Standards, common platforms allowing products to work together, are ubiquitous in our economy. They allow consumers to know that their plugs will fit into outlets and that their phones will connect to wireless networks. But imagine that a company (1) has a patent needed to use a standard, (2) promises to license the patent on reasonable terms, and then (3) reverses course, seeking to block the product or charge an exorbitant price.
In such a case, the users of the standard are stuck. They have invested in technologies based on the standard. And they may be forced to pay a price reflecting not the added value of the technology but the costs of switching to a new technology. In other words, they are subject to “patent holdup.”
The concerns threatened by patent holdup have consistently been acknowledged by officials in Republican and Democratic administrations. A unanimously adopted 2007 report of the antitrust agencies explained the difference between a patentee’s power ex ante—when “multiple technologies may compete to be incorporated into the standard”—and ex post—when “the chosen technology may lack effective substitutes,” allowing patentees to “extract higher royalties.” Similarly, the Federal Trade Commission (FTC) unanimously endorsed a 2011 report that highlighted how “an entire industry” could be “susceptible” to the “particularly acute” concern of holdup, which can result in “higher prices” and “discourage standard setting activities and collaboration, which can delay innovation.”
These longstanding bipartisan concerns are front and center in Carl Shapiro and Mark Lemley’s excellent article, The Role of Antitrust in Preventing Patent Holdup. In contrast, they were neglected by the recently departed Assistant Attorney General (AAG) of the Department of Justice’s (DOJ’s) Antitrust Division, Makan Delrahim.
This Essay does four things. First, it explains why the Shapiro/Lemley article is important, particularly to situate today’s patent holdup debates in the larger context of transaction cost economics. Second, it raises questions related to standards organizations’ rules and antitrust law that the authors could more fully consider. Third, it offers additional support showing the radical nature of Delrahim’s position. Finally, it opines on the nature of academic debate and its effect on antitrust enforcement.
The Shapiro/Lemley article is important in situating the contemporary patent holdup discussion in the context of the decades-long history of transaction cost economics. This section discusses general holdup, patent holdup, and patent “holdout.”
A. General Holdup
The authors begin by discussion transaction cost economics. They quote Nobel Prize economist Oliver Williamson, who highlighted the significance of asset specificity, which occurs when “the identity of the parties matters for the continuity of a relationship” and “assets cannot be redeployed to alternative uses or users without loss of productive value.” Williamson emphasized the “fundamental transformation that occurs when parties make relationship-specific investments,” as “ex ante competition can be replaced by ex post monopoly.” Holdup results as “the owner of a key asset can charge more than the asset is worth ex ante if the buyer has made asset-specific investments that will be lost unless the parties agree on terms of trade.” As a result, “a party that makes substantial investments, the value of which relies heavily upon the actions of another party, is vulnerable to exploitation. . .and thus may have lessened incentives to invest.”
Shapiro and Lemley explain that “[m]arket participants will structure their relationships as best they can to avoid or minimize the inefficiencies associated with opportunism.” In particular, the authors point to “[t]hree mechanisms” that “stand out as common responses to the problem of holdup:” (1) vertical integration, “which aligns interests by placing both parties to the relationship inside a single firm;” (2) long-term contracts, “which ideally can be designed to protect the party making the specific investments while rewarding the other party based on its ex ante superiority over alternatives;” and (3) flexibility, “whereby the party making the investments shifts from specific investments toward more general investments in order to reduce its reliance on the other party.”
Many who argue against antitrust enforcement in the standards context claim there are few, if any, examples of holdup. For example, in a February 2018 letter to Delrahim, thirteen academics and former judges and government officials contended that “no empirical study has demonstrated that a patent owner’s request for injunctive relief after a finding of a defendant’s infringement of its property rights has ever resulted either in consumer harm or in slowing down the pace of technological innovation.” The signers also argued that “there is no sound empirical basis for claims of a systematic problem of opportunistic ‘patent holdup’” because the concept has not been supported “in the decade or more since that theory was first propounded.”
In contrast, Shapiro and Lemley helpfully explain that as a result of vertical integration, long-term contracts, and flexibility, “actual ex post holdups represent failures by market participants to efficiently structure their relationships.” The success of these tools, though, “does not mean that holdup is not a problem.” Rather, “scholars studying the holdup problem widely agree that the general theory of holdup is very well supported empirically without expecting, much less demanding, a body of empirical work measuring actual holdups.”
B. Patent Holdup
Building on the general harms of holdup are characteristics exacerbating patent holdup. The authors list multiple factors that “make it very difficult for firms developing new products in the information technology and communications sector to protect themselves from patent holdup”: (1) broad patents with vague boundaries, (2) uncertainty about future product attributes, (3) the absence of an independent invention defense in patent law, (4) “probabilistic patents” characterized by uncertainty, (5) patent pendency lags that expose an implementer to liability because of changes between the initial application and ultimate publication, and (6) patent thickets.
As a result, users in high-tech industries subject to standardization are not able to “easily identify the single firm that owns strong, clear patents that are likely to be asserted” and enter into “ex ante licensing contract[s].” The authors explain that, in contrast to “the typical bilateral holdup situation studied in the transaction cost economics literature,” long-term contracts are not likely, as the implementer will “have difficulty even identifying all of the (possibly thousands of) patents that might be asserted in the future” against its product, “many of which might not issue until the [manufacturer] is well down the road in its development process.”
Even for the patents that can be identified in advance, the authors explain how patent holdup presents an acute instance of the general holdup problem because of the “substantial transaction costs associated with each of the three mechanisms normally used to avoid holdup.” First, “[v]ertical integration is rarely a good solution for patent holdup,” as one manufacturer’s acquisition of patents likely to be asserted against it threatens other manufacturers, a concern exacerbated because of the sheer number of patents and inability to coordinate due to potential antitrust liability.
Second, “[l]ong-term ex ante patent licensing contracts intended to cover future products involve substantial transaction costs,” as manufacturers are “likely to have a relatively poor sense of what its future products will look like when it first begins developing them,” which is a problem given that that is “when it must begin making substantial specific investments” and that it will often not be possible to identify the patents that might be asserted against its future products, as well as the scope of those patents. These problems lead to “very few companies developing complex products in the information, technology and communications area” being “able, as a practical matter, to ‘clear’ their products by entering into ex ante licensing arrangements with most or all of the parties holding patents that might later be asserted against their new products.”
Third, “[r]etaining flexibility during the development process so as to dodge possible infringement claims for the resulting product can be exceedingly difficult, especially given the large number of patents and their vague boundaries.” And “even if such flexibility could be achieved, it might be very costly in terms of reduced product performance or the need to deploy additional engineering resources,” which is “especially true when the patent is a SEP [standard essential patent], since standardization is critical to many IT technologies.”
In other words, “each of the three basic mechanisms for mitigating patent holdup . . . faces greater obstacles when it comes to patent holdup in the high-tech sector than it does for more traditional types of holdup.” And these obstacles are compounded by the challenges of quantifying patent holdups, such as: (1) the typical inability (because of confidential licensing terms) to observe the ex post price, (2) the embedding of license fees in complex agreements, (3) the inability to “observe the ex ante price, making it difficult if not impossible to measure the magnitude of the holdup,” and (4) the difficulty of relying on litigated cases and the inability to observe settlement terms.
The authors also take on the claim, often cited by those critical of antitrust enforcement, that there is a “lack of empirical evidence” of patent holdup. Shapiro and Lemley criticize one oft-cited paper, which questions the concept of patent holdup based on the comparison of “rates of change of quality-adjusted prices in ‘SEP-reliant’ industries [such as cell phones] with ‘similar’ non-SEP-reliant industries [such as cars].” But as the authors correctly point out, “[t]his exercise does not address the relevant hypothesis: whether SEP holdup increased the price of cellular phones from what it otherwise would have been.” And they question the whole enterprise, noting that “[t]he quality-adjusted prices of pharmaceuticals have risen much faster than automobiles over the same period of time,” but that does not offer “proof that pharmaceuticals are subject to a patent holdup problem.”
The authors also step back to question the “limited relevance” of this line of inquiry because of, as discussed above, the observed instances of patent holdup representing “only the tip of the iceberg in terms of the social costs of patent holdup.” And they raise the concern that “the vast majority of these papers have been funded by Qualcomm and other patent holders seeking to weaken the institutions designed to control patent holdup, increase their leverage in licensing negotiations, and thus increase their ability to monetize their patents.”
Unlike the detours taken by the anti-enforcement crowd, the authors “observe the telltale signs of actual patent holdup” based on the presence of three conditions, at least the first two of which are common in the high-tech sector: (1) the independent development of a new product, (2) significant investments specific to patents asserted against that product, and (3) the lack of protection from patent holdup. Speaking volumes, the “near-universal recognition” of the importance of FRAND commitments in the industry “is strong evidence that companies view holdup as a problem they must build institutions to avoid.”
C. Patent Holdout
Taking on another mantra of the anti-enforcement position, Shapiro and Lemley address the “supposedly parallel story of ‘patent holdout,’” which signifies patent owners “being deprived of the fruits of their R&D investments by implementers who copy their technology but refuse to pay.” The authors find that such an idea is “incoherent as a theoretical matter and rejected as an empirical matter.” Empirically, they explain that patent holdout fails to “explain what goes on in the technology industry unless it means failing to predict which of 500,000 patents, many of which you cannot see, will someday be asserted against technology you have developed yourself even though you have never heard of the inventor and they never built anything.” Conceptually, they explain that an inventor’s reward is based on “the incremental value of its invention, not the amount of money expended to achieve that invention or the risk involved” and that the patent owner, unlike the alleged infringer, “has the law on its side and can therefore shut down the defendant’s conduct.”
- Areas for Additional Development
The authors’ arguments are compelling. And their topic is so important that additional detail on two issues could be worth considering, perhaps in future work: standards organizations’ rules and the appropriate antitrust analysis.
A. Standard-setting Organizations
First, standard-setting organizations (SSOs) could benefit from more guidance on how to craft rules addressing patent holdup. For example, the authors praise the “recent salutary changes” at the Institute of Electrical and Electronics Engineers (IEEE) SSO. But they “take as given existing SSO policies, vagueness and all.” Future work could consider a menu of rules that SSOs could employ to prevent holdup. 1
For example, when might changes like those adopted by the IEEE be appropriate? In a 2015 Business Review Letter, the DOJ Antitrust Division signed off on four of these changes: (1) precluding injunctive relief unless the implementer “fails to participate in, or to comply with the outcome of, [a court] adjudication”; (2) limiting the royalty to exclude the value resulting from inclusion in the standard; (3) making clear that a patent holder adopting a commitment to license on fair, reasonable, and nondiscriminatory (FRAND) terms “cannot refuse to license its patents”; and (4) “permit[ting] a licensor to require a potential licensee to grant back a license to its own patents essential to the same standard, so that the licensor is not precluded from implementing the standard.”
Alternatively, how would the authors treat an SSO rule, like that adopted by the VITA association (which the DOJ agreed not to challenge in 2006), that “declare the maximum royalty rates and most restrictive non-royalty terms that the VITA member company . . . will request for any such patent claims that are essential to implement the . . . standard”? Or what about royalty-free licensing of SEPs, which has been incorporated into standards including USB and Bluetooth?
As a potential solution to the problems presented by patent holdup, guidance on these rules could be helpful.
B. Antitrust Framework
The authors also could more fully explore the contours of an antitrust cause of action. They appropriately note antitrust’s attention to compatibility standards in the past two decades, acknowledging how these standards promote competition and how SSOs can limit patent holdup through FRAND rules. They also accurately pinpoint how Delrahim led the DOJ Antitrust Division to dramatically reverse course, employing letters to SSOs, withdrawals from policy statements, and the filing of amicus briefs—including unprecedented ones attacking the FTC—to abandon a bipartisan consensus on the issue.
The authors also uncover the tension between the two prongs of Delrahim’s attack: (1) that “hold-up is fundamentally not an antitrust problem” and that (2) “antitrust shouldstop SSOs from trying to prevent SEP holdup.” Shapiro and Lemley explain that “to the extent that antitrust can step back in some settings, that is only possible because the market participants have recognized and responded effectively to the patent holdup problem by requiring reasonable licensing terms, and because the courts have enforced that requirement in contract or patent law.” The authors synthesize the problem: “While on the one hand Delrahim says that we don’t need antitrust because contract and equity will solve the patent holdup problem, on the other hand he is advocating policies that make it harder for contract and patent law to solve that very problem.” The authors point out that this “assault on SSOs and FRAND policies may actually necessitate more antitrust intervention in standard-setting . . . to protect a functioning market from SEP holdup.”
Shapiro and Lemley call on courts to “enforce reasonable SSO policies that target SEP holdup” and, when patent owners engage in holdup, ask antitrust “to step in to protect competition and consumers.” But they could go further in explaining their view of what an appropriate framework would look like. For example, do patentees that obtain or maintain monopoly power as a result of breaching a FRAND commitment engage in monopolization? Do FRAND breaches demonstrate exclusionary conduct in harming competitors in a manner that results in price increases or innovation harms for consumers? What about a transfer of a patent to a third party to evade a FRAND obligation? Is any of this conduct protected by the Noerr–Pennington doctrine that immunizes petitioning conduct? Attention to these questions could be useful.
- Radical Position
The authors capture the radical nature of Delrahim’s position. They note that he trumpets patent holdout, downplays patent holdup, and seeks to undo the landmark Supreme Court decision recognizing the propriety of a damages remedy, eBay v. MercExchange, in order to “giv[e] a patent owner the right to an automatic injunction.” The authors also outline many of the actions Delrahim took in furtherance of his agenda.
In this section, I supplement their argument, underscoring the drastic nature of Delrahim’s position by adding detail on his adoption of a property view not consistent with that field of law and his unjustified withdrawal from the reasonable 2013 statement of the DOJ and U.S. Patent and Trademark Office (PTO) on standard essential patents.
One of Delrahim’s most aggressive positions is that patents provide absolute property rights. He emphasizes the “core of what it means to hold an IP right – namely, the right to exclude.” He claims that patents are “a form of property, and the right to exclude is one of the most fundamental bargaining rights a property owner possesses.” And he asserts that patent rights “function best if an owner retains a right to exclude,” with the “[d]epriv[ation] . . . of this right . . . skew[ing] the bargain away from the free-market incentive scheme that the Constitution and Congress have established.”
Delrahim seeks support in the Constitution’s text, which provides that “Congress shall have the Power . . . to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.”He asserts that “the authors of the Constitution not only used the word ‘right,’ but . . . also preceded it with the equally important word ‘exclusive.”’ And he states that patent law “offer[s] incentives for holders of valid patents to seek the greatest rewards possible for their inventions.”
Delrahim’s aggressive position, however, is not supported by property law. Hornbook law does not give property owners absolute rights to exclude. There are at least 50 doctrines (such as adverse possession, easements, eminent domain, nuisance, and zoning) that limit property owners’ rights. Just to pick one example, landowners cannot exclude others from entering their land to save lives or property or to avoid some other serious harm. Relatedly, in upholding the inter partes review process for administratively reconsidering patents, the Supreme Court held that “[p]atents convey only a specific form of property right—a public franchise.”
Delrahim’s discussion of exclusive rights granted to patentees as a type of natural property rights also ignores the uncontroversial utilitarian framework for the patent grant. The Supreme Court has long made clear the primacy of the utilitarian justification. Exclusive rights exist not to bestow on patentees a moral right to a reward but to promote the best interests of society. That is why patents, like other forms of intellectual property, are subject to doctrines (like novelty, nonobviousness, the written description and enablement disclosure requirements, and a limited 20-year term) that ensure that protections for market competition balance patents’ incentive effects. Relatedly, it tells only half the story to focus on the incentive relevant to the initial invention while ignoring follow-on innovation, which is just as important and may be significantly undermined when patent owners abuse their FRAND obligations.
B. An Example
One example Shapiro and Lemley offer of Delrahim’s extreme position is his unceremonious 2018 withdrawal from the 2013 DOJ/PTO statement. In undertaking this maneuver, Delrahim stated that: a commitment to license on FRAND terms “should not create a compulsory licensing scheme”; that any discussion of injunctions should include not only patent holdup but also “holdout;” and that patent holders have a “full constitutional right to seek an injunction.” A closer look at the 2013 Statement’s reasonableness shows the radical nature of Delrahim’s gambit.
The critiques Delrahim raises are not faithful to the 2013 Statement, which highlighted the balance of “promoting both appropriate compensation to patent holders and strong incentives for innovators to participate in standard-setting activities.” 2 The Statement emphasized innovation, with the agencies “strongly support[ing] the protection of [IP] rights” and patent holders making FRAND promises “receiv[ing] appropriate compensation that reflects the value of the technology contributed to the standard.” The Statement recognized the “importan[ce] for innovators to continue to have incentives to participate in standard-setting activities” and for “technological breakthroughs in standardized technologies to be fairly rewarded.” And the Statement reasonably refused to support “one-size-fits-all mandates for royalty-free or below-market licensing,” which would “undermine the effectiveness of the standardization process and incentives for innovation.”
In addition to recognizing the importance of innovation, the Statement addressed injunctions. But its framework was a lot more reasonable than Delrahim intimated. For starters, it acknowledged that “[i]n some circumstances, the remedy of an injunction or exclusion order may be inconsistent with the public interest.” This concern was “particularly acute” when an exclusion order is inconsistent with a FRAND promise, such as where a patent owner “use[s] an exclusion order to pressure an implementer of a standard to accept more onerous licensing terms than the patent holder would be entitled to receive” pursuant to the FRAND commitment. And the concern was “exacerbated when patents are sold or otherwise transferred by their owners,” as holdup is more likely when the FRAND obligation does not “travel with a transferred patent.”
At the same time, the Statement recognized that an exclusion order “may still be an appropriate remedy,” such as when “the putative licensee is unable [to] or refuses to take” a FRAND license. For example, the licensee could “refuse to pay what has been determined to be a [FRAND] royalty” or “refuse to engage in a negotiation.” The Statement explained that such a refusal could take the form of “a constructive refusal to negotiate,” like when an implementer “insist[s] on terms clearly outside the bounds of what could reasonably be considered” to be FRAND terms “in an attempt to evade the putative licensee’s obligation to fairly compensate the patent holder.” And an exclusion order could be appropriate if “a putative licensee is not subject to the jurisdiction of a court that could award damages.”
In short, the Statement’s treatment of injunctions was balanced and structured. As naturally accompanies a promise to reasonably license patents, patent owners will not always be able to obtain injunctions, especially when such relief is inconsistent with their FRAND promises. On the other hand, the Statement set forth several categories in which the patent owner was able to obtain an injunction. And these categories included settings in which the implementer, lacking good faith, constructively refused to negotiate. Perhaps that is why the Federal Circuit relied on the Statement in overturning a district court that had “applied a per se rule that injunctions are unavailable for SEPs” and delineating when “an injunction may be justified.” In other words, the most important patent court relied on the Statement to explain why injunctions are warranted! Delrahim’s jettisoning of this reasonable compromise speaks volumes. And as I have explained elsewhere, given the reasonable nature of the 2013 Statement, the statement issued in 2019 by the DOJ, PTO, and National Institute of Standards and Technology (NIST) rests on a rickety foundation of omissions, strawmen, and generalities.
- Academic Debate
Stepping back to the biggest picture, the final point involves the nature of academic debate. Shapiro and Lemley have written a thoughtful and persuasive article that situates patent holdup in the literature of transaction cost economics. They have shown the weaknesses, and radical nature, of Delrahim’s approach. And they have made a compelling claim for antitrust’s role in the standard-setting context.
But what if all of that is not enough?
What if the only thing that matters in enforcement today is the existence of unsupported assertions and scholarship on Delrahim’s side? And what if much of this output is the result of—as Mark Lemley and Timothy Simcoe put it—Qualcomm’s “extraordinary” funding that includes “the creation of entire centers [and] scholarly papers?”
I received a first-hand example of this in 2018. In response to Delrahim’s speeches, I, along with former FTC Chairman Tim Muris, sent a letter on behalf of 77 former government officials and academics. The letter offered eight critiques of the position Delrahim had articulated. The critiques were based on:
- (1) the bipartisan recognition of the anticompetitive harms of patent holdup;
- (2) courts’ and SSOs’ recognition of the holdup problem;
- (3) the more serious antitrust concern presented by patent holdup than holdout;
- (4) the existence of an antitrust case from patentees obtaining or maintaining monopoly power by breaching a FRAND promise;
- (5) patents’ inability to provide an unqualified right to exclude;
- (6) appropriate caselaw precluding an automatic injunction in cases of patent infringement;
- (7) patents’ justification based on utilitarianism, not natural property rights; and
- (8) FRAND promises not reflecting compulsory licensing.
Delrahim responded the next day. But he did not address the critiques. In fact, he did not answer even one. He stated only that he “also received a letter from a number of antitrust and intellectual property scholars, including federal judges, in support of the United States’ policies.” And he attached the letter, which did not respond to any of the critiques. That response, of course, does not answer the question. The existence of support for one’s position does not substantively engage the debate.
This is a concerning development. The eight critiques, two years later, are still on the table. Robust debate would require an engagement of these fundamental flaws in Delrahim’s position.
The same concern confronts Shapiro and Lemley’s article, which persuasively (1) situates holdup in the long history of transaction cost economics, (2) shows why holdup is not likely to be observed, and (3) explains why patent holdup is a particularly concerning example of the phenomenon. But even if the article is correct, what if it’s ignored? What if Delrahim simply points to the papers supporting his position? That is not a debate. And that does not form the basis for justified antitrust enforcement.
On an issue so central to the economy, government enforcers have an obligation, especially when embarking on an abrupt detour from a longstanding, bipartisan approach, to justify their actions.
Shapiro and Lemley have written an important article that situates the robust debate about patent holdup in the history of transaction cost economics. As we go forward, the merits of this and similar arguments should matter in determining the appropriate treatment of anticompetitive standards-based behavior.