Hot on the heels of the UK House of Lords’ inquiry into Open Access, the Business, Innovation and Skills Committee of the House of Commons has begun its own inquiry. This morning I submitted my own evidence. Here it is.

[It’s not too late to make your own submission. It doesn’t have to be as long as this: just let the government know your attitude regarding the parts of the question that concern you most. Again, details are here. Get writing!]

—

Business, Innovation and Skills Committee

Inquiry into the Government’s Open Access Policy

Dr. Michael P. Taylor

7th February 2013

I am an honorary research associate in the Department of Earth Sciences at the University of Bristol. A computer programmer by profession, I completed a Ph.D in my spare time in 2009, and continue to publish research in vertebrate palaeontology. Although my Bristol affiliation gives me reasonable access to paywalled literature, the mechanics of access are cumbersome as I work exclusively off-campus. My unusual circumstances give me a perspective from both inside and outside academia.

I speak only for myself, not for the University of Bristol.

Executive summary

1. The UK currently leads the world in Open Access, and as a result enjoys a significant citation advantage; the rest of the world is following.
2. The purpose of funding research is to benefit society as a whole. When publishers’ interests conflict with those of broader society, the government must serve society.
3. The inclusion of publishers on the Finch Committee represents a conflict of interest. Some recommendations of the Finch Report were compromised by this conflict of interest.
4. Gold OA has a real advantages over Green OA in that it presents a single Version Of Record. But the Finch Report does not give proper weight to Green OA.
5. The form of Green OA mandated by the RCUK policy is badly compromised in allowing restrictive licences and delayed access that reduce the value of the research.
6. There is no evidence that Green OA negatively affects subscription revenue.
7. CC BY allows commercial re-use by design, in order to obtain the best return on research investment. It explicitly prohibits plagiarism. It does not infringe author’s rights, in fact allowing authors to retain more rights than the prevailing copyright-transfer regime.
8. The use of CC BY means that research does not merely allow us to know more, but to do more. It is not anti-commercial, it is anti-monopoly (and so facilitates an efficient market).
9. Gold OA APCs vary greatly between publishers: traditional publishers are far more expensive than new OA-only publishers. Much lower APCs than the £1500–£2000 quoted by the Finch Report are achievable with no loss of publication quality. Government funding should cover only a base APC of perhaps £1000 to encourage downward pressure on prices.
10. The government must make decisions on the basis of what benefits the UK as a whole, not what benefits any single industry. The government should allow both Gold and Green OA; should require the CC BY licence, whichever route is taken; should tolerate no embargo on Green OA; and should not fully fund exploitatively high APCs.

Principles

1. First, I enthusiastically welcome the government’s clearly stated commitment to Open Access (OA). There is no question that the free availability of research outputs will have a significant positive effect – not only by accelerating further research, but also in the form of practical improvements in health, education and industry.

2. I am also delighted to see the UK leading the world in Open Access. RCUK’s pre-Finch OA policy was one of the world’s earliest and most significant; the Finch Report established a clarity of vision not previously seen in any national OA policy; and BIS’s public commitment to OA in all government-funded work was a world first. As noted below, there are strong pragmatic reasons for the UK to maintain a position of leadership. But even leaving these aside, the symbolic value of leadership in open access can hardly be overstated.

3. Equally, it has been hugely encouraging to see the world following the UK’s lead. In the days after the announcement of the UK policy, similar declarations followed from several European countries; and most importantly, the €80M “Horizon 2020” programme of the European Commission also announced an Open Access policy. [Note added 18 May 2015: Eek! I only just noticed my error here: Horizon 2020 is a €80 billion program, so I dropped three orders of magnitude.] Global research is not a zero-sum game: the UK’s gain is not other countries’ loss or vice versa. As more countries open up their research outputs, the whole world benefits.

4. Against this backdrop, implementation issues must be discussed with a simple but important principle in mind: who is publicly-funded research for? When the question is stated explicitly, the answer is immediately obvious: it is for the public that funds it – for the citizens whose health, education and economic prospects are all improved by Open Access.

5. Unfortunately, the implementation strategies recommended by the Finch Report are not those that would most benefit the public, but are slanted towards the interests of academic publishers. This is because, as noted on page 113 of the Finch Report, three of the fourteen working members of the group represent publishers. The involvement of publishers in deciding the UK’s publishing policy is mystifying, as it represents a clear conflict of interest.

6. There is no question that in the process of research, publishers provide important services; but so do the providers of IT infrastructure and manufacturers of laboratory equipment, and their input was not sought in formulating policy. Neither should publishers have been consulted. The proper approach would have been for researchers, librarians, university administrators, funders, medics, educators and businesses to have worked out what policy and what strategy would best serve them in their goal of performing, disseminating and exploiting research; and then to negotiate with publishers (as with IT service providers and lab-equipment manufacturers) to obtain the necessary services at the best prices.

7. It is because of publisher involvement in the Finch Committee that the recommendations of the Report are skewed towards the interests of that one small group at the expense of citizens. All of the flaws in the recommendations in the report are directly attributable to this.

Specific issues

The Government’s acceptance of the recommendations of the Finch Group Report ‘Accessibility, sustainability, excellence: how to expand access to research publications’, including its preference for the ‘gold’ over the ‘green’ open access model

8. There are good and legitimate reasons to prefer Gold over Green. Most importantly, there is no possibility of confusion over which is the version of record under the Gold model. However, the Finch Report goes much further than expressing a preference for Gold, by almost entirely omitting Green from its discussion. (The word “green” appears only three times in the 140 pages of the report and one of these is a reference to the HM Treasury Green Book.)

9. Perhaps as a result of this, the Green OA provisions of the revised RCUK policy are much weaker than its Gold OA provisions. In particular, while RCUK-funded work that is published as Gold OA must use the very permissive Creative Commons Attribution licence (CC BY), whereas work published as Green OA may have restrictive non-commercial clauses inserted. The inclusion of non-commercial clauses greatly diminishes the value of work, as discussed in the next section.

10. The Finch Report also allows embargoes – i.e. delays in availability – when research is published as Green OA, and the RCUK policy follows this (although it shortens the delays). Allowing such delays necessarily entails delaying all the benefits of Open Access, and thereby retards the progress of research, medicine, education and industry. There is no justification for these delays: even if they benefit publishers by enabling them to avoid subscription cancellations, the interests of publishers must be outweighed by those of all the other stakeholders.

11. In any case, counter-intuitively there is no evidence that Green OA hurts subscription revenue at all. The JISC/European-funded PEER project¹, after nearly four years’ work, concluded that “there is no evidence that self-archiving has harmful effects on journal viability”. This is the only large-scale analysis of this issue to have been undertaken, and the only solid data we have to go on: publishers’ statements about effects of Green OA on subscription revenue are guesses, not informed by data.

12. In conclusion, the full benefit of Green OA will be realised only if Green articles are licenced using CC BY and if they are made available from the date of publication rather than after an embargo. If these changes have a negative effect on publishers (which all the evidence says they will not), then that is regrettable; but it would not be a reason to delay access to, and reduce the utility of, publicly funded research.

Rights of use and re-use in relation to open access research publications, including the implications of Creative Commons ‘CC-BY’ licences

13. The CC BY licence embodies the original definition of the term “Open Access” by the Budapest Open Access Initiative (BOAI)², and has been widely adopted as the open-access licence of choice by respected OA publishers such as BioMed Central (BMC) and the Public Library of Science (PLOS). It ensures that the published works can be used by the widest spectrum of organisations – not only to facilitate further pure research, but also to be used in education and to catalyse innovation in industry

14. Recently some regrettable misunderstandings of the CC BY licence have been promulgated, notably in an anti-Finch letter from editors of 21 history journals³. This claimed that the use of CC BY “means that commercial re-use, plagiarism, and republication of an author’s work will be possible, subject to the author being ‘credited’ (but it is not clear in what way they would be credited). We believe that this is a serious infringement of intellectual property rights and we do not want our authors to have to sign away their rights in order to publish with us.” I will address these misunderstandings in turn.

15. “commercial re-use”: yes, by design the CC BY licence makes this possible. Contrary to the assumption of the history-journal editors, this is not a bug but a feature. The goal of the UK government policy is to benefit the UK in general, including its many commercial concerns.

16. “plagiarism”: it is flatly wrong to say that CC BY encourages plagiarism. Plagiarism is the use of another person’s work without acknowledgement; but the CC BY licence explicitly does require acknowledgement. It is not possible to plagiarise and be within the terms of this licence.

17. “republication of an author’s work”: yes, the CC BY licence does allow this. So do the agreements currently in use by most publishers, which allow the publisher to anthologise and otherwise reuse author’s works with or without their consent. Again, this is a good thing: it enables the full value of the work to be realised, whether in an educational or commercial setting.

18. “it is not clear in what way they would be credited”: on the contrary, the text of the CC BY licence⁴ plainly states “You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work)”.

19. “serious infringement of intellectual property rights”: quite the reverse: the CC BY licence codifies the standard scholarly practice whereby authors’ works are used in various ways with full attribution but without further financial compensation. (After all, scholars have already been paid for their work.)

20. “we do not want our authors to have to sign away their rights in order to publish with us”: this statement betrays a surprising ignorance of standard journal requirements, including those of many of the journals whose editors signed the letter. These journals require the author to sign copyright over to the publisher, a far greater violation of intellectual property rights than in the open-access model.

21. The goal of the CC BY licence is to allow the full value of a work to be realised. By contrast, the non-commercial (NC) clause that the current RCUK policy allows in the case of Green OA outright prohibits use of the work in many important situations – e.g. in health, education and commercial enterprises. Worse still, the uncertainty introduced by a non-commercial clause has a chilling effect, stifling uses that might be allowed but which the innovator does not want to risk having to defend in court. There is no justification for a ban on commercial use of research publications. We fund research not only in order to know more, but in order to do more.

22. A particularly important application of Open Access articles is that massive numbers of articles can be “mined” by computer programs far more efficiently than any human can read them. Such techniques are already in use in a limited way and proving their value: for example, in discovering and synthesising new chemical formulae⁵. However, such work is greatly impeded by the difficulty of obtaining permission from publishers, and assurances that they will not claim ownership of the results. The use of a permissive licence such as CC BY disposes of such concerns and encourages innovative uses of research data that we cannot yet envisage.

23. Eric Kansa, of the University of California at Berkeley, summed up the misunderstandings and realities of fully open licences such as CC BY very efficiently: “While Open Access is not anti-commercial, it is anti-monopoly”. Monopolies lead to inefficient markets and poor economic utilisation. Permissive licences on publicly-funded work fix that.

24. In conclusion: for the benefit of the country as a whole, the government should renew its commitment to the use of the CC BY licence on all the Gold OA research that it funds, and extend this to Green OA publications as well.

The costs of article processing charges (APCs) and the implications for research funding and for the taxpayer

25. Simple calculations⁶ conservatively estimate the overall cost to the worldwide academic community of an average paywalled article at £3307 (£4.96 billion total paywall revenue each year for 1.5 million articles). Naively, an APC needs only be cheaper than this to achieve a net economic improvement. In practice, APCs need to be significantly cheaper because of the difficulty of making a transition to Gold OA while subscriptions are still being paid.

26. At present, APCs for Gold OA are under strong price pressure. I will review some key data points as they were one year ago, then consider recent developments.

• Typical APCs from traditional publishers (Elsevier “sponsored article”, Springer Open Choice) are in the region of $3000 (about £1950)
• The open-access megajournal PLOS ONE charges $1350 (about £862)
• The broad survey of Solomon and Björk (2012) found an average Gold OA fee of $906 (about £579).

27. In the last year, significant developments have changed the landscape. In roughly chronological order:

• The new high-impact journal eLife, sponsored by the Howard Hughes Medical Institute, the Max Planck Society and the Wellcome Trust, is initially waiving all APCs.
• The new for-profit open-access megajournal PeerJ introduces a new model where each author pays $99 (about £63) for a lifetime membership, which covers one publication each year at no further charge. (A $299 (~£189) lifetime membership covers unlimited publications.)
• The new Forum of Mathematics journals will waive all APCs for the first three years, after which they may charge £500 or may find a different revenue stream and retain a zero APC.
• The open-access humanities megajournal SAGE Open recently dropped its APC to $99.

28. Other new open-access ventures continue to emerge, including the Episciences initiative to create overlay journals for arXiv, and the Open Library of Humanities.

29. As low as these prices are, a surprising finding by Stuart Shieber of Harvard is that 70% of all open-access journals charge no APC at all⁷. Others have independently found similar results.

30. In the light of all this, it is surprising that the Finch Report suggests a typical APC of “£1500-£2000”. The only explanation for this is that the bulk of the APCs in the sample that this range was calculated from was biased by many high-priced papers in traditional publishers’ journals. If this is so, then a cultural change is required in researchers. There is no reason why the majority of papers should not be published in outlets with much lower APCs. (In most cases the new journals also offer other benefits, such as unlimited colour figures and supplementary information, and the ability to include audio and video in the published work.)

31. A shift away from traditional journals to newer “born open-access” journals is likely to be accelerated if there is some pressure on researchers to overcome their innate conservatism and choose venues accordingly. Researchers may find it more convenient to keep submitting to the same journals they have used previously, publishing their work as Gold OA with the APCs covered by funders. But this convenience, while attractive in the short term, will inhibit long-term change.

32. For this reason, I would support a cap on the APCs that will be paid by funders: for example, RCUK could elect to fund the first £1000 of each APC, leaving authors who insist on using more expensive journals to top up the fee from their own grants. The resulting downward pressure on prices would ensure that a true market in Gold OA provision emerges.

33. In some quarters, such a scheme might be described as a curtailment of academic freedom. It is not. The phrase “academic freedom” refers to the freedom to choose what to study and what opinions to express – not what venues to publish the results in. A researcher who freely elects what to work on is just as free, whatever journal his paper appears in.

34. In conclusion, government funding should cover the costs of modest APCs; but there is no need for them to continue to support very expensive journals as far cheaper alternatives are now becoming available at no loss of quality.

The level of ‘gold’ open access uptake in the rest of the world versus the UK, and the ability of UK higher education institutions to remain competitive.

35. Some critics of the Finch Report and RCUK’s updated OA policy have complained that by making British-funded research outputs Open Access, these policies will put the UK’s researchers at a competitive disadvantage with the rest of the world, because our research grants will have to bear the cost of APCs while those of others do not.

36. This complaint is true, but misleading. In the first place, all the signs are that the world as a whole is moving towards universal Open Access, so any period of Britain alone paying APCs will likely be short.

37. But more importantly, by making their work Open Access, British researchers place themselves at a huge advantage in terms of the visibility of their work. Numerous studies have now been performed on the Open-Access Citation Advantage. A good summary is found in the meta-analysis of Swan (2010)⁸. Swan surveyed 31 studies of the OACA, showing that 27 of them found an advantage of between 45% and 600%. I analysed the final table of that report, averaging the citation advantages given for each of ten academic fields (using the midpoints of ranges when given), and found that on average open-access articles are cited 2.76 times as often as non-open.

38. In conclusion, having made a bold beginning, the UK government should push vigorously on with its open-access plans, yielding benefits for medicine, education and industry, and giving its academics a competitive advantage over the rest of the world.

Conclusions

39. We must remember that what is best for the country as a whole will not necessarily be best for each group that is affected. In this case, it is clear that the UK will benefit from Open Access that is immediate and uses permissive licences, whether achieved by the Gold or Green route. It is likely that publishers will (rightly or wrongly) fear some damage to their business as this change is made. But the interests of one small group – publishers – must not be allowed to compromise decisions made on behalf of all other stakeholders. In particular, the government of the UK is beholden to its citizens, not to the publishing industry. The government must make decisions that promote the welfare of citizens rather than decisions that suit any one business.

40. Note that publishers who find the revised Green-OA terms unacceptable will be at liberty to decline offers of manuscripts resulting from Government-funded research. (In practice, this is unlikely to happen: rather than forego the opportunity to publish publicly funded research, publishers will simply accept the loss of their government-granted monopoly on the commercial exploitation of this research.)

41. It is not the job of government to concern itself with challenges faced by the publishing community at the expense of other stakeholders. This assertion does not arise from hostility to publishers, but from a simple recognition of who the government serves.

Footnotes