Editor’s Note: Today’s post is by Mariëlle Prevoo, open knowledge librarian Maastricht University Library, Ron Aardening, scholarly communication officer Maastricht University Library, and Ingrid Wijk, director Maastricht University Library.
Imagine a university invoicing all graduating students for both the costs of their study program and the tuition fees of their peers who dropped out along the way. While this situation would strike most as unfair, something analogous happens in the world of scholarly publishing through the charging of open access fees. In this post we will explore how restructuring APC (Article Processing Charge) pricing can lead to fairer cost allocation in scholarly communication.
Open Access publishing is here to stay
Under open access (OA), the public has immediate access to scholarly output free of charge. In the case of author-pays Gold OA, scholars are charged for publishing the results of their research, turning around the traditional library subscription business model by 180 degrees, from a consumer-paid to a supplier-paid model. Funding agencies, governments, universities, and the publishing industry are exploring ways to change scholarly publishing to OA; for example, the Dutch government and association of universities (VSNU) have set a target of 100% OA in 2020.
From subscription to Article Processing Charge
The term processing charge is somewhat misleading, as authors who submit a paper that gets rejected in whatever phase of the process do not have to pay. In other words, rejected articles do not contribute to covering their processing costs.
What is processing in this context? The publishing process can be broken down into three stages a paper can go through:
- it starts with the submission and desk review;
- (in case of a positive outcome) peer reviews; and
- (if peer review was also successful, potentially after revision) further editing and publication.
Each of the stages involves costs for the publisher, for staff salary and overheads, as well as many of the technologies and systems employed. In addition to the processing costs involved in each of the phases, there are other costs involved for the publisher (see for example “Open Access: The true cost of science publishing” by Richard van Noorden). These costs, of course, need to be covered as well, but for reasons of clarity and simplicity, we do not specify them here.
Many discussions have already taken place about how APCs might lead to increased pressure on editors to accept lower quality articles to improve profits. We want to propose an alternative to the current application of APCs that benefits everyone in the scholarly publishing chain.
A high proportion of drafted articles are never published and will drop out somewhere in the publishing process (e.g., only 39% of observational studies with safety outcome(s) registered on ClinicalTrials.gov were published at least 30 months after the study completion). Research by Michael Kovanis (2015) and colleagues indicated that 15% of publications result from the first submission, 47% from the second submission and 20% from the third submission. Many authors are using (or should we say abusing?) the peer review system over and over again without paying for any of the associated costs. For now, most of these costs are paid by subscribers, but in an OA world, the costs would be covered by authors who do successfully publish in a given journal. This arrangement penalizes authors who careful choose where to submit their work while benefitting those who employ a more scattershot approach. Some authors even use journal peer review in place of soliciting feedback from colleagues or carefully crafting the article prior to submission.
In the current APC model, the fee is only charged for published articles. The costs (and profits) of the publisher are therefore covered only by the publishing researchers, with high APCs as a result, particularly in selective journals. If we look at the value of the work done by the publisher in each phase of the publishing process, the publishing costs can be distributed more fairly. And, just as important, a researcher will think twice before submitting a paper to a ‘long shot’ journal.
A better APC model
The idea to split the APC into a submission and a publication fee has been introduced before, e.g. by the Wellcome Trust in 2004, by Mark Ware Consulting in 2009, and in posts in The Scholarly Kitchen by David Crotty in 2016 and Tim Vines in 2018. These publications also address the unfairness of charging authors of published articles for the costs generated by rejected articles. Here, we propose further disentangling these costs: a peer-review fee.
An example: Assume a journal that journal publishes 100 articles, and the current APC needed to cover costs and allow the journal to run a surplus for development and unexpected circumstances is €2,000 (total revenue of €200,000). Let’s further assume that the desk rejection is 1/3 and the acceptance rate (after initial desk review) is 15% (i.e., the overall acceptance rate is 67%*15%=10%). Based on these assumptions, adjusted calculations can be made for a fair division of costs over the separate steps in the process. For instance, having published 100 articles, a total of 1000 papers have been submitted. To earn the same level of revenue, the publisher can charge a submission fee of €75; €150 for reviewing and another €250 for actual publication. In this situation, the researcher needs to pay in total €475 if the article is published (see Figure), a fraction of the €2000 APC under the current model. Transaction costs on the side of the publishers will increase under this new model, because more invoices need to be sent out and handled. However, even if those costs are charged to authors as well, the total costs for one published article will still be significantly lower than in the current situation. Moreover, this adapted APC model can help to avoid free-riding and aiming unrealistically high as a researcher when choosing a journal.
We know from our personal experiences as librarians that scholars only look for a budget to cover the APC once their manuscript has been accepted for publication. For high APCs, they can often feel somewhat reluctant to ask their supervisor, principal investigator or the library how the costs can be covered. In the APC model that we propose, these scholars would have to ask for a budget at the beginning of the submission phase. Having this discussion early on in the publication process takes some of the pressure off, because the amount of money that the scholar is asking for is much lower. Moreover, it can be seen as expectation management, because the party providing the budget knows that more costs are to be expected if and when the manuscript gets accepted.
Journals that have experimented with submission fees before found that submission numbers went down significantly (33.5% in the case studied here) when introducing a submission fee. A decreased number of submissions offers the benefit of decreased costs – less time and effort spent on review as well as lower usage of the often costly technologies involved. Prices charged for various fees would need to be reduced to reflect these reduced costs.
It is not clear whether such a decrease would happen in a situation where the submission fee is introduced by an OA journal that is currently using a traditional OA payment system with an APC charge. In that situation the introduction of the submission fee would lead to a substantial reduction in charges paid for a published article, which may attract more authors, offsetting any decrease caused by the submission fee. The lower costs can strengthen a journal’s competitive position as compared to other journals in the field. If a reduction in the number of submissions is still observed, however, the finding of Nwachukwu and colleagues that there was no change in the characteristics of submitted papers is important to take into account. If that is the case, the assumption that 1000 submissions are needed for 100 publications would remain unchanged.
A shift from publication-based charging to process-based charging may lead to better allocation of scholarly communication costs, lower charges per published article, and probably also a better consideration about the choice whether to publish and in which journal. The time seems ripe to introduce such a new APC model now that more and more journals are flipping to an OA model. Plan S requires authors to publish open access (and thus only submit to fully-OA journals), and there is a change going on in research assessment methods (i.e., from metrics-focused to narrative-focused). In particular, Coalition S is seeking pricing transparency for authors, and our proposed structure offers exactly that.
Looking at the stakeholders that are involved in scholarly communication, there is much to gain by changing the way scholarly communication is priced. The benefits for researchers are both transparency in the prices of delivered services by a publisher and lower costs per published article. Moreover, a researcher is triggered by the new APC model to make risk assessments and decide where to submit his/her manuscript: will it still be in that high Impact Factor journal with a long lead time and a small chance of being published? Or will more transparency and a change in the pricing model lead to less focus on Impact Factors, making other criteria more important when choosing an outlet? If so, this might help to shorten the time between first submission and publication, and reduce the hidden costs of rejection and the pressure on both the publishing and the peer-review process (in the example, assuming 2 reviewers per article, over 1300 reviews need to be made for 100 published articles in the end), leaving time for other important tasks.
Of course, we cannot ignore the fact that a potential roadblock in the transition to a new APC model is that journals might be less likely to introduce submission and peer-review fees if their competitors do not charge such fees. However, when the total article processing costs are taken into account, introducing these fees might be beneficial for a journal’s competitiveness as it enables authors to publish OA at a lower cost without sacrificing quality. Because the literature on the introduction of submission fees is not (yet) based on OA journals, we cannot predict which of the two scenarios is more likely to happen: a decline in the number of submitted manuscripts or a stabilization or increase because of the lower total APC costs ahead. In an earlier post on The Scholarly Kitchen, Tim Vines suggested to let authors choose between the new APC model (including a submission fee), or not paying a submission fee but paying a (higher) APC at acceptance of the paper. This suggested approach is not only a good solution for the transitional phase to the new APC model, but can also provide valuable insights into the effects of introducing a submission fee for an OA journal.
To us, changing the APC model now seems a much better idea than to wait until that ship has sailed while we end up in an OA world with undesirably high publication (as opposed to processing) charges. If you agree, join us in getting this message across to funders so they can publicly voice their support, and to publishers to help persuade them of the enormous benefits of transparency and fair pricing.