The basic problem is that Spotify and other streaming services are trying to license music using a set of arcane procedures and institutions that, in some cases, haven’t changed in a century, to use on a platform that didn’t exist even a few years ago. While it is difficult to understand the full details of this system, it is easy to understand that it is a mess. Given that streaming services now account for more than 60% of music revenues and digital downloads account for another 20%, the gap between the highly regulated world of music licensing and the realities of the market will grow increasingly large.

There have been some recent efforts to modernize music licensing. But all too often these efforts fall back on more regulation rather than a greater reliance on market processes. They are just slightly newer choruses of the same old song.

The Properties of Music

Music is a classic example of a good with non-trivial initial fixed (“first-copy”) costs and low (often near zero) marginal costs. To create music, composers and songwriters must devote effort to creating a song. Similarly, embodying that song in a sound recording requires the production efforts of musicians and/​or vocalists, as well as recording technicians, producers, etc. But once the song exists in recorded form (or even as sheet music), the marginal costs of distributing it to users are relatively low—and in the digital age they are realistically close to zero.

If copies of the song can be distributed without restriction, competitive markets are likely to drive the price of the distributed song to (near) zero. In that case, the creators will be unable to recoup their costs, which will discourage the initial creation. This is a familiar tension between the short-run desirability of allowing information (and music is clearly “information”) that already exists (which has all of the properties of a “public good”) to be sold or distributed at its marginal costs and the long-run desirability of providing incentives for the information to be created in the first place.

The solution to this problem is to give creators a property right—a right to license (and thereby also to exclude)—that would allow them to recoup their costs and earn a profit (if, of course, their creation meets a market test). And, indeed, there is such a property right in the United States and most other countries: copyright. In the United States, music has been covered by copyright since 1831. Initially, music copyright just protected composers and/​or the publishers of sheet music against unauthorized copying of that sheet music. Copyright protection was extended to cover public performances of music in 1897. A decade later, music copyright (in 1909) was extended to cover “mechanical” reproductions such as piano rolls, physical recordings, and (as of the mid-1990s) digital downloads. And in the mid-1990s, a new law established a separate copyright to cover the creators (i.e., the recording artists and/​or the recording companies to whom the artists might assign their rights) of sound recordings.

Under current law, copyright lasts for the life of the creator plus 70 years or, if the copyright is owned by a corporation, 95 years after publication. After the expiration of the copyright, the music enters the public domain, and anyone can freely copy it.

The Underpinings of Well-funitoning Markets—and the Problems for Music

Well-defined, enforceable property rights are a necessary condition for well-functioning markets. Low transactions costs are also important. And, as public real estate registry rolls have illustrated, having a central registry of who owns what rights can enhance the functioning of markets where the goods and services that embody those rights are traded.

Unfortunately, the characteristics of music—plus some idiosyncratic features of the U.S. market—have impeded freely functioning competitive markets in music licensing and distribution. Instead, the markets have been heavily regulated.

If music copyright owners are going to enforce their ownership rights, they need to be able to sign licensing contracts with users, monitor their music use, collect the appropriate royalties, and enforce the property right against infringers. But the infrastructure for such efforts is unlikely to be complementary with songwriting or song-recording talents. And in a country the size of the United States—with tens of thousands of music distribution outlets (e.g., terrestrial radio stations, TV stations, theaters, nightclubs, bars, restaurants, music halls)—the contracting, monitoring, royalty collecting, and enforcement efforts would be substantial, especially in the pre-digital era. In the digital era, streaming services need to obtain rights to tens of millions of songs to operate successfully.

Unsurprisingly, intermediaries have sprung up to help with these transactions costs. For songwriters and composers, music “publishers” began undertaking the intermediary function when sheet music was the primary way that music was distributed. For performing artists, it has been the recording companies (which are frequently described as “record labels” or even just as “the labels”). But the economies of scale in contracting, monitoring, royalty collecting, and enforcement by the early 20th century appeared to be beyond even the largest music publishers, and the first of a series of collective “performing rights organizations” (PROs) to carry out these functions—the American Society of Composers, Authors, and Publishers (ASCAP)—was formed in 1915. It was followed in 1930 by the Society of European Stage Authors and Composers (SESAC) and by Broadcast Music Inc. (BMI) in 1939. Most recently, a new PRO, Global Music Rights (GMR), was formed in 2013. ASCAP and BMI dominate this area.

Similarly, after “mechanical rights” became protected, the National Music Publishers Association in 1927 formed a PRO-like subsidiary, the Harry Fox Agency, to carry out these functions with respect to mechanical rights.

It was no accident that, when Congress created the copyright for sound recordings in the 1990s, it anticipated that a PRO would be needed. But rather than allow the market to develop one or more PROs for those rights, it authorized the Librarian of Congress to designate the PRO. SoundExchange, which was originally a subsidiary of the Recording Industry Association of America and was subsequently spun off as a nonprofit organization, is that designated PRO.

Perhaps surprisingly, despite the presence of (literally) a handful of PROs, a comprehensive database of who owns what rights with respect to music distribution has not come into existence. The existence of “blanket licenses” that are made available by the PROs—which allow a user to distribute all of the music that is in a PRO’s repertoire—has traditionally reduced the need for such a database. The institution of the blanket license has clearly reduced transactions costs. But the absence of that database has recently impeded efforts by digital streaming services, such as Pandora, to negotiate effectively with individual music publishers for distribution rights. It also makes it more difficult to identify and pay rights holders, which at least partly accounts for the nonpayment problem and consequent lawsuits against Spotify and other streaming services.

One other special institutional feature of music licensing is worth noting: in the area of mechanical rights and for some categories of distribution of sound recording rights, compulsory licenses are required by statute. This means that, so long as a music distributor meets a modest set of requirements, an owner of mechanical or sound recording rights must grant a license to the distributor. But, of course, the question of “at what price?” remains. Arm’s-length bargaining in the shadow of such a requirement is difficult at best, and it is no accident that regulated rather than market-based pricing dominates the pricing of these rights.

The Court as Regulator

Centralizing performance rights functions in a few large PROs reduced transactions costs, but it also led to the accretion of market power. ASCAP became a locus of market power for the selling of music performance rights vis-à-vis the broadcasters and other distributors of the music by becoming the common agent for thousands of otherwise-competing composers/​songwriters and music publishers. This arrangement drew a government response: the Antitrust Division of the U.S. Department of Justice sued ASCAP in 1934 and again in 1941, arguing that ASCAP’s collective setting of royalty rates for its thousands of otherwise-competing members constituted a violation of the Sherman Act.

These suits were eventually settled with a consent decree in 1941. The decree allowed ASCAP to continue functioning as a collective licensor for its members, an implicit recognition of the PRO’s role in reducing the contracting and monitoring transactions costs for its members. However, the decree—including subsequent modifications—placed restraints on ASCAP’s actions. Perhaps in hopes of not foreclosing a more competitive market, the decree precluded ASCAP from requiring exclusivity from its members—i.e., members of ASCAP are free to license their works outside of the ASCAP framework. Prospective licensees also must have the option of licensing individual pieces of music from ASCAP rather than being required to take a blanket license. However, distributors do typically obtain blanket licenses for the entire catalogue of works from ASCAP and BMI, and from the smaller SESAC and GMR.

The ASCAP decree specified that the Federal District Court for the Southern District of New York would arbitrate disputes when prospective licensees and ASCAP could not agree on terms. A similar suit against BMI in the early 1940s was settled with a similar consent decree. As a result, the Southern District is often described as “the rate court” for these disputes, and two judges in the Southern District—one for ASCAP adjudications and one for BMI adjudications—have become the de facto regulators of these license terms—including pricing—for musical compositions.

Royalty rates for the SESAC and GMR music catalogues are negotiated directly with users. There are no government consent decrees for those latter two PROs that apply when the parties fail to agree on terms.

Development of New Distribution Technologies

As the analog era gave way to the digital era, Congress attempted to modernize music licensing by enacting the Digital Performance Right in Sound Recordings Act (DPRA) in 1995 and the Digital Millennium Copyright Act (DMCA) in 1998. These acts expanded copyright protection to public performances of sound recordings through new digital audio transmissions, including the then-emerging satellite services that were the predecessors to SiriusXM and subsequent internet-based streaming services such as Pandora and Spotify.

The Copyright Royalty Board (CRB), consisting of three administrative judges appointed by the Librarian of Congress, establishes rates for sound recording performance rights for distributors such as satellite radio and non-interactive streaming services like Pandora. Different statutory standards are applicable to different categories of services; for example, the CRB’s rates for SiriusXM are different than those for Pandora. The licenses are compulsory: the rights owners cannot refuse a licensing request at the CRB-determined rates. SoundExchange collects and distributes the royalties and also participates in CRB proceedings on behalf of the copyright holders.

Interactive digital services like Spotify are exempted from the CRB process and instead negotiate directly with the performing artists and labels.

Terrestrial radio broadcasting does not currently require sound recording licenses, although legislation to change that has been introduced in Congress. The argument for exempting terrestrial broadcast radio from paying this category of royalties was that the playing of music on terrestrial radio was a form of free promotion for the records and their artists (or the record labels that had assembled and paid the artists) and thus the latter group(s) were already being compensated by the broadcasters.

The Licensing Market

While the music distribution system has grown more competitive and changed dramatically with the introduction of new technologies, the music licensing system hasn’t kept pace. Using any piece of music or recording requires obtaining multiple licenses, each of which has its own rules. Some licenses are compulsory, some are not. Most royalty rates are set in administrative or judicial proceedings. Some are established by direct negotiation between rights holders and licensees.

The market for music rights largely consists of three regulated markets: one for musical composition public performance rights; a second for sound recording performance rights; and a third for mechanical reproduction rights. For example, Spotify and other “interactive” streaming services must obtain licenses to sound recording and composition performance rights, as well as “mechanical licenses,” in order to stream a song legally. Mechanical rights—the rights at issue in the recent spate of lawsuits—include the right to reproduce and distribute copyrighted songs through permanent digital downloads, interactive streams, or other media. Mechanical licenses are compulsory, and their rates are determined in CRB proceedings.

The Music Modernization Act

Several music licensing proposals are now pending in Congress, the most significant of which is the Music Modernization Act (MMA), which recently passed the House of Representatives. The MMA is intended to solve the problem that is at the root of the Spotify and similar lawsuits, which involves nonpayment of royalties for mechanical rights. The bill has received widespread support from most parts of the industry, including music publishers and songwriters (who own and license these mechanical rights), and digital interactive streaming services (who purchase the licenses).

Ameliorating the nonpayment problem, as the MMA aims to do, would be an improvement over the status quo. However, the MMA reinforces many of the long-standing aspects of music licensing that hinder competition: compulsory licenses, blanket licensing by a music collective, and regulated rates by the CRB. It does little to create a more competitive licensing regime with more direct negotiation between rights holders and licensees, which should be the long-run goal. This is indeed the same old song.

Central to the MMA is a Musical Licensing Collective (MLC), which would be designated by the Register of Copyrights (similar to the earlier designation of SoundExchange for sound recording rights). The MLC would be empowered to grant blanket mechanical licenses for interactive streaming or digital downloads of musical works. It would be responsible for collecting royalty payments from music services and distributing those payments to songwriters and publishers.

Would such a monopoly collective have too much market power? The MLC would perform the same functions that existing PROs perform with respect to public performance rights for musical works. But those PROs—ASCAP, BMI, SESAC, and GMR—compete with each other. New PROs can and do enter this market.

The MLC would also perform the functions with respect to mechanical licenses that firms like the Harry Fox Agency, Music Reports Inc., and newer companies such as Audiam and Loudr, already do. Would these companies be driven out by the MLC?

The litigation resulting from nonpayment of royalties suggests dissatisfaction with the way this function is now being performed. However, there is no reason to believe that a government-certified nonprofit organization, facing no competition, would perform better.

In the absence of alternatives, the quality of service to the publishers and songwriters is likely to suffer. The MMA legislation envisions a process for reviewing the MLC’s performance every five years and possibly designating a new MLC, but this represents a limited form of competition and in practice may be difficult to implement.

The MLC also would be charged with creating and managing a database of mechanical rights. A music database is a potentially pro-competitive outcome of the MMA. Accurate and easily obtainable information on who owns what is a necessary precondition for a more competitive licensing regime with more direct bargaining between rights holders and licensees. However, combining the database with the other functions of the MLC would add to its market power.

Creating a useful database may require a push from the government, such as is provided in the MMA. However, a number of nongovernmental groups are already working on this problem. The Open Music Initiative—a venture of the Berklee College of Music, the MIT Media Lab, and others—is developing an open source protocol for identifying music rights holders using blockchain technology. Other initiatives, such as Dot Blockchain Music, are developing music rights management systems based on blockchain. SoundExchange has put together a database of the songs for which royalties are being held in escrow by the Copyright Office because the licensee could not identify who to pay. They will make the database available to publishers for free.

The danger is that a government-sanctioned effort would crowd out these private efforts and eliminate competition between various databases and database technologies. Ultimately, without competition, the MLC is more likely to settle on an inferior technology.

This discussion raises the question of whether the economies of scale in maintaining a music database are large enough that it should be considered a natural monopoly. If so, we might be less worried about the MLC emerging as a monopoly. However, we should still be concerned about the process of gaining a monopoly position through a government certification process rather than market competition on the merits. The presence of a number of private-sector entities working in this area suggests that competition—either within a multi-firm market or as competition for the market—would produce good results.

Even if policymakers determine that the government should provide assistance in some form to the creation of a music rights database, it doesn’t follow that the database function should be combined with the license administration function. And even if the database were a monopoly, the licensing function—which includes administering payments to rights holders— could and probably should be competitive. Entities that perform that function—such as ASCAP and BMI—are agents of the publishers and songwriters, who would benefit from competition for their business.

Under the MMA structure, it’s not clear if publishers or songwriters would be able to withdraw from the MLC and bargain directly with the streaming services. The existence of the compulsory license would make that difficult.

The MMA envisions a database operated by the MLC, which presumably is only concerned with mechanical licenses. If the MLC’s database becomes the definitive database, shouldn’t it also include information on other rights, such as performance and sound recording rights? Wouldn’t a comprehensive database increase licensing efficiency? Admittedly, a more comprehensive database would be more difficult to assemble because it would require data from a much broader scope of entities.

Competition Is Attainable and Already Occurring

The goal of more competition is attainable. There are numerous examples of direct negotiation in this otherwise highly regulated sector. Interactive streaming services already negotiate directly for sound recording rights, which suggests that direct negotiation could also work for other rights, including the mechanical rights that are the subject of the MMA. A recent CRB rate proceeding provided evidence of negotiated contracts between record labels and non-interactive services Pandora and iHeartMedia. Pandora has negotiated an agreement with Music and Entertainment Rights Licensing Independent Network (“Merlin”), which represents thousands of independent music labels. iHeartMedia has negotiated agreements with both major and independent labels. The elimination of existing compulsory licensing provisions would surely encourage more negotiation.

Pandora has negotiated contracts for composition public performance rights with the major publishers. The smaller PROs—SESAC and GMR—are not subject to the antitrust consent decree and therefore are unregulated.

New companies that act as agents for rights holders are springing up. Merlin, mentioned above, represents thousands of independent record labels. Kobalt has developed a digital rights management platform that helps artists and publishers collect royalties from streaming services. Kobalt collects royalties directly for 8,000 artists, including Paul McCartney, and 500 publishers, including Disney. It may well be that in a digital era, the monitoring and other PRO-like functions that are important to rights holders are easier to accomplish and require less scale than was needed in the pre-digital world.

Music Licensing Reform: Singing the Same Old Song

Whenever an opportunity for pro-competitive reform of music licensing arises, policymakers seem to revert to the traditional regulatory model that discourages competition. They never miss an opportunity … to miss an opportunity. The MMA—with its reliance on compulsory licensing, blanket licensing by a marketing collective, and regulated rates—is the latest of several recent examples.

Another example involves ASCAP and BMI, which have operated since 1941 under antitrust consent decrees that encourage blanket licensing of their entire catalogues and provide for regulated rates by the antitrust rate court. But music publishers now want the ability to withdraw their catalogues partially from the PROs in order to negotiate directly with the new streaming music platforms. With that aim, the publishers initiated a review of the consent decrees by the Justice Department’s Antitrust Division. The Division concluded in 2016 that it would not seek to modify the decrees to permit partial withdrawal. This would have been an opportunity to encourage a more competitive market with more direct bargaining between rights holders and distributors.

Yet another example involves the payment of royalties by terrestrial radio. As indicated above, unlike every other music platform, old-fashioned terrestrial radio stations—by statute—pay no royalties (for sound recording rights) to record companies on the theory that radio stations provide a promotional service. Some in Congress are attempting to do away with this zero-price rule. But rather than allowing market forces to determine the royalty for radio play, the Fair Play Fair Pay Act proposes to incorporate radio into the current regulatory system with CRB-determined royalties.

The current highly regulated system for licensing music is largely an artifact of the analog era of music distribution and performance. The digital technologies now available make it possible for competition to replace much of the traditional regulatory structure for music licensing and rate determination.

Unfortunately, this won’t happen until lawmakers sing a different song. The tune and lyrics ought not to be hard to learn; after all, Congress—after many decades of requiring spectrum regulation—endorsed the flexible markets that competitive auctions have brought for wireless spectrum usage. Perhaps Congress just needs to sing that song in a different key.