You may recall that the Open Journal of Astrophysics recently published a paper by Syksy Räsänen of the University of Helsinki. I invited Syksy to write a blog post on the Open Journal for Astrophysics about why he chose to publish with us, and I’m delighted to say that his post is now available there and, with the author’s permission, I am reproducing it below on this blog. It’s also available at Syksy’s own blog . It’s quite a long post, but there is some very interesting information in it, which will probably surprise you!
—0—
The bad news:: the scientific community can no longer afford commercial science journals.
The good news: the scientific community no longer needs commercial science journals.
The bottom line: open internet archives and overlay journals are the solution.
Scientific journal publishing is in crisis. Already 25 years ago librarians referred to the rising costs of journals as a “doomsday machine”. In 2012, Harvard University Library announced that it can no longer afford scientific journals, warning that publishers had created a “fiscally unsustainable” situation. The library took the unprecedented step of asking faculty to resign from publications that keep articles behind paywalls.
In its 2015 Open Access Policy White Paper, the Max Planck Digital Library assessed the annual revenue of scientific journal publishing as 7.6 billion euros. Divided by an estimated 1.5-2.0 million published articles, they arrived at a cost of €3800 to €5000 per article. The International Association of Scientific, Technical and Medical Publishers has estimated $10 billion in revenue and 2.5 million articles annually, which gives €3500 per article. According to them, 68-75% of the costs are borne by academic libraries.
These are enormous sums, funnelled from science to the pockets of large corporations. (Some journals are published by scientific societies, but this doesn’t change the overall picture.)
To put the numbers in perspective, the total construction cost of the Large Hadron Collider at CERN –the largest scientific experiment in the history of humanity–was around 5 billion euros, or 500 million euros per year. The cost of science publishing would cover the construction of 10 to 20 top-of-the-line successor experiments to the LHC. It is equivalent to the salaries and overheads of 150,000 to 200,000 postdoctoral researchers per year. This is likely more than the combined number of postdocs in the United States and the European Union.
Scientific publishing is a strange business. Scientists carry out research for free, write it up for free, give the article to the publisher for free, manage the peer review process as editors (some editors get paid), do peer-review for the publisher for free – and then the scientists’ institutions pay the publisher so that they can read the articles. In some cases scientists even pay the publisher to have their articles published. (Here “free” simply means the scientists are not paid by the publisher, but rather by their university or research institute – in the end, mostly by taxpayers.)
Is there any other industry where corporations pay nothing for the raw materials, have negligible processing costs and enjoy a captive market that automatically buys everything they produce? This setup leads to record profits. For example, the profit margin of Elsevier’s Scientific, Technical & Medical publishing division in the years 1991-2013 was consistently over 30%, and in 2013 it was 39%. For comparison, in the same year Apple, Google and Microsoft had profit margins of 22%, 20% and 28%, respectively. No wonder Robert Maxwell, a pioneer of scientific journal publishing, described the business as a “perpetual financing machine”.
In the past, publishers at least had the expenses of printing and shipping journals. With the internet, this cost has disappeared for many journals. However, prices have not come down – quite the opposite. Thankfully, the same technological advances that have made commercial journals so profitable have also rendered them unnecessary.
Journals used to be needed for registering and communicating research, for archiving it (via paper journals in libraries), and for performing quality control via peer review. In cosmology and particle physics, the first three services have been provided by the internet archive arXiv since 1991. Depositing articles to arXiv is free, and they can be read for free in perpetuity. In 2018, 140,616 articles appeared on arXiv, and its estimated total expenditures were $1,915,997, or 12€ per article. Other disciplines, such as economics and biology, have followed suit, either starting new categories on arXiv or setting up their own archives such as bioRxiv. There is no reason why this model cannot be extended to all fields of science.
A lot of discussion around the cost of journals centred on open access, so it is important to emphasise one thing: the issue is not open access. That problem has been solved by arXiv 27 years ago. The question is how to organise peer review in a cost-effective manner. This is where overlay journals come in.
The idea of overlay journals is simple: they are journals that concentrate on the only thing journals are needed for anymore, namely peer review. As articles appear on arXiv (or other online archives) anyway, there is no need to duplicate their work. An overlay journal has a website where papers (typically already available on arXiv) can be submitted. Peer review is conducted as usual, and in case of acceptance, the final version of the paper is updated on arXiv, with a journal reference and DOI link to the journal website.
Custom toolkits for overlay journals have been designed for more than 10 years, for example in the astrophysics RIOJA project. These days the necessary software is also available off the shelf from Scholastica. The service costs $99 per month plus $10 per submitted article, and getting a DOI from Crossref costs $1 per published article. As in the case of arXiv, the costs scale well with the number of papers. If a journal publishes 100 articles annually and has a 2/3 rejection rate, the cost works out to €38 per article – about 100 times less than the sum currently paid for article publishing.
Given that open archives and overlay journals could save 7 to 9 billion euros every year, why haven’t they already replaced commercial journals?
Unfortunately, the existence of a more optimal configuration does not automatically lead the community to shift there. Instead, people respond to individual incentives, and scientists are no exception. The publishing peer Robert Maxwell noted that “scientists are not as price-conscious as other professionals, mainly because they are not spending their own money”. As journal fees are paid centrally, there is little motivation for an individual researcher to change their publication pattern. Established journals are seen to provide a quality stamp that is necessary in the competition over positions and grants. Also, founding an overlay journal requires an investment of time that does not necessarily yieldproportionate professional rewards.
So researchers as individuals have rational reasons for not changing the system. What about libraries and scientific consortia that are struggling to bring the costs down? Unfortunately, institutional efforts have often concentrated on the narrow problem of getting journals to accept payment from the author rather than the reader (i.e. open access). However, the main issue is not whether the money is paid by the author or the reader (or rather their institutions), but what they are paying for.
Thus, for example, the SCOAP3 consortium has become part of the problem by providing life support to commercial journals. It has arranged to pay publishers vast amounts of money, entirely incommensurate with the actual costs, to make articles open access, guaranteeing them a steady stream of revenue.
The open access initiative Plan S launched last September is more ambitious. Particularly noteworthy is the commitment to provide incentives to establish new open access journals and platforms. However, open archives are only “acknowledged because of their long-term archiving function and their potential for editorial innovation”, not as publishing channels of their own right when paired with overlay journals. A lot depends on how the initiative will be implemented, but for now the scheme seems to focus on the old-fashioned aim of getting commercial publishers to convert journals to open access.
According to the Max Planck Digital Library White Paper, switching all commercial journals to an open access model would drop the cost of publication to between €1,100 and €2,000 per article. This would cut expenditures by a factor of 2 to 5, saving billions of euros every year – and continuing to waste billions of euros every year.
It is striking that the principles of Plan S contain no commitment to maximising the returns on public money and optimising financial sustainability, even though this is the heart of the matter. Instead, it supports the continuation of the commercial publishing model in co-operation with corporations, whose interests are at odds with those of the scientific community. We do not need to reform the business model of scientific journal publishing, we have to abandon it.
Open access consortia should start supporting a publishing model that begins from the needs of the scientific community and aims to fulfil them in an economical manner, while helping to make the transition as smooth as possible. This involves communicating with scientists about the costs of corporate publishing, following and expanding on Harvard’s example of calling on scientists to use the power of their labour (often given to the publishers for free) to change the situation, and providing incentives and support to establish and publish in overlay journals. Scientists, in turn, need to re-evaluate their brand loyalty to established journals, and give appropriate career merits for time spent on changing the publishing system.
Open archives and overlay journals are not a utopian solution for the future. Discrete Analysis, Open Journal of Astrophysics and others are publishing already. They are a proven model for open access publishing and quality peer review in a modern, cost-effective manner tailored to the needs of the scientific community. The sooner they become the new standard, the more money we will save for science.
In the Dark 