82 CLMLR 1600	Page 1
(Cite as: 82 Colum. L. Rev. 1600)

Columbia Law Review

December, 1982

*1600 FAIR USE AS MARKET FAILURE: A STRUCTURAL AND ECONOMIC ANALYSIS OF THE

BETAMAX CASE AND ITS PREDECESSORS

Wendy J. Gordon [FNa1]

In the recent and much publicized Universal City Studios, Inc. v. Sony Corp. of America (Betamax) case, [FN1] the Court of Appeals for the Ninth Circuit held that persons who make videotapes of copyrighted television programs in the privacy of their homes should be considered to be copyright infringers. [FN2] Basic to the court's reasoning was a misunderstanding of the "fair use" doctrine. Called "the most troublesome doctrine in the whole law of copyright," [FN3] "fair use" renders noninfringing certain uses of copyrighted material that might technically violate the statute, but which do not violate the statute's basic purposes.

The Ninth Circuit took the position that "fair use" could only protect users who were productive second authors, and not persons who were making ordinary or "intrinsic" use like the home videotapers. [FN4] In doing so, the circuit court rejected explicitly the economically oriented approach to fair use adopted by the Court of Claims in Williams & Wilkins Co. v. United States. [FN5]

In its opinion in Williams & Wilkins, which was affirmed by an equally divided Supreme Court in 1975, [FN6] the Court of Claims held that massive photocopying efforts by libraries could in some circumstances constitute fair use. [FN7] The Supreme Court has recently granted certiorari in the Betamax case, [FN8] *1601 and may take this occasion to choose between the Court of Claims's approach and the more restrictive doctrinal views of the Ninth Circuit.

This Article suggests that the Court of Claims opinion in Williams & Wilkins, while flawed in many of its aspects, [FN9] did not err in seeking a base for fair use in structural and economic considerations. Specifically, it will be argued that the fair use doctrine, though sometimes called an "equitable rule of reason" for which no definition is possible, [FN10] has at bottom three straight-forward concerns. Where (1) defendant could not appropriately purchase the desired use through the market; [FN11] (2) transferring control over the use to defendant would serve the public interest; [FN12] and (3) the copyright owner's incentives would not be substantially impaired by allowing the user to proceed, [FN13] courts have in the past considered, and should in the future consider, defendant's use "fair." While other approaches to fair use may legitimately be advanced, much of fair use depends on the resolution of these concerns.

This Article will suggest that, contrary to the Ninth Circuit's view, the "productive" status of a user is at best merely a secondary indicator that these concerns may be satisfied. [FN14] On a more fundamental level, the Article will illustrate how the courts and Congress have employed fair use to permit uncompensated transfers that are socially desirable but not capable of effectuation through the market. [FN15] The market approach will provide a guide both to ascertaining where the public interest might lie in a given case, and to identifying those occasions on which a court may appropriately substitute its evaluation of the public interest for its usual refusal to second-guess the copyright owner. Through analysis of the Williams & Wilkins [FN16] and Betamax [FN17] cases, the Article will also indicate how a market approach can serve as a means for applying fair use to newly emerging uses of copyrighted works made possible *1602 by developing technologies. Overall, by unifying the various traditional fair use factors into one economic model, the Article aims to serve as an aid to predicting fair use results and as a guide to future development of the doctrine.

I. COPYRIGHTS AND MARKETS

A. Origins of the American Copyright System and the Doctrine of Fair Use

The Constitution gives Congress power to grant to authors and inventors exclusive rights over their works in order to "promote the Progress of Science and useful Arts." [FN18] Beginning in 1790, Congress passed a series of copyright statutes; [FN19] the general goal of these statutes has been to establish an incentive for authors to create, by providing them an avenue for obtaining remuneration. [FN20] The ultimate goal is not author remuneration, however, but the advancement and dissemination of culture and knowledge. [FN21]

In most cases, the incentive and dissemination purposes of copyright will not conflict. Authors will be encouraged to produce and distribute new works because the copyright laws give authors the means of being paid for their efforts. Madison foresaw exactly this result when he advocated a congressional copyright and patent power in The Federalist Papers: "The public good," he wrote, "fully coincides . . . with the claims of individuals." [FN22] Nevertheless, tensions are possible, and over time various copyright doctrines have evolved to guard against the possibility that the author's right of control over his works could defeat rather than serve the public interest in dissemination. "Fair use" is recognized as one such doctrine; it seeks to accommodate the author's need for remuneration and control while recognizing that in specific instances the author's rights must give way before a social need for access and use. [FN23] In other words, the fair use doctrine allows an individual, in certain circumstances, to make use of at least a part of an author's work without obtaining that author's consent or recompensing the author for that use.

Historically, it has been no simple matter to decide the appropriate boundaries of an author's control. [FN24] As noted above, the fair use doctrine in *1603 particular has been called "the most troublesome in the whole law of copyright," [FN25] and has traced a quicksilver course of judicial development. Congress, in codifying the fair use exception in the Copyright Act of 1976, refused to formulate a specific test for determining when a particular use is fair use. Instead Congress gave statutory recognition to a list of factors that courts have looked to in making their fair use determinations. [FN26] Four factors are enumerated in the statute:

(1) the purpose and character of the use, including whether such use is of a commercial nature or is for non-profit 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 and value of the copyrighted work. [FN27]

Congress recognized that even for the courts that had developed the criteria, the factors were "in no case definitive or determinative" but rather "provided some guage [sic] for balancing the equities;" [FN28] the statutory list is nonexhaustive. [FN29] In addition, Congress encouraged further evolution of the fair use doctrine in the courts. [FN30]

*1604 It is not clear how much weight should be given to any one of the four factors, what additional factors should be considered, or whether any one of the factors is a sine qua non for a finding of fair use. [FN31] Though the factors do implicitly direct courts to identify both the social need for the use and the possible impact on the author's economic expectations, the ambiguity of the fair use doctrine and its statutory formulation obscure the underlying issues and make consistency and predictability difficult to achieve. Some underlying principles, more specific and thus more informative than the identification of basic incentive/dissemination tensions, must be divined beneath the factors to permit their coherent application.

An economic analysis of the copyright system will serve to illuminate these underlying principles. Since the copyright system creates private property in creative works so that the market can simultaneously provide economic incentives for authors and disseminate authored works, it may be useful to begin--if not to end--the fair use inquiry by using economic principles of market failure to identify the situations in which these two goals are likely to conflict. The same economic considerations that guide the copyright system as a whole can also be used to suggest modes of resolving the conflicts, providing a method of applying fair use both to achieve desirable dissemination and to avoid the erosion of incentives. This economic view of fair use can provide a *1605 coherent approach to application of the traditional factors and their emerging companions.

B. An Overview of the Market Model

Markets are systems for consensual exchange of owned goods. Private ownership is, of course, only one of many ways in which persons can be encouraged to make productive use of resources, [FN32] and our legal, economic and social systems respond with other modes of resource control when the markets fail to generate economically desirable outcomes [FN33] or when using the market process would threaten other social goals. [FN34] This Article will show that fair use should be interpreted as a mode of judicial response to market failure in the copyright context, and that the presence or absence of the indicia of market failure provides a previously missing rationale for predicting the outcome of fair use cases.

1. The Market System. Economists in the tradition of Adam Smith defend our society's primary dependence on markets by arguing that individual transactions in the marketplace serve both social needs and the needs of the individual persons participating. [FN35] Among other contentions, they suggest that the monetary value a person places on a resource will reflect the value that the person's use of the resource will bring to society, so that voluntary transfers between individuals will create a socially desirable pattern of resource allocation. [FN36]

*1606 To understand this approach, consider the example of movie producers who wish to purchase the motion picture rights to a best-selling novel. Smithian economists would posit that each producer's ability to raise funds from investors depends on the amount of revenue that his or her moviemaking is expected to generate. Among producers of varying levels of skill, the producer best able to use the book to satisfy consumer tastes will be in a position to raise the most funds and thus to offer the highest bid. Similarly, the author or other owner of copyright in the novel will sell the rights only if the revenues he could anticipate by exploiting the work himself would be less than the purchaser's bid. Control over the resource will therefore gravitate through consensual transfers to the person in whose hands the resource can best be used to satisfy consumer desires. [FN37] In this way value, defined as "human satisfaction as measured by aggregate consumer willingness to pay for goods and services," [FN38] will be maximized.

*1607 Economists agree that several "conditions of perfect competition" must be satisfied if individual market transactions are to result in the maximization of value. When these conditions are absent, reliance on consensual bargains to achieve socially desirable results may be inappropriate. [FN39] One such condition is that all costs and benefits must be "internal" to the transactions that generate them, in the sense that the costs or benefits must be borne by persons with decision-making power in a given transaction and not by persons external to it. [FN40] For example, if a resource user is in a field where much of the social benefit produced by his activities does not translate into compensation to him, then he may generate "external benefits" not reflected in his income. In such a case, his willingness to pay for the resource might understate his ability to use it in a way that serves social needs. [FN41]

Perfect competition also requires perfect knowledge; [FN42] for example, consumers must know the qualities and characteristics of all available products, [FN43] as well as the prices and locations of the various sellers. Consider the impact of imperfect knowledge on the movie producer example advanced earlier. If reviewers give consumers inadequate information about the characteristics of movies produced, the money spent on theater tickets may not reflect the choices that would in fact best satisfy consumer tastes. [FN44] As a result, a skillful producer might be unable to raise capital in a way that would accurately *1608 reflect his ability to use resources in satisfaction of consumer tastes, [FN45] so that his ability to pay for the rights to a novel would understate his ability to use the novel to serve public needs. In a perfect market, his failure to outbid another producer would have been a reliable indicator that the movie rights would be most valuable in the other producer's hands. In a market where information is incomplete or unreliable, however, a failure to outbid the other producer may have no such significance.

Finally, then, perfect competition depends on the absence of transaction costs. [FN46] If all desirable transfers are to occur, it must be costless to obtain knowledge, [FN47] costless to locate all persons affected by a transaction, costless to dicker with them over prices and terms, and costless to maintain an enforcement mechanism to ensure that the bargain is adhered to. [FN48] This condition of perfect competition, like the others, is never fully present; some transaction costs are inevitable. [FN49] When the transaction costs outweigh the net benefits that the parties would otherwise anticipate from a transfer, then the presence of the transaction costs may block an otherwise desirable shift in resource use. [FN50]

The legal system acts in diverse ways to increase the probability that these and other conditions for perfect competition will be present. For example, to prevent consumer ignorance, truth-in-lending and labeling disclosure laws may be enacted, [FN51] and restrictions on truthful advertising may be struck down. [FN52] Similarly, antitrust laws prevent any one buyer or seller from gaining monopoly control that would distort competition. [FN53] When the market does not work perfectly, "a decision will often have to be made on whether market transactions or collective fiat is most likely to bring us closer to the . . . result *1609 the 'perfect' market would reach." [FN54] One possible response of the law might be the creation of governmental agencies to make resource allocation decisions by regulation. Another might be to set up incentives that encourage the achievement of efficient results even in the absence of markets.

The latter provides the conventional economic explanation for the standard of reasonable care used in negligence law. [FN55] In a market transaction, each party is forced to take account of the other's demands through bargaining. In unintentional accidents, a market bargain between the participants is obviously impossible, yet it is desirable for them to take into account the effect of their actions on each other. Negligence law gives a potential defendant the incentive to consider the other person's possible injury. If the costs of the precautions neglected by the defendant are less than the discounted accident costs the precautions would have prevented, [FN56] the verdict will be for the plaintiff. [FN57] If, on the other hand, the court views the defendant's pre-accident behavior as cost-justified, the plaintiff's injuries will go unrecompensed. [FN58]

A defendant who deliberately bypasses the market is not likely to find the court willing to act as a resource-allocating mechanism; the market guarantees compensation to the owner, and usually provides a less expensive and more reliable way to determine whether a use is value-maximizing than does the deliberation of a judicial stranger to the transaction. [FN59] A court generally should engage in balancing costs and benefits only if market failure has left it the only institution able to do so. [FN60] This is one reason why, for many intentional torts, the economic desirability of defendant's action is irrelevant to *1610 determining liability; in such cases the defendant could have identified, and thus bargained with, the plaintiff or class of plaintiffs involved. [FN61]

2. Markets in Copyright. Copyright and patent law create ownership rights in intellectual property, with the primary goal of generating monetary incentives for the production of creative works, thereby "promot[ing] the Progress of Science and useful Arts." [FN62] If the creators of intellectual productions were given no rights to control the use made of their works, they might receive few revenues and thus would lack an appropriate level of incentive to create. Fewer resources would be devoted to intellectual productions than their social merit would warrant. [FN63]

Economists ordinarily characterize intellectual property law as an effort to cure a form of market failure stemming from the presence of "public goods" characteristics. [FN64] A public good is often described as having two *1611 defining traits. First, it is virtually inexhaustible once produced, in the sense that supplying additional access to new users would not deplete the supply available to others. Second, and more important for the instant purposes, persons who have not paid for access cannot readily be prevented from using a public good. [FN65] Because it is difficult or expensive to prevent "free riders" from using such goods, public goods usually will be under- produced if left to the private market. [FN66] A familiar example of a public good is national defense. [FN67] Since it is not possible to use a radar early-warning network in a way that discriminates between one person who has paid for defense and his neighbor who has not, a less than optimal amount of national defense will be produced if its purchase is left to the usual consensual market mechanisms of voluntary purchase. Some sort of compulsory payment, such as taxation, and central decision-making may be necessary to eliminate free riders and obtain the socially desirable amount of defense. [FN68]

Books and inventions exhibit certain public goods characteristics. [FN69] Once the literary work or the discovery embodied in the invention is made available to the public, the sequence of words or the discovery might be used by countless consumers without exhausting the supply. Any number of persons can simultaneously use the newly invented process or reprint the literature without physically depriving others of its use. Physical control, therefore, does not offer its usual potential as a mode of inexpensive enforcement for excluding free riders. [FN70]

Though taxation and centralized purchasing might provide a satisfactory solution for some public goods problems, such an approach is inappropriate *1612 for much intellectual property. A democratic society demands decentralized and diverse creation in the intellectual sphere; freedom from state control is essential lest freedom of expression be curtailed by fear of governmental reprisal. [FN71] Thus, for works of expression, the public goods problem is addressed by another method. Statutes create special property rights for authors; they can sell the physical copies of their works and at the same time retain legal control over the reproduction and certain other uses of the work embodied in those copies. [FN72] In other words by the law provides a means for excluding nonpurchasers. [FN73] Copyright law therefore allows a market for intellectual property to function.

In addition, the copyright law makes it easy to proceed through consensual market transfer. The requirement that a notice of copyright be placed on all publicly distributed copies facilitates identification of those works that are not in the public domain and cannot be used without purchase of a copyright license, and also facilitates identification of the works' owners. [FN74] The registration of copyrights in the Copyright Office [FN75] makes owners relatively easy to locate, and gives further information on the validity and duration of the copyright claimed. To discourage departure from the market system, Congress has made certain copyright violations criminal, [FN76] and has established statutory damages and other devices [FN77] to lighten a plaintiff's enforcement task. The copyright statute thus facilitates the functioning of the consensual *1613 market [FN78] in four ways: it creates property rights, lowers transaction costs, provides valuable information, and contains mechanisms for enforcement.

When the market functions, no finding that the defendant acted "unreasonably" or against the public interest is requisite for judging culpability for the tort of copyright infringement. Like the intentional tortfeasors discussed in the preceding section, [FN79] copiers of creative works ordinarily can identify and bargain with copyright owners. If copies are made without permission, the court will not use a "reasonableness" test to second-guess whether the copyist's production was in the public interest. In the ordinary copyright case, the court assumes that the defendant could have, and therefore should have, proceeded through the market. [FN80]

Copyright markets will not, however, always function adequately. Though the copyright law has provided a means for excluding nonpurchasers and thus has attempted to cure the public goods problem, and though it has provided mechanisms to facilitate consensual transfers, at times bargaining may be exceedingly expensive or it may be impractical to obtain enforcement against nonpurchasers, or other market flaws might preclude achievement of desirable consensual exchanges. [FN81] In those cases, the market cannot be relied on to mediate public interests in dissemination and private interests in remuneration. [FN82] In extreme instances, Congress may correct for market distortions by imposing a regulatory solution such as a compulsory licensing scheme. [FN83] Thus, to avoid threatened monopolistic control over the manufacture of piano rolls and other mechanical recordings, Congress provided that any person who wished could make and sell recordings of copyrighted music, so long as he paid to the copyright owner an amount determined under the statute. [FN84] But the broad brush of this regulatory solution is too sweeping for most cases.

*1614 Fair use is one label courts use when they approve a user's departure from the market. A useful starting place for analysis of when fair use is appropriate is therefore an identification of when flaws in the market might make reliance on the judiciary's own analysis of social benefit appropriate. [FN85] By making such an identification, a measure of coherence can be brought to the doctrine of fair use. As will be seen, there are certain "conditions of perfect competition" [FN86]--or assumptions about how a proper transactional setting should look--whose failure is particularly likely to trigger in the courts an unwillingness to rely on the owner's market right to achieve dissemination. [FN87]

II. THE THREE PART TEST FOR DETERMINING FAIR USE

Fair use should be awarded to the defendant in a copyright infringement action when (1) market failure is present; (2) transfer of the use to defendant is socially desirable; and (3) an award of fair use would not cause substantial injury to the incentives of the plaintiff copyright owner. The first element of this test ensures that market bypass will not be approved without good cause. The second element of the test ensures that the transfer of a license to use from the copyright holder to the unauthorized user effects a net gain in social value. [FN88] The third element ensures that the grant of fair use will not undermine the incentive-creating purpose of the copyright law. The test will now be explored in detail.

A. The First Element of the Test: Market Failure

Because courts in the copyright area ordinarily assume that reliance on the market will serve social purposes, an economic judgment that transfer of a use to defendant will bring a net social benefit should not alone be sufficient *1615 to negate the tort of infringement. If the work is socially more valuable in the buyer's hands, then the economic model has suggested that he will be able to raise sufficient funds to purchase permission from the owner. In other words, to propose that fair use be imposed whenever the "social value . . . outweighs any detriment to the artist," [FN89] would be to propose depriving copyright owners of their right in the property precisely when they encounter those users who could afford to pay for it. Though a transfer to such user might be socially desirable, there is no need to compel it through fair use. Such transfer will in theory occur voluntarily, through purchase. Further, if the parties could arrive at mutually agreeable terms for a transfer, such an agreement would be a more reliable indicator of the transfer's value-maximizing quality than would a court's distant judgment. [FN90]

An economic justification for depriving a copyright owner of his market entitlement exists only when the possibility of consensual bargain has broken down in some way. Only where the desired transfer of resource use is unlikely to take place spontaneously, or where special circumstances such as market flaws impair the market's ordinary ability to serve as a measure of how resources should be allocated, is there an economic need for allowing nonconsensual transfer. Thus, one of the necessary preconditions for premising fair use on economic grounds is that market failure must be present. Those types of market failure that are revealed in the fair use cases and statute will be discussed in a later section of this Article. [FN91]

B. The Second Element of the Test: Balancing Injury and Benefit

If market failure is present, the court should determine if the use is more valuable in the defendant's hands or in the hands of the copyright owner. [FN92] One way of accomplishing that goal is to simulate the market inquiry. If, when the "market failure" were cured, the price that the owner would demand is lower than the price that the user would offer, a transfer to the user will increase social value. [FN93]

A court may have difficulty in determining what price would have been offered or demanded. For example, as will be shown, fair use is often found where defendant's use of the work is noncommercial and yields "external *1616 benefits," that is, benefits to society that go uncompensated. [FN94] In the presence of such market failure, the price that the defendant user would offer for use of the work will often understate the real social value of his use. The courts in fair use cases frequently make intuitive estimates of social value. [FN95]

The court's ability to estimate social value may have the assistance of some objective measures. For example, plaintiff's minimum price may depend on the loss he would experience from ceding control of the use to defendant, and evidence of out-of-pocket losses or of interrupted profits may be available. In addition, defendant's profits may reflect profits the plaintiff could have earned but for the unauthorized use. [FN96] Where the injury suffered by plaintiff is small, it is likely that the market price that he would have demanded is small, and that a transfer to defendant would have taken place if there had been no "failed market" conditions. Not surprisingly, courts are much more likely to find fair use where the injury to plaintiffs is small. [FN97]

When a court weighs plaintiff's loss against the benefit of defendant's use, it is making a comparison similar to that made by the participants in market transactions. [FN98] It may be said, then, that fair use implies the consent of the copyright owner by looking to whether the owner would have consented under ideal market conditions. Dean Prosser notes that the locution of "implied consent" is used in the law of intentional torts even where "consent does *1617 not exist, and there is no act which indicates it." [FN99] As an example of such implied consent, Dean Prosser refers to the plaintiff who undergoes an emergency operation while unconscious. There the surgeon who operates to save the plaintiff's life is not guilty of a battery because the plaintiff's consent to the operation "is 'implied' under the circumstances." [FN100] To justify imposing consent on the patient, Dean Prosser suggests the following explanation: " T he defendant is privileged because he is reasonably entitled to assume that, if the patient were competent and understood the situation, he would consent . . . ." [FN101] Thus, Dean Prosser has chosen a hypothetical condition relevant to the ultimate issue of consent in the context of battery: what would the patient consent to "if he were competent" rather than unconscious?

Are there any such hypothetical conditions relevant to the ultimate issues in copyright law? This Article suggests that the perfect market assumptions provide a guide to such conditions, because it is only in a functioning market that consensual transfers will lead to socially desirable resource use. [FN102] Are there any exigencies--such as the imminence of an unconscious patient's death--that will justify resort to hypothetical consent instead of relying solely on a plaintiff's actual consent? This Article suggests that market failure creates such an exigency because, when market failure is present, it is impossible or undesirable to make dissemination of creative works depend solely on actual consent. Thus, where transfers will not occur because of market failure, courts should ask what the copyright owner would have consented to if he and the user had bargained in a functioning market situation. It might be argued that the copyright owner should receive compensation for the use in the same way that the patient receives the benefit of the operation. However, implied consent here serves to indicate that the social welfare is served, while in the example of the medical patient it served to indicate that individual welfare is *1618 served; further, implied consent is only the second stage of a three-part inquiry. The third stage of the inquiry has a direct concern with the economic welfare of the copyright owner. [FN103]

C. The Third Element of the Test: The Substantial Injury Hurdle

The third and final part of the test is designed to maintain the appropriate balance between the incentive and dissemination interests discussed earlier. Fair use should be denied whenever a substantial injury appears that will impair incentives.

1. Complete Market Failure. When no incentive purpose would be served by giving plaintiff protection, and where no disincentive would be created by allowing defendant free use, logic suggests that the courts should then allow fair use. [FN104] Assume, for example, that prohibitively high transaction costs obtain in a particular area of use, so that copyright owners and potential users find that the costs of locating and bargaining with each other exceed whatever profit they might expect to gain from the transaction. Under such circumstances, no transactions will occur. Therefore, enforcement of market entitlements would not benefit the copyright owner, and would certainly harm the potential copyright user who is denied access, as well as those who might benefit from the use. Given the presence of a complete market failure here, judicial refusal to enforce the owner's right of control may be the only way to allow use of the work. [FN105] And since a refusal to enforce in these circumstances would not deprive the owner of any revenues he would otherwise receive, there is no injury to incentives that might militate against a grant of fair use.

2. Intermediate Cases of Market Failure. In cases of complete market failure, fair use appears to be justified upon satisfaction of the first two parts of the fair use test, i.e., identification of market failure and determination that there is a net social benefit from allowing defendant's use. There may also be intermediate cases of market failure, however, where the market cannot be relied upon to generate all desirable exchanges, but where some such transactions would be possible. In such instances, giving fair use to a class of users would enable some persons who would otherwise purchase the use to obtain that use for free. Thus fair use could cause some injury to relevant incentives because, for some users, fair use would substitute for purchase. [FN106] The first two parts of the test would no longer serve to accommodate the competing interests.

In instances of intermediate market failure, both enforcement (a finding of infringement) and nonenforcement (a finding of fair use) have dangers. *1619 The danger from enforcement is that desirable transfers may be prevented. The danger from giving fair use is that incentives may be undermined. To resolve the conflict, the defendant seeking fair use treatment should be required to surmount an additional hurdle: fair use should be denied if it would leave the plaintiff copyright owner facing substantial injury to his incentives. [FN107]

The third branch of the test diverges from the negligence model summarized earlier. [FN108] In negligence, the decision as to the wrongfulness of defendent's conduct usually depends on a comparison of costs and benefits; if the cost of defendant's precautions would have been greater than the decrease in accident costs produced, then the defendant is not negligent for having declined to take the precautions. [FN109] The absolute level of damages caused by the accident under that kind of analysis is irrelevant. [FN110] Under the third part of this fair use test, however, fair use will be denied if a substantial level of injury, judged on an absolute scale, would be generated by defendant's use, even if the comparison of injury and benefit showed that defendant's use would have created a net social benefit.

The substantial injury hurdle serves several functions. First, it preserves the incentive system at the core of copyright. Second, it reflects a recognition that judgments the courts make about whether a defendant's use is value-maximizing are rough approximations. [FN111] The substantial injury hurdle provides some additional protection for copyright owners against the possibility of a bad estimation. [FN112] Third, awarding copyright owners a veto whenever their injury is substantial gives some guarantee that the fair use system will not put them at an intolerable disadvantage. Even if authors are viewed solely as instruments of social good, [FN113] their demoralization as individuals will decrease the production of valuable works. [FN114] The substantial injury hurdle can also help courts avoid the danger of making otherwise curable market failures permanent through the grant of fair use. To explain this last function more fully, it is necessary to look at the nature of the relevant injury.

*1620 Injury is relevant to both the second and third parts of the fair use test, but different types of injury are relevant to each part. On the question of balancing plaintiff's harm against defendant's benefit, courts should look to the injury caused to the copyright owner by the specific use made by defendant. The narrow inquiry reveals whether transferring the use to the defendent gives rise to a net social benefit. In order to prevent substantial injury to incentives, however, the court should also inquire into the extent of the losses likely to follow in the market as a whole from a grant of fair use, [FN115] both from this defendant and from other similarly situated persons. When a determination is made that a certain kind or practice of use is "fair," the practices of persons other than the specific defendant may be affected. As the Senate Report to the 1976 Copyright Act states, "Isolated instances of minor infringements, when multiplied many times, become in the aggregate a major inroad on copyright that must be prevented." [FN116] Thus, the inquiry into substantial injury should include consideration of cumulative harm. [FN117]

This inquiry should also include harm that has not yet occurred but is likely to occur. [FN118] Both the loss of revenues anticipated under the market structure prevailing at the time of suit, and the loss of those revenues that would be generated under whatever market structure would follow upon a grant of infringement, should be relevant. [FN119] For example, transaction costs to obtain permission to use certain materials might be prohibitively high at one point in time, yet in some circumstances a clearinghouse system might be set up to simplify the process of purchasing permission and thus allow a market to function. To award fair use without regard to the possibility of imminent change in the market structure might be to make permanent an otherwise curable market failure and thus potentially to insulate a new and valuable use from the stimulus of consumer demand.

Whether a market failure is curable, and whether such a cure would follow upon a finding of infringement and generate substantial revenues, are difficult factual questions. They must, however, be faced; the courts should limit their grants of fair use to those occasions in which the market cannot be *1621 relied upon to allow socially beneficial uses to occur. This point is particularly important for new technologies, such as photocopying [FN120] and videotaping. [FN121] When a new use for copyrighted works becomes available to the public, market mechanisms may take time to develop. At early stages of use, the transaction costs that would be involved for a user to purchase permission to use, or for the copyright owner to seek enforcement against nonpaying users, might well exceed whatever gain the parties might otherwise expect from the transaction. A custom therefore may develop under which users proceed without permission. [FN122]

As the quantity of use grows, the copyright owners may wish to set up collection and enforcement mechanisms, including such market devices as clearinghouses. [FN123] In order to persuade users to proceed through the device, however, the copyright owners might well need a judicial declaration that the uncompensated use, previously minor and left unfettered, constituted an infringement of copyright. Such owners would be able to make a showing of substantial injury only if a court were willing to consider whether substantial revenues would follow after a finding of infringement.

The risk from granting fair use without reference to such probable injury should be obvious. New technologies will make certain copyrighted works more valuable, as, for example, the invention of cinema increased the value of those books suitable for adaptation to the screen. If copyright protection is denied because of an otherwise curable market failure, then the additional revenues that would have flowed from the new technological use will not appear. If the authors' revenues fail to reflect the additional value that new technology gives to such works, then insufficient resources may be drawn into their creation. To argue that copyright owners need receive no compensation for additional uses of their works [FN124] would overlook the possibility that such *1622 compensation may change the patterns of production in a desirable way. Stimulation of such response is, after all, a basic goal of copyright law.

D. Alternatives to Fair Use

Although this Article has advocated the three-part test as a useful approach for resolving the conflict between the incentive and dissemination goals of copyright law, other market-oriented approaches may also be useful in achieving this goal. For example, when the cause of the market failure lies in high transaction costs, economics does not dictate a choice solely between awarding fair use or giving the copyright owner complete control through injunction and damages. Instead, if the law permits, the court might choose to refuse an injunction yet grant the copyright owner monetary relief. [FN125] A monetary remedy could largely obviate any threatened injury.

*1623 It may be costly to society to give an author injunctive control over a work in a market failure situation since that might completely prohibit a defendant's use. [FN126] Enforcement limited to a damage remedy, award of profits, or reasonable royalty [FN127] might provide a mechanism for transferring revenues from user to author while simultaneously allowing utilization of the work. Such a court-dictated transfer might carry far less economic and social cost than would rules that force a choice between forbidding all involuntary transfers or leaving involuntary transferors without a remedy. By imposing a "price" for the use, the court award might itself "cure" market failure. [FN128]

It may be asked, however, whether courts should abandon the traditional all- or-nothing choice between enforcement and fair use bequeathed to them by fair use precedent. There are limitations on judicial expertise; whether the courts themselves are the appropriate institutions to "cure" market failure by inventing methods of compulsory transfer or by setting copyright prices is a very real question. [FN129] Further, an objectively determined "price" is always less satisfactory to a property owner than a price of his own choosing and is usually a less reliable indicator of value for society's purposes as well. [FN130] Were the courts to feel free to substitute one for the other, an accelerating erosion of the market system could be triggered. From the point of view of copyright owners (and thus from the point of view of society's need to maintain overall incentives), a system that permitted certain limited uncompensated takings to occur, as long as they did not cause substantial injury, might be preferable to a system in which compensation was guaranteed but only after the fact.

In addition, copyright law has traditionally made compulsory licenses (the equivalent of "reasonable royalty" determinations) a matter for legislative action. [FN131] Compulsory licenses have been hard-fought political issues, [FN132] *1624 for whose resolution Congress would appear to provide the most appropriate forum. Thus, although judicial regulation of copyright markets is not unknown, [FN133] and although Congress is often slow to act in this area, [FN134] it may be unwise to advocate judicial adoption of an alternative to fair use that asks courts essentially to restructure markets or to set prices for the use of copyrighted material.

E. Burden of Proof

Much debate today centers on whether the defendant or plaintiff should have the burden of proving injury. [FN135] The copyright statute does not state where the burden of proving injury or any other fair use factor should be placed, and Congress apparently preferred to leave this issue to the courts. [FN136] *1625 A market approach can provide some insight into how the burden of proof should be allocated.

As discussed earlier, [FN137] the copyright law appears to be premised on the assumption that functioning markets exist for copyright works. When such markets exist, injury will normally follow upon infringement and, even if the precise dimensions of that injury are hard to prove, it makes sense that lack of explicit proof of injury will be excused. [FN138] The congressional decision to provide plaintiffs the option of minimum statutory damages [FN139] reflects this sentiment and suggests that the plaintiff should have no initial burden of disproving fair use. [FN140] Instead, the defendant should have the initial burden of proving that market failure exists and that it will continue to exist even if a judgment of liability were to be tendered. In addition, if his use serves no social purpose, then he could not satisfy the second part of the fair use test and further consideration of his fair use claim would be unwarranted. Therefore, he also should be required to prove that his use serves some social purpose, [FN141] and to demonstrate the nature and extent of the public interest that is served. [FN142]

*1626 When severe market failure is present, injury to the copyright owner may not follow from infringement. [FN143] If the proof of severe market failure offered by the defendant leads the court to doubt that injury will be present, it is legitimate for the court to demand evidence from the copyright owner that such injury would indeed follow. The burden of going forward with proof of injury should then shift to plaintiff. The harm he has suffered or anticipates is an area that should be peculiarly within plaintiff's knowledge. [FN144]

Fair use has been variously regarded as a "defense," [FN145] as a "privilege," [FN146] or as a use that is noninfringing. [FN147] Given the ambiguous nature of fair use, it might be argued that the court could appropriately demand some proof of injury from plaintiff even in the first instance. [FN148] This Article does not go so far; it suggests that, on a case-by-case basis, a court should ask plaintiff to come forward with evidence of probable harm when the court has drawn an inference, from defendant's proof of market failure or from other sources, that no injury is likely. Moreover, since it is defendant who seeks to show that the conditions of his use are so unlike those contemplated by the copyright scheme that fair use is warranted, the ultimate burden of persuasion on this issue should rest with him.

F. Summary

An analysis of the economic functions served by copyright and of the internal dynamics of the copyright statute has suggested a three-part test for fair use: first, does a reason to mistrust the market appear?; second, is the transfer to defendant value-maximizing, as determined by weighing plaintiff's injury against defendant's social contribution?; third, if both the first and second conditions are satisfied, would a grant of fair use cause substantial injury? If it would not, and if the prior conditions are satisfied, then fair use should be awarded. Defendant user should prove the existence of market failure and the social merit of his use. If defendant meets these burdens and *1627 raises a significant doubt as to whether harm is likely, then plaintiff should come forward with a showing of injury.

If the market failure in a given case is likely to be cured (and substantial revenues generated) by the parties' own actions following a finding of infringement, fair use should be denied. However, if the only possible "cure" is through a complex, court-imposed structure of monetary remedies and limited injunctive relief, both fair use tradition and the inherent limitations of the judicial process suggest that the court should hesitate before attempting such a task. It may be appropriate to leave the responsibility for restructuring the parties' economic relationships to the parties or to Congress. Analytically, however, it should be recognized that the market considerations underlying the three-part test could trigger a damage remedy or market restructuring, rather than fair use, in a court that is inclined to engage in such intervention.

The three-part test proposed here, while not previously articulated in so many words, reveals itself in many of the decisions reached by the courts in the copyright area, and, as will be seen, serves to unify the various factors that courts and the statute have made relevant to fair use determinations. The remainder of this Article will explain how this test finds expression in the cases and statute, will analyze the Williams & Wilkins opinion, which gave the fullest expression to the test, and will examine how the Betamax court could have proceeded had it accepted these criteria.

III. EVIDENCE OF A MARKET APPROACH IN EXISTING AUTHORITIES

This Article has suggested that a court will ordinarily not grant a defendant fair use treatment unless the facts of the case give reason to mistrust the market. [FN149] This section will review cases and authorities in which recurring patterns of such mistrust appear to have guided courts in the fair use area. The discussion will show that courts and, in following their lead, Congress, have at times grappled with a market approach, that such an approach is reflected in the traditional fair use factors, and that, were the courts to embrace the market approach more fully, a more effective and consistent use of the traditional factors would result.

A. Inquiries Into Market Breakdown Under the Traditional Fair Use Approach

1. Market Barriers: New Technologies and Other Applications. As previously discussed, the impossibility or difficulty of achieving a market bargain is a factor that may justify a grant of fair use. The relevance of market barriers to fair use is implicitly reflected in the legislative history of section 107 of the Copyright Act. The Senate Report to the new copyright act states that "[a] key, though not necessarily determinative, factor in fair use is whether or not the work is available to the potential user," [FN150] so that the out-of-print status *1628 of a copyrighted work may help to justify fair use. [FN151] This is consistent with a market approach, since markets cannot form where goods are unavailable. [FN152] Similarly, the Guidelines for Educational Fair Use in the House Report single out for fair use treatment instances of classroom photocopying in which bargains are particularly unlikely to occur because the teacher's use is spontaneous, individual and unsystematic. [FN153] One of the prerequisites for making multiple copies under the Guidelines is that " t he inspiration and decision to use the work and the moment of its use for maximum teaching effectiveness are so close in time that it would be unreasonable to expect a timely reply to a request for permission." [FN154]

A particular type of market barrier is transaction costs. As long as the cost of reaching and enforcing bargains is lower than anticipated benefits from the bargains, markets will form. If transaction costs exceed anticipated benefits, however, no transactions will occur. [FN155] Thus, the confluence of two variables is likely to produce a market barrier: high transaction costs and low anticipated profits. New technologies are likely to present both high transaction costs and, where uses by individual scholars or in individual homes are at issue, correspondingly low anticipated profits. This may explain why the "personal," "individual" nature of copying has been held relevant to fair use, [FN156] and why "home use" may be relevant to the reach of copyright law. [FN157] Consider, for example, the impact of the photocopy machine or the tape recorder. Each makes it possible for individuals to make use of copyrighted works in new and potentially valuable ways. From the point of view of the individual user, the anticipated "profit" is likely to be small, so his use will be *1629 easily discouraged by transaction costs. Also, the technology's novelty may mean that the participants have no established market channels to rely on, so that the purchase of permission is likely to be cumbersome and expensive. High transaction costs and low per-transaction profits will converge. From the point of view of the copyright owner, the costs of enforcement against a diffuse group of individuals might outweigh anticipated receipts. A custom of use without payment will easily arise in such contexts unless the transaction costs of seeking permission or of enforcement are in some way reduced.

In such situations, transaction costs are likely to prevent at least some value-maximizing transfers from occurring [FN158] if the copyright is enforced. At the extreme, enforcing the owner's rights might eliminate the use, and thus bring no income to the owner and deprive society of the benefit of the technology. For this reason, new technologies may become the subject of fair use treatment.

New technologies do not always involve market barriers, however, and thus do not always merit fair use consideration. In a case where a county educational program was videotaping educational television programs, the systematic and centralized nature of the copying and the various market alternatives that were present made license and purchase agreements quite possible. [FN159] The court there denied fair use treatment at least in part because of this possibility. [FN160] Thus it is the absence of a market and the prospect that such a market may form that is important, not the technological nature of the use.

The role of transaction costs also explains those cases that rely on the copyright owner's apparent or likely consent in granting fair use. [FN161] Where a transfer is likely to be in the mutual interest of both owner and user, the courts *1630 appear unwilling to deter such transfers by imposing the costs of obtaining actual consent. [FN162]

2. Externalities, Nonmonetizable Interest, and Noncommercial Activities. An analysis of the limitations of markets can also illuminate the special status that certain uses, such as scholarship, have in fair use tradition. The costs and benefits of the parties contracting for the uses often differ from the social costs and benefits at stake, so that transactions leading to an increase in social benefit may not occur. [FN163] Thus, for example, a critic of the Warren Commission's investigation of the Kennedy assassination might write a "serious, thoughtful and impressive" [FN164] book that will further public interest more than the revenues of his book alone would indicate. One might say that publication of his book gives an "external benefit" to persons who might gain knowledge from the public debate sparked by the book without having purchased the book itself. Similarly, teaching and scholarship may yield significant "external benefits"; all of society benefits from having an educated citizenry and from advances in knowledge, yet teacher salaries and revenues from scholarly articles are arguably smaller than such benefit would warrant. When a defendant's works yield such "external benefits," the market cannot be relied upon as a mechanism for facilitating socially desirable transactions. [FN165]

*1631 In cases of externalities, then, the potential user may wish to produce socially meritorious new works by using some of the copyright owner's material, yet be unable to purchase permission because the market structure prevents him from being able to capitalize on the benefits to be realized. Though such inability would not itself justify fair use, it may signal to the court that it should investigate whether the social costs of relying on the market are unacceptably high. It is therefore not surprising that section 107 of the Copyright Act, which addresses fair use, lists several uses that potentially exhibit positive externalities, such as "teaching," "scholarship," and "research," among the uses for which fair use may be given. [FN166]

Section 107 also directs the courts to consider "the purpose and character of the use, including whether the use is of a commercial nature or is for nonprofit educational purposes." [FN167] Where the defendant does not seek to earn profits, it may be argued that his willingness and ability to pay for the copyrighted resources he uses will not provide an accurate measure of the public interest served by his use. Distinctions between profit and nonprofit entities or commercial and noncommercial uses must, however, be employed with great caution. Henry Hansmann has suggested that nonprofit organizations can be fully participating members of the market process. [FN168] Conversely, even commercial uses can face market failure. [FN169]

Distrust of the market may also be triggered when defendant's activities involve social values that are not easily monetized. When defendant's use contributes something of importance to public knowledge, political debate, [FN170] or human health, [FN171] it may be difficult to state the social worth of that contribution as a dollar figure. If the defendant's interest impinges on a first amendment interest, relying upon the market may become particularly inappropriate; constitutional values are rarely well paid in the marketplace and, while the citizenry would no doubt be willing to pay to avoid losing such values, it is awkward at best to try to put a "price" on them. [FN172] Not surprisingly, *1632 it has been suggested that fair use be granted when first amendment issues are involved. [FN173]

While in all of the cases described in this section--those involving external benefits, noncommercial uses, and nonmonetizable values--reason exists to distrust the market, it may be particularly difficult to determine whether the breakdown is substantial enough to frustrate the purposes for enforcing copyrights. What one deals with here may be not only traditional market failure, in the sense that conditions of perfect competition have failed, but also a court's perception that the criterion of economic "value" is itself flawed. [FN174] This concern is not illegitimate, but it should not be extended to make the copyright law an instrument of income redistribution. The courts should thus take care that they do not tax copyright owners to subsidize impecunious but meritorious users under the guise of maximizing value. [FN175] Only when the public interest to be served is great, and the damage to the owner small, does the need for this caution diminish.

3. Anti-Dissemination Motives. Section 107 places first among the purposes for which fair use is appropriate "criticism" and "comment," uses that a copyright owner might be reluctant to license. [FN176] Similarly, the treatment of burlesques and satires, which can be considered types of commentary, has been a volatile subject of fair use law. These uses share a type of market failure that helps to explain their fair use treatment and that is particularly important in a field where advancement of knowledge is the ultimate goal. The case law has tended to grant fair use treatment where copyright owners seemed to be using their property right not for economic gain but to control the flow of information.

The usual economic assumption is that the owner of a resource will either exploit that resource himself, or will sell it to someone else who will. The *1633 owner of a copyright, however, may not be willing to exploit all of the possible derivative works over which his copyright would ordinarily give him control. Even if money were offered, the owner of a play is unlikely to license a hostile review or a parody [FN177]