Competition law promotes rivalry as a means of organizing economic behavior. The main substantive provisions of competition law-or antitrust law, as it is known in the United States (US)-prohibit certain agreements in restraint of trade, conduct to maintain or acquire monopoly power, and mergers. This chapter surveys the intersection of antitrust law with intellectual property (IP).
Government enforcers and private plaintiffs have advanced a wide range of antitrust claims against IP rights holders. These include, among others, the alleged abuse of a standard setting process by a patentee, joint price setting by means of a blanket license for copyrighted musical works, and anticompetitive settlements of patent litigation. These examples arise distinctively in the context of IP. They can be viewed as a kind of special-purpose antitrust law, although the principles applied are more general. In other cases against rights holders, the IP overlay adds nothing distinctive.
This chapter examines two types of antitrust claims. Collusion claims challenge forms of cooperation that reduce competition, often by softening competition among the cooperating firms. Exclusion claims challenge the improper preservation of incumbency through self-entrenching conduct. The categories overlap because firms can collude in order to exclude. The treatment here is necessarily selective. 1 In the interest of space, mergers are largely ignored. 2 The focus is patent law and copyright law, where most cases arise, at the expense of trademark and other forms of IP. The examples are drawn mainly from the US, with some discussion of European developments. Cases that are particularly important or revealing about the interaction of IP with antitrust law are emphasized.
The case selections serve a second goal, which is to uncover whether and how IP rights alter the substantive scope of antitrust law. At first blush, the conflict between IP and antitrust is obvious. IP is a temporary monopoly intended to suppress rivalry. Antitrust law, by contrast, prohibits certain conduct that tends to improperly create or preserve a monopoly or achieve a similar effect through contract. Moreover, the point of IP is to enable high prices as an inducement to innovate, while a major antitrust goal is to keep prices down close to cost.
IP places significant limits on antitrust law. The assertion of a valid, infringed patent against a rival is legal, even if the resulting exclusion raises prices and reduces output. Beyond this core exclusionary right, questions arise about how far the privilege extends. A broader privilege increases the size of the innovator's reward, potentially at the expense of economic values protected by antitrust law. One message of this chapter is that, beyond the core, IP restricts antitrust law less than one might expect. This message is broadly in keeping with the IP Guidelines issued by US regulators, which proceed from the proposition that "standard antitrust analysis applies to [IP]." 3 Moreover, the restrictions that do exist are sometimes subtle.
Section 2 of this chapter describes two pathways by which IP limits antitrust law. First, IP limits the scope of antitrust by operation of statute. Some specific IP provisions arguably permit particular conduct that ordinarily violates antitrust law. For example, US patent law grants a right to refuse to deal, which might immunize certain refusals that otherwise violate antitrust law. Second, IP might limit antitrust law as a matter of policy. At its core, antitrust law is governed by economic analysis. The economic basis for IP-most prominently, that temporary monopoly provides an ex ante inducement to innovate-might factor within antitrust's cost-benefit calculus. The result, again, could be to permit conduct that is ordinarily prohibited.
Section 3 analyzes collusion claims against IP rights holders, including market division, settlements of patent litigation, price-restricted licenses, blanket licenses, and patent pools. Section 4 addresses exclusion by means of sham litigation, fraudulent acquisition of IP rights, assertion of IP rights to weaken competition in a second market, refusals to deal, product design, manipulation of a standard setting organization, and collective exclusion. Sections 5 and 6 turn to the reciprocal question of how competition principles limit the exercise of IP rights. Section 5 discusses misuse, a limitation on IP exploitation that mirrors antitrust law in important respects. The chapter concludes with a brief assessment of how antitrust law limits the exercise of IP rights by reinforcing limitations on the scope and duration of the right.
Whether and how IP alters antitrust law are questions that have attracted significant commentary. 4 As noted in the Introduction, IP may alter the operation of ordinary antitrust law either by operation of statute or as a matter of economic policy. This section considers these pathways in turn.
IP law and antitrust law overlap in their coverage, and in some instances IP displaces antitrust law by operation of statute. As the US Supreme Court remarked, "[t]he patent laws which give a . . . monopoly on 'making, using, or selling the invention' are in pari materia with the antitrust laws and modify them pro tanto." 5 This displacement is unsurprising. Limited exemptions from antitrust law are fairly common to effectuate (for example) the policies behind legalized export cartels, labor law, and agricultural policy. The same is true for IP.
The strongest case for a limitation comes from a specific IP provision directed to particular conduct. For example, the US Patent Act grants an exclusive right to "mak[e], us[e], . . . or sel[l]" the patented invention. 6 Copyright law similarly grants an exclusive right to distribute a copyrighted work. 7 These provisions are the source of the core limitation discussed in the Introduction, the assertion of a valid right against a rival infringer that secures the rival's exclusion. Other IP provisions might similarly be read as particularized limitations on antitrust law. Taken together, the statutory provisions define an "exclusionary scope" to the IP right. 8 Applying the scope-of-the-right concept beyond the core case, however, is not straightforward. Three issues arise. First, a particular IP entitlement may not include the privilege to act anticompetitively. By way of analogy, owning a hammer grants certain exclusive benefits in the use of the hammer, but not the right to break storefront windows. 9 Consider the following Patent Act provisions and their debated connection to antitrust law:
Territorial restrictions: A "patentee . . . may . . . grant and convey an exclusive right . . . to the whole or any specified part of the United States." 10 Patent holders read the provision to exempt territorial restrictions-for example, ) (the assertion of an "absolute and unfettered right to use its intellectual property as it wishes . . . is no more correct than the proposition that use of one's personal property, such as a baseball bat, cannot give rise to tort liability").
permission to sell a patented invention in the state of New York but nowhere else. Antitrust plaintiffs counter that the provision has nothing to do with competition. 11 Presumption of validity: A patent enjoys a presumption of validity. 12 Patent holders read the provision to allow payments to a rival to abandon litigation challenging the patent's validity. Antitrust plaintiffs counter that the presumption is merely a starting point for allocating burdens of proof in a patent infringement proceeding.
Refusals to license: A patentee is exempt from a finding of "misuse or illegal extension of the patent right . . . by reason of having . . . refused to license or use any rights to the patent." Patent holders read the provision to permit refusals that foreclose competition in a second market. Antitrust plaintiffs counter that the provision merely avoids a finding of patent misuse, not antitrust liability.
Second, the conduct may offend an explicit statutory limit on the IP right. For example, IP rights are limited in time. Conduct that extends a right beyond its expiration date is a poor candidate for avoiding ordinary antitrust law. A third complication arises when conduct expressly covered by an IP provision is paired with conduct that is not. For example, a patentee might not merely refuse to sell a patented product, but also condition the sale on purchase of a second product. Is this conduct within the scope of the patent? Closely related, does the "greater" right to refuse to license entirely include the "lesser" right to license only on particular conditions, or to some customers but not others? The overall effect of these qualifications is to give a question-begging cast to determinations that particular conduct either is or is not within the scope of the IP right.
Beyond specific provisions, the very existence of a conflicting statutory scheme might argue in favor of narrower antitrust liability. In particular, IP policy might be an input into the cost-benefit analysis that underpins antitrust law, thereby altering the antitrust result as a matter of economic policy. This is a 11 WF Baxter, "Legal Restrictions on Exploitation of the Patent Monopoly: An Economic Analysis" (1966) 76 Yale LJ 267, 349-352, concludes that § 261 is a mere "demarcation of the dividing line between interests sufficiently extensive to be transferred by assignment and those which are to be transferred by license."
second pathway by which IP might alter the ordinary operation of antitrust law.
To see how this might work, we must take a closer look at those policies. The touchstone of antitrust law is "consumer welfare." 14 The definition of that phrase, however, is contested. One working definition is the net benefit that consumers derive from the products and services they consume. An important source of consumer welfare is low prices. Lower prices imply a higher net benefit-higher "consumer surplus," in the jargon. Thus, one account of antitrust emphasizes low prices as the central goal of antitrust law. On this view, the prohibition of (say) cartels is justified because cartels raise prices, which reduces consumer surplus.
The consumer welfare approach was applied in Federal Trade Commission v Actavis, where the Supreme Court considered the antitrust treatment of certain settlements of patent litigation. 15 The Court explained that the harm from such conduct arises when a firm "maintain[s] and . . . share[s] patent-generated monopoly profits," because when this happens, "the consumer loses." This consumer loss-higher prices-is the cognizable harm. Part of the consumer loss takes the form of revenue transferred to producers as extra profit. Additional loss takes the form of "deadweight loss" wherein consumers, who value the good more than its marginal cost of production, are deflected instead to less desirable alternatives.
These harms are depicted in Figure 1. Suppose particular conduct-a cartel of vitamin producers, for example-shifts the price and quantity from point O to point X. After the conduct, consumer surplus is reduced by the sum of Areas A and B. Area A is the deadweight loss resulting from the conduct. Area B represents the transfer from consumers to producers.
An alternative goal for antitrust law is total welfare. Total welfare is the sum of consumer benefit and producer profits. A total welfare standard credits producer profits as an offset to consumer harms. (Unfortunately, the phrase "consumer welfare" sometimes refers to total welfare, a confusing and misleading usage.) Thus, a transfer from consumers to producers-considered a harm in the consumer welfare analysis-is merely neutral on a total welfare view, whereas the deadweight loss counts as a harm. Disregarding transfers narrows the scope of antitrust liability, compared to a consumer welfare standard. 16 For example, in Figure 1, the relevant harm is limited to Area A.
The analysis thus far focuses on the price and cost of existing goods. In addition to these static effects, conduct may have a dynamic effect on the development of new or improved products and services. Reduced innovation has greater long-term significance for the economy than high prices. 17 New products and services drive economic growth, and economic analysis suggests that technological change ultimately dominates price effects in its long-run contribution to welfare. 18 Thus, the dynamic benefits and costs of particular conduct are of significant concern to competition policy.
Taking dynamic effects into account might argue in favor of a measure of latitude to the innovator in some cases, thereby undercutting the static analysis. The prospect of higher prices is an ex ante inducement to innovate. As the Supreme Court stated in Verizon v Trinko:
The mere possession of monopoly power, and the concomitant charging of monopoly prices, is not only not unlawful; it is an important element of the free-market system. The opportunity to charge monopoly prices-at least for a short period-is what attracts 'business acumen' in the first place; it induces risk taking that produces innovation and economic growth. 19 Monopoly is not only an inducement but also potentially a platform for further innovation. Joseph Schumpeter argued that monopolists are first-rate innovators. 20 They possess advantages in capacity such as superior personnel, 16 A further difference arises when conduct increases efficiency to the benefit of producers. For example, a merger may cut costs, with the savings kept by the merging parties rather than being passed onto consumers. Such retained producer benefits count, on a total welfare view, in favor of the conduct. greater financial resources (particularly if capital markets are imperfect and hence future innovations are funded out of current cash flows), and the freedom to make long-term plans. Moreover, they are particularly well positioned (and hence incentivized) to capture the profits from further improvements.
We have now reached the point where the economics of competition policy and the economics of IP intersect. An award of IP is often justified as a means to induce innovation. IP might also be justified as a platform for further innovation by granting a territory, a "prospect," within which the rights holder can invest in further exploitation without fear of being undercut. 21 If monopoly is indeed an inducement or platform for innovation-with or without IP-area B appears in a different light. Before, the transferred profits were treated as either a static harm or neutral transfer. In the dynamic frame, they are also a source of benefit. Area B is a valuable inducement, for which we accept the deadweight loss in area A as a necessary evil.
The pro-monopoly perspective on innovation, as reinforced by IP policy, might be taken to alter the internal cost-benefit calculus of antitrust law. In particular, the dynamic benefits of innovation, to the extent they require protected profits, might furnish an argument for narrowing the scope of antitrust law. The result is to permit a shift from point O to point X that would otherwise be prohibited. As one implementation of this idea, we might selectively permit those practices that have a sufficiently high ratio of increased innovation compared to the reduction in static welfare. 22 Implementing such a calibrated approach would be extremely challenging in practice. After all, selective permission brings no benefit where the innovation would be elicited without the additional reward. 23 Isolating the effect of selective permission is also difficult given the multiple sources of private reward, including IP, first-mover advantages, economies of scale, and network effects. 24 The celebration of monopoly as a source of innovation, moreover, is undercut by important counterarguments. For example, new entrants in a particular market are an important source of innovation-perhaps more important than innovation by incumbents. A substantial literature supports this conclusion. As demonstrated in an influential article by Kenneth Arrow, the incumbent's incentive to innovate is dimmed when the innovation cannibalizes an existing product, compared to the incentive of an outsider with no stake in the status quo. 25 An incumbent also may be hamstrung by its own earlier success. It may prefer to continue selling a high-end product, ignoring down-market demand for a stripped-down offering. 26 Finally, as Judge Learned Hand memorably explained in United States v Aluminum Co of America (Alcoa), monopoly may prove to be a "narcotic":
Many people believe that possession of unchallenged economic power deadens initiative, discourages thrift and depresses energy; that immunity from competition is a narcotic, and rivalry is a stimulant, to industrial progress; that the spur of constant stress is necessary to counteract an inevitable disposition to let well enough alone. Reconciling the pro-monopoly and pro-competition perspectives on innovation is a challenge. One proposed reconciliation is that innovation is maximized at intermediate levels of concentration-that innovation as a function of concentration is an "inverted U." 28 Arguments that competition benefits innovation reinforce the static approach, supplying an additional reason for vigorous antitrust enforcement. These arguments apply whether or not IP is at stake. They might apply most powerfully, however, to certain settings where IP is important. In settings where innovation is to a great degree cumulative, broad rights to early innovators tend to inhibit the development of improvements. This argument, typically presented as a basis for weakening or narrowing IP rights to early innovators, might also suggest a strengthening of antitrust on the margin as applied to incumbents in industries where innovation is cumulative.
This comparison of IP and competition principles has purchase when the defendant has market power, ie, the ability to charge a price higher than the competitive price. 29 In many antitrust matters, market power is a key element of the plaintiff's case. IP is one potential source of market power. A patent might impede other firms' entry into a particular market. Copyrights and trademarks can serve as barriers to entry as well. However, most IP rights fail to confer market power. The legal power to exclude does not necessarily imply the ability to exclude rivals in any economically meaningful sense, any more than the exclusive possession of real property necessarily implies any meaningful market power.
This might seem an obvious point, but US courts had adopted a shortcut by which market power was presumed from the possession of a patent or copyright. This presumption varied in its strength, either furnishing an inference of market power that defendants could rebut with other evidence, such as the demonstrated availability of close substitutes, or else allowing a plaintiff to dispense with a market power showing entirely. 30 In any event, both versions of the presumption were squarely rejected in the 2006 Independent Ink case.
Antitrust challenges against IP rights holders can be divided along the traditional collusion/exclusion dichotomy of antitrust law. Collusion is considered here; exclusion is the subject of Section 4. The discussion of collusion focuses on cooperation by rightsholders and their rivals that reduces competition among them. 32 Examples include cartels, market division arrangements, collusive licenses, and joint ventures.
At the core of antitrust law is a set of prohibitions that exist without any case-specific inquiry into the economic effect of the conduct. For example, cartels that agree on price or quantity, bid rigging, and market division by geography are deemed illegal "per se." These forms of collusion are condemned, notwithstanding the argument that price fixing, if permitted, would induce more innovation in a particular industry. Such an argument is off the table. Per se liability reflects a judgment, often left unstated, that the offset is rare, small where it exists, difficult to reliably determine, and subject to frequent and costly error.
The main alternative regime to per se liability, the so-called "rule of reason," requires plaintiffs to show anticompetitive effect and affords defendants the opportunity to offer procompetitive justifications for the conduct. Many "horizontal" agreements between rivals are considered under the rule of reason, whether or not IP is at issue. So are most "vertical" agreements between sellers and buyers, or IP licensors and licensees.
A recurring issue in IP licensing transactions is a market division by geography or field of use. For example, in Palmer v BRG of Georgia, Inc, the Supreme Court considered an agreement reached between BRG and Bar/Bri, two providers of written materials and lectures to prepare new lawyers for the bar exam. Bar/Bri agreed not to compete in the state of Georgia. BRG took an IP license and agreed not to compete outside of Georgia. 34 The Court condemned this naked market division as a per se antitrust violation. The anticompetitive effect was clear: removal of an existing competitor from the market, with the consequence of higher prices and reduced output. There was no special dispensation for-or indeed significant attention to-the IP license at the heart of the case. In a naked market division such as Palmer, the IP overlay makes no difference.
Not all horizontal agreements, of course, are merely an opportunity to suppress output. Horizontal agreements can have a variety of procompetitive effects. For example, suppose a patentee produces for one set of end-users and licenses production for a second set of end-users to another firm. Rather than taking a competitor off the market (the situation in Palmer), this field-of-use license likely expands output. The agreement is effectively vertical, and vertical IP licensing arrangements are generally procompetitive, encouraging exploitation of a seller's products or IP. This procompetitive effect is emphasized by the IP Guidelines 35 and reflected in case law. 36 One difference between geographical and field of use divisions is statutory. Courts have read a Patent Act provision to expressly permit territorial divisions, 37 notwithstanding the critique discussed in Section 2. There is reason to doubt that this extra statutory protection makes an important practical difference, however, as a naked territorial division among competing patentees likely would be condemned nevertheless. 38
An unusual form of market division is presented by certain settlements of patent litigation. Rather than dividing the market by geography or customer, the division is by time. As the Supreme Court has explained the fact pattern:
Company A sues Company B for patent infringement. The two companies settle under terms that require (1) Company B, the claimed infringer, not to produce the patented product until the patent's term expires, and (2) Company A, the patentee, to pay B many millions of dollars. Because the settlement requires the patentee to pay the alleged infringer, rather than the other way around, this kind of settlement agreement is often called a 'reverse payment' settlement agreement. 39 The two core elements of the reverse payment claim are (1) a large payment, in some form, from the patentee to the alleged infringer, and (2) a requirement that the alleged infringer refrain from competing. For example, a patentee might pay a competitor several hundred million dollars to abandon its proposed market entry and stay out of the market until patent expiration. 35 Reverse payment cases have arisen mainly in the pharmaceutical industry, where a competing, low-price generic drug maker frequently seeks entry prior to the expiration of one or more patents. A reason for its prevalence there is that, to be successful, the reverse payment strategy must overcome an important practical challenge. Where there exist numerous, similarly situated rivals, it is less effective to make payments to just one of them, because the others will enter anyway or demand payments of their own. In pharmaceuticals, competitive entry is costly and often requires specialized knowledge.
In the US, an industry-specific regulatory scheme set up by the Hatch-Waxman Act makes settlement with reverse payment more tempting, because generic challengers are unequally positioned to challenge the patent. First challengers-so-called first filers-have a special incentive to challenge the patent because they are potentially eligible for 180 days of generic exclusivity before other generics can come in. Meanwhile, later challengers are often practically unable to challenge the patent due to a regulatory bottleneck that inhibits approval while the first-filer's exclusivity is pending. Later filers are also discouraged by "acceleration clauses" contained in first filer settlements, which permit the first filer to enter early if the later filer somehow gets to market. These features reduce the incentive of later filers to pursue early entry.
In the US, antitrust challenges by the FTC and private plaintiffs have been filed against reverse payment settlements as to more than 20 drugs. 40 The settlements are not limited to the US, however. The European Commission and UK competition authority have taken action against reverse payment settlements as well. 41 In Actavis, the US Supreme Court held that reverse payment settlements are illegal under the rule of reason if the patent holder is paying to avoid competition. Note that an ordinary patent settlement that contains a compromise on entry dates does not include a payment and does not raise concern under antitrust law. In an ordinary patent settlement, late entry dates delay the consumer benefits of entry, but they are bad for the alleged infringer, too. The infringer therefore can be expected to fight for an earlier entry date. The effect is different when the settlement does not rest solely on the strength of the patentthat is, whatever delayed entry the patentee could secure based on its probability 40 Many of these challenges are discussed in CS Hemphill, "An Aggregate Approach to Antitrust: Using New Data and Rulemaking to Preserve Competition" (2009) 109 Columbia L Rev 629, and CS Hemphill, "Unjustified Delays in Generic Drug Competition", OECD Expert Paper (6 June 2014) 3. 41 As of June 2016, settlements as to four drugs.
of winning the infringement suit. Instead, the patentee makes an additional payment to sweeten the deal. The payment secures less competition than the patentee could expect by asserting the patent alone in litigation or settlement. The payment, in other words, is for additional generic delay, compared to what is legitimately achieved by the patent alone. The logic applies not only to "weak" patents in which the challenger has a high probability of success, but also to patents where the patentee was likely (albeit not certain) to win the patent litigation. 42 In reaching its result, the Court rejected statutory and policy arguments favoring patent exceptionalism. Defendants had advanced a distinctive version of the statutory scope-of-the-patent argument-that a patentee's conduct is privileged within its nominal scope. Nominal scope refers to whatever competition that would be displaced by a patent that is valid and infringed. For example, entry by Company B one day before patent expiration is within the nominal scope; entry one day after is not. The idea was that, unless and until the patent is adjudicated to be invalid or not infringed, the patentee is free to reach market division agreements that replicate the patent's nominal scope.
One of the key policy arguments was the ex-ante inducement account discussed in Section 2. Preserving the patentee's profits through a payment, rather than risking loss in an infringement action, increases the patentee's profits, thereby increasing the ex ante incentive to innovate. An innovator might argue that the additional profits induce enough incremental innovation to make the practice beneficial overall.
In its opinion rejecting these arguments, the Court drew two principles from previous cases. The first was the need for a specific IP provision that limits competition in this particular context. The Court perceived no such provision, effectively rejecting the defendant's argument that the patentee's exclusive right to make, use, and sell the invention provided an adequate legal hook. The statutory presumption of validity 43 did not provide such a hook, either. Indeed, the Court noted a patent law policy that pointed the other way-that invalid patents should be challenged wherever possible, "so the public will not continually be required to pay tribute to would-be monopolists without need or justification." 44 The second principle was whether "competition is impeded to a greater degree" compared to previously approved restraints, essentially an inquiry into economic policy. 45 Here, the Court saw the restraint as quite unlike previously approved restraints, and hence more suspect. On this point, the Court might have added that permitting a settlement with payment would be a relatively ineffective way to confer an additional incentive to innovate. In exchange for receiving a reprieve from competition, the patentee must make a sizable payment. This payment reduces the patentee's profits and thus reduces whatever incremental incentive might be achieved through the settlement.
Other patentees have had greater success in avoiding antitrust liability by invoking the Patent Act. The most controversial example is United States v General Electric (GE). 46 GE licensed a light bulb patent to rival Westinghouse. The license permitted GE to set Westinghouse's resale price. The result was a cartel, as GE established the market-wide price and quantity. Nevertheless, the Supreme Court approved the condition, concluding that the condition was within the patentee's core right to make, use, or sell. 47 The GE rule might be defended as an example of the greater including the lesser. The essential premise is that under the Patent Act, GE is entitled to earn a monopoly profit rather than a competitive profit by asserting a valid and infringed patent and thereby excluding Westinghouse. If so, then perhaps it is free to earn that same profit instead by sharing the market within a cartel set up by license. 48 There might even be a social benefit from such licensing if the licensee has lower costs or superior market access.
Yet the GE rule is problematic where the premise is false. For example, if the patent covers only a minor aspect of the invention, assertion of the patent will not secure monopoly profits. Instead, the licensee can be expected to insist on a low royalty or else avoid the patent at low cost. Lower royalties tend to be passed along to consumers in the form of lower prices. Similarly, if patent 45 The source of this language is United States v Line Material Co 333 US 287 (1948). 46 United States v General Electric 272 US 476 (1926). 47 Simpson v Union Oil interprets the GE result as a "pro tanto" modification of the kind discussed in Section 2.1.
48 S Maurer and S Scotchmer, "Profit Neutrality in Licensing: The Boundary between Antitrust Law and Patent Law" (2006) 8 American Law and Economic Rev 476, defends the GE rule partly on this ground, arguing that a patentee is entitled to the reward it would earn if it were an efficient producer of the invention. validity or infringement is uncertain, the royalty rate must be low, or else the licensee will litigate. Once again, the patentee is unable to achieve monopoly profits through a high royalty. By contrast, the GE rule permits a much higher cartel markup to be charged. 49 GE is a high-water mark for patent exceptionalism. Actavis effectively limited GE to an agreement between a single licensor and a single licensee. 50 This dictum was unsurprising given a series of earlier cases limiting the GE rule and imposing antitrust liability. 51 Moreover, even in the single licensor/single licensee case, the GE rule may not apply when the licensee is the impetus of the restraint, for example, because the licensee seeks a price floor to protect itself from rival price-cutters. 52
One common fact pattern involves blanket licenses that pool the rights of numerous competing rights holders. Songwriters have the exclusive right under copyright law to authorize radio broadcasts and other public performances. 53 Thus, radio stations (among many other entities) require public performance licenses to a large number of works. To handle this demand, songwriters grant licenses to an intermediary, a performing rights organization such as the American Society of Composers, Authors and Publishers (ASCAP). (ASCAP is not the only US performing rights organization, but I ignore the others for simplicity.) ASCAP pools individual licenses and repackages them as a blanket license that is made available to radio stations and other public performers. The blanket license is a valuable new good, not possible without collective action, and offers immediate and easy access to a large repertoire of songs. ASCAP sets the price for the blanket license.
49 Hovenkamp (n 4) 525-527. Maurer and Scotchmer offer an alternative, innocent explanation that a GE-like condition, setting a price for the licensor and licensee alike, controls licensor opportunism.
51 For example, in the Line Material case, the Supreme Court declined to apply the GE rule to multiple cross-licensing patentees, noting that "[n]othing in the patent statute specifically gives a right to fix" a licensee's sale price-an argument that applies to GE too. For additional limitations on the GE rule, see Hovenkamp et al (n 1) § § 31.2-3; American Bar Association, Intellectual Property and Antitrust Handbook (2nd ed, ABA 2015) 78-80. 52 Interstate Circuit, Inc v United States 306 US 208 (1939). Interstate Circuit is a copyright case, and the licensor and licensee there were not competitors, but the point is general.
At the same time, the decision to collectively offer and price the works raises an antitrust claim of horizontal price fixing among the rights holders. The blanket license reduces price competition among the songwriters. The concern is that ASCAP might set a higher price to (say) broadcasters than if each songwriter set its price independently. Note that this concern is avoided if each rights holder sets its own price, and the organization simply administers the collection of fees. An example of this alternative approach is the Copyright Clearance Center, which administers copyrights for academic journals in the Due to the risk of collusion, ASCAP has operated within the restrictions of an antitrust consent decree since 1941.
In the early years of ASCAP's operations, the collusion problem was particularly severe. Songwriters granted exclusive public performance licenses to ASCAP, which eliminated price competition among the songwriters entirely. Under pressure from the Justice Department, ASCAP shifted to nonexclusive licenses, thereby permitting songwriters to negotiate individually with music users. 54 Enforcing nonexclusivity has been the subject of subsequent action by the Justice Department. 55 Even without exclusivity, the blanket license makes price competition less valuable to and hence less vigorous among songwriters. The availability of a blanket license dampens, though it does not prohibit, price competition among songwriters for per-song, per-use negotiations with users.
One important antitrust question frequently presented under the rule of reason is whether an alternative arrangement exists that achieves the benefits of centralized administration without suppressing price competition among songwriters. For example, could ASCAP offer a per-song, per-use option to music users that preferred the à la carte approach, and if so, should antitrust require that? In a series of cases in the 1970s, courts declined to require performing rights organizations to offer a per-use option, in part because it was too difficult at the time to monitor compliance with per-use licenses. 56 Given subsequent changes in the frontier of practicality, an open question is whether an antitrust challenge, if pursued today, might yield a different result. 57 Music users' demand for access to a large catalog of songs affects the antitrust treatment of music copyrights in a second way. A typical piece of recorded music is covered not only by a copyright in the composition, discussed earlier, but also a copyright in the sound recording. In the US, sound recordings enjoy a limited public performance right, restricted to digital distribution. Thus, digital distribution platforms, unlike radio stations, must secure a license to the sound recording. 58 The ownership of sound recordings has become more concentrated over time. Today, the most valuable sound recordings are held principally by three "major" music labels: Universal, Sony, and Warner. An important antitrust question is whether the mergers producing this degree of consolidation should have been permitted, given concerns about higher prices charged to licensees. A further concern is the enhanced ability to act as "kingmaker," potentially suppressing innovative business models.
In response, music labels have argued that although they are competitors for some purposes, they are offering complementary products when it comes to digital distribution through an interactive service. A distributor's access to one label's music makes access to the others more valuable. In fact, each is effectively a must-have, because a user of an interactive service expects music from all three labels. When complements are separately owned, each firm charges a higher price, failing to fully take into account the prices charged by the other firms. 59 The result is higher prices and inefficiency. A standard result in antitrust economics is that common ownership of complements is beneficial to remove this effect. Thus, a merger of two major labels might be viewed, seemingly paradoxically, as good news in this respect. An argument of this kind 56 BMI v CBS, Inc 441 US 1 (1979); BMI v Moor-Law, Inc 527 F Supp 758 (US District Court (District of Delaware) 1981).
57 For further discussion, see CS Hemphill, "Less Restrictive Alternatives in Antitrust Law" (2016) 116 Columbia L Rev 927. 58 The discussion in text applies primarily to interactive streaming services (such as Spotify) that must negotiate for these rights. Other services are subject to a compulsory license or are exempt.
59 For a discussion of this "double marginalization" issue in the context of IP rights, see C Shapiro, "Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard Setting" in AB Jaffe, J Lerner & S Stern (eds), Innovation Policy and the Economy 1 (MIT Press 2000) 119.
appears to have supported the FTC's decision not to challenge a merger that reduced the number of major labels from four to three. As the FTC explained, both merging labels were already must-haves for interactive streaming services, so consolidation would not make matters worse (and indeed, in theory, might help). 60 Complementarity is also a crucial starting point for understanding patent pools. Patent pools bring together multiple technologies in a single convenient package. Where the patents are complementary-for example, both patents are necessary to the implementation of a technology-their joint provision is highly procompetitive. Inclusion of a substitute patent, by contrast, raises serious concerns by suppressing competition between rival technologies. Patent pools may also exclude would-be licensors (from inclusion in a standard) and licensees (from use of the technology), issues taken up in the next section.
This section considers exclusion claims, the other major branch of antitrust law. The plaintiff asserts that an IP right holder has engaged in improper selfentrenching conduct. The bounds of propriety, however, are often unclear. After all, the ordinary successful assertion of a patent or copyright excludes would-be competitors. It would eviscerate IP law if such assertions frequently triggered an antitrust violation.
Some cases avoid the dilemma. For example, the antitrust defendant might lack a valid and infringed IP right in the first place, yet press a nonmeritorious IP claim. Or an IP rightsholder with a valid right might overreach, suppressing competition as to a second product not covered by the IP right. This section reviews these scenarios before turning to cases that present the dilemma in a sharper fashion.
One set of exclusion claims asserts that an IP holder has suppressed competition by falsely asserting an IP claim, either by pursuing "sham litigation" or in its acquisition oft the IP right. For example, in Professional Real Estate Investors v Columbia Pictures Industries, 61 videodiscs of copyrighted movies to hotel guests to watch in their rooms. Motion picture studios held the copyrights, which they had licensed to a payper-view cable system that competed with the rental scheme. The studios filed suit for copyright infringement, asserting a violation of their public performance rights. Courts eventually rejected the copyright claim. The defendant counterclaimed in antitrust, asserting that the copyright suit was designed to exclude its competing rental operation.
The Supreme Court denied antitrust liability, insisting that the suit must be not only subjectively motivated by exclusion of a competitor, but also "objectively baseless." Simply losing the copyright suit was not enough for objective baselessness. The case reflects a concern, rooted in the constitutional right to petition, for ensuring that rights holders have an adequate opportunity for redress.
Turning to IP acquisition, the Walker Process case holds that if the patent (or, in rare cases, a copyright) was acquired by fraud, its subsequent assertion to exclude rivals can be the premise for an antitrust claim. 62 Sometimes Walker Process claims are paired with sham litigation, because the patentee's assertion against a rival takes the form of a lawsuit. But not all Walker Process cases involve litigation, and in any event the technical focus is different-the bad conduct is in the acquisition of the patent, rather than the conduct of the litigation. Also required in these cases is a demonstration that the conduct was actually capable of harming competition, a showing of market power.
Walker Process claims pose an unusual risk to a patentee. They allow a plaintiff to challenge the validity or enforceability of a patent, even if the plaintiff lacks standing to bring such a challenge under patent law. For example, a direct purchaser of a patented good lacks standing to challenge the patent under patent law. By contrast, antitrust law confers standing on direct purchasers, including in Walker Process cases. 63
A second set of exclusion cases involves leverage-the worry that an IP holder can use an IP-protected monopoly in one market to gain an improper advantage and weaken competition in a second market. The classic fact pattern involves tying, wherein a firm conditions the purchase of one product or service on the purchase of another. The condition may be accomplished through agreement: "if you want my patented product A, you must also take my product B." A similar effect may be achieved by offering the product only in a package, including by inseparably fusing them together.
A famous example of tying is International Salt Co. v United States. 64 International Salt leased a patented machine called the Saltomat used in the canning process to inject salt into the canned product. The company required a customer, as a condition of leasing the Saltomat, to buy all its salt tablets from International Salt as well. The Supreme Court condemned this arrangement as an illegal tie. Although International Salt was permitted to restrain others from making, vending, or using the patented machines, . . . the patents confer no right to restrain use of, or trade in, unpatented salt. By contracting to close this market for salt against competition, International has engaged in a restraint of trade for which its patents afford no immunity from the antitrust laws. In the Court's view, the impact on a second market took the conduct outside the scope of the patent.
Under somewhat dated judicial precedents, tying is subject to condemnation without full rule of reason evaluation, provided several technical conditions are met. However, tying is often beneficial, resulting in net efficiencies through cost savings or quality improvement. Meanwhile, concerns about leverage and resulting foreclosure are often overblown. Given the frequent presence of procompetitive justifications and frequent absence of any motive to gain an advantage in a second market, commentators have suggested that tying should receive more lenient treatment.
The so-called "one monopoly profit" result dispels some of the concerns about leverage. 65 For example, suppose that the competitive price for a one-year lease on a Saltomat machine and a year's worth of salt is $200 each, and a profit-maximizing firm selling both machines and salt would charge $1000 for the combination. The total profits for the monopolist are $600. Meanwhile, if the salt market is competitive, a Saltomat monopolist could simply charge $800 for the machine, again earning $600. In this example, the Saltomat monopolist earns nothing by selling salt as well. There is no reason to tie, because the monopolist earns a single monopoly profit either way. 64 International Salt Co v United States 332 US 392 (1947).
65 WS Bowman, "Tying Arrangements and the Leverage Problem" (1957) 67 Yale LJ 19.
One motivation for tying is price discrimination. Customers vary in their intensity of use. One Saltomat user might do a lot more canning than another. Alternatively, users might be uncertain about their intensity of use, and prefer to "pay as they go" through metered pricing. Often, the producer sells the tying product at cost, thus promoting adoption and earning all its profits from the tied product. The welfare effects of price discrimination are complex; depending on the facts, permitting discrimination can either increase or decrease output.
The one monopoly profit result does not always hold. For example, leverage can result in foreclosure where there is independent demand for the tied good without the tying good, and some prospect of increasing market power in the tied good market by depriving sales to a rival. 66 To return to the Saltomat example, there is clearly separate demand for salt not used in conjunction with a Saltomat. The foreclosure story founders, though, on International Salt's inability to deprive rival salt sellers of economies of scale by retaining "captive" salt customers who need the Saltomat. Control of the second market is also a form of defense. The incumbent is not seeking extra profits in the second market, but instead seeks to make entry in the first market more difficult by forcing a rival to enter into both at the same time.
One variant of tying cases arises uniquely in the IP context: when both products are IP rights, and the rights are only offered together in a package. Examples include ASCAP's blanket license and package licenses of patents. Such contracts may produce important efficiencies. For example, licensing all the rights may be cheaper than licensing just a few if piecemeal monitoring is expensive. These arrangements have generally received lenient treatment. 67 An exception is the "block booking" of a slate of copyrighted films licensed by a movie studio to an exhibitor, which has been prohibited per se on concerns about leverage, despite procompetitive explanations offered by commentators such as economies of scale in selling the package. 68 66 MD Whinston, "Tying, Foreclosure, and Exclusion" (1990) 80 American Economic Rev 837. 67 For an argument that blanket licenses can suppress competitive entry into the provision of intermediary services, see RC Picker, "Unbundling Scope-of-Permission Goods: When Should We Invest in Reducing Entry Barriers?" (2005) 72 University of Chicago L Rev 189.
68 FA Hanssen, "The Block Booking of Films Reexamined" (2000) 43 Journal of Law and Economics 395.
A third category of exclusion takes the form of a refusal to deal with a rival-either a refusal to license IP or a refusal to sell a product or service containing the IP. A rights holder's refusal is, in economic terms, closely related to a tie. For example, suppose a machine manufacturer refuses to sell a patented replacement part to an independent servicer of the machines, and as a consequence the servicer is unable to service the machines. The concern is that the manufacturer has used its patent on the part to gain an advantage for its own servicing business. Both the prospect of foreclosure and the various responsesthe one monopoly profit critique, efficiencies, and opportunity for price discrimination-mirror the discussion of tying in Section 4.2.
Despite these similarities, antitrust law is more lenient toward refusals. Under the Trinko case, a refusal to deal is not generally actionable. 69 Monopoly is not a status offense. A monopolist must engage in affirmative anticompetitive conduct to be liable for an antitrust violation. Trinko also cast a skeptical eye toward essential facilities claims, an antitrust doctrine requiring access to the withheld facility under certain circumstances, stating that the Supreme Court had never recognized such an antitrust claim.
The reason for leniency is reflected in the quotation from the Trinko case discussed in Section 2-that the prospect of monopoly attracts risk-taking and "business acumen." As Judge Hand explained in Alcoa-shortly after decrying monopoly as a narcotic-monopoly is justified as a reward for "skill, foresight and industry" in order to induce innovation. This justification tolerates some loss of static consumer welfare: "the successful competitor, having been urged to compete, must not be turned upon when he wins." This result thus reflects the attention to dynamic effects and ex ante incentives discussed in Section 2. This safe harbor for incumbents is not limited to IP.
A significant exception arises from selective refusals, in which the incumbent sells a good or service to some customers but not others. 70 In particular, an incumbent might decide to make profitable sales to ordinary customers that are not competing with it, yet withhold the otherwise generally available product from a competitor. Such selective withholding is evidence of a profit sacrifice-ie, that dealing would be profitable aside from the tendency to foreclose-and liability is possible if other elements of an antitrust claim are met.
However, under the Xerox case, patent law arguably expands the incumbent's insulation from antitrust attack. 71 Copier maker Xerox refused to sell patented parts to independent servicers, which prevented them from competing with Xerox's own servicing business. The Federal Circuit acknowledged the possible advantage gained in the service market due to this conduct, but nevertheless denied liability. In support of its conclusion, the court noted the Patent Act provision exempting refusals to license from a finding of "misuse or illegal extension of the patent right." The court reached the same conclusion as to copyrighted software, pointing to the copyright holder's exclusive right to distribute copyrighted works. 72 Xerox's refusal was selective, as it continued to sell parts to end-users while refusing to sell to independent servicers. Given that selective refusals to deal are not ordinarily privileged, Xerox arguably adds selective refusals to the set of permissible conduct for IP holders. This is a significant expansion of the incumbent's opportunity to use an IP right to gain an advantage in a second market. It bears emphasis that Xerox is a lower-court case that may conflict with other opinions more inclined toward liability. 73 It is unclear how the Supreme Court might ultimately resolve the tension.
This result creates a difference between a conditional refusal by an IP holder, as in a tying case, and a selective refusal. The right to refuse does not imply a right to conditionally refuse, because such an argument "fail[s] to distinguish between the rights which are given to the inventor by the patent law and which he may assert against all the world through an infringement proceeding and rights which he may create for himself by private contract which, however, are subject to the rules of general, as distinguished from those of the patent, law." 74 The restrictive private contract, an important hook for antitrust enforcement, is missing from a selective refusal, opening the path to argue that selective refusals constitute a lesser included statutory right.
Beyond selective refusals, there is an argument that Xerox goes further. The court expressly acknowledged a patentee's potential antitrust liability for sham litigation and fraud on the patent office (discussed earlier in Section 4.1). The court also recognized a further category of liability, associated with the Supreme Court's statement in another case that "power gained through some natural and legal advantage such as a patent, can give rise to liability if a seller exploits his dominant position in one market to expand his empire into the next." 75 That statement might be read as a restatement of the "Leverage and Foreclosure" category discussed in Section 4.2. However, Xerox hinted that this type of claim might be limited to tying, 76 which leaves out leverage and foreclosure accomplished by other means, such as exclusive dealing. 77 Overall, however, such a drastic contraction of antitrust liability seems unlikely. 78 In Europe, by contrast, a refusal to license IP may be subject to liability under certain circumstances, even if the refusal is unconditional and nonselective. For example, in the Magill case, Magill sought to develop a new television guide that integrated the copyrighted listings of all three major Irish TV networks. 79 The networks refused to license the listings. The European Court of Justice affirmed the Commission's determination that, under "exceptional circumstances," a refusal is an abuse of a dominant position under European competition law.
Later cases such as IMS and Microsoft help specify what counts as exceptional circumstances: an unjustified refusal that blocks the entrant's development of a new product or "technical development" that is feasible only with a license. 80 The focus is on innovations, not just copycats of the goods of the IP rights holder. Recall the arguments canvassed in Section 2, that outsiders are a potent source of innovation, as a basis for expansive competition-law 75 Eastman Kodak Co v Image Technical Services Inc 504 US 451 (1992) (internal quotation omitted). 76 For example, the court emphasized that the quotation came from a tying case (also involving copiers), and at another point restated the categories of liability as sham litigation, fraud on the patent office, and tying. 77 For example, a seller might engage in exclusive dealing by contractually requiring that the buyer purchase no products from the seller's rival. 78 Among other problems, the rule would fall afoul of the Clayton Act, which imposes liability for exclusive dealing, even as to patented goods: see 15 USC § 14. enforcement and narrower IP rights. The emphasis on innovative outsiders, as a premise for the incumbent's liability under IMS and Microsoft, is an implementation of the Arrovian perspective.
At the same time, there may be space under European law to balance that effect against the potential negative effect on the incumbent's innovation incentive. In Microsoft, the European Commission acknowledged the possibility that negative effects on the ex ante incentive to innovate might justify the conduct and thereby furnish a sufficient basis to withhold liability. The Commission's approach thus took the form of a "balancing incentives" approach, weighing and comparing the incentives furnished to each. 81 The Court of First Instance avoided considering this approach by recharacterizing what the Commission had done. 82 Thus, although Microsoft has received attention primarily as an expansion of liability, it contains the seeds for the Commission to refrain from challenging conduct on appropriate facts.
A fourth category of exclusion cases features a newly designed product that is incompatible with a rival's product, thereby excluding the rival from the market. The simplest case is the introduction of a new and improved product. For example, consider the antitrust case brought against Tyco, a manufacturer of medical monitors used in combination with sensors. 83 Tyco introduced a new monitor and sensor design, in which part of the functionality was shifted from the monitor to the sensor. The plaintiff manufactured generic sensors that were compatible with the old design but not the new one. An appeals court rejected antitrust liability, concluding that merely introducing an improved product raises no antitrust issue even if it excludes rivals. This conclusion is not specific to patents. It applies to any genuine new product innovation, though a patent might be evidence that a genuine improvement exists.
A more complex case is presented when the introduction of the improved product is accompanied by additional conduct that induces or even forces consumers to accept the new product. Discontinuation or withdrawal of the old product is a common fact pattern. In the medical monitor case, Tyco withdrew the old technology when it introduced the new monitor and sensor technology. The court concluded that Tyco had not "effectively force[d]" consumers to take the new product, because it faced substantial competition from other monitor makers, and hence was not liable.
These issues have repeatedly arisen in the pharmaceutical industry. Branded drug makers frequently develop or acquire new versions of an existing drug, such as a tablet instead of a capsule. They may also develop (or acquire) alternative therapies in the same therapeutic class. Where the new product has stronger or longer-lived exclusivity, the branded firm has an incentive to shift patients and doctors from the old product to the new one, rather than face competition from generic drug makers on the old product. A key element of competitive entry is the ability of the generic firm, under US drug regulatory law, to "piggyback" on the branded firm's new drug approval. Moreover, state law requires or permits substitution of the generic drug, and substitution is further encouraged by private payors.
A "product switch" or "product hop" by the brand can be accomplished in various ways. One strategy is to withdraw the old product from the market entirely, and can include withdrawing regulatory approval or repurchase of inventory of the old product. One "softer" strategy is to heavily promote the new product, through marketing and detailing visits to doctor's offices, while keeping the old product on the market. Or a firm might set a high price for the old drug, relative to the new one, to encourage a switch. A successful hop leaves the generic drug with relatively few customers. Product hops are most effective when they occur prior to regulatory approval of the generic version of the old drug. In that case, there is no foothold for automatic substitution to take place with a pre-existing base of consumers.
A recent example concerns an Alzheimer's drug called Namenda. The branded drug maker sought to switch patients and doctors from a twice-a-day version of Namenda, which faced imminent generic entry, to a once-a-day extended release version. At first the drug maker encouraged adoption of the new drug through "soft" promotional efforts, but when that strategy failed, it announced an effective withdrawal of the old drug. A US appeals court concluded that the hard switch, even to an improved product, constituted illegal exclusion. The key issue for the court was consumer coercion that resulted from withdrawing the old product. According to the facts found, the withdrawal was profitable only by virtue of preventing generic competition, which revealed bad intent on the drug maker's part. 84 What about the Patent Act? The defendant invoked both the patentee's exclusive right to "mak[e], us[e], . . . or sel [l]" and the refusal-to-license provision successfully relied upon in Xerox. The court rejected this argument. The court concluded that the defendant was not only exercising a patent right, but combining it with conduct outside the patent grant, namely new product introduction. Hence the combination was outside the patent's scope. A second problem was that the combined effect of the conduct was to achieve a temporal extension of the patent, thereby evading a key limitation of the patent right.
A final area of unilateral exclusion arises in the context of standard-setting organizations (SSOs). An SSO is an industry group that sets the technical specification for a product or service, such as mobile wireless service. The existence of an SSO is generally procompetitive by promoting interoperability. Interoperability gives rise to network effects from common usage, economies of scale, and competitive markets for components.
At the same time, certain conduct by SSO members raises antitrust concerns. One prominent allegation is that a technology contributor manipulates the standard-setting process by misrepresenting the existence or licensing terms of pertinent IP rights. For example, a rights holder might fail to disclose that a patent covers a technology proposed for inclusion in the standard. Or the rights holder might falsely represent its willingness to license the patented technology on "fair, reasonable, and nondiscriminatory" (FRAND) or "reasonable and nondiscriminatory" (RAND) terms. Once the technology is adopted, the rights holder is able to "hold up" technology users and thereby extract a higher price.
Where deception or misrepresentation is alleged, the clearest case for an antitrust claim arises when the SSO would have chosen a different technology, but for the misrepresentation. A recurring question is whether a more limited effect suffices to state a claim-that the firm extracts a higher price than contemplated during the standard-setting process, even though the technology would have been adopted in any event. US courts appear to be split on the question of whether a different choice of technology is an essential element of the claim. 85 Even if a mere extraction of a high price falls short of an antitrust claim, it may be actionable as a form of fraud or breach of contract. The availability of these alternative claims depends upon, inter alia, the nature of the IP holder's representations to the standard-setting process and the specificity of required representations as set out in the SSO's rules.
One common fact pattern has arisen repeatedly in the smartphone industry, particularly in the ongoing patent war between Android and Apple. A patent holder agrees to FRAND terms and has its technology incorporated into the standard. Despite the FRAND commitment associated with the "essential" patent, royalty negotiations with a user of the technology break down, and the patent holder seeks an injunction or exclusion order. The patent holder's ultimate goal is to extract a higher royalty 86 or, potentially, to foreclose a rival.
Seeking an injunction might be regarded as a form of deception (because the rights holder misled the SSO as to its willingness to license), or else as a breach of the FRAND commitment and hence a form of ex post opportunism. In the Huawei case, the European Court of Justice concluded that seeking an injunction against a willing licensee can constitute an abuse of dominance and identified circumstances where such conduct could be improper. 87 In the Motorola Mobility matter, the FTC secured an SEP holder's commitment not to pursue injunctive relief. 88 Both Huawei and Motorola Mobility look to neutral third parties to set the license terms if the parties cannot agree, rather than seeking an injunction. Meanwhile, commentators have offered proposals to alter the terms of SSOs to limit the pursuit of injunctions. 89 Appeals (3rd Circuit) 2007) does not appear to insist that the choice of a different standard was essential. On the other hand, the Broadcom allegations satisfied the Rambus causation standard, because plaintiff alleged that the SSO would not have adopted the technology, absent the deception. 86
All of the exclusion claims so far have focused on exclusion by a dominant firm. But exclusion can instead be accomplished by multiple non-dominant firms acting in concert. For example, the members of an SSO can manipulate the standard-setting process to choose one standard over another. In Allied Tube, steel conduit makers faced a sharp challenge from an alternative, plastic polyvinyl chloride (PVC) conduit, which was superior in important respects. PVC makers sought incorporation into a widely used building standard. When the matter reached a vote, steel interests packed the meeting with new members to vote against the inclusion of PVC. Here, the anticompetitive intent and effect were clear. In other cases, it is difficult to determine whether the choice was made for a valid technical reason or instead to preserve incumbency. 90 A second form of collective exclusion is boycott. Consider, for example, the Fashion Originators' Guild of America (FOGA) case. 91 In the 1930s, US fashion designers were afflicted by "style pirates," who copied and sold versions of original designs at a much lower price. These designs were not protectable under copyright. To make up for the lack of protection, the designers set up a private IP system. Designers registered their creations with the Guild, which monitored some 12,000 cooperating retailers who had agreed not to buy knockoffs. The Guild developed a sophisticated process for identifying violations, and offending retailers were boycotted by Guild members. The Supreme Court condemned the arrangement as an illegal restraint of trade. 92 The Court refused to consider the Guild's argument that the scheme was necessary to make up for the lack of IP protection. 90 SSOs pose other potential harms as well. [1] Rather than excluding certain technologies, the SSO might exclude certain types of licensees. Hence a major focus of competition policy toward SSOs is that they license all would-be licensees. [2] SSO rules prohibiting the assertion of IP rights might themselves be exclusionary. [3] Beyond collective exclusion, the existence of the SSO poses some risk of cartelization, while curbing the risk of monopolization. For a detailed account of SSO competition claims, see MA Lemley, "Intellectual Property Rights and Standard-Setting Organizations" (2002) 90 California L Rev 1889.
91 CS Hemphill and J Suk, "The Fashion Originators' Guild of America: Self-Help at the Edge of IP and Antitrust" in RC Dreyfuss and J Ginsburg (eds), Intellectual Property at the Edge: The Contested Contours of IP (Cambridge University Press 2014) 159.
92 Fashion Originators' Guild of America v FTC 312 US 457 (1941). A Korean design registration system for kitchenware was permitted in Metro Industries v Sammi Corp 82 F 3d 839 (US Court of Appeals (9th Circuit) 1996) on the ground that plaintiff failed to demonstrate any anticompetitive effect.
In the FOGA case, the agreement was explicit. Consider, by contrast, several excluders acting in parallel without explicit or provable agreement among them. No single firm is dominant, as is generally the case for monopolization or abuse of dominance liability. An agreement among the excluders, an important premise for restraint of trade cases, is also missing. 93 A potential gap in enforcement is therefore presented by parallel exclusion by multiple non-dominant firms. 94 European law recognizes liability for "collective dominance," but collective dominance is only seldom applied to collective exclusion. 95 Instead, each firm is treated as individually dominant. Magill is an apt example. The Commission took the view that each of the networks had a dominant position, instead of sharing a jointly dominant position. 96
The previous sections address how IP law alters the ordinary operation of antitrust law. We may also ask reciprocally how competition principles limit the exercise of IP rights. Historically, these limits have been set not only by antitrust law but also by IP law, which incorporates competition principles through the doctrine of misuse. Misuse is a defense to enforcement of an IP right where the rights holder seeks to evade the limits of the right. The consequence of a misuse finding is unenforceability of the right, without any additional liability for damages. In this sense, misuse is antitrust law "lite."
There has been a steady back-and-forth between antitrust law and misuse, as concepts from one are brought over into the other. Today, the two are in rough congruence and enforce similar limits on the scope and duration of the IP grant. In principle, misuse places some additional limits on the rights holder that go beyond antitrust law. This broader coverage is perhaps unsurprising given 93 In some instances, a contract between the excluder and a customer or supplier satisfies the legal requirement of agreement.
94 For an extended discussion of collective, parallel exclusion, see Hemphill and Wu (n 17). 95 misuse's lesser consequences. 97 However, in the patent context, where misuse has received the most attention, there is significant and steady pressure to bring misuse and antitrust into even closer alignment.
The most important example of the back-and-forth is in tying. Tying doctrine began as a form of misuse. For example, in the Motion Picture Patents case, 98 a patent pool led by the Edison interests held a patent for threading film into a movie projector. The pool conditioned use of the projector upon use of Edison film. The Supreme Court held that the patent was unenforceable until the misuse was purged. 99 Patent law's internal hostility to evasions of scope has been incorporated into antitrust law. In this respect, IP law has intensified antitrust enforcement, rather than limiting it. One further aspect of the misuse cases has been to treat, as improper extension, conditional sales that serve the goal of profit extraction, rather than exclusion. For example, in Mercoid Corp v Mid-Continent Investment Co., the patent covered a combination of unpatented elements. 100 Identifying infringement by users was difficult, so the patentee sought to earn profits by controlling sales of one of the elements, a switch especially suited for use in the patented combination. Accordingly, it sued a competing switch maker for contributory infringement of the patent. The Supreme Court condemned this strategy as misuse, viewing it an improper extension of the patent to condition in this manner the use of the patent on use of an unpatented product. This extraction strategy was explicitly restored by statutory amendment. Today, selling a "nonstaple" article-an article whose primary use infringes the patent-subjects the seller to contributory infringement liability, and a patentee is free to use this strategy. 101 An important source of convergence between misuse and antitrust is the requirement that a patentee must have market power in the tying market to be found guilty of misuse, and the mere possession of a patent does not establish market power. The insufficiency of patent possession was established first in misuse. 102 As noted in Section 2, the later Independent Ink case recognized a similar limitation in antitrust cases.
Misuse is persistently broader than antitrust in its hostility to a particular form of temporal extension. Under the Brulotte rule, a patentee may not contract for royalties after patent expiration. 103 The rule was originally based on competition policy concerns. The concern was that such a rule would permit an end-run on the limited term, thereby restricting competition beyond the expiration of the patent. This scheme appeared to confer a kind of leverage from the years of the patent term into the years beyond the patent term.
The Brulotte rule is misguided as a matter of policy. First, anticompetitive extension is implausible. Generally speaking, the patentee only possesses a certain amount of market power by virtue of a patent with fixed duration. A longer royalty just spreads the profits more thinly over a longer period. That arrangement, moreover, can have a procompetitive effect, by aligning incentives between licensor and licensee, reducing the initial outlay for a cash-strapped licensee, and in some cases reducing the total amount of deadweight loss. 104 In 2015, the Supreme Court declined to overturn the Brulotte rule. The Court resisted two sorts of pressure for convergence. First, the patentee had sought antitrust-like treatment, a case-by-case ("rule of reason") approach instead of a flat ban, in accordance with the antitrust-like basis for the rule. Second, the patentee had urged a relaxed approach to stare decisis, in accordance with antitrust cases such as Independent Ink that alter the law to better match current economic understanding. The Court responded that for patent law, unlike antitrust law, economic policy considerations had less force, and the pull of stare decisis was more powerful.
This chapter has identified several limits that IP places on the ordinary operation of antitrust law. These limitations are fairly narrow, and largely limited to patent law. For example, under the GE rule, a patentee may specify the resale price of a single competing licensee. Selective refusals to license are permitted under Xerox, which goes beyond the latitude already granted to nonselective refusals to deal. 104 For an explanation of the last benefit, see I Ayres and P Klemperer, "Limiting Patentees' Market Power Without Reducing Innovation Incentives: The Perverse Benefits of Uncertainty and Non-Injunctive Remedies" (1999) 97 Michigan L Rev 985.
Meanwhile, antitrust law constrains the exercise of IP rights. The main effect, which antitrust law shares with misuse doctrine, is to rigorously enforce limitations on the scope of the grant. Clear, specific grants of authority are preserved against liability, but not accompanying conduct that lacks a similarly explicit endorsement. A rights holder may not condition access to its right to gain an advantage in a second market, as in tying cases. Nor may it outstay its welcome by restraining competition temporally past its inherent force, as in the reverse payment cases. Evasion of internal IP limit tends to favor antitrust liability, and as the development of tying doctrine suggests, may even amplify it.
The inclusion of IP policy in the internal cost-benefit calculation of antitrust law has not produced a sharply distinctive body of IP-specific antitrust law. Instead, we see substantial harmonization. As to exclusion, antitrust has internalized the argument about ex ante incentives to innovate, an argument not unique to IP, as exemplified in Trinko; this has encouraged harmonization as to exclusion. As to collusion, substantial harmonization has been achieved by largely rejecting the application of IP exceptionalist arguments, either as a matter of statutory interpretation or economic policy.