必要專利許可是專利權人主導的整體連貫行為，其排除、限制競爭效果應當整體考量

B. 必要的約束機制

C. 「無歧視許可」承諾

D. 應當嚴格理解「無歧視許可」承諾：以本身違法為背景，以不存在Less Restrictive Means為標準

(三)  違反「無歧視」承諾的合同之訴，以及與構成歧視行為的反壟斷法/競爭法控訴的異同

A. 合同之訴與反壟斷之訴

B. 在合同之訴中，Birss錯誤地將對競爭的傷害引入「無歧視」承諾

C. 即便從反壟斷/競爭法對歧視的規制的角度看，Birss對secondary-line injury所要求的「對競爭的傷害」的證明標準也過於苛刻；與Selna截然不同，後者直接推定

D. 歐洲法院對secondary-line injury所要求的「對競爭的不利影響」標準發生了很大變化

E. Birss所要求的「實質性」差異標準是錯誤的

六、   強迫性保密協議掩蓋了歧視行為，對競爭具有不利影響，應當嚴格規制；它也違反「公平合理無歧視許可」承諾

七、   標準必要專利許可歧視行為缺乏正當理由

(一)  正當理由的證明責任由被告/被申請人承擔

(二)  Cost-Justification Defense

(三)  The Changing-Conditions Defense

(四)  Functional Discounts

(五)  The Meeting-Competition Defense

(六)  中國法律規章中差別待遇的正當理由

八、   展望

附錄：

A.       Group Buying

B.       美國Robinson-Patman Act具體條文：

 

二  競爭法所定義的歧視行為

 

「歧視行為」的準確定義，對於反壟斷法/競爭法規制，和違反「無歧視」承諾的合同之訴都具有非常重大的價值。確實，「歧視行為」概念的爭議，嚴重製約了相應法律條文發揮其應有作用。我國立法、司法和學術上都有將交易相對方的自身特徵等引入歧視的定義中的錯誤觀點，這引起了歧視行為、合理理由和對競爭的不利影響等概念的混亂，並將擾亂反壟斷法/競爭法對歧視行為的規制結構。故此，本章擬釐清歧視行為的概念。

 

（一）競爭法對歧視行為實施同樣的規制，對標準必要專利歧視性許可的規制應當更加嚴格

 

有人主張：既然法律允許專利權人不許可他人、並且可以要求未經許可的實施者停止侵害行為，那麼專利權人在發出許可的時候當然就可以談任何價碼了，包括價格歧視，因為許可他人好歹增加了競爭者的數量，總比不許可任何人「反競爭效果」要小一些，所以應當允許歧視性的許可。筆者認為這種觀點過於簡單和絕對，理由如下：

 

A.      反壟斷法和專利法均明確禁止歧視行為，行使智慧財產權過程中的歧視行為沒有、也不應該按例外情況對待

《專利法》第四十八條第二項明確規定：專利權人行使專利權的行為被依法認定為壟斷行為，為消除或者減少該行為對競爭產生的不利影響的，可以給予實施發明專利或者實用新型專利的強制許可。《反壟斷法》第六條、第十七條第六項、第五十五條和第三條禁止濫用智慧財產權，對條件相同的交易相對人實行差別待遇，這構成壟斷行為；《反壟斷法》第五十五條規定：「經營者依照有關智慧財產權的法律、行政法規規定行使智慧財產權的行為，不適用本法；但是，經營者濫用智慧財產權，排除、限制競爭的行為，適用本法。」

若專利權人主張雙方是平等自願協商達成許可的，不存在「濫用」情節，則其首先需要證明被許可人具備完全信息並具備與專利權人相匹敵的談判地位，這在對歧視行為的規制中非常重要，詳見第五、第六章。

 

B.      普遍認為對智慧財產權行使行為進行反壟斷規制應採用與其他財產性權利相同的規制標準

根據U.S.Department of Justice 與 the FederalTrade Commission在2017年聯合發布的Antitrust Guidelines for the Licensing of Intellectual Property，其General Principles第一條即是：The Agencies apply the same generalantitrust principles to conduct involving intellectual property that they applyto conduct involving any other form of property.Theexercise of intellectual property rights is thus neither particularly free fromscrutiny under the antitrust laws, nor particularly suspect under them. [1]

 歐盟也闡明了同樣的原則。

我國《關於濫用智慧財產權的反壟斷指南(徵求意見稿)》第一條「執法原則」即為：「對智慧財產權行使行為進行反壟斷規制，採用與其他財產性權利相同的規制標準，遵循《反壟斷法》的基本分析框架，同時考慮智慧財產權的特點。」

 

C.       專利法的宗旨和目的禁止歧視性許可，該行為應當受反壟斷法/競爭法對歧視同樣的規制

專利權具有法定性，只有法律明確授予了專利權人的才屬於其權利範圍。另外，除非法律已經明確做出例外規定，其他法律當然也規制專利權的行使，作為市場經濟憲法的反壟斷法當然也規制著專利權的行使。

《專利法》第一條開宗明義：「為了保護專利權人的合法權益，鼓勵發明創造，推動發明創造的應用，提高創新能力，促進科學技術進步和經濟社會發展，制定本法。」筆者的理解是：專利法從社會整體而不是個別專利權人的角度，以「促進科學技術進步和經濟社會發展」為宗旨，「用」專利技術而不是「不讓用」是專利法的目的；專利法賦予專利權人的權利，是以法律有明確規定為限，不得濫用，《反壟斷法》第五十五條規定：「經營者濫用智慧財產權，排除、限制競爭的行為，適用本法。」

對於專利權人與實施者生產不同商品的情況，因為在實施者的商品市場上，專利權人未充分實施其專利，實施者應當在合理條件下獲得許可，否則可以申請強制許可。

在專利權人自己生產、使用的情況下，專利權通常允許專利權人不許可他人生產、使用專利技術；專利權人當然也可以許可他人使用，不過，專利法第六十條、第六十五條規定了許可費和賠償額的確定方法，並沒有授予專利權人單方面漫天要價、為所欲為的權利，相反，專利法賦予了實施者確定公平價格的途徑；最重要的是，這些情況都要受到關於構成壟斷或使用不充分的強制許可的限制。

對於專利權人與實施者或者數個實施者都生產相同商品或使用相同方法，彼此處於競爭關係的情況，反壟斷法禁止歧視行為。下文對Secondary-LineInjury的論述，應予以重視。

 

D.     對標準必要專利歧視性許可的規制應當更加嚴格

標準必要專利的專利權人做出了「公平合理無歧視許可」的承諾，無論從合同法、反壟斷法、禁反言或者誠實信用原則等任何角度出發，專利權人都沒有拒絕許可或進行歧視性許可的權利了，無論其原本是否曾有過這樣的權利。例如，Level discrimination，詳見下文論述。具體理由見第五章。

 

（二）中國、美國和歐盟反壟斷法/競爭法規制歧視行為的條文

 

A.       無「歧視」的法律規則比「公平」、「合理」的要具體明確

「[F]air and reasonable」 may cover a range, but if one firmis at the top end of the range and another is at the bottom, the discriminationbetween the two is problematic for the firm receiving the less favorable terms.In this respect, firms may care more about non-discrimination than aboutreasonableness. That is, all firms might be paying unreasonably high rates, butat least they would not be competitively disadvantaged relative to one another.[2]

「合理」主要是商業判斷，法律只能界定一個大致的範圍。「公平」是終極性的法律追求，涉及到太多複雜問題。相對來說，反壟斷法/競爭法對「歧視」的規制要明確具體得多，什麼構成「歧視」、怎樣算對競爭具有不利影響，法律和案例提供了具體明確的指引。反壟斷法和專利法還明確禁止智慧財產權許可中的歧視行為。認為「無歧視」的承諾沒有明確內涵的說法，是難以成立的。

反壟斷法之外，中國明確規範「歧視」行為的法律規章數以百計，憲法、刑法、勞動法、水電煤氣等公用事業管理法律規章均有明確規定。

難怪有學者認為：nondiscrimination can provideprotection against strategic action and distortions of competition and should,in principle, be simpler to implement than an ex ante hypothetical negotiation.[3]My proposal calls for a shift of emphasis from the 「fair and reasonable」 prongof FRAND, which is often inherently ambiguous, to the 「non-discrimination」prong, which if clearly defined can provide meaningful protection against expost holdup if bilateral negotiations between rights holders and industrymembers occur before firms and consumers make investments that are specific toa standard.[4]

 

B.      中國《禁止濫用市場支配地位行為暫行規定》第十九條

《禁止濫用市場支配地位行為暫行規定》第十九條明文禁止具有市場支配地位的經營者沒有正當理由，對條件相同的交易相對人在交易條件上實行差別待遇。　　

因此，我國智慧財產權許可歧視行為的構成要件包括：

1)   許可人具有市場支配地位；

2)   對條件相同的許可交易；

3)   實行差別待遇；

4)   沒有正當理由。

其中，第四項要件「沒有正當理由」，應當首先由被申請人/被告承擔證明具有「正當理由」的責任；然後由申請人/原告有針對性地進行逐條反駁；然後由法院以是否存在「LessRestrictive Means」為標準判斷該歧視行為的必要性，從而判斷正當理由是否成立。

中國反壟斷法的一個不足之處在於：僅明確規制具有市場支配地位的企業濫用其市場支配地位的歧視行為，沒有對於強勢買方要求歧視性優惠而對其與其他買方的競爭關係造成不利影響的歧視行為做出明確規定。應借鑑美國Robinson-PatmanAct做出相應補充，詳見下文Secondary-Line Injury部分。

 

C.      歐盟Section 102(c) of the TFEU

In the European Union, price discrimination by a dominant firmis forbidden under section 102(c) of the TFEU: Any abuse by one or moreundertakings of a dominant position within the internal market or in asubstantial part of it shall be prohibited as incompatible with the internalmarket in so far as it may affect trade between Member States. Such abuse may,in particular, consist in: (c) applying dissimilar conditions to equivalenttransactions with other trading parties, thereby placing them at a competitivedisadvantage. [5]

The CJEU clarified that the prohibition of discriminationapplies not only if the dominant undertaking discriminates against its own(actual or potential) competitors but also if discrimination is carried outvis-à-vis its commercial counterparts in upstream or downstream markets. [6]

The CJEU explained in Huawei v. ZTE that 「an undertaking togrant licenses on FRAND terms creates legitimate expectations on the part ofthird parties that the proprietor of the SEP will in fact grant licenses onsuch terms.」 The Court went on to find that, given these legitimateexpectations, a refusal by a patent holder to grant a license on FRAND termscould constitute an abuse under European competition law (Article 102 of theTFEU).

 

D.     美國反壟斷法對專利許可歧視行為的規制歷史悠久、案例豐富、學術探討充分，引領世界潮流

A patent license may be found to violate the antitrust lawsor constitute a patent misuse if the patentee (1) licenses or offers to licenseat unequal rates, (2) to similarly situated licensees, (3) without a validbusiness reason for the unequal rates, and (4) the license rates are asufficiently important production expense that the unequal rates cause asubstantial adverse effect on competition. [7]

In the mid-1960s, a series of decisions involving the shrimpprocessing industry announced the disturbing proposition that a patent ownerwho negotiates different royalties to different licensees, and therebyadversely affects competition between the licensees, commits patent misuse andmay even incur antitrust liability.[8]The cases concerned the same basic set of facts. The Peelers Co. Owned valuablepatents on machinery used in the peeling and processing of shrimp. A companywith which it was affiliated, Grand Caillou, was engaged in the business ofprocessing shrimp along the Texas Gulf Coast, a region that had begun toreceive heavy competition from the fledgling Pacific Northwest shrimp industry.The Peelers Co. Licensed processors in both regions, but charged theNorthwestern users a significantly higher royalty than that charged GrandCaillou and other Gulf Coast users.

In an administrative proceeding challenging the arrangement,the Federal Trade Commission held that the patentee’s discriminatory royaltypractice violated § 5 of the FTC Act. The Commission relied upon a conspiracytheory in which it reasoned that the real 「purpose」 underlying thediscriminatory rate was to 「protect and foster」 Grand Caillou’s interests as ashrimp canner by having Peelers issue licenses 「inhibiting the shrimp cannerspacking the … Shrimp of the Northwest.」 On appeal, the Fifth Circuit adoptedCommissioner Elman's argument that the case turned on 「the duty of a lawfulmonopolist to conduct its business in such a way as to avoid inflictingcompetitive injury on a class of customers.」

Applying this standard to the facts, the court concludedthat the evidence adequately demonstrated Peeler’s exercise of its patentmonopoly in such a manner as to cause substantial competitive injury to competingusers of its patents. Similar reasoning was employed by the other courts thataddressed the basic Peelers facts, resulting in findings of either patentmisuse, antitrust violations, or both[9].

FTC had brought average of 40 Robinson-Patman cases per yearfrom 1937 to 1971, but after 1980, FTC instituted only 1-2 cases per year.[10]從1982年1月1日到2010年3月31日, 美國法院共收到345件Robinson-Patman私人訴訟，其中268件屬於secondary line price discrimination.[11]Private litigation continues unabated, Salop and White(1988) estimate that 18.1percent of all private federal antitrust suits filed between 1973 and 1983 involvedRobinson-Patman claims.

From World War II through the 1970s, courts issued more thanone hundred decrees ordering patent holders to license their patents on termsthat were fair, reasonable and non-discriminatory. These orders were typicallyissued as remedies in antitrust cases involving the perceived abuse of patents,the licensing orders arising in these cases shed much-needed light on questionsonly now re-emerging in the FRAND debate. Despite the differing origins ofpatent licensing orders (remedies for anticompetitive conduct) and FRANDcommitments (voluntary promises), the large body of interpretation and analysisof the content and procedural elements of these historical patent licensingorders by courts, agencies, and private firms can help to illuminate theinterpretation and enforcement of their close kin, FRAND commitments, today. Theredo not appear to be many, if any, remedial patent licensing orders entered inantitrust cases after 1975. This trend is consistent with the overall declinein antitrust enforcement that began in the late 1970s with the rise of theChicago School law and economics movement and the publication of Robert Bork’s influentialtext, The Antitrust Paradox.[12]

 These actions soughtto break up arrangements among groups of competitors that tended to restrictentry to markets and otherwise distort competition and to open the playing fieldto all market participants. For example, in U.S.v. Hartford-Empire, the Supreme Court approved an order requiring a groupof competitors in the glassmaking industry to charge 「uniform reasonableroyalties … Without discrimination or restriction」 on licenses of their glassmakingmachinery patents.[13]

Hartford-Empire had a lasting impact on the remedial decreesissued by courts in antitrust and other cases through the 1970s. In United States v. National Lead Co., thedefendants were found to have violated Section 1 of the Sherman Act through acomplex patent pooling arrangement in the market for titanium compounds andpigments. As a remedy, the district court ordered the defendants 「to grant toany applicant … A non-exclusive license under any or all of the patents … At a uniform,reasonable royalty.」[14]

 The court inRudenberg offered an insightful rationale for the 「nondiscrimination」 componentof FRAND commitments that in seeking the commitment from Dr. Rudenberg, 「[t]heAttorney General undoubtedly was seeking to further what he might havedescribed as an open-door policy assuring equality of opportunity to all whomight have use for inventions disclosed in patents. That policy was intended togive the same chance to all who are or may be in competition regardless ofwhether they have or lack large funds and influential connections.」[15]

 

E.     美國section 2(a) of the Robinson-Patman Act對歧視行為規制的原則和方法對於專利許可具有重要指導作用，它已經通過§ 5 of theFTC Act適用於執法和司法實踐

In the United States, the price discrimination prohibition isfound in section 2(a) of the Robinson-Patman Act:

It shall be unlawful for any person engaged in commerce, inthe course of such commerce, either directly or indirectly, to discriminate inprice between different purchasers of commodities of like grade and quality … Wherethe effect of such discrimination may be substantially to lessen competition ortend to create a monopoly in any line of commerce, or to injure, destroy, orprevent competition with any person who either grants or knowingly receives thebenefit of such discrimination, or with customers of either of them.

Businesses should keep in mind some of the basic practicesthat may be illegal under the Act. These include: [16]

Price differences in the sale of identical goods that cannotbe justified on the basis of cost savings or meeting a competitor's prices; or

Promotional allowances or services that are not practicallyavailable to all customers on proportionately equal terms.

雖然美國聯邦貿易委員會官方網站表示: 「TheRobinson-Patman Act applies to commodities, but not to services, and topurchases, but not to leases.」 [17]但是，什麼構成歧視、什麼是歧視的反競爭效果等等，不會僅僅因為涉及了專利就截然不同。更進一步，同一件專利技術本身嚴格滿足of the samegrade and quality，而不像有形商品一樣容易引進差別特徵，該專利許可中的歧視行為更應該嚴格限制，尤其是當該技術使用於與同一標準相兼容的同種商品的時候。法律經濟學權威Posner認為這是 「an artificial and seemingly irrationallimitation on its scope.」 [18]加拿大競爭法報告也認為「The exclusion … Of transactions otherthan sales of articles is an apparent anachronism which fails to take intoaccount … The enormous and growing importance of transactions involvingservices and intellectual property rights.」[19]如本文第一章第四節所介紹，印度根據Section 4 in the Competition Act, 做出決定，認為愛立信的專利許可費率是歧視性的，但該條文同樣只是針對「purchaseor sale of goods or services」。[20]可見將規制歧視的一般條款適用於智慧財產權許可，在全球範圍內具有普遍共識。21世紀的中國反壟斷執法、立法和司法在規制濫用智慧財產權排除限制競爭行為時，完全可以借鑑Robinson-PatmanAct的原則和方法。

 

F.      歧視行為的法律定義：條件相同的交易，差別待遇

上述我國、歐盟和美國反壟斷法/競爭法條款均沒有就「歧視行為」給出明確定義，而是直接指明了其兩項要素：條件相同的交易，差別待遇。只要滿足該兩項要素，即受到各該條款的規制。學術界公認此即對歧視行為的規制。

In FTC v.Anheuser-Busch Inc., the Supreme Court held that 「discrimination」 is nothing more than a 「difference」 in price.對這句話的理解，請見下文。

 

G.     對條件相同的交易實行差別待遇即構成歧視行為。不必考慮對競爭的影響或者是否存在正當理由，更不必考慮交易相對人自身的特點

判斷是否存在反壟斷法/競爭法下的歧視行為，著眼點是價格信號、市場競爭機制和交易結構本身；一旦存在差別待遇，只要交易針對同類標的或者交易相對人彼此之間存在競爭關係即可，無需考慮該行為對競爭的具體影響或正當理由，更不應當考慮交易相對方自身具體特點所帶來的影響；至於該行為對競爭是否具有不利影響，是否存在正當理由，或者因為交易相對方自身的具體特點等原因該行為是否不被反壟斷法/競爭法所禁止等，在「正當理由」、「對競爭的影響」等法律評價環節再逐步考慮，以在清晰準確的概念基礎上，構建起穩定嚴謹的法律適用和推理的流程與結構；尤其不能糾結於具體交易相對人的自身特點，因為這樣的自身特點實在是千差萬別、錯綜複雜。這樣才能理解美國最高法院在 FTC v. Anheuser-Busch Inc.一案中的良苦用心, theSupreme Court held that 「discrimination」is nothing more than a 「difference」 in price. It is not that inquiries aboutthe seller’s intent or the effect of the discrimination on competition areirrelevant; rather, these inquiries are relevant later, in determining whetherthe discrimination is one that is unlawful.Mr.Chief Justice Warren said that 「we now conclude … [that] it is only by equatingprice discrimination with price differentiation that §2(a) can be administered asCongress intended.」 The contrary view, requiring an analysis of the effect oncompetition at the 「discrimination」 stage, was deemed to be unworkable andinconsistent with the purpose of the Act: 「『Inevitably every legal controversy overany price difference would shift from the detailed governing provisions—「injury,」 costjustification, 「meeting competition,」 etc.—over intothe 「discrimination」 concept for ad hocresolution from specifically pertinent statutory text.』」[21]法律經濟學權威Posner也承認：「The termdiscrimination in section 2(a) has been interpreted to mean a difference inprice.」 [22]

經濟學在定義價格歧視時也沒有考慮其對競爭的影響或者交易相對方自身的具體特點，Pricediscrimination is defined as selling the same product to different customersfor different prices when those differences are not justified by differences inthe cost of production or supply.[23] 

a)        概念不清晰將擾亂法律適用和推理的流程與結構

例如，有人認為：Apurchaser does not take advantage of a lower price or a discount which isfunctionally available on an equal basis, it has been held that either no pricediscrimination has occurred, or the discrimination is not the proximate causeof the injury. The existence of a functionally available lower price defeatsthe requirement that there be a discrimination in price in order to establish aprima facie price discrimination case. In this sense, the doctrine offunctional availability really does not operate so much as a defense to an otherwisevalid claim of price discrimination; rather, the existence of the functionallyavailable lower price prevents the plaintiff from establishing one of thenecessary requirements of a price discrimination claim.

在MathewEnterprise, Inc. V. Chrysler Group LLC案件中，JuryInstruction 23: 「In regard to element 3, discrimination in price, Stevens Creekmust prove… (2) that the lower prices were functionally unavailable to StevensCreek…In order for Stevens Creek to prove unlawful price discrimination, itmust prove that the incentive program was functionally unavailable to it. Thisrequires Stevens Creek to prove that the incentives would not have beenpractically available had Stevens Creek engaged in commercially reasonableefforts to achieve them. 」

 

一審法院認為「Anaffirmative defense is an assertion raising new facts and arguments that, iftrue, will defeat plaintiff’s claim, even if all allegations in [the] complaintare true.」In contrast, a defense that points outa defect in the plaintiff’s prima facie case is not an affirmative defense.

一審法院認為Becausecourts routinely recognize that the functional availability doctrine 「negates」two essential elements of an RPA claim, the Court cannot agree that thedoctrine is an affirmative defense. 並引用如下案例：「[I]f thelower price would have been available to the nonfavored buyer in the samemarket where the favored buyer made his purchase, the probability ofcompetitive injury due to the fact that the nonfavored buyer paid more for theproduct is not the result of price discrimination, but of the nonfavoredbuyer’s failure to take advantage of the opportunity, equally available to him,of buying at the same low prices.」[24]

 

二審法院認為In Texaco v. Hasbrouck, the Supreme Courtnoted with approval that 「the burden of proof remains with the enforcementagency or plaintiff in circumstances involving functional discounts sincefunctional pricing negates the probability of competitive injury, an element ofa prima facie case of violation.」 [25]

It would be nonsensical for the plaintiff or enforcementagency to bear the burden of proof on functional discounts at trial but notbear the burden as to functional availability, as they both operate to defeat aplaintiff’s case in chief in the same manner.

二審法院甚至認為functionalavailability essentially negates the discrimination element of the plaintiff’s§ 13(a) claim.

 

H.    歐盟與美國反壟斷法/競爭法對歧視行為規制的側重點不同

Article 82(c) focuses on the pricing behavior of powerfulsellers while the core concern of the Robinson-Patman Act is upon thepurchasing behavior of powerful buyers. U.S law represents an historic desireto protect small retail merchants from the competition of powerful buyers. WhenCongress enacted the Robinson-Patman Act, it sought explicitly to eliminate acompetitive advantage that it believed the large chain stores unfairly heldover the traditional smaller retailers. European law, on the other hand,appears focused on protecting initial sellers from the competition of theirpowerful rivals. [26]筆者認為美國反壟斷法對歧視行為的規制歷史悠久、案例豐富、學術探討充分，引領世界潮流，更為值得我國借鑑學習。包括歐盟也在借鑑學習美國，並對法律條文進行相應解釋：As far assecondary line injury price discrimination is concerned, Article 82(c) is theappropriate legal basis. [27]

 

（三）條件相同的交易

 

反壟斷法/競爭法對歧視行為的規制中，中國「條件相同的交易相對人」、美國「goods of like grade and quality」、歐盟「equivalent transactions with othertrading parties」以及標準必要專利許可違反「無歧視」承諾合同糾紛中的「similarlysituated firms」含義均相同。

 

A.      我國「條件相同的交易相對人」與美國goods of like grade and quality、歐盟equivalent transactions with othertrading parties的含義應協調一致，應按「條件相同的」「交易」的「相對人」而不是「條件相同的」「交易相對人」來理解

美國Robinson-PatmanAct是針對The goods must be of "like gradeand quality"的情形；歐盟section102(c) of the TFEU是針對equivalent transactions with other trading parties的情形，著眼點均是交易結構和市場競爭機制本身，交易針對同一種標的、交易相對人彼此之間存在競爭即可，而不是著眼於具體交易相對人的自身特徵。Discriminationamong licensees is a violation of FRAND licensing principles. It is incorrectto suggest that discrimination is permitted as long as a company is not「similarly situated」 to another. For example, it would not be appropriate for asmall new market entrant to face discriminatory licensing demands as comparedto larger, existing competitors, as such approaches would restrict competitionand market entry.

英國法院寬鬆界定similarlysituated firms的做法，頗為值得學習：「In myjudgment the FRAND rate ought to be generally non-discriminatory in that it isdetermined primarily by reference to the value of the patents being licensedand has the result that all licensees who need the same kind of licence will becharged the same kind of rate.」「I am not convinced that (b), the identity ofthe licensee, should be a strong factor in determining what comparables areuseful for determining the FRAND rate aside from the hard edgednon-discrimination point addressed below. FRAND is supposed to eliminate holdup as well as hold out. Different licensees will have differing levels ofbargaining power. That is another way of saying their ability to resist hold upand their ability to hold out will vary. It would be unfair (anddiscriminatory) to assess what is and is not FRAND by reference to this andother characteristics of specific licensees. In my view, it would not be FRAND,for example, for a small new entrant to the market to have to pay a higherroyalty rate than an established large entity. Limiting comparable licences tothose where Huawei or a similar company like Samsung is the licensee istherefore unjustified.」[28]

美國的界定同樣非常寬鬆。In theleading case of FTC v. Borden Co., the Supreme Court held that physically orchemically identical products sold under nationally advertised and privatelabelsare of ―like grade and quality—even if the advertised brandhas achieved significant consumer preference over the private labels. TheCourt, after examining the language and legislative history of Section 2(a),rejected any construction of the statute that would equate 「grade」 with 「consumerpreferences, brand acceptability, or what customers think of it and are willingto pay for it」, and held that both grade and quality are to be determined 「bythe characteristics of the product itself」. TheCourtsuggested, however, that the economic factors and consumer preferences inherentin the brand names could be considered under the more flexible 「injury」 and「cost justification」 provisions of the statute.

The 『ND』 of FRAND is necessary to ensure that a standardimplementer is not commercially penalized by having to pay a higher license feeto an SEP holder than other similarly-situated standard implementers with whichit competes [i]n downstream product markets (e.g. Computers, tablets,smartphones, etc.).[29]

The licensee『situation』 is then determined by a number ofcharacteristics, such as the firms』 particular use for the licensed IP (and henceits valuation of that IP).[30][C]ompeting firms are similarly situated if ex ante they expect to obtain thesame incremental value from the patented technology compared with the next bestalternative available to be incorporated into the standard. Firms in differentindustries, for example, such as a handset manufacturer and a maker of wirelessheart monitors, might make devices that obtain different incremental valuesfrom a patented technology and do not compete with one another, and thus canpay different rates under this interpretation.[31]

Transactions with the SEP holder need not be identical to besufficiently comparable for analyzing discrimination in an SEP license.[32]「Solong as the distinctions between the plaintiff and the proposed comparators arenot 『so significant that they render the comparison effectively useless,』 thesimilarly-situated requirement is satisfied.」[33]Applied to the broader circumstances of a plaintiff’s employment and that ofhis proffered comparator, a requirement of complete or total identity ratherthan near identity would be essentially insurmountable.[34]

但我國反壟斷法律規章對歧視行為的界定，都採用了「條件相同的交易相對人」這一表述，可能引發誤解。《反價格壟斷規定（徵求意見稿）》曾經對「條件相同」進行過定義，是指經營者與有關相對人交易等級和質量相同的同種商品時，在下列方面相近或相同：交易方式；交易環節；交易數量；貨款結算；售後服務等。上述定義在《反價格壟斷規定》中被刪除。《關於禁止濫用智慧財產權排除、限制競爭行為的規定》、《工商行政管理機關禁止濫用市場支配地位行為的規定》、《反壟斷法》均未對「條件相同的交易相對人」進行界定。可見什麼是歧視行為合適的比較基礎，是一個非常困難的、很有爭議的話題。《禁止濫用市場支配地位行為暫行規定》第十九條第二款規定：「條件相同是指交易相對人之間在交易安全、交易成本、規模和能力、信用狀況、所處交易環節、交易持續時間等方面不存在實質性影響交易的差別。」這一規定相比《反價格壟斷規定（徵求意見稿）》，增加了「規模和能力」、「信用狀況」等屬於交易相對方自身特點的因素，這些因素千差萬別、紛繁複雜，筆者認為如果解釋適用過程中不夠謹慎，可能就打開了潘多拉魔盒，有將對歧視行為的反壟斷法規制給一筆勾銷掉的可能性，甚至與「童叟無欺」的公道南轅北轍、與「看人下菜碟」的勢利難解難分，這當然不是立法者的本意，畢竟，大江南北的無論大超市還是小便利店，都已經明碼標價了。請揣摩下文美國法院在TCL v. Ericsson案件中對於「similarly situated firms」的分析。

我國反壟斷法對歧視行為的規制應當與上述美、歐觀點協調一致，第十九條應解釋為：「條件相同」是用於限定「交易」，而不是「相對人」的；之所以提到「相對人」，因為是他而不是「交易」承受了「差別待遇」，這僅僅是行文和修辭上的方便，不應做違背法理的解讀。

 

B.      在標準必要專利許可違反「無歧視」承諾的合同之訴中，「similarly situated firms」含義與上述反壟斷法/競爭法的「條件相同的交易相對人」、goods of like grade and quality、equivalent transactions with other trading parties也相同

The parties in UnwiredPlanet International Ltd v. Huawei Technologies Co. Agreed that the nondiscriminationrequirement of ETSI's FRAND commitment has the same basic meaning as Article102(c) TFEU. 「Both sides approached this issue [of interpreting 'nondiscrimination'in the FRAND commitment] on the basis that concepts such as similarly situatedparties, equivalent/comparable transactions, and objective justifications, werethe same under the non-discrimination limb of FRAND as they are in competitionlaw.」[35]

針對標準必要專利許可中的「無歧視」承諾合同糾紛，美國法院在TCL v. Ericsson案件中對於「similarly situated firms」採用了與反壟斷法/競爭法相同的判斷方法[36]，其含義與反壟斷法/競爭法的一致：

The Court concludes that for purposes of license comparisonsthe analysis should include all firms reasonably well-established in the worldmarket. This implies a necessarily wide spectrum, and correctly so for several reasons.First, ETSI contemplates facilitating competition in the market, particularlyfrom emerging firms. Second, excluding from the analysis the largest firms inthe market would have the effect of insulating them, and further contributingto their dominant positions, by imposing a barrier in the form of higher ratesfor those not at the top end of the market. (See TT Mar. 1, 2017, pp.171:22-173:25.) By the same token, TCL overstates the nature of the concern forsmall and medium sized firms. Third, permitting Ericsson to define similarly situated very narrowly by picking and choosing criteria with no relation to its SEPs or the FRAND commitment would effectively allow Ericsson to read the non-discrimination prong out of the FRAND commitment. In defining similarly situated firms, there is a similar threadamong all experts in that they look to firmsusing the same technology and at a similar level in the value chain.(Ordover Decl. ¶ 61; TT Mar. 1, 2017, p. 104 (Teece); id., p. 6- 7 (Kennedy).)

The Court also believes that similarly situated should be broadly interpreted because the mobile phone market has been extremely dynamic over the last decade. In 2007, thesix largest companies ranked by U.S. market share were, in order, Motorola,Samsung, LG, Nokia, Blackberry, and Apple. (Teece Decl. ¶ 163, Figure 16.)Within a decade Motorola, Nokia, Blackberry, and even Ericsson’s own handsetdivision would be shuttered or divested, events which Brismark acknowledged noindustry observer would have ever predicted. (Brismark Decl. ¶ 61.) TCL itselffirst entered the U.S. market in 2011, and within six years was the fourthlargest manufacturer in the U.S. by market share. (Cistulli Decl. ¶ 3.) Thevolatility of the handset market over last decade requires the Court toexercise a broad view of who will be similarly situated to TCL over the course ofthe five-year license which the Court adopts.

In determining which firms are similarly situated to TCL,the Court’s task is to identify other reasonably well-established firms in theglobal market. Certain factors obviously matter, such as the geographic scope of the firm, the licenses required by the firm, and a reasonable sales volume. The Court does not believe that factors such as the firm’s overall financial success or risk, brand recognition, the operating system of their devices, or the existence of retail stores have any bearing on whether Ericsson’s royalty rates for its SEPs are discriminatory.

A local king is a company that sells most or all of its devicesin a single country, often the same country where it is headquartered and manufacturesthe devices. Local kings are not similarly situated to global firms for tworeasons. First, their sales largely occur in one country, while a singlecountry will generally account for a relatively small percentage of the globalfirm’s sales. Because the global firm will be dealing with differentmarketplaces, different regulatory environments, and consumers with differenttastes and preferences, a global firm is unlikely to be similarly situated to alocal king. Second, local kings receive a different license from Ericsson. Alocal king only needs license to Ericsson’s SEPs in one jurisdiction, andEricsson is bound to limit its offer to a rate that reflects the SEP strength ofits portfolio in that jurisdiction. However, for global firms, Ericssonasserted that it provides a license at a global blended rate which averages outthe higher rates Ericsson could charge in some countries with the lower ratesit could charge in countries with weaker or non-existent patent protections.(Brismark Decl. ¶ 55.) Thus, a license between Ericsson and a local king doesnot reflect the rate that a global firm like TCL would have to pay.

標準實施者相對於專利權人的談判地位優劣、標準實施者在標準兼任產品市場的競爭地位的強弱等均不屬於「無歧視」的合理理由。

Two firms are not differently situated with respect to FRANDjust because one of the two licensees is suing the patent holder.[37]

 

 

 

 

三、標準必要專利許可中的若干歧視行為

 

本章擬考察通信標準必要專利許可中若干重要歧視行為。

It is likely that the non-discrimination analysis will inmany cases remain focused largely on monetary terms of FRAND licenseagreements, with only those non- monetary terms of particular importance to theparties and circumstances at hand entering the analysis on an as needed basis.

 

（一）按整機售價百分比計算許可費

 

Most devices using standards have far greater functionalitythan just implementation of a particular standard. If profits fromcomplementary innovations could have been achieved if some technology had beenput into the standard other than an asserted SEP, then one should be concerned aboutthe implementer being held up over those profits. In an ex ante negotiation, a patentholder’s demand to share in the fruits of a manufacturer’s innovativeactivities unrelated to the patent holder’s contributions to the standard wouldnot have succeeded. Thus, any such ex post demands are not 「fair andreasonable.」 These concerns can be addressed in part by the choice of royaltybase. In considering the combined effect of the royalty base and rate, we donot want to imply that the mathematical trade-off between them createsindifference between the two elements of the formula. Rather, when both thebase and rate vary significantly between heterogeneous products, it may becomeincreasingly difficult to determine and apply FRAND-compliant royalty rates.One method we favor for dealing with product heterogeneity takes advantage ofthe fact that standards are sometimes embedded within particular components ofdownstream products such as, for cellular standards, baseband chipsets.Imposing royalties at the component level is attractive both for its relativesimplicity and because it still rewards the innovator with increased revenuesin the event of growth of the volume of sales or if the price of the componentincreases. Another virtue is that charging a fee based on a component priceensures that the license terms can potentially be independent of the end use towhich the chipset is put. Thus, unless the price of the component varies withthe price of the mobile device, all devices will be burdened with the same licensecost. This outcome is consistent with the non-discrimination provisions ofFRAND. Whether the base used is the price of a component or the price of adownstream product, the patent holder benefits from downstream innovationbecause such innovation increases sales of devices and, therefore, ofcomponents. Thus, although a component royalty base limits the patent holder’sability to obtain part of the profits flowing to the downstream manufacturersresulting from their investments in innovation, it does allow the patent holderto benefit from the expansion in sales driven by those downstream innovations.In particular, using a royalty base of the component can help prevent hold-upwhen it is difficult to determine and apply FRAND-compliant rates across ahighly differentiated range of devices. Often, devices that incorporatesignificant differentiating IP are highly heterogeneous, while components whichimplement standardized technology are less so. With respect to the value of thetechnology, the standard is implemented in the component. Our concern aboutusing the price of the downstream device is that its market price can reflect agreat deal of value to consumers derived from other sources—value that a SEPholder can try to expropriate through holdup. That is, although some componentsof a device may work better in conjunction with a standard, many others aresimply unrelated. It makes economic sense to collect the license fees fromcomponents that implement and directly benefit from the standard. To the extentthat an end product is more valuable to consumers if it complies with astandard, this differential value is unlikely to be fully attributable to the SEPs.The portion of the incremental benefit that could be attributed specifically toa patent holder’s intellectual property is the amount above the base level thatcould be obtained by using unpatented technology in the standard. In any event,the appropriate compensation for SEPs that contribute an incrementalimprovement to a standard cannot be the entire amount that a customer would payfor the functionality of the entire standard. As a practical matter, it may beeasier to evaluate the contribution of a given piece of IP to the economicvalue of a component that embodies it rather than to the economic value of thefinal product. As a result, it could be easier to gauge the FRAND rate usingthe component, rather than the final product, as a base. In addition, scores ofstandards comprising thousands of declared SEPs are often incorporated indifferent devices. Because of the concern with royalty stacking (the Cournotcomplements problem) described later, the greater the number of SEP owners thatattempt to collect royalties on their SEPs (no matter how small theircontribution to one standard implemented in one component of the end product),the higher the total royalty burden is likely to be. This problem can be atleast partially addressed by limiting the royalty base to the component. Thisapproach is consistent with the FRAND principle that a SEP holder should earnreasonable compensation while still protecting against the exploitation ofmarket power by the SEP holder attributed to standardization. [38]

 

印度ICC的初步裁決觀點與此類似，如上文所述。

In Microsoft v. Motorola, the trial judge, Judge Robart, opined:「a patent’s royalty rate should be based on the importance of the patent to thestandard and to the implementer’s product. Under this analysis, this royaltyrate would fluctuate little, if at all, based on the end selling price of theproduct. Accordingly, if 0.8 cents per unit is a reasonable royalty rate for a$200.00 Xbox, then it should be a reasonable royalty rate for an Xbox sellingfor $400.00 that uses the patented technology in the same manner.」[39]This is a significant aspect of his ruling, as it appears to rule out percentage-basedroyalties based on the selling price of the end-user product as beinginconsistent with (his view of) RAND.

In In re Innovatio, Judge Holderman awarded damages of 9.56cents per Wi-Fi chip, rather than a percentage based damage amount.

 

A.      Apportionment

Apportionment is the requirement that a patentee『s recoveryfor infringement be based on the economic value of the claimed invention asopposed to the economic value of the product or service that incorporates theinvention. [40]

「The patentee,」 … 「must in every case give evidence tendingto separate or apportion the defendant's profits and the patentee's damagesbetween the patented feature and the unpatented features, and such evidencemust be reliable and tangible, and not conjectural or speculative; or the mustshow, by equally reliable and satisfactory evidence, that the profits anddamages are to be calculated on the whole machine, for the reason that the entirevalue of the whole machine, as a marketable article, is properly and legallyattributable to the patented feature.」[41]

Damages awarded for patent infringement 「must reflect thevalue attributable to the infringing features of the product, and no more.」[42]This principle—apportionment—is 「the governing rule」 「where multi-componentproducts are involved.」 [43]Consequently,to be admissible, all expert damages opinions must separate the value of theallegedly infringing features from the value of all other features. [44]Between 1854 and 1915, the Supreme Court decided more than two dozensignificant patent damages cases and in each, reiterated the rule the apatentee must apportion in all cases.[45]

「The essential requirement」 for reliability under Daubert 「isthat the ultimate reasonable royalty award must be based on the incrementalvalue that the patented invention adds to the end product.」 [46][Q]ualitativetestimony that an invention is valuable—without being anchored to aquantitative market valuation—are insufficiently reliable.[47]

Apportionment requires that a reasonable royalty forinfringement of a SEP be based on the value added by the SEP as distinguishedfrom the value added by the standard that adopted the claimed technology.[48]When a patented technology is included in a standard, demand for thetechnology, and, thus, the patent, naturally increases. The value of a SEPpatent, therefore, is due in part to the value of the technology and in partthe adoption of the technology as a standard. Because, under apportionmentprinciples, the SEP owner is entitled to compensation based only on the valueof the invention, the SEP owner has the burden of apportioning the value of theinvention from the value added by its adoption as a standard.[49]

Apportionment serves two purposes. First it protects aninfringer against having to pay as damages an amount exceeding its totalprofits on an infringing product as a result of the product incorporating twoor more patented inventions.74 Second, it prevents a patentee from expandingthe scope of its exclusive rights beyond those granted by the patent. If themeasure of damages be the same whether a patent be for an entire machine or forsome improvement in some part of it, then it follows that each one who haspatented an improvement in any portion of a [machine] may recover the wholeprofits arising from the skill, labor, material, and capital employed in makingthe whole machine, and the unfortunate mechanic may be compelled to pay treblehis whole profits to each of a dozen or more several inventors of some smallimprovement in the [machine] he has built. It is 「grave error to instruct ajury 『that as to the measure of damages the same rule is to govern, whether thepatent covers an entire machine or an improvement on a machine.』」 [50]

A convoyed sale is a sale made with the sale of the patenteditem while a derivative sale is one made as a result of the sale of thepatented item, but at a later time. Because convoy and derivative sales are notthemselves infringing, they generally should not be included in a royalty baseand, thus, are normally pertinent only to the royalty rate.[51]

 

B.      Smallest Salable Patent-Practicing Component

[I]t is generally required that royalties be based not onthe entire product, but instead on the 「smallest salable patent-practicingunit.」[52]

In any case involving multi-component products, patenteesmay not calculate damages based on sales of the entire product, as opposed tothe smallest salable patent-practicing unit, without showing that the demandfor the entire product is attributable to the patented feature. Therefore, GPNEmay not claim the entire accused iPhones and iPads as the smallest salablepatent-practicing units for damages purposes solely because GPNE claimed a「node」 having a processor that can perform the invented signaling steps ratherthan just the processor itself. [53]

The requirement that a patentee identify damages associatedwith the smallest salable patent-practicing unit is simply a step towardmeeting the requirement of apportionment. Where the smallest salable unit is,in fact, a multi-component product containing several non-infringing featureswith no relation to the patented feature …, the patentee must do more toestimate what portion of the value of that product is attributable to thepatented technology. To hold otherwise would permit the entire market valueexception to swallow the rule of apportionment. [54]Thesmallest salable unit must be closely tied to the patent to suffice, and furtherapportionment is required 「even when the accused product is the smallestsalable unit … if the smallest salable unit is still a multi-component productencompassing non-patent related features.」[55]When using a multi-component product as a royalty base, even if it is the smallestsalable unit, a patentee must still show that the patented feature drives demandfor the entire product.[56]「Outlook」 is the smallest saleable unit, but it has many features (or「components」). That one 「Outlook」 feature infringes does not mean that thesales of 「Outlook」 thus become the royalty base … The use of a saleable unitthat is greater than the patented feature is going to introduce Uniloc error.[57]Because expert relied the smallest saleable unit rule, EMVR did not apply; butbecause the patented technology was used in only two features, it was ancillaryto the core functionality of the console; accordingly, jury award based on expert『stestimony was excessive.[58]

 

C.      Entire Market Value Rule[59]

The entire market value rule is a narrow exception to thisgeneral rule derived from Supreme Court precedent in Garretson.[60]With respect to determining patent infringement damages, precedents from theSupreme Court and the United States Court of Appeals for the Federal Circuit donot allow consideration of the entire market value of accused products forminor patent improvements simply by asserting a low enough royalty rate. Forthe entire market value rule to apply, the patentee must prove that thepatent-related feature is the basis for customer demand.[61]

The entire market value rule allows a patentee to assessdamages based on the entire market value of the accused product only where thepatented feature creates the 「basis for customer demand」 or 「substantiallycreate[s] the value of the component parts.」[62]

Where small elements of multi-component products are accusedof infringement, calculating a royalty on the entire product carries aconsiderable risk that the patentee will be improperly compensated fornon-infringing components of that product. Care must be taken to avoidmisleading the jury by placing undue emphasis on the value of the entireproduct.[63] Disclosureof the end product's total revenue 「cannot help but skew the damages horizonfor the jury, regardless of the contribution of the patented component to thisrevenue.」 [64]Itis not enough to merely show that the disc discrimination method is viewed asvaluable, important, or even essential to the use of the laptop computer. Noris it enough to show that a laptop computer without an ODD practicing the discdiscrimination method would be commercially unviable. Were this sufficient, aplethora of features of a laptop computer could be deemed to drive demand forthe entire product.Even if the laptop computer at issuewould not be commercially viable without the patented feature, that would notbe enough to satisfy the entire market value rule; were it otherwise, theentire value of the computer could be attributed to any number of featureswithout which the computer could not be sold. [65]

The entire market value rule in the context of royaltiesrequires adequate proof of three conditions:[66]

The infringing components must be the basis for customerdemand for the entire machine including the parts beyond the claimed invention;

The individual infringing and noninfringing components mustbe sold together so that they constitute a functional unit or are parts of acomplete machine or single assembly of parts; and

The individual infringing and noninfringing components mustbe analogous to a single functioning unit. It is not enough that the infringingand noninfringing parts are sold together for mere business advantage. Notably,these requirements are additive, not alternative ways to demonstrateeligibility for application of the entire market value rule.

 

D.     Hold-Up The Profit On That Complementary Good

A patent holder may also be able to act strategically in thestandard-setting context to take advantage of firms that ex ante have the samealternatives as their rivals, but are more profitable. For example, supposethat the firms are not all identical because one firm also sells acomplementary product that uses the patented technology. Further suppose thatthe differentiated firm makes a large profit on that complementary good, andwould have made that same large profit had an alternative technology beenincorporated into the standard instead of the patented technology in question.Notice now that even if every firm’s incremental profit increases by the exactsame amount because the patented technology is used in the standard, the patentholder can extract the entire profit of the most profitable firm (that producesthe complementary good), even though the additional profit from selling thecomplementary good is unrelated to the patent. This is a form of hold-up, asdiscussed in the prior section, applied in a discriminatory way. This type ofstrategic behavior involving unequal treatment of rival firms that derive thesame incremental value from the patent is inefficient. It creates bothincentives to get patented technology adopted into standards far in excess ofits social value, and disincentives for firms to innovate when theirinnovations must use the standard, such as by developing goods complementary tothe standard.

 

（二）Level Discrimination

Level discrimination的問題在通信標準必要專利許可中非常突出，某些標準必要專利權人拒絕向晶片製造商籤訂專利許可協議，其目的主要是向不熟悉相關技術的通訊設備製造企業索取高額許可費，尤其是按照整機售價百分比索取許可費。

Most SSOs 「require that licenses be granted onnondiscriminatory terms, preventing an IP owner from closing the standard to particularcompetitors.」[67]

In 2002, for example, JEDEC’s president, John Kelly,publicly supported the 「no-refusal」 interpretation when, in an overview ofJEDEC’s patent policy submitted to the Federal Trade Commission (FTC) in itsRambus case, he said that 「[n]ondiscriminatory, in essence, means 『open to allcomers.』」 Nondiscriminatory license terms, he said, 「do not discriminateagainst any prospective licensee on the basis of corporate identity, history,demographics, etc.」[68]The TIA Guidelines offer a similar interpretation of the nondiscriminationrequirement: 「[a]n example of conduct that would constitute discrimination is awillingness to license all applicants except for competitors of the licensor.」[69]A FRAND commitment 「waives the patent holder’s right to refuse to license itsIP rights to anybody seeking such a license.」[70]

 

This general view has been supported by courts including theU.S. Court of Appeals for the Ninth Circuit, which held in Microsoft v.Motorola that the SEP holder, in its declarations to the ITU, promised to「grant a license to an unrestricted number of applicants on a worldwide, non-discriminatorybasis and on reasonable terms and conditions to use the patented material necessary」to practice the ITU standards. This language admits of no limitations as to whoor how many applicants could receive a license (i.e., it specifies an「unrestricted number of applicants」).[71]A similar commitment to universal access to SEPs was adopted by the IEEE inrecent amendments to its intellectual property policy.[72]Likewise, in its 2016 remedial order against Qualcomm, the KFTC required thatQualcomm grant FRAND licenses under its SEPs to any component manufacturerrequesting one.

In each remedial patent licensing decree that was reviewed forContreras』article,the defendant patent holders were required to grant licenses to 「allapplicants」—every person or firm that requested such a license. The reasoningbehind this requirement is straightforward: the decrees were typicallyfashioned to remedy distortions in the market caused by the defendants』anticompetitive conduct. The favored way of doing so was by removing patentbarriers to free competition through licensing to all interested parties on anon-discriminatory basis. In the decree entered in Rudenberg the courtrecognized the benefit of establishing 「an open-door policy assuring equalityof opportunity to all who might have use for inventions disclosed in patents.」 [73]

Early U.S. remedial orders such as those in TerminalRailroad and Hartford-Empire used non-discrimination mandates to ensure broadmarket access to relevant resources and patented technologies. Thus, in dozensof these orders, patent holders were required to grant licenses to 「allapplicants」 requesting such a license. This requirement arose because thecourts, and the DOJ that brought the suits, believed that open licensing to themarketplace would promote competition and remedy improper market concentration.For the reasons discussed above, it is likely that SDO FRAND commitments, inorder to achieve goals of broad adoption of standards, also seek to maximizeaccess to patents covering those standards.

 

Any firm should be able to obtain a license to SEPs coveredby a FRAND commitment. In particular, the patent holder should not be able torestrict licensing based on a firm’s location in the supply chain. We notethat, if a patent holder offers licenses to only downstream firms and refusesto license upstream firms, it presumably expects this strategy to be moreprofitable than licensing upstream or not offering a license. This scenariogives rise to a serious concern that the patent holder may be attempting, byexercising hold-up, to extract part of the downstream firms』 profits fromfunctionality unrelated to the patent holder’s patented technology. As apractical matter, it is likely to be easier for the patent holder to exercisehold-up and strategic discrimination against end product manufacturers than againstcomponent manufacturers. There are several reasons for this.

It would likely be more difficult for a judge or anarbitrator to determine whether the royalty claims of a SEP holder arereasonable when the end products are highly complex and contain multiplecomponents, many of which may contain numerous SEPs. That is, attempting todetermine the incremental value of patents for each end product manufacturerwill likely be more costly and uncertain than determining that value to anupstream component manufacturer. Similarly, it may be more difficult for an endproduct manufacturer to use comparable licenses to avoid holdup ordiscriminatory licensing terms because there is likely greater heterogeneityamong end products than among components. Finally, the risks of litigating toobtain a FRAND rate may be greater for end product manufacturers than forcomponent manufacturers because a trier of fact might be more willing to give asmall royalty on a large base (the value of the end product) than a largeroyalty on a small base (the value of the component), even when the formerresults in a higher total payment. Hence licensing upstream potentially lowersthe likelihood that the patent holder will try to exercise hold-up.

 In any event, webelieve that the FRAND commitment requires the SEP holder to license any firmusing the standard and seeking a license, regardless of the level ofproduction. This is particularly important as upstream firms may be impacted bythe SEP holders』 actions downstream. For example, if a downstream manufactureris sued for infringement by a SEP holder, the downstream manufacturer may seekindemnity from an upstream supplier. Even if it does not seek an indemnity, theupstream supplier will be impacted by the changes in demand for its products shoulda SEP holder exercise hold-up against downstream firms. If an upstream firm isdenied a license of its own, it may thus be itself impacted by a hold-up andviolation of FRAND, but have no recourse. [74]

The SEPs holder must make an 「unlimited」 number of licensesavailable. It cannot 「pick and choose」 among interested parties, licensing some(e.g., its allies) and refusing to license others (e.g., its rivals). Also, itcannot auction off a limited number of licenses to the 「highest bidders.」 [75]

 

CWA2[76]認為Any company that seeks a SEP license toimplement the standardized technology in their product should be entitled to alicense. Refusal to grant a SEP license to a requesting licensee isdiscriminatory in nature and is, therefore, inconsistent with FRAND licensingprinciples.

CWA2的CorePrinciple 2即為: A FRANDlicense should be made available to anybody that wants one to implement therelevant standard. Refusing to license some implementers is the antithesis ofthe FRAND promise. In many cases, upstream licensing can create significantefficiencies that benefit the patent holder, the licensee and the industry.

CWA2強調：On itsface, the FRAND promise does not restrict licensing to any particularsub-group, but instead is applicable to all potential licensees. Rather, aFRAND commitment is a commitment to license any potential licensee that seeks alicense.

This is not merely a contractual issue, but may also berequired by the competition and antitrust laws. If a patent holder could 「pickand choose」 potential licensees, then it could control who does and does notsucceed in the market. Such behaviours do not appear necessary to validate thelegitimate business interests of SEP owners, whereas they may significantlyharm FRAND licensing interests for companies that utilize standardizedtechnologies and consumers that rely on those technologies.

For example, the ETSI Directives state expressly that allmembers and third parties have the right, under the ETSI IPR Policy, 「to begranted licenses on fair, reasonable and non-discriminatory terms and conditionsin respect of a standard」.[77]As the European Commission has noted:

In order to ensure effective access to the standard, the IPRpolicy would need to require participants wishing to have their IPR included inthe standard to provide an irrevocable commitment in writing to offer tolicense their essential IPR to all third parties on fair, reasonable andnon-discriminatory terms. [...] FRAND commitments can prevent IPR holders frommaking the implementation of a standard difficult by refusing to license ...After the industry has been locked-in to the standard…[78]

This approach was supported by the ECJ, when it noted thatcompanies committing to offer FRAND licenses created 「legitimate expectationson the part of third parties」 that a license would be granted, and thisstatement was recently reiterated in the Commission’s SEP Communication.[79]European courts likewise have found a discrimination inherent in an SEP owner’schoice not to offer licenses to some supply chain participants, limiting theavailability of injunctions against downstream customers.[80]

Courts in the United States and other internationaljurisdictions have likewise found that the FRAND obligation cannot be reconciledwith refusals to license to some market participants. For example, one recentU.S. federal court decision found that 「ETSI’s IPR policy, in fact, plainlystates that any willing licensee is entitled to license [a SEP declarant’s]intellectual property at a FRAND rate.」[81]Or as another US appeals-level Court stated, 「[t]o mitigate the risk that a SEPholder will extract more than the fair value of its patented technology, many SDOsrequire SEP holders to agree to license their patents on 『reasonable andnon-discriminatory』 or 『RAND』 terms. Under these agreements, an SEP holdercannot refuse a license to a manufacturer who commits to paying the RAND rate.」[82]「Motorola,in its declarations to the ITU, promised to 『grant a license to an unrestrictednumber of applicants on a worldwide, non-discriminatory basis and on reasonableterms and conditions to use the patented material necessary』 to practice theITU standards. This language admits of no limitations as to who or how manyapplicants could receive a license…」[83]And theFederal Circuit held that：「[T]helicensor’s established policy and marketing program to maintain his patentmonopoly by not licensing others to use the invention [is not relevant for SEPs].... Because of Ericsson’s RAND commitment [...] It cannot have that kind ofpolicy for maintaining a patent monopoly.」[84]One United States District Court, addressing contract and competition lawclaims, has held that, as a matter of law, a FRAND promise obligates the SEPowner to licenses its patents to companies seeking a license, includingcomponent suppliers, and that such a requirement is consistent with historicalindustry practices, held that：「If a SEPholder could discriminate against modem chip suppliers, a SEP holder couldembed its technology into a cellular standard and then prevent other modem chipsuppliers from selling modem chips to cellular handset producers. Suchdiscrimination would enable the SEP holder to achieve a monopoly in the modemchip market and limit competing implementations of those components ….」[85]

Likewise, the US competition agencies have endorsed thisapproach in their official actions resolving pending cases. Says that：「By making a FRAND commitment, a SEP holder voluntarilychooses to license its SEPs to all implementers of the standard on fair andreasonable terms.」[86]And FRAND「oblige[s] SEP owners: [] to make the patent in question available to allinterested third parties.」[87]

In describing the ETSI IPR Policy, ETSI’s Director-Generalwho oversaw its development, Karl-Heinz Rosenbrock, has described in detail howand why ETSI adopted its policy of requiring licensing to anyone that seeks alicense and is willing to pay a FRAND rate.[88]As Mr. Rosenbrock states, 「[t]he ETSI IPR Policy allows every company thatrequests a license to obtain one, regardless of where the prospective licenseeis in the chain of production and regardless of whether the prospectivelicensee is active upstream or downstream.」 As Mr. Rosenbrock furthercatalogues, this willingness to license at all levels of the supply isconsistent with historical practices among companies that have traditionallyoffered licenses to SEP technologies.[89]

In recent years, some companies announced a purported rightto refuse licenses to some companies in the supply chain – usually thecomponent or module level companies that are the most familiar with thestandardized technologies. Such refusals to license are problematic for anumber of reasons, and may ultimately lead to charging higher-than-FRANDroyalties to downstream companies based on the value and features that thosecompanies themselves create.

By contrast, there are clear and compelling reasons why alicensing approach focused on upstream companies may – depending on thespecific circumstances – create more efficient and more fair licensingprocesses. From an efficiency standpoint, in many industries there may be onlya handful of suppliers of standard-compliant components, as compared tohundreds or thousands of downstream companies utilizing those products. Bylicensing those handful of suppliers, a patent owner may be able to efficientlylicense a large portion of the industry. From a fairness perspective, often theupstream suppliers have the most information and experience regarding thestandard and its various technologies. Downstream companies, on the other hand,may adopt more of a 「plug and play」 approach, purchasing standardizedcomponents that can be easily incorporated into the downstream devices theymake and sell. In those circumstances, a downstream company may besubstantially less prepared to evaluate the alleged essentiality or validity ofthe subject patents, or to evaluate appropriate FRAND licensing terms.

 

（三）Volume or Quantity Discounts

Volume discounts are discounts based on the volume ofpurchases made by customers and are typically based on the total volume ofpurchases during a given period of time. Quantity discounts are based on theamount bought in a single transaction.

如本文第一部分所述。Somepatent pools, notably the 802.11 patent pool administered by Via Licensing andthe H.264 patent pool administered by MPEG-LAoffersignificant volume discounts, with firms selling high volumes of licensedproducts paying as little as 9% of the per-unit royalty paid by firms sellingsmaller volumes of licensed products. It is at least debatable whether suchsubstantial quantity discounts, for which the smaller volume sellers cannoteffectively qualify, satisfy the 「nondiscrimination」 aspect of FRAND. Suchsubstantial volume discounts are difficult to justify based on the cost oflicensing.[90]

 

A.      Specific Cost Savings Resulting From The Larger Purchases

美國法院在TCLv. Ericsson一案明確表示：Salesvolume alone does not justify giving lower rates to otherwise similar firms.[91]While volume discounts are not per se illegal, they cannot be used as ajustification for a lower price unless the seller can show specific costsavings resulting from the larger purchases. While the courts do not require anexact mathematical correlation between the cost saved and the actual pricingdifferential—they do require something fairly close. 關於cost-justification defense，詳見本文第七章。

研發和維護智慧財產權的成本主要是過去已經發生了的，並且是確定數額的，智慧財產權許可不存在類似有形產品生產環節的規模經濟；籤訂許可合同的成本是一次性的，而且，相比於與具有買方實力的行業巨頭進行許可談判的金錢和時間成本，專利權人與中國標準實施者的談判成本要低很多，智慧型手機專利大戰主要發生在行業巨頭之間，即為明證；審計許可費是否足額支付的成本相比許可費來講金額很小，而且一年甚至數年僅支出一次，並且通常由許可人承擔，僅在被許可人嚴重少報的情況下，審計費用才由被許可人承擔；通信領域專利許可費往往採用按季度付費的模式，拖欠許可費的風險不大。這些分析均表明標準必要專利權人對不同標準實施者提供Volume orQuantity Discounts的行為缺乏正當理由，實質上是對中小企業的壓榨，對競爭具有排除和限制效果。

The judgment of the CFI in Michelin II suggests thatquantity rebates not based on such efficiencies are not economically justifiedand thus should be found discriminatory within the meaning of Article 82(c).[92]

The European Commission and the European courts areincreasingly targeting price discrimination as an abuse under Article 82’sgeneral clause. In general, European authorities see selective price cutting assubject to the prohibition against 「abuses」 of dominant position because theyview it as a tool for deterring entry by rivals or for forcing them to exit themarket, and thus as a device for obtaining or preserving a dominant position. Inthe EU caselaw, both predatory pricing and selective price cutting are seen asfostering dominance and thus as a threat to competitive market conditions.  The European Commission explicitly declaredthat selective price reductions need not be at below-cost levels in order toconstitute abuse. [93]

The aim of the Robinson-Patman Act is to prevent a largebuyer from gaining discriminatory preferences over the small buyer because ofthe large buyer’s greater purchasing power. 智慧財產權許可中銷量折扣正好是Robinson-PatmanAct立法初衷所要規制的場景，其合理性需要嚴格審視。

When a dominant firm’s criteria for granting such a rebatereveal that it is not cost-related, the rebate may be viewed more like afidelity or target rebate impeding rivals from selling to the dominant firm’scustomers.[94]Moreover, discount systems in which the discounts are computed on sales over along reference period have been deemed to have similar exclusionary effectsbecause the value of the discount significantly increases over the length ofthe period, and thus exerts growing pressure on the buyer to remain with its currentsupplier. [95]

 

B.      Equally And Practically Available To All Purchasers

To avoid Robinson-Patman liability, the discount must bemore than theoretically available to competing customers. They must be equallyand practically available to all purchasers; purchasers must also know of them.A discount program which applies to all purchasers, will still violate RobinsonPatman if all purchasers do not know of it or cannot achieve the sales orquantity levels in order to actually obtain the discount due to their size,purchasing requirements or history.

The functional-availability defense does not apply unless(1) competing purchasers know about the availability of the lower price; and(2) competing purchasers can functionally, not just theoretically, obtain thelower price.

The lower price must be functionally available to most, ifnot all, of the competing purchasers. Thus, if a seller structures the discountprogram so that only a few companies (typically, those with large purchasingpower) can realistically take advantage of the discount, it is likely that aprice discrimination plaintiff will be able to establish that the discount wasnot functionally available. A discount that is available in theory will notsuffice; the discount should be functionally available to the majority of thepurchasers. The NAS warns that such menus of license choices could becomediscriminatory if they were structured, for example, to advantage only verylarge users of a standardized technology.[96]

The theory behind the functional availability doctrine isthat where a discount program is made available to competing purchasers, thefact that one purchaser utilizes the program and the other does not, fails togive rise to a [Robinson-Patman Act] claim because any price disparity is not aresult of discriminatory pricing but, instead, the buyer’s use of the discountprograms.[97]Wherethe plaintiff dealer was treated the same as all other dealers of themanufacturer with respect to discounts for the same customers and products oflike grade and quality, courts find no actionable price discrimination. If thesame discounts or promotional programs are functionally available to allcompeting dealers, any discrimination and competitive advantage suffered by theplaintiff is attributed not to the defendant’s program but the plaintiff’sfailure to take advantage of its opportunity to receive those prices.

While some courts have concluded that the existence offunctionally available lower price negates the injury to competitionrequirement in a secondary-line case, not all courts follow this view.

 

（四）Loyalty Rebates

 

The term 「loyalty rebates」 refers to discounting rewardssales above a certain level with a lower price on a firm’s entire purchases ofa product over a specified period (often a year).

The European authorities generally view loyalty rebatesgranted by dominant firms as unlawful. The Court of Justice tends to seeloyalty rebates as an anticompetitive weapon directed against competitors of theseller offering those rebates. Loyalty rebates are anticompetitive because theymake it increasingly difficult for rivals, including possible new entrants intothe industry, to sell to the customers of a firm that is offering thoserebates.As the buyer’s purchases increase, the rebate on past purchasesincreases in amount, eventually growing to the point where the per-unit priceof incremental units is negative. The rapid fall in the unit price of additionalpurchases is referred to as a 「suction effect,」 apparently referring to theincreasing incentive of the buyer to continue purchasing from the samesupplier. With the rebate scheme, the incentive for a customer to remain「loyal」 to the supplier increases as the customer’s purchases approach thetarget amount. In the typical exclusive supply contract, the customer is boundby contract to remain loyal to the supplier. The customer can break thecontract, but will have to pay damages if it does. The normal damages would bethe seller’s lost profits. Since the seller would have already earned its profitson its sales up to the time of the breach, the profits are those that theseller would lose from future sales that have been diverted to a rival.

Hoffmann-La Roche case established the principle that adominant company could not tie customers to obtain 「all or most」 of theirrequirements exclusively from itself, even if it did so at their request. Thesame applied to granting fidelity-or loyalty-rebates that were conditional on acustomer obtaining all or most of its requirements from the dominant firm.Anticompetitive foreclosure effects could be assumed from such a rebate, and nofurther analysis was required.  

The ECJ ruled on the question of discrimination by holdingthat: the effect of a fidelity rebate is to apply dissimilar conditions to equivalenttransactions with other trading parties in that two purchasers pay a differentprice for the same quantity of the same product depending on whether theyobtain their supplies exclusively from the undertaking in a dominant positionor have several sources of supplies.[98]

 TFEU did not seek toensure that less efficient businesses stayed on the market. Legitimatecompetition on the merits may by definition lead to the exit of less efficientcompetitors. Article 102 TFEU therefore prohibited a dominant firm from,amongst other things, adopting pricing practices that had an exclusionaryeffect on competitors who were as efficient as itself. First, the CJ has notoverruled the presumption that exclusivity or loyalty rebates are abusive.However, this presumption now appears to be easily nullified-assuming thedominant company brings evidence during the investigation that its rebates arenot capable of leading to foreclosure, the Commission must then look closely ateffects. Underpinning this is the Court's starting point that Art 102 TFEUpermits pricing conduct by a dominant firm that can drive less efficientcompetitors from the market. The presumption should, therefore, have become onein name only.

The U.S. cases that have dealt with loyalty rebates viewthem as raising issues under the monopolization or attempted monopolizationclauses of §2 of the Sherman Act.[99]

 

（五）Cross-Licenses

A group of competitors may work together to disadvantageanother group of competitors, e.g., later entrants or firms with no SEPs. Cross-licensesbetween patent holders enable patent-rich firms to favor one another and disadvantagepatent-poor firms. [100]

筆者認為，擁有大量標準必要專利的專利權人，作為參與了標準制定的主體，其已經做出了「公平合理無歧視許可」的承諾，其有責任並且有能力在交叉許可中明確指出它們各自的標準必要專利在整體標準中分別所佔的份額，尤其是許可費的份額，並使其他標準實施者和專利權人知悉。理由如下文第五章所述。

 

[1]相應原文請見：https://www.justice.gov/atr/IPguidelines/download

[2] Rebecca Allensworth, Casting a FRAND Shadow: The Importance ofLegally Defining 『Fair and Reasonable』 and How Microsoft v. Motorola Missed theMark, 22 TEX. INTELL. PROP. L.J. 235 (2013–2014).

[3] Simcoe, T., Shampine, A. Economics of Patents and Standardization:Network Effects, Hold-up, Hold-out, Stacking. In Jorge Contreras (Eds.),"The Cambridge Handbook of Technical Standardization Law", CambridgeUniversity Press, page 121.

[4] Gilbert, 「Deal or no deal? Licensing negotiations in standardssetting organizations,」 Antitrust Law Journal, Vol. 77 No. 3 (2011), pp.855–888 at 859.相應原文請見https://escholarship.org/uc/item/6kv798tf

[5]相應原文請見：https://eur-lex.europa.eu/eli/treaty/tfeu_2008/art_102/oj

[6] MEO - Servicos de Comunicacoes e Multimedia SA v Autoridade da Concorrencia(C-525/16) EU:C:2018:270; [2018] Bus. L.R. 1155; [2018] 4 WLUK 286 (ECJ (2ndChamber)). 24, .

[7] Christopher B. Roberts, Antitrust and Patent Misuse Considerationsof Royalty Rate Differences Among Patent Licensees, 65 Journal of the PatentOffice Society 344, 1983.

[8] Laitram Corp. v. King Crab, Inc., 244 F. Supp. 9, 146 U.S.P.Q.(BNA) 640 (D. Alaska 1965), opinion supplemented, 245 F. Supp. 1019, 147U.S.P.Q. (BNA) 136 (D. Alaska 1965); LaPeyre v. F.T.C., 366 F.2d 117, 151U.S.P.Q. (BNA) 79 (5th Cir. 1966); Peelers Co. v. Wendt, 260 F. Supp. 193, 151U.S.P.Q. (BNA) 378 (W.D. Wash. 1966).

[9] Laitram Corp. v. King Crab, Inc., 244 F. Supp. 9, 146 U.S.P.Q.(BNA) 640 (D. Alaska 1965), opinion supplemented, 245 F. Supp. 1019, 147U.S.P.Q. (BNA) 136 (D. Alaska 1965) (finding patent misuse but not an antitrustviolation); Peelers Co. v. Wendt, 260 F. Supp. 193, 151 U.S.P.Q. (BNA) 378(W.D. Wash. 1966) (finding a Sherman Act § 2 violation).

See also Honeywell Inc. v. SperryRand Corp., 180 U.S.P.Q. (BNA) 673, 1974-1 Trade Cas. (CCH) P 74874, 1973 WL903 (D. Minn. 1973) (antitrust violation not proven, but court suggested aframework for finding a Sherman Act § 1 violation in discriminatory patentroyalty situations); Ansul Co. v. Uniroyal, Inc., 306 F. Supp. 541, 163U.S.P.Q. (BNA) 517 (S.D. N.Y. 1969), aff』d in part, rev』d in part and remanded,448 F.2d 872, 169 U.S.P.Q. (BNA) 759, 170 U.S.P.Q. (BNA) 549, 1971 Trade Cas.(CCH) P 73568, 1971 Trade Cas. (CCH) P 73638 (2d Cir. 1971) (a patentee may not「destroy a licensee in the marketplace」 by charging discriminatory royaltyrates).

[10] Scott Martin & Irving Scher, The Robinson-Patman Act: Sellers』and Buyers』 Violations and Defenses, 1649 P.L.I./CORP. 553, 561 (2008)

[11] Ryan Luchs, Tansev Geylani, Anthony Dukes, Kannan Srinivasan,(2010) The End of the Robinson-Patman Act? Evidence from Legal Case Data.Management Science 56(12):2123-2133.

[12] Contreras , Jorge L. 2015. 「 A Brief History of FRAND: AnalyzingCurrent Debates in Standard Setting and Antitrust through a Historical Lens ,」80 Antitrust Law Journal 39 .

[13] 46 F. Supp. 541 (N.D. Ohio 1942), order modified, 323 U.S. 386(1945), modified in response to a request for clarification, 324 U.S. 570(1945)

[14] 63 F. Supp. 513 (S.D.N.Y. 1945), aff』d, 332 U.S. 319 (1947)

[15] 81 F. Supp. 42 (D. Mass. 1948)

[16]相應原文請見：https://www.ftc.gov/tips-advice/competition-guidance/guide-antitrust-laws/price-discrimination-robinson-patman

[17]相應原文請見：https://www.ftc.gov/tips-advice/competition-guidance/guide-antitrust-laws/price-discrimination-robinson-patman

[18] Richard A Posner, The Robinson-Patman: Act Federal Regulation ofPrice Differences, American Enterprise Institute for Public Policy Research,1976. At 35.

[19] J. Antony VanDuzer, Anticompetitive Pricing Practices and theCompetition Act - Theory, Law and Practice (VanDuzer Report) (Ottawa:Competition Bureau, 1999). At 30.

[20]相應條款原文請見：https://indiankanoon.org/doc/1780194/

[21] FTC v. Anheuser-Busch, Inc., 363 U.S. 536 (1960).

[22]Richard A Posner, The Robinson-Patman: Act Federal Regulation ofPrice Differences, American Enterprise Institute for Public Policy Research,1976. At 35.

[23] Carlton , Dennis and Jeffery Perloff . 2004 . Modern IndustrialOrganization (4th Edition). Pearson; Posner , Richard . 2001 . Antitrust Law:An Economic Perspective . Chicago : University of Chicago Press, at 79– 80.

[24] Tri-Valley Packing Ass』n v. F.T.C., 329 F.2d 694, 703–04 (9th Cir.1964)

[25] Texaco Inc. v. Hasbrouck, 496 U.S. 543, 561 & n.18 (1990)

[26] Daniel J. Gifford, Robert T. Kudrle. The Law and Economics of PriceDiscrimination in Modern Economies: Time for Reconciliation? UC Davis LawReview， Vol. 43. 全文請見：https://lawreview.law.ucdavis.edu/issues/43/4/articles/43-4_Gifford-Kudrle.pdf

[27] Damien Geradin & Nicolas Petit, Price Discrimination Under ECCompetition Law: Another Antitrust Doctrine in Search of Limiting Principles? Journalof Competition Law and Economics, 2(3), 479–531.

[28]該案2988 Judgment第175段http://www.bailii.org/ew/cases/EWHC/Patents/2017/2988.pdf

[29] Damien Geradin, FRAND Arbitration: The Determination of Fair,Reasonable and Non-Discriminatory Rates for SEPs by Arbitral Tribunals, 3 CPIAntitrust Chron., Summer 2016, at 1, 9

[30] Anne Layne-Farrar, Nondiscriminatory Pricing: Is Standard SettingDifferent?, 6 J. Competition L. & Econ. 815, (2010)

[31] Dennis W. Carlton & Allan L. Shampine, An EconomicInterpretation of FRAND, 9 J. Competition L. & Econ. At 546 (2013).

[32] J. Gregory Sidak, Fair and Unfair Discrimination in Royalties forStandard-Essential Patents Encumbered by a FRAND or RAND Commitment, theJournal on Innovation Criterion, Vol.2, 2017.

[33] Coleman v. Donahoe, 667 F.3d 835, 846 (7th Cir. 2012)

[34] Lee v. Kansas City S. Ry. Co., 574 F.3d 253, 260 (5th Cir. 2009)

[35]引用自該案2988 Judgment第487段http://www.bailii.org/ew/cases/EWHC/Patents/2017/2988.pdf

[36]相應原文請見TCL v. Ericsson一案的Amended Memorandum of Findings of Fact and Conclusions of Law第56-59頁。部分字體加粗顯示是筆者所為。https://www.courtlistener.com/recap/gov.uscourts.cacd.584255/gov.uscourts.cacd.584255.1962.0_1.pdf

[37] Janusz Ordover and Allan Shampine, Implementing the FRANDCommitment, The antitrust source, October 2014.

[38] Janusz Ordover and Allan Shampine, Implementing the FRANDCommitment, The antitrust source, October 2014.

[39]「Findings of Fact and Conclusions of Law,」Microsoft v. Motorola, U.S. District Court, Western District of Washington atSeattle, Para. 617; available at http://www.scribd.com/doc/138032128/13-04-25-Microsoft-Motorola-FRAND-Rate-Determination

[40] E.g., Power Integrations, Inc. v. Fairchild Semiconductor Int『l,Inc., 904 F.3d 965, 978, 128 U.S.P.Q.2D (BNA) 1121, 2018 U.S. App. LEXIS 26842(Fed. Cir. Jul. 3, 2018); Ericsson, Inc. v. D-Link Sys., 773 F.3d 1201, 1226,113 U.S.P.Q.2d (BNA) 1001 (Fed. Cir. 2014)

[41] Garretson v. Clark, 111 U.S. 120

[42] Ericsson, Inc. v. D-Link Sys., Inc., 773 F.3d 1201, 1226 (Fed. Cir.2014).

[43] Commonwealth Sci. & Indus. Research Organisation v. Cisco Sys.,809 F.3d 1295

[44] VirnetX, Inc. v. Cisco Sys., Inc., 767 F.3d 1308, 1329 (Fed. Cir.2014).

[45] Those cases include:

Dowagiac Mfg. Co. v. Minnesota MolinePlow Co., 235 U.S. 641, 35 S. Ct. 221, 59 L. Ed. 398, 1915 U.S. LEXIS 1822(1914);

Westinghouse Electric & Mfg. Co.v. Wagner Electric & Mfg. Co., 225 U.S. 604, 32 S. Ct. 691, 56 L. Ed. 1222,1912 U.S. LEXIS 2108, 41 L.R.A. (n.s.) 653 (1912);

Warren v. Keep, 155 U.S. 265, 15 S.Ct. 83, 39 L. Ed. 144, 1894 U.S. LEXIS 2271 (1894);

Keystone Mfg. Co. v. Adams, 151 U.S.139, 14 S. Ct. 295, 38 L. Ed. 103, 1894 U.S. LEXIS 2042 (1894);

Sessions v. Romadka, 145 U.S. 29, 12S. Ct. 799, 36 L. Ed. 609, 1892 U.S. LEXIS 2120 (1892);

McCreary v. Pennsylvania Canal Co.,141 U.S. 459, 12 S. Ct. 40, 35 L. Ed. 817, 1891 U.S. LEXIS 2535 (1891);

Crosby Steam Gage & Valve Co. v.Consolidated Safety Valve Co., 141 U.S. 441, 12 S. Ct. 49, 35 L. Ed. 809, 1891U.S. LEXIS 2534 (1891);

Hurlbut v. Schillinger, 130 U.S. 456,9 S. Ct. 584, 32 L. Ed. 1011, 1889 U.S. LEXIS 1768 (1889);

Tilghman v. Proctor, 125 U.S. 136, 8S. Ct. 894, 31 L. Ed. 664, 1888 U.S. LEXIS 1921 (1888);

Yale Lock Mfg. Co. v. Sargent, 117U.S. 536, 6 S. Ct. 934, 29 L. Ed. 954, 1886 U.S. LEXIS 1871 (1886);

Dobson v. Hartford Carpet Co., 114U.S. 439, 5 S. Ct. 945, 29 L. Ed. 177, 1885 U.S. LEXIS 1778 (1885);

Thomson v. Wooster, 114 U.S. 104, 5S. Ct. 788, 29 L. Ed. 105, 1885 U.S. LEXIS 1739 (1885);

Black v. Thorne, 111 U.S. 122, 4 S.Ct. 326, 28 L. Ed. 372, 1884 U.S. LEXIS 1765 (1884);

Garretson v. Clark, 111 U.S. 120, 4S. Ct. 291, 28 L. Ed. 371, 1884 U.S. LEXIS 1764 (1884);

Manufacturing Co. v. Cowing, 105 U.S.253, 26 L. Ed. 987, 15 Otto 253, 1881 U.S. LEXIS 2115 (1882);

Elizabeth v. Pavement Co., 97 U.S.126, 24 L. Ed. 1000, 7 Otto 126, 1877 U.S. LEXIS 1761 (1878);

Cawood Patent, 94 U.S. 695, 24 L. Ed.238, 4 Otto 695, 1876 U.S. LEXIS 1930 (1877);

Blake v. Robertson, 94 U.S. 728, 24L. Ed. 245, 4 Otto 728, 1876 U.S. LEXIS 1934 (1877);

Burdell v. Denig, 92 U.S. 716, 23 L.Ed. 764, 2 Otto 716, 1875 U.S. LEXIS 1809 (1876);

Tremolo Patent, 90 U.S. 518, 23 L.Ed. 97, 23 Wall. 518, 1874 U.S. LEXIS 1331 (1875);

Littlefield v. Perry, 88 U.S. 205, 22L. Ed. 577, 21 Wall. 205, 1874 U.S. LEXIS 1358 (1875);

Packet Co. v. Sickles, 86 U.S. 611,22 L. Ed. 203, 19 Wall. 611, 1873 U.S. LEXIS 1473 (1874);

Philp v. Nock, 84 U.S. 460, 21 L. Ed.679, 17 Wall. 460, 1873 U.S. LEXIS 1385 (1873);

Mowry v. Whitney, 81 U.S. 620, 20 L.Ed. 860, 14 Wall. 620, 1871 U.S. LEXIS 1028 (1872);

Suffolk Mfg. Co. v. Hayden, 70 U.S.315, 18 L. Ed. 76, 3 Wall. 315, 1865 U.S. LEXIS 714 (1866);

New York v. Ransom, 64 U.S. 487, 16L. Ed. 515, 23 HOW 487, 1859 U.S. LEXIS 798 (1860);

New York v. Ransom, 64 U.S. 487, 490,16 L. Ed. 515, 1859 U.S. LEXIS 798, 23 HOW 487 (1859); and,

Seymour v. McCormick, 57 U.S. 480, 14L. Ed. 1024, 16 How. 480, 1850 U.S. LEXIS 1562 (1854).

[46] Ericsson, 773 F.3d at 1226. In short, apportionment.

[47] Commonwealth Sci. & Indus. Research Organisation v. Cisco Sys.,809 F.3d 1295

[48] Ericsson, Inc. v. D-Link Sys., 773 F.3d 1201, 1232, 113 U.S.P.Q.2d(BNA) 1001 (Fed. Cir. 2014).

[49] Ericsson, Inc. v. D-Link Sys., 773 F.3d 1201, 1233, 113 U.S.P.Q.2d(BNA) 1001 (Fed. Cir. 2014).

[50] Seymour v. McCormick, 57 U.S. 480, 490,491, 14 L. Ed. 1024 (1854).

[51] Micro Chem., Inc. v. Lextron, Inc., 317 F.3d 1387, 1393, 65U.S.P.Q.2d (BNA) 1532, (Fed. Cir. 2003); Rite-Hite Corp. v. Kelley Co., 56 F.3d1538, 1549 n.9, 35 U.S.P.Q.2d (BNA) 1065 (Fed. Cir. 1995) (en banc).

[52] LaserDynamics, 694 F.3d at 67

[53] GPNE Corp. v. Apple, Inc., 2014 U.S. Dist. LEXIS 53234

[54] VirnetX, Inc. v. Cisco Sys., 767 F.3d 1308

[55] Rembrandt Soc. Media, LP v. Facebook, Inc., 2013 U.S. Dist. LEXIS171127, at *23–*24 (E.D. Va. Dec. 3, 2013)

[56] Network Prot. Scis., LLC v. Fortinet, Inc., 2013 U.S. Dist. LEXIS138890, at *21 (N.D. Cal. Sept. 26, 2013)

[57] AVM Techs., LLC v. Intel Corp., 2013 U.S. Dist. LEXIS 1165, at *7(D. Del. Jan. 4, 2013)

[58] Tomita Techs. USA, LLC v. Nintendo Co., Ltd., 2013 U.S. Dist. LEXIS116486, at *23–*24 (S.D.N.Y. Aug. 13, 2013)

[59] Cornell Univ. v. Hewlett Packard Co., JudgeRader

[60] LaserDynamics, 694 F.3d at 67.

[61] Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292

[62] Rite-Hite Corp. v. Kelley Co., 56 F.3d 1538

[63] Ericsson, 773 F.3d at 1226

[64] 632 F.3d 1292, 1320 (Fed. Cir. 2011).

[65] LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51

[66] Cornell Univ. v. Hewlett-Packard Co., 609 F. Supp. 2d 279

[67] Mark A. Lemley, Intellectual Property Rights and Standard-SettingOrganizations, 90 Calif. L. Rev. at 1901, (2002).

[68] John J. Kelly, An Overview of the JEDEC Patent Policy, FTC Dkt. No.9302, at 4 (Mar. 26, 2002), http://www.rambus.org/legal/ftc/Evidence/CX0449.PDF

[69] TIA, Guidelines to the Telecommunications Industry AssociationIntellectual Property Rights Policy § 5, at 5 (May 1, 2014) , http://www.tiaonline.org/sites/default/files/pages/IPRPGL-2014May01_0.pdf

[70] Mario Mariniello, Fair, Reasonable and Non-Discriminatory (FRAND)Terms, 7 J. Competition L. & Econ. 523, 525 (2011).

[71] Microsoft Corp. v. Motorola, Inc. (9th Cir. 2012 , p.884).

[72] IEEE- SA 2015, Sec. 6.2(b).

[73] Jorge L. Contreras, A Brief History of FRAND: Analyzing CurrentDebates in Standard Setting and Antitrust Through a Historical Lens, 80Antitrust L.J. 39, 80 (2015)

[74] Janusz Ordover and Allan Shampine, Implementing the FRANDCommitment, The antitrust source, October 2014.

[75] David J. Teece, Edward F. Sherry, Peter C. Grindley, On the「non-discrimination」 aspect of FRAND licensing: A response to the IndianCompetition Commission’s recent orders, IIMB Management Review (2018) 30, 10–26.

[76]由the Fair Standards Alliance and ACT和The App Association發起、並由連同包括Apple Inc.在內的五十餘家機構推動和批准的Core Principles and Approaches for Licensing of Standard EssentialPatents，以下簡稱CWA2。相應原文請見：ftp://ftp.cencenelec.eu/EN/News/WS/2019/CWA_SEP/CWA95000.pdf

[77] ETSI Directives, Guide on Intellectual Property Rights, at Sec.1.4.

[78] EC Horizontal Guidelines, supra, at note 12, paragraphs 285-287;see also EC Technology Transfer Guidelines, para. 261.

[79] C-170/13 Huawei Techs. Co. v. ZTE Corp., [2015] E.C.R. 477, para.53; see also EC SEP Communication.

[80] LG Düsseldorf, Urt v 1.7.2018 - 4c O 81/17, available at https://www.justiz.nrw.de/nrwe/lgs/duesseldorf/lg_duesseldorf/j2018/4c_O_81_17_Urteil_20180711.html.

[81] Order Den. Anti-Suit Inj. at 31, Apple Inc. v. Qualcomm, Inc., CaseNo. 3:17-cv-00108-GPC-MDD (N.D. Cal. Sept. 7, 2017), ECF No. 141.

[82] Microsoft Corp. v. Motorola, Inc., 795 F. 3d 1024, 1031 (9th Cir.2015).

[83] Microsoft Corp. v. Motorola, Inc., 696 F.3d 872, 884 (9th Cir.2012)

[84] Ericsson, Inc. v. D-Link Sys, Inc., 773 F.3d 1201, 1230 (Fed. Cir.2014)

[85] Fed. Trade Comm』n v. Qualcomm Inc., No. 17-CV00220-LHK, 2018 WL 5848999at *12-13 (N.D. Cal. Nov. 6, 2018)

[86] U.S. Fed. Trade Comm』n, Letter to Commentators, Motorola MobilityLLC & Google Inc., Docket No. C-4410 (July 23, 2013) available at https://www.ftc.gov/sites/default/files/documents/cases/2013/07/130724googlemotorolaletter.pdf.

[87] Order Den. Anti-Suit Inj. at 31, Commission Decision in CaseAT.39985 - Motorola - Enforcement of GPRS Standard Essential Patents C(2014)2892 final, 29 April 2014, para. 55

[88] Karl Heinz Rosenbrock, Licensing At All Levels Is The Rule UnderThe ETSI IPR Policy (Nov. 3, 2017), available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3064894;Karl Heinz Rosenbrock, Why the ETSI IPR Policy Requires Licensing to All, http://www.fair-standards.org/wp-content/uploads/2017/08/Why-the-ETSI-IPR-Policy-Requires-Licensing-to-All_Karl-Heinz-Rosenbrock_2017.pdf.

[89] See, e.g., Countercls. & Affirmative Defense ¶ 53, BroadcomCorp. v. Qualcomm Inc., No. 05-3350 (MLC) (JJH) 2008 WL 2140801 (D.N.J. Feb.29, 2008) (「Qualcomm, which owns a large portion of the intellectual propertycovering CDMA technology, operates a pro-competitive licensing model, in whichit offers licenses on fair, reasonable and non-discriminatory terms to anyinterested company.").

[90] David J. Teece, Edward F. Sherry, Peter C. Grindley, On the「non-discrimination」 aspect of FRAND licensing: A response to the IndianCompetition Commission’s recent orders, IIMB Management Review (2018) 30, 10–26.

[91]相應原文請見TCL v. Ericsson一案的Amended Memorandum of Findings of Fact and Conclusions of Law第61頁。https://www.courtlistener.com/recap/gov.uscourts.cacd.584255/gov.uscourts.cacd.584255.1962.0_1.pdf

[92] CFI, Manufacture franc¸aise des pneumatiques Michelin v. Commission(Michelin II ), 30 September 2003, T-203/01 at §§98 and 100.

[93] Daniel J. Gifford, Robert T. Kudrle. The Law and Economics of PriceDiscrimination in Modern Economies: Time for Reconciliation? UC Davis LawReview， Vol. 43. 全文請見：https://lawreview.law.ucdavis.edu/issues/43/4/articles/43-4_Gifford-Kudrle.pdf

[94] Case T-219/99, British Airways Plc. v. Comm』n, E.C.R. II-5917, ¶¶246-47 (2003).

[95] See Case 322/81, NV Michelin v. Comm』n, E.R.C. 3461, ¶¶ 81-82(1983); see also Commission Decision 92/163/EEC, Tetra Pak II, 1992 O.J. (L72/1) 174 (1991) (barring aggregation on quantity discounts).

[96] NAS 2013 , 65.

[97] General Auto Parts Co. v. Genuine Parts Co., 2007-1 Trade Cas.(CCH) ¶ 75707, 2007 WL 704121 (D. Idaho 2007), judgment aff』d, 293 Fed. Appx.515, 2008-2 Trade Cas. (CCH) ¶ 76312 (9th Cir. 2008).

[98] ECJ, Hoffmann-La Roche v. Commission, 13 February 1979, 85/76,ECR[1979]-461

[99] Daniel J. Gifford, Robert T. Kudrle. The Law and Economics of PriceDiscrimination in Modern Economies: Time for Reconciliation? UC Davis LawReview， Vol. 43. 全文請見：https://lawreview.law.ucdavis.edu/issues/43/4/articles/43-4_Gifford-Kudrle.pdf

[100] Janusz Ordover and Allan Shampine, Implementing the FRANDCommitment, The antitrust source, October 2014.

特別說明：

第一，本文引用了大量文獻，它們是本文的精華。有時候沒有標明出處，筆者對此深表歉意。

第二，本文只是草稿，不代表筆者最終觀點。於2019年8月31日完成。

第三，本文僅是學術研究，但成一家之言，以拋磚引玉，求百家爭鳴，繁榮學術討論。本文與筆者作為律師現在、未來或過去辦理的任何具體案件的觀點均無關，更不反映任何客戶的立場。

第四，基於我國巨額智慧財產權許可費逆差，筆者自然站在了被許可人的角度閱讀文獻，因此本文或許必然存在偏見，請讀者諸君務必注意。