Note on Coase and Music

Henry H. Perritt, Jr.

Why should the law interfere with the exchange of music?  Music is art.  Those who generate music and those who enjoy it should be empowered in a free society to decide for themselves how they will generate it and enjoy it.  Coase teaches that rational actors will make good decisions about allocating their resources in the absence of law.  If a singer is willing to sing a song for a dollar and someone wants to listen to his song and is willing to pay a $1.50 for it, they will reach an agreement under which the listener will pay something between $1.00 and $1.50 to the singer, who will then perform for the listener. Of course the singer and the listener have to find each other before they can bargain.  The cost of finding each other is what is called a “transaction cost,” which the Coase Theorem explicitly does not take into account. 

It also may be that the singer is willing to sing for a dollar but is concerned that someone will steal his song, neither giving him credit for it nor preserving the market for future paying listeners for his art.  In other words, the listener may be able to sing also and, having heard the song, may appropriate it and sing it for others who are just as happy to pay the original listener as to pay the original singer. 

It is not impossible for the original two parties to take this concern into account in their bargaining.  The original singer may say to the original listener, “I will pay you $.25 for your promise not to tell anyone the lyrics of my song or to sing it to anyone else.  The listener, more interested in listening than in singing, agrees.  The two parties offset the original payment for the singing against the payment for not appropriating the song and end up with a bargain under which the singer sings and the listener only listens and pays somewhere between $.75 and $1.25. 

Of course the original listener may break his promise and pass out the lyrics and sing the song in the public park.  Detecting such a breach of his promise and punishing him for it represent additional transaction costs. 

In addition, music is—or at least recorded music is--what is known by economists as a “public good.”  Public goods are distinguished from private goods by two characteristics:  non-excludability, and non-rivalness.  Non-excludability means that when the singer sings the song, anyone nearby can hear it.  The singer may be unable to exclude those who haven’t paid unless he builds a soundproof wall.  Non-rivalness[1] means that the singer still has the song even after he has performed it and the original listener may sing it for someone else and still keep it in his head.  Recorded music possesses the non-rival and non-excludable characteristics strongly.  With the benefit of digital technology it is impossible for the one who produces an original recording to prevent anyone who possesses that recording from making a perfect copy, copy protection aside.  It is thus impossible for the original recorder to exclude those who have not paid him for it from enjoying the recording.  In addition, anyone who possesses the original recording has little disincentive not to duplicate it because he can keep it even as he gives the perfect copy to his friend.. 

Performed music is not quite so much a public good as recorded music.  Performed music is excludable.  One can perform—and performances typically occur—in an enclosed space from which those who have not paid can be excluded.  In that context, performed music is excludable and thus lacks one of the two defining characteristics of a public good.  Also, attending a performance is more than simply hearing the music; it is an experience.  One cannot internalize the performance and thus “keep” it while passing it on to someone else in exactly the same way one can duplicate a recording and pass along the duplicate. So performed music also exhibits less of the second defining characteristic: non rivalness.

Nevertheless for some forms both public good characteristics exist to some degree. Public goods characteristics are externalities—things a simple model of private contract have difficulty making an enforceable part of the deal.

Because of these transaction costs, performers will make fewer deals than they would in a transaction-cost-free world and will insist on higher prices because they fear that listeners will steal their songs.  Listeners will make fewer deals and insist on paying lower prices than they would in a transaction-cost-free world because of search costs among other things.[2] 

There is one more transaction cost that needs to be considered:  free riding.  The parable in the preceding paragraphs already touched on free riding.  If an original listener internalizes a song and performs it for others, the original listener has in obtained a free ride on whatever investment the singer made in creating the music and lyrics.  The listener who then becomes a singer gets a free ride on the creative effort, although he cannot get a free ride on the original singer’s voice quality or vocal performance ability.  

In a world in which music is only performed live and not recorded, the transaction costs are tolerable. Designing a legal regime to manage the transaction costs and promote optimal resource allocation with respect to creation and consumption of music would be a relatively low priority.  But the possibility of recording music and passing along the recordings dramatically changes the economic equation.  As this section has already noted, recorded music possesses the two defining characteristics of a public good much more strongly than does performed music. 

Not only that, the original listener—the one who breached his contract with the original singer-- is not the only one who can duplicate the recording.  Anyone who has it can duplicate it.  That opens up the original singer to the possibility of free riding not only by his original listener, with whom he can, after all, make a contract prohibiting whatever he wants to prohibit, but also by everyone else in the world who gets possession of the sound waves representing his music in circumstances that permit them to record it.  And, of course, the original singer is not in a position to enter into contracts with all of those unidentifiable third parties. 

Also, the economic harm from free riding is greater because, if the original singer produces recordings as well as performances, he has to invest not only in his voice and put forth creative effort, he has to has the capital goods and skill to make acceptable recordings.  Depending on the available technology, it may be possible for free riders to duplicate the recording with much less expensive capital equipment and far less skill than is necessary to make the original recording in the first place.[3]  The third party now gets a free ride not only on the voice training and creative effort of the original singer but also on that part of the original singer’s capital investment not necessary for mere duplication of the recording.[4] 

If the original singer is commercially motivated and invested in voice training, put forth the creative effort, and bought the capital equipment and skilled labor necessary to produce recorded music in order to sell it to as many consumers as possible, his business model is threatened by the possibility of free riding.  A third-party free rider, not having to make the same investment as the original singer, can make money by selling the original singer’s work at a price less than the original singer must charge to recover a reasonable return on his investment. 

The result, once everyone understands all this, will be the underproduction of recorded music by those to seek at least enough revenue to recover their costs.  At the limit, no one will record music; they will only perform it live.  The quantity of music that would be performed and consumed through recordings will be subtracted from the goods available to society. 

Both technology and law can contribute to avoiding this artistic catastrophe.  Technology can make it more difficult to duplicate recorded music except through equipment or knowledge possessed by the original singer, thus increasing the cost and thereby reducing the incidence of free riding. 

The law also can reduce various kinds of transaction costs, including free riding.  At the most basic level, the law can make the contracts between the original singer and the original listener enforceable, thereby reducing transaction costs associated with enforcement of the original deal.  More ambitiously, the law can impose a duty on anyone in possession of a song not to make it available to others without the permission of the original singer and without performing whatever terms the original singer may insist upon for subsequent availability of his artistic work.  For example, the law might give the original singer a statutory or common law right in the music and lyrics and impose a duty on anyone else not to reproduce, distribute, publicly perform, publicly display or prepare derivative works of the original song.[5]  Such legal regimes reduce the risk of free riding by imposing potential costs on free riders—litigation costs and the prospect of fines or damages for violating their duties.  Even if not everyone violating the duties gets caught, there is some non-zero probability of getting caught and punished. Rational persons tempted to engage in free riding will calculate some expected value of the costs imposed by law.  The law also could go further, and seek to reduce other transaction costs, such as search costs, although it might do this indirectly by creating incentives (reducing the transaction costs) associated with private voluntary efforts to make it easier for singers and listeners to find each other and make deals.

The Coase paradigm not only assumes zero transaction costs, it ignores questions of distributive justice. A society where distributive justice prevails is not only one in which resources are allocated efficiently; it is one in which each member enjoys resources she deserves, in a moral sense. Distributive-justice considerations are sources of argument as much as they are sources of principles for law. Does each member of society deserve, in a moral sense, to be able to enjoy music? Does each person with musical talent deserve to be able to adapt music that already exists? Does each musician deserve to have his art recognized as his creation and to be able to make a living off of making music? Affirmative answers to each of these questions are plausible, but affirmative answers to all of the questions offer little guidance as to how copyright law should strike balances among competing claims.


[1] Non-rivalness also is known as non-depletability.

[2] See Litman at 28 (recording industry and its “bookkeeping tricks” barely manage to subsidize its searching for an identifying the “musicians whose work will prove to be worth listening to”).

[3] The original singer needs recording and mixing equipment and CD mastering equipment (if he delivers recorded music on CDs). The free rider only needs a CD “burner,” which comes as standard equipment on most PCs.

[4] William A. Landes offers the same explanation of the economics of free riding, albeit outside the music context. See William M. Landes, Copyright, Borrowed Images, and Appropriation Art: An Economic Approach, 9 Geo Mason L. Rev. 1, 4-6 (2000).

[5] 17 U.S.C. § 105 does exactly this.