Editor’s Note: Today’s post is by Bernd Pulverer. Bernd studied in Cambridge and London, followed by research positions in Toronto, Seattle and Innsbruck. Bernd was Senior Editor at Nature until 2002, Editor in Chief of Nature Cell Biology and, since 2009, of The EMBO Journal as well as Head of Scientific Publications at EMBO; he serves on the boards of bioRxiv, DORA, the PREMIER Project, CMMC and ESOF 2020.

Much has been said about Plan S, both encouraging and critical. The relative preponderance of criticism – notably by academics, who ought to be key beneficiaries – has been frustrating to the plan’s  backers, who see it as a long-overdue concerted effort to support a global transition to open access (OA) without paywalls that hamper research progress and without excessive publishing profit. Those of us who support the principle of OA –  and the life sciences journals that I represent at EMBO Press certainly do –  struggle to respond to Plan S. We are conflicted, as we want to be part of initiatives that support full gold OA as a standard. On the other hand, our priority is to preserve attributes of scholarly dissemination we consider core to the scientific process: the efficient sharing of high quality, reproducible research in a format that optimally serves science. This inevitably includes a high degree of selectivity based on peer review and a hands-on editorial process, as well as progressive publishing policies and platforms. Crucially, in the life sciences this selectivity is based on two key attributes of research papers:

  • the quality of the data and strength of the conclusions
  • the interest and importance of the research

We can only support initiatives that preserve this selectivity.

There is an inherent economic tension in OA based on author payments, as selectivity either reduces income or concentrates costs on the subset of published authors. The high article processing charges (APCs) of selective journals have to be, with some notable exceptions, diverted from research funding and risk excluding authors from their first-choice journals. Submission fees are theoretically a means to distribute costs equitably, but tend to be discounted as a viable financial model. Models relying exclusively on author payments have other shortcomings: commentaries, review and journalistic articles are not financeable through APCs. Plan S representatives have actually acknowledged that, stating that a journal may sell such content through a site license under Plan S. More importantly, there is no concrete financial support by Plan S for Open Science platforms (including preprints and databases), crucial infrastructure and quality control mechanisms (e.g., ORCID, CrossRef, PubPeer), which are currently at least partially supported by journal fees.

Replacing the access barrier to readers with another one to authors is not an option for us. Funders need to step in to remove both barriers by paying for selective, high quality publication irrespective of the cost and they must reach across geography and discipline. Using Plan S to cut costs would undermine core functions of quality publishing. Plan S proponents should focus in the first instance not on the profit margins of commercial publishers, but on the real costs of quality publishing. The money is already in the system, although much of it not under the control of the funders and the shift of the financial burden to research heavy countries represents a challenge, which is in fact better addressed by ‘Publish & Read’ arrangements such as Projekt DEAL. Those concerned by the perceived high cost of publishing might consider that at least in the life sciences, the costs represent a small fraction of overall research expenditure.

Plan S promised to grapple with all of these challenges – has it succeeded? The jury is out and emotions continue to run as high as the stakes are high: there is a lingering concern that if Plan S remains a parochial effort by few, albeit influential, funders, this may have been the last hurrah for a full OA conversion of scientific publishing. There is also concern that Plan S may actually have unintended side effects such as decreasing the actual number of gold OA papers (not journals!) by excluding hybrid journals, that it may disadvantage researchers supported by Plan S funders by restricting their journal choice, or that it may play into the hands of the dominant commercial publishing houses at the expense of smaller independent publishers, learned societies and academic institutions. Plan S recently released revised guidance (succinctly summarized by Alison Mudditt) in an attempt to address some of the concerns that have been raised. In my view key points remain unresolved.

Charge caps and gold-plated OA

Plan S continues to attempt to change too many variables simultaneously in a complex system — as a consequence it risks throwing out the baby with the bathwater to undermine broad adoption. Rather than focus on maximizing the number of gold OA papers published, the core philosophy of Plan S remains that OA can only be achieved by flipping entire journals to full OA (after a short, binding transition period). Plan S does include as a second option the posting of papers that have been enhanced through journal administered quality control processes on a parallel OA repository. This option presents a work-around for journals with a limited number of Plan S supported papers. While it will not require an OA flip for journals that regard this as an unsustainable option, it won’t be a sustainable model for journals with a significant fraction of Plan S papers. Indeed, Plan S signatories may hope hybrid journals will ultimately flip as rising OA uptake outcompetes subscription papers, but those with few European papers won’t grow OA content to such levels. Also, I would question the economic wisdom of investing into parallel platforms that essentially re-host journal content (one might call this gold-plated green OA). If such platforms can take over journal quality control processes, so be it, but as of now, this has rarely happened. Journal OA conversion is in principle more attractive and aligned with the aims of the initiative – it can certainly work if Plan S adapts to forego one key aim of its founding fathers: to cap costs at a specific charge barrier that remains to be defined even in v2 of the guidance. I outline alternatives below.

One size fits all economics undermine Open Science

The concept of a one-size-fits-all cost cap for a research paper may appear attractive at first, as it is easy to implement and police, and as it seems to help rein in those pesky profit-hungry publishers. Alas, it is misguided to base Plan S economics on averages. Setting a cap based on average costs implies that charges above the average are solely based on excessive profit expectations, rather than the real costs of high-quality selection and editorial services (in excess of $5K at EMBO Press, which is not-for-profit). Plan S has been keen to emphasize support for what it terms quality. Quality requires selectivity, which increases the costs per published paper. Thus, the revised guidance still runs the risk that high-quality publishing will be undermined, while lower selectivity OA will be supported in a sustainable way. To make matters worse, it appears likely that the admissible cost is going to be based on current average charges at OA journals across disciplines with widely divergent editorial processes and publication costs. This paradigm ignores that OA journals have proliferated in the lower selectivity sector of the market – after all, that is where APC-based OA currently works best.

The journals that are perceived by the research community to add most value to the system are highly selective and thus expensive. Just like an expensive microscope that may at first sight appear to be an excessive investment, but that may ultimately contribute much more to scientific progress than the same investment distributed across cheaper infrastructure, it is in my view essential to stop pursuing a one-size-fits-all solution for publishing. Different disciplines have entirely different needs and financial support. We need financial transparency to account for the real costs of specific publishing services and here Plan S has certainly set the right tone. Cost transparency will help the community decide what is a meaningful service to them – ultimately it should be up to those that execute research progress to have a decisive voice in shaping a system that aids them best. I am certain this will be based on OA, but not on OA at any cost.

Making Plan S a success

The Plan S initiators rightly identified that the funders have tremendous power to change how research is performed and shared – they are firmly upstream in the ecosystem of the scientific process. They have the power to decisively address many of the ills in scholarly research, including enhancing reproducibility in the biosciences by mandating data reporting standards (including FAIR data) and reducing the overreliance on journal metrics for research assessment, as pointed out by DORA. I strongly agree with this premise and encourage Plan S funders to more broadly consult with community standards initiatives and progressive journals that have already forayed into this area in order to set consistent field-specific mandates that researchers and journals can follow.

As we outlined in our feedback to Plan S, we see two attractive variants for a more granular financial model that support the stated goals of Open Science and quality:

  • Plan S would define clear quality attributes of journals that can be audited. Specific per-paper costs are awarded based on each attribute. This can be refined by discipline to set up a matrix of specific costs for specific services for a specific discipline. These attributes should, in my view, not include journal metrics, but do need to include acceptance rates. An emphasis should be placed on open science services (preprint posting, source data deposition) and progressive editorial policies (open/transparent peer review, appeals process, pre-publication research integrity screening, differentiated corrections/retractions policy, data and preprint citation, structured methods). Journals would deliver auditable reports on these attributes and charge caps awarded based on that.
  • Set up a competitive funding scheme that journals can apply to. The funding would again take into account the above quality and selectivity criteria. Decision would be made by independent experts, much like research grants. A very similar system was proposed recently by Adriano Aguzzi.

These schemes certainly require more heavy lifting than simple top-down mandates and charge thresholds, but they are required to make a complex system adapt to OA. I feel passionately that Plan S must make good on its important claim that its transformation is to reach beyond OA to Open Science. This requires new infrastructure, such as the Source Data project developed by EMBO. New infrastructure needs new investment, not cost cutting. Funds can be sourced by reducing support for journals that are not following progressive policies, but extra investment will be necessary and this will be more than justified by the truly dramatic cost savings that the dissemination of more reproducible research promises.

As Plan S is currently framed, many journals are likely to decide that the fraction of their papers subject to Plan S is sufficiently small to allow them to post this content on green OA platforms to achieve Plan S compatibility. Those of us who believe that a properly financed transition to gold OA is the right goal are left with little, if anything, tangible to give us the confidence to flip our journals to full OA. In my view, we should not lose the focus on the bigger picture, that is to support the most useful forms of dissemination at the cost required. Open Access is just one step on a road to Open Science, albeit an important one.


8 Thoughts on "Guest Post: Plan S Version 2 and the Cost of Quality"

Thanks much for the well-reasoned commentary in what has sometimes been a shouting stage.
This makes me wonder if the present open access compromises in the US isn’t so terrible. OA is presently a muddle of green OA with increasingly liberal publisher rules, preprint/author version postings, CHORUS to make publicly funded works publicly accessible, hybrid journals, gold OA, and the underrated trick of writing to the author and asking for a copy. It is a muddle, but I wonder if it’s not mostly successful at getting at least recent literature in the open, with reasonable effort on the part of the reader. Backfiles are a notable exception.

You can’t make this stuff up! Scions of industry placing price controls and limiting methods of distribution! I thought all that went out with the fall of the USSR and China’s opening of China. Biomed Central was started as a for profit business and still is while PloS is still living off the good aegises of benefactors. We have green, gold, platinum and I hear rumors of heliotrope not to mention ochre and my favorite chartreuse. All in an effort to force academics to publish via avenues they don’t want to travel upon.

If you put diamonds and gold in a mechanical watch, of course it becomes more expensive.
However, scholarship wants and needs digital watches for 10% of the price of mechanicals with diamonds and gold – and hooked up to a time server to make them more accurate to boot.

I suggest institutions provide their scholars with digital, accurate watches – let’s see how many still want to pay from their own money for the gold and diamond inaccurate watches.

The problem with your argument is that the value of a journal comes from the rigour if its peer review process. And this is expensive because top experts won’t give their time for free to hugely profitable multinational corporations… oh wait.

Yep, most reviewers work for free. So do a lot of editors. The money from publishing currently goes goes to corporations, and some of it to scholarly societies, although not much. Plan S will not change this distribution of benefits unless they favor more radical options – a really big cutback in APCs affecting corporate profits, or financial flows or subsidies from the big publishers to volunteer and non-for-profit journals [like mine]. So we can pay the archiving costs that that require, and maybe xml coding if that is forced on us. Dream on.

Science has generally changed through incremental steps. Likewise, scholarly scientific publishing has mostly developed along an evolutionary arc. For instance, the shift from print to online journals was fairly gradual. Here, as pointed out by the author, people are trying to change (drastically) a complex system by fiat. Not surprisingly, it is not going according to plan. Even Pulverer’s suggested solution bumps into the complexity of the system where he proposes a “matrix of specific costs for specific services for a specific discipline”. Working at Mathematical Reviews and MathSciNet, where we index and curate the research literature in the mathematical sciences, I have learned that the boundaries between specific disciplines are quite blurry. Where, for instance, would the matrix put these journals: Bulletin of Mathematical Biology, European Biophysics Journal, Proceedings of the Royal Society A (Mathematical, Physical, and Engineering Sciences), or, for that matter, Science and Nature?

To amplify a point made by Pulverer, at the recent ICIAM in Valencia, there was much discussion by the leaders of various mathematical societies about the potential harmful effect that Plan S could have on the societies, for whom publishing revenues enable them to provide a host of other services, such as conferences and professional development. This unintended consequence bears resolving before full implementation of the plan.

I can offer no specific solutions or alternatives, other than to recommend that we try to steer the development of scholarly publishing rather than try to change vehicles while traveling at speed.

My apologies for the tangential nature of this question, but I’m curious to know if anyone knows if the major databases like JSTOR, MUSE, EBSCOHost have been brought into the Plan S discussions in any significant way. What I’ve seen in terms of business model reactions to Plan S and OA more generally (read and publish, publish and read, subscribe to open) seem to only take into account replacing revenue that’s lost from direct subscriptions with libraries and doesn’t acknowledge losses that would occur related to licensing. Perhaps this is a picayune question, but I suspect it’s relevant particularly to smaller publishers and university presses that may have distribution agreements with these organizations.

A delayed response: We are a very small (one journal) publisher. We are connected to MUSE through the intermediary of our UP distributor, which tells us that MUSE has not communicated any response to Plan S. This is not picayune for us, as the majority of our (admittedly tiny) revenue comes from MUSE and JSTOR downloads. Libraries and individuals who have access to MUSE/JSTOR have largely canceled their direct subscriptions. I am under the impression this is typical for the humanities. I have heard no commentary on this and would welcome it!

Comments are closed.