Professor Paul Heald on Copyright Extension and Early 20th-Century Music Compositions
In 1988, works from the early 20th century began entering the public domain, a process that continued until Congress passed legislation extending copyrights in 1998. Were the musical compositions "freed" during this window performed any more (or less) often as a result?
At this Thursday's Faculty "Works in Progress" (WIP) talk, Professor Paul Heald (JD '88 and currently visiting from the University of Georgia) presented his most recent work on "Testing the Over- and Under-Exploitation Hypotheses: Bestselling Musical Compositions (1913-32) and Their Use in Cinema (1968-2007)." Regular readers of the blog will be somewhat familiar with Prof. Heald's work through his recent posts here but, as this paper's title clearly indicates, it deals with intellectual property, his area of greatest expertise.
In the paper and accompanying empirical study, Prof. Heald looked at a group of the most valuable musical compositions made between 1913 and 1932, and the frequency with which those compositions appeared in films released after 1968. This data set provides a interesting natural experiment because of quirks in the history of US copyright law: works made in the 1913-1922 period are now in the public domain (1913 works fell into the public domain in 1988, after 75 years), while works made between 1922 and 1932, if their copyrights were renewed, are protected by the 1998 "Sonny Bono" Copyright Term Extension Act (CTEA).
By looking at the relative use of these two classes of (arguably) otherwise-similar works over time in films, Prof. Heald hoped to find evidence for or against two arguments for copyright: that it prevents under-exploitation or over-exploitation. Briefly, the under-exploitation story says that works that fall into the public domain, though still valuable, will be less available than they would be if they were still protected by copyright, since they have no owner interested in their promotion or maintenance. For example, arguments were made before Congress in the debate surrounding the CTEA that old films would deteriorate without owners interested in protecting them.
The overexploitation argument, on the other hand, posits that public domain works may become either "worn out" from overuse or be subjected to "debasing" uses that "recode" the meaning of the work. The most frequent example given in the literature seems to be hypothetical pornographic works featuring Disney characters, availability of which might damage or destroy the value of those characters. One audience member mentioned another, somewhat less bizarre example: restrictions placed by Ira Gershwin and his heirs on performances of his opera Porgy and Bess (requiring the work be performed with black actors in the lead roles), which would be impossible without copyright protections.
To briefly summarize the results of Prof. Heald's work, he found
that compositions did appear in movies at a greater rate as they
transitioned to the public domain. However, compositions that remained
under copyright (post-1922 works) also appeared more often over the
same time period, a result Heald ascribes to either increasing general
popularity of works from the same early 20th-century time period, or
characteristics of the IMDB data
used for the study (it is possible that it is more comprehensive for
more recent films). This larger trend, Prof. Heald argues, subsumes any
increase in exploitation of works as they fell into the public domain.
In other words, works do not appear to become unavailable "orphans" as
their copyright expires, undermining the case for
Further, while works are more widely available after their copyright
expires, the difference is not significant, similarly undermining the
parallel case for over-exploitation. In short, copyright extension is
doing little to protect these works or make them more widely available
(though, as Prof. Heald mentioned, his study does not address whether
copyright was necessary to incentivize their creation in the first
The audience of faculty members offered a lot of comments and constructive criticism. The most interesting contributions focused on two possible limitations of the study.
One commenter noted that licensing of songs for use in movies is somewhat more complex in practice than simply obtaining a license for the composition, in ways that might obscure the results of the study. Frequently, producers also want rights to a specific recording of the song (for example, a movie producer might want not only to include "I've Got the World on a String" in the picture, but specifically the 1952 Frank Sinatra recording - nothing else will do). Further, the recording will often be protected by copyright even if the original composition is not. This could obscure any effects arising from the public domain status of the composition. The commenter argued - and Prof. Heald broadly agreed - that a study that analyzed works differently based on whether the composition itself, chosen sound recording, both, or neither were in the public domain would be an improvement. Limitations in the data set used and the fact that other data sources are not readily available would make this difficult, however.
Another commenter noted that since the two groups of works studied (those on either side of the 1922 CTEA divide) were otherwise similar, substitution between them might cause their prices and rates of exploitation to converge, regardless of copyright status. Copyrighted works for which other works that have recently entered the public domain are good substitutes would become cheaper, and in general we should expect to see them appear in films at roughly the same rate regardless of copyright status. Since this is, broadly, what we see in Prof. Heald's results, the commenter noted that it might be that the data simply show substitution effects (plus some statistical noise), rather than anything concrete about the effects of copyright term extension. This could be seen as a form of selection bias, or simply a limitation of the study methodology.
This criticism seemed both superficially compelling and difficult to dismiss. Prof. Heald's best response was to argue that the public domain and copyrighted compositions studied were not generally good substitutes for each other, since (as illustrated in the example above), those making films often have a specific piece of music in mind. The availability of another (cheaper) public domain work, Prof. Heald speculated, probably has little effect on demand for each copyrighted work. While this, if true, would certainly deflect the commenter's challenge, it is possible that it proves too much. If the compositions are so individualized that they are never good substitutes, with each as a "market of one," it seems like the value of copyrighted works as a control group would be diminished. Whether different songs are good substitutes for each other or not also remains an open empirical question, the answer to which seems critically important to the study and its results.
Despite these concerns, the faculty seemed very interested in Prof. Heald's work, and eager to offer their comments. I share their enthusiasm, and look forward to the final version of the paper. Personally, I wish research like this and Prof. Heald's earlier work on books had been available when the CTEA was considered in 1998, and I hope it will provide some impetus to reexamine that policy.