Saturday, 17 November 2018

How strong are the objections to Plan S?

After a coalition of European science funding agencies announced their Plan S initiative for open access, a number of researchers wrote an open letter criticizing the move, under the title “Reaction of Researchers to Plan S: Too Far, Too Risky”. To summarize, they fear that Plan S would increase costs, lower quality, and restrict academic freedom. In order to evaluate how seriously these fears should be taken, let me start with a 5-point analysis of the issues, before discussing the open letter’s specific concerns.

Point 1: Traditional journals are overpriced by an order of magnitude.


Overall, publishers earn about 3800-5000 euros per article they publish. The true costs of organizing peer review are much lower: a few hundred euros per article at SciPost or PeerJ. The difference is therefore not mainly due to commercial publishers’ profits, but rather to large inefficiencies at all traditional publishers. (Have a look at the salaries of the American Chemical Society’s executives.)

Some publishers also use the money for subsidizing other activities, such as conferences (learned socities) or science journalism (Nature, Science). While these activities are valuable, they are not the main purpose of scientific publishing, and should not determine its future.

Point 2: No open access, no affordable journals.


The inefficiencies and overpricing of traditional publishers can seem puzzling: with so many journals and publishers around, how can the competition allow this? However, in a subscription-based system, you have no effective competition: when you need to read a given article, you cannot look elsewhere for a substitute. Authors choose where to publish, readers pay: no competition. With Gold open access, authors choose and pay, so prices can play a role in their choices. Actually, any form of true open access eliminates each publisher’s effective monopoly over its articles, and allows prices to decrease.

(I write true open access in order to exclude delayed open access after an emabargo: this only makes the monopolies temporary, rather than eliminating them. Delayed open access should not be identified with Green open access, which can exist in the absence of embargos.)

Moreover, an open access publishing model should cost less than a subscription system, if only because building and defending paywalls is costly. Publishers sometimes pretend otherwise, and try to earn extra money for making articles openly accessible, a tactic called double dipping. That this tactic works is proof enough that the existing system is perverse and needs radical reform.

Point 3: When it comes to quality, the existing system performs poorly.


There is widespread perception that scientific research is undergoing a “reproducibility crisis”, and that most published studies are false. An important part of the problem is that articles are often judged by the journals they appear in. And some of the most prestigious journals, such as Science and Nature, follow criteria that are more journalistic than scientific when selecting articles.
While this is not directly related to open access (or the lack thereof), this should be kept in mind when contemplating reforms such as Plan S. We are not talking of improving a well-functioning system, but of reforming a dysfunctional system that may not be self-correcting.

Point 4: Researchers are prisoners of the system, and most of them don’t care.


Researchers are used to working for journals, where they publish without remuneration, and work as peer reviewers and editors, most often for free. A priori there is nothing wrong with that, as researchers do not count on journals for earning their money. The system becomes objectionable when journals abuse their position, in particular when they require researchers to give away copyright to their works, and proceed to strongly enforce said copyright in their own favour.

A particularly draconian case is brought to us by Elsevier, the notoriously rapacious publisher: Elsevier has so far refused to join the Initiative for Open Citations, a form of data opening that is manifestly not a threat to existing business models. Of course, Elsevier has its own reasons, more related to its ambition of controlling research data and workflow, than to its legacy publishing business.

Elsevier’s behaviour has made it the target of the most widespread boycott of an academic publisher to date. This boycott has been joined by only 17000 researchers, and has had little noticeable effect. While giving away one’s copyright is often unavoidable, working for Elsevier as a reviewer can easily be renounced without harming one’s career. Boycotting Elsevier as a reviewer is the least a researcher can do, if he is concerned about how academic publishing works. The relative failure of the boycott demonstrates that the vast majority of researchers are not concerned.

Point 5: Progress is slow, and may well stop.


After decades of efforts, the open access movement has achieved relatively little in terms of making articles openly accessible, and nothing in terms of reducing costs. One may be tempted to conclude that the endeavour is futile and the cost reductions illusory. However, the success of the pirate open access website Sci-Hub shows otherwise, by giving a glimpse of a possible universal open access future, and exerting a strong downward pressure on subscription prices. Thanks to Sci-Hub, it is indeed possible to conveniently access the literature without paying a subscription, and this has tremendously helped consortiums such as the German DEAL or the French Couperin in their negotiations with publishers.

Inevitably, Sci-Hub is under legal attack from publishers, and so is the social networking site ResearchGate, where researchers share their article. The American crackdown on Sci-Hub involves drastic measures that target not only Sci-Hub itself, but also a large number of Internet intermediaries: academic publishers such as the American Chemical Society and Elsevier are at the forefront of establishing a legal censorship regime for the Internet.

In Europe, the establishment of censorship rather comes from the news and entertainment industry, and takes the shape of Articles 11 and 13 of the draft copyright directive. The principles behind these articles (taxing links and filtering uploads) would destroy arXiv, GitHub and Wikipedia. Rather than narrowing these principles to news and entertainment, the directive’s authors have carved exceptions that will save today’s existing platforms – but not possible open platforms that do not exist yet.

Therefore, the present conjoncture is as good as it will get for a transition to open access, with Sci-Hub widely accessible, and Internet relatively open. The conjoncture is likely to get worse, possibly quickly. This may be a now-or-never moment.

Plan S.


In this context comes Plan S: a coordinated attempt by research funding bodies to induce a flip to open access in the publishing system. Previously, individual funding bodies were content to mandate that the research they fund be openly accessible: it is now apparent that such mandates can achieve their limited goals, but neither change the system nor decrease publication costs. This is the rationale for the more radical features of Plan S, in particular banning hybrid journals.

I will now discuss the various objections to Plan S from the open letter.

Objection 0: Plan S goes too fast.


Plan S aims to be implemented by 2020, which can seem frightfully soon for such a radical flip. However, there is nothing technically complicated in creating a new open access journal, flipping an existing journal to an open access model, or creating a new preprint archive. All this has been done many times before: in the hybrid system we now have, most players already have the required technical infrastructure. What prevents publishers from flipping is their understandable desire to milk the subscription cow for as long as possible.

A gradual transition could a priori appear smoother and less disruptive: but the transitional situation has been going on for decades, and is in danger of becoming permanent. And in the transitional situation, double dipping can make costs not only higher than in a pure open access situation, but also higher than in the traditional subscription model. Moreover, the open letter rightly complains that researchers who are subject to Plan S could be disfavoured in the prevailing perverse system of career incentives: and indeed there is a kind of chicken-or-egg problem with flipping both the publishing system and the assessment system. But making the flip more gradual only worsens this problem.

So there are good reasons for going fast, even before taking into account Point 5: that the present conjoncture is good, but may well degrade quickly.

Objection 1: The complete ban on hybrid (society) journals of high quality is a big problem.


For Plan S, the ban on hybrid journals is a feature, not a bug: it amounts to burning the bridges to the subscription system, and signals a commitment to a complete flip to open access. Hybrid journals allow double dipping: the reason why a partial flip to open access does not reduce costs, although a complete flip would.

The problem is that this outlaws many existing journals. For an individual researcher in the existing publication and assessment system, to publish open access while renouncing hybrid journals can mean career suicide. However, raising this objection means ignoring that Plan S is a drive to change the whole system, by a globally significant coalition of funders. This objection assumes that existing journals and assessment methods do not change: in other words, that Plan S fails. This objection therefore depends Objection 2, which predicts such a failure.

But again, it would be quite easy for existing high-quality journals to become compliant with Plan S. Failing that, creating new high-quality open access journals is nowadays difficult but not impossible, as the examples of SciPost (physics) or PeerJ (biology) show. This will become much easier in the context of Plan S, with many good researchers having to renounce their usual journals and seek other venues. It is not reasonable to complain about the lack enough high-quality open access journals, before the advent of strong incentives for creating them.

The most serious objection to the ban on hybrid journals is actually tactical: that the ban could be circumvented by creating mirror journals.

Objection 2: We expect that a large part of the world will not (fully) tie in with Plan S.


The scenario is now that Plan S would fragment the research community. Part of Europe would adopt a Plan S-compliant publishing system, and the rest of the world would keep the existing system. They would ignore the Plan S-compliant journals, ruining the careers of researchers who publish there.

This piece of dystopian fiction first assumes that Plan S fails to induce a general flip of the publishing landscape towards open access. The second assuption is that there is no progress on research assessment, that the San Francisco Declaration on Research Assessment is widely ignored (including by its signatories), and that researchers continue to be evaluated according to the journals they publish in. And the third assumption is that Plan S-compliant journals will not acquire the necessary prestige.

While the likelihood of the first two assumptions is hard to evaluate, the third is certainly not plausible. With the involvement of the Europe’s most prestigious research funder (the ERC), Plan S-compliant publishing is set to play a major role in Europe. And there is no reason why the rest of the world would ignore most European research, just because it does not appear in the usual journals. A fragmented publishing system does not imply a fragmented research community.

Objection 3: The total costs of scholarly dissemination will likely rise instead of reduce under Plan S.


This assertion runs contrary to the elementary economics of the game (see Points 1 and 2), and the open letter does not provide a basis for it.

Objection 4: Plan S ignores the existence of large differences between different research fields.


Since Plan S means a rapid transition to open access, it would of course have more effect on fields that are less advanced in this respect. But it is not clear why some fields need less open access, or a longer transition. Plan S is a rather minimal set of principles for ensuring true open access and eliminating some of the abuses of the existing publishing system: this leaves much room for different open access models. By adopting best practices from other fields without waiting, fields such as chemistry would be spared agonizing hesitations and transitions.

Note added Nov. 19th: In a comment to this post, Leonid Schneider points out that the EMBO journal "employs full time data experts to analyse images for manipulations". This is an example of a field-specific, justified expense, that goes beyond the costs of the lean journals that I cited in Point 1. While this may not justify the journal's 5200$ APC, it would be good to take such costs into account when implementing Plan S.

Objection 5: Plan S is a serious violation of academic freedom.


Freedoms come with purposes, responsibilities, and limits. You are not free to sell yourself into slavery, to work for less than the minimum wage, or to take dangerous drugs, for the good reason of protecting yourself and others. Similarly, giving away your copyright to publishers, putting your work behind paywalls, and playing the game of a perverse assessment system, should not be allowed in the name of academic freedom. For a more detailed discussion of academic freedom in the context of Plan S, see Marc Couture’s guest post on Richard Poynder’s blog.

Don’t complain, act!


Why does the open letter worry so much about the “ranking and standing” of Plan S-subjected researchers, when the debate is about journals? The letter’s authors seem to accept the entrenched practice of judging researchers by the journals they publish in, although this is widely denounced as perverse. But it is not possible to significantly reform the publishing system without upsetting this practice, at least temporarily. If careers continue to be determined by numbers of articles in Nature or Science, then it is game over for open access and affordable publishing.

When it comes to open science, chemistry seems to lag behind fields such as physics and biology. The field’s leading publisher, the American Chemical Society, did not even join the Initiative for open citations. But chemists could welcome the opportunity to catch up. If you had no telephone and were offered a mobile phone, would you insist on installing a landline first?

The open letter has hundreds of signatories. Surely one could find among them enough well-respected researchers for building the editorial board of a new open access, affordable, high quality, generalist chemistry journal. They would not even need to do it from scratch: they could just start a new division of PeerJ or SciPost. Assuming of course that they really support open access, as the open letter claims in its first sentence.