Compulsory Licensing of Music, Generally
A compulsory license, to state the obvious, is a way to obtain the rights to copyrighted content without negotiating directly with the rights holder; instead, the user gets to take advantage of a rate set by a statutory scheme.
The big place this happens today is music licensing. As Victor puts it in the first line of his beautifully written piece, "[t]he United States copyright system generally favors free-market licensing negotiations. ... [B]ut when music distributors like iTunes, Spotify, or Pandora want to obtain licenses in order to disseminate a song, they can sometimes take advantage of compulsory licenses outlined in the Copyright Act." (1). The rate can be fixed at a set price (e.g. a 2% fixed rate) or set by a rate-setting entity. In this case, that's the Copyright Royalty Board (CRB), which is the successor to the Copyright Arbitration Review Panel (CARP) and before that the Copyright Royalty Tribunal (CRT).
Origins of the Compulsory License for Music
Congress originally created the compulsory licensing option as a response to disruptions caused by new technologies that made it possible to easily copy and mechanically replay music. (20-25). Under the 1909 Act, in an earlier version of today's Section 115, Congress established the so-called "mechanical license" for musical compositions, which used a fixed rate of 2 cents per copy. (Nice!) Victor asserts that the mechanical license was a "condition" given to users and the public in exchange for expansion of copyright law to cover copying on mechanical devices capable of replaying music. (20-22).
Here is a direct quote, which I use both to avoid having to re-write, and to illustrate Victor's writing:
This was new terrain for U.S. copyright law, which had never been applied to such “mechanical” reproductions before. Some record and piano roll producers began paying precautionary royalties to music publishers, but the law was unsettled as to whether the unlicensed creation of a record or piano roll infringed the musical composition copyright owner’s exclusive rights. The Supreme Court finally addressed the question in White-Smith Music Pub. Co. v. Apollo Co, and concluded that piano rolls did not implicate copyright because they are not “copies within the meaning of the copyright act.”
The Court, however, invited Congress to extend copyright to mechanical reproductions, if it so chose. Congress accepted this invitation in the 1909 Copyright Act, which explicitly granted copyright owners the right to exclude use “of instruments serving to reproduce mechanically the musical work.” The 1909 Act, however, applied a “condition” to this extension of copyright protection: the compulsory licensing scheme that is now known as the mechanical license or Section 115 license, and set the compulsory rate at a fixed 2-cent royalty per copy.(22).
Not surprisingly, the initial 2-cent per copy royalty rate ultimately proved unworkable. Or at any rate, people in industry complained that it was unfair given the realities of the economy, such as inflation. (24). In response, when it overhauled the Copyright Act in 1976, Congress decided "to replace the fixed rate with one administered by a rate-setting entity." (25).
The Conundrum of Rate Setting
This is where the conundrum of rate setting arises. If a rate setting entity is suddenly responsible for determining how much users, and thereby consumers, must pay to obtain access to music, instead of negotiating directly with rights owners, what policy should the rate-setting entity use to set rates?
We might think the rate should simply be set at what it would have been, had a direct negotiation taken place. After all, the compulsory license is not a statement of non-infringement, like fair use. It is, rather, an alternative to direct negotiation.
Scholars like Robert Merges have long viewed transaction costs as the main hurdle to ensuring fair and efficient access to IP-protected works, and have frequently applied the classic Calabresi/Melamed model to IP licensing. Viewing compulsory licensing as a mechanism in which government's role is limited to facilitating efficient licensing in the face of high transaction costs, in particular the multiplicity of dispersed music copyright owners, would therefore seem logical. (That said, Merges does not see a compulsory license as the best way; he prefers collective rights organizations that operate in the private sector without government involved at all).
Wendy Gordon has gone further, arguing that even fair use, which is a statement of non-infringement, should be seen as a response to high transaction costs. Victor tactfully notes that "Gordon’s early work in this field is frequently contested, including by Gordon herself" (13). It seems circular to argue that whether an act of copying constitutes copyright infringement should itself be determined by whether the user could have gotten a license.
An easy example illustrates how ludicrous this is. Jacob could set up a method for licensing the rights to write a blog post commenting on his article. If I did not pay the license, I would be infringing simply by writing this post and quoting liberally from his article. This can't be right. Surely, transaction costs can't be the whole story. Fair use is about more than facilitating a market transaction. It is also a statement that the value of the copyist's work outweighs the harm to the rights holder, if any. The value of the former is too great to let copyright law stifle it, etcetera, etcetera.
The "Incentives/Access Tradeoff" Model vs. The "Transaction Costs" Model
In a creative move, Victor builds on these debates surrounding the proper role of fair use, and argues that a similar tension exists when it comes to deciding how to set rates in compulsory licensing.
He argues that in the early years, legislators saw compulsory licensing's role as similar to that of "fair use." In Victor's words, compulsory license was configured as a
mechanism for modulating the tension between copyright’s two competing priorities: financially incentivizing creators to produce works and providing the public with access to those works[.](2). Throughout the article, Victor calls this model for compulsory licensing the “incentives/access tradeoff" model.
Precisely what rate the incentives/access tradeoff would set is itself highly debatable. On one account, if I wanted to obtain a compulsory license to make a cover of a popular song, I would receive a rate that seeks to enhance the ability of average people to pay for the right to do so. On an account that is more attuned to distributive justice, the rate would be set so that the poorest in society can obtain access.
But this is an aside and not central Victor's project. Instead, Victor is interested in comparing the “incentives/access tradeoff" model with another model: the "transaction costs" model. As alluded to above, the transaction costs model views the compulsory license as a way to achieve the idealized "market price" for a work in the face of high transaction costs, such as a multiplicity of dispersed rights owners.
The transaction cost model of compulsory licensing strikes me as similar to the concept of the hypothetical "reasonable royalty," used across IP regimes to set damages in certain circumstances. It seeks to determine, by assessing a variety of factors, what the rights holder and the licensee would actually have agreed to if they sat down to hammer it out in a negotiation. They didn't. But if they had, what would the licensee have paid? According to the Calabresi/Melamed model, following what a "willing buyer" and "willing seller" would consider a fair price will, all else being equal, achieve the most economically efficient outcome.
The 801(b) Factors
The best evidence of Victor's view that compulsory licenses were intended to serve the incentives/access tradeoff is that, starting under the 1976 Act, compulsory license rate setting for music was conducted pursuant to a factor based approach codified in 17 U.S.C. § 801(b). Under this scheme, the Copyright Royalty Tribunal (CRT) (and eventually the CRB) was given power to set rates for the mechanical license every 10 years. (25). The CRT was supposed to set the mechanical royalty rate based on consideration of the "801(b) factors/objectives":
(A) To maximize the availability of creative works to the public;
(B) To afford the copyright owner a fair return for his creative work and the copyright user a fair income under existing economic conditions;
(C) To reflect the relative roles of the copyright owner and the copyright user in the product made available to the public with respect to relative creative contribution, technological contribution, capital investment, cost, risk, and contribution to the opening of new markets for creative expression and media for their communication;
(D) To minimize any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.(25-26) (quoting former 17 U.S.C. § 801(b)).
I added the emphasis to the first and third factors because they are key to Victor's argument. As he puts it, "the 801(b) factors correspond in many respects to an approach to compulsory licensing that emphasizes the incentives/access tradeoff, rather than merely transaction costs remediation. In particular, the first factor emphasizes that the goal of rate-setting regulators should be to maximize the availability of creative works 'to the public.' "(28). In addition, he observes that the third factor appears to direct the CRT to take into account "the relative roles of the copyright owner and the copyright user" in, among other things, communicating the work to the public. In Victor's interpretation, this invites balancing between the value of incentives for copyright owners, and the value of access added by copyright users, such as player piano manufacturers and, today, music streaming services. (28).
Eschewing the 801(b) Factors in Favor of a "Willing Buyer and Willing Seller" Approach
The history surrounding creation of the CRT and institution of rate setting procedures pursuant to the 801(b) factors entails significant statutory developments and frustratingly complex minutiae. A few of these minutiae include the difference between rights to use a musical composition versus a sound recording; the difference between a reproduction right versus a public performance right; and the bizarre distinction Congress made between compulsory licensing of interactive versus non-interactive streaming services. Victor tackles the task of explaining these complexities. (37-51). I highly recommend his explanation for those who teach IP and tend to breeze over/avoid compulsory licenses.
But only a few developments are necessary to understand Victor's basic argument. He asserts that when Congress passed the Digital Millennium Copyright Act (DMCA) in 1998, amending the Section 114 compulsory license for certain public performances of sound recordings (34-35), and again when Congress passed the Music Modernization Act (MMA) last year, overhauling the entire compulsory licensing regime for both sound recordings and musical compositions (here is Tyler Ochoa's analysis via Eric Goldman's blog), Congress replaced the "access/incentives tradeoff" approach to compulsory license rate setting with a “willing buyer and willing seller" approach. The upshot is that "in future rate-setting proceedings, the rates of both the [Section 115] mechanical license and the Section 114 license will be set using an ostensible market-mimicking standard." (42).
The two key statutory provisions are currently codified as follows. Section 115 now provides, with respect to musical compositions, in relevant part,
"[t]he Copyright Royalty Judges shall establish rates and terms that most clearly represent the rates and terms that would have been negotiated in the marketplace between a willing buyer and a willing seller."17 U.S.C. § 115(c )(1)(F).Section 114 now provides, with respect to the sound recording compulsory license, in relevant part,
"[t]he Copyright Royalty Judges shall establish rates and terms that most clearly represent the rates and terms that would have been negotiated in the marketplace between a willing buyer and a willing seller."17 U.S.C. § 114(f)(1)(B).
At a normative level, he argues this shift is harmful to the copyright system. Viewing the compulsory music licensing regime as similar to other private ordering-based solutions to transaction cost problems, "miss[es] the main justification for compulsory copyright licensing, both descriptively and normatively." (70). Instead, he argues, the compulsory music licensing regime should be considered more like those internal copyright doctrines that serve to "ensur[e] balance within the copyright system," such as fair use, first sale, term limits, and the idea-expression dichotomy. (70-71).
Some Comments
Victor's argument is compelling to me, having not researched music compulsory licensing independently of reading his article and writing this post. I assume the biggest pushback he will receive is the point made earlier—that compulsory licensing is not like fair use; it is not intended to do anything more than find the market price that the parties would have reached if not for the high transaction costs. Putting that larger question aside, I have two quick comments.
1. How Different Is the Market-Set Rate from the Incentives/Access Rate?
First, this paper raises the perennial problem of the counterfactual. Victor suggests that when regulators are explicitly told to balance between incentives and access in rate setting, rather than told to use rates suggested by free-market licensing deals, the result will often be lower rates, and that rates were lower under the 801(b) factors than they would have been if Congress had instead ordered rate setting to proceed under a "willing buyer and a willing seller" approach from the start.
But is this true? Victor quotes Matthew Sag and Peter DiCola's assertion that, in some circumstances, music rates set by arbitrators "were far too high " (62) (quoting DiCola and Sag). But without hard evidence, it is hard to assess the validity of Victor's conclusion that "[h]ad these proceedings been governed by the 801(b) factors, the outcome might have been different." (62). There are a lot of moving pieces that go into rate setting, especially when using a factor-based test, so I just wonder how different rates really are if calculated from a market-set perspective than from an incentives/access perspective. Yes, the first 801(b) factor was "[t]o maximize the availability of creative works to the public[,]" But there were also other factors that the rate setting entity could consider, like "[t]o afford the copyright owner a fair return for his creative work...", and so on (25-26) (quoting former 17 U.S.C. § 801(b)). For comparison, when calculating a hypothetical "reasonable royalty" under the Georgia-Pacific factors in assessing patent damages, courts consider many factors to decide what a hypothetical market negotiation would look like. It is anyone's guess how different a hypothetical "reasonable" royalty is from a hypothetical "access-friendly" royalty in any given case.
I would also be remiss if I didn't point out that, unsurprisingly, commentators on the copyright system are in disagreement on whether rates are too high or in fact too low. For example, in a report prepared in the fall of 2018 entitled "The Market for Performance Rights in Sound Recordings: Bargaining in the Shadow of Compulsory Licensing," Professor Mark Schultz asserts, based on a study of sound recording rates around the world, that "rates are probably not too high. For one thing, we can be confident that current rates are not sufficiently high to significantly impair the use of music..." So it's all a matter of perspective, I suppose.
2. Are Music Listeners' Interests an Externality?
Second, I am curious whether there is an externalities story here that would be an alternative way to frame Victor's argument. In his telling, some people see the goal of compulsory licensing as replicating the rate that the copyright owner and the licensee would reach if transaction costs were lower. But markets don't work well if there are externalities involved. It seems to me that in many situations, the consumer of copyrighted music (the Spotify listener, for instance) is effectively an external party to the transaction that occurs between the copyright owner and the distributor (Spotify). They are merely "the public" contemplated by the 801(b) factors, rather than the "user" who is obtaining the license. The interests of the end consumer, including their desire to "consume" music as well as their potential desire to produce new things with it, are likely not fully accounted for in the deal, or not as well as they would be if they were actually involved in the negotiation. Thus, perhaps one could argue there is a market failure that necessitates deviating from a market price.
Victor notes the prevalence of the law-and-economic concept of externalities, a.k.a. "spillovers," in discussions surrounding fair use, observing that
[t]he narrow transaction-costs-focused version of fair and the more expansive approach embraced by [Glynn] Lunney and others [such as Abraham Bell & Gideon Parchomovsky, William W. Fisher, Lydia Pallas Loren, Brett M. Frischmann & Mark A. Lemley] rest on different conceptions of how copyright markets fail, though both are grounded in economic theory. The former sees transactions costs as the only meaningful barrier to otherwise efficient allocations of copyright goods in the marketplace, whereas the latter considers factors like deadweight loss and externalities as relevant to fair use’s role in ensuring a well-functioning copyright system.
(16).
I would love to see, in this or in future works, Victor's take on how the possible presence of externalities should influence the way in which compulsory copyright licenses are set. This could raise very interesting questions and achieve more granularity in conceptualizing the "incentives/access" tradeoff in the compulsory licensing context.
I would love to see, in this or in future works, Victor's take on how the possible presence of externalities should influence the way in which compulsory copyright licenses are set. This could raise very interesting questions and achieve more granularity in conceptualizing the "incentives/access" tradeoff in the compulsory licensing context.
