Vol. 98, No. 4

Fair Use and Fair Price

Roy Baharad* & Gideon Parchomovsky†
Fair Use, Intellectual Property, IP Law
Articles

In this Article, we present and develop a new justification for the fair use doctrine. The accepted lore among copyright law scholars is that fair use is a means for overcoming a market failure in the form of high transaction costs. According to this view, the doctrine sanctions unauthorized use of copyrighted works in cases where transaction costs hinder voluntary, mutually beneficial exchanges.

Departing from conventional wisdom, we argue that the fair use doctrine serves as an important empowerment even in fully functional markets. Fair use enables users to secure more favorable licensing terms from copyright owners by endowing users with a threat point in their negotiations. Without fair use, users would have to pay the price demanded by copyright owners or not use the work. With fair use, many users can credibly assert that their intended use of copyrighted content is privileged by the fair use doctrine and thus they can use the desired content without authorization. The fair use doctrine, therefore, gives users leverage in their negotiations with copyright owners.

We illustrate our thesis by applying it to the landmark fair use decisions of the Supreme Court, including the recent ruling in Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, that determined the bargaining standpoint of users for decades to come. We demonstrate the distributive effects of the Supreme Court’s fair use jurisprudence and explain how the Court must act to preserve and augment the empowering effect of fair use. The theory presented in this Article proves that the reach of fair use goes well beyond market failures and that the impact of the doctrine is much more significant than previously thought.

INTRODUCTION

Considered by many as the most important doctrine in our copyright law system,¹ fair use sanctions certain unauthorized uses of copyrighted works that would otherwise constitute a copyright infringement. Under the fair use doctrine, when a use is considered fair, the user is relieved of liability and need not pay compensation to the copyright owner whose content they used. Therefore, fair use may be conceptualized as a doctrine that confers upon deserving users a private taking power over copyrighted content that can be exercised at a zero price.²

Since its inception in the English common law and equity courts, the fair use doctrine has never ceased to fascinate theorists and students.³ The existence of the doctrine raised two critical challenges for scholars. First, under what circumstances should a use be considered fair? Second, why should fair users be fully relieved of the duty to compensate copyright owners?

The answer to both these questions has been provided by Professor Wendy Gordon. In a pathbreaking article authored almost forty years ago, Gordon conceptualized fair use as a means for overcoming a market failure in the form of high transaction costs.⁴ Gordon persuasively argued that courts should recognize fair use when three cumulative conditions obtain: first, high transaction costs prevent voluntary market exchange between copyright owners and users; second, the allegedly fair use is socially beneficial; and, third, a fair use finding would not unduly undermine incentives to create.⁵ Gordon’s key insight was that when transaction costs are prohibitive, there will be no voluntary trade between copyright owners and users. In this scenario, allowing users to use copyright content for free benefits the user without harming copyright owners, for the latter would not be able to collect payments from users as transaction costs bar voluntary exchanges. Gordon correctly submitted that under these circumstances, allowing users to use copyright content for free, by classifying their use as fair, is welfare enhancing.

By tying fair use to the level of transaction costs, Gordon, at once, provided a cogent defense of the fair use doctrine and exposed its vulnerability. Professor Tom Bell and other scholars pointed out that in an interconnected world, where technological advancements constantly lower transaction costs, there may no longer be a need for fair use, and at a minimum, courts’ willingness to recognize fair uses should diminish.⁶ Importantly, this skeptical view of fair use has not been universally endorsed by other scholars, who advanced other utilitarian and non-utilitarian justifications for fair use.⁷

In this Article, we develop a new justification for fair use that is radically different from prior theorizing. Our theory seeks to complement and reinvigorate the theories of scholars who view fair use as an essential component of our copyright system. Yet, our outlook places fair use on a very different ground from past scholarship. We contend that fair use’s most significant yet overlooked role is to facilitate bargaining between copyright owners and users, even when transaction costs are low. We argue that fair use helps users not only in those unrepresentative and rare cases where transaction costs prevent consensual transactions between copyright owners and users, but also in the more common case where transaction costs are low or nonexistent. Fair use does this by improving the bargaining power of users and giving them leverage or a threat point vis-à-vis copyright owners.

Although fair use has been traditionally considered an open-ended and unpredictable doctrine,⁸ two recent developments have infused a certain level of predictability into this area of the law. First, a close reading of fair use cases uncovers, what we call, two fair use clusters: transformative uses and uses that yield a substantive public benefit.⁹ Second, an empirical study by Professor Barton Beebe established a meaningful positive correlation between two of the statutory factors that courts are instructed to weigh in making fair use determinations—the purpose of the use and the effect of the use on the market for the copyrighted work.¹⁰ This emergence of clusters of uses or activities that have a legitimate claim for fair use is a welcome development for users, as it manifests in increased licensing leverage.

Consider transformativeness. Transformativeness covers a wide range of uses. While the level of transformativeness varies among uses, all users who engage in transformative uses of copyrighted works have a colorable fair use claim. Obviously, not all of them would win a fair use ruling in court, but each can credibly argue in negotiations with copyright owners that their use would be found fair with a certain probability. Since transformativeness, as recently established by the Supreme Court, “is a matter of degree,”¹¹ the level of transformativeness can be represented on a unit interval, essentially reflecting its probability of enjoying a fair use defense. This probability, in turn, immediately translates into licensing leverage when negotiating with a monopolistic rightsholder.

To illustrate, imagine that Anne plans to make a transformative use of a photograph in which Bob owns a copyright. Assume that there is 0.5 probability that Anne’s use would be found fair if she uses Bob’s photograph without his permission and a lawsuit ensues. Assume further that Bob typically demands $100 per license from users who wish to use his photograph. Anne, however, owing to her potential fair use claim, should be able to secure a license for $50—a price that reflects the strength of her fair use claim.

Now consider a different user, Carol, whose planned use of Bob’s work is slightly less transformative than Anne’s. The probability of Carol’s use being fair is only 0.3. Yet, all things being equal, she, too, should be able to receive a license from Bob for a lower price than his original asking price—$70, instead of $100. As we shall demonstrate throughout this Article, these examples are representative.

The same is true for a host of other transformative users—users who can claim that their derivative works bestow a significant benefit on the public, and users who appropriate only a small portion from copyrighted works. All of them have a certain individual probability of succeeding on a fair use claim when sued by the copyright owner. The license price each would be able to negotiate depends on the strength of their fair use claim, as well as on their bargaining power. But critically, each should be able to negotiate a license for a price lower than the asking price of the copyright owner. We use a formal model to demonstrate this result and identify its robustness over a wide range of scenarios. As we will show, even a small probability of obtaining a fair use ruling may well change the bargaining outcome between users and copyright owners.

The Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith decision that was recently issued by the Supreme Court provides a powerful example of our theory.¹² In this case, the Supreme Court had to determine whether fourteen unauthorized silkscreens and two unauthorized pencil drawings that were produced based on a Lynn Goldsmith photograph qualify as fair uses of the photograph.¹³ At the heart of the matter lies the definition of the term “transformativeness.” In rejecting the foundation’s fair use claim, the Court held that a transformative work must have a fundamentally different and new artistic purpose and character.¹⁴ This definition is clearly narrower than that endorsed by the Supreme Court in Campbell v. Acuff-Rose Music, Inc., according to which a transformative work is one that has a different purpose or conveys new message or meaning.¹⁵ The Supreme Court’s adherence to a rather narrow interpretation of transformativeness will carry far-reaching implications not only for the parties to the case, but also for a multitude of users who negotiate with copyright owners.

It bears emphasis that negotiation is often a prelude to litigation. In fact, many of the celebrated fair use cases, including Google LLC v. Oracle America, Inc.,¹⁶ Authors Guild, Inc. v. Google Inc.,¹⁷ and Campbell,¹⁸ were filed only after the parties tried, and failed, to negotiate a consensual agreement. Furthermore, even after a case is brought to court, the litigating parties attempt to negotiate a settlement while the case is pending. Indeed, such was the case in Oracle and Authors Guild. Naturally, not all negotiations result in a successful outcome, and some copyright disputes will inevitably end up in court. But it should be underscored that in any regular case, in which the existence of transaction costs does not hinder trade, the rightsholder-user negotiation becomes the primary apparatus for allocating use of copyrighted content; litigation is nothing but a complementary mechanism. This understanding implies that fair use theory must account for the doctrine’s impact on the negotiation process.

Our analysis shows that irrespective of its actual invocation in court, the fair use doctrine provides considerable benefits to users by improving their bargaining position. In economic parlance, the theory advanced by this Article views fair use as empowering significant categories of users by improving what negotiation theorists call the “best alternative to a negotiated agreement” (“BATNA”).¹⁹ By diminishing the BATNA of copyright holders and bolstering that of users, the fair use doctrine redesigns the licensor-licensee relationship to promote the use and distribution of copyrighted content.

The present Article thus conceptualizes fair use not as a mechanism that overcomes transaction-costs-related market failures, but rather, as a doctrine that rests the foundations for just and efficient bargaining framework. In this regard, we analyze the behavior of relevant economic actors that operate “in the shadow” of the fair use doctrine.²⁰

We develop our argument in accordance with the following structure. In Part I, we explain the fair use doctrine and discuss the theories that have been developed to justify it. In Part II, we introduce a new theory of fair use and explicate how it differs from the extant literature. In Part III, we present the normative attractiveness of our theory by highlighting its positive welfare and distributive implications. A short Conclusion ensues.

I. Fair Use: Law and Theory

A. The Fair Use Doctrine

The fair use doctrine made its first appearance in the U.S. in the 1841 case of Folsom v. Marsh.²¹ According to scholars, however, its origins are far more ancient. Professor Matthew Sag, for example, suggests that the fair use doctrine “predate[d] Folsom v. Marsh by at least 100 years.”²² Professor William Patry went even further and dated the doctrine back to the year 1710 when the Statute of Anne was enacted.²³ Other researchers trace the roots of the fair use doctrine to the common law and natural rights conceptions.²⁴

In 1976, the fair use doctrine was codified and became part of the Copyright Act. 17 U.S.C. § 107 (“section 107”), in which the fair use doctrine is enshrined, opens with a preamble that offers a non-exhaustive list of presumptively fair uses, including, “criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research.”²⁵ Then, it proceeds to enumerate four factors that courts ought to consider when making fair use determinations: “(1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; [and] (4) the effect of the use upon the potential market for or value of the copyrighted work.”²⁶

It is important to note that the codification of the fair use doctrine did not purport to change its nature as an equitable doctrine. In the accompanying House Report, Congress referred to the fair use doctrine as “an equitable rule of reason.”²⁷ The statutory formulation of the doctrine, therefore, preserved the flexibility and open-endedness that have become the hallmark of the fair use doctrine. As Judge Pierre Leval wrote in his 1990 classic article on fair use:

What is most curious about this doctrine is that neither the decisions that have applied it for nearly 300 years, nor its eventual statutory formulation, undertook to define or explain its contours or objectives. . . . [They] furnish little guidance on how to recognize fair use. The statute, for example, directs us to examine the “purpose and character” of the secondary use as well as “the nature of the copyrighted work.”²⁸

Even more remarkable was Judge Leval’s admission that although “courts have treated the definition of the doctrine as assumed common ground[, t]he assumption of common ground is mistaken. Judges do not dineshare a consensus on the meaning of fair use.”²⁹ Indeed, the courts’ inability to converge on a common understanding of fair use has frustrated copyright scholars, leading Professor Larry Lessig to conclude that the fair use doctrine amounts to nothing other than “the right to hire a lawyer to defend your right to create.”³⁰

While scholars have bemoaned the uncertainty that shrouds the fair use doctrine³¹ and, over the years, have advanced various proposals in order to cabin the unpredictability of fair use,³² in this Article, we take a different tack. We argue, contrary to conventional wisdom, that the inherent uncertainty of the fair use doctrine might actually help, rather than harm, users. The possibly virtuous effect of uncertainty has been overlooked by theorists as they have focused exclusively on the litigation arena. We, by contrast, are interested in the effect of fair use outside of the courtroom, in negotiations between copyright owners and users. We develop this argument fully in Part II of this Article, as part of our transactional model of fair use. But before elaborating on the potential virtues of uncertainty, it behooves us to complete our discussion of the development of the fair use doctrine and highlight some critical recent developments.

Four years after Judge Leval’s 1990 portrayal of fair use as a helplessly underminable doctrine, the fair use landscape was reshaped. The turning point was the Supreme Court’s 1994 decision in Campbell v. Acuff-Rose Music, Inc.. In finding 2 Live Crew’s version of Roy Orbison’s and Bill Dees’s copyrighted song “Oh Pretty Woman” a fair use, the Court stated that “the more transformative the new work, the less will be the significance of other factors, like commercialism, that may weigh against a finding of fair use.”³³ As we will show, lower courts were quick to follow the language and spirit of the Campbell opinion. Since Campbell, transformativeness has become the currency of the fair use realm.³⁴

The importance of transformativeness to fair use determinations was highlighted by Judge Leval in his classic article.³⁵ Yet, it was the Supreme Court’s endorsement of transformativeness in Campbell that officially conferred upon it an elevated status.³⁶ Subsequent cases have substantially increased the weight granted to transformativeness in fair use determinations. Among others, courts held that a challenged work is worthy of protection whenever it “contains significant transformative elements,”³⁷ is “sufficiently transformative,”³⁸ or is endowed with a “patently transformative character.”³⁹ It has been further emphasized that a central purpose of fair use investigations is to establish “whether and to what extent the new work is ‘transformative.’ ”⁴⁰ In other instances, courts submitted that copyright infringement, notwithstanding the commercial character of the work in question, may not be determined given that the original work is “used for a transformative purpose.”⁴¹

The effect of the rise of transformativeness can be best seen in the context of fair use cases involving appropriation art. Appropriation art is an art form predicated on the use of existing objects with subtle modifications.⁴² Cases on appropriation art include the Second Circuit’s famous Blanch v. Koons⁴³ and Cariou v. Prince.⁴⁴ In both cases, the court was persuaded of the sufficient transformation that the process of appropriation can embed. In Blanch, the court explained that Jeff Koons’s appropriation of Blanch’s photograph “was intended to be—and appears to be—‘transformative.’ ”⁴⁵ In Cariou, the court was more hesitant but nonetheless concluded that Richard Prince’s appropriating work “could be transformative even without commenting on Cariou’s work or on culture, and even without Prince’s stated intention to do so.”⁴⁶

A closely related development in fair use jurisprudence that can also be traced back to Campbell involves parodies. A parody is defined as a derivative work whose purpose is to criticize or comment on a preexisting copyrighted work.⁴⁷ After classifying 2 Live Crew’s unauthorized rendition of “Oh Pretty Woman” as a parody, the Court likewise ruled that parodic uses are especially transformative and therefore have an especially strong claim for fair use.⁴⁸ The Court added that copyright owners are unlikely to authorize parodies that are of unflattering commentary or mockery to their works,⁴⁹ and thus, the production of parodies critically depends on fair use.

The Supreme Court’s ruling in Campbell has been understood by lower courts and commentators to create something of a safe harbor for parodies.⁵⁰ As Professor Pamela Samuelson observed, “[n]otwithstanding the Court’s unwillingness in Campbell to presume that parodies are fair, every subsequent parody case has been adjudged a fair use.”⁵¹ For example, in Suntrust Bank v. Houghton Mifflin Co., the Eleventh Circuit ruled that Alice Randall’s “The Wind Done Gone,” a critical literary account of Margaret Mitchell’s “Gone With the Wind,” was a parody of the original and was therefore a fair use of it.⁵² Likewise, in Burnett v. Twentieth Century Fox Film Corp., the Court ruled that an audiovisual work produced by the creators of Family Guy, which poked fun at the figure of Carol Burnett, constituted a fair use on account of its parodic nature.⁵³ Similarly, in Brownmark Films, LLC v. Comedy Partners, the Seventh Circuit held that a music video by the creators of South Park which parodied the plaintiff’s music video has “obvious transformative value.”⁵⁴ Recently, the Second Circuit openly acknowledged that in the aftermath of Campbell, “parody, which ‘needs to mimic an original to make its point,’ . . . is routinely held transformative.”⁵⁵

Another line of cases, beginning with Kelly v. Arriba Soft Corp.,⁵⁶ highlighted another key determinant of fair use: social benefit. Importantly, this category, too, grew out of the Supreme Court’s Campbell decision, with its emphasis on transformativeness. In Campbell, the Court suggested that a “parody has an obvious claim to transformative value” because “it can provide social benefit, by shedding light on an earlier work, and, in the process, creating a new one.”⁵⁷ The fair use cases that thrust public benefit to the forefront of the fair use analysis predominantly involved users from the technology sector. In Kelly, the Ninth Circuit ruled that the public display of photographs in the form of thumbnails as part of the operation of a visual search engine constituted fair use.⁵⁸ In reaching this conclusion, the court mentioned the benefit conferred on the public by the appellee’s search engine.⁵⁹ Approximately four years later, in 2007, in Perfect 10, Inc. v. Amazon.com, Inc., the Ninth Circuit was asked to revisit the issue, when an adult content company sued Google, alleging that its authorized display of its copyrighted photos as thumbnails in response to users’ searches constituted copyright infringement.⁶⁰ Finding that Google’s use was fair, the Ninth Circuit emphasized the “significant public benefit” of Google’s search system, explaining:

[A] search engine provides social benefit by incorporating an original work into a new work, namely, an electronic reference tool. Indeed, a search engine may be more transformative than a parody because a search engine provides an entirely new use for the original work, while a parody typically has the same entertainment purpose as the original work.⁶¹

The social benefit factor quickly found its way to the decisions of other circuits.⁶² It played an important role in the Second Circuit’s ruling in Authors Guild v. Google, Inc..⁶³ The case was brought after Google decided to establish a large, searchable digital repository of literary works by scanning the books in several libraries around the world, including the libraries of Harvard University, The University of California, Stanford University, The University of Michigan, Columbia University, Princeton University and the New York Public Library.⁶⁴ The goal was to allow users to search the content of the books, yet the content of books could not be copied and only snippets from books were displayed to users in response to their searches.⁶⁵ Nonetheless, after prolonged negotiations between the parties failed to yield a settlement, the Authors Guild sued Google for copious copyright infringements.⁶⁶ In reaching the conclusion that Google’s use was fair, the district court dedicated a full section of its decision to a discussion of the myriad public benefits arising from Google’s use.⁶⁷ On appeal, the Second Circuit affirmed the district court’s decision, awarding Google an important legal victory, but toned down the importance of the public benefits provided by Google.⁶⁸

The final imprimatur of public benefit as a key determinant of fair use was given by the Supreme Court in its 2021 decision in Google LLC v. Oracle America, Inc.⁶⁹ In a 6-2 decision, the Supreme Court ruled that Google’s unauthorized appropriation of 11,500 lines of Oracle’s Java Application Programming Interface (“API”) for the Android operating system constitutes fair use.⁷⁰ Writing for the majority, Justice Breyer noted the added value created by Google’s use for third parties:

Here Google’s use of the Sun Java API seeks to create new products. It seeks to expand the use and usefulness of Android-based smartphones. Its new product offers programmers a highly creative and innovative tool for a smartphone environment. To the extent that Google used parts of the Sun Java API to create a new platform that could be readily used by programmers, its use was consistent with that creative “progress” that is the basic constitutional objective of copyright itself.⁷¹

Importantly, it was not only the courts that helped dispel some of the doctrinal mist that enveloped the fair use doctrine for centuries. Academics, too, have managed to point to overarching principles that affect fair use determinations. Two such efforts are worthy of special note.

In two separate projects, the first published in 2008⁷² and the second in 2020,⁷³ Professor Barton Beebe empirically examined fair use case law to find correlations between various fair use factors and case results. His method allowed him to pierce the judicial rhetoric and examine which statutory factors are outcome determinative in fair use cases. In his 2008 article, which surveyed the opinions from all fair use cases issued between 1978 and 2005, Beebe noted that “[i]t appears . . . that courts and commentators have exaggerated the influence of transformativeness doctrine on our fair use case law.”⁷⁴ However, in his 2020 article that analyzed the opinions from all fair use cases issued between 2005 and 2019, Beebe reports that “while the transformativeness test appeared to be waning in influence by 2005, it has since recovered its previous level of influence, even in the lower-profile, workaday fair use opinions that make up the majority of the data.”⁷⁵ Along similar lines, Clark Asay, Arielle Sloan and Dean Sobczak have empirically established that if courts perceive a certain use as transformative, it would almost invariably qualify for fair use protection, which implies that transformativeness is essentially a sufficient condition for enjoying the fair use defense.⁷⁶ These recent findings suggest an important confluence between judicial rhetoric and reality.

The Supreme Court’s recent decision in Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith⁷⁷ unfolded the most recent development in fair use jurisprudence. Issued in May 2023, the ruling is the first application of fair use in the arts and entertainment industry since the 1994 Campbell ruling. The Warhol case involves a series of Prince’s photos taken by photographer Lynn Goldsmith in 1981 to accompany a Vanity Fair article concerning Prince.⁷⁸ Unbeknownst to Goldsmith, the magazine solicited Andy Warhol to create a stylized painted version of the photograph that became known as the “Orange Prince” silkscreen.⁷⁹ Vanity Fair published the Orange Prince portrait in its November 1984 edition, and Goldsmith was co-credited for this work.⁸⁰ Orange Prince, however, was only one of sixteen painted versions created by Warhol, collectively known as Warhol’s “Prince Series.”⁸¹ Pursuant to its completion, the Prince Series has been routinely displayed in museums and galleries.⁸² After Prince’s passing in 2016, Condé Nast, the official publisher of Vanity Fair, published a commemorative magazine entitled “The Genius of Prince,” with the Orange Prince portrait embellishing its cover.⁸³ Alleging copyright infringement, Goldsmith argued that notwithstanding her initial 1984 licensing agreement with Condé Nast to use her photo one time as an “artistic reference,” she was unaware of the Orange Prince silkscreen—as well as of the Prince Series at large—until its reintroduction as part of the commemorative 2016 edition.⁸⁴

Once Goldsmith learned of the aforementioned facts, she brought an infringement suit against the foundation. The Southern District of New York granted the foundation’s motion for a preliminary ruling that Warhol’s work, though it incorporated Goldsmith’s photograph, was sufficiently transformative to qualify as a fair use.⁸⁵ The ruling was then overturned by the Second Circuit.⁸⁶ Finding that Warhol’s portrait incorporated no significant addition or alteration upon Goldsmith’s original photograph, the court concluded that the portrait was infringing.⁸⁷ In a 7-2 decision, the Supreme Court elected to affirm the Second Circuit’s conservative interpretation of transformativeness, thus rejecting the appeal and siding with Goldsmith.⁸⁸ The Warhol case thus marks a potential deviation from status quo interpretation of the fair use doctrine. We analyze the implications of the rulings to our theory in Part II, below.

The preceding discussion teaches that the Supreme Court’s approach to fair use has interjected a certain degree of certainty and predictability into the doctrine. This is not to say that fair use has become a clear doctrine—far from it—but it can be generally stated that in the aftermath of Campbell, parodic uses, sufficiently transformative uses, and uses that produce significant public benefits are likely to be found fair. As we will explain in Part II, the creation of these fair use clusters, or silos, is of vital importance to our theory of fair use. But before introducing our own theory of fair use, we must give credit to prior theorists and discuss their contributions. It is to this task that we next turn.

B. Theoretical Justifications of the Fair Use Doctrine

As befits a doctrine of its significance, scholars have advanced several theories to justify fair use. In this Section, we review the leading theoretical justifications of fair use. It should be emphasized at the outset that our goal is not to discredit other theories or even criticize them. As we will show, our theory of fair use complements existing theories by elucidating a central function of fair use that has hitherto evaded other scholars. The goal of the proceeding discussion is twofold: first, we wish to map the theoretical landscape of fair use, so we can precisely locate our own theory within it. Second, we seek to show how our justification of fair use interacts with the extant theoretic literature.

The most dominant theory has been put forth by Wendy Gordon.⁸⁹ In an immensely influential article, Gordon argued that the fair use doctrine is a mechanism for allowing the use of copyrighted content when voluntary transactions between copyright owners and users are barred by high transaction costs.⁹⁰ Specifically, Gordon postulated that fair use should be recognized when three cumulative conditions are met: (1) high transaction costs prevent consensual bargaining between copyright owners and users; (2) the unauthorized use is socially desirable; and (3) legitimizing the disputed use would not undermine incentives to create.⁹¹

To illustrate the operation of Gordon’s theory, imagine a student who wishes to quote a copyrighted manuscript. Assume that the author of the manuscript charges $20 for the requested use and that is also the maximum price that the student is willing to pay. However, transaction costs—defined as the cost of identifying the counterparty to the transaction, negotiating and formalizing an agreement with them, and enforcing the agreement—would bar the exchange from taking place. Under these circumstances, allowing the student to quote the manuscript without permission would make them (and society at large) better off without harming the author. The author, in our example, could not receive payment from the student, not because they refused to pay, but rather owing to the fact that the level of transaction costs made payment impossible. Hence, the author stands to lose nothing if the student’s use is considered fair and the student receives a benefit of $20. Society, too, is benefitted by the award of fair use to the student since the realm of creativity is enriched, while incentives to create future works are not harmed.

Gordon’s approach grounded fair use in economic theory, proving that the doctrine can be welfare enhancing when applied properly. However, the tie Gordon created between fair use and high transaction costs has proven to be a double-edged sword. As the title of her article, Fair Use as a Market Failure suggests, Gordon justified fair use as a means of overcoming a market failure in the form of high transaction costs. This was a great strength of Gordon’s theory, but also a potential weakness. Critically, Gordon’s theory established a direct correlation between the level of transaction costs and the cases in which fair use should be recognized. The problem is that improvements in telecommunications and computer technologies have dramatically lowered transaction costs since Gordon published her article in 1982, putting a lot of strain on Gordon’s theory. Critics of Gordon’s theory suggested that in an interconnected world, fair use has outlived its “raison d’être” and all uses must be secured via payment.⁹² It bears emphasis that Gordon responded to her critics by broadening and sharpening the role of market failure in her account. Yet, Gordon’s account remains inextricably related to the concept of market failure, and it conceives fair use as a mechanism that operates in non-transactional settings, i.e., settings in which voluntary transactions are vitiated.⁹³

A slightly different justification of fair use that complements Gordon’s original justification has been offered by Ben Depoorter and Francesco Parisi. Depoorter and Parisi correctly pointed out that technological advancements do not lower all transaction costs.⁹⁴ The problem of strategic holdouts remains, even in the face of technological advancements. The holdout problem is especially acute for users, such as creators of documentary works, who need to secure permission from multiple copyright owners. In such cases, each copyright owner possesses veto power over the planned use and may strategically exercise it to the detriment of users. The fair use doctrine allows users to carry out their creative projects despite strategic attempts by copyright owners to hamper the enterprise.⁹⁵ Depoorter’s and Parisi’s account is both persuasive and elegant, but “it potentially limits the usefulness of the market-failure theory to only cases in which a user must clear multiple rights and has no other alternatives.”⁹⁶

Two additional justifications of fair use focus on allocative efficiency. The first, associated with William Fisher, calls on courts to use fair use to promote the goals of copyright protections.⁹⁷ To this end, Fisher asks courts to think of all possible uses of copyrighted works. Then, for each use, he calls on courts to design what can be called an “efficiency ratio,” with the numerator representing the profit an author could realize if fair use is denied and the denominator representing the loss to society if the copyright owner refuses to license the work.⁹⁸ A high efficiency ratio suggests, per Fisher, that the use is probably unfair, while a low one indicates that the use is fair.⁹⁹ Fisher sets the cutoff at the use with the highest marginal aggregate social gain (gain to society from creation minus loss to society from monopoly control over uses).¹⁰⁰ As Fisher himself admits, his approach to analyzing fair use, at least with respect to some sorts of works, “is nearly coterminus [sic] with economic analysis of the copyright system as a whole.”¹⁰¹ Fisher likewise acknowledges that his approach is probably impractical in light of the informational burden it imposes on judges.¹⁰²

A different allocative justification, termed the “dual-grant theory” of fair use, was constructed by Professor Abraham Bell together with one of this article’s authors.¹⁰³ The dual-grant theory maintains that in designing our copyright system, Congress created two blocs of uses, allocating one to authors and the other to the public. Each group was given the uses which it values most. Accordingly, authors received “standard commercial uses,” while the public received “uses of highly dispersed social value,”¹⁰⁴ such as political speech, and uses that promote the pursuit of science, knowledge and truth.¹⁰⁵ By contrast to Fisher’s theory of fair use, which requires courts to make individual case-by-case determinations of fair use, the dual-grant theory calls on courts to make fair use determinations based on the category of uses to which the challenged use belongs. This, in turn, renders the approach a lot more practical but less precise compared to Fisher’s framework of analysis.

II. Fair Use as Licensing Leverage

As our discussion in Part I demonstrates, extant theories of fair use perceive it as a mechanism for bypassing the market and a substitute to owner-user bargaining. Importantly for the purpose of our analysis, all existing justifications of fair use focus exclusively on the courts. The underlying assumption in all four accounts is that fair use serves users exclusively in litigation or as a tool for overcoming negotiation breakdowns or high transaction costs that prevent negotiations from occurring ab initio.

The justification we develop in this Part focuses on the effect of fair use on enabling voluntary transactions between copyright owners and users. By contrast to prior justifications of fair use, we show that the main function of fair use is to facilitate bargaining, rather than replace it. We call this effect “the hidden function of fair use.” As we demonstrate, the fair use doctrine allows users to secure more favorable licensing terms from copyright owners. Importantly, our theory does not compete with any of the prior theories. Rather, it complements all four of them. This complementarity suggests that the effect of fair use is broader and deeper than previously believed.

A. The Theory

In their acclaimed bestseller on negotiation theory, Getting to Yes, Professors Roger Fisher and William Ury famously stress that “the relative negotiating power of two parties depends primarily upon how attractive to each is the option of not reaching agreement.”¹⁰⁶ They therefore submit that a sine qua non¹⁰⁷ for A’s bargaining advantage is that their benefit from a state of disagreement exceeds B’s, since, in such case, B would be willing to sacrifice more in order to reach an agreement, which diverts the terms of agreement formation in A’s favor. Normally, the party with the upper hand in a bargaining setting is the one who possesses a stronger BATNA.¹⁰⁸ To exemplify, suppose that A is interested in selling B a used car, which B values at $10,000. Realizing this, A can require up to $10,000 to secure an agreement between the two. Assume now that C offers a similar car for sale, in exchange for only $5,000. C’s entry affects the negotiations between A and B: the maximum price that B would be willing to pay for A’s car now drops to $5,000. C’s entry, by providing B with an alternative to negotiating with A, enhances B’s BATNA.

The fair use doctrine has a similar effect to that of C’s entry in the previous example: it elevates users’ BATNA and thereby confers upon them significant leverage in negotiations with copyright owners. A helpful way to see this effect of fair use is to think of it as a call option the law gives to users over copyrighted content with a strike price of zero if their use is found fair. Recall that a fair use finding means that a user is not only free to use copyrighted content without permission, but also that they are allowed to do so free of charge.

To illustrate, suppose that Ella, an artist, is interested in using Francine’s copyrighted work for a transformative purpose and contacts her to acquire authorization. Assume that Ella values Francine’s work at $8,000 and is willing to pay Francine up to this sum in order to secure authorization. If Ella were to use Francine’s work without permission, she would be sued for copyright infringement and a court would order her to pay Francine $10,000 in damages.¹⁰⁹

To see the effect of fair use on negotiation dynamics, consider first a world without fair use. In this world, users can only use copyrighted content permissively. Any attempt at bypassing the market would be remedied by damages, an injunction, or both. Under a legal regime that does not recognize fair uses, Ella has no feasible alternative to negotiating with Francine. Using the work without authorization is not a viable option from Ella’s perspective, as it represents a negative net value (-$2,000): while Ella would receive a benefit of $8,000 from her use, she ought to pay $10,000 in damages. Therefore, in economic parlance, Ella has no credible threat of using without authorization: both parties know that if negotiations fail, Ella will just have to forgo her planned use. Consequently, Ella is willing to pay any price up to $8,000 (her valuation of Francine’s work), eliciting a positive benefit, instead of zero. Francine, in turn, would take advantage of her monopolistic status as an exclusive rightsholder and accord by setting up an asking price of $8,000—the maximum Ella is willing to pay.¹¹⁰ In such a world, therefore, Ella confronts an asking price of $8,000, which means that Francine gets to pocket the entire bargaining surplus.

Now consider a world with fair use. Assume that, given the high transformativeness of Ella’s intended use, there is a 0.5 probability that a court will find the use fair. This means that if negotiations fail, Ella is better off using Francine’s work without authorization than with forgoing the use: using the work without permission provides her with a benefit of $8,000 while the expected costs are only $5,000 (0.5 × $10,000 = $5,000). Critically, the introduction of fair use changes the expected value of unauthorized use from -$2,000 to $3,000 ($8,000 – $5,000). The existence of the fair use doctrine, thus, dramatically empowers users vis-à-vis copyright owners. This implies that Ella’s threat of using Francine’s work without permission becomes credible: if Francine’s asking price would exceed $5,000, Ella would walk off the negotiation table and use the work without authorization. This is because any asking price that crosses this threshold makes her benefit from authorization less than $3,000, which should lead her to refuse to pay this amount and, instead, use the work without authorization. In such a world, therefore, Ella confronts an asking price of $5,000.¹¹¹ We summarize the results of our examples in Table 1, below.

Table 1. The Effect of Fair Use on Users’ Licensing Leverage
World	Ella’s Benefit from Using	Expected Cost of Unauthorized Use	Expected Value of Unauthorized Use	Ella’s Value of Not Using	Ella’s BATNA	Francine’s Asking Price
Without Fair Use	$8,000	$10,000	-$2,000	0	Non-existent	$8,000
With Fair Use	$8,000	$5,000	$3,000	0	Increased	$5,000

Assume next that Ella’s desired use is not only highly transformative, but a clear parody, thus representing a 0.8 probability of a fair use finding by a court. In that case, there is only a 0.2 probability that a court will find Ella liable for copyright infringement and award Francine damages. This, in turn, renders the expected value of unauthorized use for Ella even higher: $8,000 – 0.2 × $10,000 = $6,000. In light of this fact, Ella will never accept any asking price that surpasses $2,000. Realizing this, Francine would agree to license the work to Ella for $2,000. The results are summarized in Table 2.

Table 2. The Effect of Increased Fair Use on Users’ Licensing Leverage
World	Ella’s Benefit from Using	Expected Cost of Unauthorized Use	Expected Value of Unauthorized Use	Ella’s Value of Not Using	Ella’s BATNA	Francine’s Asking Price
With Increased Fair Use Probability	$8,000	$2,000	$6,000	0	Further increased	$2,000

The logic that underlies this finding is as follows. When facing a copyrighted work, any user entertains a trichotomous choice: (1) they may pay a licensing fee and use the work unhinderedly; (2) they may avoid using it altogether; or (3) they may refrain from licensing, use the work without acquiring authorization, and face the expected costs of a copyright infringement lawsuit. Users’ BATNA essentially depends on the feasibility of the third alternative. Herein lies the significance of fair use. Fair use bestows upon users the power to credibly threat to use copyrighted content without authorization and thereby leads rightsholders to adjust their asking price downwards. A world without fair use, by contrast, would allow any rightsholder to prevail in an infringement suit against any unauthorized user, which obviates the user’s third alternative. In such a world, users are left with options (1) and (2): they can either pay copyright owners the licensing fees they post or refrain from using the work altogether.

The fair use doctrine not only engenders alternative (3) but also makes it viable. As in the example discussed above, knowing that Ella reaps a positive expected value of $3,000 from unauthorized use, Francine realizes that for any asking price that exceeds $5,000 (leaving Ella with a benefit of less than $3,000), Ella has a credible threat of not taking the deal and using the work without permission.

B. The Model

Our theory can be generalized in a simple formal fashion. Let v denote a given user’s valuation of a given copyrighted work. Let A denote the owner’s asking price. A is the maximum licensing fee that the user finds attractive, i.e., the maximum price under which the user does not have a credible threat to engage in unauthorized use as a substitute for acquiring authorization. Denote by p the probability that a court will find that the user’s unauthorized use is protected by the fair use doctrine, and denote by D the damages the court is expected to award the rightsholder if fair use protection is not granted.

The owner’s asking price is given either by the user’s valuation of the work, v, or, if the user obtains a positive expected value from an unauthorized use (namely, if v – (1 –p)D > 0), by deducting this sum from v:

Note that in a world without fair use, p = 0. This means that if v < D, namely, as long as the court is expected to award damages that exceed the user’s valuation of the copyrighted work, unauthorized use is never a worthwhile alternative to licensing. The user thus possesses no leverage vis-à-vis the owner, and the latter will therefore require payment of v, the user’s valuation of the work, in order to grant authorization.

The fair use doctrine invariably increases the expected value of unauthorized use: raising the value of p from p = 0 to p > 0, which results in a higher p – (1 –p)D. Furthermore, for some threshold values of and , it increases v – (1 –p)Dto satisfy v – (1 –p)D > 0, and in such case, the existence of the fair use doctrine results in positive expected value from unauthorized use. From the user’s perspective, this means that v > (v – (1 –p)D), and thus A, their negotiated authorization price, decreases. Figure 1 exemplifies the relationship between A, the asking price, and p, the probability of fair use, for given values of v and D that uphold D > v.

Figure 1. The Effect of Fair Use on Copyright Pricing

This stylized model unfolds the technical bargaining mechanism that underlies our theory. Thus far, the accepted lore perceived fair use as offering users an effective bypass from licensing negotiations—but this statement can be confidently made only when fair use is granted with certainty (p = 1), which is hardly the regular case. The model emphasizes that even under the conventional setting where a fair use ruling is uncertain, the doctrine plays a critical role in the market for copyrighted content—the uncertainty translates into decreased asking price by monopolistic rightsholders, hence advantaging users’ bargaining standpoint.

We now turn to introduce evidence from licensing practices that supports our theory, and then we discuss the normative implications.

C. Illustrations

Our theoretical exposition of the hidden function of fair use finds support in real world cases. In this Section, we provide various examples that substantiate our theoretical predictions. These cases show that users are aware of the bargaining leverage they can get from the fair use doctrine and take advantage of its more favorable licensing terms.

Oracle and the Future of Licensing in Technologies

As noted, our theory likewise applies to providers of technological applications that generate a benefit to the public. A case in point can be found in Justice Thomas’s dissenting opinion in Google LLC v. Oracle America, Inc..¹¹² In support of his view that Google’s use of Oracle’s code was not fair, he noted that Google’s use created a reality where “device manufacturers no longer saw much reason to pay to embed the Java platform.”¹¹³ Justice Thomas proceeded to emphasize that:

[B]efore Google released Android, Amazon paid for a license to embed the Java platform in Kindle devices. But after Google released Android, Amazon used the cost-free availability of Android to negotiate a 97.5% discount on its license fee with Oracle. Evidence at trial similarly showed that right after Google released Android, Samsung’s contract with Oracle dropped from $40 million to about $1 million.¹¹⁴

The sharp decline in Oracle’s revenues from licensing its Java code provides a powerful illustration of the impact of fair use on market transactions. As the figures show, the bargaining leverage of Oracle in all of its licensing transactions critically depended on the licensees’ assessment of the strength of Google’s actual fair use claim and their own hypothetical fair use claims. For although Google was at the forefront of this legal battle, many other technology companies were similarly situated to Google insofar as their status as fair users.¹¹⁵ As the sentiment that Google would prevail gained purchase among technology companies, the revenues of Oracle sharply decreased, precisely as our theory predicts.

In analyzing the Supreme Court’s decision, commentators have noted that “[i]n the future, this decision may prompt more disrupters to use fair use as a shield in releasing new products or services that build off of older functional technologies, or otherwise influence negotiations as some potential licensees may find the value of certain functional code to be devalued by the Oracle.”¹¹⁶

Documentary Filmmakers

Copyright scholars tend to refer to the documentary filmmakers’ population as a primary beneficiary of fair use. Documentary filmmakers must incorporate prior works, many of which are subject to copyright protection. Hence, the fair use doctrine is critical to the operation of the industry. It may come as no surprise, therefore, that documentary filmmakers are cognizant of their ability to leverage on fair use in negotiations with copyright owners. A 2020 study by the Center for Media and Social Impact (“CMSI”) concerning the state of the documentary field¹¹⁷ indicates that 76% of all participating U.S. documentary directors and producers have utilized or leveraged the fair use doctrine in the making of their most recent film.¹¹⁸ This corresponds to other surveys, wherein 70% of the filmmakers have rated their understanding of fair use as “good or excellent,”¹¹⁹ while 73% of those with more than a decade of experience found fair use a “very useful” doctrine.¹²⁰ Likewise, an elaborate manual published by the Archive Valley company, which provides archival services for documentarists, introduces the underlying strands of fair use by explaining to authors that “if you meet the fair use guidelines . . . you can use footage for free,” but complements this straightforward statement by informing authors that “[e]ven if you know that you do want to license footage, knowing your rights about fair use can put you in a stronger negotiating position.”¹²¹

Finally, it seems that the leverage fair use bestows upon filmmakers changes the boundaries of the bilateral owner-user bargaining. Crucially, a strong fair use claim appears to pave authors’ way to efficiently negotiate with relevant third parties, too. In this regard, 95% of filmmakers have reported to persuade broadcasters regarding the applicability of fair use to their case when equipped with a lawyer’s letter.¹²² 99% responded identically when asked about negotiations with insurers.¹²³ This means that fair use not only leverages users when bargaining with rightsholders, but, as importantly, alleviates the entire chain of negotiations necessary for pursuing one’s artistic vision.

Parodies and Satires

The story of the musician “Weird Al” Yankovic, who is known for his humoristic commercial adaptations of popular music, provides yet another real-world example of the effect of ambiguity on broadening the population of users that can enjoy increased licensing leverage in light of fair use. Whether Yankovic’s spoof songs are in fact parodies¹²⁴ or satires¹²⁵ is a controversy that has never been, and probably never will be, settled, as Yankovic always acquires authorization from the owners of the rights to the original songs¹²⁶—be it because of industry norms¹²⁷ or in order to be on the “safe side” of copyright law. Interestingly, however, Yankovic’s practice has been described as “[l]icensing [w]ith [l]everage.”¹²⁸

Yankovic, whose parodic use adjoins satire, bargains for authorization under a fairly reasonable possibility of fair use protection. And Yankovic is not alone. Other artists who produce adaptations of copyrighted works have the same leverage that he does: they possess a colorable fair use claim.

It should be underscored that this broadening effect is not limited to musical adaptations. Rather, it applies to copyrighted subject matter categories, from videoclips to literary works. As importantly, the effect is not confined to adaptations that straddle the line separating parodies and satires. In the case of transformative uses, the effect is much broader. Consider, for example, the case of Gregg Gillis, better known as the popular D.J. Girl Talk, whose foremost artistic contribution has been described as “mak[ing] danceable musical collages out of short clips from other people’s songs.”¹²⁹ Gillis insists that his sampling work is protected by the fair use doctrine, and therefore, as opposed to Yankovic, he consistently refuses to solicit authorization from rightsholders to the original works.¹³⁰ Markedly, while many have accused Gillis of infringing upon their exclusive rights, a lawsuit has yet to be filed.¹³¹ In short, thus, it appears that both Gillis and his accusers believe that the musician enjoys a solid case, garnering a sufficiently high fair use probability that in turn nullifies the credibility of threats to sue. This translates into a reservation price of zero.

Warhol and the Future of Transformativeness

Another exemplification of the utilization of the transformativeness threshold into licensing leverage can be driven from Cariou v. Prince, in which the Second Circuit defined transformativeness as any alteration of the original’s “expression, meaning, or message,” finding that even appropriation art can be considered fair use.¹³² Under this expansive definition of fair use, any user who adapts a work can raise a fair use defense. Obviously, not every adaptation would be ruled fair. Similarly, not every appropriation artist will meet the same faith as Richard Prince. But every transformative user has bargaining leverage thanks to this broad definition.

As noted, however, the Supreme Court’s recent decision in Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith diverges from preceding rulings on transformativeness that advantaged users, and it seems to reallocate bargaining power between users and rightsholders.¹³³ Owing to Warhol, we might confront a shift toward a new era that introduces a new equilibrium in the market for copyrighted content.

Notwithstanding their absence from both popular discourse and academic scholarship on fair use, the Justices—both members of the majority and dissenters—were far from oblivious to this outcome. First, this understanding manifests in Justice Sotomayor’s articulation of the majority opinion. Justice Sotomayor first warns us that the broad interpretation of transformativeness would favor users with utterly disproportionate leverage over the original rightsholder. Upon recognizing that transformativeness “is a matter of degree,”¹³⁴ Justice Sotomayor maintains that:

[Holding for the Plaintiff] would potentially authorize a range of commercial copying of photographs, to be used for purposes that are substantially the same as those of the originals. As long as the user somehow portrays the subject of the photograph differently, he could make modest alterations to the original . . . and claim transformative use.¹³⁵

Justice Sotomayor’s words implicitly capture our theory, acknowledging the simple fact that the degree of transformativeness that suffices for the Court to bestow a fair use defense would shape market interactions between owners and users. Determining that changes reminiscent to those made by Andy Warhol are not significant enough to meet this threshold, particularly when the use is of commercial nature, Justice Sotomayor precludes users from leveraging a fair use claim and consequently—when viewed via an economic lens—submitting that their intended use does not deserve a reduced licensing fee.

More explicit are the statements made in Justice Kagan’s dissenting opinion. Justice Kagan distinctly highlights the pivotal role of the Supreme Court in allocating bargaining surplus between owners and users, noting that by refusing to side with Warhol’s deeds, the Court would frustrate users’ ability to license under favorable terms:

Still more troubling are the consequences of today’s ruling for other artists. If Warhol does not get credit for transformative copying, who will? And when artists less famous than Warhol cannot benefit from fair use, it will matter even more. . . . [A]s our precedents show, licensors sometimes place stringent limits on follow-on uses, especially to prevent kinds of expression they disapprove. And licensors may charge fees that prevent many or most artists from gaining access to original works.¹³⁶

The transactional role of the fair use doctrine and its effect on licensing negotiations has been properly identified and adequately considered. Despite that, the Court has upset the longstanding understanding of market actors, potentially requiring the market to form a new equilibrium that is less favorable to users. This leads us to conclude that the reallocation of bargaining power that we expect to witness in the near future—together with the inevitable increase of copyright pricing—is the Court’s fully conscious decision, rather than an unforeseen economic side effect.

III. Normative Implications

In this Part, we explain how the licensing leverage created by fair use ameliorates the efficiency losses and inequities associated with copyright protection. In particular, we demonstrate the doctrine’s ability to increase the number of voluntary transactions between copyright owners and users and to redistribute wealth from rightsholders to users.

A. Countervailing Allocative Inefficiencies

Being quintessential public goods, intellectual works cannot be supplied efficiently by the free market.¹³⁷ In the absence of a legal prohibition on copying, users would be able to copy expressive works with impunity. In such a world, the market price of copyrighted works would rapidly drop to zero, and so would the incentives to create and supply intellectual goods. Unable to recoup their initial investment in the creation of original content, authors may well decide to put their creative skills to rest. The need to maintain authors’ incentives to create warrants legal intervention that would grant them exclusive control over the distribution of their works.¹³⁸

Yet, the legal exclusivity copyright law bestows upon authors creates several costs. Chief among them is monopolistic pricing.¹³⁹ Two major effects are customarily attributed to monopolistic pricing: allocative inefficiency and distributional inequity.

The allocative inefficiency arising from exclusivity is well established.¹⁴⁰ Monopolistic pricing invariably creates a deadweight loss.¹⁴¹ This loss emanates from the fact that some consumers who were willing to pay the competitive price for the good or service would not be able to pay the higher, monopolistic price, and would simply forgo the good or service. As a result, certain welfare-enhancing transactions that would have occurred in a competitive market would not take place under monopolistic pricing.¹⁴² To illustrate, suppose that a good is offered by a monopolistic manufacturer for $20. Assume, however, that had the production process been perfectly competitive, the same good would have been offered for a price of $12. In this example, the deadweight loss is represented by the foregone transactions of the buyers who value the good at more than $12 but less than $20. As economist William McEachern points out, this phenomenon “is called the deadweight loss of monopoly because it is a loss to consumers but a gain to nobody.”¹⁴³

Perhaps even more worrisome than its adverse effect on allocative efficiency is the distributional inequities imposed by monopolistic pricing. Since the monopolistic price is higher than the competitive price, it works to transfer wealth from consumers to the monopolist. Monopolistic pricing, in other words, allows monopolists to capture consumer surplus.¹⁴⁴ To see this, let’s return to the above example, and consider a consumer who values the good at $22. Recall that the monopolist sets the price at $20, whereas in a competitive market, the price of the good would be $12. Under perfect competition the consumer would have paid $12 for the good, thus deriving a surplus of $10 ($22 – $12) from the transaction. Under monopolistic pricing the surplus of the consumer shrinks to $2 ($22 – $20), with the $8 lost to the consumer being transferred to the monopolist—all relative to a competitive market.

To understand how fair use remedies the allocative inefficiencies resulting from the exclusivity granted by copyright protection, consider a copyright owner who offers their work for a fixed license fee, which represents the profit-maximizing amount for them. As opposed to the general analysis unfolded in Part II, we now assume that, for whatever reason, price differentiation is impracticable from the rightsholder’s perspective.¹⁴⁵ Our theory submits that under such circumstances, the mere existence of fair use for certain users creates a positive externality for other users who have no fair use claim.¹⁴⁶ In this respect, the number of voluntary transactions may increase, decline, or remain unchanged, but what bears emphasis is that fair use avails the entire universe of users, including those users who would never persuade the court that their use is considered fair.

To see this counterintuitive outcome, suppose that the relevant users are five parodists and two satirists. For this example, assume that the parodists’ valuation of using the work is $10,000, while the satirists’ valuation is, say, $6,000.¹⁴⁷ For simplicity’s sake, assume that the damages in case of infringement are likewise $10,000. It is easy to see that if the monopolist has to determine a fixed, undifferentiated asking price, this price will be either $10,000 (thus authorizing only parodists) or $6,000 (thus authorizing all users, satirists included). The profit-maximizing price in that case is $10,000, since 5 × $10,000 > 7 × $6,000. Therefore, only parodists manage to obtain authorization from the owner, while satirists are precluded from using the work in light of their unwillingness to pay $10,000, neither as licensing fees nor as damage awards.

Things change dramatically, however, in a world with fair use. To begin, suppose any parodist holds 0.5 probability of obtaining a fair use ruling in their favor, whereas satirists have no chance of convincing the court that they are entitled to a fair use ruling. But note, while satirists will rarely enjoy a fair use ruling in their favor, they in fact manage to obtain user authorization in a world in which the fair use doctrine exists. In this world, the owner can never sell at $10,000: if they were to stick to this asking price, all parodists are incentivized to use the work without permission, as the expected value of using without authorization is $5,000 (0.5 probability multiplied by $10,000 in damages). Therefore, the two relevant pricing schemes are either $6,000 (authorizing only satirists) or $5,000 (authorizing all users). The

rightsholder is expected to set an asking price of $5,000 since 7 × $5,000 > 2 × $6,000.¹⁴⁸

If, on the other hand, any parodist holds 0.9 probability of obtaining fair use, this increased probability simply results in the rightsholder focusing exclusively on negotiations with satirists, setting an asking price of $6,000 as 7 × $1,000 < 2 × $6,000.

Interestingly enough, this example illustrates not our primary argument that fair use bolsters users’ bargaining power, but rather, that if rightsholders are subject to a fixed licensing fee, the doctrine may avail even users with zero probability of obtaining fair use protection. The hitherto undiscussed impact of fair use on copyright pricing allows them to enjoy copyrighted content they would not have in a world without fair use.

It should be noted that while parodists in the latter example will not be willing to pay the fixed price required by the rightsholder for authorization, they will manage to reach a favorable settlement if the copyright owner decides to sue them for infringement. Due to their enhanced probability of enjoying a fair use ruling, a settlement agreement is expected to feature extremely low payment on their behalf.¹⁴⁹

B. Redistributing Bargaining Surplus

In addition to improving allocative efficiency by increasing the number of users who enjoy owner authorization, fair use also has the effect of ameliorating the distributive distortions arising from copyright protection. As already noted, without fair use, rightsholders are positioned to capture the lion’s share of the users’ bargaining surplus by setting up prices that approximate users’ maximal valuations. In our previous example, even the five parodists who are willing to pay $10,000 for the right to use a work would have to fork over to the copyright owner—who sets an asking price of $10,000—all of their bargaining surplus.

This unfortunate outcome exemplifies monopolistic holdout. More specifically, monopolistic rightsholders manage to extract increased payments particularly from those users who consider the work extremely valuable. Pertinent examples include young technological entrepreneurs who are in need of particular copyrighted software in order to advance their startup initiative, or documentary filmmakers at the dawn of their career who are required to negotiate with a myriad of rightsholders in order to actualize their cinematic vision.

These users, and others, confront a significant barrier erected by copyright owners, who strategically hold out to extract exorbitant licensing fees. On our theory, the fair use doctrine counterbalances such strategic behavior. In addition to improving licensing terms for all users who possess a fair use claim (and, as exemplified earlier, even for those who do not), the doctrine helps reduce monopolistic holdouts in negotiations. This insight is not merely theoretical. It can be seen at work in the aftermath of the Supreme Court’s recent decision in Google LLC v. Oracle America, Inc..¹⁵⁰ The ruling has won commentators’ accolades for lowering entry barriers for smaller startup initiatives: consistent with our analysis, the ruling eroded the strategic incentive of powerful software owners to prevent smaller competitors, seeking interoperability with industry standards, from entering the market.¹⁵¹ Smaller developers can currently draw on the Google decision to acquire what is needed from a preexisting application programing interface, to provide consumers an innovative, transformative technological program.¹⁵²

The Warhol decision withheld the user-advantaging trend that bestows fair use protection even on mildly transformative users. Ample criticism has been directed against the Court since the decision was handed, primarily spotlighting the confinement of artistic freedom inevitably imposed by their ruling.¹⁵³ But one crucial aspect keeps eluding the commentators. As our analysis shows, the Court’s decision not only restricts artistic expression, but also diminishes the bargaining power of users in negotiations with rightsholders and thereby increases the price of copyrighted content. A narrow view of fair use makes one miss this aspect.

C. The Role of Damage Awards

To equip users with the full force of fair use, it is imperative to ensure that other copyright law doctrines may chill the effect this Article identifies. In particular, exaggerated damage awards may undesirably offset the economic virtues of the fair use doctrine. This is an especially grave concern in the copyright domain, as copyright owners typically sue for statutory damages. As Ben Depoorter noted, “[o]nce infringement has been established, a plaintiff may elect a statutory damage award. In doing so, 17 U.S.C. § 504(c) of the 1976 Copyright Act relieves the copyright holder from the burden of providing any evidence whatsoever of actual harm. Among developed Western democracies, the U.S. copyright statutory framework is exceptional.”¹⁵⁴ Indeed, under the current regime, statutory damages for watching an illegal livestream may amount to $150,000.¹⁵⁵ Oren Bracha and Talha Syed have similarly pointed out that “[t]he most troubling [aspects of statutory damages] are those cases in which massive supracompensatory damages are inflicted on ordinary individuals who are unable to spread the impact of such awards through limited liability or dispersed ownership.”¹⁵⁶

Apart from the usual pitfalls scholars associate with excessive copyright damages—namely, overdeterrence and suboptimal use of copyrighted works—we wish to emphasize its adverse impact on the licensing leverage that fair use provides to users. To see this, return to the main example on which we draw to convey our conception of fair use. Assume that rather than $10,000, Francine is expected to enjoy $50,000 in damages if she prevails in court. In this scenario, as in the case in a world without fair use, Ella is devoid of any feasible alternative to negotiations: the value of unauthorized use to her is 0.5 × (-$50,000) + $8,000 = -$17,000, which means that even in a world with fair use, she has no credible threat of using Francine’s work without authorization. Under these circumstances, the fair use doctrine does not bestow upon Ella a credible threat (and licensing leverage thereof) as long as her probability of prevailing in court falls short of 0.84.

By the same token, courts’ commitment to relatively modest awards is expected to significantly enhance the impact of fair use on users’ bargaining standpoint. Modest awards imply that even a user with a low probability of persuading the court of their entitlement to fair use will enjoy licensing leverage in negotiations with copyright owners. Taking this idea to the extreme, suppose that Francine is expected to be awarded $5,000 if she prevails in court. Recall that Ella values the use of Francine’s work at $8,000. In this case, Francine’s a-priori asking price is reduced irrespective of fair use: even without the doctrine, Ella will not be willing to pay any licensing fee that surpasses $5,000, since this would make unauthorized use more beneficial. Under the fair use doctrine, this price will drop even further.

Excessive damage awards imply that even users with a strong fair use claim might rationally refrain from an unauthorized use to avoid the insurmountable expected loss in future copyright litigation. We, therefore, call upon judges and legislators to determine damages in copyright cases when accounting for the possible erosion of users’ negotiation standpoint. In other words, we submit that by exhibiting restraint in the determination of damages, decisionmakers can bolster the hidden function of fair use; yet, by superfluously awarding damages to rightsholders, they might nullify this function in its entirety.

D. The Latent Virtue of Ambiguity

Our theory of fair use also puts ambiguity in a different light. A recurring criticism of the fair use jurisprudence concerns its uncertainty. The ambiguity of the fair use doctrine has been accused of overdeterring individuals from using copyrighted content. As James Gibson argued, even users with a strong fair use claim may rationally elect to license copyrighted content or refrain from using it altogether in order to avoid the inherent uncertainty that attends the fair use doctrine.¹⁵⁷ Losing on a fair use claim comes at a high price: a court may order the unlucky defendant to pay high statutory damages. Anticipating this, users may rationally choose not to assert their fair use claim in court, and instead, buy a license from the copyright owner. Gibson termed this dynamic “rights accretion.”¹⁵⁸ Furthermore, he described it as an ongoing process by which copyright holders broaden the scope of their protection.¹⁵⁹

Our theory points to a potentially countervailing effect. The uncertainty that characterizes the fair use doctrine may actually help users in negotiations with copyright owners. This is because the inherent ambiguity embedded in the fair use doctrine enhances the licensing leverage of users, relative to a bright-line-rule fair use doctrine. To demonstrate the argument, we first provide a brief description of the ambiguity inherent in extant fair use jurisprudence, and then we point out its advantages compared to a world in which the fair use ambiguity does not exist.

In its current form, the fair use doctrine is an open-ended standard, rather than a rule.¹⁶⁰ As the time-honored distinction suggests, rules are well-defined legal commands that offer a precise definition of a proscribed action or conduct. Conversely, standards are legal provisions that employ a more general, obscure description, thus subjecting the lawfulness of a particular behavior to judicial discretion ex post. While standards allow the legal system to function with greater flexibility, rules provide actors with a greater degree of clarity and certainty. As has been noted before:

[T]he ex post guidance provided by courts is often confined to the specifics of the case at hand and does little to clarify the realm of legitimate behavior for other actors. The unpredictability associated with standards affects not only wrongdoers, but also law-abiding citizens who wish to act in accordance with the law but cannot readily discern what acts are permissible.¹⁶¹

As noted above, fair use is a prototypical standard.¹⁶² Fair use jurisprudence is ridden with terms that may purport to be synonymous to fair use, but in fact provide little or no ex ante clarity with respect to their applicability to a given case. Parodic use, for example, was tautologically defined by the Supreme Court as anything that may be reasonably perceived as a parody.¹⁶³ The concept of transformative use has likewise eluded any consistent definition,¹⁶⁴ and the “public benefit” test, to which courts increasingly reference, is inherently uninformative.¹⁶⁵ All of this implies that even when the user possesses a strong claim, fair use protection is hardly guaranteed. Nor can it be completely ruled out even in the presence of users with relatively weak claims. In its current conception as a quintessential standard, therefore, the user and the owner confront a given probability that fair use will be recognized. But as noted earlier, this probability—created by the malleable, open-ended environment in which fair use determinations are being made—may render a user’s threat of resorting to unauthorized use credible, which in turn creates a favorable bargaining framework that allows them to secure authorization with reduced licensing fees.

To see this, compare the current fair use regime with an alternative hypothetical regime in which fair use is governed by rules, namely, all permitted uses are well-defined on an ex ante basis. This alternative design leaves no room for uncertainty as to the defense’s applicability: the probability that fair use will be recognized in a given case is either zero or one. While allowing clarity and complete reliance, we wish to stress the adverse effect that such dichotomy carries on parties’ bargaining. Begin with cases wherein fair use will be conferred upon users with complete certainty. It is evident that whenever a fair use ruling is guaranteed, users are unwilling to pay for licensing, and negotiations will never take place. On the other hand, if there is no chance for a user to enjoy a fair use judgment, they are essentially confronting a world without fair use, hence the doctrine fails to bestow licensing leverage upon them.

This comparison of a standard-based versus rule-based fair use illuminates the surprising virtue of the current design of the fair use doctrine. To be sure, each user individually, if asked, would say they prefer to obtain fair use with certainty, but the actual comparison that needs to be considered here is whether, behind a veil of ignorance, the right to use copyrighted content without permission should be allocated with certainty to a distinct, small class of users, or with uncertainty to a class of a broader scope. Users may reasonably subscribe to the latter. As noted earlier, a well-defined fair use doctrine not only allows fewer users to invoke the right in court, but more importantly, it also deprives all the rest from the licensing leverage they would have enjoyed under uncertain fair use, thereby subjecting them to the choice of paying the monopolistic asking price or avoid using.

Indeed, the current design of fair use facilitates a bargaining framework in which users enjoy discounts in asking prices on account of the possibility of a future fair use judgment in their favor. Conversely, a clearly and fully specified fair use doctrine will have no effect on owner-user bargaining: it advances fair use as a mere substitute to negotiations, rather than a complement. If parties foresee a fair use ruling with probability 1, bargaining will not take place; if a fair use judgment is not feasible, then parties essentially reside in a world without fair use, and the doctrine does not therefore impact their negotiated licensing terms. Either way, predicating the fair use doctrine as a system of well-defined bright line rules precludes its function in assisting users.

This insight corresponds to a handful of scholarly works that have stressed the merit of standards, compared to rules, at facilitating trade. The 1995 works of Jason Scott Johnston,¹⁶⁶ and Ian Ayres and Eric Talley,¹⁶⁷ have independently demonstrated that when it comes to bargaining over an entitlement, uncertainty is bliss. Both articles showed that standards have the effect of splitting entitlements between parties, which thereby enhance bargaining and negotiations. When entitlements are split, each party can buy or sell their share of the entitlement to the other.¹⁶⁸ The same is true of fair use. Fair use has the effect of splitting rights in expressive content between rightsholders and potential fair users.¹⁶⁹ Accordingly, in negotiations between them, the copyright owner sells to the potential fair user the right to use the work while the potential fair user sells away their privilege to assert fair use.

Interestingly, fair use’s ambiguity not only aids users with a strong fair use claim, but it also broadens the scope of user population that enjoys licensing leverage. To see this, consider the widely accepted distinction between parodies and satires, which prevails since Campbell.¹⁷⁰ Traditionally, it has been argued that this distinction favors parodists by exempting them from any authorization requirements, while disadvantaging satirists who enjoy no similar privilege.¹⁷¹ But this argument, too, overlooks the hidden benefit that users at large—parodists and satirists alike—can reap from fair use. To see this, it is necessary to take a step back and picture a world without fair use protection granted to parodies. In such a world, the author of the original work holds the upper hand by exercising complete control over the licensing process and can exclusively decide which subsequent works to authorize. The vestment of fair use privileges on parodies spills over to hybrid works that combine critical commentary on the original work as well as more general statements that amount to satire.¹⁷² This, in turn, means that satirists, too, reap some benefit from fair use. The effect of fair use on satirists is never zero. Because there is no clear distinction between parodies and satires, the fair use doctrine—on account of this very ambiguity—bestows bargaining leverage not only on parodists, but also on satirists.

Conclusion

In this Article, we uncovered a crucial, yet overlooked, function of the fair use doctrine: the empowerment of users in negotiations with copyright owners. The fair use doctrine endows users with a credible threat to leave the negotiation table and use the work without permission. This credible threat allows users to force copyright owners to lower their licensing fees, thus enhancing the total number of uses and increasing the share of the bargaining surplus kept by users. Our theory maintains that the principal effect of fair use is to facilitate market transactions involving copyrighted content, and not to bypass them as other theorists have suggested. Our analysis thus shows that the impact of fair use in the copyright domain is far greater than previously thought. Importantly, our theory of fair use is fully consistent with preexisting scholarly contributions; it complements them, not competes them. The addition of our theory to the existing literature on fair use demonstrates the full prowess of fair use in protecting users interests and needs in the copyright world.

98 S. Cal. L. Rev. 761

Download

Roy Baharad *

* Doctoral Candidate, University of Chicago Law School.

Gideon Parchomovsky†

Robert G. Fuller, Jr. Professor of Law, University of Pennsylvania Law School; Edward Silver Professor, Faculty of Law, The Hebrew University of Jerusalem. We thank Shiri Alon, Clark Asay, Ian Ayres, Shyamkrishna Balganesh, Oren Bar-Gill, Stefan Bechtold, Abraham Bell, Bari Britvan, Mala Chatterjee, Melissa Eckhause, Shira Ephron, Kristelia Garcia, Yehonatan Givati, Ehud Guttel, Thomas Haley, Lital Helman, Laura Heymann, Justin Hughes, Adi Libson, Peter Menell, David Nimmer, Shmuel Nitzan, Sarah Polcz, Matthew Sag, Peter Siegelman, Christopher Sprigman, Alex Stein, Alfred Yen, Eyal Zamir, and participants at the American Law and Economics Association Annual Conference and at the Seventh Copyright Scholarship Roundtable at Columbia Law School, for valuable comments and suggestions. We are especially indebted to Joseph Fishman for his enlightening remarks. For excellent research assistance, we thank Ron Bechar, Guy Cohen, Rachel Hujsa, Ruoyu Ji, Noam Kozlov, Ariel Melitz, Jeremy Rapaport-Stein, Ido Schlesinger, Yoav Stoler, Jude Taragin, Amir Tzur and Miriam Weinstock. Parchomovsky thanks the Israel Science Foundation for research support.