Welcome to the SPARC Open Access Newsletter, issue #165
September 2, 2012
by Peter Suber

Read this issue online


SOAN is published and sponsored by the Scholarly Publishing and Academic Resources Coalition (SPARC).


Tectonic movements toward OA in the UK and Europe

Because this article is long, I'm including a table of contents:

1. Three major OA announcements from the UK on the same day
2. Some recent history as context for these announcements
3. Basics of the new RCUK policy
4. Basics of the Finch recommendations
5. General agreement between the RCUK policy and Finch recommendations
6. Appreciation of the large-scale shift to OA in the UK
7. Some consequences for journals and authors
8. Responding to publisher fears of green OA
9. Objections and recommendations
10. Announcements from Europe the day after the UK announcements

(1) Three major OA announcements from the UK on the same day

On July 16, 2012, three announcements transformed open access (OA) policy in the UK.

First the Research Councils UK (RCUK), announced a new OA policy to replace the one it originally adopted in 2006. The new policy will take effect April 1, 2013.

New RCUK policy, July 16, 2012

Guidance to accompany the new RCUK policy

2006 RCUK policy

Later in the morning, David Willetts, UK Minister of Universities and Science, announced that the government had accepted most of the June OA recommendations from the panel he had appointed the previous September. The panel is officially called the Working Group on Expanding Access to Published Research Findings, but informally called the Finch Group, after its convenor, Janet Finch.

Willetts response to the Finch recommendations, July 16, 2012

Finch report itself: "Accessibility, sustainability, excellence: how to expand access to research publications: Report of the Working Group on Expanding Access to Published Research Findings," June 2012

Finally, the Higher Education Funding Council for England (HEFCE) announced plans to require OA to research submitted to the next Research Excellence Framework (REF) in 2014.

HEFCE announcement, July 16, 2012

(2) Some recent history as context for these announcements

Because the HEFCE didn't announce a policy, but merely plans to develop one, here I'll focus on the RCUK policy and Finch recommendations.

But before I do, note that these three major announcements were not the only large, recent OA announcements, even if we limit our attention to the UK. If it looks as though the three announcements on July 16 took the UK to a tipping point, the run-up to those announcements will make the increasing momentum forcefully clear.

* On June 21, 2012, the Royal Society released its recommendations on open data.

David Willetts commissioned the RS to write this report at the same time he commissioned the Finch Group to write a report on OA publications.

The RS report recommends, among other things, that "[c]ommon standards for sharing information are required to make it widely usable" and that "[p]ublishing data in a reusable form to support findings must be mandatory."

Also see the very good article on the data recommendations from Geoffrey Boulton, the chair of the RS group producing the report.

One week after the Royal Society released its major report on open research data, the UK Cabinet Office released its own major report on open government data.

* On June 28, the Wellcome Trust strengthened its OA mandate in two ways. First, it will impose sanctions for non-compliance. Second, when it pays a publication fee to a fee-based OA journal, it will require CC-BY or the equivalent. It previously required an open license but accepted CC-BY-NC or the equivalent. The new compliance-building measures take effect immediately, and even apply retroactively to articles published since October 2009. The new libre standard will take effect early next year.

* On June 28, the UK Cabinet announced that Sir Mark Walport would be the new Chief Scientific Adviser to the UK government, starting April 1, 2013. Walport is currently the Director of the Wellcome Trust, and has been the prime mover behind its exemplary OA policies. (Note that the Trust was also represented on the Finch Group, though not by Walport.)

(Another piece of people news: On June 18, Timothy Gowers became Sir Timothy Gowers. He may have been knighted for his work in mathematics --he is a Fields medalist-- but the Times Higher Supplement gave less space to his mathematics than to his activism for OA.)

* On July 26, the UK Department for International Development announced an OA mandate to take effect November 1. It requires green or gold OA, with a preference for gold and a willingness to pay for it. If authors don't choose gold they "must" choose green. The policy also encourages the use of CC-BY licenses and discourages the use of CC-BY-NC-ND. The DFID policy was announced 10 days after the RCUK policy, and will take effect five months earlier.

Here's some recent history from outside the UK:

* On May 14, the UN Human Rights Council released a report in which Special Rapporteur Farida Shaheed recommended (74.c) that "States...promote open access to scientific knowledge and information on the Internet" and (74.d) that "Universities, research and funding institutions  adopt mandatory open-access policies for journals and repositories of research...." The report links both recommendations to "the right to enjoy the benefits of scientific progress and its applications...the right to participate in cultural life, as well as other human rights."

* On May 15, the Global Research Council (GRC) announced that its members, nearly 50 public funding agencies from nearly 50 countries, will work toward a common understanding on how to define research integrity and how to promote OA. The GRC aims for a consensus statement on each topic by May 2013.

* In late June, Denmark announced a national OA mandate that can be satisfied by depositing in an OA repository (green) or publishing in an OA journal (gold). The comprehensive new policy applies to the Danish Council for Independent Research, the Danish National Research Foundation, the Danish Council for Strategic Research, the Danish National Advanced Technology Foundation, and the Danish Council for Technology and Innovation.

* On July 1, a new green OA mandate took effect at Australia's National Health and Medical Research Council (NHMRC). As the result of new leadership, the Australian Research Council (ARC) is now considering an OA mandate as well.

* On July 3, Germany's largest public funder, the Deutsche Forschungsgemeinschaft (DFG), issued a major strategy paper on OA. It wants to provide OA to all DFG-funded research, and like the RCUK policy and Finch recommendations, it prefers gold OA to green. In addition, it will support the conversion of TA journals to OA, and the development of OA books.

* On July 13, the European Research Council announced that it would start depositing articles in UK PubMed Central (UKPMC). Because UKPMC now has three participants from outside the UK (Telethon Italy, the Austrian Research Fund, an ERC), the repository will change its name to Europe PMC.

* Finally, on July 17, the very day after the three major announcements from the UK, the European Commission released three major OA announcements of its own. However, because these came after the UK announcements, and because they are large, I'll discuss them separately in Section 10 below.

(3) Basics of the new RCUK policy. (I'll add more detail in Section 5.)

The new RCUK policy requires green or gold OA for all RCUK-funded research, starting April 1, 2013.

RCUK-funded authors "must" publish in RCUK-compliant journals. A journal is RCUK-compliant if it offers a suitable gold option or a suitable green option. It need not offer both.

To offer a suitable gold option, a journal must provide immediate (unembargoed) OA to the version of record from its own web site, under a CC-BY license, and must allow immediate deposit of the version of record in an OA repository, also under a CC-BY license. It may but need not levy an Article Processing Charge (APC).

To offer a suitable green option, a journal must allow deposit of the peer-reviewed manuscript (with or without subsequent copy editing and formatting) in an OA repository not operated by the publisher. It must allow non-commercial reuse, hence the equivalent of a CC-BY-NC or more liberal license. It may require an embargo of up to six months from the date of publication for articles in the natural sciences and up to 12 months in the social sciences and humanities. (More precisely, the embargo may be up to 12 months for work funded by the AHRC and the ESRC, and up to six months for work funded by any other Research Council.) The journal must not charge a fee for this option.

The RCUK will give block grants to universities to pay APCs on behalf of faculty who publish RCUK-funded research in journals offering a suitable gold option and charging an APC. Universities will create publication funds for redistributing these block grants, and use their own criteria to decide which APCs to pay and which not to pay. (More on this in Section 7.)

The policy also includes a data provision, though not quite an open-data mandate. Articles arising from RCUK-funded research "must" include "a statement on how the underlying research materials – such as data, samples or models – can be accessed."

The policy applies to authors whose research is "wholly or partially funded" by the RCUK.

The new policy replaces a 2006 policy which mandated green OA and offered more direct ways to pay APCs on behalf of grantees. The new policy drops the green OA mandate and pays APCs solely through block grants to universities.

(4) Basics of the Finch recommendations. (I'll add more detail in Section 5.)

* The very first Finch recommendation (Finch report, p. 7) is that OA journals should be "the main vehicle" for publishing new research. The report recommends OA repositories only for theses and dissertations, grey literature, data, and preservation (p. 8). The preference for gold over green is explicit and emphatic.

The third recommendation (p. 7) calls for libre OA. But it's not as specific as the RCUK in specifying which open licenses are acceptable, and it's not as liberal as RCUK in calling for rights to commercial use. Instead, it recommends "policies to minimise restrictions on the rights of use and re-use, especially for noncommercial purposes."

The fifth recommendation (p. 7) calls for "walk-in access to the majority of journals to be provided in public libraries across the UK." Of course even if the number of scholarly journals available at public libraries rose from a negligible sliver to a majority, this would not be OA. The fact that it's better than nothing doesn't change the fact that it's designed not to provide OA but to make up for the lack of OA.

The ninth recommendation (p. 8) calls for the development of "the infrastructure of subject and institutional repositories...particularly in providing access to research data and to grey literature, and in digital preservation."

The tenth and last of the major recommendations (p. 8) warns against short embargoes on green OA that might create "undue risk to valuable journals that are not funded in the main by APCs." Although most OA journals are not funded by APCs, this is an elliptical reference to subscription or toll-access (TA) journals. The same recommendation warns that "[r]ules [about embargoes] should be kept under review in the light of the available evidence as to their likely impact on such journals." The implication is that evidence of harm would justify longer embargoes. There is no hint that continuing lack of evidence of harm would justify shorter embargoes, and no hint that short embargoes might be justified anyway, for example, to improve access to research.

* One minor recommendation (p. 9) calls on the government to reduce the "VAT [value-added tax] burden" on digital journals. Though minor, it's worth mentioning because it's the only Finch recommendation that David Willetts, the UK Minister of State for Universities and Science, did not accept. As he explained in his response, he couldn't accept it because the VAT is determined in part by agreements with the EU beyond his control.

* The Finch Group expects the full transition to open access to cost £50-60 million/year ($79-95 million/year) of which £38 million/year ($131 million/year) would go toward paying APCs. The rest would pay for green OA infrastructure and renewed licences for non-OA journals during the transition to OA.

This £50-60 million/year would come funding agencies and universities, whose budgets are already squeezed. The government is not proposing a new allocation to cover these costs.

* Unlike the RCUK policy, the Finch report includes long passages of explanation and justification. As a result, we not only see the Finch Group subordinate green to gold, see something of its thinking.

Some of its reasons for preferring gold to green OA are based on real virtues of gold. Unfortunately, however, many of its reasons for preferring gold to green OA are based on a distorted and jaundiced view of green.

The report says  (8.9) that green OA isn't libre, or cannot be libre, which is false. For a host of counter-examples and a sense of their momentum, see my June 2012 article on the rise of green libre OA.

The report says (8.24) "that – beyond the relatively narrow range of subjects and disciplines that support large-scale repositories – the impact of repositories on researcher behaviour has so far been limited." But on the contrary, green OA exceeds gold OA in every field other than medicine, biology, and biochemistry.

The report says (8.26), correctly, that repositories don't provide their own peer review but can and do host articles peer-reviewed elsewhere. However, when evaluating green OA as a policy option, it asks about the adequacy of repositories "_by themselves_" (emphasis in original), hence, without peer review. It doesn't ask about the adequacy of repositories as they are commonly used today. The authors of the report may mean that requiring green OA for articles peer-reviewed by journals would be bad policy. But if so, that's a conclusion, not a neutral question for framing the inquiry. Moreover, it begs the question against the 50+ funding agencies who have thoughtfully decided to require green OA for peer-reviewed manuscripts. It also paternalizes the majority (59%) of TA publishers who voluntarily allow green OA for their peer-reviewed manuscripts.

The report implies that all or most TA publishers that allow green OA demand embargoes (8.28), which is false. It's true that all or most green OA mandates from funding agencies permit embargoes, and it may be true that the Finch Group itself would like to demand embargoes, in order to protect TA publishers. But if we look at the green policies of TA publishers themselves, it's clear that most don't think they need that kind of protection. Publishers who do demand embargoes on self-archiving, such as the Nature Publishing Group, are notable as exceptions to the general permission for immediate self-archiving.

Although the Finch report favors gold over green, it misrepresents gold OA as well.

For example, it says more than once (p. 6, Point 3.10 at p. 23, and Point 8.27 at p. 95) that all or most OA journals charge APCs. But the opposite is the case: most (70%) do not. This is elementary, and has been known since 2005.

It says flatly (3.10) that APCs are "paid by authors". But this is usually false. According to the 2011 Study of Open Access Publishing (SOAP), the APC is usually paid by the author's funding agency (59%) or employer (24%), not by authors out of pocket (12%).

The report also picks a high estimate for the average APC. As The Guardian put it, "the Finch report's transition cost of £50–60m is made up of '£38m on publishing in open access journals, £10m on extensions to licences and £3-5m on repositories.' ...The £38m is based on an average APC 'between £1.5k and £2k', which is nonsense. Even at the low end of this range, £1,500 is nearly double the $1,350 (£870) charged by PLoS ONE, and almost three times as much as the $906 (£585) found [by David Solomon and Bo?Christer Björk] as the average of 100,697 articles in 1,370 journals."

The report seems comfortable with an estimate of £1.5-2k per APC either because the Wellcome Trust pays that much (6.11) or because the 2011 Open Road report recommends paying no more than £2k per APC. The Cambridge Economic Policy Associates (CEPA),which updated the Open-Road estimate for the Finch Group, came in (7.14) with a slightly lower estimate of £1,450 per APC.

If as a result of the Finch report, the UK offers to pay £1.5-2k per APC, then fee-based OA journals with lower fees will raise their fees to avoid leaving money on the table. If the true average APC is closer to $906 (£585), as David Solomon and Bo?Christer Björk have determined, that will mean a lot of fee raising.

(5) General agreement between the RCUK policy and Finch recommendations

Many people initially misinterpreted the RCUK policy, including me. We saw it as more green-friendly than it actually is and, as a result, we saw it in conflict with the Finch recommendations.

Among those who joined me, at first, in seeing notable differences between the RCUK policy and Finch recommendations were Stevan Harnad (the RCUK will "stay the course in which they were already leading -- mandatory Green OA...")...

...the Research Libraries UK ("neither the Finch report nor the Government response gives sufficient weight to the role of the well-established network of repositories in the UK....[But] we...strongly support the revised open access policies announced by RCUK which fully recognise the importance of Green OA...")...

...and Robert Campbell, Senior Publisher at Wiley-Blackwell and a member of Finch Group (parts of the RCUK policy "differ radically...from the Finch Group's recommendations....").

When I realized my mistake, I contacted Mark Thorley for more background and detail. Mark is the convenor of the RCUK Research Outputs Network (RON), the group responsible for developing and implementing the RCUK Open Access policy. After a long phone call, I wrote up my understanding, we revised it together, and I posted it with his permission to my blog on August 22. Here I'll repeat some of that clarification of the policy.

I contacted Mark originally to talk about what I thought were important differences between the RCUK policy and the Finch recommendations on (1) embargoes, (2) open licenses, and (3) the role of green OA. I knew that the RCUK disbursed public funds but was independent of the government. I wanted to learn more about that independence, and in particular whether the RCUK felt free to depart from the Finch recommendations, especially after David Willetts, the UK Minister of State for Universities and Science, accepted all the major Finch recommendations on behalf of the UK government.

Although our call was long, time was short and we only had time to talk about embargoes and the role of green OA.

1. First Mark pointed out that the RCUK participated in the Finch Group, and took the Finch recommendations into account when finalizing its own revised OA policy.

2. Mark didn't agree with me that the RCUK differed from the Finch Group on embargoes, open licenses, or the role of green OA.

On embargoes, I pointed out that the RCUK demanded a six month embargo on green OA and that the Finch Group thought that any embargo shorter than 12 months was "unreasonable". In reply, Mark drew a distinction. The RCUK is justified in refusing long embargoes on green OA when it offers to pay for gold OA. On his reading, the Finch Group agrees with that position, and only calls short embargoes unreasonable when demanded by funders unwilling to pay for gold OA.

3. On the role of green, Mark said that the RCUK had the same preference for gold as the Finch Group. The reason is that both groups want libre OA under CC-BY licenses, for example to support text-mining and to enable immediate OA without any publisher imposed embargo. Requiring immediate libre green might hurt publishers, but paying for immediate libre gold would not.

I pointed out that the RCUK was willing to distribute peer-reviewed research articles through repositories, while the Finch Group recommended repositories only for theses and dissertations, grey literature, data, and preservation. On his reading, the Finch Group may expect that the primary role for repositories will be for theses, grey literature, and data. But the Finch Group would definitely accept green OA for research articles when a journal offered no gold option.

According to Mark, the RCUK and Finch Group share this position: When publicly-funded researchers publish in a journal with a suitable gold option, then those authors should pursue that gold option. If the journal offers no suitable gold option but does offer a suitable green option, then grantees should pursue the green option instead. If a given journal offers no suitable gold or green option, then the journal is not RCUK-compliant and researchers must look for another journal.

When a journal offers both suitable green and suitable gold options, the researcher may choose either one. At the same time, however, institutions will develop local policies to manage their publication funds, and these local policies may affect a researcher's choice between green and gold.

Because the RCUK is still working out the details of its mechanism for funding APCs, and hopes to announce more in September, Mark would not comment on the likely value of grants to universities. However, he concedes that some faculty requests for APC funds might be rejected, for example, because a given journal appears to be a fraudulent operation or because a given author has already tapped into the fund too many times that year. But he was clear that these will be university decisions independent of any RCUK policy.

He added that journals offering a suitable gold OA option would probably not want to offer a suitable green option as well. Hence, as more journals start offering gold options to make themselves eligible for RCUK funding, many that permit green OA today may stop permitting green, or might only provide a green option with an embargo period to be too long to be compliant with the RCUK policy. Hence, authors turned down for APC funding may not have a green option to exercise at a given journal, even if those authors and their universities wanted to exercise it.

I mentioned the rights-retention OA policies at funders like the Wellcome Trust and the NIH, and at universities like Harvard and MIT. He said that universities could always adopt a policy of that kind. Doing so, of course, would create a standing green OA option regardless of a journal's own publishing contract. On the one hand, he acknowledged that the RCUK policy is currently silent on rights-retention green OA mandates. On the other hand, referring to such policies at Wellcome and NIH, he added that "this might well be something we would consider in the future, but for the moment it is up to institutions to recognize the benefits of rights-retention in helping manage their intellectual property."

4. If there are differences between the RCUK policy and the Finch recommendations, they are minor. The RCUK will go forward with its current policy, and has no plans to revise it to conform more closely to the Finch report.

Note that the RCUK is using its blog to explain and discuss various aspects of its OA policy.

I wasn't the only one to see that the similarities between the RCUK policy and Finch recommendations were being misinterpreted and underestimated. After my phone call to Mark Thorley, but before I posted the results, Graham Taylor of the Publishers Association, who wrote a helpful comparison of the two positions.

(6) Appreciation of the large-scale shift to OA in the UK

By one route or another, the UK has started a massive shift to OA for publicly-funded research. Whatever we think of the means, the end is a momentous good. The resulting OA research will be good for all the reasons we have urged over the years: sharing knowledge, accelerating research, maximizing the usefulness of research, and maximizing the return on the public's investment in that research.

In addition, it's very promising to see the UK, a world leader in research, determined to be a world leader in OA as well. It's equally promising to see a world leader in OA move past the "why OA" and "whether OA" questions to nitty-gritty implementation questions.

No one should be surprised that nitty-gritty implementation questions trigger friction, criticism, and opposition. (Everyone loves the idea of universal education until we have to agree on how to do it.) The implementation details may even make us question whether this particular route will really maximize the return on the public's investment in research.

I have some serious reservations of my own about the RCUK policy and Finch recommendations, and I'll spell them out below.

But all criticism should be framed by appreciation for the goal, and for the national commitment to the wide and deep change required to reach the goal.

I don't believe that the RCUK and Finch Group are using all that we've learned over the past two decades about making a rapid, inexpensive, and effective transition to OA. But there's no mistaking their commitment to making that transition.

The Finch report says (p. 5) that "the principle that the results of research that has been publicly funded should be freely accessible in the public domain is a compelling one, and fundamentally unanswerable." This is true and important. It has seldom been put more succinctly or asserted by a more influential body.

(7) Some consequences for journals and authors

* Most peer-reviewed journals are not RCUK-compliant today. Most OA journals use no open licenses at all, let alone CC-BY, and most journals that allow self-archiving allow gratis-only self-archiving, not self-archiving under CC-BY-NC licenses.

Unfortunately, we don't yet have an authoritative source for the fact that most journals that allow self-archiving only allow gratis self-archiving. But the Directory of Open Access Journals (DOAJ) is an authoritative source for how many peer-reviewed OA journals use open licenses.

As of yesterday (September 1, 2012), only 2,359 or 29% of the 8,098 journals in the DOAJ used some kind of CC license.

Only 956 or 11.8% of the journals in the DOAJ had the SPARC Europe Seal of Approval, which requires CC-BY.

These low numbers are an embarrassment to the OA movement. But I and others have often pointed them out. They're easy to find and should be widely known.

We can assume that the RCUK and Finch Group know these numbers and are trying to engineer a change in journal policies. However, the "guidance" accompanying the new RCUK policy says (p. 2) that journals not living up to the RCUK standard are a "minority", not a majority.

The Finch report (p. 11) was explicit about the intent to nudge journals toward gold OA business models: "Our recommendations and the establishment of systematic and flexible arrangements for the payment of APCs will stimulate publishers to provide an open access option in more journals."

David Willetts endorsed this goal, and told the Publishers Association in May that it's time to adapt and change. "To try to preserve the old model is the wrong battle to fight."

* Journals will have two incentives to become RCUK-compliant before the policy takes effect next April. One is to publish high-quality RCUK-funded research. The other is to become eligible for RCUK funding for APCs. To become eligible to publish RCUK-funded research, either option will do, green or gold. To become eligible for APC funding, the journal must offer a suitable gold option --and charge an APC for exercising the option.

Note that there are two incentives to offer a gold option and only one to offer a green option, and that a gold option gives a journal all the RCUK benefits that a green option gives, and more, while a green option gives a journal nothing that a gold option doesn't give. It's not hard to predict that more journals will see reasons to offer gold options than green options.

But among journals that don't currently offer either option, how many will become RCUK-compliant by offering at least one? We don't know yet. If the answer is that most or all will become RCUK-compliant, then we'll have to ask whether the RCUK can provide enough funding for all the resulting APCs. If the answer is that most will *not* become RCUK-compliant, then the policy will seriously limit the freedom of authors to submit work to the journals of their choice. (More in this dilemma in Section 8.)

* By requiring grantees to publish in compliant journals, without knowing in advance how many journals will be compliant, the new RCUK policy is ready to limit the freedom of authors to submit work to the journals of their choice.

This is another change from the 2006 policy, which "reaffirm[ed] [the RCUK's] long-standing position that authors choose where to place their research for publication."

To put this in perspective, the Wellcome Trust and NIH policies (among others) tell grantees that if a given publisher will not allow OA on the Wellcome or NIH terms, then they must look for another publisher. I've often defended those policies, and acknowledged that the argument against limiting author freedom applies more strongly to university policies than funder policies.

I've also pointed out that, in practice, 100% of surveyed publishers now accommodate the NIH policy, in effect allowing NIH-funded authors perfect freedom to submit work to the journals of their choice.

Hence, I don't bring up this limitation on grantee choice as an objection. I bring it up because the main reason why all OA mandates are green today, and none gold, is that gold OA mandates would force authors to publish in the peer-reviewed journals that are OA (30% today), or the even smaller subset that are libre OA under CC-BY licenses (12% today), seriously limiting their freedom.

The RCUK policy is not a gold OA mandate, or not a simple one, because in some circumstances it can be satisfied with green. But it deliberately steers authors toward OA journals and in that sense approaches a gold OA mandate. When the author's journal offers no suitable green option, the policy becomes a definite gold OA mandate. (Moreover, it's a gold policy with incentives for journals not to offer suitable green green options; more below.)

If faced with the usual objection to gold OA mandates, I can imagine two ways it might respond. First, publicly-funded authors should not publish in journals that "continue to maintain access policies that are fundamentally incompatible with the principle of unrestricted access to publicly funded research" and should "be expected to select from among [compliant] journals when choosing where to publish their research."  (This is from the "guidance" document at p. 2; notice that this rationale covers the green part of the RCUK mandate, and not just the gold.) Second, the RCUK hopes that its policy and its funding will themselves increase the number of compliant journals. Insofar as that works, it will reduce the limitations on author freedom.

* The RCUK is offering strong incentives for journals to become compliant, and a significant number of journals will act on them. I expect to see two kinds of journals start to offer suitable APC-based gold options that currently do not do so: non-hybrid TA journals and no-fee OA journals.

Non-hybrid TA journals will add APC-based gold options, or convert from TA to hybrid OA, in order to collect fees they are not now collecting. Nothing I've seen in the RCUK policy or Finch recommendations even prohibits double-dipping (charging APCs for OA articles and subscriptions for all articles, including the OA articles). Adding a double-dipping hybrid option is an easy move for a journal to make, and it's easy money.

No-fee OA journals will add APC-based gold options, or convert from no-fee OA to fee-based OA, for the same reason, to collect fees they are not now collecting.

If I'm right, we'll see a decline in full-TA journals and a corresponding rises in hybrid OA journals. And we'll see a decline in no-fee OA journals and a corresponding rise in fee-based OA journals.

Both the RCUK policy and Finch report welcome the conversion of TA journals to hybrid (or full) OA. But I've seen no sign that they welcome, or even foresee, the conversion of no-fee OA journals to fee-based OA journals. On the contrary, we saw in Section 4 that the Finch report falsely assumed that all or most OA journals charge APCs.

* The RCUK and Finch Group agree (see Section 5) that funders offering to pay for gold OA are justified in refusing to permit embargoes on green OA longer than six months in the natural sciences, or longer than 12 months the social sciences and humanities. The RCUK is now offering to pay for gold OA, and we'll soon know about the HEFCE. Hence, it certainly appears that long embargoes on green OA are coming to an end. But this is a Pyrrhic victory if green OA itself is coming to an end.

* If a journal has a suitable gold option and a suitable green option, then authors may pick either. But journals would vastly prefer authors to pick the gold option, especially if the journal charges an APC. It would bring in revenue and the green option would not. Hence, journals have a strong incentive to drop their green options. Or they have a strong incentive to make sure their green options fail to qualify as RCUK-compliant. For example, they might require long embargoes or bar CC-BY-NC licenses.

This is not cynical speculation. Mark Thorley of the RCUK also predicts that journals with a suitable gold option may not want to support a suitable green option; see Section 5.

* By paying APCs solely through block grants to universities, the RCUK allows universities to bring their own criteria to bear on which APCs to pay and which not to pay. The hope is that this will introduce some price competition among APCs.

(Note that the Finch Group supports the RCUK plan to pay APCs through university publication funds, filled by funding agencies. See "key action" no. viii at p. 9.)

Committees running the publication funds at UK universities will see the APCs charged by different journals. If some APCs are higher than others, for roughly the same quality, impact, and prestige, the committees will likely favor the lower fees. Because the budgets are finite, they may refuse to fund the higher fees, which will send a signal to those journals.

I hope it works that way. But just as we don't know how many non-compliant journals will become compliant, we don't know how universities will make APC funding decisions. Some high-prestige journals may charge high fees, and some committees will think them worth paying. We all know the temptations. Whether certain high fees are fair or bloated, some committees may be willing to pay them, even if that means paying fewer fees overall.

Some universities will follow a mixed strategy, sometimes favoring low fees, and sometimes accepting high ones, depending on the author, the journal, the percentage of the budget left for that year, and a host variables reflecting internal campus politics.

But no matter how this shakes out, some authors will be denied. Perhaps they requested funding for a journal of insufficient quality, impact, prestige. Perhaps they requested an excessive one-time fee, or perhaps they requested too many fees in the same year. Or perhaps faculty from their department requested too many fees in the same year, and the committee is under pressure to preserve some cross-campus equity.

What do authors do when denied APC funding? If their work was accepted by a journal with a suitable gold option, and no suitable green option, then the authors must go for the gold. Without APC funding, either they'll have to pay the fee out of pocket or withdraw their paper.

If their work was accepted by a journal with a suitable gold and a suitable green option, then authors are free to choose green. But what if fewer and fewer journals offer suitable green options? (More in Section 9.)

(8) Responding to publisher fears of green OA

* A recurring theme in the Finch report is the need to protect subscription publishers from risk. We saw it above (Section 4) in the Group's warning against short embargoes, even though short embargoes would clearly advance research. The need to protect non-OA publishers from risk comes up more often, and receives higher priority, than the need to improve access to research.

For other examples from the report, see Point xviii at p. 10; Point 1.8 at p. 16; Point 6.10 at p. 61; Point 7.50 at p. 81; Point 7.70 at p. 88n179; Point 8.5.v at pp. 90-91; Point 8.5.vi at p. 91; Point 8.32 at p. 97; Point 8.33.x at p. 98; Point 8.48 at p. 102; and Point 9.12 at p. 106.

Some of these passages directly warn against risks to subscription publishers by moving too far too fast with OA initiatives. Some warn against risks to "quality" publishing instead. But in context these warnings serve as premises in arguments to protect subscription publishing, without acknowledging the existence of quality OA publishing.

The report repeatedly equates the sustainability of quality publishing with the sustainability of incumbent publishers, and repeatedly makes the protection of conventional non-OA publishers a precondition of any advance toward OA.

* It's hard to avoid noticing that conventional publishers were the best-represented stakeholders in the Finch Group. The Group consisted (p. 113) of five publishers, four researchers or university administrators, three funders, two librarians, one observer, and one secretary. Three of the publishers were society publishers and none were predominantly-OA publishers.

It's as if the Group's mission were not to identify the public interest in distributing the peer-reviewed results of publicly-funded research, but instead to identify a compromise acceptable to publishers. Both are worth finding, I should add. But they are distinct. For clear thinking and better public policy, the two should have been kept distinct and pursued in sequence.

The proper first step in a process commissioned by the government on behalf of the public is to start by identifying the public interest in distributing the peer-reviewed results of publicly-funded research. The group doing so should be staffed for that job and include only sector representatives inclined to put the public interest above private interests. The group's recommendations could then be released for public comment. That would give publishers and other private interests a fair chance to air their views at length. If the public comments persuade the government that compromise is necessary or justified, then that would be the time to investigate and formulate a compromise. It makes no sense to start the process with private interests in the room and seek accommodation with them before seeking the public interest. It makes no senses to assume in advance that compromising the public interest would be necessary or justified.

It's as if the government saw its responsibility not as public leadership but as private mediation. The government should be a champion of the public interest, even when it conflicts with some well-funded private interests.

* Subscription-based publishers have long feared that rising levels of green OA would trigger cancellations. The Finch report goes out of its way to repeat these fears and protect subscription-based publishers from cancellations.

But the report indulges those fears without weighing any of the evidence that the fears are self-serving and overstated.

The report doesn't cite the fact the green OA approaches 100% in physics, and that two major physics publishers --American Physical Society and the Institute of Physics-- have said in public that they have seen no cancellations they can attribute to green OA. (The Managing Director of the IoP was a member of the Finch Group.)

It doesn't cite the fact that at three separate Congressional hearings on OA (September 2008, July 2010, and March 2012), lawmakers asked publisher-witnesses directly whether the green-mandating OA policy at the NIH had caused cancellations or other harm. In all three cases publishers were unable to cite evidence of harm.

It doesn't cite the many conventional publishers who not only permit green OA but positively encourage it, such as the IEEE's Signal Processing Society and the Nature Publishing Group. The NPG, for example, has encouraged it since 2005 and reported in public that "author self-archiving [is] compatible with subscription business models."

It doesn't cite the conventional journal publishers who endorse the green-mandating NIH policy, including the American Association for the Advancement of Science (AAAS).

It doesn't cite the conventional publishers who endorse the green-mandating Federal Research Public Access Act, such as Rockefeller University Press.

It doesn't cite the conventional publishers who voluntarily do more to provide green OA than the NIH policy requires, for example, by providing OA to the published editions of articles by NIH-funded authors (not just to the unedited peer-reviewed manuscripts), by depositing their articles in PubMed Central under open licenses rather than all-rights-reserved copyrights, and by participating in PubMed Central (in many cases with embargo periods shorter than the NIH-allowed 12 months, and in many cases even when authors were not NIH-funded).

It doesn't cite the fact that 100% of surveyed publishers accommodate the green-mandating NIH policy. If publishers face risks when publishing NIH-funded authors, nevertheless no surveyed publishers have decided that these risks exceed the benefits.

It doesn't cite the fact that while green OA levels continue to rise, so do sales and profits at Elsevier.

Finally, the report is especially concerned about risks to society publishers, as if society publishers were intrinsically committed to subscription business models (and as if green OA necessarily jeopardized those models). The report doesn't cite the fact that that hundreds of societies already publish hundreds of OA journals. At the time the Finch Group was deliberating, Caroline Sutton and I had long since released our November 2007 list of 425 societies publishing 450 full or non-hybrid OA journals. We had already released our updated December 2011 list of 530 societies publishing 616 full OA journals. The latest version of the list shows 609 societies publishing 702 full OA journals.

The large-scale Publishing and the Ecology of European Research (PEER) study released its final report on June 18, 2012, too late to be considered by the Finch Group. But there's still time for UK policy-makers to reassess publisher fears of green OA in light of its results. The PEER study is the largest to date on the question whether green OA causes harm to subscription publishers. It was coordinated by the International Association of Science, Technical and Medical Publishers, a consistent opponent of green OA policies. Nevertheless, Norbert Lossau, Scientific Coordinator of OpenAIRE and member of the PEER Executive Committee, summarized the results this way: "the economic research of the PEER project could not find any evidence for the hypothesis that self-archiving affects journal viability."

The assessment of financial analyst Claudio Aspesi in March 2012 also came too late to affect the Finch recommendations. But again we can hope that it's not to late to use them to put publisher fears into perspective:  "I find it somewhat difficult to believe that any academic or research library has cancelled a single subscription to a medical or life sciences journal because it would be able to access, twelve months later, a relatively modest percentage of the articles published in any one issue of that journal. To argue otherwise invites disbelief and cynicism...[I]f Elsevier and the other publishers demand or oppose changes in public policy, they should provide factual evidence that they will be harmed."

* In short, publisher fears of green OA have been overstated for years. Many successful non-OA publishers have repudiated these fears. People who consult the evidence can answer these fears. And when policy-makers ignore these fears, publishers adapt.

* Nevertheless, publisher fears are merely overstated, not groundless. They're evidence-free today, and disregard a large and growing body of counter-evidence, but they may not be evidence-free forever. Public policy should not indulge them, but public policy should monitor the effects of green OA on subscription publishers.

If rising levels of green OA do start to cause cancellations, for example, in fields outside physics, then we can decide what to do about it. We can act in light of the evidence, whatever it turns out to be. We can weigh the demonstrable degree of harm to publishers against the demonstrable degree of benefit to research, researchers, research institutions, and taxpayers. We can see to what extent the publishers experiencing cancellations are doing their best to adapt to the opportunities of the digital age, and to what extent they are laggards at adaptation who deserve no public assistance, especially at the expense of researchers and taxpayers.

In short, we needn't let fear of harm serve as evidence of harm, and we needn't assume without discussion that even evidence of harm to subscription publishers would justify compromising the public interest in public access to publicly-funded research.

Policy-makers must take seriously the argument that green OA mandates could be justified even if they do eventually cause cancellations. The case for this "even if" argument can be long or short. It's essentially the argument for OA itself, which UK policy-makers already accept. By no coincidence, it's also the argument that public agencies should put the public interest ahead of private interests, which oddly they do not seem to accept. But in either form, the argument is essential to avoid the mistake of letting public agencies make insurance for publishers a higher priority than access to publicly-funded research.

(9) Objections and recommendations

For a very abbreviated version of these recommendations, my editorial in BMJ, "Ensuring open access for publicly funded research," August 8, 2012.

* The policy creates incentives for journals to drop their green options, or to modify their green options so that they are not "suitable" under RCUK criteria. Journals will want APC revenue from the gold option, and will want to drop or disqualify their green option in order
to steer authors toward the gold option.

Moreover, most conventional publishers don't already offer a suitable green option. And most don't want to if it means adopting relatively short embargoes and allowing CC-BY-NC or more liberal licenses on repository copies. Hence, even without the incentive to steer authors toward the revenue-producing gold option, suitable green options will be rare.

One remedy is to require RCUK-compliant journals to offer both gold and green options, not just one.

Another, better remedy is for the RCUK to adopt (or re-adopt) a rights-retention green OA mandate of its own. It could require RCUK-funded authors to retain certain non-exclusive rights and use them to authorize green OA. That would create a "standing green option" regardless of what publishers decide to offer on their own. The green policies at the Wellcome Trust and NIH do exactly that. The RCUK policy could add funding and incentives for gold OA on top of this green foundation. The Wellcome Trust does that as well.

If the RCUK doesn't adopt a rights-retention green OA mandate, then universities can create the same effect for their own faculty by adopting rights-retention green OA policies like those at Harvard and MIT. The advantage is that universities needn't wait for the RCUK to act. The disadvantage is that all UK universities must take separate action to cover the same ground that the RCUK could cover by itself.

Through funder or university policies, we can assure green OA right now. We can do it for research in every field. We can do it without limiting the freedom of authors to submit work to the journals of their choice. We can do it before we figure out how to pay for a sustainable transition to gold OA. And we can do it in the background even after we provide funding and incentives for gold OA.

* The RCUK policy creates incentives for journals to create APC-based gold options.

Insofar as this makes full-TA journals convert to hybrid OA, it advances OA. However, hybrid OA journals create other problems (more below)

Insofar as it makes no-fee OA journals convert to fee-based OA, it's a setback. It could rid the world of no-fee OA journals, which are more viable in the humanities and social sciences than fee-based journals, and more viable in developing countries. They provide needed relief from business-model monoculture, and even offer gold options to natural scientists in the UK who cannot find funding for APCs (more below).

It's hard to remedy the threat to no-fee OA journals. I can recommend that any funded policy to support fee-based OA journals should also support no-fee OA journals. (One obvious reason: The vast majority, 70%, of peer-reviewed OA journals are no-fee.) But that's easier said than done. Fee-based journals live on fees, which funders can agree to fund, while no-fee journals generally live on subsidies, which are difficult to distribute in article-sized increments. The current RCUK solution induces no-fee OA journals to convert to fee-based OA journals and accept support in the form of APCs.

The full solution will require a coordinated effort by research funders and universities to support no-fee OA journals at least as favorably as they support fee-based OA journals. Even apart from the RCUK policy, serious attention to this problem is long overdue.

* The RCUK policy will pay APCs to hybrid OA journals even when the journals double dip, that is, even when they collect APCs on their OA articles and charge subscription fees for the same articles. This is a waste of taxpayer money and a giveaway to bad actors. The longer it continues, the more it will encourage double dipping.

The remedy is simple: Do not allow publication funds to pay APCs at double-dipping hybrid journals. This means that every RCUK-compliant hybrid journal must promise to reduce its subscription price in proportion to the uptake of its gold option, and must provide public data to prove it is doing so. I've heard publishers complain that making a truly proportional reduction in the subscription price is tricky and onerous. I'm sure it is. But some publishers already do it, and putting this burden on the remaining hybrid publishers is better than subsidizing their double payments with taxpayer money.

If it's too difficult to support single-dipping hybrid OA journals without supporting double-dipping hybrids, then just stop supporting hybrids altogether. As long as the policy creates incentives for journals to change their business models, it should encourage full OA and discourage hybrid OA. Among other things, that would connect support for gold OA with relief for library budgets.

For example, this is the position of the Compact for Open-Access Publishing Equity (COPE), which encourages universities to pay for gold OA, but only from full OA journals.

For Danny Kingsley, permitting payments to double-dipping hybrids is the primary problem with the current UK plan.

* Because the RCUK will only pay APCs through block grants to universities, researchers not affiliated with universities will not be eligible for that APC funding. There will be other sources of APC funds, but (for example) it appears that the Wellcome Trust will only pay APCs for Wellcome-funded authors. RCUK-funded authors unaffiliated with universities will be out of luck. Even exemplary university-based green policies will not help these authors.

The problem will be worse if journals follow the RCUK incentives to drop or fail to adopt suitable green options.

* For university-affiliated researchers, either there will be enough money to pay the APCs on every new article arising from RCUK-funded research or there won't.

If there will, this will be a very expensive transition to OA, much more expensive than a green OA mandate.

If there won't be enough money, then several painful consequences follow. First, this will still be an expensive transition to OA, more expensive than a green OA mandate. Second, some authors will be denied APC funding for budgetary reasons. That's what it means to say there won't be enough money. Some university publication funds may have enough money to cover all the articles published by their faculty in a given year, but others will not. Perhaps no university publication funds will have enough.

On top of this, some authors will be denied APC funding for non-budgetary reasons (see Section 5), for example because the journals in which they published fall below the committee's standard for quality, prestige, or impact.

When authors are denied APC funding, for budgetary or non-budgetary reasons, they will either have to pay APCs out of pocket or withdraw their accepted articles and submit them elsewhere. They could resubmit their articles to no-fee OA journals, but we're already seen that those journals will be endangered.

They could resubmit their articles to journals with higher quality, or lower fees, or both, trying to fix whatever problem it was that caused the local funding committee to deny them funds. But we can't assume they will all succeed. Some will still be denied. If the problem is that they or their department already tapped into the publication fund too many times that year, resubmitting their papers elsewhere won't fix the problem.

These authors may choose green OA over gold, at least if their journals offer suitable green options. But we've seen that journals with suitable green options will also be endangered.

It appears that authors who are denied APC funding, who can't afford to pay the APC out of pocket, and who can't find journals with suitable green options, will have to go unpublished. If there are any TA-only journals left, they will be off-limits because they will not be RCUK-compliant. If authors buckle and pay the APC out of pocket, they will still be worse off than they are today and worse off than they would be under a green OA policy.

I'm not objecting that fee-based gold OA isn't worth funding. Immediate OA is better than embargoed OA, and libre OA is better than gratis OA. If a green policy can't get us immediate libre OA, then these extra benefits are worth paying for. I'm objecting that there won't be enough funding for everyone, and there won't be good solutions for those who are left out.

The funding problem gets worse if some journals inflate their APCs, and some university publication funds indulge those journals even some of the time. The lack of good solutions gets worse if journals follow the new incentives to drop their green options

When a given article cannot become gold OA, for budgetary or non-budgetary reasons, then green OA for that article is currently contingent, not assured. But it should be assured, not contingent.

Part of the remedy here is a green OA mandate built into the RCUK policy. The RCUK should require rights retention for green OA, like the Wellcome Trust and NIH, in order to assure that authors may deposit their articles in an OA repository regardless of what publishers allow on their own. The policy could require deposit in an OA repository even when authors pursue a gold option, or it could stipulate that gold OA makes green OA unnecessary. (That question is secondary here.) But it must require green OA, at least when authors are denied APC funding, and it must make green OA permissible, through author-retained rights, in order to remain effective against recalcitrant publishers.

Publishers will remain free to refuse to publish RCUK-funded authors. In principle, some may do so, stranding authors who would otherwise exercise the RCUK-based green option I'm recommending. But in principle the same is true at the NIH, and in practice not a single surveyed publisher refuses to publish NIH-funded authors. The wide scope of the RCUK policy suggests that we'd see the same result there. Publishers will not want to exclude themselves from that much high-quality research. But even if some publishers do refuse to publish RCUK-funded authors, fewer authors will be stranded by that step than by the current approach of omitting a green OA mandate to pick up authors who fall through the gold cracks.

* How will authors respond when their institution denies their request for APC funding?

Of course they'll be frustrated and angry. But where will their direct their anger? They could blame the committee running the university publication fund, or they could blame the RCUK policy. And some anger will probably hit both targets. I worry that many angry faculty will blame OA itself, emboldened by the widespread false belief that all OA is gold OA. I worry that many will think: "We never had these problems before the massive national shift to OA, so the problem must lie with OA itself." We run the risk of turning researchers from friends of OA to foes. Once they are angry, it will be very difficult, and very late, to explain that the problem only arises under a gold OA policy, not under a green OA policy.

A June 21 article in the Times Higher Education carried the headline, "Open access may require funds to be rationed." The headline sounds alarmist, and certainly no green OA policy would require rationing funds. But a program favoring gold over green, offering to pay for gold, and offering no new funds to cover the program, could well require rationing. The anger will not only harm OA itself. It will harm authors who need publication for promotion and tenure. And it will harm all researchers who would have benefited from the results that would have been, but will not be, published.

* The mechanisms to create price competition may do too little to control APCs and prevent profiteering. Clearly, they will work to some extent. University publication funds will not pay high fees when they could pay low fees, at least not for the same quality and prestige. But these funds will have their reasons to pay high fees on occasion, perhaps on many occasions. Sometimes higher fees will reflect higher costs (such as higher rejection rates), but sometimes higher fees will not reflect costs at all but prestige and publisher probing to see what the market will bear.

Again, I'm not objecting that gold isn't worth funding. This time the objection is that taxpayers may end up paying more for gold OA than they have to.

Even if the mechanism to assure price competition works as hoped, it will carry a price. In order to send publishers a signal that their fees are too high, some authors will have to be denied funds. If universities are ever willing to pay high fees at high-prestige journals, then even more authors will be denied funds.

I don't pretend to know the full remedy to this problem. But a green OA mandate alongside funding and incentives for gold is part of the remedy. That would allow the APC funding to fall short of demand without allowing the growing OA corpus to fall short of the growing national research output.

If policy-makers agree after a time that taxpayers are paying more for gold OA than they have to, then a green OA mandate would let them scale back the funds for gold OA, and the expectations for gold OA, without scaling back on OA itself.

* I've long argued that when enough journals are fee-based OA, then there will be price competition for those fees. If we don't see price competition yet, one reason is that not enough fee-based OA journals are competing for authors (or enough price-conscious authors or price-conscious fee payors). The RCUK policy will certainly create a critical mass of fee-based OA journals competing for price-conscious fee payors. But it may undermine this part of the remedy by offering public funds (via university committees) for APCs rather than allowing market prices to emerge. Hence, another part of the remedy here may be to hold off on trying to fund a nation's worth of APCs until rising levels of green OA have their own effect.

That is, another purpose for a green OA mandate is to serve as Phase One of a two-part transition to OA. As we approach the point when the bulk of publicly-funded research is routinely made green OA, some publishers will be able to keep enough subscribers and some will convert to OA. Among those converting to OA, some will charge APCs and some will not. Because the bulk of APC payments will not be assured by public funds, but from money freed up at universities by the cancellation or conversion of subscription journals, journals will have to set fees low enough attract submissions and high enough to pay their expenses. Roughly speaking, that's the recipe for a market-based APC.

If this type of green OA was embargoed and gratis (which is probable), and if we wanted immediate and libre (which is certain), then that would be the time to introduce Phase Two. Phase One only needs to last long enough to make OA the default for new funded research, to cultivate publisher adaptation to the new normal, and to fine-tune the immediate/libre/gold OA plans for Phase Two.

In June 2012, too late for the RCUK and Finch Group to consider, Alma Swan and John Houghton released a report to the UK Open Access Implementation Group making the case that green OA allows a significantly less expensive transition to OA than gold. "[F]or all the sample universities during a transition period when subscriptions are maintained, the cost of adopting Green OA is much lower than the cost of Gold OA - with Green OA self-archiving costing institutions around one-fifth the amount that Gold OA might cost, and as little as one-tenth as much for the most research intensive university sampled. In a transition period, providing OA through the Green route would have substantial economic benefits for universities, unless additional funds were released for Gold OA, beyond those already available through the Research Councils and the Wellcome Trust...."

One advantage of the two-phase transition is saving serious money on Phase One (because it's green, not gold) and saving serious money on Phase Two (because we won't overpay for APCs). But even if we skip the two-phase transition, all the other arguments for a green OA mandate above would still apply. And notice that the primary counter-argument --risk to publishers-- would not apply. The funds and incentives for gold OA in the new RCUK policy will steer most authors to fee-based OA journals. Those taking the green option will be very few in number, which will reduce the risks to publishers accordingly. Adding a green OA mandate now will be even more risk-free than adding one earlier would have been.

* I'm not recommending a green-only policy. I support gold OA and I support paying for it. I acknowledge that (today) gold makes it easier than green to eliminate embargoes and ensure libre OA, and I strongly want to eliminate embargoes and ensure libre. More, I supporting demanding immediate libre OA in exchange for paying any part of the cost of publication. Green and gold are complementary, and I support a dual or mixed policy in order to get the advantages of each. My summary objection to the Finch recommendations and current RCUK policy is that they don't take sufficient advantage of green and, in the case of the Finch report, do not even acknowledge the advantages of green. As a result, the current RCUK/Finch policy will likely pay more than necessary, make the transition slower than necessary, leave a regrettable percentage of publicly-funded research non-OA, and put the business interests of publishers ahead of the access interests of researchers.

There are ways to encourage gold while requiring green, and even to prefer gold while requiring green. But that's not what the current policy does. I want funders to pay for gold OA when they can afford to do so, provided they already require green OA and require the rights to make it permissible. The largest problems I see with the RCUK policy could be solved it it added (or restored) a green OA mandate at the bottom layer of the policy, and laid its gold funding and incentives on top of that.

Mark Thorley of the RCUK told me (Section 5) that a rights-retention green mandate like the one at the Wellcome Trust or NIH is "something we would consider in the future." I find that very encouraging. I recommend such a policy sooner rather than later.

* The desire to reap the advantages of immediate OA and libre OA is a good reason to look to gold rather than green, at least today. But it's not a good reason to omit a green mandate. A supplementary green mandate would not interfere with those benefits in the slightest. It would simply ensure OA when we can't arrange for those benefits. We will get those benefits at the same cost and same speed if we add a green OA mandate. Hence, if there are reasons not to add a green mandate, we must look elsewhere.

If the objection is that a green OA mandate creates risks for publishers, see Section 8. It's time to stop putting insurance for publishers ahead of access to research. That's now the problem to solve, not the solution.

Alma Swan makes the point very well in the SPARC Europe response to the Finch report: "The explicit commitment in the report to preserving the status quo with respect to publisher revenues may appear to be in the interests of research but it is not. It is anti-competitive and quashes innovation in the system at a moment where value for money and the need for new, more effective ways of helping many constituencies exploit new knowledge have never been greater....[Later, by supporting long embargoes and NC licenses, both without evidence and both at the request of publishers] the report sacrifices the interests of other sectors of society by protecting the subscription based publishing  industry."

* The models I'm recommending are not fringe, novel, or untested. The Wellcome Trust and NIH are two funders (among many) with rights-retention green OA mandates. The Wellcome Trust started with a rights-retention green mandate, and gradually added funding and incentives for gold. But it never stopped mandating green. That's the right way to combine green and gold.

Another model is the new OA mandate at the UK's own DFID (see Section 2). It prefers gold to green, offers to pay for gold, requires libre when it pays for gold, prefers full OA journals to hybrid, and requires green for grantees who don't choose gold.

For yet another model, see what the European Commission is setting up for Horizon 2020. It requires OA, and requires it to be green if it isn't gold; it is still considering when to pay for gold; and for green, it uses the same embargo periods as the RCUK. For more details, see the next section.

(10) Announcements from Europe the day after the UK announcements

On July 17, the day after the three large OA announcements in the UK, the European Commission released three major documents on European OA policy.

One document announcement said the EC will "[e]stablish open access to scientific publications as a general principle for all EU funded projects in Horizon 2020." It invited "[r]esearch stakeholder organisations...to: ...[a]dopt and implement open access measures for publications and data resulting from publicly funded research..."

See the "Draft Communication on a reinforced European Research Area partnership for excellence and growth."

A second document summarized the pilot green OA mandate for 20% of the EU research funding program for 2007-2013 (called the Seventh Framework Programme, or FP7) and explained how the Commission plans to extend the OA mandate to 100% of the research funding program for 2014-2020 (called Horizon2020). Note that Horizon2020 represents 80 billion Euros --roughly 100 billion dollars-- of publicly-funded research. "In Horizon 2020, both the 'Green' and 'Gold' models are considered valid approaches to achieve open access. All projects will be requested to immediately deposit an electronic version of their publications (final version or peer-reviewed manuscript) into an archive in a machine-readable format. This can be done using the 'Gold' model (open access to published version is immediate), or the 'Green' model. In this case, the Commission will allow an embargo period of a maximum of six months, except for the social sciences and humanities where the maximum will be twelve months....The eligibility of 'Gold' open access publishing costs will be maintained in Horizon 2020. The Commission will also consider whether and under what conditions open access publication fees can be reimbursed after the end of the grant agreement....In addition, the Commission will...set up a  pilot scheme on open access to and re-use of research data generated by projects in selected areas of Horizon 2020...."

See the "Draft Communication Towards better access to scientific information: Boosting the benefits of public investments in research."

A third document recommended that members states mandate OA for publicly-funded research. Note that 92% of publicly-funded research in Europe is funded by the separate member states (thanks to Carl-Christian Buhr). "The European Commission...hereby recommends that member states...Define clear policies for the dissemination of  and open access to  scientific publications resulting from publicly funded research....Ensure that, as a result of these policies:...there should be open access to  publications resulting from publicly funded research as soon as possible, preferably immediately and in any case no later than six months after the date of publication, and  twelve months for social sciences and humanities....Ensure that research funding institutions responsible for managing public research funding and academic institutions receiving public funding implement the policies....Define clear policies for the dissemination of  and open access to  research data resulting from publicly funded research....Ensure that, as a result of these policies: research data that result from publicly funded research become publicly accessible, usable and re-usable through digital e-infrastructures...."

See the "Draft Recommendation on access to and preservation of scientific information."

Also see:
--The press release on the three documents above.  Excerpt: "As a first step, the Commission will make open access to scientific publications a general principle of Horizon 2020....As of 2014, all articles produced with funding from Horizon 2020 will have to be accessible: [1] articles will either immediately be made accessible online by the publisher ('Gold' open access) - up-front publication costs can be eligible for reimbursement by the European Commission; or [2] researchers will make their articles available through an open access repository no later than six months (12 months for articles in the fields of social sciences and humanities) after publication ('Green' open access). The Commission has also recommended that Member States take a similar approach to the results of research funded under their own domestic programmes. The goal is for 60% of European publicly-funded research articles to be available under open access by 2016."
--The video of the press conference releasing the three documents above, July 17, 2012.
--The statement of Neelie Kroes, Vice-President of the European Commission responsible for the Digital Agenda, at the press conference, July 17, 2012.
--The blog post of Neelie Kroes on the new EC policies.
--The call for proposals to help implement the new policies. See pp. 25-26.

* The EC support for green OA has been noticed in the UK: "A spokesman for the [European] Commission confirmed that despite a passage in one of its documents stating that immediate open access is 'preferable', it did not favour gold. This was lamented by David Willetts, the universities and science minister, who pledged to 'encourage the Commission and member states to put a greater emphasis on gold' because of its 'potential to make the greatest contribution to future economic growth'. A spokeswoman for Research Councils UK confirmed that the research councils did prefer gold despite their flexible framework. The Russell Group responded to the new policy by restating its anxiety about transition costs and urging the government to 'reconsider the green option ahead of a full international transition to gold'."

It has also been noticed by publishers:  "STM [International Association of Scientific, Technical & Medical Publishers] is disappointed that the Commission decided not to follow the pragmatic lead of the UK Finch Access Group in preferring Gold to Green Open Access and in opting for minimum Green embargo periods of not less than 12 months when APCs cannot be paid.

Reed Elsevier stock did not fall the day of the three dramatic UK announcements. But it fell 2% the next day after the three dramatic EC announcements. (If you like, you could interpret this as a reaction to the combined UK-EC actions.)

Finally, it was noticed by European OA organizations.

From SPARC Europe: "It is very good news that the intention is to build on the  OpenAIRE  infrastructure, with its central repository and network of National Open Access Desks. OpenAIRE is delivering the outputs from the Open Access pilot in FP7 and will now continue to be the home for outputs from Horizon 2020 research. This is a model system, harvesting relevant material from the institutional repositories of researchers across the European Union, making it openly available and adding value in the process: it should be emulated in other nations and regions."

From Enabling Open Scholarship (EOS): "The Communication confirms commitment to OpenAIRE, the repository that collects the 20% of FP7-funded research outputs currently covered by the existing Open Access pilot. In future, OpenAIRE will be used for all EU-funded research under the Horizon 2020 framework programme."

From the Mediterranean Open Access Network (MedOANet): "MedOANet...considers important the clarification on behalf of the Commission that both self-archiving into open access repositories ('green' open access) and open access publishing ('gold' open access) are valid approaches to achieve the aim of full open access to publications resulting from Horizon 2020."

* Why did the EC not tilt toward gold, as the UK did? Of course there are good arguments for giving green a significant role. But the question is why those arguments carried more weight in Europe than in the UK. I don't know the answer. But here's one factor.

A May 2011 survey from the EC turned up this result: "[A]lmost 70 % of respondents with an opinion think that it is better to use self-archiving rather than open access publishing to satisfy the open access requirement in FP7."

* The EC policy for Horizon2020 will create much more OA literature than the RCUK policy. EC funding subject to an OA mandate is much larger than UK funding subject to an OA mandate.

Horizon2020 will disburse €80 billion ($100 billion) over its six year span, or about €13.3 billion/year ($16.6 billion/year). By contrast, the RCUK has an annual budget of about £3 billion ($4.7 billion). The annual Horizon2020 research budget is about 3.5 times larger than the annual RCUK budget.

This 3.5 multiplier will nearly disappear when the budgetary impact of the RCUK is supplemented by that of the HEFCE. The HEFCE budget is £7.291 billion/year ($11.5 billion/year). The combined annual budgets of the RCUK and HEFCE come very close to the annual budget of Horizon2020 ($16.2 billion v. $16.6 billion, respectively).

On the other hand, the EC is not only mandating OA for 100% of the Horizon2020 budget. It's also recommending similar policies for member states, and 92% of publicly-funded research in Europe is funded by the member states. (Thanks to Carl-Christian Buhr for this figure.) If all the member states follow the EC policy for Horizon2020, then their impact on OA will be more than 12 times greater than Horizon2020 impact alone. Even if just one-quarter of the member states, by budget, follow suit, their impact will be more than three times greater than the Horizon2020 impact alone.

* If other countries are scanning for models to follow, and we know that many are, some will follow the EC model and some will follow the UK model. I have no idea how this will net out. But it seems to me that there are five primary variables.

1. Green policies are faster and less expensive than gold policies.

2. The EC policy involves more money, more research, more publications, and more OA than the UK policy. That suggests it will have more impact.

3. The EC policy includes a strong recommendation to the member states. That gives it a significant head start in influencing other national policies.

4. The UK policy takes effect one year earlier, April 2013, rather than April 2014.

Despite the UK's lack of the green speed advantage, the UK expects a faster transition. David Willetts told The Guardian that "he expected a full transformation to the open approach over the next two years."

The EC says that its "goal is for 60% of European publicly-funded research articles to be available under open access by 2016."

5. Between the two, most conventional publishers would prefer the UK approach, and most OA advocates and researchers the EC approach.

Stay tuned.


* Postscript

Disclosure: I participated in David Willetts' roundtable discussion in March 2011...

...which led to the creation of the Finch Group in September 2011.


Five years ago in SOAN

SOAN for July 2, 2007

* The lead essay in that issue:  "Problems and opportunities (blizzards and beauty)"

Excerpt:  "I'd like to think that we'd pursue OA even if historical circumstances gave us only one of these motivations [solving problems and seizing opportunities] rather than both at once --that is, if we suffered from access problems and had no new technology to exploit, or if we had a spectacular new technology to exploit but no particular problem to solve.  But we have both and we should acknowledge it more often....There are...beautiful opportunities to seize.  There's the fact that the internet emerged just as journal subscription prices were reaching unbearable levels.  There's the fact that the internet widens distribution and reduces costs at the same time.  There's the fact that digital computers connected to a global network let us make perfect copies of arbitrary files and distribute them to a worldwide audience at virtually no cost.  There's the fact that unrestricted access to digital files supports forms of discovery and processing impossible for paper texts and DRM-clamped digital files.  There's the fact that for 350 years, scholars have willingly (even eagerly) published journal articles without payment, a custom that frees them to consent to OA without losing revenue.  There's the fact that OA is already lawful and doesn't require copyright reform, even if it would benefit from reforms of the right kind.  There's the fact that OA is within the reach of authors acting alone and needn't wait for publishers, legislation, or markets.  There's the fact that, for researchers acting on their own, the goal of OA is even easier to accomplish than the goal of affordable journals....Here's a less obvious but even more fundamental opportunity.  Knowledge is "non-rivalrous" (to use a term from the economics of property).  That means we can share it without dividing it, and consume it without diminishing it.  My possession and use of some knowledge doesn't exclude your possession and use of the same knowledge.  By contrast, familiar physical goods like land, food, and machines are all rivalrous.  To share them, we must take turns or settle for portions. We're very fortunate that knowledge is non-rivalrous....But for all of human history before the digital age, writing has been rivalrous....Digital texts, however, are non-rivalrous....Digital writing is the first kind of writing that does not reduce recorded knowledge to a rivalrous object....We take advantage of this gift when we post information online and permit free access and unrestricted use for every user with an internet connection.  But if we charge for access, enforce exclusion, create artificial scarcity, or prohibit essential uses, then we treat the non-rivalrous digital file like a rivalrous physical object, dismiss the opportunity, and spurn the gift...."

SOAN for August 2, 2007

* The lead essay in that issue:  "Progress toward an OA mandate at the NIH, one more time"

Excerpt:  "The most important effect of this language [approved by the House and Senate Appropriations Committees and by the full House], clearly, is to make the NIH's weak request a strong requirement.  This is the significant, long-overdue reform....The bill is not everything we wanted but it's a significant, unmistakable gain on the most important front --the mandate-- and it's not a loss or retreat on any front. I wish the bill had shortened the embargo.  Any embargo is a compromise with the public interest, and longer embargoes are more harmful in medicine than in other fields.  But I'd much rather have a mandate than a shortened embargo, if we had to choose.  The reason is simply that a short embargo without a mandate isn't really short, since there would be no enforceable deadline for ending the embargo and providing OA.  Moreover, we don't have to choose.  Shortening the embargo can be our next goal...."

SOAN for September 2, 2007

* The lead essay in that issue:  "Will open access undermine peer review?"

Excerpt:  "Publisher lobbyists who object that high-volume OA archiving [or "green" OA] will undermine peer review are never specific in explaining why or how.  But they seem to be thinking about a three-link chain of causation:  high-volume OA archiving will cause libraries to cancel journal subscriptions, which will in turn cause journals to lose revenue, which will in turn cause the journals to fold up and therefore to cease providing peer review.  At least that's the form of the objection to which I'll respond here...."


Ten years ago in SOAN

Ten years ago, SOAN was called FOSN (Free Online Scholarship Newsletter).  Here are excerpts from two issues 10 years ago this quarter.

* FOSN for August 8, 2002

Excerpt:  "Institutional eprint archiving is currently undergoing a unprecedented surge of acceptance and support.  Years of patient work by many people at many institutions around the world have slowly assembled the pieces, spread the word, impressed the skeptics, and created a critical number of interoperable archives....For these purposes, eprint archiving has three components:  (1) the software for building the archives, Eprints for large institutional or disciplinary archives and Kepler for smaller individual "archivelets", (2) the Open Archives Initiative metadata harvesting protocol, the standard for making the archives interoperable, and (3) the decision by universities and laboratories to launch archives and fill them with the research output of their faculty....Here are the major developments on these three fronts going back only six months.  If you've been following the progress of the [OA] movement for any number of years, you'll agree that no other single idea or technology in the movement has enjoyed this density of endorsement and adoption in a six month period...."

Excerpt:  "At first the FOS News blog [later renamed Open Access News] was an experiment to see whether it could carry some of the burden formerly carried by the newsletter.  It has definitely succeeded.  I love the way that I can post items immediately, open the door for other contributors to post items, give every item a unique URL, and let users choose whether to read the results on the web, by email, or through RSS syndication. So here's where things stand in the evolution of the FOS Newsletter....News items with and without short comments will go straight to the FOS News blog.  If you signed up for the newsletter in order to get these news items, then you should start visiting the blog regularly, add it to your RSS newsfeed aggregator, or sign up for email delivery (on the blog page). The newsletter will no longer carry these items...."

* FOSN for September 15, 2002

Excerpt:  "In the United States, the two areas of law that I cover most in [this newsletter] --copyright and civil liberties-- have changed fundamentally in the very recent past.  The changes are unusually rapid and unusually radical.  They face dissent, but from unusually few citizens and unusually few courts.  This is an ominous combination.  Both areas of law have a constitutional basis, and both have drifted far from what were formerly their settled constitutional standards.  In both, Congress adopted radical rules that repeal rights of Americans.  In both, the changes are so egregious by constitutional standards that courageous federal courts should overturn practically all of them.  But most courts have so far been acquiescent....Civil liberties law took a sharp right turn after September 11.  The clear rationale was to detect and avert terrorism.  Whether the threat of terrorism justifies every provision in the new enactments is very far from clear, but at least the danger is clear and the necessity of a response is clear.  The rationale for the copyright revolution is the internet --not infringers or criminals who use the internet, but the internet itself.  There have always been infringers and criminals, but a new and terrifying danger arose when they (and the rest of us) gained access to a worldwide network of universal Turing machines which supports the free distribution of perfect copies to huge numbers of people.  This is the feature of the internet that makes [OA] possible.  It is something new under the sun, and we've barely begun to realize its beneficial consequences.  It is already a momentous public good, and has potential for much greater good.  But for the IP industry and Congress, this feature of the net is the equivalent of terrorism, a momentous harm, a disaster justifying the violation of first principles....[T]he DMCA anti-circumvention clause errs on one side when it should err on neither side.  Where the problem is right versus right (fair use versus copyright protection), then we must find a way to honor both or at least balance the two.  By contrast, in airport security the problem is right versus convenience --hence, it's justifiable to err on one side.  From this point of view, we see that in the DMCA the IP industry is doing exactly what infringers are doing, namely, erring on their own side rather than seeking the difficult yet constitutionally required balance."


Coming this quarter

Here are some important OA-related events coming up in the next three months.

* November 1, 2012. New OA policy takes effect at the UK Department for International Development.

* November 2012. European Parliament votes on Horizon2020 program.

* OA-related conferences in September 2012
* OA-related conferences in October 2012
* Events celebrating Open Access Week, October 22-28, 2012
* OA-related conferences in November 2012

* Other OA-related conferences



* My new book is out:  Open Access, MIT Press 2012.

I've created a "book home page" for updates, supplements, and other notes.


This is the SPARC Open Access Newsletter (ISSN 1546-7821), written by Peter Suber and published by SPARC.  The views I express in this newsletter are my own and do not necessarily reflect those of SPARC or other sponsors.

SPARC home page for the Open Access Newsletter and Open Access Forum

SPARC Open Access Newsletter, archived back issues

To subscribe or unsubscribe

Open Access Overview

Open Access Tracking Project

Open Access News blog (2002-2010)

Peter Suber

My other writings on OA

Harvard Open Access Project

Meet me on Google+

SOAN is licensed under a Creative Commons Attribution 3.0 United States License.

Return to the Newsletter archive