Jim Gibson is right that courts should be wary of letting the mere presence of licensing improperly foreclose the defense of fair use. As he says, a court in a copyright infringement case should not treat the existence of a market for licenses of the work as a factor weighing against the defendant’s claim of fair use until the court has examined “why [that] licensing market exists.” However, Gibson fails to distinguish the varying reasons licensing might be relevant to a fair use determination. As a result, the solution he proffers—attributing relevance to licensing only in markets free of excessive risk aversion—is both over- and under-inclusive. Ironically, his approach also overstates the importance of the existence of markets to the proper determination of fair use questions. In the following, I hope to clarify the role that market existence should play in fair use.

Gibson’s article is interesting and well-argued, and its thesis is straightforward. Since fair use results can’t be guaranteed in advance, and since many creative people can’t distribute their work without the cooperation of insurers and other stakeholders who refuse to tolerate even small liability risks, many movie-makers, musicians, and other artists purchase licenses to copy where a license isn’t legally necessary—where the fair use doctrine would have provided shelter from liability had the copyist chosen to rely on the doctrine. Yet courts weighing a defense of fair use sometimes ask whether a defendant’s use occurs within a copyright owner’s normal and expected market, and sometimes weigh the presence of existing markets against fair use.

The result may be a destructive feedback loop: potential defendants buy licenses out of an excess of caution; their unwillingness to test the fair-use waters then shrinks the scope of fair use for the next round of litigants; the shrinkage once embedded in precedent then causes even more excessive caution; and so on. As Gibson wittily reminds us, the shampoo bottle reads: “Lather, rinse, repeat.” If taken literally, one would never get out of the shower.

It is easy to overstate the danger. Just as most of us have the common sense to know that two lathers are enough, many cases contain articulate judicial warnings not to over-rely on licensing data. For example, the Second Circuit has warned that at least for transformative uses, “a copyright holder cannot prevent others from entering fair use markets merely ‘by developing or licensing a market . . . .’” But Gibson highlights a threat to fair use that is nevertheless real.

Gibson’s primary solution is to give courts tools with which to identify markets that are “rife with risk aversion,” and to argue that markets so identified should be irrelevant to the fair-use inquiry. Unfortunately, however, his article implies the converse: that a licensing practice should be relevant to deciding the fair use inquiry, so long as the market arises out of a justifiable (rather than an excessively risk-averse) expectation of liability. But, as I will show, sometimes the absence of risk aversion should not make a market’s existence relevant—and sometimes the presence of risk aversion does not make a market’s existence irrelevant.

Fair use raises the question of whether a particular use should be placed under a copyright owner’s control. The issue should embrace two inquiries: whether (1) in general or (2) in the particular instance, the liberty to engage in the contested copying should be a commodity only available through permission and purchase. The defendant has no fair use claim if market institutions are the proper locus for allocating the use, and if the particular market is working well. By contrast, the defendant’s claim of fair use should prevail (1) if the use is one that is of a general type not well allocated through markets, or (2) if the particular market is deeply flawed.

At least that is my view. It is acknowledged as such, if somewhat oddly restated, in Gibson’s article.

How does Gibson’s notion of excessive risk aversion fit into this overall structure? I turn first to those contexts where Gibson’s point has the most purchase. I then turn to those contexts where it has no purchase at all.

Where Risk Aversion May Be Relevant

On the general question of commodification, we know that there are some things (like the freedom to criticize) that simply shouldn’t be sold on any commercial market, whether or not such a market could feasibly arise. To paraphrase a very wise comment that I remember as Bob Gorman’s: even if all ideas could be bought and sold, that wouldn’t make you want to force the public to pay before using them. There are strong reasons for not treating everything as a commodity. For example, incursions of commercialism and bureaucracy can dampen creativity in many domains.

The courts that ask if a commercial market has arisen in a contested use do not adequately explain why they do so. Such courts persist in asking, “Is there injury to a market for licensing the contested use?” even though many commentators and even some judges recognize the inquiry as circular. Focusing on concerns regarding commodification helps to explain the turn to such market evidence and to indicate its proper limits.

The legal system continually engages in the difficult task of drawing dividing lines between “commodity markets,” on the one hand, and alternative realms such as “public domain” or “gift,” on the other. The lines are inevitably fuzzy. Courts faced with this issue often try to identify community norms concerning whether a particular practice should be commodified. This is part of what happens in fair use cases. For example, as Lloyd Weinreb points out, when the Supreme Court decided in the Betamax case that fair use permitted VCR owners to record broadcast TV shows for later personal viewing, the Court was probably informed by public feelings about privacy and proper behavior.

The word “norm” has many meanings. It can indicate, among other things, “typical” group behavior, or a group’s belief as to how its members “should” behave, or the requirements of an overarching “best” view of ethics and law. Usually, a judge is interested in discovering the overarching best view, but in the process she will often try to reach a kind of reflective equilibrium with community beliefs. In turn, typical behavior ordinarily serves as some evidence of the community’s beliefs.

So, should courts regard licensing behavior as some evidence that the community believes the use in question should be undertaken only with a license? Not necessarily. We all know instances in which community norms have been antithetical to morality. Even in contexts where the norms themselves are inoffensive, it is often unwise for law to adopt them. Further, the community is broader than the market participants; neither all potential users nor all people affected by a fair use ruling have expressed their preferences through the market.

Moreover, behavior that responds to law is very unreliable as a guide to attitude. When people obey the law, their behavior may merely show that they do not object strongly enough, or feel safe enough, to engage in disobedience. And here is where Gibson’s point enters: behavior becomes an increasingly less reliable indicator of attitude as fear increases. Therefore, if a licensing market arises out of excessive risk aversion, an exceptionally broad chasm opens up between behavior and beliefs: licensing behavior becomes a particularly unreliable guide to what even market participants think the rules should be.

Thus, in cases where the courts turn to the community for information on the general question of whether a use shouldn’t happen without a license, the presence of Gibsonian risk aversion is relevant. Gibson’s article rightly plants a red flag in such contexts, marking them as situations in which courts should be especially careful not to rely upon custom as evidence of community beliefs. Here Gibson’s tests for identifying risk-aversive markets are useful.

Where Risk Aversion May Be Irrelevant

If the defendant’s type of use is generally suitable for commodification, copyright will still not deserve enforcement unless the defendant faces a market that functions well enough to direct resources to socially desirable ends. There are many market defects that can make Adam Smith’s “invisible hand” falter, and thus validate fair use in individual cases. One such defect is the presence of transaction costs so high that licensing is impractical.

On the factual question of impracticality, the existence of a market shows that for at least some players, transaction cost barriers are surmountable, perhaps because new institutions (such as collective rights societies) have evolved. Such evidence is necessarily incomplete—what constitutes a “low” transaction cost to a player like Google may be an immense barrier to ordinary users—but the actual licensing pattern does tell us something about the cost of finding bargaining partners, dickering over fees, arranging for monitoring, and the like.

Transaction costs excuse a defendant’s market bypass only if such costs are high. If a transaction cost barrier has been lowered—such as through the creation of a collective rights organization—it doesn't matter if this happened because licensees were excessively risk-averse or for some other reason. So on the factual issue of whether the defendant could practicably have licensed, excessive risk aversion should be irrelevant.

A Concluding Thought

Some of the force of Gibson’s article may spring from an underlying worry that nags at many copyright scholars: that fair use might be misunderstood as justified by the presence of transaction costs, and that as transaction costs shrink, markets will become more feasible and fair use will disappear. So Gibson’s article functions as one of many salutary efforts showing that the bare existence of markets should not foreclose fair use. While I applaud such articles, it’s important to remember that however refined the courts’ use of existing markets becomes, it should never constitute more than a small portion of the fair use inquiry, for two reasons. First, the absence of transaction costs merely removes one of many possible bases for fair use. I believe that a careful rereading of my early work on transaction-cost theory will make that clear. Second, any reference to “customary markets” is fraught with danger. Custom is unlikely to provide probative evidence about community beliefs when people are merely responding to what they fear the law will do. Any licensing done out of fear of liability—whether reasonable or excessively risk-averse—is a poor guide to what a community believes.

Wendy J. Gordon is Professor of Law and Paul J. Liacos Scholar in Law at Boston University School of Law, and can be reached at This email address is being protected from spam bots, you need Javascript enabled to view it . She thanks Jessica Litman and Lloyd Weinreb for their helpful comments, and Daniel Bahls for his excellent research assistance.

Preferred Citation: Wendy J. Gordon, The ‘Why’ of Markets: Fair Use and Circularity, 116 Yale L.J. Pocket Part 371 (2007), http://yalelawjournal.org/2007/4/25/gordon.html.

© 2007 by Wendy J. Gordon.