Open access and the reality of getting from here to there.

Open access (OA) publishing* has long been touted as an important reform needed in academic publishing. OA is when an academic journal article is published under an open license like CC-BY or CC-BY-SA** or public domain and is made available to readers without a paywall. The benefits of OA are obvious. Anybody anywhere can read a scientific paper without having to pay or have a subscription. It is hard to disagree that on some fundamental level science publishing should be out in the open like this.  And on a practical level researchers without university affiliations and in countries with libraries that cannot afford subscriptions will clearly benefit. And it might solve the problem of journal subscription prices increasing at a rate much faster than inflation and breaking the backs of libraries since subscriptions go away. Despite its benefits, OA is not the world we are in today – the traditional model has been focused on a reader pays (subscription) model. But I have come to think that the forces aligned behind OA have become strong enough to push us part way there. So how do we get from one model to another?

To be clear up front, this is not about whether OA is good or bad, I have my own (of course balanced and nuanced!) opinion*** . But here I just want to stipulate that OA is a viable worthy model. Instead, I want to talk about what the transition from today to full OA will or could look like. This is suddenly very timely given several recent events in ecological publishing:

• First, OA has received a huge push forward with the release of Plan S. This is a consortium of research funding agencies that have banded together to insist that the work they fund be published OA by 2020 (specifically gold OA, hybrid or delayed OA is unacceptable*). These funding agencies including the ERC (European Research Commission) and many individual European countries’ funding agencies (e.g. France, Netherlands, Norway, Ireland, UK) and several private trusts (Gates foundation, Wellcome Trust). The German, Chinese and Swedish research councils have also endorsed the principles but their commitment to execute them is less firm.
• Second, one of ecology’s own, long-established journals (Diversity and Distributions or DD) has made a switch from a hybrid (primarily subscription) model to a gold OA model effective October 8, 2018. The APC (article processing charge) is $2,200 (with a 20% discount for members of the International Biogeography Society).  The initial proposal included waivers only for authors in very poor countries, but, when pushed by the editorial team, Wiley adopted the policy that “Ability to pay the APC should not be a barrier to the publication of important science. Authors without funding for publication charges will be provided with a waiver of the APC.” Such a waiver policy is common for a non-profit society journal APC, but it is among the more generous available from a for-profit publisher. The publishers will also be opening up all back-issues as OA. (NB: I am well aware of further recent controversy around this journal but that is not on the topic of this post and I will stick to the OA aspects here).
• Third, Dan Bolnick’s blog post on just how much he has spent on APC in recent years also opened eyes. And pointed out that even if you have grants, APC charges can add up enough that they can realistically reduce funds spent on actually doing (as opposed to publishing research). A year’s worth of APC charges can put a serious dent in even, say, a Canadian NSERC grant that is typically a few $10,000s or a US NSF small grant of $135,000 (which boils down to probably $30,000 spent in each of 3 years after overhead).

Across these three events there is a clear trend towards more OA. There are clear OA upsides in these events. But these events have also forced ecologists to look at the reality of OA, and to start to notice some of the downsides as well. Specifically and prominently, APC charges that are often over $2,000 dollars and usually over $1,200 (and sometimes over $5,000 even in non-predatory OA journals like Nature Communications)****, which means that many, many people could not publish in such an OA journal. Including those who invested sweat equity as reviewers or past associate editors. An APC (without a proper waiver policy) exposes a strong line between those who have a big grant and those who do not have a grant or who have a poorly funded grant. Longstanding fears around OA begin to feel real such as:

• OA exacerbates the divide between academics who have or have not (specifically those having a rich-country grant can publish easier, which makes it easier to get another grant, which makes it easier to publish more, and etc.). Those without large grants are left out in the cold. Although many are quick to acknowledge the challenges in relatively poorer countries (including many important contributors to the literature) where large grants flat out don’t exist, the reality is also that many very good, frequently publishing individual researchers in the US and UK spend much of their careers without a grant too and most Canadian researchers are funded but at levels that don’t admit for a lot of OA publishing. While not everybody in continental Europe has a grant, I think it is a fair characterization that right now grant funding rates are higher and there is much more institutional funding for infrastructure-like needs such as APC. I think it is not a coincidence that the heart of the push for OA has been continental Europe.
• Once APC charges get above a few hundred dollars, it is not realistic to use an individual paying out of their own pocket as a backup plan. An occasional desperate early career researcher might pay $2,000, but they are the ones least able to pay. And of course nobody should have to pay even a few hundred dollars out of their own pocket to publish.
• The only alternative to grant-funded OA is a departmental or university fund and these have been scarce to materialize despite a lot of talk. If you had to obtain $2,000 today to publish OA from your own university, what are the odds you could do it? And then do it again next year?
• The ethical downsides of having a journal funded by authors hovers over the discussion of OA as well (although not specifically over any of the three news events I mentioned above). The specter of predatory OA publishers with embarassingly porous peer review is part of a larger shades-of-gray spectrum exploring what will happen if you give authors the power to fund the journal that decides whether or no tot publish that author.

This is not to argue that OA is bad (or good). What really interests me, and what I think proponents on all sides of the OA debate need to start thinking about, is what the transition from a more traditional subscriber model to an OA model looks like. As the Plan S and Diversity and Distribution news stories and Bolnick post show, OA is gaining momentum. Those who could in the past ignore OA no longer can. So the question of what a half-way world with a mix of traditional and OA journals looks like is a timely question (it is the world of the next 5 years at least).

And I think the answer is that it looks exactly and literally like moving from one fitness peak to another, with a fitness valley in between.

The 100% reader pays (left) peak is the traditional world (e.g. academic publishing in 1990). Library subscriptions and individual author subscriptions as part of their society memberships fund most of the costs of publication. Color figure charges or per text page charges might have been charged to authors, but those were small (several hundred dollars) and structured to incentive authors to behave responsibly (limit color figures and page counts). This model was viable for decades.

The 100% author pays (OA) peak on the right is presumably equally viable. Libraries could drop the enormous (ever increasing at a rate much faster than inflation) subscription fees they pay. Instead universities (and other institutions) could redirect all of that subscription money to pay the APC charges for their researchers.  In principle this could be a zero-sum game where we spend the same amount to publish as we do today, but it is all OA. Thus this demonstrably a feasible approach.

But if we continue the analogy of evolution on an adaptive landscape, we cannot just jump from one peak to other (unless we are a saltationist). Very much like evolution, our complex adaptive publishing ecosystem may have a long-term optimality principle, but it moves in small steps and is based on the fitness of individual actors.  If we lived in a dictatorship or even an oligopoly of research funding groups, perhaps we could jump peaks. But the reality is that there are many dozens of national government funding agencies (in the US alone we have four major players even at the national government level: NSF, NIH, NASA, and USDA with NOAA, SBA and others also providing funding). And then there are regional, state, and provincial agencies and private trusts. And a good deal of research is funded directly by individual universities (and other employers of scientists like government departments). And much existing reading of scholarly literature is funded by university libraries. No major US, Canadian, Australian, Japanese etc government research funding agency has announced plans to participate in Plan S. Nor have any states, provinces or most private trusts or NGOs. And what about less-rich countries (that are an important and growing part of the author pool) that now has to pay for their researchers to publish? I cannot conceive of a world where all of these hundreds (thousands?) of organizations manage to unite and agree on an instantaneous and painless transition to a shared vision of the future (whatever it might be). It is inconceivable that we will not enter an accidental, unplanned, many-actors muddle. We are doing that as we speak.

And once in a mixed world, strong inertial forces will keep us in the middle. As long as there are still free to publish-in (i.e. old model subscriber-based) journals, universities will not have to step up and fund their researchers for APC chargers. This has been one of the major failures of the move to OA to date. Universities and governments have been very eager to adopt policies to mandate OA but have been much worse at finding funding to pay for it. Indeed, many universities have seemed naive about just how much it would cost to move to OA – throwing a tiny fraction of the money needed out there and thereby unintentionally creating fierce internal competition for limited OA publication funds. Plan S calls for funding APC for researchers, but it provides few details about how to make this happen when the funding agency didn’t provide the funding. Again, in theory, if universities stopped paying subscription fees and started paying APC fees, it could be a zero sum game, but I cannot imagine how this could be coordinated.

Equally, as long as subscription models are viable, publishers (and societies) will be slow to risk switching journals that have been profitable for a long time to OA. Many society journals and many for-profit publishers of must-read journals (e.g. Nature) are not signalling any enthusiasm for dropping subscription fees and switching in a couple of years to OA. OA has been much more a new journal phenomenon than a switching existing journals phenomenon (the example of DD notwithstanding). Plan S calls for subscription journals to switch to OA, but again it does not provide any details of how this will happen.

The result for some time is going to be a mixed ecosystem where some journals are subscription based and some are gold OA with hefty APC charges. Could this be good for individual researchers? To put it simply, I don’t think so. I think this will be hell. The mixed state is likely to exaggerate all of the rich-get-richer issues of OA. Those with money will know they can get in an OA journal while those without will have to compete to get into what might be a dwindling pool of free-to-publish subscriber-based journals (which in any case is by definition a smaller pool of journals than those with grants can get into). Even for a well-funded researcher, one is going to have start carefully managing “this paper was funded by a Plan S agency, so it has to go to gold OA,  and this paper was not funded by a Plan S agency so do I really want to pay $3000 to publish it?” And on this 10 author paper my funding was Plan S so it has to go gold OA, but the first and senior author didn’t have funding and aren’t in a country that pays OA fees, so am I going to pay $3,000 for a paper where I am 5th out of 10 authors? Getting published in a good journal is hard enough without having to play these games! Its hard to imagine a world where anything dominates choices about which journal to submit to more than impact factor (wrong-headed though that is), but the Plan S world just might push the OA/author-pays vs traditional/reader pays criterion higher than IF!

So it seems to me that if we want to get to the OA peak, we necessarily must descend into the connecting valley to get across (and unlike in evolution it is not a foregone conclusion that we will climb back out of the valley and reobtain a peak). And it seems to me inarguable that this valley has a noticeably lower relative fitness for researchers (and universities and publishers). And it seems to me that we will spend a long time in this suboptimal valley.

Again this is not a knock on OA, nor those trying to get there.The Plan S push is well intentioned and may ultimately prove to be judged brilliant in hindsight******, but in the short-term it has committed us irrevocably to leaving the left peak and hoping we make it across to the right peak. The Diversity and Distributions story I believe is also well-intentioned (if imperfectly executed). These show that even when the destination is good (OA with a good waiver policy or funding for APC*****), change is always painful and disruptive, and creates new winners and losers in the new world. And so we all need to get real serious real fast about the transition or we could find everything worse for almost everybody.

I want to be clear that this painful valley pinches all players. It is pinching publishers, universities and granting agencies as well as researchers and scholarly readers. Nobody knows where we will end up. Everybody is going to feel pain during the transition. This is not a case of one group forcing its goals on another. And indeed the point is no one entity (or even whole stakeholder group) can drive this change alone. Instead everybody knows they are not in control of events and, so, is trying to hedge their bets. Ironically – and this is a key point – hedging bets prolongs the time we spend in the suboptimal valley and prolongs the pain for everybody even if bet hedging is individually rational.

What do you think? Have I painted an overly pessimistic picture? Is a half-way world actually painful as I claimed or could such a diverse publishing model be a healthy ecosystem for all the stakeholders (authors, readers, publishers, universities, libraries, funding agencies)? Or is the OA peak worth the pain to get there? How could we minimize the time (and pain) in transition? Or should we throw on the brakes to forcing OA?

---

*OA actually comes in several flavors or “colors”. Gold OA is when a whole journal functions on an OA (author pays, no paywall to access model). Hybrid OA is when the journal has a reader-pays subscription but individual authors can pay to bump their own article up to OA. Platinum OA is when a 3rd party funds the costs and neither author nor reader pays. Green OA is when the journal is behind a paywall but authors can put their PDF on their own website (legally). Delayed OA is when an article is behind a paywall for a fixed period of time (6-18 months usually) and then becomes OA.

**Note that no serious OA journal uses the CC-BY-NC or CC-BY-SA-NC (non-commercial flavors). There are good reasons for this; NC is very problematic. And even SA puts a serious crimp in the spirit of open access. Most OA is just CC-BY.

*** My opinion is that the OA debate is in the end about who pays (author vs. reader) and that other issues like how much you pay, who you pay, and ancillary policies like Green OA (posting a pdf on your website), deposition with JSTOR (which is not technically OA because it has small sliding fees), etc are all much more important than OA.

**** there are some notable exceptions to high APC charges like PeerJ and Frontiers in Biogeography which are a few $100s and Evolutionary Ecology Research which is $1,000 but has very many routes to waivers including if your library subscribes to the journal (http://www.evolutionary-ecology.com/advice/AdviceFrames.html). Which all said (if I want to be in I-told-you-so mode) highlights my opinion in the previous footnote that how much you pay matters as much or more than who pays.

***** the split between addressing APC charges by waivers vs. researcher funding is a whole addition muddle that is going to cause a lot of pain. If Plan S countries like, e.g., France step up and fund APC charges how incentivized are they going to be to push for good waiver policies? Where does a lack of focus on waiver policies leave researchers without APC funding?

****** Plan S also includes demands for full access to all code and data and proper metadata. Surely good things, but also not trivial things. Separate efforts have been working at (and making progress at) open code and open data for years but are not yet there. I cannot imagine how one initiative can enforce all of these issues.

Full disclosure: I am editor-in-chief of Global Ecology and Biogeography which is a traditional hybrid OA (primarily subscriber-paid) model. Diversity and Distributions is a sister journal that I work closely with. Both are owned by Wiley. All opinions in this blog are my own.