Last week, having quickly digested the executive summary of the Finch Report on open access (OA), I told you it was complicated. I’ve now read the report in its entirety, along with a large swathes of blogospheric commentary. I’m still decidedly of the view that it’s complicated but I wanted to think through some of the initial responses. In particular, I’d like to try to address the vexed issue of costs.
There are positive elements in the report, as I have already noted. It was warmly welcomed by science minister David Willetts and the Publishers Association, but elsewhere the reactions have been cooler. For a sampler, have a look at Cameron Neylon’s thoughtful commentary, Stevan Harnad‘s trenchant critique, and accounts of the concerns of repository managers, RLUK and universities such as UCL.
In the main, there are two features of the report’s recommendations that have caused most disappointment.
The first is the lack of clear support for subject or institutional repositories as a vehicle for green OA. By this route papers published in subscription journals can be made freely available. It’s a compromise because the version of the paper that is placed in the repository lacks the professional formatting provided by the journal and typically can only be deposited 6-12 months after publication; these conditions are designed to allow the publisher to recover costs.
The green route is popular among advocates of open access because it is inexpensive to authors and has been taken up by many more of them than gold OA. In the UK about 35% of papers are published by green open access, while only 5% go down the more expensive gold route, where payment of an author processing charge (APC) of $1000-$5000 enables the paper to be made freely available upon publication. However, rather than promoting this access model, the Finch working group suggests that repositories should be developed more as places to archive research data and to grey literature (reports, working papers, PhD theses etc.).
The second disappointment is the high estimated additional costs that the report judges may be required to manage the transition to full open access by the gold route. Commentary has fixated on the recommendation that an extra £50-60m per year be found from public monies to cover transitional costs. This looks like an extraordinarily heavy price, especially when you consider that the total annual cost of OA charges and subscriptions to academic institutions in the UK currently runs to about £175m. The problem is exacerbated because the working group can give no clear indication of how long the transition might last.
These are serious difficulties and have been raised by serious people. But I don’t see them as insurmountable.
While green OA may be cheaper to authors, it is far from perfect. In its current form, green OA does not necessarily guarantee text and data mining rights. Moreover, some publishers resist deposition in subject-based rather than institutional repositories, a barrier that tends to fragment the literature and thereby creates its own barriers to access. But in the long term there is an even bigger problem with green open access: it is likely to become the victim of its own success.
Green OA is sustainable at present because repositories aren’t good enough to pose a significant threat to publishers’ subscription revenues. Institutional repositories are relatively hard to find because their collections are not indexed by PubMed (though Google Scholar can help) and their contents are patchy, in no small part because of unreliable support from authors unaware of the facilities for deposition that their institutions provide.
My guess is that publishers like it that way; it’s better for business. But if with proper development repositories were to emerge as the go-to place for the research literature, they would soon start to hit sales of subscriptions and thereby undermine the journals on which they rely for the organisation of peer review.
This is a conundrum from which there is no easy escape, at least not by the green OA route alone, which to my mind explains why the Finch working group has opted for Gold OA as the primary route for enhancing access.
Such a strategy can only work if the price is right. The working group’s estimate of a £50-60m surcharge every year during the transition includes £38m to cover APCs, which they suggest should come from higher education and research budgets. This level of excess cost hasn’t gone down well at a time when budgets are already under severe strain; Harnad sees the Finch figures as a wanton gift publishers.
But the preoccupation with the headline figure has obscured the real issue with these costs: they are stunningly imprecise.
Let’s look at how they were worked out. The £38m estimate is derived from a projection that is midway between optimistic and pessimistic forecasts for how the move to OA will pan out; the projections take account of possible variations in APCs and in the relative speed of adoption of OA in the UK and the rest of the world. The optimistic scenario (lower APCs and faster uptake in the rest of the world) has a predicted transitional cost of £0 (yes, zero — there might even be savings), whereas a pessimistic assessment suggests the costs may be closer to £70m (see paragraphs 7.22, 7.23 of the report).
That’s quite a span. It may be larger still since working group has made no assessment of whether the subscriptions currently charged by publishers represent value for money. As has been made clear here and elsewhere, the 35% profit margins earned from research publishing by the likes of Elsevier are indicative of a market that is insufficiently exposed to the heat of competition. (Nor do they sit well with a business model that relies to a large extent on authors and reviewers providing their work for free).
Within that combined imprecision it seems to me there is scope for the government to make gold OA work cost effectively as it looks to formulate its policy from Finch’s recommendations.
Although Harnad has expressed the hope that the UK might ignore the report, I see in it an opportunity that might well chime with the government’s natural inclination to apply market forces to extract value for money. It is already clear that the disruption of academic publishing caused by internet-propelled moves to open access has yet to play out completely. As the Finch report itself acknowledges, the transfer of costs from subscriptions to APCs should create a more transparent, more efficient market.
The government has to pick up on this. Dare I suggest that it could influence the market by deliberately limiting any funds made available to cover transition expenses? Further downward pressure on costs can be made if the government reiterates that the Research Excellence Framework should be blind to impact factors, which are expensive and increasingly of questionable value, and by demanding that future REFs will only consider papers that are open access. The government, and the scientific community, will also have to work hard to roll out OA across the world; there are encouraging signs that this is already happening.
David Willetts has my sympathy as he grapples with the tortuous detail of the Finch report. It is complicated. But I hope he knows that it offers him the chance to make a real difference to the accessibility of the research literature, and the value that can be extracted from it.
Thanks for post Stephen, interesting as ever. Putting potential hazard of authors paying to publish to one side and issue of increasing an already complex proposition, might your suggestion “that [the UK government] could influence the market by deliberately limiting any funds made available to cover transition expenses?” not simply represent and invite budget tightening, and therefore reduced freedom for authors in choice of where they can publish?
Believe me, as someone who remains very closely involved with Science is Vital, I dithered a while over that particular question.
But the fact is that, since a major portion of the funding to cover APCs and subscription charges in the UK comes from the public purse, the government has always been involved in determining the level of funding for publishing. I remain very much in favour of author choice but not at any cost.
It may be impossible to escape the lure of impact factors, which for some high-ranking journals are likely to raise the cost of APCs (because of high rejection rates), but I would like to see moves away from a scientific culture that seems bedazzled by them. Frankly, I don’t know what the long term solution will look like but if we can place greater emphasis on ensuring that quality science is rapidly disseminated and papers are judged on their own merits (rather than the aggregate merit of any journal), we will be moving in the right direction. I look forward to the introduction of eLife into the high-end marketplace with interest.
Thanks Stephen, another excellent post. I think you’re right about Green OA being preferable and non-threatening for publishers while uptake remains relatively low- and this current low uptake is the reason why Stevan Harnad is always banging on about institutional mandates to deposit papers.
I would take issue with your comment that “Institutional repositories are relatively hard to find because their collections are not indexed by PubMed”- this might be an issue for biomedical research papers, but for papers outside that area it’s much less of an issue. Green OA papers are very findable- and the software on which they are hosted optimises this findability in Google, Google Scholar, Bing etc. See for example this Google search for a paper in the repository I manage- the top two hits are to the repository paper, above the Mendeley hit. It’s also notable that the publisher version from CUP is nowhere to be seen. Our stats bear out the fact that people are finding and downloading our green OA material- we receive over 100 downloads a day. Really big green OA repositories like Glasgow’s receive something in the order of 20,000 downloads a month.
Thanks for the comment Neil – it’s good to have the perspective of someone from outside the life sciences, which is my only experience. Maybe we are behind the curve but it’s only lately that I have realised the potential of Google Scholar to ferret these things out. Though as more and more people realise the utility of these repositories, I imagine publishers will get more nervy about their impact.
I haven’t read the whole thing, but the conclusions lead me to agree 100 percent with Steven Harnard. The report gives publishers a licence to continue to make huge profits from taxpayers’ money and the free labour of academics.
Finch should be ashamed for not having the courage to address the crucial question of what it should really cost to publish a paper on the web. My guess is something like $300. If universities and funding agencies refused to pay more than this, the face of scientific publishing would be changed for the better.
Well, it’s not up to the report to issue licences, but rather the government, which was the thrust of my post. Clearly there was the hand of publishers on the tiller affecting the steering of the final version of the report but given the composition of the committee, that’s no great surprise. It all boils down to what different people mean by sustainable. Like you I suspect that there is a sustainable price point for gold OA which is well below the current average.
Interesting post – but curious push to have REF blinded to impact factor and yet contingent on OA publication.
I would of course agree with supporting author choice but not at any cost, and would personally extend this to quality of publications (which is not necessarily connected only to perceived impact and novelty, but his too doesn’t preclude the utility of good filters).
Not every choice is a better one, in this field or otherwise, and if authors are to be given the means to express their own choices in a sustainable way then some transition – and funding to facilitate gold OA in one of the many outlets available – seems unavoidable.
There is pressure here for all publishers to raise their game and reduce costs as far as is practicable, but if this is just a race to the bottom and choice and quality (among the other attributes of good publishing) are ignored then I don’t see how science or scholarship will be advanced.
Competing interest: I’m a publisher, like working for and with authors and readers, and do believe in doing things the right way (first time, preferably!)
I don’t see any realistic prospect of a ‘race to the bottom’ — science is too competitive for that.
As regards the push on the REF – well, it’s already supposed to be blind to impact factors. The trouble is no-one believes that this is workable in practice. But we do, as a community, have a troubled relationship with impact factors (see more thoughts here). I would rather we had better systems of article-level evaluation. Peer review goes so far, but if we can develop more reliable post-publication metrics, that may go some way to easing the hold of IFs (and should accelerate publication by cutting down on the business of working your way down the IF rankings during the manuscript submission/rejection cycle).
You will of course have read Nature’s editorial on this subject.
http://www.nature.com/nature/journal/v486/n7404/full/486439a.html
(Disclosure: I am an editor at Nature.)
Indeed I have. I thought it very good!
I can’t claim any credit for it (!) – the point of that editorial that struck me most is that making things ‘free’ costs money. Even databases that maintain repositories of data (and/or publications) need staff to curate them. I am not convinced that some of the more messianic advocates of OA understand this.
You won’t get any argument from me on those points. There are some in the vanguard who probably want to go downy he arXiv route and rely entirely on post-publication peer-review. That would certainly be much cheaper but I can’t see the mainstream research community having much enthusiasm for that solution, partly due to conservatism and partly to the intrinsically competitive nature of science which will always stimulate a demand for ways to make comparisons and evaluations.
There’s a danger of getting a distorted view of the blogosphere. Though there has been much excitement here about open access in the past 6 months, and some of that has rippled out into common rooms, I suspect for many it remains a low-grade issue. Obviously I’d like that to change but I am aware of the challenge.
You’ve probably see this, Stephen, but also welcoming the report are those nice people at Elsevier.
Given Elsevier’s status (deserved or not) as a kind of touchstone for all that is perceived to be wrong with scientific publishing, I suspect there will be a lot of academics – even among those whose reaction to the whole OA issue involves a certain amount of eye-rolling – who will think:
Indeed. Hence the remark about ‘continued agitation’ at the end of my piece in Research Fortnight… I don’t think we’re anywhere near the end of this business.
Thanks for this thoughtful piece. One of the central points about encouraging a culture change in favour of Gold OA is that it should encourage an effective, functional market. Authors will be able to choose the best value publishing solution for their research, weighing impact, discoverability, prestige etc against cost to their research programme. This should drive prices, and those controversial publishers’ margins, down to their optimal level: you can’t legislate for these on any other basis. That is why Finch is correct not to stipulate pricing levels for APCs.
Hi Steve. Good comment. Furthermore, could Gold OA and subscriber model systems not co-exist in tandem and each community/market decide what works best for them? Mandates don’t seem to belong in the equation; would remove aspects of freedom.
(Disclosure: I’m one of ‘those nice people’ at Elsevier. A publisher in health sciences. Great to have chance to join this debate: agitation and mandates, not so much!)
I confess I don’t really understand the publisher preoccupation with mandates, particularly for publicly-funded research. An a university-employee and grant holder, I am in receipt of significant amounts of public money to carry out my research; I am also heavily reliant (currently) on the payment of library subscriptions via HEFCE, which is also funded from the public purse. I have no problem whatsoever with mandates from funders laying down conditions that are designed to promote access to any work that I produce.
I would like to see the counter-argument. It seems to me the objection to mandates is part of an effort to maintain a comfortable status quo. The argument about ‘author freedom’ doesn’t bear examination. There is already plenty of author freedom in terms of OA destinations for work.
That’s right – there are a lot of OA destinations already, so choice exists. So called mandates – depending on their strictness – could put researchers in a bind IF there is no funding to back up this option.
Then again, authors are entitled to choose their preferred publication venue – their education is a duty of the state and university an opportunity to advance the common good (not just through publication!) which most graduates will pay off through fees/taxation.
Research funding is somewhat separate from this foundation, but even then less than 1% project budgets go on publications (recent source, I don’t have to hand). There lies a question here as to whether the ingenuity and dedication of a researcher in their research, in advancing science, belongs to the latest funder through whole/part funding. Their expression of this knowledge in articles is the point of contention – it’s easy to see that data and process outputs can be allocated in terms of ownership, it’s the thought and expression though that I would see as resting solely with the individual (or group of individuals if collaborating), and so it’s their choice (and good OA publishing is a good choice, which can be facilitated with appropriate funding arrangements, just as any repositories for associated info will take funding).
Each researcher should have their own choice here (not just one from above), and so while some outputs can easily be made available only to the funder or according to their wishes, whether articles or research reports – that go beyond description and through interpretation and inspiration – relating to the work, or even just the first/most prominent, should follow a funder mandate or the researcher’s own preference is the question. This is a question I feel each researcher should ask of themselves – avenues for decisions will then follow!
Choice is out there, and with commitment, competition will drive this ecosystem forward.
(All my own opinion, naturally!)
As a publisher I have no problems with mandates provided they are accompanied by a commitment to fund Gold OA, so that there is a business model to support whatever services the research community requires of publishers. (See also my earlier post of 8:07 pm yesterday which I should have made part of this reply strand.) The Wellcome Trust’s OA policy, for example, is a good one in this respect, and merits imitation by other funders.
Edited your comment to add a link to the comment you mention.
Hi Stephen
Re opposition to mandates, it’s a matter of one size not fitting all journals and failure to include publisher in decision. Further here http://www.elsevier.com/wps/find/intro.cws_home/newmessagerwa
Clealry there is something about the peer-review and publication process that groups like NIH and Wellcome are interested in, otherwise why not just do themselves?
(Per Ian’s comment: these opinions are my own and not necessarily those of my employer.)
Thanks for the link but it doesn’t really clarify the matter for me.
I appreciate the point that we still have a very mixed economy of research funding and modes of publication. That’s an evolving situation. I guess my views are coloured by the fact that my experience is all within the life sciences, which are relatively well off in terms of funding for research and publication. As I said, given the high levels of public investment in this business, I don’t have a problem with stipulations being made about publication. I’m still unclear why we should consult publishers about mandates applied by funders.
I take Ian’s point about situations where there is no funding being different but his argument about ownership seems to overlook extant provisions for protecting IP (and agreeing the split between researcher, institution and funder) before publication.
True that publishers don’t come into the picture much before publication, but for the value of the scholarly record they are responsible for maintaining.
They do very much so on publication and protect that IP (copyright or licence – agreed between author and publisher). Investment in the (not as straightforward as some make out) process and the long term have consistently come from subs/purchase, and this business model can move to open access where appropriate if researchers are given the funding options and choice.
The principles are clearer than the practice here, as many different areas operate in different ways given a whole range of complex issues, from subject area (including pure and applied science, as well as arts and humanities), to size and funding availability, and community needs/acceptance.
I think the direction here needs to focus on enabling the researcher, rather than merely obliging funders, universities or publishers. Give them the opportunity and power to make their own principled/practical decisions and to act on them, with mechanisms for publication, feedback and measurement advanced and enhanced accordingly.
Forgive the pun, but it isn’t an open and shut case for all, though I would hope we could agree the aim is to achieve a more competitive ecosystem based on quality and independence, and built to last.
Hello Andrew. I think that if if comes to a matter of public policy, it makes sense to encourage a move towards Gold OA, because it will ultimately deliver better value for money (for the reasons above) and also because it scales better with research funding than subscriptions do. That said, it will be a complex and unstable matter to make this transition, and and there may be some types of content less suited to an OA model, such as review articles and outputs from research for which there was no grant. For these reasons, the subscription model may be with us for a while yet.
Disclaimer/disclosure: although my career has been in publishing, these opinions are my own and not necessarily those of any organization that I have worked for or may work for in the future…