In a nutshell, these funders (and others that may join them) are banding together to impose Open Access (OA) publishing on their researchers. Most contentiously, the funders plan to cap the Article Processing Charges (APCs) that they are willing to cover, making it impossible for their researchers to publish in journals that charge relatively high amounts. On the journal side, those with APCs above the cap must either accept more (i.e., lower quality) articles or cut expenses by doing less review and editing. A CC BY license is required, and publishing an OA article in a hybrid journal is also apparently banned, such that these funders’ researchers are effectively unable to publish in over 85% of existing journals.
If other funders get involved, there’s a possibility that they’ll use their collective influence over researcher behavior to ram through major changes in how we fund peer review and publication. The current focus of the discussion is on APCs (i.e., charges applied only to accepted articles), but seeing as we’re taking a hammer to the whole system we should contemplate some alternatives too.
Why? APCs have the unfortunate feature that the authors pay for the assessment of all the other submissions that ended up being rejected. Manuscripts rejected from multiple OA journals thus contribute to the APCs of several different authors. Is it fair for authors of good articles to pay for the peer review of others’ lower quality work? Moreover, journals that do lots of peer review to find one acceptable article have higher APCs, as illustrated by the figure below.
*acceptance rate estimates from public statements and informed guesswork
While these are rough estimates, the fitted line is APC = 850 + 350*(1/Acceptance Rate). This choice of equation is not a coincidence: first, the inverse of acceptance rate represents the number of articles the journal has to assess before it finds one that can be published, and $350 is a decent ballpark for the cost of a round of peer review (e.g. $250 in 2011 [=$280 in 2018] from here, other estimates here).
The $850 represents an amount that for these journals (presumably) covers the costs associated with the accepted article (typesetting, proofing, hosting, etc.), along with the overheads associated with all the other things that publishers do. This amount may not be accurate for some journals – for example, PLOS Biology is said to run at a deficit at its current APC and is subsidized by revenues from PLOS ONE. But bear with me.
Breaking down an APC and how it relates to submission numbers and acceptance rates suggests another way to cover publication costs: a submission fee of $350 and a publication fee of $850 would generate the same revenue as the current APCs at these journals. (This approach has been suggested before, see here, here, and here.)
Most importantly, this fee structure means authors of accepted articles only pay a total of $1200 for publication, no matter how selective the journal. The fee would be $1200 for PLOS ONE and $1200 for Nature Communications. Even if an article is rejected from two journals before being accepted at a third, the total cost is $1900, which is still less than a typical APC.
This solves one of OA’s thorniest problems – how to run a high-quality, highly-selective journal without charging an enormous and impossible APC. The proposed fee structure even works for a journal with a 1% acceptance rate: a $350 submission fee and a $850 publication fee would equate to an APC of $35,850, which matches Sir Philip Campbell’s famous/infamous estimate that Nature would charge $30,000 to $40,000 for an OA article.
The big obstacle to adopting submission fees has always been first-mover disadvantage: authors don’t particularly like the prospect of submission fees and may favor journals that don’t charge them. Any journal introducing submission fees thus risks deterring authors. Interestingly, this didn’t seem to be the case at the Journal of Bone and Joint Surgery, where the introduction of a $250 submission fee led to a fall in low quality submissions, but the volume of higher quality articles remained steady.
If Plan S becomes a reality then ‘business as usual’ goes out the window, as publishers will find themselves scrambling to ensure that as many journals as possible are a) Open Access, and b) charging an acceptably low APC. One or more publishers opting to achieve (a) and (b) by introducing submission fees would affect so many journals that the first mover disadvantage becomes a non-issue.
As one might expect, having submission fees paid directly by funders would also significantly mitigate the first-mover effect, and helpfully the direct payment of publication fees is already part of Plan S.
The introduction of submission fees also affects submission patterns, most notably by steering away articles with only slim chances of acceptance. This effect would change the calculations in the figure above: fewer low quality articles arrive, so acceptance rates rise and hence reduce income from submission fees. However, a drop in submissions also means a drop in costs, as those weak articles no longer need to be processed through the system.
This effect on submissions should be more pronounced for a single journal introducing submission fees when their competitors did not, as authors can take their speculative submissions to the no-cost journals. However, if Plan S (or its equivalent) prompts the widespread introduction of submission fees, authors may not easily find a suitable journal that does not charge fees. Since they need to submit their article somewhere to get it published, the decrease in submissions may be smaller than feared.
A drop in submissions is not fatal for the submission fee math – even a 30% fall can be accommodated by raising the submission fee to $500, or by raising the submission fee to $400 and the publication fee to $1100. Journals would adjust their submission fees within some reasonable range depending on their brand perception, current levels of submissions, and a desire to remain competitive with other journals in the field. Journals with high fees may even be able to signal the higher quality of their review and publishing process.
Submission fees have other useful properties. First, they are ‘pay as you go’ for peer review: they penalize authors who submit low quality articles over and over to different journals, and reward those who prepare their work to a high standard and submit it to the most appropriate outlet.
Similarly, submission fees counteract the perverse incentives created when authors receive financial rewards for publishing in high impact journals, which is a major driver for inappropriate submissions. If they had to pay each time, would as many authors take a wild stab at getting their incremental work published in a top journal, then working their way, journal by journal, down the Impact Factor rankings until they reach an appropriate level?
Submission fees also bring peer review into line with lots of other services that cost money regardless of whether you succeed or fail, such as professional exams or even dental check-ups. Viewed through this lens, the ‘no win, no fee’ approach of APCs seems like an anomaly.
Clearly, there are many other aspects to consider (particularly unintended consequences), but if the funders are going to force a change in the financial structure of peer review and publication, we should give submission fees serious consideration: they’re a logical way to cap the cost of Open Access while maintaining the current ecosystem’s ability to filter articles by perceived importance.