ANTITRUST ANALYSIS OF FRAND LICENSING POST-FTC v. QUALCOMM

By Aminta Raffalovich and Steven Schwartz¹

A great deal of recent attention is being paid to the Ninth Circuit’s decision in Federal Trade Commission ("FTC") v. Qualcomm.² On one side of the debate are those, like the FTC and various amici, who reject the view expressed by the Ninth Circuit and consider the licensing behavior by Qualcomm as a quintessential antitrust violation. That argument views Qualcomm’s licensing behavior as a blatant example of the illegal exercise of monopoly power.

On the other side of the debate are those, like the Antitrust Division of the Department of Justice and still other amici, who decry the District Court’s decision³ as an inappropriate extension of the antitrust laws into a non-antitrust arena and applaud the Ninth Circuit’s decision as drawing reasonable boundaries around the application of antitrust laws.

What complicates the assessment of the decision in Qualcomm is the context in which the behavior at issue occurs. The patent overhang and, in particular, the standard-setting and essential-patent context make the analysis different—and arguably more challenging—than a typical analysis of pricing or licensing behavior in a vertical setting.

However, while arguably more challenging, the fundamental elements of the analysis are largely the same as for any other antitrust analysis. The analysis necessarily includes an assessment of the extent of the market power held by the target firm, whether the market power flows from an effort to monopolize the market, the impact of the behavior on competition, and other familiar questions.

I. WHAT IS MARKET POWER?

Our discussion begins with market power.⁴ It is axiomatic among antitrust practitioners on the economic and legal sides that an assessment of the extent of the market power enjoyed by the relevant entities is central to an antitrust analysis. Unfortunately, there is often confusion about what market power means, and what it does not. To economists, market power is a non-pejorative term that simply means that a firm can influence the amount of goods (or services) it sells by altering the price of the product or service.⁵ Noting that a firm has some degree of market power is not, by itself, a negative statement because any time a firm faces a downward sloping demand curve, i.e., a situation in which a price increase leads to a reduction in quantity demanded and vice versa, that firm has some market power. Since most goods are sold in markets in which firm demand curves slope downward, most sellers, therefore, have at least some market power. From an antitrust perspective, the issue is whether a firm has so much market power—call it "monopoly power"—that the firm has the ability to sustain prices above and output below competitive levels.⁶ Thus, a key question in an antitrust analysis is whether a firm (or collection of firms) has sufficient market power to rise to the level of "monopoly power."

[Page 138]

From an economic perspective, market power manifests itself in the ability to set price above marginal cost—that is, the cost incurred to produce an additional unit of a good (or service). In a perfectly competitive market, where a seller cannot influence the quantity they sell by adjusting the selling price, the sales price is equal to marginal cost. In that situation, the demand curve faced by the seller is horizontal, and not downward sloping, and the profit-maximizing price will be equal to marginal cost.⁷ Accordingly, if a firm’s price is above its marginal cost, it is not operating in a perfectly competitive market.⁸ The firm’s demand curve will slope downward and, hence, it has market power.⁹

[Page 139]

In practice, of course, markets generally are not perfectly competitive. Economic evidence and antitrust tradition acknowledge this. For this reason, there is generally no expectation that prices will be set to be equal to marginal cost and, thus, the mere fact of a deviation of price from marginal cost is not cause for antitrust concern. Typically, it is only deviations of price from marginal cost that are substantial and persistent that are issues for antitrust analysis. Thus, from an antitrust perspective, it is a reasonable shorthand to be concerned with "market power" that results in the ability to set and sustain price substantially above marginal cost, i.e., monopoly power. As stated by an Assistant Attorney General for Antitrust:

Antitrust enforcers are not in the business of price control. We protect competitive process, not a particular result, and particularly not a specific price. In fact, if a monopoly is lawfully obtained, whether derived from IP rights or otherwise, we do not even object to setting a monopoly price.¹⁰

This is a crucial point: it is when monopoly power is obtained or maintained unlawfully that monopoly prices are a matter for antitrust inquiry. If that monopoly power is obtained or maintained through the exercise of superior business skill or acumen, for example, or through effective and thoughtful innovation that results in a competitively superior product, the resulting exercise of market power—even monopoly power—is not an issue of antitrust concern. This point is especially critical to remember in the context of lawful monopolies, e.g., a monopoly over a patented technology.¹¹ This is where the analysis of Qualcomm begins.

II. THE RELATIONSHIP BETWEEN PATENTS AND MARKET POWER

Patents present a challenging situation in economics and antitrust precisely because a patent confers on the patent owner a legal monopoly on the technology taught in the patent. That means that no entity is allowed to practice that technology without a license from the patent owner. However, it is also the case that such a legal monopoly does not necessarily allow the patent owner to exercise market power, much less monopoly power, either with respect to the patent (e.g., licensing) or products using the patented technology (i.e., with respect to pricing, output, and other terms and conditions of sale). The patent monopoly and the ability to exercise monopoly power with respect to the patent are often mistakenly viewed as different sides of the same coin, but they are not. To understand why this is so, consider how a patent owner may exercise a degree of market power above and beyond what they could do, absent the patent.¹²

[Page 140]

Patents confer market power to the extent that they allow the patent holders to set prices at a higher level, that is, farther above marginal cost, than would otherwise be the case, absent the patent.¹³ The patent system allows this so that the patent owner has the opportunity to recover the cost of innovation and earn a return on that innovation investment. Indeed, it is precisely the prospect of being able to exercise some greater degree of market power that is the incentive to the innovator that produces the innovation.¹⁴ In this sense, it is by design that patents are intended to create the opportunity for firms to enjoy greater market power and earn higher profits than would be the case, absent the patent.

As a practical matter, the economic value of a patent is no greater than the amount of money people are willing to pay for the advantage conferred by the patented technology over the next-best alternative. This is true in the product marketplace, where the seller of a product with a patented feature (1) can set prices above competitors’ prices to the extent that the market (i.e., consumers/users) perceives an advantage attributable to the patented feature that it is willing to pay for and (2) is constrained in setting that premium by the size of that perceived advantage. In other words, where a patent allows a product to incorporate features, performance attributes or some other meaningful differentiator that consumers value, consumers will pay "more" for that product. However, the amount of "more" will depend on how much better or advantageous those features are. Where the differences are small or large, the "patent price premium" will be correspondingly small or large, all else equal.

The same situation holds in the marketplace for technology, where the price or "royalty" a patent holder can reasonably obtain in exchange for a license to the patented technology is constrained by the benefits the patented technology confers on a potential licensee who incorporates that technology into their product(s). In other words, a patent holder seeking to license their technology will, all else equal, receive a higher royalty the greater the advantage to the licensee from being able to utilize the patent relative to the next-best alternative. For patent practitioners, this is a familiar line of reasoning; indeed, the assessment of the appropriate "reasonable royalty" in a patent damages setting accounts for the relative advantage of the patent versus alternative technologies (among other things).¹⁵

The value of a patent also depends on the institutional setting in which the patent is used and licensed. One example is when patents are considered essential to the standards set by a standard-setting organization ("SSO"). An SSO is an organization whose principal role is to develop, coordinate, and promulgate technical standards to facilitate the operation of firms producing products covered by the standard. The practical effect of standards set by an SSO is to create uniformity with respect to things like terminology, specifications, and so forth. One example of products subject to standards is medical devices that work the same way and display information in the same way, regardless of manufacturer.

[Page 141]

Compliance with and incorporation of standards into manufactured products typically requires access to patented technology. When there are patents that are considered essential to being able to comply with the standard, patent owners can agree to designate their patents as "standard-essential." Standard-essential patents ("SEPs") are patents that teach technology that must be used to comply with the standard. To ensure that compliance is possible, the owner of an SEP is often required to license the SEP to all comers on fair, reasonable, and non-discriminatory ("FRAND") bases.¹⁶ This agreement to license on FRAND terms is between the patent owner and the SSO.¹⁷

As a general matter, whether a patent is an SEP or not, patent owners will often license their patents to others in return for payment in the form of a royalty. That royalty can be paid as a lump sum, in which the license is "bought and paid for" at the time of the license. Royalties can be "running"—that is, paid based on the sales of the products (or services) sold that use the technology embodied in the licensed patent(s). In that case, the royalty is commonly specified as some percentage ("the royalty rate") of some measure of the sales value of the product sold ("the royalty base"). The royalty is set by negotiation; the outcome of that negotiation reflects a variety of factors, including, as noted above, the alternatives to the patented technology. All else equal, when there are viable (albeit imperfect) alternatives to the patented technology, royalties will be relatively lower than in a situation where the licensee has no access to viable alternatives.

Where licensing in a standard-setting context differs from the usual case is in that royalty-setting process; the typical royalty-setting analysis is different for SEPs. By its very nature, an SEP designation means that the covered technologies are essential to an industry standard (i.e., must be practiced) and, thus, there are no alternatives to the technology once the standard has been implemented.¹⁸ That is the case, as a tautological matter.

Because compliance with the standard requires a license to the SEPs, absent some rules governing licensing, the patent owner would be able to exploit the SEP status and extract relatively larger royalties from all of the adherents to the standard, all else equal. The patent owner could behave opportunistically and, if there were no rules governing the terms on which licenses had to be granted, the SEP holder could engage in "patent hold-up." Patent hold-up, as it is commonly used, is the exploitation of firms that produce products governed by a standard and that are locked-in to the technology embodied in the standard. That is a very specific source of market power that would allow the SEP owner to obtain higher royalties than might otherwise be the case. Hold-up, all else equal, is likely either to slow the pace at which technology is adopted through a standard or lead to higher prices of the products meeting the standard as a result of the higher royalties resulting from hold-up. That is, at least in part, the economic rationale for the FRAND obligation. The ability of firms to obtain a license under FRAND conditions is critical to the acceptance of a standard and the use of the technologies in SEPs.

[Page 142]

Minimizing the risk of hold-up is important for the standard-setting process to work and for an SSO to function; a key purpose of the FRAND requirement is to guard against patent hold-up.¹⁹ While SSOs require SEP holders to commit to licensing their patented technologies on FRAND terms, two key points must be understood. First, the obligation to license an SEP on FRAND terms is a contractual obligation. The Patent Act does not impose this obligation. It exists because the SEP owner agrees to comply with the rules of the SSO and to license pursuant to FRAND terms. Indeed, several decisions have held that a firm’s violation of its FRAND commitment is a breach of contract.²⁰ The second point is that there is no unique definition of "FRAND." There are no explicit criteria to determine whether SEP licenses are either individually or collectively FRAND. That creates one of the challenges in evaluating the competitive impact of licensor behavior in an SEP/FRAND setting.

III. ANTITRUST ANALYSIS OF QUALCOMM’S LICENSING POLICY

A. General Principles

It is well-recognized among economists and antitrust lawyers that the mere possession of market power and the associated ability to exercise some degree of control over prices and set prices above marginal cost, in and of itself, does not give rise to antitrust liability. This is true as a matter of economic theory and empirical analysis. As a theoretical matter, control over prices and the ability to price above marginal cost does not equate to the ability to sustain anticompetitive outcomes. Factors such as conditions of entry, elasticity of demand, and, more generally, the competitiveness of the markets in which products are sold influence whether a firm has sufficient market power to sustain anticompetitive outcomes. Economists have also studied a variety of markets where competition is vigorous while, at the same time, competing firms have market power. Examples include consumer goods and grocery items and categories where product differentiation gives rise to some degree of market power for sellers, yet competition within categories and product groups is often quite intense.

[Page 143]

As we understand it, the law generally reaches the same conclusion. For example, the Supreme Court has recognized that market power can be the outcome of the competitive process antitrust law is designed to protect. For example, in Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, the Supreme Court stated:

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. To safeguard the incentive to innovate, the possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct.²¹

Thus, as we noted above, to give rise to antitrust liability, market power must be large enough to permit conduct that allows a firm to sustain prices above and output below competitive levels, that is, to the "monopoly power" level. A critical focus, then, for purposes of evaluating antitrust liability, is the effect of the challenged conduct on the firm’s market power—i.e., the firm’s ability to charge prices above the competitive level.²² Does the behavior enhance market power? Does it elevate market power to a level of monopoly power, where non-competitive outcomes are sustainable? Are there competitive mechanisms that defeat the effort to exercise this monopoly power? Those are among the key questions. With patents, those questions are equally important and the answers no less challenging.

For purposes of illustration, assume Firm A has monopoly power over a product and charges prices above the competitive level (i.e., supracompetitive prices). Now assume that Firm A obtains a patent on technology used in the product and further increases prices. Is there an antitrust issue associated with the price increase resulting from the patent grant? The simple answer is "not necessarily." Firm A had monopoly power over the product both before and after the grant of the patent. The patent did not necessarily change Firm A’s ability to charge supracompetitive prices. This does not mean, however, that the ability to further increase prices as a result of the patent grant is never an antitrust issue. Indeed, that may be the issue.

B. The FTC v. Qualcomm Decision

Qualcomm is a publicly traded, semiconductor company headquartered in San Diego, California that specializes in telecommunication technology. Its principal ventures include developing, designing, and selling baseband processors and other semiconductor devices, or "chips," used in cellular handsets (i.e., cellphones, tablets, etc.). Qualcomm sells its chips to cellular handset makers—called original equipment manufacturers or "OEMs"—like Apple, Samsung, and Huawei. In addition to its chip business, Qualcomm generates substantial revenue by licensing its intellectual property portfolios. The company owns numerous patents covering the technologies required for most cellular handsets to communicate on an operator’s network. Licensing intellectual property has been lucrative for Qualcomm, through its business unit Qualcomm Technology Licensing ("QLT"), with over $7.6 billion in revenues and earnings before taxes of $6.5 billion in the fiscal year ending in September 2016.²³

[Page 144]

In January 2017, the FTC filed suit in the Northern District of California against Qualcomm, alleging that Qualcomm’s conduct with respect to certain of its licensing practices violated Sections 1 and 2 of the Sherman Act, and thus Section 5 of the FTC Act, 15 U.S.C. § 45(a).²⁴ Specifically, the FTC alleged that Qualcomm unlawfully maintained its monopoly in two chip markets—3G code division multiple access ("CDMA") and 4G long-term evolution ("LTE")—with a set of interrelated policies that excluded competitors, reduced both the incentives and the abilities of competitors to innovate, and resulted in supracompetitive licensing fees for Qualcomm’s SEPs, which were passed on to the prices paid by consumers for cellphones and tablets.²⁵

Qualcomm was a leading developer in the 2G-CDMA telecommunication standard and an active participant in developing the standards for 3G-CDMA and 4G-LTE.²⁶ As a result, Qualcomm holds a substantial share of all SEPs for the 2G-CDMA, 3G-CDMA, and the 4G-LTE markets.²⁷ In addition, Qualcomm holds SEPs in cellular and non-cellular technologies that include location services, camera and multimedia, and user interface technologies.

Unlike many of its competitors with SEPs in telecommunications technology, Qualcomm is active not just in those technology markets, but in markets for products that employ that telecommunication technology, including cellular hardware and chips. Modem chips, which Qualcomm began selling in 1996, are what allow cellular handsets to communicate with each other over a network. By all metrics and for an extended period of time, Qualcomm has been a very successful player in the modem chip market, achieving as much as a 96% share of the worldwide CDMA modem chip market and up to 89% of the premium LTE modem chip market.²⁸ The FTC alleged that Qualcomm has monopoly power with respect to those chips and maintains that power through its alleged anticompetitive business practices.

[Page 145]

The FTC identified a set of interrelated Qualcomm policies spanning both markets that it claimed violated antitrust law. One important policy at issue was Qualcomm’s practice to not sell chips to any OEM that would not also pay royalties for licenses to Qualcomm’s patent portfolios, which include SEPs and non-SEPs. This practice is called the "no license, no chips" policy.²⁹ Qualcomm requires OEMs that purchase chips to obtain a license to the patented technologies, rather than licensing the patents directly to chip manufacturers that must develop products compatible with the current standard of telecommunication technology. Qualcomm argued that by passing over competing chipmakers and only licensing to those further downstream, it was simply exerting more control over its intellectual property by avoiding patent exhaustion.³⁰

The issue presented by the FTC was that because OEMs needed to purchase chips, and because Qualcomm dominates in the 3G-CDMA and 4G-LTE chip markets, OEMs were often forced to work with Qualcomm to meet their chip needs. According to the FTC, if an OEM were denied access to Qualcomm’s chips, as was threatened under the "no license, no chips" policy, it would be at a severe disadvantage in efforts to design and sell critically important premium-tier phones and phones for use on CDMA networks.³¹ Thus, OEMs had little leverage to negotiate the royalty and license terms demanded by Qualcomm. Further, once the OEM agreed to pay the royalty, it was owed on all cellular handsets the OEM produced, regardless of the manufacturer of the chip employed in that device. Additionally, and perhaps most importantly, the FTC alleged that the royalty rates charged for the licenses were supracompetitive and not within Qualcomm’s FRAND commitments.

The FTC argued that because Qualcomm’s licensing fees were supra-FRAND, this resulted in a cascade of harms that affected rival chipmakers, OEMs, and end-consumers. The FTC reasoned that by requiring end-producers to pay these supra-FRAND licensing fees, it effectively made it cheaper for the OEM to purchase and use Qualcomm’s suite of chips, including both those chips in which Qualcomm had a monopoly and those chips in which Qualcomm had several competitors. In antitrust terms, this policy "raises rival’s costs."

[Page 146]

Under the FTC’s theory, because Qualcomm is capturing supracompetitive royalties in the technology market, a rival chipmaker would either have to lower the price of their chip, making it more attractive for a customer to purchase it, or increase their chip prices to maintain an acceptable level of profit, but making them a weaker competitor in the chip market. Regardless of the actual pricing behavior adopted by the rival chipmakers to compensate, the FTC was clear that this supra-FRAND royalty makes the chip market less attractive for a new company to enter while squeezing the profits of an existing company enough to make them likely to exit.

Further, according to the FTC’s theory, since Qualcomm owned a significant share of SEPs and had monopoly power in the two chip markets, 3G-CDMA and 4G-LTE chips, the "no license, no chips" threat allowed Qualcomm to successfully charge high license fees for its intellectual property portfolios that included the SEPs. An expert for the FTC testified that Qualcomm’s royalty rates for its SEPs were supra-FRAND. Additionally, the FTC argued that, just as OEMs were forced to agree to less-than-ideal licensing terms due to the threat of losing access to Qualcomm chips, the same threat stopped OEMs from challenging those license terms with the SSO or in court.

After trial at the District Court, Judge Lucy Koh agreed with the FTC and barred Qualcomm from continuing its "no license, no chips" policy. Judge Koh stated:

Qualcomm’s practice of refusing to sell modem chips until an OEM signs a patent license agreement, and Qualcomm’s associated threats, generate and sustain Qualcomm’s unreasonably high royalty rates. Because Qualcomm receives royalties on any handset sale, even when that handset contains a rival’s modem chip, Qualcomm’s unreasonably high royalty rates impose an artificial and anticompetitive surcharge on the price of rivals’ modem chips. At times, Qualcomm has even charged OEMs higher royalty rates when OEMs purchase rivals’ chips than when OEMs purchase Qualcomm’s chips, which further harms rivals.³²

The decision was appealed to the Ninth Circuit, which vacated Judge Koh’s ruling in August 2020. Regarding the "no license, no chips" policy, the Ninth Circuit found that Qualcomm’s policies are "chip supplier neutral."³³ The Ninth Circuit explained:

Because rival chip manufacturers practice many of Qualcomm’s SEPs by necessity, Qualcomm offers these companies what it terms "CDMA ASIC Agreements," wherein Qualcomm promises not to assert its patents in exchange for the company promising not to sell its chips to unlicensed OEMs. These agreements, which essentially function as patent-infringement indemnifications, include reporting requirements that allow Qualcomm to know the details of its rivals’ chip supply agreements with various OEMs. But they also allow Qualcomm’s competitors to practice Qualcomm’s SEPs royalty-free.
Qualcomm reinforces these practices with its so-called "no license, no chips" policy, under which Qualcomm refuses to sell modem chips to OEMs that do not take licenses to practice Qualcomm’s SEPs. Otherwise, because of patent exhaustion, OEMs could decline to take licenses, arguing instead that their purchase of chips from Qualcomm extinguished Qualcomm’s patent rights with respect to any CDMA or premium LTE technologies embodied in the chips. This would not only prevent Qualcomm from obtaining the maximum value for its patents, it would result in OEMs having to pay more money (in licensing royalties) to purchase and use a competitor’s chips, which are unlicensed. Instead, Qualcomm’s practices, taken together, are "chip supplier neutral"—that is, OEMs are required to pay a per-unit licensing royalty to Qualcomm for its patent portfolios regardless of which company they choose to source their chips from.³⁴

[Page 147]

C. Are Supra-FRAND Royalties Anticompetitive?

To establish a violation of Section 2 of the Sherman Act, there must be (1) monopoly power in a relevant antitrust market, and (2) anticompetitive conduct.³⁵ For purposes of our discussion, we assume that Qualcomm has monopoly power in both CDMA and premium LTE modem chips, and that each of these constitutes a relevant antitrust market.³⁶ That leaves the question of whether, within those relevant markets, there is anticompetitive conduct. Here, the challenged conduct is Qualcomm’s "no license, no chips" policy.³⁷

Of the several examples of harm put forth by the FTC, the assertion undergirding all of them was that Qualcomm’s royalty rates for its SEPs were supra-FRAND. To demonstrate this, the FTC relied on a licensing expert for the analysis of Qualcomm’s licensing rates. According to the testimony, Qualcomm’s historical licensing rates were near 5%.³⁸ The licensing expert testified that based on his analyses, these rates were well above FRAND, and that appropriate rates should have been under 1%.³⁹ Qualcomm opposed the FTC’s FRAND analyses and conclusions; however, Qualcomm did not put forth its own FRAND analysis.⁴⁰

[Page 148]

If we take the FTC’s FRAND analyses as true for purposes of this discussion, there is a further issue that is raised by the Ninth Circuit. If Qualcomm is charging a supra-FRAND rate, that is, in the first instance, a breach of the FRAND obligations it undertook when identifying its patents as SEPs for the standards. Thus, the initial question is whether that is a violation of the contract that exists between the SSO and Qualcomm. The Ninth Circuit concluded that the issue was a contract one, and the extension into antitrust law was unwarranted.

But many antitrust cases arise out of contracts. For example, price fixing cases often arise out of contracts.⁴¹ What makes this case different than other contract situations?

Ultimately, that question turns on a point that even the FTC’s economic expert, Dr. Carl Shapiro, conceded.⁴² That is, if the Qualcomm licensing royalty rate met the FRAND criteria, the predicate for the FTC’s allegation against Qualcomm falls away. Importantly, the FTC challenged neither the grant of Qualcomm’s patents nor their incorporation in industry standards. Absent the challenged conduct, Qualcomm still holds its SEPs and is entitled to royalties. Under the FTC’s theory, it is a breach of the contract between the SSO and Qualcomm—and not the contract itself—that produces the harm claimed by the FTC.

In other contract situations, the question of whether there has been an antitrust violation does not depend on whether the conduct breached a particular agreement. In Qualcomm, however, the theory advanced by the FTC was premised on the existence of a contractual breach.

[Page 149]

Still, if we put aside the breach of contract issue, it is not clear that Qualcomm’s licensing policy actually harmed competition in the relevant markets.⁴³ Thus, it is uncertain, at best, whether it is appropriate to consider such behavior an antitrust violation. To see this, suppose that Qualcomm charged a much lower rate, in line with the FTC’s proposed FRAND rate. OEMs would still have to pay this royalty rate on each handset that practices Qualcomm’s SEPs, but the OEMs would be capturing more of the profits than they had in the supra-FRAND world. Rival chipmakers must still compete with Qualcomm, who benefits from licensing revenue. Indeed, it is difficult to imagine how "no license, no chips" can be considered a threat to chip competition in this setting.

This does not mean, however, that supra-FRAND royalties can never be anticompetitive. There can be situations where supra-FRAND royalties create opportunities for an SEP holder to engage in anticompetitive behavior that would not be possible if the firm were bound by its FRAND commitment. This may be particularly concerning in cases where the SEP holder is also competing elsewhere in the supply chain—as was the case in Qualcomm. While every case requires a thorough examination of the facts at hand, we provide a few examples of circumstances where supra-FRAND royalties combined with the "no license, no chips" policy may be anticompetitive from an economic perspective. While these circumstances were hinted at in Qualcomm, we argue that none were sufficiently demonstrated by the FTC.

Predatory pricing. It is theoretically possible for an SEP holder to drive out its competitors in a complementary market by pricing the complementary good (here, modem chips) at artificially low levels, given that the SEP holder can "make up the difference" via its supra-FRAND SEP price. This kind of pricing behavior is sometimes referred to as "predatory pricing."⁴⁴ Indeed, the Supreme Court has recognized predatory pricing as anticompetitive where: (1) "the prices complained of are below an appropriate measure of its rival’s costs"; and (2) the alleged predator had "a dangerous probability" "of recouping its investment in below-cost prices."⁴⁵

[Page 150]

While we are not privy to all the facts of the case, if Qualcomm indeed effectively lowered its chip price in a manner that was predatory or with the intent of ultimately driving out competition, this could be problematic from an antitrust perspective. For example, if Qualcomm priced its chips at or below marginal cost, other competing chip manufacturers may not have been able to effectively compete in the market, resulting in the elimination or reduction of chip competition. However, it appears the FTC did not provide sufficient evidence that this kind of pricing existed, or that it was the result of Qualcomm’s supra-FRAND pricing. Indeed, the FTC’s economic expert claimed the opposite—that Qualcomm had maintained artificially high prices in the chip market.⁴⁶ According to the Ninth Circuit, "Here, not only did the FTC offer no evidence that Qualcomm engaged in predatory pricing, the district court’s entire antitrust analysis is premised on the opposite proposition: that Qualcomm ‘charge[s] monopoly prices on modem chips.’"⁴⁷

Behavior that punishes customers who challenge the SEPs or the associated royalties. The patent system works when it allows for debate over the validity, infringement, and value of a firm’s patents. Indeed, one of the hallmarks of the system is that it incentivizes marketplace participants to challenge assessments made by an increasingly overburdened patent office (which receives hundreds of thousands of applications every year),⁴⁸ creating a system of checks and balances that we believe should readily work together to promote valuable innovation. This is especially true in the context of SEPs, given the greater potential for abuse relative their non-essential counterparts.⁴⁹

Patent challenges play an important role in preventing the exercise of unjustified market power. The Supreme Court has recognized that "[l]icensees may often be the only individuals with enough economic incentive to challenge the patentability of an inventor’s discovery. If they are muzzled, the public may continually be required to pay tribute to would-be monopolists without need or justification."⁵⁰ Indeed, discussions surrounding the role of patent challenges on competition have been growing following the Supreme Court’s decision in FTC v. Actavis, Inc.⁵¹ In that case, the Supreme Court held that reverse payment agreements—those that some pharmaceutical companies pay to prevent patent challenges from generic manufacturers seeking to enter the market—may violate antitrust laws.⁵² Several legal scholars have argued that patent challenge clauses, a contractual element that prohibits patent licensees from challenging the validity of licensed patents, should receive antitrust scrutiny.⁵³

[Page 151]

Again, while we are not privy to all the facts of the case, if Qualcomm’s threats to withhold chips from any OEM that challenged the license agreements proved effective, this could be deemed to violate antitrust law. However, it appears the FTC did not sufficiently demonstrate this. The Ninth Circuit stated that "it appears that OEMs have been somewhat successful in ‘disciplining’ Qualcomm’s pricing through arbitration claims, negotiations, threatening to move to different chip suppliers, and threatened or actual antitrust litigation. These maneuvers generally resulted in settlements and renegotiated licensing and chip-supply agreements with Qualcomm, even as OEMs continued to look elsewhere for cheaper modem chip options."⁵⁴

Tying. It is possible for an SEP holder to drive out its competitors in a complementary market by making the grant of a license to the SEPs (the tying product) contingent on purchasing the complementary good (the tied product—here, modem chips). This is often referred to as "tying." From an economic perspective, tying can generate anticompetitive effects.⁵⁵ Tying can enable a firm to leverage market power in the tying product market by forcing buyers to purchase the tied product with the tying product.⁵⁶ When implemented in this way, tying can reduce competitors’ access to the market for the tied product and reduce buyers’ choices among competing products, thereby harming competition. Tying can deny competitors’ access to the tied product market not because the party imposing the tying arrangement has a better product or a lower price, but because of market power in the tying product market.

While the FTC did not present a tying claim as such, the Ninth Circuit directly addressed this issue, stating that "[i]f Qualcomm were to refuse to license its SEPs to OEMs unless they first agreed to purchase Qualcomm’s chips (‘no chips, no license’), then rival chip suppliers indeed might have an antitrust claim under both §§ 1 and 2 of the Sherman Act based on exclusionary conduct."⁵⁷ The Ninth Circuit determined that this was not the case:

[U]nlike a hypothetical "no chips, no license" policy, "no license, no chips" is chip-neutral: it makes no difference whether an OEM buys Qualcomm’s chip or a rival’s chips. The policy only insists that, whatever chip source an OEM chooses, the OEM pay Qualcomm for the right to practice the patented technologies embodied in the chip, as well as in other parts of the phone or other cellular device.⁵⁸

[Page 152]

D. Does It Matter Where the Royalty Is Collected in the Production Chain?

In addition to challenging Qualcomm’s licensing rates as anticompetitive, the FTC challenged Qualcomm’s practice of collecting royalties at the OEM stage of the production chain as anticompetitive. We thus turn our focus to this economic and competitive question: does it matter that Qualcomm collects its royalties from OEMs rather than rival chip manufacturers and, if so, how does it matter?

As a matter of economic theory, the total revenues of a given industry are equal to the sum of the payments to the inputs plus the profits earned by all firms in the production chain. In equilibrium, the profits earned by any firm, wherever their location in the production chain, are bounded by the demand curve for the final product and the payments to other inputs. The reasoning behind this is that no input is of any value unless it is used to produce the thing consumers value—the final product. Thus, any attempt to value a given input must account for consumer demand for the final product, the payments to all inputs in the production chain, and the profits earned by all firms in the production chain. No stage of the production chain is independent of the others. There is no "fixed pie" associated with any individual stage of the production chain.

In the context of smartphones, the source and limit of all profits in the production chain is the value consumers assign to smartphone features and functionalities. While the technologies and components that enable and manufacture those features and functionalities have value in that they contribute to production, neither the technologies nor the components have value by themselves. This means that the royalties earned by the holders of the SEPs that enable smartphones to work are constrained by the difference between what consumers are willing to pay for a smartphone with the patented features and all other costs of producing a smartphone across the entire production chain. This is true regardless of where the royalties are earned in the production chain.

This is also true when the provider of an input enjoys monopoly power. When an input provider exercises monopoly power in one stage of the production chain and charges monopoly prices, the distribution of revenues and profits across the production chain changes, but final market prices still determine the overall revenues and profits. Thus, the monopolist, and the rest of the vertical market players, will still be constrained by the amount that consumers pay for the final product.

In Qualcomm, the FTC took the stance that collecting royalties at the OEM stage of the production chain was anticompetitive because it imposed an "artificial surcharge" on rival chip manufacturers’ sales. The FTC’s argument continues as follows: because rival chip manufacturers and OEMs were sharing the additional costs from the supra-FRAND royalty rates, this squeezed the profits rival chip manufacturers could appropriate from the sale of each chip. Therefore, to remain profitable enough to continue to participate in the market, rivals had to raise the price they charged OEMs for their chips, thereby making them less competitive against Qualcomm. The FTC’s economic expert testified that this harm to competing chip manufacturers would trickle down and affect end-consumers; he argued that with higher chip prices from competitors, Qualcomm would then also have an incentive to raise (not lower) its chip prices to OEMs⁵⁹—a harm that would likely trickle down to end-consumers.

[Page 153]

If Qualcomm were to license to chip manufacturers, the increased cost to chip manufacturers from paying for the licenses would likely be passed along to OEMs, at least in part. Thus, OEMs would still incur some, if not all, of the cost of licensing fees. Further, it appears that despite its once larger market presence, Qualcomm is competing with other chip manufacturers on price in the chip market.⁶⁰ Based on our understanding of the facts, and absent additional circumstances such as those discussed in Section III.C, we conclude that it does not matter where Qualcomm collects its royalties.⁶¹ Our conclusion might be different outside of the standard-setting context if Qualcomm forced OEMs to pay royalties on all sales, regardless of the chips used if the rival manufacturers did not infringe Qualcomm’s patents. In this setting, it is reasonable to expect that such behavior could be competitively problematic.⁶²

[Page 154]

IV. CONCLUSION

The Qualcomm case presents a number of interesting and challenging economic questions that are fundamental to an antitrust assessment of licensing behavior for patents, especially SEPs. From both a legal and economic standpoint, the setting of the case is unique, and gives rise to a number of difficulties in interpreting and analyzing antitrust harm. This is further demonstrated by how controversial both decisions have been and the discussions of the impact the decisions may have on future antitrust cases.

The Ninth Circuit determined that an alleged failure to meet FRAND licensing obligations pursuant to an agreement with an SSO is, first and foremost, a contractual issue. The decision is an important cautionary note that not every licensing activity immediately rises to an antitrust issue, even if it affects pricing and other terms and conditions of sale. The decision also cautions against theories of anticompetitive harm that are premised on the existence of an unproven contractual breach.

Even more fundamentally, the antitrust analysis of licensing behavior must consider the actual market conditions and a correct counterfactual analysis. When, as is the case here, every handset sold by an OEM, for example, embodies the technology covered in the Qualcomm SEP portfolio, someone is going to pay a royalty, and the antitrust analysis must reflect that. For the end result to favor the FTC, it would have had to convincingly show that (1) the licensing rates were supra-FRAND and caused by the alleged behavior, and (2) the alleged supra-FRAND rates harmed competition, not just competitors.

[Page 155]

——–

Notes:

^1. Aminta Raffalovich, J.D., M.S. is a Vice President at Intensity, LLC specializing in the economics of technology and innovation. She serves as a trusted advisor and economic expert in complex commercial litigation and disputes.

Steven Schwartz, Ph.D. is a Managing Director at Intensity, LLC with over 35 years of economic consulting experience. He has been retained as an economic expert in numerous litigation and non-litigation matters and has provided testimony in numerous Federal and State courts, as well as before the U.S. International Trade Commission and the U.S. Tax Court.
Stephanie Khoury, Ph.D. and Kira Stearns, Ph.D., Economists at Intensity, LLC, contributed to the preparation of this article.
The views expressed are those of the authors and may not reflect the views of Intensity, LLC, its clients, or any of its or their affiliates. This article is for general information purposes and is not intended to be and should not be taken as legal advice.

^2. FTC v. Qualcomm Inc., 969 F.3d 974 (9th Cir. 2020).

^3. FTC v. Qualcomm Inc., 411 F. Supp. 3d 658 (N.D. Cal. 2019).

^4. In the discussion below, we largely ignore the issues surrounding market definition. We recognize the centrality of the relevant market to an antitrust analysis, but we choose here to focus on the issues surrounding market power and the impact of the licensing behavior on competition.

^5. See, e.g., N. Gregory Mankiw, Principles of Microeconomics 279—280 (6th Ed. 2012) .

^6. The focus on whether such price increases or quantity decreases can be sustained is a critical element. If a firm raises price and enough alternatives are available such that enough consumers switch their purchases and render the price increase unprofitable, the firm did not successfully exercise monopoly power. This relationship between a firm and its ability to sustainably increase price has led to the development of a hypothetical monopolist test, the small but significant and non-transitory increase in price, or "SSNIP," test. In essence, the test examines whether a hypothetical monopolist can increase prices (typically, an increase of 5% to 10% is used) and do so profitably, without losing sales to other products or competitors. This test was originally developed for use in mergers, but has been applied more generally to antitrust market definitions. See U.S. Department of Justice and the Federal Trade Commission, Horizontal Merger Guidelines 8—10 (2010); Steven Salop, et al., Antitrust Economics for Lawyers 6 (2017); John D. Harkrider, Operationalizing the Hypothetical Monopolist 1, 11, available at https://www.justice.gov/sites/default/files/atr/legacy/2007/08/30/202598.pdf.

^7. Profits are maximized when a firm’s marginal revenue is equal to its marginal cost. In a perfectly competitive market, because the firm’s demand curve is horizontal, price will equal marginal revenue. Thus, when a firm is profit maximizing, price = marginal revenue = marginal cost. See, e.g., Dennis Carlton & Jeffrey Perloff, Modern Industrial Organization 56-59, 66-67 (3rd Ed. 2000).

^8. In a case where demand curves slope downward, the profit-maximizing price will exceed marginal revenue. Thus, when profits are maximized, price > marginal revenue = marginal cost. It is not necessarily the case that the amount of market power can be inferred directly from the price-cost margin.

^9. Economists have a variety of ways to measure market power. One well-known measure—the price-cost margin—is frequently used. This margin—also known as the Lerner Index—is expressed as: (price — marginal cost) / price. When price and marginal cost are equal, the Lerner Index is zero. As the gap between price and marginal cost (the price-cost margin) grows, the index rises. Its upper bound is one, in cases where marginal cost is zero. In practice, of course, prices are readily observable, but marginal costs are more difficult to measure. As a practical matter, economists will often approximate the Lerner Index by using average costs as a proxy for marginal costs.

^10. R. Hewitt Pate, Competition and Intellectual Property in the U.S.: Licensing Freedom and the Limits of Antitrust 8 (2005) (remarks before the 2005 Competition Workshop, Florence, Italy), available at http://www.usdoj.gov/atr/public/speeches/209359.pdf.

^11. To be sure, a monopoly over a technology is, in the first instance, narrow. It provides a legal right to exclude others from the use of that technology. But, it does not necessarily translate into any meaningful market power in a product market in which a product (or products) implementing the patented technology is sold.

^12. For purposes of this discussion, we are assuming downward sloping demand curves and, thus, some degree of market power for the seller. The question we focus on is the incremental market power attributable to the patent.

^13. For further discussion, see, e.g., Illinois Tool Works Inc. v. Independent Ink, Inc., 547 U.S. 28, 38—46 (2005).

^14. Innovation—research and development activity generally—is risky. The patent monopoly and the prospect of being able to exercise increased market power because of the patent (though not the uncertainty) creates the prospect for supra-normal returns to the patent holder and extra returns specifically attributable to the patent. Those returns to successful patents provide the return on all R&D investment projects, whether successful and not.

^15. The determination of a "reasonable royalty" typically involves consideration of the factors described in Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970), which are often referred to as "the Georgia-Pacific factors." Georgia-Pacific factor 9 is: "The utility and advantages of the patent property over the old modes or devices, if any, that had been used for working out similar results." Id. at 1120.

^16. See, e.g., Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297, 313 (3d Cir. 2007); Jorge L. Contreras, A Brief History of FRAND: Analyzing Current Debates in Standard Setting and Antitrust through a Historical Lens, 80 Antitrust Law Journal 39, 42 (2015).

^17. See generally Joshua D. Wright, SSOs, FRAND, and Antitrust: Lessons from the Economics of Incomplete Contracts, 21 George Mason Law Review 791—809 (2014).

^18. See, e.g., Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297, 314 (3d Cir. 2007).

^19. This has been recognized by various courts, including the Ninth Circuit. See, e.g., Microsoft Corp. v. Motorola, Inc., 696 F.3d 872, 876 (9th Cir. 2012) ("Many SSOs try to mitigate the threat of patent holdup by requiring members who hold IP rights in standard-essential patents to agree to license those patents to all comers on terms that are ‘reasonable and nondiscriminatory,’ or ‘RAND.’ . . . For example, consider the ITU, whose H.264 standard is implicated in this appeal. The ITU’s Common Patent Policy (the ‘ITU Policy’) provides that ‘a patent embodied fully or partly in a [standard] must be accessible to everybody without undue constraints.’ Anyone who owns a patent declared essential to an ITU standard must submit a declaration to the ITU stating whether it is willing to ‘negotiate licenses with other parties on a non-discriminatory basis on reasonable terms and conditions.’ If a ‘patent holder is not willing to comply’ with the requirement to negotiate licenses with all seekers, then the standard ‘shall not include provisions depending on the patent.’").

^20. E.g., id. at 889; In re Innovatio IP Ventures, LLC Patent Litig., 921 F. Supp. 2d 903, 923 (N.D. Ill. 2013); see also Realtek Semiconductor Corp. v. LSI Corp., 946 F. Supp. 2d 998, 1005, 1008 (N.D. Cal. 2013).

^21. 540 U.S. 398, 407 (2004) (emphasis in original).

^22. See, e.g., United States v. Grinnell Corp., 384 U.S. 563, 570—571 (1966); ABA Section of Antitrust Law, Antitrust Law Developments 240—241 (8th Ed. 2017).

^23. Complaint at 5, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Feb. 1, 2017).

^24. Brief of the Federal Trade Commission at 16-17, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (9th Cir. Nov. 22, 2019).

Section 5 of the Federal Trade Commission Act, 15 U.S.C. § 45(a), prohibits "unfair methods of competition," including violations of Sections 1 and 2 of the Sherman Act, among other antitrust law.

^25. Complaint at 2, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Feb. 1, 2017).

^26. Cellular standards are described as evolving over generations. Second generation ("2G") technology is usually associated with the early 1990s. Fourth generation ("4G") technology was first established in the United States in 2011.

^27. Complaint at 12-14, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Feb. 1, 2017).

^28. FTC v. Qualcomm Inc., 411 F. Supp. 3d 658, 687, 694 (N.D. Cal. 2019). Recent estimates suggest Qualcomm controls a 79% share of the CDMA chip market. Qualcomm alleges that its share in the LTE market is below 50%. See Opening Brief for Appellant Qualcomm Inc. at 118, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (9th Cir. Aug. 23, 2019).

LTE chips are those chips compatible with a 4G cellular standard.

^29. The FTC claimed that the "no license, no chips" policy led to competitive harm in conjunction with interrelated Qualcomm policies, including: (i) incentive payments, (ii) refusal to license to rivals, and (iii) exclusive dealings. See Complaint at 2—3, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Feb. 1, 2017).

^30. Under the doctrine of patent exhaustion, the purchaser of a component containing the licensed patented technology does not also need to obtain a license (i.e., there is no "double licensing"). For a discussion of the doctrine of patent exhaustion, see, e.g., Quanta Computer, Inc. v. LG Electronics, Inc., 533 U.S. 617, 625—638 (2008).

Qualcomm also argued that it is appropriate to license only to OEMs because it is their products (i.e., cellular handsets), and not the chips, that practice the standards associated with Qualcomm’s SEPs.

^31. Complaint at 18, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Feb. 1, 2017).

^32. FTC v. Qualcomm Inc., 411 F. Supp. 3d 658, 698 (N.D. Cal. 2019).

^33. FTC v. Qualcomm Inc., 969 F.3d 974, 985 (9th Cir. 2020).

^34. Id. at 984—985.

^35. Id. at 990.

While the FTC claimed violations of both Section 1 and Section 2 of the Sherman Act, for purposes of our discussion of the "no license, no chips" policy, we focus on Section 2. Further, while the FTC claimed violations under both sections of the Sherman Act, the suit was brought under Section 5 of the FTC Act, which allows the FTC to challenge any potential anticompetitive conduct, including but not limited to, violations of Sections 1 and 2 of the Sherman Act.

^36. This was alleged by the FTC and not disputed by Qualcomm.

^37. We understand that the FTC alleged that other conduct by Qualcomm, in conjunction with the "no license, no chips" policy, harms competition. We focus on the "no license, no chips" policy herein.

^38. Trial Tr. at 1011:5—8, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Jan. 14, 2019) ("Q. Can you please briefly explain Qualcomm’s historical licensing practices? A. Yes. I mean, historically, they have charged 5 percent for CDMA based devices with—well there are caps also.").

^39. Trial Tr. at 1011:9-12, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Jan. 14, 2019) ("Q. And do you have an opinion as to whether Qualcomm’s licensing rates are consisted with its FRAND obligations? A. Yes. In my opinion, they’re far too high to be consisted with their FRAND obligations.").

In his unsealed testimony, the FTC’s licensing expert states various numbers from his analyses that range in value depending on methodology and the relevant devices, but all fall below 1%. For example, for LTE multimode devices, he states that his analysis showed that a FRAND rate would be a maximum of 0.67%. See FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK, Trial Tr., at 1013:3-6 (N.D. Cal. January 14, 2019) ("Q. And what were the results of your analysis? A. So for the top down analysis for LTE multimode, I had a minimum royalty rate of 0.52%, a median royalty rate of 0.58 percent, and a maximum royalty rate of 0.67%.").

^40. Opening Brief for Appellant Qualcomm Inc. at 93-94, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (9th Cir. Aug. 23, 2019); see also Trial Tr. at 1097:2-5, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Jan. 14, 2019) ("Q. Did counsel for-did Qualcomm put forth any expert analysis correcting your analysis or quantitatively showing what would happen if you used a different methodology? A. No they did not.").

^41. See, e.g., Catalano, Inc. v. Target Sales, Inc., 446 U.S. 643, 647 (1980).

^42. Trial Tr. at 1198:2-25, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Jan. 15, 2019) ("Q. Okay. So point one: If the court were to determine that the evidence doesn’t support the existence of a supra-FRAND royalty, then the rest of your analysis can be ignored; right? It depends on that as a starting point? . . . A. Then the economic consequences of the surcharge would not come into play. I think that’s logical. . . . Yes, the building block for, for this discussion of the effects is that Qualcomm was able to achieve a royalty surcharge by applying the chip leverage.").

^43. We do not address whether Qualcomm’s licensing policy is or could be competitively problematic outside of these markets or under a different market definition herein. We note, however, that the Ninth Circuit did not reject the possibility of anticompetitive effects to OEMs and their customers, stating:

[E]ven assuming that a deviation between licensing royalty rates and a patent portfolio’s ‘fair value’ could amount to ‘anticompetitive harm’ in the antitrust sense, the primary harms the district court identified here were to the OEMs who agreed to pay Qualcomm’s royalty rates—that is, Qualcomm’s customers, not its competitors. These harms were thus located outside the ‘areas of effective competition’—the markets for CDMA and premium LTE modem chips—and thus had no direct impact on competition in those markets.
FTC v. Qualcomm Inc., 969 F.3d 974, 999—1000 (9th Cir. 2020) (emphasis in original).

Antitrust scholars have questioned this aspect of the Ninth Circuit’s decision, arguing that higher prices for OEMs and their customers—and not just harm to competitors—concerns antitrust laws. See, e.g., Herbert Hovenkamp, FRAND and Antitrust, 105 Cornell Law Review 1683, 1685—87 (2020).

^44. For further discussion, see, e.g., Phillip Areeda & Donald F. Turner, Pricing and Related Practices Under Section 2 of the Sherman Act, 88 Harvard Law Review 697 (1975).

^45. Brooke Group v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 222—224 (1993).

^46. It is worth noting that it appears Dr. Shapiro’s claim of increased chip prices by Qualcomm was not shared by the FTC in its complaint, or by a group of 46 law and economics scholars that submitted an amicus brief in support of the FTC following the reversal by the panel. The amici state that "the panel ignored the district court’s factual findings that Qualcomm in its licensing agreements offset patent royalty surcharges by providing chip rebates when the OEM purchased Qualcomm chips." Brief of 46 Amici Curiae Law and Economics Scholars in Support of Petition for Rehearing En Banc, 15, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (9th Cir. Oct. 5, 2020).

^47. FTC v. Qualcomm Inc., 969 F.3d 974, 1001 (9th Cir. 2020).

^48. U.S. Patent and Trademark Office, U.S. Patent Statistics Chart: Calendar Years 1963-2019, available at https://www.uspto.gov/web/offices/ac/ido/oeip/taf/us_stat.htm.

^49. It is unsurprising, then, that SEPs are litigated more often than non-SEPs. These litigations indicate that a large number of SEPs are not, in fact, essential. See Mark A. Lemley & Timothy Simcoe, How Essential Are Standard-Essential Patents?, 104 Cornell Law Review 607-642 (2019).

^50. Lear, Inc. v. Adkins, 395 U.S. 653, 670 (1969).

^51. 133 S. Ct. 2223 (2013).

^52. Id. at 2227, 2233.

^53. See, e.g., Michal S. Gal & Alan D. Miller, Patent Challenge Clauses: A New Antitrust Offense?, 102 Iowa Law Review 1477 (2017); Thomas K. Cheng, Antitrust Treatment of the No Challenge Clause, 5 NYU Journal of Intellectual Property and Entertainment Law 437 (2016).

^54. FTC v. Qualcomm Inc., 969 F.3d 974, 1001—1002 (9th Cir. 2020).

^55. Abbott, Alden F. & Joshua D. Wright, Antitrust Analysis of Tying Arrangements and Exclusive Dealing, in Antitrust Law and Economics 183 (Keith N. Hylton ed. 2010).

^56. David S. Evans & Michael Sallinger, Why Do Firms Bundle and Tie? Evidencefrom Competitive Markets and Implications for Tying Law, 22 Yale Journal on Regulation 37, 38—39 (2005)

^57. Qualcomm, 969 F.3d at 1002.

^58. Id. at 1002—1003.

^59. Dr. Shapiro testified:

Why will Qualcomm’s all-in prices go up? Well, there’s two reasons. First, they’ve got rivals who are weakened and charging higher all-in prices because they’re having to pay the surcharge. So as a general rule, when your competitors raise their price, it tends to give a firm incentive to raise his price as well in oligopolies, and in addition to that, Qualcomm is less keen to compete to win business from its competitors because even when it loses, it gets a royalty surcharge.
And so that’s another reason why Qualcomm will not be as aggressive in pricing and will tend to raise its all-in prices due to the royalty surcharge.

Trial Tr., 1142:2-1144:18, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (N.D. Cal. Jan. 15, 2019).

^60. See, e.g., FTC v. Qualcomm Inc., 969 F.3d 974, 983-984 (9th Cir. 2020).

^61. There are, of course, situations where a monopolist in an upstream market can leverage that power in a downstream market to increase profit above the monopoly profit level. When, why, and how this happens is beyond the scope of this article. See, e.g., Robert H. Bork, The Antitrust Paradox: A Policy at War with Itself (1978); Michael D. Whinston, Tying Foreclosure, and Exclusion, 80 American Economic Review 837 (1990); Einer Elhauge, Tying, Bundled Discounts, and the Death of the Single Monopoly Profit Theory, 123 Harvard Law Review 397-481 (2009).

^62. If this were the case, it would be economically similar to the allegations against Microsoft in Caldera Inc. v. Microsoft Corp., 87 F. Supp. 2d 1244 (D. Utah 1999). In that case, it was alleged that Microsoft was charging licensing fees for its software on all processors, even those that were not running Microsoft’s operating system. Id. at 1249-1250. This meant that for those processors not running Microsoft’s operating system, customers of the processor were paying two licensing fees—one to Microsoft and one to the licensor of the operating system actually running on said processor. Id. at 1250.

There have been arguments, both by the FTC and the amici in support of the FTC, that this case mirrors that surcharge, since OEMs pay royalties on all chips, no matter if they were produced by Qualcomm or its rivals. However, Qualcomm differs in that the licensing fees are in fact owed on all chips in the market as they are SEPs, while in Caldera the additional fee imposed by Microsoft was for nothing of value on those processors not running Microsoft’s operating system. An amicus curiae brief in support of Qualcomm states:

In Caldera, however, "[the] effect of [a per-processor licensing] arrangement was that an OEM who chose to install [a competing system] would pay two royalties on the same machine." 87 F. Supp 2d at 1250. Thus, the per-processor arrangement could serve as a disincentive for OEMs to purchase or invest in competing systems. Here, by contrast, OEMs pay for use of Qualcomm’s SEPs that are essential to every cellular device produced, regardless of which supplier’s chip is used.

Brief of the United States of America as Amicus Curiae in Support of Appellant and Vacatur, 17-18, FTC v. Qualcomm Inc., No. 5:17-cv-00220-LHK (9th Cir. Oct. 5, 2020).