This post is part of a series of evidence summaries for the 21 for 2021 project, a CREATe project within the AHRC Creative Industries Policy and Evidence Centre (PEC). The 21 for 2021 project offers a synthesis of empirical evidence catalogued on the Copyright Evidence Portal, answering 21 topical copyright questions for the 21st century. In this post, Paul Heald (Albert J. Harno & Edward W. Cleary Chair in Law at the University of Illinois College of the Law) reviews the empirical evidence relating to an optimal copyright term.
Under the Berne Convention, the minimum term of copyright protection is life-of-the-author plus 50 years. Many countries, including the US and the EU have terms of life-of-the-author plus 70 years, while others, such as Mexico, extend protection for life plus 100 years. For historical reasons, the term for pre-1978 US works is 95 years, so long as certain domestic formalities were adhered to.
The extreme length of the copyright term, as opposed to the standard length of patent or plant variety protection (both 20 years) has generated significant empirical inquiry and criticism. Most of the work has come from economists conducting cost-benefit analysis on the effect of long copyright terms. The primary economic justification for copyright assumes that legal protection is needed to incentivize the production of new works which would otherwise go uncreated. Secondary economic justifications include stimulating the creation of derivative works (like movies from books) and incentivizing the commercialization and distribution of works once they are created. Most empirical research queries whether current copyright terms are longer or shorter rather than optimal.
Note: This blog, and most the research cited, uses an economic definition of “optimal” as the length of legal protection that maximizes social welfare by balancing incentives to create against the deadweight costs associated with creation of exclusive rights (often loosely referred to as “monopoly costs”).
This blog will discuss the empirical research that addresses several key questions related to copyright term length: 1. Does the long term of copyright incentivize or deter new creations? 2. Does the long term of copyright increase or decrease access to works? 3. Does the long term of copyright incentivize or deter the creation of derivative works? 4. Are negative pricing effects associated with the long term of copyright?
Given almost universal international adherence to the Berne Convention, shortening the length of the copyright to less than life-of-the-author plus 50 is a political impossibility, and the empirical research discussed in this blog has not influenced legislators to shorten the term of copyright in any jurisdiction. However, the data collected bears on several questions that governments presently considering: 1. Should a legislature extend copyright protection beyond the Berne minimum of life-plus-fifty? 2. If a legislature extends protection beyond the Berne minimum, should it adopt a registration requirement? 3. Should a legislature grant authors a greater right to reacquire their copyrights back from publishers? 4. Should a legislature adopt the “rule of the shorter term” (which denies protection to foreign works that have fallen into the public domain in their home country)? 5. Should a legislature allow compulsory licensing of orphan works who have unfindable owners due, in part, to the long length of copyright?
Using a variety of different empirical methodologies, researchers have generally concluded that the current term of copyright is far too long and that legislatures should consider all Berne-compliant options to minimize the significant costs imposed by long copyright terms.
Debates and recent evolutions
The modern debate over the proper length for the copyright term has its origins with the passage of the 1998 Sonny Bono Copyright Term Extension Act in the US. This law extended the term of protection for post-1978 works from life-of-the-author plus 50 to life-plus-70, twenty years beyond the Berne Convention minimum, and added 20 years to the life of existing pre-1978 works. The EU had already extended protection to life-plus-70, but opposition was somewhat muted there due to a stronger artists rights tradition and a lack of reliance on economic and empirical methodologies at the time. The first shots in the debate were fired by seven Nobel Prize winning economists, Akerloff, et al (2002), in their submission to Congress as it considered term extension. Addressing prospective extension of the copyright term to life plus 70, the economists opined that no additional incentives to create would result from the legislation; therefore, the deadweight losses of extension surely outweighed the non-existent benefit.
As a legal challenge to the extension wound its way through the US judicial system, culminating in the landmark decision in Eldred v. Ashcroft (2003) upholding the constitutionality of the statute, defenders of the US extension offered several justifications, none of them based on data. The debate eventually focused on the economic wisdom of extending the term of existing works (because no plausible incentive-to-create rationale could be offered for works already created).
In the wake of the legislation and the court challenge, Landes & Posner (2003a) argued forcefully that an extended copyright term would maintain incentives for the commercialization and distribution of existing works. Prominent law professor Arthur Miller (p.14) asserted that incentives to produce derivative works are diminished when works fall into the public domain. Liebewitz & Margolis (2005) questioned, in response to Akerloff, whether copyrighted works would be “produced as regularly or as well as they would be” after their copyright term ends. Finally, anticipating the argument that public domain status means lower prices, Liebowitz (2008a) asserted that copyrighted and public domain books would cost the same.
Apart from Liebowitz’s pricing paper, support for the US term extension came in the form of economic intuition rather than empirical research, setting the stage for the hunt for data evidenced in the research discussed below.
The debate over the proper length of copyright exploded in the EU in 2005, as the EU commission proposed to extend the term of copyright protection for sound recordings in 2005 (CIPIL (2006)). Worldwide, the issue is continually raised in trade negotiations as the US trade representative insists that US trading partners move to life-plus-70 (both prospectively and retroactively). Notably, when the US withdrew from the Trans-Pacific Partnership negotiations, its provisions extending the copyright term in all members states were withdrawn.
Existing evidence and research agendas
Since measuring the effect of copyright term length on creativity is difficult, and because most economists conclude that the required life-plus-fifty is already sufficient to stimulate substantial creativity, scholarship has tended to focus on the relative costs and benefits of copyright protection.
A. Optimality. Direct evidence concerning the optimal length of copyright in economic terms is necessarily incomplete. For example, experts agree that the length or protection should vary by type of work (software v. movies) and the extent to which the sunk costs of creation can be recouped without a property right (for example, by using a paywall or by selling advertising). Certainly, a multi-million-dollar film needs a greater monetary incentive than the creation of a home video. For this reason, abstract discussions of term length have had little impact on policy. See Yuan (2006), Yoon (2002), and Rappaport (1998).
Nonetheless, the work of Pollock (2009), and others, is instructive. Developing an algorithm using standard assumptions as to discount rate and diminishing demand over time (“cultural decay”), Pollock concluded that the optimal term of copyright is around fifteen years and concluded with a 99.9% confidence rate that the optimal term is less than 50 years. Heald (2020) has recently updated Pollock’s insight in graphical terms by depicting the time necessary for a publisher to capture the present value of a copyright.
The graph shows that, assuming a 3% discount rate and depreciation rates suggested by NBER research (Soloveichik (2013) and 2014), publishers can expect to earn nearly 100% of expected revenue from a work in its first 30 years, suggesting that deadweight losses caused by longer protection would outweigh any additional creative incentives. Garcia and McCrary (2019) suggest an even shorter optimal term based on their study of the revenue streams of recorded music: 5-10 years. Garcia, Hicks, and McCrary (2020) find an even shorter economic life for blockbuster recordings. Ku, Sun, & Fan (2015), examining the correlation between copyright term extension and registration rates in the US, are unsure of term length optimality. For an early and highly sceptical discussion of the proper copyright term, see Breyer (1970).
Giorcelli and Moser (2015) concluded in a narrow study of Italians operas from 1770-1990 that an increase in term length may have been a factor in stimulating the production of more operas, but they do not opine on the term length optimality.
B. Access to Existing Works. Early defenders of copyright term extensions argued that legal protection had benefits beyond providing creative incentives. They suggested that copyright protection should increase the distribution and commercialization of creative works. Virtually all subsequent empirical work suggests instead that copyright reduces public access to works. The chart below from Heald (2014a) shows an increase in the availability of books as they fall into the public domain, and a substantial suppression of availability of books still under copyright original published in the 1930s-19990s
Similarly, Heald (2020a) found that Canadian novels were more likely to be in print after they fell into the public domain. Imke Reimers (2019) also studied books and confirmed that the expiration of copyright status increased the availability of new editions. Studying ebooks in Australian, New Zealand, US, and Canadian public libraries, Flynn, Giblin, and Petitjean (2019) found that availability increased after the copyright term for a title expired. Cuntz (2020) found changes in copyright status significantly increase the number of total performances individual operatic works received on stage once a copyright expired. Macgarvie, McKeon, & Watson (2018) found that expiration of copyright for EU sound recordings from the 1950s and early 1960s was associated with a 141-247% increase in re-releases. Biasi & Moser (2018) found that the nullification of German copyrights during WWII resulted in a significant increase in the availability of various scientific publications. Erickson, Heald, & Kretschmer (2015) found that author pages on Wikipedia were more likely to have images when photographic copyrights had expired. Heald (2008a) found that popular US songs falling into the public domain were not less likely to be exploited in movies than newer copyrighted songs. This finding was confirmed in Heald, et al (2012). Heald (2007) found that bestselling books published from 1913-22 were more likely to be in print than their copyrighted bestselling counterparts published from 1923-32.
C. Effect on Derivative Works. Less research has been done on the effect of copyright term length on incentives to create derivative works. In theory, the owner of a copyrighted work might have a greater incentive to engage in follow-on creation than another party mining the public domain. Two recent studies have found that the expiration of the copyright term in a work leads to increased creation of derivative versions. Flynn, Giblin, & Petitjean (2019) found that ebooks were more likely to be made from public domain titles, concluding that “exclusive rights do not appear to trigger investment in works that have low commercial demand.” Buccafusco & Heald (2012) compared US bestselling novels from 1913-22 (all in the public domain) with bestselling novels from 1923-32 (all under copyright) and found that a significantly higher percentage of the public domain novels had been made into audiobooks and ebooks.
D. Price Effects. An early paper by Liebowitz (2008a) suggested that copyright term length might have no effect on the price of copyrighted books, but several subsequent papers have shown significant discrepancies between the prices of copyrighted and public domain works. For example, Heald (2020a) found that the price of Canadian bound and ebook editions dropped significantly after their copyright terms expired. This was consistent with his prior research on book pricing in South Africa and the US: Heald (2020b) & Heald (2007). Flynn, Giblin, and Petitjohn (2019) also found libraries paid a higher price for access to copyrighted ebooks in Australia, New Zealand, Canada, and the US. In an historical study, Li, Macgarvie, & Moser (2015) found a fifty percent increase in book prices in the UK associated with the doubling of the copyright term in 1814.
E. Orphan Works. The longer the term of copyright, the more difficulty potential licensees have in finding copyright owners. Several jurisdictions have responded by offering compulsory licences when a reasonable search fails to identify an owner. One important exploration of this topic is Favale, Homberg, Kretschmer, Mendis & Secchi (2013) which compares various national schemes and conducts a simulated rights clearance exercise.
Future directions for research
Although the universal adherence to the Berne Convention and its life-plus-50 minimum term discourages research on optimal term lengths for various types of copyrighted works, the examination of the effect of long copyright terms remains salient to several different policy discussions. The following issues are likely to provide directions for future research.
First, the US is particularly aggressive in trying to convince its trading partners to extend copyright protection beyond the Berne minimum, to life-plus-70. Existing research in the US, Canada, and South Africa suggests that extension may affect availability of works, the development of derivative works, and prices. Data from as-yet-unstudied jurisdictions would certainly be relevant in the course of future negotiations.
Second, some countries, like Canada, have already agreed to extend the copyright term to life-plus-70 but are considered reducing potential costs by requiring copyright owners to register their works. Such a requirement might address the orphan works problem and increase the number of (unregistered) works entering the public domain. Very little research has been done on the interaction between term length and registration requirements.
Third, research by Heald (2017) suggests that US authors are sometimes better than their initial publishers at keeping works in-print and available to the public. Countries like Canada and South Africa are currently considering giving authors an inalienable right to re-obtain their copyrights at year 25 after a transfer. Research examining what authors’ do with their copyrights after reversion may help policy makers craft a more efficient balance of rights.
Fourth, some countries, like the US, apply their term of copyright to all works, even those which are no longer protected in their home countries due to expiration of a shorter copyright term. Under the Berne Convention, a member state is not obligated to apply its copyright term so generously. The sort of research discussed above that bears generally on the cost/benefit analysis of copyright protection should inform a jurisdiction’s decision to adopt the “rule of the shorter term.”
Fifth, very little research has yet been done to quantify one of the most significant costs of long copyright terms–the proliferation of orphan works. Clearly, the harder to find a copyright owner, the higher the transaction costs involved in obtaining a license. Empirical data is lacking on the extent of the orphan works problem and its economic effects.
Finally, most of the studies cited above have been conducted in US, UK, and Canadian book and music markets. Copyright encompasses software, architectural works, film, sculpture, drama, and dance. Are existing studies applicable to other disciplines? Are they applicable to other countries with different legal systems and borrowing norms? Expanding existing methodologies to diverse media and cultures (Asia would be a good place to start) creates exciting opportunities for future research.