The Life and Death of an Open Access Journal: Q&A with Librarian Marcus Banks

Librarians have been at the forefront of the open access movement since the beginning, not least because in 1998 the Association of Research Libraries (ARL) founded the Scholarly Publishing and Academic Resources Coalition (SPARC). Today SPARC is arguably the world’s most active and influential OA advocacy organisation.

Marcus Banks

It is important to note that librarians’ interest in open access grew primarily out of their frustration with the so-called “serials crisis” — the phenomenon that has seen the cost of scholarly journals consistently grow at a higher rate than library serials budgets.

SPARC’s initial strategy, therefore, was to encourage the growth of new low-cost, non-profit, subscription journals able to compete with the increasingly expensive ones produced by profit-hungry commercial publishers. As SPARC’s then Enterprise Director Rick Johnson wrote in 2000, “In 1998, after years of mounting frustration with high and fast-rising commercial journal prices, a group of libraries formally launched SPARC to promote competition in the scholarly publishing marketplace. The idea was to use libraries’ buying power to nurture the creation of high-quality, low-priced publication outlets for peer-reviewed scientific, technical, and medical research.”

In the wake of the 2002 Budapest Open Access Initiative (an event attended by Johnson), however, SPARC began to focus more and more of its efforts on open access. The assumption was that this would not only allow research to be made freely available, but finally resolve the affordability problem faced by the research community. As the BOAI text expressed it, “the overall costs of providing open access to this literature are far lower than the costs of traditional forms of dissemination.”

Ironically, despite their high profile advocacy for open access many librarians have proved strangely reluctant to practice what they preach, and as late as last year calls were still being made for the profession to start “walking the talk”.  

On the other hand, many librarians have embraced OA, particularly medical librarians. In 2001, for instance, the Journal of the Medical Library Association (JMLA) began to make its content freely available on the Internet. And in 2003 Charles Greenberg, then at the Yale University Medical Library, launched an open access journal with BioMed Central called Biomedical Digital Libraries (BDL). One of the first to join the editorial board (and later to take over as Editor-in-Chief) was Marcus Banks, who was then working at the US National Library of Medicine.

Four years later, however, BDL became a victim of BMC’s decision to increase the cost of the article-processing charges (APCs) it levies. This meant that few librarians were able to afford to publish in the journal any longer, and submissions began to dry up. Despite several attempts to move BDL to a different publishing platform, in 2008 Banks had to make the hard decision to cease publishing the journal.

What do we learn from BDL’s short life? In advocating for pay-to-publish gold OA did open access advocates underestimate how much it costs to publish a journal? Or have publishers simply been able to capture open access and use it to further ramp up what many believe to be their excessive profits? Why has JMLA continued to prosper under open access while BDL has withered and died? Was BDL unable to compete with JMLA on a level playing field? Could the demise of BDL have been avoided?  What, if anything, does the journal’s fate tell us about the future of open access?

I discuss these and other questions with Banks below. The issue of affordability, it seems to me, is particularly apposite, as librarians are having to confront the harsh truth that, far from reducing the costs of scholarly communication, open access appears more likely to increase them.

It turns out that Banks has an interesting perspective on this issue. As he puts it, “At the risk of frustrating many librarian colleagues, I must say that the framing of open access as a means of saving money has been and remains a serious strategic error.”

He adds, “A fully open access world may not save any money and could cost more than we pay now — this world would include publication charges as well as payments for tools that mined and sorted the now completely open literature. That’s fine with me, because in this world we’d be getting better value for money.”

The interview begins …

RP: Can you say something about your background and career to date?

MB: I have been a librarian since 2002. My first position after earning my Masters of Library and Information Science was as an Associate Fellow at the US National Library of Medicine (NLM), from 2002-2004. During this time NLM was developing PubMed Central (PMC) as a freely accessible digital archive of biomedical literature.

Growth at PMC was slow, as deposits to it were voluntary — this was years before PMC became the required repository under the terms of the NIH Public Access Policy. Publishers rightly worried that a fully open access archive would challenge their business model, a concern that persists today.

Watching this debate unfold raised my awareness of the various agendas in scholarly publishing, as well as of the potential for open access publishing to expand the reach of biomedical literature.

RP: What are you doing currently?

MB: My most recent position was as the Director of Library/Academic & Instructional Innovation at Samuel Merritt University in Oakland, California. Since then my wife and I have returned to the Chicago area for both personal and professional reasons. I am currently pursuing employment while building a consulting practice devoted to transformation in scholarly communication. Even with “gainful employment” I would continue the consulting.

RP: You said that the growing debate about scholarly communication made you aware of the potential for open access publishing. You were later involved in the creation of an open access journal called Biomedical Digital Libraries, which I think was launched in 2004 but ceased operations in 2007. Can you say what your role at the journal was, why the journal was created, and why it did not succeed?

MB: Charles Greenberg, then at the medical library of Yale, launched Biomedical Digital Libraries (BDL) at the Medical Library Association meeting in May 2003. It was an open access title published by BioMed Central (BMC). His first task was to recruit an editorial board, and I joined in as an Associate Editor. Our first papers appeared in 2004. As Charlie moved on to other projects, I became co-editor and then sole Editor-in-Chief in 2006.

Charlie really believed that librarians, as vocal proponents of open access, should practice what they preach by launching a digital open access journal. At this time our association publication, the Journal of the Medical Library Association (JMLA), was becoming open access and thus was also "walking the walk." But JMLA kept a print production schedule even though articles also appeared online (this is not the case anymore, there are early release online articles).

We offered rapid peer review and publication as soon as the article was ready. We also had a larger core author base than JMLA — data mining papers from computer scientists would have been very welcome in BDL. That said, the majority of our authors came from academic libraries.

BMC published BDL for 4 years. During the beginning of that tenure many academic institutions provided generous memberships that fully, or mostly, covered the author's cost of publication in BMC titles. Librarians do not generally have large research budgets that can cover OA publishing fees, so this was vital to our primary author base. (JMLA publication was then, and is still, free of cost to authors.) BMC’s institutional membership costs were affordable, and predictable, in those first years. The costs were flat each year, and easy to budget.

Then BMC went to a scaled model, in which the cost of an institutional membership would rise in proportion to how many BMC publications from a given institution appeared each year. (This has likely changed significantly by now, this was the situation in 2006-2007). The more you publish, the more you pay, and there was no way in advance to know what the final bill would be.

This caused library directors, who generally paid for the BMC memberships, to cancel them in large numbers. Once those memberships went away so too did publication subsidies. Our primary author base, which was not in a position to pay more than $1,000 per paper to publish, dried up. The journal stopped publication on BMC as of the end of 2007. 

Although in some respects this is a “failure,” we did some interesting things. Along the way we went to open rather than anonymous peer review, which is still a fairly novel approach. We also published papers about “altmetrics” years before that term existed. It was a good run.

RP: I know that in 2007 a number of institutions cancelled their BMC membership, including Yale, citing “skyrocketing” increases. Do you think perhaps when they began to advocate for pay-to-publish gold OA open access advocates underestimated the cost of publishing a journal, or are those who argue that publishers overcharge nearer the mark?

MB: I’d have to side with the argument that publishers overcharge, including open access publishers. The PeerJ innovation shows that smaller charges can sustain a quality journal, and most of the journals listed in the DOAJ have no publication charges and yet they exist.

The fact that traditional publishers absorbed OA titles into their portfolios — like Springer with BioMed Central in 2008 — shows that there is a profit margin in OA. This is not completely bad; the world needs commerce, and at least the profitable OA titles are available to all online readers while the profitable subscription titles are restricted. But I’m not convinced that the author fees charged by some OA publishers reflect the actual costs of publication.

What do we learn?

RP: To what extent do you feel that BDL’s fate was a consequence of the field in which it published — librarianship?

MB: Not much. Many librarians have tenure obligations and are required to publish. JMLA continues to thrive, as do other library journals. In a way BDL was an outlier publication in BMC, a library-focused journal on a platform that generally published hard science.

RP: After it ceased operating as a BioMed Central journal you tried to transfer BDL to an Open Journal Systems platform, and then to the University of California’s eScholarship platform. Why did neither of these attempts succeed either?

MB: These attempts occurred in 2008. Their main advantage was no author fees for publication. But by this time I had maintained my own blog for three years, almost as long as my time with BDL. I was finding blogging to be a more effective means of sharing insights and research about librarianship, which at the time sported a vibrant "biblioblogosphere." So I argued that library journals should evolve into blogs, something that is still possible today although the credentialing power of formal publication cannot be discounted.

Given my circa-2008 feelings about the potential of blogs I became disinterested in propping up the journal model. Meanwhile, our editorial board members had moved on to other projects.

RP: Where do you stand today on the issue of whether journals should be replaced by blogs?

MB: Conceptually I still believe that a well-managed community blog could serve the same function as a journal, and disseminate new insights more quickly than a journal. Culturally, though, there is still tremendous branding and credentialing power from getting a paper in the “right” journal.

Given this I see blogs as complementary or additive to journals for academic publishing, at least at the present time. That said, in some cases blog posts will also be part of the scholarly record. OCLC has done great work recently on preparing a Framework for the Evolving Scholarly Record, which makes this same point about how new content types are now part of the record.

RP: Your vison of journals evolving into blogs is subject-specific right? It wouldn’t be realistic to publish, say, general medical papers on blogs?

MB: Indeed, receptivity to non-standard scholarly items such as blog posts is subject-specific. General medical journals are likely to assume, and value, their established forms for years to come. I must note, though, that one of the principal justifications for this established form is because of the ostensible power of peer review to sort the wheat from the chaff. This is despite evidence that pre-publication peer review can be too cautious and slow.

As Kahneman and Tversky demonstrated, we are more afraid of losing what we know (in this case, peer review) than we are excited about embracing what we do not yet understand (in this case, different ways to do science).

RP: How did you plan to cover the costs of publishing BDL on OJS?

MB: We were hosted by Scholarly Exchange, which initially offered hosting free for a year (as I recall) and then with affordable rates after that. Within that first year I made the switch to the University of California eScholarship platform, which was permanently free as long as it had sponsorship by a University of California entity (which we had from UC San Francisco.)

It appears that Scholarly Exchange is still operating, although their web site has not been updated for a few years.

RP: What do we learn from the failure of BDL to prosper in terms both of open access and of scholarly communication more generally?

MB: Although leading OA titles like those from BMC or PLOS charge author fees, many of the titles that appear in the Directory of Open Access journals do not, as I noted earlier. I also pointed out that there are now new funding models, such as PeerJ’s lifetime memberships.

So the question of financing for open access titles — and what exactly this financing supports (is it the actual publication? or the shareholders of companies with OA in their portfolios?) — is an unsettled area that will probably always be vexed. At least the traditional subscription model, for all of its shortcomings, is able to sustain an enterprise.

That said, the PeerJ innovation shows there is enough energy behind OA to be entrepreneurial. And I feel that within the next decade or two immediate OA to biomedical content, with no publication embargos, will be required by law. If this is true the funding streams to support such a law will materialise.

Costs

RP: Based on your experience then, how much would you say its costs to publish an open access journal?

MB: I’ll need to dodge this question somewhat, as my experience as the editor at BDL was removed from the balance sheets available to BMC. All I knew was that I had a set number of author fee waivers I could grant each year, which I generally used for authors from less-wealthy nations. I relied on the balance of authors to come from institutions that mostly (or ideally, completely) subsidised publication.

Beyond my experience, the reported figures for publishing articles open access vary widely. Outsell’s average cost for a publishing an OA article in 2011 was $660, but during that same year Nature reported that it would cost $30,000-$40,000 per paper to go completely OA. This figure beggars belief, no matter how selective a journal may be.

Less insidious are claims from many publishers that it is hard to isolate the mean cost of publishing an open access paper, as these costs are absorbed within complex business entities and/or cross-sudsidised by other parts of a publishing enterprise (as at PLOS, where PLOS One papers underwrite papers in other journals.)

But even this explanation is unsatisfactory. My view aligns with Andrew Odlyzko’s, which is that article publication costs of $400/article should be sufficient. Any figure appreciably higher than that is likely a reflection of a publication’s perceived non-substitutability in the market, and not the actual costs of publication.

RP: I assume Nature’s figures are based not on how much it costs to publish a journal, but on what Nature estimates it would need to charge in order to maintain its current revenues if it migrated to an OA model. In 2004 Richard Charkin, then CEO of Macmillan, put it this way to British MPs: “in order to replace [Nature’s] revenues you would have to charge the author somewhere between £10,000 and £30,000 because the costs of editorial design and support are so high.” Be that as it may, you said that most of the journals in the DOAJ have no publication charges. I wonder how indicative that is of how OA publishing is likely to develop in the future. First, to get an accurate picture one would presumably need to count the number of papers published that require payment of an APC, not the number of journals? Second, do you not think that we will see the traditional commercial publishers take control of the OA publishing market — that is certainly how I interpret the recent blog post by the Wellcome Trust’s Robert Kiley. (See Table 3 in particular) What is your view?

MB: Charkin’s quote assumes that all of the incumbent practices from the print-based publishing model must be maintained in their entirety, and in perpetuity, even in a digital-only publishing environment. The Odlyzko paper I cited above proposes many ways that digital publishing can be made more cost-effective and efficient.

As to Kiley’s post I read it somewhat differently — his main critique is with the practices of hybrid journals, which charge much higher average APCs than fully open access journals. That conclusion reinforces the similar finding of Solomon and Bjork in their Wellcome study one year ago. I think that, if every journal were fully open access, APCs would be lower on average and thus funder dollars would stretch further.

But maybe that is a distinction without much consequence, as your main point is that commercial publishers are likely to exploit the move to open access as a vibrant profit center. I agree, and in principle this does not concern me. At least in this universe publisher profits would support immediately available OA literature, rather than the two-tiered, embargoed system we have now.

That said, in my perfect universe the prices publishers charged to produce open access articles would be regulated just like electricity or water rates. In the main bioscience publishers are offering research that derives from government or non-profit funds that are meant to improve public health, which would be the legal basis for such a regulatory scheme.

Of course such a prospect is very dim, given the political climate against government regulation that is particularly virulent in the US right now. And even in a pro-regulatory environment publishers would point to their multinational character as proof of being impossible to corral since capital knows no borders.

Short of regulation, funder pressure to drive down the cost of APCs (perhaps via APCs caps at hybrid journals, as Kiley references in his post) may work. And if that does work, regulation won’t be needed.

RP: What do you think the implications would be if the transition to open access saw no reduction in the overall cost of scholarly communication? I know you have argued that open access is not about solving the serials crisis, but if we are to believe librarians scholarly communication is in the grip of a serious, and long-standing, affordability problem. If open access cannot offer cost savings then presumably that problem could only intensify in an OA environment surely?

MB: At the risk of frustrating many librarian colleagues, I must say that the framing of open access as a means of saving money has been and remains a serious strategic error.

As Stewart Brand said, information wants to be free and expensive at the same time. Publishing only on the free side of the equation has enabled librarians to be painted as free-loaders who don’t wish to pay for valuable products. It has also spawned publisher propaganda that the role of librarians should simply be to increase their budget bids so that everything can continue as normal.

A fully open access world may not save any money and could cost more than we pay now—this world would include publication charges as well as payments for tools that mined and sorted the now completely open literature. That’s fine with me, because in this world we’d be getting better value for money.

Sense of injustice?

RP: Just to confirm: JMLA is now an open access journal? I could not find a web page for the journal?

MB: JMLA is currently open access. The web site is here.

RP: The site you point to is a section of PubMed Central. So the journal doesn’t actually have its own site; it is publishing directly on PubMed Central. This, as you may know, has attracted criticism, notably from Kent Anderson on The Scholarly Kitchen. Anderson makes a number of points, but for our purposes let’s just consider his claim that this has allowed the MLA to save “between $200,000 and $500,000 (or more) in hosting costs” over a ten-year period. If he is right, it suggests to me that BDL was at a significant disadvantage from the start. After all, if its primary competitor had to pay no hosting costs then it was not really a level playing field for BDL was it? Or would you question Anderson’s claim?

MB: Speaking as an author with several publications in JMLA, I see the merit in Anderson’s claim that JMLA has unequal access to PMC servers vis-à-vis the access provided to other publishers. I’ll have to trust Anderson’s figures regarding the saved hosting costs, as he is a publisher and I am not.

That said, though, Anderson’s claim of JMLA “freeloading” can also be applied to subscription publishers who benefit extensively from the access-control mechanisms set up by universities. As the library director at Samuel Merritt University (SMU) I coordinated closely with my colleagues in Information Technology to ensure that only SMU affiliates had access to our licensed content.

We proactively alerted alumni that they could not access this content, because our licenses did not permit it. If any holes opened up within our walled resources, we closed them. This was all at our and IT’s expense; it is not as though publishers paid SMU for a “licensed content control specialist.”

In his piece Anderson goes to great lengths to detail an analogy about getting into a concert even though you have not paid for a ticket — something that may not cause identifiable harm, but is still unfair. At the outer limit of the analogy, a police officer is granting him the free ticket and then driving him to the concert too. What a sweet deal.

Alluding to JMLA, Anderson says, “Everyone else has to pay for these things, but I don’t. I’m literally getting a free ride.” Publishers enjoy such a free ride every day, in the form of uncompensated labour at universities to preserve subscription-based access arrangements.

RP: Actually I was more interested in your perspective as the former Editor-in-Chief of a journal that was not able to compete with JMLA. Assuming Anderson’s claims are correct, do you not feel a sense of injustice over the favoured treatment afforded JMLA? Might BDL have managed to survive if it had also been able to use PubMed Central as its publishing platform?

MB: Yes, if BDL had been on PMC it may have lasted longer and possibly still exist today. But it also could have lasted longer if we had attracted a wider author base of data scientists etc. I feel no sense of injustice. BDL was an experiment at the dawn of widespread open access publishing, and we learned some useful lessons from it.

As some commenters noted in response to Anderson’s post, the NLM has long had a unique relationship with MLA. To me this explains JMLA’s arrival on PMC; there was no smoking-gun plot of the kind Anderson conjures. He is on stronger ground with his criticisms of PMC’s more recent arrangement with eLife. However, by now PMC is the archive of record for the NIH Public Access Policy.

If NLM/PMC are tipping the scales in favour of open access they are on solid policy ground to do so. I’ve long marvelled at the fact that anything written by a government employee in the scope of their employment is public domain (with the exception of classified material, of course). But anything written by a government independent contractor — which is effectively what an NIH grantee is — can be privatised.

RP: From what you said earlier, I assume JMLA does not charge an APC. How then does it fund its operations?

MB: JMLA does not charge an APC. Any costs of its production are paid from Medical Library Association funds, which are principally derived from member dues and annual conference revenues.

RP: When I contacted Charles Greenberg about BDL he stressed that he did not have an opinion about JMLA and why it went open access, but he gave much the same reasons as you for the failure of BDL. He then said, “I think it would not surprise anyone that grants that cover open access APC charges do not go to librarians, though librarians are partners on more grants, such as CTSA grants that did not exist at the time of BDL's demise.” We could note that librarians are not the only group in academia who are discovering that grants to cover open access APC charges are not going to go to them. Do you think we could we see a growing number of researchers unable to publish their papers as a result? That is what this blog post by professor of behavioural and evolutionary ecology at UCLA Peter Nonacs would seem to portend?

                

MB: Nonacs’s post points to the structural deficiency of the author pays publishing model. As he notes, in the open access environment he now finds himself in the “new poor” along with (among others) librarians who wish to publish.

The only systemic solution is a dedicated funding stream for publishing charges that comes from an institution or funder’s research budget; in proportion to the overall budgets that fund research open access publishing charges are small.

This is much easier said than done given the realities of institutional budgeting practices. Nonetheless there is enough money overall to support immediate, non-embargoed open access — certainly so in the biosciences.

Licensing

RP: I believe BDL published its papers under a CC BY licence. As you will know, licensing has become a contentious issue in the OA movement. Is it the CC BY licence you favour for scholarly research? If so, why? If not, why not?

MB: My preference would be for authors to choose the license that applies to their papers rather than adopt a blanket license adopted by their publishers. Publishers, and librarians, are effectively agents of authors. The intellectual determination regarding under what conditions work can be re-used should rest with those who created it.

RP: Many argue that only scholarly content that has been released under a CC BY licence can claim to be open access. You do not agree?

MB: I do not agree. My view aligns with that of Peter Suber in his Very Brief Introduction to Open Access: “Open-access (OA) literature is digital, online, free of charge, and free of most copyright and licensing restrictions.”

RP: I note that, while JMLA indicates that the copyright in the papers it publishes belongs to the authors, it does not specify a licence. Indeed the PDF files appear to have no copyright notice at all (e.g. here). That would seem to imply “all rights reserved” would it not? Or at least anyone wanting to republish or reuse them in any way would have to assume that wouldn’t they?

MB: That’s a reasonable assumption, but not true in JMLA’s case. The license is effectively CC-BY-NC, as shown here in the author copyright form. The crucial clause:  “Copyright in all articles appearing in the Journal of the Medical Library Association is owned by their authors. Readers may copy articles without permission of the copyright owners, as long as the author and the Medical Library Association are acknowledged in the copy and the copy is used for educational, not-for-profit purposes.”

RP: I wonder if it is reasonable to expect that anyone wanting to republish, reuse, or perhaps text mine, open access papers should have to hunt down an author copyright form (or contact individual authors) in order to establish whether they can do what they want to do with a paper. I think JMLA is not the only OA journal to be vague or unforthcoming about licensing. Should not all OA journals be totally transparent about what can and cannot be done with the papers they publish? That, after all, was partly the reason for developing the Creative Commons licences.

MB: I agree with you completely. This information should be more concisely and consistently displayed on all open access articles.

Lacunae?

RP: Some commentators have suggested that as open access publishers experiment with new business models a number of them will inevitably fail. This, they say, could see parts of the scholarly record disappear — which is what Jeffrey Beall argues has already happened with some of the journals published by VictorQuest Publications. (The same concern was raised here). In the case of BDL, back copies have been archived on the BMC site, and I think they are also available in PMC. But do you think we could see a spate of open access journals fail, and that this could see lacunae in the scholarly record? Or is the danger no greater than with traditional subscription journals in your view?

MB: I am wary of Jeffrey Beall’s critiques. His useful service of identifying predatory publishers is now marred by his categorical and strident rejection of the value of all open access publishing, even the non-predatory sort. Of course he can espouse this view, but his manner of doing so generates more heat than light.

That said, there is a genuine concern that open access publications with short digital shelf lives could be lost to history (as happened with BDL, the one and only article we published on Scholarly Exchange is now lost to time. You are correct that all the BMC articles are in PMC.)

There is probably a greater risk of this happening with open access titles than with subscription-based publications, which generally have longer track records and/or are managed by profitable companies with a business interest in maintaining this record. This is not an intrinsic fault of open access as a concept though, but rather a reflection of the inherent risks of attempting to challenge and change the long-established business model for scholarly publishing.

RP: OA is a hotly contested topic. In a recent blog post you argued that both publishers and open access advocates distort the truth when making their respective cases. As you put it, “Open access advocates elide the filtering and credentialing role of publication in particular journals, which saves researchers very valuable time in deciding what to read. Subscription-based publishers obfuscate the fact that the marginal cost of posting an article is zero, in order to perpetuate a scarcity-based economy that was necessary for printed materials but is now antiquated.” You also point out that this jostling and obfuscation has now been going on for 15 years without common ground being reached by the two warring factions. How do you see this playing out? Do you, for instance, believe that open access is — as Stevan Harnad put it 18 years ago, “The Inevitable and the Optimal”? If so, how will it eventually be achieved and how much longer will it take?

MB: I agree with Harnad that OA is inevitable, as funder mandates will continue to reduce and eventually eliminate embargoes. This may take a decade, maybe two — hopefully less than either. I have no crystal ball but that writing is on the wall.

That said, I do not believe that OA to the scholarly literature is optimal. It’s just a start. Allow me to quote from that same Medium piece if I may:

“Open access advocates are right that information wants to be free, and publishers are right that information wants to be valuable. That value manifests when information becomes insight. This is value worth paying for. This is how publishers could thrive even in a completely open access world.

“In such a world scholarly papers would be considered a natural resource, like the air. Nobody can meter access to the air, but anyone can build an airplane to glide through it and build a business from there. Let’s glide along ourselves, and move from scarcity-based to insight-based economics for scholarly publishing.”

RP: Thank you for answering my questions. And good luck for the future.