At the 2001 meeting that launched the Budapest Open Access Initiative (BOAI) the newly-fledged OA movement outlined two strategies for making the scholarly literature freely available. Later dubbed green OA and gold OA, these are now the two primary means of providing open access, and both types have been mandated by research funders in the UK. For instance, in 2013 Research Councils UK (RCUK) introduced an OA policy that favours gold open access, and in 2014 the Higher Education Funding Council for England (HEFCE) announced what is essentially a green OA policy, which will come into force next year. So how does the future for open access look?
Just to remind ourselves: With gold OA, researchers publish their papers in an open access journal and the publisher makes them freely available on the Internet as a natural part of the publication process. With green, OA researchers continue to publish in subscription journals, but then self-archive a version of their work in an open repository, either a central repository like PubMed Central, or an institutional repository. Meanwhile, the official version of the paper (version of record) remains behind a subscription paywall on the publisher’s site.
BOAI did not specify that OA journals should levy an article-processing charge (APC), but while OA advocates point out that most OA journals do not charge a fee, the reality (unless something changes) is that the pay-to-play model is set to dominate OA publishing.
Importantly, this means that although BOAI attendees assumed OA publishing would be less costly than traditional subscription method, use of the APC will make scholarly publishing more expensive, certainly during the transition to open access (which could last indefinitely).
And to the chagrin of OA advocates, much of the revenue generated by APCs is currently being sucked up by traditional publishers like Elsevier and Wiley, especially through the use of hybrid OA.
In reviewing the figures for 2013-2014, for instance, Wellcome’s Robert Kiley reported that Elsevier and Wiley “represent some 40% of our total APC spend, and are responsible for 35% of all Trust-funded papers published under the APC model.” (74% of the papers concerned were published as hybrid OA).
The story is similar at RCUK. As the Times Higher noted in April: “Publishers Elsevier and Wiley have each received about £2 million in article processing charges from 55 institutions as a result of RCUK’s open access policy.” In total RCUK paid out £10m, which is in addition to the subscription fees universities are already paying.
In effect, it would seem, traditional publishers are in the process of appropriating gold OA, and doing so in a way that will not only ensure they maintain their current profit levels, but that will likely increase them. And the profits of scholarly publishers, OA advocates argue, are already obscenely high.
But green OA advocates maintain that this is not inevitable, and have long argued that if implemented wisely, and strategically, open access can squeeze out the excessive costs of scholarly publishing, and so reduce publisher profits. However, they insist, this will only happen if researchers self-archive their subscription papers rather than opt for pay-to-publish. If researchers do this, they say, publishers will have to compete with repositories for access provision, and so will be compelled to downsize their operations. This in turn will put downward pressure on costs (and thus any publishing fees). Only at the point where these costs have fallen, argue green OA advocates, should researchers consider paying to publish.
But green OA has its own issues. Indeed, critics argue that is has an incurable Achilles Heel. Specifically, since researchers invariably have to assign copyright (or at least exclusive publishing rights) to publishers when taking the subscription route, green OA will always be a hostage to self-archiving embargoes. That is, most publishers prohibit self-archived papers being made OA for a period of time after publication — generally between 6 months and 4 years.
Moreover, these critics argue, when publishers feel any kind of threat from self-archiving they will naturally seek to limit and eventually emasculate it — by, for instance, introducing embargoes where they do not currently exist, by lengthening existing embargoes, or by complicating and obfuscating their self-archiving rules in order to deter researchers from opting for green OA — as theys have sought to do on a number of occasions already (e.g. here, here and here).
In an attempt to neutralise publisher embargoes, therefore, in 2006 green OA advocate Stevan Harnad proposed what he called the Immediate-Deposit/Optional-Access (ID/OA) Mandate. This type of OA policy requires all mandated researchers to deposit their papers in an open repository at the point of acceptance by a journal. When depositing, however, they can specify whether the work is made available on an open access basis, or (where an embargo applies) on a closed access basis. Harnad dubbed this “Almost OA”.
Fundamental to the ID/OA mandate is the so-called “copy request” or “fair dealing” button. When implemented in a repository this enables closed access papers to be made freely available to anyone who requests a copy. To do this they simply click on the button in the bibliographic record, and the repository software then forwards the request to the author who, explains Harnad, “can click once to comply with the request”.
If the author approves the request the repository software then sends the requester a copy of the document, or more usually a private link to the full text.
The ID/OA mandate combined with the “copy request” button, explains Harnad, was “specifically formulated” to ensure that providing immediate access to papers is “immune from any delays or embargoes (based on publisher policy or copyright restrictions).”
Over the past nine years green OA advocates have advocated tirelessly for institutions and funders to introduce ID/OA mandates, and for institutional repositories to implement the “copy request” button. They also recommend that, in order to ensure compliance, mandates should be tied to the evaluation procedures used by the institution’s promotion and tenure committee.
And this advocacy has borne fruit, initially with the pioneering ID/OA mandate introduced at the University of Liège.
But it was last year’s announcement of HEFCE’s “Policy for open access in the post-2014 Research Excellence Framework” (REF) that green OA advocates claim as their greatest victory.
The HEFCE policy states that, “to be eligible for submission to the post-2014 REF, authors’ final peer-reviewed manuscripts must have been deposited in an institutional or subject repository on acceptance for publication. Deposited material should be discoverable, and free to read and download, for anyone with an internet connection.”
Importantly, HEFCE adds: “authors can comply with the policy by making a ‘closed’ deposit on acceptance. Closed deposits must be discoverable to anyone with an Internet connection before the full text becomes available for read and download (which will occur after the embargo period has elapsed). Closed deposits will be admissible to the REF.”
Many OA advocates have expressed unhappiness with HEFCE’s decision to allow closed deposits, and disappointment that it will permit publisher embargoes of up to 12 months for STEM subjects and 24 months for HSS subjects. But Harnad insists that this is not problematic, because the button can free closed deposit papers on request, and so moot any publisher embargo.
The HEFCE policy in conjunction with the “copy request” button, he has written, will “detoxify embargoes” and “plant the seeds for their speedy extinction, by depriving publishers of the power to delay access-provision with their embargoes.”
But will the HEFCE policy live up to Harnad’s promise? Specifically, can we expect it to “plant the seeds” for the extinction of embargoes?
Set in concrete?
Currently, the signs are not great. Eleven months after HEFCE’s announcement (and doubtless in response to it), Elsevier published a set of new sharing and hosting policies that, far from signalling the extinction of publisher embargoes, would seem more likely to set them in concrete.
Most significantly, where previously authors were permitted to make papers they deposit in their institutional repository freely available from day one, henceforth Elsevier-published papers can only be made OA after the expiration of the specific journal’s embargo — as Harnad was quick to complain.
But if Harnad is right to argue that the button moots any publisher embargo does this matter?
It would seem so, for two reasons.
First, it is far from clear that researchers routinely respond to copy requests. Certainly that would seem to be the conclusion of a 2010 study undertaken by a group of open access advocates (including Harnad).
While by no means a detailed or extensive study (reporting as it does on the use of the button in just three institutional repositories) the findings are not encouraging. Approval rates [i.e. when authors responded positively to a copy request] in these institutions varied from 27% (University of Minho), through 47% (University of Southampton) to 60% (University of Stirling).
Moreover, these rates have subsequently fallen at the University of Minho, to just 23% last year. We don’t know whether Southampton and Stirling have seen a similar drop, but the director of the University of Minho’s documentation services Eloy Rodrigues reports that researchers quickly get “tired” of using the button.
Second, it has yet to be satisfactorily established that it is lawful to use the “copy request” button.
Harnad and fellow OA advocates argue that its use is covered by the fair use/fair dealing rules associated with copyright. But not everyone agrees. For instance, a researcher at a US university told me recently that the button has not been implemented in his repository because the university lawyer thinks it is unlawful.
Either way, Elsevier now seems keen to outlaw use of the button (as commentators were quick to point out). In its new hosting policy, for instance, the publisher states that manuscripts should “Not be used to substitute for services provided directly by the journal, for example article aggregation, systematic distribution via e-mail lists or list servers or share buttons…”
But can publishers outlaw the button? When I put the question to Rodrigues earlier this year he replied, “It is at least very questionable that publishers would have any solid legal ground to act against the button use, and, on the other hand, it would give them very bad publicity. So, from a cost-benefit point of view, I think the button is not a high priority for publishers.”
In fact, suggests Harnad, the “copy request” button is not a “share button”. Rather it is, “a one-on-one eprint request, from one requestor, to the author, with one click each. It is not automatic. Nor is it article aggregation, systematic distribution or a ‘share button’ (as in research-index, academia.edu — and Mendeley, till Elsevier bought it!) The lawyers are just trying to use and include every menacing word they don't understand.”
Harnad’s advice to researchers worried about Elsevier’s attempt to ban the button, therefore is: “just ignore it.”
Is Elsevier really targeting the “copy request” button with its new policy? More importantly, does it believe it to be unlawful to use the button? “It’s not an easy and straight forward question you ask, although it is a good one, and we too are thinking this through,” replied Elsevier. “Our policies do permit private sharing of accepted manuscripts by repositories during their embargo period, and the share button can facilitate this. The policies also prohibit systematic distribution in this way. So at low volume, on non-commercial repositories, to users who want access for non-commercial purposes I don’t think we would have any problem. A challenge is distinguishing this use case from others.”
We can safely assume that Elsevier (and other scholarly publishers) do not like the “copy request” button, and that they would be happy to see it to go away. The problem publishers face, of course, is that repositories are black boxes so far as the workings of the copy request button is concerned. They don’t know how it is being used, by whom, and for what purposes. But if evidence turned up suggesting that papers released via the button were being widely distributed and/or used for commercial purposes they would surely take action to prevent it, and perhaps seek to have the button outlawed in toto.
The problem for researchers, by contrast, is that Elsevier’s move — combined with the lack of clarity over its legal status — makes using the button appear more risky. As Harnad et al noted in their article about the button, researchers “are fearful of what they do not know, or do not understand.”
And the problem for the OA movement is that, since the button is viewed as an integral and essential part of the ID/OA mandate, there must now be greater concern about the likely efficacy of the HEFCE policy. After all, OA advocates maintain that embargo-delayed access (even when it is as short as six months) — “is next to useless” and “not open access in any common-sense interpretation of the term.” (See comment #4 here for instance).
But what are HEFCE’s views on these matters? The spokesman I contacted began by asserting that HEFCE’s policy is not in fact an ID/OA policy. As he put it, “Our policy requires that open access is granted as soon as possible after deposit, and no later than the embargo maxima set out in our policy (12 months for REF Main Panels A&B, 24 months for C&D). This is different from the model ID/OA policy described by Stevan Harnad, which states that ‘only depositing itself needs to be mandated’ and ‘setting the access privileges to the full-text can be left up to the author’, with ‘open access strongly encouraged, but not mandated’.”
He added that the wording in the HEFCE policy stating that closed deposits will be admissible to the REF “refers mainly to papers that are still under embargo at the point they are submitted to the REF.”
And that, of course, is the point of implementing the button.
With this thought in mind, I asked if HEFCE believes that the button is lawful, and whether it sanctions its use. “We have not sought legal advice on the status of the ‘copy request’ button,” the HEFCE spokesman replied. “We note that some commentators hold that ‘fair dealing’ provisions within UK copyright law would cover its use, but we don’t believe that its legal status has been tested in the courts.”
He added, “It is for institutions, and for those that operate subject repositories, to decide whether to implement the ‘copy request’ button in their systems. This is neither a stipulation of our policy, nor a matter that requires our approval.”
In short, HEFCE is unwilling to commit itself on use of the button, and Elsevier is keeping its options open.
This puts institutions in a difficult place: they will have to decide for themselves whether implementing the copy request button is lawful. And in today’s increasingly risk-averse culture, we must question whether university lawyers in the UK would be any more willing to sanction its use than the lawyer in the US university I referenced above. What we don’t know, of course, is how many repository managers have actually raised the question with their legal departments.
The hassle factor
In the meantime, gold OA advocates are arguing that Elsevier’s move heralds the death of green OA. As PLOS co-founder Michael Eisen was quick to note — in a blog post to which he attached the inflammatory title “the inevitable failure of parasitic green open access” — by changing its policy Elsevier has reminded us that green OA always had a “fundamental logical flaw”. He added: “It should always have been clear that the second Elsevier saw green OA as an actual threat, they would no longer side with the angels. And that day has come.”
Eisen continued, “I hope IRs will continue to grow and thrive. Stevan and other green OA advocates have always been right that the fastest — and in many ways best — way for authors to provide open access is simply to put their papers online. But we can [no] longer pretend that such a model can coexist with subscription publishing. The only long-term way to support green OA and institutional repositories is not to benignly parasitize subscription journals — it is to kill them.”
What Eisen ignores here is the fact that in pioneering use of article-processing charges PLOS (along with fellow OA publisher BioMed Central) created the enabling environment that has allowed subscription publishers to appropriate gold open access. As such, we can expect the current oligopoly to continue to dominate scholarly publishing, and in an undesirable way.
But if Eisen is right to predict that green OA is set to fail, it is unlikely to be a direct consequence of publisher embargoes, or legal uncertainty over the “copy request” button. Rather it will be because libraries are finding it increasingly difficult to manage the escalating number of ever more complex (and sometimes incompatible) institutional and funder mandates, the constantly changing self-archiving policies of publishers (and the shifting embargo periods) plus widespread confusion over article versions. All of this is creating a horrendous and expensive bureaucratic headache, and one which universities are ill equipped to cope with. At the same time, they face the nightmare of trying to manage the payment of hundreds or thousands of APCs.
All too aware of this, publishers appear to have come up with a cunning plan. On cue, Elsevier is trialling what it calls its Institutional Repository Program. Universities taking part in the program will have access to an API that provides a) metadata on Elsevier articles, b) tools for checking full text entitlements, and c) access to the best available version of an article. (See this chart for more detail). In practice, what Elsevier is offering is to take on much of the work required to manage green OA.
Of course, Elsevier is not doing this out of the goodness of its heart. The quid pro quo seems to be that libraries hand back access provision to publishers. By doing so, they can leave decisions about when a paper becomes OA, and which version to use, to publishers. And they don’t need to engage in the expensive process of creating their own metadata for the papers their faculty produce, because publishers will feed their metadata to co-operating repositories.
The benefit for publishers here is that this allows them to take control of green OA — to appropriate it much as they are appropriating gold OA. And the more control they have over green OA, the more power they have to manage and direct the transition to open access. Specifically, they will be able to increasingly weaken green OA, whilst pushing researchers and funders towards gold OA. And the first step in this process is presumably to persuade libraries that institutional repositories need only link to the “version of record” on the publisher’s site, they do not need to host the full-text themselves.
In other words, even when papers become open access it will be the publisher who controls that access, with repositories acting merely as portals to the content, not the source of it.
Clearly, not all institutions will want to partner with publishers in this way, and institutional repositories will doubtless continue to host the full-text of other document types. But as the complexities of managing green OA come more sharply into focus many libraries will find the option of being able to outsource most of the work to publishers hard to resist.
This suggests that it will be the hassle factor of managing self-archiving that will kill off green OA, not publisher embargoes as such, or uncertainty over the button.
The long game
This is not just speculation, or publisher wish fulfilment. Libraries are already moving in this direction. For instance in February, after taking part in the pilot of Elsevier’s Institutional Repository Program, the Dean of University Libraries at the University of Florida, Judith Russell, explained in a webinar how she and her library colleagues have come to change their views about the role of a repository vis-à-vis scholarly papers.
As she put it, they now see it as “a metadata repository and a vehicle for discovery”. As such, it will “not necessarily deliver the full text of the article.”
One benefit of partnering with publishers in this way, she added, is that it “reduces the burden on the author of having to get the manuscript copy to us … [since] ... we can rely on Elsevier to deliver both the metadata and the access to the content. So we have [also] reduced the effort that we would make chasing authors for their manuscript.”
It was this pilot, explained Alicia Wise on Twitter that “informed the change” to Elsevier’s sharing and hosting policies.
Meanwhile, we can see publishers pushing in the same direction in the US with the CHORUS project.
In the UK, HEFCE would doubtless point out that its OA policy requires not just that metadata is deposited in repositories, but the full text too. But if in a year or so — when it has become apparent that complying with its policy is proving too difficult and expensive for research institutions to manage effectively — universities and/or publishers turn up at HEFCE’s door and point out that publishers are better equipped to manage green OA, and that they are willing to feed institutional repositories with the metadata they need for REF compliance purposes, would they not be likely to get a sympathetic hearing?
Likewise if — when it has become clear just how difficult it is proving to comply with the requirements of the US OSTP Memorandum — publishers go to the federal agencies subject to the memorandum and suggest that, rather than spending millions of dollars building their own repositories (or piggy-backing on PubMed Central), they could instead use the CHORUS service to direct users to scholarly papers on publisher’s web sites, would they not get a sympathetic hearing?
This is surely the long game publishers are playing: appropriate gold OA in a way that preserves their profits, while simultaneously seek to appropriate green OA in order to control it, and then gradually phase it out, thus ensuring a transition to a pay-to-publish environment that best suits their needs, and at a cost based on their asking price.
But we need to ask: would this not be the best outcome? Does it matter how open access is achieved, as long as it is achieved?
Actually, it does matter. As we noted, one of the main promises of the OA movement was that open access would solve the affordability problem that has held universities in its iron fist for several decades now — the so-called “serials crisis”. Pay-to-publish gold OA may seem like a good solution, but if it proves as expensive as (or more expensive than) subscription publishing, how will the research community afford it?
True, gold OA allows universities to shift some of the cost burden on to the shoulders of funders, but not everyone has a funder like RCUK willing to pay for APCs. As Aaron McCollough put it recently, “If we aren’t careful, current access inequities (insufficient library purchasing power) could be replaced by a different access gap (exclusion of authors working at institutions that can’t afford APCs).”
And even if the affordability problem is not as serious as many claim, is it right that publishers should be able to continue plundering taxpayer’s money while offering so little value in return?