Copyright without Creators

In the scholarly literature on copyright, the term “publishers” is often used to denote both literary publishers and a broader range of distribution intermediaries. I use the generic term “intermediaries” to cover the full range of screening, publishing, distribution and marketing entities in creative goods markets. Note further that I use the term “artists” interchangeably with “creators” or “authors” to denote individuals who are the source of creative inputs.

I am not the first to identify the role played by copyright in supporting distributors’ incentives, but to my knowledge it has never been fully elaborated as the primary justification for copyright without reference to authorial incentives. The sole exception is Mosoff (2013), who proposes a publisher-oriented approach to copyright (with a special application to scholarly publishing). Nadel (2004) and Bambauer (2008) provide more extensive analysis of the connection between copyright and distribution incentives but conclude that copyright cannot be justified on that basis. Ku (2002) and Zimmerman (2003) recognize copyright’s role in supporting publisher’s incentives; however, Ku argues that this argument is now obsolete due to technological changes. For earlier relevant discussions, see Kaplan (1967:8–9, 75); Goldstein (1990–1991:109–113); Chafee (1945:509–510). A familiar line of argument identifies large media interests as the primary beneficiary of copyright law but relies on that observation to argue that copyright law has been distorted so as to advance those private interests at the expense of the public. Unlike these prior contributions (with the exception of Mosoff (2013)), I argue that (i)copyright can be justified on the basis of distributors’ (or more broadly, intermediaries’) incentives, (ii) intermediaries are not obsolete in digital content markets; and (iii) intermediaries’ interests are not inherently misaligned with the public interest.

For perhaps the original and best statement of copyright skepticism, see Plant (1934, reprinted 1974). For widely cited modern examples, see Breyer (1970); Lessig (2004).

For exceptions, see Bambauer (2008); Nadel (2004); Zimmerman (2003); Ku (2002).

Note that I am using the term “romantic” in this case to mean an individual motivated by non-economic considerations. This differs from the term, “romantic author,” as used by some copyright scholars to refer to the concept of the unique creative genius (Jaszi and Woodmansee, 1994:6; Woodmansee, 1984). For further discussion, see Hughes (1998).

This is true, for example, of the leading economic model of copyright in Landes and Posner (2003).

In the following discussion, I concentrate on the tasks typically required to source, produce and distribute content in “non-digital” content markets. Subsequently, I revisit how these tasks (and associated costs) have changed in digitized content markets. See infra Part 5.1.

For example: Avengers, released in 2012, had a reported budget of $220 million (Stewart, 2012); The Dark Knight, released in 2012, had a reported budget of $230 million (Fritz andKaufman, 2012); Disney’s John Carter, released in 2012, had a $250 million budget (Graser, 2012).

For similar figures, see Vogel (2011) (stating that marketing costs can rise to $100,000 for a standard release and exceed $500,000 for a major artist); Allen (2010:201) (stating that a label’s costs to “market and promote a single easily reach $1 million”).

Waldfogel (2012b) offers a dissenting view, arguing that artists are increasingly able to secure sales through exposure as a result of Internet radio play and online social media.

In 1986, out of a total population of approximately 25,000 new hardbound trade titles, there were less than 200 best-sellers, which generated an estimated nearly $1 billion in sales out of a total of $1.7 billion. Put differently: less than 1% of all trade books published in 1986 accounted for almost 60% of total sales (Liebowitz and Margolis, 2003).

Since the value of the good is not objectively verifiable, this dilemma cannot be resolved by a warranty contract that would refund payment subject to consumer satisfaction (obviously, that contract would never result in a payment being owed by consumers, who would always claim to be dissatisfied).

For similar views, see Caves (2002:178–180, 185–186). For application of this type of model to the fashion market, see Barnett (2005).

For discussion, see Caves (2002:189–200).

The extent to which the presence of a star actually improves the likelihood of a successful release remains unresolved. Using a sample of 2000 films released from 1984 to 1996, De Vany (2004 §§ 4.3.2, 4.5.1, 4.5.2) finds that, on average, a star significantly increases a movie’s higher least revenue (that is, a star constrains the lower tail portion of the revenue distribution) and slightly increases a movie’s chance of making a profit, in each case relative to a movie without a star. This is consistent with other findings in the empirical literature showing that actors positively impact opening performance (Elberse, 2007:102, 120), as well as popular observations in Hollywood that “stars help the movie to open.”

For a fuller discussion, see Caves (2002:157–160).

It might be argued that an artist can diversify his or her investment to some extent across multiple projects. Even assuming this to be true, an individual’s risk-diversification capacities cannot plausibly approach the diversification capacities of a large corporation, which has access to deep internal and external capital markets by which to construct a portfolio of tens to hundreds or more of creative projects. At a minimum, a large entity offers a lower-cost risk-diversification vehicle relative to an individual artist and is therefore the socially preferred mechanism by which to achieve this function.

For other evidence that U.S. publishing suffered from a lack of capital and risk-bearing capacities, see Charvat (1959:42–43).

This hedging effect is reflected literally by cross-collateralization clauses in producers’ contracts, which sometimes specify that a producer is only entitled to a share of the profits once the film’s revenues have been allocated to cover both the costs of that particular film as well as a certain portion of the costs attributed to other films (Vogel, 2011:123).

The major studios and parents are as follows: Paramount Pictures (subsidiary of Viacom); Sony Pictures (subsidiary of the Sony group); Twentieth Century Fox (subsidiary of News Corporation); Warner Bros. (subsidiary of Time-Warner); Universal (subsidiary of Comcast); Walt Disney Studios (subsidiary of the Walt Disney Company).

On record labels’ ability to pool risk, see Schultz (2009).

Sony Music Entertainment is a subsidiary of the Sony corporate group and Universal Music Group is a subsidiary of Vivendi, a media conglomerate. The final member of the Big Three, Warner Music Group, is currently owned by Access Industries, an industrial conglomerate.

For similar thoughts, see Hardy (1988:183–185); Birnhack (2009:133–134).

Dates for the aforementioned extensions are, respectively, as follows: Act of Feb. 3, 1831, ch. 16, sec. 1, 4 Stat. 436, 436; Act of Aug. 18, 1856, ch. 169, 11 Stat. 138, 138–139; Act of Mar. 3, 1865, ch. 126, §1, 13 Stat. 540, 540; Act of July 8, 1870, ch. 230, §86, 16 Stat. 198, 212; Act of Mar. 4, 1909, ch. 320, §1(e), Pub. L. No. 60–349, 35 Stat. 1075, 1075–1076; Townsend Amendment of 1912, ch. 356, §5(l), Pub. L. No. 62–303, 37 Stat. 488, 488; Sound Recording Amendment of 1971, Pub. L. No. 92–140, 85 Stat. 391 (1971).

The paucity of classics is reflected in part by the fact that when renewals were required to maintain the term of a copyright, only a small percentage of works were renewed despite a low fee (Rappaport, 1998).

One report finds that record labels’ catalogs generate expected cash returns of 7–20% annually (Satariano, 2009).

This concern was particularly salient at the time of the most recent extension of copyright – the Sonny Bono Copyright Term Extension Act, enacted in 1998. One of the stated motivations behind the extension was to enable U.S. copyright holders to enjoy a recent term extension under European copyright laws. U.S. copyright holders would not have been protected in those foreign jurisdictions upon expiration of the then-shorter term under U.S. copyright law. The 1998 term extension therefore both directly and indirectly enhanced the internal financing capacities of U.S. distribution intermediaries, who derive a large portion of their revenues from foreign markets. I thank Paul Goldstein for bringing this point to my attention.

For exceptions, see Liebowitz and Margolis (2003); Landes and Posner (2002).

This type of objection is pursued in Balganesh (2009:1589–1591, 2010:676–681).

See, e.g., Folsom v. Marsh, 9 F. Cas. 342 (C.C.D. Mass. 1841); Woolsey v. Judd, 11 How. Pr. 49 (N.Y. Super. Ct. 1855); Grigsby v. Breckinridge, 65 Ky. 480 (App. Ct. 1867).

Act of July 8, 1870, ch. 230, §86, 16 Stat. 198, 212 (extending copyright to dramatizations and translations); Act of Mar. 3, 1891, ch. 565, 26 Stat. 1106 (clarifying that copyright extends to dramatizations).

Copyright Act of 1909, ch. 320, §1(b), 35 Stat. 1075.

17 U.S.C. § 106(2) (2006).

17 U.S.C. §1(b) (repealed 1976).

Nadel (2004) argues that the derivative right, and the associated expansion in expected profits on a blockbuster hit, induces copyright holders to favor popular works over more “economically marginal” works. There are three vulnerabilities in this argument. First, it is unclear why reducing total expected returns (by weakening copyright) would lead producers to favor “economically marginal” works. The opposite would seem to be the case given that producers would then favor the lowest-risk projects that appeal to the broadest population; conversely, expanding expected returns (by increasing copyright) provides producers with additional profits that can be invested in higher-risk “artistic” projects that appeal to niche audiences. For similar views, see Goldstein (1990–1991:113). Second, the winner-take-all nature of mass-cultural markets would persist even in the absence of copyright, given that it is largely attributable to low-cost reproduction technologies (which undermine the commercial value of all “runner-up” artists) and the evaluation difficulties inherent to experience goods (which induce consumers to use stars as a proxy for quality). Third, available evidence largely rejects the widely shared intuition that concentration suppresses cultural diversity. SeeAlexander (1996) (finding that moderately concentrated creative markets exhibit greater product diversity relative to both more and less concentrated markets); Alexander (1997) (same); Einstein (2004) (finding that programming diversity in network television remained largely unchanged after repeal of the “financial interest and syndication” rules, which resulted in further industry consolidation); Lopes (1992) (finding no relationship between market concentration and product diversity in the popular music market).

For similar observations, see Bracha (2008:248–249); Lemley (1996:882–883).

For a description of this type of criticism expressed during deliberations over the 1976 revision to the Copyright Act, see Van Houweling (2010:606).

For related thoughts, see Birnhack (2009:128–131).

Historical research shows that the 1909 Act’s presumption codified judicial tendencies that predate the Act (Fisk, 2003).

Note that Hardy’s study was limited to cases adjudicated through 1987. The governing decisions on joint authorship in the prominent Second and Ninth Circuits, both of which have been issued since the period covered by Hardy’s study, are consistent with this trend insofar as both set standards that assist large intermediaries in defending against claims of joint authorship by non-employee contributors. See Aalmuhammad v. Lee, 202 F.3d 1227 (9th Cir. 1999); Childress v Taylor, 945 F.2d 500 (2d Cir. 1991).

17 U.S.C. §101 (2006) (defining “work made for hire” as “a work specially ordered or commissioned for use (1) as a contribution to a collective work; (2) as a part of a motion picture or other audiovisual work; (3) as a translation; (4) as a supplementary work; (5) as a compilation …”). On the manner in which the work-made-for-hire provision enables publishers to avoid fragmentation of rights over a single creative work, see Van Houweling (2010:598).

See Act of Mar. 4, 1909, ch. 320, s 42, 35 Stat. 1075, 1084; see also 3 Nimmer & Nimmer §10.01 [A] (2010) (stating that the rules on assignment and divisibility made it “impossible to assign anything less than the totality of rights commanded by copyright”).

For further discussion, see U.S. CONGRESS, HOUSE REPORT NO. 94–1476, NOTES AND REVISIONS, 17 U.S.C. §201; Van Houweling (2010:564, 601–602).

For similar thoughts, see Landes and Posner (2003:39); Fisher (1998:1237–1239).

See supra note 7.

Some sources indicate that recording costs are higher in genres that involve complex arrangements that necessitate the assistance of a professional production engineer (Gordon, 2011:280).

The commission assessed by end-user Internet intermediaries (e.g., iTunes) depends on the distributor counterparty. Trade commentary often mentions that iTunes and similar websites give preferential terms to major labels (see e.g., Christman, 2011).

These promotional tasks include (i) producing a music video (which involves costs beyond the budget of most individuals (Davis and Laing, 2006:210)); (ii) promoting music to music critics, and hundreds of radio shows and other media outlets; (iii) bulk buying of advertising on television; and (iv) negotiating on a bulk basis with online retailers and other online distribution services (Statement of Martin Mills, 2012).

The percentage share of large labels approaches 90% in the “AAA Commercial” programming format – that is, commercial “pop” music stations. The percentage share for non-major (“indie”) labels only declines relative to the average in the Country and AAA Noncommercial (e.g., college radio stations) formats (Thomson, 2009). Waldfogel (2012b) finds that independent labels have garnered an increasing share of radio play on Internet radio stations for the period 2006–2011.

For a current example, see Grossman (2009) (reporting that unknown author attracted attention by self-publishing and then secured contract with a traditional publisher to promote the book heavily). For older examples, see Bowker (1996:30) (describing success of Christmas Story and The Celestine Prophecy, best-sellers that were self-published in 1993 and 1994, respectively, and then widely distributed by conventional publishers in 1995).

To be precise, the findings tend to indicate that the “tail” of the demand curve lengthens but flattens – that is, there are more purchases of niche products but those products are usually only purchased once or twice – while the “head” of the curve becomes even “fatter” – that is, there is even more intense clustering around a small number of hits (Elberse, 2008).

SeePitt (2010a:112, 116–117, 2010b).

SeeMichaels (2008) (reporting a study of iTunes downloads in the United Kingdom in 2008 by the UK collective licensing agency, which found that 85% of total inventory did not sell at all and the bulk of all downloads derived from a small group of hits); Page and Garland (2007) (finding that, in peer-to-peer file sharing, there is a dominant superstar effect, insofar as most users cluster around a small number of tracks (the top 5% of downloads constitute 80% of all downloads), and a weak long-tail effect, insofar as the vast remainder of tracks are swapped at least once but still do not attract significant attention).

Consistent with this assertion, Elberse (2008) finds that the long-tail effect in music downloads is most pronounced in the case of high-intensity users, which correspond to informed consumers (who presumably do not require the star proxy to assess product quality). Given uninformed consumers’ rational preference to imitate the consumption habits of other better-informed consumers, which are used as a proxy for content quality, an intermediary’s resources can be deployed to cultivate consumption by well-informed consumers. That in turn precipitates more widespread consumption among less well-informed consumers, and so on. The result is the familiar superstar effect in mass-cultural markets.

Copyright skeptics sometimes point to current star popular artists like Justin Bieber, who initially garnered attention through a YouTube video, or Adele, who is represented by an independent label, to “demonstrate” that the support of a major label is unnecessary. But even that anecdotal evidence is faulty: those artists were subsequently signed by, or indirectly contracted with, major record labels (Bieber by Universal, Adele by Sony), who provided the necessary marketing, distribution and tour support to catapult these artists to stardom. For multiple examples, see Levine (2011:51–53).

See supra note 47.

For other sources, see Gordon (2011:188) (interview with independent label executive) and 252 (interview with artist manager stating that a major is required to secure radio play) and 278 (interview with president of independent record label); King (2009:153) (stating that funding a commercial radio campaign can cost approximately $500,000 and is therefore beyond the budget of an individual artist).

This point is emphasized correctly by Lemley (2011). My argument departs from Prof. Lemley’s insofar as he takes the view that commercialization incentives are not a persuasive basis for intellectual property rights. The argument is that, assuming the underlying content has already been produced, conferring legal exclusivity on the commercializing entity constitutes an unjustified subsidy that is inconsistent with the free-market allocation of social resources. Gordon (2013) (and Justice Breyer in his dissent to Golan v. Holder (S. Ct. 2012)) makes a related point, arguing that there is no distinguishing characteristic of cultural distribution, as compared to other types of economic activity, that justifies providing legal exclusivity. There are three responses. First, the absence of legal or technological exclusivity would expose any mass-cultural distributor to a free-rider threat: third parties would allow the distributor to expend costs on supporting and marketing a large pool of unsuccessful releases until the occasional hit arises. Anticipating this outcome, the distributor declines to invest (or, as discussed subsequently, will be biased toward bundled forms of content distribution). Second, consistent with the fundamental link between property rights and market-based resource allocation, legal or technological exclusivity is required to make markets in trading and allocating content rights (including financing secured by those rights), which in turn drives creative properties toward the most efficient set of commercializing entities. Hence, providing cultural distributors with copyright protection is consistent with the market allocation of resources that is our default arrangement in all other areas of economic activity. Third, as I argue subsequently, even if there exists some revenue model under which production, distribution and marketing costs could be supported without legal exclusivity, there is no assurance that that model is the most efficient option. With robust (but waivable) copyright, the market can select freely among all feasible transactional choices. For further discussion of these points, see infra Part 6.2.1.

Note that, in the absence of tax-funded or philanthropic transfers, cultural markets almost never rely on entirely open content-delivery models (Barnett, 2010).

iTunes offers DRM-free downloads through its higher-priced “iTunes Plus” option. See “iTunes Stores: iTunes Plus Frequently Asked Questions (FAQ)”, avail. at http://support.apple.com/kb/ht1711.

For detailed description of these types of license terms, see Creative Commons, “About the licenses”, available at http://www.creativecommons.org.

Gordon (2013:10, n. 24) makes a related observation, noting that copyright discourages publishers from resorting to “expensive self-help options.”

It might be fairly objected that copyright itself limits the feasible range of production and distribution structures, as illustrated by injunctions against certain online file-sharing sites. However, it is almost certainly the case that it is less costly for the market to reduce the strength of copyright protections supplied by the state – either by explicit or by implicit waiver – than to create copyright equivalents where none exist. If that is true, then the set of transactional structures under copyright is larger relative to the set that would exist without it. To my knowledge, most examples of “privately created” intellectual property rights are either (i) norms-based property rights systems limited to a small number of homogenous repeat-play firms (such as the knowledge-sharing norms that prevailed among leading U.S. semiconductor firms during the decades following World War II (Teece, 2000, App. A); or, less commonly, (ii) elaborate contractual and policing mechanisms, which are in turn exposed to significant antitrust risk (such as the Fashion Originators’ Guild, a privately run system for policing fashion piracy, which was shut down by an antitrust prosecution, see Fashion Originators Guild of America v. Fed. Trade Cmm’n, 312 U.S. 457 (1941)).

As I have emphasized elsewhere, what matters is the incremental size of those premia and costs relative to an alternative state of affairs in which content holders migrate to some other instrument by which to regulate access and extract premia. There are multiple plausible scenarios under which that incremental value may be roughly neutral or even negative (Barnett, 2009).

The following evidence is merely suggestive. In a work in progress, additional evidence is being gathered and incorporated into a more complete study of the effects of digitization on the popular music industry.

It is sometimes argued that the vast majority of artists never received significant royalties on the sale of recorded music, and therefore always principally obtained revenues through live performance. Even if true, this ignores the fact that, even if the sale of recorded music did not directly benefit a performing artist through royalty payments, the intermediary supplied an up-front cash advance and engaged in promotional efforts to sell records, which indirectly promoted the sale of concert tickets. Without significant revenues from recorded-music sales, the incentive to make those promotional efforts disappears.

Author’s calculations based on data provided by the Recording Industry Association of America (for all recorded music sales) and Pollstar (for major concert ticket sales in North America).

Sources: Recording Industry Association of America (recorded music sales, all media); Pollstar (gross revenues for major concerts in North America). All figures adjusted for inflation using constant 2012 dollars.

For some of the reasons mentioned, above, Schultz (2009) and Day (2011) have similarly expressed skepticism with respect to the viability or optimality of a performance-based funding model for the music market.

Waldfogel (2012b) reports Nielsen data showing an increase in the number of albums released from 2000 through 2009 and a decline thereafter through 2010. Lunney (2012) reports no decline in the number of new artists among the “top 50” releases through 2010. Other data cited by Waldfogel (2012b) suggest that the increased number of releases is due to an increase in the number of albums that are either self-released (by the artist) or released by an independent label. There are three qualifications to this finding. First, for a large number of the releases, the entity type is unknown. Second, given contractual and ownership interests between major labels and certain “indies,” it is often unclear how to appropriately categorize an entity as an independent or a major. For further discussion, see Christman (2011). Third, an increase in the number of albums released does not necessarily translate into a welfare improvement, given the fact that average quality or “production values” may have declined as both barriers to entry into the distribution market, and expected maximal economic returns, have both fallen. Waldfogel (2012a) proposes quality metrics to address this difficult point. In an alternative approach, Lunney (2012) measures the originality of new releases by identifying the percentage constituted by “covers” of existing songs.

For more exhaustive analysis of this evidence, see Connolly and Krueger (2007, 2006).

Sources: Pollstar (top 100 grossing concerts in North America); Recording Industry Association of America. Figures for CD prices derived by author from data provided by Recording Industry Association of America. All figures adjusted for inflation using constant 2012 dollars.

All calculations based on information provided by Pollstar (for concert revenues) and the Oxford Music Online database (for age of lead performer and year of first album). In a minority of cases, the age of the lead performer or year of first album was confirmed through other sources. In tours involving two star performers, I selected the average of these values for each performer. In tours involving “music festivals” with large numbers of performers, I used the values for the performer most prominently associated with the festival (usually as indicated by the name of the festival). In the case of a handful of tours that were not primarily musical in nature (e.g., Cirque du Soleil), I omitted that tour and, for that year, added the next highest-ranked musical tour to the “top 20” sample. Connolly and Krueger (2006, Figs. 4.3c, 4.4) provide data on the distribution of ticket revenues among top performers for 1982–2003. Consistent with the tendencies identified above, they find a similar increasing skew of concert ticket revenues to leading performers starting in the 1980s. Further research is required to identify whether this skew effect was exacerbated following 1999 and whether it is best attributed to the rise of mass piracy, other factors or a combination thereof.

An important alternative account is that a significant portion of the concertgoing audience (the “baby boomer” generation) is aging and, as a result, artists that are popular with that segment enjoy an increasing share of total ticket sales. Note that this account can be reconciled with the explanation set forth above. If concert ticket prices are rising as artists seek to make up for revenue on recorded-music sales now lost to piracy, then younger budget-constrained fans will attend fewer concerts while older and less budget-constrained fans will attend more concerts, which in turn favors artists that are popular with the latter segment of the concertgoing public.

See supra notes 51–54 and accompanying notes.