Abstract
Open Access mandates in Europe raise the question if the priority is to reduce publishing costs, or the overdue conversion to Open Science communication. At risk are not only high‐quality journals, but also community institutions and international research collaboration.
Nineteen years after the launch of the first Open Access (OA) journals in the life sciences (BioMed Central in late 1999, PLOS Biology in 2003, followed by EMBO Press's first OA journal Molecular Systems Biology in 2005), 3–10% of the literature is published in OA Journals (also known as “Gold OA”) OA. Notably, PLOS Biology, Molecular Systems Biology, EMBO Molecular Medicine, and eLife remain among a relatively small group of highly selective OA journals. Repeated calls for universal OA, starting with the PLOS call of 2000 and the subsequent Budapest and Berlin declarations have led to a steady rise in OA journals and the broad adoption of OA publishing options at the majority of subscription journals (often called the “hybrid model”, which allows OA based on demand). Some large commercial publishers like Springer Nature have bought into OA platforms and highlight their OA commitment, yet a wholesale OA conversion of the scientific literature remains elusive. Why?
There are several reasons:
Payment of author‐contributed “article processing fees” (APC) is the predominant financial model for OA, yet few funders and institutions specifically set aside money for this—partially because money earmarked for publications remains sequestered in long‐term subscription arrangements. From a researcher's perspective, it is hard to justify diverting scarce resources for research to a publication process if the subscription model provides a free alternative to them.
The bulk of the publishing cost is due to editorial and peer review selection, quality control, and enhancement of manuscripts. Thus, the cost of APCs has to scale with a journal's degree of selectivity. It is not surprising that there are only a few highly selective OA journals.
“Green OA” has been touted as an alternative to “Gold OA”: The manuscript is posted on institutional or international platforms with an open access and reuse licence but with an embargo (often 6 months) and usually not in the final version published in the journal. Publishers have almost universally accommodated Green OA, and many regard this as a viable alternative to full Gold OA, even if it represents a compromise.
There has been limited coordination between funders and research institutions in setting OA policies. Many respect academics’ freedom to choose where to publish.
OA—at any cost?
The disadvantage of Green OA is that large sums are invested in setting up parallel platforms to journals, which often host preliminary versions of papers locally without integration, and which are consequently not widely used. Surprisingly, the economic impact of the fragmented and often derivative repository landscape that runs in parallel to journal platforms has barely been studied. Local repositories usually lack standardization and interoperability, and Green OA articles typically lack the integral metadata and linking structures of the published journal article. Access to global OA repositories such as PubMed Central is increasing, although this diverts it from the scholarly version of record—the journal‐published research paper. Preprint servers such as bioRxiv provide a Green OA implementation that adds value by accelerating the dissemination of research findings (see Pulverer(2016)EMBOJ 35:2617–2619).
Gold OA is certainly the most efficient and effective mechanism for OA publication and best addresses the goal of sharing findings in an accessible, reusable, and consistent manner. As long as the existing publishing costs can be diverted from site‐licence costs to OA publication, the overall costs of Gold OA should be covered. However, Gold OA is not necessarily the most equitable system: The risk is that one flawed paywall—that to the reader—is replaced by another indefensible one—that to authors. Since a much smaller number of authors than readers have to shoulder the whole cost of publishing, it accentuates imbalances in funding across nations, research disciplines, and laboratories.
While an author‐based APC model has increased the visibility of the per‐paper charges, and with that the potential for a self‐regulating free market to limit low “value for money” in publishing, it has generally not increased the transparency of what these charges pay for. Nonetheless, assuming authors have adequate funding for publication, this could make for a new economy where authors choose a journal not merely based on its perceived reputation, but also the value added through selection, improvement of the research, dissemination, and archiving. Alas, in reality we are not in a truly free market, as journal name matters tremendously for career advancement. As launch members of the San Francisco Declaration on Research Assessment (DORA), we have called attention for some years to the flagrant misapplication of pseudo‐quantitative journal metrics (most notably the journal impact factor; Pulverer (2015) EMBO J 34: 1601–1602) and journal name to research assessment. Where a research article is published makes a dramatic difference in terms of the value that the scientific community, research institutions, and funders assign to it, and thus, for the rewards its authors receive in terms of funding and career progression. Thus, for now journal costs and indeed OA remain a secondary criterion compared to journal reputation, at least for those who can afford the costs.
Policy makers and funders have realized the inefficiency and bias of the “author pays” model for Gold OA and are now shifting to so‐called “Read & Publish” models, where funders, research institutions, or national consortia pay centrally for Gold OA, usually on a simple per‐paper basis. This makes for a buyer's market, where large consortia can dictate the costs of publishing. This is fine as long as there is a clear understanding of, and focus on, quality over quantity. In times of concerns about reproducibility, spin, and serial examples of compromised research integrity, the value of journals that apply state of the art quality control mechanisms, that ensure balanced, transparent and detailed reporting of research findings should be obvious. Journals have taken on a crucial role as the last, and often main, quality checkpoint in the scientific process. Alas, instead of quality, OA consortia appear to focus on cost reduction, often citing the large difference between the average cost of papers in OA journals to those in subscription or hybrid journals (see Schimmer et al (2015)). Those who conclude that the current average costs of Gold OA APCs provide evidence for a legitimate ceiling to OA charges apparently willfully ignore that OA journals are, on average, less selective and run on lower overheads than highly selective, value‐added journals. Emphatically, OA journals of course do not have to be at all of lower quality (as epitomized by the journals listed above), but the APC business model scales best at lower selectivity and is therefore enriching for it (an excess of this is the booming trade of predatory publishers).
The cost of publishing
Leaving aside the often dramatic profit associated with commercial publishing (note: all EMBO Press publications are wholly owned by EMBO, a not‐for‐profit organization that invests all journal income in support for scientists)—and despite frequent claims otherwise—the actual costs of publishing a research paper are ultimately similar if they are published OA or via a subscription model. Only the mechanism of settling the costs may change and with that the visibility of charges that were previously sequestered in opaque multi‐year subscription arrangements between libraries and large publishers.
The fees charged by this journal for OA ($5,200) are substantial because selection is for quality as well as the interest and importance of the work—12% of submitted papers are published, raising the unit costs. To put the cost in perspective, consider the total research investment that underlies the average bioscience research paper—the limited analysis available suggests charges are in the order of 1–3% of research spending. To put the acceptance rate into perspective, consider that it is similar across a broad range of selective journals and in fact similar to award rates of competitive grants and positions. This strong systemic selectivity will likely only decrease if fewer scientists compete for limited funds and positions. It may seem like an easy way to reduce costs is to reduce staff overheads by devolving more of the work to academic researchers, but note that this will incur similar costs, albeit indirectly via research salaries rather than OA fees.
More OA journals or more OA?
Librarians who have suffered through years of reduced budgets while facing chronic increases in subscription fees are understandably relieved that a concerted effort in Europe now aims to catalyze Gold OA conversion of journals: cOAlitionS (also known as PlanS), an initiative supported by 13 national funders, the EC, and the Wellcome and Gates foundations, and a number of national initiatives such as the DEAL consortium of about 200 German Science Organizations mandate Gold OA with capped charges.
The centrally funded “Read & Publish” models that DEAL and other national consortia pursue allow negotiation of cost reductions by brute force and generally fail to differentiate between different research outputs (every published unit is eligible for the same compensation). Notably, these negotiations are focused on the largest for‐profit publishers, which are more able to absorb cost reductions across their broad journal portfolios than small non‐profit community publishers, which are left high and dry.
PlanS derives from a position that Gold OA is fundamental principle of the scientific process and requires top‐down mandates since the “hybrid model” has failed as a transitional model to full OA. PlanS confirms the importance of quality in the dissemination of research (including COPE standards), author access to journals irrespective of their financial means and author copyright retention. PlanS requires linking to underlying data and code, open citations and it regards Open Science platforms beyond journals as legitimate means to share research (after peer review and revision and without embargo). EMBO and EMBO Press agree with these goals (see http://embo.org/news/articles/2018/response-to-science-europe-s-open-access-plan). However, PlanS has at the same time set constraints on the OA conversion: article charges are to be capped (although additional guidance issued later states that ‘fair and reasonable’ publication costs will be determined by an independent study ‘to help inform the potential standardization of fees’), and, as written in the current guidance, hybrid journals, that is journals with optional Gold OA, are effectively to be excluded after a transition period.
EMBO Press has three OA journals that demonstrate that OA publishing is in principle possible for high‐quality, highly selective journals. However, two EMBO Press publications, EMBO Reports and The EMBO Journal, for now rely on optional OA. Why? Because we cannot countenance raising costs beyond 5,000€/paper, as it would exclude researchers with limited funding. Note also that APC‐based OA does not support reviews and commentary in our journals. EMBO Press is poised and ready to go fully OA if the conditions are right—that is, if we receive adequate financial support for a high‐level editorial and peer review process, without excluding authors from the journal.
PlanS—OA catalyst or diversion from OS?
The goal of these initiatives should ideally be to increase the proportion of Gold OA papers. Instead, the stated goal of PlanS is to increase the proportion of OA journals. This could precipitate a number of unintended consequences, including a lower OA rate, the potential exclusion of European authors from up to 80% of journals, discouraging collaboration with European researchers, replacement of high value‐added journals by quantity publishing, or the demise of long‐standing community institutions reliant on journal income. Most importantly, forcing down costs of publishing is likely to discourage journals from innovating to embrace Open Science (including hosting and curating of source data, reproducible protocols and reagent identifiers, deep interlinking between datasets, papers and preprints, and advanced search technology).
The re‐energized discussions around OA and the sense that broad adoption will only happen when mandated by funders make for a fresh breeze of change. We should seize the momentum to go one step further to the more important goal of Open Science—all the way from bench to published page and repository. Publishing has admittedly been lagging behind digital innovation elsewhere, partially due to high‐profit margins and lack of incentives to change. PlanS should catalyze a change to Open Science, not stop at the goal of OA at a cost that will hamper progress to Open Science by which research will be rendered far more impactful and useful. For science, an excellent return on investment beckons.
Note: A subsequent editorial will summarize how we embrace Open Science publishing principles at the EMBO publications.
The EMBO Journal (2018) 37: e101215