by Dr George R Barker
The Productivity Commission (the Commission) recently released a report recommending radical changes to copyright law that would substantially weaken copyright, claiming that “Australia’s copyright system has expanded over time, often with no transparent, evidence-based policy analysis demonstrating the need for, or quantum of, new rights” (pg. 29). The problem, however, is that the Commission is wrong to claim that there has been a strengthening of copyright, and that there is no evidence-based policy analysis available to support such a strengthening of copyright (were it to have occurred). Indeed, the Commission is instead guilty of the exact behavior it is complaining about: namely making radical policy recommendations to weaken copyright that have no evidence basis at all.
Some stakeholders, large companies and private advocacy groups, however, are currently engaged in a concerted effort to obtain a major weakening of copyright. They have promoted faulty and misleading interpretations of data to support their claims.
It seems the Commission in conducting its current inquiry into Australia's intellectual property arrangements may have been unduly influenced by the sheer weight of current anti-copyright dogma. As a result the Commission has made some radical recommendations, including the adoption of a broad US-style fair use exception to copyright, which would limit its coverage, and a radical limitation on the term of copyright, which would limit its duration.
The PC took submissions on its issues paper in November, and released its current draft based on the submissions at the end of April 2016. Google’s submission to the Commission inquiry in November (the Submission) specifically cited five reports to support the weakening of copyright now proposed by the Commission. The studies cited in the Submission have, however, been discredited, containing fundamental errors in empirical research, making them unfit for policy-making1. Moreover, the evidence in the studies appears to contradict the claims made in the Submission. In particular:
However, the studies’ authors fail to draw attention to these results, which if they had done so, would undermine their conclusions. Generally, as outlined below the claims from the five studies referred to in the Submission are unreliable.
The key claim from the 2012 Singapore fair use study, authored by Ghafele and Gibert, is that Singapore’s introduction of an open ended fair dealing clause2 that closely resembled the US fair use law provided a significant stimulation of growth in “private copying technology industries,” generating an additional 11% growth in total value-add in private copying. This claim is incorrect and misleading for at least three reasons, which, along with other errors, render the claim from the Singapore Study irrelevant to policy making: First, 95% of Singapore’s output of private copying technology is exported and could not have been affected by the domestic legal change, and the type of growth that would have been necessary within the domestic 5% is clearly impossible. Second, the study’s aggregation of a so-called private copying industry is the product of subjectively aggregating a number of diverse industries that have only an incidental relation to private copying.3 Finally, as Dr. George S. Ford of The Phoenix Center demonstrated, the authors of the study violated “the most basic and necessary assumption” of the “difference in differences” methodology that they used and made numerous other methodological errors, causing him to assess their study as “junk science.”4 Ford was unable to replicate their results even after correcting a number of key errors in their methodology.
The Australian study authored by Lateral Economics for the Australian Digital Alliance focuses on two main claims. First, the study claims that industries that rely on copyright exceptions and limitations contribute more to the Australian economy ($182 billion or 14% of GDP) than those that rely on copyright ($98 billion or 10% of GDP). Second, the study offers a “very conservative estimate” that if US-style fair use copyright exceptions were introduced to Australia the “value added or the welfare gain to the Australian economy would be $593 million higher” per annum after 10 years. We prove that both these claims are unfounded, and contradicted by Lateral Economics’ own data presented in the report.
This report and earlier reviews show that their industry groupings are redundant and not relevant to the key question, and that the claimed increase in real growth by 1% (or around $600 million) from introducing flexible fair use policies over three years is groundless. Lateral Economics’ own data shows that fair use exceptions in the US are associated with a lower rate of growth of value-add in what it calls copyright exceptions industries in the US, compared to the same industries in Australia. In addition, the evidence from the Singapore study shows that US-style fair use exceptions was associated with a large decline in the growth rate of the core copyright industries.The Gibert Lisbon Council Study
Google claims the paper authored by Gibert for the Lisbon Council found that copyright exceptions enhance economic growth, jobs and prosperity. However, due to numerous fundamental errors detailed in this paper and by other reviewers,5 the study cannot be relied upon. The Lisbon paper examines correlations in eight countries between selected economic outcomes and an index which the authors claim measures the scope and flexibility of exceptions to exclusive rights (SFEER Index) for each country. However, errors in the economic data used, make the statistical tests meaningless. The source economic data used, for example, was in nominal terms, and expressed in five different currencies. Consequently irrelevant variations in the nominal values, and in cross-country exchange rates over time make it impossible to draw any conclusions. Moreover the SFEER Index is based on the subjective and incomplete review of a limited number of the relevant features of copyright exceptions in the eight countries. As a result, the index cannot be replicated, which implies that the study is unscientific. The correlations the Lisbon study reports are further based on at most eight and as few as five observations. This is too few to generate reliable statistical results. The Lisbon Council’s conclusions are moreover based on the statistical significance of less than 5% of the 462 tests run, suggesting “cherry picking of results,” with the small number of “successes” in turn simply explained by random variation.
The Sky is Rising Report, authored by Mike Masnick (of the highly anti-copyright blog, Techdirt), for the Computer and Communications Industry Association (CCIA) states that “there has been an explosion in creative output over the past couple of decades.” While technological advances have, in fact, facilitated measuring an increase in the creation of creative content, this fact does not demonstrate a case for weakening copyright, without a clear method for separating the effects of digitization and the effects of copyright protection.
Further, the evidence cited in the Masnick report is misleading. For example, it claims “Americans have greater access to music than ever before in history, and in 2006, it was estimated that consumers downloaded about one billion tracks per month.”6 But the report fails to mention that much of that downloading was unpaid, or downloaded illegally, and that these illegal downloads have been shown to displace legitimate sales. The report incorrectly claims that music sales have hit “record highs in 2011 and 2012.” This is inaccurate. Music sales revenue (including both physical sales and digital downloads) has decreased by more than 40% from 1999 to 2013 both in the US and internationally (inflation adjusted). Similarly for movies, the report argues that “box office remains the largest revenue driver. Sales of physical home entertainment media (DVD, Blu-ray) will continue to decline, and will be the only category of video entertainment that does so.” However, the report fails to note that the declining category (DVD, Blu-ray) was previously the main source of revenues, and the most affected by piracy, being more vulnerable to copying than the more experiential movie theater viewing. For television, the report argues that average TV viewing amounts to approximately nine hours per day, an obvious error. Actually these nine hours include listening to radio and other types of media uses.7 More generally the report simply fails to mention the well documented evidence that piracy harms media sales, and the growing evidence on consequent falls in investment and employment in creative sectors as a result.
Finally, A Law and Economics Analysis of Fair Use Differences Comparing the US and UK by Antony Dnes from 20118 is cited by Google for the proposition that fair use appears to allow “innovations to emerge rapidly in the US,” “of high value to consumers” “without generating significant damage to copyright holders.” Upon review, however, the Dnes study is not an empirical study and provides no evidence for these propositions. Rather it is a review of the legal doctrines in the US and the UK. In fact as Dnes notes,
“No-one quite knows whether the US fair-use and UK fair-dealing approaches are truly distinct, or what the full economic significance is of any distinctions; these issues are assessed here through an analysis of laws and cases.” Dnes Study P7
Also, although Google correctly points out that the Dnes report was prepared for the UK’s Hargreaves Review of Intellectual Property and Growth, it is important to point out that in his final report Hargreaves does not explicitly rely on, or cite the Dnes report for any key propositions, although it was included in his reference list. Hargreaves instead concluded fair use was not a good idea for the UK and that fair use's benefits were overstated.9
For all of the above reasons, discussed in more detail in my full report, the argument advanced that broader copyright exceptions will promote productivity and economic growth is not based on sound research. The evidence base for what the Commission is proposing is in short non-existent and its recommendations need to be either revisited by the Commission - or rejected outright by the Government in the interest of all Australians.
The Commission has sought further information and feedback on its recommendations and one can only hope it decides to revisit its position.
Dr George Barker is Director of the Centre for Law and Economics at the Australian National University (ANU), a Visiting Fellow at the London School of Economics and a Visiting Fellow of the British Institute of International and Comparative Law London. He has testified in Europe, North America and the Asia Pacific before ministers, courts and regulatory agencies on intellectual property law, competition law, market design, regulatory policy, and on corporate, contract and tort law liability and damages.
Read the complete research paper Claims to Expand Copyright Exceptions Driven by “Bad Science” here.
1. See Dr George S. Ford (2016) “The Economic Impact of Expanding Fair Use in Singapore: More Junk Science for Copyright Reform” Perspectives: Phoenix Centre for Advanced Legal and Economic Public Policy Studies http://phoenix-center.org/perspectives/Perspective16-01Final.pdf; Dr. George S. Ford (2015) The Lisbon Council’s 2015 Intellectual Property and Economic Growth Index: A Showcase of Methodological Blunder, Phoenix Center http://phoenix-center.org/perspectives/Perspective15-03Final.pdf. George Barker & Ivan Png 2013, Unfair Evidence on Fair Use, 3June, available at https://law.anu.edu.au/news/cle/unfair- evidence-fair-use; George R Barker 2013, ‘Agreed Use and Fair Use: The Economic Effects of Fair Use and Other Copyright Exceptions’ paper presented to the 2013 Annual Congress of the Society for Economic Research on Copyright Issues (SERCI), MINES ParisTech, Paris (France)
2. Clause III.35. Fair dealing in relation to works was adopted as an amendment to The Singapore Copyright Act (Ch. 63) - Act 2 of 1987 on 1 January 2005. Ghafele and Gibert claim that clause III.35 “strongly echoes §107. Limitations on exclusive rights: fair use of the Copyright Law of the USA. The clause outlines the various factors courts must consider when making a retroactive judgment about fair use… The language of the clause embeds a degree of flexibility into the law in terms of determining whether or not a particular use of a copyrighted work is fair, and thus non-infringing. It is in stark contrast to alternative systems, such as those found in Europe, that specifically define what constitutes a non-infringing use of a copyrighted work by exhaustively listing the activities that are exempt from infringement liability. Exhaustive lists denote a very strong, but limited, boundary for what constitutes a non-infringing use.” Ibid pp3-4
3. George R. Barker (2013)
4. George S Ford (2016) pp 6, 12 and 13
5. George S. Ford (2015)
6. Masnick cites “Hiatt, Brian and Serpick, Evan. (2007). The Recording Industry’s Decline. Rolling Stone Magazine” for this claim but this article is not available. Instead one can find Brian Hiatt and Evan Serpick “Music Biz Laments Worst Year Ever” Rolling Stone January 13, 2006 which notes “sales of digital tracks jumped from 141 million in 2004 to 353 million in 2005, and sales of digital albums rose from 5.5 million to 16.2 million” this indicates Masnick’s billion per month number must include illegal downloads see http://www.rollingstone.com/music/news/music-biz-laments-worst-year-ever-20060113#ixzz46Fy0LZ2D
8. A Law and Economics Analysis of Fair Use Differences Comparing the US and UK, Antony W. Dnes, 2011 http://webarchive.nationalarchives.gov.uk/20140603093549/ http://www.ipo.gov.uk/ipreviewdocj.pdf
9. Hargreaves page 45 paragraph 5.16